British Journal of Dermatologylib.ajaums.ac.ir/booklist/339455.pdf · C. Mitteldorf R. Stadler H.P....

March 2007 - Vol. 156 Issue 3 Page 411-612

Snippets

RESEARCH SNIPPETS

pages xv–xv

Editorial

Has teledermatology in the U.K. finally failed?

J.S.C. English and D.J. Eedy pages 411–411

Review article

Maturity of teledermatology evaluation research: a systematic literature review

N. Eminović, N.F. de Keizer, P.J.E. Bindels and A. Hasman pages 412–419

Cutaneous biology

Nine procaspases are expressed in normal human epidermis, but only caspase-14 is fully processed

A.-A. Raymond, M.-C. Méchin, R. Nachat, E. Toulza, R. Tazi-Ahnini, G. Serre and M. Simon pages 420–427

Testosterone metabolism to 5α-dihydrotestosterone and synthesis of sebaceous lipids is regulated by the peroxisome proliferator-activated receptor ligand linoleic acid in human sebocytes

E. Makrantonaki and C.C. Zouboulis pages 428–432

The antiwrinkle effect of topical concentrated 2-dimethylaminoethanol involves a vacuolar cytopathology

G. Morissette, L. Germain and F. Marceau pages 433–439

Clinical and laboratory investigations

Expression of p16, CD95, CD95L and Helix pomatia agglutinin in relapsing and nonrelapsing very thin melanoma

L.A. Fearfield, J.M.G. Larkin, A. Rowe, R. A'Hern, C. Fisher, N. Francis, R. MacKie, B. McCann, M.E. Gore and C.B. Bunker pages 440–447

Inactivation of the CDKN2A and the p53 tumour suppressor genes in external genital carcinomas and their precursors

N. Soufir, S. Queille, M. Liboutet, O. Thibaudeau, F. Bachelier, G. Delestaing, B.C. Balloy, J. Breuer, A. Janin, L. Dubertret, C. Vilmer and N. Basset-Seguin pages 448–453

Serum chemokine profile in patients with bullous pemphigoid

H. Nakashima, M. Fujimoto, N. Asashima, R. Watanabe, Y. Kuwano, N. Yazawa, N. Maruyama, H. Okochi, A. Kumanogoh and K. Tamaki pages 454–459

Resolution of inflammatory acne vulgaris may involve regulation of CD4+ T-cell responses to Propionibacterium acnes

H.E. Wilcox, M.D. Farrar, W.J. Cunliffe, K.T. Holland and E. Ingham pages 460–465

Tight junction components occludin, ZO-1, and claudin-1, -4 and -5 in active and healing psoriasis

S. Peltonen, J. Riehokainen, K. Pummi and J. Peltonen pages 466–472

Cutaneous mucinosis associated with dermatomyositis and nephrogenic fibrosing dermopathy: fibroblast hyaluronan synthesis and the effect of patient serum

M. Edward, L. Fitzgerald, C. Thind, J. Leman and A.D. Burden pages 473–479

Sweet's syndrome: a spectrum of unusual clinical presentations and associations

C.Y. Neoh, A.W.H. Tan and S.K. Ng pages 480–485

Cutaneous side-effects in patients with rheumatic diseases during application of tumour necrosis factor-α antagonists

H-H. Lee, I-H. Song, M. Friedrich, A. Gauliard, J. Detert, J. Röwert, H. Audring, S. Kary, G.-R. Burmester, W. Sterry and M. Worm pages 486–491

Changes in skin barrier function following long-term treatment with moisturizers, a randomized controlled trial

I. Buraczewska, B. Berne, M. Lindberg, H. Törmä and M. Lodén pages 492–498

Oral treatment with probiotic Lactobacillus johnsonii NCC533 (La1) for a specific part of the weaning period prevents the development of atopic dermatitis induced after maturation in model mice, NC/Nga

R. Inoue, A. Nishio, Y. Fukushima and K. Ushida pages 499–509

Contact dermatitis and allergy

Compositae dermatitis from airborne parthenolide

E. Paulsen, L.P. Christensen and K.E. Andersen pages 510–515

Dermatological surgery and lasers

Comparative study of photodynamic therapy vs. CO2 laser vaporization in treatment of condylomata acuminata, a randomized clinical trial

K. Chen, B.Z. Chang, M. Ju, X.H. Zhang and H. Gu pages 516–520

Epidemiology and health services research

Teledermatology in the U.K.: lessons in service innovation

T.L. Finch, F.S. Mair and C.R. May pages 521–527

The Family Dermatology Life Quality Index: measuring the secondary impact of skin disease

M.K.A. Basra, R. Sue-Ho and A.Y. Finlay pages 528–538 Erratum article printed in: British Journal of Dermatology 156:4 p. 791 doi: 10.1111/j.1365-2133.2007.07825.x

Photobiology

Repeated low-dose ultraviolet (UV) B exposures of humans induce limited photoprotection against the immune effects of erythemal UVB radiation

J. Narbutt, A. Lesiak, A. Sysa-Jedrzejowska, A. Wozniacka, A. Cierniewska-Cieslak, J. Boncela, C. Jochymski, W. Kozlowski, A. Zalewska, M. Skibinska and M. Norval pages 539–547

Therapeutics

A pilot study of the safety and efficacy of picolinic acid gel in the treatment of acne vulgaris

M.P. Heffernan, M.M. Nelson and M.J. Anadkat pages 548–552

Treatment of refractory erosive oral lichen planus with extracorporeal photochemotherapy: 12 cases

A.D. Guyot, D. Farhi, S. Ingen-Housz-Oro, A. Bussel, N. Parquet, C. Rabian, H. Bachelez and C. Francès pages 553–556

Concise communication

Reporting of study design in titles and abstracts of articles published in clinically oriented dermatology journals

R. Ubriani, N. Smith and K.A. Katz pages 557–559

Case reports

Solitary plaque mycosis fungoides on the penis responding to topical imiquimod therapy

L.Y.T Chiam and Y.C. Chan pages 560–562

Paraneoplastic pemphigus without an underlying neoplasm

G.T. Park, J.H. Lee, S.J. Yun, S.C. Lee and J.B. Lee pages 563–566

Allogeneic bone marrow transplantation in a 7-year-old girl with congenital erythropoietic porphyria: a treatment dilemma

S.M. Taibjee, O.E. Stevenson, A. Abdullah, C.Y. Tan, P. Darbyshire, C. Moss, H. Goodyear, A. Heagerty, S. Whatley and M.N. Badminton pages 567–571

Gene corner

Galli–Galli disease is an acantholytic variant of Dowling–Degos disease

E. Sprecher, M. Indelman, Z. Khamaysi, J. Lugassy, D. Petronius and R. Bergman pages 572–574

Correspondence

A survey of approach to wound healing by secondary intention among British Society for Dermatological Surgery members

A. Lally and R. Turner pages 575–576

Acute necrotic pancreatitis induced by severe hypercalcaemia due to tacalcitol ointment

C. Knackstedt R. Winograd A. Koch F. Abuzahra C. Trautwein and H.E. Wasmuth pages 576–577

Raised limb bands in infancy: a post-traumatic aetiology?

M.M. Marque B. Guillot G. Le Gallic and D. Bessis pages 578–579

Human papillomavirus type 5 infection in a patient with Hailey–Hailey disease successfully treated with imiquimod

C-C. Chan H-Y. Thong Y-C. Chan and Y-H. Liao pages 579–581

Perioral dermatitis in a patient with myasthenia gravis following systemic corticosteroid treatment

J.M. Goss K.M. Nord M.R. Olarte M.E. Grossman pages 582–582

Confluent and reticulated papillomatosis successfully treated with amoxicillin

R.F. Davis and K.E. Harman pages 583–584

Folliculotropic mycosis fungoides with CD30+ large-cell transformation in a young woman: beneficial effect of bexarotene

C. Mitteldorf R. Stadler H.P. Bertsch and C. Neumann pages 584–586

Lichen scrofulosorum caused by Mycobacterium szulgai: a new cause of a tuberculide reaction

G.L. Ross, H. Chong, T. Collyns, D.M. Gascoyne-Binzi and R.P.E. Sarkany pages 586–587

Long-term efficacy of topical tacrolimus on oral lesions of chronic graft-versus-host disease

J.-C. Fricain, V. Sibaud, N. Swetyenga, R. Tabrizi, F. Campana and A. Taïeb pages 588–590

Minocycline-induced pigmentation of pre-existing capillaritis

V. Madan and J.T. Lear pages 590–591

Anti-desmoglein-1 antibodies are prevalent in Tunisian patients with hydatidosis and leishmaniasis

M. Kallel Sellami, M. Zitouni, W. Tombari, M. Ben Ayed, O. Abida, L. Laadhar, M. Mokni, B. Fezza, H. Turki, I. Mokhtar, A. Ben Osman, R. Kamoun Mohamed, P. Joly, F. Tron, D. Gilbert, H. Masmoudi, S. Makni and (members of the 'Franco-Tunisian Group of Survey and Research on Tunisian endemic pemphigus') pages 591–593

Needle-free anaesthesia prior to botulinum toxin type A injection treatment of palmar and plantar hyperhidrosis

A. Benohanian pages 593–596

Increased human papillomavirus type 31 DNA load in a verrucous high-grade intraepithelial neoplasia of a human immunodeficiency virus-infected patient with extensive bowenoid papulosis

A. Kreuter, N.H. Brockmeyer, H. Pfister, P. Altmeyer, U. Wieland and For the German Competence Network HIV/AIDS pages 596–598

Telangiectatic cutaneous metastasis from carcinoma of the prostate

S. Reddy, R.H. Bang and M.E. Contreras pages 598–600

Isolated recurrence of vesicobullous incontinentia pigmenti in a schoolgirl

S. Darné and A.J. Carmichael pages 600–602

Infantile systemic hyalinosis: a case report and mutation analysis in a Chinese infant

Y-C. Huang, Y-Y. Xiao, Y-H. Zheng, W. Jang, Y-L. Yang and X-J. Zhu pages 602–604

Scytalidium infection associated with dyskeratosis congenita

J. Garg, R. Tilak, A.K. Gulati, S. Singh, P. Prakash and A. Garg pages 604–606

Pityriasis rubra pilaris and hypothyroidism. Efficacy of thyroid hormone replacement therapy in skin recovery

V. Orlandini, O. Cogrel, M.S. Doutre, C. Beylot and M. Beylot-Barry pages 606–607

Nicorandil-induced peristomal ulcers: is nicorandil also associated with gastrointestinal fistula formation?

S. Ogden, Y. Mukasa, C.C. Lyon and I.H. Coulson pages 608–609

Variability in the clinical pattern of cutaneous side-effects of drugs with systemic symptoms: does a DRESS syndrome really exist?

S.H. Kardaun, A. Sidoroff, L. Valeyrie-Allanore, S. Halevy, B.B. Davidovici, M. Mockenhaupt and J.C. Roujeau pages 609–611

Book review

Textbook of Pediatric Dermatology 2nd edition

D.K.B. Armstrong pages 611–611

News and Notices

News and Notices

pages 611–611

Corrigenda

Corrigenda

pages 612–612

Corrigenda

pages 612–612

Corrigenda

pages 612–612

RESEARCH SNIPPETS DOI 10.1111/j .1365-2133.2007.07787.x

Testosterone metabolism to 5a-dihydrotestosterone and synthesis of sebaceouslipids is regulated by the peroxisome proliferator-activated receptor ligandlinoleic acid in human sebocytesDespite clinical evidence that androgens stimulate sebaceous lipids, experiments in vitrohave shown ineffectiveness. This contradiction led to the assumption that cofactorsmay be required for lipid regulation. In this study, the combination of testosteroneand the natural peroxisome proliferator-activated receptor (PPAR)-a/d ligand linoleicacid exhibited a synergistic effect on lipid synthesis and on 5a-dihydrotestosteronelevels in cultured sebocytes. Furthermore, the 5a-reductase type I inhibitor LY191704only unfolded its action under concurrent cell exposure to both compounds. Thesedata suggest a catalytic PPAR effect on cellular testosterone activation and pointtowards a new mode of androgen involvement in sebum production and probably ofthe pathogenesis of sebaceous gland-associated diseases such as acne.Makrantonaki E, Zouboulis CC. Testosterone metabolism to 5a-dihydrotestosterone and synthesis of sebaceous lipids is regulatedby the peroxisome proliferator-activated receptor ligand linoleic acid in human sebocytes. Br J Dermatol 2007; 156:428–32.

Serum chemokine profile in patients with bullous pemphigoidIn bullous pemphigoid (BP), autoantibodies initiate a cascade of inflammatory responsethat causes subepidermal blisters. Nakashima et al. report serum levels of chemokines,including IP-10, MIG, MIP-1a, MIP-1b, RANTES, eotaxin, MCP-1, MCP-2, MCP-3 andGRO-a, in patients with BP. Serum levels of IP-10 and MCP-1 were significantlyincreased in patients with BP compared with healthy controls. Furthermore, serumlevels of IP-10, MIG, MCP-1 and eotaxin in patients with BP significantly increasedwith the affected area. This study suggests that, while various chemokines are involvedin the pathomechanism of BP, MCP-1 and IP-10 have a significant contribution inleucocyte recruitment.Nakashima H, Fujimoto M, Asashima N et al. Serum chemokine profile in patients withbullous pemphigoid. Br J Dermatol 2007; 156:454–59.

Resolution of inflammatory acne vulgaris may involve regulation of CD4+ T-cellresponses to Propionibacterium acnesPropionibacterium acnes is strongly implicated in the pathogenesis of inflammatory acne butits role remains unclear. Wilcox et al. have analysed the CD4+ T-cell response to P. acnesand bacterial heat shock proteins in patients with acne, individuals with acne that hadpreviously resolved (‘resolved individuals’) and nonacne controls by lymphocytetransformation and limiting dilution analysis (LDA). A significantly lower proportionof resolved individuals demonstrated a single-hit LDA response to P. acnes. This mayrepresent negative regulation of the CD4+ T-cell response to P. acnes in these subjects andsuggests that such regulation may be important in the spontaneous resolution of acne.Wilcox HE, Farrar MD, Cunliffe WJ et al. Resolution of inflammatory acne vulgaris may involve regulation of CD4+ T-cellresponses to Propionibacterium acnes. Br J Dermatol 2007; 156:460–65.

Compositae dermatitis from airborne parthenolideAnecdotal clinical and plant chemical reports suggest Compositae dermatitis to be trulyairborne. Feverfew (Tanacetum parthenium) and its main allergen parthenolide areEuropean sensitizers suspected of causing airborne contact allergy. Among 12individuals allergic to feverfew/parthenolide, eight had positive patch-test reactionsto a high-volume air sampler filter extract, collected around feverfew plants.Subsequently, parthenolide was detected in the extract at a concentration of510 ng mL)1, and the most sensitive person tested positive to 8Æ1 ng parthenolide.Thus, the nonvolatile parthenolide is carried on airborne plant particles in allergenicamounts and thereby induces dermatitis on exposed skin in some feverfew-allergicpatients.Paulsen E, Christensen LP, Andersen KE. Compositae dermatitis from airborneparthenolide. Br J Dermatol 2007; 156:510–15.

(a) (b)

(c)

(e)

(d)

EDITORIAL DOI 10.1111/j .1365-2133.2007.07750.x

Has teledermatology in the U.K. finally failed?

Although two papers published in this issue provide us with

additional information about the current state of teledermatol-

ogy services, we probably remain some way off a final evalua-

tion of the current role of this technology in the provision of

dermatology services in the NHS. Over the past decade many

clinicians in the U.K. have expressed concerns over the devel-

opment of teledermatology and this was probably reflected by

the fact that the Action on Dermatology in 2001 had only three

pilot sites out of 15 devoted to looking at teledermatology as a

way of improving access for patients to secondary care. One of

these never started and the status of the other two is unknown.

It was thought by the Department of Health and Government

policy advisers that the use of digital photography and

improvements in electronic communications such as store-

and-forward technology would provide a potential panacea for

specialties like dermatology in an effort to decrease outpatient

waiting times. The reason for this seems simple: the images

obtained from digital photography are of such high quality

that why couldn’t images of skin conditions be e-mailed to a

dermatologist and save the patient a hospital visit? This would

reduce costs and decrease waiting times, but clinical effective-

ness and cost-effectiveness of such services remain largely

untested. If only dermatology were that simple! In an article

on palpation of the skin, Dr Cox recently summarized some of

the reasons, obvious to dermatologists, why this approach is

simplistic.1 They include the fact that it is often necessary to

palpate the skin in order to help make the diagnosis, one can-

not judge the patient’s concerns and anxieties just from a series

of digital images, other areas of pathology are missed that are

not photographed and often the images are of poor quality,

thus making a diagnosis difficult.

In a recent NHS R&D Technology Assessment, patients

referred to a 2-week wait clinic were invited to have a series

of digital photographs, with and without dermoscopy, imme-

diately before their face-to-face consultation.2 An unexpectedly

high proportion (33%) of referrals proved to have a malig-

nancy or a severely dysplastic lesion, with almost 22% having

a malignant melanoma or squamous cell carcinoma, possibly

reflecting the rise in incidence of skin cancers reported else-

where. If the highest level of clinician confidence had been

applied, no cancers would have been missed, but only 20% of

patients would have avoided an outpatient appointment. It

was concluded that it is unlikely that this approach can dra-

matically reduce the need for conventional clinical consulta-

tions while still maintaining clinical safety.2

The first paper in this issue is a literature review of teleder-

matology evaluation and research.3 The authors carried out a

systematic review of the literature over the last 40 years and

concluded that most of the studies published are feasibility

studies rather than phase III or phase IV studies. They found

that compared with other specialties in telemedicine, teleder-

matology seems to be a ‘mature application’. However, more

research on clinical outcomes such as preventable referrals and

time to recovery would be needed to prove that teledermatol-

ogy is indeed a promising and cost-saving technology.

The second paper reports qualitative research into telederma-

tology in the U.K. and is an extremely thorough study invol-

ving interviewing various stakeholders in teledermatology

development within the NHS.4 One of the authors’ conclusions

was that the original policy vision of how teledermatology

would be used as a technological answer for long waiting lists

and consultant shortages has failed to be realized. They also

found that teledermatology services were initially viewed as a

diagnostic service but then increasingly gained acknowledge-

ment as a triage and management service.

The successful implementation of teledermatology as a rout-

ine service requires understanding of and paying great atten-

tion to the interplay between social and technical aspects of

teledermatology. This means it is not a quick, simple fix for

dermatology. However, it is uncertain that the politicians and

policy makers who lead developments in the Health Service

will listen.

J . S .C . ENGL I SH

D . J . E EDY*

Department of Dermatology,

Queen’s Medical Centre,

Nottingham University Hospitals

NHS Trust, Nottingham

NG7 2UH, U.K.

*Department of Dermatology,

Craigavon Area Hospital Group,

Portadown BT63 5QQ, U.K.

E-mail: [email protected]

References

1 Cox NH. Palpation of the skin – an important issue. J R Soc Med2006; 99:598–600.

2 Bowns IR, Collins K, Walters SJ, McDonagh AJG. Telemedicine indermatology: a randomised controlled trial. Health Technol Assess 2006;

10(43):1–58.3 Eminovic N, de Keizer NF, Bindels PJE, Hasman A. Maturity of tele-

dermatology evaluation research: a systematic literature review. Br J

Dermatol 2007; 156:412–19.4 Finch TL, Mair FS, May CR. Teledermatology in the U.K.: lessons in

service innovation. Br J Dermatol 2007; 156:521–7.

� 2007 The Authors

Journal compilation � 2007 British Association of Dermatologists • British Journal of Dermatology 2007 156, pp411 411

REVIEW ARTICLE DOI 10.1111/j .1365-2133.2006.07627.x

Maturity of teledermatology evaluation research:a systematic literature reviewN. Eminovic, N.F. de Keizer, P.J.E. Bindels* and A. Hasman

Departments of Medical Informatics and *General Practice, Academic Medical Centre, Meibergdreef 15, 1105 AZ Amsterdam, The Netherlands

CorrespondenceN. Eminovic.


Accepted for publication15 August 2006

Key wordsevaluation, maturity, phases, systematic review,

teledermatology, telemedicine

Conflicts of interestNone declared.

Summary

Background There is a growing interest in teledermatology in today’s clinical prac-tice, but the maturity of the evaluation research of this technology is still unclear.Objectives This systematic review describes the maturity of teledermatology evalu-ation research over time and explores what kind of teledermatology outcomemeasures have been evaluated.Methods Systematic review of literature found in Medline database (1966 up toApril 2006). A telemedicine evaluation strategy consisting of four consecutiveresearch phases (parallel to drug and diagnostics evaluation research) extendedwith a fifth postimplementation phase was used to classify all included studies bytwo independent reviewers. In addition, main characteristics (store-and-forwardor real-time, study design, outcome measures) were registered.Results Three hundred and forty-five papers were systematically selected from Med-line, and 244 papers were excluded. For two randomized controlled trials (RCTs),multiple papers in phase III were found. After correcting for this, 99 studiesremained included (11 phase I, 72 phase II, two phase III, six phase IV, eightpostimplementation phase). The number of phase II studies is the largest and stillgrowing, while other phases are much less represented. Diagnostic accuracy wasthe most often used outcome measure and was found in phase I, II and IV. Store-and-forward teledermatology has been evaluated more since 2001, but mostphase IV studies (RCTs, including cost aspects) are on real-time teledermatology.Conclusions Most teledermatology evaluation studies are classified as feasibility stud-ies (phase II). The number of phase III and IV studies remains low through theyears. Compared with other specialties in telemedicine (i.e. telesurgery, telepaedi-atrics), teledermatology seems to be a mature application. However, more evalu-ation studies with a focus on clinical outcomes such as preventable referrals ortime to recovery are needed to prove that teledermatology indeed is a promisingand cost-saving technology.

Teledermatology can be defined as the use of imaging and

telecommunication technologies to provide skin services by a

dermatologist to another health professional (general practi-

tioner, nurse, other specialist) or directly to a patient. Tele-

medicine in general enables health professionals to

communicate with each other and with their patients over a

geographical or physical distance. Together with other visually

oriented specialties, dermatology is one of the clinical special-

ties in which telemedicine is most applied.

Two types of teledermatology can be distinguished: store-

and-forward (SAF) and real-time (RT) teledermatology.1 The

SAF variant uses asynchronous data transfer technology (e.g.

e-mail) while RT teledermatology is based on synchronous

data transfer technologies (e.g. videoconferencing software).

RT teledermatology requires both communicating parties (e.g.

physician, dermatologist and patients) to be available at the

same time, which is not necessary when using an SAF variant.

This variation in teledermatology is not only technology

based, but also has impact on organizational and clinical char-

acteristics of teledermatology.

Although telemedicine and teledermatology are increasingly

becoming integrated in local regular healthcare systems, there

Additional information added after online publication 5th January2007: Preliminary results for this study were presented at the conference

Medical Informatics in Europe (MIE), August 2006, Maastricht, theNetherlands.


412 Journal Compilation � 2006 British Association of Dermatologists • British Journal of Dermatology 2007 156, pp412–419

is lack of valid scientific evidence showing positive effects,

which is mainly due to the low quality of the studies.2–5

Implementation and evaluation of teledermatology seem to be

performed in parallel while medical interventions are usually

first fully investigated through different research phases before

being implemented.

The primary aim of this review is to describe the maturity

status of teledermatology evaluation research over time, for

both SAF and RT, as a function of the four consecutive tele-

medicine evaluation phases defined by Holle and Zahlmann.6

These four phases are comparable with the four phases of

drug research. The secondary aim of the study is to explore

the outcome measures used in the different evaluation phases.

Materials and methods

Search strategy

We identified English language published teledermatology

studies by searching Medline database (1966 up to April

2006) using various search queries combining the key words

‘teledermatology’,‘dermatology’,‘telemedicine’,‘skin’,‘e-health’

and ‘electronic mail’. As we were only interested in ori-

ginal full papers reporting on the evaluation of a specific

teledermatology service, we excluded literature reviews, com-

ments, abstracts, letters and editorials. The second step inclu-

ded the manual screening of the selected full papers by two

independent reviewers. Papers which were not about derma-

tology but about another specialty (e.g. radiology, pathology)

and papers where the evaluation of a specific teledermatology

service was not the primary aim were excluded. Papers on a

telemedicine application for several specialties were included

only if the results on dermatology were separately reported.

Literature reviews1,7 were screened to retrieve possibly missed

references. Papers from conference proceedings were excluded

if a full journal paper on the same study was obtained in the

selection procedure.

Classification of evaluation studies

Included studies were assigned independently to one of the

four phases of the Holle and Zahlmann strategy.6 With per-

mission of the authors, we added an extra phase: the postim-

plementation phase. This fifth phase should be considered as

an extra category and not specifically as a continuation of the

four phases. Table 1 shows the five phases which are deter-

mined by the study design, the number of participants in the

study and the setting of the study.

We distinguished two main study designs: observational

and intervention studies. Intervention studies were further

divided into: uncontrolled trial; nonrandomized controlled

trial with the same patients as control group (e.g. a patient

has been seen through teledermatology first and face-to-face

afterwards); nonrandomized controlled trial with other

patients as control group; and randomized controlled trial

(RCT).

In cases where multiple papers on one study had been pub-

lished, the study was classified once in one phase. However, if

multiple papers on one study were classified in different

phases (for example, clinical outcomes in phase III and costs

in phase IV), than the study was classified in each phase separ-

ately. When different phases of a study were described in one

paper, the paper was classified only once, in the highest

phase.

We predefined the following outcome measures: diagnos-

tic accuracy, diagnostic reliability, patient satisfaction with

provided care, patient and doctor satisfaction with the sys-

tem, costs, preventable referrals, quality of images, learning

effect in doctors and delay in treatment. When measuring

diagnostic accuracy the telediagnosis is compared with a

Table 1 Four phases of Holle and Zahlmann,6 and added fifth phase (‘postimplementation’) with their characteristics and main criteria used inthis review to classify teledermatology studies

Phase Study design Usual participants Main criteria used by the reviewers

I Exploratory, small interventionstudies

Researchers and project members,simulated patients

Experimental setting, i.e. teleconsultationdid not result in a diagnosis or

treatment planII Feasibility studies, can be controlled

intervention studies, but rarelyinvolve a separate control group

Specially trained and highly

motivated potential users,real or simulated patients

Field or experimental studies with

potential users

III RCT Unselected sample of users withminimal training, real patients

RCT on clinical outcomes

IV (Simulation) cost studies usingthe benefits results from an RCT

Not always applicable,see also phase III

Costs are compared betweenconventional dermatological care and

teledermatology and/or with any kindof benefits resulting from an RCT

Postimplementation Telemedicine part of the regularcare, observational study

Actual users of telemedicine Teledermatology is fully integratedin the regular care

RCT, randomized controlled trial.


Journal Compilation � 2006 British Association of Dermatologists • British Journal of Dermatology 2007 156, pp412–419

Maturity of teledermatology evaluation research, N. Eminovic et al. 413

gold standard diagnosis (face-to-face diagnosis, biopsy etc.).

Diagnostic reliability is diagnostic agreement between differ-

ent examiners who determined a diagnosis using the same

or a different, but not gold standard, diagnostic method.

Patient satisfaction with the teledermatology system refers to

how the patients experience a new technology such as tele-

dermatology, while patient satisfaction with care refers to

how the patients feel about the health care delivered in

general.

All discrepancies in classifying the selected papers between

the reviewers were solved by consensus. Interobserver reliabil-

ity regarding the classification into phases was calculated to

determine the validity of the classification procedure.

Results

The Medline search resulted in 345 unique references. Of

these studies, 181 studies were included and 164 excluded

(see Fig. 1). Forty-five papers were about another specialty

(radiology, pathology) or about all kinds of medical condi-

tions including dermatology as a subset but without report-

ing separate results. After reading the full papers, 80 papers

were excluded. Most of these papers did not report on an

evaluation of a specific teledermatology service (n ¼ 46).

Five studies needed to be excluded as they were in some

way duplicate publications: conference proceedings and a full

version paper. Finally, 101 papers were included for the clas-

sification into five phases. It appeared that in two situations

two related papers on the same RCTs were classified in the

same phase, phase III. Following the rules, these studies were

classified once, making the total number of classified studies

as 99.

The two reviewers agreed about the classification of the

studies into phases in 84% of cases (j ¼ 0Æ7).

Phase trends

In 1996 two teledermatology evaluation studies were pub-

lished. These first teledermatology studies were a phase I8 and

phase II study9 (Fig. 2). Phase I studies have been published

sporadically with a total of 11 studies up to 2006.

Seventy-two studies were assigned to phase II. In 1998, 10

phase II studies were published while only one phase I and

one postimplementation study were published. The same high

number of phase II studies was found again in 2003 and

2005.

The number of phase III studies was limited to two, pub-

lished in 2002 and 2004.10,11 Phase IV and the postimple-

mentation phase were represented by six and eight studies,

respectively. In 2000 three cost studies were published12–14

and another two were published in 2001.15,16 The last pub-

lished phase IV study dates from 2003.17

Two large groups of trials were found: U.K. Multicentre

Teledermatology Trials14,18–21 and Health Waikato Tele-

dermatology Trials.13,22,23 These trials investigated RT

Fig 1. Two steps of the literature selection

procedure. The first step of selection

procedure was based on title, abstract and

paper characteristics, while the second step

was based on the full paper. RCT, randomized

controlled trial; td, teledermatology; conf.,

conference.

Fig 2. Cumulative number of studies per phase and over the years.



414 Maturity of teledermatology evaluation research, N. Eminovic et al.

teledermatology and covered different phases. The U.K. trials

covered phases I–IV, while the Waikato trials covered phases

II–IV.

Outcome measures and study design

The diagnostic accuracy was the most common outcome

measure used in 53 studies (Table 2). In seven of 11 phase

I studies, the quality of images was investigated. The two

phase III RCTs investigated delay in treatment and prevent-

able referrals as clinical outcome measures, but also inclu-

ded patient and doctor satisfaction with the system. Central

themes of the eight postimplementation phase studies were

patient and doctor satisfaction with the teledermatology

system.

In total, there were 43 intervention studies controlled by

the same group of patients. An uncontrolled study design was

applied in 30 studies, mainly in phase II (21 of 72), but also

in phase I (six of 11). We found four RCTs in phase II which

were small and not focusing on clinical outcomes and could

therefore not be classified as phase III or IV.24–27

Store-and-forward vs. real-time teledermatology

The majority of the included studies (63%) investigated the

SAF variant of teledermatology (Fig. 3). The RT variant was

investigated in 29% of the studies, with the majority per-

formed before 2001. Since 2001 more SAF applications were

the subject of evaluation compared with RT teledermatology

applications. The combination of the two technologies was

evaluated twice.14,28 In six studies it was not clear what kind

of technology was used. Three of these six studies were phase

I studies in which technical aspects and quality of images

were evaluated.29–31 The other three studies were two recent

phase II32,33 and one postimplementation phase study.34

SAF technology was investigated in all phase III studies,

while four of six phase IV studies evaluated RT12,13,15,16 and

one SAF teledermatology.17

Table 2 Outcome measures and study designs in five phases

Phase n (SAF, RT, both) Outcome measures (number of studies) Study design (number of studies)

I8,21,29–31,38–43 11 (5, 3, 0)

3 unknownSAF or RT

Diagnostic accuracy (5)

Diagnostic reliability (3)Quality of images (7)

Doctor satisfaction with the system (3)Preventable referrals (1)

Confidence in the findings (1)Technical aspects (2)

Observational (2)

Intervention controlled bythe same patients (3)

Uncontrolled (6)

II9,18–20,22–28,32,33,44–102 72 (47, 18, 1)2 unknown

SAF or RT

Diagnostic accuracy (47)Diagnostic reliability (17)

Quality of images (12)Patient satisfaction with the system (16)

Doctor satisfaction with the system (11)Patient satisfaction with care (5)

Costs (5)Preventable referrals (12)

Learning effect GPs (3)

Confidence in the findings (1)Time needed for consultation (8)

Technical aspects (3)

RCT (4)Intervention controlled by

other patients (7)Intervention controlled by the

same patients (40)Uncontrolled (21)

III10,11 2 (2, 0, 0) Patient satisfaction with the system (2)

Doctor satisfaction with the system (2)Patient satisfaction with care (2)

Preventable referrals (1)Delay in treatment (2)

RCT (2)

IV12–17 6 (1, 4, 1) Diagnostic accuracy (1)Costs (6)

Preventable referrals (3)Time needed for consultation (1)

RCT (6)

Postimplementation34,103–109 8 (4, 3, 0)1 unknown

SAF or RT

Patient satisfaction with the system (2)Doctor satisfaction with the system (2)

Costs (2)Delay in treatment (1)

Time needed for consultation (1)

Observational (5)Uncontrolled (3)

SAF, store-and-forward; RT, real-time; GP, general practitioner; RCT, randomized controlled trial.




Discussion

Teledermatology has been the subject of evaluation studies for

10 years now. The results of our review show that evaluation

of teledermatology takes place in different research phases.

Despite a growing implementation of teledermatology in

today’s clinical practice, there are no clear trends towards

phase III and IV studies in more recent years in comparison

with the first years of teledermatology evaluation.

The number of phase II studies grew continuously over the

past 10 years and is still growing. Phase III, IV and V studies

are much less represented. There are different possible expla-

nations for this. Phase I studies, usually reporting on technical

issues such as, for example, quality of images, are often not

published separately as scientific journal papers (possible pub-

lication bias). Technical evaluations are also more likely to be

published in technical journals which are not covered by the

Medline database.

Only two RCTs could be classified in phase III and six in

phase IV. Performing high-level evidence studies such as RCTs

in telemedicine is often seen as difficult as telemedicine is a

complex intervention depending on local variations in admin-

istration and organisation.35,36 Characteristics of teledermatol-

ogy such as its relatively low risk, low investment costs and

the possibility of using it in addition to regular care are likely

to speed up its implementation despite the lack of scientific

research on the pros and cons. However, detailed and sophis-

ticated evaluation of a new technology requires many

resources and should be performed only after some basic but

important evaluation steps have been carried out in phase I

and II. Furthermore, for certain nonclinical outcome measures

of teledermatology (e.g. diagnostic accuracy), a phase II study

should provide sufficiently strong evidence and a phase III or

IV study is not expected. These issues are likely to reduce the

push to perform RCTs in teledermatology.

Postimplementation studies are, like phase I studies, not

often published in scientific journals. Commercial telederma-

tology suppliers probably perform this kind of evaluation fre-

quently, but mainly for their own use without academic

ambitions. In our review we obtained only Medline indexed

literature. Further in-depth analysis of phase I or postimple-

mentation phase studies would require obtaining literature

databases of technical journals and nonpublicly available

reports.

More detailed and sophisticated evaluation of any new tech-

nology is expected to be carried out after some basic evalu-

ation steps have been performed, such as in phase I and II. As

no or limited scientific evidence about the impact of teleder-

matology on clinical outcome has been demonstrated,4 more

detailed evaluation studies (phase III and IV) are needed.

Diagnostic accuracy is the most often evaluated outcome

measure in the teledermatology studies included in our

review. This result was not surprising as this outcome measure

is strongly related to the visual aspect of teledermatology.

With the changing technology (i.e. different cameras with dif-

ferent possibilities and settings) and large variations in teleder-

matology settings, this outcome is likely to be studied

repeatedly in most teledermatology evaluations. However, it is

regrettable that important clinical outcomes such as prevent-

able referrals and delay in treatment are not more often inves-

tigated, especially in phase III and IV studies.

Earlier studies showed that RT teledermatology was more

clinically efficient than the SAF variant.14 However, SAF tele-

dermatology is less expensive and under certain circum-

stances more cost effective, making it more attractive than

the RT variant. This increasing interest in SAF has also been

shown in our review as RT was especially popular in the

early evaluations and SAF has been much more popular over

the past 5 years. However, most phase IV studies that we

found were about RT. Therefore we support Whited’s con-

clusion that there is a lack of economic evaluation of SAF

teledermatology.7

An earlier study concluded that teledermatology was

maturing because of the reasonably good quality and quan-

tity of the published literature when compared with other

specialties (i.e. telesurgery, telepaediatrics).37 Hailey et al.4

found for teledermatology the strongest indication for bene-

fits of telemedicine applications. Although these conclusions

might be true when compared with other telemedicine

applications, our review shows that there is no clear trend in

maturity of teledermatology evaluation research over the past

10 years as there is a lack of phase III and IV studies. At this

moment, we found no indications for the maturation of tele-

dermatology evaluation research or that the number of phase

III and phase IV studies is likely to increase in the near

future. The timing of our study might be too early to meas-

ure an increase of phase III and phase IV studies as these

Fig 3. Cumulative number of store-and-forward (SAF) vs. real time

(RT) teledermatology evaluation studies in the past 10 years.




studies are often time consuming in both performing and

reporting. Although teledermatology seems to be a valuable

application, we believe that there is a need for more phase

III and IV studies to provide a high level of evidence on

positive clinical outcomes and cost effectiveness of telederma-

tology prior to its implementation and wide use in clinical

practice.

References

1 Eedy DJ, Wootton R. Teledermatology: a review. Br J Dermatol2001; 144:696–707.

2 Whitten PS, Mair FS, Haycox A et al. Systematic review of costeffectiveness studies of telemedicine interventions. BMJ 2002;

324:1434–7.3 Mair F, Whitten P. Systematic review of studies of patient satisfac-

tion with telemedicine. BMJ 2000; 320:1517–20.4 Hailey D, Roine R, Ohinmaa A. Systematic review of evidence for

the benefits of telemedicine. J Telemed Telecare 2002; 8 (Suppl. 1):1–30.

5 Hersh WR, Helfand M, Wallace J et al. Clinical outcomes resultingfrom telemedicine interventions: a systematic review. BMC Med

Inform Decis Mak 2001; 1:5.6 Holle R, Zahlmann G. Evaluation of telemedical services. IEEE Trans

Inf Technol Biomed 1999; 3:84–91.7 Whited JD. Teledermatology research review. Int J Dermatol 2006;

45:220–9.8 Crump WJ, Levy BJ, Billica RD. A field trial of the NASA Telemed-

icine Instrument Pack in a family practice. Aviat Space Environ Med1996; 67:1080–5.

9 Jones DH, Crichton C, Macdonald A et al. Teledermatology in theHighlands of Scotland. J Telemed Telecare 1996; 2 (Suppl. 1): 7–9.

10 Whited JD, Hall RP, Foy ME et al. Teledermatology’s impact ontime to intervention among referrals to a dermatology consult

service. Telemed J E Health 2002; 8:313–21.

11 Collins K, Bowns I, Walters S. General practitioners’ perceptionsof asynchronous telemedicine in a randomized controlled trial of

teledermatology. J Telemed Telecare 2004; 10:94–8.12 Wootton R, Bloomer SE, Corbett R et al. Multicentre randomised

control trial comparing real time teledermatology with conven-tional outpatient dermatological care: societal cost-benefit analysis.

BMJ 2000; 320:1252–6.13 Oakley AM, Kerr P, Duffill M et al. Patient cost-benefits of realtime

teledermatology – a comparison of data from Northern Irelandand New Zealand. J Telemed Telecare 2000; 6:97–101.

14 Loane MA, Bloomer SE, Corbett R et al. A comparison of real-timeand store-and-forward teledermatology: a cost-benefit study. Br J

Dermatol 2000; 143:1241–7.15 Loane MA, Oakley A, Rademaker M et al. A cost-minimiza-

tion analysis of the societal costs of realtime teledermatologycompared with conventional care: results from a randomized

controlled trial in New Zealand. J Telemed Telecare 2001; 7:233–8.

16 Loane MA, Bloomer SE, Corbett R et al. A randomized controlledtrial assessing the health economics of realtime teledermatology

compared with conventional care: an urban versus rural perspec-tive. J Telemed Telecare 2001; 7:108–18.

17 Whited JD, Datta S, Hall RP et al. An economic analysis of a storeand forward teledermatology consult system. Telemed J E Health

2003; 9:351–60.18 Gilmour E, Campbell SM, Loane MA et al. Comparison of telecon-

sultations and face-to-face consultations: preliminary results of a

United Kingdom multicentre teledermatology study. Br J Dermatol1998; 139:81–7.

19 Loane MA, Corbett R, Bloomer SE et al. Diagnostic accuracy andclinical management by realtime teledermatology. Results from

the Northern Ireland arms of the UK Multicentre TeledermatologyTrial. J Telemed Telecare 1998; 4:95–100.

20 Loane MA, Bloomer SE, Corbett R et al. Patient satisfaction withrealtime teledermatology in Northern Ireland. J Telemed Telecare

1998; 4:36–40.21 Loane MA, Gore HE, Corbett R et al. Preliminary results from the

Northern Ireland arms of the UK Multicentre Teledermatology

Trial: effect of camera performance on diagnostic accuracy. J Tele-med Telecare 1997; 3 (Suppl. 1): 73–5.

22 Oakley AM, Astwood DR, Loane M et al. Diagnostic accuracy ofteledermatology: results of a preliminary study in New Zealand.

NZ Med J 1997; 110:51–3.23 Oakley AM, Duffill MB, Reeve P. Practising dermatology via tele-

medicine. NZ Med J 1998; 111:296–9.24 Kvedar JC, Menn ER, Baradagunta S et al. Teledermatology in a

capitated delivery system using distributed information architec-ture: design and development. Telemed J 1999; 5:357–66.

25 Lim AC, Egerton IB, See A et al. Accuracy and reliability ofstore-and-forward teledermatology: preliminary results from the

St George Teledermatology Project. Australas J Dermatol 2001; 42:247–51.

26 Mann T, Colven R. A picture is worth more than a thousandwords: enhancement of a pre-exam telephone consultation in

dermatology with digital images. Acad Med 2002; 77:742–3.27 Qureshi AA, Brandling-Bennett HA, Giberti S et al. Evaluation of

digital skin images submitted by patients who received practicaltraining or an online tutorial. J Telemed Telecare 2006; 12:79–82.

28 Baba M, Seckin D, Kapdagli S. A comparison of teledermatologyusing store-and-forward methodology alone, and in combination

with web camera videoconferencing. J Telemed Telecare 2005;11:354–60.

29 Vidmar DA, Cruess D, Hsieh P et al. The effect of decreasing digi-tal image resolution on teledermatology diagnosis. Telemed J 1999;

5:375–83.30 Wiltgen M, Gerger A, Wagner C et al. Evaluation of the influence

of image compression to the automatic discrimination of histo-logical images of skin lesions. Methods Inf Med 2004; 43:141–9.

31 Waldron KJ, Enedah C, Gladstone H. Stiffness and texture percep-tion for teledermatology. Stud Health Technol Inform 2005; 111:

579–85.

32 Persaud DD, Jreige S, Skedgel C et al. An incremental cost analysisof telehealth in Nova Scotia from a societal perspective. J Telemed

Telecare 2005; 11:77–84.33 Hicks LL, Boles KE, Hudson S et al. Patient satisfaction with tele-

dermatology services. J Telemed Telecare 2003; 9:42–5.34 Aas IH. Changes in the job situation due to telemedicine. J Telemed

Telecare 2002; 8:41–7.35 Bashshur R, Shannon G, Sapci H. Telemedicine evaluation. Telemed

J E Health 2005; 11:296–316.36 Kaplan B. Evaluating informatics applications – some alternative

approaches: theory, social interactionism, and call for methodolo-gical pluralism. Int J Med Inform 2001; 64:39–56.

37 Krupinski E, Nypaver M, Poropatich R et al. Telemedicine/tele-health: an international perspective. Clinical applications in tele-

medicine/telehealth. Telemed J E Health 2002; 8:13–34.38 Provost N, Kopf AW, Rabinovitz HS et al. Comparison of conven-

tional photographs and telephonically transmitted compresseddigitized images of melanomas and dysplastic nevi. Dermatology

1998; 196:299–304.




39 Wallace S, Sibson L, Stanberry B et al. The legal and risk managementconundrum of telemedicine. J Telemed Telecare 1999; 5 (Suppl. 1):

8–9.40 Chen K, Lim A, Shumack S. Teledermatology: influence of zoning

and education on a clinician’s ability to observe peripherallesions. Australas J Dermatol 2002; 43:171–4.

41 Seidenari S, Pellacani G, Righi E et al. Is JPEG compression ofvideomicroscopic images compatible with telediagnosis? Compar-

ison between diagnostic performance and pattern recognition onuncompressed TIFF images and JPEG compressed ones. Telemed J E

Health 2004; 10:294–303.

42 Lanschuetzer CM, Pohla-Gubo G, Schafleitner B et al. Telepatholo-gy using immunofluorescence/immunoperoxidase microscopy.

J Telemed Telecare 2004; 10:39–43.43 Lamminen H, Ruohonen K, Uusitalo H. Visual tests for measuring

the picture quality of teleconsultations for medical purposes. Com-put Methods Programs Biomed 2001; 65:95–110.

44 Piccolo D, Smolle J, Wolf IH et al. Face-to-face diagnosis vs teledi-agnosis of pigmented skin tumors: a teledermoscopic study. Arch

Dermatol 1999; 135:1467–71.45 Tait CP, Clay CD. Pilot study of store and forward teledermatolo-

gy services in Perth, Western Australia. Australas J Dermatol 1999;40:190–3.

46 White H, Gould D, Mills W et al. The Cornwall dermatology elec-tronic referral and image-transfer project. J Telemed Telecare 1999; 5

(Suppl. 1): 85–6.47 Whited JD, Hall RP, Simel DL et al. Reliability and accuracy of der-

matologists’ clinic-based and digital image consultations. J Am AcadDermatol 1999; 41:693–702.

48 Barnard CM, Goldyne ME. Evaluation of an asynchronous telecon-sultation system for diagnosis of skin cancer and other skin

diseases. Telemed J E Health 2000; 6:379–84.49 Piccolo D, Smolle J, Argenziano G et al. Teledermoscopy – results

of a multicentre study on 43 pigmented skin lesions. J TelemedTelecare 2000; 6:132–7.

50 Krupinski EA, LeSueur B, Ellsworth L et al. Diagnostic accuracyand image quality using a digital camera for teledermatology.

Telemed J 1999; 5:257–63.51 High WA, Houston MS, Calobrisi SD et al. Assessment of the

accuracy of low-cost store-and-forward teledermatology consult-ation. J Am Acad Dermatol 2000; 42:776–83.

52 Lewis K, Gilmour E, Harrison PV et al. Digital teledermatology forskin tumours: a preliminary assessment using a receiver operating

characteristics (ROC) analysis. J Telemed Telecare 1999; 5 (Suppl. 1):

57–8.53 Williams TL, Esmail A, May CR et al. Patient satisfaction with tele-

dermatology is related to perceived quality of life. Br J Dermatol2001; 145:911–17.

54 Williams T, May C, Esmail A et al. Patient satisfaction with store-and-forward teledermatology. J Telemed Telecare 2001; 7 (Suppl. 1):

45–6.55 van den Akker TW, Reker CH, Knol A et al. Teledermatology as a

tool for communication between general practitioners and derma-tologists. J Telemed Telecare 2001; 7:193–8.

56 Taylor P, Goldsmith P, Murray K et al. Evaluating a telemedicinesystem to assist in the management of dermatology referrals. Br J

Dermatol 2001; 144:328–33.57 Jolliffe VM, Harris DW, Morris R et al. Can we use video images

to triage pigmented lesions? Br J Dermatol 2001; 145:904–10.58 Jolliffe VM, Harris DW, Whittaker SJ. Can we safely diagnose pig-

mented lesions from stored video images? A diagnostic comparisonbetween clinical examination and stored video images of pigmented

lesions removed for histology. Clin Exp Dermatol 2001; 26:84–7.

59 Scheinfeld N, Fisher M, Genis P et al. Evaluating patient acceptanceof a teledermatology link of an urban urgent-care dermatology

clinic run by residents with board certified dermatologists. Skinmed2003; 2:159–62.

60 Oliveira MR, Wen CL, Neto CF et al. Web site for training non-medical health-care workers to identify potentially malignant skin

lesions and for teledermatology. Telemed J E Health 2002; 8:323–32.61 Rashid E, Ishtiaq O, Gilani S et al. Comparison of store and for-

ward method of teledermatology with face-to-face consultation.J Ayub Med Coll Abbottabad 2003; 15:34–6.

62 Pak HS, Harden D, Cruess D et al. Teledermatology: an intraobserv-

er diagnostic correlation study, part II. Cutis 2003; 71:476–80.63 Pak HS, Harden D, Cruess D et al. Teledermatology: an intraobserv-

er diagnostic correlation study, part I. Cutis 2003; 71:399–403.64 Eminovic N, Witkamp L, Ravelli AC et al. Potential effect of

patient-assisted teledermatology on outpatient referral rates. J Tele-med Telecare 2003; 9:321–7.

65 Du Moulin MF, Bullens-Goessens YI, Henquet CJ et al. Thereliability of diagnosis using store-and-forward teledermatology.

J Telemed Telecare 2003; 9:249–52.66 Shapiro M, James WD, Kessler R et al. Comparison of skin biopsy

triage decisions in 49 patients with pigmented lesions and skinneoplasms: store-and-forward teledermatology vs face-to-face

dermatology. Arch Dermatol 2004; 140:525–8.67 Chao LW, Cestari TF, Bakos L et al. Evaluation of an internet-based

teledermatology system. J Telemed Telecare 2003; 9 (Suppl. 1): 9–12.68 Coras B, Glaessl A, Kinateder J et al. Teledermatoscopy in daily

routine – results of the first 100 cases. Curr Probl Dermatol 2003;32:207–12.

69 Hockey AD, Wootton R, Casey T. Trial of low-cost teledermatolo-gy in primary care. J Telemed Telecare 2004; 10 (Suppl. 1): 44–7.

70 Oztas MO, Calikoglu E, Baz K et al. Reliability of web-based tele-dermatology consultations. J Telemed Telecare 2004; 10:25–8.

71 Caumes E, Le Bris V, Couzigou C et al. Dermatoses associated withtravel to Burkina Faso and diagnosed by means of teledermatolo-

gy. Br J Dermatol 2004; 150:312–16.72 Mahendran R, Goodfield MJ, Sheehan-Dare RA. An evaluation of

the role of a store-and-forward teledermatology system in skin can-cer diagnosis and management. Clin Exp Dermatol 2005; 30:209–14.

73 Salmhofer W, Hofmann-Wellenhof R, Gabler G et al. Wound tele-consultation in patients with chronic leg ulcers. Dermatology 2005;

210:211–17.74 Blum A, Hofmann-Wellenhof R, Luedtke H et al. Value of the

clinical history for different users of dermoscopy compared with

results of digital image analysis. J Eur Acad Dermatol Venereol 2004;18:665–9.

75 Klaz I, Wohl Y, Nathansohn N et al. Teledermatology: qualityassessment by user satisfaction and clinical efficiency. Isr Med Assoc

J 2005; 7:487–90.76 See A, Lim AC, Le K et al. Operational teledermatology in Broken

Hill, rural Australia. Australas J Dermatol 2005; 46:144–9.77 Phillips CM, Burke WA, Shechter A et al. Reliability of dermatol-

ogy teleconsultations with the use of teleconferencing technology.J Am Acad Dermatol 1997; 37:398–402.

78 Loane MA, Gore HE, Corbett R et al. Effect of camera performanceon diagnostic accuracy: preliminary results from the Northern Ire-

land arms of the UK Multicentre Teledermatology Trial. J TelemedTelecare 1997; 3:83–8.

79 Reid DS, Weaver LE, Sargeant JM et al. Telemedicine in Nova Sco-tia: report of a pilot study. Telemed J 1998; 4:249–58.

80 Berman B, Elgart GW, Burdick AE. Dermatopathology via a still-image telemedicine system: diagnostic concordance with direct

microscopy. Telemed J 1997; 3:27–32.




81 Phillips CM, Burke WA, Allen MH et al. Reliability of telemedicinein evaluating skin tumors. Telemed J 1998; 4:5–9.

82 Lowitt MH, Kessler II, Kauffman CL et al. Teledermatology andin-person examinations: a comparison of patient and physician

perceptions and diagnostic agreement. Arch Dermatol 1998;134:471–6.

83 Lesher JL Jr, Davis LS, Gourdin FW et al. Telemedicine evaluationof cutaneous diseases: a blinded comparative study. J Am Acad Der-

matol 1998; 38:27–31.84 Nordal EJ, Moseng D, Kvammen B et al. A comparative study of

teleconsultations versus face-to-face consultations. J Telemed Telecare

2001; 7:257–65.85 Lamminen H, Lamminen J, Ruohonen K et al. A cost study of tele-

consultation for primary-care ophthalmology and dermatology.J Telemed Telecare 2001; 7:167–73.

86 Lamminen H, Tuomi ML, Lamminen J et al. A feasibility study ofrealtime teledermatology in Finland. J Telemed Telecare 2000; 6:102–

7.87 Chan HH, Woo J, Chan WM et al. Teledermatology in Hong

Kong: a cost-effective method to provide service to the elderlypatients living in institutions. Int J Dermatol 2000; 39:774–8.

88 Demiris G, Edison K, Vijaykumar S. A comparison of communica-tion models of traditional and video-mediated health care deliv-

ery. Int J Med Inform 2005; 74:851–6.89 Marcin JP, Nesbitt TS, Cole SL et al. Changes in diagnosis, treat-

ment, and clinical improvement among patients receiving tele-medicine consultations. Telemed J E Health 2005; 11:36–43.

90 Granlund H, Thoden CJ, Carlson C et al. Realtime teleconsultationsversus face-to-face consultations in dermatology: immediate and

six-month outcome. J Telemed Telecare 2003; 9:204–9.91 Perednia DA, Wallace J, Morrisey M et al. The effect of a tele-

dermatology program on rural referral patterns to dermatolo-gists and the management of skin disease. Medinfo 1998; 9:

290–3.92 Whited JD, Mills BJ, Hall RP et al. A pilot trial of digital imaging

in skin cancer. J Telemed Telecare 1998; 4:108–12.93 Harrison PV, Kirby B, Dickinson Y et al. Teledermatology – high

technology or not? J Telemed Telecare 1998; 4 (Suppl. 1): 31–2.94 Zelickson BD, Homan L. Teledermatology in the nursing home.

Arch Dermatol 1997; 133:171–4.

95 Lyon CC, Harrison PV. A portable digital imaging system indermatology: diagnostic and educational applications. J Telemed

Telecare 1997; 3 (Suppl. 1): 81–3.96 Kvedar JC, Edwards RA, Menn ER et al. The substitution of digital

images for dermatologic physical examination. Arch Dermatol 1997;133:161–7.

97 Pak HS, Welch M, Poropatich R. Web-based teledermatology con-sult system: preliminary results from the first 100 cases. Stud Health

Technol Inform 1999; 64:179–84.98 Tucker WF, Lewis FM. Digital imaging: a diagnostic screening

tool? Int J Dermatol 2005; 44:479–81.

99 Moreno-Ramirez D, Ferrandiz L, Bernal AP et al. Teledermatologyas a filtering system in pigmented lesion clinics. J Telemed Telecare

2005; 11:298–303.100 Knol A, van den Akker TW, Damstra RJ et al. Teledermatology

reduces the number of patient referrals to a dermatologist. J Tele-med Telecare 2006; 12:75–8.

101 Moreno-Ramirez D, Ferrandiz L, Galdeano R et al. Teledermatosco-py as a triage system for pigmented lesions: a pilot study. Clin Exp

Dermatol 2006; 31:13–8.102 Krupinski EA, Engstrom M, Barker G et al. The challenges of fol-

lowing patients and assessing outcomes in teledermatology. J Tele-med Telecare 2004; 10:21–4.

103 Burgiss S, Julius C, Watson H et al. Telemedicine for dermatologycare in rural patients. Telemed J 1997; 3:227–33.

104 Phillips CM, Burke WA, Bergamo B et al. Review of teleconsult-ations for dermatologic diseases. J Cutan Med Surg 2000; 4:71–5.

105 Bergmo T. A cost-minimization analysis of a realtime teledermatol-ogy service in northern Norway. J Telemed Telecare 2000; 6:273–7.

106 Weinstock MA, Nguyen FQ, Risica PM. Patient and referring pro-vider satisfaction with teledermatology. J Am Acad Dermatol 2002;

47:68–72.107 Krupinski E, Barker G, Rodriguez G et al. Telemedicine versus

in-person dermatology referrals: an analysis of case complexity.Telemed J E Health 2002; 8:143–7.

108 Oakley AM, Rennie MH. Retrospective review of teledermatologyin the Waikato, 1997–2002. Australas J Dermatol 2004; 45:23–8.

109 Massone C, Soyer HP, Hofmann-Wellenhof R et al. Two years’experience with web-based teleconsulting in dermatology. J Tele-

med Telecare 2006; 12:83–7.




CUTANEOUS BIOLOGY DOI 10.1111/j .1365-2133.2006.07656.x

Nine procaspases are expressed in normal humanepidermis, but only caspase-14 is fully processedA.-A. Raymond, M.-C. Mechin, R. Nachat, E. Toulza, R. Tazi-Ahnini,* G. Serre and M. Simon

UMR 5165 CNRS – University of Toulouse III, Institut Federatif de Recherche Claude de Preval, IFR30 (INSERM-CNRS-Universite Paul Sabatier-Centre

Hospitalier Universitaire de Toulouse), Faculte de Medecine, 37 allees Jules Guesde, 31073 Toulouse, France

*Division of Genomic Medicine, University of Sheffield, Medical School, Sheffield, U.K.

CorrespondenceMichel Simon.



Key wordscaspase, differentiation, epidermis, protease,

psoriasis


M.-C.M. and R.N. participated equally in this

work.

Summary

Background During normal stratification of the epidermis, keratinocytes undergo acomplex programme of terminal differentiation. This programmed cell deathresults in corneocyte accumulation to form the stratum corneum (SC). Terminaldifferentiation and apoptosis share numerous features such as elimination of nu-clei and organelles, changes in cell shape, and activation of transglutaminases andproteases. Caspases are cysteine proteases that play a central role in apoptosis.Therefore they may also be involved in the terminal differentiation of keratino-cytes.Objectives To identify the caspases expressed in normal human epidermis and todefine their pattern of expression and activation.Methods We analysed mRNAs from human epidermis by reverse transcription–polymerase chain reaction (RT–PCR), skin cryosections by immunohistologicalmethods, and epidermal protein extracts by Western blotting.Results The mRNAs encoding caspase-1, -2, -3, -4, -6, -7, -8, -9, -10 and -14were detected by RT–PCR. Accordingly, the immunohistological analyses showedclear expression in the epidermis of the corresponding proteins except caspase-2and caspase-8, with only a weak expression of caspase-9. Moreover, procaspase-1, -2, -3, -4, -6, -7, -9, -10 and -14, and the fully processed caspase-14, wereimmunodetected in total epidermis extracts. However, only procaspase-1 and theprocessed caspase-14 were detected in extracts of superficial SC. In addition tothese two proteins, procaspase-4 was detected in extracts of superficial SCobtained from lesional psoriatic skin.Conclusions This study, the first exhaustive description of caspase expression andprocessing in normal human epidermis, indicates that in vivo granular keratino-cytes express nine procaspases, and in addition the activated form of caspase-14.This confirms that only caspase-14 is involved in keratinocyte differentiation, andsuggests that keratinocytes are ready to induce apoptosis in response to cutaneousdamage.

Apoptosis involves a signalling pathway in which cysteinyl

aspartate-specific proteinases, called caspases, play a central

role.1 These cytoplasmic enzymes are synthesized as zymogens

(procaspases), comprising a prodomain, as well as a large

(p20) and a small (p10) subunit. The proforms are activated

by proteolytic cleavage between their subdomains (by other

proteases or by themselves thanks to their autocatalytic activ-

ity), which leads to the formation of a mature caspase, a tetra-

mer consisting of two p20/p10 heterodimers. Most of the 15

caspases reported until now can be divided into initiator (-1,

-2, -4, -5, -8, -9, -10, -11, -12 and -13) and effector caspases

(-3, -6, -7 and -14). The former display a long prodomain

whereas the prodomain of the latter is made up of fewer than

40 amino acids. Caspase-11 (murine), -13 (bovine) and -15

(described in various other mammalian species) do not exist in

humans.2–4 Upon binding to signal-transducing molecules, ini-

tiator caspases activate downstream effector caspases which

finally cleave various cellular substrates (structural proteins

such as actin, laminin or components of cellular junctions, and

regulatory proteins such as a DNAse inhibitor),5,6 thereby



inducing the demise of the cell. Caspase-14 is peculiar as it has

a restricted pattern of expression, as opposed to all other cas-

pases identified, and is not activated during apoptosis.7

The skin is one of the first lines of defence against the envi-

ronment. The epidermal protective barrier is established by

keratinocytes that undergo a complex programme of terminal

differentiation, from the stratum basale towards the surface of

the skin. During the differentiation programme, keratinocytes

stop bcl-2 expression, show transglutaminase and protease

activation, rearrange their cytoskeleton, change their shape

and lose their nucleus and organelles. This leads to epidermal

terminal differentiation being considered as a ‘planned cell

death pathway’.8 However, epidermal differentiation differs

from classical apoptosis in that it does not result in the frag-

mentation of cells into apoptotic bodies and their subsequent

phagocytosis, but induces the formation of dead cornified cells

called corneocytes. Stacking of the corneocytes in the upper

epidermis forms a resistant barrier, the stratum corneum (SC).

We nevertheless assumed that epidermal keratinocytes use

some apoptosis elements during their differentiation, in par-

ticular one or several caspases. This seems to be the case for

the terminal differentiation of rodent lens epithelial cells,9

erythrocytes10 and T cells,11,12 or during postnatal develop-

ment of rat cerebellum.13

There are few reports on any relationships that might exist

between caspases and keratinocyte terminal differentiation.

When keratinocytes, cultured in conditions that induce differ-

entiation, are treated with the broad-spectrum caspase inhibi-

tor Z-VAD-FMK, SC development has been shown to be

perturbed.14 Some studies have been carried out on caspase

expression in the epidermis, but the results are as yet incom-

plete and sometimes in contradiction. Caspase-1, -2, -3, -4

and -7 mRNAs have been detected by reverse transcription–

polymerase chain reaction (RT–PCR) in cultured human kera-

tinocytes but those of caspase-5 and -6 have not.15 In mouse

epidermis, caspase-2, -3, -6, -8, -9 and -14 mRNAs have been

detected but those of caspase-1, -7, -11 and -12 have not.16

Caspase-1 and -3 have been partially purified from human

corneocyte extracts.15 It has been suggested that procaspase-3

is activated during late keratinocyte differentiation.14 How-

ever, processed forms of caspase-3, -6 and -7 have not been

detected in human epidermis.7 More data are available about

caspase-14 which is only expressed in the skin,17 choroid

plexus, retinal pigment epithelium and thymic Hassall’s bod-

ies.18 Its expression, conserved among mammals,19 was repor-

ted in adult and embryonic human epidermis to be restricted

to the differentiated keratinocytes, where the processed forms

(p17 and p11) were detected.7 In particular, active caspase-14

was observed in SC extracts.20,21 Immunogold cytochemistry

localized caspase-14 to the amorphous material of keratohy-

aline granules and to the keratin bundles. It is associated with

nuclear remnants, corneodesmosomes and intracellular matrix

of corneocytes.19,22 Caspase-14 is also expressed in hair folli-

cles and sebaceous glands of human epidermis.20

Our objective was to analyse by RT–PCR the caspases

expressed in normal human epidermis, to define their pattern

of expression using immunohistochemistry and indirect

immunofluorescence, and to determine by Western blotting

which caspases are processed.


Human tissues

All described studies concerning human tissues/subjects were

conducted in accordance with the Declaration of Helsinki Prin-

ciples, and with the ethical guidelines of the University of

Toulouse III, the University of Sheffield and the French Minis-

try of Research and Technology. Each patient provided

informed consent to testing. Plastic surgery specimens of nor-

mal human breast or abdominal skin, obtained from patients

undergoing plastic surgery without any history of skin dis-

eases, were kindly provided by Professor J.-P. Chavoin (Plastic

Surgery Department, Rangueil Teaching Hospital, Toulouse,

France). Samples of SC were obtained from volunteers with

normal skin and without any history of genetic skin disorders

or atopy, and from lesional and nonlesional skin of three

patients with psoriasis recruited in Sheffield.

Analysis of mRNA expression by reverse

transcription–polymerase chain reaction

Human epidermis cDNA, prepared as previously described23

and tested for the absence of genomic contamination, was

used as a template for PCR analysis with specific caspase pri-

mer pairs, chosen in reference to previously published

works.24,25 The specificity of the primers was confirmed using

the BLAST program.26 The sequence of primers and their posi-

tion on caspase cDNA are shown in Table 1. Positive control

reactions with Jurkat cell cDNA (Biochain Institute Inc.,

Hayward, CA, U.S.A.) were carried out in parallel. Glycer-

aldehyde-3-phosphate dehydrogenase (GAPDH) primers were

used to confirm normalization of the cDNAs. PCR reactions

were performed in 25-lL volumes using 0.75 U of Goldstar

Taq DNA polymerase (Eurogentec, Seraing, Belgium) with the

supplied buffer, 1.5 mmol L)1 MgCl2, 5 mmol L)1 of each

deoxyribonucleoside triphosphate, 5 pmol of each primer and

1 lL of each cDNA sample. The PCR conditions were as fol-

lows: 94 �C for 2 min, followed by 35 or 45 cycles at 94 �Cfor 1 min, between 52 �C and 62 �C (52 �C for caspase-3, -5

and -6, 56 �C for caspase-4, -7, -10, -12, -14 and GAPDH,

58 �C for caspase-8, 60 �C for caspase-9, 62 �C for caspase-1

and -2) for 1 min, and 72 �C for 1 min, and a final elonga-

tion at 72 �C for 1 min.

Extraction of proteins from human epidermis and

superficial stratum corneum

Dermoepidermal cleavage was performed by heat treatment,

and the epidermis extracted in a Tris–ethylenediamine tetra-

acetic acid (EDTA) buffer containing Nonidet P-40 as previ-

ously reported.27



Expression of caspases in human epidermis, A.-A. Raymond et al. 421

Superficial SC samples were obtained by a tape-stripping

procedure.28 Proteins were dissolved in a Tris–EDTA buffer

supplemented with 2% sodium dodecyl sulphate (SDS) and

50 mmol L)1 dithiothreitol as previously reported.29

Protein concentrations were measured using the Bio-

Rad protein assayTM (Bio-Rad Laboratories GmbH, Munich,

Germany).

Antibodies

The characteristics of the antibodies used in this study are

summarized in Table 2.

Protein electrophoresis and immunoblotting

Proteins were separated by SDS–polyacrylamide gel electro-

phoresis (PAGE) on 12Æ5% acrylamide gels and electrotrans-

ferred to nitrocellulose membranes. The membranes were

stained with Ponceau Red, then heated at 70 �C for 45 min in

order to cross-link the small proteins, incubated for 30 min in

blocking buffer [Tris-buffered saline (40 mmol L)1 Tris-HCl

pH 8, 0Æ9% NaCl) supplemented with 2Æ5% low-fat dried milk

or fish gelatin, and 0Æ05% Tween-20], and probed with anti-

bodies as previously described.27 Immunoreactivities were

revealed with the ECLTM Western blotting kit as described by

the manufacturer (Amersham Biosciences, Little Chalfont,

U.K.).

Immunohistology

Immunohistochemistry was performed on unfixed 5–7-lm

thick cryosections of breast or abdominal skin samples

obtained from several donors. The primary antibodies diluted

in phosphate-buffered saline (PBS) were applied for 1 h at

37 �C or overnight at 4 �C (Table 2). For detection of cas-

pase-2, -3 and -14 we used the Histostain-Plus kit, as

described by Zymed Laboratories Invitrogen (Cergy Pontoise,

France), the manufacturer. Anticaspase-4, -6, -7 and -10 anti-

bodies were revealed with a horseradish peroxidase-conju-

gated rabbit antigoat IgG secondary antibody (Tebu-bio, Le

Perray en Yvelines, France) and 3-amino-9-ethylcarbazole as

the substrate.

For indirect immunofluorescence, cryosections were

blocked with PBS containing 2% bovine serum albumin

(BSA) and 0Æ05% Tween-20 and then incubated with pri-

mary antibodies diluted in PBS containing 1% BSA and

Table 1 Primer sequences

Caspase-1 Upstream 5¢-AAC CCA GCT ATG CCC ACA TCC-3¢ (96–116)Downstream 5¢-TTA ATG TCC TGG GAA GAG GTA-3¢ (953–933)

Caspase-2 Upstream 5¢-GTT ACC TGC ACA CCG AGT CAC G-3¢ (557–578)Downstream 5¢-GCG TGG TTC TTT CCA TCT TGT TGG TCA-3¢ (790–764)

Caspase-3 Upstream 5¢-GAA TAT CCC TGG ACA ACA-3¢ (157–174)Downstream 5¢-ACG CCA TGT CAT CAT CAA-3¢ (618–601)

Caspase-4 Upstream 5¢-GGT CAT CAT TGT CCA GGC-3¢ (792–809)

Downstream 5¢-CCA TTG TGC TGT CTC TCC-3¢ (994–977)Caspase-5 Upstream 5¢-ACC ACA TGC TAA AGA ACA-3¢ (81–98)

Downstream 5¢-CGA TTT GCA GAA GAG GTT-3¢ (434–417)Caspase-6 Upstream 5¢-ACC CGC AGG TTT TCA GA-3¢ (249–265)

Downstream 5¢-CAT GAG CCG TTC ACA GT-3¢ (715–699)Caspase-7 Upstream 5¢-AGC CTG GGT TTT GAC GTG-3¢ (385–402)

Downstream 5¢-ACC GTG GAA TAG GCG AAG-3¢ (755–738)Caspase-8 Upstream 5¢-TAT GGC ACT GAT GGA CAG GA-3¢ (1323–1342)

Downstream 5¢-AGA TCA CGC CAC TGC ACT C-3¢ (1882–1864)Caspase-9 Upstream 5¢-AAG ACC ATG GGT TTG AGG TG-3¢ (970–989)

Downstream 5¢-TGT CGT CAA TCT GGA AGC TG-3¢ (1519–1500)Caspase-10 Upstream 5¢-AGA AGT CCA GCT CAG CCT-3¢ (352–369)

Downstream 5¢-ACT CGG CTT CCT TGT CTA-3¢ (813–796)Caspase-12 Upstream 5¢-GGA CCA AGC ACT GGG ATC A-3¢ (795–813)

Downstream 5¢-TAC CTG AGT TGC TTC TTA TGA GT-3¢ (1168–1146)Caspase-14 Upstream 5¢-ATA TGA TAT GTC AGG TGC CCG-3¢ (83–102)

Downstream 5¢-CCG AAA CAT GTG TTC CAG AG-3¢ (170–151)GAPDH Upstream 5¢-TAG CCC AGG CAC TTG GAG CC-3¢ (299–319)

Downstream 5¢-CGG TTT AAG TAA CAG TAT GG-3¢ (1121–1101)

Primer location on cDNA sequences (in bp) is written in brackets. GenBank references ofcDNAs used: NM033293 (caspase-1), BC002427 (caspase-2), NM032991 (caspase-3),

BC017839 (caspase-4), NM004347 (caspase-5), BC004460 (caspase-6), BC015799(caspase-7), NM033355 (caspase-8), NM001229 (caspase-9), NM032977 (caspase-10),

AF464193 (caspase-12) and NM012004 (caspase-14). GAPDH, glyceraldehyde-3-phos-phate dehydrogenase.



422 Expression of caspases in human epidermis, A.-A. Raymond et al.

0Æ05% Tween-20 for 1 h at 37 �C or overnight at 4 �C(Table 2). The sections were washed twice in PBS contain-

ing 0Æ05% Tween-20 and twice in PBS and then incubated

for 1 h at room temperature with the secondary antibody

[tetramethylrhodamine isothiocyanate-conjugated goat anti-

mouse IgG (Zymed), fluorescein isothiocyanate (FITC)-con-

jugated goat antirabbit IgG (Zymed) or FITC-conjugated

rabbit antigoat IgG (Tebu-bio)] diluted in PBS containing

1% BSA and 0Æ05% Tween-20. After two washes in PBS

containing 0Æ05% Tween-20 and two washes in PBS, the

sections were finally mounted in an antifading solution

(Fluoroprep; BioMerieux, Marcy-l’Etoile, France). For nega-

tive controls, sections were incubated with nonimmune

sera, with irrelevant rabbit antipeptide antibodies or without

primary antibodies.

Results

At least nine caspase genes are expressed in normal

human epidermis

To test for the expression of caspase genes, RT–PCR analysis

was performed on cDNA from human epidermis with primer

pairs specific for each human caspase. In a first 35-cycle

amplification, the mRNAs encoding caspase-1 (858 bp), -2

(234 bp), -3 (462 bp), -4 (203 bp), -6 (467 bp), -7

(371 bp), -9 (550 bp), -10 (462 bp) and -14 (88 bp) were

detected. However, only weak signals were observed for cas-

pase-1, -2, -6, -7 and -10. Therefore a 45-cycle amplification

was performed. The results obtained (Fig. 1) confirmed the

previous data; in particular the mRNAs encoding caspase-5

and -12 were not amplified. Moreover, an amplicon of

560 bp corresponding to caspase-8 was barely detected in

only two of four cDNA samples. Parallel negative and positive

PCR controls were performed without any cDNA and with

cDNA from Jurkat cells, respectively.

Caspase-14 is the only fully processed caspase in normal

human epidermis

To explore the expression of caspases at the protein level,

extracts of epidermis obtained from either abdominal or breast

skin were separated by SDS–PAGE and immunodetected using

anticaspase antibodies (Table 2). The same results were

Table 2 Anticaspase antibodies

Caspase Type Specificity Dilution for WB Dilution for IHE Dilution for IIF

1 SA101, rabbit polyclonal p20 and precursor 1 : 1000 Not carried out 1 : 320

2 F-7, mouse monoclonal p10 and precursor 1 : 500 1 : 10–1 : 160 1 : 10–1 : 1603 H-277, rabbit polyclonal p20 and precursor 1 : 1000 1 : 160 1 : 160

4 N-15, goat polyclonal p20 and precursor 1 : 500 1 : 40 1 : 206 A-16, goat polyclonal p10 and precursor 1 : 500 1 : 20 1 : 10

7 C7724, rabbit polyclonal p20 and precursor 1 : 500 1 : 40 1 : 808 C-20, goat polyclonal p20 and precursor 1 : 500 Not carried out 1 : 10

9 F-7, mouse monoclonal p10 and precursor 1 : 250 Not carried out 1 : 1010 C8351, rabbit polyclonal p20 and precursor 1 : 1000 Not carried out 1 : 40

14 H-99, rabbit polyclonal p20 and precursor 1 : 5000 1 : 80 1 : 80

WB, Western blot; IHE, immunohistoenzymology; IIF, indirect immunofluorescence. All anticaspase antibodies were purchased from SantaCruz Biotechnology (Santa Cruz, CA, U.S.A.) (200 lg mL)1) except anticaspase-1 which came from Biomol International (Plymouth Meet-

ing, PA, U.S.A.) (100 lg mL)1), and anticaspase-10 and anticaspase-7 which came from Sigma-Aldrich (L’Isle d’Abeau, France). The letterand number correspond to the catalogue references.

Fig 1. The expression of 10 caspase genes was detected by reverse

transcription–polymerase chain reaction (PCR) in normal human

epidermis. PCR reactions were performed with human epidermis

cDNA (lane 1), without any cDNAs (lane 2) and with Jurkat cell

cDNA (lane 3), using primers specific to each caspase as indicated.

The PCR results obtained using primers specific to caspase-8 gene and

cDNA from two different individuals are shown. The size of PCR

products, analysed on agarose gels, is indicated on the left.




obtained with both types of extracts. We detected the pro-

forms of the caspases corresponding to the cDNAs detected in

RT–PCR [procaspase-1 (45 kDa), -2 (48 kDa), -3 (33 kDa),

-4 (43 kDa), -6 (34 kDa), -7 (32 kDa), -9 (46 kDa), -10

(59 kDa) and -14 (28 kDa)] except caspase-8. Under the con-

ditions used, only weak expression of procaspase-9 was im-

munodetected, in comparison with the others. However,

strong expression of procaspase-8 and -9 was detected in sim-

ilar extracts of cultured human keratinocytes (data not

shown). Caspase-1, -2, -3, -6, -7, -9 and -10 were present

only as unprocessed proenzymes. The anticaspase-4 antibody

recognized both the proenzyme and a 35-kDa protein prob-

ably corresponding to a processing intermediate as proposed

by Kamada et al.30 or resulting from alternative splicing,

whereas the fully processed 20-kDa form was not detected. In

the case of caspase-14, we detected both the proenzyme and

the p17 fragment, indicating full processing of this caspase

(Fig. 2).

We then analysed by immunoblotting the presence of

caspases in SC extracts obtained by tape-stripping from five

healthy individuals. Only procaspase-1 and the p17 fragment

of caspase-14 were detected in all SC extracts (Fig. 3).

Caspases are detected in all layers of the normal

epidermis

To identify the sites of caspase expression in normal human

epidermis, breast and abdominal skin cryosections were ana-

lysed by immunohistoenzymology and/or indirect immuno-

fluorescence using the same antibodies. Identical results were

obtained with both techniques, and on sections from both

anatomical sites. Control experiments were always negative.

No labelling was observed with the anticaspase-2 and anticas-

pase-8 antibodies. The anticaspase-1, -4, -6 and -10 antibodies

displayed the same pattern of staining. They labelled the entire

epidermis except the SC with an increased staining intensity at

the stratum granulosum level. Moreover, sweat glands were

labelled by anticaspase-1, -4 and -10 antibodies and muscular

fibres, only observed on sections of breast skin, were stained

by the anticaspase-1 antibody (Fig. 4 and data not shown).

The anticaspase-3 antibody labelled the entire epidermis and

Fig 2. The only fully processed caspase in normal human epidermis is

caspase-14. Twenty micrograms of proteins extracted from normal

human epidermis of breast skin were immunoblotted with anticaspase

antibodies (lane 1) or without primary antibodies (lane 2). Arrows

and arrowheads point to proteins specifically recognized by each

antibody. The position of molecular mass standards (kDa) is indicated

on the right. Note that the antigoat immunoglobulin secondary

antibody reacts nonspecifically with the heavy chains of human

immunoglobulins (50-kDa bands; panels Casp-4, -6 and -8), and that

the antimouse immunoglobulin secondary antibody reacts

nonspecifically with the light chains of human immunoglobulins

present in the extracts as monomers or heterodimers (25- and 75-kDa

bands; panel Casp-2).

Fig 3. Caspase-1 and -14 are immunodetected in extracts of normal

superficial stratum corneum. Equal amounts of proteins (20 lg) of

samples obtained from five different normal individuals using a tape-

stripping procedure were used for immunoblotting with anticaspase-1

and anticaspase-14 antibodies, as indicated. The p17 fragment of

caspase-14 is indicated on the right, and the position of molecular

mass standards (kDa) on the left.

Fig 4. Immunofluorescence staining of caspases in human epidermis.

Normal human skin cryosections were analysed by indirect

immunofluorescence using the anticaspase antibodies, as indicated,

and an irrelevant affinity purified antipeptide rabbit antibody (negative

control). Note that sweat gland myoepithelial cells [caspase-1 (b)]

and arrector pili muscle cells [caspase-1 (c)] are labelled by the

anticaspase-1 antibody. The dotted lines indicate the dermoepidermal

junction. Scale bars ¼ 75 lm.




muscular fibres (data not shown). The anticaspase-7 antibody

stained with the same intensity the entire epidermis except the

SC (Fig. 4). Only a labelling of some, usually granular, kera-

tinocytes was observed with the anticaspase-9 antibody (data

not shown). When using the anticaspase-14 antibody, we

observed labelling of the entire epidermis, including the SC as

described previously.7,31

Caspase-14 is also the only fully processed caspase

detected in psoriatic stratum corneum

To evaluate a potential role for caspases in psoriasis, we ana-

lysed their presence in SC samples obtained from lesional and

nonlesional psoriatic skin (three patients) using the same anti-

bodies. As observed in normal skin, procaspase-1 (data not

shown) and the p17 processed fragment of caspase-14

(Fig. 5) were detected. Additionally, the proform (45 kDa)

and smaller forms of caspase-4 (including the previously

observed 35-kDa band) were observed in the extracts of le-

sional psoriatic SC. However, the fully processed 20-kDa form

of caspase-4 was not detected (Fig. 5).

Discussion

We analysed the expression of caspases in human skin using

different techniques including PCR on cDNA from human epi-

dermis, Western blotting of epidermis extracts and immuno-

histological analysis of skin cryosections.

We detected the mRNAs encoding caspase-1, -2, -3, -4, -6,

-7, -9, -10 and -14. In agreement, all the corresponding pro-

forms of the proteases were immunodetected. In some cases

(two of four different cDNAs), weak expression of caspase-8

mRNA was detected but we were unable to immunodetect the

corresponding protein. In agreement with these results, cul-

tured human keratinocytes have been shown to express the

same mRNAs except those encoding caspase-6.15,17,32–35 The

expression in mouse keratinocytes and epidermis of caspase-2,

-3, -6, -8 and -14 mRNA but not -7 and -12 was also

observed.16 Our results confirm and extend a previous study

showing the presence of procaspase-3, -6, -7 and -14 in

mouse epidermis.7

Most of the caspases were localized in all the epidermal

layers, with a more intense immunostaining of the granular

layer for caspase-1, -4, -6 and -10. Caspase-9 was only detec-

ted in some, usually granular, keratinocytes. This explains the

low detection of this caspase using Western blotting. The

anticaspase-8 antibody did not detect any proteins on skin

sections, confirming the Western blotting results. This strongly

suggests that caspase-8 expression is controlled at the post-

transcriptional level. Concerning caspase-2, although the

proform was detected in epidermal extracts, we did not

successfully obtain staining of human skin using immuno-

histological techniques with two different antibodies, one

from Santa Cruz Biotechnology (Santa Cruz, CA, U.S.A.;

Table 2) and one from Sigma-Aldrich (L’Isle d’Abeau, France;

C7349), both specific for the precursor and p10 fragment.

This could be due to a low sensitivity of the method or more

probably to inaccessibility of the antigen.

Surprisingly, the only activated caspase seems to be caspase-

14, as detected by Western blotting of epidermis extracts. This

indicates that the amount of fully processed caspases is too

low to be immunodetected or more probably that most cas-

pases are not activated at all. This result raises the following

question: why do keratinocytes express unprocessed procas-

pases? We could speculate about an anticipation of a cutane-

ous stress harmful for the survival of keratinocytes or for the

Fig 5. Caspase-14 and procaspase-4 are detected in extracts of

psoriatic superficial stratum corneum (SC). Equal amounts of proteins

(20 lg) of samples obtained from the SC of three normal individuals

and from both lesional (L) and nonlesional (NL) SC of three patients

with psoriasis were separated by sodium dodecyl sulphate–

polyacrylamide gel electrophoresis, and transferred to membranes.

The membranes were stained with Ponceau Red and used for

immunoblotting using antibodies to caspase-14 and caspase-4, and an

isotype control antibody (anticaspase-6), as indicated. The position of

molecular mass standards (kDa) is indicated on the left. The p17

fragment of caspase-14, the proform (Casp-4) and the p35 fragment

of caspase-4 (arrowhead) are indicated on the right.




integrity of the skin barrier, e.g. ultraviolet (UV) irradiation.

Indeed, caspase-1, -3, -8 and -9 have been shown to be acti-

vated in human keratinocytes during UV-induced apopto-

sis,32–34,36,37 this induction being considered as a protective

function against skin cancer. In agreement, the intense immuno-

staining of the granular layer observed for most caspases

might be related to the fact that cutaneous stresses first affect

the most superficial layers of the epidermis. All classes of cas-

pases seem to be ready to be activated: inflammatory caspase-1

and -4, initiator caspase-2, -9 and -10, and effector caspase-3,

-6 and -7.

Caspase-14 was suggested to play a role in keratinocyte ter-

minal differentiation. In the same way, different caspases were

shown to be involved in cell differentiation, e.g. caspase-3 in

lens epithelium differentiation,9,38 and caspase-1 in neuronal

differentiation of PC12 cells.39 Here, we show that fully proc-

essed caspase-14 is detected at the SC surface. This suggests

that the enzyme might be important for the late steps of

terminal differentiation, e.g. during desquamation. Indeed,

immunoelectron microscopy showed that it was localized near

corneodesmosomes.22 For desquamation to occur, corneodes-

mosome components must be proteolysed. We previously

demonstrated the involvement of proteases of the kallikrein

family in the degradation of the extracellular components.40

We also suggested that intracellular components (plakoglobin

and the intracellular part of desmoglein 1) must also be degra-

ded, but the involved proteases have not been identified. In

agreement with this hypothesis, intracellular and transmem-

branous components of desmosomes are cleaved by caspases

after induction of apoptosis in the keratinocyte cell lines

HaCaT and A431.6,41 To test for a role of caspase-14, we

activated human recombinant procaspase-14 by granzyme B

in vitro and incubated an extract of epidermis containing plako-

globin and desmogleins with the active caspase-14. However,

we did not observe cleavage of these proteins (data not

shown).

Finally, we looked for a potential disruption of caspase

expression in psoriatic epidermis. Western blot analysis of SC

samples revealed, as in normal epidermis, the presence of

procaspase-1 and caspase-14. It is surprising that the level of

caspase-14 in the superficial SC of psoriatic skin (both lesional

and nonlesional skin) seems to be not so different from its

level in the SC of normal skin21 (this work). Indeed, a reduc-

tion in the expression of the caspase is generally observed

using immunohistochemistry in the parakeratotic SC of psori-

atic skin.7,42,43 This could be explained in part by the well-

known masking of antigens/epitopes in the hyperkeratotic SC.

This result also suggests that despite a reduction in the level of

procaspase-14 in the noncornified keratinocytes of psoriatic

epidermis, a peculiar mechanism of regulation ensures the

same level of processing and accumulation of the active

caspase-14 in both normal and diseased tissue. In addition to

these two caspases, the expression of procaspase-4 was also

noticed in the SC of lesional psoriatic skin. This could be linked

to the aberrant expression of other apoptosis-related molecules

in the psoriatic skin such as ICAD (inhibitor of caspase-3-

related DNase), Bcl-2 or Bax.44 Alternatively, this result sug-

gests a possible involvement of caspase-4 in psoriasis physio-

pathology. In particular, the caspase-4 gene is one of the

numerous interferon-induced genes,45–47 interferon-c being an

important cytokine in psoriasis. Moreover, the caspase-4 gene

was recently shown to be upregulated in psoriasis.47

In conclusion, keratinocytes express in the epidermis the

proforms of the initiator caspase-1, -2, -4, -9 and -10, and of

the effector caspase-3, -6 and -7. Therefore, they are probably

ready to respond to major stresses and to induce apoptosis.

Moreover, they express an active caspase-14, which is prob-

ably necessary for the last steps of terminal differentiation but

the exact function of which remains to be discovered.

Acknowledgments

We thank Professor A.-P. Arrigo for the gift of anticaspase-3

antibodies, Professor J.-P. Chavoin for providing us with

human skin and Dr M. Guerrin for reading the manuscript.

This study was supported by grants from the Centre National

de la Recherche Scientifique, the European Regional Develop-

ment Funds, the European Social Funds, the Institut National

de la Sante et de la Recherche Medicale, and from the Societe

de Recherche Dermatologique.

References

1 Oliver L, Vallette FM. The role of caspases in cell death and differ-entiation. Drug Resist Updat 2005; 8:163–70.

2 van de Craen M, Vandenabeele P, Declercq W et al. Characteriza-tion of seven murine caspase family members. FEBS Lett 1997;

403:61–9.3 Koenig U, Eckhart L, Tschachler E. Evidence that caspase-13 is not

a human but a bovine gene. Biochem Biophys Res Commun 2001;285:1150–4.

4 Eckhart L, Ballaun C, Uthman A et al. Identification and characteri-zation of a novel mammalian caspase with proapoptotic activity.

J Biol Chem 2005; 280:35077–80.5 Shi Y. Mechanisms of caspase activation and inhibition during

apoptosis. Mol Cell 2002; 9:459–70.6 Weiske J, Schoneberg T, Schroder W et al. The fate of desmosomal

proteins in apoptotic cells. J Biol Chem 2001; 276:41175–81.7 Lippens S, Kockx M, Knaapen M et al. Epidermal differentiation

does not involve the proapoptotic executioner caspases, but is asso-ciated with caspase-14 induction and processing. Cell Death Differ

2000; 7:1218–24.8 Gandarillas A. Epidermal differentiation, apoptosis, and senescence:

common pathways? Exp Gerontol 2000; 35:53–62.9 Ishizaki Y, Jacobson MD, Raff MC. A role for caspases in lens fiber

differentiation. J Cell Biol 1998; 140:153–8.10 De Maria R, Zeuner A, Eramo A et al. Negative regulation of ery-

thropoiesis by caspase-mediated cleavage of GATA-1. Nature 1999;401:489–93.

11 Alam A, Cohen LY, Aouad S et al. Early activation of caspases dur-ing T lymphocyte stimulation results in selective substrate cleavage

in nonapoptotic cells. J Exp Med 1999; 190:1879–90.12 Kennedy NJ, Kataoka T, Tschopp J et al. Caspase activation is

required for T cell proliferation. J Exp Med 1999; 190:1891–6.13 Oomman S, Finckbone V, Dertien J et al. Active caspase-3

expression during postnatal development of rat cerebellum is not




systematically or consistently associated with apoptosis. J CompNeurol 2004; 476:154–73.

14 Weil M, Raff MC, Braga VM. Caspase activation in the terminaldifferentiation of human epidermal keratinocytes. Curr Biol 1999;

9:361–4.15 Takahashi T, Ogo M, Hibino T. Partial purification and characteri-

zation of two distinct types of caspases from human epidermis.J Invest Dermatol 1998; 111:367–72.

16 Kuechle MK, Predd HM, Fleckman P et al. Caspase-14, a kera-tinocyte specific caspase: mRNA splice variants and expression

pattern in embryonic and adult mouse. Cell Death Differ 2001;

8:868–70.17 van de Craen M, van Loo G, Pype S et al. Identification of a

new caspase homologue: caspase-14. Cell Death Differ 1998;5:838–46.

18 Lippens S, VandenBroecke C, van Damme E et al. Caspase-14 isexpressed in the epidermis, the choroid plexus, the retinal pigment

epithelium and thymic Hassall’s bodies. Cell Death Differ 2003;10:257–9.

19 Alibardi L, Tschachler E, Eckhart L. Distribution of caspase-14 inepidermis and hair follicles is evolutionarily conserved among

mammals. Anat Rec A Discov Mol Cell Evol Biol 2005; 286:962–73.20 Eckhart L, Declercq W, Ban J et al. Terminal differentiation of

human keratinocytes and stratum corneum formation is associatedwith caspase-14 activation. J Invest Dermatol 2000; 115:1148–51.

21 Fischer H, Stichenwirth M, Dockal M et al. Stratum corneum-derivedcaspase-14 is catalytically active. FEBS Lett 2004; 577:446–50.

22 Alibardi L, Dockal M, Reinisch C et al. Ultrastructural localization ofcaspase-14 in human epidermis. J Histochem Cytochem 2004;

52:1561–74.23 Guerrin M, Ishigami A, Mechin MC et al. cDNA cloning, gene

organization and expression analysis of human peptidylargininedeiminase type I. Biochem J 2003; 370:167–74.

24 Ito A, Uehara T, Tokumitsu A et al. Possible involvement of cyto-chrome c release and sequential activation of caspases in ceramide-

induced apoptosis in SK-N-MC cells. Biochim Biophys Acta 1999;1452:263–74.

25 Fischer H, Koenig U, Eckhart L et al. Human caspase 12 hasacquired deleterious mutations. Biochem Biophys Res Commun 2002;

293:722–6.26 Altschul SF, Madden TL, Schaffer AA et al. Gapped BLAST and PSI-

BLAST: a new generation of protein database search programs.Nucleic Acids Res 1997; 25:3389–402.

27 Simon M, Montezin M, Guerrin M et al. Characterization and purifi-

cation of human corneodesmosin, an epidermal basic glycoproteinassociated with corneocyte-specific modified desmosomes. J Biol

Chem 1997; 272:31770–6.28 Lundstrom A, Serre G, Haftek M et al. Evidence for a role of cor-

neodesmosin, a protein which may serve to modify desmosomesduring cornification, in stratum corneum cell cohesion and desqua-

mation. Arch Dermatol Res 1994; 286:369–75.29 Guerrin M, Simon M, Montezin M et al. Expression cloning of

human corneodesmosin proves its identity with the product of theS gene and allows improved characterization of its processing dur-

ing keratinocyte differentiation. J Biol Chem 1998; 273:22640–7.

30 Kamada S, Washida M, Hasegawa J et al. Involvement of caspase-4(-like) protease in Fas-mediated apoptotic pathway. Oncogene 1997;

15:285–90.31 Eckhart L, Ban J, Fischer H et al. Caspase-14: analysis of gene struc-

ture and mRNA expression during keratinocyte differentiation.Biochem Biophys Res Commun 2000; 277:655–9.

32 Takahashi H, Nakamura S, Asano K et al. Fas antigen modulatesultraviolet B-induced apoptosis of SVHK cells: sequential activation

of caspases 8, 3, and 1 in the apoptotic process. Exp Cell Res 1999;249:291–8.

33 Sitailo LA, Tibudan SS, Denning MF. Activation of caspase-9 is

required for UV-induced apoptosis of human keratinocytes. J BiolChem 2002; 277:19346–52.

34 Assefa Z, Garmyn M, Vantieghem A et al. Ultraviolet B radiation-induced apoptosis in human keratinocytes: cytosolic activation of

procaspase-8 and the role of Bcl-2. FEBS Lett 2003; 540:125–32.35 Chaturvedi V, Sitailo LA, Bodner B et al. Defining the caspase-con-

taining apoptotic machinery contributing to cornification in humanepidermal equivalents. Exp Dermatol 2006; 15:14–22.

36 Hurwitz SA, Spandau DF. Quantitative analysis of UVB-inducedapoptosis in human epidermis. Exp Dermatol 2000; 9:185–91.

37 Takasawa R, Nakamura H, Mori T et al. Differential apoptotic path-ways in human keratinocyte HaCaT cells exposed to UVB and

UVC. Apoptosis 2005; 10:1121–30.38 Zandy AJ, Lakhani S, Zheng T et al. Role of the executioner caspases

during lens development. J Biol Chem 2005; 280:30263–72.39 Vaisid T, Kosower NS, Barnoy S. Caspase-1 activity is required for

neuronal differentiation of PC12 cells: cross-talk between the cas-pase and calpain systems. Biochim Biophys Acta 2005; 1743:223–30.

40 Caubet C, Jonca N, Brattsand M et al. Degradation of corneodesmo-some proteins by two serine proteases of the kallikrein family,

SCTE/KLK5/hK5 and SCCE/KLK7/hK7. J Invest Dermatol 2004;122:1235–44.

41 Dusek RL, Getsios S, Chen F et al. The differentiation-dependent des-mosomal cadherin desmoglein 1 is a novel caspase-3 target that

regulates apoptosis in keratinocytes. J Biol Chem 2006; 281:3614–24.42 Pistritto G, Jost M, Srinivasula SM et al. Expression and transcrip-

tional regulation of caspase-14 in simple and complex epithelia.Cell Death Differ 2002; 9:995–1006.

43 Lippens S, Kockx M, Denecker G et al. Vitamin D3 induces caspase-14 expression in psoriatic lesions and enhances caspase-14 process-

ing in organotypic skin cultures. Am J Pathol 2004; 165:833–41.44 Takahashi H, Manabe A, Ishida-Yamamoto A et al. Aberrant expres-

sion of apoptosis-related molecules in psoriatic epidermis. J Dermatol

Sci 2002; 28:187–97.45 Chawla-Sarkar M, Lindner DJ, Liu Y-F et al. Apoptosis and interfer-

ons: role of interferon-stimulated genes as mediators of apoptosis.Apoptosis 2003; 8:237–49.

46 Ahn EY, Pan G, Vickers SM, McDonald JM. IFN-gamma upregulatesapoptosis-related molecules and enhances Fas-mediated apoptosis

in human cholangiocarcinoma. Int J Cancer 2002; 100:441–51.47 Kulski JK, Kenworthy W, Bellgard M et al. Gene expression profil-

ing of Japanese psoriatic skin reveals an increased activity in mo-lecular stress and immune response signals. J Mol Med 2005;

83:964–75.





Testosterone metabolism to 5a-dihydrotestosterone andsynthesis of sebaceous lipids is regulated by theperoxisome proliferator-activated receptor ligandlinoleic acid in human sebocytesE. Makrantonaki* and C.C. Zouboulis*�

*Laboratory of Biogerontology, Dermatopharmacology and Dermatoendocrinology, Institute of Clinical Pharmacology and Toxicology, Charite University Medicine

Berlin, Campus Benjamin Franklin, Garystrasse 5, 14195 Berlin, Germany

�Departments of Dermatology and Immunology, Dessau Medical Center, Auenweg 38, 06847 Dessau, Germany

CorrespondenceChristos C. Zouboulis.



Key words5a-reductase inhibitor, androgens,

lipids, peroxisome proliferator-activated

receptors, sebocytes


Summary

Background Despite the clinical evidence that androgens stimulate sebaceous lipids,androgens in vitro have shown no similar effects. This contradiction led to theassumption that cofactors may be required for lipid regulation and peroxisomeproliferator-activated receptor (PPAR) ligands were suggested to be adequatecandidates.Objectives The influence of testosterone and linoleic acid, a PPAR ligand, as singleagents and in combination with of LY191704, a 5a-reductase type I inhibitor,was examined on 5a-dihydrotestosterone (5a-DHT) synthesis and lipid contentin human SZ95 sebocytes.Methods Cell proliferation and viability were measured by the 4-methylumbelliferylheptanoate fluorescence assay and by the Boehringer Lactate Dehydrogenase Assaykit, respectively. 5a-DHT enzyme-linked immunosorbent assay was used for thedetection of 5a-DHT synthesis in cell supernatants after treatment, whereas lipidproduction was documented by means of the Nile red lipid microassay and fluor-escence microscopy.Results Testosterone promoted 5a-DHT synthesis (P < 0Æ001), whereas linoleicacid increased sebaceous lipids (P < 0Æ001). The combination of testosterone andlinoleic acid exhibited a synergistic effect on the synthesis of 5a-DHT (P < 0Æ01vs. testosterone) and sebaceous lipids (P < 0Æ01 vs. linoleic acid). Furthermore,LY191704 reduced 5a-DHT and sebaceous lipid levels (P < 0Æ01 and P < 0Æ001in comparison with testosterone/linoleic acid, respectively). Cell proliferationand viability remained unchanged under treatment with all compounds tested.Conclusions These data suggest a catalytic effect of PPAR ligands on cellular testo-sterone activation by 5a-reduction and the importance of the latter for theregulation of sebaceous lipids.

Among the numerous endocrine signals that affect skin, those

of androgens play a predominant role. High testosterone levels

have been implicated in enhanced sebaceous gland activity in

humans1,2 and consequently in diseases marked by seborrhoea,

such as acne vulgaris. Enhanced sebaceous gland activity is

attributed to the potent androgen 5a-dihydrotestosterone

(5a-DHT)3 as sebaceous gland cells possess all necessary

enzymes for conversion of testosterone to 5a-DHT.4

The effects of testosterone and 5a-DHT are mediated by

binding to the nuclear androgen receptor (AR), also expressed

in human sebaceous gland cells.5 The AR is a member of the

steroid superfamily of ligand-dependent transcriptional factors.

5a-DHT binds to the AR with greater affinity than testosterone

and the 5a-DHT/AR complex appears to be more stable6 and,

therefore, more effective.

In contrast to the in vivo observations, in vitro experiments

with human sebocytes have shown that testosterone affects

proliferation in a dose-dependent manner7,8 but not lipid syn-

thesis.9,10 This contradiction led to the assumption that cofac-

tors may be required for the induction of the entire so-called

androgenic influence of the sebaceous gland.11 Current

research has indicated that peroxisome proliferator-activated



receptors (PPARs) and their ligands could be the primary can-

didates.9,10 PPARs regulate multiple lipid metabolism genes in

mitochondria, peroxisomes and microsomes, all prominent in

sebocyte cytoplasm.9,10

In the experiments presented here we examined the influ-

ence of testosterone and linoleic acid (LA), a PPAR-a/d lig-

and12 previously shown to induce lipid synthesis of human

sebocytes in vitro,9,10 as single agents and in combination on

5a-DHT synthesis and lipid content in SZ95 immortalized

human sebaceous gland cells.8 Furthermore, the mechanism of

testosterone 5a-reduction and its influence on lipid synthesis

was examined by adding the 5a-reductase type I inhibitor

LY19170413 to the experimental setting.


Cell culture

Immortalized human SZ95 sebocytes8 were maintained in

Sebomed� medium (Biochrom, Berlin, Germany) supplemen-

ted with 10% fetal calf serum, 5 ng mL)1 recombinant

human epidermal growth factor, 1 mmol L)1 Ca2+ and

50 lg mL)1 gentamicin. The cells were incubated with testo-

sterone (2 · 10)8 mol L)1) (Sigma, Munich, Germany), LA

(10)4 mol L)1) (Sigma) and with the 5a-reductase inhibitor

LY191704 (10)8 mol L)1), provided by Schering AG (Berlin,

Germany). All cells used for the experiments were between

passages 26 and 28.

Cell proliferation

Cell proliferation was assessed after incubation of 1500 SZ95

sebocytes/well in 96-well plates for 48 and 96 h by the

4-methylumbelliferyl heptanoate fluorescence assay, measured

automatically as previously described.14

Detection of cell viability

SZ95 sebocytes were incubated at a density of 20 000 cells/

well in 96-well plates for 48 h and then treated with the

active compounds. After 24 h the supernatants were collected

and centrifuged to remove cell detritus. Finally, 100 lL of the

supernatant was used to measure lactate dehydrogenase release

employing the Boehringer Lactate Dehydrogenase Assay kit

(Mannheim, Germany) and adhering to the manufacturer’s

instructions.

5a-dihydrotestosterone enzyme-linked immunosorbent

assay

After a 48-h adhesion period and a 24-h incubation of SZ95

sebocytes with the active compounds in 24-well plates, the

supernatants were collected, centrifuged in order to remove

cell debris and then the 5a-DHT released was measured by a

5a-DHT enzyme-linked immunosorbent assay (IBL, Hamburg,

Germany) according to the manufacturer’s instructions.

Nile red microassay

After a 48-h adhesion period and a 24-h treatment of 20 000

SZ95 sebocytes/well in 96-well plates with the active com-

pounds, the Nile red microassay for evaluation of lipid content

was performed as previously described:15 485 nm excitation

and 565 nm emission filters were used for neutral lipids

detection and 540 nm excitation and 620 nm emission filters

for polar lipids.

Fluorescence microscopy

SZ95 sebocytes were seeded on 18-mm coverslips in culture

dishes at an approximately 40% confluence. After a 48-h ad-

hesion period, SZ95 sebocytes were treated for 24 h, whereas

control cells received no treatment. Then, the cells were

washed twice with phosphate-buffered saline (PBS), and

100 lL of a 100 lg mL)1 Nile red solution in PBS was added.

The plates were then incubated at 37 �C for 10 min, and the

fluorescence released was determined by Axiophot SIP 31182

fluorescence microscope (Carl Zeiss, Gottingen, Germany).

Statistical analysis

Values represent the mean values ± SD of at least three experi-

ments. Statistical significance was calculated by the two-sample

independent-groups t-test. Mean differences were considered

to be significant when P < 0Æ05.

Results

Preliminary experiments

Testosterone (2 · 10)8 mol L)1), LA (10)4 mol L)1) and

LY191704 (10)8 mol L)1) concentrations used and the treat-

ment duration (up to 96 h) were chosen in preliminary

experiments in order not to influence SZ95 sebocyte prolifer-

ation and viability. They were tested as single compounds and

in combination (data not shown). The testosterone levels

administered were similar to those circulating in healthy

young men.

Regulation of 5a-dihydrotestosterone synthesis

in SZ95 sebocytes

As expected, 5a-DHT synthesis was markedly increased by

treatment with testosterone (P < 0Æ001), while it was not

influenced by LA treatment (Fig. 1). Interestingly, 5a-DHT

synthesis was further increased in SZ95 sebocytes by treatment

with testosterone and LA in comparison with testosterone

(P < 0Æ05) or LA (P < 0Æ001) alone (Fig. 1).

Amplification of lipid synthesis in SZ95 sebocytes

Testosterone alone had no effect on sebocyte lipid content,

while LA increased significantly neutral (sebocyte) lipids



Androgens and PPAR ligands, E. Makrantonaki and C.C. Zouboulis 429

(+ 523%, P < 0Æ001 vs. control) and polar (membrane)

lipids (+ 50%, P < 0Æ001 vs. control) (Fig. 2). The com-

bination of testosterone and LA exhibited a synergistic effect

on the amplification of neutral lipids (+ 599%, P < 0Æ001

vs. control and + 14Æ5%, P < 0Æ01 vs. LA) but not polar

lipids (+ 59%, P < 0Æ001 vs. control and nonsignificant vs.

LA) (Fig. 2).

The lipid content of the treated SZ95 sebocytes was com-

pared after staining with the Nile red dye and visualization of

the fluorescence by fluorescence microscopy (Fig. 3). In con-

trast to untreated SZ95 sebocytes (Fig. 3a), lipid accumulation

was enhanced in LA-treated cells (Fig. 3c), while testosterone

alone showed no effect (Fig. 3b). Treatment with testosterone

and LA induced synergistically the lipid synthesis (Fig. 3d),

corresponding to the results of the Nile red microassay.

LY191704 reduces 5a-dihydrotestosterone and lipid

synthesis in SZ95 sebocytes treated with testosterone

and linoleic acid

The 5a-reductase type I inhibitor LY191704 reduced 5a-DHT

synthesis in SZ95 sebocytes treated with testosterone and LA

() 29%, P < 0Æ01 vs. testosterone/LA treatment) (Fig. 4c) but

did not affect 5a-DHT synthesis in SZ95 sebocytes treated

with either testosterone or LA (Fig. 4a,b). In addition, neutral

lipids were decreased after treatment of SZ95 sebocytes with

all three compounds () 30%, P < 0Æ001 in comparison with

testosterone and LA) (Fig. 3e, 5c) but were not affected after

incubation with LY191704 and either testosterone or LA

(Fig. 5a,b).

Discussion

Several studies have demonstrated that there is an association

between local overproduction of active androgens and skin

disorders such as acne. Patients with acne produce higher rates

of testosterone and 5a-DHT in their skin than healthy individ-

uals.16 In addition, only a few patients with androgenic dis-

orders exhibit hyperandrogenaemia, an observation which

indicates the predominance of peripheral tissue events for the

occurrence of clinical signs.17

In order to gain greater insights into the peripheral andro-

gen activation and activity we examined the effects of testos-

terone at physiological levels and of LA, as single agents and

in combination, on 5a-DHT production and lipid synthesis

in SZ95 sebocytes. Interestingly, the combination exhibited

Fig 1. 5a-dihydrotestosterone (5a-DHT) synthesis in SZ95 sebocytes

after a 24-h treatment in 24-well plates with testosterone (T) and/or

linoleic acid (LA), determined by enzyme-linked immunosorbent

assay. SZ95 sebocytes treated with T + LA produced significantly

higher levels of 5a-DHT in contrast to cells treated only with T

(P < 0Æ05). Control cells without treatment were set at 0% and the

production of 5a-DHT in the treated cells was calculated as percentage

of control. Values represent the mean of three experiments ± SD

(***P < 0Æ001).

Fig 2. Lipid synthesis in SZ95 sebocytes after a 24-h treatment in

96-well plates with testosterone (T) and/or linoleic acid (LA),

determined by Nile red microassay. The production of neutral (light

columns) and polar lipids (dark columns) is illustrated. T showed no

effect on lipid synthesis, whereas LA increased significantly the

production of neutral (P < 0Æ001 vs. control) and polar lipids

(P < 0Æ001 vs. control). The combination of T and LA exhibited a

synergistic effect on neutral lipids (P < 0Æ001 vs. control and P < 0Æ01

vs. LA). Lipid synthesis in control cells without treatment was set at

0% and the lipid production in the treated cells was calculated as

percentage of control. Values represent the mean of three

experiments ± SD with 10 wells evaluated for each data point in each

experiment (***P < 0Æ001).

(a) (b)

(c)

(e)

(d)

Fig 3. Lipid synthesis in SZ95 sebocytes without treatment (a), after a

24-h treatment with testosterone (T) (b) and/or linoleic acid (LA)

(c,d) and LY191704 (e), shown by means of fluorescence

microscopy. The production of sebaceous lipids is illustrated after

staining of SZ95 sebocytes with 100 lg mL)1 Nile red dye. Treatment

only with T showed no effect (b), while LA enhanced lipid

accumulation (c). The combination of T + LA acted synergistically on

lipid synthesis (d). In SZ95 sebocytes treated with

T + LA + LY191704 (e), lipid synthesis was inhibited in contrast to

SZ95 sebocytes treated with T + LA (d).



430 Androgens and PPAR ligands, E. Makrantonaki and C.C. Zouboulis

an unexpected synergistic effect on 5a-DHT production and

synthesis of sebaceous lipids. This is, to our knowledge, the

first work indicating that LA ‘activates’ the stimulatory effect

of testosterone on sebaceous lipogenesis that would be

expected according to in vivo data. Androgens alone have

been found ineffective on sebaceous lipid synthesis in previ-

ous in vitro experiments.9,10 This illustrates that cofactors such

as LA are required for regulation of sebaceous lipids by

androgens and may be responsible for the pathogenesis of

sebaceous gland-associated diseases with excessive sebum

production, such as acne.18 Indeed, Rosenfield et al.9 have

previously demonstrated the interaction of 5a-DHT with

PPAR ligands in inducing differentiation of sebocyte-like rat

preputial cells and lipid synthesis. However, in this study,9

PPAR-c agonists, such as BRL, were shown to be more

effective in stimulating lipid synthesis in the presence of

androgens, whereas no effect was seen with the PPAR-a/dagonist LA. This discrepancy is probably due to the fact that

in our experiments human sebaceous cells (SZ95 sebocytes)

were used, for which recent results have also demonstrated

that PPAR-a is the most important PPAR that regulates lipid

synthesis and inflammation.11,19 In addition, PPAR-a, d, c1and c2 have been shown to be expressed at mRNA and pro-

tein levels in SZ95 sebocytes,10 whereas LA has been docu-

mented to stimulate intracellular accumulation of sebaceous

lipids.9,10

Further to corroborate the importance of testosterone

5a-reduction in the stimulation of sebaceous lipid synthesis, we

performed similar experiments adding LY191704 [8-chloro-

4-methyl-1,2,3,4,4a,5,6,10b-octaahydro-benzo(f)quinolin-3(2H)-

one], a potent, selective, nonsteroidal inhibitor of human

steroid 5a-reductase type I,13 which catalyses the conversion

(a)

(c)

(b)

Fig 4. 5a-dihydrotestosterone (5a-DHT) synthesis in human sebocytes incubated with LY191704 (LY) and testosterone (T) (a) and/or linoleic

acid (LA) (b,c), determined by enzyme-linked immunosorbent assay. LY showed a significant inhibitory effect on 5a-DHT synthesis (P < 0Æ01)

only in the presence of T + LA (c). LY + T (a) and LY + LA (b) showed no effect on 5a-DHT production. As controls were used cells treated

only with T (a), LA (b) and T + LA (c), respectively, and the production of 5a-DHT was calculated as percentage of control. Values represent the

mean of three experiments ± SD (**P < 0Æ01).

(a)

(c)

(b)

Fig 5. Lipid synthesis in SZ95 sebocytes after a 24-h treatment in 96-well plates with LY191704 (LY) and testosterone (T) (a) and/or linoleic

acid (LA) (b,c), determined by Nile red microassay. The production of neutral (light columns) and polar lipids (dark columns) is illustrated. LY

inhibited lipid synthesis in SZ95 sebocytes (P < 0Æ001) only in the presence of T + LA (c). LY + T (a) and LY + LA (b) showed no inhibitory

effect on lipid production. As control were used cells treated only with T (a), LA (b) and T + LA (c), respectively, and the production of lipids

was calculated as percentage of control. Values represent the mean of three experiments ± SD with three wells evaluated for each data point in

each experiment (***P < 0Æ001).



Androgens and PPAR ligands, E. Makrantonaki and C.C. Zouboulis 431

from testosterone to 5a-DHT in peripheral tissues by an

NADPH-dependent reaction.

The isozyme 5a-reductase type I is expressed predominantly

in skin. It is present in the cytoplasm and cell membrane com-

partment in skin cells20 and particularly in facial sebocytes,4

illustrating the key role of sebaceous gland cells in androgen

metabolism.

In accordance with in vivo studies, treatment with LY191704

and either testosterone or LA did not show any inhibitory

effect on 5a-DHT and sebaceous lipid synthesis.21 However,

LY191704 showed its inhibitory action when cells were

exposed concurrently to testosterone and LA. Hence, we hypo-

thesize that the 5a-reductase type I can be inhibited only after

being activated by LA. The simultaneous presence of the pre-

cursor molecule testosterone is of great importance for this

action. Further studies are needed to clarify this hypothesis.

In conclusion, our results suggest that lipid production is

regulated by the PPAR-a/d ligand LA in human sebocytes and

that PPAR ligand-induced 5a-reduction of testosterone is,

indeed, a mechanism for amplifying sebaceous lipid synthesis.

Acknowledgments

This work was supported by research grants of the Deutscher

Dermatologen Kongreß 1995 – Verein zur Forderung der Der-

matologie eV and of Schering AG, Berlin, Germany.

References

1 Pochi PE, Strauss JS. Sebaceous gland response in man to the

administration of testosterone, delta-4-androstenedione, and dehy-droisoandrosterone. J Invest Dermatol 1969; 52:32–6.

2 Giltay EJ, Gooren LJ. Effects of sex steroid deprivation/administra-tion on hair growth and skin sebum production in transsexual

males and females. J Clin Endocrinol Metab 2000; 85:2913–21.3 Zouboulis CC, Degitz K. Androgen action on human skin – from

basic research to clinical significance. Exp Dermatol 2004; 13 (Suppl. 4):5–10.

4 Fritsch M, Orfanos CE, Zouboulis CC. Sebocytes are the key regula-tors of androgen homeostasis in human skin. J Invest Dermatol 2001;

116:793–800.5 Fimmel S, Saborowski A, Orfanos CE et al. Development of efficient

transient transfection systems for introducing antisense oligonucleo-tides into human epithelial skin cells. Horm Res 2000; 54:306–11.

6 Anderson KM, Liao S. Selective retention of dihydrotestosterone byprostatic nuclei. Nature 1968; 219:277–9.

7 Akamatsu H, Zouboulis CC, Orfanos CE. Control of human sebo-cyte proliferation in vitro by testosterone and 5-alpha-dihydrotesto-

sterone is dependent on the localization of the sebaceous glands.J Invest Dermatol 1992; 99:509–11.

8 Zouboulis CC, Seltmann H, Neitzel H et al. Establishment and char-acterization of an immortalized human sebaceous gland cell line

(SZ95). J Invest Dermatol 1999; 113:1011–20.9 Rosenfield RL, Deplewski D, Kentsis A, Ciletti N. Mechanisms of

androgen induction of sebocyte differentiation. Dermatology 1998;

196:43–6.10 Chen W, Yang CC, Sheu HM et al. Expression of peroxisome prolif-

erator-activated receptor and CCAAT/enhancer binding proteintranscription factors in cultured human sebocytes. J Invest Dermatol

2003; 121:441–7.11 Zouboulis CC, Eady A, Philpott M et al. What is the pathogenesis of

acne? Exp Dermatol 2005; 14:143–52.12 Kliewer SA, Forman BM, Blumberg B et al. Differential expression

and activation of a family of murine peroxisome proliferator-activated receptors. Proc Natl Acad Sci USA 1994; 91:7355–9.

13 Hirsch KS, Jones CD, Audia JE et al. LY191704: a selective, nonste-roidal inhibitor of human steroid 5 alpha-reductase type 1. Proc Natl

Acad Sci USA 1993; 90:5277–81.14 Zouboulis CC, Garbe C, Krasagakis K et al. A fluorometric rapid

microassay to identify anti-proliferative compounds for humanmelanoma cells in vitro. Melanoma Res 1991; 1:91–5.

15 Zouboulis CC, Seltmann H, Hiroi N et al. Corticotropin-releasinghormone: an autocrine hormone that promotes lipogenesis in

human sebocytes. Proc Natl Acad Sci USA 2002; 99:7148–53.16 Sansone G, Reisner RM. Differential rates of conversion of testo-

sterone to dihydrotestosterone in acne and in normal humanskin – a possible pathogenic factor in acne. J Invest Dermatol 1971;

56:366–72.17 Orfanos CE, Adler YD, Zouboulis CC. The SAHA syndrome. Horm

Res 2000; 54:251–8.18 Kligman AM. An overview of acne. J Invest Dermatol 1974; 62:268–

87.19 Alestas T, Ganceviciene R, Fimmel S et al. Enzymes involved in the

biosynthesis of leukotriene B(4) and prostaglandin E(2) are activein sebaceous glands. J Mol Med 2006; 84:75–87.

20 Chen W, Zouboulis CC, Fritsch M et al. Heterogeneity and quantita-tive difference of type 1 5 alpha-reductase expression in cultured

skin epithelial cells. Dermatology 1998: 196 51–2.

21 Leyden J, Bergfeld W, Drake L et al. A systemic type I 5 alpha-reductase inhibitor is ineffective in the treatment of acne vulgaris.

J Am Acad Dermatol 2004; 50:443–7.



432 Androgens and PPAR ligands, E. Makrantonaki and C.C. Zouboulis


The antiwrinkle effect of topical concentrated2-dimethylaminoethanol involves a vacuolar cytopathologyG. Morissette, L. Germain* and F. Marceau

Centre de Recherche en Rhumatologie et Immunologie, Centre Hospitalier Universitaire de Quebec, Quebec QC, G1V 4G2, Canada

*Laboratoire d’Organogenese Experimentale, Hopital du Saint-Sacrement du CHA, Quebec QC, G1S 4L8, Canada

CorrespondenceFrancois Marceau.


Accepted for publication21 September 2006

Key words2-dimethylaminoethanol, triethanolamine, vacuolar

adenosine triphosphatase, vacuolar cytopathology


Summary

Background The ‘cosmeceutical’ agent 2-dimethylaminoethanol (DMAE) is a tertiaryamine found in high concentration in numerous topical antiwrinkle preparations.Objectives We hypothesized that a 337 mmol L)1 (3%) DMAE reservoir applied tothe skin could reproduce the cytopathology induced by other amines by main-taining a millimolar drug concentration within a certain depth of the skin layers,and that vacuolar cell expansion could account for the very rapid effect on theapparent skin fullness.Methods Morphological and functional assays were applied to cultured rabbit der-mal fibroblasts treated with tertiary amines in vitro. A morphological verificationof the vacuolization caused by topical DMAE was also attempted in vivo using theinner skin of the rabbit ear and in vitro using primary cultures of human cutane-ous epithelial cells.Results Fibroblasts responded to DMAE (2Æ5–10 mmol L)1) by massive vacuoliza-tion (0Æ5–4 h; phase contrast observations). Triethanolamine, another chemicalfrequently used topically, was also active in this respect (10 mmol L)1). The vac-uolar adenosine triphosphatase inhibitor bafilomycin A1 prevented DMAE- ortriethanolamine-induced vacuolization; adding bafilomycin A1 or cell washoutslowly reversed the established vacuolization induced by DMAE. Further effectsof DMAE in cultured fibroblasts included a moderate cytotoxicity (10 mmol L)1)that was abated by bafilomycin A1 cotreatment, a concentration-dependent mito-tic arrest (2Æ5 mmol L)1) and transient and mild effects on cell ploidy. The epi-dermis of the rabbit external ear was significantly thickened and exhibited clearperinuclear swelling indicative of vacuolization in response to 3% DMAE (1 h;paraffin tissue sections). Cultured human cutaneous epithelial cells responded toDMAE by vacuolization (inhibited by bafilomycin A1 cotreatment).Conclusions The vacuolar cytopathology induced by concentrated organic aminesmay be the cellular basis of the antiwrinkle effect of DMAE.

Despite their wide use, the pharmacology and toxicology of

agents used in cosmetology are rarely documented in a

detailed manner in the scientific literature. The ‘cosmeceutical’

agent 2-dimethylaminoethanol (DMAE) is an ingredient of

numerous antiwrinkle creams; in the lay press, it has been

dubbed ‘an instant anti-aging face-lift.’1 Some objective clin-

ical information is available about DMAE: a placebo-controlled

trial consisting of a 337 mmol L)1 (3%) DMAE facial gel

applied for 16 weeks showed efficacy in attenuating wrinkles

of various sizes and improving lip shape and the fullness of

skin.2 The agent was well tolerated during the trial and its

1-year extension. Another randomized, double-blind, split-face

study using 3% DMAE showed a nonsignificant trend towards

increased skin firmness.3 Other related aminoethanol deriva-

tives (di- and triethanolamine) are also widely exploited in

topical formulations.4

The mode of action of DMEA in dermatology is unknown.

There have been speculations about possible interference with

cholinergic neurotransmission (perhaps by analogy with the

antiwrinkle effect of botulinum toxin) and about a mild anti-

inflammatory effect.2 DMEA is a tertiary amine (pKa 8Æ88) and

many cultured cell types respond to a millimolar concentration

of such compounds by a massive vacuolization. This is a form

of cytopathology that is not fully understood but is explained

by ion trapping of the charged organic amines in acidic com-

partments of the cells and secondary osmotic swelling of the


Journal Compilation � 2007 British Association of Dermatologists • British Journal of Dermatology 2007 156, pp433–439 433

organelles.5,6 The reaction is relatively rapid (1–2 h for most

amines) and is driven by vacuolar adenosine triphosphatase

(V-ATPase), as shown by it being fully prevented by treatment

with the specific inhibitor bafilomycin A1; morphological ana-

lyses have shown that it involves at least the trans-Golgi.5,6

Massive vacuolization induced by amine drugs such as pro-

cainamide often involved supratherapeutic concentrations and

is of toxicological interest. Notable exceptions are the local

anaesthetics (lidocaine, etc.) that, used as supramillimolar

solutions applied in confined anatomical areas, cause such

reactions in vivo at sites of injection and in cultured cells of

various types.5,7–9 For some amines, such as triethylamine and

procainamide, the sustained vacuolization response of cultured

cells is associated with only a low frequency of cytotoxicity

that is abated by concomitant bafilomycin A1 treatment, a

mitotic arrest, an inhibition of protein translocation along the

cell secretory pathway and some inhibition of cell migra-

tion.5,6 Other agents that could be considered as substituted

triethylamines (e.g. metoclopramide, chloroquine) also caused

vacuolization, but were more toxic at millimolar concentra-

tions probably as a result of additional actions at other bio-

chemical and/or subcellular levels.5,6,10

We hypothesized that a 337 mmol L)1 (3%) DMAE reser-

voir applied to the skin could reproduce the cytopathology

induced by other amines by maintaining a millimolar drug

concentration within a certain depth of the skin layers, and

that vacuolar cell expansion could account for the very rapid

effect on the apparent skin fullness. This was tested in cultured

rabbit dermal fibroblasts and human epidermal cells. Another

tertiary amine widely used in cosmetics, triethanolamine,11

has been compared with DMAE. A histological verification of

the vacuolization caused by 3% DMAE was also attempted

using the inner skin of the rabbit ear.


Drugs

All drugs were obtained from Sigma-Aldrich (St Louis, MO,

U.S.A.).

Cells and microscopic techniques

The institutional Review Board approved the animal studies

and the animals were cared for in accordance with the Guide

to the Care and Use of Experimental Animals (Canadian Coun-

cil on Animal Care). Rabbit dermal fibroblasts were grown

from explants of fat-free subcutaneous tissue removed from

male New Zealand white rabbits (1Æ5–2 kg). The culture med-

ium was Dulbecco’s modified Eagle’s medium (Invitrogen,

San Diego, CA, U.S.A.) supplemented with 10% fetal bovine

serum and antibiotics.12 An institutional research committee

approved the anonymous use of primary cultures of human

epidermis, harvested and maintained as described.13 Six-well

plates were seeded with 20 000 irradiated 3T3 feeder

cells cm)2 and 10 000 keratinocytes cm)2 at passage 2, as

described.13 Drugs were added to the complete culture med-

ium for each cell type; bafilomycin A1 was initially dissolved

in dimethyl sulphoxide (final concentration of solvent

< 0Æ1%).

The vacuolization of cultured cells, photographed in phase

contrast such that clear vacuoles could be observed on a dark

background, was quantified as the percentage of individual

cell pixels above a set threshold, after increasing the contrast

of the picture in a standardized manner (Adobe Photoshop,

version 6.0; Adobe Systems Inc., San Jose, CA, U.S.A.). These

numerical values were averaged and compared using nonpara-

metric statistical tests.

Cytotoxicity assay

Subconfluent fibroblasts plated in 35-mm Petri dishes were

maintained in their regular culture medium; test drugs or

combinations were added for 4 or 24 h. After this time, the

medium was removed and centrifuged at 200 g for 5 min.

The pellet containing some detached cells was resuspended in

reaction medium comprising 1 mL of fresh 12Æ5 lg mL)1 flu-

orescein diacetate (Sigma-Aldrich) in phosphate-buffered sal-

ine diluted from a 5 mg mL)1 stock in filtered acetone. The

cells were reintroduced into the Petri dish and incubated for

30 min at 37 �C in the dark; propidium iodide (100 lL of

0Æ5 mg mL)1) was then added, the dishes were further incu-

bated for 3–5 min at room temperature and then observed

(epifluorescence 100·) without rinsing, using two sets of

excitation–emission filters (the red fluorescence indicates pro-

pidium iodide uptake by nonviable cells, the green fluores-

cence shows the hydrolysis of the fluorescein ester by viable

cells). The proportion of nonviable cells (adherent or not)

was established by counting.

Other cell-based assays

A fibroblast proliferation assay was applied as previously des-

cribed for smooth muscle cells.5 For the ploidy assay, fibro-

blasts maintained in complete culture medium were treated

with drugs for various time periods (up to 48 h), detached

using trypsin–ethylenediamine tetraacetic acid, fixed with 95%

ethanol, stained with propidium iodide and analysed using the

EPICS XL cytofluorometry apparatus (Beckman Coulter, Fuller-

ton, CA, U.S.A.; excitation 488 nm, emission 620 nm); results

were analysed using the System II software (version 3.0; Beck-

man Coulter).

Effect of 2-dimethylaminoethanol application on rabbit

ear skin

Freshly prepared emulsions of DMAE (3%) in heavy mineral

oil (Fisher Scientific, Nepean, ON, Canada) were applied to

the inner bare surface of the external ear of conscious rabbits.

The contralateral ear was the control (treated with the oil

vehicle). Animals were killed by CO2 inhalation after 1 h.

The procedure has no observable impact on the behaviour of



434 2-Dimethylaminoethanol-induced vacuolization, G. Morissette et al.

animals and no consistent macroscopic effect on the skin. The

treated areas were dissected and multiple tissue samples were

fixed and processed for paraffin sections (haematoxylin and

eosin stain).

Results

Massive fibroblast vacuolization induced by

2-dimethylaminoethanol

Rabbit dermal fibroblasts are spindle-shaped cells that respond

to HCl-buffered DMAE (pH 7Æ4; structure in Fig. 1c) by a

massive vacuolization over a period of 4 h (phase contrast

observations at 4 h, Fig. 1a; graphical representation of digit-

ized images, Fig. 1b). A period of 4 h was chosen for stan-

dardized observations, but a robust morphological response

was observed as early as 30 min (DMAE 10 mmol L)1;

data not shown). A concentration response study showed that

the 1 mmol L)1 level was not usually active to induce vacuoli-

zation, but that a concentration-dependent response occurred

in the 2Æ5–10 mmol L)1 range. At 10 mmol L)1, the cells

usually exhibited a more or less retracted morphology, some

being rounded and detached. In massively vacuolated cells,

nuclei were often located at the periphery and exhibited a

condensed appearance, as previously observed with smooth

muscle cells treated with other amines.5,6 Triethanolamine

(structure, Fig. 1c; pKa 7Æ8) is related to DMEA but is more

polar; it was less active than DMEA at inducing vacuolization

of the fibroblasts (Fig. 1a,b). However, the response to

10 mmol L)1 of each agent was of comparable magnitude.

Not only did bafilomycin A1 prevent DMAE- or triethanol-

amine-induced vacuolization (P < 0Æ001, Mann–Whitney test

for the 10 mmol L)1 concentration of each amine; Fig. 1b),

but it also slowly reversed the established morphological alter-

(a) (b)

(c)

(d)

Fig 1. Massive vacuolization of rabbit dermal fibroblasts induced by 2-dimethylaminoethanol (DMAE) or triethanolamine [(EtOH)3N]. (a) Phase

contrast observation of cells treated for 4 h with buffered (pH 7Æ4) DMAE or (EtOH)3N introduced into the regular culture medium (original

magnification · 100). (b) Quantification of the vacuolization as a percentage of individual cell-surface pixels above a set threshold following

image treatment (see Materials and methods). Values are means ± SEM of 106–236 determinations. (c) Structure of DMAE and (EtOH)3N.

(d) Bafilomycin A1 (300 nmol L)1) reverses the vacuolized appearance of cells pretreated with DMEA (10 mmol L)1; treatment schedules as

indicated). The control at the left of each data set has received the dimethyl sulphoxide vehicle of bafilomycin A1 for the same period.



2-Dimethylaminoethanol-induced vacuolization, G. Morissette et al. 435

ations induced by DMAE (10 mmol L)1; Fig. 1d), suggesting

that V-ATPase is continuously functioning to maintain a pro-

ton gradient that attracts the amine and water in specific

organelles. Cell washout with fresh culture medium also

reversed the vacuolization induced by DMAE (pretreatment for

8 h) in a manner dependent both on time and on DMAE con-

centration (Fig. 2). Thus, the morphological effect of the

amine was not significant at 4 h after washout in cells pre-

treated with 2Æ5 mmol L)1, whereas those exposed to the

higher concentration level (10 mmol L)1) were still vacuolar;

at 16 h post-washout, vacuolization was not significant for the

10 mmol L)1 DMAE concentration (digitized morphological

analysis, Fig. 2b).

Treatment with DMAE (5 or 10 mmol L)1 only) for 24 h

was associated with a significant cytotoxicity that was also pre-

vented by bafilomycin A1 cotreatment (Fig. 3). Treatment with

DMAE for 4 h was not cytotoxic despite extensive vacuolization

at some concentrations (Fig. 3). For comparison, the DNA-

binding antimitotic agent actinomycin D showed no cyto-

toxicity at 4 h, but induced an extensive mortality at 24 h that

was not abated by bafilomycin A1 (Fig. 3). Similar findings

were made using rabbit vascular smooth muscle cells, except

that DMAE-induced cytotoxicity was more extensive (74% of

cells). Cotreatment of smooth muscle cells with bafilomycin

(a)

(b)

Fig 2. Regression of 2-dimethylaminoethanol (DMAE)-induced vacuolization following cell washout with fresh culture medium. Rabbit dermal

fibroblasts were pretreated for 8 h with DMAE or its saline vehicle, then were observed, washed and observed again after further incubation at

37 �C, as indicated. (a) Sample photographic record. (b) Quantification of the vacuolization, as in Figure 1b. Values are means ± SEM of 67–141

determinations.

Fig 3. Cytotoxicity assay applied to dermal fibroblasts maintained in

the presence of 2-dimethylaminoethanol (DMAE) (4 or 24 h): effect

of drugs or drug combinations on fibroblast mortality. The controls

are common for DMAE and actinomycin D (indicated as the ‘zero’

concentration of DMAE). Proportion of nonviable cells established

using 177–474 cells per experimental point.




A1 was again very effective in preventing cell death (reduced

to 3%, data not shown).

Organic amines at concentrations that elicit vacuolization,

such as procainamide (2Æ5 mmol L)1) and bafilomycin A1,

exert an antiproliferative effect on vascular smooth muscle.5,6

The same applied to DMAE when used at a noncytotoxic con-

centration (2Æ5 mmol L)1; Fig. 4a). After treatment for 48 h

with the cytotoxic agent actinomycin D, the cell number in

Petri dishes was lower than the number of seeded cells

(50 000), but this was not the case for the noncytotoxic con-

centrations of DMAE or procainamide.

In a previous study, procaine at a concentration that pro-

duced vacuolization inhibited the mitosis of hepatoma cells or

fibroblasts by retarding the late S and G2 phases of the cell

cycle.14 DMAE (2Æ5 mmol L)1) was only transiently active in

this respect in rabbit fibroblasts (Fig. 4b), as the effect of this

amine on ploidy was neutral for treatments longer than 6 h.

Sub-G1 peaks, indicative of apoptosis, were small and uninflu-

enced by DMAE in these experiments (data not shown).

In other cell cultures, bafilomycin A1 treatment (24 h,

300 nmol L)1) had a more sustained effect on fibroblast

ploidy: the proportion of cells in S phase was 55% (33% in

controls); in G1/G0 phase was 41% (54% in controls); and in

G2/M phase was 4% (12Æ8% in controls; n ¼ 2). Under the

same experimental conditions, colchicine exerted a profound

reduction in the proportion of cells in G1/G0 phase with an

increase in the proportion of cells in G2/M phase (Fig. 4c;

a control for cell reactivity).

In vivo 2-dimethylaminoethanol application

Measurements were performed using a photographic record

(paraffin tissue sections, 1000·) of the inner bare skin of the

rabbit external ear treated for 1 h with 3% DMAE or vehicle

(Fig. 5; n ¼ 2). The epidermis was significantly thickened by

DMAE treatment (Fig. 5a), and recognizable cellular structures

located in the suprabasal layers of the epidermis were more

frequent (arrowheads in Fig. 5a). Furthermore, DMAE-treated

skin exhibited numerous examples of clear perinuclear areas

indicative of vacuolization (koilocytosis, arrowheads; statistical

analysis, Fig. 5b) in the epidermal cells as a response to 3%

DMAE. In the affected cells, nuclei appeared somewhat con-

densed and the cell volume was enlarged. The tissues were

otherwise unremarkable (no signs of inflammation).

Effect of 2-dimethylaminoethanol on human epidermal

cells

As the in vivo experiments indicated the epidermis as a major

target for topically administered DMAE, the effect of a

2Æ5 mmol L)1 concentration of the amine was verified in pri-

mary cultures of human skin epithelial cells (4 h treatments,

Fig. 6). DMAE induced intracellular vacuoles of various sizes,

many of which were very small and difficult to photograph.

Bafilomycin A1 cotreatment prevented this morphological

response yet had no effect alone (Fig. 6).

Discussion

Virtually all cultured cell types respond within hours to a

wide array of organic amines by massive cell vacuolization

explained by ion trapping in acidic organelles and subse-

quent osmotic swelling. DMAE and triethanolamine, agents

used in cosmetics, are active in the millimolar concentration

range, as demonstrated with rabbit dermal fibroblasts and

human keratinocytes from the epidermis and hair follicles.

(a)

(b) (c)

Fig 4. (a) Effect of 2-dimethylaminoethanol

(DMAE), actinomycin D (actino D) or

procainamide (PA) on the proliferation of

rabbit dermal fibroblasts. Approximately

50 000 cells were seeded in Petri dishes

3 days before the counts and the drugs were

introduced for the last 48 h. The controls are

common for all drugs (indicated as the ‘zero’

concentration of DMAE). Values are

means ± SEM of triplicate determinations.

Ploidy in rabbit dermal fibroblasts treated

with (b) DMAE (2Æ5 mmol L)1) or

(c) colchicine (pooled results of experiments

conducted with 10 or 100 lmol L)1) for

various time periods. Values are means ± SEM

of three different determinations based on

different primary cell lines.




The central role of V-ATPase in this phenomenon, previ-

ously established for triethylamine and procainamide-related

drugs,5,6 also applies to DMAE and triethylamine, which are

more hydrosoluble and exhibit higher concentration thresh-

olds for the morphological response. However, the active

DMAE concentration (2Æ5 mmol L)1) is considerably lower

than that found in topical preparations (as described above),

supporting the relevance of this cellular response. The mild

cytotoxicity of DMAE in fibroblasts might derive from an

extreme vacuolization reaction, perhaps with vacuole rupture,

as both activities are abated by the V-ATPase inhibitor bafilo-

mycin A1. However, the correlation between vacuolization

and cytotoxicity varies for each amine and is clear only for

the less toxic drugs.6 Another common response to amines

that induce vacuolization is mitotic arrest with no persistent

effect on ploidy; it has previously been observed that mitotic

arrest occurs only at a concentration of procainamide that

induces vacuolization, and the same holds true for DMAE

(Fig. 4a). Bafilomycin A1 itself induces a complete mitotic

arrest in cultured smooth muscle cells,6 as well as in several

tumour-derived cell lines in a recent study.15 The latter

report documents that this effect was not apoptotic; it was

parallel to an increase in S and G2/M phases that was not

sustained over a continuous treatment period of 48 h. DMAE

only transiently exhibited this type of effect on ploidy,

although the agent effectively blocked cell division. Thus, the

Golgi-disrupting action of concentrated amine drugs may be

the basis of the mitotic arrest without overt cytotoxicity, as

discussed elsewhere.6

(a)

(b)

Fig 5. Histology of the inner bare skin of the rabbit external ear (paraffin sections). The skin was treated in vivo with either 3% 2-dimethyl-

aminoethanol (DMAE) for 1 h or the oil vehicle. (a) Epidermal thickness. Left: sample photographic record (haematoxylin and eosin; original

magnification · 1000). Arrowheads point to recognizable cellular structures in the suprabasal cell layers; these structures were more common in

the photographic record from DMAE-treated skin. Right: comparison of epidermal thickness in 19 control and 28 DMAE-treated fields. Values are

means ± SEM; *P < 0Æ001 (Student’s t-test). (b) Koilocytosis in the epidermal cells in DMAE-treated tissues. Left: photographic record.

Arrowheads point to koilocytic cells. Right: average proportions of koilocytic cells in the whole photographic record as a function of treatment.

Values are means ± SEM of proportions established by two independent observers. Both sets of data indicate a highly significant effect of DMAE

(P < 0Æ001, v2 test).

Fig 6. Morphological effect of 2-dimethylaminoethanol (DMAE)

(2Æ5 mmol L)1, 4 h) on primary cultures of human cutaneous

epithelial cells (phase contrast, original magnification · 100). The

vacuolization response was prevented by cotreatment with bafilomycin

A1 (300 nmol L)1). DMSO, dimethyl sulphoxide.




Rabbit skin has a thin stratum corneum and epidermis, and

the rate of transdermal drug absorption is generally faster than

for human skin.16 Within the limitations of the model, topical

DMAE (3%) was found to increase the epidermal thickness in

1 h in rabbit ears. The relationship between the histological

observations and the cellular cytopathology characterized in

cultured cells is indicated by the increased frequency of appar-

ently vacuolar and dilated cells in the DMAE-treated epidermis.

Thus, it is plausible that a reservoir of concentrated DMAE

applied to the skin elicits cellular expansion through massive

vacuolization for an extended period of time, and this could

lead to the improved appearance of wrinkled skin. It is of

great interest that the application of EMLA cream, a topical

local anaesthetic formulation containing 2Æ5% lidocaine and

2Æ5% prilocaine, has recently been found to increase the thick-

ness of the human epidermis.17 Indeed, these local anaesthet-

ics can be considered as substituted triethylamine derivatives,

several of which induce cell vacuolization effectively; specific-

ally, lidocaine has been shown to be active in this respect

(2Æ5 mmol L)1).5 Thus, epidermal thickening induced by

EMLA cream could be mechanistically similar to the antiwrin-

kle effect of DMAE.

The termination of DMAE action presumably involves sys-

temic absorption following the vanishing of the reservoir

applied to the skin; this was modelled here by cell washout in

in vitro experiments and led to a relatively rapid regression of

vacuolization in fibroblasts. Of note, vacuolization regressed

less rapidly in cells treated with the higher concentration level

of DMAE (10 mmol L)1, Fig. 2), which is considerably lower

than that applied to the skin. An ethanolamine closely related

to DMAE, N-methyldiethanolamine, was sequestered in the skin

and only slowly released into the bloodstream following topical

administration to rats,18 supporting the possibility of a reser-

voir formation by ion trapping for extended periods of time in

the skin cells. To our knowledge, no information is available

on the systemic absorption of DMAE, but animal toxicology

data for triethanolamine suggest efficient transdermal absorp-

tion, low systemic toxicity or carcinogenicity, rapid elimination

in urine and ‘cloudy swelling’ (which may be synonymous

with koilocytosis) in liver and renal tubule cells at high doses.4

Diethanolamine, also commonly used in cosmetic formulations

(1–25%), caused fetal wastage, decreased mitosis and increased

apoptosis in fetal brain tissue when applied to the skin of preg-

nant mice.19 The vacuolar cytopathology induced by concentra-

ted organic amines and its consequences (mitotic arrest,

inhibition of the secretory pathway, minor cytotoxicity) may

not be dissociable from the improvement of the skin appear-

ance that is rapidly produced by topically administered DMAE.

Acknowledgments

We thank Ms Johanne Bouthillier for technical help, Dr Marc

Pouliot (CHUQ-CHUL) for facilitating the access to micro-

scopic equipment and Amelie Lavoie and Dr Danielle Larouche

(LOEX, Hopital du Saint-Sacrement) for help with human

keratinocyte culture. This work was supported by the Cana-

dian Institutes of Health Research (operating grant MOP-

74448, Canada Graduate Scholarships Doctoral Award to

G.M.).

References

1 Perricone N. The Wrinkle Cure: Unlock the Power of Cosmeceuticals for Supple,

Youthful Skin. Emmaus, PA: Rodale Books, 2000.2 Grossman R. The role of dimethylaminoethanol in cosmetic derma-

tology. Am J Clin Dermatol 2005; 6:39–47.3 Uhoda I, Faska N, Robert C et al. Split face study on the cutaneous

tensile effect of 2-dimethylaminoethanol (deanol) gel. Skin ResTechnol 2002; 8:164–7.

4 Knaak JB, Leung HW, Stott WT et al. Toxicology of mono-, di-,and triethanolamine. Rev Environ Contam Toxicol 1997; 149:1–86.

5 Morissette G, Moreau E, C-Gaudreault R, Marceau F. Massive cellvacuolization induced by organic amines such as procainamide.

J Pharmacol Exp Ther 2004; 310:395–406.6 Morissette G, Moreau E, C-Gaudreault R, Marceau F. N-substituted

4-aminobenzamides (procainamide analogs): an assessment ofmultiple cellular effects concerning ion trapping. Mol Pharmacol

2005; 68:1576–89.7 Kim T, Holley GP, Lee JH et al. The effects of intraocular lidocaine

on the corneal endothelium. Ophthalmology 1988; 105:125–30.8 Atilla H, Tekeli O, Can B et al. Effects of intracameral lidocaine on

ocular tissues. Clin Experiment Ophthalmol 2003; 31:73–7.9 Michalik M, Pierzchalska M, Pabianczyk-Kulka A et al. Procaine-

induced enhancement of fluid-phase endocytosis and inhibition ofexocytosis in human skin fibroblasts. Eur J Pharmacol 2003; 475:1–

10.10 Fan C, Wang W, Zhao B et al. Chloroquine inhibits cell growth

and induces cell death in A549 lung cancer cells. Bioorg Med Chem

2006; 14:3218–22.11 Stott WT, Waechter JM, Rick DL et al. Absorption, distribution,

metabolism and excretion of intravenously and dermally adminis-tered triethanolamine in mice. Food Chem Toxicol 2000; 38:1043–51.

12 Marceau F, Tremblay B. Mitogenic action of bradykinin and des-Arg9-bradykinin on cultured fibroblasts. Life Sci 1986; 39:2351–8.

13 Robitaille H, Proulx R, Robitaille K et al. The mitogen-activatedprotein kinase kinase kinase dual leucine zipper-bearing kinase

(DLK) acts as a key regulator of keratinocyte terminal differenti-ation. J Biol Chem 2005; 280:12732–47.

14 Henics T, Wheatley DN. Vacuolar cytoplasmic phase separation incultured mammalian cells involves the microfilament network and

reduces motional properties of intracellular water. Int J Pathol 1997;78:343–54.

15 McSheehy PMJ, Troy H, Kelland LR et al. Increased tumour extra-cellular pH induced by bafilomycin A1 inhibits tumour growth

and mitosis in vivo and alters 5-fluorouracil pharmacokinetics. Eur JCancer 2003; 39:532–40.

16 Anderson JA, Henck JW. Toxicity and safety testing. In: The Biologyof the Laboratory Rabbit (Manning PJ, Ringler DH, Newcomer CE,

eds). San Diego, CA: Academic Press, 1994; 449–66.17 Tahir A, Webb JB, Allen G, Nancarrow JD. The effect of local

anaesthetic cream (EMLA) applied with an occlusive dressing onskin thickness. Does it matter? J Plast Reconstr Aesthet Surg 2006;

59:404–8.18 Leung HW, Ballantyne B, Frantz SW. Pharmacokinetics of N-methyl-

diethanolamine following acute cutaneous and intravenous dosingin the rat. J Toxicol Cutaneous Ocul Toxicol 1996; 15:343–53.

19 Craciunescu CN, Wu R, Zeisel SH. Diethanolamine alters neurogen-esis and induces apoptosis in fetal mouse hippocampus. FASEB J

2006; 20:1635–40.




CLINICAL AND LABORATORY INVESTIGATIONS DOI 10.1111/j .1365-2133.2006.07581.x

Expression of p16, CD95, CD95L and Helix pomatiaagglutinin in relapsing and nonrelapsing very thinmelanomaL.A. Fearfield,*�� J.M.G. Larkin,� A. Rowe,* R. A’Hern,� C. Fisher,� N. Francis,§ R. MacKie,– B. McCann,**M.E. Gore� and C.B. Bunker*�

*Department of Dermatology, Imperial College School of Medicine (START Laboratories), Chelsea and Westminster Hospital, London, U.K.

�Department of Dermatology, Royal Berkshire Hospital, London Road, Reading RG1 5AN, U.K.

�Royal Marsden Hospital, Fulham Road, London, U.K.

§Department of Histopathology, Charing Cross Hospital, London, U.K.

–Department of Dermatology, University of Glasgow, Glasgow, U.K.

**Department of Histopathology, Norfolk and Norwich Hospital, Colney Lane, Norwich, U.K.

CorrespondenceLonise A. Fearfield.


Accepted for publication14 July 2006

Key wordsbiological markers, prognosis, skin neoplasms


Summary

Background The incidence of malignant melanoma is increasing worldwide andpatients are being diagnosed earlier with thinner primary lesions. Most patientswith very thin melanoma (Breslow thickness < 0Æ76 mm) are cured by surgerybut 2–18% relapse locally or with distant metastases.Objectives The objective of this study was to establish potential new prognosticmarkers in very thin melanoma.Methods We identified a group of subjects with relapsing very thin primary cuta-neous melanoma and a matched control group who had not relapsed. We inves-tigated the expression of p16, Helix pomatia agglutinin (HPA), CD95 and CD95ligand (CD95L) by immunohistochemistry on paraffin-embedded tissue sectionsfrom the subject group, their subsequent metastases and the control group.Results Reduced p16 expression was significantly associated with relapse in verythin melanoma (P ¼ 0Æ0129). Loss of p16 expression was also found in 76% ofmetastases. There was no significant association between HPA, CD95 or CD95Lexpression and subsequent relapse.Conclusions This work is the first to show a significant loss of p16 in relapsing verythin melanoma.

Melanoma has been increasing in incidence in caucasian popu-

lations around the world.1 This has been attributed by some

authors to an increased number of thin tumours.2,3 Although

the prognosis for patients presenting with very thin melanoma

(Breslow thickness < 0Æ76 mm) is good with an estimated 5-

year survival of 93Æ2%,4 no reliable prognostic markers exist

to identify those patients at increased risk of relapse. Identifi-

cation of such patients would allow close clinical follow-up

and consideration of entry into trials of adjuvant therapy.

Mutations of the p16 tumour suppressor gene [also called the

cyclin-dependent kinase inhibitor-2A (CDKN2A), p16INK4A or

MTS1 gene] have been identified in many different tumours

including melanoma, pancreatic adenocarcinoma, oesophageal

carcinoma and various haematological malignancies. The gene

has been mapped to human chromosome region 9p215 and

encodes the 40-kDa p16 protein. The p16 protein is a kinase

inhibitor that blocks the CDK4 and CDK6 kinases. These kinases

phosphorylate the retinoblastoma protein (pRb) allowing pro-

gression through the G1–S cell-cycle checkpoint. Loss of p16

consequently allows cells to escape cell-cycle arrest in the G1

phase. Immunohistochemical studies have generally shown a

correlation between reduced p16 protein expression and the

progression of melanoma6–9 although this association has not

been investigated specifically in very thin melanoma.

The CD95/CD95 ligand (CD95L) (or FAS/FAS ligand) system

is involved in various immune functions10–12 including cyto-

toxic T-cell-mediated apoptosis.13,14 A number of solid tumours

express CD95L15 and hepatocellular tumour cells expressing

CD95L are able to kill infiltrating T lymphocytes, thereby

evading immune attack.16 Furthermore, CD95L expressed by

melanoma cell lines is functionally active against FAS-sensitive

A20 B lymphoma cells.17 A plausible association between the

expression of CD95L and tumourigenicity in melanoma there-

fore exists and this is generally supported by immunohisto-

chemical studies on melanoma samples from patients.18,19

However, only small numbers of very thin melanoma are



reported in the literature and differences in methodology and

patient populations make comparison between published stud-

ies difficult.

Although the identity of the ligand for the Roman snail Helix

pomatia agglutinin (HPA) lectin in malignant cells is unknown,

binding of the lectin to tissue sections from breast, colon and

stomach carcinomas has been associated with a poor progno-

sis.20–23 Few studies have looked at the expression and bio-

logical significance of HPA binding in melanoma although it

has been reported that the extent of HPA lectin binding by

three human melanoma cell lines is related to their potential

for metastasis formation in an animal model,24 and HPA bind-

ing to primary melanoma sections has been reported as a mar-

ker for the risk of relapse in a clinical study.25 However, no

published studies have investigated the relationship between

HPA binding and risk of relapse in very thin melanoma.

In this study we have characterized a group of patients with

primary cutaneous relapsing very thin melanoma (Breslow

thickness < 0Æ76 mm) and a matched control group who have

not relapsed. We have investigated the expression of the po-

tential prognostic markers p16, CD95, CD95L and HPA by

immunohistochemistry on paraffin-embedded tissue sections

from the subject group, their subsequent metastases and the

control group and correlated marker expression with clinical

outcome.


Subject recruitment

Ethical approval for the study was obtained both locally from

the Royal Marsden Hospital (RMH) Research Ethics Committee

and from the North Thames Multicentre Research Ethics Com-

mittee (MREC). Subjects (alive and deceased) with relapsing

very thin melanomas (Breslow thickness < 0Æ76 mm) were

identified from: (i) a clinical database at RMH spanning 1990–

1998; (ii) a histopathology database at RMH spanning January

1980 to July 1999; and (iii) via clinicians in the U.K. Melanoma

Study Group. The sample size was determined by the number of

subjects available. Patients with more than one primary cuta-

neous melanoma were excluded. Forty-five subjects with relap-

sing very thin primary cutaneous melanoma were initially

identified. Twenty-five of these subjects were excluded from the

data set after clinicopathological review due to insufficient bi-

opsy of a large melanoma (n ¼ 2), histologically unconfirmed

metastatic disease (n ¼ 1), vulval melanoma (n ¼ 1), thick

melanomas erroneously entered into the database as thin (n ¼2), original tissue blocks not available (n ¼ 7) and Breslow

thickness > 0Æ76 mm on review of original histological material

(n ¼ 12). Controls were identified as patients with very thin

melanomas that had not relapsed after at least 5 years of follow-

up from the RMH, Charing Cross and Chelsea and Westminster

Hospitals histopathology databases and the U.K. Melanoma

Study Group. Controls were selected to match individually the

subjects in terms of age, sex, melanoma primary site and hist-

ology. Forty-eight controls were initially identified. Fifteen

controls were subsequently excluded after review due to inad-

equate initial biopsy (n ¼ 1), death from other causes within

5 years of diagnosis of primary melanoma (n ¼ 3), follow-up

information not available (n ¼ 5), more than one primary

melanoma (n ¼ 2), no primary tissue available for re-assess-

ment (n ¼ 2) and Breslow thickness > 0Æ76 mm (n ¼ 2). Ori-

ginal paraffin-embedded blocks of primary tumours and

metastases were obtained from participating institutions.

Immunohistochemistry

Standard methods were used to prepare 3-lm sections. Heat-

mediated antigen retrieval by pressure cooker or water bath

method was performed prior to staining for CD95L and p16

using antigen unmasking solution (Vector Laboratories, Bur-

lingame, CA, U.S.A.). Proteolytic pretreatment with porcine

trypsin (1 mg mL)1; Sigma-Aldrich Chemical Co, Poole, U.K.)

for 20 min at 37 �C was performed prior to staining with

HPA. The following primary antibodies were incubated over-

night at 4 �C: mouse monoclonal IgG2a antihuman p16; rabbit

polyclonal IgG antihuman CD95L and mouse monoclonal IgG1

antihuman CD95 (all Santa Cruz Biotechnology, Santa Cruz,

CA, U.S.A.). HPA (Sigma-Aldrich) was used at 10 lg mL)1

followed by antibiotin monoclonal mouse antibody (Vector

Laboratories), both for 1 h at room temperature. Antibiotin

monoclonal mouse (biotin–HPA), anti-p16 and anti-CD95

antibodies were detected with the mouse IgG Vectastain� and

CD95L with the rabbit IgG Vectastain� ABC alkaline phospha-

tase kits and Vector Red alkaline phosphatase solution was

used as the cromophore (all Vector Laboratories).

Analysis of immunostaining

Immunostaining of sections was measured in three ways:

immune reactivity (p16, CD95 and CD95L), percentage of

positive cells or nuclei (HPA and p16, respectively) and stain-

ing index (p16). For the assessment of immune reactivity, at

least two nonadjacent sections were used for each antibody

from both primary tumours and metastases. Staining intensity

of primary tumours from subjects and controls was scored

‘blinded’ and metastases ‘unblinded’ by two investigators

(L.A.F. and A.R.). Tumours were scored according to the in-

tensity of stain compared with the normal adjacent epidermis,

sebaceous and eccrine glands (Table 1). Intensity was scored

throughout a section and summed to produce an overall score,

Table 1 Quantitation of immune reactivity

GradeIntensity of staining compared with the normal adjacentepidermis and/or sebaceous glands

0 Weak: absent to less than a quarter of the intensity

1 Mild: quarter to less than half the intensity2 Moderate: half to less than three-quarters of the intensity

3 Strong: three-quarters to less than equal to the intensity4 Very strong: equal or more intense



p16, HPA and CD95/CD95L in very thin melanoma, L.A. Fearfield et al. 441

i.e. 10% of a nonuniformly stained tumour might be graded

0, 70% graded 2 and 20% graded 4 for an overall score of

2Æ2 [(10/100 · 0) + (70/100 · 2) + (20/100 · 4)]. For

the assessment of percentage of positive cells or nuclei, 5–10

high power fields (150 · 150 lm) were selected randomly

and scored throughout each representative section for the

number of HPA-positive cells and the number of p16-positive

nuclei. A staining index, combining immune reactivity and

percentage of positive cells, adapted from Straume et al.,9 was

calculated as follows: staining index ¼ staining intensity

grade · percentage of positive cells grade (Table 2).

Statistical methods

Nonparametric tests were used to compare continuous variables

due to the non-normality of the data. The Mann–Whitney U-test

was used to compare the potential markers in subjects at base-

line and controls, and the Wilcoxon signed-rank sum test was

used to compare the paired values in subjects at baseline with

their metastases. Median differences and their 95% confidence

intervals (CIs) were calculated. Analyses were carried out using

statistical programmes written in Microsoft Excel.

Results

Forty-five subjects with relapsing very thin primary cutaneous

melanoma (Breslow thickness < 0Æ76 mm) and 48 controls

with nonrelapsing very thin primary cutaneous melanoma were

initially identified. Twenty-five subjects and 15 controls were

excluded from the data set after clinicopathological review (see

Materials and methods). The clinicopathological features of

patient and control groups included in the study are summar-

ized in Table 3. The only statistically significant difference

(Mann–Whitney U-test) between the subjects and controls with

regard to the clinicopathological parameters was the mitotic

rate: significantly more controls had mitoses of < 1 mm)2

(P < 0Æ01). Sixty-four percent of controls were Clark level II or

less compared with 40% of subjects but this was not significant.

Exclusions of subjects and controls after review therefore did

not result in notable differences between the two groups and

Table 3 demonstrates the comparability of the two groups.

p16

Immunostaining for p16 was performed on sections from 18

subjects, their 17 subsequent metastases and 28 controls (Fig. 1). Nuclear staining in sebaceous and eccrine glands was

used as a positive internal control as in other studies.9 Three

indices were used to quantitate immunostaining: immune re-

activity, percentage of p16-positive cells and staining index

(Table 4).

Subjects vs. controls

Median immune reactivity scores were 1Æ9 [interquartile range

(IQR) 1Æ3–2Æ1] for subjects and 2Æ4 (IQR 2Æ1–3Æ0) for controls

(median difference )0Æ7, 95% CI )1Æ2 to )0Æ2, P ¼ 0Æ01).

Table 2 Calculation of staining index

Percentage of

positive cells

Percentage of

positive cells grade

Immune

reactivity

Staining

intensity grade

0 0 0–1 0< 10% 1 > 1–2 1

10–50% 2 > 2–3 2> 50% 3 > 3–4 3

Table 3 Summary of clinicopathological characteristics of subjects andcontrols

Subjects Controls

Total number 20 33Sex, female, n (%) 13 (65) 22 (67)

Age, years, median (range) 47 (23–63) 44 (22–84)Site, n (%)

Head and neck 1 (5) 1 (3)Arm 4 (20) 8 (24Æ2)

Leg 7 (35) 12 (36Æ4)Trunk 8 (40) 12 (36Æ4)

Breslow mm, median (range) 0Æ55 (0Æ2–0Æ7) 0Æ5 (0Æ2–0Æ7)Clark level, n (%)

I 1 (5) 0II 7 (35) 21 (64)

III 12 (60) 8 (24)IV 0 4 (12)

V 0 0

VGP, n (%) 16 (80) 26 (79)Regression, n (%) 11 (55) 16 (48)

Mitoses per mm2, n (%)0 7 (35) 6 (18)

< 1a 9 (45) 27 (82)2 4 (20) 0

Tumour type, n (%)Superficial spreading 17 (85) 32 (97)

Lentigo 1 (5) 1 (3)Nodular 1 (5) 0

Unclassified 1 (5) 0TIL, n (%)

No 6 (30) 10 (30Æ3)Mild 7 (35) 12 (36Æ4)

Moderate 0 7 (21Æ2)Brisk 7 (35) 4 (12Æ1)

aP ¼ 0Æ004 for the subject vs. the control group. TIL, tumour-

infiltrating lymphocytes; VGP (vertical growth phase), the focalappearance within the radial growth phase of a new cell popula-

tion growing as an expanding spheroidal nodule with metastaticpotential; regression, the complete or partial disappearance of

the primary lesion spontaneously, heralded by TILs, melanin-containing macrophages, vascular ectasia and at end stage, fibro-

sis and scarring; brisk, TIL noted throughout the substance ofthe VGP or infiltrating across the entire base of the VGP; non-

brisk, TIL present in one or more foci of the VGP; absent,lymphocytes present but not infiltrating the melanoma or any

part of the VGP.



442 p16, HPA and CD95/CD95L in very thin melanoma, L.A. Fearfield et al.

The median percentage of p16-positive cells was 28Æ8% (IQR

16–35Æ3%) for subjects and 35Æ5% (IQR 32–48%) for controls

(median difference )11Æ5%, 95% CI )20% to )1Æ5%, P ¼

0Æ02) and median staining index was 2 (IQR 2–4) for subjects

and 4 (IQR 2–6) for controls (median difference )2, 95% CI

)2 to 0, P ¼ 0Æ03).

Fig 1. Representative sections (· 250) stained for p16, CD95, CD95 ligand (CD95L) and Helix pomatia agglutinin (HPA) to demonstrate mild/

moderate [(a) p16, (c) CD95, (e) CD95L and (g) HPA] and moderate/intense staining [(b) p16, (d) CD95, (f) CD95L and (h) HPA].




Subjects vs. metastases

The median immune reactivity score for metastases was 0Æ8(IQR 0–1Æ4, median difference –0Æ9, 95% CI )1Æ3 to )0Æ6,

P < 0Æ01 vs. subjects). The median percentage of p16-positive

cells was 0Æ5% for metastases (IQR 0–7Æ5%, median difference

)23Æ2%, 95% CI )30Æ3% to )13%, P < 0Æ001 for subjects vs.

metastases). Median staining index was 0 for metastases (IQR

0–2, median difference )2, 95% CI )2 to )2, P < 0Æ001 for

subjects vs. metastases). Furthermore, 22% of primary mela-

noma sections had a staining index of zero in comparison

with 76% of metastases.

These results demonstrate, in all scoring systems, a signifi-

cant reduction in p16 staining in very thin melanoma in sub-

jects who subsequently relapsed in comparison with controls

who had not relapsed. Further significant loss of p16 staining

was seen in metastases in comparison with the corresponding

primary tumours.

CD95 and CD95L

Tissue blocks were available for sectioning in 18 subjects,

their 17 subsequent metastases and 32 controls for CD95 and

in 18 subjects, 17 subsequent metastases and 30 controls for

CD95L. Due to diffuse immunostaining it was not possible to

count individual cells positive for CD95 and CD95L and so

only immune reactivity was measured; normal epidermis and

adnexal structures adjacent to tumours were used as internal

controls (Fig. 1, Table 5). For CD95 staining, median immune

reactivity score was 1Æ1 (IQR 1Æ0–1Æ6) for subjects, 1Æ5 (IQR

1Æ0–1Æ8) for controls and 1Æ5 (IQR 0Æ7–2Æ3) for metastases

(P not significant for all comparisons). For CD95L staining,

median immune reactivity score was 2Æ3 (IQR 1Æ8–2Æ6) for

subjects, 1Æ8 (IQR 1Æ4–2Æ3) for controls and 1Æ9 (IQR 1Æ6–2Æ7)

for metastases (P not significant for all comparisons). These

results show no significant association between staining for

CD95 or CD95L in very thin melanoma and the likelihood of

subsequent relapse and no significant difference in CD95 or

CD95L expression between primary very thin melanomas and

their subsequent metastases.

Helix pomatia agglutinin

HPA immunostaining was performed on sections from 15

subjects, their 15 metastases and 27 controls. The intensity of

immune reactivity was uniform throughout the sections and

therefore only the percentage of positive cells stained per high

power field was recorded (Fig. 1, Table 5). The median per-

centage of positive cells per high power field was 40 (IQR

33–50) for subjects, 27 (IQR 21–43) for controls and 21

(IQR 0–68) for metastases. The median difference in percent-

age of positive cells per high power field was 10% for subjects

vs. controls (95% CI )3Æ5 to 20, P ¼ 0Æ10) and )17Æ3% for

subjects vs. metastases (95% CI )24 to 5Æ5, P not significant).

These results demonstrate no significant association between

staining for HPA in very thin melanoma and the likelihood of

subsequent relapse and no significant difference in HPA

expression between primary very thin melanomas and their

subsequent metastases.

Discussion

In this study we have shown that the expression of p16 is sig-

nificantly reduced in relapsing very thin melanoma compared

with nonrelapsing controls and in relapsing primary mela-

noma compared with subsequent metastases in terms of the

immune reactivity score, percentage of p16-positive nuclei

and the combined staining index. This is the first study to spe-

cifically analyse various immunohistochemical markers in this

important although rare group of patients that at present have

no other significant prognostic indicators available to them

except Breslow thickness. However, the limitations of this

study include the small number of subjects and the fact that

the analysis was retrospective although the databases were

prospectively maintained. As the study is small, even quite

large differences between subjects and controls will not reach

statistical significance. Relapsing very thin melanoma is rela-

tively rare and as a result small numbers are inevitable; how-

ever, the number of subjects reported here is comparable to

other similar published studies.26–29

The results reported here are similar to other studies show-

ing that p16 expression is frequently reduced in relapsing

sporadic melanoma. For example, in a series of 102 primary

invasive cutaneous melanomas, the 5-year survival rate for

Table 4 Summary of p16 staining scores in subjects, controls andmetastases; medians and interquartile ranges are shown

p16 Subjects Controls Metastases

Immune reactivity 1Æ9 (1Æ3–2Æ1) 2Æ4 (2Æ1–3Æ0)a 0Æ8 (0–1Æ4)b

Percentage

positive cells

28Æ8 (16–35Æ3) 35Æ5 (32–48)c 0Æ5 (0–7Æ5)d

Staining index 2 (2–4) 4 (2–6)e 0 (0–2)f

aP ¼ 0Æ01 for subjects vs. controls. bP < 0Æ01 for subjects vs.

metastases. cP ¼ 0Æ02 for subjects vs. controls. dP < 0Æ001 forsubjects vs. metastases. eP ¼ 0Æ03 for subjects vs. controls.fP < 0Æ001 for subjects vs. metastases.

Table 5 Summary of CD95, CD95L and HPA staining scores in

subjects, controls and metastases; medians and interquartile ranges areshown

Subjects Controls Metastases

CD95: immunereactivity

1Æ1 (1Æ0–1Æ6) 1Æ5 (1Æ0–1Æ8) 1Æ5 (0Æ7–2Æ3)

CD95L: immunereactivity

2Æ3 (1Æ8–2Æ6) 1Æ8 (1Æ4–2Æ3) 1Æ9 (1Æ6–2Æ7)

HPA: % positivecells

40 (33–50) 27 (21–43) 21 (0–68)

P not significant for comparisons between subjects, controls andmetastases for all scores.




patients with a significant reduction or loss of p16 staining

was zero compared with a 5-year survival rate of 67% in

those patients with moderate or strong p16 staining.7 In a fur-

ther series of 202 vertical growth phase primary melanomas

and 68 corresponding metastases, absent or minimal nuclear

staining significantly predicted poor patient survival with 37%

and 67% estimated 10-year survival rates for subjects with

absent or present p16 expression, respectively.9 However, our

report is the first to show that loss of p16 expression is a poor

prognostic factor in very thin melanoma.

The loss of p16 expression in melanoma metastases that we

report has been observed previously. In our study, 76% of

metastases had a staining index score of 0 for p16. A similar

proportion of metastases (77%) with a staining index of 0–1

has been reported in one previous study9 and a further report

demonstrated absent p16 expression in 44% of metastatic

melanoma lesions compared with 9% of primary invasive

melanomas.6 Loss of p16 in metastases in comparison with

primary tumours suggests that subclones of p16-deficient

tumour cells may be more likely to escape from the primary

tumour and to survive and proliferate at distant sites.

A role for p16 mutations in familial melanoma is well

recognized;30 however, its role in sporadic melanoma is less

clear. Mutations in the p16 gene have been identified in up to

33% of sporadic melanomas31–38 and therefore it is likely to

be an important factor in a proportion of sporadic melanomas.

A higher frequency of mutations is found in melanoma cell

lines. A number of other mechanisms of p16 inactivation such

as large heterozygous and homozygous deletions and silencing

by methylation of CpG islands in the promoter region have

been characterized. Their frequency and role in pathogenesis

and progression of melanoma are still unclear.

No significant differences between subjects, controls or

metastases were found in the immune reactivity scores for

CD95 or CD95L. However, the immune reactivity scores for

CD95 were generally reduced in all subjects and we have also

demonstrated that CD95L is expressed in most subjects and

controls with very thin melanoma. A previous study39 found

CD95L expression in only two of 12 very thin melanomas; a

further study reported the staining of frozen sections from pri-

mary melanomas (Breslow thickness 1Æ8–8Æ6 mm) and subse-

quent cutaneous metastases, demonstrating CD95L expression

in only 10% of primaries and 55% of the cutaneous metasta-

ses.19 Maeda et al.18 performed staining on paraffin-embedded

tissue from various benign melanocytic lesions, melanoma in

situ, stage I melanoma (Clark level 2 or 3), advanced mela-

noma (Clark level 4 or 5) and lymph node metastases. All

advanced melanomas and lymph node metastases of cutaneous

origin showed CD95L expression but melanoma in situ, six

very thin melanomas, benign or dysplastic naevi did not. This

latter discrepancy may be explained, in part, by the different

scoring methods used, as tumours were scored positive in the

Maeda et al. study only if the intensity of the CD95L signal

was greater than that of keratinocytes. In contrast, we have

compared signal intensity between tumour and epidermis,

sebaceous and eccrine glands. However, Sprecher et al.40 did

find CD95L expression in normal melanocytic naevi, Spitz

naevi and melanomas—including very thin lesions.

Three ideas have been advanced to explain how alterations

in the CD95/CD95L system may account for the escape from

immune surveillance by melanoma cells, which leads to diffi-

culties in interpretation of CD95 and CD95L immune staining

in melanoma. Experimental data interpreted to support all

three hypotheses have been published. The first such idea is

that CD95 may be downregulated and CD95L upregulated

on melanoma cells.41,42 The consequence of this would be

the induction of apoptosis in incoming CD95-expressing T

lymphocytes. The two other hypotheses that have been

suggested are: first, that CD95L may be downregulated, CD95

upregulated or normally expressed and that melanoma cells

may survive by escaping autocrine suicide/apoptosis;43,44 and

second, that expression of CD95 and CD95L may be normal

but the receptor may be nonfunctional or be blocked, for

example by soluble CD95, again resulting in escape from

apoptosis.45–47

We have shown that HPA binding is higher in relapsing

very thin melanomas than in controls, although the difference

is not statistically significant. This result is compatible with a

study24 that demonstrated that the extent of lectin binding by

three human melanoma (LOX, FEMX-1 and SESX) cell lines

was related to their ability to form metastases in nude mice.

Furthermore, Thies et al.25 reported a series of 100 patients

with cutaneous malignant melanoma followed up for at least

10 years. Tumour sections were stained with several lectins in-

cluding HPA. Kaplan–Meier analysis of time to first metastasis

showed a positive correlation between HPA binding and meta-

stasis formation and multivariate analysis demonstrated that

HPA binding to primary melanoma was an independent mar-

ker for the risk of metastasis. Small numbers and methodolo-

gical differences in the current study may in part account for

the lack of a significant association. For example, we used bi-

otinylated HPA as per the method of Schumacher et al.22

whereas Thies et al.25 used a native unconjugated HPA; differ-

ences in the fixing and processing of samples within our study

may also have caused variation in HPA binding between

samples (see Schumacher et al.48).

Our results support the hypothesis that expression of N-ace-

tyl-galactosamine (GalNAc)/-glucosamine (GlcNAc) residues

on the cell surface of malignant cells, as detected by HPA

binding, are involved in the process of metastatic spread.22,25

As the expression of HPA-binding glycoconjugates by primary

tumours is associated with metastatic potential, the metastases

themselves might also be expected to be invariably positive

for HPA binding. Investigations of HPA binding to metastases

have clearly demonstrated that this is not the case; for exam-

ple, about 20% of breast cancer metastases are negative for

HPA binding.49,50 Although the specific HPA-binding glyco-

conjugate associated with metastatic behaviour has yet to be

elucidated, some explanations have been advanced to explain

the increased expression of HPA-binding saccharides in malig-

nant cells. One possibility is the neo-expression of GalNAc/-

GlcNAc binding sites.25 Another possibility is that terminal




carbohydrate residues are lost in malignant cells with the con-

sequent unmasking of subterminal HPA-binding regions.51

In conclusion, we have shown that there is no significant

association between HPA, CD95 or CD95L expression and

subsequent relapse although reduced p16 expression is shown

for the first time to be significantly associated with relapse in

very thin melanoma.

References

1 Marks R. Epidemiology of melanoma. Clin Exp Dermatol 2000;25:459–63.

2 Dennis LK. Analysis of the melanoma epidemic, both apparent andreal: data from the 1973 through 1994 surveillance, epidemiology,

and end results program registry. Arch Dermatol 1999; 135:275–80.3 MacKie RM, Bray CA, Hole DJ et al. Incidence of and survival from

malignant melanoma in Scotland: an epidemiological study. Lancet2002; 360:587–91.

4 Clark WH Jr, Elder DE, Guerry D 4th et al. Model predicting sur-vival in stage I melanoma based on tumor progression. J Natl Cancer

Inst 1989; 81:1893–904.5 Kamb A, Gruis NA, Weaver-Feldhaus J et al. A cell cycle regulator

potentially involved in genesis of many tumor types. Science 1994;264:436–40.

6 Reed JA, Loganzo F Jr, Shea CR et al. Loss of expression of thep16/cyclin-dependent kinase inhibitor 2 tumor suppressor gene in

melanocytic lesions correlates with invasive stage of tumor pro-gression. Cancer Res 1995; 55:2713–18.

7 Straume O, Akslen LA. Alterations and prognostic significance ofp16 and p53 protein expression in subgroups of cutaneous mela-

noma. Int J Cancer 1997; 74:535–9.8 Sparrow LE, Eldon MJ, English DR et al. p16 and p21WAF1 protein

expression in melanocytic tumors by immunohistochemistry. Am JDermatopathol 1998; 20:255–61.

9 Straume O, Sviland L, Akslen LA. Loss of nuclear p16 protein

expression correlates with increased tumor cell proliferation (Ki-67) and poor prognosis in patients with vertical growth phase

melanoma. Clin Cancer Res 2000; 6:1845–53.10 Nagata S. Apoptosis by death factor. Cell 1997; 88:355–65.

11 Nagata S. Human autoimmune lymphoproliferative syndrome, adefect in the apoptosis-inducing Fas receptor: a lesson from the

mouse model. J Hum Genet 1998; 43:2–8.12 Van Parijs L, Abbas AK. Homeostasis and self-tolerance in the

immune system: turning lymphocytes off. Science 1998; 280:243–8.13 Stalder T, Hahn S, Erb P. Fas antigen is the major target molecule

for CD4+ T cell-mediated cytotoxicity. J Immunol 1994; 152:1127–33.

14 Hanabuchi S, Koyanagi M, Kawasaki A et al. Fas and its ligand in ageneral mechanism of T-cell-mediated cytotoxicity. Proc Natl Acad Sci

USA 1994; 91:4930–4.15 Walker PR, Saas P, Dietrich PY. Tumor expression of Fas ligand

(CD95L) and the consequences. Curr Opin Immunol 1998; 10:564–72.

16 Strand S, Hofmann WJ, Hug H et al. Lymphocyte apoptosis inducedby CD95 (APO-1/Fas) ligand-expressing tumor cells – a mechan-

ism of immune evasion? Nat Med 1996; 2:1361–6.17 Hahne M, Rimoldi D, Schroter M et al. Melanoma cell expression

of Fas(Apo-1/CD95) ligand: implications for tumor immuneescape. Science 1996; 274:1363–6.

18 Maeda A, Aragane Y, Tezuka T. Expression of CD95 ligand in mel-anocytic lesions as a diagnostic marker. Br J Dermatol 1998;

139:198–206.

19 Terheyden P, Siedel C, Merkel A et al. Predominant expression ofFas (CD95) ligand in metastatic melanoma revealed by longitudinal

analysis. J Invest Dermatol 1999; 112:899–902.20 Leathem AJ, Brooks SA. Predictive value of lectin binding on

breast-cancer recurrence and survival. Lancet 1987; 1:1054–6.21 Okuyama T, Maehara Y, Kakeji Y et al. Interrelation between

tumor-associated cell surface glycoprotein and host immuneresponse in gastric carcinoma patients. Cancer 1998; 82:1468–75.

22 Schumacher U, Higgs D, Loizidou M et al. Helix pomatia agglutininbinding is a useful prognostic indicator in colorectal carcinoma.

Cancer 1994; 74:3104–7.

23 Schumacher U, Adam E. Lectin histochemical HPA-binding patternof human breast and colon cancers is associated with metastases

formation in severe combined immunodeficient mice. Histochem J1997; 29:677–84.

24 Kjonniksen I, Rye PD, Fodstad O. Helix pomatia agglutinin bindingin human tumour cell lines: correlation with pulmonary metastases

in nude mice. Br J Cancer 1994; 69:1021–4.25 Thies A, Moll I, Berger J et al. Lectin binding to cutaneous malig-

nant melanoma: HPA is associated with metastasis formation. Br JCancer 2001; 84:819–23.

26 Gromet MA, Epstein WL, Blois MS. The regressing thin malignantmelanoma: a distinctive lesion with metastatic potential. Cancer

1978; 42:2282–92.27 Paladugu RR, Yonemoto RH. Biologic behavior of thin malignant

melanomas with regressive changes. Arch Surg 1983; 118:41–4.28 Ronan SG, Eng AM, Briele HA et al. Thin malignant melanomas

with regression and metastases. Arch Dermatol 1987; 123:1326–30.29 Vilmer C, Bailly C, Le Doussal V et al. Thin melanomas with un-

usual aggressive behavior: a report on nine cases. Melanoma Groupof French Federation of Cancer Centers. J Am Acad Dermatol 1996;

34:439–44.30 Tucker MA, Goldstein AM. Melanoma etiology: where are we?

Oncogene 2003; 22:3042–52.31 Gruis NA, Weaver-Feldhaus J, Liu Q et al. Genetic evidence in

melanoma and bladder cancers that p16 and p53 function in separ-ate pathways of tumor suppression. Am J Pathol 1995; 146:1199–

206.32 Healy E, Sikkink S, Rees JL. Infrequent mutation of p16INK4 in

sporadic melanoma. J Invest Dermatol 1996; 107:318–21.33 Maelandsmo GM, Florenes VA, Hovig E et al. Involvement of the

pRb/p16/cdk4/cyclin D1 pathway in the tumorigenesis of spor-adic malignant melanomas. Br J Cancer 1996; 73:909–16.

34 Kumar R, Lundh Rozell B, Louhelainen J et al. Mutations in the

CDKN2A (p16INK4a) gene in microdissected sporadic primarymelanomas. Int J Cancer 1998; 75:193–8.

35 Fujimoto A, Morita R, Hatta N et al. p16INK4a inactivation is notfrequent in uncultured sporadic primary cutaneous melanoma.

Oncogene 1999; 18:2527–32.36 Peris K, Chimenti S, Fargnoli MC et al. UV fingerprint CDKN2a but

no p14ARF mutations in sporadic melanomas. J Invest Dermatol1999; 112:825–6.

37 Piccinin S, Doglioni C, Maestro R et al. p16/CDKN2 and CDK4gene mutations in sporadic melanoma development and progres-

sion. Int J Cancer 1997; 74:26–30.38 Ruiz A, Puig S, Lynch M et al. Retention of the CDKN2A locus

and low frequency of point mutations in primary andmetastatic cutaneous malignant melanoma. Int J Cancer 1998; 76:

312–6.39 Ekmekcioglu S, Okcu MF, Colome-Grimmer MI et al. Differential

increase of Fas ligand expression on metastatic and thin or thickprimary melanoma cells compared with interleukin-10. Melanoma

Res 1999; 9:261–72.




40 Sprecher E, Bergman R, Meilick A et al. Apoptosis, Fas and Fas-lig-and expression in melanocytic tumors. J Cutan Pathol 1999; 26:

72–7.41 Aragane Y, Maeda A, Cui CY et al. Inhibition of growth of mela-

noma cells by CD95 (Fas/APO-1) gene transfer in vivo. J Invest Der-matol 2000; 115:1008–14.

42 Soubrane C, Mouawad R, Antoine EC et al. A comparative study ofFas and Fas-ligand expression during melanoma progression. Br J

Dermatol 2000; 143:307–12.43 Ouhtit A, Gorny A, Muller HK et al. Loss of Fas-ligand expression

in mouse keratinocytes during UV carcinogenesis. Am J Pathol 2000;

157:1975–81.44 Owen-Schaub LB, van Golen KL, Hill LL et al. Fas and Fas ligand

interactions suppress melanoma lung metastasis. J Exp Med 1998;188:1717–23.

45 Ferrarini M, Imro MA, Sciorati C et al. Blockade of the Fas-triggeredintracellular signaling pathway in human melanomas is circumven-

ted by cytotoxic lymphocytes. Int J Cancer 1999; 81:573–9.

46 Mouawad R, Khayat D, Soubrane C. Plasma Fas ligand, an inducerof apoptosis, and plasma soluble Fas, an inhibitor of apoptosis, in

advanced melanoma. Melanoma Res 2000; 10:461–7.47 Shin MS, Park WS, Kim SY et al. Alterations of Fas (Apo-1/CD95)

gene in cutaneous malignant melanoma. Am J Pathol 1999;154:1785–91.

48 Schumacher U, Adam E, Brooks SA et al. Lectin-binding propertiesof human breast cancer cell lines and human milk with particular

reference to Helix pomatia agglutinin. J Histochem Cytochem 1995;43:275–81.

49 Schumacher U, Kretzschmar H, Brooks S et al. Helix pomatia lectin

binding pattern of brain metastases originating from breast cancers.Pathol Res Pract 1992; 188:284–6.

50 Brooks SA, Leathem AJ. Expression of n-acetyl galactosaminylatedand sialylated glycans by metastases arising from primary breast

cancer. Invasion Metastasis 1998–99; 18:115–21.51 Fenlon S, Ellis IO, Bell J et al. Helix pomatia and Ulex europeus lectin

binding in human breast carcinoma. J Pathol 1987; 152:169–76.





Inactivation of the CDKN2A and the p53 tumour suppressorgenes in external genital carcinomas and their precursorsN. Soufir, S. Queille,* M. Liboutet,* O. Thibaudeau,* F. Bachelier,� G. Delestaing,* B.C. Balloy,� J. Breuer,§A. Janin,� L. Dubertret,� C. Vilmer� and N. Basset-Seguin*�

Laboratoire de Biochimie Hormonale et Genetique, Hopital Bichat Claude Bernard, 46 rue Henri Huchard, 75018 Paris, France

*INSERM U532, �Department of Dermatology and �Department of Pathology, Hopital Saint-Louis, Paris, France

§Department of Medical Microbiology, St Barts and the Royal London Hospital Medical School, London, U.K.

CorrespondenceNadem Soufir.


Accepted for publication25 May 2006

Key wordsCDKN2A, genital carcinoma, human

papillomavirus, p53


C.V. and N.B-S. have codirected the work.

Summary

Background p53 has been extensively studied in external genital carcinoma (EGC),and is frequently inactivated, but little is known about the role of the CDKN2Atumour suppressor gene in the oncogenesis of EGC.Objectives To investigate the role of CDKN2A and p53 in the pathogenesis of EGCsand their precursor lesions vulval intraepithelial neoplasia (VIN3), penile intra-epithelial neoplasia and lichen sclerosus (LS).Methods By means of CDKN2A and p53 mutation screening (single-strand conform-ational polymorphism analysis and sequencing), methylation analysis of alternat-ive CDKN2A promoters (methylation-specific polymerase chain reaction) and p53immununochemistry, we analysed eight invasive EGCs (five from vulva and threefrom penis) and 25 precancerous lesions (two undifferentiated VIN3 and 23 vul-val/penile lesions of LS) from 33 patients.Results p53 mutations (mainly transversions) and CDKN2A mutations (includingone hot spot) were present in 75% and 50% of invasive tumours, respectively,but were absent in all precancerous lesions. Remarkably, all CDKN2A-mutatedtumours also harboured a p53 mutation. CDKN2A or p53 mutations wereobserved more frequently in LS-derived EGCs than in human papillomavirus-derived EGCs (P ¼ 0Æ053). A positive anti-p53 staining, but without p53 muta-tions, was also detected in 30% of LS lesions, suggesting a p53 stabilization inresponse to inflammation and carcinogenic insult. Methylation of p16INK4a andp14ARF promoters was not a frequent mechanism of CDKN2A inactivation.Conclusions Our study shows a high prevalence of co-inactivating mutations of p53and/or CDKN2A genes in EGC, that seem to occur preferentially in LS-derivedtumours and late in oncogenesis.

Squamous cell carcinomas (SCCs) of the external genital

region (vulva and penis) [mucous external genital carcinoma

(EGC)] and their precursor lesions can be separated into two

main subtypes that differ in terms of pathological and clinical

presentations.

Human papillomavirus (HPV)-associated carcinomas are of

basaloid or warty type, whereas tumours unrelated to HPV are

usually keratinizing and differentiated, usually associated with

lichen sclerosus (LS).1,2 Precancerous high-grade external

genital lesions are defined as vulval or penile intraepithelial

neoplasia (VIN3 or PIN3) and can also be classified as

HPV-dependent (undifferentiated VIN3 or PIN3) or HPV-

independent lesions (differentiated VIN3 or PIN3, mainly

associated with LS lesions).3,4

The role of the p53 gene has been extensively studied in

EGC.5–12 In EGC associated with oncogenic HPV infection,

inactivation of the p53 protein occurs through binding with

HPV E6 protein and subsequent degradation. In HPV-negative

carcinomas, mainly developed in the context of genital LS le-

sions,13 p53 is frequently inactivated by somatic mutations.7

Degradation of the p53 protein by viral oncoprotein leads to

the absence of p53 protein detection in HPV-positive EGC,

whereas somatic mutations of the p53 gene are often associ-

ated with increased stability of the mutant inactive p53 pro-

tein in HPV-negative EGC.

In contrast, little is known on the role of the CDKN2A gene

in EGC oncogenesis. CDKN2A encodes two tumour suppressor

proteins p16INK4a and p14ARF, that control cell growth



through the Rb-CDK4 and p53 pathways, respectively,14 and

is frequently inactivated in human cancer.15 Inactivation of

CDKN2A can occur by deletion, point mutation or methylation

of the two CpG-rich promoter regions.

CDKN2A is an important regulator of the proliferation of

both normal and tumorigenic squamous epithelial cells.16

p16INK4a accumulates in normal human keratinocytes under-

going replicative senescence17 and is inactivated during kera-

tinocyte immortalization,18 and p14ARF is also implicated in

cellular senescence19 and immortalization.20,21

To our knowledge, only few studies have been performed

on the role of CDKN2A in EGCs, and none has compared this

with the status of p53. In this work, we have analysed in a

series of EGCs and precancerous lesions the status of CDKN2A

and that of p53. The importance of the two main pathways

involved in the control of cell proliferation in the transform-

ation of external genital lesions is discussed.


Patients and DNA extraction

Thirty-three skin specimens were collected from 33 patients in

an outpatient department dedicated to genital diseases. All pre-

senting patients with HPV, benign LS lesions or malignant

tumours were asked to participate in the study with their

informed consent (Dermatology Department, Hopital Saint-

Louis, Paris, France). Skin samples included eight invasive

SCCs of the external genital mucosa (five from vulva and three

from penis), two undifferentiated VIN3 and 23 vulval or

penile LS. Invasive SCC was defined as an invasive proliferation

made of sheets of atypical, sometimes dyskeratotic, squamous

cells, with a variable number of concentric layers and kerati-

nizing pearls. Undifferentiated in situ neoplasia (VIN3 or PIN3)

was defined as full-thickness atypia with disordered matur-

ation and atypical mitosis throughout the epithelium. LS was

characterized by a band of hyalinized collagen in the upper

dermis with a variable lymphocytic infiltrate. Genital carci-

noma was histologically and clinically defined as HPV-related

when associated with lesions of undifferentiated VIN3 or PIN3

and as LS-derived when associated with typical LS lesions.

Tissues were divided into two parts. One part was snap fro-

zen in liquid nitrogen and stored at )80 �C. The other part

was fixed in formalin, embedded in paraffin, and subjected to

histopathological assessment performed by the same patholo-

gist. DNA was prepared from frozen samples as previously de-

scribed.22

Immunohistochemistry

Skin samples fixed in 10% formalin were deparaffinized, and

stained after antigen retrieval in a microwave oven at 450 W

for 15 min in Tris buffer (pH 7Æ3) with monoclonal antibody

D-07 (code M7001; Dako, Glostrup, Denmark), which reacts

with both wild-type and mutant forms of the human p53 pro-

tein, at a dilution of 1 : 50, and revealed by avidin–biotin-

coupled immunoperoxidase staining and diaminobenzidine. A

p53-mutated tumour was always added as a positive control.

Staining was nuclear, and was negative when omitting the pri-

mary antibody. Mayer’s haematoxylin was used for counter-

staining. p53 positivity was evaluated by the percentage of

p53 nuclear staining on two independent microscopic fields

(magnification ·200) including at least 100 cells. Slides were

independently examined by two pathologists.

Single-strand conformational polymorphism analysis and

sequencing

The coding sequence and flanking intronic sequences of exon

1a, exon 1b and exon 2 of the CDKN2A gene were analysed

by polymerase chain reaction (PCR)–single-strand conforma-

tional polymorphism analysis as previously described. Primer

sequences for exons 1a and 2 were previously published.23,24

Exon 1b was analysed through two overlapping PCR products

generated by using the following primers: P14F1 5¢-TCAGGG-

AAGGGCGGGTGCG-3¢, P14R1 5¢-GCCGCGGGATGTGAACCA-3¢(PCR product 245 bp), P14F2 5¢-GCCGCGAGTGAGGGTTTT-

3¢, P14R2 5¢-CACCGCGGTTATCTCCTC-3¢ (PCR product

257 bp). Exons 4–9 of the p53 gene were amplified using the

same conditions as previously described.23 Products from 50-

lL PCR reactions from reamplified shifted bands and tumoral

genomic DNAs were purified using an Amicon Microcon 100

(Millipore, Beverly, MA, U.S.A.) and sequenced using a dye-

labelled terminator on an automated sequencer 310 (Applied

Biosystems, Foster City, CA, U.S.A.) according to the manufac-

turer’s instructions.

p16INK4a and p14

ARF promoter methylation analysis

Methylation analysis of p16INK4a and p14ARF promoters was

performed as methylation-specific PCR assay (MSP).25 In brief,

DNA modification by bisulphite converts exclusively unmeth-

ylated cytosines to uracil. Subsequent PCR amplification with

primers specific for unmethylated vs. methylated DNA reveals

the methylation status of investigated DNA sections.

Bisulphite-modified DNA was purified using a commercially

available PCR purification kit according to the manufacturer’s

recommendations (Qiagen, Valencia, CA, U.S.A.). Modified and

purified DNA was precipitated by ethanol for 12 h, resuspended

in 100 lL of water and used immediately or stored at )20 �C.

Primer pairs for PCR amplification are given in Table 1 and

were purchased from MWG-Biotech (Courtaboeuf, France).

Amplification was performed for 35 cycles (denaturing

95 �C/5 min, annealing 60 �C/90 s, extension 72 �C/60 s)

and was concluded by a final 8-min extension at 72 �C. A

positive control using universally methylated (or unmethylat-

ed) DNA was added for each reaction. All specimens were

analysed at least twice. A negative lane was added for each

run. Methylation was confirmed if at least two experiments

demonstrated an unequivocal MSP amplification product. In

addition, all methylation patterns were confirmed by sequen-

cing the PCR products of the corresponding bisulphite DNAs.



CDKN2A and p53 in external genital carcinomas, N. Soufir et al. 449

Human papillomavirus typing

HPV typing was performed blindly by J.B., without informa-

tion regarding the type of lesion analysed, as previously de-

scribed for the detection of HPV in genital lesions. The primer

pair MY09–MY11 was originally designed by Manos et al.26

for the detection of HPV in genital lesions. To increase the

level of sensitivity, a second established primer, GP6, was

seminested within MY11.27,28

Statistical methods

We used the v2 test (Fisher’s exact unconditional test when

necessary) to evaluate the association between the occurrence

of CDKN2A and p53 mutations in external genital tumours. We

also compared the frequency of the association of mutations

in p53 and CDKN2A between external genital tumours and

sporadic ultraviolet (UV)-induced skin carcinomas, and com-

pared the HPV16 status between VIN3/PIN3-related tumours

or lesions and LS lesions. Two-sided tests were computed, and

P < 0Æ05 was considered statistically significant. The SAS soft-

ware (SAS Institute, Cary, NC, U.S.A.) was used for statistical

analysis.

Results

Thirty-three patients were included in the analysis. Eight

patients had invasive EGCs (five women and three men; mean

age at diagnosis 72 years, range 59–88). Of these carcinomas,

four were LS-derived EGC as associated with typical LS lesions,

three were HPV-related EGC as associated with undifferentiat-

ed VIN3 or PIN3 lesions, and one was an EGC of undeter-

mined origin as no associated lesion was seen in pathological

samples. In addition, 25 patients with precancerous lesions

were studied, including two women with undifferentiated

VIN3 and 23 patients with LS (21 women and two men). The

mean age at diagnosis was 58Æ6 years.

Seven p53 mutations were observed in exons 5, 7 and 8 of

the p53 gene, in six of eight tumours (75%), one tumour

harbouring two different mutations (Table 2). All mutations

were of missense type. Overall, p53 mutations were present

in all LS-derived carcinomas (four of four), but in only one of

three of the HPV-derived carcinomas (P ¼ 0Æ053). No p53

mutation was observed in the precancerous lesions (undiffer-

entiated VIN3 or LS lesions).

Two different CDKN2A mutations were found in four of

eight invasive tumours (50%): two LS-derived tumours, one

HPV-derived tumour and one of undetermined origin. These

mutations were C fi T transitions located in exon 2 of

CDKN2A and had different consequences on p16INK4a and

p14ARF reading frames. Both were p16INK4a nonsense muta-

tions (Arg58Ter in one case, Arg80Ter in three cases), leading

to a truncated p16INK4a protein. In addition, these mutations

also induced p14ARF amino acid substitutions at residues

located either on conserved codons or in the nucleolar local-

ization domain of the p14ARF protein (Table 2). Remarkably,

all CDKN2A-mutated tumours also harboured a mutation of the

p53 gene. No CDKN2A mutations were found in precancerous

lesions, but a previously described p16INK4a exon 2 poly-

morphism was observed in two cases (Table 2).

Methylation of the p16INK4a gene promoter was not detec-

ted in any of the eight tumours analysed but was present in

one of 11 LS lesions studied. Methylation of the p14ARF gene

promoter was observed in one tumour that was neither

mutated for CDKN2A nor for p53, and was also observed in

one of 11 LS lesions studied (Table 2).

A strong (> 50%) anti-p53 staining was observed in two of

five p53-mutated tumours but in none of three HPV-derived

tumours without p53 mutation (Fig. 1a). In addition, a posi-

tive anti-p53 staining ranging from 10% to > 50% was seen

in seven of 19 LS lesions (Fig. 1b).

HPV16 was detected in four of eight invasive tumours:

three of three undifferentiated VIN3/PIN3 lesion-derived

tumours and one tumour of undetermined origin. HPV16 was

also detected in the two undifferentiated VIN3. In contrast,

HPV was not detected in the two LS-derived carcinomas ana-

lysed (Table 2). HPV16 was statistically more frequent in un-

differentiated VIN3/PIN3-related tumours or in VIN3 lesions

Table 1 Polymerase chain reaction (PCR)primers for methylation-specific PCR of the

CDKN2A genePrimer

set

PCRproduct

size (bp)

Annealingtemperature

(�C)

P16-Ma TTATTAGAGGGTGGGGCGGATCGCGACCCCGAACCGCGACCGTAA

150 60

P16-Ua TTATTAGAGGGTGGGGTGGATTGTCAACCCCAAACCACAACCATAA

151 60

P14-Mb GTGTTAAAGGGCGGCGTAGCAAAACCCTCACTCGCGACGA

122 60

P14-Ub TTTTTGGTGTTAAAGGGTGGTGTAGTCACAAAAACCCTCACTCACAACAA

132 60

Primer sequences are written 5¢ fi 3¢. M represents methylation-specific primers and U,

unmethylation-specific primers. aPrimers used by Herman et al.25 bPrimers used by Estelleret al.35



450 CDKN2A and p53 in external genital carcinomas, N. Soufir et al.

Tab

le2

CDKN

2Aan

dp5

3m

utat

ions

,m

ethy

lation

stat

usof

p16IN

K4aan

dp1

4ARF

prom

oter

s,p5

3im

mun

oche

mistry

(IH

C)

and

hum

anpa

pillo

mav

irus

(HPV

)ty

ping

ofex

tern

alge

nita

lca

rcin

omas

(EGCs)

and

thei

rpr

ecur

sor

lesion

s

Patien

t

Age

(yea

rs)

Sex

Dia

gnos

is

p53

IHC

p53

sequ

ence

chan

ge

p53

amin

oac

id

chan

ge

CDKN

2Ase

quen

ce

chan

ge

p16IN

K4aam

ino

acid

chan

ge

p14A

RF

amin

oac

id

chan

ge

HPV

stat

us

Met

hylation

p16IN

K4a

Met

hylation

p14A

RF

Inva

sive

tum

ours

188

FEG

C+

undi

fVIN

3N

D–

––

––

16

––

282

MEG

C+

undi

fPI

N3

ND

Tgc

fiGgc

Ex5

Cys

135G

lyCga

fiTga

Ex2

Arg

80Ter

Pro9

4Leu

16

––

365

FEG

C+

undi

fVIN

3–

––

––

–16

+6

–+

459

MEG

Cun

d50

%aa

Gfi

aaC

Ex5

Lys1

32Asp

Cga

fiTga

Ex2

Arg

80Ter

Pro9

4Leu

16

––

560

FEG

C+

LS>

80%

Ggg

fiTgg

Ex8

Gly

356T

rp–

––

––

–

672

FEG

C+

LSN

DGgc

fiAgc

Ex7

Gly

245S

erCga

fiTga

Ex2

Arg

80Ter

Pro9

4Leu

ND

ND

ND

771

MEG

C+

LSN

DcG

cfi

cAc

Ex5

Cgg

fiTgg

Ex7

Arg

175C

ys

Arg

248T

rp

Cga

fiTga

Ex2

Arg

58Ter

Pro7

2Leu

ND

ND

ND

879

FEG

C+

LSN

DtG

cfi

tTc

Ex7

Cys

252P

he

––

––

––

Prec

ance

rous

lesion

s9

71F

undi

fVIN

3–

––

––

–16

+6

––

1046

Fun

difVIN

3–

––

––

–16

––

1163

FLS

––

––

––

–N

DN

D

1279

FLS

––

––

––

–+

–13

64F

LS–

––

––

–N

DN

DN

D

1466

FLS

––

––

––

ND

–+

1545

FLS

>10

%–

––

––

16

––

1669

FLS

––

–c.44

2G

fiA

Ex2

Ala1

48Th

r–

ND

––

1762

FLS

––

––

––

––

–

1884

FLS

ND

––

––

–6

ND

ND

1946

FLS

>10

%–

––

––

ND

ND

ND

2062

FLS

>50

%–

––

––

ND

––

2136

FLS

>50

%–

––

––

ND

ND

ND

2261

MLS

ND

––

––

–N

D–

–

2365

FLS

10%

––

––

–6

––

2466

FLS

>50

%–

––

––

ND

ND

ND

2540

FLS

––

––

––

ND

ND

ND

2668

FLS

30%

––

––

–16

ND

ND

2763

FLS

––

––

––

–N

DN

D28

67F

LS–

––

––

–N

D–

–

2950

FLS

––

––

––

ND

ND

ND

3062

FLS

ND

––

c.44

2G

fiA

Ex2

Ala1

48Th

r–

16

+6

––

3154

FLS

–N

DN

DN

DN

DN

DN

D–

–32

48M

LSN

D–

––

––

ND

ND

ND

3328

FLS

––

––

––

ND

ND

ND

The

mut

ated

base

isw

ritten

inup

per

case

.Si

gnifi

cant

resu

lts

are

show

nin

bold

face

,an

dpo

lym

orph

ism

sar

esh

own

initalic

s.LS

,lic

hen

scle

rosu

s;Ex

,ex

on;N

D,no

tde

term

ined

;un

difVIN

3or

PIN

3,un

differ

entiat

edvu

lval

orpe

nile

insit

une

oplasia;

und,

unde

term

ined

.




(five of five) than in LS lesions (three of nine; P ¼ 0Æ0157,

Fisher’s exact test ¼ 0Æ031).

Discussion

In this work we analysed the status of the CDKN2A and p53

genes in benign and malignant external genital lesions. Our

results show for the first time and distinctively on a small ser-

ies of invasive tumoral samples two distinct molecular patterns

according to the origin of the EGC: mutations of both the p53

gene and the CDKN2A locus in 100% of LS derived-tumours

and a lower frequency (33%) of p53 or CDKN2A mutations in

undifferentiated VIN/PIN3-related tumours.

p53 has been extensively studied in EGCs. In our study,

p53 mutations were statistically more frequent in LS-derived

carcinomas than in HPV-derived carcinomas. Although our

work included a small series of carcinomas, our results are in

light with previous findings, particularly a meta-analysis that

revealed significantly more p53 mutations in HPV-negative

vulval SCC than in HPV-positive vulval SCC.29

In vulval carcinomas, p53 mutations were mostly transver-

sions (Table 2), in contrast to the p53 mutational spectrum

observed in UV-induced skin carcinomas, which comprises

mostly single or double transitions at dipyrimidic sites

(C fi T or CC fi TT).30 No LS lesions (n ¼ 23) showed any

evidence of p53 mutation, highly suggesting that p53 muta-

tions are a late event in external genital tumours, which has

already been proposed by several authors. This is in contrast

with UV-induced skin carcinomas, for which p53 mutations

are believed to be early events as they are usually present in

chronically sun-exposed skin. However, p53 immunostaining

was present in seven of 19 (37%) LS lesions, with a staining

that was nuclear and mostly basal and suprabasal (Fig. 1). Im-

munostaining intensity varied among these positive samples

but was always > 10% of cells analysed (see Table 2). This

observation was previously reported and was considered to be

an important feature of vulval LS compared with LS from

other sites;11 it was attributed to overexpression of the TP53

wild-type protein reflecting stress response to inflammation

and carcinogenic insult. In our series, the p53 positivity in LS

lesions was not associated with mutations of the p53 gene,

therefore favouring the hypothesis of wild-type protein stabil-

ization, the significance of which awaits further studies.

The CDKN2A gene has been much less studied in EGC. A

decreased expression of p16INK4a protein and p16INK4a promo-

ter methylation have been reported in vulval carcinomas and

their precursors.31,32 In our series, we found CDKN2A muta-

tions in four of eight invasive EGCs (50%), with mutations af-

fecting both p16INK4a and p14ARF. A hot spot (Arg80Ter)

present in three tumours was previously described in a mela-

noma cell line33 and in nonmelanoma skin cancers (NMSCs)

occurring in patients with xeroderma pigmentosum,24 but not

in sporadic NMSC. No mutation was detected in LS, suggest-

ing that, as for the p53 gene, inactivation of CDKN2A by muta-

tions is a late phenomenon in vulval carcinogenesis.

Methylation of CDKN2A promoters was rare in our series

compared with gene inactivation by point mutation. This con-

trasts with a recent work that reported a much higher propor-

tion of p16INK4a gene promoter methylation in vulval

carcinoma as well as in LS lesions (68% and 42Æ8%, respect-

ively).32 This difference might be due (i) to the high propor-

tion of CDKN2A-mutated tumours in our series, in which

methylation of promoters and mutations of CDKN2A were not

associated, and (ii) to differences in sensitivity or specificity of

gene promoter methylation detection. This awaits further an-

alysis. However, we also detected methylation of p16INK4a and

p14ARF promoters in LS lesions, which could mean that epi-

genetic inactivation of the CDKN2A locus most probably repre-

sents an early event, insufficient for malignant transformation,

that may occur in clinically benign lesions such as LS.32 All

CDKN2A-mutated tumours also harboured a mutation in the

p53 gene, suggesting that co-inactivation of both pathways is

important in EGCs, and may be indicating a cooperative effect.

In previous studies we observed a similar finding in cutaneous

SCC from patients with xeroderma pigmentosum,24 but not in

sporadic skin carcinomas.23 In animal models, a cooperative

effect between inactivation of CDKN2A and p53 gene inactiva-

tion in cancer development has been demonstrated.34 p16INK4a

mutations destabilize the G1/S checkpoint by inactivating pRb

and releasing E2F transcription factors that directly activate

p14ARF. Induction of p14ARF then stabilizes p53 and permits

p53-mediated apoptosis. p14ARF or p53 mutants would inhibit

p53-mediated apoptosis. Therefore mutations in both genes

could confer a selective cellular growth advantage, allowing

mutations in other genes to accumulate and promoting tu-

morigenesis.

In conclusion, our study shows a high prevalence of inac-

tivating mutations of p53 and/or CDKN2A genes preferentially

observed in EGC derived from LS (100%), that seem to

occur late in oncogenesis as they were not present in LS

lesions. The carcinogenic insult responsible for the mutations

involving these two crucial genes still remains to be eluci-

dated.

References

1 Hording U, Junge J, Daugaard S et al. Vulvar squamous cell carci-

noma and papillomaviruses: indications for two different etiolo-gies. Gynecol Oncol 1994; 52:241–6.

(b)(a)

Fig 1. p53 Immunohistochemistry in lichen sclerosus (LS) lesion (a)

and LS-derived vulval carcinoma (b). Original magnification · 400.



452 CDKN2A and p53 in external genital carcinomas, N. Soufir et al.

2 Santos M, Montagut C, Mellado B et al. Immunohistochemical stain-ing for p16 and p53 in premalignant and malignant epithelial le-

sions of the vulva. Int J Gynecol Pathol 2004; 23:206–14.3 Ridley CM. Nomenclature of non-neoplastic vulval conditions. Br J

Dermatol 1986; 115:647–8.4 Vilmer C, Cavelier-Balloy B, Nogues C et al. Analysis of alterations

adjacent to invasive vulvar carcinoma and their relationship withthe associated carcinoma: a study of 67 cases. Eur J Gynaecol Oncol

1998; 19:25–31.5 Lee YY, Wilczynski SP, Chumakov A et al. Carcinoma of the vulva:

HPV and p53 mutations. Oncogene 1994; 9:1655–9.

6 Hietanen SH, Kurvinen K, Syrjanen K et al. Mutation of tumor sup-pressor gene p53 is frequently found in vulvar carcinoma cells. Am

J Obstet Gynecol 1995; 173:1477–82.7 Pilotti S, D’Amato L, Della Torre G et al. Papillomavirus, p53 alter-

ation, and primary carcinoma of the vulva. Diagn Mol Pathol 1995;4:239–48.

8 Castren K, Vahakangas K, Heikkinen E, Ranki A. Absence of p53mutations in benign and pre-malignant male genital lesions with

over-expressed p53 protein. Int J Cancer 1998; 77:674–8.9 Ngan HY, Cheung AN, Liu SS et al. Abnormal expression or mutation

of TP53 and HPV in vulvar cancer. Eur J Cancer 1999; 35:481–4.10 Gastrell FH, McConnell DT. Human papillomavirus and vulval

intra-epithelial neoplasia. Best Pract Res Clin Obstet Gynaecol 2001;15:769–82.

11 Vanin K, Scurry J, Thorne H et al. Overexpression of wild-type p53in lichen sclerosus adjacent to human papillomavirus-negative vul-

var cancer. J Invest Dermatol 2002; 119:1027–33.12 Rolfe KJ, MacLean AB, Crow JC et al. TP53 mutations in vulval

lichen sclerosus adjacent to squamous cell carcinoma of the vulva.Br J Cancer 2003; 89:2249–53.

13 Leibowitch M, Molinie V. Vulvar dermatoses and intra-epithelialneoplasia. Classification, criticism and suggestions. Ann Pathol 1996;

16:334–42.14 Larsen CJ. Contribution of the dual coding capacity of the

p16INK4a/MTS1/CDKN2 locus to human malignancies. Prog CellCycle Res 1997; 3:109–24.

15 Sherr CJ. Principles of tumor suppression. Cell 2004; 116:235–46.16 Enders GH, Koh J, Missero C et al. p16 inhibition of transformed

and primary squamous epithelial cells. Oncogene 1996; 12:1239–45.17 Loughran O, Clark LJ, Bond J et al. Evidence for the inactivation of

multiple replicative lifespan genes in immortal human squamouscell carcinoma keratinocytes. Oncogene 1997; 14:1955–64.

18 Hara E, Smith R, Parry D et al. Regulation of p16CDKN2 expression

and its implications for cell immortalization and senescence. MolCell Biol 1996; 16:859–67.

19 Dimri GP, Itahana K, Acosta M, Campisi J. Regulation of a senes-cence checkpoint response by the E2F1 transcription factor and

p14(ARF) tumor suppressor. Mol Cell Biol 2000; 20:273–85.

20 Kamijo T, Zindy F, Roussel MF et al. Tumor suppression at themouse INK4a locus mediated by the alternative reading frame

product p19ARF. Cell 1997; 91:649–59.21 Sherr CJ, Weber JD. The ARF/p53 pathway. Curr Opin Genet Dev

2000; 10:94–9.22 Daya-Grosjean L, Robert C, Drougard C et al. High mutation fre-

quency in ras genes of skin tumors isolated from DNA repair defici-ent xeroderma pigmentosum patients. Cancer Res 1993; 53:1625–9.

23 Soufir N, Moles JP, Vilmer C et al. p16 UV mutations in humanskin epithelial tumors. Oncogene 1999; 18:5477–81.

24 Soufir N, Daya-Grosjean L, de La Salmoniere P et al. Association

between INK4a-ARF and p53 mutations in skin carcinomas ofxeroderma pigmentosum patients. J Natl Cancer Inst 2000;

92:1841–7.25 Herman JG, Graff JR, Myohanen S et al. Methylation-specific PCR: a

novel PCR assay for methylation status of CpG islands. Proc Natl AcadSci U.S.A. 1996; 93:9821–6.

26 Manos MM, Waldman J, Zhang TY et al. Epidemiology and partialnucleotide sequence of four novel genital human papillomaviruses.

J Infect Dis 1994; 170:1096–9.27 Snijders PJ, van den Brule AJ, Schrijnemakers HF et al. The use of

general primers in the polymerase chain reaction permits thedetection of a broad spectrum of human papillomavirus genotypes.

J Gen Virol 1990; 71:173–81.28 van den Brule AJ, Snijders PJ, Raaphorst PM et al. General primer

polymerase chain reaction in combination with sequence analysisfor identification of potentially novel human papillomavirus geno-

types in cervical lesions. J Clin Microbiol 1992; 30:1716–21.29 Rosenthal AN, Hopster D, Ryan A, Jacobs IJ. Immunohistochemical

analysis of p53 in vulval intraepithelial neoplasia and vulval squa-mous cell carcinoma. Br J Cancer 2003; 88:251–6.

30 Basset-Seguin N, Moles JP, Mils V et al. TP53 tumor suppressorgene and skin carcinogenesis. J Invest Dermatol 1994; 103:S102–6.

31 Chan MK, Cheung TH, Chung TK et al. Expression of p16INK4 andretinoblastoma protein Rb in vulvar lesions of Chinese women.

Gynecol Oncol 1998; 68:156–61.32 Lerma E, Esteller M, Herman JG, Prat J. Alterations of the p16/Rb/

cyclin-D1 pathway in vulvar carcinoma, vulvar intraepithelial neo-plasia, and lichen sclerosus. Hum Pathol 2002; 33:1120–5.

33 Pollock PM, Pearson JV, Hayward NK. Compilation of somatic mu-tations of the CDKN2 gene in human cancers: non-random distribu-

tion of base substitutions. Genes Chromosomes Cancer 1996; 15:77–88.34 Sharpless NE, Alson S, Chan S et al. p16(INK4a) and p53 deficiency

cooperate in tumorigenesis. Cancer Res 2002; 62:2761–5.

35 Esteller M, Cordon-Cardo C, Corn PG et al. p14ARF silencing bypromoter hypermethylation mediates abnormal intracellular local-

ization of MDM2. Cancer Res 2001; 61:2816–21.





Serum chemokine profile in patients with bullouspemphigoidH. Nakashima, M. Fujimoto,* N. Asashima, R. Watanabe, Y. Kuwano, N. Yazawa, N. Maruyama,� H. Okochi,�A. Kumanogoh§ and K. Tamaki

Department of Dermatology, Faculty of Medicine, University of Tokyo, Tokyo, Japan

*Department of Dermatology, Kanazawa University Graduate School of Medical Science, 13-1 Takaramachi, Kanazawa, Ishikawa 920-8641, Japan

�Division of Biostatistics, School of Pharmaceutical Science, Kitasato University, Tokyo, Japan

�Department of Regenerative Medicine, Research Institute, International Medical Center of Japan, Tokyo, Japan

§Department of Immunopathology, Research Institute for Microbial Diseases, Osaka University, Osaka, Japan

CorrespondenceM. Fujimoto.



Key wordsautoimmunity, bullous pemphigoid, chemokine,

multiplex protein analysis system, pemphigus

vulgaris, Th1/Th2


Summary

Background Bullous pemphigoid (BP) is an autoimmune inflammatory disease caus-ing blister formation at the dermoepidermal junction. Cutaneous infiltration ofactivated CD4+ T cells and eosinophils is an early event in blister formation dur-ing the disease process, suggesting that the trafficking of circulating leucocytesthrough the sites of inflammation is crucial in the pathogenesis of the disease.While the accumulated evidence suggests that some cytokines are involved in thepathogenesis, there have been few reports about serum chemokine profiles inpatients with BP.Objectives To determine serum profiles of various chemokines and their clinical as-sociation in patients with BP.Methods Concentrations of 10 chemokines – interferon (IFN)-c-inducible protein-10 (IP-10), monokine induced by IFN-c (MIG), macrophage inflammatory pro-tein (MIP)-1a, MIP-1b, RANTES, eotaxin, monocyte chemoattractant protein(MCP)-1, MCP-2, MCP-3 and growth-regulated oncogene-a – were measured inserum samples from 38 patients with BP, 16 with pemphigus vulgaris (PV) and17 normal controls using a sandwich immunoassay-based multiplex protein arraysystem.Results While there was no significant increase in any serum chemokine levels inpatients with PV, serum levels of IP-10 and MCP-1 were significantly increasedin patients with BP compared with healthy controls. Furthermore, serum levelsof IP-10, MIG, MCP-1 and eotaxin in patients with BP increased significantlywith disease severity as determined by the area affected.Conclusions These observations suggest that an elaborately orchestrated network ofchemokines, especially MCP-1 and IP-10, contributes to the pathomechanism ofBP.

Bullous pemphigoid (BP) is an autoimmune subepidermal

blistering disease, usually occurring in the elderly, which is

characterized by large, tense blisters. The production of auto-

antibodies directed to a 180-kDa hemidesmosomal protein

(BP180) of the basement membrane zone is the initiating

event of the pathomechanism.1 Once the autoantibodies bind

to the basement membrane, a cascade of inflammatory events

occurs, resulting in subsequent blister formation at the dermo-

epidermal junction. During this process, cytokines and chem-

okines are considered to play crucial roles in inflammatory

cell recruitment, deposition and perpetuation.2 Therefore, clar-

ifying the serum and local levels of cytokines and chemokines

may help in understanding the immune dysregulation in the

pathomechanism.

It has been demonstrated that various cytokines including

interleukin (IL)-1b, IL-2, IL-4, IL-5, IL-6, IL-8, interferon

(IFN)-c and tumour necrosis factor-a are increased in sera

of patients with BP.3–8 By contrast, while serum levels of

several chemokines including eotaxin,9–11 RANTES (regulated

on activation, normal T-cell expressed and secreted),6 thy-

mus and activation-regulated chemokine (TARC)12 and mu-

cosae-associated epithelial chemokine13 have been reported,



the involvement of most chemokines in the disease remains

unknown. In this study, we utilized a recently developed

protein array system and assessed the serum levels of 10

chemokines: IFN-c-inducible protein-10 (IP-10), monokine

induced by IFN-c (MIG), macrophage inflammatory protein

(MIP)-1a, MIP-1b, monocyte chemoattractant protein

(MCP)-1, MCP-2, MCP-3, growth-regulated oncogene-a(GRO-a), eotaxin and RANTES.


Subjects

Serum samples obtained from 38 patients with BP (17 women

and 21 men; mean age 71Æ0 years), 16 patients with pemphi-

gus vulgaris (PV; nine women and seven men; mean age

55Æ7 years) and 17 healthy individuals (eight women and nine

men; mean age 61Æ7 years) were examined. BP and PV were

diagnosed on the basis of clinical features, and findings of skin

biopsy examination by light microscopy, direct immunofluo-

rescence and indirect immunofluorescence. Patients with BP

and PV had not been treated with oral corticosteroids, mino-

cycline/tetracycline or immunosuppressive drugs. As an index

of disease severity, the percentage of the body surface area

covered with skin lesions (oedema, erythema, blisters, ero-

sions) was estimated when the serum samples were obtained.

Patients with < 20%, 20–40% and > 40% of their body sur-

face affected were classified into group 1 (mild), 2 (inter-

mediate) and 3 (severe), respectively. Institutional review

board approval and informed consent from all patients and

controls were obtained. The samples collected were stored at

)80 �C until use.

Multiplex chemokine assay

The 10 chemokines measured simultaneously were IP-10, MIG,

MIP-1a, MIP-1b, RANTES, eotaxin, MCP-1, MCP-2, MCP-3 and

GRO-a. A sandwich immunoassay-based protein array system

(Biosource, Camarillo, CA, U.S.A.), which contains dyed micr-

ospheres conjugated with a monoclonal antibody specific for a

target protein, was used. Serum samples were thawed and run

in duplicate. Antibody-coupled beads were incubated with the

plasma sample after which they were incubated with biotinylat-

ed detection antibody before finally being incubated with

streptavidin-conjugated phycoerythrin. A broad sensitivity

range of standards (Biosource) ranging from 1Æ95 to

32 000 pg mL)1 was used for the quantification of a dynamic

wide range of chemokine concentrations and to provide the

greatest sensitivity. Standards included all the chemokines

assayed. This captured immunoassay was then read by the Bio-

plex array reader (Bio-Rad Laboratories, Hercules, CA, U.S.A.)

which uses Luminex fluorescent bead-based technology (Lum-

inex Corporation, Austin, TX, U.S.A.) with a flow-based dual

laser detector with real-time digital signal processing to facili-

tate the analysis of up to 100 different families of colour-coded

polystyrene beads and to allow multiple measurements of the

sample, ensuring the effective quantification of chemokines.

The concentrations of chemokines analysed in these assays were

quantified using a calibration or standard curve. A regression

analysis was performed to derive an equation that was then

used to predict the concentration of the unknown samples.


Statistical differences between measured values in patients with

BP, patients with PV and controls were analysed using the

Steel–Dwass’ multiple comparison test. Trends in the measured

values according to the affected area (disease severity) were

analysed by the two-sided Jonckheere–Terpstra test. Correla-

tions between each serum chemokine level were analysed

using Spearman’s correlation coefficient by rank test. P < 0Æ05

was considered statistically significant.

Results

Serum levels of 10 chemokines in patients with bullous

pemphigoid and pemphigus vulgaris

Concentrations of 10 chemokines, IP-10, MIG, MIP-1a,

MIP-1b, RANTES, eotaxin, MCP-1, MCP-2, MCP-3 and GRO-a,

were simultaneously measured in the sera from patients with

BP, patients with PV and healthy controls. Serum levels of IP-10

and MCP-1 were significantly increased in patients with BP

compared with healthy controls (P ¼ 0Æ011 for IP-10 and P ¼0Æ014 for MCP-1, Fig. 1). Serum levels of MIG, MIP-1a, MIP-1band eotaxin were also slightly higher in patients with BP than in

controls, although the differences were not significant. By con-

trast, all chemokines except for eotaxin showed comparable lev-

els between patients with PV and healthy controls (Fig. 1).

Serum eotaxin levels in patients with PV were higher than those

in controls although the difference was not significant (Fig. 1).

When chemokine levels higher than the mean + 2 SD of

the control serum samples were considered to be elevated, IP-

10 level was elevated in 39%, MIG in 29%, MIP-1a in 18%,

MIP-1b in 29%, eotaxin in 32% and MCP-1 in 36% of

patients with BP (Table 1). In patients with PV, 38% had ele-

vated eotaxin level, 25% had elevated MCP-1 level and 19%

had elevated IP-10 level. While an increased RANTES level was

observed in one control sample, all the other chemokine levels

in control samples were less than the mean + 2 SD levels. Col-

lectively, patients with BP showed more prominent abnormal-

ities in chemokine levels than patients with PV.

Positive correlation between serum levels of monokine

induced by interferon-c and macrophage inflammatory

protein-1 and disease severity

To assess the association of chemokine levels with disease

severity, patients with BP were classified into three groups by

their affected percentage body surface area. Significant disease

severity-related increases in serum levels of IP-10, MIG,

MCP-1 and eotaxin (P ¼ 0Æ008 for IP-10, P ¼ 0Æ0007 for



Chemokine profile in BP, H. Nakashima et al. 455

MIG, P < 0Æ0001 for MCP-1 and P ¼ 0Æ005 for eotaxin by the

Jonckheere–Terpstra test) were present (Fig. 2). By contrast,

serum levels of MIP-1a, MIP-1b, MCP-2, MCP-3, GRO-a and

RANTES showed no such tendency (data not shown). Thus,

serum levels of MIG, IP-10, MCP-1 and eotaxin appear to

reflect the disease severity.

Fig 1. Comparison of serum levels of interferon (IFN)-c-inducible protein-10 (IP-10), monokine induced by IFN-c (MIG), macrophage

inflammatory protein (MIP)-1a, MIP-1b, RANTES, eotaxin, monocyte chemoattractant protein (MCP)-1, MCP-2, MCP-3 and growth-regulated

oncogene-a (GRO-a) in patients with bullous pemphigoid (BP), pemphigus vulgaris (PV) and normal controls (Control). The measured values

from individual patients were plotted by dots. The boxes and whiskers indicate the median ± (or +) 25% and maximum/minimum values in

each group. A broken horizontal line indicates the cut-off value (mean + 2 SD of the control samples). *P < 0Æ05 by Steel–Dwass’ multiple

comparison test. Significant differences were also confirmed using the Mann–Whitney U-test with Bonferroni correction.



456 Chemokine profile in BP, H. Nakashima et al.

Correlation between serum chemokine levels

We further analysed the correlation between serum levels of

measured chemokines. Serum IP-10 levels in patients with BP

correlated positively with serum levels of MIG (r ¼ 0Æ385;

P ¼ 0Æ018), MCP-1 (r ¼ 0Æ328; P ¼ 0Æ041) and eotaxin (r ¼0Æ345; P ¼ 0Æ031). Serum eotaxin levels also correlated posi-

tively with serum concentration of MCP-1 (r ¼ 0Æ490; P ¼0Æ001). There was a strongly positive correlation between

serum levels of MIP-1a and MIP-1b (r ¼ 0Æ979, P < 0Æ0001).

Thus, several chemokines were concomitantly elevated, which

suggests that these cooperatively contribute to the disease

development.

Discussion

The aim of this study was to determine the profile of various

chemokines (IP-10, MIG, MIP-1a, MIP-1b, RANTES, eotaxin,

MCP-1, MCP2, MCP-3 and GRO-a) in sera from patients with

BP. While no chemokine was significantly elevated in sera

from patients with PV, patients with BP exhibited significantly

increased serum levels of IP-10 and MCP-1 compared with

healthy controls. Serum levels of IP-10, MIG, MCP-1 and eo-

taxin in patients with BP also showed a significant increase

with disease severity as determined by affected area. Addition-

ally, there were positive correlations between IP-10 and MIG,

IP-10 and eotaxin, IP-10 and MCP-1, eotaxin and MCP-1, and

MIP-1a and MIP-1b. Collectively, a variety of chemokines is

likely to contribute to the development of BP cooperatively.

Pathogenesis of the inflammatory skin disease is often rela-

ted to the T helper (Th) 1/Th2 balance. BP is considered to

be dominantly mediated by Th2 responses. Consistently, Kaki-

numa et al.12 have reported that serum levels of TARC, a major

Th2 chemokine, were elevated in patients with BP. While

there has been no literature describing Th1 chemokines of BP

sera, IP-10 and MIG measured in this study are considered as

chemoattractants of Th1 lymphocytes.14 Serum IP-10 levels

were significantly elevated in patients with BP compared with

controls. Also, there was a trend for MIG levels in sera from

patients with BP to increase in association with the affected

area. By contrast, serum levels of MIP-1a and MIP-1b, which

are also known as Th1 chemokines,15,16 showed neither signi-

ficant differences between patients with BP and controls nor

trends to increase with increasing disease severity. Positive

correlations between IP-10 and MIG levels suggest that IP-10

and MIG may work at adjacent parts of intricate inflammatory

cascades in the development of BP. Also, considering that in-

creased serum MIG levels were limited to patients with severe

disease, MIG may especially contribute to exacerbation of BP.

Alternatively, insufficient sensitivity of MIG detection in this

Table 1 Frequency of elevated samples of each chemokine

IP-10 MIG MIP-1a MIP-1b Eotaxin MCP-1 MCP-2 MCP-3 RANTES GRO-a

BP (n ¼ 38) 15 (39) 11 (29) 7 (18) 11 (29) 12 (32) 13 (36) 1 (3) 2 (6) 3 (8) 9 (24)

PV (n ¼ 16) 3 (19) 3 (19) 1 (6) 1 (6) 6 (38) 4 (25) 1 (6) 0 1 (6) 3 (19)Control (n ¼ 17) 0 0 0 0 0 0 0 0 1 (6) 0

BP, bullous pemphigoid; PV, pemphigus vulgaris; IP-10, interferon (IFN)-c-inducible protein-10; MIG, monokine induced by IFN-c; MIP,

macrophage inflammatory protein; MCP, monocyte chemoattractant protein; GRO-a, growth-regulated oncogene-a. Values are numbers(percentages) of the patients who showed elevated serum levels of each chemokine (> mean + 2 SD that of controls).

Fig 2. Serum levels of interferon (IFN)-c-inducible protein-10 (IP-10), monokine

induced by IFN-c (MIG), monocyte

chemoattractant protein (MCP)-1 and eotaxin

in patients with bullous pemphigoid classified

by affected area and in controls. Patients with

< 20%, 20–40% and > 40% of their body

surface affected were classified as group 1

(mild), 2 (intermediate) and 3 (severe),

respectively. The boxes and whiskers indicate

the median ± (or +) 25% and maximum/

minimum values in each group.




system may interfere with significant differences between

serum levels of patients with BP and controls. None the less,

serum IP-10 and MIG levels may also serve as indicators for

disease severity in BP.

Interestingly, T cells positive for CXC chemokine receptor

3, which is preferentially expressed on Th1-type cells, have

been demonstrated to exist beneath the bullae.17 Furthermore,

Budinger et al.18 have described that the autoreactive T cells,

which participate in the blister formation of BP and recognize

the extracellular domains of BP180, possess both Th2 and Th1

cytokines. Taken together, not only Th2 cells but also Th1

cells may be important for pathogenesis of BP.

BP lesions exhibit prominent eosinophil infiltration. Eotaxin

is considered to be a potent chemoattractant and activator of

eosinophils,19 and a main attractant of Th2 lymphocytes.20,21

Although serum eotaxin levels were not significantly increased

in patients with BP, serum eotaxin levels were associated with

disease severity and were also positively correlated with IP-10

and MCP-1. Correlation between eotaxin and these chemok-

ines implies that there may be a similar process for infiltration

of eosinophils and another subset of leucocytes including Th1

lymphocytes. Some previous reports have demonstrated eleva-

ted levels of eotaxin in sera and blister fluid as well as in-

creased expression in skin lesions of patients with BP.9–11 The

enhanced expression of eotaxin indicates that this chemokine

may also play an important role in the pathomechanism of

this autoimmune bullous skin disease by recruiting and activa-

ting eosinophils and lymphocytes.

MCP-1 is a chemoattractant and activator of monocytes,

lymphocytes and basophils, but lacks activity on neutrophils

and eosinophils.22,23 Serum MCP-1 levels were elevated in

patients with BP, and exhibited a tendency to increase accord-

ing to the disease severity. Furthermore, MCP-1 levels were

correlated positively with IP-10 and eotaxin, which act on

other types of inflammatory cells than those regulated by

MCP-1. Therefore, correlation between MCP-1 and these

chemokines suggests that there may be a similar process for

infiltration of different subsets of leucocytes utilizing a variety

of chemokines. While there is no literature regarding serum

MCP-1 levels in BP, some studies have demonstrated that

expression of CCR2, the chemokine receptor to which MCP-1

binds, is increased in chronically inflamed skin lesions of both

atopic dermatitis and psoriasis.24 Therefore, MCP-1 may play

an essential role in prolonged inflammation of BP lesions.

Serum RANTES levels did not show significant differences

between patients with BP and controls. RANTES is known as a

chemoattractant for monocytes, T lymphocytes and eosinoph-

ils.25 Our data are consistent with the findings of D’Auria et

al.6 Furthermore, they demonstrated that the RANTES concen-

trations in blister fluid were < 1/100 of those in the sera.

Therefore, RANTES may be rapidly degraded in the blister

fluid. Alternatively, RANTES derived from platelets may

increase the serum concentration.

In summary, the current study indicates that patients with

BP have a complex disorder of various chemokine levels in

their sera. This supports that various immune pathways are

involved in the pathogenesis of BP. It was also remarkable that

the chemokine profile of the sera from patients with BP

showed heterogeneity, which may suggest that the different

profiles correspond to distinct disease subsets or severity, or

may reflect different stages of the disease. Nonetheless, it

appears that especially MCP-1 and IP-10, then MIG and

eotaxin, play relatively more important roles than the other

chemokines in BP. Further profiling of immune mediators

including chemokines will elucidate the precise mechanism of

the disease development.

References

1 Labib RS, Anhalt GJ, Patel HP et al. Molecular heterogeneity of

the bullous pemphigoid antigens as detected by immunoblotting.J Immunol 1986; 136:1231–5.

2 Pastore S, Mascia F, Mariotti F et al. Chemokine networks in inflam-matory skin diseases. Eur J Dermatol 2004; 14:203–8.

3 Tamaki K, So K, Furuya T et al. Cytokine profile of patients withbullous pemphigoid. Br J Dermatol 1994; 130:128–9.

4 Sun CC, Wu J, Wong TT et al. High levels of interleukin-8, solubleCD4 and soluble CD8 in bullous pemphigoid blister fluid. The re-

lationship between local cytokine production and lesional T-cell ac-tivities. Br J Dermatol 2000; 143:1235–40.

5 Inaoki M, Takehara K. Increased serum levels of interleukin (IL)-5,IL-6 and IL-8 in bullous pemphigoid. J Dermatol Sci 1998; 16:152–7.

6 D’Auria L, Pietravalle M, Mastroianni A et al. IL-5 levels in theserum and blister fluid of patients with bullous pemphigoid: cor-

relations with eosinophil cationic protein, RANTES, IgE and diseaseseverity. Arch Dermatol Res 1998; 290:25–7.

7 D’Auria L, Mussi A, Bonifati C et al. Increased serum IL-6, TNF-

alpha and IL-10 levels in patients with bullous pemphigoid: rela-tionships with disease activity. J Eur Acad Dermatol Venereol 1999;

12:11–15.8 Ameglio F, D’Auria L, Bonifati C et al. Cytokine pattern in blister

fluid and serum of patients with bullous pemphigoid: relationshipswith disease intensity. Br J Dermatol 1998; 138:611–14.

9 Frezzolini A, Teofoli P, Cianchini G et al. Increased expression ofeotaxin and its specific receptor CCR3 in bullous pemphigoid. Eur J

Dermatol 2002; 12:27–31.10 Shrikhande M, Hunziker T, Braathen LR et al. Increased coexpres-

sion of eotaxin and interleukin 5 in bullous pemphigoid. Acta DermVenereol (Stockh) 2000; 80:277–80.

11 Wakugawa M, Nakamura K, Hino H et al. Elevated levels of eotaxinand interleukin-5 in blister fluid of bullous pemphigoid: correl-

ation with tissue eosinophilia. Br J Dermatol 2000; 143:112–16.12 Kakinuma T, Wakugawa M, Nakamura K et al. High level of

thymus and activation-regulated chemokine in blister fluid andsera of patients with bullous pemphigoid. Br J Dermatol 2003;

148:203–10.13 Kagami S, Kakinuma T, Saeki H et al. Increased serum CCL28 levels

in patients with atopic dermatitis, psoriasis vulgaris and bullouspemphigoid. J Invest Dermatol 2005; 124:1088–90.

14 Bonecchi R, Bianchi G, Bordignon PP et al. Differential expressionof chemokine receptors and chemotactic responsiveness of type 1

T helper cells (Th1s) and Th2s. J Exp Med 1998; 187:129–34.15 Koch AE, Kunkel SL, Harlow LA et al. Macrophage inflammatory

protein-1 alpha. A novel chemotactic cytokine for macrophages inrheumatoid arthritis. J Clin Invest 1994; 93:921–8.

16 Tanaka Y, Adams DH, Hubscher S et al. T-cell adhesion induced byproteoglycan-immobilized cytokine MIP-1 beta. Nature 1993;

361:79–82.



458 Chemokine profile in BP, H. Nakashima et al.

17 Sallusto F, Lenig D, Mackay CR et al. Flexible programs of chemo-kine receptor expression on human polarized T helper 1 and 2

lymphocytes. J Exp Med 1998; 187:875–83.18 Budinger L, Borradori L, Yee C et al. Identification and characteriza-

tion of autoreactive T cell responses to bullous pemphigoid antigen2 in patients and healthy controls. J Clin Invest 1998; 102:2082–9.

19 Ponath PD, Qin S, Post TW et al. Molecular cloning and characteri-zation of a human eotaxin receptor expressed selectively on eosin-

ophils. J Exp Med 1996; 183:2437–48.20 Sallusto F, Mackay CR, Lanzavecchia A. Selective expression of the

eotaxin receptor CCR3 by human T helper 2 cells. Science 1997;

277:2005–7.21 Gerber BO, Zanni MP, Uguccioni M et al. Functional expression of

the eotaxin receptor CCR3 in T lymphocytes co-localizing with e-osinophils. Curr Biol 1997; 7:836–43.

22 Yoshimura T, Robinson EA, Tanaka S et al. Purification and aminoacid analysis of two human glioma-derived monocyte chemoat-

tractants. J Exp Med 1989; 169:1449–59.23 Proost P, Wuyts A, van Damme J. Human monocyte chemotactic

proteins-2 and -3: structural and functional comparison with MCP-1. J Leukoc Biol 1996; 59:67–74.

24 Vestergaard C, Just H, Baumgartner Nielsen J et al. Expression ofCCR2 on monocytes and macrophages in chronically inflamed skin

in atopic dermatitis and psoriasis. Acta Derm Venereol (Stockh) 2004;84:353–8.

25 Alam R, Stafford S, Forsythe P et al. RANTES is a chemotactic and

activating factor for human eosinophils. J Immunol 1993;150:3442–8.





Resolution of inflammatory acne vulgaris may involveregulation of CD4+ T-cell responses to PropionibacteriumacnesH.E. Wilcox, M.D. Farrar, W.J. Cunliffe,* K.T. Holland and E. Ingham

Skin Research Centre, Institute of Molecular and Cellular Biology, Garstang Building, Faculty of Biological Sciences, University of Leeds, Leeds LS2 9JT, U.K.

*Department of Dermatology, Leeds General Infirmary, Leeds LS1 3EX, U.K.

CorrespondenceMark D. Farrar.


Accepted for publication11 June 2006

Key wordsacne vulgaris, heat shock proteins, limiting dilution

analysis, lymphocyte transformation,

Propionibacterium acnes


Summary

Background Propionibacterium acnes has been strongly implicated in inflammatory acne.However, its role in the disease is unclear. It has been hypothesized that animmune response to P. acnes and/or P. acnes heat shock proteins (HSPs) may playa role in the pathogenesis of inflammatory acne.Objectives To compare the cell-mediated immune response to P. acnes and HSPs inacne patients, nonacne controls and individuals with resolved acne.Methods The proliferative response of peripheral blood mononuclear cells (PBMC)from acne patients, resolved acne donors and healthy controls to P. acnes, P. acnesHSP60 and HSP70, and mycobacterial HSPs was assessed by lymphocyte trans-formation assay (LTA). The proliferative response of purified CD4+ T cells wasfurther analysed by limiting dilution analysis (LDA). Contingency tables (G-test)were used to analyse the proportion of individuals in each group showing apositive proliferative response for LTA or data fitting single-hit kinetics forLDA.Results Analysis of stimulation of PBMC with P. acnes, P. acnes HSP60 and HSP70 inthe LTA showed the proportion of positive responders to be independent of sub-ject group. However, the proportion of acne patients with a positive response tomycobacterial HSPs was significantly higher than those for the other subjectgroups. Analysis of LDA data showed the proportion of resolved donors withresponses to P. acnes fitting the single-hit kinetics model to be significantly lowerthan those of the other groups. There were no significant differences in responsesto other antigens.Conclusions The significantly lower proportion of resolved donors demonstrating asingle-hit kinetics response to P. acnes by LDA may represent negative regulationof the CD4+ T-cell response to P. acnes in these subjects.

Acne vulgaris is a multifactorial, spontaneously resolving pleo-

morphic skin disease, characterized by a variety of nonin-

flamed and inflamed skin lesions. Propionibacterium acnes has been

strongly implicated in the pathogenesis of the disease due to

the clinical improvement seen following a reduction in bacter-

ial population density with antibiotic therapy and the finding

that a lack of clinical response was associated with the pres-

ence of antibiotic-resistant propionibacteria.1,2 However, the

P. acnes factors that contribute to the disease are still not

known. The bacterium has been shown to have mitogenic as

well as antigenic activity3 and the inflammatory infiltrate

within early lesions is predominantly CD4+ T cells.4,5

Antibodies in sera of acne patients have been shown to recog-

nize a greater number of P. acnes protein antigens than those in

sera from healthy controls.6 Antigens specifically involved in

acne have not been identified but we have previously hypothes-

ized that heat shock proteins (HSPs) may have a role to play.7–9

These proteins are highly conserved throughout prokaryotes

and eukaryotes, with roles in protein folding and the cellular

stress response.10,11 HSPs are thought to contribute to inflam-

matory diseases through the induction of cytokine production

by a variety of cells and these proteins are frequently dominant

antigens in bacterial infections.12–14 Conditions within some

pilosebaceous follicles may be such that P. acnes are subject to

physiological stress and upregulate production of HSPs leading

to a cell-mediated inflammatory response to these proteins.

Studies of the cell-mediated immune (CMI) response to

P. acnes and P. acnes products have been carried out but the CMI



response to P. acnes HSPs has not been investigated. Previous

investigations of cellular immunity to P. acnes assayed by leuco-

cyte migration inhibition showed that only individuals with

severe acne could be considered as demonstrating CMI to the

bacterium.15 Investigations employing the lymphocyte trans-

formation assay (LTA) provided conflicting results with

one study demonstrating a significantly greater proliferative

response in acne patients compared with nonacne controls16

and a second study demonstrating a significantly greater

response only after resolution of the disease following a course

of therapy.17 Acne is most often a naturally self-limiting dis-

ease and important information could be gained by the inclu-

sion of a group of individuals with resolved acne in further

studies into immune responses to P. acnes and its products.

Therefore this study was undertaken to assess the specific CMI

response to P. acnes and P. acnes HSPs of cells from the peripheral

blood of acne patients, nonacne controls and individuals with

resolved acne. Responses to HSPs from mycobacteria, which

are closely related to propionibacteria but are not implicated in

acne, were also assessed.

Patients, materials and methods

Acne patients, control subjects and resolved donors

Three groups of subjects were studied—acne patients, nonacne

controls and resolved donors. Acne patients were outpatients of

the Department of Dermatology, Leeds General Infirmary, and

suffered from mild to severe acne as assessed by the Leeds tech-

nique.18 The nonacne controls were healthy individuals who

had no previous history of acne. The resolved donor group

consisted of individuals who had suffered from mild to moder-

ate acne at some period in their life but were now acne-free.

To determine the resolved acne status of these individuals,

volunteers were asked to complete a brief questionnaire

accompanied by a set of photographs representing acne

grades. None of the volunteers in any group were taking anti-

microbial therapy during or 6 weeks prior to the study. For

the LTA, the subject groups consisted of 12 patients (age

range 14–24 years), 10 controls (age range 20–24 years) and

10 resolved donors (age range 18–36 years). For the limiting

dilution analysis (LDA), the subject groups consisted of 11

patients (age range 14–23 years), 12 controls (age range 21–

29 years) and 13 resolved donors (age range 18–36 years).

Volunteers were asked if they had received the BCG vaccin-

ation. Not all of the samples from the donors were tested

against all antigens due to the availability of donors, the will-

ingness to donate the required volume of blood and the num-

ber of cells recovered from each blood sample (n ‡ 10).

Ethical approval for this project was obtained from the local

Ethics Committee.

Culture media

Total peripheral blood mononuclear cells (PBMC) were main-

tained in Medium A, which consisted of RPMI-1640 supple-

mented with 20 mmol L)1 HEPES, 60 lg mL)1 penicillin,

100 U mL)1 streptomycin, 2 mmol L)1L-glutamine and 10%

(v/v) fetal calf serum. Purified CD4+ T cells were maintained

in Medium B consisting of RPMI-1640 supplemented with

60 lg mL)1 penicillin, 100 U mL)1 streptomycin, 2 mmol L)1

L-glutamine and 7Æ5% (v/v) autologous serum. Purified

CD14+ cells were maintained in Medium C, which consisted

of Medium B supplemented with 25 U mL)1 recombinant

interleukin (IL)-2 (Roche Diagnostics Ltd., Lewes, U.K.).

PBMC and purified cells were grown at 37 �C in 5% (v/v)

CO2 in air. P. acnes was routinely grown in TYEG broth (10%

w/v tryptone, 5% w/v yeast extract, 1% w/v glucose) and on

reinforced clostridial agar (Oxoid, Basingstoke, U.K.) at 34 �Cin an atmosphere of 80% (v/v) N2, 10% (v/v) CO2, 10%

(v/v) H2.

Propionibacterium acnes and heat shock proteins

Propionibacterium acnes strain P37 was grown to the mid-expo-

nential phase. Bacterial cells were harvested, washed three

times in phosphate-buffered saline (PBS) and stored in 1%

(v/v) formalin in PBS at 4 �C. Prior to use, bacterial cells were

washed three times in PBS, resuspended at 1 · 109 cells mL)1

and sonicated to disaggregate clumps. P. acnes HSP60 and

HSP70 were overproduced in Escherichia coli and purified as pre-

viously described.8 Mycobacterium bovis BCG HSP65 was obtained

from the World Health Organization antigen bank, Germany.

Mycobacterium smegmatis HSP70 was obtained from Sigma (Poole,

U.K.).

Lymphocyte transformation assay

Peripheral venous blood (10 mL) was collected from each

subject into tubes containing 100 U preservative-free heparin

then diluted with an equal volume of Hank’s balanced salt

solution (HBSS) supplemented with 60 lg mL)1 penicillin

and 100 U mL)1 streptomycin. The diluted sample was

layered onto Lymphoprep� (Nycomed, Oslo, Norway) and

centrifuged at 800 g for 20 min. PBMC were harvested from

the Lymphoprep�–HBSS interface, washed three times with

antibiotic-supplemented HBSS and diluted to a final concentra-

tion of 1 · 106 cells mL)1 in culture medium A. Six replicate

volumes (100 lL) of PBMC were co-cultured with equal vol-

umes of formalized P. acnes (1 · 107 cells mL)1), each HSP

(0Æ5 lg mL)1), phytohaemagglutinin (PHA, 10 lg mL)1) or

Medium A only in a 96-well plate. Cells were cultured for

3 days with PHA, 6 and 9 days with P. acnes, and 7 and 9 days

with each HSP.

Culture periods and antigen amounts had been optimized in

preliminary experiments. High-activity [3H]thymidine (0Æ25 lCi)

was then added to each well and cells incubated for a

further 4 h. DNA was extracted from cells and harvested onto

glass fibre filter mats using a semi-automatic cell harvester.

The amount of [3H]thymidine incorporated into cellular DNA

was determined by scintillation counting. Results were

expressed as the stimulation index calculated as the mean



CD4+ T-cell regulation in acne resolution, H.E. Wilcox et al. 461

counts per minute (cpm) of the test divided by the mean cpm

of the negative control. A stimulation index of ‡ 3Æ0 was con-

sidered to be a positive response.

Limiting dilution analysis

Isolation of CD14+ monocytes and CD4+ T cells

CD14+ monocytes to be used as antigen-presenting cells

(APCs) and CD4+ T cells required for LDA were isolated from

PBMC by magnetic cell sorting (MACS�). PBMC were separ-

ated from whole blood as described for LTA, then monocytes

positively selected using CD14+ Microbeads (Miltenyi Biotec

Ltd, Bisley, U.K.) according to the manufacturer’s instructions.

The remaining cell fraction was then used for purification of

CD4+ T cells using the CD4+ T-cell isolation kit (Miltenyi

Biotec Ltd) according to the manufacturer’s instructions. Puri-

fied cells were quantified and viability assessed by trypan blue

dye exclusion.

Assay setup

Isolated CD4+ T cells were diluted to 1Æ25 · 106, 1Æ0 · 106,

7Æ5 · 105, 5Æ0 · 105, 2Æ5 · 105, 1Æ25 · 105 and 0 cells mL)1

in Medium B. Replicate (18, 24 or 30) 10 lL volumes of each

dilution were dispensed into wells of Terasaki microculture

plates. Isolated APCs were diluted to 3 · 105 cells mL)1 in

Medium C alone, Medium C with antigen (P. acnes or each

HSP), or Medium C with PHA, and 10-lL aliquots added to

the wells containing the CD4+ T cells. The Terasaki plates

were inverted and cultured for 3 days. High-activity

[3H]thymidine (0Æ25 lCi) was then added to each well and

incubation continued for 4 h. Microcultures were harvested

using a semi-automatic cell harvester and the amount of

[3H]thymidine incorporated into cellular DNA determined by

scintillation counting. The number of cultures positively

responding to each stimulus was calculated for each concen-

tration of CD4+ T cells. A culture was deemed to be positive

if the cpm of the stimulated CD4+ T cells was equal to, or

greater than, the mean cpm of the unstimulated CD4+ T cells

plus three times the SD. Antigen-specific CD4+ T-cell precur-

sor frequencies were then calculated.

Statistical analyses

The number of subjects whose cells responded to each anti-

gen in the LTA was analysed using the G-test (R · C and

2 · 2 tables).19 For the LDA, precursor frequency was esti-

mated by linear regression using the least squares method,

in which the negative slope of the line is equal to the

reciprocal value of the cell dose containing an average of

one precursor cell.20,21 A v2 test for goodness of fit was

performed to determine whether the data conformed to the

single-hit model.20 A P-value of > 0Æ05 indicated an accept-

able fit to the model. For presentation of the data, the

negative logarithm of the fraction of nonresponding cultures

was plotted against the number of CD4+ T cells per

culture.

Results

Peripheral blood mononuclear cell proliferative response

to Propionibacterium acnes and heat shock proteins

Following stimulation of PBMC with mycobacterial HSPs,

58Æ3% of acne patients, 10Æ0% of resolved donors and 18Æ2%

of controls responded to M. bovis BCG HSP65; 66Æ7% of acne

patients, 20Æ0% of resolved donors and 9Æ1% of controls

responded to M. smegmatis HSP70 (Fig. 1). All patients, resolved

donors and controls had a positive response to PHA (data not

shown). The proportion of positive responders to each antigen

within each group was analysed using the G-test for independ-

ence. The number of subjects whose cells responded to each

antigen was not independent of whether the subject was a

patient, resolved donor or a control. There was a significantly

(P < 0Æ05) higher proportion of positive responders to HSP65

and HSP70 in the acne patient group compared with the

respective resolved acne donor group or control group for that

antigen. The BCG vaccination status of each volunteer was

ascertained. There was no significant difference in the propor-

tion of BCG-vaccinated individuals between the acne, resolved

and control groups (data not shown).

A high proportion of individuals in each group had a signi-

ficant PBMC response to P. acnes (66Æ7% of acne patients,

80Æ0% of resolved donors, 45Æ5% of controls). However, only

a small percentage responded to either P. acnes HSP60 (10Æ0%

of acne patients, no resolved donors or controls) or P. acnes

HSP70 (10Æ0% of acne patients, 20Æ0% of resolved donors,

10Æ0% of controls; Fig. 1). G-test analysis showed the propor-

tion of positive responders to P. acnes, P. acnes HSP60 or P. acnes

Fig 1. Proportion of acne patients, resolved donors and control

individuals with positive peripheral blood mononuclear cell (PBMC)

proliferative response to each antigen. PBMC from each donor were

cultured with antigen for up to 9 days and proliferation measured

by [3H]thymidine incorporation. A stimulation index ‡ 3Æ0 was

considered positive. Data were analysed by G-test of independence.

*Indicates a significant difference (P £ 0Æ05) in the proportion of

positive responders. Mb, Mycobacterium bovis BCG; Ms, Mycobacterium

smegmatis; Pa, Propionibacterium acnes; HSP, heat shock protein.



462 CD4+ T-cell regulation in acne resolution, H.E. Wilcox et al.

HSP70 within each group to be independent of whether the

subject was an acne patient, resolved donor or control.

Limiting dilution analysis

Mean precursor frequencies were calculated for those data that

conformed to the model of single-hit kinetics. For responses

to P. acnes, mean precursor cell frequencies for acne patients,

controls and resolved donors were one in 112890, 133836

and 135208, respectively. For responses to P. acnes HSP60, fre-

quencies were one in 124774, 510788 and 87884, respective-

ly. Frequencies for responses to P. acnes HSP70 were one in

147036, 262036 and 144243, respectively. For responses to

M. bovis HSP65, frequencies were one in 266368, 89683 and

116037, respectively. Finally, for responses to M. smegmatis

HSP70, frequencies were one in 93497, 244318 and 213560,

respectively. The mean precursor frequencies could not be

analysed for any difference between groups as the number of

donors in each group that resulted in data that fitted the

single-hit kinetics model was too small.

As a failure of the data to fit the single-hit kinetics model

could indicate regulation of the CD4+ T-cell responses, the

proportion of subjects with data that fitted the model was

recorded for each antigen (Fig. 2) and the results analysed

with the G-test of independence. The proportion of subjects

with data that fitted the single-hit kinetics model following

LDA of P. acnes-specific CD4+ T cells was not independent of

subject group, with a significantly lower number of resolved

donors (8%) having LDA data that fitted the single-hit kinetics

model compared with acne patients (36%) and controls

(50%). In contrast, the proportion of acne patients, controls

and resolved donors with data that fitted the single-hit kinetics

model following LDA of CD4+ T-cells specific for P. acnes

HSP60, P. acnes HSP70, M. bovis BCG HSP65 or M. smegmatis

HSP70 was independent of subject group. As with the LTA,

there was no significant difference in the proportion of BCG-

vaccinated individuals between the acne, resolved and control

groups (data not shown).

Owing to the increased complexity of the curves generated

by the data that did not conform to the single-hit kinetics

model, interpretation of the results was difficult. However, a

number of general patterns emerged. Several samples had data

points that appeared to form a nonlinear negative slope

(Fig. 3a) with the data points forming a rough curve as they

decreased in value. In other cases, sample data appeared to

result in a dominant positive slope (Fig. 3b). The negative

section visible on these charts occurred only between two data

points, due to the inclusion of the zero cell concentration at

the origin. Further dilutions between these points may have

revealed more of a negative slope. Other sample data appeared

to have a distinct negative slope at the lower cell concentra-

tions and a positive slope at the higher cell concentrations, as

illustrated in Figure 3c. Finally, some of the samples had data

points that formed a jagged (zigzag) pattern between adjacent

points or over a number of points as illustrated in Figure 3d.

Discussion

The early developing inflammatory lesions in acne vulgaris are

characterized by a predominantly CD4+ T-lymphocytic cellular

infiltrate and there is evidence to suggest that this represents a

type IV hypersensitivity reaction.4,5,22 Epitopes derived from

P. acnes components are thought to be the main driver of this

inflammatory response although the identity of these antigens

is unknown. We have previously hypothesized that HSPs may

have a role due to their association with other inflammatory

conditions.7–9 Therefore, this work aimed to identify differ-

ences in the proliferation of PBMC and isolated CD4+ T cells

from acne patients, controls and resolved individuals in

response to P. acnes, P. acnes HSPs and the related mycobacterial

HSPs. For both LTA and LDA, the BCG vaccination status was

compared between groups as this may have affected the results

of these assays. No significant difference was found in the vac-

cination status of each group.

Data from the LTAs showed that the proportion of acne

patients with a positive response to M. bovis BCG HSP65 and

M. smegmatis HSP70 was significantly higher than that for con-

trols or resolved donors. There was no significant difference

in the proportion of individuals with a positive response to

P. acnes or P. acnes HSPs between the three donor groups. The

increased proliferative response to mycobacterial HSPs may be

as a result of the presentation of immunoreactive species-

specific epitopes not present in the P. acnes HSPs or a lack of

regulation to shared epitopes in acne patients. P. acnes and

mycobacterial HSPs demonstrate 45–50% identity to their

human homologues at the amino acid level. The homology

between the bacterial HSPs is greater with an identity of 78%

between the P. acnes and mycobacterial HSP60 and 67% for

HSP70.8 Species-specific regions of these proteins may there-

fore be responsible for the differences in responses.

Fig 2. Proportion of acne patients, resolved donors and control

individuals with limiting dilution analysis data fitting the single-hit

kinetics model for determination of precursor frequency to each

antigen. Purified CD4+ T cells were cultured for 3 days with antigen,

proliferation measured by [3H]thymidine incorporation and precursor

frequencies calculated. Data were analysed by G-test of independence.

*Indicates a significant difference (P £ 0Æ05) in the proportion of

positive responders. Mb, Mycobacterium bovis BCG; Ms, Mycobacterium

smegmatis; Pa, Propionibacterium acnes; HSP, heat shock protein.




These results do not necessarily indicate that P. acnes HSPs

do not have a role to play in acne. Individuals should have the

potential to respond to P. acnes-specific epitopes of the HSPs

but this response appeared to be regulated in all three groups

of individuals in this study. Regulation of the response to

P. acnes HSPs in resolved individuals and controls may explain

their lack of acne. Regulatory cells may inhibit inflammatory

immune responses to P. acnes HSP60 and may represent per-

ipheral tolerance to this antigen as a mechanism of preventing

cross-reactivity with human HSPs.23 The low proportion of

patients with acne responding to P. acnes HSPs may have been

due to a degree of regulation having already developed in

these individuals who were at the upper end of the age range

(when acne is most likely to resolve). A further explanation

for these results is that this study used cells from peripheral

blood and the regulation of responses locally in the skin may

occur only at a later stage in the disease process. Regulation

could be detected in peripheral blood prior to resolution of

the disease. Therefore, further studies are required using

younger volunteers, who have only recently developed acne,

to determine if P. acnes HSPs have a role in the disease.

Analysis of T-cell responses by LDA allows the frequency

of T cells responding to a particular antigen to be deter-

mined and can also reveal the activity of regulatory cells.

The proliferative response of purified CD4+ T cells from

individuals in each subject group to P. acnes, P. acnes HSPs

and mycobacterial HSPs was tested by LDA. Mean precursor

frequencies could be calculated only for those data that fit-

ted the single-hit kinetics model. From the results it can be

seen that the proportion of samples that did fit this model

was small. Therefore, differences between groups could not

be analysed. The failure of data to fit single-hit kinetics can

indicate regulatory effects within the cell population being

tested. Therefore, the proportion of subjects in each group

that demonstrated single-hit kinetics was analysed by the

G-test as for LTA. The finding that a significantly lower

number of resolved donors had LDA responses to P. acnes

that fitted single-hit kinetics compared with other groups

could indicate the presence of a regulatory subset of CD4+

T cells within the purified CD4+ T-cell population. This is

an important finding and gives support to our hypothesis

that downregulation of the CD4+ T-cell-mediated immune

(a) (b)

(d)(c)

Fig 3. Examples of titration curves obtained following limiting dilution analysis of CD4+ T cells with P. acnes and heat shock proteins. Dilutions

of purified CD4+ T cells were stimulated with antigen and proliferation was measured by [3H]thymidine incorporation. Graphs are the negative

logarithm of the fraction of nonresponding cultures plotted against the number of CD4+ T cells per culture. (a) Nonlinear negative slope,

(b) dominant positive slope, (c) V-shaped with negative slope at the lower cell concentrations and positive slope at the higher concentrations,

(d) zigzag. The linear regression line (dashed line) was calculated by the least squares method and the goodness of fit of the data to the single-hit

kinetics model was assessed using the v2 test.



464 CD4+ T-cell regulation in acne resolution, H.E. Wilcox et al.

response to P. acnes is associated with the spontaneous reso-

lution of acne.

Differences in the shape of LDA titration curves is not

unique to this study. Zigzag, V-shaped, saw-tooth curves and

other descriptions have been reported.20,24,25 Concave nega-

tive slopes (Fig. 3a) have been described as multi-hit and

reflect the need for an interaction between two or more cells

of the same type in order to produce a measurable positive

response. Positive slopes can result from interaction between a

high frequency of stimulatory cells and a lower frequency of

inhibitory (regulatory) cells as demonstrated in Figure 3b. The

presence of regulatory cells in the cell populations tested could

therefore account for these nonlinear plots. No particular pat-

tern was dominant and the data indicate that the response to

P. acnes is complex and varies between individuals.

Further analysis of CD4+ T-cell responses to P. acnes is

required to define in greater detail the proliferative response

and the involvement of different T-cell subsets. If the presence

of regulatory T cells was confirmed to have a role in acne

resolution then this would have important clinical implica-

tions. The identification of those P. acnes antigens important in

mediating the inflammatory response, and to which this

response is regulated in resolution, would represent a signifi-

cant step in the development of an immunotherapy for acne

whereby the induction of tolerance to the antigen(s) could be

a therapeutic option.

Acknowledgments

This study was funded by the Leeds Foundation for Dermato-

logical Research.

References

1 Thiboutot D. Acne: an overview of clinical research findings.Adv Clin Res 1997; 15:97–109.

2 Eady EA, Cove JH, Holland KT, Cunliffe WJ. Erythromycin-resistantpropionibacteria in antibiotic-treated acne patients: association with

therapeutic failure. Br J Dermatol 1989; 121:51–7.3 Jappe U, Ingham E, Henwood J, Holland KT. Propionibacterium acnes

and inflammation in acne: P. acnes has T-cell mitogenic activity. Br JDermatol 2002; 146:202–9.

4 Norris JF, Cunliffe WJ. A histological and immunocytochemicalstudy of early acne lesions. Br J Dermatol 1988; 118:651–9.

5 Jeremy AHT, Holland DB, Roberts SG et al. Inflammatory events areinvolved in acne lesion initiation. J Invest Dermatol 2003; 121:20–7.

6 Holland KT, Holland DB, Cunliffe WJ, Cutcliffe AG. Detectionof Propionibacterium acnes polypeptides which have stimulated an

immune response in acne patients but not in normal individuals.Exp Dermatol 1993; 2:12–16.

7 Ingham E. The immunology of Propionibacterium acnes and acne.Curr Opin Infect Dis 1999; 12:191–7.

8 Farrar MD, Ingham E, Holland KT. Heat shock proteins and inflam-matory acne vulgaris: molecular cloning, overexpression and puri-

fication of a Propionibacterium acnes GroEL and DnaK homologue. FEMSMicrobiol Lett 2000; 191:183–6.

9 Farrar MD, Ingham E. Acne: inflammation. Clin Dermatol 2004;22:380–4.

10 Georgopoulos C, Welch WS. Role of the major heat shock proteinsas molecular chaperones. Annu Rev Cell Biol 1993; 9:601–34.

11 Lindquist S, Craig EA. The heat-shock proteins. Annu Rev Genet 1988;

22:631–7.12 Galdiero M, Cipollaro de l’Ero G, Marcatili A. Cytokine and adhe-

sion molecule expression in human monocytes and endothelialcells stimulated with bacterial heat shock proteins. Infect Immun

1997; 65:699–707.13 Marcatili A, Cipollaro de l’Ero G, Galdiero M et al. TNF-a, IL-1a,

IL-6 and ICAM-1 expression in human keratinocytes stimulated invitro with Escherichia coli heat-shock proteins. Microbiology 1997;

143:45–53.14 Kaufmann SHE, Vath U, Thole JER et al. Enumeration of T-cells

reactive with Mycobacterium tuberculosis organisms and specific for therecombinant mycobacterial 64 kilodalton protein. Eur J Immunol

1987; 17:351–7.15 Gowland G, Ward RM, Holland KT, Cunliffe WJ. Cellular immun-

ity to P. acnes in the normal population and patients with acnevulgaris. Br J Dermatol 1978; 99:43–7.

16 Puhvel SM, Amirian D, Weintraub J, Reisner RM. Lymphocytetransformation in subjects with nodulo cystic acne. Br J Dermatol

1977; 97:205–11.17 Karvonen SL, Rasanen L, Cunliffe WJ et al. Delayed hypersensitivity

to Propionibacterium acnes in patients with severe nodular acne andacne fulminans. Dermatology 1994; 189:344–9.

18 Burke BM, Cunliffe WJ. The assessment of acne vulgaris: the Leedstechnique. Br J Dermatol 1984; 111:83–92.

19 Sokal RR, Rohlf FJ. Tests of independence. In: Biometry: The Principlesand Practice of Statistics in Biological Research (Sokal RR, Rohlf FJ, eds),

3rd edn. New York: WH Freeman and Co., 1995.20 Lefkovits I, Waldmann H. Limiting Dilution Analysis of Cells of the Immune

System. Oxford: Oxford University Press, 1999.21 Taswell C. Limiting dilution assays for the determination of immu-

nocompetent cell frequencies. I. Data analysis. J Immunol 1981;126:1614–9.

22 Layton AM, Morris C, Cunliffe WJ, Ingham E. Immunohistochemi-

cal investigation of evolving inflammation in lesions of acnevulgaris. Exp Dermatol 1998; 7:191–7.

23 Hammerling GJ, Schonrich G, Ferber I, Arnold B. Peripheral toler-ance as a multi-step mechanism. Immunol Rev 1993; 133:93–104.

24 Dozmorov IM, Lutsenko GV, Sidorov LA, Miller RA. Analysis ofcellular interactions in limiting dilution cultures. J Immunol Methods

1996; 189:183–96.25 Dozmorov IM, Eisenbraun MD, Lefkovits I. Limiting dilution analy-

sis: from frequencies to cellular interactions. Immunol Today 2000;21:15–18.





Tight junction components occludin, ZO-1, and claudin-1,-4 and -5 in active and healing psoriasisS. Peltonen,* J. Riehokainen,*� K. Pummi*� and J. Peltonen*�

*Department of Dermatology, University of Turku, PL 52, 20521 Turku, Finland

Departments of �Medical Biochemistry and �Anatomy, University of Turku, Kiinamyllynkatu 10, 20520 Turku, Finland

CorrespondenceSirkku Peltonen.



Key wordsclaudin, involucrin, occludin, psoriasis, tight

junction, ZO-1


Summary

Background Cells of the granular layer are interconnected by tight junctions (TJs) innormal epidermis. The structural proteins of epidermal TJs include occludin, ZO-1,and claudin-1 and -4.Objectives Our aim was to correlate the expression of TJ components with keratino-cyte differentiation using psoriasis as a model of premature keratinization.Methods The distribution of TJ proteins was evaluated in the skin of nine patientswith psoriasis. Punch biopsies were taken from perilesional skin, from active pso-riasis plaques, and from healed, previously lesional locations. The punch biopsieswere analysed using indirect immunolabelling for ZO-1, occludin and claudin-1,-4 and -5. In addition, epidermal samples were analysed by reverse transcrip-tion–polymerase chain reaction for claudin-1, -4 and -5 mRNAs.Results Claudin-5 was localized to the granular cell layers of normal control skinas well as perilesional and lesional psoriatic epidermis. This was unexpected, asprevious studies have not detected claudin-5 in the epidermis. Occludin andZO-1 were expressed in the granular cell layer in psoriatic perilesional epidermis.In the psoriasis plaques, ZO-1 and occludin were detected in a wider zoneextending from the granular layer to the middle spinous cell layers. In healedpsoriasis plaques, the expression of occludin and ZO-1 resumed a normal-lookingprofile, being restricted to the upper epidermis only. Claudin-1 and -4 did notshow marked changes in psoriasis compared with normal skin.Conclusions The results demonstrate claudin-5 in normal epidermis and psoriaticskin, and abnormal distribution of occludin and ZO-1 in psoriasis plaques. Clin-ical healing of aberrant keratinization is associated with restoration of the normaldistribution of occludin, ZO-1 and also involucrin.

Molecular characterization of numerous tight junction (TJ)

proteins1–3 and the availability of antibodies to various TJ

components have focused increasing attention on TJs in var-

ious tissues, and have resulted in recognition of TJs also in

stratified epithelia such as the epidermis.4–7 Subsequently, TJs

have been studied in human diseases, including those affecting

the epidermis.5,8–12 Immunoelectron microscopy and immu-

nohistochemistry have shown that TJs are located in the gran-

ular cell layer and that they contain occludin, ZO-1, ZO-2 and

claudin-1 and -4.4–7 The above findings, and results obtained

from transgenic mice, have prompted re-evaluation of the role

of TJs in the regulation of the paracellular diffusion barrier of

the epidermis.13–18 The role of TJs in the formation of the

epidermal barrier has been emphasized by experiments using

transgenic mouse models for claudin-1 and claudin-6.14,15

The results showed that claudin-1 knockout mice and mice

overexpressing claudin-6 had epidermal barrier dysfunc-

tion.14,15 On the other hand, occludin knockout mice did not

show signs of affected barrier function or of morphological

alterations in the ultrastructure of intestinal TJs, indicating a

crucial function for claudins in the function of TJ seals.17 The

role of claudins in the assembly of TJs has further been under-

lined by studies showing that ZO-1–3 can be recruited to

claudin-based TJs in the absence of occludin.19

We and others have previously described altered localization

of TJ proteins in conditions where the horny layer is missing

or abnormal, such as in developing skin, healing wounds, or

psoriasis.5,6,10–13 Observations on psoriatic skin or the hyper-

proliferative zone at the border of healing blisters also indicate

that aberrant keratinocyte differentiation is linked with chan-

ges in the expression profile of TJ components. More specific-

ally, ZO-1 and occludin have been detected in the acanthotic



spinous cell layers in psoriasis lesions.5,11 Previous results have

also suggested that the expression of occludin may be used as

a differentiation marker for keratinocytes.11 To date, the

expression of claudin-1, -4 and -5 has not been analysed in

psoriasis. The present work adds to our knowledge of psoriasis

and TJs by extending the study to healing of psoriasis and

expression of claudin-1, -4 and -5. This study also correlates

the expression of TJ proteins with the expression of involu-

crin, another differentiation marker.


Biopsy samples

Nine patients with psoriasis, five women and four men, age

range 35–67 years, were included in the study. The study

took place at the Department of Dermatology, Turku Univer-

sity Central Hospital, with the approval of the joint Ethic

Committee of the University of Turku and the Turku Univer-

sity Hospital, and with written consent of the patients. All

patients had an activation of previously stable plaque-type pso-

riasis. Punch biopsies, 3 mm in diameter, were taken from:

(i) a typical scaly psoriasis plaque, (ii) perilesional skin at

~7 mm distance from the edge of the biopsied lesion, and

(iii) the location of the original lesion after therapy, 6–

8 weeks later. The treatment included potent corticosteroid

cream for 1–2 weeks, followed by narrowband (wavelength

311 nm) ultraviolet B phototherapy for 15–18 sessions during

5–6 weeks. If the plaque showed no scaling even though it

displayed slight erythema or postinflammatory pigmentation it

was considered as a healed plaque. Three control samples were

obtained from otherwise healthy patients undergoing plastic

surgery. The skin samples were frozen in liquid nitrogen and

stored at )70 �C.

Primary antibodies

The following primary antibodies were used: affinity-purified

rabbit polyclonal antibodies to human claudin-1 (51–9000),

claudin-5 (34–1600), ZO-1 (61–7300) and occludin (71–

1500), and mouse monoclonal antibodies to human claudin-4

(32–9400), claudin-5 (35–2500) and ZO-1 (33–9100), all

from Zymed Laboratories (South San Francisco, CA, U.S.A.).

Mouse monoclonal antibody to involucrin (MS-126-P0) was

obtained from NeoMarkers (Fremont, CA, U.S.A.). Nonimmu-

nized mouse (2025) and rabbit (2027) IgG were purchased

from Santa Crutz Biotechnology (Santa Cruz, CA, U.S.A.).

Indirect immunofluorescence labelling

Punch biopsy samples were cut into 7-lm cryosections on

silanated glass slides and fixed in 100% methyl alcohol at

)20 �C for 10 min. To prevent nonspecific binding, the sam-

ples were preincubated in 1% bovine serum albumin (BSA) in

phosphate-buffered saline (PBS) for 15 min. Antibodies were

diluted in 1% BSA in PBS, and incubated on the samples for

20 h at 4 �C. Following six 5-min washes in PBS, the samples

were incubated with secondary antibodies for 1 h at 20 �C.

The samples were washed in PBS and mounted with Glycergel

(Dako, Glostrup, Denmark). Secondary antibodies for indirect

immunofluorescence (IIF) were tetramethylrhodamine isothio-

cyanate (TRITC)-conjugated swine antirabbit (R0156) and

rabbit antimouse IgG (R0270) from Dako, or Alexa Fluor

488-conjugated goat antimouse IgG (A-11029) from Molecu-

lar Probes (Eugene, OR, U.S.A.). In double labellings, involu-

crin antibody was mixed with polyclonal antibodies to ZO-1

or occludin. After the incubation and washes, TRITC-conju-

gated swine antirabbit IgG was mixed with Alexa Fluor 488-

conjugated goat antimouse IgG (A-11029). In control immu-

noreactions, primary antibodies were replaced with 1% BSA in

PBS or nonimmunized mouse or rabbit IgG.

Microscopy

IIF labellings were examined and photographed using a Leica

DMRB fluorescence microscope, equipped with a Leica DC

300F camera and Leica IM50 (version 1.20) software (Leica

Microsystems AG, Heerbrugg, Switzerland). Resolutions of

images were 1950 · 1545 pixels and images were saved in

tiff format. Confocal laser scanning microscopy was carried

out using a Zeiss LSM 510 META confocal microscope

equipped with argon ion and helium–neon lasers (Zeiss, Jena,

Germany) and LSM 3.0 software. The objective magnification

used was · 20 (numeric aperture 0Æ5), · 40 (oil immersion,

numeric aperture 1Æ3) or · 63 (oil immersion, numeric aper-

ture 1Æ4). For excitation of TRITC the 543 line was used. The

resolutions of original images were 1024 · 1024 pixels and

images were saved in tiff format and assembled into panels

using Adobe Photoshop or Corel Draw software.

RNA isolation and reverse transcription–polymerase

chain reaction

To isolate RNA from whole skin, a 4-mm punch biopsy was

taken from the upper arm of an otherwise healthy woman.

RNA was isolated by acid guanidinium isothiocyanate–chloro-

form–phenol extraction.20 Epidermal samples for RNA isola-

tion were obtained from three healthy controls (age range

35–49 years). In analogy, five psoriasis plaques from five

patients with psoriasis were chosen for sampling. Four of the

patients (men aged 31–45 years) had plaque-type psoriasis

that had been untreated for several months. The fifth patient

(a woman aged 41 years) had guttate-type psoriasis. The

extensor aspect of the lower arm close to the elbow, which is

a typical location for psoriasis plaques, was first wiped with

70% ethanol. Then the normal skin of the control individuals

and a psoriasis plaque of patients with psoriasis were gently

scraped using a disposable curette, carefully avoiding bleeding,

on an area of about 10 cm2. The corneal layer and the epider-

mal keratinocytes that were detached by scraping were imme-

diately submerged from the curette into the RNAlater RNA

Stabilization Reagent (76104; Qiagen, Valencia, CA, U.S.A.).



Tight junctions in psoriatic skin, S. Peltonen et al. 467

RNA was then isolated using RNeasy Mini Kit (74104;

Qiagen) according to the protocol provided by the manufac-

turer. The RNA was transcribed into single-stranded DNA in a

20-lL reaction containing 200 U of SuperScript II Reverse

Transcriptase (18064–022; Invitrogen, Carlsbad, CA, U.S.A.),

1 lmol L)1 of each of the four deoxynucleotides, Random

Hexamers (C1181; Promega, San Luis Obispo, CA, U.S.A.),

20 U of RNasin� (N2111; Promega) and 0Æ01 mol L)1 di-

thiothreitol in the first-strand buffer. The reaction was incuba-

ted at 42 �C according to the first-strand synthesis protocol

provided by the manufacturer. The cDNA yielded by reverse

transcription (RT) reactions was used as a template for poly-

merase chain reaction (PCR). PCRs for claudin-1, -4 and -5

and glyceraldehyde-3-phosphate dehydrogenase were per-

formed using a MultiGene-12 RT–PCR profiling kit (PH-083B;

SuperArray Bioscience Co., Frederick, MD, U.S.A.) detecting

claudin-1–11, according to the protocol provided by the

manufacturer.

Results

Immunohistochemical mapping of five TJ proteins, claudin-1,

-4, -5, ZO-1 and occludin, was performed in skin samples of

nine patients with psoriasis and three control samples. Punch

biopsies of apparently normal perilesional skin, psoriasis

plaque, and the same psoriasis plaque in the clinically healed

phase, were analysed. In perilesional epidermis the localization

of occludin, ZO-1 and claudin-1 and -4 corresponded to that

described earlier for normal skin (Figs 1, 2).4–6 Specifically,

occludin and ZO-1 were located to the granular cell layer

(Fig. 2a,g) while claudin-1 was distributed to all living cell

layers, the basal cell layer being only faintly labelled (Fig. 1a).

Claudin-4 was mainly localized to the upper epidermis

(Fig. 1d). Unexpectedly, all samples from patients with psori-

asis, both lesional and perilesional, displayed fine intercellular

labelling for claudin-5 in the cell–cell contact zones of the

granular cell layer (Fig. 1g–i). To confirm the existence of

claudin-5 in epidermis, the same psoriasis samples were im-

munolabelled for polyclonal rabbit antibody to claudin-5. This

resulted in similar findings as obtained with the monoclonal

antibody (Fig. 3a). The parakeratotic foci and their surround-

ings were especially intensely labelled (Fig. 3a,b). The pres-

ence of claudin-5 in psoriatic epidermis prompted us to

perform further labelling for skin samples of healthy patients

with monoclonal and polyclonal antibodies to claudin-5. Con-

focal laser scanning microscopy revealed thin lines of positive

labelling for claudin-5 in the granular cell layer (Fig. 3c,d). As

described earlier,6,19 claudin-5 was also detected in the dermal

blood vessel walls in psoriasis and control samples. Control

immunoreactions of normal skin and psoriasis samples with

isotype-matched mouse and rabbit antibodies did not show

labelling in intercellular contact zones in normal or psoriatic

epidermis (Fig. 3e–h).

Further to confirm the expression of claudin-5 in psoriatic

epidermis, RT–PCR was performed using RNA isolated

from normal skin, normal epidermis and psoriatic epidermis.

(a) (b) (c)

(d) (e) (f)

(g) (h) (i)

Perilesional skin

Claudin1

Claudin4

Claudin5

Psoriasis plaque Healed psoriasis plaque

Fig 1. Indirect immunofluorescence labelling

of perilesional skin (left column), psoriasis

plaque (middle column) and healed psoriasis

plaque (right column) for claudin-1, -4 and

-5. (a–c) Claudin-1 labels the granular and

spinous cell layers, the basal cell layer being

mostly negative in all samples. (d–f) Claudin-

4 is mainly expressed in the granular cell

layer but is detectable also in spinous cell

layers. (g–i) Claudin-5 is expressed only in

the granular cell layer in perilesional, lesional

(arrows) and healed psoriatic epidermis, in

addition to being present in dermal blood

vessel walls. Scale bars ¼ 50 lm.



468 Tight junctions in psoriatic skin, S. Peltonen et al.

A commercial kit designed for the detection of claudins was

utilized for PCR and showed expression of claudin-1 and -4

in all samples. Claudin-5 was detected in the whole skin sam-

ple and three of the five psoriasis samples, but not in any of

the control epidermis samples (Fig. 4).

In the psoriasis plaque, the distribution of claudin-1 was

extended to all living cell layers, excluding the basal layer,

thus resembling the distribution profile described for healthy

skin (Fig. 1b). Anticlaudin-4 antibody labelled more numer-

ous spinous cell layers in psoriasis plaque than in normal skin

(Fig. 1e). In analogy to our previous results, ZO-1 and occlu-

din in psoriasis plaques displayed a wider distribution than in

control or perilesional epidermis. Specifically, ZO-1 and occlu-

din were detected in psoriasis plaques in the granular and in

the middle spinous cell layers (Fig. 2b,h).

The expression of involucrin, an early differentiation mar-

ker, was compared with the localization of ZO-1 and occlu-

din. Double immunolabellings showed that in perilesional

epidermis, involucrin, ZO-1 and occludin colocalized in the

granular cell layer (Fig. 2d,j). The expression of involucrin by

granular cells is in agreement with previous findings showing

that involucrin can be expressed by granular cells in psoriasis.

The results showed that, in psoriasis plaques, involucrin dis-

played a more extended shift of expression to the spinous

layer than ZO-1 or occludin. In more detail, the expression of

ZO-1 and occludin was localized to the middle spinous cell

layers while involucrin was detected also in the lower spinous

cell layers (Fig. 2e,k).

Healing of psoriasis was associated with the restoration of

the expression of involucrin, ZO-1 and occludin exclusively in

the granular cell layers (Fig. 2f,l). The expression patterns of

claudin-1, -4 and -5 remained essentially the same in the

healed and in the active psoriasis plaques (Fig. 1c,f,j,i).

Discussion

The results of the present study demonstrate markedly altered

expression profiles of occludin and ZO-1 in the epidermis of

psoriasis plaques compared with normal skin. Interestingly,

claudin-5 was detected in the epidermis of normal skin,

(a) (b) (c)

(d) (e) (f)

(g) (h) (i)

(j) (k) (l)

Normal perilesional skin Psoriasis plaque Healed psoriasis plaque

Occludin

Occl+INV

ZO1+INV

ZO1

Fig 2. Indirect immunofluorescence labelling

of perilesional skin (left column), psoriasis

plaque (middle column) and healed psoriasis

plaque (right column) for ZO-1, occludin and

involucrin. (a,d,g,j) In perilesional skin, ZO-1

and occludin are expressed in the granular cell

layer, corresponding to the localization of

involucrin. (b,e,h,k) In psoriasis plaque, ZO-1

and occludin are spread to the upper spinous

cell layers while expression of involucrin is

detectable even in the lowest spinous cell

layers. (c,f,i,l) In healed psoriasis plaque, the

expression of ZO-1 and occludin is mainly

detected in the granular cell layer, along with

involucrin. However, tight junction proteins

can still be seen scattered between some

uppermost spinous cell layers. Scale bars ¼50 lm.




psoriasis plaques and perilesional skin. This is an unexpected

finding as previous studies have reported that normal epidermis

is negative for claudin-5, while claudin-5 is a component of TJs

of dermal endothelial cells.6,7 In the present study, claudin-5

was expressed more intensely in the upper epidermis of the pso-

riasis plaque than in the normal epidermis. Specifically, most

intense labelling was detected in and near the vicinity of the

parakeratotic foci in psoriatic epidermis. In perilesional and

healed epidermis claudin-5 was strictly restricted to the granular

cell layer, the distribution resembling that of occludin. Claudin-

5 immunolabelling was detectable only using a microscope with

high-quality immunofluorescence optics and confocal laser

scanning modality. As the amount of claudin-5 in the epidermis

is very low compared with claudin-1 and -4, it is understand-

able that epidermis has been considered to be negative for clau-

din-5. The localization of claudin-5 to the cell–cell contact sites

of the granular cells suggests that claudin-5 can be a component

of the epidermal TJs in epidermis.

(a)

(c) (d)

(e) (f)

(g) (h)

Fig 3. Indirect immunofluorescence labelling of lesional psoriatic skin

and normal control skin for claudin-5 and nonimmunized mouse

and rabbit IgG. (a,b) Claudin-5 in psoriasis plaque as demonstrated

using polyclonal rabbit antibody to claudin-5. (a) Indirect

immunofluorescence and (b) phase contrast image showing claudin-5

in the granular cell layer in the vicinity of a focally parakeratotic area.

(c) Normal skin labelled with a polyclonal antibody to claudin-5.

(d) Normal skin labelled with a monoclonal antibody to claudin-5.

Both antibodies reveal claudin-5 in the granular cell layer.

(e–h) Control labellings using IgG from nonimmunized mouse and

rabbit. (e,f) Psoriasis plaque and (g,h) normal skin labelled for rabbit

(e,g) or mouse (f,h) IgG do not show specific labelling patterns. Scale

bars ¼ 50 lm (a,b,e–h), 25 lm (c,d).

(a)

Wholeskin

GAPDH

Cl-5Cl-4Cl-1

Normalepidermis

Psoriaticepidermis

Psoriaticepidermis

(b)

(c)

(d)

Fig 4. Reverse transcription–polymerase chain reaction (PCR) for

claudin (Cl)-1, -4 and -5 and glyceraldehyde-3-phosphate

dehydrogenase (GAPDH) in samples of (a) normal whole skin,

(b) normal epidermis and (c,d) two psoriatic epidermis samples.

PCR for claudin-1 and -4 and GAPDH shows PCR products with an

expected size of 160–210 bp, claudin-5 yielding a larger fragment.




For RT–PCR, RNA was isolated from the material scraped

from the epidermis with a disposable curette. To our know-

ledge, this technique to obtain cells for RNA isolation from

the upper epidermis has not been published previously. RT–

PCR expectedly revealed claudin-1 and -4 in all epidermal

samples. However, claudin-5 was detected only in samples

from active, untreated psoriatic skin and control samples of

normal skin including dermal blood vessels. This suggests that

the expression level of claudin-5 in normal epidermis was

below the detection level of this technique.

The function of claudins has been related to the size-select-

ive sieving of molecules by TJs. For example, endothelial cells

contributing to the blood–brain barrier express at least two

distinct claudins: claudin-5 and -12. In claudin-5-null mice,

the brain endothelium is size-selectively leaky for small pro-

teins (<~800 Da), but not for large proteins.21,22 It is feasible

to speculate that the upregulation of claudin-5 in psoriatic epi-

dermis may serve to compensate for the dysfunction of the

lipid barrier by adding to the tightness of the barrier against

small molecules.

The barrier function in psoriasis is impaired potentially due

to different reasons: (i) alterations in the lipid structures of

the cornified cell layer23–25 and (ii) abnormal differentiation

of keratinocytes leading to prematurity of the cornified cell

envelope.26–28 Previous results have shown that barrier dys-

function leads to premature expression of involucrin.29 In

addition, several other differentiation markers, such as small

proline-rich proteins, cystatin A and transglutaminase 1, are

more conspicuously expressed in psoriasis while late differen-

tiation markers such as profilaggrin and loricrin are dimin-

ished or absent.29,30

In a recent report, experimental overexpression of claudin-6

in transgenic mouse epidermis induced the expression of clau-

din-5 and -8 in the same tissue.16 These mice also displayed

aberrations in epidermal development including hyperkeratosis

and atypical patterns of keratin expression.15,16 The results of

the present and previous studies thus suggest that there is

cross-talk between the different components of the epidermal

barrier: a defect in the lipid barrier can alter expression pro-

files of genes encoding TJ components and other proteins of

the epidermis.

Our earlier results have shown the spreading of ZO-1 to the

spinous cell layers in psoriasis plaques.5 In addition to being

located in TJs, ZO-1 may also be distributed with adherens

junctions in spinous cell layers as ZO-1 can associate with

a-catenin at the adherens junction and occludin at the TJ.31,32

The results of the present study also showed that healing of

psoriasis is associated with the reconstitution of expression of

ZO-1 and occludin to the granular cell layer.

The expression of claudin-1 is not altered in psoriasis, in

the perilesional psoriatic skin, or in the healed plaque, com-

pared with normal skin. Claudin-1 and -4 are abundantly

expressed in the intercellular contacts of upper epidermis.

These claudins may be associated with TJ-like structures that

have been described in the suprabasal cell layers of several

stratified epithelia.32 Claudin-4 was detected in more numer-

ous spinous cell layers in psoriasis than in normal epider-

mis, which may be explained by the increased number of

spinous cell layers in psoriasis. The expression profile of TJ

proteins in the perilesional skin resembled that in normal

skin.

Taken together, the results of the present study show over-

expression of three different types of TJ proteins which, in

combination, are sufficient to form functional TJs. The biolo-

gical significance of the potentially increased number of TJs

may relate to the defective diffusion barrier of the horny layer

in psoriasis plaques. Thus, the expression of occludin and

ZO-1 in acanthotic spinous cell layers and of claudin-5 in the

granular cell layer suggests a compensatory mechanism for TJs

under conditions when the horny layer is not fully function-

ing as a barrier.

The results of the present study highlight the fact that the

expression of the different TJ proteins is separately regulated

but, on the other hand, coordinated expression of selected TJ

components may serve to compensate for the defective barrier

function observed in psoriasis. The results of the present study

also demonstrate that changes in the expression profiles of TJ

components studied here are transient and are associated with

active psoriasis, with the exception of claudin-5 which is

expressed in perilesional and healed psoriasis. These findings

on altered gene expression profiles of TJ proteins may reflect

differentiation/dedifferentiation of epidermal cells in psoriasis,

or compensatory responses to impaired barrier function in

psoriasis plaques. The present study provides evidence for

both lines of thinking.

Acknowledgments

This work was supported by the Department of Dermatology,

Turku University Central Hospital, and Finnish Society of Der-

matology. We thank Maria Alanne and Jouko Sandholm for

assistance with confocal microscopy.

References

1 Stevenson BR, Keon BH. The tight junction: morphology to mole-

cules. Annu Rev Cell Dev Biol 1998; 14:89–109.2 Tsukita S, Furuse M, Itoh M. Structural and signalling molecules

come together at tight junctions. Curr Opin Cell Biol 1999; 11:628–33.3 Tsukita S, Furuse M. Pores in the wall: claudins constitute tight

junction strands containing aqueous pores. J Cell Biol 2000;149:13–16.

4 Morita K, Itoh M, Saitou M et al. Subcellular distribution of tightjunction-associated proteins (occludin, ZO-1, ZO-2) in rodent skin.

J Invest Dermatol 1998; 110:862–6.5 Pummi K, Malminen M, Aho H et al. Epidermal tight junctions:

ZO-1 and occludin are expressed in mature, developing, and affec-ted skin, and in vitro differentiating keratinocytes. J Invest Dermatol

2001; 117:1050–8.6 Brandner JM, Kief S, Grund C et al. Organization and formation of

the tight junction system in human epidermis and cultured kera-tinocytes. Eur J Cell Biol 2002; 81:253–63.

7 Morita K, Miyachi Y. Tight junctions in the skin. J Dermatol Sci2003; 31:81–9.




8 Schneeberger EE, Lynch RD. The tight junction: a multifunctionalcomplex. Am J Physiol Cell Physiol 2004; 286:C1213–28.

9 Sawada N, Murata M, Kikuchi K et al. Tight junctions and humandiseases. Med Electron Microsc 2003; 36:147–56.

10 Malminen M, Koivukangas V, Peltonen J et al. Immunohistochemi-cal distribution of the tight junction components ZO-1 and occlu-

din in regenerating human epidermis. Br J Dermatol 2003; 149:255–60.

11 Yoshida Y, Morita K, Mizoguchi A et al. Altered expression ofoccludin and tight junction formation in psoriasis. Arch Dermatol Res

2001; 293:239–44.

12 Morita K, Tsukita S, Miyachi Y. Tight junction-associated proteins(occludin, ZO-1, claudin-1, claudin-4) in squamous cell carcinoma

and Bowen’s disease. Br J Dermatol 2004; 151:328–34.13 Morita K, Furuse M, Yoshida Y et al. Molecular architecture of tight

junctions of periderm differs from that of the maculae occludentesof epidermis. J Invest Dermatol 2002; 118:1073–9.

14 Furuse M, Hata M, Furuse K et al. Claudin-based tight junctions arecrucial for the mammalian epidermal barrier: a lesson from clau-

din-1-deficient mice. J Cell Biol 2002; 156:1099–111.15 Turksen K, Troy TC. Permeability barrier dysfunction in transgenic

mice overexpressing claudin 6. Development 2002; 29:1775–84.16 Troy TC, Rahbar R, Arabzadeh A et al. Delayed epidermal permeab-

ility barrier formation and hair follicle aberrations in Inv-Cldn6mice. Mech Dev 2005; 122:805–19.

17 Saitou M, Furuse M, Sasaki H et al. Complex phenotype of micelacking occludin, a component of tight junction strands. Mol Biol

Cell 2000; 11:4131–42.18 Bazzoni G, Dejana E. Keratinocyte junctions and the epidermal bar-

rier: how to make a skin-tight dress. J Cell Biol 2002; 156:947–9.19 Itoh M, Furuse M, Morita K et al. Direct binding of three tight

junction-associated MAGUKs, ZO-1, ZO-2, and ZO-3, with theCOOH termini of claudins. J Cell Biol 1999; 147:1351–63.

20 Chomczynski P, Sacchi N. Single-step method of RNA isolation byacid guanidinium thiocyanate–phenol–chloroform extraction. Anal

Biochem 1987; 162:156–9.

21 Morita K, Sasaki H, Furuse K et al. Expression of claudin-5 in der-mal vascular endothelia. Exp Dermatol 2003; 12:289–95.

22 Nitta T, Hata M, Gotoh S et al. Size-selective loosening of theblood–brain barrier in claudin-5-deficient mice. J Cell Biol 2003;

161:653–60.23 Elias PM, Menon GK. Structural and lipid biochemical correlates of

the epidermal permeability barrier. Adv Lipid Res 1991; 24:1–26.24 Motta S, Monti M, Sesana S et al. Abnormality of water barrier

function in psoriasis. Role of ceramide fractions. Arch Dermatol1994; 130:452–6.

25 Fartasch M. Epidermal barrier in disorders of the skin. Microsc Res

Tech 1997; 38:361–72.26 Ishida-Yamamoto A, Iizuka H. Differences in involucrin immuno-

labeling within cornified cell envelopes in normal and psoriaticepidermis. J Invest Dermatol 1995; 104:391–5.

27 Ishida-Yamamoto A, Eady RA, Watt FM et al. Immunoelectronmicroscopic analysis of cornified cell envelope formation in normal

and psoriatic epidermis. J Histochem Cytochem 1996; 44:167–75.28 Ishida-Yamamoto A, Takahashi H, Iizuka H. Immunoelectron

microscopy links molecules and morphology in the studies ofkeratinization. Eur J Dermatol 2000; 10:429–35.

29 Ekanayake-Mudiyanselage S, Aschauer H, Schmook FP et al. Expres-sion of epidermal keratins and the cornified envelope protein

involucrin is influenced by permeability barrier disruption. J InvestDermatol 1998; 111:517–23.

30 Iizuka H, Takahashi H, Honma M, Ishida-Yamamoto A. Unique kera-tinization process in psoriasis: late differentiation markers are abol-

ished because of the premature cell death. J Dermatol 2004; 31:271–6.31 Muller SL, Portwich M, Schmidt A et al. The tight junction protein

occludin and the adherens junction protein alpha-catenin share acommon interaction mechanism with ZO-1. J Biol Chem 2005;

280:3747–56.32 Langbein L, Grund C, Kuhn C et al. Tight junctions and composi-

tionally related junctional structures in mammalian stratified epi-thelia and cell cultures derived therefrom. Eur J Cell Biol 2002;

81:419–35.





Cutaneous mucinosis associated with dermatomyositis andnephrogenic fibrosing dermopathy: fibroblast hyaluronansynthesis and the effect of patient serumM. Edward, L. Fitzgerald,* C. Thind,* J. Leman� and A.D. Burden*

Section of Squamous Cell Biology and Dermatology, Division of Cancer Sciences and Molecular Pathology, Robertson Building, University of Glasgow, Glasgow

G12 8QQ, U.K.

*Department of Dermatology, Western Infirmary, Dumbarton Road, Glasgow, U.K.

�Falkirk and District Royal Infirmary, Majors Loan, Falkirk, U.K.

CorrespondenceMichael Edward.



Key wordsdermatomyositis, hyaluronan, mucinosis,

nephrogenic fibrosing dermopathy


Summary

Background Dermal mucin is an amorphous gelatinous substance composed primar-ily of hyaluronan (HA) and sulphated glycosaminoglycans (GAGs). In primarycutaneous mucinosis, accumulation of mucin is a characteristic feature of lichenmyxoedematosus, scleromyxoedema and reticular erythematous mucinosis. Sec-ondary mucinoses are disorders where mucin deposition is an additional finding,and deposition is associated with lupus erythematosus, dermatomyositis, sclero-derma and granuloma annulare. The underlying cause of the abnormal mucindeposition is unknown. An increasing number of cases of a fibromucinousscleromyxoedema-like disorder associated with renal dysfunction, recently termednephrogenic fibrosing dermopathy (NFD), is being reported.Objectives To examine the synthesis of sulphated GAGs and HA by fibroblastsderived from uninvolved and involved skin of a patient with dermatomyositisand two patients with NFD, and the effect of patient serum.Methods GAGs were quantified by a radiometric assay and HA was determined byan enzyme-linked HA-binding protein assay.Results We found that fibroblasts derived from active lesions of NFD synthesizeelevated levels of GAGs, and in particular HA, compared with normal controls,while serum from the patient with dermatomyositis and the two patients withNFD stimulates GAG synthesis, including HA synthesis, by both control andpatient fibroblasts.Conclusions Fibroblasts from patients with active NFD are either activated to syn-thesize elevated levels of HA or contain another cell type, possibly derived fromcirculating fibrocytes. In both disorders, there is additionally a serum-derived fac-tor that stimulates production of sulphated GAGs and HA by fibroblasts.

The cutaneous mucinoses are a highly heterogeneous group

of disorders in which elevated levels of mucin are found in

the skin, predominantly in the dermis. Dermal mucin is an

amorphous gelatinous substance that is thought to be

composed primarily of hyaluronan (HA) and sulphated

glycosaminoglycans (GAGs), and is normally visualized

histochemically by staining tissue sections with alcian blue at

pH 2Æ5.

Primary cutaneous mucinoses are characterized by mucin

deposition as the major histological feature, such as in lichen

myxoedematosus (papular mucinosis), scleromyxoedema,

reticular erythematous mucinosis, and many more. Secondary

mucinoses are disorders where mucin deposition is an

additional finding, and deposition is found associated with

disorders including lupus erythematosus, dermatomyositis,

scleroderma and granuloma annulare.1 Many of these primary

and secondary mucinoses have considerable overlap in their

characteristics, but the underlying cause of the abnormal

mucin deposition is unknown.

Increased levels of serum immunoglobulins in mucinoses

such as lichen myxoedematosus, scleromyxoedema and Graves

disease-associated pretibial myxoedema suggest that these anti-

bodies may be involved in the increased mucin deposition.

However, it has been reported that fibroblast mucin



production is not affected by removal of the IgG paraprotein

from lichen myxoedematosus patient serum2 or removal of

the autoantibodies in Graves disease-associated pretibial myx-

oedema.3 This suggests that other factors present in serum are

responsible. Pandya et al.4 have demonstrated that fibroblasts

from a patient with papulonodular mucinosis associated with

systemic lupus erythematosus produced more GAGs than con-

trol cells, and in addition, GAG synthesis was increased in

control cells exposed to patient serum.

Recently there have been reports of scleromyxoedema-like

cutaneous changes appearing following renal dialysis,5,6 now

termed nephrogenic fibrosing dermopathy (NFD), and the

two patients from whom we have obtained biopsies and

serum both receive dialysis, although it likely that it is renal

failure rather than dialysis that is the trigger as some lesions

develop in patients not on dialysis.7 NFD fibromucinosis may

also occur in patients with liver disease,8 suggesting that there

may be an identical aetiological factor in renal and liver dis-

ease. In patients on peritoneal dialysis, levels of transforming

growth factor (TGF)-b1 in dialysate increase while levels of

basic fibroblast growth factor (bFGF) increase in dialysate

during peritonitis.9 In addition, several studies have documen-

ted an interaction between TGF-b1 and bFGF resulting in

stimulation of skin fibroblasts and irreversible fibrosis.10

Reports of mucinosis associated with dermatomyositis are

rare, although mucin deposits in cutaneous lesions are prob-

ably frequent when adequate staining techniques are

employed.11

Here we have examined the synthesis of sulphated GAGs

and HA by fibroblasts derived from a patient with dermato-

myositis and from two patients with NFD, and the effect of

patient serum on GAG synthesis by patient and control fibro-

blasts.


Patient with dermatomyositis

A 31-year-old woman presented with a slightly itchy rash on

the flexural aspects of the forearms and anterior chest of

6 months’ duration. This consisted of small coalescing, skin-

coloured dermal papules. A skin biopsy revealed dermal

accumulation of mucin without any epidermal changes or

inflammatory infiltrate. A diagnosis of papular mucinosis was

made.

On review several weeks later she had developed a viola-

ceous rash involving the forehead, eyelids, cheeks and chin.

Plaques consistent with Gottron’s papules were seen overlying

metacarpophalangeal (MCP) and interphalangeal joints of both

hands in addition to a violaceous rash streaking down the

extensor tendons. Nail fold telangiectases were also evident.

The papular eruption remained unchanged. Biopsy of a Got-

tron’s papule overlying the fifth MCP joint showed smudging

of the dermoepidermal junction and a chronic inflammatory

cell infiltrate within the dermis. Staining with alcian blue was

positive for mucin. Clinical examination revealed no apparent

muscle weakness. Creatine kinase was within the normal

range. Magnetic resonance imaging of the pelvic girdle

musculature did not reveal any evidence of an inflammatory

myopathy. Antinuclear antibody was not significantly raised at

a titre of 1 : 40. A diagnosis of amyopathic dermatomyositis

was made. Extensive screening including pelvic examination,

Pap smear, computed tomography of the abdomen and pelvis,

and serum tumour markers failed to uncover any underlying

internal malignancy. Both the dermatomyositis and papular

mucinosis have been controlled by a prolonged course of oral

steroids.

Nephrogenic fibrosing dermopathy patient 1

A 49-year-old woman presented with nephrotic syndrome

with rapid progression to advanced renal failure over a

2-month period. She had a history of ulcerative colitis for

which she was receiving treatment with mesalazine. Renal

biopsy showed focal segmental glomerulonephritis. Within

weeks of starting haemodialysis, she developed induration of

the skin of the anterior abdominal wall, anterior upper thighs,

antecubital fossae and dorsal forearms. Skin biopsy showed an

increased number of fibroblasts in the dermis, fibrosis, and

the deposition of extracellular material that stained positively

with alcian blue. Her skin condition has remained largely

unchanged over 4 years of observation.

Nephrogenic fibrosing dermopathy patient 2

A 61-year-old woman developed acute on chronic renal fail-

ure associated with renal calculi requiring haemodialysis.

Shortly after beginning haemodialysis, she developed pain and

stiffness in the hands associated with marked swelling and

induration of the skin. Skin biopsy revealed dermal accumu-

lation of mucin. Over the subsequent 4 years, the skin

improved slightly without any specific intervention.

Controls

Fibroblasts were established from forearm punch biopsies

taken from a healthy 29-year-old woman and a healthy

42-year-old man. Serum was prepared from blood taken from

these two healthy control volunteers and from two women

(aged 52 years and 59 years) on dialysis who did not exhibit

any fibrotic disorders.

Fibroblasts

Fibroblasts were established from patient and control biopsies

using the explant technique. Tissue was finely chopped, placed

in a 25-cm2 flask with 1Æ5 mL medium [minimum essential

medium (MEM)/10% fetal calf serum (FCS); Invitrogen, Paisley,

U.K.], and incubated at 37 �C for 3 days before addition of

a further 3 mL of medium. Once established, fibroblasts

were routinely subcultured and expanded before freezing.

Fibroblasts were subsequently used between passages 5 and 9.



474 Nephrogenic fibrosing dermopathy and hyaluronan, M. Edward et al.

Patient and control serum

Blood from controls and patients was withdrawn into poly-

propylene tubes, and allowed to clot overnight at 4 �C. The

serum was separated from the clot, and centrifuged at

2000 g for 20 min and filtered through 0Æ45 lm filters prior

to freezing.

Glycosaminoglycan analysis

Total GAG synthesis (sulphated GAGs and HA) was deter-

mined by incubating the cells with 3H-glucosamine and 35SO4

(Amersham Biosciences, Amersham, U.K.). Twenty-four-well

plates containing confluent patient or control skin fibroblasts

were washed with 1Æ0 mL serum-free (SF)-MEM/well, and

incubated with SF-MEM for 2 h prior to radiolabelling to

ensure there was no residual FCS present. Then 0Æ6 mL of

medium containing either SF-MEM (for comparison of GAG

synthesis by different fibroblasts) or appropriate patient or

control serum, 3H-glucosamine (5 lCi mL)1) and 35SO4

(20 lCi mL)1) was added to triplicate wells and the cultures

incubated for 24 h at 37 �C, after which the contents of each

well, together with 2 · 100 lL phosphate-buffered saline

(PBS) washes, were transferred to 1Æ5-mL Eppendorf tubes.

The cells were detached with trypsin and counted in a Coulter

counter. A sample of the harvested medium was diluted 1 in

15 with SF-MEM, and retained for subsequent HA determin-

ation. To the remaining radioactive medium was added

100 lL of carrier GAG solution (3Æ5 mg chondroitin sulphate,

3Æ0 mg mL)1 HA in H2O; Sigma Chemical Company, Poole,

U.K.), followed by boiling for 5 min. Pronase (100 lL of

3 mg mL)1 50 mmol L)1 Tris–HCl pH 7Æ6; Sigma) was

added, and the samples incubated for 24 h at 50 �C followed

by heat inactivation and centrifugation at 13 000 g for 5 min.

To the supernatant was added 1/5 volume of 10% cetylpyridi-

nium chloride (CPC) in 0Æ03 mol L)1 NaCl, and the mixture

was incubated for 20 min at 40 �C, followed by centrifuga-

tion at 13 000 g for 5 min. The resultant precipitate was

washed twice with 0Æ1% CPC in 0Æ03 mol L)1 NaCl, and the

pellet finally dissolved in 2 mol L)1 NaCl. Samples were taken

and suspended in Optiphase HiSafe 3 scintillant, and radioacti-

vity determined using a Packard Tricarb scintillation counter.

Hyaluronan assay

HA present in various media was determined using the Corge-

nix HA test kit that is based on an enzyme-linked HA-binding

protein (HABP) assay. Briefly, diluted media samples and

appropriate controls were incubated in HABP-coated micro-

wells, allowing HA present in the samples to react with the

immobilized binding protein. After the removal of unbound

material by washing, HABP conjugated with horseradish per-

oxidase solution was added to form complexes with the

bound HA. Following another washing step, the chromogenic

substrate tetramethylbenzidine with hydrogen peroxide was

added to develop a colour reaction. The intensity of colour

was measured in an enzyme-linked immunosorbent assay plate

reader at 450 nm. Levels of HA were determined from a refer-

ence curve obtained using reference HA solutions provided

with the kit.


Comparisons were performed using one-way ANOVA and Dun-

nett’s post hoc test.

Results

Glycosaminoglycan synthesis by control fibroblasts and

fibroblasts derived from uninvolved and involved patient

skin

It is apparent from Figure 1a that fibroblasts derived from

both involved and uninvolved skin of the patient with derma-

tomyositis did not demonstrate any significant difference in

GAG synthesis, as measured by 3H-glucosamine and 35SO4

incorporation, from that of two control fibroblast cultures.

The fibroblasts derived from NFD patient skin did, however,

exhibit elevated levels of both 3H-glucosamine and 35SO4

incorporation into GAGs, suggesting that synthesis of sulphat-

ed GAGs and possibly HA was somehow stimulated in these

patient fibroblasts. While the fibroblasts from NFD uninvolved

skin exhibited elevated GAG synthesis, the greatest increase

was observed with the fibroblasts from involved skin, where

NFD patient 2 fibroblasts exhibited a 3Æ1-fold increase in3H-glucosamine incorporation and a 3Æ3-fold increase in35SO4 incorporation compared with control 1 fibroblasts.

Levels of HA were also determined using a specific HABP

plate assay. Again, the fibroblasts from the patient with

dermatomyositis did not differ in their expression of HA from

the control fibroblasts, while the fibroblasts from NFD unin-

volved and involved skin synthesized elevated levels of HA,

particularly in the cells derived from involved skin where a

10Æ7-fold increase was observed in patient 2 fibroblasts com-

pared with control 1 fibroblasts (Fig. 1b).

Effect of patient serum on glycosaminoglycan synthesis

by control fibroblasts

It has previously been suggested that the serum of patients

with pretibial myxoedema and papulonodular mucinosis asso-

ciated with systemic lupus erythematosus may contain certain

factors responsible, at least in part, for the observed fibrosis/

mucinosis.3,4 To determine if the serum from our patients

with dermatomyositis and NFD contained any GAG-stimula-

ting activity, the effect of serum from these patients and from

two normal control volunteers was examined on GAG synthe-

sis by normal fibroblasts. Two additional serum samples from

unaffected patients on dialysis were also included, as the two

patients with NFD were on dialysis. A modest but significant

(P < 0Æ01) stimulation of GAG synthesis, as measured by3H-glucosamine incorporation by the control fibroblasts



Nephrogenic fibrosing dermopathy and hyaluronan, M. Edward et al. 475

exposed to the dermatomyositis and NFD patient serum, was

observed when compared with normal and control dialysis

patient serum (Fig. 2a). This suggests that sera from both the

patient with dermatomyositis and the two patients with NFD

contain factors capable of stimulating GAG synthesis by fibro-

blasts. Samples taken from the radiolabelled medium were

analysed for HA. To ensure that possible variations in HA con-

tent of the serum did not influence the levels of synthesis by

the fibroblasts, HA levels in the serum were determined, and

subtracted from that present in the medium following incu-

bation with the control fibroblasts. Figure 2b demonstrates

that the serum from the patients with dermatomyositis and

NFD contained potent HA-stimulating activities, with the

NFD patient 1 serum stimulating a 2Æ75-fold increase in HA

synthesis compared with the control 1 fibroblasts (P < 0Æ01).

Notably, one of the patients with NFD (230 ng HA mL)1

serum) and one of the dialysis control patients (242 ng

HA mL)1 serum) had elevated levels of serum HA compared

with the controls (mean 17Æ6 ng HA mL)1 serum). Elevated

levels of serum HA occur in a number of diseases including

liver disease,12 arthritis13 and certain cancers.14

Effect of patient serum on glycosaminoglycan synthesis

by fibroblasts from uninvolved patient skin

Fibroblasts established from uninvolved skin of the patients

with NFD exhibited elevated levels of GAG synthesis compared

with control fibroblasts in the presence of control serum,

while fibroblasts from uninvolved skin of the patient with der-

matomyositis did not exhibit any elevated GAG synthesis

(a)

(b)

Fig 1. Glycosaminoglycan (a) and hyaluronan (b) synthesis by

fibroblasts from patients with dermatomyositis and nephrogenic

fibrosing dermopathy. Glycosaminoglycan synthesis was determined

by measuring the incorporation of 3H-glucosamine (solid bars) and35SO4 (cross-hatched bars) into cetylpyridinium chloride-precipitable

material from the medium of control (Con1 and Con2),

dermatomyositis uninvolved and involved (Derm un and Derm in),

and nephrogenic fibrosing dermopathy uninvolved and involved (NFD

un and NFD in) fibroblasts. Hyaluronan from the corresponding

cultures was determined using a highly specific plate assay utilizing a

biotinylated hyaluronan-binding protein. Values are mean ± SD for

three separate cultures. ANOVA, F-test P < 0Æ001; Dunnett’s post hoc test

*P < 0Æ05, **P < 0Æ01 in comparison with Con1.

(a)

(b)

Fig 2. Effect of serum from patients with dermatomyositis and

nephrogenic fibrosing dermopathy on normal fibroblast

glycosaminoglycan (a) and hyaluronan (b) synthesis.

Glycosaminoglycan synthesis was determined by measuring the

incorporation of 3H-glucosamine into cetylpyridinium chloride-

precipitable material, and hyaluronan synthesis was determined by

measuring the levels present in the culture medium utilizing a

hyaluronan-binding plate assay. Dermatomyositis (Derm) and

nephrogenic fibrosing dermopathy (NFD1 and NFD2) patient sera

were compared with normal control sera (Con1 and Con2) and with

control dialysis patient sera (Dial1 and Dial2). Values are mean ± SD

for three separate cultures. ANOVA, F-test P < 0Æ001; Dunnett’s post hoc

test **P < 0Æ01 in comparison with Con1.




(Fig. 3a). Fibroblasts derived from uninvolved skin of the

patients with dermatomyositis and NFD all exhibited signifi-

cantly higher levels of GAG synthesis compared with the

control cultures when incubated in the presence of the corres-

ponding patient serum. It is therefore apparent that fibroblasts

from the patient with dermatomyositis are not activated to

synthesize increased levels of GAG in the presence of control

serum, but do respond to the dermatomyositis patient serum

with a 2Æ3-fold (P < 0Æ01) increase in 3H-glucosamine incor-

poration and a 1Æ4-fold increase in incorporation of 35SO4. In

contrast, the fibroblasts from uninvolved NFD skin exhibited

elevated levels of both 3H-glucosamine and 35SO4 in the

presence of control serum, and even higher levels of GAG

synthesis in the presence of the corresponding NFD patient

serum, where NFD patient 1 fibroblasts exhibited a 4Æ3-fold

increase in 3H-glucosamine incorporation and a 2Æ8-fold

increase in 35SO4 incorporation (P < 0Æ01). Similarly, HA

synthesis was significantly increased in uninvolved fibroblasts

from patients with NFD in the presence of control and corre-

sponding NFD patient serum (Fig. 3b), with the greatest

increase being observed in the presence of patient serum (9Æ9-

fold, P < 0Æ01). The fibroblasts from uninvolved dermatomyo-

sitis skin showed a slight but not significant increase in HA

compared with control fibroblasts in the presence of control

serum, while HA synthesis was stimulated 4Æ4-fold in the pre-

sence of dermatomyositis patient serum compared with the

normal control (P < 0Æ05).

Effect of patient serum on fibroblasts from involved

patient skin

The pattern of GAG stimulation in fibroblasts from involved

skin was similar to that observed with fibroblasts from unin-

volved skin (Fig. 4a). Fibroblasts from involved dermatomyo-

sitis skin exhibited a similar level of GAG synthesis to that of

control fibroblasts in the presence of control serum, but

enhanced their synthesis significantly in the presence of serum

from the patient with dermatomyositis. Again, the fibroblasts

from involved NFD skin exhibited elevated GAG synthesis in

the presence of both control and corresponding NFD patient

serum, with the greatest increase being observed with the

NFD patient serum (5Æ1-fold increase in 3H-glucosamine

incorporation, and 3Æ7-fold increase in 35SO4 incorporation in

NFD patient 2 compared with control fibroblasts; P < 0Æ01).

The HA synthesis was similar to the 3H-glucosamine incorpor-

ation (Fig. 4b), with the greatest increase in HA synthesis

being observed in fibroblasts from NFD patient 2 in the pres-

ence of NFD patient serum (17Æ5-fold compared with control

fibroblasts in control serum; P < 0Æ01).

Discussion

Mucinous lesions are probably more widespread than currently

reported, primarily due to fixation techniques in which one of

the main molecules involved, HA, is depleted. Mucin is nor-

mally visualized histochemically by staining tissue sections with

alcian blue at pH 2Æ5. However, this technique requires to be

modified as fixation of tissue in 10% formalin in PBS does not

fix HA, which can subsequently be lost by simple PBS washing.

In addition, alcian blue does not readily stain HA at pH 2Æ5, so

most of the observed staining with alcian blue is presumably

attributable to sulphated GAGs.15 The most effective and speci-

fic method of visualizing HA in tissue is by utilizing a biotinyl-

ated HABP, isolated from articular cartilage. Indeed, Neudecker

et al.16 have recently shown that scleromyxoedema-like lesional

biopsies of patients in renal failure are associated with marked

deposition of HA, particularly in the papillary dermis, as visual-

ized using a biotinylated HABP.

Dermatomyositis is an inflammatory disorder of both mus-

cle and skin; however, in practice, dermatomyositis represents

(a)

(b)

Fig 3. Effect of patient serum on glycosaminoglycan (a) and

hyaluronan (b) synthesis by fibroblasts from uninvolved skin of

patients with dermatomyositis and nephrogenic fibrosing dermopathy.


incorporation of 3H-glucosamine (solid bars) and 35SO4 (cross-

hatched bars) into cetylpyridinium chloride-precipitable material, and

hyaluronan synthesis was determined by measuring the levels present

in the culture medium utilizing a hyaluronan-binding plate assay.

Dermatomyositis (Derm) and nephrogenic fibrosing dermopathy (NFD

1 and NFD 2) fibroblast glycosaminoglycan and hyaluronan synthesis

was examined in the presence of patient and control serum. Values

are mean ± SD of three separate cultures. ANOVA, F-test P < 0Æ001;

Dunnett’s post hoc test *P < 0Æ05, **P < 0Æ01 compared with control

fibroblasts in the presence of control serum.




a disease continuum ranging from inflammatory myopathy in

isolation to cutaneous disease in the absence of muscle

involvement. There are several reports documenting the occur-

rence of macroscopic mucinous plaques in dermatomyositis

and, indeed, our patient exhibited papular mucinosis. The

underlying cause of the mucinosis is unknown, but immuno-

logically mediated fibroblast stimulation has been postulated

in secondary mucinosis.17 The presence of a tumour has fre-

quently been associated with dermatomyositis,18 and in such

cases it may be that the tumour is releasing factors into serum

that promote the development of the mucinous lesions. Exten-

sive screening of our patient including pelvic examination,

Pap smear, computed tomography of the abdomen and pelvis,

and tumour markers including a-fetoprotein, chorionic gon-

adotrophin, carcinoembryonic antigen and C125 failed to

uncover an underlying internal malignancy over a 3-year per-

iod of clinical observation. It does appear likely that the muc-

inous lesions are caused, at least in part, by the presence of

certain serum-derived factors as we have demonstrated that

dermatomyositis patent serum stimulates both sulphated GAG

and HA synthesis by fibroblasts from both control and

involved skin. In contrast, the patient fibroblasts show similar

levels of GAG synthesis as control fibroblasts in the presence

of control serum, suggesting that these fibroblasts are not acti-

vated, nor contain a population of cells that overexpress GAGs.

There has been a report of epidermal changes in dermatomyo-

sitis patients with mucinosis, in whom there is moderate atro-

phy of the epidermis,11 while a lack of CD44, a receptor for

HA, in mouse epidermis, leads to an abnormal accumulation

of HA in the dermis.19 This suggests that disruption of kera-

tinocyte–HA interactions may affect both epidermal prolifer-

ation and HA synthesis within the dermis.

In contrast, the fibroblasts derived from involved and unin-

volved NFD patient skin show elevated levels of sulphated

GAG and HA synthesis in addition to patient serum actively

stimulating GAG synthesis. Again, it is unclear what the

serum-derived active factors are, or why the fibroblasts consti-

tutively synthesize excess levels of GAGs. The cells may be

permanently activated by certain factors, or some infiltrating

cell type may be actively synthesizing excess collagen and

GAGs and contributing to the development of the lesions. A

possible candidate cell type is the circulating CD34+ fibro-

cyte.20 Several cell types express CD34, including embryonic

fibroblasts, endothelial cells and bone marrow stromal cells,

but the haematological origin of fibrocytes within a tissue is

confirmed by the expression of the panleucocyte antigen

CD45.21 In addition to CD34, and CD45, the coexpression of

type I procollagen has also been utilized to identify fibro-

cytes.22 These circulating fibrocytes are likely to play an im-

portant role in wound healing as CD34+ cells appear at the

inflammatory stage, and apparently decrease with maturation

of the wound.23 However, it appears that the decreased

expression of CD34 is a result of differentiation of the fibro-

cytes into myofibroblasts, and subsequent elevated collagen

synthesis and contractile properties, and this is supported by

the observation that cultured fibrocytes are induced to express

a-smooth muscle actin and downregulate expression of CD34

in the presence of TGF-b1.24

Why the fibrocytes infiltrate the skin to create lesions char-

acteristic of the dermopathy is unknown. Are there certain

chemokines expressed in the skin attracting the fibrocytes, or

are the fibrocytes stimulated by some factor within the circula-

tion, and why do they apparently switch off in normal acute

wound healing as the wound matures, yet produce excess col-

lagen and GAGs in NFD lesions? Neudecker et al.16 have sug-

gested that circulating low molecular weight HA may play a

role in NFD. While liver is the main organ involved in deg-

radation of high molecular weight HA, the kidney may be

involved in degradation of lower molecular weight HA which,

following renal failure, may accumulate in the circulation.25

(a)

(b)

Fig 4. Effect of patient serum on glycosaminoglycan (a) and

hyaluronan (b) synthesis by fibroblasts from involved skin of patients

with dermatomyositis and nephrogenic fibrosing dermopathy.


incorporation of 3H-glucosamine (solid bars) and 35SO4 (cross-

hatched bars) into cetylpyridinium chloride-precipitable material, and

hyaluronan synthesis was determined by measuring the levels present

in the culture medium utilizing a hyaluronan-binding plate assay.

Dermatomyositis (Derm) and nephrogenic fibrosing dermopathy (NFD

1 and NFD 2) fibroblast glycosaminoglycan and hyaluronan synthesis

was examined in the presence of patient and control serum. Values

are mean ± SD of three separate cultures. ANOVA, F-test P < 0Æ001;

Dunnett’s post hoc test *P < 0Æ05, **P < 0Æ01 compared with control

fibroblasts in the presence of control serum.




HA is a molecule with a multitude of biological functions

depending upon its molecular size, with high molecular weight

HA being anti-inflammatory, and low molecular weight frag-

ments inflammatory.26 These low molecular weight fragments

may therefore induce an inflammatory wound healing reaction

resulting in fibrosis,27 possibly mediated by circulating fibro-

cytes. This hypothesis does not, however, explain why only a

few patients with renal failure develop NFD, or why the first

case was identified only in 1997. Recent reports suggest a cor-

relation between the development of NFD and exposure to the

magnetic resonance angiography contrast agent, gadolinium.28

The two patients with NFD described here were exposed to

gadolinium approximately 2 months prior to the development

of fibrotic lesions, thus further supporting the possibility that

gadolinium is acting as the trigger for the development of

NFD. Current studies are under way to examine in more detail

the relationship between gadolinium and NFD.

Acknowledgments

We gratefully acknowledge financial assistance from the British

Skin Foundation.

References

1 Rongioletti F, Rebora A. Cutaneous mucinoses: microscopic criteriafor diagnosis. Am J Dermatopathol 2001; 23:257–67.

2 Harper RA, Rispler J. Lichen myxedematosus serum stimulateshuman skin fibroblast proliferation. Science 1978; 199:545–7.

3 Cheung HS, Nicoloff JT, Kamiel MB et al. Stimulation of biosyn-thetic activity by serum of patients with pretibial myxedema.

J Invest Dermatol 1978; 71:545–7.4 Pandya AG, Sontheimer RD, Cockerell CJ et al. Papulonodular muci-

nosis associated with systemic lupus erythematosus: possible mech-anisms of increased glycosaminoglycan accumulation. J Am Acad

Dermatol 1995; 32:199–205.5 Cowper SE, Robin HS, Steinberg SM et al. Scleromyxedema-like

cutaneous disease in renal dialysis patients. Lancet 2000; 356:1000–1.6 McNeill AM, Barr RJ. Scleromyxedema-like fibromucinosis in a

patient undergoing hemodialysis. Int J Dermatol 2002; 41:364–7.7 Cassis TB, Jackson JM, Sonnier MD, Callen JP. Nephrogenic fibro-

sing dermopathy in a patient with acute renal failure never requi-

ring dialysis. Int J Dermatol 2006; 45:56–9.8 Banno H, Takama H, Nitta Y et al. Lichen myxedematosus associ-

ated with chronic hepatitis C. Int J Dermatol 2000; 39:212–15.9 Mlambo NC, Hylander B, Brauner A. Increased levels of transform-

ing growth factor beta 1 and basic fibroblast growth factor inpatients on CAPD. A study during non-infected steady state and

peritonitis. Inflammation 1999; 2:131–9.10 Takehara K. Growth regulation of skin fibroblasts. J Dermatol Sci

2000; 24:S70–7.

11 Del Pozo J, Almagro M, Martinez W et al. Dermatomyositis andmucinosis. Int J Dermatol 2001; 40:120–4.

12 Guechot J, Poupon RE, Poupon R. Serum hyaluronan as a markerof liver fibrosis. J Hepatol 1995; 22:103–6.

13 Emlen W, Niebur J, Flanders G, Rutledge J. Measurement of serumhyaluronic acid in patients with rheumatoid arthritis: correlation

with disease activity. J Rheumatol 1996; 23:974–8.14 Wilkinson CR, Bower LM, Warren C. The relationship between

hyaluronidase and hyaluronic acid concentration in sera fromnormal controls and from patients with disseminated neoplasm.

Clin Chim Acta 1996; 256:165–73.

15 Lin W, Shuster S, Maibach HI, Stern R. Patterns of hyaluronanstaining are modified by fixation techniques. J Histochem Cytochem

1997; 45:1157–69.16 Neudecker BA, Stern R, Mark LA, Steinberg S. Scleromyxedema-like

lesions of patients in renal failure contain hyaluronan: a possiblepathophysiological mechanism. J Cutan Pathol 2005; 32:612–15.

17 Igarashi M, Aizawa H, Tokudome Y, Tagami H. Dermatomyositiswith prominent mucinosis skin change. Histochemical and biochemi-

cal aspects of glycosaminoglycans. Dermatologica 1985; 170:6–11.18 Leow YH, Goh CL. Malignancy in adult dermatomyositis. Int J

Dermatol 1997; 36:904–7.19 Kaya G, Rodrigez I, Jorcano JL et al. Selective suppression of CD44

in keratinocytes of mice bearing an antisense CD44 transgene dri-ven by a tissue-specific promoter disrupts hyaluronate metabolism

in the skin and impairs keratinocyte proliferation. Genes Dev 1997;11:996–1007.

20 Quan TE, Cowper S, Wu S-P et al. Circulating fibrocytes: collagen-secreting cells of the peripheral blood. Int J Biochem Cell Biol 2004;

36:598–606.21 Bucala R, Spiegel LA, Chesney J et al. Circulating fibrocytes define a

new leukocyte subpopulation that mediates tissue repair. Mol Med1994; 1:71–81.

22 Schmidt M, Sun G, Stacey MA et al. Identification of circulatingfibrocytes as precursors of bronchial myofibroblasts in asthma.

J Immunol 2003; 171:380–9.23 Aiba S, Tagami H. Inverse correlation between CD34 expression

and proline-4-hydroxylase immunoreactivity on spindle cells notedin hypertrophic scars and keloids. J Cutan Pathol 1997; 24:65–9.

24 Abe R, Donnelly SC, Peng T et al. Peripheral blood fibrocytes: dif-ferentiation pathway and migration to wound sites. J Immunol

2001; 166:7556–62.25 Lipkin GW, Forbes MA, Cooper EH, Turney JH. Hyaluronic acid

metabolism and its clinical significance in patients treated by con-

tinuous ambulatory peritoneal dialysis. Nephrol Dial Transplant 1993;8:357–60.

26 Termeer CJ, Hennies U, Voith T et al. Oligosaccharides of hyaluro-nan are potent activators of dendritic cells. J Immunol 2000;

165:1862–70.27 Noble PW. Hyaluronan and its catabolic products in tissue injury

and repair. Matrix Biol 2002; 21:25–9.28 Grobner T. Gadolinium – a specific trigger for the development

of nephrogenic fibrosing dermopathy and nephrogenic systemicfibrosis. Nephrol Dial Transplant 2006; 21:1104–8.





Sweet’s syndrome: a spectrum of unusual clinicalpresentations and associationsC.Y. Neoh, A.W.H. Tan and S.K. Ng

National Skin Centre, 1 Mandalay Road, Singapore 308205, Singapore

CorrespondenceDr C.Y. Neoh.



Key wordsinfections, imatinib, neutrophilic dermatosis,

neutrophilic panniculitis, Sweet’s syndrome,

unusual presentations

Conflict of interestNone declared.

Summary

Background Sweet’s syndrome (SS) is the prototypic neutrophilic dermatosis. Firstdescribed in 1964, the characterization of new clinical associations, unique histo-pathological findings and clinical variants have stimulated much interest and dis-cussion recently. However, the prevalence of these unusual variants and clinicalassociations within a single cohort of patients, has not been described.Objectives To describe and evaluate the prevalence of unusual clinical and histo-pathological features, as well as the clinical associations of SS seen in patientsfrom the National Skin Centre, Singapore.Methods This is a retrospective study of all consecutive cases of SS seen at our cen-tre over a 5Æ5-year period (June 1999–December 2004). Data on associated sys-temic diseases was obtained from the medical records and matched withinformation from the National Cancer Registry, Singapore. Patients not activelyfollowed up for more than 3 months were contacted for their updated healthstatus.Results Thirty-seven patients were identified. Ten (27%) had non-idiopathic SS.These were associated with haematological disorders, connective tissue disorders,infections or a drug. Twenty-nine patients (78%) had at least one atypical clinicalor histopathological feature. Atypical clinical features included bullous lesions, SSwith hand involvement or neutrophilic dermatoses of the hands and the con-comitant existence of subcutaneous SS with pyoderma gangrenosum. SS was thepresenting feature in three patients with infections caused by atypical organisms,including Mycobacterium chelonae, Penicillium species and Salmonella type D. Uniquehistopathological variants included subcutaneous SS and lesions containing anadmixture of mature and immature neutrophils. Subcutaneous neutrophilicinflammation seemed to be more common in patients with an underlying haem-atological disorder. This group of patients also had a lower mean haemoglobinlevel.Conclusions Unusual clinical and histopathological variants of SS described in theliterature are similarly encountered in our cohort of patients, with some featuresbeing more common than others. We highlight and discuss some unique clinicaland histopathological observations seen in our patients with SS.

Sweet’s syndrome (SS) is the prototypic neutrophilic dermato-

sis. Its typical features consist of pyrexia, elevated neutrophil

counts, painful red plaques and a predominantly mature neu-

trophilic dermal infiltrate. Since its original description in

1964,1 many subsequent reports have expanded the clinical

and pathological features of this condition. Of special import-

ance is the identification of new malignancies and/or medica-

tions associated with SS and the description of variants such as

neutrophilic dermatosis of the hands. New histopathological

variants like histiocytoid SS, SS-associated leukaemia cutis and

neutrophilic panniculitis have also attracted particular attention

in recent years. Taking these evolving and new concepts into

account, we review patients diagnosed with SS, at the National

Skin Centre, Singapore over 5Æ5 years and highlight the atyp-

ical presentations and clinical associations of SS.

Patients and methods

Consecutive patients attending the National Skin Centre, a ter-

tiary dermatological referral centre in Singapore, between June



1999 and December 2004, diagnosed with SS, were included

in this study. The diagnosis of SS was established through a

history of the abrupt onset of erythematous tender plaques

with corresponding dense neutrophilic infiltrates on histo-

pathology and a rapid therapeutic response to systemic corti-

costeroids.2,3 Medical and histopathological records were

reviewed. Patients with atypical clinical and histopathological

features, as well as those with underlying systemic diseases

were identified. Those who had not been followed up for

more than 3 months were contacted by phone for their updated

health status. The patients were also matched with data from

the National Cancer Registry, Singapore. The National Cancer

Registry is a comprehensive registry established in January

1968, with cancer notifications derived from medical

practitioners, pathology records, hospital records and death

certificates.

Results

Thirty-seven patients were diagnosed with SS during the study

period. They were aged 14–78 years (mean: 46 years). The

female : male ratio was 1Æ5 : 1. The mean duration of follow-

up was 14 months. Ten (27%) patients had nonidiopathic SS.

SS was associated with haematological malignancies in five

patients [two acute myeloid leukaemia (AML), one chronic

myeloid leukaemia (CML) and two myelodysplastic syndrome

(MDS)]. One patient with MDS had concomitant Penicillium

species bone marrow infection. SS was drug-induced (imati-

nib) in one of the patients with AML. Nontuberculous myco-

bacterial infections were found in two patients (Table 1).

Three patients had associated connective tissue disorders (CTD)

[undifferentiated CTD, systemic lupus erythematosus (SLE) and

SLE with autoimmune thrombocytopenic purpura (AITP)].

Eighty per cent of patients with idiopathic SS and 75% of

patients with nonidiopathic SS exhibited atypical features.

Unusual clinical presentations and associations

Neutrophilic dermatosis of the dorsal hands. Seven patients (19%) pre-

sented with lesions on the hands (Fig. 1). Involvement of the

dorsum of the hands was an isolated event in two patients

only. The rest had concomitant involvement of the lower

limbs, face or trunk. Of these, three patients had typical SS

lesions while two had bullous lesions. Histopathological find-

ings in all cases showed dense, neutrophilic infiltrates. Vasculi-

tis was seen in one specimen. No patients had associated

systemic diseases at presentation or on follow-up. All patients

were successfully treated with oral prednisolone, colchicine or

topical corticosteroids.

Sweet’s syndrome with bullous or pustular lesions. Eighteen patients

(49%) were documented with either bullous or pustular lesions

(Fig. 2). The dominant morphology could not be ascertained in

most patients because of the retrospective nature of this study.

Of these, five had associated underlying systemic conditions

(SLE with AITP, undifferentiated CTD, AML, group D Salmonella Tab

le1

Patien

tsw

ith

infe

ctio

n-as

soci

ated

Swee

t’s

synd

rom

e

Patien

tAge

Sex

Ass

ocia

ted

infe

ctio

nO

ther

med

ical

cond

itio

nsSw

eet’s

synd

rom

ele

sion

sTre

atm

ent

Out

com

eof

Swee

t’s

synd

rom

e

141

Male

Gro

upD

Salm

onell

ace

rvic

ally

mph

aden

itis

Pulm

onar

yMyc

obacteriu

mka

nsasii

infe

ctio

non

trea

tmen

tw

ith

rifa

mpi

cin,

ison

iazi

d,et

ham

buto

lpa

st12

mon

ths

Eryt

hem

atou

spl

aque

sw

ith

pustul

esIn

trav

enou

sam

pici

llin,

gent

amic

in,m

etro

nida

zole

Res

olve

d

247

Male

Penicillium

spec

ies

Mye

lody

splastic

synd

rom

ew

ith

refrac

tory

anae

mia

Eryt

hem

atou

spl

aque

s

ofSS

with

pyod

erm

aga

ngre

nosu

m

Intrav

enou

sce

ftaz

idim

e,

clox

acill

in,pr

edni

solo

ne.

Patien

ttran

sfer

red

toho

me

coun

try

for

trea

tmen

t

No

new

SSle

sion

s

ontran

sfer

356

Fem

ale

Myc

obacteriu

mch

elona

e

cerv

ical

lym

phad

enitis

2Rec

urre

ntge

nita

lhe

rpes

sim

plex

infe

ctio

n.

Pulm

onar

yM

ycob

acteriu

mtuberculosis

infe

ctio

n,co

mpl

eted

trea

tmen

t6

mon

ths

befo

re

curr

entpr

esen

tation

Ves

icul

arer

ythe

mat

ous

plaq

ues

Rifam

pici

n,clar

ithr

omyc

in

pred

niso

lone

Res

olve

d



Unusual clinical feature in Sweet syndrome, C.Y. Neoh et al. 481

and Mycobacterium chelonae cervical lymphadenitis).4 Of five

patients with haematological disorders, only one patient with

AML presented with bullous lesions. Biopsies were performed

in all patients and consistently showed marked dermal oedema

with dense dermal neutrophilic infiltrates, compatible with SS.

Concomitant multiple neutrophilic dermatoses. A 47-year-old Indones-

ian man presented concurrently with lesions of SS on the

limbs and pyoderma gangrenosum (Fig. 3) on his chest. He

was subsequently diagnosed with MDS with refractory anae-

mia, as well as bone marrow infection with Penicillium species

(Table 1).

Imatinib-induced Sweet’s syndrome. A 36-year-old woman with a

1-year history of CML presented with fever and recurrent ten-

der nodules on the limbs. Imatinib mesylate was started

1 month before the appearance of SS. Spontaneous resolution

occurred after imatinib was stopped. Upon re-institution of

imatinib at progressively increasing doses, there was a

recurrence of SS. The lesions responded remarkably to oral

prednisolone but recurred upon stopping this treatment.

Sweet’s syndrome associated with infections. Three patients had SS

associated with infections (Table 1). These infections include

group D Salmonella lymphadenitis, Penicillium species bone mar-

row infection and Mycobacterium chelonae lymphadenitis. The first

two patients tested negative for human immunodeficiency

virus (HIV) infection while the last patient declined the test.

Unusual histopathological presentations

Subcutaneous neutrophilic inflammation. Subcutaneous neutrophilic

inflammation was seen in seven patients (Fig. 4). In six

patients, there was an overlying dermal inflammatory infiltrate

comprising mainly neutrophils. The only patient with isolated

diffuse lobular neutrophilic infiltrates within the subcutis

presented with typical erythematous plaques and was later

diagnosed with MDS. Two other patients had associated haem-

atological disorders (one MDS, one CML). One had an under-

lying undifferentiated connective tissue disease.

Sweet’s syndrome with infiltrates containing neutrophils and immature white

cells. Two of three patients with a mixture of mature neu-

trophils and immature white cells on histopathology (Fig. 5)

were diagnosed with acute myeloid leukaemia and MDS,

respectively. All three patients presented acutely with skin

lesions resembling SS with fever, in contrast to the more

Fig 2. Bullous lesions of Sweet’s syndrome.

Fig 1. Neutrophilic dermatosis of the dorsal hands.

Fig 3. Ulcerative lesion of pyoderma gangrenosum occurring in a

patient with concomitant Sweet’s syndrome.



482 Unusual clinical feature in Sweet syndrome, C.Y. Neoh et al.

indolent lesions of leukaemia cutis. The last patient had a

repeat biopsy on a fresh SS lesion, 6 months after the initial

presentation, which did not show any atypical white cells.

Comprehensive investigations for haematological malignancy

are negative to date.

Vasculitis. Fibrinoid necrosis of the blood vessels within a

dense, neutrophilic infiltrate was seen in two specimens from

fresh SS lesions (Fig. 6). Both patients responded to treatment

with oral prednisolone and topical corticosteroids.

Features defining patients with Sweet’s syndrome

associated with haematological disorders

Subcutaneous neutrophilic inflammation was seen in three

of five patients with an underlying haematological disorder.

The mean haemoglobin level in the patients with associ-

ated haematological disorders was 8Æ6 g dL)1 compared with

13Æ0 g dL)1 in patients without.

Discussion

Sweet’s syndrome was originally described by Dr Robert

Douglas Sweet in 1964.1 In 1986, several diagnostic criteria

for SS were proposed by Su and Liu.5 It was subsequently

revised by von den Driesch,2 who classified them into two

categories: major and minor. The clinical and histopathologi-

cal concepts of SS have evolved over the years and are still

constantly being updated. We highlight in this study some

unusual features of SS seen in our patients.

The mean age of presentation of 46 years in our patients

falls within the previously reported age range of 30–60 years.6

The female : male ratio appears to be lower at 1Æ5 : 1. This

ratio remains fairly similar to that from a previous epidemi-

ological study of SS in Singapore.3 It is known that the

female : male ratio for malignancy-associated SS is 1 : 1 com-

pared with 4 : 1 in idiopathic SS.2 However, since the propor-

tion of malignancy-associated SS in both series from Singapore

was not much more than idiopathic SS, the lower female :

male ratio is unlikely to be attributable to this. SS may poss-

ibly be more common in Asian males from Singapore, in con-

trast with those seen in Western countries.

Neutrophilic dermatosis of the dorsal hands is a recently

described disorder that is now recognized to be a unique,

localized variant of SS with predominant or isolated hand

involvement.7,8 This relationship is supported by the similar

female preponderance of both conditions and their identical

incidence of associated haematopoietic and inflammatory dis-

eases.8 Initially termed as pustular vasculitis of the hands,9,10

it was later revised after recognizing that vasculitis was an

inconsistent finding and of secondary importance in its diag-

nosis. Frequently misdiagnosed as a cutaneous infection, these

signs were often found to be significantly associated with an

occult malignancy or systemic inflammatory disorder.8,9,11

Although none of our patients had associated disorders, larger

cohort studies are needed to establish a definite association

between this entity and systemic diseases.

Fig 5. Infiltrate of mature neutrophils admixed with immature white

cells (haematoxylin and eosin, original magnification · 600).

Fig 6. Fibrinoid necrosis of blood vessels with extravasation of red

blood cells. Dense neutrophilic infiltrate in upper field (haematoxylin

and eosin, original magnification · 400).

Fig 4. Lobular infiltrates of neutrophils between adipocytes:

Neutrophilic panniculitis (haematoxylin and eosin, original

magnification · 200).




Despite the rare incidence of the bullous variant of SS,12

almost half of our patients appear to have presented with this.

The inclusion of the more commonly encountered pseudoves-

icular variant of SS due to pronounced oedema, may have

caused this group to be over-represented. Only a handful

of these patients had photographic documentation of their

lesions, making the verification of truly bullous lesions diffi-

cult. Patients with haematological disorders in our series did

not present more commonly with the bullous variant of SS,

contrary to previous reports.13,14 Another uncommon presen-

tation is the successive appearance of multiple neutrophilic

dermatoses in the same patient.15 One of our patients presen-

ted concurrently with pyoderma gangrenosum and SS.

Together, these cases provide evidence that these skin dis-

orders form a spectrum of entities, rather than isolated

diseases.16

Drug-induced SS has been reported with all-trans retinoic

acid,17 carbamazepine,18 trimethoprim-sulfamethoxazole,19

granulocyte colony-stimulating factor20 and recently with

imatinib.21–23 We report another patient who developed SS

after imatinib. Imatinib mesylate (Gleevec�, Novartis, Basel,

Switzerland) is approved by the United States Food and Drug

Administration (FDA) for treatment of CML, hypereosinophilic

syndrome, systemic mastocytosis, gastrointestinal stromal

tumours and dermatofibrosarcoma protuberans.24 Despite the

wide use of imatinib, reports of imatinib-induced SS have been

described only in CML patients. As SS can be associated with

CML itself, the time course of appearance of SS in relation to

any drugs must be examined carefully in any CML patient,

before attributing the lesions to a particular medication.

Sweet’s syndrome has been well described in association

with nontuberculous mycobacterial infections.25,26 Similarly,

reports of SS occurring together with Salmonella infections are

available.27,28 Penicilliosis, however, has not been described in

association with SS, despite it being a common opportunistic

pathogen in Southeast Asia.29 It is interesting that in the three

patients with infections associated with SS (Table 1), there

were either recurrent, multiple infections or infections second-

ary to atypical organisms, raising the possibility of an underly-

ing immunocompromised state. As demonstrated by our two

patients with SS in association with lymphadenitis, an underly-

ing atypical infection should be considered, particularly if no

clinical response to conventional treatment is observed.

Although SS is traditionally classified as a neutrophil-

predominant dermatosis without vasculitis, the presence of

vasculitis should not preclude SS. Seen in two of our patients,

vasculitis is considered an epiphenomenon occurring second-

ary to extensive neutrophilic dermal inflammation in SS.30 The

phenomenon in which immature granulocytes are found

mixed with mature neutrophils in SS has been termed SS-

associated leukaemia cutis.31 These immature white cells are

hypothesized to constitute the incipient presence of a specific

leukaemic infiltrate (primary leukaemia cutis) or may repre-

sent the recruitment of immature myeloid precursor cells in

response to chemotactic factors stimulated by inflammatory SS

lesions (secondary leukaemia cutis).31

Distinguishing true leukaemia cutis from neutrophilic derma-

toses in a patient with a haematological disorder can be diffi-

cult, since dermal infiltrates of immature white cells can also be

found in the latter, more benign condition.32–34 It is believed

that in cases of neutrophilic dermatoses occurring in the setting

of a myeloid malignancy, the mature neutrophils are in fact

derived from the same clonal differentiation of myeloblasts.35 A

common origin of these cells explains why the presence of leu-

kaemia cutis and neutrophilic dermatoses in haematological

malignancies, both herald an overall poorer prognosis for the

malignancy.35 A series of SS patients with biopsy specimens

containing immature myeloid cells resembling histiocytes has

also recently been described.36 The presence of these histopath-

ological variants make thorough clinicopathological correlation

and meticulous follow-up vital for patients with SS.

The entity of neutrophilic panniculitis is contentious. Whe-

ther it is a distinct neutrophilic dermatoses or subcutaneous

SS, is debatable. It has been proposed as a specific entity char-

acterized by: (i) lobular panniculitis on histopathology, (ii) a

significant association with myelodysplasia or (iii) the good

response to treatment with steroids.37 However, some con-

sider it to be subcutaneous SS, in which pathological changes

are localized either within the adipose tissue or involving both

the dermis and the subcutaneous fat.38 Neutrophilic panniculi-

tis, being a broad descriptive term, encompasses differential

diagnoses including alpha-1 antitrypsin deficiency syndrome,

infections, pancreatitis, rheumatoid arthritis and factitial pan-

niculitis.38 From our findings, we feel that an underlying

haematological malignancy should be considered when there

is subcutaneous involvement in SS and/or anaemia.

SS is an intriguing dermatosis with protean clinical associa-

tions and histological findings which are constantly being

revised. In this study, we have illustrated and reinforced these

evolving concepts and variants of SS with patients from Singa-

pore. Clinical awareness and recognition of these unique

features are essential in the initiation of appropriate investiga-

tions and follow-up of these patients.

References

1 Sweet RD. An acute febrile neutrophilic dermatosis. Br J Dermatol1964; 76:349–56.

2 von den Driesch P. Sweet’s syndrome (acute febrile neutrophilicdermatosis). J Am Acad Dermatol 1994; 31:535–56.

3 Goh CL, Alora M, Kohar Y. Sweet’s syndrome in a skin clinic inSingapore: an epidemiology study. Ann Acad Med Singapore 1996;

25:222–7.4 Theng TSC, Chan YC, Leow YH, Tan SH. Sweet’s syndrome associ-

ated with Mycobacterium chelonae and herpes simplex virus infections:

a case report. Ann Acad Med Singapore 2003; 32:411–14.5 Su WPD, Liu HNH. Diagnostic criteria for Sweet’s syndrome. Cutis

1986; 37:167–74.6 Moschella SL. Neutrophilic dermatoses. In: Dermatology (Bolognia JL,

Jorizzo JL, Rapini RP, eds), 1st edn. London: Elsevier Limited,2003; 411–23.

7 Galaria NA, Junkins-Hopkins JM, Kligman D, James WD. Neutro-philic dermatosis of the dorsal hands: pustular vasculitis revisited.




484 Unusual clinical feature in Sweet syndrome, C.Y. Neoh et al.

8 Walling HW, Snipes CJ, Gerami P, Piette WW. The relationshipbetween neutrophilic dermatosis of the dorsal hands and Sweet

syndrome: report of 9 cases and comparison to atypical pyodermagangrenosum. Arch Dermatol 2006; 142:57–63.

9 DiCaudo DJ, Connolly SM. Neutrophilic dermatosis (pustular vas-culitis) of the dorsal hands: a report of 7 cases and review of the

literature. Arch Dermatol 2002; 138:361–5.10 Strutton G, Weedon D, Roberson I. Pustular vasculitis of the hands.

J Am Acad Dermatol 1995; 32:192–8.11 Weenig RH, Bruce AJ, McEvoy MT et al. Neutrophilic dermatosis of

the hands: four new cases and review of the literature. Int J Dermatol

2004; 43:95–102.12 Voelter-Mahlknecht S, Bauer J, Metzler G et al. Bullous variant of

Sweet’s syndrome. Int J Dermatol 2005; 44:946–47.13 Chan HL, Lee YS, Kuo TT. Sweet’s syndrome: clinicopathologic

study of eleven cases. Int J Dermatol 1994; 33:425–32.14 Cohen PR, Kurzrock R. Sweet’s syndrome and cancer. Clin Dermatol

1993; 11:149–57.15 Vazquez Garcia J, Almagro Sanchez M, Fonseca Capdevila E.

Multiple neutrophilic dermatoses in myelodysplastic syndrome.Clin Exp Dermatol 2001; 26:398–401.

16 Davies MG, Hastings A. Sweet’s syndrome progressing to pyoder-ma gangrenosum: a spectrum of neutrophilic skin disease in

association with cryptogenic cirrhosis. Clin Exp Dermatol 1991;16:279–82.

17 Cox NH, O’Brien HAW. Sweet’s syndrome associated with trans-retinoic acid treatment in acute promyelocytic leukemia. Clin Exp

Dermatol 1994; 19:51–2.18 Sitjas D, Cuatrecasas M, De Moragas JM. Acute febrile neutrophilic

dermatosis (Sweet’s syndrome). Int J Dermatol 1993; 32:261–8.19 Walker DC, Cohen PR. Trimethoprim-sulfamethoxazole-associated

acute febrile neutrophilic dermatosis: case report and review ofdrug-induced Sweet’s syndrome. J Am Acad Dermatol 1996; 34:918–

23.20 Shimizu T, Yoshida I, Eguchi H et al. Sweet syndrome in a child

with aplastic anaemia receiving recombinant granulocyte colony-stimulating factor. J Pediatr Hematol Oncol 1996; 18:282–4.

21 Ayirookuzhi SJ, Ma L, Ramshesh P, Mills G. Imatinib-inducedSweet syndrome in a patient with chronic myeloid leukaemia. Arch

Dermatol 2005; 141:368–70.22 Liu D, Seiter K, Mathews T et al. Sweet’s syndrome with CML cell

infiltration of the skin in a patient with chronic phase CML whiletaking imatinib mesylate. Leuk Res 2004; 28 (Suppl. 1):S61–3.

23 Scheinfeld N. Imatinib mesylate and dermatology part 2: a review

of the cutaneous side effects of imatinib mesylate. J Drugs Dermatol2006; 5:228–31.

24 Krause DS, Van Etten RA. Tyrosine kinases as targets for cancertherapy. N Engl J Med 2005; 353:172–87.

25 Choonhakam C, Chetchotisakd P, Jirarattanapochai K, Mootsika-pun P. Sweet’s syndrome associated with non-tuberculous myco-

bacterial infection: a report of five cases. Br J Dermatol 1998;139:107–10.

26 Chen HH, Hsiao CH, Chin HC. Successive development of cutane-ous polyarteritis nodosa, leucocytoclastic vasculitis and Sweet’s

syndrome in a patient with cervical lymphadenitis caused by Myco-bacterium fortuitum. Br J Dermatol 2004; 151:1096–100.

27 Florez A, Sanchez-Aguilar D, Roson E et al. Sweet’s syndrome asso-

ciated with Salmonella enteritidis infection. Clin Exp Dermatol 1999;24:239–40.

28 Zillikens D, Goldstein RK, Elsner P et al. Sweet’s syndrome associ-ated with Salmonella typhimurium infection. Acta Derm Venerol 1991;

71:77–9.29 Wong SSY, Wong KH, Hui WT et al. Differences in clinical and

laboratory diagnostic characteristics of penicilliosis marneffei inhuman immunodeficiency virus (HIV)- and non-HIV-infected

patients. J Clin Microbiol 2001; 39:4535–40.30 Cohen PR. Skin lesions of Sweet syndrome and its dorsal hand

variant contain vasculitis: an oxymoron or an epiphenomenon?Arch Dermatol 2002; 138:400–2.

31 Cohen PR, Kurzrock R. Sweet’s syndrome revisited: a review ofdisease concepts. Int J Dermatol 2003; 42:761–78.

32 Urano Y, Miyaoka Y, Kosaka M et al. Sweet syndrome associatedwith chronic myelogenous leukaemia: demonstration of leukaemic

cells within a skin lesion. J Am Acad Dermatol 1999; 40:275–9.33 Morgan KW, Callen JP. Sweet’s syndrome in acute myelogenous

leukaemia presenting as periorbital cellulitis with an infiltrate ofleukaemic cells. J Am Acad Dermatol 2001; 45:590–5.

34 Pozo J, Martinez W, Pazos JM et al. Concurrent Sweet’s syndrome andleukaemia cutis in patients with myeloid disorders. Int J Dermatol

2005; 44:677–80.35 Vignon-Pennamen MD, Aracting S. Sweet’s syndrome and leukae-

mia cutis: a common skin homing mechanism? Dermatology 2003;206:81–4.

36 Requena L, Kutzner H, Palmedo G et al. Histiocytoid Sweet syn-drome: a dermal infiltration of immature neutrophilic granulo-

cytes. Arch Dermatol 2005; 141:834–42.37 Sutra-Loubet C, Carlotti A, Guillemette J et al. Neutrophilic pannic-

ulitis. J Am Acad Dermatol 2004; 50:280–5.38 Cohen PR. Subcutaneous Sweet’s syndrome: a variant of acute feb-

rile neutrophilic dermatosis that is included in the histopathologic

differential diagnosis of neutrophilic panniculitis. J Am Acad Dermatol2005; 52:927–8.





Cutaneous side-effects in patients with rheumatic diseasesduring application of tumour necrosis factor-a antagonistsH-H. Lee, I-H. Song,* M. Friedrich, A. Gauliard,* J. Detert,* J. Rowert, H. Audring, S. Kary,* G.-R. Burmester,*W. Sterry and M. Worm

Department of Dermatology and Allergy and *Department of Rheumatology and Clinical Immunology, Charite-Universitatsmedizin Berlin, D-10117 Berlin,

Germany

CorrespondenceMargitta Worm.



Key wordsatopic eczema, biologics, psoriasis, rheumatic

disease, tumour necrosis factor-a antagonists

Conflicts of interestDuring the time of recording cutanous adverse

events under the patients’ treatment with TNF

antagonists, S.K. was employee of the Charite

University Medicine, Clinic for Rheumatology and

Clinical Immunology. Since September 2004, S.K.

has been working as contractor for Abbott GmbH

& Co. KG in Germany. This status has not

influenced the presentation of data or the discussion

in this article.

H.H.L. and I.H.S. contributed equally to this

study.

Summary

Background Patients with rheumatic diseases receiving antitumour necrosis factor(TNF)-a-based treatment may develop cutaneous reactions.Objectives To analyse the new onset or aggravation of skin lesions in patients witha rheumatic disease during treatment with TNF-a antagonists.Methods We conducted a prospective analysis of 35 of 150 patients with a longhistory of rheumatic disease, including rheumatoid arthritis, ankylosing spondy-litis (Bechterew’s disease) and psoriatic arthritis, to test for the development ofcutaneous manifestations during anti-TNF-a (infliximab, adalimumab or etaner-cept) treatment.Results Chronic inflammatory skin diseases such as psoriasis and eczema-like mani-festations represented the majority of cases (16 of 35). Cutaneous infectionscaused by viral, bacterial and fungal agents were also observed in many patients(13 of 35). Skin diseases such as dermatitis herpetiformis, leucocytoclastic vascu-litis and alopecia occurred in single cases only.Conclusions We observed a broad, diverse clinical spectrum with a majority ofchronic inflammatory and infectious skin diseases. However, we did not identifyindividual risk factors and a discontinuation of the anti-TNF-a treatment was notnecessary if adequate dermatological treatment was performed. The onset ofcutaneous side-effects in anti-TNF-a-based treatments should be determined bynationwide registries.

Rheumatic diseases are chronic inflammatory autoimmune dis-

orders and may lead to physical disability. The most common

form is rheumatoid arthritis (RA), which affects approximately

1% of adults worldwide.1 Other forms of arthritis include

ankylosing spondylitis (AS, Bechterew’s disease) and psoriatic

arthritis (PsA). Tumour necrosis factor (TNF)-a plays a central

role in the pathogenesis of rheumatic diseases and has become

a target molecular structure for antibody or TNF receptor

(TNF-R)-based treatment in the past few years. Several rand-

omized controlled clinical trials have been performed in which

biologics against TNF-a and/or its receptor showed a high

clinical efficacy and a good tolerability.2–7

The following biologics targeting the action of TNF-a have

been registered since 2002 by the U.S. Food and Drug Admin-

istration: infliximab (Remicade�; licensed by Centocor Inc.,

Horsham, PA, U.S.A.) is a chimeric IgG1j monoclonal anti-

body with approximately 10% mouse sequence in the variable

antigen-binding region. It binds with high affinity to the

soluble and transmembrane forms of TNF-a and inhibits

the binding of the cytokine to its receptor. By contrast,

adalimumab (Humira�; licensed by Abbott Laboratories,

Abbott Park, IL, U.S.A.) is a recombinant, fully human IgG1

monoclonal antibody that binds specifically to human TNF-aand neutralizes the activity of this cytokine. Etanercept

(Enbrel�; licensed by Immunex Corp., Thousand Oaks, CA,

U.S.A.) is a subcutaneously administered biological response

modifier and is a dimeric fusion protein consisting of the

extracellular ligand-binding portion of the human TNF-R. This

molecule is linked to the Fc portion of human IgG1 and pre-

vents the interaction of both TNF-a and TNF-b with their

respective receptors. Currently, these and other biologics

against TNF-a are in various stages of clinical studies and/or

are in the stage of approval in dermatology, especially for pso-

riasis and PsA.8 This indicates that anti-TNF-a antibodies are

effective for the immunological treatment of psoriasis. How-

ever, only approximately 30% of treated patients respond suf-

ficiently. It is not yet clear which are the responding subtypes.

On the other hand, infections, the occurrence of autoanti-

bodies and hypersensitivity reactions have been reported

during TNF-a-targeted treatment.9,10 Some recent single case



reports have demonstrated cutaneous side-effects in patients

with a rheumatic disease during anti-TNF-a treatment.11–15

Such reports included the onset of skin lesions such as hyper-

keratosis, eczema, blistering and dry skin. Also, local reactions

at the injection side (urticaria and pruritus) were detected

during clinical studies.16 However, larger studies evaluating

cutaneous side-effects in these patients are not available.

In this prospective analysis, we examined 35 patients with

an active rheumatic disease receiving anti-TNF-a antibodies.

The clinical work-up shows that a wide spectrum of chronic

inflammatory but also infectious skin disorders may develop

in patients receiving anti-TNF-a treatment.


In total, 150 adult patients who were sequential attendees at

the clinic and who had an active rheumatic disease were trea-

ted with different TNF-a antagonists in the Department of

Rheumatology and Clinical Immunology (Charite-Universitats-

medizin Berlin, Berlin, Germany). During the treatment, 35 of

150 patients developed skin eruptions (Table 1) and were

referred to the Department of Dermatology and Allergy

between November 2002 and November 2004.

A detailed history and dermatological examination was per-

formed in all 35 patients. Eleven of these patients underwent

a skin biopsy for histological examination, and seven had an

additional diagnostic work-up regarding type I and/or type IV

sensitizations.

Results

Examination of the 35 patients (18 women and 17 men, aged

19–76 years) receiving TNF-a antagonists revealed a broad

spectrum of skin diseases. These included the new onset or

Table 1 Details of the study population (n ¼ 35), including demographics, rheumatic disease, antitumour necrosis factor (TNF)-a treatment, and

time between treatment initiation and development of the skin eruptions

Patient Sex

Age

(years)

Rheumatic

disease

Anti-TNF-atreatment Skin eruption

Time after initiation of

anti-TNF-a treatment

New onset or

exacerbation

1 F 71 RA Etanercept Palmoplantar pustular psoriasis 10 months New onset2 F 44 RA Etanercept Psoriasis vulgaris 6 months New onset

3 M 39 RA Etanercept Eczematous skin lesion 12 months New onset4 F 41 AS Infliximab Atopic dermatitis-like eruption 1 month New onset

5 M 39 NC Adalimumab Psoriasis vulgaris 6 months Exacerbation6 F 49 RA Adalimumab Perivascular neutrophilic dermatitis 1 week New onset

7 M 45 PsA Infliximab Acute staphylococcal infection 6 months New onset8 F 74 RA Adalimumab Herpes simplex virus II infection 4 months New onset

9 M 64 RA Adalimumab Pityriasis versicolor 18 months New onset10 F 68 RA Infliximab Eczematous skin lesion 4 months New onset

11 M 19 AS Infliximab Psoriasis-like eruption 2 months New onset12 M 28 AS Infliximab Pityriasis versicolor 9 months New onset

13 F 23 RA Adalimumab Alopecia areata 7 months Exacerbation14 F 45 RA Adalimumab Microbial eczema 6 months New onset

15 F 62 RA Adalimumab Dermatitis herpetiformis 6 months Exacerbation16 M 51 AS Infliximab Eczematous skin lesion 10 months New onset

17 F 39 RA Adalimumab Androgenetic alopecia 8 months New onset18 F 44 RA Etanercept Tinea 6 months New onset

19 F 39 AS Infliximab Microbial eczema 6 months New onset20 M 63 RA Etanercept Psorasis vulgaris 2 weeks Exacerbation

21 F 34 AS Infliximab Microbial eczema 4 months New onset22 F 76 RA Adalimumab Pityriasis rosea 26 months New onset

23 M 63 RA Infliximab Dermatitis sicca 1 month New onset24 F 52 RA Etanercept Eczematous skin lesion 10 months New onset

25 F 54 RA Adalimumab Palmoplantar pustular psoriasis 8 months New onset26 F 27 AS Infliximab Psoriasis-like lesion 17 months New onset

27 M 45 AS Infliximab Palmoplantar pustular psoriasis 4 months New onset28 M 43 AS Infliximab Pityriasis versicolor 1 month New onset

29 M 46 AS Infliximab Eczematous skin lesion 2 weeks New onset30 M 56 RA Adalimumab Eczematous skin lesion 1 week New onset

31 M 34 AS Infliximab Eczematous skin lesion 10 months New onset

32 M 51 RA Infliximab Microbial eczema 11 months New onset33 M 57 AS Infliximab Microbial eczema 5 months New onset

34 M 64 AS Infliximab Tinea 10 months New onset35 F 40 RA Adalimumab Tinea 13 months New onset

AS, ankylosing spondylitis; NC, nonclassified; PsA, psoriatic arthritis; RA, rheumatoid arthritis.



Side-effects of TNF-a antagonists in rheumatic disease, H-H. Lee et al. 487

worsening of pre-existing chronic inflammatory and infectious

skin diseases (Fig. 1). An association between a particular

anti-TNF-a treatment and a particular skin disease was not

observed. Twenty of 35 patients had RA (14 women, six

men), 13 of 35 patients had AS (four women, nine men),

one man had PsA and one man had a nonclassified rheumatic

disease (Fig. 2).

Treatment modalities

The anti-TNF-a treatment modalities used to treat the patients

with rheumatic diseases are summarized in Table 2. Seventeen

patients received infliximab, two of these in combination with

methotrexate (MTX) and one in combination with lefluno-

mide (LEF). Twelve patients were treated with adalimumab,

one of these in combination with MTX, one with LEF, and

two with MTX and LEF together. The other six patients

received etanercept, one of these in combination with LEF.

Frequent occurrence of chronic inflammatory skin

diseases

A total of 16 of 35 patients presented with inflammatory skin

eruptions, with eight of 16 showing psoriasis-like skin lesions

(Fig. 3a) and eight of 16 showing eczema either as atopic der-

matitis (AD)-like eruptions or as eczematous skin lesions

(Fig. 3b).

Of the eight patients with psoriasis-like skin lesions, five

had the typical clinical pattern of psoriasis vulgaris and three

exhibited palmoplantar pustular psoriasis. Two of these eight

patients had a history of psoriasis vulgaris, and six had lesions

for the first time.

Of the eight patients with eczema, the skin lesions were

located primarily on the extremities within the antecubital and

popliteal fossae, but also in the face or trunk area. A moderate

to severe itch was always a concomitant symptom in these

patients.

Infectious skin diseases

Of the 35 patients, 13 presented with a new onset of an

infectious skin disease during anti-TNF-a antibody treat-

ment. Three of the 13 patients showed typical skin eruptions

16

13

6

0

4

8

12

16

20

Chronic inflammatoryskin disease

Infectious skin disease Other skin disease

Nu

mb

er o

f pat

ien

ts

Fig 1. Distribution of skin diseases (n ¼ 35). Chronic inflammatory

skin disease: psoriasis-like lesions (n ¼ 8), atopic dermatitis-like

eruptions (n ¼ 8). Infectious skin disease: pityriasis versicolor (n ¼ 3),

tinea corporis (n ¼ 3), microbial eczema (n ¼ 5), herpes simplex

virus infection (n ¼ 1), acute cutaneous staphylococcal superinfection

(n ¼ 1). Other skin disease (n ¼ 1, respectively): dermatitis

herpetiformis, pityriasis rosea, dermatitis sicca, alopecia areata,

androgenetic alopecia, leucocytoclastic vasculitis.

14

46

9

110

4

8

12

16

20

RA AS PsA NK

Nu

mb

er o

f pat

ien

ts

f m

Fig 2. Diagnosis and gender distribution in the study population

(n ¼ 35). RA, rheumatoid arthritits; AS, ankylosing spondylitis;

PsA, psoriatic arthritis; NK, nonclassified rheumatic disease.

Table 2 Dosage of tumour necrosis factor (TNF)-a antagonists as recommended by the manufacturer and current treatment of rheumatic diseaseof the study population (n ¼ 35)

TNF-a antagonist

Number and sex of patients with rheumatic disease

RA (n ¼ 20) AS (n ¼ 13) PsA (n ¼ 1) NC (n ¼ 1)

M F M F M F M F

Infliximab (Remicade�) 3 mg kg)1 i.v. (8 weeks) 2 1a 9 4b 1b – – –

Adalimumab (Humira�) 40 mg s.c. (2 weeks) 2c 9a,b – – – – 1c –Etanercept (Enbrel�) 2 · 25 mg s.c. (1 week) 2a 4 – – – – – –

AS, ankylosing spondylitis; i.v., intravenous; NC, nonclassified; PsA, psoriatic arthritis; RA, rheumatoid arthritis; s.c., subcutaneous.aOne patient in combination with leflunomide (LEF). bOne patient in combination with methotrexate (MTX). cOne patient in combinationwith MTX and LEF.



488 Side-effects of TNF-a antagonists in rheumatic disease, H-H. Lee et al.

resulting from the yeast Pityrosporum ovale, causing pityriasis

versicolor in all patients (Fig. 3c), whereas tinea corporis was

diagnosed in three of 13 patients. One of the 13 patients had

a reactivation with herpes simplex virus that appeared clinic-

ally as small vesicles on the right-upper abdomen. Five of the

13 patients developed a bacterial infection of the skin present-

ing clinically as microbial eczema, and one patient had an

acute cutaneous staphylococcal superinfection with acute

erythematous swelling of the right ear.

Other skin diseases

Six of the 35 patients showed various other skin eruptions.

One 39-year-old woman developed androgenetic alopecia and

telogen hair loss 8 months after commencing adalimumab

treatment. A 23-year-old woman developed alopecia areata

immediately after beginning LEF monotherapy; after the first

injection of adalimumab 1 year later, the disease activity

increased dramatically. A 62-year-old woman had dermatitis

herpetiformis, which was diagnosed many years before the

anti-TNF-a treatment but flared 6 months after initiation of

the treatment with adalimumab. The development of a leuco-

cytoclastic vasculitis was observed in a 49-year-old woman,

and a 76-year-old woman developed pityriasis rosea. Finally, a

63-year-old man had extremely dry skin that presented clinic-

ally as dermatitis sicca.

Discussion

TNF-a is a pleiotropic proinflammatory cytokine and plays,

among others, a central role in the pathogenesis of rheumatic

diseases. High levels of TNF-a have been found in the syn-

ovial fluid of patients with RA and AS,17–19 but also in those

with PsA and in psoriasis skin lesions.20

In this report we present 35 of 150 patients with rheumatic

disease receiving anti-TNF-a-based treatment experiencing the

onset of cutaneous lesions. The data show a broad spectrum

of skin eruptions. Patients with lesions typical of chronic

inflammatory skin diseases represent the majority, but infec-

tions and several other skin manifestations were also observed.

No allergic or toxic skin reaction occurred in the patients, but

instead a diverse clinical pattern that might be related to a

genetic predisposition.

Psoriasis and AD are common chronic inflammatory skin

disorders that are characterized by the infiltration of leuco-

cytes, the activation of skin-resident cells and an increased

production of numerous cytokines, chemokines and inflamma-

tory molecules. Nearly half (16 of 35) of the patients with

rheumatic diseases and skin eruptions developed either AD- or

psoriasis-like lesions.

There is strong evidence that psoriasis is an immune-related

disease.21 Many studies have argued for a key role of T cells

in the pathogenesis of psoriasis, but recent data also point to a

potential role of innate inflammatory triggers. Lymphocytes in

the lesions are mainly skin-homing CD4+ and CD8+ T cells

with an activated, memory/effector type-1 immunopheno-

type.22

Activated dendritic cells (DCs) and infiltrating T cells deter-

mine a type-1 inflammatory pathway that links the activation

of multiple inflammatory genes to the release of interferon

(IFN)-c from T cells, interleukin (IL)-12 from the activated

DCs and, as described recently, IL-23 from keratinocytes.23

(a)

(b)

(c)

Fig 3. (a) Patient with psoriasis-like skin lesions. (b) Patient with

eczematous eruptions in both popliteal fossae. (c) Patient with an

infectious skin disease (pityriasis versicolor) during treatment with

antitumour necrosis factor-a biologics.




TNF-a can stimulate inflammation in the type-1 pathway

through activating effects on DCs or by increasing transcrip-

tion of type-1 genes.

AD is a multifactorial and polygenetic skin disease charac-

terized by T helper 2 cells, mast cells and eosinophils produ-

cing IL-4, IL-5, IL-10 and IL-13.24 Acute eczematous skin

lesions show an early infiltration of CD4+, but also CD8+,

T cells in the dermis producing IL-4. In the late stage of

AD, IFN-c is the major cytokine triggering the chronic inflam-

mation.

The genetic background of psoriasis and AD has been exten-

sively studied and several susceptibility genes have been des-

cribed.25 Recent genome-wide linkage scans have identified

multiple loci linked to each other. An overlap between psori-

asis and AD susceptibility loci on chromosomes 1q21, 3q21,

17q25 and 20p12 suggest that some susceptibility factors are

found within genes or gene families with common effects on

epithelial immunity. However, many genes from these loci are

not yet identified or tested.26 Genes linked to psoriasis include

HLA-C, SLC9A3R1, NAT9, RAPTOR and SLC12A8; those linked to

AD include SPINK5, MS4A2 and PHF11.27

The identification of RUNX1 as an RA-associated gene has

been published recently.28 RUNX1 is a haematological tran-

scriptional regulator responsible for acute myelogenic leukae-

mia. On the other hand, one susceptibility factor for psoriasis

has been suggested to be the defective regulation of SLC9A3R1

(a phosphoprotein implicated, in association with other mole-

cules, in the regulation of membrane dynamism for synapse

formation and T-cell activation) or NAT9 (a new member

of the N-acetyl-transferase superfamily encoded by a still

unknown gene) by RUNX1. Thus, RUNX1 might establish a

genetic relationship between RA and psoriasis.

The other large group of patients (13 of 35) identified in

this study is characterized by the onset of cutaneous infectious

diseases during anti-TNF-a treatment. This observation is rela-

ted to the role of TNF-a in direct defence against infections.

Interestingly, fungal and viral skin infections were most fre-

quently observed in this study. However, these infectious skin

eruptions were mild and easily controlled by the appropriate

pharmacological treatment.

Recently, the use of TNF-a antagonists has been associated

with new cases of active tuberculosis worldwide.29,30 It is

known that TNF-a plays a central role in the cellular defence

mechanisms by monocytes/macrophages against infection

caused by Mycobacterium tuberculosis. Treatment with anti-TNF-aagents suggests they pose a threat of increased risk of onset of

not only tuberculosis but also of other (opportunistic) infec-

tions. However, owing to the different mode of action and

the specific pharmacodynamics of the TNF-a antagonists, the

role of the actual risk regarding infectious diseases may be not

as expected. This assumption has to be proven by future clin-

ical studies.

The other skin diseases developed by six of 35 patients

appeared only as single cases. We documented one case each

of dermatitis herpetiformis, pityriasis rosea, dermatitis sicca,

leucocytoclastic vasculitis, androgenetic alopecia and alopecia

areata, respectively. In regard to alopecia areata, it is known

that TNF-a induces the loss of hair; therefore, the paradoxical

side-effect of development of alopecia areata following anti-

TNF-a treatment is still a matter of discussion.31

Several other reports have also referred to adverse skin reac-

tions during treatment with anti-TNF-a antagonists. Interest-

ingly, all cases occurred during the application of infliximab.

Devos et al. described six patients receiving infliximab for RA

or Crohn’s disease, four of whom developed leucocytoclastic

vasculitis, lichenoid drug reaction, acute folliculitis or superfi-

cial granuloma, respectively, and two developed perniosis-like

eruptions.15 Vergara et al. reported four patients (three with

RA, one with AS) with an interface dermatitis pattern.13 In

addition, a positive patch test with infliximab suggested

T-cell-specific sensitization. AD-like eruptions were noticed by

Wright:14 in two patients, the eczematous skin lesions

occurred with accentuation in the antecubital fossae and were

assessed as nonallergic reactions to the infliximab therapy.

Chan et al. reported one patient with acute pustular dermatitis

following infliximab infusions, accompanied with necrosis of

the subcutaneous tissue, a systemic group A b-haemolytic

streptococcal infection and other severe systemic illness.11 The

diagnosis of necrotizing fasciitis was confirmed. Furthermore,

Kent et al. referred to one patient with RA who developed bul-

lous skin lesions that were neither typical of pemphigus vul-

garis nor of bullous pemphigoid despite a common clinical

pattern.12 Finally, Sfikakis et al. recently reported five patients

(three of them with a rheumatic disease) who developed pso-

riasis induced by anti-TNF-a treatment.31

Taken together, our data show that a wide range of differ-

ent skin lesions might occur during anti-TNF-a treatment.

These are based on either the exacerbation of a pre-existing

skin disease or indicate the new onset of the skin disease.

Chronic inflammatory or infectious skin diseases were

most frequently observed. However, these diseases are also

frequent in the general population. If a patient receiving

anti-TNF-a treatment is referred to a dermatologist, any skin

disease should be considered and a thorough work-up is

necessary to determine the exact diagnosis. In none of our

patients was a discontinuation or termination of the anti-

TNF-a treatment necessary and all skin diseases were treated

regularly.

Finally, our data indicate the need for long-term observa-

tions of patients treated with anti-TNF-a agents, for instance

through the introduction of nationwide registries.

References

1 Gabriel SE. The epidemiology of rheumatoid arthritis. Rheum Dis ClinNorth Am 2001; 27:269–81.

2 Maini R, St Clair EW, Breedveld F et al. Infliximab (chimeric anti-tumour necrosis factor alpha monoclonal antibody) vs. placebo in

rheumatoid arthritis patients receiving concomitant methotrexate: arandomised phase III trial. ATTRACT Study Group. Lancet 1999;

354:1932–9.3 Weinblatt ME, Kremer JM, Bankhurst AD et al. A trial of etanercept,

a recombinant tumor necrosis factor receptor: Fc fusion protein, in



490 Side-effects of TNF-a antagonists in rheumatic disease, H-H. Lee et al.

patients with rheumatoid arthritis receiving methotrexate. N Engl JMed 1999; 340:253–9.

4 Bathon JM, Martin RW, Fleischmann RM et al. A comparison ofetanercept and methotrexate in patients with early rheumatoid

arthritis. N Engl J Med 2000; 343:1586–93.5 Lipsky PE, van der Heijde DM, St Clair EW et al. Infliximab and

methotrexate in the treatment of rheumatoid arthritis. Anti-TumorNecrosis Factor Trial in Rheumatoid Arthritis with Concomitant

Therapy Study Group. N Engl J Med 2000; 343:1594–602.6 Mease PJ, Goffe BS, Metz J et al. Etanercept in the treatment of pso-

riatic arthritis and psoriasis: a randomised trial. Lancet 2000;

356:385–90.7 Furst DE, Schiff MH, Fleischmann RM et al. Adalimumab, a fully

human anti tumor necrosis factor-alpha monoclonal antibody, andconcomitant standard antirheumatic therapy for the treatment of

rheumatoid arthritis: results of STAR (Safety Trial of Adalimumabin Rheumatoid Arthritis). J Rheumatol 2003; 30:2563–71.

8 Friedrich M, Philipp S, Sabat R et al. Use of biologicals in psoriasis.Z Rheumatol 2003; 62:439–49.

9 Weisman MH. What are the risks of biologic therapy in rheuma-toid arthritis? An update on safety. J Rheumatol Suppl 2002; 65:33–8.

10 Braun J, Sieper J. Role of novel biological therapies in psoriaticarthritis: effects on joints and skin. BioDrugs 2003; 17:187–99.

11 Chan AT, Cleeve V, Daymond TJ. Necrotising fasciitis in a patientreceiving infliximab for rheumatoid arthritis. Postgrad Med J 2002;

78:47–8.12 Kent PD, Davis JM III, Davis MD, Matteson EL. Bullous skin lesions

following infliximab infusion in a patient with rheumatoid arthri-tis. Arthritis Rheum 2002; 46:2257–8.

13 Vergara G, Silvestre JF, Betlloch I et al. Cutaneous drug eruption toinfliximab: report of 4 cases with an interface dermatitis pattern.

Arch Dermatol 2002; 138:1258–9.14 Wright RC. Atopic dermatitis-like eruption precipitated by

infliximab. J Am Acad Dermatol 2003; 49:160–1.15 Devos SA, van den Bossche N, de Vos M, Naeyaert JM. Adverse

skin reactions to anti-TNF-alpha monoclonal antibody therapy.Dermatology 2003; 206:388–90.

16 Schaible TF. Long term safety of infliximab. Can J Gastroenterol 2000;14:29C–32C.

17 Braun J, Bollow M, Neure L et al. Use of immunohistologic andin situ hybridization techniques in the examination of sacroiliac

joint biopsy specimens from patients with ankylosing spondylitis.Arthritis Rheum 1995; 38:499–505.

18 Saxne T, Palladino MA Jr et al. Detection of tumor necrosis factoralpha but not tumor necrosis factor beta in rheumatoid arthritis

synovial fluid and serum. Arthritis Rheum 1988; 31:1041–5.19 Chu CQ, Field M, Feldmann M, Maini RN. Localization of tumor

necrosis factor alpha in synovial tissues and at the cartilage-pannusjunction in patients with rheumatoid arthritis. Arthritis Rheum 1991;

34:1125–32.20 Asadullah K, Docke WD, Volk HD, Sterry W. The pathophysio-

logical role of cytokines in psoriasis. Drugs Today 1999; 35:913–24.

21 Bos JD, de Rie MA, Teunissen MB, Piskin G. Psoriasis: dysregula-

tion of innate immunity. Br J Dermatol 2005; 152:1098–107.22 Friedrich M, Krammig S, Henze M et al. Flow cytometric charac-

terization of lesional T cells in psoriasis: intracellular cytokineand surface antigen expression indicates an activated, memory/

effector type 1 immunophenotype. Arch Dermatol Res 2000; 292:519–21.

23 Piskin G, Sylva-Steenland RM, Bos JD, Teunissen MB. In vitro and insitu expression of IL-23 by keratinocytes in healthy skin and psori-

asis lesions: enhanced expression in psoriatic skin. J Immunol 2006;176:1908–15.

24 Leung DY, Bieber T. Atopic dermatitis. Lancet 2003; 361:151–60.25 Werfel T, Morita A, Grewe M et al. Allergen specificity of skin-

infiltrating T cells is not restricted to a type-2 cytokine pattern inchronic skin lesions of atopic dermatitis. J Invest Dermatol 1996;

107:871–6.26 Cookson WO, Ubhi B, Lawrence R et al. Genetic linkage of child-

hood atopic dermatitis to psoriasis susceptibility loci. Nat Genet2001; 27:372–3.

27 Bowcock AM, Cookson WO. The genetics of psoriasis, psoriaticarthritis and atopic dermatitis. Hum Mol Genet 2004; 13:R43–55.

28 Helms C, Cao L, Krueger JG et al. A putative RUNX1 binding sitevariant between SLC9A3R1 and NAT9 is associated with suscepti-

bility to psoriasis. Nat Genet 2003; 35:349–56.29 Keane J, Gershon S, Wise RP et al. Tuberculosis associated with inf-

liximab, a tumor necrosis factor alpha-neutralizing agent. N Engl JMed 2001; 345:1098–104.

30 Gardam MA, Keystone EC, Menzies R et al. Anti-tumour necrosisfactor agents and tuberculosis risk: mechanisms of action and clin-

ical management. Lancet Infect Dis 2003; 3:148–55.31 Sfikakis PP, Iliopoulos A, Elezoglou A. Psoriasis induced by anti-

tumor necrosis factor therapy – a paradoxical adverse reaction.Arthritis Rheum 2005; 52:2513–18.





Changes in skin barrier function following long-termtreatment with moisturizers, a randomized controlled trialI. Buraczewska,*� B. Berne,* M. Lindberg,� H. Torma* and M. Loden�

*Department of Medical Sciences, Dermatology and Venereology, University Hospital, Uppsala, Sweden

�ACO HUD NORDIC AB, Box 622, SE-194 26 Upplands Vasby, Sweden

�Department of Medicine, Occupational and Environmental Dermatology, Karolinska Institute and Occupational and Environmental Health, Stockholm Centre for

Public Health, Stockholm County Council, Stockholm, Sweden

CorrespondenceIzabela Buraczewska.


Accepted for publication2 October 2006

Key wordscapacitance, epidermis, long-term treatment,

moisturizers, skin barrier function, transepidermal

water loss

Conflicts of interestI.B. and M.L. are employed by ACO HUD

NORDIC AB, which provided funding for the

research presented in this article. No other authors

have anything to declare.

Summary

Background Moisturizers are commonly used by patients with dry skin conditionsas well as people with healthy skin. Previous studies on short-term treatmenthave shown that moisturizers can weaken or strengthen skin barrier function andalso influence skin barrier recovery. However, knowledge of the effects on skinbarrier function of long-term treatment with moisturizers is still scarce.Objectives To investigate the impact of long-term treatment with moisturizers onthe barrier function of normal skin, as measured by transepidermal water loss(TEWL) and susceptibility to an irritant, and to relate those effects to the compo-sition of the designed experimental moisturizers.Methods Volunteers (n ¼ 78) were randomized into five groups. Each group treat-ed one volar forearm for 7 weeks with one of the following preparations:(i) one of three simplified creams, containing only a few ingredients in order tominimize the complexity of the system; (ii) a lipid-free gel; (iii) one ordinarycream, containing 5% urea, which has previously been shown to decrease TEWL.The lipids in the simplified creams were either hydrocarbons or vegetable trigly-ceride oil, and one of them also contained 5% urea. After 7 weeks, treated andcontrol forearms were exposed for 24 h to sodium lauryl sulfate (SLS) using apatch test. TEWL, blood flow and skin capacitance of both SLS-exposed andundamaged skin were evaluated 24 h after removal of patches. Additionally, a24-h irritancy patch test of all test preparations was performed on 11 volunteersin order to check their possible acute irritancy potential.Results Changes were found in the barrier function of normal skin after 7 weeksof treatment with the test preparations. The simplified creams and the lipid-freegel increased TEWL and skin response to SLS, while the ordinary cream had theopposite effect. One of the simplified creams also decreased skin capacitance. Alltest preparations were shown to be nonirritant, both by short-term irritancypatch test and by measurement of blood flow after long-term treatment.Conclusions Moisturizers influence the skin barrier function of normal skin, asmeasured by TEWL and susceptibility to SLS. Moreover, the effect on skin barrierfunction is determined by the composition of the moisturizer. The ingredientswhich influence the skin barrier function need to be identified, and the mechan-ism clarified at the molecular level.

Moisturizers are commonly used by patients with dry skin

conditions, and their application is also a widespread habit

among people with healthy skin. However, there is still a

lack of knowledge about the effects of moisturizers on skin

barrier function, especially after long-term treatment. Studies

conducted so far, both in healthy and diseased skin, show

that some moisturizers may weaken skin barrier func-

tion—observed as changes in transepidermal water loss

(TEWL) and susceptibility to irritants—while others may

strengthen it.1–6 Moisturizers can also influence skin barrier

recovery.1,7 It can be assumed that differences in composi-

tion account for the differences in effect. Lipid content and



natural moisturizing factors such as urea have been suggested

to be most important.

The aim of the present study was to investigate the impact

of long-term treatment with moisturizers on the barrier func-

tion of normal skin, as measured by TEWL, and susceptibility

to an irritant, and to relate those effects to the composition of

the moisturizers used. The moisturizers chosen for the study

were one ordinary cream, which has previously been shown

to decrease TEWL,1 and three simplified creams containing

only a few ingredients in order to minimize the complexity of

the system.

The lipids used in the simplified creams were either pure

hydrophobic hydrocarbons derived from mineral oil (paraffin

and isohexadecane) or a more polar vegetable oil consisting of

triglycerides and sterols (canola oil), in order to see the differ-

ence between two contrasting types of lipids. Urea was added

to one composition.

The duration of the treatment was chosen to be 7 weeks,

which is longer than the application time in the majority of

similar studies performed to date. An irritancy patch test of all

test preparations was also performed in order to check their

possible acute irritancy potential.


Test preparations

Five test preparations were used in the study: one ordinary

cream, hereafter called Complex cream; three simplified

creams emulsified with a long-chain polymer, with a lipid

phase consisting of either the hydrocarbons isohexadecane and

paraffin (Hydrocarbon cream) or a vegetable triglyceride oil,

canola oil (Canola cream and Canola/urea cream); and one

lipid-free gel consisting of the polymer used in simplified

creams (Polymer gel). The polymeric emulsifier in the simpli-

fied creams and the Polymer gel was acrylates/C10-30 alkyl

acrylate crosspolymer. The creams were oil-in-water (o/w)

emulsion, and all test preparations had their pH adjusted to

� 5. The Canola/urea cream also contained urea. Table 1 pre-

sents detailed compositions along with the number of volun-

teers testing each preparation.

Participants and experimental design of the long-term

treatment study and the irritancy test

The study consisted of two parts: a long-term treatment study

with the test preparations and an irritancy test.

Volunteers (n ¼ 78; 58 women and 20 men, aged 25–

60 years) participated in the long-term treatment study, and

11 volunteers (eight women and three men, aged 27–

46 years) in the irritancy test. Both studies were double-blind

and randomized. Informed consent was obtained from all

volunteers and the study was approved by the Regional Ethi-

cal Review Board (Uppsala, Sweden). Volunteers were not

allowed to use any skin care products on their forearms, nei-

ther during the test period nor in the three preceding days.

In the long-term treatment study, volunteers treated one

volar forearm twice daily for 7 weeks with one test prepar-

ation, leaving the other forearm untreated to serve as the con-

trol. The treatment side was randomized. After 7 weeks, on

day –1, both volar forearms, treated and control, were exposed

to sodium lauryl sulfate (SLS) using a patch test: 50 lL of 1%

aqueous solution of SLS (purity ‡ 99%, Sigma-Aldrich, St

Louis, MO, U.S.A.) in large Finn chambers, Ø 12 mm (Epitest

Ltd OY, Tuusula, Finland).8 The chambers were fixed to the

skin on the volar aspect of the forearm with adhesive tape

Table 1 Composition of test preparations and numbers of volunteers testing them

Test preparation Lipids Emulsifiers Urea Other ingredients pHNo. ofvolunteers

Complex creama 20%: Capric/caprylic

triglyceride, canola oil,cetearyl alcohol,

paraffin, glycerylstearate

1Æ3%: PEG-100 stearate,

carbomer, polysorbate 60

5% Water, propylene glycol,

glyceryl polymethacrylate,dimethicone, sodium

lactate, methylparaben,propylparaben,

lactic acid, citric acid

5 15

Hydrocarbon

cream

40%: Isohexadecaneb

(20%), paraffinc (20%)

0Æ4%: Acrylates/C10-30 alkyl

acrylate crosspolymerd0% Water 5 16

Canola cream 40%: Canola oile 0Æ4%: Acrylates/C10-30 alkyl


Canola/urea

cream

40%: Canola oild 0Æ4%: Acrylates/C10-30 alkyl


Polymer gel 0% 0Æ4%: Acrylates/C10-30 alkyl

acrylate crosspolymerd0% Water, methylparabenf 5 16

aCanoderm� kram 5%, ACO HUD NORDIC AB, Stockholm, Sweden; bArlamol HD, Uniqema, Gouda, the Netherlands; cMerkur White OilPharma, Merkur Vaseline, Hamburg, Germany; dPemulen TR-2, Noveon Inc., Cleveland, OH, U.S.A.; eAkorex L, Karlshamns AB, Karlshamn,

Sweden; fNipagin M, Clariant, Clariant International, Pontypridd, U.K.



Long-term treatment with moisturizers, I. Buraczewska et al. 493

(Scanpore, Norgeplaster, Oslo, Norway). They were removed

after 24 h, on day 0. On removal of the patches, the skin was

rinsed gently with water and patted dry with a paper towel.

Evaluations were performed on treated and control forearms

24 h after removal of patches (day 1) on the SLS-exposed skin

area and on an area of undamaged (non-SLS-exposed) skin a

few centimetres below it. SLS-exposed skin and undamaged

skin areas were assessed for transepidermal water loss (TEWL)

and blood flow, and undamaged skin was also evaluated for

skin capacitance.

Volunteers participating in the irritancy test had the five test

preparations, water, and 1% aqueous solution of SLS (pur-

ity ‡ 99%, Sigma-Aldrich) applied on the skin on the volar

aspect of forearms for 24 h to assess their acute irritancy

potential using a patch test: 50 lL in large Finn chambers,

Ø 12 mm (Epitest Ltd OY).8 The chambers were fixed to the

skin with adhesive tape (Scanpore, Norgeplaster). They were

removed after 24 h, and the skin was then rinsed gently with

soap and water and patted dry with a paper towel.

Visual scoring was performed 1 h and 24 h after removal

of patches, and TEWL was measured 24 h after removal of

patches.

Evaluations

TEWL was quantified using DermaLab (Cortex Technology,

Hadsund, Denmark) according to published guidelines for

TEWL measurements.9 Skin electrical capacitance was meas-

ured with a Corneometer 820 (Courage and Khazaka GmbH,

Cologne, Germany). Cutaneous blood flow was measured with

a laser Doppler flowmeter (Periflux PF1d, Perimed, Stockholm,

Sweden)10 equipped with a special multifibre probe (PF 113

integrating probe, Perimed).

Visual scoring ranged from 0 to 4 according to the Euro-

pean Society of Contact Dermatitis (ESCD) guidelines on clin-

ical scoring of acute SLS irritant reactions,8 where erythema,

oedema and scaling were taken into consideration. The scale

was as follows: 0 ¼ negative: no damage; 0Æ5 ¼ doubtful:

very weak erythema or minute scaling; 1 ¼ weak: weak ery-

thema, slight oedema, slight scaling and/or slight roughness;

2 ¼ moderate: moderate degree of erythema, oedema, scaling

and/or roughness; 3 ¼ strong: marked degree of erythema,

oedema, scaling, roughness; 4 ¼ very strong/caustic: as 3,

with necrotic areas.

Calculations and statistics

The results are shown using boxplots (Figs 1–3). The box is

defined by the upper (Q3) and lower (Q1) quartiles and the

median is marked by a subdivision of the box. The whiskers

are the lines that extend from the top and bottom of the box

to the lowest and highest observations that are inside the

region defined by the following limits: lower limit, Q1 – 1Æ5(Q3–Q1); upper limit, Q3 + 1Æ5 (Q3–Q1). Outliers are points

outside of the lower and upper limits and are plotted with

asterisks.

In long-term treatment with moisturizers, results from

measurements of TEWL, skin capacitance and blood flow of

treated forearms are presented as a percentage ratio to the cor-

responding values obtained from control forearms, which

serve as 100%.

Moreover, relative increases of TEWL in SLS-exposed skin

area in comparison with TEWL in undamaged skin on the

same forearm were calculated for treated and untreated fore-

arms to estimate the difference in skin reactivity to SLS after

treatment, according to the following formula: relative increase

(%) ¼ (vSLS/vUND) · 100, where vSLS represents the value

measured on SLS-exposed skin and vUND represents the corres-

ponding value from undamaged skin on the same forearm.

In the irritancy test, results from visual scoring are

tabulated as frequencies. The percentage of total possible

score (% TPS) was calculated as follows: % TPS ¼ (0 · n0 +

0Æ5 · n0Æ5 + 1 · n1 + 2 · n2 + 3 · n3 + 4 · n4)/[4 · (n0 +

n0Æ5 + n1 + n2 + n3 + n4)] · 100, where ni represents the

Complex creamPolymer gelCanola/urea creamCanola creamHydrocarbon cream

200

175

150

125

100

75

50

TEWL

Capacitance

P =

0·0

18

P =

0·0

11

P =

0·0

07

P =

0·4

14

P =

0·0

16

P =

0·5

72

P =

0·0

17

P =

0·4

53

P =

0·0

06

P =

0·1

67

Fig 1. TEWL and skin capacitance of

undamaged skin treated with test preparations

for 7 weeks. Values are presented as

percentage ratio of the values obtained from

the corresponding control areas, which serve

as 100%. P-values relate to differences from

control areas. The results are presented as box

plots with the median value as a line across

the box and the first quartile value at the

bottom (Q1) and the third at the top (Q3).

The whiskers are lines that extend from the

top and bottom of the box to the lowest and

the highest observation within a defined

region, with outliers plotted as asterisks

outside this region. Reference line is given at

100%. n ¼ 15 for Canola cream and Complex

cream and 16 for Canola/urea cream,

Hydrocarbon cream and Polymer gel.



494 Long-term treatment with moisturizers, I. Buraczewska et al.

number of volunteers with a visual score of i. TEWL is presen-

ted as a percentage ratio of the corresponding value obtained

from water.

An explorative data analysis was performed using a Wil-

coxon signed rank test on paired data to analyse differences

between results of TEWL, visual scoring, skin capacitance and

blood flow for all test preparations. A Kruskal–Wallis test of

equality of medians was used to investigate the differences

between the simplified creams.

Minitab� statistical software, Release 14 for Windows

(Minitab Inc., State College, PA, U.S.A.), was used for calcula-

tions and plots. The level of significance was set at P < 0Æ05.

Results

Treatment of normal skin for 7 weeks with Hydrocarbon

cream, Canola cream, Canola/urea cream and Polymer gel

increased TEWL in comparison with the untreated controls,

while treatment with Complex cream, conversely, decreased

TEWL. In addition, skin capacitance was decreased in skin

treated with Hydrocarbon cream, in comparison with the con-

trol, while no difference was found for the remaining test

preparations (Fig. 1).

No signs of irritation measured as changes in blood flow

were found on the areas treated with the test preparations in

comparison with controls (data not shown).

After exposure to SLS, skin treated with Hydrocarbon

cream, Canola cream, Canola/urea cream and Polymer gel had

significantly higher TEWL than SLS-exposed skin on untreated

forearms, while skin treated with Complex cream had lower

TEWL than control (Fig. 2). However, the relative increase in

TEWL after SLS exposure was the same in skin treated with

Hydrocarbon cream, Canola cream, Polymer gel and Complex

cream as in untreated control skin. Skin treated with Canola/


1000

900

800

700

600

500

400

300

200

100

Treated

Control

P = 0·394 P = 0·478 P = 0·002 P = 0·587 P = 0·094

Fig 3. Relative increase of TEWL after SLS

exposure in comparison with TEWL of

undamaged skin area on the same forearm,

calculated for the skin areas treated with test

preparations and untreated control areas. For

explanation of the boxplots, see Figure 1.


350

300

250

200

150

100

50

0

TEWL

Blood flow

P =

0·0

09

P =

0·0

36

P =

0·0

21

P =

0·0

33

P =

0·7

12

P =

0·0

01

P =

0·0

01

P =

0·0

45

P =

0·0

44

P =

0·0

34

Fig 2. Transepidermal water loss (TEWL) and

blood flow of sodium lauryl sulfate (SLS)-

exposed skin treated with test preparations,

in relation to control areas (not treated with

moisturizers but also exposed to SLS). For

explanation of the graph, see Figure 1.




urea cream had a higher relative increase of TEWL than the

untreated control skin (Fig. 3).

Blood flow closely followed the pattern of TEWL, except in

skin treated with Polymer gel, where blood flow did not differ

significantly from the control. In contrast to other test prepa-

rations, skin treated with Complex cream had lower blood

flow after SLS exposure than the corresponding untreated fore-

arm (Fig. 2).

The Kruskal–Wallis test revealed no differences between the

simplified creams (Hydrocarbon cream, Canola cream and

Canola/urea cream) in their effects on TEWL of undamaged

skin (P ¼ 0Æ685), TEWL of SLS-exposed skin (P ¼ 0Æ687) and

blood flow of SLS-exposed skin (P ¼ 0Æ833). However, there

was a difference in impact on skin capacitance (P ¼ 0Æ017).

The occlusive irritancy patch test showed that Canola cream,

Canola/urea cream, Hydrocarbon cream and Polymer gel were

no more irritant than water, evaluated both by visual scoring

and TEWL measurements. Hydrocarbon cream produced elevat-

ed visual scoring 24 h after removal of patches, but not after

1 h. Moreover, Complex cream was less irritant than water in

visual scoring 1 h after removal of patches, and showed no

difference from water at 24 h, either in visual scoring or

TEWL. SLS, used as a reference, gave significantly higher

results than water (Table 2).

Discussion

In this study we found changes in the barrier function of nor-

mal skin after 7 weeks of treatment with test preparations, as

observed by changes in TEWL and skin response to SLS.

Simplified creams with different types of lipids (Hydro-

carbon cream, Canola cream and Canola/urea cream), as well

as Polymer gel, increased TEWL after long-term treatment.

After exposure to SLS, skin treated with these test preparations

had higher TEWL compared with untreated control skin.

Blood flow was also elevated, except for Polymer gel. There

were no differences between the three simplified creams in

impact on TEWL and blood flow. However, looking at the

relative increase of TEWL after SLS exposure in treated skin, a

significant increase was found only for Canola/urea cream. In

accordance with previous findings, treatment with Complex

cream had the opposite effect, with a decreased TEWL and

skin response to SLS.1 However, there was no difference from

control in the relative increase of TEWL in comparison with

untreated skin after SLS exposure.

Skin capacitance was decreased after usage of Hydrocarbon

cream, and remained unchanged, compared with controls, for

all other test preparations. This result suggests that the

increased TEWL is caused by impairment of the skin barrier

and not by elevated hydration of the stratum corneum.11 Test

preparations were shown to be not irritant, both by short-

term irritancy patch test and by measurement of blood flow

after long-term treatment.

Our findings are in accordance with previously performed

studies in which 2- or 3-week treatments with moistur-

izers influenced skin barrier function. Held et al.3 showed Tab

le2

Irrita

tion

pote

ntia

lof

test

prep

arat

ions

,gi

ven

asre

sults

ofvi

sual

scor

ing

and

tran

sepi

derm

alw

ater

loss

(TEW

L).Visua

lsc

orin

gha

sa

cate

gory

scale

from

0to

4.N

umbe

rof

volu

ntee

rsw

ith

the

resp

ective

scor

ean

dpe

rcen

tage

tota

lpo

ssib

lesc

ore

(TPS

)is

give

n.TEW

Lat

the

24-h

read

ing

ispr

esen

ted

asa

perc

enta

gera

tio

tore

sults

obta

ined

for

wat

er.M

edian

valu

ean

dfirs

tan

dth

ird

quar

tile

sar

egi

ven

(n¼

11).

P-va

lues

relate

todi

ffer

ence

from

wat

er.

Scor

e

Wat

erH

ydro

carb

oncr

eam

Can

ola

crea

mCan

ola/

urea

crea

mCom

plex

crea

mPo

lym

erge

l1%

SLS

1h

24h

1h

24h

1h

24h

1h

24h

1h

24h

1h

24h

1h

24h

02

104

43

117

1111

113

110

0

0Æ5

61

17

40

20

00

40

42

13

06

04

02

00

04

01

62

00

00

00

00

00

00

63

30

00

00

00

00

00

00

04

00

00

00

00

00

00

00

TPS

(%)

13Æ6

1Æ1

14Æ8

8Æ0

13Æ6

06Æ

80

00

13Æ6

034

Æ129

Æ5P-

valu

e0Æ

800

0Æ03

61Æ

000

1Æ00

00Æ

091

1Æ00

00Æ

009

1Æ00

01Æ

000

1Æ00

00Æ

033

0Æ00

4

TEW

L,%

(Med

ian,

Q1–

Q3)

–10

0–

97Æ6

(92Æ

7–10

2Æ2)

–10

1Æ4

(89Æ

0–11

2Æ3)

–94

Æ4(9

2Æ9–

102Æ

7)–

98Æ2

(91Æ

4–10

5Æ2)

–10

3Æ1

(90Æ

3–11

3Æ1)

–46

3Æ1

(298

Æ7–6

07Æ7

)

P-va

lue

––

–0Æ

398

–0Æ

760

–0Æ

307

–0Æ

683

–0Æ

563

–0Æ

004




impairment of skin barrier function by a cream with a high

lipid content (70%); 4 weeks of treatment increased skin sus-

ceptibility to SLS but did not change TEWL of undamaged

skin. The same cream also increased susceptibility to SLS after

only 5 days of treatment,12 and to nickel in nickel-allergic vol-

unteers after 7 days of treatment.13 The opposite results were

found by Loden et al.,1,2,4,5 who showed that 2- or 3-week

treatments of normal or atopic skin with moisturizers contain-

ing urea either decreased TEWL and susceptibility to SLS or

had no effect, but did not impair the skin barrier function.

The ordinary cream with 20% lipids was chosen as it has

been shown to decrease TEWL in short-term studies.1 A num-

ber of studies have shown that moisturizers with high lipid

content can affect barrier function.3,12,13 In the simplified

creams, we used as high a quantity of lipids as possible in the

formulations; with the emulsifier chosen, the maximum

amount turned out to be 40%. Two kinds of lipid were used

in the simplified creams, either pure hydrophobic hydrocar-

bons derived from mineral oil (paraffin and isohexadecane) or

a more polar vegetable oil consisting of triglycerides and

sterols (canola oil), in order to compare the effects of two

contrasting types of lipid. These creams were stabilized with a

polymeric emulsifier, which is believed not to penetrate and

influence skin barrier function due to its large molecular

size.14 In contrast, the Complex cream contained a mixture of

various fatty substances, of both vegetable and mineral origin,

and was emulsified with a combination of polymers and low-

molecular-weight emulsifiers. It also contained urea, which

was shown in a recent placebo-controlled study to reduce

TEWL and decrease skin susceptibility to SLS.2 Because the

Complex cream itself decreased TEWL,1 the same amount of

urea was incorporated into one of the simplified creams. In

addition, the polymeric emulsifier in a gel formulation was

used as a reference.

In the present study, the test preparations interfered with

the normal function or structure of the skin. The mechanism

of their action is not known, but there are several possibilities

to consider, and they can have both direct and indirect effects

on the skin barrier, stratum corneum.

Increased TEWL and dryness can be caused by exposure to

irritants. However, the patch test of the test preparations did

not reveal increased irritation compared with water. Moreover,

there were no signs of irritation, as measured with blood

flow, after 7 weeks of usage.

Water is the main component in the majority of moisturiz-

ers, and water itself may contribute to dryness and increased

TEWL. Prolonged contact with water disrupts the intercellular

lipid lamellae structure in the stratum corneum.15 Although

water evaporates quickly after application,16,17 the effect on

the skin barrier of 7 weeks of a twice-daily exposure is

unknown.

Models of the skin barrier (for review see18) emphasize that

barrier function depends on the structural organization of the

stratum corneum. Lipids in moisturizers may remain on the

skin surface or enter the skin,19 and more physiological lipids

may penetrate into the epidermis and affect skin barrier struc-

ture and recovery.20–22 The three simplified creams investi-

gated in this study, all of which had a negative effect on the

skin barrier, contained high proportions of lipids, 40%, but

the type of lipid was not of importance for the effect. At the

same time, Complex cream with a lower level of lipids, 20%,

produced the opposite result. The results for the simplified

creams may thus be explained by the impact of lipids, but it

does not explain the impairment of the skin barrier obtained

after treatment with Polymer gel.

Several emulsifiers have structures similar to the intercellular

bilayer lipids, and it is conceivable that such emulsifiers might

influence the organization of these lipids.14 As the impairment

of the skin barrier was also seen after application of the Poly-

mer gel, this negative effect could well be due to the emulsi-

fying polymer used in the simplified creams and the Polymer

gel, acrylates/C10-30 alkyl acrylate crosspolymer. The possible

impact of monomers, which may be present in low concentra-

tion, on lipid structures must be taken under consideration.

However, the Complex cream also contained a polymeric

emulsifier, carbomer, which is similar in structure to acry-

lates/C10-30 alkyl acrylate crosspolymer, and emulsifiers of

lower molecular weight—cetearyl alcohol, polysorbate 60,

glyceryl stearate and PEG-100 stearate—without producing a

negative impact on the skin barrier.

Humectants such as urea, glycerine and propylene glycol

may penetrate the skin.23,24 It has been suggested that the

decreased TEWL and lower response to SLS after treatment

with the Complex cream is due to its urea content.1,5 How-

ever, the addition of urea to the Canola/urea cream did not

improve skin barrier function, and even made the skin more

susceptible to SLS. The urea might have favoured the absorp-

tion of potentially damaging ingredients in the formulation or

allowed for increased penetration of SLS during the 24-h

exposure. Other studies with moisturizers containing urea

also showed various effects on skin barrier (reviewed in Loden

et al.5,25,26).

Formation of the constituents of the skin barrier, as well as

its regulation, is controlled by epidermal enzymes. Among

these, special attention is drawn towards lipid-processing

enzymes, enzymes taking part in the formation of the corni-

fied envelope of corneocytes, and enzymes which are involved

in the process of desquamation.27–31 If moisturizers or their

ingredients affected the activity or turnover of these enzymes,

they could influence skin barrier function.

Even though changes in TEWL and susceptibility to irritants

after use of moisturizers have been shown by a number of

studies,1–6 this subject has not yet been investigated thor-

oughly. To our knowledge, our study is the first to examine

the effect of moisturizers on normal skin after such a long

period of application. Moreover, our test preparations were

designed to minimize complexity and thus facilitate analysis.

We have shown that moisturizers influence the skin barrier

function of normal skin, as measured by TEWL and suscepti-

bility to SLS. Moreover, that influence depends on the compo-

sition of the moisturizer. A complex interaction of lipids and

other components of the cream with the structure of the




stratum corneum, and effects on keratinocytes, can be

assumed. Changes in TEWL following SLS exposure are prob-

ably due to an increased (or decreased) bioavailability of SLS

due to impaired barrier function.

These results raise many questions about the nature of the

observed influence. More knowledge about the effect of mois-

turizers on skin and its barrier function is needed, in order to

improve treatment of various skin disorders and assist in the

design of preparations to target skin problems efficiently and

safely.

Acknowledgments

This study was supported by ACO HUD NORDIC AB, Stock-

holm, Sweden. We would like to thank Carl Wessman and

Gudrun Artz for the great help they have provided.

References

1 Loden M. Barrier recovery and influence of irritant stimuli in skintreated with a moisturizing cream. Contact Dermatitis 1997; 36:256–

60.2 Loden M, Barany E, Mardahl P et al.. The influence of urea treatment

on skin susceptibility to surfactant-induced irritation: a placebo-controlled and randomized study. Exog Dermatol 2004; 3:1–6.

3 Held E, Sveinsdottir S, Agner T. Effect of long-term use of moistur-izer on skin hydration, barrier function and susceptibility to

irritants. Acta Derm Venereol 1999; 79:49–51.4 Loden M, Andersson AC, Lindberg M. Improvement in skin barrier

function in patients with atopic dermatitis after treatment with amoisturizing cream (Canoderm). Br J Dermatol 1999; 140:264–7.

5 Loden M, Kuzmina N, Nyren M et al. Nickel susceptibility and skinbarrier function to water after treatment with a urea-containing

moisturizer. Exog Dermatol 2004; 3:99–105.6 Ganemo A, Virtanen M, Vahlquist A. Improved topical treatment of

lamellar ichthyosis: a double-blind study of four different creamformulations. Br J Dermatol 1999; 141:1027–32.

7 Held E, Lund H, Agner T. Effect of different moisturizers on SLS-irritated human skin. Contact Dermatitis 2001; 44:229–34.

8 Tupker RA, Willis C, Berardesca E et al. Guidelines on sodium lau-ryl sulfate (SLS) exposure tests. A report from the Standardization

Group of the European Society of Contact Dermatitis. Contact Derma-titis 1997; 37:53–69.

9 Pinnagoda J, Tupker RA, Agner T et al. Guidelines for transepider-mal water loss (TEWL) measurement. A report from the Standard-

ization Group of the European Society of Contact Dermatitis.Contact Dermatitis 1990; 22:164–78.

10 Bircher A, de Boer EM, Agner T et al. Guidelines for measurementof cutaneous blood flow by laser Doppler flowmetry. A report

from the Standardization Group of the European Society of ContactDermatitis. Contact Dermatitis 1994; 30:65–72.

11 Loden M, Lindberg M. Product testing—testing of moisturizers.

In: Bioengineering of the Skin: Water and the Stratum Corneum. (Elsner P,

Berardesca E, Maibach HI, eds), Boca Raton: CRC Press, Inc.,1994; 275–89.

12 Held E, Agner T. Effect of moisturizers on skin susceptibility toirritants. Acta Derm Venereol 2001; 81:104–7.

13 Zachariae C, Held E, Johansen JD et al. Effect of a moisturizer onskin susceptibility to NiCl2. Acta Derm Venereol 2003; 83:93–7.

14 Barany E, Lindberg M, Loden M. Unexpected skin barrier influencefrom nonionic emulsifiers. Int J Pharm 2000; 195:189–95.

15 Warner RR, Stone KJ, Boissy YL. Hydration disrupts human stra-tum corneum ultrastructure. J Invest Dermatol 2003; 120:275–84.

16 Blichmann CW, Serup J, Winther A. Effects of single application

of a moisturizer: evaporation of emulsion water, skin surfacetemperature, electrical conductance, electrical capacitance, and

skin surface (emulsion) lipids. Acta Derm Venereol 1989; 69:327–30.

17 Loden M. The increase in skin hydration after application of emol-lients with different amounts of lipids. Acta Derm Venereol 1992;

72:327–30.18 Norlen L. Molecular skin barrier models and some central prob-

lems for the understanding of skin barrier structure and function.Skin Pharmacol Appl Skin Physiol 2003; 16:203–11.

19 Ghadially R, Halkier-Sorensen L, Elias PM. Effects of petrolatum onstratum corneum structure and function. J Am Acad Dermatol 1992;

26:387–96.20 Mao-Qiang M, Brown BE, Wu-Pong S et al. Exogenous nonphysio-

logic vs. physiologic lipids. Divergent mechanisms for correctionof permeability barrier dysfunction. Arch Dermatol 1995; 131:809–

16.21 Wertz PW, Downing DT. Metabolism of topically applied fatty acid

methyl esters in BALB/C mouse epidermis. J Dermatol Sci 1990;1:33–7.

22 Man MQ, Feingold KR, Elias PM. Exogenous lipids influencepermeability barrier recovery in acetone-treated murine skin. Arch

Dermatol 1993; 129:728–38.23 Wellner K, Wohlrab W. Quantitative evaluation of urea in stratum

corneum of human skin. Arch Dermatol Res 1993; 285:239–40.24 Brinkmann I, Muller-Goymann CC. An attempt to clarify the influ-

ence of glycerol, propylene glycol, isopropyl myristate and a com-bination of propylene glycol and isopropyl myristate on human

stratum corneum. Pharmazie 2005; 60:215–20.25 Loden M. Role of topical emollients and moisturizers in the treat-

ment of dry skin barrier disorders. Am J Clin Dermatol 2003; 4:771–88.

26 Loden M. Do moisturizers work? J Cos Derm 2004; 2:141–9.

27 Elias PM. Stratum corneum defensive functions: an integrated view.J Invest Dermatol 2005; 125:183–200.

28 Candi E, Schmidt R, Melino G. The cornified envelope: a model ofcell death in the skin. Nat Rev Mol Cell Biol 2005; 6:328–40.

29 Zeeuwen PL. Epidermal differentiation: the role of proteases andtheir inhibitors. Eur J Cell Biol 2004; 83:761–73.

30 Rawlings AV, Harding CR. Moisturization and skin barrier func-tion. Dermatol Ther 2004; 17 (Suppl. 1): 43–8.

31 Rawlings AV, Matts PJ. Stratum corneum moisturization at themolecular level: an update in relation to the dry skin cycle. J Invest

Dermatol 2005; 124:1099–110.





Oral treatment with probiotic Lactobacillus johnsoniiNCC533 (La1) for a specific part of the weaning periodprevents the development of atopic dermatitis inducedafter maturation in model mice, NC/NgaR. Inoue, A. Nishio, Y. Fukushima* and K. Ushida

Laboratory of Animal Science, Kyoto Prefectural University, Shimogamo, Sakyo-ku, Kyoto 606-8522, Japan

*Nutrition Business Group, Nestle Japan Manufacturing, Ltd, Shinagawa, Tokyo 140-0002, Japan

CorrespondenceDr Kazunari Ushida.



Key wordsatopic dermatitis prevention, faecal IgA,

immunomodulation, Lactobacillus johnsonii

NCC533, probiotics


Summary

Background The inhibitory effect of probiotic bacteria on atopic dermatitis has beenshown in human infants, but the mechanism is still unclear.Objective This study aimed to show the effects of the administration of a probioticduring the weaning period in mouse models on production of the intestinalsecretory IgA (sIgA) and on the development of atopic dermatitis (AD) in laterlife.Methods The effects of the administration of Lactobacillus johnsonii NCC533 (La1) dur-ing weaning were evaluated using a mouse model (Balb/c). The weaning periodof mice was divided into three phases according to the evolution of faecal IgA.La1 was administered in each phase and the evolution of the faecal IgA was esti-mated. In the next experiment, the effect of the administration of La1 in phase 2on host immunity after maturation was further assessed by using the modelNC/Nga mouse for human AD.Results Administration of La1 in each phase showed a distinct effect on the pro-duction of sIgA. But sIgA production was only positively stimulated when La1was administrated in phase 2. The development of AD induced by mite antigenfrom 6 weeks old was significantly prevented by the primary administration ofLa1 in phase 2. AD-like lesions were significantly milder than those of the con-trol mice, and histological observations showed an almost normal appearance ofthe epidermis and upper dermis of the mice treated with La1.Conclusion This study suggested that the primary administration of La1 in a specificpart of the weaning period is effective in preventing or inhibiting the develop-ment of AD after maturation by modulating or accelerating the gut immuneresponse.

The intestinal microbiota of monogastric animals is comprised

of more than 500 distinct bacterial species, which play an im-

portant role in host nutrition and health by promoting the

nutrient supply, and shaping and maintaining normal mucosal

immunity.1 Evidence that the intestinal microbiota contributes

significantly to host physiological functions, in particular,

immunological functions, has been derived from studies on

germ-free and conventional animals. The gut immune system

of germ-free mice is certainly maldeveloped compared with

that of conventional mice.2,3

The intestinal microbiota is quantitatively the most import-

ant source of microbial stimulation on the mucosal immune

system and presumably provides a primary signal for driving

the postnatal maturation of the immune system;4 thus, it is

suggested that the intestinal microbiota is significantly associ-

ated with the development of both food allergies and other

immune diseases.5 The bacterial stimulation derived from the

colonization of intestinal commensal bacteria is certainly

required for driving Th2-type skewed neonatal immunity

towards the Th1-type and it is a crucial step in the develop-

ment of immunity for the prevention of allergies.6 However,

bacterial transmission and colonization have been highly

reduced as a result of modern high standards for hygiene with

regard to the neonate, and the high prevalence of allergies in

developed countries has been attributed to this phenomenon

for several years (hygiene hypothesis).7



In support of the significance of the intestinal microbiota in

regard to allergies, Sudo et al.8 reported that germ-free mice

failed to establish oral tolerance, whereas Bifidobacterium infantis

mono-associated mice developed oral tolerance normally. They

also reported the interesting finding that the reconstitution of

the intestinal microbiota of germ-free mice by B. infantis was

effective only in the early life of mice but not in old age. This

suggested that early life is extremely important for the intro-

duction of bacterial antigen stimulation in order to shape neo-

natal immunity to a correct balance of cytokines.

In a previous study, our findings agreed with this concept,

indicating that there is a period around weaning that is pre-

sumably critical for the maturation of the host immune sys-

tem in terms of immune deviation.9 In this period, the

amount of luminal secretory IgA (sIgA), which mediates the

primary immunological defence line in the mucosal immune

system, decreased drastically and remained at a low level for

several days. On the other hand, the diversity of the intesti-

nal microbiota increased remarkably in this period, as if

responding to the decrease of amount of the sIgA. In this

period, a very large number of bacterial antigens are pre-

sumably transported to immune inductive sites, such as

Peyer’s patches and thus the mucosal immune system may

be actively stimulated.

From these observations, we proposed that this period

could provide an opportunity for the development of toler-

ance against allergies by stimulating and/or modulating the

host immune system via the gut immune system. The oral

administration of the probiotic Lactobacillus johnsonii NCC533

(La1) was selected for this purpose. A stronger immunomod-

ulatory effect was expected from this strain than from other

probiotic strains because, unlike most other probiotics,

this strain has the ability to attach to intestinal epithelial cells

by supporting lipoteichoic acid and the elongation factor

Tu (EF-Tu), both of which strongly stimulate immune

functions.10–12

One objective of this study was to find an efficient way to

stimulate the immune system and/or to modulate it with the

administration of the probiotic La1 in early life using a mouse

model; the evolution of faecal IgA was used as a reference

marker in the development of gut immune systems. Another

objective was to estimate the efficacy of La1 administration in

early life as a means to prevent allergies in maturity. In

experiment 1, an effort was made to determine the most effi-

cient time during the weaning period to administer La1. The

NC/Nga mouse, the model mouse for atopic dermatitis

(AD),13 was used in experiment 2 to estimate the effects of

the administration of La1 on the prevention or inhibition of

AD.


Experiment 1

Definition of the experimental phases for the administration. From the

evolution of faecal IgA in mice, which was demonstrated in a

previous study,9 three experimental phases for the administra-

tion of La1 were defined (Fig. 1a). A period when the mice

are aged 18–20 days represented phase 1. Mice from 20–

22 days old and from 22–24 days old were in phase 2 and

phase 3, respectively.

Animals. Six pregnant BALB/c mice were obtained from a

commercial supplier (Japan Clea Co., Ltd., Osaka, Japan).

Pups delivered from these mice were used as the donors of

fresh faeces. Five pups were selected randomly from each

litter, and a total of 30 pups were divided into six experi-

mental groups (five pups in each group). They were

housed under conventional circumstances in our facilities

and not separated until weaning at 21 days of age. After

weaning, male and female mice were housed separately and

fed on a standard pellet diet (CE-2, Japan Clea Co. Ltd).

Preparation of Lactobacillus johnsonii NCC533. Lactobacillus johnsonii

NCC533 (La1) was obtained from the Nestle Japan Co., Ltd.

(Tokyo, Japan). La1 was precultured overnight at 37 �C with

5 mL of anaerobic GAM broth (Nissui, Tokyo, Japan) supple-

mented with 1% glucose. Then, 200 lL of the precultured

La1 was transferred into the fresh medium, incubated at

37 �C for 20 h, and centrifuged at 10 000 g and 4 �C for

1 min. The bacterial pellet of La1 was washed twice with

phosphate-buffered saline (PBS) and collected by centrifuga-

tion in the same manner as described above. The pellet was

resuspended in 1 mL of PBS and used immediately for admin-

istration. This procedure was repeated every day during the

administration period.

Administration of Lactobacillus johnsonii NCC533. Five pups from

one of the six experimental groups were gavaged with

100 lL each of resuspended La1 (1Æ5 · 1011 cfu) on all

3 days of phase 1 and designated as the P1-La1 group

(Fig. 1a). Five other pups from the remaining experimental

groups were administered PBS in phase 1 and designated as

the P1-control group. This method was repeated in phases

2 and 3, and the P2-La1, P2-control, P3-La1 and P3-control

groups were established. The dose was always administered

at 10 a.m.

Sample collection and measurement of total IgA. Fresh faeces were col-

lected from all 30 pups at ages 18–24, 27, 30, 33, 36 and

40 days (total 360 samples). The IgA concentration in these

samples was measured as described elsewhere.8 Briefly, the

faeces were homogenized in PBS and their supernatants

were subjected to an enzyme-linked immunosorbent assay

(ELISA) using a mouse IgA ELISA Quantitation kit (Bethyl

Laboratories, Montgomery, TX, U.S.A.).

Induction of allergic diarrhoea. At 6 weeks of age, all the mice in

this experiment were systemically primed by ovalbumin

(OVA) for the induction of allergic diarrhoea as described in

Kweon et al.14 From the 7th to 10th intubation, the incidence

of diarrhoea was evaluated for 2 h after intubation.



500 Primary probiotic La1 for AD prevention, R. Inoue et al.

Experiment 2

Animals and administration of Lactobacillus johnsonii NCC533. As the

administration in phase 2 showed a particular effect on

the faecal IgA evolution in experiment 1, the administra-

tions were performed in phase 2 in experiment 2. Six preg-

nant NC/Nga mice were obtained from a commercial

supplier (Japan SLC, Shizuoka, Japan). Pups delivered from

these mice were used in this experiment. Pups were divided

into two groups and administered either La1 or PBS in

phase 2 as described in experiment 1 (Fig. 1b). They were

housed and reared as described in experiment 1 but were

fed another commercial mouse chow different from experi-

ment 1 (Labo MR stock; Nihon Nosan Kogyo, Tokyo,

Japan).

Collection of faeces and measurement of faecal IgA. The collection of

fresh faeces and measurement of faecal IgA were performed as

described in experiment 1.

Induction of human atopic dermatitis-like lesion. At 6 weeks of age, all

mice were shaved to remove the hair from the upper back.

Twenty micrograms of antigen from human mites (Mite

Extract Dermatophagoides farinae, LSL, Tokyo, Japan) diluted in

200 lL of PBS was swabbed onto the shaved area three times

a week using a cotton swab (Fig. 1b). Shaving the hair from

the region where the antigen was applied was performed dur-

ing each of the applications. This method is according to

Matsuoka et al.15 The values estimated by the following meth-

ods were evaluated for males and females separately because

Matsuoka et al.15 demonstrated that there was a sex difference

in the symptoms of the AD-like lesion.

Evaluation of skin lesions. In a preliminary experiment, it was esti-

mated that, in the conditions under which the animals were

housed in this experiment, the symptoms of AD-like lesions

began to develop from the 7th to the 9th application of the

mite antigen. Therefore, the severity of the induced skin

lesions was scored from the 7th application of the mite antigen.

20 22

Weaning

0 21 40

Faecal sample collection period

Phase 1

Phase 2

Phase 3

Day after birth 18 20

PBS

La1

20 22

PBS

La1

P1-control

P1-La1

0 21 4018

0 21 4018 22 24

PBS

La1

Weaning

Weaning

P2-control

P2-La1

P3-control

P3-La1

Control

La1

0 21 4018

PBS

La1

42 87

Weaning

Faecal sample collection period Mite antigen application period(three times per week)

Experiment 1

Experiment 2

Day after birth

dissection

(a)

(b)

Fig 1. Experimental schedule for

(a) experiment 1 and (b) experiment 2. The

administration period for each phase is shown

with a heavy line, and the shaded region

indicates the weaning period (from days 18

to 24 after birth). The deltas indicate the

application of mite-extracted antigen on the

back of NC/Nga mice. Each experimental

group was treated with either PBS or La1

(1Æ5 · 1011 cfu), as indicated with arrows.



Primary probiotic La1 for AD prevention, R. Inoue et al. 501

Scoring was performed as described in Matsuoka et al.15 by a

volunteer who was unaware of the experiment details.

Scratching behaviour. Scratching behaviour was observed after

each application of the mite antigen. It was noted from the

7th application as well as the scoring of skin lesions. Points

were scored as described in Matsuoka et al.15

Dissection and sample collection. The mite antigen was applied for

5 weeks (a total of 15 times). At the time when a 16th application

would have been administered, no mite antigen was applied;

instead the mice were anaesthetized, killed and dissected. Our pre-

liminary study indicated that the AD-like lesions were the most

severe at the time of the 15th application of the mite antigen.

Each mouse was anaesthetized by an intraperitoneal injec-

tion of 50 lL of pentobarbital sodium (Schering-Plough,

Osaka, Japan). The region to which antigen had been applied

was photographed using a Nikon Cool Pix 990 camera

(Nikon, Tokyo, Japan). The skin from that region was then

collected and fixed with 10% neutral formalin for the histo-

logical analysis. Venous blood was collected from the inferior

vena cava with a 21-gauge needle and a 1-mL syringe.

Measurement of total IgE in serum. The concentration of total IgE in

serum was measured using a YAMASA IgE EIA Kit (YAMASA,

Tokyo, Japan) according to the manufacturer’s instructions.

Estimation of the IgG1/IgG2a ratio in serum. The concentrations of

total IgG1 and IgG2a in serum were measured using the

Mouse IgG1 and IgG2a ELISA kit (Bethyl Laboratories) accord-

ing to the manufacturer’s instruction.

Histological analyses of antigen-applied skin. The fixed skin was

embedded in paraffin, and 3-lm thick serial paraffin sections

were prepared. Sections were stained with either haematoxylin

and eosin or toluidine blue. The number of mast cells stained

with toluidine blue was counted from the upper dermis to the

dermal muscle in a field of 2 · 2 mm under a microscope.

Eight fields were randomly selected for counting the mast

cells, and the average density mm)2 was estimated.


A normality test and F-test were performed to assess the nor-

mality of the data of faecal IgA on each experimental day

using Statcel (Seiunsha Inc., Tokyo, Japan) with Excel 2003

for Windows (Microsoft, Seattle, U.S.A.). When significance

was detected by either test, the Mann–Whitney U-test was per-

formed (control vs. the group that received La1 on each day)

with Statview for Windows (SAS Institute Inc., Cary, NC,

U.S.A.). When normality of the faecal IgA data was confirmed,

a Student’s t-test was performed with Excel 2003.

Skin scores, scratching scores and the IgG1/G2a ratio were

analysed with the Mann–Whitney U-test (control vs. the group

receiving La1 on each day). The number of mast cells was ana-

lysed by the Student’s t-test (control vs. the group receiving La1).

Results

Experiment 1

Effect of the administration of La1 in the three phases during weaning on the

evolution of faecal IgA. The P1-La1 group had a lower amount of

faecal IgA throughout the experimental period after the admin-

istrations (Fig. 2a). On day 19, the amount of faecal IgA in P1-

La1 was significantly lower than that in the P1-control

(0Æ395 ± 0Æ013 vs. 0Æ584 ± 0Æ087 mg g)1 faeces; P < 0Æ01).

Afterwards, the amount in each group decreased rapidly until

weaning. The faecal IgA in the P1-control group showed a

minimum amount on day 22 (0Æ042 ± 0Æ011 mg g)1 faeces),

and that in P1-La1 on day 23 was (0Æ044 ± 0Æ017 mg g)1 fae-

ces). After showing the minimum amount, the amount of faecal

IgA began to recover in both groups. However, the recovery of

the faecal IgA amount occurred more rapidly in the P1-control

group than in the P1-La1 group. The faecal IgA amount in the

P1-control was significantly higher from day 27 onward until

the end of the experimental period (0Æ704 ± 0Æ032 vs.

0Æ34 ± 0Æ099 mg g)1 faeces on day 40; P < 0Æ01).

The evolution of faecal IgA in the P2-La1 group was different

from that in the other seven groups, including the P2-control

(Fig. 2b). Although the P2-control showed a significantly

higher amount of faecal IgA than P2-La1 on day 21, the faecal

IgA amount in P2-La1 was higher than that in the P2-control

for 4 days after the last La1 administration and was signifi-

cantly higher on days 22, 23 and 27. Unlike the other seven

groups, the evolution of the faecal IgA in the P2-La1 group

did not show a rapid decrease and recovery during the wean-

ing period.

Both the P3-La1 and control groups showed a similar

faecal IgA evolution during the experimental period

(Fig. 2c). The P3-La1 group tended to show a slightly higher

faecal IgA amount than the P3-control group after the admin-

istrations, but a significant difference was detected only on

day 40 (1Æ613 ± 0Æ108 vs. 1Æ086 ± 0Æ133 mg g)1 faeces;

P < 0Æ05).

Induction of allergic diarrhoea. Despite the high incidence of allergic

diarrhoea reported by Kweon et al.,14 whose study showed

100% induction by the 9th intubation, diarrhoea after 10

intubations of OVA was induced on only five of all 30 mice

in this experiment. Only two of those five mice showed severe

symptoms similar to those described by Kweon et al.14 The

shape of the distal colon contents was evaluated by dissection,

as in Kweon et al.;14 however, all mice that did not exhibit

symptoms maintained a healthy shape of their intestinal con-

tents, as did untreated mice (images not shown).

Experiment 2

The evolution of faecal IgA. Figure 3 shows the evolution of faecal

IgA in NC/Nga mice. Male mice in the La1 group tended to

have a higher amount of faecal IgA from days 27 to 33 than

those in the control group, and it was significantly higher on




day 27 (0Æ145 ± 0Æ075 vs. 0Æ031 ± 0Æ020 mg g)1 faeces;

P < 0Æ01). Female mice of the La1 group showed a very simi-

lar evolution of faecal IgA to that of P2-La1. In the female La1

group, the decrease of faecal IgA stopped after the first admin-

istration, while the faecal IgA of the control group kept

decreasing until day 22. The amount of faecal IgA in the

female La 1 group tended to be higher than that of the control

female group throughout the experimental period, and signi-

ficant differences were detected on days 22 and 27

(0Æ037 ± 0Æ039 vs. 0Æ001 ± 0Æ001 mg g)1 faeces on day 22,

P < 0Æ05; 0Æ076 ± 0Æ064 vs. 0Æ028 ± 0Æ065 mg g)1 faeces on

day 27, P < 0Æ05).

Appearance of an atopic dermatitis-like lesion. Almost all male and

female control mice had symptoms of AD-like lesions, which

were represented as skin scores (Fig. 4). Excoriation and

haemorrhage were seen on the region to which antigen was

applied in six of eight male mice (75%) and eight of nine

female mice (89%) in the control group but on two of 11

male (18%) and two of five female mice (40%) in the La1

0

18 20 22 24 26 28 30 32 34 36 38 40

18 20 22 24 26 28 30 32 34 36 38 40

18 20 22 24 26 28 30 32 34 36 38 40

50

100

150

200

250

300(a)

(b)

(c)

0

50

100

150

200

250

300

0

50

100

150

200

250

300

Day after birth

*

*

* * * * *

***

*

*

Fig 2. Effect of oral La1 administration on the evolution of faecal

IgA from suckling to maturity. (s), La1-administered mice (n ¼ 5).

(d), PBS-administered mice (n ¼ 5). The oral administration was

conducted in (a) phase 1, (b) phase 2, and (c) phase 3. Data are shown

by the percentage mean value for the average IgA amount on day 18.

Bars represent the standard deviation (*P < 0Æ05, La1 vs. control).

0

50

100

150

200

250

300

0

50

100

150

200

250

300

*

*

Male

Female

*

*

Day after birth

18 20 22 24 26 28 30 32 34 36 38 40

18 20 22 24 26 28 30 32 34 36 38 40

Fig 3. Effect of oral La1 administration on the evolution of

faecal IgA in male and female NC/Nga mice from suckling to

maturity. (s), La1-treated mice (male: n ¼ 8, female: n ¼ 9).

(d), PBS-administered mice (male: n ¼ 11, female: n ¼ 5). The oral

administration was conducted in phase 2. Data are shown by the

percentage mean value for the average IgA amount on day 18. Bars

represent the standard deviation (*P < 0Æ05, La1 vs. control).




group, respectively (Fig. 4). Only one male mouse in the La1-

administered group showed lesions as severe as those of mice

in the control group. The skin scores of the La1 group were

clearly lower than those of the control group on every day

they were estimated, except for the score of male mice on the

10th day of application.

Scratching behaviour. Scratching scores were evaluated to clarify

the frequency of scratching behaviour. Male mice of the La1

group had lower scratching scores than those of the control

group throughout the evaluation period, except for the scores

on the 8th day of application (from the 7th to the 15th appli-

cations; data not shown). The scratching score of the control

(b) Male Lal

(c) Female control

(d) Female La1

(a) Male control

Fig 4. Skin features of the region treated with

antigen in NC/Nga mice after 15 applications.

(a) Control male mice; the male mice

administered PBS in phase 2 (days 20–22

after birth), (b) La1 male mice; the male

mice administered La1 in phase 2, (c) Control

female mice; the female mice administered

PBS in phase 2, (d) La1 female mice; the

female mice administered La1 in phase 2.




male mice increased gradually and peaked at the 12th applica-

tion (26Æ5 ± 17Æ7). It then decreased to 17Æ8 ± 18Æ3 at the

13th application and remained around 18 afterwards. The

scratching score of La1 male mice remained around 15

throughout the evaluation period, except at the 7th, 9th and

15th applications (7th: 7Æ5 ± 5Æ7, 9th: 7Æ8 ± 9Æ1, 15th:

10Æ8 ± 11Æ9).

The scratching scores of La1 female mice were slightly

higher than those of control female mice from the 7th to

10th applications (Fig. 5). However, the scratching scores of

control female mice drastically increased from the 10th to the

11th applications (from 6Æ7 ± 4Æ2 to 21Æ9 ± 10Æ0). They con-

tinued to increase afterwards, although a slight decrease was

observed at the 15th application. As for the La1 female mice,

the scratching score increased gradually, but it was lower than

that of the control female mice from the 10th application to

the end of the evaluation.

IgE concentration and IgG1/G2a ratio in serum. IgE in some serum

samples, especially the samples from the La1 group, was not

detectable even when serum diluted 10-fold less was used for

the assay (Table 1). The concentration of serum IgE detected

tended to be higher in the control group than in the La1

group, although its value varied widely from mouse to

mouse. IgE was detected in five of eight mice (62%), and the

mean value of the concentration was 447Æ5 ± 612Æ2 ng mL)1

except for undetectable samples. In contrast to this, it was

detected in only two of 11 mice (18%) of the La1 group

(71Æ1 and 15Æ0 ng mL)1). In female mice, serum IgE was

detected in almost all control mice (eight of nine mice;

89%), and the mean value of the concentration was

391Æ6 ± 342Æ2 ng mL)1 except for one undetected sample.

The concentrations of serum IgE from La1 female mice,

which were detected in two of five mice (40Æ0%), were 84Æ8and 56Æ8 ng mL)1.

The IgG1/G2a ratio in male mice of both control and La1

showed very similar values (2Æ37 ± 1Æ45 vs. 2Æ30 ± 1Æ54).

However, the IgG1/G2a ratio in female mice was significantly

different between the control and the La1 group. Female mice

in the control group showed a very high IgG1/G2a ratio

(5Æ68 ± 6Æ62), whereas the ratio of the La1 female mice was

1Æ82 ± 0Æ62 (P < 0Æ05).

Histological analysis of the skin treated with antigen. Figure 6 shows

the histological status of antigen-treated skin of representative

individuals from each experimental group. A thickened upper

dermis and separation of the epidermis layer were observed in

both male and female mice in the control group. However,

the treated skin of almost all mice in the La1 group had a

practically normal histological status. There was no specific

sign of AD on the treated skin of the La1 group and the

untreated NC/Nga mice.

Mast cell density of the skin to which antigen was applied. The number

of mast cells in the treated region was counted using toluidine

blue staining, and the densities of mast cells in each group

were estimated. From the male mice in the control group,

97Æ2 ± 33Æ2 cells mm)2 of treated skin were detected. In the

case of the La1 male mice, a lower number of mast cells than

in the control males were detected (88Æ0 ± 29Æ0 cells mm)2,

P > 0Æ05 vs. control male mice. The La1 female mice had a

significantly lower density of mast cells than did the female

control mice. The mast cell density of La1 female mice was

91Æ2 ± 19Æ7 cells mm)2, while that of female mice in the

control group was 145Æ9 ± 49Æ7 cells mm)2 (P < 0Æ05; con-

trol vs. La1 group).

0

10

20

30

40

50

60

7 8 9 10 11 12 13 14 15

*

Antigen application

e r o c s g n i h c t a r c S

Fig 5. Scratching scores of NC/Nga female mice from the 7th to the

15th mite antigen application. The scratching scores were evaluated

according to Matsuoka et al.15 Data are shown by the mean value, and

bars represent the standard deviation. (s), La1 treated mice. (d), PBS

treated mice, see Fig. 3 (*P < 0Æ05, La1 vs. control).

Table 1 Serum total IgE concentration of NC/Nga mice after 15 miteantigen applications (ng mL)1)

Control La1

Male Female Male Female

1511Æ3 940Æ0 71Æ1 84Æ8427Æ5 677Æ5 15Æ0 56Æ8134Æ8 673Æ8 ND ND115Æ0 452Æ5 ND ND

48Æ8 134Æ8 ND NDND 127Æ5 ND ND

ND 98Æ8 NDND 27Æ5 ND

ND NDND

ND

Control mice were treated with PBS in phase 2 (days 20–22after birth), and La1 mice were treated with La1 during the

same period. Total IgE was estimated with a YAMASA IgE EIAKit (YAMASA), and 10-times diluted serum was assayed. ND,

not detectable.




(a)

(b)

(c)

(d)

Non-treated NC mice

Female Lal

Female control

Male Lal

Male control

Fig 6. Histological status of the treated skin in NC/Nga mice after 15 antigen applications. Images of the skin from the same region of nontreated

age-matched NC/Nga mice are shown as a reference. The bar on the right side of each image represents the thickness of upper dermis and arrows

in the images indicate the separation of the epidermis layer. (a) Control male mice, (b) La1 male mice, (c) Control female mice, (d) La1 female,

see Fig. 4.




Discussion

Unlike systemic immunity, the mucosal immune system and

gut immune system, especially, play critical roles in the sup-

pression of the immune responses to nonhostile antigens.16

Thus, it has been suggested that the stimulation and/or modu-

lation of the gut immune system by the administration of pro-

biotics are effective for the alleviation of food allergies as well

as atopic diseases.17 This study focused on the administration

of probiotics during the developmental process, and the effects

were assessed using a mouse model because the immune

system during the developmental process is supposed to be

more sensitive to probiotics than a mature system.

In experiment 1 of this study, the probiotic L. johnsonii

NCC533 (La1) was administered in the developmental process

of the mouse model, and the most effective period for the

administration was estimated. The La1 administration in phase

2 had a positive effect on the production of sIgA. It has been

suggested that the oral administration of probiotic bacteria

stimulates IgA production not only in the intestine but also in

serum. However, most studies demonstrated a significant

increase of the pathogen-specific sIgA in either the intestine or

serum.18–20 Moreover, some studies have reported that the

oral administration of probiotic did not affect the amount of

total faecal IgA.21,22 A few studies have demonstrated that

there was a significant increase of the total faecal IgA as a

result of the administration of probiotic bacteria in humans

and in mice,23,24 but those increases of the faecal IgA were

also achieved in early life as in this study. These facts indicate

that the stimulation of oral La1 administration in phase 2, the

period around weaning, was effective for the production of

sIgA in the intestine. In this period, the intestine of pups must

be highly vulnerable to foreign antigens as the supplemen-

tation of the maternal defensive molecules is stopped by

weaning, as indicated by the drastic decrease of faecal IgA.9

Furthermore, quite a few of the self-sIgA are produced in the

intestine of pups.9 Thus, the administered La1 should have

easily reached the immune inductive site and then enhanced

the development of the gut immune systems, including self

sIgA production. The immune cells for IgA production seem

to be already deployed in Peyer’s patches at this period,25

although they are not producing a large quantity of sIgA

because there is less antigen stimulation by food or bacteria

until weaning.5,26

In contrast with this, the administration in phase 1 surpris-

ingly showed negative effects on the production of sIgA after

administration. This effect was probably caused specifically in

this period. At this time, the pup intestines were still protected

by the maternal defensive molecules.9,27,28 Thus, the adminis-

tered La1 must have been prevented from reaching the

immune inductive site either partly or mostly by maternal

defensive molecules, such as sIgA and lactoferrin. Further-

more, an insufficient number of immune cells for proper

immune responses, such as CD4+, are deployed in the intes-

tine during this period.29 Therefore, immune responses might

have been improperly induced by some La1 that reached the

immune inductive site while escaping from the maternal

defensive molecules. The La1 administration in phase 3 did

not show much positive effect on IgA production. Increasing

sIgA in this period9 may have affected the viability of the

administrated La1 because self-secreted sIgA might play a role

in the prevention of bacterial colonization.9

Early life is quite important for the immune development

of T-helper cells.8 It is, therefore, quite likely that these differ-

ent immunomodulatory effects of La1 administration in each

phase have respective effects on the development of immune

diseases, such as allergies and autoimmune diseases, after

maturation. However, nearly all of the mice unexpectedly

showed tolerance to the induced allergic diarrhoea. The toler-

ance is probably not caused by La1 administration but, rather,

is characteristic of the mice used in this experiment. These

mice were completely conventional whereas the mice in the

reference study were SPF.14 The mice in this experiment cer-

tainly had more antigenic stimulation to their gut immune

system from early life than did the SPF mice. Thus, their

immune system skewed more towards the Th1 type than that

of the SPF mice, regardless of the La1 administration. The Th1

type skewed the immune system of the mice in this experi-

ment and probably inhibited the development of allergic diar-

rhoea. This result implicitly supported the hygiene hypothesis

proposed by Strachan.7

In experiment 2, we proposed that the immunomodulatory

effect of the La1 administration in early life is applicable for

the prevention and/or inhibition of AD because the mucosal

immune system is deeply connected with systemic immunity.5

In fact, Kalliomaki et al.,30,31 in a randomized placebo-

controlled trial, suggested that probiotic therapy might be effi-

cient for the treatment of atopic dermatitis. An NC/Nga

mouse, which is a model for atopic diseases, was used in

experiment 2. This mouse develops a human AD-like lesion

triggered by a mite antigen under conventional circum-

stances.13 The La1 administration was done in phase 2 in this

experiment because administration in this phase had the most

significant effect on the sIgA production in experiment 1.

In fact, the administration of La1 in phase 2 also showed a

positive effect on the production of IgA in NC/Nga mice in

both male and female mice (Fig. 3). The induction of an

AD-like lesion by the mite antigen application was successful

(Figs 4, 6), and remarkable inhibitory effects on the lesion

were observed as a result of the primary administration of La1

in phase 2. The antigen-treated skin of nearly all of the mice

in the La1 group showed a practically normal appearance and

histological status. Involvement of serum IgE level and mast

cell infiltration into antigen-applied skin also suggested that

the mice in the control group had more severe allergic symp-

toms than that of the mice in La1 group.

The sex difference of the inhibitory effect by the La1

administration was observed on the IgG1/G2a ratio and

density of mast cells. The sex difference of lesion severity in

this mouse model has been reported in several studies13,15,32

although the precise reason for it is still unknown. Female

mice tend to show more severe symptoms when lesions are




induced as they were in this experiment.15 The difference

in lesion severity may be attributed to the fact that the

inhibitory effect in male mice was smaller than that in

females. As in the study by Matsuoka et al.,15 our results

showed that the symptoms of AD-like lesions in female

mice, the scratching scores, and the density of mast cells,

for instance, were more severe than those in male mice.

The higher inhibitory effects in female mice than in male

mice, therefore, were presumably achieved by the primary

administration of La1.

This study strongly suggested a strong link between the gut

immune system and AD and proposed a preventive method

for AD using a lactic acid bacterium that possesses the ability

to adhere to intestinal mucosa. The inhibitory effect of the

primary administration of La1 on AD was derived from the

efficient stimulation and/or modulation of the gut immune

systems, as represented by the increase of sIgA production.

Although the precise relation between gut and systemic

immune system has not yet been revealed, an immunomodu-

lation of the gut immune system by oral administered antigen

can certainly be manifested in the systemic immune system.33

For instance, it has been demonstrated by several studies that

tolerance to orally administered antigen is established not only

in the mucosal but also in the systemic immune system.34,35

Thus, the immune system skewed towards Th1 type cells or

anergy to nonharmful antigens established firstly in gut by

La1 administration may have also been achieved in the sys-

temic immune system through this immunological relation.

Dendritic cells or/and T-regulatory cells may play a role in

this immunological relation because they seem to possess the

ability to establish a tolerance both in mucosal and systemic

immunity.36–38

Acknowledgments

The authors thank Dr Keizo Nakayama and Dr Takamitsu

Tsukahara for sharing their knowledge of histopathology. A

part of this study was supported by the Applied Science Board

for Lactic Acid Bacteria (Tokyo, Japan).

References

1 Xu J, Bjursell MK, Himrod J et al. A genomic view of the human-Bacteroides thetaiotaomicron symbiosis. Science 2003; 299:2074–6.

2 Umesaki Y, Setoyama H, Matsumoto S et al. Differential roles ofsegmented filamentous bacteria and clostridia in development of

the intestinal immune system. Infect Immun 1999; 67:3504–11.3 Umesaki Y, Setoyama H. Structure of the intestinal flora respon-

sible for development of the gut immune system in a rodentmodel. Microbes Infect 2000; 11:1343–51.

4 Bjorksten B. The gastrointestinal flora and the skin. Is there a link?Pediatr Allergy Immunol 2001; 12 (Suppl. 14):51–5.

5 Brandtzaeg PE. Current understanding of gastrointestinal immuno-regulation and its relation to food allergy. Ann N Y Acad Sci 2002;

964:13–45.6 Rook GA, Stanford JL. Give us this day our daily germs. Immunol

Today 1998; 19:113–16.

7 Strachan DP. Hay fever, hygiene, and household size. Br Med J1989; 299:1259–60.

8 Sudo N, Sawamura S, Tanaka K et al. The requirement of intestinalbacterial flora for the development of an IgE production system

fully susceptible to oral tolerance induction. J Immunol 1997;159:1739–45.

9 Inoue R, Ushida K. Development of the intestinal microbiota inmice and its possible interactions with the evolution of the luminal

IgA in the intestine. Exp Anim 2005; 54:437–45.10 Granato D, Bergonzelli GE, Pridmore RD et al. Cell surface-associ-

ated elongation factor Tu mediates the attachment of Lactobacillus

johnsonii NCC533 (La1) to human intestinal cells and mucins. InfectImmun 2004; 72:2160–9.

11 Schiffrin EJ, Rochat F, Link-Amster H et al. Immunomodulation ofhuman blood cells following the ingestion of lactic acid bacteria.

J Dairy Sci 1999; 78:491–7.12 Haller D, Bode C, Hammes WP et al. Non-pathogenic bacteria elicit

a differential cytokine response by intestinal epithelial cell/leuco-cyte co-cultures. Gut 2000; 47:79–87.

13 Matsuda H, Watanabe N, Geba GP et al. Development of atopicdermatitis-like skin lesion with IgE hyperproduction in NC/Nga

mice. Int Immunol 1997; 9:461–6.14 Kweon MN, Yamamoto M, Kajiki M et al. Systemically derived large

intestinal CD4(+) Th2 cells play a central role in STAT6-mediatedallergic diarrhea. J Clin Invest 2000; 106:199–206.

15 Matsuoka H, Maki N, Yoshida S et al. A mouse model of the atopiceczema/dermatitis syndrome by repeated application of a crude

extract of house-dust mite Dermatophagoides farinae. Allergy 2003;58:139–45.

16 Mayer L. Mucosal immunity. Pediatrics 2003; 111:1595–600.17 Isolauri E. Probiotics in human disease. Am J Clin Nutr 2001;

73:1142S–6S.18 Kaila M, Isolauri E, Soppi E et al. Enhancement of the circulating

antibody secreting cell response in human diarrhea by a humanLactobacillus strain. Pediatr Res 1992; 32:141–4.

19 Link-Amster H, Rochat F, Saudan KY et al. Modulation of a specifichumoral immune response and changes in intestinal flora mediated

through fermented milk intake. FEMS Immunol Med Microbiol 1994;10:55–63.

20 Yasui H, Kiyoshima J, Ushijima H. Passive protection against rota-virus-induced diarrhea of mouse pups born to and nursed by dams

fed Bifidobacterium breve YIT4064. J Infect Dis 1995; 172:403–9.21 Pessi T, Sutas Y, Hurme M, Isolauri E. Interleukin-10 generation in

atopic children following oral Lactobacillus rhamnosus GG. Clin Exp

Allergy 2000; 30:1804–8.22 Roller M, Rechkemmer G, Watzl B. Prebiotic inulin enriched with

oligofructose in combination with the probiotics Lactobacillus rhamno-sus and Bifidobacterium lactis modulates intestinal immune functions in

rats. J Nutr 2004; 134:153–6.23 Fukushima Y, Kawata Y, Hara H et al. Effect of a probiotic formula

on intestinal immunoglobulin A production in healthy children.Int J Food Microbiol 1998; 42:39–44.

24 Fukushima Y, Kawata Y, Mizumachi K et al. Effect of bifidobacteriafeeding on fecal flora and production of immunoglobulins in lacta-

ting mouse. Int J Food Microbiol 1999; 46:193–7.25 Kramer DR, Cebra JJ. Early appearance of ‘natural’ mucosal IgA

responses and germinal centers in suckling mice developing in theabsence of maternal antibodies. J Immunol 1995; 154:2051–62.

26 Fagarasan S, Honjo T. Intestinal IgA synthesis: regulation of front-line body defenses. Nat Rev Immunol 2003; 3:63–72.

27 Kelleher SL, Lonnerdal B. Immunological activities associated withmilk. Adv Nutr Res 2001; 10:39–65.




28 Hanson LA. Human milk and host defense: immediate and long-term effects. Acta Paediatr Suppl 1999; 88:42–6.

29 Hoshi H, Horie K, Tanaka K et al. Patterns of age-dependentchanges in the numbers of lymph follicles and germinal centers in

somatic and mesenteric lymph nodes in growing C57Bl/6 mice.J Anat 2001; 198:189–205.

30 Kalliomaki M, Salminen S, Arvilommi H et al. Probiotics in primaryprevention of atopic disease: a randomized placebo-controlled trial.

Lancet 2001; 357:1076–9.31 Kalliomaki M, Salminen S, Poussa T et al. Probiotic and prevention

of atopic disease: 4-year follow-up of a randomized placebo-

controlled trial. Lancet 2003; 361:1869–71.32 Sugiura K, Hiramoto K, Shamoto M et al. Immunological cell situ-

ation in the skin of atopic model mice. J Eur Acad Dermatol Venereol2004; 18:148–52.

33 Ogra PL, Faden H, Welliver RC. Vaccination strategies for mucosalimmune responses. Clin Microbiol Rev 2001; 14:430–45.

34 Chase MW. Inhibition of experimental drug allergy by priorfeeding of the sensitizing agent. Proc Soc Exp Biol 1946; 61:257–

9.35 Mattingly JA, Waksman BH. Immunologic suppression after oral

administration of antigen. J Immunol 1978; 121:1878–83.36 Viney JL, Mowat AM, O’Malley JM et al. Expanding dendritic cells

in vivo enhances the induction of oral tolerance. J Immunol 1998;160:5815–25.

37 Williamson E, Westrich GM, Viney JL. Modulating dendritic cells

to optimize mucosal immunization protocols. J Immunol 1999;163:3668–75.

38 Sakaguchi S Regulatory T cells: key controllers of immunologicself-tolerance. Cell 2000; 101:455–8.




CONTACT DERMATITIS AND ALLERGY DOI 10.1111/j .1365-2133.2006.07674.x

Compositae dermatitis from airborne parthenolideE. Paulsen, L.P. Christensen* and K.E. Andersen

Department of Dermatology, Odense University Hospital, Sdr. Boulevard 29, DK-5000 Odense C, Denmark

*Department of Food Science, Danish Institute of Agricultural Sciences, Research Centre Aarslev, Kirstinebjergvej 10, DK-5792 Aarslev, Denmark

CorrespondenceEvy Paulsen



Key wordsairborne, allergic contact dermatitis, Compositae,

parthenolide, plants, Tanacetum parthenium


The studies were carried out at the Department of

Food Science, Danish Institute of Agricultural

Sciences, Research Centre Aarslev and the

Department of Dermatology, Odense University

Hospital.

Summary

Background Compositae dermatitis confined to exposed skin has often been consid-ered on clinical grounds to be airborne. Although anecdotal clinical and plantchemical reports suggest true airborne allergy, no proof has been procured.Feverfew (Tanacetum parthenium) is a European Compositae plant suspected of caus-ing airborne contact allergy, and its most important allergen is the sesquiterpenelactone (SQL) parthenolide (PHL).Objectives The aims of this study were to (i) assess the allergenicity of feverfew-derived monoterpenes and sesquiterpenes and their oxidized products in fever-few-allergic patients and (ii) re-assess the role of PHL and other SQLs in airbornecontact allergy.Patients and methods Feverfew-allergic patients were patch tested with extracts andfractions containing volatile monoterpenes and sesquiterpenes as well as extractsof airborne particles from flowering feverfew plants, obtained by fractionation ofether extracts, dynamic headspace and high-volume air sampler (HIVAS) tech-nique, respectively.Results Among 12 feverfew-allergic patients, eight had positive patch-test reactionsto a HIVAS filter extract, while two tested positive to a headspace extract. Subse-quent analysis of the HIVAS extract by gas chromatography and mass spectro-metry detected PHL in a concentration of 510 ng mL)1 in the HIVAS extract.Testing with a dilution series of PHL showed positive reactions down to 8Æ1 ngin selected patients. None of the 12 patients tested positive to monoterpenes orsesquiterpenes, whether they were oxidized or not.Conclusions The clinical results have proved that some feverfew-allergic patients aresensitive to airborne particles released from the plant, and isolation of PHL fromthe particle-containing HIVAS extract in allergenic amounts is strong evidence ofPHL as the responsible allergen.

The classical form of Compositae dermatitis is an airborne pat-

tern eczema, first described in the U.S.A. as weed dermatitis

in the 1930s, and 30–40 years later in India as Parthenium

dermatitis and in Australia as Australian bush dermatitis.1–3

The hypothesis of airborne allergens has been much debated

since then, and although the clinical features, the allergenicity

of dried plant parts and positive patch-test reactions to a filter

from a high-volume air sampler (HIVAS) as well as to air-dried

clothing are suggestive of true airborne contact allergy, no def-

inite proof has been established.1–7 Furthermore, a European

equivalent to airborne Compositae dermatitis has been ques-

tioned because of the humid, colder, climatic conditions.8

Feverfew [Tanacetum parthenium (L.) Schultz-Bip.] is a frequent

sensitizer in European Compositae-allergic patients.9,10 On the

basis of case reports, this plant, which is native to south east

Europe and partly naturalized from gardens in Northern Eur-

ope, has been considered to be a cause of airborne contact

dermatitis (ABCD).11,12 Its main allergenic sesquiterpene lac-

tone (SQL) is parthenolide (PHL, Fig. 1),11 which may consti-

tute up to 0Æ9% of dry weight in aerial parts of the plant.13,14

However, experimental studies using HIVAS and dynamic

headspace (DHE) techniques around feverfew plants failed to

detect both particle-bound and volatile PHL.14 The volatiles

emitted from feverfew plants were mainly monoterpenes, but

as only a minority of feverfew-allergic patients tested positive

to these, their contribution to airborne Compositae dermatitis

seemed to be of small significance.15

On the other hand, it has been established, notably in fra-

grance research, that oxidized monoterpenes are often stronger

allergens than the nonoxidized substances.16 Therefore, the aims



of this study on airborne Compositae dermatitis were to

(i) assess the allergenicity of feverfew-derived monoterpenes

and sesquiterpenes and their oxidized products in feverfew-

allergic patients and (ii) re-assess the possible role of PHL and

other SQLs in view of more accurate and sensitive gas chromato-

graphy–mass spectrometry (GC-MS) methods available with

lower detection and quantification thresholds compared with

GC-MS techniques used previously for detection of airborne

SQLs.14

Patients, materials and methods

Patients

Adults, who by previous patch testing in our clinic had had at

least 2+ reactions to feverfew extract 1% petrolatum (pet.)

and PHL 0Æ1% pet., were invited to participate in the study.

Exclusion criteria were pregnancy, treatment with systemic

steroid or other immunosuppressive or -modulating drugs as

well as recent ultraviolet (UV) radiation therapy.

Patch-test materials were applied to the patient’s back for

48 h using Finn Chambers� (Epitest Ltd, Helsinki, Finland)

on Scanpor� (Scanpor Alpharma AS, Oslo, Norway), and read-

ings were taken on day 3 according to the criteria of the

International Contact Dermatitis Research Group.17

All participants gave their written, informed consent, and in

general the study followed the Declaration of Helsinki proto-

cols and was approved by the local ethics committee (case no.

20040123).

Plant material and parthenolide content

Twelve specimens of T. parthenium (L.) Schulz Bip. were grown

in a greenhouse at 17–23 �C in 20-L pots and watered daily.

Plants were grown under natural light conditions. The concen-

tration of PHL (Fig. 1) in the aerial parts of the feverfew

plants was �0Æ2% of dry weight as determined by analytical

reversed-phase high performance liquid chromatography using

acetonitrile–water as mobile phase.13,14

Analysis for parthenolide in high-volume air sampler

filters

A HIVAS (Air sampler #15000, The Science Source, Waldo-

boro, ME, U.S.A.) was placed in a closed chamber of

1Æ5 · 1Æ5 · 1Æ5 m in a greenhouse together with eight flower-

ing feverfew plants at a distance of 15–20 cm from the plants.

The sampler was mounted with filter disks capturing plant

particles down to 4 lm in size, and equipped with a cap pro-

tecting the filter from direct contact with the plants so only

airborne particles released from the plants were collected. The

air sampler collected airborne particles 8 h a day over a 16-h

time period (from 06Æ00 to 22Æ00) with a 30-min pause each

hour in order to avoid overheating. The sampling of airborne

particles was performed at 18–20 �C over a 6-day period. The

average flow rate through the air sampler was measured to

215 L min)1. The HIVAS filter was changed once a day and

kept in a 250-mL blue-cap flask at )24 �C until analysis. In

total six HIVAS filters were obtained, which were randomly

divided into two groups with four (group I) and two (group

II) HIVAS filters, respectively.

The HIVAS filters in groups I and II were extracted with

50 mL 96% ethanol, filtered, and the eluates combined and

concentrated in vacuo (30 �C) to 50 mL each for testing as is on

Compositae-sensitive patients and/or analysis. Only the HIVAS

extract obtained from the group I filters was tested on patients

and after testing, both the group I and II filter extracts were

analysed by GC-MS for PHL and other SQLs. GC-MS analysis

was performed on a Trace DSQ single quadrupole mass spec-

trometer operated at 70 eV and directly coupled to a Trace GC

Ultra gas chromatograph (both, Thermo Finnigan, Austin, TX,

U.S.A.) equipped with a split/splitless PTV injector (split flow

10 mL min)1, 200 �C) and a Zebron capillary column

(30 m · 0Æ25 mm internal diameter, film thickness df ¼0Æ25 lm, ZB-1, Phenomenex, Torrance, CA, U.S.A.). Helium

was the carrier gas at a constant column head pressure of

10 psi. The temperature of the column and transfer line was

200 �C. GC-MS was performed on 2 lL HIVAS extract. Com-

pounds in the HIVAS extract were separated by the following

oven temperature gradient: 32 �C for 1 min, programmed to

260 �C at 10 �C min)1 followed by constant temperature for

20 min. The mass spectrometer was operated in full scan

mode over a mass range from 39 to 650 amu (1 scan g)1)

and analysed for PHL and other SQLs. PHL was the only SQL

detected in the HIVAS extracts, and in order to improve the

sensitivity for detection of PHL by GC-MS a selected ion mon-

itoring (SIM) acquisition method was developed. PHL was

detected by GC-MS SIM by setting the mass selective detector

to repeatedly scan for the following ions m/z 145, 190 and

248 (molecular ion). PHL was repeatedly identified in the

HIVAS extracts in full and SIM scan modes by comparison

with an authentic standard of PHL (Sigma-Aldrich, Steinheim,

Germany).

Quantification of PHL in the HIVAS extracts was per-

formed in SIM mode using PHL as external standard. The

calibration curve was determined to be linear in the concen-

tration range 50–1000 ng mL)1 (r2 ¼ 0Æ988). The detection

limit of PHL was determined from the calibration curve to

be �20 ng mL)1 [Signal to noise (S/N) ¼ 10] in SIM

scan mode and the quantification limit to be 50 ng mL)1

(S/N ¼ 3).

O

O

Parthenolide (PHL)

O

Fig 1. Chemical structure of the sesquiterpene lactone parthenolide

(PHL) present in the aerial parts of Tanacetum parthenium (feverfew).



Dermatitis from airborne parthenolide, E Paulsen et al. 511

Isolation of volatiles from feverfew

Volatiles were isolated from aerial parts of feverfew plants by

DHE and column chromatography (CC), respectively, for test-

ing. Volatiles emitted from feverfew were collected from the

closed chamber mentioned above by DHE technique as

described previously with a few alterations.14 Volatiles were

collected directly from the chamber in a 100-mL pipette

equipped with 50 g Porapak Q 50–80 mesh (Waters, Milford,

MA, U.S.A.), and eluted with diethyl ether, concentrated to a

final volume of 300 lL14 and dissolved in 5 mL 96% ethanol.

Four DHE collections were made and the extracts combined to

one headspace extract (20 mL), which was used for testing

and for oxidation experiments. The combined headspace

extract was analysed by GC and GC-MS as described previous-

ly.14 The total concentration of the volatiles in the ethanol

headspace extract was > 1% and used for testing as is.

Volatiles were extracted with 2 L of diethyl ether from

800 g aerial parts of feverfew plants for 60 min. The diethyl

ether extract was concentrated in vacuo (30 �C) to 10 mL. Vola-

tiles were isolated from the extract by silica gel CC using an

n-hexane-diethyl ether gradient. Fractions containing volatiles

only and no PHL were combined, the solvent removed in vacuo

(30 �C), and the residue dissolved in 40 mL 96% ethanol for

testing and oxidation.

Oxidation by ultraviolet radiation of volatiles isolated by

dynamic headspace and column chromatography

The combined DHE extract and the extract obtained by CC

were purged with O2 and oxidized for 30 min using a UV-

immersion lamp reactor (Sigma-Aldrich) equipped with a

400-mL reaction vessel made of borosilicate glass. The photon

source was provided with a TQ 150 high-pressure Hg lamp

with linkage and condenser made of silica (Sigma-Aldrich).

Analysis of the oxidized extracts by GC-MS revealed quantita-

tive, but no qualitative, differences in the volatile profile com-

pared with the nonoxidized extracts.

Results

Clinical studies

Twelve people (11 female, one male, mean age 54 years) with

contact sensitization from feverfew and PHL were recalled for

patch testing with extracts and fractions containing volatile

monoterpenes and sesquiterpenes as well as HIVAS filter

extracts of airborne particles from feverfew plants. The results

of the initial patch testing, including repeated testing with the

crude feverfew extract and PHL, are summarized in Table 1.

Contrary to former results, the subjects were mainly positive

to airborne fractions: eight of 12 tested positive to an extract

from HIVAS filters, and two of 12 had positive reactions to an

extract collected by the DHE technique. On the other hand, no

one tested positive to the monoterpene and sesquiterpene

fraction. Ten consecutive eczema patients tested negative to

the HIVAS filter extract as well as to the oxidized and nonoxi-

dized headspace extracts. After PHL had been detected in the

extract from the HIVAS filters in minute amounts, it was

decided to retest some of the eight people with positive reac-

tions to the crude extract in order to determine the threshold

elicitation concentration of PHL. Two dilution series of PHL,

incorporated in a dry film using the TRUE Test� technology,

were kindly provided by Mekos Laboratories AS (Hillerød,

Table 1 Results of patch testing with oxidized and nonoxidized extracts from feverfew, collected by HIVAS, DHE or direct extraction, in 12

feverfew-allergic patients

Patient no. 1 2 3 4 5 6 7 8 9 10 11 12

Sex/age (years) F/48 F/55 F/42 F/59 F/53 F/48 M/79 F/47 F/61 F/55 F/52 F/59

AllergenFeverfew 1% pet. ++ ++ ++ +++ ++ +++ ++ ++ ++ +++ ++ ++

PHL 0Æ1% pet. +++ ++ ++ +++ +++ +++ ++ +++ ++ +++ +++ +++l-Carvone 1% pet. – – – – – – 0 + – 0 foll. – 0 foll.

l-Carvone 5% pet. 0 foll. 0 0 – – – 0 0 – ++ 0 0Nonoxidized monoterpenes and sesquiterpenes – – 0 – – – – – 0 – – –

Oxidized monoterpenes and sesquiterpenes – – 0 – – – 0 – 0 0 foll. – –HIVAS filter extract ++ + ++ – – + ++ – ++ ++ 0 ++

Nonoxidized DHE extract – – 0 foll. – – – ++ – – + – –Oxidized DHE extract – – 0 foll. – – – + 0 – – – 0 foll.

HIVAS filter under Scanpor� – – – – – – – – – – – –HIVAS filter in FC under Scanpor� 0 – – – – – – – – – – –

Colophonium 1500 lg cm)2 + – – – – – ++ + – – – –FM I 450 lg cm)2 ++ – – ++ ++ ++ + – – ++ ++ –

Feverfew p.s. No No No Yes Yes No Yes ? No Yes No YesABCD suspected by patient No Yes No No Yes No Yes No Yes Yes Yes Yes

F, female; M, male; PHL, parthenolide; HIVAS, high-volume air sampler; DHE, dynamic headspace; FC, Finn Chamber; FM I, fragrance mix

I; p.s., primary sensitizer; ABCD, airborne contact dermatitis; foll., follicular reaction; 0, doubtful positive reaction.



512 Dermatitis from airborne parthenolide, E Paulsen et al.

Denmark). The results are shown in Table 2: two of five per-

sons tested positive to 30 and 10 ng PHL, corresponding to

24Æ3 and 8Æ1 ng per patch as a TRUE Test� allergen occupies

0Æ81 cm2.

Six patients were sensitized by gardening, and three of these

reported feverfew to be one of the primary sensitizers,

whereas four were occupationally sensitized as gardeners/

greenhouse workers or sandwich makers. The last two pati-

ents were both privately and occupationally sensitized by

gardening.

Chemical studies

The particles and volatiles collected by HIVAS and DHE tech-

niques, respectively, were derived from several feverfew plants

in a greenhouse instead of single specimens to imitate the nat-

ural outdoor conditions and ensure a sufficient amount of any

volatile allergen for patch testing.

PHL was detected using GC-MS SIM by repeatedly scanning

for the following ions m/z 145, 190 and 248 (molecular ion),

which are some of the characteristic and/or major ions in the

mass spectrum of PHL. The retention time for PHL was

23Æ9 min and was repeatedly identified in the combined

HIVAS extracts in full and SIM scan mode by comparison with

an authentic standard of PHL as illustrated in Figure 2.

PHL in the combined HIVAS extracts used for testing

(group I) was determined to be 510 ng mL)1 HIVAS extract

corresponding to an average release of �800 ng PHL h)1 or

100 ng PHL h)1 per feverfew plant. The presence of PHL was

also detected in the group II HIVAS extracts that were not

used for testing. The concentration of PHL in this combined

HIVAS extract corresponded to an average release of �310 ng

PHL h)1 or 40 ng PHL h)1 per feverfew plant, thus confirm-

ing that PHL was released over the whole collection period,

but, as expected, not released uniformly.

Table 2 Results of patch testing with two dilution series ofparthenolide

Patient no. 2 3 7 9 12

Sex/age (years) F/55 F/42 M/79 F/61 F/59PHL concentration

0Æ010 mg cm)2 NT NT NT NT +++0Æ0033 mg cm)2 ++ ++ NT ++ ++

0Æ0011 mg cm)2 ++ ++ NT ++ +++0Æ00037 mg cm)2 0 foll. + foll. NT ++ ++

0Æ00012 mg cm)2 0 foll. – NT ++ ++

0Æ060 lg cm)2 – – + foll. ++ ++0Æ030 lg cm)2 – 0 foll. 0 + ++

0Æ010 lg cm)2 – – – 0 +0Æ0030 lg cm)2 – – – – 0 foll.

0Æ0010 lg cm)2 – – – – –

PHL, parthenolide; F, female; M, male; 0, doubtful positivereaction; foll., follicular reaction; NT, not tested.

0 5 10 15 20 25 30 35 40

0 5 10 15 20 25 30 35 40

Retention time (min)

Rel

ativ

e ab

unda

nce

Rel

ativ

e ab

unda

nce

PHL

PHL

Parthenolide (PHL) standard (500 ng ml–1)

High-volume air sampler (HIVAS) extract

0

10

20

30

40

50

60

70

80

90

100

0

10

20

30

40

50

60

70

80

90

100

Fig 2. Gas chromatography–mass

spectrometry (GC-MS) selected ion

monitoring (SIM) chromatogram of a

standard sample of parthenolide (PHL) in a

concentration of 500 ng mL)1 and the

combined high-volume air sampler (HIVAS)

extract, respectively. PHL was detected at

23Æ9 min in SIM and full scan mode (m/z

39–650) in both samples. The selected ions in

the GC-MS SIM acquisition method were m/z

145, 190 and 248 [molecular ion (M)+].




Analysis of the combined headspace extract by GC-MS

revealed the monoterpenes camphene, p-cymene, camphor,

trans-chrysanthenol and trans-chrysanthenyl acetate as the major

volatiles, and no PHL was detected in the extracts in accord-

ance with previous investigations.14

The major terpenes in the monoterpene and sesquiterpene

extract were camphor and trans-chrysanthenyl acetate as

shown by GC-MS,14 although the concentrations of oxygen-

ated monoterpenes such as bornyl acetate were higher in this

extract compared with the headspace extract. The total con-

centration of the volatiles in this ethanol extract was > 2%

and tested as is.

Discussion

Although Hjorth et al.,4 in their frequently quoted paper on

Compositae dermatitis in Danish patients, emphasized the air-

borne pattern, later studies have shown a greater diversity of

clinical features in European Compositae-allergic patients,

which is probably why the existence of airborne Compositae

dermatitis in Europe has been doubted.8,18,19

Concerning feverfew, anecdotal reports have been suggest-

ive of the plant as a cause of ABCD based on clinical features

and information on exposure.11,20

However, this is the first study where several feverfew-aller-

gic patients consistently showed positive patch-test reactions

to airborne fractions that were subsequently proved to contain

PHL. This allergen is not volatile, so its presence must be due

to plant particles containing traces of PHL.14 PHL is probably

synthesized in glandular trichomes on the leaf surface as it has

been reported in other allergenic Compositae species.21 The

concentration of PHL in the HIVAS extract of 510 ng mL)1

would allow an amount of �8 ng per Finn Chamber�, pro-

viding that the chamber contained 15 lL. However, the parti-

cle-bound PHL was probably unevenly distributed, and

furthermore, the extract was not applied by a micropipette,

but by capillary tubes, which would lead to a higher concen-

tration of PHL in the Finn Chamber�.

The fact that some of the subjects were very sensitive to

the allergen, reacting positively down to 8Æ1 ng, is strong

evidence of PHL as the allergen causing the reactions to the

HIVAS extract. The average amount of PHL released from a

single test plant in this study was 100 ng h)1, meaning that

a border of feverfew plants may release milligrams of the

allergen, which would be sufficient to elicit reactions in

moderately sensitive persons: among the five subjects tested

with the PHL dilution series, the least sensitive tested posi-

tive to 0Æ0011 mg cm)2 corresponding to < 1 lg per test

area. Furthermore, the feverfew plants used in this study

contained 0Æ2% PHL of dry weight in the aerial parts, but

concentrations of up to 0Æ9% PHL of dry weight in leaves

has been reported previously as well as concentrations

exceeding 1Æ2% PHL of dry weight in leaves of post-flower-

ing feverfew plants.13 So the release of PHL under real-life

conditions may be even higher as those reported in the pre-

sent study.

The fact that PHL elicits patch-test reaction on normal skin

on the back of some PHL-sensitive patients in a dose of 10 ng

in a TRUE Test� patch makes it one of the most powerful

contact allergens in the environment.22 Dermatitic skin on

exposed areas may be even more sensitive. PHL has been isol-

ated from a number of other Compositae plants such as, e.g.

tansy (T. vulgare), Anthemis cretica, Inula aschersoniana, Ambrosia and

Arctotis spp. as well as members of the Magnoliaceae family,23

and furthermore there is evidence of true cross-reactivity

between PHL and the allergens of the popular Chrysanthemum

indicum L. [Dendranthema indicum (L.) Desm.].24 The implications

are that it may be quite difficult to avoid this strong allergen

and closely related compounds.

The degree of sensitivity to PHL varied in the five subjects

tested with the dilution series (Table 2): the most sensitive

patient (no. 12), who was diagnosed with Compositae allergy

in 2004, reported feverfew as a primary sensitizer and still

had problems with summer-exacerbated eczema. On the other

hand, the second most sensitive person (patient no. 9) was

diagnosed with Compositae allergy in 2003, probably sensiti-

zed by gardening, but not feverfew, and her hand eczema was

well controlled now.

The relation between the patients’ suspicion of ABCD and a

positive reaction to the HIVAS extract was not unambiguous,

but in five of the eight patients the two parameters were coin-

cident. Feverfew as a primary sensitizer was not associated

with positive reactions to the HIVAS extract.

As expected, PHL was not detected in the headspace extracts

and neither were other SQLs. The positive reactions to nonoxi-

dized/oxidized extracts in two patients thus could not be

explained by their SQL allergy, but patient no. 7 had a colo-

phonium and patient no. 10 a fragrance allergy, and it is

possible that their headspace extract reactions are due to cross-

reactivity between volatile terpenes from feverfew and

terpenes in colophonium and fragrance mix I. Cross-reactivity

between fragrance terpenes, colophonium and Compositae

plant extracts has been reported before, although it is not

invariably seen in all patients as documented in this study.25

Likewise, there was no association between positive reactions

to fragrance mix/colophonium and the extract from the

HIVAS filters. Also, all people in this study tested negative to

the monoterpene and sesquiterpene fraction isolated directly

from feverfew plants, whether the fraction was oxidized or

not, emphasizing that these are not primary allergens.

Carvone was included in the patch-test series due to a pre-

viously reported association between sensitization to this mon-

oterpene and Compositae extracts.26 Two of the 12 subjects

tested positive (Table 1), and the allergy was of old relevance

in one and unknown relevance in the other, making true

cross-reactivity a possibility in the latter.

In conclusion, the clinical part of this study has proved

that feverfew-allergic patients are sensitive to airborne

and—to a lesser degree—volatile fractions from feverfew,

and the subsequent isolation of the SQL PHL from the parti-

cle-containing airborne fraction in amounts sufficient to elicit

positive patch-test reactions in the same patients is strong



514 Dermatitis from airborne parthenolide, E Paulsen et al.

evidence of PHL as the allergen responsible for ABCD in

feverfew-allergic patients. In accordance with previous stud-

ies, monoterpenes and sesquiterpenes seem to be of minor

importance.14,15

Acknowledgments

Kgl. Hofbundtmager Aage Bangs Fond is gratefully acknow-

ledged for financial support and Mekos Laboratories AS, Hil-

lerød, Denmark for providing the parthenolide dilution series.

References

1 Shelmire B. Contact dermatitis from vegetation. Patch testing and

treatment with plant oleoresins. J S Med Assoc 1940; 33:337–46.2 Lonkar A, Jog MK. ‘Epidemic’ contact dermatitis due to Parthenium

hysterophorus (Compositae family of plants). Report of 350 cases.Contact Derm Newsl 1972; 11:291–2.

3 Burry JN, Kuchel R, Reid JG, Kirk J. Australian bush dermatitis:Compositae dermatitis in South Australia. Med J Aust 1973;

1:110–16.4 Hjorth N, Roed-Petersen J, Thomsen K. Airborne contact dermatitis

from Compositae oleoresins simulating photodermatitis. Br J Derma-tol 1976; 95:613–20.

5 Hausen BM. Kompositenallergie. Ein Beitrag zur Identifizierung derKontaktallergene, atiologischen Diagnostik und aktuellen Bedeu-

tung. Allergologie 1979; 2:143–7.6 Schumacher MJ, Silvis NG. Airborne contact dermatitis from

Ambrosia deltoidea (triangle-leaf bursage). Contact Derm 2003; 48:212–

16.7 Srinivas CR. Transmission of Parthenium dermatitis by clothing.

Arch Dermatol 2005; 141:1605.8 Schmidt RJ. Compositae. Clin Dermatol 1986; 4:46–61.

9 Geier J, Hausen BM. Epikutantestung mit dem Kompositen-Mix.Allergologie 2000; 23:334–41.

10 Paulsen E, Andersen KE, Hausen BM. Sensitization and cross-reac-tion patterns in Danish Compositae-allergic patients. Contact Derm

2001; 45:197–204.11 Mensing H, Kimmig W, Hausen BM. Airborne contact dermatitis.

Hautarzt 1985; 36:398–402.12 Hausen BM, Vieluf IK. Allergiepflanzen/Pflanzenallergene. Handbuch und Atlas

der allergieinduzierenden Wild-und Kulturpflanzen, 2nd edn. Landsberg/Mun-chen: Ecomed Verlagsgesellschaft mbH, 1997.

13 Awang DVC, Dawson BA, Kindack DG. Parthenolide content offeverfew (Tanacetum parthenium) assessed by HPLC and 1H-NMR

spectroscopy. J Nat Prod 1991; 54:1516–21.14 Christensen LP, Jakobsen HB, Paulsen E et al. Airborne Compositae

dermatitis: monoterpenes and no parthenolide are released fromflowering Tanacetum parthenium (feverfew) plants. Arch Dermatol Res

1999; 291:425–31.15 Paulsen E, Christensen LP, Andersen KE. Do monoterpenes released

from feverfew (Tanacetum parthenium) plants cause airborne Composi-tae dermatitis? Contact Derm 2002; 47:14–18.

16 Matura M, Skold M, Borje A et al. Selected oxidized fragrance ter-

penes are common contact allergens. Contact Derm 2005; 52:320–8.17 Wahlberg JE. Patch testing. In: Textbook of Contact Dermatitis (Rycroft

RJG, Menne T, Frosch PJ, Lepoittevin J-P, eds), 3rd edn. Berlin andHeidelberg: Springer Verlag, 2001:435–68.

18 English J, Norris P, White I, Cronin E. Variability in the clinicalpatterns of Compositae dermatitis. Br J Dermatol 1989; 121 (Suppl.

34): 27.19 Paulsen E, Andersen KE. Compositae dermatitis in a Danish derma-

tology in 1 year (II). Clinical features in patients with Compositaecontact allergy. Contact Derm 1993; 29:195–201.

20 Diepgen TL, Haberle M, Baurle G. Fallstricke in der Berufsderma-tologie: Das aerogene Kontaktekzem auf Pflanzen. Dermatosen 1989;

37:23–5.21 Hausen BM, Spring O. Sunflower allergy. On the constituents of

the trichomes of Helianthus annuus L (Compositae). Contact Derm 1989;20:326–34.

22 Johansen JD, Andersen KE, Svedman C et al. Chloroatranol, anextremely potent allergen hidden in perfumes: a dose–response

elicitation study. Contact Derm 2003; 49:180–4.23 Fischer NH, Olivier EJ, Fischer HD. The biogenesis and chemistry

of sesquiterpene lactones. III. Germacranolides. In: Progress in theChemistry of Organic Natural Products 38 (Herz W, Grisebach H, Kirby

GW, eds). Wien: Springer-Verlag, 1979.24 Schulz KH, Hausen BM, Wallhofer L, Schmidt-Loffler P. Chrysanth-

emen-Allergie. 2. Experimentelle Untersuchungen zur Identifizie-rung der Allergene. Arch Derm Forsch 1975; 251:235–44.

25 Paulsen E, Andersen KE. Colophonium and Compositae mix asmarkers of fragrance allergy: cross-reactivity between fragrance ter-

penes, colophonium and Compositae plant extracts. Contact Derm2005; 53:285–91.

26 Paulsen E, Andersen KE, Carlsen L, Egsgaard H. Carvone: an over-looked contact allergen cross-reacting with sesquiterpene lactones?

Contact Derm 1993; 29:138–43.




DERMATOLOGICAL SURGERY AND LASERS DOI 10.1111/j .1365-2133.2006.07648.x

Comparative study of photodynamic therapy vs. CO2 laservaporization in treatment of condylomata acuminata,a randomized clinical trialK. Chen, B.Z. Chang, M. Ju, X.H. Zhang and H. Gu

Institute of Dermatology, Chinese Academy of Medical Sciences and Peking Union Medical College, 14 Jiangwangmiao Road, Nanjing 210042, China

CorrespondenceHeng Gu.



Key words5-aminolaevulinic acid, condylomata acuminata,

photodynamic therapy, randomized clinical trial


Summary

Background Most conventional therapies for condylomata acuminata (CA) are trau-matic and have high recurrence rates.Objectives To investigate the efficacy and safety of topical application of 5-amino-laevulinic acid (ALA) photodynamic therapy (PDT) for the treatment of CA.Methods Sixty-five patients with CA were allocated into the treatment (ALA-PDT)group and treated with 20% ALA solution under occlusive dressing for 3 hfollowed by irradiation with the helium–neon laser at a dose of 100 J cm)2

and a power of 100 mW. Another 21 CA patients were allocated into the con-trol group and treated with the CO2 laser. The treatment was to be repeated1 week later if the lesion was not completely removed after the firsttreatment.Results After one treatment, the complete removal rate was 95% in the ALA-PDTgroup and 100% in the control group. After two treatments with ALA-PDT, thecomplete removal rate in the treatment group was 100%. The recurrence rate forALA-PDT group was 6Æ3% which was significantly lower than that in control group(19Æ1%, P < 0Æ05). Moreover, the proportion of patients with adverse effects in theALA-PDT group (13Æ9%) was also significantly lower than that in control group(100%, P < 0Æ05). The side-effects in patients treated with ALA-PDT mainlyincluded mild burning and/or stinging restricted to the illuminated area.Conclusions The present study shows that topical application of ALA-PDT is a simp-ler, more effective and safer therapy with a lower recurrence for treatment of CAcompared with conventional CO2 laser therapy.

Condylomata acuminata (CA) are caused by human papilloma-

virus (HPV). Based on the national sexually transmitted disease

(STD) surveillance system, it has been ranked as the third

most prevalent STD in China. More and more attention has

been paid to this medical condition mainly because it has

been believed to be related to genital carcinogenesis. Until

now, there is no medicine that exerts direct anti-HPV effects.

Conventional treatment of CA usually includes excision, phys-

ical therapies, topical medication treatment and immunothera-

pies. However, most conventional treatment strategies are

traumatic, need treatment of long duration and have a high

rate of relapsing episodes.1

Topical application of photodynamic therapy (PDT) with

5-aminolaevulinic acid (ALA) for the treatment of cutaneous

cancer was first described by Kennedy et al.2 Since then, the

application of ALA-PDT in clinical practice has become more

and more popular, and it can be used not only for treatment

of cutaneous cancers and/or precancerous lesions but also for

the treatment of nonmalignant skin disorders, such as psoria-

sis, viral infections and diseases of epidermal appendages.3,4

Recently, studies have showed that ALA-PDT is an attractive

technique in the treatment of CA.5–8 In order to explore

further the efficacy and safety of ALA-PDT on CA, a clinical

trial to evaluate the topical application of ALA-PDT was con-

ducted among 86 patients with CA from December 2004 to

December 2005.

Subjects and methods

Subjects

The patients attending the Institute of Dermatology, Chinese

Academy of Medical Sciences, during the period from Decem-

ber 2004 to June 2005 were invited to participate in the



study. All patients provided written informed consent. The

diagnosis of CA was made based on clinical examination in

addition to the acetowhitening test. The age range of the

patients was 18–50 years. All of the wart lesions were located

on the external genitals or distal urethra. The patients were

excluded from the study if they had been treated with any

systemic immunotherapies within the preceding 4 weeks and/

or any topical antiviral drugs within the preceding 2 weeks,

had other STDs, or were pregnant or breastfeeding. Patients

presenting with vaginal and/or cervical condylomata were also

excluded from the study. The eligible patients were randomly

allocated into the ALA-PDT group or CO2 laser (control)

group in an allocation ratio of 3 : 1.

5-aminolaevulinic acid-photodynamic therapy

A 20% ALA (Fudan Zhangjiang Bio-Pharm Co. Ltd, Shanghai,

China) solution (w/v) was prepared by dissolving ALA in

sterile 0Æ9% NaCl immediately prior to its application. The

external genital skin around the lesions was washed with an

aqueous 0Æ4% chlorhexidine solution. A thin cotton swab or

an absorbent ball was soaked in the solution and was gently

inserted into the distal urethra or put on the external genitals,

to be in contact with and cover the warts and the adjacent

normal skin (5-mm border). Then small amounts of 20% ALA

were dropped on to the surface of the absorbent ball every

30 min. The lesions were occluded with food-grade cling film

and covered with thick gauze for light protection. Patients

were asked to lie still until illumination was performed. The

time interval between drug application and illumination was

3 h. Light irradiation of 100 J cm)2 at 100 mW was applied

to the lesion and the adjacent normal skin (5-mm border)

using a cylindrical helium–neon laser fibre (Yage Laser Appar-

atus Co. Ltd, Wuhan, China) emitting a 632Æ8-nm laser light.

In patients with multiple lesions, the warts were irradiated

one by one. If the warts were not totally removed after

1 week, a repeat treatment was given once a week for max-

imum of 3 weeks. The lesions were evaluated after each ALA-

PDT treatment session. The follow-up was conducted at 4, 8

and 12 weeks after the treatment finished.

CO2 laser treatment

After routine sterilization, the lesion was given topical anaes-

thesia with lidocaine. The lesions and their adjacent normal

skin (5-mm border) were vaporized with CO2 laser (Shanghai

Laser Institute, Shanghai, China) one by one until all the

lesions were completely removed. If the lesions were still

there after the first treatment, a repeat treatment was given

once a week for a maximum of 3 weeks. The follow-up was

conducted at 4, 8 and 12 weeks after the treatment finished.

Assessment

Laboratory-based evaluation was conducted for all the patients

for routine check of blood and urine, liver function and renal

function before and after the treatment. The number, location

and area of the warts and the adverse effects were evaluated

for each patient at each follow-up.


The indicators of efficacy were the clearance rate of wart le-

sions and the recurrence rate in those patients whose lesions

had completely cleared. Lesions were designated ‘complete

response’ if all (100%) of the lesions disappeared. A recur-

rence was defined as occurrence of the lesions after a complete

response. The clearance rate ¼ (the number of the warts

which had been cleared completely/the number of the

warts) · 100%.

The number of the warts in the two groups before treat-

ment was evaluated with the Wilcoxon matched pairs test.

The number of the warts after treatment and the adverse

effects in the two groups were evaluated with the v2 test.

Results

General characteristics

A total of 86 patients were enrolled into the study in which 65

patients were allocated into the ALA-PDT group and 21 patients

into the control group. Among the ALA-PDT group, 44 (67Æ7%)

were males and 21 (32Æ3%) females with a mean age of

35Æ2 years (SD ¼ 10Æ4 years). The duration of disease was

91Æ9 ± 90Æ5 (mean ± SD) days, and the number of the warts

was 1Æ5 ± 0Æ7 (mean ± SD). For this study, the locations of

warts were chosen as those found in the distal urethra in 17

(26Æ2%) patients and on the external genitalia in 48 (73Æ9%)

patients. Twenty-one patients (15 men and six women) with a

mean age of 35Æ9 (SD ¼ 9Æ3) years were allocated into the CO2

laser (control) group. The mean duration of disease in this

group was 89Æ9 ± 80Æ3 (mean ± SD) days. The average number

of warts was 1Æ9 ± 0Æ9 (mean ± SD). Five (23Æ8%) patients had

lesions located in the distal urethra and 16 (76Æ2%) in the exter-

nal genitalia rather than the distal urethra.

Clinical efficacy

In the ALA treatment group, a complete response was

observed among 15 of 17 (88Æ2%, 89Æ5% of the total warts)

patients with urethral CA (Fig. 1a,b) and in 45 of 48 (93Æ8%,

96Æ3% of the total warts) patients with lesions located in other

areas 1 week after the first treatment. All 65 patients achieved

complete response after two PDT treatments. In the CO2 laser

treatment group, a complete response was found in all 21

patients after one treatment (Table 1).

Follow-up for patients with complete response lasted for

12 weeks. Of these patients, four relapsing patients were

found in each group, resulting in a recurrence rate of 6Æ3%

and 19Æ1% for the ALA-PDT and control groups, respectively,

which was significantly different between these two groups

(P < 0Æ05).



PDT vs. CO2 laser vaporization in treatment of CA, K. Chen et al. 517

Safety evaluation

In all patients with ALA-PDT treatment, excellent tissue heal-

ing was achieved without the occurrence of any infection,

ulcers, scarring or urethral malformations. The mean time of

healing ranged from c. 7 to c. 10 days. The rate of adverse

reaction was 13Æ9% in the ALA-PDT group. Most patients

experienced a mild burning and/or stinging sensation during

irradiation, and a few patients experienced erythema, mild

oedema and erosion. The rate of adverse reaction was 100%

in the CO2 laser group. The side-effects mainly included pain,

erosion, ulcer, hyperpigmentation and hypopigmentation.

Urethral malformations occurred in two of five patients with

urethral CA. The rate of adverse reaction was significantly

higher in the CO2 laser group than in the ALA-PDT group

(P < 0Æ05).

Discussion

The conventional treatment of CA is usually targeted for

removing the macroscopic lesions and for avoiding their

recurrence. The commonly used modalities for CA recently

include three categories: topical remedies (0Æ5% podophyllo-

toxin, 5% fluorouracil); physical therapies (CO2 laser therapy,

cryotherapy, electrocautery) and/or surgical excision; immu-

notherapy, including systemic or topical immunoregulants.

Most conventional therapies are likely to be painful and lead

to topical tissue damage or scarring. Urethral malformations

can occur with laser or cryotherapy in patients with urethral

CA. In addition, it is troublesome that most conventional

treatments are associated with a high recurrence rate.1

ALA-PDT is a new technique introduced in the 1990s. The

mechanisms of cellular damage mediated by ALA-PDT include

a variety of biochemical and molecular reactions, creating a

complex process that leads to tissue destruction at both macro-

structural and microstructural levels. ALA is selectively

absorbed by tumour cells and rapidly proliferating cells, and

transformed to endogenous protoporphyrin IX (PpIX) after

the exogenous application of ALA. The PpIX is then activated

by red light, leading to the formation of singlet oxygen,

which leads to the killing or destruction of tumour cells and

proliferative cells.9 ALA-PDT is a well-established treatment

regimen for superficial epithelial skin tumours, but it is evi-

dent that inflammatory diseases of the skin and virus-induced

lesions can also profit from PDT. Some basic investigations

proved that ALA-PDT could inactivate the active viral particles.

In the study by Smetana et al.,10 ALA was added to cells that

Fig 1. Urethral condylomata acuminata. Clinical appearance (a)

before and (b) 1 week after photodynamic therapy using

5-aminolaevulinic acid.

Table 1 Response rates in patients with condylomata acuminata after one or two cycles of ALA-PDT or CO2 laser therapy

Number oftreatments

Location of lesions

TotalUrethral orifice Other area

ALA-PDT

(np ¼ 17, nw ¼ 19)

CO2 laser

(np ¼ 5, nw ¼ 7)

ALA-PDT

(np ¼ 48, nw ¼ 81)

CO2 laser

(np ¼ 16, nw ¼ 34)

ALA-PDT

(np ¼ 65, nw ¼ 100)

CO2 laser

(np ¼ 21, nw ¼ 41)

1 17 (89Æ47%) 7 (100%) 78 (96Æ30%) 34 (100%) 95 (95%) 41 (100%)2 2 (10Æ53%) – 3 (3Æ70%) – 5 (5%) –

Total 19 (100%) 7 (100%) 81 (100%) 34 (100%) 100 (100%) 41 (100%)

The numbers of the warts in the two groups before treatment were evaluated with the Wilcoxon matched pairs test. The numbers of thewarts after treatment in the two groups were evaluated with v2 test. In both the ALA-PDT and CO2 laser therapy groups, a similar complete

response was achieved for the warts in different location after completing all the treatments (P > 0Æ05). ALA-PDT, 5-aminolaevulinic acid-photodynamic therapy; np, number of patients; nw, number of wart lesions.



518 PDT vs. CO2 laser vaporization in treatment of CA, K. Chen et al.

were infected with the human immunodeficiency virus. Subse-

quent exposure to red light PDT drastically reduced the virus

titre. In guinea pigs infected with herpes simplex virus (HSV),

administration of ALA and exposure of the lesions to red light

shortened the duration of the appearance of vesicles from

more than a week to a few days and reduced the HSV titre in

the lesions.

In recent years, a few studies on ALA-PDT treatment for CA

have been conducted. A report by Wang et al.5 describes 164

patients with urethral CA who were treated with ALA-PDT.

The complete response rate was 95%, and the recurrence rate

was 5% after 6–24 months of follow-up. Stefanaki et al.6 used

ALA-PDT in the treatment of 12 male patients with CA.

Lesions were treated with 20% ALA and irradiated at the opti-

mal time with a dose of 70 J cm)2 or 100 J cm)2 light. The

overall cure rate was 72Æ9%, 12 months after treatment. Fehr

et al.7 and Uhlir et al.8 also treated the patients with CA with

ALA-PDT and achieved a satisfactory efficacy. CA, a benign

epidermal hyperplasia, is an infectious disease caused by HPV

infection.

ALA-PDT for the treatment of CA aims directly at two

points. Firstly, a large amount of ALA can permeate into the

warts through an abnormal stratum corneum, which is then

distributed and accumulated selectively in the hyperplastic tis-

sue.11–13 Subsequently, the ALA-PDT can cause the necrosis

and apoptosis of the keratinocytes infected by HPV selectively

and finally cleans up the warts.14 Secondly, in addition, ALA-

PDT may have some direct antiviral activity. The photosensitizer

can bind to virion surface glycoproteins, and thus impede the

early steps of the infection cycle.15

In this study, ALA-PDT for the treatment of CA also

showed a satisfactory therapeutic effect: the complete

response rate was 95% after one ALA-PDT treatment. Among

them, the complete response rates of urethral CA and the

external genitalia CA were 89Æ5% and 96Æ3%, respectively.

All patients (100%) achieved complete response after two

ALA-PDT treatments. The therapeutic effect of ALA-PDT was

as effective as the CO2 laser. Determination of the reasonable

ALA density, the optimal laser exposure time and the proper

total amount of laser irradiation is the key to obtaining the

optimal therapeutic results of ALA-PDT. It was reported that

patients with urethral CA were given topical 10% ALA fol-

lowed by PDT once a week, and the complete response rates

were 27% and 73%, respectively, after one and two ALA-

PDT treatments.5 In our study using 20% ALA-PDT for the

treatment of 17 patients with urethral CA, the complete

response rates were 89Æ5% and 100%, respectively, after one

and two ALA-PDT treatments. This prompts us to consider

that it is possible to improve the therapeutic effect with the

increase of ALA density. Ideally, irradiation should occur

after sufficient accumulation of intracellular PpIX to destroy

the lesions and to be relatively sparing of the adjacent nor-

mal skin. In this study, in accordance with the previous in-

vestigation (the optimal treatment time was between 30 min

and 3 h after ALA application),16 20% ALA was applied for

3 h followed by light exposure. In the study by Frank and

Bos,17 the lesions of seven patients with CA were treated

with 20% ALA for 14 h followed by irradiation with a dose

of 100 J cm)2. Only four patients obtained overall cure

3 months after treatment. The low therapeutic effect may

correlate to irradiation at an unsuitable time. A light over-

dosage often led to increased tissue damage18 and a light un-

derdosage could not achieve a satisfactory clinical therapeutic

effect. The irradiation in topical ALA-PDT was applied at a

light dose of between 60 and 250 J cm)2.19 In addition, the

exposed degree of the lesion during the therapeutic process

is also one of the factors influencing the therapeutic effect.

In this study, the complete response rate was 89Æ5% and

96Æ3% in the patients with urethral CA and in those with CA

in other areas, respectively, after one PDT treatment. Com-

pared with the latter, the former had an inferior therapeutic

effect. Such a lower effect can be correlated to the difficulty

of exposure of the lesion to treatment.

Conventional treatment usually targets only visible wart le-

sions, and high recurrence rates (10–65%) result partly from

the location of HPV as far as 1 cm from the clinical border of

the warts.12 In the study, the recurrence rate in ALA-PDT

group (6Æ3%) was much lower than that in CO2 laser treat-

ment group. The reason for the lower recurrence rate after

ALA-PDT may be related to simultaneous treatment of micro-

scale and subclinical lesions.7

Conventional treatments are often painful, and sometimes

lead to traumatic infection and scarring. In this study, urethral

malformations occurred in two of five patients with urethral

CA in the CO2 laser group. Compared with conventional treat-

ment, topical ALA-PDT is a safe new treatment for CA. The

most frequent side-effect is only a mild burning and/or sting-

ing sensation during irradiation.

Based on the present study, it can be concluded that topical

application of ALA-PDT is a simpler, more effective and safer

therapy with a lower recurrence for treatment of CA compared

with conventional CO2 laser therapy.

References

1 Krogh GV. Clinical relevance and evaluation of genitoanal

papilloma virus infection in the male. Semin Dermatol 1992; 11:229–40.

2 Kennedy JC, Pottier RH, Pross DC. Photodynamic therapy withendogenous protoporphyrin IX: basic principles and present

clinical experience. J Photochem Photobiol B Biol 1990; 6:143–8.3 Lui H, Anderson RR. Photodynamic therapy in dermatology: recent

developments. Dermatol Clin 1993; 11:1–13.4 Nelson JS, Jeffes EWB, Weinstein JL et al. Topical 5-aminolevulinic

acid for the photodynamic therapy of psoriasis and actinic keratos-es. Lasers Surg Med Suppl 1995; 7:43.

5 Wang XL, Wang HW, Wang HS et al. Topical 5-aminolaevulinicacid-photodynamic therapy for the treatment of urethral condylo-

mata acuminata. Br J Dermatol 2004; 151:880–5.6 Stefanaki IM, Georgiou S, Themelis GC et al. In vivo fluorescence

kinetics and photodynamic therapy in condylomata acuminata. Br JDermatol 2003; 149:972–6.

7 Fehr MK, Hornung R, Degen A et al. Photodynamic therapy ofvulvar and vaginal condyloma and intraepithelial neoplasia using



PDT vs. CO2 laser vaporization in treatment of CA, K. Chen et al. 519

topically applied 5-aminolevulinic acid. Lasers Surg Med 2002; 30:273–9.

8 Uhlir M, Jirsa M, Otcenasek M et al. Photodynamic therapy ofbenign viral vulvar lesions. Ceska Gynekol 2002; 67:320–3.

9 Loh CS, MacRobert AJ, Bedwell J et al. Oral vs. intravenous admin-istration of 5-aminolaevulinic acid for photodynamic therapy. Br J

Cancer 1993; 68:41–51.10 Smetana Z, Malik Z, Orenstein A et al. Treatment of viral infections

with 5-aminolevulinic acid and light. Lasers Surg Med 1997; 21:351–8.

11 Svaasand LO. Dosimetry model for photodynamic therapy with

topically administered photosensitizers. Lasers Surg Med 1996;18:139–49.

12 Schultz RE, Miller JW, MacDonald GR et al. Clinical and molecularevaluation of acetowhite genital lesions in men. J Urol 1990;

143:920–3.13 Barry B. Properties that influence percutaneous absorption. In:

Dermatological Formulations: Percutaneous Absorption (Barry B, ed.). NewYork: Marcel Dekker, 1983; 145–6.

14 Oleinick NL, Morris RL, Belichenko I. The role of apoptosis inresponse to photodynamic therapy: what, where, why and how.

Photochem Photobiol Sci 2002; 1:1–21.15 North J, Neyndorff H, King D, Levy JG. Viral inactivation in blood

and red cell concentrations with benzoporphyrin derivative. BloodCells 1992; 18:129–40.

16 Ross EV, Romero R, Kollias N et al. Selectivity of protoporphyrin IXfluorescence for condylomata after topical application of 5-amino-

laevulinic acid: implications for photodynamic treatment. Br J Der-matol 1997; 137:736–42.

17 Frank RG, Bos JD. Photodynamic therapy for condylomata acumi-

nata with local application of 5-aminolevulinic acid. Genitourin Med1996; 72:70–1.

18 McCaughan JS, Guy JT, Hicks W et al. Photodynamic therapy forcutaneous and subcutaneous malignant neoplasms. Arch Surg 1989;

124:211–16.19 Peng Q, Warloe T, Berg C et al. 5-Aminolevulinic acid-based

photodynamic therapy: clinical research and future challenges.Cancer 1997; 79:2282–308.



520 PDT vs. CO2 laser vaporization in treatment of CA, K. Chen et al.

EPIDEMIOLOGY AND HEALTH SERVICES RESEARCH DOI 10.1111/j .1365-2133.2006.07608.x

Teledermatology in the U.K.: lessons in service innovationT.L. Finch, F.S. Mair* and C.R. May

Centre for Health Services Research, University of Newcastle upon Tyne, 21 Claremont Place, Newcastle upon Tyne NE2 4AA, U.K.

*Department of General Practice and Primary Care, University of Glasgow, 1 Horselethill Road, Glasgow G12 9LX, U.K.

CorrespondenceTracy Finch.


Accepted for publication27 June 2006

Key wordsinnovation, teledermatology, telemedicine


Summary

Background Teledermatology has the potential to revolutionize the delivery ofdermatology services by facilitating access to specialist services at a distance. Inthe U.K. over the previous decade there have been numerous attempts at introdu-cing and using teledermatology; however, the development of teledermatology asroutine service provision remains limited.Objectives To identify factors that promote successful use of teledermatology as apart of routine service provision.Methods A longitudinal qualitative study of teledermatology, drawing on data fromin-depth semistructured interviews; observations of systems in practice; and pub-lic meetings. Data were analysed collectively by the research team using estab-lished qualitative analytical techniques to identify key thematic categories. Thesample consisted of teledermatology services within the U.K. (n ¼ 12) studiedover 8 years (1997–2005). Individual participants (n ¼ 68 interviews) were con-sultant dermatologists, researchers, teledermatology nurses, administrators,patient advocates, general practitioners and technologists.Results The analysis compared services that did or did not become part of routinehealthcare practice to identify features that supported the normalization of tele-dermatology. Requirements for using and integrating teledermatology into prac-tice included: political support; perceived benefit and relative commitment thatoutweighs effort; pragmatic approaches to proving efficacy and safety; perceptionof risk as being ‘manageable’ on the basis of professional judgement; high levelsof flexibility in practice (in terms of individuals, technology and organization);and reconceptualizing professional roles.Conclusions Successful implementation of teledermatology as a routine servicerequires greater understanding of and attention to the interplay between socialand technical aspects of teledermatology, and how this is accommodated both byhealthcare professionals and the organizations in which they work.

Teledermatology is the delivery of dermatology services ‘at a

distance’ through the use of information and communication

technology.1 It represents part of a general shift in medicine

towards the increasing use of technology to address problems

of inefficiencies in healthcare provision in the context of in-

creasing demand for services, and to overcome inequities in

access to services across geographical regions.2 In the U.K.

over the past decade the potential of telehealthcare has gener-

ated much interest among dermatology service providers, who

have been grappling with the problem of lengthy waiting lists

for routine dermatological appointments.3

Dermatology has, from the outset, been hailed as a medical

specialty particularly well suited to telehealthcare, for it relies

to a large extent on visual information for diagnosis.1,4 Tele-

dermatology has been practised both in real time through the

use of live interactive videoconferencing (synchronous), and

as store-and-forward systems involving the transmission of

still images and textual information for remote assessment and

subsequent review at a later point in time (asynchronous).

There has been considerable research globally5–7 and within

the U.K.8,9 to demonstrate and evaluate the potential of tele-

dermatology. In the U.K., efforts were initially focused on real-

time teledermatology; however, store-and-forward systems

have become increasingly favoured as being more cost-

efficient and convenient for healthcare providers.8,10 Teleder-

matology has also been used for various purposes, including

triage,11 diagnostic and management services,5,9,12 and

advice/opinion for primary care practitioners.3 It has been set



up in a number of ways, including as direct referral from pri-

mary care using images and clinical history sent to secondary

care dermatology services for assessment/opinion and for

triage of referrals; and as facilitating community-based clinics

led by nurses or general practitioners (GPs) with a special

interest in dermatology (using teledermatology for consultant

support). Teledermatology has generally demonstrated high

levels of concordance in diagnosis and management plans

compared with face-to-face consultations.4,7,13,14 Studies asses-

sing patients’ views of teledermatology tend to report positive

results,15–19 although many such studies lack analysis of the

needs of different dermatology patient groups.20 The derma-

tology community itself has responded variably to the idea of

teledermatology, ranging from pockets of enthusiasm to resist-

ance and concerns about perceived threats to the professional

practice of dermatology and the doctor–patient relationship.21

Teledermatology, despite its potential for facilitating service

provision, does raise important questions for professional

practice and about the way in which services must be recon-

figured to accommodate new ways of working. In telehealth-

care generally, there is a problem of pilot and demonstration

projects that fail to become a normal part of healthcare deliv-

ery.22 Introducing a new technical system in itself raises prob-

lems and tensions around new technologies, professional roles

and identities, and integrating old and new approaches to

healthcare provision, because they invariably change the ways

that work is organized and undertaken.23,24 Where service de-

velopments based on the introduction of new technology are

combined with efforts at conducting evaluation research, mak-

ing telehealthcare ‘work’ in clinical practice becomes addition-

ally complicated.25

This paper aims to further research on teledermatology

beyond evaluation studies to provide an understanding of fac-

tors that contribute to the success of teledermatology in practice,

through a large-scale qualitative study of teledermatology in

the U.K. that spans the last 8 years.


Participants and settings

Between 1997 and 2005, we have undertaken a longitudinal

qualitative study of teledermatology services, both ‘real’ and

‘experimental’, and their development in depth, and over

time. The data collected covers representative instances of

National Health Service (NHS) provision of teledermatology in

the U.K., consisting of 12 known services dating back from

1994 to the present: one government sponsored (T1); a com-

mercial provider (no longer in operation) (T4); research stud-

ies that were discontinued (T2, T3, T5, T6); GP-based

systems still in operation (T7, T8, T10); and nurse-led sys-

tems still in operation (T9, T11, T12). The services that were

developed as research studies included one that was based on

real-time videoconferencing, and three that used store-and-

forward. All of the research studies aimed to assess diagnostic

concordance and the utility of teledermatology for patient

management. The teledermatology services that remain in op-

eration are all based on store-and-forward systems, and

include some that operate as nurse and/or GP with special

interest-led clinics with a store-and-forward link for consultant

review and/or triage, and others that are set up for use by

GPs either to access consultant opinion/advice or for triage

purposes. We cannot provide further details of the telederma-

tology services because Multicentre Research Ethics Committee

(MREC) approval of the study, and consent provided by indi-

vidual participants, was granted on the condition that we do

not disclose any details about the sites and services that might

identify individual respondents. Given the relatively small

number of services in operation, even small details (such as

locations, dates etc.) are likely to identify respondents and as

such have had to be omitted.

For the purpose of the analysis, the services we have stud-

ied are referred to throughout the Results section as ‘normal-

ized’ or ‘non-normalized’. In this context, ‘normalized’ is

used to refer to services that have become a routine compon-

ent of dermatology services in a particular area, even if this is

on a relatively small scale26 in relation to the total number of

patients managed. ‘Non-normalized’ refers to pilot projects or

research studies that were discontinued after a brief period of

time.

Interview respondents for the study were selected on the

basis of known experience in teledermatology or for their

patient advocacy roles in relation to dermatology. Appropriate

ethical and research governance approval was obtained, and

stringent procedures were followed to ensure the anonymity

of individual participants and the services to which they are

linked.

Data collected included key informant interviews (n ¼ 68);

observations of public meetings (n ¼ 2); observation; and tex-

tual analysis of project documents. All interviews were audio-

taped, transcribed, and then submitted for data analysis.

Interviewees included consultant dermatologists (19),

researchers (17), teledermatology nurses (11), service admin-

istrators (8), patient advocates (4), technologists (6) and GPs

(3). These figures represent total interviews rather than partici-

pants, as some provided multiple interviews over time.

Two full public meetings were also included: British Associ-

ation of Dermatologists Conference about teledermatology

(June 2001); and ‘How to Improve Dermatology Services:

Lessons from Action on Dermatology’ (April 2004). Partici-

pants were interviewed up to four times (typically once or

twice). Interviews were conducted either in-person, at a site

convenient for the respondent, or over the telephone. A semi-

structured interview guide was used for all interviews, but

was varied according to the specific role of interviewee and

context of interview. Across all interviews, however, we

explored: aspects of the development of the service (from idea

through to practice); organizational context and clarification

of roles of key participants; purpose of the particular service

development and evaluation (if appropriate); and experiences

of service development and utilization (both positive and

negative).



522 Teledermatology in the U.K., T.L. Finch et al.

Ethical approval for this study was granted by the Northern

and Yorkshire MREC, and all employing NHS Trusts of indi-

vidual clinical respondents, as required by the Department of

Health Research Governance Framework for Health and Social

Care (2001, revised 2005).

Data analysis

Data for analysis consisted of interview transcripts. However,

contemporaneous field notes and project documentation were

used to support and strengthen the development and interpret-

ation of the interview data. Data collection and analysis pro-

ceeded in an iterative fashion and a constant comparative

approach to data analysis was used.27 All authors participated

in ‘data clinics’, during which transcripts were coded and key

themes were identified and agreed.

Results

The results of this study reveal the degree of modifications in

practice that are required in order to make teledermatology

operate successfully. Key themes in the process of normaliza-

tion included: the policy context; perceived benefit and related

commitment; evidence gathering and both proving safety and

managing risk; reorganizing services; and issues surrounding

professional roles and boundary crossing. These themes and

some of the associated factors that emerge as promoting or

inhibiting successful innovation in the context of teledermato-

logy services are illustrated below.

Policy context

New services are not promoted in a vacuum. The policy context

in which new services are introduced may either facilitate or

inhibit new initiatives. This is certainly true of the teledermatol-

ogy services we studied, as most of these were conceived

through high-level central or local health policy support (see

Table 1). However, where such support was not sustained, and

thus not translated into resources, services failed to normalize. It

has been suggested that such policy directives may increase an

organization’s willingness to adopt new health technologies but

may still do little to aid the organization’s ability to integrate

new technological systems into existing practices successfully.28

Perceived benefit and related commitment

The introduction of teledermatology services was often chal-

lenging (see Table 2), and the full impact of these services on

working practices only became apparent to participants once

they started working with the system in practice, and began

the process of identifying and resolving problems in order to

make the teledermatology system work. The technical systems

thus required new ways of working: unless the objectives of

teledermatology were clear, respondents questioned whether

the additional effort needed was worth it (Table 2). The tele-

dermatology services that became normalized were those for

which the perceived benefits, for example saving patients’

and/or health professionals’ travelling time and costs, or redu-

cing waiting times, clearly outweighed the effort and commit-

ment required to make the system work in the local setting.

The issue of the need for widespread and ongoing support by

all of those contributing to such service reconfigurations was a

recurring theme.

Table 1 Policy background

‘I mean this policy, I presume it was from the top, Downing

Street or wherever these politics come from. I suppose it musthave been because I had this commissioner geezer come into

my office and say you have got to do some [teledermatology],but that’s effectively what he said…’ (Consultant Dermatologist

– T11)‘I have been trying to locate regional policy leads in England to

talk about telemedicine and then I found out that they’ve alldisappeared because it’s all been completely reorganized.’

(Telemedicine Service Manager – T1)

Table 2 Perceived benefit and related commitment

Normalized Non-normalized

‘Oh definitely I had many tears… at the beginning because I feltvery isolated too. Here I was trying to, my IT knowledge was very

poor, I’d hardly touched a computer [it all] seemed such a newand foreign system to me as well as trying to get to grips with

dermatology. I did feel a bit isolated because I didn’t actually haveanyone to run to next door to help me and my IT problems and

things tended to have to be sorted out over the phone and some-times would take hours. So it was really quite scary at the begin-

ning, and, but amazingly, I mean at times I felt, I thought I can’tdo this, I’m going to pack this in, I can’t do it, but after I would

say about six months, as I got more familiar with it and obviouslymy knowledge increased it became much easier and now one can

see the benefits of the service, it’s great, it’s really good.’ (Tele-dermatology Nurse – T9)

Interviewer: Well that all comes down to that old chestnut about what was it for,that was never quite clear…‘No, it’s never been defined as to ‘‘this service is specifically forthis’’. It hasn’t, it’s just been ‘‘we’ve got some money, we’ll do

this and see how it goes, get the waiting list down’’. Some areasof the project were wanting to get waiting lists down and that

really I think is all we were looking at, others were looking atpatients, others were looking at fulfilling the fact that we should

be advancing the technology and I found it very difficult to say‘‘we’re doing this teledermatology because…’’. In the end I didn’t

know why we were doing it.’ (Teledermatology Nurse – T5)



Teledermatology in the U.K., T.L. Finch et al. 523

Evidence gathering, ‘proving’ safety and managing risk

With an increasing emphasis on evidence-based practice,

those developing teledermatology services appeared acutely

aware of the need to produce some level of ‘evidence’ to sup-

port their continued activities and to demonstrate that their

system was a safe mode of practice. Innovations in healthcare,

because they require sometimes quite profound change and

adaptation in existing practices, are perceived as posing new

risks and presenting fresh dilemmas for clinicians, thus much

time and effort was expended in trying to gather evidence to

prove the safety and utility of the services we studied. Formal

evaluation studies tended to be perceived as time consuming

and/or difficult to conduct within a service setting, and some

of those involved clearly felt the effort spent on such evalu-

ations was not entirely justifiable, or impeded the achieve-

ments of service objectives (Table 3). Integration of

experimental and real NHS services was a problem.25 How-

ever, those developing teledermatology services were focused

on the potential risks and responsibilities posed by such new

ways of working, and this was something that participants

often built into the design of the system (technically or soci-

ally), or addressed through ongoing modification of the

system (or participants’ use of it) in practice. The telederma-

tology services that were sustained in practice tended to be

those for which potential risk was acknowledged and

addressed by building safeguards into the system (along with

deference to professional judgement and responsibility). The

providers of such services also emphasized ‘audit’ in terms of

close monitoring of the effects of using the system on patient

outcomes, rather than formal (scientific) ‘evaluation’ (see

Table 3). Subsequently, as such services came into regular use

alongside existing services, the potential risks were accorded

less importance and teledermatology was seen to be little dif-

ferent from existing dermatology practices.

Reconfiguring services

In order for teledermatology services to become accepted in

practice, there was a need for acknowledgment of an ongoing

process of modification (see Table 4). However, as evident

particularly in the earlier teledermatology services we studied,

the hardware and software were frequently overemphasized

(especially image quality) to the neglect of the crucially im-

portant issue of how the service needed to be adapted by its

users to make it workable in practice. If equipment was sim-

ply ‘parachuted’ into an existing clinical context, without ade-

quate and ongoing support, the chances for successful

introduction and integration of the teledermatology system

into practice were severely compromised.

Professional roles and boundary crossing

The introduction of technological systems within clinical prac-

tice can promote profound changes in professional roles, and

affect the organization of professional knowledge and practice.

The introduction of such systems requires the acknowledg-

ment of frequently invisible investment in ‘workability’ by

health professionals. The results of this study reveal how this

process of achieving workability in teledermatology required

changes in professional roles and boundaries of responsibility.

At the individual level, professionals such as nurses and con-

Table 3 Evidence gathering, proving safety and managing risk


‘[…] we are all slightly obsessed by, well in theory we’re obsessedwith evidence-based medicine and actually in practice, we just get

on and do it [yeah], if treatment works we just go and use it […].We have evaluated our project, sort of, nowhere near as thoroughly

or in any great academic way [as others, we] never received anyfunding and did not have the time to do it, so we did set out certain

criteria however before we started the project, we would see XXXpatients minimum in the first year, we wanted to see waiting lists

come down and so on, and all these things have happened […] wehave seen a drop in the number of patients being referred up to the

hospital so I think it is working.’ (Consultant Dermatologist – T11)

‘… going back to the ethics bit in relation to the GPs is the morecomplicated the research procedures become the more difficult it’s

going to be for actual practising practitioners to be involved. Allthe actual procedures interfere with the actual technology, you

could have argued with what we’re doing that if you hadn’t hadto make the allowances for informed consent the whole thing

could have been done in a different way again, the delays wouldhave been less, so there’s a delay related to that allowing people

to consider the whole thing and take away the written informa-tion and talk to people about it.’ (Project Lead – T6)

‘Yes, to start with, I reviewed all the patients but we very rapidly

found that most patients were better by the time I saw them, so wewere able to show that it was possible not only to make a diagnosis,

to provide a management and follow-up plan, arrange investiga-tions, which would be available by the time I saw them. We were

also able rapidly to see that it was possible to make an accuratediagnosis or differential diagnosis very quickly.’ (Consultant Derma-

tologist – T9)

‘I don’t know about [colleague] but I suspect there’s a learningcurve, in dermatology a lot of what you learn is spot diagnosis,

there are lots of books on this, is a picture a little bit of informa-tion, what’s the diagnosis, we all can do the spot diagnosis but

when it comes to actually deciding when it’s a real patient and it’snot just to get 10/10 or 2/10 in a test, it might be a bit different.

But I am certainly viewing it a bit like that, it’s going to be spotdiagnosis and working out management plans in a way you would

for that sort of test, whether it be a test that you mark your owntest that you personally do for yourself I don’t know, about how

we can cope with volume.’ (Teledermatology Consultant – T5)




sultants needed to make changes or adjustments to their (tra-

ditionally perceived) professional roles in order to make the

new systems ‘work’, in terms of achieving the best possible

outcomes for patients. Normalization of the service was asso-

ciated with acknowledgment and acceptance of such changes

in responsibility, as illustrated in Table 5. However, concerns

were sometimes raised about whether teledermatology services –

and telehealthcare services more generally – were attempting

to shift responsibility for care from secondary to the primary

care sector (Table 5). The capacity of primary care to absorb

and deal effectively with the potential added burdens associ-

ated with the introduction of teledermatology was an issue

raised by some participants.

Discussion

In this study, we have been able to observe closely, and over

time, how different approaches to innovating the delivery of

Table 5 Professional roles and boundary crossing


‘I think because really it’s a nurse-led service, and is it working […]

it’s helping us to keep our waiting list manageable, patients like it,and the nurses are there to nurse, the doctors diagnose and pre-

scribe whereas nurses are increasingly prescribing but don’t diag-nose, they nurse people and make people better, they tell them how

to put the creams on and so on, so they do spend time with them,so I think it is working from that point of view, in that’s what the

patients need […] what it’s done it’s increased dermatologicalexpertise throughout the [region], that’s what it was about is in-

creasing our capacity to help patients with dermatological condi-tions, and to some extent it doesn’t really matter what their

qualifications are, providing they are able to do that […]. I thinkmost people accept that dermatological nursing is, in it’s own right,

a very satisfactory way of dealing with people with skin diseases.’(Consultant Dermatologist – T11)

‘If telemedicine works in one area how many more areas will

there become […] and are they [GPs] really as equipped to dealwith this addition to their, to their [existing] role, that may be

something which some GPs might consider to be just one step toofar you know. Telemedicine for dermatology one minute then

ENT then, then you know. And [then it’s] ‘‘what’s the hospitaldoing then?’’ ’ (Teledermatology Project Manager – AOD pilot,

non-normalized)

‘I think it’s [my role] still limited in teledermatology, because wecan’t, if you’re in a clinic with a doctor and a doctor says to a

patient you have got eczema, you have got psoriasis, here’s a pre-scription, I can then go in to the patient and say here’s your pre-

scription, you’ve got eczema and I can spend ten minutes talkingto the patient, going through the treatment saying this is how you

should use your treatments. When you’ve got a patient in teleder-

matology I know has got eczema, but the doctor hasn’t made adiagnosis, so I’m in a way more confined because I can’t give def-

inite treatment advice even though I know it would benefit thepatient while they were waiting. I could say to a patient, ‘‘Go and

see a GP and see if he’ll give you some Ulavate,’’ but I had to stopmyself. I felt as if I wanted to say so much more, wanted to do so

much more but I had to be so careful in case I gave wrong infor-mation, only because it hadn’t been given that seal of approval as

that is what the patient’s got.’ (Teledermatology Nurse – T5)

Table 4 Reconfiguring services


‘In my honest opinion I do feel that [organizational issues are]

going to be a problem for anybody who sets [teledermatology] upbecause face-to-face consultations are a must and you should have

side by side the technology and investment in technology andhuman resources as well to do all these consultations...’ (Consultant

Dermatologist – T7)

‘No, we purchased [the teledermatology system] off the shelf,through a tendering process, but we worked with a fairly small,

flexible organization who were willing to try and adapt the system,they couldn’t adapt it completely to our own needs but they were

very good at listening to what we wanted because we were […]quite a large pilot for them. We had a very, really useful two-way

communication system set up and they took on board an awful lotof the suggestions we made in order to make the system work bet-

ter in the way that we particularly used it…’ (Teledermatology Pro-

ject Manager – T12)

‘[…] the equipment was a bit of a pain sometimes when it broke

down. I do know that initially, they were very attentive to startwith.’

Interviewer: Well, it was in development, wasn’t it?

‘But from about two thirds of the way through the interest had

waned so that the girls would be stuck there with nothing work-ing and the patient still there, and then to have to say to patients,

‘‘I’m terribly sorry,’’ so that perhaps wasn’t good, we want com-mitment of the people properly to the maintenance of the equip-

ment and everything else.’ (Teledermatology Nurse Manager – T5)




dermatology services through the introduction of telederma-

tology have succeeded or failed and the contributing factors

underlying these outcomes. The study is unique (not only in

relation to teledermatology but for telehealthcare generally) in

that it analyses the development of most of the known tele-

dermatology services on a national level, over a significant

period of time (8 years), thus providing a sound analysis of

the differences between success and failure in implementation

efforts in this context.

Various factors promoting the successful implementation

of teledermatology were identified, and include: (i) a per-

ception of the benefits of new services that outweighed the

effort and commitment required to make such services

work; (ii) a willingness to rely on demonstrated utility in

practice rather than formal scientific evidence; (iii) allowing

health professionals to make modifications to the technical

system, rather than imposing new practices around fixed

technological systems; (iv) accommodating uncertainty and

allowing risk management through professional judgement

rather than protocol; and (v) flexibility in division of labour

and responsibility across nursing and medical domains.

‘Successful’ teledermatology services are based on working

relationships (nurses, consultants, GPs) underpinned by trust

and confidence in professional judgement, combined with

flexibility within the service to allow for modifications as

required to achieve follow-up and continuity of care for

patients. The original ‘policy’ vision of how teledermatology

would be utilized, as a technological fix for long waiting

lists and consultant shortages, failed to be realized. Never-

theless, several teledermatology services have survived and

proved successful, occasionally playing a major role in ser-

vice delivery, while at other times serving as a useful

adjunct to traditional services. It is also important to note

that while teledermatology was at the outset often viewed

as a diagnostic service it appears to have gained increasing

acknowledgment as a triage and management service.3 Flexi-

bility with regard to how such services are deployed and

utilized, based on local requirements and circumstances, and

perceived benefits relative to the necessary commitment to

making such services work, has proven essential for teleder-

matology to have a role in service provision. Cross-sector

and professional support for such initiatives was also a cru-

cial factor.

This study demonstrates that, when introducing telederma-

tology services, the ways in which services become reconfig-

ured must be given careful thought and consideration, in

terms of understanding not only what is required of the

technological system to enable professionals to use it, but

also acceptance on the part of administrators that the service

will need to be redesigned and modified in practice to fit in

with the local context of service provision, and to meet the

particular needs of the local community. The lessons learnt

from following the history and development of teledermatol-

ogy services merit consideration by those involved in the de-

velopment and implementation of other telehealthcare

innovations.

Acknowledgments

We wish to acknowledge the material and intellectual contri-

bution of Dr Maggie Mort to the ongoing programme of

research on which this paper is based, and for her comments

on the original manuscript. We thank all of our study partici-

pants for giving their time to participate in this study, and to

Ms D. Mukadam and Ms Cheryl Wiscombe for administrative

support. Funding was supplied by Department of Health

(grant ICT/032) and Economic and Social Research Council

(grants L218 25 2067 and RES 000270084).

References

1 Burdick AE, Berman B. Teledermatology. Adv Dermatol 1997; 12:19–

45.2 Preston J, Brown FW, Hartley B. Using telemedicine to improve

healthcare in distant areas. Hosp Community Psychiatry 1992; 43:25–32.3 Mallett RB. Teledermatology in practice. Clin Exp Dermatol 2003;

28:356–9.4 Lewis K, Gilmour E, Harrison PV et al. Digital teledermatology for

skin tumours: a preliminary assessment using a receiver operatingcharacteristics (ROC) analysis. Journal of Telemedicine and Telecare 1999;

5:S57–8.5 Oakley AMM, Astwood DR, Loane M et al. Diagnostic accuracy of

teledermatology: results of a preliminary study in New Zealand.N Z Med J 1997; 110:51–3.

6 Perednia DA, Wallace J, Morrisey M et al. The effect of a telederma-tology program on rural referral patterns to dermatologists and the

management of skin disease. Medinfo 1998; 9:290–3.7 Tait CP, Clay CD. Pilot study of store and forward teledermatology

services in Perth, Western Australia. Australas J Dermatol 1999;40:190–3.

8 Eedy DJ, Wootton R. Teledermatology: a review. Br J Dermatol2001; 144:696–707.

9 Wootton R, Bloomer SE, Corbett R et al. Multicentre randomisedcontrol trial comparing real time teledermatology with convention-

al outpatient dermatological care: societal cost-benefit analysis. BMJ2000; 320:1252–6.

10 Loane MA, Bloomer SE, Corbett R et al. A comparison of real-timeand store-and-forward teledermatology: a cost-benefit study. Br J

Dermatol 2000; 143:1241–7.

11 White H, Gould D, Mills W et al. The Cornwall dermatology elec-tronic referral and image transfer project. Journal of Telemedicine and

Telecare 1999; 5:S85–6.12 Gilmour E, Campbell SM, Loane MA et al. Comparison of telecon-

sultations and face-to-face consultations: preliminary results of aUnited Kingdom multicentre teledermatology study. Br J Dermatol

1998; 139:81–7.13 High WA, Houston MS, Calobrisi SD et al. Assessment of the accur-

acy of low-cost store-and-forward teledermatology consultation.J Am Acad Dermatol 2000; 42:776–83.

14 Whited JD, Hall RP, Simel DL et al. Reliability and accuracy of der-matologists’ clinic-based and digital image consultations. J Am Acad

Dermatol 1999; 41:693–702.15 Collins K, Walters S, Bowns I. Patient satisfaction with telederma-

tology: quantitative and qualitative results from a randomized con-trolled trial. J Telemed Telecare 2004; 10:29–33.

16 Sibson L, Dunn R, Evans J et al. The virtual mole clinic: preliminaryresults from the Plymouth skin cancer screening study using tele-

medicine. Med Inform Internet Med 1999; 24:189–99.




17 Jones DH, Crichton C, Macdonald A et al. Teledermatology in theHighlands of Scotland. J Telemed Telecare 1996; 2:7–9.

18 Lowitt MH, Kessler II, Kauffman CL et al. Teledermatology and in-person examinations. Arch Dermatol 1998; 134:471–6.

19 Loane MA, Corbett R, Bloomer SE et al. Diagnostic accuracy andclinical management by realtime teledermatology. Results from the

Northern Ireland arms of the UK Multicentre TeledermatologyTrial. J Telemed Telecare 1998; 4:95–100.

20 Williams TL, May CR, Esmail A et al. Patient satisfaction with tele-dermatology is related to perceived quality of life. Br J Dermatol

2001; 145:911–17.

21 Finch T, May C, Mort M, Mair F. Telemedicine, telecare, and thefuture patient: innovation, risk and governance. In: New Technologies

in Health Care: Challenge, Change and Innovation (Webster A, Wyatt S,eds). Basingstoke: Palgrave Macmillan, 2006; 84–96.

22 Clough K, Jardine I. Telemedicine: five years on – what progress?Br J Healthcare Comput Inf Manage 2003; 20:21–3.

23 Sicotte C, Lehoux P. Teleconsultation: rejected and emerging uses.Methods Inf Med 2003; 42:451–7.

24 Lehoux P, Sicotte C, Denis JL et al. The theory of use behind tele-medicine: how compatible with physicians’ clinical routines? Soc Sci

Med 2002; 54:889–904.25 Finch T, May C, Mair F et al. Integrating service development with

evaluation in telehealthcare: an ethnographic study. BMJ 2003;327:1205–9.

26 May CR, Harrison R, Finch T et al. Understanding the normaliza-tion of telemedicine services through qualitative evaluation. J Am

Med Inform Assoc 2003; 10:596–604.

27 Strauss A. Qualitative Analysis for Social Scientists. Cambridge: CambridgeUniversity Press, 1987.

28 Taylor SM, Elliot K, Taylor S. Community-based heart health pro-motion: perceptions of facilitators and barriers. Can J Public Health

1998; 89:406–9.




EPIDEMIOLOGY AND HEALTH SERVICES RESEARCH DOI 10.1111/j .1365-2133.2006.07617.x

The Family Dermatology Life Quality Index: measuring thesecondary impact of skin diseaseM.K.A. Basra, R. Sue-Ho* and A.Y. Finlay

Department of Dermatology, Cardiff University, Wales College of Medicine, Heath Park, Cardiff, CF14 4XN, U.K.

*Department of Microbiology, Wycombe General Hospital, Queen Alexandra Road, High Wycombe, Bucks, HP11 2TT, U.K.

CorrespondenceM.K.A. Basra.

E-mail: [email protected];

[email protected]


Key wordsdermatology, family, health-related quality of life,

quality of life, secondary impact

Conflicts of interestA.Y.F. is joint copyright owner of the

Dermatology Life Quality Index (DLQI) and

Family Dermatology Life Quality Index (FDLQI).

A.Y.F.’s department gains some income from the

use of the DLQI. A.Y.F. gives paid consultancy

advice to several pharmaceutical companies but this

study has not involved any pharmaceutical

company and has not received any external

funding. M.K.A.B. is a joint copyright owner of

the FDLQI.

This study was conducted at the University

Hospital of Wales, Cardiff, U.K.

Summary

Background Skin diseases are known to have a major impact on the lives of patientsand their families. Many instruments are available to measure the health-relatedquality of life (HRQoL) of patients but no measure has been developed so far toquantify the secondary impact on family members of the patients.Objectives To develop and validate a dermatology-specific instrument to measurethe adverse impact on the HRQoL of family members of patients with skindisease.Methods Detailed semi-structured interviews were conducted with family membersof patients to identify different aspects of HRQoL affected. An initial draft versionof the questionnaire based on the main topic areas was pilot tested to assess theface and content validity. A 10-item questionnaire, the Family Dermatology LifeQuality Index (FDLQI), was finalized after modifications to the draft question-naire based on feedback from families and dermatology professionals and onitem reduction. Psychometric evaluation was conducted on a new cohort of fam-ily members (n ¼ 132) who completed the FDLQI and the patients (n ¼ 109)who completed the Dermatology Life Quality Index (DLQI).Results Fifty-nine different aspects of family members’ HRQoL were identifiedfrom the analysis of the interviews, which were categorized into main topicareas. Factor analysis of 10 items of the final questionnaire revealed two factorsand together these explained 60% of the common variance. The FDLQI demon-strated high internal consistency (Cronbach’s a ¼ 0Æ88) and test–retest (intraclasscorrelation coefficient ¼ 0Æ94) reliabilities. The responsiveness of the instrumentto change was shown by significant change in the family members’ FDLQI scoresin cases where patients’ clinical condition either improved or worsened. Con-struct validity was assessed by testing a number of a priori hypotheses. A strongcorrelation was seen between the family members’ FDLQI scores and patients’DLQI scores (r ¼ 0Æ69), a significantly higher FDLQI score was seen for inflam-matory skin diseases compared with noninflammatory diseases/isolated lesions(P < 0Æ0001), and there was a positive relationship between the family members’FDLQI scores and patients’ disease severity (r ¼ 0Æ49).Conclusions The FDLQI is simple and practical and seems to have the potential tobe used as an additional outcome measure in clinical practice and evaluationresearch.

Skin disease can have a major impact on patients’ lives1,2 and

a number of patient-completed instruments3–6 have been

described to measure this impact. Quality of life (QoL) of

individuals is closely related to the QoL of those around them

such as partners and in some situations the QoL of a partner

may be more impaired than the patient.7 Family caregivers

may experience a major impact on their lives such as physical

and mental exhaustion, social disruption, marital problems

and financial implications.7 Efforts to improve the QoL of

individuals should therefore also address the QoL of the fam-

ilies or caregivers.8 The concept of Family QoL9 has been

described as being a situation where ‘the family needs are met



and family members enjoy their life together as a family and

have the chance to do things that are important to them’.

Family QoL is considered to be an important outcome of ser-

vices and policies,10 and should be analysed independently

and additionally to the QoL of the patient.11 In some specialit-

ies, for example paediatrics, the QoL of the family is regarded

as a part of the patient’s health.12 However, in dermatology,

the awareness of the importance of this impact of disease on a

patient’s family is relatively recent.

Because the lives of some people with skin disorders are

severely disrupted and because family members are involved

in caregiving, it would be expected that families’ lives would

also be affected.13 The families of children with eczema have

lower QoL than families of healthy children14,15 and taking

care of a child with moderate to severe atopic eczema has

been shown to be more stressful than caring for a child with

insulin-dependent diabetes mellitus.16 Family QoL is related to

the severity of eczema17 and can actually improve following

specialist clinical care.18 The impact of a patient’s skin disease

on family members has been studied in only a small number

of dermatological conditions19–22 with most attention focused

on atopic eczema.

Disease-specific instruments to measure the secondary

impact of skin disease on the family have been described for

atopic eczema.15,23,24 Although disease-specific instruments are

good for tapping specific aspects of a certain disease, they do

not allow generalizability.25 However, a generic instrument is

useful not only for comparisons across different dermatoses

and monitoring interventions, but also in meta-analysis

research. In the same way that dermatology-specific measures

such as the Dermatology Life Quality Index (DLQI)5 or Skin-

dex6 can be used across a wide range of skin diseases, there is

a need for a generic dermatology-specific instrument that could

be used to measure the secondary impact on health-related

QoL (HRQoL) of families across different skin diseases. If it

were possible to quantify this secondary impact of skin disease,

the awareness of the secondary impact would be enhanced,

encouraging a clinical focus on addressing these problems. The

existence of such a tool would allow its use as an appropriate

outcome measure in clinical and health service research.

This paper describes the development and evaluation of an

instrument to quantify the impact of a patient’s skin disease

on family members—the Family Dermatology Life Quality

Index (FDLQI). The aim was to construct and validate a gen-

eric dermatology-specific tool for the family members of

patients, applicable to a wide range of skin diseases, which

should be simple and user-friendly for clinical use and could

serve as an additional outcome measure in evaluation research

or clinical trials.


This open prospective study was approved by the South East

Wales Local Research Ethics Committee and completed in a

number of stages. Written informed consent was given by

participants at all stages.

Study sample

The main study participants were the immediate family mem-

bers accompanying patients attending the dermatology outpa-

tient clinic of the University Hospital of Wales, Cardiff.

Inclusion criteria for the family members were age above

18 years, ability to understand and read the English language,

having a close relationship with the patient and living in the

same household. Family members were excluded if they or

the patients had any other significant illness or disability that

impaired their quality of life, in order to avoid confusion of

that impact with dermatology-related HRQoL.

Interviews and item generation

Detailed semi-structured interviews were conducted with 50

family members or partners of patients with different skin dis-

eases.26 Participants were encouraged to describe all the ways

that their lives had been affected by living with a patient with

skin disease. Enquiries were then made into those aspects of

the family member’s QoL27 that were not mentioned by the

participant but were considered important by the interviewer,

using a checklist of possible HRQoL domains. From the con-

tent analysis of interview transcripts, 59 aspects of family

members’ affected HRQoL were identified. These were categ-

orized into main topic areas by a consensus of the investiga-

tors (Fig. 1). A preliminary draft questionnaire based on the

main topic areas of family members’ HRQoL was formulated

which contained 19 items, each given a four-point response

and score format, i.e. not at all/not relevant ¼ 0, a little ¼ 1,

quite a lot ¼ 2 and very much ¼ 3. All questions asked about

the QoL impact over the last 1 month.

Pilot test

The purpose of the pilot test was to test the content and face

validity of the draft questionnaire and to select appropriate

items for the final questionnaire. A new cohort of 20 subjects

who were immediate family members of patients attending

the dermatology clinic was recruited. Each participant was

given the draft questionnaire to complete and then asked to

comment on the clarity, understandability and relevance of in-

dividual items and also to identify those aspects of family

impact that were not covered in the questionnaire. At the

same time, opinion was sought from 14 dermatology staff

members regarding the practicality and face validity of the

questionnaire.

Finalization of the questionnaire

The final selection of the items was based on the results of the

pilot test and initial statistical analysis of the responses to

assess the internal consistency of the items. Some of the items

were rewritten or revised and some were amalgamated to

enhance relevance and practicality and reduce ambiguity. The

criteria adopted for item removal was that either the specific



The Family Dermatology Life Quality Index, M.K.A. Basra et al. 529

item was endorsed by very few respondents or had low item–

total correlation (< 0Æ20).28 Because our aim was to develop

an instrument that would be both evaluative and potentially

sensitive to change, the items that were scored high by the

majority of the subjects were retained for the final question-

naire.

The present version of the questionnaire (the FDLQI) is

self-administered, contains 10 items (see Appendix) and can

be completed in 2–3 min. Each question inquires about the

family member’s perception of a certain specific impact on

his/her HRQoL over the last 1 month, a time consistent with

the episodic nature of many dermatological conditions, and is

scored on a four-point scale (0–3). The scores of individual

items are summed to generate a total scale score with a range

of 0–30; higher total FDLQI scores indicate greater impair-

ment of the family member’s HRQoL.

Psychometric evaluation of the Family Dermatology Life

Quality Index

The technical properties of the FDLQI were assessed on a new

cohort of family members and patients with various skin con-

ditions. Eligible family members/partners were given a copy

of the FDLQI to complete while the patients (only those older

than 16 years) were requested to complete the DLQI.5,29 The

DLQI is a self-administered questionnaire and consists of 10

questions, each scored on a four-point scale. Scores are added

to yield a total score; higher scores mean greater impairment

of patient’s HRQoL. A global question (GQ) was included at

the end of both the FDLQI and DLQI which asked the

respondent about his/her subjective assessment of the patient’s

disease severity over the last month on a 0–10 visual analogue

scale; 0 indicating ‘clear’ skin and 10 ‘worst possible’. The

medical record of the patient was used to ascertain the clinical

diagnosis. Enrolled participants were also given an envelope

containing the FDLQI with the GQ to self-assess disease sever-

ity, to complete after 1 week and return by post, but only if

they thought that the patient’s skin condition had remained

stable during this period. A subset of subjects was approached

3–6 months later at one of their routine follow-up visits by

one of the investigators (M.K.A.B.) and asked whether the

patients’ skin condition had changed (improved or deterior-

ated) since the last administration of the questionnaire. If it

had, the family members/partners were given the FDLQI with

the GQ about the patient’s disease severity to complete again.

At each stage questionnaires with more than one missing item

were excluded from analysis.

Factor analysis

For an instrument to be able to use a summed total score, it is

necessary to demonstrate that the instrument shows either

strict or essential unidimensionality.30 Strict unidimensionality

indicates the presence of a single common factor whereas

essential unidimensionality shows the presence of a reasonably

dominant common factor along with other secondary minor

factors. Exploratory factor analysis31 was performed on the 10

items to identify the factor structure underlying the FDLQI

items and to determine whether expected essential unidimen-

sionality was present to support the use of a total FDLQI score.

Because we expected that the factors would be interrelated,

we used the oblique rotation produced by the Oblimin

method followed by Kaiser normalization.32

Reliability

The FDLQI was evaluated in terms of internal consistency

and test–retest reliability. Internal consistency was assessed by

means of Cronbach’s alpha (a) coefficient,33 which expresses

the degree of consistency of the items. It has been suggested

that a should be above 0Æ7034 and not more than 0Æ90.28

a coefficient values below 0Æ70 are regarded as indicative of

0 10 20 30 40 50

Psychological distress

Burden of care

Effect on social life

Effect on holidays

Effect on work load

Effect on job

Effect on routine expenditure

Effect on leisure activities

Effect on physical well-being

Effect on sleep

Need to avoid favourite activities

Dissatisfaction with health services

Need for support

Effect on nutrition

Problems due to peoples' attitude

Effect on sex life

Role of religious faith

Miscellaneous

Number of family members (out of 50) affected by each QoL aspect

Fig 1. Main family health-related quality of

life (HRQoL) categories and their frequency

among 50 families.



530 The Family Dermatology Life Quality Index, M.K.A. Basra et al.

individual items not adequately contributing to overall scale

score,28 while a very high a coefficient shows that items are

measuring a common latent trait with a high level of item

redundancy. Reproducibility (test–retest reliability) examines

the stability of the scale on repeat measurement. It was

determined by means of the intraclass correlation coefficient

(ICC) using a one-way random effects parallel model.35 To

ensure that only stable individuals were included for test–

retest, only those questionnaires (returned within 1–2 weeks)

were analysed in which the GQ rating about the disease

severity had not changed by more than one point on the 0–

10 scale.

Responsiveness to change

A subset of participants (family members) who confirmed that

patients’ skin condition had changed completed the FDLQI

(and GQ) 3–6 months later. To assess the sensitivity of the

FDLQI to change, the difference in the scores on the two

occasions was calculated as well as the difference in the dis-

ease severity GQ score. Based on the difference in the GQ

(disease severity) score, subjects were divided into two

groups: ‘improved’ (in whom the GQ score decreased) and

‘worsened’ (in whom the GQ score increased). The paired

sample t-test was used to assess the within-group difference

and the Mann–Whitney U-test was used to interpret the differ-

ence in FDLQI scores between improved and worsened

groups. The Spearman’s rank correlation coefficient was used

to determine the correlation between the change in FDLQI

score and the change in the disease severity score.

Validity

We assessed the construct validity of the instrument by testing

a number of a priori hypotheses.

Firstly, we hypothesized that there would be a moderate to

high correlation between the HRQoL of the patient and the

family. To assess this we compared the FDLQI scores of the

family members with the DLQI scores of the patients (only

those ‡ 16 years of age) using Spearman’s rank correlation;

here we expected to find strongly positive correlation.

Secondly, we hypothesized that family members of patients

with inflammatory skin diseases would have greater impair-

ment of their HRQoL when compared with the family mem-

bers of patients with noninflammatory diseases or isolated

lesions. To test this we divided the participants into two dis-

ease category groups: inflammatory (family members of

patients suffering from inflammatory skin diseases, e.g. acne,

eczema, psoriasis) and noninflammatory [family members

with patients having noninflammatory diseases or isolated

benign, premalignant or malignant lesions, e.g. naevi, viral

warts, basal cell carcinoma (BCC), solar keratoses]. Then we

compared the FDLQI scores between the two disease categories

using the Mann–Whitney U-test. Within-group analysis was

also carried out to examine the relationship between different

diseases and the family impact (e.g. psoriasis, eczema, acne).

Thirdly, we hypothesized that family impact would be rela-

ted to the patient’s disease severity.36 Here we tested the cor-

relation between FDLQI scores and the patient’s disease

severity score (GQ), as rated by the family members using

Spearman’s correlation coefficient.

All the statistical analysis was carried out on SPSS version

12.0 computer software (SPSS Inc., Chicago, IL, U.S.A.).

Results

Development of the questionnaire

Of the 51 family members recruited for interviews, one was

later found to have a significant heart problem and was exclu-

ded. The demographic characteristics of the remaining 50 par-

ticipants are shown in Table 1. The mean age of the family

members was 48Æ1 (range 24–82) years; 19 were male and

31 were female. Most of them were either one of the parents

(44%) or spouses/partners (44%) of the patients. The patients

(mean age ¼ 35Æ7 years; range 5 months–84 years) suffered

from one of the 21 skin diseases shown in Table 1. Interview

analysis identified 59 different aspects of family members’ life

quality which were adversely affected by the patients’ skin dis-

eases. These were grouped into main topic areas (Fig. 1). A

detailed description of these initial interviews is presented

elsewhere.37

The most frequently described aspects were psychological

distress (98%), burden of care (54%), social life (48%), holi-

days (46%) and housework (42%). The two least frequently

mentioned aspects of life were the effect on sex life and the

Table 1 Sociodemographic characteristics of subjects (n ¼ 50) in theinterview stage

Characteristics % Diagnoses %

Sex Eczemas 28Male 38 Psoriasis 14

Female 62 Discoid lupuserythematosus

8

Acne 6Marital status Ichthyosis 4

Married 66 Granuloma annulare 4Single 24 Squamous cell carcinoma 4

Divorced/widowed 10 Basal cell carcinoma 4Malignant melanoma 4

Relationship with the patient Keratoacanthoma 2Parent 44 Solar keratosis 2

Spouse/partner 44 Alopecia areata 2Daughter 4 Angioedema 2

Grandparent 4 Pemphigus 2

Grandchildren 2 Vitiligo 2Sibling 2 Haemangioma 2

Lichen sclerosis 2Ethnic origin Urticaria 2

Caucasian 88 Incontinentia pigmenti 2Asian 8 Mycosis fungoides 2

Afro-Caribbean 4 Dental sinus 2




role of religious faith in coping with the patients’ illness

(included here due to its major significance in some cultures

and mentioned by some subjects during the interviews); each

of these were described by only 8% of the participants. There

was no statistically significant difference between the male and

female participants in the frequency of any of the HRQoL

aspects.

The present 10-item version of the FDLQI questionnaire was

finalized after meticulous reviewing of the feedback from the

family members of the patients and dermatology experts dur-

ing the pilot test and initial psychometric testing of the 19

items. Items were removed if either the item–total correlation

was < 0Æ2028 or if it was endorsed by very few respondents.28

Evaluation of the questionnaire

Of the 141 family members/partners approached, five

declined to participate for personal reasons. Two more were

later dropped from analysis after finding from their medical

records that they had other significant health issues not men-

tioned at the time of recruitment. Responses of two subjects

were eliminated due to incomplete answers. The final psycho-

metric analysis was carried out on data from 132 family mem-

bers (of patients with 45 different skin conditions) who

completed the FDLQI and 109 of their related patients (only

those ‡ 16 years of age) who completed the DLQI. The socio-

demographic characteristics of study participants are shown in

Table 2 and the dermatological characteristics in Table 3.

There was no statistically significant difference in the age of

male and female participants [family members (P ¼ 0Æ10),

patients (P ¼ 0Æ24)].

Family Dermatology Life Quality Index score distribution

Scores for the overall scale (0–30) ranged from 0 to 26, med-

ian ¼ 7Æ0, mean ¼ 8Æ02 and SD ¼ 6Æ36, indicating the ability

of the instrument to detect subject variability on a continuum

of QoL impact (Table 4). There were no ceiling effects (i.e.

maximum score of 30 not reached) and a minimum of floor

effects with only 7Æ6% subjects scoring 0.

Using the mean score for individual items (range 0–3), the

most highly scoring items were the emotional impact (1Æ39),

burden of care (0Æ95), effect on the physical well-being of the

family member (0Æ86), extra household expenditure (0Æ82)

and problems due to people’s reaction to the patients’ skin

appearance (0Æ80) (Fig. 2). Using the percentage (out of 132)

responding positively (combining a little, quite a lot and very

much), the most commonly reported QoL aspects were emo-

tional aspect (79Æ5%), burden of care (60%), impact on phys-

ical well-being (53Æ1%) and extra household expenditure

(53%). Although for each of the 10 items the female mean

score was marginally higher than the male mean score, there

was neither a statistically significant difference in the total

FDLQI scores (M ¼ 7Æ24, F ¼ 8Æ43, P ¼ 0Æ31) nor in the

individual item scores (P ¼ 0Æ08–0Æ91) between male and

female subjects. However, there was a significant difference in

the FDLQI scores (P < 0Æ001) of family members of patients

12 years of age or younger (mean FDLQI score ¼ 12Æ4,

SD ¼ 6Æ6, n ¼ 16) compared with those with family members

older than 12 years (mean FDLQI score ¼ 7Æ4, SD ¼ 6Æ1,

n ¼ 116).

Caregivers can estimate the severity of patient’s disease ac-

curately.36,38 The family members’ assessment of the patients’

disease severity (GQ) score ranged from 0 to 10, with a mean

of 5Æ1 (SD ¼ 2Æ8). The patients’ own assessment of their dis-

ease severity (GQ) mean score was 5Æ3 (SD ¼ 2Æ8). There was

a high correlation between the family members’ and patients’

subjective assessment of the patients’ disease severity (r ¼0Æ69, Pearson’s rank correlation coefficient, P ¼ 0Æ01).

Factor analysis

Factor analysis of the data (from 132 participants) was per-

formed using principal component analysis followed by obli-

que rotation with Kaiser normalization.32 Based on the

guidelines suggested by Gorsuch,31 i.e. 5–10 cases for each

item, the sample size of our study (n ¼ 132 for a scale of 10

items) was considered adequate for factor analysis. Two

Table 2 Sociodemographic characteristics of participants in theevaluation stage

Total number of family members (n) 132Males 45

Females 87Mean age of family members (years) 49Æ2Relation with patient (%)

Parents 44Æ7Spouses/partners 45Æ5Son/daughter 9Æ1

Marital status (%)

Married 74Æ2Single 22Æ7Divorced/widowed/separated 3Æ1

Occupation (%)

Employed 50Retired 20Æ5Housewife 22Student 3Æ8Unemployed 3Æ8

Residence (%)

Urban 93Æ9Rural 6Æ1

Educational status (%)Primary 7Æ6Secondary 56Æ1Diploma/degree 34Æ8

Ethnic origin (%)Caucasian 90Æ1Asian 6Æ1Afro-Caribbean 3Æ8

Total number of patients (n) 132Males 58

Females 74Mean age of patients (years) 39Æ5




main factors were identified by retaining those factors with

eigenvalues > 1; the first factor had an eigenvalue of 4Æ78 and

accounted for about 48% of the common variance, while the

second smaller factor had an eigenvalue of 1Æ18 and accounted

for about 12% of the variance (Table 5). The two factors

together accounted for 60% of the common variance, which

is higher than the minimum recommended (i.e. 50%) for a

stable factor solution.39 The two extracted factors were signifi-

cantly interrelated (r ¼ 0Æ48), confirming that our approach

of using oblique rotation31 was appropriate. Reckase40 has

suggested that unidimensionality can be assessed through the

eigenvalues of factors. A scale is considered to exhibit unidi-

mensionality if there is one dominant factor. Another method

to conclude unidimensionality is to calculate the ratio of the

first and second factor eigenvalues. If the ratio is higher than a

critical value, such as four, as suggested by some psychometri-

cians,31,41 the scale is unidimensional. In our factor analysis,

not only was there one dominant first factor but also the ratio

between the first (4Æ78) and second (1Æ18) factor eigenvalues

was greater than the minimum recommended ratio of four for

supporting unidimensionality. Figure 3 shows the scree plot

that demonstrates the eigenvalues against each factor. As can

be seen, from the first factor onwards the line is almost flat

meaning that it is only the first factor that accounts for most

Table 3 Dermatological characteristics of patients in the validationstage

Number of patients; n 132Total number of dermatoses; n 45

Group 1 (inflammatory); n (%) 24 (53Æ3)a

Group 2 (noninflammatory),

isolated lesions—benign,premalignant and malignant);

n (%)

21 (46Æ6)a

Mean duration of patients’ diseases 7Æ7 years

(range 2 weeks–55 years)

Group 1 %b Group 2 %b

Eczemas 15Æ9 Naevi 5Æ3Psoriasis 15Æ2 Solar keratosis 5Æ3Acne 15Æ2 Basal cell carcinoma 3Æ8Leg ulcer 2Æ3 Vitiligo 1Æ5Urticaria 1Æ5 Viral wart 1Æ5Pemphigoid 3Æ0 Haemangioma 1Æ5Contact dermatitis 1Æ5 Morphoea 1Æ5Fungal infection 1Æ5 Bowen disease 1Æ5Erythema nodosum 1Æ5 Squamous cell carcinoma 0Æ8Othersd 11Æ2 Otherse 9Æ6Total 68Æ8c Total 32Æ3c

aPercentage of skin diseases. bPercentage of participants

(family members of patients with these skin diseases).cBecause of rounding of figures, percentages sometimes do

not add up to 100. dNodular prurigo, folliculitis, pompho-lyx, ichthyosis, erythema multiforme, discoid lupus

erythematosus, rosacea, pityriasis rosea, lichen planus,

seborrhoeic dermatitis, lichen sclerosis, acne inversa,chronic actinic dermatitis, epidermolysis bullosa. eAlopecia

areata, milia, haematoma, hyperhidrosis, onychomycosis,epidermoid cyst, pruritus, mycosis fungoides, porokerato-

sis, Zoon’s balanitis, pityriasis lichenoides chronica.

Table 4 Family Dermatology Life Quality Index (FDLQI) scoredistribution in various family member groups and main dermatoses

n Mean score SD Range

Overall FDLQI score 132 8Æ02 6Æ4 0–26FDLQI—inflammatory 90 10Æ0 6Æ2 1–26

FDLQI—noninflammatory 42 3Æ8 4Æ3 0–17Male family members 45 7Æ2 5Æ9 0–23

Female family members 87 8Æ4 6Æ6 0–26Eczema 20 13Æ0 5Æ9 3–25

Acne 20 8Æ9 5Æ6 1–25Psoriasis 20 6Æ7 3Æ0 3–16

Naevus 7 3Æ1 3Æ1 0–8Basal cell carcinoma 5 1Æ6 1Æ7 0–4

Solar keratosis 7 1Æ1 0Æ9 0–2

0

0·2

0·4

0·6

0·8

1

1·2

1·4

1·6

1·8

Emot

ional

Physic

al well

-bein

g

Relatio

nship

s

People

reac

tion

Social

life

Leisu

re a

ctivit

ies

Burde

n of

care

House

wor

k

Job/

study

Expen

ditur

e

FDLQI Item

Mea

n It

em S

core

(m

axim

um

= 3

) Overall

Inflammatory

Noninflammatory

Fig 2. Mean Family Dermatology Life Quality Index (FDLQI) scores

for individual items; overall and for inflammatory and

noninflammatory groups (shown in order as presented in the

questionnaire).

Table 5 Factor analysis: components and their eigenvalues with % ofvariance explained by each component

Component Eigenvalues % of variance

1 4Æ780 47Æ8012 1Æ180 11Æ797

3 0Æ864 8Æ6404 0Æ768 7Æ677

5 0Æ623 6Æ2266 0Æ488 4Æ878

7 0Æ435 4Æ3468 0Æ355 3Æ554

9 0Æ290 2Æ90410 0Æ218 2Æ178




of the variance and each successive factor accounts for smaller

and smaller amounts of the total variance. It has also been

suggested40 that unidimensionality may hold if at least 20% of

the variance is accounted for by the first factor. The variance

explained by the first factor in our principal component analy-

sis, which was about 48%, again supports the assumption of a

dominant single underlying dimension. These results support

the justification for using a single total FDLQI score. Items

were then identified that loaded on each of the two compon-

ents with standard regression coefficients of more than 0Æ50

(Table 6). Factors were labelled as Psychosocial impact, which

loaded six items (emotional impact, physical well-being,

impact on relationships, peoples’ reaction, social life and leis-

ure activities) and Physical impact, which loaded four items

(burden of care, additional housework, effect on job/study

and extra household expenditure). The assignment of a par-

ticular item to a component was based on higher loading

value. The loading of ‘physical well-being’ on ‘psychosocial

impact’ may be explained by the concept of ‘burn-out’,42

which describes the fact that prolonged exposure to stress,

and to situations where demands exceed individual’s adaptive

resources, is related to physical fatigue, emotional exhaustion

and cognitive weariness. The loading of ‘additional household

expenditure’ on ‘physical impact’ may relate to the fact that

cheaper choices often require more effort—physically, emo-

tionally and cognitively.

Reliability

Internal consistency of the instrument, measured by Cron-

bach’s a coefficient (in 132 subjects) was 0Æ88, indicating

adequate internal consistency. This was not improved by

deleting individual items (0Æ86–0Æ87), demonstrating that all

the items of the scale contribute to the total score. The value

of a was more than the minimum of 0Æ70 required for a

panel of items such as the FDLQI to be considered a scale,

showing further that the individual items were contributing

adequately to the overall scale score.34 All item-to-total score

correlations were significant at P < 0Æ001.

Of the 64 (47Æ7%) respondents who returned the second

set of completed questionnaires 1–2 weeks after the first

administration, 13 were eliminated because the patients’ dis-

ease severity score self-assessed by the family member had

changed by more than one point from the initial score. One

more questionnaire was excluded due to incomplete

responses. Final reproducibility (test–retest reliability) assess-

ment was based on data from 50 respondents (37Æ8%) in

whom the clinical severity of patients’ disease had not

changed (assessed by maximum change in GQ score of one

point). Test–retest reliability of the scale assessed by ICC, was

0Æ94, reflecting the scale’s ability to show reproducible results

in stable subjects. For individual items the value of ICC ranged

between 0Æ70 and 0Æ87 indicating substantial agreement

between the two scores.43

Responsiveness to change

Responsiveness of the scale to change was assessed on 27 fam-

ily members 3–6 months after initial FDLQI administration at

the routine follow-up visit of the patient. Only those subjects

who reported that the patient’s skin condition was improved

or worsened compared with the first administration were

given the questionnaire to complete. Of these responses, only

data from subjects in whom the subjects’ self-assessed disease

severity (GQ score) was changed by two or more points were

analysed (n ¼ 27).

In those subjects who confirmed that the patients’ condition

had improved (n ¼ 17), the mean FDLQI score decreased

from 11Æ4 to 4Æ8 (mean change ¼ 6Æ6, SD ¼ 3Æ3, Fig. 4a),

which was highly significant (P < 0Æ0001); the corresponding

decrease in global rating of disease severity (GQ score) was

from 6Æ9 to 3Æ9 (mean change ¼ 3Æ0, SD ¼ 1Æ3). In subjects

who reported that the patients’ skin condition had worsened

Component number

10987654321

Eig

enva

lue

5

4

3

2

1

0

Fig 3. Scree plot showing eigenvalue against each factor; from the

first factor onwards, the line is almost flat showing each successive

factor accounting for smaller and smaller amounts of the total

variance; 48% of the common variance is explained by the first factor.

Table 6 Factor analysis: structure matrix after rotation showing the

loadings of different items on two components

FDLQI item 1 2

Emotional 0Æ797 0Æ312

Physical well-being 0Æ708 0Æ631Relationships 0Æ775 0Æ315

People’s reaction 0Æ753 0Æ322Social life 0Æ772 0Æ408

Leisure activities 0Æ773 0Æ419Burden of care 0Æ367 0Æ824

Extra housework 0Æ340 0Æ861Job/study 0Æ519 0Æ578

Extra expenditure 0Æ599 0Æ608

FDLQI, Family Dermatology Life Quality Index.




(n ¼ 10), the mean FDLQI score increased from 8Æ1 to 15Æ6(mean change ¼ 7Æ5, SD ¼ 6Æ5, P < 0Æ01, Fig. 4b) and the

corresponding change in the global disease severity rating was

an increase from 4Æ9 to 7Æ7 (mean change ¼ 2Æ8, SD ¼ 2Æ1).

The difference in the FDLQI score change between the

improved and worsened subjects grouped together was also

significant (P < 0Æ001). There was a high correlation between

the magnitude of change in the FDLQI score and the change

in GQ score (r ¼ 0Æ74, Spearman’s rank correlation coeffi-

cient, at significance level of 0Æ01).

Validity

Construct validity of the FDLQI was assessed by testing the a

priori hypotheses. There was strong correlation between the

families’ FDLQI scores and patients’ DLQI scores (r ¼ 0Æ69,

Spearman’s rank correlation, P ¼ 0Æ01) assessed in 109 paired

subjects and patients (‡ 16 years of age) (Fig. 5). The overall

mean DLQI score was 8Æ12, SD ¼ 7Æ3, range ¼ 0–29, n ¼109. This reflects that the overall level of impairment of the

HRQoL of the family member is related to that of the patient.

The mean FDLQI score for the inflammatory group

(mean ¼ 10Æ0, n ¼ 90) was much higher than for the non-

inflammatory/isolated lesion group (mean ¼ 3Æ8, n ¼ 42,

P < 0Æ0001) using the Mann–Whitney U-test. The difference

was also significant across all the individual FDLQI items

(Fig. 2) except for peoples’ reactions (P ¼ 0Æ1). This result

suggests that the magnitude and nature of the family impact

differed between the two clinical groups and most of the

items of the FDLQI were able to discriminate between the two

clinical groups. Within-group analysis for two disease categor-

ies showed that the FDLQI scores were able to discriminate

between different dermatoses (Table 4). The family impact

within the inflammatory group was seen to differ (P ¼ 0Æ001)

between psoriasis (mean ¼ 6Æ7, SD ¼ 2Æ9, n ¼ 20), eczema

(mean ¼ 13Æ1, SD ¼ 5Æ9, n ¼ 20) and acne (mean ¼ 8Æ9,

SD ¼ 5Æ6, n ¼ 20). Similarly for the noninflammatory/isol-

ated lesions group, the FDLQI scores were clearly low as

expected, e.g. benign naevi (mean ¼ 3Æ1, SD ¼ 3Æ1, n ¼ 7),

BCC (mean ¼ 1Æ6, SD ¼ 1Æ7, n ¼ 5) and solar keratosis

(mean ¼ 1Æ1, SD ¼ 0Æ9, n ¼ 7).

The FDLQI scores and disease severity of the patient (GQ

score), as assessed by the family member, were moderately

correlated (r ¼ 0Æ49, P ¼ 0Æ01, n ¼ 132, Spearman’s rank

correlation) in the postulated direction confirming the hypoth-

esis that family impact of skin disease is positively associated

with the disease severity of the patient.

Discussion

The impact of many skin diseases is not limited to the patient

but may extend to the rest of the family.44 Family members

play a central role in the care of such patients, especially those

with inflammatory skin disease, and so family impact data are

potentially important components of the measurement of the

overall burden of skin disease. In dermatology, measurement

of this ‘secondary impact’ of skin diseases on the patients’ fam-

ily members has largely been ignored except for atopic derma-

titis. Families may not get the level of psychosocial support that

they need, further affecting medical and economic outcomes.

11·35

4·76

6·88

3·88

0

2

4

6

8

10

12

1 2Time point

Sco

re

FDLQI score

Disease severity(GQ score)

8·1

15·6

4·9

7·7

0

2

4

6

8

10

12

14

16

18

1 2Time point

Sco

re

FDLQI score

Disease severity(GQ score)

(a)

(b)

Fig 4. Sensitivity of Family Dermatology Life Quality Index (FDLQI)

to change. (a) Sensitivity to improvement in patients’ clinical

condition: both FDLQI and disease severity (global question, GQ)

scores have decreased on repeat measurement (1 ¼ first

administration, 2 ¼ second administration after 3–6 months).

(b) Sensitivity to worsening in patients’ clinical condition: both

FDLQI and disease severity (GQ) scores have increased on repeat

measurement (1 ¼ first administration, 2 ¼ second administration

after 3–6 months).

FDLQI (n = 109) DLQI (n = 109)

0

5

10

15

20

25

30

Sco

re

Fig 5. Score distribution for Family Dermatology Life Quality Index

(FDLQI) (completed by the family members) and Dermatology Life

Quality Index (DLQI) (completed by the patients) with median,

quartiles, extreme values and outliers.




This paper describes the development and initial validation

of an instrument which is specific for the family members/

partners of patients with a wide range of skin diseases. The

evaluation of the FDLQI was conducted to meet the basic cri-

teria28,45 necessary for an instrument to be scientifically valid

and hence used with confidence.

The FDLQI measures family impact over the previous

1 month, a duration consistent with the episodic nature of

many skin diseases3 and a reasonable period to attain stability

after a change in therapy.6 The results of exploratory factor

analysis support the assumption of unidimensionality and

hence justify summing individual scale items to yield an over-

all scale score. The score ranged from 0 to 26, reflecting the

measure’s sensitivity to variation in the family impact. To be

sensitive to absence or presence of QoL impairment, QoL in-

strument scale scores should not have large floor or ceiling

effects.46 A scale having a ceiling effect of more than 20%

suggests that items with a higher range of impact measure-

ment are needed to differentiate moderate cases from severe.

In the same way, a floor effect of more than 20% is consid-

ered to be insensitive to important lower levels of impact on

QoL.46 In this context, the FDLQI, having a minimal floor

effect (7Æ6%) and no ceiling effect, would be expected to per-

form well in studies measuring change. The FDLQI demon-

strates content validity on the grounds that its item content

was based on the most frequently raised issues47 described by

the patients’ families, confirmed during the pilot testing as

well as by the consensus of experts in dermatology.

The reliability of the scale was clearly demonstrated; intern-

al consistency and test–retest reliability were reasonably ade-

quate with Cronbach’s a of 0Æ88 and ICC of 0Æ94. When an

instrument is required for monitoring individuals or in clinical

trials, an internal consistency reliability of Cronbach’s a in

excess of 0Æ85 is needed.48 A high ICC value indicates that the

scale was reproducible in stable subjects. Although our sample

for responsiveness was not large, we were able to show sensi-

tivity of the scale to change in subjects’ HRQoL with respect

to change in clinical severity of patients’ skin disease. The

FDLQI has demonstrated evidence of validity based on success-

fully testing the a priori hypotheses about its construct. We

were able to demonstrate that a family’s HRQoL was not only

related to the patient’s HRQoL, but it was also related to the

clinical severity of the patient’s disease in the expected direc-

tion and as shown by others.17 In fact, one study of the

family impact of atopic eczema36 demonstrated that the

caregiver’s assessment of severity of the child’s eczema was

the strongest predictor of the secondary family impact. The

known group validity of the FDLQI was also demonstrated by

its ability to differentiate between groups on the basis of the

nature of skin disease as predicted, i.e. the inflammatory skin

disease group having significantly higher score than the non-

inflammatory/isolated lesions group. One of our findings was

that eczema causes more impairment of family’s HRQoL than

psoriasis (P < 0Æ0001); this complements the previous finding

that HRQoL of eczema patients is more affected than that

of psoriasis patients.5,29 This possibly reflects the relatively

younger and more family-dependent age of onset in eczema

patients as well as the more itchy nature of eczema than

psoriasis.

There are several limitations of our study. There was a

relatively small sample size and recruitment of convenience

samples, which implies that scores and prevalence estimates

could be applied only to those who participated in the study.

Further studies in different settings and different populations

should be carried out to confirm our findings on family

impact. Although the preliminary analysis has provided some

evidence of the probable unidimensionality of the instru-

ment, we feel more work, as suggested by Mazzotti et al.,49

is required to confirm this issue. Further evidence needs to

be collected of the responsiveness of the instrument to clin-

ical change, for example before and after hospital admission

or other major intervention. There will also be a need to

define the meaning of FDLQI scores50 as well as the minimal

important difference for overall FDLQI score. The establish-

ment of validation is an ongoing process and there is no

gold standard against which QoL instruments can be valid-

ated.51 Measuring other aspects of validity, such as conver-

gent and divergent validity, by comparing FDLQI scores with

other generic family-oriented QoL scales would increase our

confidence in the use of the instrument and further streng-

then its potential usefulness in clinical practice and observa-

tional research.

One motivation for this study has been a wish to heighten

awareness of a very major consequence of skin disease, which

up to now has been largely ignored. If a concept cannot be

measured it attracts little scientific attention; by proposing for

the first time a method applicable to all skin diseases to score

the secondary impact of skin disease on life quality of family

members, it is hoped that attention will be focused on this

important aspect of clinical dermatology.

The ability to identify and score this secondary impact may

encourage the development of new strategies to alleviate this

impact and will allow the assessment of such strategies. The

FDLQI could potentially be used as an additional outcome

measure for health service research and clinical trials. A speci-

ality-specific measure, such as the FDLQI, may complement

generic health status measures in addressing issues concerning

health services funding and the distribution of resources

between specialities.52

There was very high acceptance of and interest in the

FDLQI by respondents. Many commented on the great impact

caused to them by their family members’ skin disease and

often mentioned that this burden had never before been

acknowledged by health service providers. This very strong

feedback has encouraged us to consider that these concepts do

need to be understood, measured and acted upon.

In conclusion, the FDLQI was developed to meet the need

for a comprehensive yet simple and practical instrument for

general clinical use and as an additional outcome measure in

dermatology for research and clinical trials. It is easy to com-

plete and score and seems reliable, valid and responsive to

improvement or deterioration in HRQoL.




Acknowledgments

We wish to thank the dermatology patients, their family

members and dermatology staff for their help with this study.

In particular we wish to thank Sister Anne Thomas for advice

on ethical issues and Mrs Susan Williams for her help.

References

1 Jowett S, Ryan TJ. Skin disease and handicap: an analysis of the

impact of skin conditions. Soc Sci Med 1985; 20:425–9.2 Finlay AY, Ryan TJ. Disability and handicap in dermatology. Int J

Dermatol 1996; 35:305–11.3 Anderson RT, Rajagopalan R. Development and validation of a

quality of life instrument for cutaneous diseases. J Am Acad Dermatol1997; 37:41–51.

4 Morgan M, McCreedy R, Simpson J, Hay R. Dermatology qualityof life scales—a measure of the impact of skin diseases. Br J Dermatol

1997; 136:202–6.5 Finlay AY, Khan GK. Dermatology Life Quality Index (DLQI)—a

simple practical measure for routine clinical use. Clin Exp Dermatol1994; 19:210–6.

6 Chren MM, Lase RJ, Quin LM et al. Skindex, a quality-of-life meas-ure for patients with skin disease: reliability, validity, and respon-

siveness. J Invest Dermatol 1996; 107:707–13.7 Rees J, O’Boyle C, MacDonagh R. Quality of life: impact of chronic

illness on the partners. J R Soc Med 2001; 94:563–6.8 Park J, Hoffman L, Marquis J et al. Toward assessing family out-

comes of service delivery: validation of a family quality of life

survey. J Intellect Disabil Res 2003; 47:367–84.9 Turnbull AP, Turnbull HR, Poston D et al. Enhancing quality of life

of families of children and youth with disabilities in the UnitedStates. In: Families and People with Mental Retardation and Quality of Life:

International Perspectives (Turnbull AP, Brown I, Turnbull HR, eds).Washington, DC: American Association on Mental Retardation,

2004; 51–100.10 Bailey DB, McWilliam RA, Darkes LA et al. Family outcomes in

early intervention: a framework for program evaluation and effi-cacy research. Exceptional Children 1998; 64:313–28.

11 Verdugo MA, Cordoba L, Gomez J. Spanish adaptation and valid-ation of the Family Quality of Life Survey. J Intellect Disabil Res 2005;

49:794–8.12 Fink R. Issues and problems in measuring children’s health

status in community health research. Soc Sci Med 1989; 29:715–9.

13 Elliot BE, Luker K. The experiences of mothers caring for a childwith severe atopic eczema. J Clin Nurs 1997; 6:241–7.

14 Lapidus CS, Kerr PE. Social impact of atopic dermatitis. Med Health R2001; 84:294–5.

15 Lawson V, Lewis-Jones MS, Finlay AY et al. The family impact ofchildhood atopic dermatitis: the Dermatitis Family Impact Ques-

tionnaire. Br J Dermatol 1998; 138:107–13.16 Su JC, Kemp AS, Varigos GA et al. Atopic eczema: its impact on the

family and financial cost. Arch Dis Child 1997; 76:159–62.17 Ben-Gashir MA, Seed PT, Hay RJ. Are quality of family life and dis-

ease severity related in childhood atopic dermatitis? J Eur Acad Der-matol Venereol 2002; 16:455–62.

18 Balkrishnan R, Manuel J, Clarke J et al. Effects of an episode of spe-cialist care on the impact of childhood atopic dermatitis on the

child’s family. J Pediatr Health Care 2003; 17:184–9.19 Carroll CL, Balkrishnan R, Feldman SR et al. The burden of atopic

dermatitis: impact on the patient, family and society. Pediatr Dermatol

2005; 22:192–9.

20 Richards HL, Chong SLP, Mason DL, Griffiths CEM. The impact ofpsoriasis on healthy partners of patients with psoriasis. Br J Dermatol

2002; 147 (Suppl. 62):54.21 Miller AC, Pit-Ten Cate IM, Watson HS, Geronemus RG. Stress and

family satisfaction in parents of children with facial port-winestains. Pediatr Dermatol 1999; 16:190–7.

22 Braue A, Ross G, Varigos G, Kelly H. Epidemiology and impact ofchildhood Molluscum contagiosum: a case series and clinical review of

the literature. Pediatr Dermatol 2005; 22:287–94.23 Chamlin SL, Cella D, Frieden IJ et al. Development of the Childhood

Atopic Dermatitis Impact Scale: initial validation of a quality-of-life

measure for young children with atopic dermatitis and their fam-ilies. J Invest Dermatol 2005; 125:1106–11.

24 McKenna SP, Whally D, Dewar AL et al. International developmentof the Parents’ Index of Quality of Life in Atopic Dermatitis

(PIQoL-AD). Qual Life Res 2005; 14:231–41.25 Aaronson NK. Quantitative issues in health-related quality of life

assessment. Health Policy 1988; 10:217–30.26 Guyatt G, Bombardier C, Tugwell PX. Measuring disease-specific

quality of life in clinical trials. CMAJ 1986; 134:889–95.27 Poston D, Turnbull AP, Park J et al. Family Quality of Life: a qualit-

ative inquiry. Ment Retard 2003; 41:313–28.28 Streiner DL, Norman GR. Health Measurement Scales: A Practical Guide to their

Development and Use, 3rd edn. Oxford: Oxford University Press, 2003.29 Lewis V, Finlay AY. 10 years experience of the Dermatology Life

Quality Index (DLQI). J Invest Permatol Symp Proc 2004; 9:169–80.30 Nandakumar R, Ackerman T. Test modelling. In: The SAGE Handbook

of Quantitative Methodology for the Social Sciences (Kaplan D, ed). ThousandOaks, CA: Sage Publications, 2004; 93–105.

31 Gorsuch RL. Factor Analysis, 2nd edn. Hillsdale, NJ: LawrenceErlbaum, 1983.

32 Fabrigar LR, Wegener DT, MacCallum RC, Strahan EJ. Evaluatingthe use of exploratory factor analysis in psychological research. Psy-

chological Methods 1999; 4:272–99.33 Cronbach LJ. Coefficient alpha and the internal structure of tests.

Psychometrika 1951; 16:297–334.34 Nunnaly JC Jr. Psychometric Theory, 2nd edn. New York: McGraw-

Hill, 1978.35 Shrout PE, Fleiss JL. Intraclass correlations: uses in assessing rater

reliability. Psychol Bull 1979; 86:420–8.36 Balkrishnan R, Houseman TS, Caroll C et al. Disease severity and

associated family impact in childhood atopic dermatitis. Arch DisChild 2003; 88:423–7.

37 Basara MKA, Finlay AY. The family impact of skin diseases: the

greater patient concept. Br J Dermatol 2007; in press.38 Housman TS, Patel M, Camacho F et al. Use of a self-administered

eczema severity index by parent caregivers: results of a validationstudy. Br J Dermatol 2002; 147:1192–8.

39 Streiner DL. Figuring out factors: the use and misuse of factoranalysis. Can J Psychol 1994; 39:135–40.

40 Reckase M. Unifactor latent trait models applied to multifactortests: results and implications. J Educ Stats 1979; 4:207–30.

41 Hattie J. Methodology review: assessing unidimensionality of testsand items. Applied Psychological Measurement 1984; 20:1–14.

42 Shirom A, Melamed S, Tocker S et al. Burnout, mental and physicalhealth: a review of the evidence and a proposed explanatory

model. Int Rev Ind Organ Psychol 2005; 20:269–309.43 Donner A, Eliasziw M. Sample size requirements for reliability

studies. Stats Med 1987; 6:441–8.44 Finlay AY. Skin disease disability: measuring its magnitude. Keio J

Med 1998; 47:131–4.45 Guyatt G, Feeny DH, Patrick DK. Measuring health related quality

of life. Ann Intern Med 1993; 118:622–9.




46 Bullinger M, Anderson R, Cell AD. Developing and evaluatingcross-cultural instruments: from minimal requirements to optimal

models. Qual Life Res 1993; 2:451–9.47 Kaplan RM, Bush JW, Berry CC. Health status: types of validity and

the index of well-being. Health Serv Res 1976; 11:478–507.48 Weiner EA, Stewart BJ. Assessing Individuals. Boston: Little Brown,

1984.49 Mazzotti E, Barbaranelli C, Picardi A et al. Psychometric properties

of the Dermatology Life Quality Index (DLQI) in 900 Italianpatients with psoriasis. Acta Derm Venereol 2005; 85:409–13.

50 Hongbo Y, Thomas CL, Harrison MA et al. Translating the science

of quality of life into practice: what do Dermatology Life QualityIndex scores mean? J Invest Dermatol 2005; 125:659–64.

51 Guyatt GH, Kirshner B, Jaeschke R. Measuring health status: whichare the necessary measurement properties? J Clin Epidemiol 1992;

45:1341–5.52 Katsambas A. Quality of life in dermatology and the EADV. J Eur

Acad Dermatol Venereol 1994; 3:211–14.

Appendix: the Family Dermatology Life QualityIndex (FDLQI)

Name: ................................................ FDLQI Score

Relationship with patient: .........................................

Patient’s diagnosis (if known): ....................................... Date:

...........................

• The questions relate to the impact of your relative/partner’s

skin disease on your quality of life over the last month.

• Please read the questions carefully and tick one box for

each.

1. Over the last month how much emotional distress have you

experienced due to your relative/partner’s skin disease (e.g.

worry, depression, embarrassment, frustration)?

Not at all/Not relevant u A little u A lot u Very much u

2. Over the last month how much has your relative/partner’s

skin disease affected your physical well-being (e.g. tiredness,

exhaustion, contribution to poor health, sleep/rest disturb-

ance)?

Not at all/Not relevant u A little u Quite a lot u Very much u


skin disease affected your personal relationships with him/her

or with other people?


4. Over the last month how much have you been having

problems with other peoples’ reactions due to your relative/

partner’s skin disease (e.g. bullying, staring, need to explain to

others about his/her skin problem)?



skin disease affected your social life (e.g. going out, visiting or

inviting people, attending social gatherings)?



skin disease affected your recreation/leisure activities (e.g. holi-

days, personal hobbies, gym, sports, swimming, watching TV)?


7. Over the last month how much time have you spent on look-

ing after your relative/partner (e.g. putting on creams, giving

medicines or looking after their skin)?


8. Over the last month how much extra housework have you

had to do because of your relative/partner’s skin disease (e.g.

cleaning, vacuuming, washing, cooking)?



skin disease affected your job/study (e.g. need to take time off,

not able to work, decrease in the number of hours worked, hav-

ing problems with people at work)?



skin disease increased your routine household expenditure (e.g.

travel costs, buying special products, creams, cosmetics)?


Thank you for completing the questionnaire.

� MKA Basra, AY Finlay. Cardiff University 2005.

The FDLQI is copyright. For more information please contact

the authors.




PHOTOBIOLOGY DOI 10.1111/j .1365-2133.2006.07670.x

Repeated low-dose ultraviolet (UV) B exposures of humansinduce limited photoprotection against the immune effectsof erythemal UVB radiationJ. Narbutt, A. Lesiak, A. Sysa-Jedrzejowska, A. Wozniacka, A. Cierniewska-Cieslak,* J. Boncela,� C. Jochymski,�W. Kozlowski,� A. Zalewska, M. Skibinska§ and M. Norval–

Department of Dermatology and *Centre for Molecular and Medical Biophysics, Medical University of Lodz, Krzemieniecka 5, 94-017 Lodz, Poland

�Center of Medical Biology, Polish Academy of Sciences, Lodz, Poland

�Department of Pathology, Military Medical Institute, Warsaw, Poland

§Basildon Hospital, Basildon, Essex, U.K.

–Medical Microbiology, University of Edinburgh Medical School, Edinburgh, U.K.

CorrespondenceJoanna Narbutt.



Key wordscyclooxygenase-1, cyclooxygenase-2, cytokines,

photoprotection, thymine dimers, ultraviolet-

induced immunosuppression


Summary

Background Exposure of human subjects to ultraviolet (UV) B radiation causes im-munosuppression. Most experiments to date have not tested the effects of lowdaily doses of UVB radiation.Objectives To ascertain whether photoprotection against several UV-inducedimmune effects might develop following repeated exposure.Methods Groups of approximately 30 healthy individuals were given whole-bodyUVB irradiation on each of 10 consecutive days with 0.7 minimal erythema dose,or whole-body irradiation as before followed by a single erythemal UVB dose ona small body area, or irradiated only with a single erythemal UVB dose on asmall body area, or were not irradiated. They were sensitized with diphenyl-cyclopropenone (DPCP) 24 h after the final dose, and skin biopsies collected toassess cytokine mRNA expression and the number of cells with thymine dimersand expression cyclooxygenase (COX)-1 and COX-2.Results The contact hypersensitivity (CHS) response to DPCP was significantlylower in the three irradiated groups compared with the unirradiated controls,while cutaneous interleukin (IL)-1b, IL-6, IL-10 and tumour necrosis factor-amRNAs, COX-1 and COX-2 and thymine dimers were all significantly higher.When the single erythemal UVB dose was given following the repeated lowexposures, a slight downregulation in cytokine expression and thymine dimerformation was indicated.Conclusions The repeated low doses of UVB protected to a limited extent againstthe effects of an erythemal UVB dose on cytokine expression and thymine dimerformation, but not on CHS or COX enzymes.

Solar ultraviolet (UV) radiation reaching the surface of the

earth contains predominantly UVA wavelengths (315–

400 nm) with a smaller component of UVB wavelengths

(280–315 nm). As UVB is more effective than UVA at indu-

cing biological damage, UVB is thought to contribute towards

about 80% of the harmful effects associated with sun expos-

ure.1 One of these effects is suppression of cell-mediated

immune responses. The process involved is complex.2–4 It is

initiated by chromophores in the upper layers of the skin that

are capable of absorbing the radiation. One such chromophore

is DNA5 with the commonest photoproduct being cyclobutane

pyrimidine dimers (thymine dimers, dTT). These dimers are

repaired, but this process is slow with an estimated half-life of

33 h.6 The DNA damage leads to the increased synthesis of

several cytokines.7–11 In addition, an upregulation in the

production of platelet activating factor,12 histamine and

prostaglandin E2 (PGE2)13,14 occurs. Cyclooxygenase (COX)

enzymes are the first in a series that converts arachidonic acid

into prostaglandins and thromboxane. Two isoforms of COX

are found:15 COX-1 is constitutively expressed in almost all

cell types including keratinocytes, while COX-2 is induced in

many tissues in response to stimuli such as injury or inflam-

mation. UV radiation can stimulate COX-2 expression in basal

keratinocytes.16



Modulation of the contact hypersensitivity (CHS) response is

frequently used to evaluate immunity in human subjects follow-

ing UV exposure.17 In most cases the sensitizer, a hapten such as

diphenylcyclopropenone (DPCP), is applied to the skin within a

few days of the UV radiation, with the elicitation phase taking

place several weeks later. To measure local immunosuppression,

the hapten is placed directly on an irradiated skin site. A primary

allergic response (PAR) can develop within 1–2 weeks of sensi-

tization in many individuals who have not previously come in

contact with the particular sensitizer. This response is thought

to indicate the onset of hapten-specific sensitization.18

To date, most information regarding UV-induced immuno-

suppression in human subjects has been gathered from study-

ing the induction phase of CHS following exposure of a small

area of the body to a single UV dose, often > 1 minimal ery-

thema dose (MED).19–21 However, such a protocol does not

mimic natural conditions as most people are exposed to the

sun for short periods of time on a daily basis over the sum-

mer months with many experiencing occasional burning UV

exposures on limited areas of the body. There is the possibility

that, as the skin adapts to the chronic UV radiation by epider-

mal thickening and tanning, photoprotection against the

effects of UV on the immune system may develop. In the pre-

sent study, we aimed to test this possibility by repeatedly

irradiating volunteers with suberythemal UVB, followed by a

single erythemal UVB dose on a small body area. CHS, cytokine

and COX expression, and DNA damage were then assessed in

comparison with unirradiated subjects or others irradiated

with the single erythemal dose only.


Subjects

The study included 140 healthy subjects of either phototype II

or III, as assessed by Fitzpatrick score.22 These are the most

commonly occurring skin types in Poland and include subjects

who do not burn easily and develop a tan on sun exposure.

The volunteers were recruited following advertising in the

local newspapers and within the University of Lodz. They

were without any skin or other disease and were not receiving

any medication. Subjects exposed to high doses of sunlight or

sunlamps within 2 months prior to the study, or previously

sensitized with DPCP, were excluded. The experimental pro-

cedures were conducted in the winter months to decrease any

influence of natural sunlight. Each volunteer gave written

informed consent before entry into the study and underwent a

thorough physical examination and full blood count. The

experimental plan was approved by the Local Ethics Commit-

tee of the Medical University of Lodz and was conducted

according to the Declaration of Helsinki principles.

Phototesting and ultraviolet B irradiation

Phototesting of each volunteer was undertaken approximately

1 week before the study began using a Waldmann Medizin-

technik UV 109 device (Waldmann Medizintechnik, Villingen-

Schwenningen, Germany) containing TL-12 tubes (Philips,

Eindhoven, the Netherlands) emitting 58.6% UVB and 41.4%

UVA. The test was performed with an incremental dose series

on six squares (1 · 1 cm) on the back. The MED was defined

as a just perceptible erythema 24 h later. From this value, 0.7

MED was calculated for each individual although it is recog-

nized that, due to undulations on the body surface and varia-

tions in MED with site, some areas may receive slightly higher

and others slightly lower doses.

The volunteers were divided into five groups as shown in

Table 1: 40 nonirradiated individuals served as the control

group (group A), 30 subjects were given whole-body UV

irradiation for 10 consecutive days with a dose of 0.7 MED

(group B), 30 subjects were given whole-body UV irradiation

for 10 consecutive days with a UV dose of 0.7 MED followed

24 h later by a single UV dose of 3 MED (left buttock,

10 · 10 cm) (group C), 30 subjects were irradiated with a

single UV dose of 3 MED (left buttock, 10 · 10 cm) (group

D) and 10 individuals were irradiated on the left buttock

(10 · 10 cm) with a single UV dose of 4 MED (group E).

The TL-12 lamps, with an erythema effectiveness irradiance

for UVB (280–315 nm, biologically weighted with the CIE

erythema action spectrum) of 0.64 mW cm)2, were used for

the whole-body exposures, and the phototesting device was

used to irradiate the buttock. Measurement of the intensity of

the lamps was performed using a type 1 UV meter calibrated

against a spectrophotometer (Waldmann Medizintechnik). For

erythema, pigmentation, PAR and CHS measurements, all 140

subjects were included. For the cytokine mRNAs (n ¼ 40),

Table 1 Characteristics of subjects in groups A–E

Group

Number of

volunteers

Mean age,

years (range) Sex, F/M Phototype, II/III

MED, J cm)2,

mean (range)

A: unirradiated 40 25.0 (18–36) 17/23 19/21 0.18 (0.09–0.21)B: 10 · 0.7 MED, whole body 30 30.0 (19–38) 19/11 14/16 0.16 (0.09–0.19)

C: 10 · 0.7 MED, whole body(C1) + 3 MED, 10 · 10 cm (C2)

30 27.8 (19–32) 16/14 18/12 0.15 (0.07–0.17)

D: 3 MED, 10 · 10 cm 30 29.4 (18–37) 16/14 11/19 0.14 (0.09–0.19)E: 4 MED, 10 · 10 cm 10 30.6 (19–34) 4/6 5/5 0.14 (0.07–0.19)

MED, minimal erythema dose.



540 Photoprotection following repeated UVB exposures, J. Narbutt et al.

COX proteins (n ¼ 21) and dTT (n ¼ 21), not all the subjects

in each group were used on ethical grounds, as these assess-

ments required the collection of skin biopsies, up to three per

person in some instances.

Erythema and pigmentation

Erythema and pigmentation were quantified using the UV

Optimise 555 device (Chromo-Light, Espergaerde, Denmark)

before and after irradiation, but before sensitization, on the

buttock skin in the UV-exposed groups (groups B–E). The

mean values of three readings were calculated.

Sensitization and elicitation of contact hypersensitivity

Sensitization and elicitation of CHS were performed using

DPCP (Fluka Chemie GmbH, Buchs, Switzerland) as described

by Narbutt et al.23 In brief, the subjects were sensitized on the

left irradiated (groups B–E, 24 h after the final UVB exposure)

or nonirradiated (group A) buttock skin. The sensitization site

was assessed for the development of a PAR, including its

length of time, severity and size. Three weeks later, all the

volunteers received an antigenic challenge on the unirradiated

upper inner left arm skin using a series of 20 lL acetone con-

taining 0.4, 0.8, 1.6, 3.2 and 6.4 lg DPCP. The highest con-

centration of DPCP was applied only if no PAR was observed.

The CHS response was evaluated after 48 h by a subjective vis-

ual scoring system: 0, no reaction; 1, macular erythema; 2,

erythema with infiltration; 3, erythema with infiltration and

papules or vesicles; 4, bullous reaction. This score was deter-

mined by one dermatologist who did not know the group

assignations.

Assessment of interleukin (IL)-1b, IL-6, IL-10 and tumour

necrosis factor-a mRNA expression by relative

quantitative reverse transcription–polymerase chain

reaction

The control samples consisted of 3-mm punch biopsies from

10 subjects in group A, 15 in group C (before irradiation

began) and 15 in group D (before irradiation began), making

40 in total. The irradiated samples consisted of two more

biopsies from the same 15 individuals in group C (C1: 24 h

after 10 days of 0.7 MED UVB, and C2: 24 h after 3 MED

UVB) and one more biopsy from the same 15 individuals in

group D (24 h after 3 MED UVB). The samples (50 mg) were

dissolved in Fenozol reagent and total mRNA prepared using a

total RNA Prep Plus Kit (A&A Biotechnology, Gdynia, Poland).

The first strand of cDNA was synthesised according to the

manufacturer’s instructions (Promega, Madison, WI, U.S.A.)

using (dT)20 primer with 2 lg of total RNA. Then mRNAs

for tumour necrosis factor (TNF)-a, interleukin (IL)-10, IL-6

and IL-1b were determined by relative quantitative reverse

transcription–polymerase chain reaction (RT–PCR) using

b-actin mRNA as an internal standard,24 using the following

primers: 5¢-GTAGCCCATGTTGTAGCAAACC-3¢ and 5¢-GAG-

GACCTGGGAGTAGATGAGG-3¢ for TNF-a, 5¢-GAGAACAGCT

GCACCCACTTCC-3¢ and 5¢-CTGGGTCTTGGTTCTCAGCTTGG-

3¢ for IL-10, 5¢-ACCTGAACCTTCCAAAGATGG-3¢ and 5¢-GACT

GCAGGAACTCCTTAAAGC-3¢ for IL-6, 5¢-CTCGCCAGTGAAAT-

GATGG-3¢ and 5¢-GCATCTTCCTCAGCTTGTCC-3¢ for IL-1band 5¢-CAGCAGATTCAAGCAGCTATGG-3¢ and 5¢-GTCTGTGG-

TGCTGATCTCATCC-3¢ for b-actin. The RT–PCR assay con-

tained 5 lL cDNA sample, 10 · Taq polymerase buffer

(Epicentre Biotechnologies, Madison, WI, U.S.A.), 2 mmol

L)1 MgCl2, 1 · PCR enhancer, deoxyribonucleoside triphos-

phate mix, and 25 pmol of each primer in 50 lL reaction vol-

ume. The samples were denatured at 95 �C for 5 min, then

cooled on ice before 1.25 U Taq DNA polymerase (Epicentre

Biotechnologies) was added. The reaction was performed in

25–30 extension cycles consisting of a 30-s denaturation step

at 94 �C, a 30-s annealing step at 60 �C and a 40-s polymer-

ase extension step at 72 �C. Finally, each reaction was termin-

ated with a 10-min elongation step at 72 �C. The final

products were electrophoresed in 7% polyacrylamide gels

using the genetic size marker 100 bp DNA Ladder (Promega).

Bands were visualized by UV radiation, and the results were

recorded photographically and analysed densitometrically

using LKB Ultrascan XL Enhanced Laser Densitometer (Phar-

macia LKB, Uppsala, Sweden). Concentrations of mRNAs were

normalized in each sample relative to b-actin mRNA. Usually,

four to seven bands were analysed for each sample and the

mean value was calculated. The mean ± SD was then calcu-

lated for each group.

Expression of cyclooxygenase (COX)-1 and COX-2, and

assessment of DNA damage (thymine dimers)

Three-millimetre biopsies were collected from buttock skin

of seven unirradiated individuals (group A), seven individu-

als in group C (C1: 24 h after 10 days of 0.7 MED UVB,

and C2: 24 h after 3 MED UVB), and seven individuals in

group D (24 h after 3 MED UVB). The samples were put in

liquid nitrogen immediately, and stored at )80 �C until

analysis.

For COX-1 and COX-2 assessment, 4-lm frozen sections

were collected on to Superfrost Plus slides (Menzel, Braun-

schweig, Germany), fixed in acetone for 10 min and air

dried. Endogenous peroxidase was blocked by incubating for

5 min with 3% hydrogen peroxide before washing in phos-

phate-buffered saline (PBS) and incubation for 30 min in 10%

normal horse serum in Tris–HCl–PBS and 1% bovine serum

albumin (BSA). After rinsing, the sections were incubated

overnight at 4 �C with the COX-1 monoclonal antibody

(mAb) (dilution 1 : 40, NCL-COX-1, clone 12E12; Novocas-

tra, Newcastle upon Tyne, U.K.) or COX-2 mAb (dilution

1 : 50, NCL-COX-2, clone 4H12; Novocastra). Detection

was performed with LSAB+ System-HRP (Dako Cytomation,

Glostrup, Denmark). High-sensitivity diaminobenzidine (Dako

Cytomation) chromogenic substrate system was used. The

slides were counterstained with haematoxylin and dehydrated.

Negative controls were performed by omitting the primary



Photoprotection following repeated UVB exposures, J. Narbutt et al. 541

antibody. The presence of brown colour at the sites of the tar-

get antigens (in cellular membrane, cytoplasm) indicated posi-

tive immunoreactivity. The number of COX-1+ or COX-2+

cells in the epidermis was counted by a computer-image pro-

gram (SIS Analysis; Olympus, Tokyo, Japan). At least two

sections were examined from each subject at · 400 magnifica-

tion, and the mean ± SD number of positive cells mm)2 for

each group calculated.

For the assessment of dTT, 4-lm cryostat sections were col-

lected on to Superfrost Plus slides (Menzel), fixed for 1 h in

4% buffered formalin at room temperature and air dried. The

sections were treated for 15 min of microwave oven heating

in Target Retrieval Solution (citrate buffer, pH 6.0; Dako Cyto-

mation) at 800 W and then transferred to distilled water.

Endogenous peroxidase was blocked by incubation for 30 min

with 0.3% hydrogen peroxide in methanol. Slides were

washed in PBS and incubated for 30 min with 10% normal

rat serum in Tris–HCl–PBS and 1% BSA. After rinsing, the sec-

tions were incubated overnight at 4 �C with the dTT mAb

(dilution 1 : 4000, clone KTM53; Kamiya Biomedical Com-

pany, Seattle, WA, U.S.A.). Detection, colour development and

counterstaining were performed as outlined above for COX.

Negative controls were performed by omitting the primary

antibody. Stained sections were examined at · 400 magnifica-

tion throughout the whole of the epidermis. Presence of

brown colour at the target antigen in cellular nucleus indicat-

ed positive immunoreactivity. The number of dTT+ nuclei

was counted by a computer-image program (SIS Analysis). At

least two sections were examined from each subject and the

mean ± SD number of positive cells mm)2 for each group

calculated.


The Mann–Whitney test, Wilcoxon pair test and v2 test were

used to analyse the results, with P < 0.05 being considered

statistically significant.

Results

Repeated suberythemal ultraviolet (UV) B exposure

caused erythema and pigmentation development, and

was slightly protective against the erythemal effects of a

subsequent high UVB dose

Clinical examination revealed no visible erythema after

10 days of suberythemal whole-body UVB but a slight

pigmentation was evident. Irradiation with 3 or 4 MED UVB

on the small body area caused well-defined oedematous ery-

thema. Table 2 shows the detailed results for the measurement

of erythema and pigmentation in groups B–E. In all four

groups, the erythema values were significantly higher

(P < 0.005) following the UV exposure than before irradia-

tion. In group C, after 10 days of irradiation with suberythe-

mal UVB followed by irradiation with 3 MED on the small

body area, the erythema value was significantly lower than in

group D (P < 0.0001) and group E (P < 0.0001). Suberythe-

mal UVB on 10 consecutive days led to a significant increase

in pigmentation (P ¼ 0.001 for group B and 0.0019 for

group C), while the additional erythemal UVB dose decreased

the pigmentation value in group C (P ¼ 0.000004). Follow-

ing erythemal UVB, pigmentation was significantly lower in

group D compared with group C (P ¼ 0.0004). The decrease

in pigmentation observed in groups D and E may have resul-

ted from the intense erythema and oedema which cause

changes in the optical properties of the skin.

Repeated suberythemal ultraviolet (UV) B exposures

suppressed the occurrence of primary allergic response

and contact hypersensitivity, and did not protect against

the immunosuppressive effects of a subsequent

erythemal UVB dose

PAR was detectable a mean of 8.8 days after sensitization.

The response ranged from erythema with definite borders

Table 2 Erythema and pigmentation following ultraviolet (UV) exposure in groups B–E (mean ± SD, n ¼ 30 in groups B–D and n ¼ 10 ingroup E)

Group B:10 · 0.7 MED,

whole body

Group C: 10 · 0.7 MED,whole body + 3 MED,

10 · 10 cm

Group D: 3 MED,

10 · 10 cm

Group E: 4 MED,

10 · 10 cm

Erythema% value light reflectance before UV 28.1 ± 5.8 26.0 ± 5.3 29.2 ± 5.8 25.2 ± 4.6

% value light reflectance after 10 · 0.7 MED 36.6 ± 7.0 37.8 ± 6.7 – –% value light reflectance after

3 or 4 MED (10 · 10 cm)

– 48.9 ± 7.4 62.6 ± 9.9 60.4 ± 7.1

Pigmentation

% value light reflectance before UV 20.6 ± 5.7 21.6 ± 7.1 21.1 ± 8.0 23.2 ± 6.2% value light reflectance after 10 · 0.7 MED 24.0 ± 5.8 24.9 ± 6.7 – –

% value light reflectance after 3 or4 MED (10 · 10 cm)

– 22.1 ± 8.3 13.9 ± 9.0 17.4 ± 8.1

MED, minimal erythema dose.




to strong erythema with oedema, and vesicles in some

cases, but without blistering. The diameter of the PAR var-

ied between 9 and 19 mm. The percentage of subjects in

the unirradiated group with PAR (60%) was significantly

higher than in all the irradiated groups (26.7%, 33.3%,

26.7% and 10% in groups B, C, D and E, respectively;

P < 0.05 for all comparisons). There were no statistical dif-

ferences between the irradiated groups B–E with respect

to the occurrence of PAR, its length of time, severity or

diameter.

The CHS response at the highest concentration of DPCP was

not analysed statistically as the eliciting dose of DPCP was

applied only in the volunteers in whom PAR did not develop.

UV radiation in all the groups suppressed the CHS response,

and it was observed at all the DPCP concentrations (P < 0.05

for all comparisons, data not shown). As analysis of the visual

assessment of the CHS revealed statistically significant differ-

ences between the unirradiated group and any of the irra-

diated groups at any of the DPCP concentrations and because,

in the majority of the irradiated volunteers, the CHS was

assessed as 0 (absent), we decided to simplify the analysis by

treating the CHS response as a binominal trait where 0 repre-

sented no reaction and 1 represented a response. These results

are shown in Figure 1. Statistical analysis of the CHS response

showed no significant differences between subjects in group B

and group C (P > 0.05), or between subjects in group C and

group D (P > 0.05).


increased the expression of interleukin (IL)-1b, IL-6,

IL-10 and tumour necrosis factor-a mRNAs in skin

biopsies, and slightly protected against the increased

expression of these cytokines induced by a subsequent

erythemal UVB dose

Following the suberythemal UVB exposures for 10 consecu-

tive days (group C1), the expression of the mRNAs of

IL-1b, IL-6, IL-10 and TNF-a increased significantly

(Fig. 2a–d). After the additional exposure to erythemal UVB

(group C2), there was a further, although smaller, increase

in IL-1b, IL-6 and IL-10 mRNA expression and a decrease in

TNF-a mRNA expression. The single erythemal UVB dose

induced a large rise in the expression of all four cytokine

mRNAs (group D). The increase was greater than that fol-

lowing the repeated suberythemal exposures and also follow-

ing the repeated suberythemal exposures plus the additional

erythemal exposure.

0102030405060708090

100%

of

sub

ject

s w

ith

no

resp

on

se

0·4 μgDPCP

0·8 μgDPCP

1·6 μgDPCP

3·2 μgDPCP

EEEEA A A AB B B B CCCC DDDD

Fig 1. Percentage of subjects with no contact hypersensitivity

response to the challenge dose of 0.4, 0.8, 1.6 and 3.2 lg

diphenylcyclopropenone (DPCP) in groups A–E.

0

10000

20000

30000

40000

50000

60000

(b)

(a)

(c)

(d)

Exp

ress

ion

of

IL-1

bet

a m

RN

A(a

rbit

rary

un

its)

P<0·000001

P=0·0000063 P<0·01

P=0·04 P=0·00048

P=0·000013

DC2C1Unirradiated

0

10000

20000

30000

40000

50000

60000

Exp

ress

ion

of

IL-6

mR

NA

(arb

itra

ry u

nit

s)

P<0·000001

P=0·0045P<0·01

P=0·017

P<0·001

P=0·000037

DC2C1Unirradiated

0

10000

20000

30000

40000

50000

60000

Exp

ress

ion

of

TN

F-a

lph

a m

RN

A(a

rbit

rary

un

its)

P<0·0001

P<0·001P=0·0045

P>0·05

P=0·017

P=0·046

DC2C1Unirradiated

0

10000

20000

30000

40000

50000

60000

Exp

ress

ion

of

IL-1

0 m

RN

A(a

rbit

rary

un

its)

P<0·000001

P=0·004P<0·01

P>0·05

P=0·0011

DC2C1Unirradiated

Fig 2. Expression of (a) interleukin (IL)-1b, (b) IL-6, (c) tumour

necrosis factor (TNF)-a and (d) IL-10 mRNAs in skin biopsies from

unirradiated subjects (n ¼ 40), subjects irradiated with 0.7 minimal

erythema dose (MED) ultraviolet (UV) B for 10 consecutive days

(C1, n ¼ 15) followed by 3 MED UVB (C2, n ¼ 15), and subjects

irradiated with 3 MED UVB (D, n ¼ 15). The cytokine expression was

normalized to the b-actin mRNA intensity and is presented in

arbitrary units. The mean ± SD for each group is shown.





increased the number of cyclooxygenase (COX)-1+ and

COX-2+ cells in skin biopsies, and did not protect

against the increased number induced by a subsequent

erythemal UVB dose

In unirradiated skin biopsies, COX-1+ cells were observed

only occasionally, mainly in the basal layer of the epidermis,

and no COX-2+ cells were seen. The number of both COX-

1+ and COX-2+ cells increased significantly following UVB

exposure (P < 0.05 for all comparisons, Fig. 3). In irradiated

skin, the COX-1+ cells were located in both the basal layers

and the upper part of the epidermis, while the COX-2+

cells were present mainly in the suprabasal layer of the epi-

dermis (Fig. 4). The additional erythemal UVB dose follow-

ing the repeated suberythemal doses (group C2) induced

almost the same number of COX-1+ and COX-2+ cells as

the single erythemal UVB dose on its own (group D)

(Fig. 3), indicating that photoprotection for this factor had

not developed.


increased the number of thymine dimer cells in skin

biopsies, and slightly protected against the increased

number induced by a subsequent erythemal UVB dose

In unirradiated skin, only a few dTT cells were observed

(15 mm)2), scattered within the epidermis, and the staining

was considered weak. The number increased significantly fol-

lowing UV exposure (P < 0.05 for all comparisons, Fig. 5).

The highest was seen after irradiation with the single erythe-

mal dose (group D) and this number was significantly greater

than that following the repeated suberythemal doses followed

by the single erythemal dose (group C2) (P ¼ 0.00001). The

dTT+ cells were found mainly in the upper parts of the epi-

dermis following the repeated suberythemal doses of UVB,

while they were present throughout the epidermis following

the single erythemal dose.

Discussion

In a previous study we attempted to simulate the natural ex-

posure of subjects with skin phototypes II/III to summer sun-

light by whole-body irradiation with 0.3 MED solar-simulated

0

100

200

300

400

500

600

Nu

mb

er o

f C

OX

-1+

an

dC

OX

-2+

cells

mm

–2

COX-2

D C2 C1 A

COX-2 COX-2 COX-2 COX-1 COX-1 COX-1 COX-1

Fig 3. Number of cyclooxygenase (COX)-1+ and COX-2+

cells mm)2 in the epidermis of unirradiated subjects (A, n ¼ 7),

subjects irradiated daily with 0.7 minimal erythema dose (MED)

ultraviolet (UV) B for 10 consecutive days (C1, n ¼ 7) followed by 3

MED UVB (C2, n ¼ 7), and subjects irradiated with 3 MED UVB (D,

n ¼ 7). The mean ± SD for each group is shown.

(a)

(b)

Fig 4. Expression of cyclooxygenase (COX)-1 (a) and COX-2 (b) in

human epidermis after 0.7 minimal erythema dose (MED) ultraviolet

(UV) B for 10 consecutive days followed by 3 MED UVB as shown by

immunohistochemistry (original magnification · 400).

0 100 200 300 400 500 600 700 800 900

1000

Nu

mb

er o

f ep

ider

mal

dT

T+

cells

mm

–2

P=0·000001

P<0·000001

D C2 C1 A

Fig 5. Number of thymine dimers (dTT+) cells mm)2 in the

epidermis of unirradiated subjects (A, n ¼ 7), subjects irradiated daily

with 0.7 minimal erythema dose (MED) ultraviolet (UV) B for 10

consecutive days (C1, n ¼ 7) followed by 3 MED UVB (C2, n ¼ 7),

and subjects irradiated with 3 MED UVB (D, n ¼ 7). The mean ± SD

for each group is shown.




radiation (SSR) daily for up to 30 days.23 Their immune

responses to DPCP were then assessed. We found that both

the PAR and the CHS were reduced. These effects were de-

pendent on the cumulative UV dose as the downregulation

became more apparent as the number of days of exposure in-

creased. Therefore photoadaptation of this immune response

had not occurred, despite the development of some pigmenta-

tion in the skin.

In the present study, broadband UVB lamps were used

rather than SSR but, as the UVB waveband is the most biolo-

gically active and has similar immunological effects as SSR, we

assume that the changes induced by such lamps mimic closely

those following exposure to natural sunlight. We wished to

establish whether photoprotection against several effects of an

erythemal UVB dose was generated by pre-exposure of sub-

jects to suberythemal whole-body UVB radiation for 10 days.

We estimate that the dose of UVB given daily was equivalent

to being in the sun for about 35 min around midday on a

clear sky summer day in mid-Europe. The repeated irradia-

tions resulted in a small but significant degree of protection

against the erythemal effects of the subsequent high UVB

dose. This might be expected as the subjects were all of photo-

type II/III (the commonest skin type in Poland) with the

ability to tan to some extent in response to solar UV radiation.

The repeated low UVB doses resulted in the suppression of

both the PAR and CHS, indicating a lack of photoadaptation

and corroborating our previous study.23 In addition, these ex-

posures did not offer any protection against the downregulat-

ing effects of a subsequent erythemal UVB dose on PAR and

CHS. In one of the few previous reports to monitor CHS after

several exposures to low doses of UVB, Cooper et al.19 showed

that irradiation of a small body area with 0.7 MED UVB on

four consecutive days caused suppression of the CHS response

in 68% of subjects. Damian et al.25 demonstrated that the recall

response to nickel in nickel-allergic volunteers was suppressed

by exposure of a small area of the skin to suberythemal SSR

and that this effect was maintained even after 4 weeks of

repeated UV exposures. Long-term UVB phototherapy given to

patients with psoriasis was reported to impair subsequent sen-

sitization to DPCP.26

Following UV radiation of human skin, there is an upregu-

lation in the expression of various cytokines,10, 27–33 induced

by various mediators such as calcitonin gene-related peptide,

a-melanocyte stimulating hormone, platelet activating factor,

histamine and PGE2.3 The vast majority of the studies monit-

oring cytokine changes has been performed either in vivo

following a single erythemal dose of UVB, or in vitro in

keratinocyte cultures. As far as we are aware, the situation

in vivo following repeated suberythemal exposures has not been

examined previously. We found that the mRNAs of the four

selected cytokines, IL-1b, IL-6, IL-10 and TNF-a, were

increased as a result of the 10 daily suberythemal UVB expos-

ures. However, the rise for each cytokine was significantly less

than that following the single erythemal UVB exposure, indi-

cating that the extent of the change was not dependent on the

cumulative dose of UVB. It was also noted that a small but

significant photoprotection against the mRNA induction of all

four cytokines occurred in the group irradiated daily for

10 days before receiving the single erythemal UVB exposure

(group C2, Fig. 2a–d) compared with the group irradiated

with the single erythemal UVB dose (group D, Fig. 2a–d).

PGE2 is the predominant prostaglandin formed in the

human epidermis, and its production is increased after UV

exposure.34 Of the two major forms of COX, COX-2 is

induced by both UVB and UVA-II radiation.35 The COX

enzymes have important effects on apoptosis and proliferation

in the skin,36–38 and PGE2 has been shown to play a critical

role in UV-induced systemic immunosuppression in mice.13,16

It was demonstrated by Athar et al.39 that the extent of COX-2

protein expression in the epidermis of UVB-irradiated mice

strongly correlated with dose; after several weeks of exposure,

COX-2 was present in all the layers of the epidermis but most

strongly in the basal layer. We found a similar pattern of

expression of COX-2 in human skin following the 10 daily

suberythemal exposures to UVB. The UVB radiation increased

COX-1 and COX-2 protein expression even when no clinical

signs of erythema were present. When the numbers of

COX-1+ and COX-2+ cells in the epidermis of subjects irradi-

ated with the single erythemal UVB dose were compared with

the numbers in subjects preirradiated with 10 daily subery-

themal doses followed by the erythemal dose, no evidence for

the development of photoprotection was obtained.

One of the major absorbers of UV radiation in the skin is

DNA, and DNA photoproducts, such as dTT, are recognized as

critical molecular triggers for local and systemic suppression

of CHS.7,40 IL-12, given either before or after the UV expos-

ure, can prevent this effect.41–44 DNA damage is associated

with many biological changes in the skin other than immuno-

modulation, including erythema and sunburn cell forma-

tion,45,46 and it also enhances melanogenesis.47 In the present

study, all the UV protocols increased the number of dTT+

epidermal cells, most after the single erythemal UVB dose.

The number in the skin of subjects irradiated with the 10

daily doses of suberythemal UVB before administration of the

single erythemal dose was about 34% less than this figure.

Thus some photoprotection against DNA damage had been

generated by the repeated pre-exposures. As erythema corre-

lates with DNA damage, the small degree of photoprotection

observed against erythema may explain this result. However,

the protection was by no means absolute as, in the repeatedly

exposed subjects, the subsequent single erythemal UV dose in-

creased the number of dTT+ cells substantially, from about

363 to about 573 mm)2. In confirmation of this result,

Sheehan et al.48 showed that dTT steadily accumulated during

12 daily suberythemal SSR exposures of volunteers with skin

types II and IV, and that a tan may not be the major factor in

photoprotection against erythema.

In conclusion, we have shown that repeated exposures of

individuals to suberythemal UVB resulted in the generation of

slight erythema and pigmentation, the suppression of PAR and

CHS, and an increase in the mRNA expression of four cyto-

kines, in COX-1 and COX-2 proteins, and in dTT. The repeated




irradiations protected to a limited extent against the subse-

quent effects of an erythemal UVB dose on erythema, cytokine

expression and dTT formation, but not on the suppression of

PAR and CHS or the expression of the COX enzymes. Thus

sun protection measures are recommended, even in subjects

who have developed a tan.

Acknowledgments

This study was funded by the European Union project QTL-

CT-2001-0022IHA-UV and the Medical University of Lodz

projects 503-1019-1 and 502-11-352. The dTT monoclonal

antibody was kindly provided by Professor H. van Loveren.

References

1 Diffey BL. Sources and measurement of ultraviolet radiation. Methods2002; 28:4–13.

2 Clydesdale GJ, Dandie GW, Muller HK. Ultraviolet light inducedinjury: immunological and inflammatory effects. Immunol Cell Biol

2001; 79:547–68.3 Ullrich SE. Mechanisms underlying UV-induced immune suppres-

sion. Mutat Res 2005; 57:185–205.4 Schwarz T. Regulatory T cells induced by ultraviolet radiation.

Int Arch Allergy Immunol 2005; 137:187–93.

5 Vink AA, Shreedhar V, Roza L et al. Cellular target of UVB-inducedDNA damage resulting in local suppression of contact hypersensi-

tivity. J Photochem Photobiol B Biol 1998; 44:107–11.6 Young AR, Chadwick CA, Harrison GI et al. The in situ repair kinet-

ics of epidermal thymine dimers and 6–4 photoproducts in humanskin types I and II. J Invest Dermatol 1996; 106:1307–13.

7 Kibitel J, Hejmadi V, Alas L et al. UV-DNA damage in mouse andhuman cells induces the expression of tumor necrosis factor a.

Photochem Photobiol 1998; 67:541–6.8 Petit-Frere C, Clingen PH, Grewe M et al. Induction of interleukin-

6 production by ultraviolet radiation in normal human epidermalkeratinocytes and in a human keratinocyte cell line is mediated by

DNA damage. J Invest Dermatol 1998; 111:354–9.9 Nishigori C, Yarosh DB, Ullrich SE et al. Evidence that DNA

damage triggers interleukin-10 cytokine production in UV-irradi-ated murine keratinocytes. Proc Natl Acad Sci U S A 1996; 93:

10354–9.10 Skov L, Hansen H, Allen M et al. Contrasting effects of ultraviolet

A1 and ultraviolet B exposure on the induction of tumour necrosisfactor-alpha in human skin. Br J Dermatol 1998; 38:216–20.

11 Brink N, Szamel M, Young AR et al. Comparative quantification ofIL-1beta, IL-10, IL-10r, TNFalpha and IL-7 mRNA levels in UV-

irradiated human skin in vivo. Inflamm Res 2000; 49:290–6.12 Walterscheid JP, Ullrich SE, Nghiem DX. Platelet-activating factor,

a molecular sensor for cellular damage, activates systemic immunesuppression. J Exp Med 2002; 195:171–9.

13 Hart PH, Jaksic A, Swift G et al. Histamine involvement in UVB-and cis-urocanic acid-induced systemic suppression of contact

hypersensitivity responses. Immunology 1997; 91:601–8.14 Shreedhar V, Giese T, Sung VW, Ullrich SE. A cytokine cascade

including prostaglandin E2, interleukin-4, and interleukin-10 isresponsible for UV-induced systemic immune suppression. J Immu-

nol 1998; 160:3783–9.

15 Smith WL, Garavito RM, DeWitt DL. Prostaglandin endoperoxidaseH synthase (cyclooxygenases-1 and -2). J Biol Chem 1996;

271:33157–60.

16 Tripp CS, Blomme EAG, Chinn KS et al. Epidermal COX-2 inductionfollowing ultraviolet irradiation: suggested mechanism for the role

of COX-2 inhibition in photoprotection. J Invest Dermatol 2003;121:853–61.

17 Damian DL, Halliday GM. Measurement of ultraviolet radiation-induced suppression of recall contact and delayed-type hypersensi-

tivity in humans. Methods 2002; 28:34–45.18 Tseng C, Hoffman B, Kurimoto I et al. Analysis of effects of ultravi-

olet B radiation on induction of primary allergic reactions. J InvestDermatol 1992; 98:871–5.

19 Cooper KD, Oberhelman L, Hamilton TA et al. UV exposure redu-

ces immunization rates and promotes tolerance to epicutaneousantigens in humans: relationship to dose, CD1a-DR+ epidermal

macrophage induction, and Langerhans cell depletion. Proc Natl AcadSci U S A 2003; 89:8497–501.

20 Kelly DA, Walker SL, McGregor JM, Young AR. A single exposureof solar simulated radiation suppresses contact hypersensitivity

responses both locally and systemically in humans: quantitativestudies with high frequency ultrasound. J Photochem Photobiol B Biol

1998; 44:130–42.21 Wolf P, Hoffmann C, Quehenberger F et al. Immune protection

factors of chemical sunscreens measured in the local contacthypersensitivity model in humans. J Invest Dermatol 2003; 121:

1080–7.22 Fitzpatrick TB. The validity and practicality of sun-reactive skin

types I through VI. Arch Dermatol 1988; 124:869–71.23 Narbutt J, Lesiak A, Skibinska M et al. Suppression of contact hyper-

sensitivity after repeated exposures of humans to low doses of solarsimulated radiation. Photochem Photobiol Sci 2005; 4:517–22.

24 Al-Nedawi KN, Czyz M, Bednarek R et al. Thymosin beta4 inducesthe synthesis of plasminogen activator inhibitor 1 in cultured

endothelial cells and increases its extracellular expression. Blood2004; 103:1319–24.

25 Damian DL, Barnetson RS, Halliday GM. Effects of low-dose ultra-violet radiation on in vivo human cutaneous recall responses. Australas

J Dermatol 2001; 42:161–7.26 Friedmann PS, White SI, Parker S et al. Antigenic stimulation during

ultraviolet therapy in man does not result in immunological toler-ance. Clin Exp Immunol 1989; 76:68–72.

27 Oxholm A, Oxholm P, Staberg B, Bendtzen K. Immunohistologicaldetection of interleukin 1-like molecules and tumor necrosis factor

in human epidermis before and after UVB-irradiation in vivo. Br JDermatol 1998; 118:369–76.

28 Chung JH, Youn SH, Koh WS et al. Ultraviolet B irradiation-

enhanced interleukin (IL)-6 production and mRNA expression aremediated by IL-1alpha in cultured human keratinocytes. J Invest

Dermatol 1996; 106:715–20.29 Barr RM, Walker SL, Tsang W et al. Suppressed alloantigen presen-

tation, increased TNF-alpha, IL-1, IL-1Ra, IL-10, and modulationof TNF-R in UV-irradiated human skin. J Invest Dermatol 1999;

112:692–8.30 Enk AH, Katz SI. Identification and induction of keratinocyte-

derived IL-10. J Immunol 1992; 149:92–5.31 Kang K, Hammerberg C, Meunier L, Cooper KD. CD11b+ macro-

phages that infiltrate human epidermis after in vivo ultravioletexposure potently produce IL-10 and represent the major secretory

source of epidermal IL-10 protein. J Immunol 1994; 153:5256–64.

32 Piskin G, Bos JD, Teunissen MB. Neutrophils infiltrating ultravioletB-irradiated normal human skin display high IL-10 expression. Arch

Dermatol Res 2005; 269:339–42.33 Strickland I, Rhodes LE, Flanagan BF, Friedmann PS. TNF-a and

IL-8 are upregulated in the epidermis of normal human skin after




UVB exposure: correlation with neutrophil accumulation andE-selectin expression. J Invest Dermatol 1997; 108:763–8.

34 Pentland AP, Mahoney M, Jacobs SC, Holtzman MJ. Enhancedprostaglandin synthesis after ultraviolet injury is mediated by

endogenous histamine stimulation. J Clin Invest 1990; 86:566–74.

35 Mahns A, Wolber R, Stab F et al. Contribution of UVB and UVA toUV-dependent stimulation of cyclooxygenase-2 expression in artifi-

cial epidermis. Photochem Photobiol Sci 2004; 3:257–62.36 Muller-Decker K, Kopp-Schneider A, Marks F et al. Localization of

prostaglandin H synthase isozymes in murine epidermal tumors:

suppression of skin tumor promotion by inhibition of prostaglan-din H synthase-2. Mol Carcinog 1998; 23:36–44.

37 Pentland AP, Schoggins JW, Scott GA et al. Reduction of UV-induced skin tumors in hairless mice by selective COX-2 inhibi-

tion. Carcinogenesis 1999; 20:1939–44.38 Pentland AP, Scott G, VanBuskirk J et al. Cyclooxygenase-1 deletion

enhances apoptosis but does not protect against ultraviolet light-induced tumors. Cancer Res 2004; 64:5587–91.

39 Athar M, An KP, Morel KD et al. Ultraviolet B (UVB)-inducedCOX-2 expression in murine skin: an immunohistochemical study.

Biochem Biophys Res Commun 2001; 280:1042–7.40 Kripke ML, Cox PA, Alas LG, Yarosh DB. Pyrimidine dimers in

DNA initiate systemic immunosuppression in UV-irradiated mice.Proc Natl Acad Sci U S A 1992; 89:7516–20.

41 Muller G, Saloga J, Germann T et al. IL-12 as mediator and adjuvantfor the induction of contact sensitivity in vivo. J Immunol 1995;

155:4661–8.42 Schwarz A, Grabbe S, Aragane Y et al. Interleukin-12 prevents

UVB-induced local immunosuppression and overcomes UVB-induced tolerance. J Invest Dermatol 1996; 106:1187–91.

43 Schwarz A, Stander S, Bernburg M et al. Interleukin-12 suppressesultraviolet radiation-induced apoptosis by inducing DNA repair.

Nat Cell Biol 2002; 4:26–31.44 Schwarz A, Maeda A, Kerneback K et al. Prevention of UV radi-

ation-induced immunosuppression by IL-12 is dependent on DNA

repair. J Exp Med 2005; 201:173–9.45 Young AR, Chadwick CA, Harrison GI et al. The similarity of action

spectra for thymine dimers in human epidermis and erythema sug-gests that DNA is the chromophore for erythema. J Invest Dermatol

1998; 111:982–8.46 Stege H, Roza L, Vink AA et al. Enzyme plus light therapy to repair

DNA damage in ultraviolet-B-irradiated human skin. Proc Natl AcadSci U S A 2000; 97:1790–5.

47 Eller MS, Ostrom K, Gilchrest BA. DNA damage enhances melano-genesis. Proc Natl Acad Sci U S A 1996; 93:1087–92.

48 Sheehan JM, Cragg N, Chadwick CA et al. Repeated ultraviolet ex-posure affords the same protection against DNA photodamage and

erythema in human skin types II and IV but is associated with fas-ter DNA repair in skin type IV. J Invest Dermatol 2002; 118:825–9.




THERAPEUTICS DOI 10.1111/j .1365-2133.2006.07621.x

A pilot study of the safety and efficacy of picolinic acid gelin the treatment of acne vulgarisM.P. Heffernan, M.M. Nelson and M.J. Anadkat

Division of Dermatology, Washington University School of Medicine, St Louis, MO, U.S.A.

CorrespondenceMichael P. Heffernan, Division of Dermatology,

Wright State University School of Medicine,

Dayton, OH 45408, U.S.A.



Key wordsacne vulgaris, picolinic acid

Conflicts of interestM.P.H. has been active in industry-sponsored

research with picolinic acid. He has conducted

reimbursed investigations for Novactyl, Inc. and

has received travel grants from Novactyl, Inc.

M.M.N. and M.J.A. have no financial conflicts of

interest to declare.

Summary

Background Cost limitations, adverse effects or lack of efficacy limit the use of cur-rent topical therapies in mild to moderate acne vulgaris.Objectives To determine the safety and efficacy of picolinic acid, a novel zinc fingertherapy, in the treatment of mild to moderate acne vulgaris.Methods Twenty subjects with mild to moderate acne vulgaris were treated at ourcentre during an open-label study with 10% picolinic acid gel (PCL-016) twicedaily to the face over 12 weeks.Results Fifteen patients completed the 12-week open-label study. A reduction of58Æ2% (P < 0Æ001) in mean total lesion count, 55Æ5% (P < 0Æ001) in meaninflammatory lesion count and 59Æ7% (P < 0Æ005) in noninflammatory lesioncount was seen in this population. No serious adverse events or clinically signifi-cant changes in laboratory values were noted.Conclusions Results from this study suggest that 10% picolinic acid gel appliedtwice daily may be safe and effective in the treatment of mild to moderate acnevulgaris.

Acne vulgaris is a widely prevalent condition that affects

nearly 80% of the U.S. population aged 11–30 years, and

most of the population at some point in their life.1–3

Patients with this condition are most bothered by their

appearance.4,5 Affected patients are prone to embarrassment,

social withdrawal, depression, anxiety, anger, scorn and stig-

matization.6–10 Higher unemployment rates have also been

demonstrated in adults with acne vulgaris compared with

those without acne.11 Patients with acne vulgaris have sim-

ilar levels of social, emotional and psychological impair-

ments as patients with chronic illnesses such as asthma,

arthritis and diabetes.12

The pilosebaceous unit is believed to be the primary site of

pathology in acne vulgaris. The distribution of acne favours

sites with the highest concentration of pilosebaceous units,

such as the face, chest or back.2,6,13,14 Four main processes

have been suggested as factors in the development of acne

vulgaris: sebaceous gland hyperplasia with increased sebum

production, altered follicular growth and differentiation, fol-

licular colonization with Propionibacterium acnes, and inflamma-

tion.13,15–18 The primary lesion of acne vulgaris, the

microcomedo, is formed as a result of sebaceous gland hyper-

plasia and alteration in follicular growth and differentiation.

Resultant noninflammatory (open and closed comedones) or

inflammatory (papule, pustule, nodule) lesions may later

evolve. Scarring often results in severe cases, but may occur in

instances of mild to moderate disease.19,20

Increasingly effective therapies have evolved in recent years

as the pathogenesis of acne vulgaris has become better under-

stood. Topical retinoids, antibiotics, benzoyl peroxide, azelaic

acid and salicylic acid are currently the mainstays of treatment

for mild to moderate acne vulgaris. Many patients fail to

improve with these agents due to cost, adverse effects leading

to noncompliance (i.e. irritation), or lack of therapeutic bene-

fit. The use of oral antibiotics or systemic retinoids increases

both the cost and the risks for adverse effects.

Picolinic acid is an intermediate metabolite of the amino

acid tryptophan. It appears to play a key role in zinc transport.

As a therapeutic agent, the molecule seems to work by per-

turbing zinc binding in zinc finger proteins (ZFPs). Picolinic

acid has antiviral and antibacterial properties in vitro and in vivo,

and also modifies the immune response alone and in conjunc-

tion with other cytokines. Picolinic acid was first evaluated in

the treatment of herpes labialis.

Here we report the results in 20 subjects with mild to mod-

erate acne vulgaris treated at our centre during an open-label,

phase I study of 10% picolinic acid gel (PCL-016).


Patients

Subjects were eligible for study participation if they were at

least 18 years of age and had a clinical diagnosis of mild to



moderate acne vulgaris of the face at baseline. Subjects were

required to have between 10 and 100 inflammatory lesions

(papules and pustules) and no more than two nodules. Sub-

jects were judged to be in generally good health as deter-

mined by the principal investigator. Subjects were excluded if

they had other active skin diseases that may interfere with

evaluation; had a beard or other facial hair that may interfere

with evaluation; had a history of an allergic reaction or signifi-

cant sensitivity to constituents of the study drug; had a poorly

controlled medical condition; were female and pregnant or

breastfeeding or considering becoming pregnant during the

study; had a history of clinically significant drug or alcohol

abuse in the last year; had participated in a clinical research

study within 30 days of enrolment; or were considered unre-

liable or unable to understand protocol directions in the esti-

mation of the investigator. Subjects with other forms of acne

such as acne rosacea, acne excoriee, chloracne, acne conglo-

bata, acne fulminans or drug-induced acne were excluded.

Subjects with moderate to severe nodulocystic acne deemed

by the investigator to require systemic treatment were also

excluded.

All subjects provided written informed consent before

inclusion into the study. The study protocol was approved by

the Institutional Review Board of the Washington University

School of Medicine.

Methods

All subjects received open-label 10% picolinic acid (PCL-016)

gel (NV-02; Novactyl, Inc., St Louis, MO, U.S.A.) to be

applied in a thin layer to affected areas of the face twice daily

for 12 weeks. Picolinic acid was stored at room temperature

at the study site. Subjects stored study medication at home at

room temperature after it was dispensed to them.

Subjects were not permitted to use any treatments for acne

vulgaris during the study. Subjects discontinued topical astrin-

gents for 1 day and other topical medications, such as antibi-

otics, antiseptics, corticosteroids and retinoids, for at least

2 weeks. Subjects discontinued systemic corticosteroids and

oral antibiotics for at least 4 weeks. Systemic retinoids were

discontinued for at least 6 months. All other systemic therapy

for acne vulgaris required discontinuation for at least

3 months. Women using oral contraception were required to

be taking the same contraceptive for at least 3 months prior to

study entry and to agree to continue this therapy until after

completion of the study.

Clinical and laboratory assessments were performed at

weeks 0, 4, 8 and 12 after the start of therapy. Clinical assess-

ments included physical examination and an acne lesion

count. A Physician Global Improvement score was also deter-

mined at the final visit. Laboratory assessments included a full

blood count, serum chemistry, urine pregnancy test (if applic-

able) and plasma picolinic acid level. Concomitant medications

and drug compliance were reviewed at each visit. All adverse

events were assessed and recorded including date and time of

onset, description, severity, time course, duration, outcome

and relationship to study drug. An additional visit was con-

ducted at week 1 following the initiation of therapy for

review of concomitant medications, adverse event query and

laboratory assessments.

Photographs of subjects who gave written informed consent

were taken at weeks 0, 4, 8 and 12. Photographs were taken

with a Nikon N80 camera. Photographs were performed to

document efficacy and were not used for statistical analysis.

The primary efficacy endpoints included improvement in

the total inflammatory lesion count, total lesion count and

global severity score at week 12 compared with baseline. Sec-

ondary efficacy endpoints included improvement in the total

noninflammatory lesion count at week 12 compared with

baseline. Safety variables evaluated included the incidence of

adverse events, abnormalities in routing laboratory monitor-

ing, and comparisons of pre- and post-treatment plasma picol-

inic acid levels. Efficacy and safety outcomes were designed to

direct planning for larger, blinded phase II–III investigations.


P-values were obtained for both per protocol and intent-to-

treat analyses using a paired t-test. Analyses performed using a

single-sample t-test yielded identical values.

Results

Baseline characteristics

Twenty subjects who met all of the inclusion criteria and none

of the exclusion criteria were enrolled in this study. Fifteen

patients completed the 12-week open-label study per protocol.

Four patients were lost to follow-up and one patient withdrew

due to a localized adverse event (discussed later). Table 1

summarizes the baseline characteristics for the participants.

Efficacy

All 15 patients who completed the study demonstrated reduc-

tion in total lesion count and noninflammatory lesion count.

Fourteen of these patients (93Æ3%) had a reduction in inflam-

matory lesion count. A reduction of 58Æ2% (from 85Æ33 to

Table 1 Baseline characteristics

Male 5Female 15

Age range (years) 20–48 (mean 29Æ6)White 12

African-American 6Hispanic 1

Asian 1Total baseline lesions 14–172 (mean 78Æ15)

Total baseline inflammatory lesions 12–58 (mean 27Æ45)Total baseline noninflammatory lesions 2–148 (mean 50Æ7)



Picolinic acid gel for acne vulgaris, M.P. Heffernan et al. 549

35Æ67; P ¼ 0Æ0001) in mean total lesion count, 55Æ5% (from

30Æ40 to 13Æ33; P ¼ 0Æ0002) in mean inflammatory lesion

count and 59Æ7% (from 54Æ93 to 22Æ33; P ¼ 0Æ0007) in non-

inflammatory lesion count was seen using per protocol analy-

sis (Fig. 1).

Similar findings were demonstrated using intent-to-treat an-

alysis for all 20 subjects enrolled into the study. A reduction

of 50Æ6% (from 78Æ15 to 38Æ60; P ¼ 0Æ0001) in mean total

lesion count, 47Æ2% (from 27Æ45 to 14Æ30; P ¼ 0Æ0001) in

mean inflammatory lesion count and 52Æ4% (from 50Æ70 to

24Æ30; P ¼ 0Æ0012) in mean noninflammatory count was seen

using intent-to-treat analysis.

A Physician Global Improvement score was determined for

each of the 15 patients completing the study at the final visit.

Seven (47%) of these patients were judged to be almost clear

(90–99% clearance) or to have marked improvement (75–

89% clearance). Three patients (20%) had moderate (50–74%

clearance) or minimal (25–49% clearance) improvement. Five

patients (33%) were judged to have no change (0–24% clear-

ance) from baseline.

Safety

Nineteen of the 20 patients enrolled were exposed to study

drug for at least 1 week, 17 for at least 4 weeks, and 15 for

all 12 weeks of the study. There were no deaths or serious

adverse events noted during this study. There were no clinic-

ally significant changes in any laboratory values (full blood

count and serum chemistry) during the study. No patients

became pregnant during the study.

One of the 20 subjects enrolled (patient 2) experienced a

drug-related adverse event leading to withdrawal from the

study. This patient described ‘burning’ after application of

study drug lasting 4 h that was ‘too much’. This was classified

by the investigator to be mild and probably related to study

drug. There were no sequelae from this event. Two other

patients described adverse events determined as probably rela-

ted to study drug. Patient 15 noted burning after application

lasting up to 1 h that resolved after the first week of therapy.

Patient 6 described a sensation of ‘facial tingling’ on the first

day of study drug application only. Table 2 summarizes each

of the documented adverse events from this study.

Plasma measurements of picolinic acid were done using a

validated assay developed by ABC Laboratories (now known as

ABC Pharma Services, Columbia, MO, U.S.A.) for Novactyl,

Inc. There were no stipulations regarding timing of blood

draws relative to last dosing. Samples were collected in vacu-

tainers containing powdered heparin. After gentle mixing, the

samples were kept on wet ice and centrifuged within 15 min

of collection. Resulting plasma was aliquoted and kept at

)20 �C until analysis. Plasma levels of picolinic acid varied

widely between patients and within patients during the study

(Table 3). The maximum level detected was 113 ng mL)1. No

correlation between plasma picolinic acid levels and the occur-

rence of adverse events was evident.

Discussion

This study was designed to assess the clinical efficacy and safety

of 10% picolinic acid gel, a novel immunomodulatory therapy

targeting ZFPs, in the treatment of mild to moderate acne

vulgaris. Results from this open-label study show picolinic acid

0.00

10.00

20.00

30.00

40.00

50.00

60.00

70.00

80.00

90.00

Week

Les

ion

Co

un

t

Total Lesions 85.33 83.73 59.40 47.40 35.67Inflammatory Lesions 30.40 30.00 20.27 15.20 13.33Noninflammatory Lesions 54.93 53.73 39.13 32.20 22.33

0 1 4 8 12

Fig 1. Mean lesion counts using per protocol

analysis.

Table 2 Adverse events

Patient Adverse event Week

Relation

to studydrug

2 Burning at application site 1 Probable

6 Tingling at application site 1 Probable15 Burning at application site 1 Probable

12 Sinusitis 5 Possible15 Allergic rhinitis 12 Possible

9 Vaginal yeast infection 8 Unlikely11 Upper respiratory viral infection 9 Unlikely



550 Picolinic acid gel for acne vulgaris, M.P. Heffernan et al.

to be highly efficacious with 100% of the subjects who com-

pleted the study per protocol achieving a reduction in total and

noninflammatory lesion count and 93Æ3% of subjects achieving

a reduction in inflammatory lesion count. A clinically signifi-

cant reduction in mean total lesion count, mean inflammatory

lesion count and mean noninflammatory lesion count was seen

using both intent-to-treat and per protocol analysis. In add-

ition, nearly half of the patients (47%) who completed the

study were almost clear or had marked improvement at the

final visit using the Physician Global Improvement score.

Picolinic acid exists in vivo as a metabolite of the essential

amino acid tryptophan. It is produced in approximately 25–

50-mg quantities daily, assuming normal dietary intake. Picol-

inic acid chelates transition metal ions (i.e. Zn2+) and is

involved in the absorption and transport of transition metal

ions.

Picolinic acid has been demonstrated to possess immune-

modulating properties. Zinc binding within ZFPs is perturbed

in the presence of picolinic acid. This leads to an alteration in

chemokine expression independently and in the presence of

other chemokines such as interferon-c.21,22 Picolinic acid has

also been shown to have antiviral properties against herpes

simplex virus type 1 and type 2. ZFPs are involved in viral

replication, viral packaging and normal cell homeostatic func-

tions. An Investigational New Drug application has recently

been submitted by the manufacturer for picolinic acid cream

in the treatment of herpes labialis.

A phase I cumulative irritant study was recently completed by

Novactyl, Inc. assessing 10% picolinic acid cream in healthy vol-

unteers. Patients applied 200-mg quantities daily under a patch

for 21 consecutive days. The result of this irritant study was that

10% PCL-016 cream was clinically comparable with BlistexTM

(Blistex Inc., Oak Brook, IL, U.S.A.).

There was only one adverse effect, burning at the applica-

tion site, which led to patient discontinuation from this study.

This was judged to be mild in intensity and probably related

to study drug by the investigator. Another patient experienced

burning after application of study drug that resolved within

1 week of the study. One patient described facial tingling pre-

sent only on day 1 of study drug application. The remaining

adverse effects discussed earlier were mild and judged as poss-

ibly related or unlikely to be related to study drug. There were

no deaths or serious adverse events noted in this study. There

were no significant laboratory abnormalities during the study.

It is estimated by the manufacturer that 10–20 mg of PCL-

016 is delivered to the skin with each application of 10%

picolinic acid gel. There was both intrapatient and interpatient

variability in plasma levels of picolinic acid during the study.

The maximum plasma level of picolinic acid detected was

113 ng mL)1. No correlation between plasma picolinic acid

levels and the occurrence of adverse events was evident. The

minimum plasma picolinic acid level achievable before adverse

effects are expected to occur, if a relation exists, remains to be

elucidated.

In conclusion, the results of this study suggest that 10% pi-

colinic acid gel applied twice daily may be safe and effective

in the treatment of mild to moderate acne vulgaris. This is the

first reported use of NV-02 in humans, and the first reported

use of a zinc finger drug in the treatment of acne vulgaris.

Future randomized, double-blind, multicentre trials will be

necessary to confirm and better define these findings.

References

1 Kraning KK, Odland GF, eds. Prevalence, morbidity, and cost ofdermatological diseases. J Invest Dermatol 1979; 73:395–513.

2 Leyden JJ. New understandings of the pathogenesis of acne. J AmAcad Dermatol 1995; 32:S15–25.

3 Cunliffe WJ, Gould DJ. Prevalence of facial acne vulgaris in lateadolescence and in adults. BMJ 1979; i:1109–10.

4 Jowett S, Ryan T. Skin disease and handicap: analysis of the impactof skin conditions. Soc Sci Med 1985; 20:425–9.

5 Shuster S, Fisher GH, Harris E, Binnell D. The effect of skin diseaseon self-image. Br J Dermatol 1978; 99 (Suppl. 16):18–19.

6 Plewig G, Kligman AM. Acne and Rosacea, 3rd edn. New York: Sprin-ger-Verlag, 2000.

7 Kellet SC, Gawkrodger DJ. The psychological and emotional impactof acne and the effect of treatment with isotretinoin. Br J Dermatol

1999; 140:273–82.8 Koo J. The psychosocial impact of acne: patients’ perceptions. J Am

Acad Dermatol 1995; 32:S25–30.9 Wu SF, Kinder BN, Trunnell N, Fulton JE. Role of anxiety and

anger in acne patients: a relationship with the severity of the dis-order. J Am Acad Dermatol 1988; 18:325–33.

10 Aktan S, Ozmen E, Sanli B. Anxiety, depression, and nature of acnevulgaris in adolescents. Int J Dermatol 2000; 39:354–7.

11 Cunliffe WJ. Acne and unemployment. Br J Dermatol 1986; 115:386(Letter).

Table 3 Picolinic acid drug levels (ng mL)1)

Patient Week 0 Week 1 Week 4 Week 8 Week 12

1 ND 13Æ6 17Æ6 11Æ3 BQL

2 BQL BQL3 ND 17Æ9 54Æ5 BQL 47Æ04 BQL 52Æ3 90Æ5 BQL BQL5 ND BQL BQL BQL BQL

6 BQL 16Æ6 24Æ3 BQL BQL7 ND 72Æ6 113Æ0 44Æ9 BQL

8 BQL BQL BQL BQL BQL9 BQL 15Æ5 BQL BQL BQL

10 BQL 96Æ0 91Æ211 BQL 21Æ2 24Æ3 BQL BQL

12 BQL 17Æ3 31Æ1 56Æ9 BQL13 BQL

14 BQL BQL BQL BQL BQL15 BQL BQL BQL BQL BQL

16 BQL 31Æ917 BQL 22Æ3 BQL BQL 7Æ9618 BQL BQL BQL BQL BQL

19 BQL BQL BQL20 BQL 39Æ8 57Æ4 BQL 29Æ3

ND, none detected on assay; BQL, below quantitative limit(< 10 ng mL)1).



Picolinic acid gel for acne vulgaris, M.P. Heffernan et al. 551

12 Mallon E, Newton JN, Klassen A et al. The quality of life in acne: acomparison with general medical conditions using generic ques-

tionnaires. Br J Dermatol 1999; 140:672–6.13 Gollnick HPM, Zouboulis CC, Akamatsu H et al. Pathogenesis and

pathogenesis-related treatment of acne. J Dermatol 1991; 18:489–99.14 Cunliffe WJ, Gollnick H. Acne: Diagnosis and Management. London: Mar-

tin Dunitz, 2001.15 Cunliffe WJ, Holland DB, Clark SM, Stables GI. Comedogenesis:

some new aetiological, clinical and therapeutic strategies. Br JDermatol 2000; 142:1084–91.

16 Cunliffe WJ, Simpson NB. Disorders of the sebaceous gland. In:

Textbook of Dermatology (Champion RH, Burton JL, Burns DA,Breathnach SM, eds), 6th edn. Oxford: Blackwell Science, 1998;

1927–84.17 Burton JL, Shuster S. The relationship between seborrhoea and acne

vulgaris. Br J Dermatol 1971; 84:600–1.

18 Leyden JJ, McGinley KJ, Mills OH, Kligman AM. Propionibacteriumlevels in patients with and without acne. J Invest Dermatol 1975;

65:382–4.19 Webster GF. Inflammation in acne vulgaris. J Am Acad Dermatol

1995; 33:247–53.20 Layton AM, Henderson CA, Cunliffe WJ. A clinical evaluation of

acne scarring and its incidence. Clin Exp Dermatol 1994; 19:303–8.21 Bosco MC, Rapisarda A, Massazza S et al. The tryptophan catabolite

picolinic acid selectively induces the chemokines macrophageinflammatory protein-1 alpha and -1 beta in macrophages. J Immunol

2000; 164:3283–91.

22 Melillo G, Cox GW, Biragyn A et al. Regulation of nitric-oxidesynthase mRNA expression by interferon-gamma and picolinic

acid. J Biol Chem 1994; 269:8128–33.



552 Picolinic acid gel for acne vulgaris, M.P. Heffernan et al.

THERAPEUTICS DOI 10.1111/j .1365-2133.2006.07647.x

Treatment of refractory erosive oral lichen planus withextracorporeal photochemotherapy: 12 casesA.D. Guyot, D. Farhi,* S. Ingen-Housz-Oro,* A. Bussel,� N. Parquet,� C. Rabian,� H. Bachelez* and C. Frances

Service de Dermatologie-Allergologie, AP-HP, Hopital Tenon, Universite Paris VI, 4 rue de la Chine, 75020 Paris, France

*Service de Dermatologie 1, �Service d’Hemapherese Therapeutique and �Service d’Immunologie Biologique, AP-HP, Hopital Saint Louis, Universite Paris VII,

Paris, France

CorrespondenceCamille Frances.



Key wordsextracorporeal photochemotherapy, oral erosive

lichen planus


Summary

Background Case reports have suggested that extracorporeal photochemotherapy(ECP) might be beneficial for the treatment of erosive oral lichen planus (OLP)recalcitrant to conventional immunosuppressive therapies.Objectives To evaluate over a long-term period the clinical efficacy and toxicity ofECP in a series of patients with refractory OLP, and to monitor peripheral bloodlymphocyte subset counts under treatment.Methods Twelve patients with refractory OLP underwent a standardized protocol ofECP. Sessions were performed twice weekly for 3 weeks, and then the treatmentschedule was adapted according to clinical benefit. The disease severity was eval-uated monthly on a clinical basis. Complete remission was defined as the absenceof any erosion and partial remission as a decrease of at least 50% of erosion sur-face. Blood cell counts with CD4+ and CD8+ lymphocyte subsets were evaluatedevery 3 months.Results All patients showed a decrease of the erosive surface; nine (75%) achieveda complete remission and three (25%) a partial remission. Seven of the eightpatients followed for more than 3 years had recurrences of erosions when ECPsessions became less frequent or were stopped. After resumption of an initiallyaccelerated regimen of ECP, all again showed partial or complete remission.Blood lymphocyte counts decreased during treatment, without statistically signifi-cant changes in CD4+/CD8+ ratio, and increased during relapse.Conclusions ECP is an effective alternative therapy in erosive OLP showing resistanceto classical treatments. The decrease in blood lymphocyte counts appears to paral-lel the clinical improvement under treatment.

Erosive oral lichen planus (OLP) is a chronic, inflammatory

disease which may be very painful with a great impact on

quality of life.1 Spontaneous remissions are rare, observed in

2Æ8–6Æ5% of cases.2–4 The first-line therapy in OLP is high-

potency topical corticosteroids.4 For patients refractory to this

treatment, topical ciclosporin or tacrolimus may represent a

valuable second-line therapy.4,5 Long-term systemic steroids

(> 0Æ5 mg kg)1 daily) are usually effective but are associated

with a well-known spectrum of side-effects and dramatic

relapses following steroid withdrawal. Local or systemic retin-

oids are usually poorly tolerated.6,7 Systemic immunosuppres-

sive agents such as azathioprine or methotrexate are frequently

unhelpful.6 Given the dramatic improvement of OLP lesions

observed in a patient treated with extracorporeal photochemo-

therapy (ECP) for an associated cutaneous T-cell lymphoma,

and preliminary short-term results obtained in a limited series

of patients,8 we initiated a long-term follow-up study to

investigate the clinical efficacy and toxicity of ECP in 12

patients with refractory erosive OLP. We also attempted to

investigate the correlations between clinical variables and the

course of blood lymphocyte counts.


Twelve patients entered the study. All suffered from histo-

logically proven, erosive OLP with an erosive surface

> 1 cm2. They were 10 women and two men (Table 1).

Their mean age at the onset of ECP was 60Æ6 years (range 40–

74). Mean duration of the disease was 7Æ5 years (range 1–

15). Three had a history of autoimmune disorders (pernicious

anaemia, type I diabetes mellitus, and primary biliary cirrhosis

with Gougerot–Sjogren syndrome, respectively). None had



chronic hepatitis C. All were refractory to local corticosteroids.

Other ineffective treatments used prior to ECP are listed in

Table 1. Erosive lesions mostly involved the tongue, cheeks

and lips. Amalgam fillings were not replaced, as the four

patients with only oral involvement had diffuse erosive OLP,

and only one or two amalgam fillings. No topographic rela-

tionship could be established between erosive OLP and amal-

gam fillings. Others had widespread disease with genital

involvement. Indeed, vulvar erosions were observed in eight

cases, and cutaneous lesions in four.

ECP was performed twice a week for the initial 3 weeks,

and then the frequency of sessions was progressively reduced

according to clinical evolution, to once every 4 months.

Finally, treatment was stopped when complete, stable remis-

sion was reached for at least 6 months. All systemic treatments

were withdrawn or were stable at least 6 months before start-

ing ECP.

At each session, mononuclear cells were extracted from two

blood masses (80 mL kg)1) using a blood cell separator

(Spectra; Gambro BCT, Denver, CO, U.S.A.). Soluble 8-meth-

oxypsoralen was added ex vivo to the cytapheresis product

(200 ng mL)1). Cells were then irradiated with ultraviolet

(UV) A at 3 J cm)2 with an UV irradiator (Vilbert-Lourmat,

Marne-la-Vallee, France), and reinfused to the patients.

Clinical follow up was performed monthly for 6 months,

then every 3 months, by the same investigator. The erosive

surface areas were estimated by measurement of the diameter

of the largest erosions. Clinical efficacy was evaluated on the

basis of the reduction in erosive surface: partial remission was

defined as a 50% or more decrease of the whole erosive sur-

face, and complete remission as a complete healing of all

lesions. Other clinical disease factors, such as erythema or

white asymptomatic lesions, were not taken into consideration.

Blood cell counts and immunophenotyping analysis of lym-

phocyte subsets were performed at each ECP session. Absolute

CD4+ and CD8+ cells were quantified in fresh blood samples

by quadruple fluorescence with beads, using a FACSCalibur

analyser (Becton Dickinson, Franklin Lakes, NJ, U.S.A.). Varia-

tions observed in absolute lymphocyte count, CD4+ and

CD8+ populations and CD4+/CD8+ ratio were statistically

analysed using Student’s t-test.

Results

Complete remission (Fig. 1) was achieved in nine patients

(75%), after a mean period from ECP onset of 11Æ5 months

(range 1–36) and a mean number of 21 sessions (range 6–

47). The other three patients achieved partial remission after a

mean period of 2Æ6 months (range 1–5) and a mean number

of 11 sessions (range 6–16).

Table 1 Description of patients with erosive oral lichen planus

Patient Sex Age (years) Duration of LP (years) Genital/skin LP Previous systemic treatments Previous topical treatments

1 F 58 6 +/) CT –

2 F 60 14 +/+ CT, acitretin, AZA –3 F 47 3 +/+ – –

4 F 54 1 +/) – Ciclosporin, tretinoin5 F 74 12 +/+ CT, AZA Ciclosporin

6 F 62 15 )/) CT –7 F 40 2 )/) – Ciclosporin

8 F 60 5 +/+ – Ciclosporin9 F 68 7 +/) Tretinoin

10 F 70 5 +/) Tretinoin11 M 60 ? )/) CT Ciclosporin

12 M 74 5 )/) CT Tretinoin

LP, lichen planus; CT, corticosteroid; AZA, azathioprine.

(b)

(a)

Fig 1. (a) Erosive lichen planus of the lip before extracorporeal

photochemotherapy. (b) Complete remission with disappearance of

erosions.



554 Treatment of refractory erosive OLP with ECP, A.D. Guyot et al.

One patient did not have recurrence of oral erosions 4 years

after ECP withdrawal. The 11 other patients relapsed: three

relapsed during a session schedule of one per month or fewer,

and eight after ECP was stopped, at a mean of 8Æ3 months

(range 5–10) after the last session.

For eight patients who relapsed, an initial accelerated regi-

men of ECP was used again, with two sessions a week, and

treatment was then progressively reduced. Clinical improve-

ment was observed in all patients (Table 2). Five had com-

plete remission after 3–7 months of treatment (mean number

of sessions: 12Æ2). Three had only partial remission after 3, 8

and 12 months of treatment (mean number of sessions: 18).

Evolution of vulvar erosions ran in parallel with that of oral

ulcerations with earlier healing. The tolerance to the treatment

was excellent. No patient presented transient hypotension,

low-grade pyrexia or an infectious complication of induced

lymphopenia. ECP had to be discontinued in one patient

because of difficult venous access.

The mean blood lymphocyte count was 1Æ713 · 109 L)1

before the onset of ECP. It progressively decreased during

treatment to reach 0Æ982 · 109 L)1 before the last session of

ECP, when clinical response was obtained. This decrease was

statistically significant (P < 0Æ01). At the time of relapse, the

mean lymphocyte count had increased to 1Æ593 · 109 L)1

(P < 0Æ05). The decrease of lymphocyte subsets at the time of

remission was predominant for CD4+ T lymphocytes (51%,

from 0Æ750 to 0Æ368 · 109 L)1, P < 0Æ01) as well as CD8+

T lymphocytes (42%, from 0Æ499 to 0Æ292 · 109 L)1,

P < 0Æ01). The variations affecting the CD4+/CD8+ ratio

were not statistically significant.

Discussion

Therapeutic management of chronic erosive lichen planus may

be problematic when current treatments are ineffective. Sup-

pression of all irritant factors is required. Removal of amalgam

dental fillings is usually recommended when oral lesions are

close to amalgam, especially in cases of a positive patch test

reaction to its constituents.9,10 In this series, no triggering or

causative factor was suggested by clinical data.

ECP has been proposed as an immunomodulatory strategy,

allowing clinical benefit in cutaneous T-cell lymphoma, mostly

Sezary syndrome, and in nonmalignant disorders such as car-

diac transplant rejection and graft-versus-host disease (GvHD).

Indeed, the interest in ECP for severe recalcitrant forms of

lichen planus has been raised in view of clinical benefits in

patients with GvHD, and of the clinical and physiopathological

similarities between lichen planus and lichenoid GvHD. Fur-

thermore, ECP appears relatively safe, without evidence for an

increased risk of opportunistic infections.11 However, ECP is

expensive and time consuming. Patients are immobilized dur-

ing half a day for each session. Adequate venous access is

required.

In the present, long-term follow-up study we confirm the

safety of ECP, as none of our patients showed any infectious

complication during study. We also show evidence for striking

efficacy, as all patients experienced a complete or partial

remission under treatment. However, the effect of ECP was

only palliative, and the remission was dependent on a persist-

ent treatment with spaced sessions. Only one patient showed

stable remission for several years after ECP discontinuation.

Others, who experienced recurrence of erosions after ECP

withdrawal, requested further treatment with ECP as they con-

sidered this to be the most effective and the most well-toler-

ated treatment. ECP was effective again on relapses. Similar

results were obtained in a recently reported case study per-

formed in four patients with erosive OLP.12

The clinical response was associated with a contraction of

peripheral blood T-lymphocyte subsets, while B-lymphocyte

and natural killer cell counts remained relatively stable. Even

more interestingly, a peripheral blood T-cell expansion was

observed prior to the onset of relapse, suggesting that homeo-

static proliferation in the T-cell compartment might be a key

event in the occurrence of relapse. However, these results

should be interpreted cautiously, as the number of patients

was insufficient to provide an accurate comparative analysis of

lymphocyte counts in patients with partial remission vs. those

with complete remission. Further studies are currently ongo-

ing, in order to determine more accurately the lymphocyte

phenotype affected by ECP in OLP.

The mechanisms underlying the efficacy of ECP in vivo are

not fully elucidated. In lichen planus, immunological reactions

mediated by T cells are probably directed against one or sev-

eral as yet unknown foreign or self-modified antigens.13 Both

activated CD4+ and CD8+ T lymphocytes are present within

lichen lesions. CD4+ cells may activate cytotoxic CD8+ cells

that cause epidermal cell damage. The predominant secretion

of interferon-c and tumour necrosis factor-a suggested a type

1 cytokine pattern.14,15

As in GvHD, apoptosis of circulating lymphocytes is not the

sole mechanism of action of ECP. In animal models of

contact hypersensitivity, these apoptotic cells may induce

Table 2 Response of oral erosions to extracorporealphotochemotherapy (ECP)

Patient

Response

to ECP

Durationof remission

(months) Relapse

Responseto ECP

resumption

1 CR 5 Yes CR2 CR 33 Yes CR

3 PR 10 Yes PR4 CR 30 Yes –

5 CR 38 Yes PR6 PR 6 Yes –

7 CR 53 No –8 CR 34 Yes PR

9 PR 18 Yes –10 CR 19 Yes CR

11 CR 15 Yes CR12 CR 2 Yes CR

CR, complete remission; PR, partial remission.



Treatment of refractory erosive OLP with ECP, A.D. Guyot et al. 555

antigen-specific regulatory T cells.16 In GvHD treated with

ECP, monocytes, which appear to be resistant to apoptotic

effects induced by ECP, have an enhanced synthesis of immuno-

suppressive cytokines such as interleukin (IL)-10 and IL-1

receptor antagonist; they differentiated into immature CD83+

dendritic cells actively phagocytosing apoptotic cells.17 ECP

induces an increase in the plasmacytoid DC2 population with

subsequent increased production of Th2 cytokine factors.17,18

In conclusion, ECP may be effective as a salvage therapeutic

option for erosive lichen planus refractory to current treat-

ments. Further prospective studies are needed in order to

determine more accurately the risk/benefit ratio, and the

mechanisms underlying tolerance induction in this context.

References

1 Lundqvist EN, Wahlin YB, Bergdahl M, Bergdahl J. Psychologicalhealth in patients with genital and oral erosive lichen planus. J Eur

Acad Dermatol Venereol 2006; 20:661–6.2 Silverman S, Gorsky M, Lozada-Nur F. A prospective follow-up

study of 570 patients with oral lichen planus: persistence, remis-sion, and malignant association. Oral Surg Oral Med Oral Pathol 1985;

60:30–4.3 Silverman S, Gorsky M, Lozada-Nur F et al. A prospective study of

findings and management in 214 patients with oral lichen planus.

Oral Surg Oral Med Oral Pathol 1991; 72:665–70.4 Conrotto D, Carbone M, Carrozzo M et al. Ciclosporin vs. clobetasol

in the topical management of atrophic and erosive oral lichen pla-nus: a double blind, randomized controlled trial. Br J Dermatol

2006; 154:139–45.5 Byrd JA, Davis MDP, Bruce AJ et al. Response of oral lichen planus to

topical tacrolimus in 37 patients. Arch Dermatol 2004; 140:1508–12.6 Cribier B, Frances C, Chosidow O. Treatment of lichen planus: an

evidence-based medicine analysis of efficacy. Arch Dermatol 1998;134:1521–30.

7 Hersle K, Mobacken H, Sloberg K, Thilander H. Severe oral lichenplanus: treatment with an aromatic retinoid (etretinate). Br J Derma-

tol 1982; 106:77–80.8 Becherel PA, Bussel A, Chosidow O et al. Extracorporeal photo-

chemotherapy for chronic erosive lichen planus. Lancet 1998;351:805.

9 Laeijendecker R, Dekker SK, Burger PM et al. Oral lichen planusand allergy to dental amalgam restorations. Arch Dermatol 2004;

140:1434–8.10 Rogers RS III, Bruce AJ. Lichenoid contact dermatitis: is inorganic

mercury the culprit? Arch Dermatol 2004; 140:1524–5.

11 McKenna KE, Whittaker S, Rhodes LE et al. Evidence-based practiceof photopheresis 1987–2001: a report of a workshop of the British

Photodermatology Group and the U.K. Skin Lymphoma Group. Br JDermatol 2006; 154:7–20.

12 Kunte C, Erlenkeuser-Uebelhoer I, Michelsen S et al. Behandlungdes therapieresistenten erosiven, oralen Lichen ruber mit extra-

korporaler Photopherese (ECP). J Dtsch Dermatol Ges 2005; 3:889–94.

13 Kastelan M, Prpic Massari L, Gruber F et al. The role of perforin-mediated apoptosis in lichen planus lesions. Arch Dermatol Res 2004;

296:226–30.14 Sklavounou A, Chrysomali E, Scorilas A et al. TNF-alpha expression

and apoptosis regulating proteins in oral lichen planus: a compara-tive immunohistochemical evaluation. J Oral Pathol Med 2000;

29:370–5.15 Khan A, Farah CS, Savage NW et al. Th1 cytokines in oral lichen

planus. J Oral Pathol Med 2003; 32:77–83.16 Maeda A, Schwarz A, Kernebeck K et al. Intravenous infusion of

syngeneic apoptotic cells by photopheresis induces antigen-specificregulatory T cells. J Immunol 2005; 174:5968–76.

17 Fimiani M, Di Renzo M, Rubegni P. Mechanism of action of extra-corporeal photochemotherapy in chronic graft-versus-host disease.

Br J Dermatol 2004; 150:1055–60.18 Heshmati F. Mechanisms of action of extracorporeal photochemo-

therapy. Transfus Apheresis Sci 2003; 29:61–70.



556 Treatment of refractory erosive OLP with ECP, A.D. Guyot et al.

CONCISE COMMUNICATION DOI 10.1111/j .1365-2133.2006.07705.x

Reporting of study design in titles and abstracts of articlespublished in clinically oriented dermatology journalsR. Ubriani, N. Smith* and K.A. Katz

Department of Dermatology, University of Pennsylvania School of Medicine, Philadelphia, PA, U.S.A.

*Department of Medicine, Christiana Hospital, Newark, DE, U.S.A.

CorrespondenceKenneth Katz, 1439B Corccrun St NW,

Washington, DC 20009, U.S.A.



Key wordscritical appraisal, epidemiology, evidence-based

medicine, medical literature, study design


Summary

Background Dermatologists may have difficulty in identifying the types of study designused in published articles, hindering their ability to appraise the literature critically.Objectives To assess the frequency with which titles or abstracts of articles pub-lished in clinically oriented dermatology journals reported the type of studydesign using standard key words, including ‘randomized control trial’, ‘nonrand-omized control trial’, ‘double-blind’, ‘placebo control’, ‘crossover trial’, ‘before–after trial’, ‘gold standard’, ‘blinded or masked comparison’, ‘cohort’, ‘inceptioncohort’, ‘validation cohort’, ‘validation sample’, ‘survey’, ‘case series’, ‘cost-effectiveness analysis’, ‘cost-benefit analysis’, ‘cost-utility analysis’, ‘cross-sectional study’ and ‘case–control’.Methods A cross-sectional study analysed articles published between December2004 and November 2005 in the ‘Epidemiology and Health Services Research’and ‘Therapeutics’ sections of the British Journal of Dermatology (BJD), in the ‘Studies’section of the Archives of Dermatology (Arch Dermatol) and in the ‘Reports’ section ofthe Journal of the American Academy of Dermatology (JAAD).Results In the BJD, 15 of 37 articles (40Æ5%, 95% confidence interval, CI 24Æ8–57Æ9%) included at least one standard key word in the title or abstract, comparedwith 43 of 87 articles (49Æ4%, 95% CI 38Æ5–60Æ4%) in the Arch Dermatol and 19of 93 articles (20Æ4%, 95% CI 12Æ8–30Æ1%) in the JAAD (P < 0Æ001).Conclusions Most articles in the three journals did not report the study design usedin the title or abstract. A consistent and clear indication of the design used instudies may better enable editors, reviewers and readers to assess critically articlespublished in clinically oriented dermatology journals.

Evidence-based medicine has increasingly attracted the interest

of dermatologists.1–8 One critical step in appraising evidence

is identifying the type of study presented in a report.2,9–11

Doing so allows the reader to determine where the report falls

in a hierarchy of evidence that gives more weight to random-

ized controlled trials, for example, than to single case

reports.3,12 It also allows the reader to assess how inherent

strengths and weaknesses of the study’s design might affect

the results and, ultimately, the conclusions presented in the

report.13,14

In our experience, identifying the type of study reported in

a paper is a major challenge for dermatologists in practice and

in training. We have also noticed that some journal

abstracts—the most frequently read and accessed parts of jour-

nal articles15—either do not report the type of study design

used, or report a design using nonstandard terminology. The

purpose of this study was to investigate the frequency with

which articles published in three clinically oriented dermatol-

ogy journals report study design using standard terminology.


We performed a cross-sectional study of articles published

between December 2004 and November 2005 in the ‘Epidemi-

ology and Health Services Research’ and ‘Therapeutics’ sections

of the British Journal of Dermatology (BJD), the ‘Studies’ section of

the Archives of Dermatology (Arch Dermatol) and the ‘Reports’ section

of the Journal of the American Academy of Dermatology (JAAD). K.A.K.

(a board-certified dermatologist with training in epidemiology

and statistics) and R.U. (a resident in dermatology) independ-

ently reviewed the abstract of each article and determined whe-

ther it contained at least one of the 19 study design-related key

words (Table 1) that were proposed in a 1987 report in the

Annals of Internal Medicine16 and are currently recommended for use



in the ‘Design’ section of a structured abstract by the Arch

Dermatol.17 The BJD and JAAD do not publish recommended

design-related key words in their author instructions. A key

word present anywhere in the title or abstract of an article was

considered acceptable. Disagreements were resolved by consen-

sus. We also performed an electronic search of all of the titles

and abstracts of the articles for one term—‘systematic review’—

not included in the list of recommended key words. A v2 test

comparing proportions of articles that included a key word in

the three journals (not including ‘systematic review’) was per-

formed using Stata 9.0 (Stata Corporation, College Station, TX,

U.S.A.).

Results

In the BJD, 15 of 37 articles (40Æ5%, 95% confidence interval,

CI 24Æ8–57Æ9%) included at least one standard key word in

the title or abstract, compared with 43 of 87 articles (49Æ4%,

95% CI 38Æ5–60Æ4%) in the Arch Dermatol and 19 of 93

abstracts (20Æ4%, 95% CI 12Æ8–30Æ1%) in the JAAD

(P < 0Æ001). Table 1 shows the frequency of use of each

standard key word in each journal. Of the 19 key words, 12

(63%), 10 (53%) and 13 (68%) were not used in any articles

in the BJD, the Arch Dermatol and the JAAD, respectively. The

term ‘systematic review’ appeared in two of 37 articles (5%)

from the BJD. This term did not appear in any of the articles

from the Arch Dermatol or the JAAD.

Discussion

Among articles from the BJD, the Arch Dermatol and the JAAD

included in this study, fewer than half included a standard

study design-related key word in the title or abstract. Articles

published in the Arch Dermatol had the highest percentage

(49%) of key word inclusion, perhaps reflecting the fact that

it is the only journal among the three to include a list of key

Table 1 Frequency of use of study design-related key words in titles or abstracts of

articles published between December 2004and November 2005 in the British Journal of

Dermatology, the Archives of Dermatology and theJournal of the American Academy of Dermatology,

classified by type of study design17Key word

Frequency of use in article titles and abstracts

British Journal

of Dermatology(37 articles)

Archives of

Dermatology(87 articles)

Journal of theAmerican Academy

of Dermatology(93 articles)

Intervention studies

Randomized control trial 8 8 4Nonrandomized control trial 0 0 0

Double-blind 7 6 4Placebo control 5 4 1

Crossover trial 0 0 1Before–after trial 0 1 0

Studies of screening and diagnostic testsGold standard 0 0 0

Blinded or masked comparison 0 0 0Studies of prognosis

Cohort 0 9 5

Inception cohort 0 0 0Validation cohort 0 0 0

Validation sample 0 0 0Descriptions of clinical features of medical disorders

Survey 3 9 7Case series 1 6 0

Studies with a formal economic evaluationCost-effectiveness analysis 0 0 0

Cost-benefit analysis 0 0 0Cost-utility analysis 0 0 0

Other studiesCross-sectional study 1 3 0

Case–control 1 1 0

Totala 26 47 22

aThe total number of key words is greater than the number of articles containing a key

word because some articles contained more than one key word in the title and abstract.The table does not include articles with the term ‘systematic review’ in the title or

abstract. ‘Systematic review’ was not in the list of recommended key words.17 Two of 37

articles (5%) of articles from the British Journal of Dermatology included this term; none fromthe Archives of Dermatology or the Journal of the American Academy of Dermatology did.



558 Study design reporting, R. Ubriani et al.

words16 in its ‘Instructions to Authors’.17 This finding also

parallels the results of a study of the quality of dermatology

abstracts that were published in 2000.18 In that study,

abstracts from the Arch Dermatol received higher quality scores

than those from the JAAD on all characteristics assessed,

including reporting of study design, and higher quality scores

than those from the BJD on seven of eight characteristics

assessed, including study design.18 The study also assessed the

use of study design-related key words in article abstracts (but

not in titles), although it was one of three factors included in

an overall assessment of study design; frequency of key word

use in each journal was not reported.18

There are several ways in which the frequency of reporting

of type of study design in journal abstracts could be increased.

Firstly, journals could raise awareness by emphasizing the

importance of reporting study design in published ‘Instruc-

tions to Authors,’ as the Arch Dermatol already does. Secondly,

journals could review and edit abstracts more strictly to ensure

that study design type is reported. JAMA adopted such an

approach in 1999,15 after preliminary results of a study dem-

onstrated deficiencies in JAMA abstracts.15,19 Within 1 year,

the quality of abstracts in JAMA had significantly improved.20

Thirdly, journals could require authors to select a type of

study design from a pull-down menu of standard key

words16,17 on an electronic submission website, just as article

type might be selected. This selection could then either be

automatically entered into an abstract or compared with the

study design reported in the submitted abstract. Finally, jour-

nals could introduce a separate element of the structured

abstract, called ‘Type of study design used’, with text limited

to standard key words.16,17 With any of these approaches,

providing definitions of key words (or links to websites that

include this information)16 would also be useful. These

approaches might also help eliminate the use of ambiguous

and confusing nonstandard words and phrases that sometimes

appear in descriptions of study design.

This study has limitations. Firstly, we examined only 12

issues from each of three major dermatology journals. Sec-

ondly, the ‘Methods’ sections of a manuscript might include

information about type of study design not found in the

abstract and, therefore, not be captured in this study. Thirdly,

we did not assess the accuracy of study design reporting.

However, given the importance of the abstract to readers and

browsers of the medical literature,15 lack of information about

study design in the title or abstract itself is still concerning.

Finally, although we did search for one design type—system-

atic review—that was not included in the list of recommen-

ded key words, the list of search terms is not exhaustive and

might not include all relevant study designs.16,17 Future lists

of recommended key words might be more effective if they

are more inclusive of research typically conducted by derma-

tologists.

Evidence-based medicine is increasingly important to der-

matologists. Improving the reporting of study design types in

dermatology journals might facilitate dermatologists’ ability to

make a critical appraisal of articles in the literature that are

relevant to care of their patients.

Acknowledgments

Dr. Katz’s fellowship was funded by a Kirschstein National

Research Service Award from the National Institutes of

Health.

References

1 Abeni D, Girardelli CR, Masini C et al. What proportion of derma-tological patients receive evidence-based treatment? Arch Dermatol

2001; 137:771–6.2 Barzilai DA, Freiman A, Dellavalle RP et al. Dermatoepidemiology.

J Am Acad Dermatol 2005; 52:559–73.3 Bigby M. Evidence-based medicine in a nutshell. A guide to find-

ing and using the best evidence in caring for patients. Arch Dermatol1998; 134:1609–18.

4 Bigby M, Williams H. Appraising systematic reviews and meta-analyses. Arch Dermatol 2003; 139:795–8.

5 Dellavalle RP, Stegner DL, Deas AM et al. Assessing evidence-baseddermatology and evidence-based internal medicine curricula in US

residency training programs: a national survey. Arch Dermatol 2003;139:369–72.

6 Straus SE, Sackett DL. Bringing evidence to the clinic. Arch Dermatol1998; 134:1519–20.

7 Williams HC. Applying trial evidence back to the patient. ArchDermatol 2003; 139:1195–200.

8 Williams HC, Strachan DP. The Challenge of Dermato-epidemiology. BocaRaton, FL: CRC Press, 1997.

9 Guyatt G, Rennie D (eds). American Medical Association. Users’ Guides tothe Medical Literature: a Manual for Evidence-Based Clinical Practice. Chicago,

IL: AMA Press, 2002.10 Sackett DL. Evidence-Based Medicine: How to Practice and Teach EBM, 2nd

edn. Edinburgh: Churchill Livingstone, 2000.11 Sackett DL, Rosenberg WM, Gray JA et al. Evidence based medicine:

what it is and what it isn’t. BMJ 1996; 312:71–2.12 Guyatt GH, Haynes RB, Jaeschke RZ et al. Users’ Guides to the

Medical Literature: XXV. Evidence-based medicine: principles forapplying the Users’ Guides to patient care. Evidence-Based Medi-

cine Working Group. JAMA 2000; 284:1290–6.13 Greenhalgh T. How to read a paper. Getting your bearings (deci-

ding what the paper is about). BMJ 1997; 315:243–6.

14 Hennekens CH, Buring JE, Mayrent SL. Epidemiology in Medicine, 1stedn. Boston, MA: Little, Brown, 1987.

15 Winker MA. The need for concrete improvement in abstractquality. JAMA 1999; 281:1129–30.

16 Ad Hoc Working Group for Critical Appraisal of the MedicalLiterature. A proposal for more informative abstracts of clinical

articles. Ann Intern Med 1987; 106:598–604.17 Archives of Dermatology (homepage on the Internet). Chicago, IL:

AMA, c2006. Instructions for Authors (about 20 screens). Availablefrom: http://archderm.ama-assn.org/misc/ifora.dtl (last accessed

22 February 2006).18 Dupuy A, Khosrotehrani K, Lebbe C et al. Quality of abstracts in 3

clinical dermatology journals. Arch Dermatol 2003; 139:589–93.19 Pitkin RM, Branagan MA, Burmeister LF. Accuracy of data in

abstracts of published research articles. JAMA 1999; 281:1110–11.20 Pitkin RM, Branagan MA, Burmeister LF. Effectiveness of a journal

intervention to improve abstract quality. JAMA 2000; 283:481.



Study design reporting, R. Ubriani et al. 559

CASE REPORT DOI 10.1111/j .1365-2133.2006.07599.x

Solitary plaque mycosis fungoides on the penis respondingto topical imiquimod therapyL.Y.T Chiam and Y.C. Chan

National Skin Centre, 1 Mandalay Road, Singapore 308205

CorrespondenceYuin-Chew Chan.



Key wordsmycosis fungoides, penis, topical imiquimod


Summary

Mycosis fungoides (MF) presenting in the genitalia is rare. We report a case oflong-standing penile MF in a young man. Commonly used treatments for limitedplaque MF include topical corticosteroids and ultraviolet light therapy. There area few anecdotal reports on the use of topical imiquimod in MF. Our patientresponded well and remained in complete remission after treatment with topicalimiquimod.

Cutaneous T-cell lymphoma (CTCL) is a rare disease with an

incidence of approximately 0Æ36 per 100 000 person-years in

the United States.1 Mycosis fungoides (MF) is the most com-

mon variant of CTCL. The lesions of MF have a predilection

for nonexposed areas. There have been reports of unilesional

MF affecting the trunk, upper extremities, inguinal region,

foot and back.2

Case report

A healthy 32-year-old Bangladeshi man was referred to our

centre by his primary health care physician for a chronic pink

plaque on his penis. It had been present for 15 years, but in-

creased growth and pain over the past year had prompted him

to seek medical attention.

Physical examination revealed a 19 · 11 mm well-demarca-

ted psoriasiform scaly pink plaque on the left glans penis

(Fig. 1). There were no other skin lesions. His testes were

normal in size. He had no enlarged lymph nodes or hepato-

splenomegaly.

Histological examination revealed laminated hyperkeratosis

in the stratum corneum, with irregular epidermal hyperplasia

(Fig. 2). The basal keratinocytes had an infiltrate of atypical

lymphocytes. The nuclei were hyperchromatic with halos.

A Pautrier’s microabscess was seen. This was consistent with

MF.

His full blood count, liver and renal function tests and chest

radiograph were normal. Peripheral blood film for Sezary cells

was negative.

He was treated with clobetasol propionate ointment for

6 weeks with no improvement. Treatment was changed to Fig. 1 Psoriasiform plaque on the glans penis.


560 Journal compilation � 2006 British Association of Dermatologists • British Journal of Dermatology 2007 156, pp560–562

imiquimod 5% cream and he was instructed to apply this on

the plaque once every 2 days. Our patient experienced local

pain during the first week of treatment. He also developed

local skin erosion in the third month of therapy, necessitating

temporary withdrawal of the topical application for a few

days. After 4Æ5 months of topical imiquimod, he was clinically

in complete remission and remained so on follow-up

6 months later.

Discussion

The differential diagnoses of a chronic nonhealing plaque on

the glans penis include psoriasis, lichen sclerosus, Bowen dis-

ease, squamous cell carcinoma, extramammary Paget disease

and MF. Differentiating between MF and lichen sclerosus can

be difficult clinically and histologically, especially if no Pautri-

er’s microabscess is seen. In these cases, T-cell rearrangement

studies may be necessary to support the diagnosis of MF.

MF involving only the genitalia is rare. A single report of

MF involving the vulva has been described.3

Long-term prognosis and survival outcome of early MF is

excellent. MF patients with clinical stage IA disease have a

favourable long-term outcome in which life expectancy is

similar to matched control populations.4

Treatment of MF is indicated to reduce symptoms, improve

clinical appearance, prevent complications and prevent pro-

gression of disease. Treatment for MF includes topical and

systemic agents, which can be administered alone or in

combination. Various topical therapies for limited plaque MF

include steroids,5 psolaren plus ultraviolet (UV) A,6 narrow-

band UVB and nitrogen mustard. Systemic therapies include

interferon, retinoids, denileukin diftitox (licensed in the

U.S.A.) and chemotherapy.

Early MF can be treated with excellent outcomes with non-

aggressive topical treatments. Systemic chemotherapy such as

CHOP has not been shown to improve survival in these

patients.

Imiquimod is a topical immunomodulator. It is a potent sti-

mulator of T-helper 1 cytokines, including interferon (IFN)-a,

IFN-c, and interleukin-12. It increases natural killer-cell activity

and also enhances antigen presentation by Langerhans cells.7

It has been approved by the Food and Drug Administration for

the treatment of genital warts and actinic keratosis. There is

growing evidence pointing to many other potential therapeutic

benefits of this drug in cutaneous neoplastic lesions.8

There have been anecdotal reports in the literature demon-

strating clinical use of imiquimod for the treatment of MF.9 In

a recent preliminary open-label study, topical imiquimod 5%

cream was associated with a 50% clinical response rate in

stage IA–IIB MF.10

Application site reactions are commonly seen in patients

undergoing treatment with topical imiquimod. The localized

skin reactions are mild to moderate, and the most frequently

reported application site reactions include pruritis, erythema,

oedema, discharge, papular rash, pain and occasionally hypo-

pigmentation or hyperpigmentation. Adverse effects were well

tolerated and largely limited to lesions that cleared with treat-

ment.

Conclusion

This is a rare case of penile MF in a young man, which was

steroid-resistant but responded well to topical imiquimod.

References

1 Weinstock MA, Gardstein B. Twenty-year trends in the reportedincidence of mycosis fungoides and associated mortality. Am J Public

Health 1999; 89:1240–4.2 Vang R, Medeiros LJ, Malpica A et al. Non-Hodgkin’s lymphoma

involving the vulva. Int J Gynecol Pathol 2000; 19:236–42.3 Kim YH, Jensen RA, Watanabe GL et al. Clinical stage 1A (limited

patch and plaque) mycosis fungoides. A long-term outcome analy-sis. Arch Dermatol 1996; 132:1309–13.

4 Zackheim HS, Kashani-Sabet M, Amin S. Topical corticosteroids formycosis fungoides: experience in 79 patients. Arch Dermatol 1998;

134:949–54.5 Querfeld C, Rosen ST, Kuzel TM et al. Long term follow-up of

patients with early-stage cutaneous T-cell lymphoma who achievedcomplete remission with psoralen plus UV-A monotherapy. Arch

Dermatol 2005; 141:305–11.

Fig 2. Pautrier’s microabscess and hyperchromatic nuclei with halos

(haematoxylin and eosin; original magnification · 40).



MF on the penis and topical imiquimod therapy, L.Y.T Chiam and Y.C. Chan 561

6 Kim YH, Martinez G, Varghese A, Hoppe RT. Topical nitrogenmustard in the management of mycosis fungoides: update of the

Stanford experience. Arch Dermatol 2003; 139:165–73.7 Sauder DN. Immunomodulatory and pharmacological properties of

imiquimod and S-27609. J Am Acad Dermatol 2000; 43 (1 Pt 2):S6–11.8 Navi D, Huntley A. Imiquimod 5 percent cream and the treatment

of cutaneous malignancy. Dermatol Online J 2004; 10:4.

9 Suchin KR, Junkins-Hopkins JM, Rook AH. Treatment of stage 1Acutaneous T-cell lymphoma with topical application of the

immune response modifier imiquimod. Arch Dermatol 2002;138:1137–9.

10 Deeths MJ, Chapman JT, Dellavalle RP et al. Treatment of patch andplaque stage mycosis fungoides with imiquimod 5 percent cream.




562 MF on the penis and topical imiquimod therapy, L.Y.T Chiam and Y.C. Chan


Paraneoplastic pemphigus without an underlying neoplasmG.T. Park, J.H. Lee,* S.J. Yun,* S.C. Lee* and J.B. Lee*

Department of Dermatology, Sungkyunkwan University School of Medicine, Samsung Medical Center, Seoul 135-710, Korea

*Department of Dermatology, Chonnam National University Medical School, 8-Hak-dong, Gwangju 501-757, Korea

CorrespondenceJee-Bum Lee.



Key wordsparaneoplastic pemphigus, underlying neoplasm


Summary

We describe a 52-year-old man with paraneoplastic pemphigus (PNP) withoutany evidence of an underlying neoplasm over an 8-year follow-up period. Hehad a chronic relapsing vesiculobullous eruption for approximately 7 years (fromApril 1998 to May 2005). Initially, scattered flaccid vesicles with crusts devel-oped on the face and trunk, which waxed and waned several times. Our patientwas diagnosed as having PNP based on immunopathological criteria for PNP, i.e.histopathological, immunoblotting and immunoprecipitation analyses. However,physical and laboratory examinations including serial blood tests with peripheralblood smear, whole-body positron emission tomography/computed tomographyand abdominal ultrasound were unable to detect any underlying neoplasm overan 8-year follow-up period.

Paraneoplastic pemphigus (PNP), which was first described by

Anhalt et al.1 in 1990, is an autoimmune blistering disease that

is characterized by the production of autoantibodies (aAbs)

mainly directed against the proteins of the plakin family.

Overlapping aAb specificity distributions have been reported

for PNP, pemphigus vulgaris (PV) and pemphigus foliaceus

(PF), suggesting a close relationship between the different

types of pemphigus.2–9 In addition, the diverse clinical mani-

festations of PNP and the lack of specificity of its histological

criteria imply that it is a pleomorphic disease entity.10,11 Inter-

estingly, despite the implication of various neoplasms in

almost all cases of PNP,1,10–16 only one case of PNP without

an underlying tumour has been reported.17 PNP was recently

reported to be caused by fludarabine, a synthetic nucleoside

analogue.18

We report a patient with PNP without any obvious underly-

ing tumour or illness as determined by physical and laboratory

examinations conducted over approximately 8 years of follow

up.

Case report

A 52-year-old man, who had been healthy prior to his first

visit to our hospital in April 1998, experienced scattered pru-

ritic vesicular eruptions and erosions on the face and trunk

(Fig. la). Six months before his first visit, he had noticed

newly developed skin eruptions of pruritic erythematous ma-

cules and tiny vesicles with crusts confined to the face, which

did not respond to conventional therapy for eczematoid

dermatitis at a private clinic. On his first visit to our hospital

he appeared healthy except for the skin lesions, which showed

scattered multiple vesiculobullae with crusts surrounded by

erythema on the face, upper chest and upper back, and ero-

sions on the lips. Mucosa of both conjuctiva and mouth (not

genital mucosa) were involved. He denied taking any medica-

tions and did not have any other disease, especially a neo-

plasm. Physical examination showed no organomegaly or

lymphadenopathy, and with the exception of an elevated IgE

level (407 IU mL)1, normal 0–100) all other laboratory tests

were negative or normal. These tests included full blood cell

count, peripheral blood smear, chest X-ray, electrolyte levels,

lactate dehydrogenase, urine analysis, prothrombin and partial

thromboplastin times, complement C3 and C4, liver enzymes,

blood urea nitrogen, creatinine, rheumatoid factor, lupus

erythematosus cell phenomenon and antinuclear antibody test.

Skin biopsy from a vesicle on the face revealed dyskeratotic

cells and basal vacuolization in the epidermis, an infiltrate

containing lymphohistiocytic cells and melanophages, and red

blood cell extravasation in the dermis (Fig. 2). Direct immu-

nofluorescence (IF) showed IgG deposits in epidermal inter-

cellular spaces (data not shown). Indirect IF using the

patient’s serum was performed with normal human foreskin

and rat bladder as substrates. IgG aAbs bound both the epithe-

lial cellular surface (ECS) and the basement membrane zone

(BMZ) of rat bladder (Fig. 3a). Similar findings were obtained

on human foreskin (data not shown). However, rat bladder

did not react with the serum of a patient with PV as a nega-

tive control (Fig. 3b). The titres of circulating aAbs were

examined by indirect IF using normal human foreskin as a

substrate. These titres were increased or decreased according

to disease activity from 1 : 1280 to 1 : 20. We also per-

formed immunoblotting using normal human foreskin epider-

mal extract. Regardless of disease severity, both 210- and

190-kDa bands corresponding to envoplakin and periplakin



were always detected on the immunoblot at a titre of 1 : 20.

Moreover, these were consistent with those demonstrated by

PNP patient serum (Fig. 4a, lanes 1 and 2). In contrast, bul-

lous pemphigoid (BP) and PV patient sera did not react with

these two bands, as was expected (Fig. 4a, lanes 3 and 4). To

confirm these immunoblotting analysis results, we performed

immunoprecipitation analysis. As shown in Fig. 4b, our

patient’s serum and PNP patient serum immunoprecipitated

with periplakin and envoplakin protein. However, PV patient

serum used as a negative control showed no reaction with

either protein (Fig. 4b). We also performed enzyme-linked

immunosorbent assay (ELISA) using recombinant purified de-

smoglein (Dsg) 1 antigen (MBL Co., Nagoya, Japan) with our

patient’s serum, and four PNP sera associated with underlying

neoplasms, four PF and three PV sera as positive controls, and

normal serum and two BP sera as negative controls. According

to the manufacturer’s protocol, the results of ELISA revealed

that our patient’s serum (anti-Dsg 1, 78 U mL)1), two of four

PNP sera, four of four PF sera and two of three PV sera

showed positive (anti-Dsg 1, ‡ 20 U mL)1) values, but that

all negative control sera showed negative (anti-Dsg 1,

< 14 U mL)1) values.

Despite thorough laboratory examinations including whole-

body positron emission tomography/computed tomography

(General Electric Medical Systems, Milwaukee, WI, U.S.A.),

abdominal ultrasound, peripheral blood smear, Venereal Dis-

ease Research Laboratory test, carcinoembryonic antigen, a-

fetoprotein, prostate-specific antigen and serum/urine immu-

(b)(a)

Fig 1. Vesicles and erosions with crusting on

the entire face, involving the mucosa of both

eyes and lips and upper chest in the active

disease state (a), and following recovery (b).

Fig 2. Histological examination revealed dyskeratotic cells and basal

vacuolization in the epidermis, a dermal infiltrate of lymphohistiocytic

cells and melanophages, and red blood cell extravasation in the dermis


(b)(a)

Fig 3. Indirect immunofluorescence of the

patient’s serum shows IgG deposits in the

transitional epithelium of rat bladder as a

substrate. (a) Our patient’s serum; (b) serum

of a patient with pemphigus vulgaris (PV) as

negative control. Original magnification

· 200.



564 PNP without an underlying neoplasm, G.T. Park et al.

noelectrophoresis, we were unable to detect any tumour over

a period of approximately 8 years. During the follow-up per-

iod, our patient’s skin lesions were treated with prednisolone,

immunosuppressant (azathioprine) and dapsone. However, his

symptoms gradually worsened and spread to both extremities,

and the treatment was stopped in November 2004. Fortu-

nately, the chronic vesiculobullous eruption improved sponta-

neously after a duration of 7 years, and since May 2005 our

patient has been healthy and off all medication (Fig. 1b). Indi-

rect IF of our patient’s serum in a state of complete remission

was examined on normal human foreskin and rat bladder. IgG

aAbs of our patient’s serum did not bind to the ECS or BMZ

of human foreskin and rat bladder at a titre of 1 : 10, and

only negligible 190- and 210-kDa bands were seen in immu-

noblotting at a titre of 1 : 20 (data not shown).

Discussion

A vesiculobullous disease termed PNP with distinct aAbs was

first described in 1990.1 A diagnosis of PNP is established in

patients who have at least four of the following five criteria

proposed by Anhalt et al.: a polymorphic eruption that affects

the skin and mucous membranes; histopathological features

that include intraepidermal acantholysis and dyskeratosis with

vacuolar changes; an intraepidermal and/or BMZ deposition of

IgG and C3 on direct IF; serum aAbs to multiple epithelia;

and the recognition of the following antigenic bands in im-

munoblotting and/or immunoprecipitation assays: 250, 230,

210 and 190 kDa.1,9,17 Amagai et al.3 have demonstrated both

by immunoprecipitation and by ELISA that PNP sera have

pathogenic antibodies against Dsg 3 and Dsg 1. Thus, the PNP

autoantibody complex is composed of anti-Dsg (Dsg 3 and

Dsg 1), antiplakin (desmoplakins I and II, BP antigen 1, en-

voplakin, periplakin, plectin) and anti-170-kDa (unidentified)

antibodies. Of the various target antigens recognized by indi-

vidual patients with PNP, the most specific serological findings

are aAbs against periplakin and envoplakin.19 These serological

findings are important because these aAbs differ from those

found in PV, BP and other autoimmune bullous diseases.

These aAbs were always detected by immunoblotting analysis

in our patient’s sera obtained during the active disease state.

This suggests that aAbs against periplakin and envoplakin are

still detectable serological markers, unless PNP is treated com-

pletely. Most PNP cases have been reported in patients with a

history of neoplasia, such as lymphoma, chronic lymphocytic

leukaemia, poorly differentiated sarcoma and Castleman’s dis-

ease.1,10–16 However, PNP is not always accompanied by neo-

plasia.17 The present case suggests that not all patients with

PNP have a definite neoplasm or even a history of one. How-

ever, we cannot rule out the possibility that other factors such

as drugs and inflammatory diseases can trigger aAbs related to

PNP in the absence of a neoplasm even though we followed

this patient to find underlying tumours during a long-term 8-

year follow-up period. In addition, patients with PF have been

shown to acquire aAbs against both 190- and 210-kDa anti-

gens of desmosomal plaques by an epitope-spreading phe-

nomenon even though aAbs against Dsg 1 were detected by

ELISA using recombinant purified Dsg 1.5

PNP can mimic a variety of dermatological diseases includ-

ing PV, BP, erythema multiforme and lichen planus.1,2,10–12,16

Owing to this clinical and histological variability, other speci-

fic examinations such as immunoblotting, immunoprecipita-

tion and indirect IF with rat bladder are required for a

definitive diagnosis of PNP.9,20 Thus far, its association with

lymphoproliferative disorders, positive indirect IF staining

with rat bladder, and positivity for envoplakin and/or peripla-

kin by immunoblotting and immunoprecipitation are gener-

ally regarded as sensitive and specific evidence for a diagnosis

of PNP.9,21

Despite no association with a neoplasm, the present case

was diagnosed by a positive indirect IF result using rat

bladder, and by immunoblotting and immunoprecipitation

analyses. Interestingly, our patient recovered from the disease

after 7 years, as is often observed for other autoimmune

bullous diseases such as PV, and has lived well without

medication, i.e. with a favourable outcome. However, the

question remains as to the source of these aAbs against the

plakin family in the absence of a neoplasm. This case indicates

that more studies are required to identify the factors that elicit

an immune reaction and induce PNP in the absence of an

underlying neoplasm.

(a) (b)

Fig 4. The serum of our patient reacted with periplakin and

envoplakin. (a) Immunoblotting (IB) analysis of the sera of various

patients. Lane 1, control paraneoplastic pemphigus (PNP) serum

reacted with 210- and 190-kDa antigens; lane 2, our patient’s serum;

lane 3, control bullous pemphigoid (BP) serum reacted with 230- and

180-kDa antigens; lane 4, control pemphigus vulgaris (PV) serum

reacted with 160- and 130-kDa antigens. (b) Immunoprecipitation

(IP) analysis of the sera of several patients. Control PNP serum, our

patient’s serum and PV serum were incubated with normal human

epidermal extract. Each precipitate was immunoblotted using

antiperiplakin and anti-envoplakin antibodies. The arrows indicate

envoplakin and periplakin, respectively. Control PNP serum and our

patient’s serum reacted with both envoplakin and periplakin, but PV

serum did not. Lower panels show endogenous periplakin and

envoplakin, respectively.



PNP without an underlying neoplasm, G.T. Park et al. 565

References

1 Anhalt GJ, Kim SC, Stanley JR et al. Paraneoplastic pemphigus. An

autoimmune mucocutaneous disease associated with neoplasia. NEngl J Med 1990; 323:1729–35.

2 Joly P, Thomine E, Gilbert D et al. Overlapping distribution ofautoantibody specificities between pemphigus vulgaris and para-

neoplastic pemphigus. J Invest Dermatol 1994; 103:65–72.3 Amagai M, Nishikawa T, Noussari HC et al. Antibodies against de-

smoglein 3 (pemphigus vulgaris antigen) are present in sera frompatients with paraneoplastic pemphigus and cause acantholysis in

vivo in neonatal mice. J Clin Invest 1998; 102:775–82.4 Joly P, Gilbert D, Thomine E et al. Immunofluorescence and immu-

noelectron microscopy analyses of a human monoclonal anti-epi-thelial cell surface antibody that recognizes a 185-kD polypeptide:

a component of the paraneoplastic pemphigus antigen complex? JInvest Dermatol 1993; 101:339–45.

5 Ghohestani R, Joly P, Gilbert D et al. Autoantibody formationagainst a 190-kDa antigen of the desmosomal plaque in two

patients with pemphigus foliaceus. Br J Dermatol 1997; 137:774–9.6 Joly P, Mokhtar I, Gilbert D et al. Immunoblot and immunoelectron

microscopic analysis of endemic Tunisian pemphigus. Br J Dermatol1999; 140:44–9.

7 Chorzelski TP, Maciejowski E, Jablonska S et al. Coexistence of pem-phigus and bullous pemphigoid. Arch Dermatol 1974; 109:849–53.

8 Korman NJ, Stanley JR, Woodley DT. Coexistence of pemphigusfoliaceus and bullous pemphigoid: demonstration of autoantibodies

that bind to both the pemphigus foliaceus antigen complex andthe bullous pemphigoid antigen. Arch Dermatol 1991; 127:387–90.

9 Joly P, Richard C, Gilbert D et al. Sensitivity and specificity of clin-

ical, histologic, and immunologic features in the diagnosis of para-neoplastic pemphigus. J Am Acad Dermatol 2000; 43:619–26.

10 Tankel M, Tannenbaum S, Parekh S. Paraneoplastic pemphigus pre-senting as an unusual bullous eruption. J Am Acad Dermatol 1993;

29:825–8.

11 Mehregan DR, Oursler JR, Leiferman KM et al. Paraneoplastic pem-phigus: a subset of patients with pemphigus and neoplasia. J Cutan

Pathol 1993; 20:203–10.12 Lam S, Stone MS, Goeken JA et al. Paraneoplastic pemphigus, cic-

atricial conjunctivitis, and acanthosis nigricans with pachydermato-glyphy in a patient with bronchogenic squamous cell carcinoma.

Ophthalmology 1992; 99:108–13.13 Fullerton SH, Woodley DT, Smoller BR, Anhalt GJ. Paraneoplastic

pemphigus with autoantibody deposition in bronchial epitheliumafter autologous bone marrow transplantation. JAMA 1992;

267:1500–2.

14 Camisa C, Helm TN, Liu YC et al. Paraneoplastic pemphigus: areport of three cases including one long-term survivor. J Am Acad

Dermatol 1992; 27:547–53.15 Berg WA, Fishman EK, Anhalt GJ. Retroperitoneal reticulum cell

sarcoma: a cause of paraneoplastic pemphigus. South Med J 1993;86:215–17.

16 Stevens SR, Griffiths CEM, Anhalt GJ et al. Paraneoplastic pemphiguspresenting as a lichen planus pemphigoides like eruption. Arch Der-

matol 1993; 129:866–9.17 Ostezan LB, Fabre VC, Caughman SW et al. Paraneoplastic pemphi-

gus in the absence of a known neoplasm. J Am Acad Dermatol 1995;33:312–15.

18 Gooptu C, Littlewood TJ, Frith P et al. Paraneoplastic pemphigus:an association with fludarabine? Br J Dermatol 2001; 144:1255–61.

19 Hashimoto T, Amagai M, Watanabe K et al. Characterization ofparaneoplastic pemphigus autoantigens by immunoblot analysis.

J Invest Dermatol 1995; 104:829–34.20 Horn TD, Anhalt GJ. Histologic features of paraneoplastic pemphi-

gus. Arch Dermatol 1992; 128:1091–5.21 Liu AY, Valenzuela R, Helm TN et al. Indirect immunofluorescence

on rat bladder transitional epithelium: a test with high specificityfor paraneoplastic pemphigus. J Am Acad Dermatol 1993; 28:696–9.



566 PNP without an underlying neoplasm, G.T. Park et al.


Allogeneic bone marrow transplantation in a 7-year-old girlwith congenital erythropoietic porphyria: a treatmentdilemmaS.M. Taibjee, O.E. Stevenson,* A. Abdullah,� C.Y. Tan,� P. Darbyshire, C. Moss, H. Goodyear,* A. Heagerty,*S. Whatley� and M.N. Badminton�

Departments of Haematology and Dermatology, Birmingham Children’s Hospital, Birmingham B4 6NL, U.K.

*Departments of Paediatrics and Dermatology, Heart of England NHS Trust, Birmingham, U.K.

�Department of Dermatology, City Hospital, Birmingham, U.K.

�Department of Medical Biochemistry, University Hospital Wales, Cardiff, U.K.

CorrespondenceSaleem Taibjee.



Key wordsallogeneic, bone marrow transplantation, congenital

erythropoietic porphyria, graft-versus-host disease,

Gunther’s disease, postinflammatory

hypopigmentation


Summary

Congenital erythropoietic porphyria (CEP, Gunther’s disease) has a very variablephenotype. In the more severely affected, bone marrow transplantation (BMT) ispotentially curative, but is not without risks. We describe a 7-year-old girl withCEP characterized by severe photosensitivity but only mild anaemia, in whomthe difficult decision to proceed with allogeneic BMT was made after discussionin a multidisciplinary team. She has shown successful engraftment, accompaniedby biochemical and clinical resolution of her metabolic disease. She remains well3 years later, the oldest patient with CEP receiving BMT to survive beyond12 months. However, she has experienced significant morbidity including floridcutaneous graft-versus-host disease with postinflammatory hypopigmentation.Her case is important in highlighting the delay in diagnosis not uncommon inthis condition and the complex decision-making process involved in proceedingwith BMT.

Congenital erythropoietic porphyria (CEP, Gunther’s disease)

is a rare autosomal recessive porphyria resulting from marked

deficiency of uroporphyrinogen III synthase (UROS) enzyme

activity (10q25.2-q26.3). It has a variable clinical phenotype,

ranging from hydrops fetalis and intrauterine death, severe

lifelong photosensitivity, skin fragility, transfusion-dependent

haemolytic anaemia and hypersplenism with curtailed life

expectancy, to only mild cutaneous features in some cases.1,2

At least 36 different UROS gene mutations have been identi-

fied in CEP3 and the estimation of functional activity of

mutant enzyme in prokaryotic expression systems has enabled

genotype–phenotype correlation.1,2,4 The biochemical disturb-

ance reflected by endogenous porphyrin profiles also predicts

phenotypic severity in patients.5

Treatment options are based on clinical severity, and

include photoprotection, b-carotene, activated charcoal, hyper-

transfusion, hydroxycarbamide and splenectomy.2,6–10 In vitro

studies indicate that retroviral-mediated gene therapy of hae-

matopoietic stem cells may be a potential treatment option in

the future.11,12 In severe cases the prognosis tends to be poor

and bone marrow transplantation (BMT) may be justified as

it is potentially curative. Successful engraftment leads to a

replacement of bone marrow with erythroblasts with normal

UROS activity. Porphyrin levels normalize and photosensitivity

subsides. It was first used in 1991 in a 10-year-old girl with

CEP who demonstrated biochemical and clinical resolution,

but died of overwhelming cytomegalovirus infection, high-

lighting the considerable risks involved.13 Since then at least

a further eight cases have been reported where either

BMT1,14–19 or cord-blood stem cell transplantation20 has been

successful in this condition (this may be an underestimate,

after identification of at least one further successful case fol-

lowing personal communication with Prof. Alain Fischer,

Hopital Necker-Enfants Malades, Paris, 2005). Most patients

were transfusion dependent prior to BMT, but with notable

exceptions.18,19 One patient subsequently required retrans-

plantation.16,19 The management dilemmas in CEP, particularly

concerning BMT, have previously been alluded to by Dawe

et al.21 where they described a 35-year-old man with severe

CEP where allogeneic BMT was considered but discounted

based on the perceived risks.

We report a 7-year-old girl with CEP where the decision to

proceed with BMT was a difficult one, involving a multidisci-

plinary team and the family.



Case report

Our patient was born to consanguineous Pakistani parents. She

had an unremarkable perinatal history. Her parents had

noticed pink-stained nappies during her infancy. Aged

8 months she had been assessed for failure to thrive and diar-

rhoea, without a clear cause identified. Aged 2 years, she was

seen by a paediatric dental surgeon regarding her discoloured

brown teeth, and at 3 years endocrine causes for her excessive

hair were excluded. She had skin changes, which had initially

been attributed to atopic eczema. The unifying diagnosis of

CEP was eventually made at age 5 years by both a senior

dermatologist and a paediatrician. In retrospect her parents

described a blistering eruption affecting the face and hands

since infancy, present all year round but worse after exposure

to sunlight. Examination showed blistering, prominent scar-

ring and hypertrichosis on the face and dorsum of hands

(Figs 1, 2), erythrodontia (Fig. 3) with pink fluorescence

under Wood’s light (Fig. 4) and discoloured toenails.

Blood and urine porphyrin analysis was characteristic for

CEP, with total urinary porphyrin 20 141 nmol L)1 (reference

range 20–320), comprising 87% uroporphyrin isomer I, and

plasma total porphyrin 444Æ6 nmol L)1 (reference ran-

ge < 11Æ2) with plasma porphyrin fluorescence emission spec-

troscopy confirming a peak at 618 nm. Faecal porphyrin

analysis showed a total porphyrin of 2143 nmol g)1 dry

weight (reference range 0–200), comprising 91% copro-

porphyrin isomer I. UROS enzyme assay confirmed a markedly

reduced red cell uroporphyrin III synthesis activity of

41 nmol mL)1 erythrocytes h)1 (mean activity in three con-

temporaneous control samples 183 nmol mL)1 h)1). Full

blood count showed a mild normocytic anaemia with haemo-

globin (Hb) 10Æ8 g dL)1 (normal 11Æ5–16Æ5) and normal

leucocyte and platelet count. Bone marrow aspiration showed

a normocellular marrow with mild dyserythropoiesis only.

No mutation could be identified after sequencing exon and

intron-exon flanking regions of the UROS gene.

The initial management included sun avoidance and sun

protection using high factor sunblock, with minimal success.

A particular factor was the family’s trips to Pakistan, which

included the Pakistani summer months. Our patient continued

to display ongoing severe photosensitivity and scarring.Fig 1. Dorsum of hands at age 5 years: scarring, milia and active

blisters.

Fig 2. Face at age 5 years: scarring, dyspigmentation and

hypertrichosis.

Fig 3. Erythrodontia.

Fig 4. Dentition showing pink fluorescence under Wood’s light.



568 Allogeneic bone marrow transplantation in a girl with CEP, S.M. Taibjee et al.

A multidisciplinary team was established, involving the derma-

tologist, paediatric haematologist, bone marrow transplant

team and general practitioner. The school and family were

both involved. Based on previous reports of successful BMT in

CEP, the long process of discussion culminated in the decision

to proceed with BMT 2 years after her initial diagnosis. Rele-

vant factors included her severe dermatological manifestations,

failure to thrive and short stature, ongoing anaemia (Hb

8Æ1 g dL)1 at 1 month prior to BMT), significant lifestyle

restrictions, poor school attendance, the views expressed by

the family itself, as well as the availability of a sibling with a

full HLA match (HLA typing: A2, A11; B37, B60; Cw6, Cw3;

DR10, DR11, DR52; DQ7, DQ1).

In January 2003, aged 7 years, she underwent allogeneic

BMT from her 4-year-old brother at Birmingham Children’s

Hospital. Conditioning consisted of fludarabine (days )14

to )10; 25 mg m)2 daily), busulphan (days )9 to )6;

3Æ5 mg kg)1 daily) and cyclophosphamide (with mesna, days

)5 to )2; 50 mg kg)1 daily). She received ciclosporin (from

day )1; 2Æ5 mg kg)1 twice daily) for graft-versus-host disease

(GVHD) prophylaxis, granulocyte colony-stimulating factor

(from day +8; 5 lg kg)1 daily) and weekly intravenous im-

munoglobulin (Vigam�; BPL, Elstree, U.K.; from day +2;

0Æ4 g kg)1 weekly). Following bone marrow infusion (defined

as day 0), her neutrophil count was first noted to be greater

than 0Æ5 · 109 L)1 on day +13, suggesting successful engraft-

ment, and this was subsequently confirmed by repeated inter-

phase fluorescent in situ hybridization analyses demonstrating

100% of cells with donor (XY) chromosome complement,

with no cytogenetic evidence of a host (XX) cell line, i.e. full

donor chimerism. She remained photosensitive for only

4 weeks post-BMT, with only occasional new blisters in this

initial period. This was reflected in the steady biochemical

improvement in her porphyrin profile (Table 1). Her anaemia

gradually resolved (Hb 12Æ8 g dL)1 at 5 months post-BMT).

Unfortunately her recovery from BMT was punctuated by a

series of problems, most notably GVHD, anorexia and psycho-

social issues. Acute GVHD commenced at week 3 post-BMT

with a typical maculopapular rash of approximately 30% sur-

face area associated with palmar-plantar erythema. It was most

florid at weeks 4 and 5 post-BMT, and then gradually resolved

with desquamation and hypopigmentation. This was accom-

panied by moderate (grade 2) gut GVHD leading to diarrhoea

and vomiting, confirmed on upper gastrointestinal endoscopy

and colonoscopy. Topical treatments comprised emollients and

antiseptics only. The initial GVHD systemic treatment included

ciclosporin (maximum 3 mg kg)1 twice daily) and predniso-

lone (maximum 1 mg kg)1 daily, or intravenous methyl-

prednisolone when noncompliant with oral medications;

maximum 2 mg kg)1 daily). Two months post-BMT, due to

ongoing gut GVHD and poor compliance with ciclosporin

(which was stopped), she was commenced on mycophenolate

mofetil (continued until 4 months post-BMT; maximum

600 mg m)2 twice daily) and FK506 (gradually weaned off

by February 2004; maximum 0Æ1 mg kg)1 twice daily). Her

diarrhoea persisted intermittently until almost 4 months post-

BMT. Nutritional problems necessitated a period of total

parenteral nutrition. Three months post-BMT, as a result of

persistent anorexia and poor compliance with medication, she

went on to require gastrostomy insertion for enteral feeding,

which was continued for several weeks.

Although there was no evidence of skin sclerosis or other

features of chronic GVHD, she developed widespread post-

inflammatory hypopigmentation and alopecia (Figs 5, 6).

Despite referral to the cosmetic camouflage service and the

input of a clinical psychologist, she remains distressed by her

Table 1 Urine, erythrocyte and plasma porphyrin results before and after bone marrow transplantation (BMT)

At diagnosis

(pre-BMT)

1 month

post-BMT

3 months

post-BMT

9 months

post-BMT

Reference

range

TUP (nmol L)1) 20 141 944 355 604 20–320TUP : creatinine ratio (nmol mmol)1) 2722 175 222 123 < 34

Erythrocyte porphyrin (lmol L)1) 7Æ8 4Æ7 Not done Not increased 0Æ4–1Æ7Plasma porphyrin (nmol L)1) 444Æ6 37Æ4 Not done 15Æ1 < 11Æ2

TUP, total urinary porphyrin.

Fig 5. Postinflammatory hypopigmentation following graft-versus-

host disease.



Allogeneic bone marrow transplantation in a girl with CEP, S.M. Taibjee et al. 569

appearance. She has been difficult to rehabilitate due to var-

ious behavioural problems. Her reluctance to mobilize resulted

in fixed contractures at the elbows and shoulders despite phy-

siotherapy. She was finally discharged from hospital 6 months

post-BMT. At the latest review, 3 years post-BMT, she is hap-

pier and integrating well into school. She still has marked

hypopigmentation, with slow but steady repigmentation, and

wears a wig. The disruption to the family unit caused by her

disease and prolonged hospital admission may have contribu-

ted to the separation of her parents and behavioural problems

in one of her brothers. However, members of her family

express no regrets with their decision to proceed with BMT.

Discussion

The decision to proceed with BMT in CEP should be based on

clinical assessment of the disease’s impact, as well as estimat-

ing the long-term prognosis wherever possible. Our patient

was mildly anaemic and not transfusion dependent. Moore

et al.22 considered the onset of dermatological changes includ-

ing scarring in the first decade of life as a poor prognostic fac-

tor, although there is little evidence for this assertion.

Freesemann et al.5 have previously demonstrated the correl-

ation between degree of porphyrin excess and disease severity

in nine patients with CEP. Total urinary porphyrins ranged

between 2291 and 101 788 nmol per 24 h, but this is hard

to correlate with the levels in our patient, expressed as

nmol mmol)1 creatinine (Table 1). In the same study, whole

blood erythrocyte porphyrins ranged from 0Æ6 to 12Æ0 lmol L)1

(median 2Æ0; reference range 0Æ09–0Æ67) which compares

with 7Æ8 lmol L)1 erythrocytes (i.e. packed red blood cells)

in our patient. Although the different assay methodologies

make a direct comparison difficult, the median baseline eryth-

rocyte porphyrin reported by Freesemann et al. was three times

the assay upper reference limit (URL), compared with 4Æ5times the URL in our patient, confirming that she had a rela-

tively severe baseline biochemical phenotype.

Genotype–phenotype correlation, using in vitro prokaryote

expression of UROS constructs to estimate functional activity

of the mutant enzyme, is extremely useful in CEP. Patients

with severe disease demonstrate mutations leading to near-

absent enzymatic activity, whereas milder cases without trans-

fusion dependence have at least one allelic mutation resulting

in significant residual UROS activity.1,2,5 In view of the paren-

tal consanguinity, a homozygous gene defect was expected in

our patient. However, despite the markedly reduced UROS

enzyme activity, a mutation was not identified using the cur-

rent gene sequencing approach which includes sequencing of

the erythroid-specific region. Using this approach our current

sensitivity of mutation detection is 75% (S.W., unpublished

data). Two other specialist centres have reported allele detec-

tion frequencies of 89–92%,2,23 although neither centre has

reported a patient in whom no mutation was identified on

either allele. This suggests other sites for mutations within this

34-kb gene, including regulatory elements which may be

located thousands of base pairs upstream or downstream from

the coding region. The role of possible modifying genes on

UROS expression is also yet to be established.

Despite immunosuppressive prophylaxis, GVHD remains a

common complication of allogeneic BMT. It appears that our

patient was the only one showing significant GVHD, but there

are no clear reasons for this. Her sibling donor was HLA iden-

tical, and the prophylactic regimen she received was similar to

that in previous reports in CEP, where all patients received

prophylaxis consisting of ciclosporin, although in some cases

combined with methotrexate or corticosteroids. Although she

is the oldest of the 10 patients with CEP now reported to have

undergone BMT, increasing age only emerges as a risk factor

for GVHD in data including mixed adult and paediatric popu-

lations. Age does not appear to be a risk factor in paediatric

cohorts alone.24 The degree of postinflammatory hypopigmen-

tation observed is likely to relate to her Fitzpatrick skin

type V.

Although our patient presented to a variety of specialties

with classical features of CEP, the unifying diagnosis was not

made until 5 years of age. This highlights the diagnostic delay

not uncommon in this condition, as the dermatological signs

may be subtle in early years and many clinicians are unfamil-

iar with this rare disease.

To our knowledge, our patient is the oldest child with CEP

undergoing BMT to survive beyond 12 months. Her case

highlights the complex decision to proceed with BMT, involv-

ing a multidisciplinary team and the family. To date, all 10 of

the patients with CEP receiving BMT reported in the literature

(and the additional unreported patient identified after personal

Fig 6. Postinflammatory hypopigmentation following graft-versus-

host disease, with gastrostomy.



570 Allogeneic bone marrow transplantation in a girl with CEP, S.M. Taibjee et al.

communication with Alain Fischer, Paris) have shown

successful resolution of their disease, although one patient

subsequently died of transplant-associated complications.

An assessment of clinical and psychosocial impact of the dis-

ease, including prognostic factors such as porphyrin levels,

enzyme assays and genotyping, should be balanced against the

potential morbidity and mortality of this intervention.

Acknowledgments

We thank the laboratory staff, Department of Medical Bio-

chemistry, University Hospital Wales, for performing bio-

chemical analyses on our patient, the dedicated staff on the

BMT unit, Birmingham Children’s Hospital, who helped our

patient through her long period of rehabilitation, and our

patient and family for their positive outlook throughout.

References

1 Tezcan I, Xu W, Gurgey M et al. Congenital erythropoietic porphy-ria successfully treated by allogenic bone marrow transplantation.

Blood 1998; 92:4053–8.2 Desnick RJ, Astrin KH. Congenital erythropoietic porphyria: advan-

ces in pathogenesis and treatment. Br J Haematol 2002; 117:779–95.3 Human Gene Mutation Database. http://www.hgmd.cf.ac.uk/ac/

gene.php?gene¼UROS (last accessed 26 October 2006).4 Desnick RJ, Glass IA, Xu W et al. Molecular genetics of congenital

erythropoietic porphyria. Semin Liver Dis 1998; 18:77–84.5 Freesemann AG, Bhutani LK, Jacob K, Doss MO. Interdependence

between degree of porphyrin excess and disease severity in con-genital erythropoietic porphyria (Gunther’s disease). Arch Dermatol

Res 1997; 289:272–6.6 Badminton MN, Elder GH. Management of acute and cutaneous

porphyrias. Int J Clin Pract 2002; 56:272–8.7 Guarini L, Piomelli S, Poh-Fitzpatrick MB. Hydroxyurea in congen-

ital erythropoietic porphyria. N Engl J Med 1994; 330:1091–2.8 Hift RJ, Meissner PN, Kirsch RE. The effect of oral activated char-

coal on the course of congenital erythropoietic porphyria. Br JDermatol 1993; 129:14–17.

9 Pimstone NR, Gandhi SN, Mukerji SK. Therapeutic efficacy of oralactivated charcoal in congenital erythropoietic porphyria. N Engl J

Med 1987; 316:390–3.10 Piomelli S, Poh-Fitzpatrick MB, Seaman C et al. Complete suppres-

sion of the symptoms of congenital erythropoietic porphyria bylong-term treatment with high level transfusions. N Engl J Med

1986; 314:1029–31.

11 Kauppinen R, Glass IA, Aizencang G et al. Congenital erythropoieticporphyria: prolonged high-level expression and correction of

the heme biosynthetic defect by retroviral-mediated gene transferinto porphyric and erythroid cells. Mol Genet Metab 1998; 65:10–

17.12 Mazurier F, Geronomi F, Lamrissi-Garcia I et al. Correction of defi-

cient CD34(+) cells from peripheral blood after mobilisation in apatient with congenital erythropoietic porphyria. Mol Ther 2001;

3:411–17.13 Kauffmann L, Evans DIK, Stevens RF, Weinkove C. Bone marrow

transplantation for congenital erythropoietic porphyria. Lancet 1991;

337:1510–11.14 Dupuis-Girod S, Akkari V, Ged C et al. Successful match-unrelated

donor bone marrow transplantation for congenital erythropoieticporphyria (Gunther disease). Eur J Pediatr 2005; 164:104–7.

15 Harada FA, Shwayder TA, Desnick RJ, Lim HW. Treatment ofsevere CEP with bone marrow transplant. J Am Acad Dermatol 2001;

5:279–82.16 Lagarde C, Hamel-Teillac D, De Prost Y et al. Allogenic bone

marrow transplantation in congenital erythropoietic porphyria.Gunther’s disease. Ann Dermatol Venereol 1998; 125:114–17.

17 Lanino E, Morreale G, Gabarino L et al. Severe congenitalerythropoietic porphyria. Complete correction after BMT from

an unrelated donor. European Group for Blood and MarrowTransplantation 2001; http://congress.akm.ch/abstract/abstract/

abt.ausg_pkt_ses?xssprache¼ENGOPT&xnkon_nr¼45&xnSESSION_NR¼1612&xnCO_PROFILE¼356788# (last accessed 26 October

2006).18 Shaw PH, Mancini AJ, McConnell JP et al. Treatment of CEP in

children by allogenic stem cell transplantation: a case reportand review of the literature. Bone Marrow Transplant 2001; 27:101–5.

19 Thomas C, Ged C, Nordmann Y et al. Correction of congenital ery-thropoietic porphyria by bone marrow transplantation. J Pediatr

1996; 129:453–6.20 Zix-Kieffer I, Langer B, Eyer D et al. Successful cord blood stem

cell transplant for congenital erythropoietic porphyria (Gunther’sdisease). Bone Marrow Transplant 1996; 18:217–20.

21 Dawe SA, Stephens AD, Peters TJ et al. Congenital eythropoieticporphyria: dilemmas in present day management. Clin Exp Dermatol

2002; 27:680–3.22 Moore MR, McColl KEL, Fitzsimmons EJ et al. The porphyrias. Blood

Rev 1990; 4:88–96.23 De Verneuil H, Ged C, Moreau-Gaudry F. Congenital erythropoietic

porphyria. In: The Porphyrin Handbook: Medical Aspects of Porphyrins

(Kadish KM, Smith KM, Guilard R, eds), Vol. 14. San Diego:Academic Press, 2003; 43–63.

24 Bolwell BJ. Are predictive factors clinically useful in bone marrowtransplantation? Bone Marrow Transplant 2003; 32:853–61.



Allogeneic bone marrow transplantation in a girl with CEP, S.M. Taibjee et al. 571

Gene Corner

Galli–Galli disease is an acantholytic variant of Dowling–Degos disease

DOI: 10.1111/j.1365-2133.2006.07703.x

Dowling–Degos disease (DDD; MIM 179850) is a rare auto-

somal dominant disorder characterized by reticulate flexural

hyperpigmentation associated with hyperkeratotic papules,

pitted perioral scars and comedo-like lesions or cysts.1 Charac-

teristic histopathological features include filiform epithelial

downgrowths of the rete ridges, typically involving the folli-

cular infundibulum, basilar hyperpigmentation and dermal

melanosis. Follicular retention cysts and perivascular mono-

nuclear infiltrates can also be present.1 Galli–Galli disease is a

term that has been used to describe patients displaying promin-

ent acantholytic changes on histology, in addition to clinical

and pathological features resembling those of DDD.2 Two muta-

tions in KRT5, encoding one of the two major basal epidermal

keratin intermediate filaments, were recently reported to under-

lie DDD in a number of European cases.3 In the present study,

we report a novel mutation in KRT5 in a patient with clinical and

histopathological features typical of Galli–Galli disease.

Case and methods

A 43-year-old woman of Arab Moslem origin was referred

because of an exceedingly pruritic papular rash involving axil-

lae, inframammary areas and groins (Fig. 1a), and suggestive

of Hailey–Haily disease (MIM 169900), Darier disease (MIM

124200) or pemphigus vegetans. She reported that her

deceased father and eldest sister were similarly affected. Clin-

ical examination revealed brownish reticulate pigmentation

and densely distributed erosive papules in all skin flexural

areas (Fig. 1a). A number of skin biopsies revealed findings

typical of DDD including acanthosis, epidermal rete ridges

with bud-like and filiform projections, a mixed dermal inflam-

matory infiltrate of variable severity, and dermal melano-

phages (Fig. 1b). Numerous foci of suprabasal acantholysis

were also observed (Fig. 1c). To address the possibility that

the patient had an autoimmune blistering disease, possibly

underlying the acantholytic changes observed on histology,

we performed indirect and direct immunofluorescence studies

(aimed at detecting the presence of anti-epidermal antibodies

in patient skin and serum). Both tests were negative. Taken

together, clinical and pathological data suggested a diagnosis

of DDD – Galli–Galli variant.

The patient provided her written and informed consent to

participate in the study according to a protocol previously

approved by the Committee for Genetic Studies of the Israeli

Ministry of Health. Genomic DNA was extracted from whole

blood samples using the salt/chloroform extraction method

and polymerase chain reaction (PCR) amplified using primer

pairs spanning the entire coding sequence as well as intron–

exon boundaries of the KRT5 gene as described elsewhere.4

Mutation p.M1? was verified using PCR–restriction fragment

length polymorphism (RFLP) analysis, exploiting the fact that

it abolishes a recognition site for endonuclease NlaIII. Briefly,

a 795-bp long DNA fragment was amplified using forward

primer 5¢-AGCTCTGTTCTCTCCAGCAC-3¢ and reverse primer

5¢-CAGTCTAATTCAGAACGTGTCC-3¢ and digested with NlaIII.

Results and discussion

We established the entire KRT5 coding sequence in our

patient, as previously described.4 DNA samples from other

family members were not available for analysis. We identified

a single deleterious change consisting of a heterozygous

T fi C transition at KRT5 cDNA position 2 (Fig. 1d). A PCR–

RFLP assay was established based upon the fact the mutation

c.T2C destroys a recognition site for endonuclease NlaIII.

Using this PCR–RFLP assay, we excluded the mutation from a

panel of 100 population-matched chromosomes (not shown).

The c.T2C mutation is predicted to disrupt the KRT5 initi-

ation codon (p.M1?) and therefore to lead to either haplo-

insufficiency by causing lack of protein translation or to the

usage of a downstream alternative initiation codon, resulting

in the synthesis of a N-deleted KRT5 mutant protein as previ-

ously reported in other monogenic disorders.5,6

Most pathogenic sequence alterations described to date in

keratin genes in general, and in KRT5 in particular are mis-

sense mutations located within or at the boundaries of the

central a-helical rod domain.7,8 Clinical and histopathological

consequences of mutations affecting these conserved regions

of epidermal keratin genes classically consist of skin blister-

ing/hyperkeratosis and keratinocyte vacuolar degeneration,

respectively, as seen in epidermolysis bullosa simplex caused

by mutations in KRT5/KRT14.8,9 In contrast, mutations affect-

ing the keratin variable head and tail domains are often non-

sense/frameshift mutations, associated with markedly atypical

phenotypes. Notable examples include mutations resulting in

KRT1 or KRT5 tail domain elongation and causing ichthyosis

hystrix of Curth Macklin10 (MIM 146590) or epidermolysis

bullosa simplex with migratory circinate erythema11 (MIM



609352), respectively; and nonsense/frameshift mutations

affecting the early N-terminal domain of KRT5 and KRT14 and

resulting in DDD3 or Naegeli–Franceschetti–Jadassohn syn-

drome12 (MIM 161000), respectively.

Acantholysis is a typical feature of a number of diseases

such as Hailey–Hailey disease and pemphigus vulgaris.13,14

However, it has also been described as a rare incidental find-

ing in numerous skin diseases such as squamous cell carci-

noma, melanocytic lesions and irritant dermatitis.15–17 This

histopathological finding always reflects abnormal function of

the desmosomal plaque, although the cause for desmosomal

dysfunction varies from one disease to another. The function

of the keratin head domain is not yet known in detail

although a direct interaction between desmoplakin, an import-

ant desmosomal protein, and the KRT5 head domain has been

demonstrated.18 It could therefore be envisaged that haplo-

insufficiency or N-terminal deletion of KRT5, which can be

predicted by all mutations described so far in DDD (Betz et al.3

and this study), may prevent the proper maturation of the

desmosomal plaques, causing abnormal cell–cell adhesion. Of

interest, one of the patients found by Betz et al.3 to carry the

founder KRT5 Ile140fs mutation had previously been reported

as affected with Galli–Galli disease,2 again suggesting a (direct

or indirect) causal relationship between KRT5 mutations and

acantholysis. In addition, abnormal location of desmocollin 3,

another component of the desmosomal plaque, has been

reported in DDD.19 Why have acantholytic changes been

reported in only a minority of DDD cases remains unex-

plained. The location and/or type of KRT5 causative mutation

is unlikely to account for the variable occurrence of this histo-

logical feature as the same mutation (Ile140fs) has been

reported both in Galli–Galli disease and in DDD.3 Independent

modifying genetic traits, including polymorphic sequence

alterations in KRT5, may possibly modulate the phenotypic

expression of mutations in DDD. In this regard, our patient

was found to carry, in addition to the causative mutation, a

number of known polymorphic nucleotide heterozygous

changes in KRT5 (IVS1+9C fi T, c.C594A, c.C594A, c.G732A,

(a) (b)

(c) (d)

Fig 1. Clinical, pathological and genetic assessment of a patient with Galli–Galli disease. (a) On examination, brownish reticulate pigmentation

and numerous erosive reddish papules are observed in the right axilla. Histopathological examination reveals (b) typical filiform elongation of the

rete ridges, melanophages in the dermis and perivascular mononuclear infiltrates (haematoxylin and eosin; original magnification ·100) as well as

(c) numerous foci of suprabasal acantholysis (haematoxylin and eosin; original magnification ·400). (d) Sequence analysis reveals a heterozygous

T fi C transition at cDNA position 2 resulting in loss of KRT5 initiating codon (upper panel). The wild-type (WT) sequence is given for

comparison (lower panel).



Galli–Galli disease is an acantholytic variant of DDD, E. Sprecher et al. 573

cA1582G, c.G1627A). Finally, acantholysis may not be invari-

ably present at all times and anatomical sites in DDD, and

therefore could occasionally be overlooked.

Regardless of the pathomechanism involved, the present

observation and a previous study3 indicate that Galli–Galli dis-

ease is merely a variant of DDD. Whether acantholysis is truly

rare in DDD, or represents a common histological feature of

this disease, which has so far been under-reported, remains to

be determined.

Acknowledgments

We are grateful to our patient for her participation in our

study. We thank Vered Friedman for her help with nucleic

acid analysis. This study was supported in part by a grant pro-

vided by the Bureau for Economic Growth, Agriculture, and

Trade, Office of Economic Growth and Agricultural Develop-

ment, U.S. Agency for International Development, under the

terms of Award No. TA-MOU-01-M21-023.

E . S PR ECH ER*��M. INDE LMAN*

Z. KHAMAY S I*

J . LUGA S SY*�D. PETRON IU S*

R . BERGMAN*�

*Laboratory of Molecular Dermatology and

Department of Dermatology, Rambam Health

Care Campus, POB 9602, Haifa 31096, Israel

�Bruce Rappaport Faculty of Medicine and

�Rappaport Institute for Research in the Medical

Sciences, Technion, Israel Institute of Technology,

Haifa, Israel


References

1 Kim YC, Davis MD, Schanbacher CF, Su WP. Dowling–Degos disease

(reticulate pigmented anomaly of the flexures): a clinical and histo-pathologic study of 6 cases. J Am Acad Dermatol 1999; 40:462–7.

2 Braun-Falco M, Volgger W, Borelli S et al. Galli–Galli disease: anunrecognized entity or an acantholytic variant of Dowling–Degos

disease? J Am Acad Dermatol 2001; 45:760–3.3 Betz RC, Planko L, Eigelshoven S et al. Loss-of-function mutations

in the keratin 5 gene lead to Dowling–Degos disease. Am J HumGenet 2006; 78:510–19.

4 Ciubotaru D, Bergman R, Baty D et al. Epidermolysis bullosa sim-plex in Israel: clinical and genetic features. Arch Dermatol 2003;

139:498–505.5 Cruts M, Gijselinck I, van der Zee J et al. Null mutations in pro-

granulin cause ubiquitin-positive frontotemporal dementia linkedto chromosome 17q21. Nature 2006; 442:920–4.

6 Paulsen M, Lund C, Akram Z et al. Evidence that translation reinitia-tion leads to a partially functional Menkes protein containing two

copper-binding sites. Am J Hum Genet 2006; 79:214–29.7 Omary MB, Coulombe PA, McLean WH. Intermediate filament pro-

teins and their associated diseases. N Engl J Med 2004; 351:2087–100.

8 Irvine AD. Inherited defects in keratins. Clin Dermatol 2005; 23:6–14.

9 Rugg EL, Leigh IM. The keratins and their disorders. Am J Med GenetC Semin Med Genet 2004; 131C:4–11.

10 Sprecher E, Ishida-Yamamoto A, Becker OM et al. Evidence for

novel functions of the keratin tail emerging from a mutation caus-ing ichthyosis hystrix. J Invest Dermatol 2001; 116:511–19.

11 Gu LH, Kim SC, Ichiki Y et al. A usual frameshift and delayed ter-mination codon mutation in keratin 5 causes a novel type of epi-

dermolysis bullosa simplex with migratory circinate erythema.J Invest Dermatol 2003; 121:482–5.

12 Lugassy J, Itin P, Ishida-Yamamoto A et al. Naegeli–Franceschetti–Jadassohn syndrome and dermatopathia pigmentosa reticularis: two

allelic ectodermal dysplasias caused by dominant mutations inKRT14. Am J Hum Genet 2006; 79:724–30.

13 Lanza A, Cirillo N, Femiano F, Gombos F. How does acantholysisoccur in pemphigus vulgaris: a critical review. J Cutan Pathol 2006;

33:401–12.14 Szigeti R, Kellermayer R. Autosomal-dominant calcium ATPase

disorders. J Invest Dermatol 2006; 126:2370–6.15 Kane CL, Keehn CA, Smithberger E, Glass LF. Histopathology of

cutaneous squamous cell carcinoma and its variants. Semin Cutan MedSurg 2004; 23:54–61.

16 Hutcheson AC, Nietert PJ, Maize JC. Incidental epidermolytichyperkeratosis and focal acantholytic dyskeratosis in common

acquired melanocytic nevi and atypical melanocytic lesions. J AmAcad Dermatol 2004; 50:388–90.

17 Naraghi ZS, Mansouri P, Mortazavi M. A clinicopathological studyon acute cutaneous lesions induced by sulfur mustard gas (yper-

ite). Eur J Dermatol 2005; 15:140–5.18 Kouklis PD, Hutton E, Fuchs E. Making a connection: direct

binding between keratin intermediate filaments and desmosomalproteins. J Cell Biol 1994; 127:1049–60.

19 Braun-Falco M, Ring J. Enhanced cytoplasmic expression of desmo-collin 3 in epidermal rete ridges of Dowling–Degos syndrome. Br J

Dermatol 2003; 149:1293–6.

Accepted for publication: 26 September 2006

Key words: Dowling–Degos disease, keratin, mutation, pigmentation

Conflicts of interest: none declared.



574 Galli–Galli disease is an acantholytic variant of DDD, E. Sprecher et al.

Correspondence

A survey of approach to wound healing bysecondary intention among British Society forDermatological Surgery members

DOI: 10.1111/j.1365-2133.2006.07657.x

SIR, Secondary intention healing is a simple and swift ap-

proach to wound healing that, when used appropriately,

results in a satisfactory cosmetic and cost-effective outcome.1

There are, however, no guidelines on how best to manage

wounds healing by secondary intention postoperatively. A

study comparing hydrocolloid occlusive dressings and conven-

tional wound management (twice daily wound cleansing and

topical application of bacitracin ointment and an adhesive

dressing) found that wounds treated with occlusive dressings

healed quicker and were less painful.2 However, this study

looked only at shave biopsy and 3-mm punch biopsy sites. A

recent Cochrane review of dressings and topical agents for

general surgical wound healing by secondary intention3 was

unable to make any firm conclusions as only small, poor-qual-

ity trials exist. We decided to investigate what choices individ-

ual dermatologists with an interest in surgery make regarding

dressings and topical agents for wounds allowed to heal by

secondary intention.

British Society for Dermatological Surgery (BSDS) members

were sent a single-page questionnaire via e-mail and were

invited to participate in this study in September 2005. Two

months later, a reminder was e-mailed to members.

One hundred and eighty-two BSDS members with an e-mail

address were invited to participate. Fifty-one (28%) replied.

Twenty-nine of 51 (57%) had one dedicated surgical session

per week with the remainder having two to four surgical ses-

sions per week. Five (10%) carried out an average of fewer

than five operations per week, while 22 (43%) undertook five

to 10 per week and 24 (47%) performed > 10 per week. Sec-

ondary intention wound healing was used by 42 (82%) BSDS

members on fewer than three wounds per week, by eight

(16%) members on three to six wounds per week, and by a

single member (2%) on more than six wounds per week. The

choices made with regard to dressing type and topical agents

for wounds healing by secondary intention are illustrated in

Fig. 1a,b. The majority (n ¼ 40; 78%) said their patients had

fewer than three follow-up visits at the dermatology depart-

ment for wound care. Thirty-six (71%) members arranged for

follow-up wound care at the patient’s general practitioner and

most (71%) gave their patients a postoperative wound care

information sheet.

The questionnaire was kept brief to encourage uptake

among BSDS members. A response rate of less than a third was

therefore disappointing as published data on response rates

to physician surveys suggest that response rates are routinely

40–50%.4 The brevity of the questionnaire we used limited

the breadth of information obtained by this survey and issues

such as the influence of patient comorbidities or wound site

and size on selection of wounds for secondary intention heal-

ing were necessarily overlooked. None the less, we can make

some conclusions from the responses received. We surveyed a

group of dermatologists with a special interest in surgery. Their

specialist interest is confirmed by their workload recorded –

almost half of those surveyed performed > 10 operations per

week. Wounds were left to heal by secondary intention by the

respondents in a minority of cases. Perhaps this is a facet of

the population of dermatologists surveyed – those with a

greater interest and experience in surgical techniques may

prefer to close wounds with grafts, flaps etc. rather than allow

healing by secondary intention. There was no consensus on

either dressing type or topical agent used for wound care,

although alginate-based dressings and topica Polyfax� (poly-

myxin B+ bacitracin) ointment (PLIVA, Petersfield, U.K.) and

Bactroban� (mupirocin) cream (GlaxoSmithKline, Uxbridge,

02468

101214161820

Num

ber

of r

eplie

s

Dressing types

02468

1012141618

Num

ber

of r

eplie

s

Topical agents

Polyfax Bactroban Other agent Nothing No reply

Alginate

(a)

(b)

Gauze Hydrocolloid Foam Combination No reply

Fig 1. (a) Types of dressing used on wounds healing by secondary

intention. (b) Types of topical agent used by dermatologists on

wounds healing by secondary intention.



U.K.) were the most popular choices. Clearly there is intercen-

tre variation and choices regarding secondary intention wound

healing are most likely to be influenced by personal preference

and experience. The fact that almost one-fifth of those

surveyed use a combination of dressing types confirms the

variation in approach to postoperative care with secondary

intention healing. A study looking at larger dermatological

wounds would be required to establish whether one dressing

type or topical preparation carries benefits over another with

regard to satisfactory outcome of surgical wounds when

factors such as scarring and postoperative infection are

assessed. Furthermore, information on indication for surgery,

site and size of wound, comorbidities and facilities for self-

wound care postoperatively are all important factors that

should be taken into consideration in any future studies.

Acknowledgments

We thank Sean Semple of the BAD office for his assistance in

e-mailing the questionnaire to BSDS members. We also thank

all BSDS members who participated in the survey.

A . LA L LY

R . TURNER


Churchill Hospital, Oxford OX3 7LJ, U.K.


References

1 Christenson LJ, Phillips PK, Weaver AL, Otley CC. Primary closurevs second-intention treatment of skin punch biopsy sites: a random-

ized trial. Arch Dermatol 2005; 141:1093–9.2 Nemeth AJ, Eaglstein WH, Taylor JR et al. Faster healing and less

pain in skin biopsy sites treated with an occlusive dressing. ArchDermatol 1991; 127:1679–83.

3 Vermeulen H, Ubbink D, Goossens A et al. Dressings and topicalagents for surgical wounds healing by secondary intention. Cochrane

Database Syst Rev 2004; 2:CD003554.4 Shosteck H, Fairweather WR. Physician response rates to mail and

personal interview surveys. Public Opin Q 1979; 43:206–17.


Part of this work was presented in poster format at the British Society

for Dermatological Surgery section of the British Association ofDermatologists Annual Meeting in Manchester, July 2006.

Acute necrotic pancreatitis induced by severehypercalcaemia due to tacalcitol ointment

DOI: 10.1111/j.1365-2133.2006.07658.x

SIR, Acute necrotic pancreatitis results from severe aseptic

inflammation leading to tissue necrosis and abdominal pain.

Pancreatic necrosis can activate systemic inflammatory path-

ways and cause subsequent multiple organ failure with high

mortality.1 Gallstones and excessive alcohol intake account

for > 75% of cases of pancreatitis.2 Hypercalcaemia is a rare

cause of acute pancreatitis.3 Hypercalcaemia is induced by a

high parathyroid hormone (PTH) production, malignant

tumours with bone metastasis or paraneoplastic production of

PTH-related peptides. Rarely, excessive oral intake or topical

use of vitamin D has been identified to induce hyper-

calcaemia.4,5 However, hypercalcaemia due to topical use of

vitamin D-containing ointments is usually mild and is not

associated with severe organ dysfunction. We report a patient

with an excessive use of a vitamin D analogue-containing

ointment leading to acute hypercalcaemia and acute necrotic

pancreatitis.

A 28-year-old man was referred to the Department of

Dermatology for acute exacerbation of psoriasis [Psoriasis Area

and Severity Index (PASI) score of 68]. At admission he

denied systemic or topical treatment with vitamin D ana-

logues. Therapy with dithranol in alternation with tacalcitol

(Curatoderm�; Hermal, Reinbek, Germany) was initiated. The

latter treatment was limited to 30% of body surface area once

daily.

On the third day after admission he developed severe

abdominal pain and vomiting and was referred to the Depart-

ment of Gastroenterology. At this time the patient admitted

that he had already used a vitamin D analogue ointment

(tacalcitol) for 3 months at a mean dosage of about 30 g per

day. Thus, the cumulative dosage was approximately 2500–

3000 g, exceeding the manufacturer’s guidelines by eight to

10 times. No alcohol intake within recent days or symptoms

of an infection were reported. The patient did not take any

other medication or vitamin supplements.

Apart from the severe psoriasis and tenderness in the middle

abdomen, physical examination was unremarkable. The labora-

tory results showed an elevation of lipase (2340 U L)1, normal

13–60) and amylase activity (1186 U L)1, normal 13–53).

Serum calcium levels were greatly increased (3.45 mmol L)1,

normal 2.10–2.60; ionized calcium 1.85 mmol L)1, normal

1.15–1.30). PTH was totally suppressed (intact PTH

< 5 ng L)1), while 1,25-dihydroxyvitamin D3 (calcitriol)

levels were elevated (60 ng L)1, normal 16–57). Notably,

serum 25-hydroxyvitamin D3 (calcidiol) levels were extremely

increased (> 100 lg L)1, normal 9–38). Angiotensin-convert-

ing enzyme level was normal.

Computed tomography of the chest and abdomen 3 days

after the onset of symptoms showed acute pancreatitis with

pancreatic oedema and areas of necrosis but no evidence of

gallstones or obstructed bile ducts (Fig. 1). There were no

signs of lymphadenopathy suspicious for sarcoidosis.

Severe hypercalcaemia due to vitamin D intoxication was

considered to be responsible for the acute pancreatitis.

A treatment with isotonic saline infusions and furosemide

to achieve a high fluid turnover was started. However, as

calcium levels did not decrease significantly a 5-day course

of oral steroids (Decortin�; Merck, Darmstadt, Germany;

100 mg once daily) was initiated. This regimen led to a



576 Correspondence

normalization of laboratory parameters and clinical recovery.

After resolution of the pancreatitis, magnetic resonance cho-

langiopancreatography demonstrated an open biliary and pan-

creatic system without any evidence of stones or tumours

(Fig. 2).

Pancreatitis is an acute inflammation of the pancreas clinic-

ally characterized by abdominal pain and elevated pancreatic

enzyme activities.1 Hypercalcaemia is a rare cause of acute

pancreatitis.3 The pathophysiological mechanisms leading to

acute pancreatitis are not well understood. Recent evidence

suggests that calcium can precipitate in the pancreatic duct

and a calcium-dependent activation of trypsinogen occurs

within the parenchyma.6,7 Accordingly, calcium infusion in

rats leads to hyperamylasaemia and morphological alterations

of the pancreas.6

The major causes for hypercalcaemia are cancer and hyper-

parathyroidism.8 Hypercalcaemia can also be induced by high

concentrations of calcitriol or calcidiol due to an increase in

intestinal calcium absorption and bone resorption.8 Vitamin

D3 derivates (e.g. tacalcitol) have been used successfully in the

treatment of psoriasis.9 Ingestion of high doses of vitamin D

or topical application for the treatment of skin diseases has

repeatedly been shown to cause hypercalcaemia.4,5 However,

severe organ damage secondary to hypercalcaemia due to an

excessive topical application has yet not been reported. There-

fore our report of acute necrotic pancreatitis due to vitamin D

intoxication by an excessive use of vitamin D-containing oint-

ment seems of great clinical importance.

It is noteworthy that the PASI score was extremely high in

our patient, indicating a very severe disease where scaling

could have influenced drug absorption. Furthermore, the

cumulative application exceeded the manufacturer’s guidelines

by eight to 10 times and long-term exposure by two to three

times.10 This is particularly important as the amount of drug

absorption by the skin is often underestimated by patients and

ointments are considered as relatively harmless in comparison

with the oral preparation. Therefore, calcium levels should

be monitored in patients using vitamin D-containing medica-

tions with severe disease or who are likely to overdose the

treatment.

C . KNA CK S T EDT

R. WINOGRAD*

A. KOCH*

F . ABUZAHRA�C. TRAUTWE IN*

H.E . WASMUTH*

Medical Department I, *Medical Department III and

�Department of Dermatology,

University Hospital Aachen, RWTH Aachen,

Pauwelsstrasse 30, 52074 Aachen, Germany

Correspondence: Hermann E. Wasmuth.


References

1 Imrie CW. Acute pancreatitis: overview. Eur J Gastroenterol Hepatol1997; 9:103–5.

2 Kusnierz-Cabala B, Kedra B, Sierzega M. Current concepts on diag-nosis and treatment of acute pancreatitis. Adv Clin Chem 2003;

37:47–81.3 Brandwein SL, Sigman KM. Milk-alkali syndrome and pancreatitis.

Am J Med Sci 1994; 308:173–6.4 Jacobus CH, Holick MF, Shao Q et al. Hypervitaminosis D associated

with milk drinking. N Engl J Med 1992; 326:1173–7.5 Bourke JF, Berth-Jones J, Hutchinson PE. Hypercalcaemia with top-

ical calcipotriol. BMJ 1993; 306:1344–5.6 Mithofer K, Fernandez-del Castillo C, Frick TW et al. Acute hyper-

calcemia causes acute pancreatitis and ectopic trypsinogen activa-tion in the rat. Gastroenterology 1995; 109:239–46.

7 Ward JB, Petersen OH, Jenkins SA, Sutton R. Is an elevated concen-tration of acinar cytosolic free ionised calcium the trigger for acute

pancreatitis? Lancet 1995; 346:1016–19.

8 Ziegler R. Hypercalcemic crisis. J Am Soc Nephrol 2001; 12:S3–9.9 Lambert J, Trompke C. Tacalcitol ointment for long-term control

of chronic plaque psoriasis in dermatological practice. Dermatology2002; 204:321–4.

10 van de Kerkhof PCM, Berth-Jones J, Griffiths CEM et al. Long-termefficacy and safety of tacalcitol ointment in patients with chronic

plaque psoriasis. Br J Dermatol 2002; 146:414–22.


Fig 2. Magnetic resonance cholangiopancreatography showing the

gall bladder, common bile duct and pancreas. There is no evidence of

gall stones or compression due to tumours.

Fig 1. Computed tomography of the abdomen showing the pancreas

surrounded by traces of necrosis.



Correspondence 577

Raised limb bands in infancy: a post-traumaticaetiology?

DOI: 10.1111/j.1365-2133.2006.07673.x

SIR, We read with great interest the article by Lateo et al.,1

recently published in this Journal, of four children with raised

limb bands. The authors argued that this entity could repre-

sent a new expression of amniotic band syndrome either

resulting from ‘exogenous’ constrictions by fibrous amniotic

bands or as the result of intrinsic developmental defects. We

report three further cases of raised limb bands and postulate

an acquired post-traumatic pathogenesis.

Patient 1. A 5-month-old girl born at term following an

uncomplicated pregnancy was referred for consultation regard-

ing linear lesions on the back of her lower legs that had first

appeared at the age of 2 months and were increasing in num-

ber. The patient’s mother noticed that the lesions were initial-

ly red and flat after wearing socks, and were still persistent.

While the redness faded, the eruption became progressively

palpable (Fig. 1a). The bands were asymptomatic and the

child was otherwise healthy. No notable trauma, except for

wearing socks, preceded their development. There was no

family history of similar lesions. Examination showed papular,

yellow, thin, linear, horizontal lesions on the mid-calves. They

were bilaterally symmetrical without any associated constric-

tion of the soft tissues or other limb abnormalities. Skin

biopsy revealed a dermal infiltrate of adipocytes along vessels

and eccrine ducts (Fig. 1b).

Patient 2. A 4-month-old boy born at term following an

uncomplicated pregnancy presented with two linear bands at

the posterior aspect of his legs. His mother reported red, flat

bands appearing below the elastic of his socks. Examination

revealed brown, raised lesions running horizontally on both

calves (Fig. 2) without any other associated abnormalities

especially on the lower limbs. No skin biopsy was performed.

There was no family history of similar lesions and there

was no evident trauma before the development of the

bands. Eighteen months later, the limb bands remained

unchanged.

Patient 3. An 18-year-old boy, with a past medical history of

mild mental retardation, was referred for consultation regard-

ing hyperpigmentated linear lesions on the back of his legs

and right thigh. The bands had been present since childhood.

Examination showed two brown, thin, raised, linear, hori-

zontal lesions on the mid-calves and one on the posterior

aspect of the right thigh; the patient also had a transverse

palmar gyrus. Skin biopsy showed an infiltrate of adipocytes

along capillaries in the upper dermis. Two years later, the

limb bands remained unchanged.

Since the first report provided by Meggit et al. in 2002,

13 cases (including the present ones) of raised limb bands

developing in infancy have been described.1–4 In addition,

two observations recently reported by Zhu et al.5 as ’congenital

curvilinear palpable hyperpigmentation’ may actually be the

same entity because of similarities of history, clinical presenta-

tion and evolution.

The cause of this eruption remains unknown although

amniotic band syndrome has been suspected because of asso-

ciated limb constrictions observed in four patients,1,2,4 shor-

tened toes in one patient2 and club foot in one patient.4

However, no oedema or constricting limb band and finger

abnormalities were present in our patients. Nevertheless, in

two cases the skin lesion was preceded by circular erythema,

horizontally localized on the mid-calves and strictly located

at the elastic band of the socks, clearly suggesting a

(a) (b)

Fig 1. Patient 1. (a) Thin, linear,

semicircular, palpable eruption running

horizontally on the mid-calf. (b) Skin biopsy

of the linear papules demonstrating a dermal

infiltrate of adipocytes along eccrine ducts and

vessels (haematoxylin and eosin; original

magnification ·100).



578 Correspondence

post-traumatic aetiology. To support this hypothesis, histology

of the skin lesions showed in two cases the ascent of adipo-

cytes into the upper dermis, commonly described in post-trau-

matic piezogenic papules of the feet. Therefore, we postulate

that postnatally acquired stricture could share the same trau-

matic mechanism as the compression described in the ’extrin-

sic’ model for pathogenesis of amniotic band syndrome and

may represent a new aetiology of raised limb bands.

M.M. MARQUE

B . GU I L LOT

G. L E GAL L I C

D . BE S S I S


University Hospital of Montpellier,

Montpellier, France


References

1 Lateo SA, Taylor AEM, Meggit SJ. Raised limb bands developing in

infancy. Br J Dermatol 2006; 154:791–2.2 Meggitt SJ, Harper J, Lacour M, Taylor AEM. Raised limb bands

developing in infancy. Br J Dermatol 2002; 147:359–63.3 Russi DC, Irvine AD, Paller AS. Raised limb bands developing in

infancy. Br J Dermatol 2003; 149:436–7.4 Dyer JA, Chamlin S. Acquired raised bands of infancy: association

with amniotic bands. Pediatr Dermatol 2005; 22:346–9.5 Zhu IY, Fitzpatrick JE, Weston WL. Congenital curvilinear palpable

hyperpigmentation. J Am Acad Dermatol 2005; 53:162–4.


Human papillomavirus type 5 infectionin a patient with Hailey–Hailey diseasesuccessfully treated with imiquimod

DOI: 10.1111/j.1365-2133.2006.07667.x

SIR, Hailey–Hailey disease, also known as familial benign

pemphigus (FBP), is a rare autosomal dominantly inherited

blistering disease characterized by recurrent vesicles and

erosions, particularly involving flexural areas. The major

underlying pathological process in FBP is acantholysis of

keratinocytes caused by mutations in the genes encoding

the Golgi secretory pathway Ca2+-ATPase (SPCA1, ATP2C1),

affecting cell–cell adhesion. It is not uncommon for

patients with FBP to have secondary bacterial infection.

Nevertheless, FBP with human papillomavirus type 5 (HPV-5)

infection has not been reported in the literature. We report

a patient with FBP with concomitant HPV-5 infection suc-

cessfully treated with topical imiquimod. To our knowledge,

this is the first reported case of FBP associated with HPV-5

infection.

A 62-year-old Taiwanese man presented with a 15-year his-

tory of recurrent itching and oozing rashes involving the scro-

tum and perineum. Two years prior to his first visit to our

clinic, he started to notice asymptomatic papular eruptions at

the perianal area. At the time, he was diagnosed and treated

for tinea cruris and eczema. Nevertheless, the lesions did not

show any response to the prescribed topical antifungal agents

or topical corticosteroids.

On physical examination, multiple, small linear erosions on

an erythematous base were present at the scrotal area

(Fig. 1a). The perianal skin was macerated with a whitish

appearance. In addition, several discrete, pea- to bean-sized,

flat-topped, brownish hyperkeratotic papules with a tendency

to coalesce were observed at the perianal area (Fig. 1b).

To clarify the nature of the lesions, skin biopsy of the peri-

anal lesions was performed. The histopathological features

revealed irregular psoriasiform acanthosis with intraepidermal

clefting and acantholysis (Fig. 2a). Interestingly, cytopathic

changes including chromatin clumping and perinuclear

koilocytic features, and enlargement of keratinocytes, were

found in the spinous layer (Fig. 2b, c). The characteristic

clinical appearance, as well as biopsy, readily confirmed the

diagnosis of FBP, whereas the cytopathic changes as shown in

the histopathological analysis were suggestive of superimposed

cutaneous warts. Viral typing by nested polymerase chain

reaction (PCR) confirmed the pathogen to be HPV-5. With

reference to previously published methods,1 we used general

consensus primers MY09/MY11 for first PCR to amplify the

corresponding part of the HPV L1 gene, followed by nested

PCR primers GP+5/GP+6. Each PCR was carried out in

GeneAmp PCR system 9700 (Perkin-Elmer, Foster City, CA,

U.S.A.). HPV typing by DNA autosequence was analysed by

GenBank online BLAST server, which confirmed the pathogen

to be HPV-5.

Further investigation revealed a positive family history of

FBP. Due to the extensive nature of the lesions, topical 5%

imiquimod cream three times weekly was prescribed. Clinical

resolution of the lesions occurred within a month and side-

effects were minimal (Fig. 1c). Imiquimod was discontinued

after 1 month of treatment and no recurrence was observed

during a 12-month follow-up period.

The differential diagnoses of hyperkeratotic, flattened pa-

pules on the anogenital region in patient with FBP include

concomitant condylomata acuminata and the verrucoid variant

Fig 2. Patient 2. Brown, linear, horizontal and palpable bands on the

posterior aspect of the legs.



Correspondence 579

of FBP. Verrucoid FBP typically has a flexural or genital distri-

bution.2 Rabinovitz and Patel have reported a case of isolated

perianal verrucoid familial benign chronic pemphigus which

was initially mistaken for condylomata acuminata.3 In

addition, Langenberg et al.4 have suggested that all verrucoid

genital lesions with the histological characteristics of Hailey–

Hailey disease may represent FBP. The morphological and

histopathological features of our patient were compatible with

verrucoid FBP. Nevertheless, the presence of HPV infection

confirmed by viral typing, as well as the good response to

topical imiquimod, favoured a diagnosis of FBP with con-

comitant HPV-5 infection, rather than a verrucoid FBP,

which has not been reported to have an association with HPV

infection.

Another interesting feature of our patient was that he

was infected with HPV-5, an oncogenic HPV commonly

(a)

(b)

(c)

Fig 1. (a) Multiple tiny, fissured erosions on an erythematous background were noted over the scrotum. (b) The perianal skin was brownish-

grey, with multiple dark brownish, flat papules scattered around the perianal area. (c) The perianal brown papules had resolved significantly at the

1-month post-treatment follow-up, leaving only mild pigmentation.

(a) (b) (c)

Fig 2. (a) Biopsy from the papular lesions showed acanthosis with suprabasal cleavage and acantholysis. (b) Characteristic ‘dilapidated brick wall’

appearance resulting from acantholytic keratinocytes. (c) Cytopathic changes with hypergranulosis, chromatin clumping and perinuclear halo

(arrowhead) were found in the overlying skin. Haematoxylin and eosin; original magnification: (a) · 100; (b) · 200; (c) · 400.



580 Correspondence

associated with a rare genetic disease, epidermodysplasia

verruciformis (EV). The general population is usually not

affected by this particular virus due to genetic protection.5

Patients with EV were found to have a defective cell-mediated

immune response to HPV infection manifested by the inhi-

bition of natural cytotoxicity and the proliferation of

T lymphocytes against HPV-infected squamous cells in EV skin

lesions.5,6

Although considered specific to EV, HPV-5 has also been

found in several benign skin conditions such as psoriasis

vulgaris and in the epidermal repair process.7,8 It has been

proposed that following immunosuppressive treatment with

phototherapy or topical corticosteroids, psoriatic skin

may become the natural reservoir for HPV-5.7 However,

EV-associated HPVs, including HPV-5, are found as a latent

infection in the hair follicles and the normal skin of

immunosuppressed and, to a lesser extent, immunocompetent

non-EV individuals, so this association with psoriasis is

probably not specific.9–11 We propose that latent infection

with HPV-5 may become temporarily activated in diseases

or conditions involving wound repair processes where exten-

sive keratinocyte migration and proliferation occur.8 HPV-5

DNA would not be detectable in healed lesions once normal

cellular immunity against EV-associated papillomaviruses is

resumed in non-EV patients who are genetically resistant to

HPV-5.

Given that there is no known defect in cell-mediated im-

munity in patients with FBP, the repeated wound repair pro-

cesses in lesions of FBP, in combination with a compromised

local immunity caused by prolonged treatment with topical

corticosteroid, might have contributed to the development of

the HPV-5-associated lesions in our patient. An alternative ex-

planation is that the verrucoid variant of FBP is associated with

HPV-5 infection, but that this specific HPV type has not previ-

ously been adequately looked for as the primers used would

normally be expected to detect mucosal HPV types rather than

beta papillomaviruses. A larger study including more patients

with the verrucoid variant of FBP is necessary to substantiate

this possibility.

We used the histopathological findings of the skin

biopsy (Fig. 2), namely, minimal inflammation of the lesional

skin, as our guide to decide the treatment of choice. The

lack of immune response suggested that immunomodu-

latory agents might be beneficial in our patient by enhancing

his local immunity against HPV infection. We therefore

treated our patient with imiquimod 5% cream, which had

been proven to be clinically effective in treating anogenital

warts. In comparison with other treatment modalities such as

electrocautery, cryotherapy, chemodestruction or excision,

topical imiquimod application is less painful, with milder

tissue destruction.12 In our patient, three times weekly

application of imiquimod 5% cream brought about significant

clinical improvement within 1 month. Side-effects were

minimal.

We propose that conditions such as FBP or autoimmune

bullous diseases with clinically refractory wound healing may

be prone to EV-associated HPV infection. In such circum-

stances, topical immunomodulatory agents are a rational treat-

ment choice, although the exact therapeutic mechanism awaits

further exploration.

C-C . CHAN

H-Y. THONG*

Y-C. CHAN�Y-H. L I AO


National Taiwan University Hospital and

National Taiwan University College of Medicine,

7 Chung-Shan South Road, Taipei 100, Taiwan

*Department of Management Science and Engineering,

School of Engineering, Stanford University,

Stanford, CA, U.S.A.

�Department of Nursing, Taichung Armed Forces

General Hospital, Taichung, Taiwan

Correspondence: Y-H. Liao.


References

1 Ko MJ, Chu CY. Disseminated human papillomavirus type 11infection in a patient with pemphigus vulgaris: confirmed by DNA

analysis. J Am Acad Dermatol 2004; 51:S190–3.2 Murakawa GJ, Kerschmann R, Berger T. Vulvar verrucous papules.

Arch Dermatol 1999; 135:204–5, 207–8.3 Rabinovitz AJ, Patel VJ. Isolated perianal verrucoid familial benign

chronic pemphigus. Cutis 1982; 30:660–2.4 Langenberg A, Berger TG, Cardelli M et al. Genital benign chronic

pemphigus (Hailey–Hailey disease) presenting as condylomas. J AmAcad Dermatol 1992; 26:951–5.

5 Majewski S, Malejczyk J, Jablonska S et al. Natural cell-mediatedcytotoxicity against various target cells in patients with epider-

modysplasia verruciformis. J Am Acad Dermatol 1990; 22:423–7.6 Cooper KD, Androphy EJ, Lowy D et al. Antigen presentation and

T-cell activation in epidermodysplasia verruciformis. J Invest Dermatol1990; 94:769–76.

7 Majewski S, Jablonska S. Possible involvement of epidermodys-plasia verruciformis human papillomaviruses in the immunopatho-

genesis of psoriasis: a proposed hypothesis. Exp Dermatol 2003;12:721–8.

8 Favre M, Majewski S, Noszczyk B et al. Antibodies to human papil-lomavirus type 5 are generated in epidermal repair processes.

J Invest Dermatol 2000; 114:403–7.9 Boxman IL, Berkhout RJ, Mulder LH et al. Detection of human

papillomavirus DNA in plucked hairs from renal transplantrecipients and healthy volunteers. J Invest Dermatol 1997; 108:712–

15.10 Boxman IL, Russell A, Mulder LH et al. Case–control study in a

subtropical Australian population to assess the relation betweennon-melanoma skin cancer and epidermodysplasia verruciformis

human papillomavirus DNA in plucked eyebrow hairs. TheNambour Skin Cancer Prevention Study Group. Int J Cancer 2000;

86:118–21.

11 Harwood CA, Surentheran T, Sasieni P et al. Increased risk of skincancer associated with the presence of epidermodysplasia verruci-

formis human papillomavirus types in normal skin. Br J Dermatol2004; 150:949–57.

12 Sauder DN. Immunomodulatory and pharmacologic properties ofimiquimod. J Am Acad Dermatol 2000; 43:S6–11.




Correspondence 581

Perioral dermatitis in a patient withmyasthenia gravis following systemiccorticosteroid treatment

DOI: 10.1111/j.1365-2133.2006.07668.x

SIR, Perioral dermatitis is a common dermatosis presenting as a

persistent erythematous eruption composed of tiny papules

and papulopustules distributed symmetrically around the

mouth and over the chin. A narrow zone of sparing around

the vermilion border of the lips is a pathognomonic feature.

Additional lesions may arise on the nasolabial folds, glabella,

periocular skin and eyelids.1–5 Facial flushing and comedones

are notably absent. Classically, perioral dermatitis develops sec-

ondary to the use of topical corticosteroids or facial cosmetics.1

Young women (age 16–45 years) are most often affected,

although cases involving children and men have been repor-

ted.1,2 Treatment requires cessation of the offending agent with

or without the addition of antibiotics such as tetracycline or

erythromycin.1–5 Without adequate treatment, the disease

tends to run a persistent and fluctuating course. While perioral

dermatitis has been closely associated with topical corticoster-

oid use, the emergence of this condition in patients taking sys-

temic corticosteroids has rarely been reported in the literature.6

We report a 37-year-old woman with no significant past

medical history who was admitted following 2 days of gener-

alized muscle weakness, blurry vision, difficulty swallowing,

and voice changes. She was diagnosed with myasthenia gravis

(MG) following a positive edrophonium test, single fibre elec-

tromyography that demonstrated severe dysfunction at the

neuromuscular junction, and positive anti-MuSK antibody.

Antibodies against skeletal muscle and acetylcholine receptor

binding were negative. A total-body computed tomographic

scan revealed no signs of malignancy. She was treated with

oral prednisone 100 mg daily, pyridostigmine (Mestinon�;

Valeant, Costa Mesa, CA, U.S.A.), and was scheduled for five

rounds of plasmapheresis.

After 3 weeks of treatment with oral prednisone and pyrido-

stigmine, the dermatology department was consulted to evalu-

ate the patient for an asymptomatic eruption of 1–2-mm

erythematous papules distributed circumferentially around her

mouth and over her chin (Fig. 1). A small zone (2–3 mm) of

sparing was noted around her lip margin. She had minimal

involvement of the nasolabial folds, and no lesions of the

glabella, eyelids or periocular region. Physical examination also

revealed 23 small (< 1 mm), erythematous papulopustules

sparsely distributed across her upper back. No lesions were

observed on her neck, chest or extremities. She denied a history

of acne, topical corticosteroid or cosmetics use, or similar erup-

tions in the past. Based on this clinical presentation, the diagno-

sis of perioral dermatitis secondary to systemic corticosteroid

use was made. The patient was educated on the aetiology of

her dermatological condition and provided with treatment

options. She declined treatment of her perioral dermatitis at this

time, opting to wait for resolution of the disease as her corti-

costeroid dose was tapered over the coming months.

While numerous studies have linked perioral dermatitis with

topical corticosteroid use, only one other report, published in

this Journal by Adams et al. in 1982, specifically explores the

emergence of perioral dermatitis in five renal transplant

patients maintained on oral prednisolone.6 In our case, we

report a young woman who developed her first perioral erup-

tion 3 weeks after beginning oral prednisone 100 mg daily for

treatment of MG. Of note, the patients reported by Adams et al.

developed their eruptions after 7–13 months of systemic corti-

costeroids, whereas our patient’s lesions appeared in the first

month of steroid therapy, which is more consistent with the

time course of steroid-induced acneiform eruptions.

J .M. GOS S

K .M. NORD*

M.R. OLART E�M.E . GROS SMAN*

Columbia University, College of Physicians

and Surgeons, New York, NY, U.S.A.

Departments of *Dermatology and

�Neurology, Columbia University,

College of Physicians and Surgeons,

New York, NY, U.S.A.

Correspondence: Marc E. Grossman.


References

1 Plewig G, Kligman AM. Perioral dermatitis. In: Acne and Rosacea, 3rdedn. Berlin: Springer-Verlag, 2000; 504–6.

2 Wilkinson DS, Kirton V, Wilkinson JD. Perioral dermatitis:a 12 year review. Br J Dermatol 1979; 101:245–57.

3 Cotterill JA. Perioral dermatitis. Br J Dermatol 1979; 101:259–62.4 Weber G. Rosacea-like dermatitis: contraindications or intolerance

reactions to strong steroids. Br J Dermatol 1972; 86:253–9.5 Hafeez ZH. Perioral dermatitis: an update. Int J Dermatol 2003;

42:514–17.6 Adams SJ, Davison AM, Cunliffe WJ, Giles GR. Perioral dermatitis in

renal transplant recipients maintained on corticosteroids and immu-nosuppressive therapy. Br J Dermatol 1982; 106:589–92.


Fig 1. A 37-year-old woman with perioral dermatitis appearing

3 weeks after the initiation of systemic corticosteroid treatment.



582 Correspondence

Confluent and reticulated papillomatosissuccessfully treated with amoxicillin

DOI: 10.1111/j.1365-2133.2006.07676.x

SIR, Further to the recently published experience of the Mayo

Clinic in the treatment of confluent and reticulated papilloma-

tosis (CRP),1 we report a case of CRP treated successfully with

amoxicillin.

Confluent and reticulated papillomatosis is an unusual skin

disorder originally described by Gougerot and Carteaud. It

consists of a scaly, macular and papular eruption typically

affecting the seborrhoeic areas of young adults. As the

name suggests, lesions become confluent in the centre, coales-

cing to form a reticulated pattern at the periphery of affected

areas.

We describe a 29-year-old female who initially presented in

2004 with a 4-year history of an asymptomatic eruption

affecting the axillae, mons pubis, submammary (Fig. 1) and

inguinal folds. She had been treated unsuccessfully with

topical steroids and antifungal agents. Differential diagnoses

included a fungal eruption, pityriasis versicolor, Darier’s

disease and Dowling–Degos disease (reticulate pigmented

(a) (b)

Fig 2. Right axilla (a) before and (b) after

treatment.

Fig 1. Confluent and reticulated

papillomatosis affecting axillae, mons pubis

and submammary areas.



Correspondence 583

anomaly of the flexures) but these were excluded by investi-

gations; skin scrapings for mycology were negative and a skin

biopsy showed mild chronic inflammation in the upper

dermis with some mild pigmentary incontinence. Wood’s

light examination was unremarkable. A skin swab and culture

of skin scrapings on chocolate agar were negative. Following

clinical and histological correlation a diagnosis of CRP was

made.

The patient was treated unsuccessfully with oral mino-

cycline 100 mg daily for 3 months, and thereafter all treat-

ment was stopped because the patient became pregnant.

During pregnancy, she was treated with amoxicillin 250 mg

(taken three times daily) for a coincidental lower respiratory

tract infection and noted dramatic improvement in her CRP.

Clinical examination confirmed complete clearance. A further

1-month course of amoxicillin 250 mg (taken three times

daily) was prescribed 3 months later for disease relapse.

Again, improvement was noted after approximately 1 week,

with clearance of the eruption within 3 weeks (Fig. 2),

confirming that the initial response was not coincidental.

Several theories for the cause of CRP have been suggested,

including an abnormal host response to Pityrosporum orbiculare

and more recently infection with the Dietzia strain of Actinomy-

cete.2 In conjunction with the microbiology team we attempted

to isolate Dietzia strain X from culturing skin scrapings on

chocolate agar but without success.

Various antibiotics have been reported to treat CRP, includ-

ing azithromycin, erythromycin and oral fusidic acid.3 A

recent retrospective analysis of 39 cases with CRP from

the Mayo Clinic reported complete clearance in 14 of

22 patients treated with minocycline therapy, partial

response in a further four of 22 patients and four cases

where a response to minocycline was not documented.1 Our

patient was initially treated with minocycline but this was not

beneficial.

To the best of our knowledge this is the first report of

successful treatment with amoxicillin in CRP.

R .F . DAV I S

K .E . HARMAN

Department of Dermatology, Leicester Royal Infirmary,

Leicester, U.K.


References

1 Davis MDP, Weenig RH, Camilleri MJ. Confluent and reticulate

papillomatosis (Gougerot–Carteaud syndrome): a minocycline-responsive dermatosis without evidence for yeast pathogenesis.

A study of 39 patients and a proposal of diagnostic criteria. Br JDermatol 2006; 154:287–93.

2 Natarajan S, Milne D, Jones AL et al. Deitzia strain X: a newlydescribed Actinomycete isolated from confluent and reticulated papillo-

matosis. Br J Dermatol 2005;153: 825–7.

3 Jang HS, Oh CK, Cha JH et al. Six cases of confluent and reticulatedpapillomatosis alleviated by various antibiotics. J Am Acad Dermatol

2001; 44:652–5.


Folliculotropic mycosis fungoides with CD30+large-cell transformation in a young woman:beneficial effect of bexarotene

DOI: 10.1111/j.1365-2133.2006.07672.x

SIR, Folliculotropic mycosis fungoides (MF) is a rare variant of

cutaneous T-cell lymphoma (CTCL). Folliculotropic MF has a

male/female ratio of 4–5 : 1. The sites of predilection are the

face, neck and upper trunk. Clinically, folliculotropic MF

shows acneiform lesions, including prominent lesional and

perilesional comedo-like eruptions, epidermal cysts, follicular

keratoses, erythematous patches and plaques. Sometimes there

may be additional mucinorrhoea and hair loss in affected

areas. Histologically, this subtype of MF is characterized by

atypical lymphocytes infiltrating predominantly the hair folli-

cles. Pautrier microabscesses may be present, as in nonfollicu-

lotropic MF.1,2 Mucinous degeneration of hair follicles may be

present. Folliculotropic MF has recently been suggested to be

more refractory to standard therapy and associated with a

worse prognosis as compared with classical MF.1 We report a

young woman with folliculotropic MF presenting a CD30+

large-cell transformation. She experienced full remission upon

treatment with low-dose bexarotene, administered for a time

period of 1 year.

At the age of 17 years our patient developed eczematous

lesions on her extremities, thought to be atopic dermatitis.

In addition, since the age of 19 years there was a progressive

development of acneiform lesions on her face with small and

larger cysts (Fig. 1a). At the age of 28 years she presented

with multiple poikilodermatous plaques containing small whi-

tish cysts. Skin biopsy showed the typical histological pattern

of folliculotropic MF (Fig. 1b). Immunohistologically the

atypical lymphocytes demonstrated a CD3+, CD4+ and CD8–

phenotype, with a few CD20+ B cells and CD30+ cells

admixed. Polymerase chain reaction analysis performed from

lesional skin showed a monoclonal rearrangement of the T-cell

receptor. Staging procedures did not reveal any lymph node

or organ infiltration of her lymphoma. Some months after

therapy with ultraviolet (UV) B 311 nm and systemic isotre-

tinoin (Roaccutane�; Roche, Mannheim, Germany) had been

started, she consecutively developed three fast-growing, ulcer-

ated cherry-sized nodules at her right elbow and at her chin

(Fig. 1c). Two nodules were excised and examined histologic-

ally. The epidermis was necrotic, covered with an overlying

crust. The dermis was filled with a dense sheet-like nodular

infiltrate, predominantly consisting of large CD3– cells, with

pale cytoplasm and prominent nucleoli, which stained positive

for CD30 (Fig. 1d). The diagnosis of CD30+ large-cell trans-

formation of folliculotropic MF was made. The patient’s whole

face and the CD30+ nodule on her chin were treated by elec-

tron beam irradiation (20 Gy). As the cysts persisted almost

unchanged in the facial area 3 months after radiotherapy

(Fig. 2a) as well as within the lesions on the trunk and



584 Correspondence

extremities, we initiated treatment with bexarotene (Targre-

tin�; Zeneus, Munich, Germany) 150–450 mg daily. After

18 weeks the first lesions showed healing. This treatment was

accompanied by a slow but steady further regression of cysts

and comedones, resulting in a nearly complete healing after a

period of 12 months (Fig. 2b) at a total dose of 129Æ98 g.

Follicular atrophy became apparent where the cystic lesions

had regressed. There was no relapse of follicular MF or

CD30+ large-cell transformed nodules during the whole treat-

ment period with bexarotene. The patient has been off therapy

for 4 months without signs of relapse.

Until now, folliculotropic MF, although a well recognized

variant of MF, has been reported only rarely. Possibly, due to

the intra- and perifollicular localization of the neoplastic

lymphocytes, follicular MF seems to be less responsive to

treatment modalities normally applied for classical MF.1 Some

authors reported a response to electron beam radiation, or

psoralen plus UVA alone or in combination with isotretinoin

and/or interferon alfa-2a.3–5 To date, to our knowledge only

two patients have been reported to be successfully treated with

oral bexarotene.6,7 Bexarotene is a selective retinoid X recep-

tor-binding molecule, which induces apoptosis in target cells.

At a dose of 300 mg m)2 daily, response rates of more then

50% are reported in early-stage CTCL.8

In our patient follicular MF underwent large-cell CD30+

transformation, presenting with rapidly growing nodules.

Transformation of MF is often associated with resistance to

therapy and poor prognosis.1 We treated our patient with local

radiation therapy, followed by therapy with bexarotene (initial

dose 150 mg daily, rising to 450 mg daily) and achieved

complete remission of plaques and cysts after 12 months.

New nodules were not encountered. It remains to be clarified

whether bexarotene, besides being active in folliculotropic MF,

will also inhibit the eruption of CD30+ transformed nodular

infiltrates in our patient. There are recent communications

indicating that transformation to a more aggressive CTCL

can occur in patients under therapy with bexarotene.9,10

C . MITT E LDORF

R . STADL ER*

H.P . B ERT SCH

C . NEUMANN


Goettingen University, von-Siebold-Str. 3,

37073 Goettingen, Germany


Medical Centre Minden, Portastrasse 7–9,

32427 Minden, Germany.


References

1 Kazakov DV, Burg G, Kempf W. Clinicopathological spectrum of

mycosis fungoides. J Eur Acad Dermatol Venereol 2004; 18:397–415.2 Flaig MJ, Cerroni L, Schuhmann K et al. Follicular mycosis fungoid-

es. A histopathologic analysis of nine cases. J Cutan Pathol 2001;28:525–30.

3 Leverkus M, Rose C, Brocker E-B, Goebeler M. Follicular cutaneousT-cell lymphoma: beneficial effect of isotretinoin for persisting

cysts and comedones. Br J Dermatol 2005; 152:176–98.4 Lacour JP, Castanet J, Perrin C, Ortonne JP. Follicular mycosis

fungoides. J Am Acad Dermatol 1993; 29:330–4.5 Willemze R, Jaffe ES, Burg G et al. WHO–EORTC classification for

cutaneous lymphomas. Blood 2005; 105:3768–85.

(a)

(c) (d)

(b)

Fig 1. Folliculotropic mycosis fungoides (MF) with CD30+ large-cell

transformation. (a) Initial clinical presentation with acneiform lesions

on the face, comprising small and larger cysts. (b) Scanning

magnification of a pretreatment skin biopsy showing the typical

histological pattern of folliculotropic MF, with epidermal cysts

(haematoxylin and eosin; original magnification · 50). Inset shows

epidermotropism of atypical lymphocytes in the cyst wall, forming

Pautrier microabscesses. (c) Fast growing, ulcerated cherry-sized

nodule on the chin. (d) Dense sheet-like nodular infiltrate,

predominantly consisting of large cells with pale cytoplasm and

prominent nucleoli which stained positive for CD30 (inset)


(a) (b)

Fig 2. Folliculotropic mycosis fungoides with CD30+ large-cell

transformation. (a) Three months after electron beam irradiation of

the face (20 Gy), and before initiating therapy with bexarotene

(Targretin�). The cysts persisted almost unchanged in the facial area.

(b) After therapy with bexarotene 150–450 mg daily for 12 months,

total dose of 129Æ98 g. Nearly complete regression of cysts.



Correspondence 585

6 Apisarnthanarax N, Ha CS, Duvic M. Mycosis fungoides with fol-licular mucinosis displaying aggressive tumor-stage transformation:

successful treatment using radiation therapy plus oral bexarotenecombination therapy. Am J Clin Dermatol 2003; 4:429–33.

7 Shistik G, Scalf LA, Fenske N, Glass LF. Follicular mucosis fungoid-es: successful treatment with oral bexarotene. J Drugs Dermatol 2004;

3:301–4.8 Duvic M, Martin AG, Kim Y et al. Phase 2 and 3 clinical trial of oral

bexarotene (Targretin capsules) for the treatment of refractory orpersistent early-stage cutaneous T-cell lymphoma. Arch Dermatol

2001; 137:581–93.

9 Kreuter A, Altmeyer P. Rapid onset of CD8+ aggressive T-celllymphoma during bexarotene therapy in a patient with Sezary

syndrome. J Am Acad Dermatol 2005; 53:1093–5.10 Bouwhuis SA, Davis MDP, el-Azhary RA et al. Bexarotene treatment

of late-stage mycosis fungoides and Sezary-syndrome: developmentof extracutaneous lymphoma in 6 patients. J Am Acad Dermatol 2005;

52:991–6.


Lichen scrofulosorum caused byMycobacterium szulgai: a new cause of atuberculide reaction

DOI: 10.1111/j.1365-2133.2007.07683.x

SIR, A 26-year-old Ghanaian man presented with a 1-year his-

tory of an asymptomatic eruption. On examination there were

uniform 2–3-mm flesh-coloured, predominantly follicular

papules, some of which had central scale as well as a few tiny

pustules (Fig. 1). The papules were scattered on the face,

arms, dorsa of the hands, trunk and legs. There was no pal-

pable lymphadenopathy. He was otherwise completely well

and asymptomatic. Chest X-ray and blood tests including a full blood count

were normal. Histological examination of one of the papules

taken from the dorsum of the hand revealed a superficial

and deep granulomatous folliculitis, with well-formed

noncaseating epithelioid granulomas containing Langhans

giant cells which closely surrounded the hair follicle

(Fig. 2a,b). No acid-fast bacilli were seen on Ziehl–Neelsen

staining, and culture was negative. Amplification by poly-

merase chain reaction (PCR) from the skin biopsy indicated

the presence of mycobacterial DNA with the sequence of

the amplified fragment being that of the atypical myco-

bacterium Mycobacterium szulgai. A Mantoux test was strongly

positive with ulceration. Human immunodeficiency virus

serology was negative.

Treatment was commenced with standard quadruple anti-

tuberculous therapy (isoniazid, rifampicin, pyrazinamide and

ethambutol) and the eruption resolved completely within

6 weeks. After 2 months he ran out of medication, and the

rash recurred within several days. It resolved again when ther-

apy was re-instituted.

(a)

(b)

Fig 2. (a) Skin biopsy of a papular lesion (haematoxylin and eosin;

low power) showing perifollicular granulomatous inflammation and

overlying lichenoid change. (b) Skin biopsy of a papular lesion

(haematoxylin and eosin; high power) showing noncaseating

granulomatous infiltrate.

Fig 1. Clinical appearance of the papular eruption in this patient.



586 Correspondence

The combination of the clinical features of a widespread

flesh-coloured, papular eruption, combined with the granulo-

matous folliculitis and a strongly positive Mantoux test, is

typical of the rare tuberculide reaction known as lichen

scrofulosorum. Tuberculides occur as the result of a vigorous

cell-mediated immune response in the skin to mycobacteria.1

Lichen scrofulosorum is the least frequently reported of

the tuberculide reactions,1 although it may be more common

than previously realised, as in a recent study of 511 patients

with tuberculosis, 39 of them had lichen scrofulosorum.2 It

generally occurs in children or young adults.1,2 Our patient

demonstrated the typical clinical features of flesh-coloured

perifollicular flat-topped or scaly papules, and pustules ran-

ging from 0Æ5 to 3 mm in diameter, particularly occurring

on the trunk.2 The histology is of noncaseating granulomas

surrounding hair follicles and sweat ducts.1 Lichen scrofuloso-

rum is usually associated with M. tuberculosis infection, gener-

ally of lymph nodes or bone rather than the lungs,1,2

although no internal focus of infection is found in 12–28%

of patients.2 Lichen scrofulosorum has also been reported to

occur along with other tuberculide reactions in some

patients.3 In addition to M. tuberculosis, lichen scrofulosorum

can be caused by other mycobacteria, including M. avium4 and

M. bovis (from bacille Calmette–Guerin vaccination).5 Lichen

scrofulosorum has not previously been reported in association

with M. szulgai infection. The response to treatment of tuber-

culous and nontuberculous forms of lichen scrofulosorum is

usually good, with resolution occurring within 1–4 months

of appropriate antimycobacterial treatment, as seen in our

patient.1 Untreated, lesions persist for months or years,

although spontaneous resolution without scarring may occur

eventually.2

In lichen scrofulosorum, mycobacteria are not visible on

microscopy of biopsied skin, and cannot be cultured from

the lesions.1,2 There are reports of identification of mycobac-

terial nucleic acid sequences from biopsied skin in erythema

induratum and papulonecrotic tuberculide using PCR.1,6 This

is the first case of the identification of mycobacterial nucleic

acid sequences from skin lesions in lichen scrofulosorum by

PCR: in two previous cases, PCR did not identify mycobacte-

rial sequences.3,6 The identification of mycobacterial nucleic

acid sequences by PCR despite negative microscopy and cul-

ture for mycobacteria from the skin lesions implies either

that the cell-mediated immune response in the skin keeps the

number of bacilli very low or that the PCR is identifying

fragments of dead mycobacteria only. Mycobacterium szulgai is a

slow-growing scotochromogenic mycobacterium widely dis-

tributed in the environment which rarely causes disease in

humans.7 It has been isolated from water, including aquaria.7

It can be distinguished from other mycobacteria by analysis

of its 16S ribosomal DNA sequence. Mycobacterium szulgai can

cause a lung disease similar to pulmonary tuberculosis, and

these patients are usually,8 but not always, immunosup-

pressed.9 It has also been reported to affect bones and

joints.1,8 Mycobacterium szulgai can cause a skin disease with

sporotrichoid spread of ulcerating lesions10 and subcutaneous

nodules.8 However, it has not previously been reported to

cause any tuberculide reactions. The mycobacterium is sensi-

tive to standard antituberculous agents,1,10 and sometimes

clarithromycin.10

This case highlights the value of PCR on skin biopsies for the

investigation of tuberculide reactions, and the commonality of

pathological responses which can be induced by a wide variety

of mycobacteria in the skin.

Acknowledgments

We thank Dr T. Harrison (Infectious Diseases Unit, St George’s

Hospital) for advice and assistance in managing the patient.

G .L . RO S S

H. CHONG*

T. COL LYNS�D.M. GASCOYNE-B INZ I�

R.P .E . SARKANY�

Skin & Cancer Foundation,

95 Rathdowne Street,

Carlton, Vic. 3053, Australia

Departments of *Cellular Pathology and

�Dermatology, St George’s Hospital,

Blackshaw Road,

London SW17 0QT, U.K.

�Department of Microbiology,

Leeds General Infirmary, Great George Street,

Leeds LS1 3EX, U.K.

Correspondence: Gayle Ross, 37 Union Road,

Surrey Hills, Vic. 3127, Australia.


References

1 Yates VM, Rook GAW. Mycobacterial infections. In: Rook’s Textbook

of Dermatology (Burns DA, Breathnach SM, Cox NH, Griffiths CEM,eds), 7th edn, Vol. 2. Oxford: Blackwell Publishing, 2004; 28.20–

28.35.2 Singal A, Bhattacharya S. Lichen scrofulosorum: a prospective study

of 39 patients. Int J Dermatol 2005; 44:489–93.3 Park YM, Hong JK, Cho SH, Cho BK. Concomitant lichen scro-

fulosorum and erythema induratum. J Am Acad Dermatol 1988;38:841–3.

4 Komatsu H, Terunuma A, Tabata N, Tagami H. Mycobacterium aviuminfection of the skin associated with lichen scrofulosorum: report

of three cases. Br J Dermatol 1999; 141:554–7.5 Park YM, Kang H, Cho SH, Cho BK. Lichen-scrofulosorum-like

eruption localized to multipuncture BCG vaccination site. J Am AcadDermatol 1999; 41:262–4.

6 Torrelo A, Valverde E, Mediero I, Zambrano A. Lichen scrofuloso-rum. Pediatr Dermatol 2000; 17:373–6.

7 Abalain-Colloc ML, Guillerm D, Salaun M et al. Mycobacterium szulgaiisolated from a patient, a tropical fish and aquarium water. Eur J

Clin Microbiol Infect Dis 2003; 22:768–9.8 Cross GM, Guill MA, Aton JK. Cutaneous Mycobacterium szulgai infec-

tion. Arch Dermatol 1985; 121:247–9.9 Gur H, Porat S, Haas H et al. Disseminated mycobacterial

disease caused by Mycobacterium szulgai. Arch Intern Med 1984; 144:1861–3.

10 Kapur N, Schuster H, Parker N et al. Severe sporotrichoid infection

with Mycobacterium szulgai. Clin Exp Dermatol 2004; 29:337–9.




Correspondence 587

Long-term efficacy of topical tacrolimuson oral lesions of chronic graft-versus-hostdisease

DOI: 10.1111/j.1365-2133.2006.07679.x

SIR, Chronic graft-versus-host disease (cGVHD) is a common

complication of allogeneic haematopoietic cell transplantation,

occurring about 100 days postallograft. cGVHD may involve

eyes, liver, intestine, skin and mucous membranes. The oral

mucosa is involved in 80% of patients with cGVHD, with

polymorphous lesions including erythema, erosions, ulcera-

tions, lichenoid striae, atrophy, superficial mucoceles and

xerostomia.1 Treatment of multiorgan cGVHD usually requires

immunosuppressive drugs, especially ciclosporin and cortico-

steroids, but oral cGVHD is often refractory to systemic treat-

ment. Therefore, adjunctive topical treatments may improve

clinical responses and prevent side-effects of systemic drugs,

especially when lesions are limited to the oral mucosa, and

can preserve the graft-versus-leukaemia effect of cGVHD. In

this way, several topical treatments such as corticosteroids

(prednisone, dexamethasone, budesonide), immunosuppres-

sants (ciclosporin, azathioprine) and phototherapy have been

used.2 The efficacy of topical tacrolimus on oral lesions of

cGVHD has recently been reported without evaluation of

recurrence after discontinuation of treatment.3,4 We report

three additional patients with oral and lip cGVHD resistant to

systemic treatments who were treated with tacrolimus oint-

ment, with a long period of follow up after discontinuation of

treatment.

Patient 1. A 41-year-old man allografted for acute myelogenous

leukaemia was treated for cutaneous, intestinal, ocular and oral

cGVHD. Systemic treatment included mycophenolate mofetil,

ciclosporin and budesonide, without local treatment for kera-

totic and ulcerative oral lesions (Fig. 1a). A topical treatment

with clobetasol propionate cream 0.5% and tacrolimus oint-

ment 0.1% three times daily was prescribed. One week later,

ulcers of the lip had resolved (Fig. 1b) while erosions of the

buccal mucosa persisted. However, the patient had not used

topical tacrolimus for the oral lesions as recommended. Two

months later, topical tacrolimus treatment was stopped due

to vomiting. Seven months later, oral and lip ulcers

relapsed despite treatment with clobetasol. Topical tacrolimus

was reintroduced and the lip lesions promptly regressed.

During the treatment, the blood level of tacrolimus was

undetectable.

Patient 2. A 58-year-old man had received allogeneic peripheral

blood progenitor cell transplantation for chronic lymphocytic

leukaemia. Five months after allograft, cGVHD occurred on

the skin, liver, eyes and mouth. Oral lesions were character-

ized by leucokeratosis, oral superficial mucoceles and ulcers of

the tongue, buccal mucosa and lips (Fig. 2a). Oral lesions

were refractory to systemic (corticosteroids, ciclosporin,

mycophenolate mofetil) and topical (prednisone, clobetasol)

treatments. The introduction of topical tacrolimus 0.1%

allowed a rapid clinical improvement in 1 month (Fig. 2b)

without detectable blood levels of tacrolimus. Topical treat-

ment was progressively withdrawn and after 1 year the oral

lesions had not relapsed (Fig. 2c).

Patient 3. A 61-year-old man had been treated for a chronic

myelogenous leukaemia by an allogeneic blood stem cell graft.

Three months after allograft, he developed liver, skin, ocular

and oral cGVHD that was treated with oral corticosteroids and

ciclosporin. Oral lesions consisted of leucokeratosis of the pal-

ate and buccal mucosa, xerostomia, and ulcers of the buccal

mucosa and lips. Topical clobetasol three times daily was pre-

scribed for 8 months, with poor results. This was replaced

with topical tacrolimus 0.1% three times daily. After 15 days

of treatment, the lip ulcers decreased in size. This treat-

ment was applied for 2 months and followed by a progressive

withdrawal over 1 month. One year later, the oral lesions

(a)

(b)

Fig 1. Patient 1. (a) Ulceration of the lip due to chronic

graft-versus-host disease. (b) After 1 week, the ulcers of the lip have

resolved with topical tacrolimus treatment.



588 Correspondence

relapsed. A topical treatment including clobetasol and tacro-

limus was reintroduced, allowing total clearance after

1 month. During the treatment, the blood level of tacrolimus

was undetectable.

Oral lesions of cGVHD are often difficult to manage. Several

small studies2 have reported an improvement of cGVHD oral

lesions using topical therapies (prednisone, dexamethasone,

budesonide, ciclosporin, azathioprine, phototherapy) with

minor side-effects (local burning, bad taste, nausea). How-

ever, the results of these open-label trials with very small

groups must be interpreted with caution. Despite the appar-

ent benefits of adjunctive topical treatments, some patients

are not improved with combined systemic and topical treat-

ments.2 More recently, topical tacrolimus has been proposed

as an effective treatment of cutaneous and oral cGVHD.3–5

Tacrolimus is a macrolide with immunosuppressive properties

due to calcineurin inhibition. The mode of action of tacro-

limus which suppresses the production of local interleukin-2,

interferon-c and tumour necrosis factor-a, all involved in

cGVHD, could explain its efficacy.5 As observed on atopic

skin, topical tacrolimus could penetrate the oral epithelium6

and concentrate in mucosal connective tissue involved in

cGVHD. Furthermore, the cellular entry of tacrolimus is

independent of the glucocorticoid receptor. This could explain

the clinical benefit of tacrolimus in corticosteroid-resistant

patients. Blood levels of tacrolimus were undetectable in our

three patients, as in other studies in oral cGVHD or lichen

planus.7 This result strongly supports a local mechanism of

action.

Follow up after treatment has shown that two patients have

relapsed, after 7 and 12 months, respectively. This suggests

for the first time in cGVHD a benefit of topical tacrolimus in

terms of recurrence but, as has been shown in oral lichen

planus, relapse indicates a palliative and not a curative effect.7

Even though epidemiological studies have not shown an

increase of cutaneous cancer after topical use of tacrolimus the

Food and Drug Administration has recently issued public

health advice to inform about the potential cancer risk based

on a few case reports and animal studies.8 Further long-term

studies are necessary to assess the safety of topical tacrolimus,

especially in the treatment of oral cGVHD, an inflammatory

disease that can lead to squamous cell carcinoma.9,10

J . -C . FR I CA IN

V . S I B AUD*

N. SWETY ENGA�R. TABR I Z I�F . CAMPANA

A. TA I E B*

Departments of Odontology and

�Maxillofacial Surgery, Pellegrin University Hospital,

Bordeaux, France


Saint Andre University Hospital, Bordeaux, France

�Department of Haematology,

Haut Leveque University Hospital, Pessac, France


References

1 Fricain JC, Sibaud V, Hafian H et al. Oral manifestations ofchronic graft-versus-host disease. Ann Dermatol Venereol 2005; 132:

1017–25.2 Imanguli MM, Pavletic SZ, Guadagnini JP et al. Chronic graft versus

host disease of oral mucosa: review of available therapies. Oral SurgOral Med Oral Pathol Oral Radiol Endod 2006; 101:175–83.

3 Sanchez AR, Sheridan PJ, Rogers RS. Successful treatment of orallichen planus like chronic graft-versus-host-disease with topical

tacrolimus: a case report. J Periodontol 2004; 4:613–19.

(a)

(b)

(c)

Fig 2. Patient 2. (a) Ulcer of the lip due to chronic graft-versus-host

disease. (b) Improvement of the lesion after 1 month of treatment

with topical tacrolimus. (c) One year after treatment with topical

tacrolimus, the lesion did not relapse.



Correspondence 589

4 Eckardt A, Starke O, Stadler M et al. Severe oral graft-versus-host-disease following allogeneic bone marrow transplantation: highly

effective treatment with topical tacrolimus. Oral Oncol 2004;40:811–14.

5 Elad S, Or R, Resnick I et al. Topical tacrolimus – a novel treatmentalternative for cutaneous chronic graft-versus-host-disease. Transpl

Int 2003; 16:665–70.6 Gupta AK, Adamiak A, Chow M. Tacrolimus: a review of its use

for the management of dermatoses. J Eur Acad Dermatol Venereol 2002;16:100–14.

7 Olivier V, Lacour JP, Mousnier A et al. Treatment of chronic erosive

oral lichen planus with low concentrations of topical tacrolimus:an open prospective study. Arch Dermatol 2002; 10:1335–8.

8 Ormerod AD. Topical tacrolimus and pimecrolimus and the risk ofcancer: how much cause for concern?. Br J Dermatol 2005;

153:701–5.9 Demarozy F, Soligo D, Lodi G et al. Squamous cell carcinoma of

the oral cavity associated with graft versus host disease: report of acase and review of the literature. Oral Surg Oral Med Oral Pathol Oral

Radiol Endod 2005; 100:63–9.10 Szeto HC, Shek TWH, Lie AKW et al. Squamous cell carcinoma of

the tongue complicating chronic graft-versus-host-disease afterallogeneic hematopoietic stem cell transplantation. Am J Hematol

2004; 77:200–2.


Minocycline-induced pigmentation ofpre-existing capillaritis

DOI: 10.1111/j.1365-2133.2007.07680.x

SIR, Excellent absorption, long half-life and lipophilic proper-

ties inducing good tissue distribution have made minocycline

an effective therapeutic option in the treatment of acne vul-

garis. However, use of minocycline has been associated with

various forms of hyperpigmentation, incidence of which is

well reported in the dermatology and rheumatology literature.

We report a patient with minocycline-induced discoloration

of capillaritis, with no evidence of dyspigmentation elsewhere.

Such localized pigmentation at the site of pre-existing capilla-

ritis appears to be a unique phenomenon, not previously

reported.

A 36-year-old man with insulin-dependent diabetes mellitus

who had received a renal transplant was referred by the renal

physicians for management of worsening nodulocystic acne.

His immunosuppressive cocktail included azathioprine, predn-

isolone and tacrolimus. Dermatological examination showed

moderately severe, tender nodulocystic acne affecting the face,

chest and back. Asymptomatic capillaritis of both lower legs

was also noted. Treatment was commenced with oral mino-

cycline modified release capsules 100 mg daily (Minocin MR�;

Lederle, Maidenhead, U.K.) in conjunction with topical 0Æ01%

tretinoin gel (Retin-A�; Janssen-Cilag, High Wycombe, U.K.).

Eighteen months later, after several defaulted dermatology

appointments the patient presented with slate-blue discoloration

of both lower legs at sites of pre-existing capillaritis (Fig. 1). No

evidence of dyspigmentation was noted at other cutaneous or

mucosal sites, and the patient denied any other symptoms of

minocycline toxicity. As his acne had improved only slightly,

treatment with minocycline was discontinued. This led to a slow

but gradual resolution of the dyspigmentation.

Dyspigmentation is a recognized adverse effect of mino-

cycline which produces three distinct types of cutaneous

pigmentation: type I, blue-black pigmentation confined to

sites of scarring or inflammation on the face; type II,

blue-grey circumscribed pigmentation of normal skin of the

lower legs and forearms; and type III, diffuse muddy brown

pigmentation of normal skin accentuated in sun-exposed

areas.

Mouton et al.1 described a fourth type of minocycline-

induced pigmentation (MIP) in two patients with circum-

scribed blue-grey pigmentation within acne scars confined to

the back. MIP is not confined to the skin and mucosa, but has

also been noted to occur in bone,2 thyroid,3 atherosclerotic

plaques,4 subcutaneous fat5 and heart valves.6

Based on histochemistry, immunohistochemistry, electron

microscopic and energy-dispersive X-ray analyses, it has been

speculated that the pigment may be a drug metabolite–

protein complex chelated with calcium, or an insoluble

minocycline–melanin complex.1 Similar observations have

been made by other authors who noted electron-dense gran-

ules in MIP skin biopsy samples showing the presence of

larger quantities of iron and smaller quantities of sulphur and

calcium.5 Despite these findings, the exact pathogenesis of

MIP remains obscure, along with the explanation for the

observation of preferential involvement of specific tissue sites

with sparing of others. Pigmentation at the site of pre-existing

capillaritis in our patient may be explained by increased red

cell extravasation and perhaps increased minocycline depos-

ition in these areas. Occurrence of MIP on the lower legs may

Fig 1. Minocycline-induced blue-grey pigmentation at the site of

pre-existing capillaritis.



590 Correspondence

be a type II variant, but the involvement of pre-existing capil-

laritis and sparing of normal skin appears to be unique.

V . MADAN

J .T . L E AR


Central Manchester and Manchester Children’s

University Hospital NHS Trust,

Manchester Royal Infirmary,

Oxford Road, Manchester M13 9WL, U.K.


References

1 Mouton RW, Jordaan HF, Schneider JW. A new type of mino-

cycline-induced cutaneous hyperpigmentation. Clin Exp Dermatol2004; 29:8–14.

2 Pandit S, Hadden W. Black pigmentation of bone due to long-termminocycline use. Surgeon 2004; 2:236–7.

3 Birkedal C, Tapscott WJ, Giadrosich K et al. Minocycline-inducedblack thyroid gland: medical curiosity or a marker for papillary

cancer? Curr Surg 2001; 58:470–1.

4 Gerson DM, Robinson M. Black pigmentation of atherosclerotic pla-ques associated with chronic minocycline therapy. Cardiovasc Pathol

2006; 15:168–70.5 Fakhfakh AC, Humbert P, Aubin F. Cutaneous pigmentation induced

by minocycline: ultrastructural analysis and X-ray microanalysis.Ann Dermatol Venereol 1992; 119:975–9.

6 Sant’Ambrogio S, Connelly J, DiMaio D. Minocycline pigmentationof heart valves. Cardiovasc Pathol 1999; 8:329–32.


Anti-desmoglein-1 antibodies are prevalentin Tunisian patients with hydatidosis andleishmaniasis

DOI: 10.1111/j.1365-2133.2006.07687.x

SIR, Pemphigus is an autoimmune blistering disease mediated

by pathogenic antibodies directed against desmoglein (Dsg) 1

and Dsg3.1,2 In Tunisia, pemphigus has particular epidemio-

logical features. It occurs as an endemic form with an inci-

dence rate of 6Æ7 cases per million per year. Case series

suggest that pemphigus, especially the foliaceus form (PF), is

frequent among young Tunisian women with an incidence

rate reaching 20 cases per million in rural areas. This could be

related to a risk factor more prevalent and/or more often

expressed in these areas.3,4

We have reported previously that anti-Dsg1 antibodies were

prevalent in Tunisian healthy subjects, which argues for the

role of a still unknown environmental factor in the occurrence

of Tunisian endemic pemphigus.5 An epidemiological case–

control study demonstrated that several environmental factors

were significantly associated to Tunisian endemic pemphigus

such as traditional cosmetics, Turkish baths and cutting up

raw poultry.6 However, infectious factors have not been yet

investigated.

Recent data demonstrated that in endemic areas of Fogo

selvagem, anti-Dsg1 antibodies were found not only in

patients with PF, but also in patients with parasitic diseases,

thus suggesting a possible involvement of these microorgan-

isms in the aetiopathogenesis of pemphigus.7

In Tunisia, hydatidosis and leishmaniasis are two endemic

parasitoses that are predominant mainly in rural areas where a

higher incidence of Tunisian endemic pemphigus has also

been reported.8 Based on these observations, we aimed to

search for a link between these two parasitoses and Tunisian

pemphigus. For this purpose, we searched for antibodies

against Dsg1 and Dsg3 in sera samples of parasitic patients.

Dsg reactivity was compared with a group of Tunisian normal

subjects.

Serum samples were obtained from 58 patients with parasi-

tic disease, 35 with hydatidosis and 23 patients with visceral

leishmaniasis. We also obtained sera from 59 patients with PF

and 152, age-, sex- and origin-matched, Tunisian normal

individuals. Anti-Dsg1 and anti-Dsg3 antibodies were detected

by a commercial enzyme-linked immunosorbent assay (ELISA)

(MBL, Nagoya, Japan) using recombinant ectodomain of Dsg1

and Dsg3, respectively. The assay method was performed

according to the manufacturer’s instructions. The cut-off

values were set at 14 and 7 for anti-Dsg1 and anti-Dsg3

ELISAs, respectively. Index values ranging from the cut-off

value to 20 were considered as weakly positive. For all posi-

tive sera, we completed with (i) IgG subclass studies by ELISA,

using antihuman IgG1, IgG2, IgG3 and IgG4 murine mono-

clonal antibodies (Sigma, Steinheim, Germany); (ii) indirect

immunofluorescence on monkey oesophagus using a fluo-

rescein-conjugated antihuman IgG antibody (The Binding Site,

Birmingham, U.K.); and (iii) immunoblot analysis using

human epidermal extract as described.9 We also tested patients

with PF and controls for antileishmania and antihydatid fluid

antibodies using an ‘in house’ ELISA as described.10,11 Plates

were coated with hydatid fluid or soluble leishmania extract

and cut-off values were calculated as the mean plus 2 SD of

assay values obtained from Tunisian healthy subjects.

For Dsg1 ELISA, 55 (93Æ22%) patients with PF had positive

values, as expected and observed in other studies. Ten sera

from healthy individuals (6Æ57%) were also positive for anti-

Dsg1 antibodies. In the infectious disease group, we found

positive results in 14 of the 35 hydatidosis sera (40%) and five

of the 23 visceral leishmaniasis sera (21Æ7%). The prevalence

of anti-Dsg1 in these two groups was significantly higher than

that of the control group (P < 10)4 and P ¼ 0Æ04, respect-

ively). Index values were higher than 20 for 11 sera from the

hydatidosis group, and for two sera from the visceral leish-

maniasis group (Fig. 1a).

For Dsg3 ELISA, only three sera from patients with hydati-

dosis (8Æ5%) and four sera from patients with visceral leish-

maniasis (17Æ3%) were positive. The prevalence of antibodies

directed against Dsg3 was significantly higher (P ¼ 0Æ004) in

sera from the visceral leishmaniasis group than the control



Correspondence 591

group (3%). However, all index values were weakly positive

in the visceral leishmaniasis group (Fig. 1b).

IgG subclass studies showed that anti-Dsg1 antibodies found

in patients with hydatidosis were IgG2 in all cases. In two

patients, additional IgG1 and IgG3 anti-Dsg1 antibodies were

also detected. In visceral leishmaniasis patients, IgG3 was the

most common subclass, followed by IgG1 and IgG2. Similar sub-

class distribution was noted for anti-Dsg3 antibodies (Table 1).

Among all positive sera (for anti-Dsg1 and or anti-Dsg3)

from parasitic patients, only two sera, from a patient with

hydatidosis and a patient with visceral leishmaniasis, showed

intercellular staining on monkey oesophagus with a titre of 1:

10. By immunoblot analysis on epidermal extract, only one

serum from a patient with hydatidosis reacted with the

160-kDa band of Dsg1.

The prevalence of antibodies against parasitic antigens in sera

samples from patients with PF did not differ significantly from

the control group. For antihydatid fluid antibodies, similar prev-

alences (4%) were detected in patients with PF and in controls.

A higher prevalence of anti-soluble leishmania extract anti-

bodies was observed (10% in patients with PF vs. 5% in con-

trols); however, the difference was not statistically significant.

In this study, we demonstrated that a significant number

of patients with parasitic diseases, such as leishmaniasis and

hydatidosis, possess anti-Dsg1 and/or anti-Dsg3 antibodies

when tested against the ectodomain of desmogleins using

both highly specific and sensitive antibodies. These anti-

bodies belong to IgG1, IgG2 and IgG3 subclasses but were

never IgG4, which is the most common in Tunisian

patients with pemphigus.5 Recent data demonstrated that in

endemic areas of Fogo selvagem, anti-Dsg1 antibodies are

prevalent in patients with parasitic diseases, thus suggesting

a possible involvement of these microorganisms in the aeti-

opathogenesis of pemphigus.7 In this study, Diaz et al.7

found that 83% of patients with onchocerciasis, 43% of

patients with leishmaniasis and 58% with Chagas disease

have anti-Dsg1 antibodies. Epitope mapping studies revealed

that all positive sera recognized epitopes on the EC5 (extra-

cellular 5) domain of Dsg1. The authors hypothesized that

this domain might contain the initial epitope(s) that trig-

ger(s) the autoimmune response in pemphigus. In our

study, the majority of positive sera gave negative staining

by indirect immunofluorescence and immunoblot analysis.

Several explanations may account for the discrepancy

between results given by ELISA and other immunological

techniques. Firstly, the solid phase ELISA may be more sen-

sitive than other assays for detecting anti-Dsg antibodies.

Secondly, and as reported by Li et al.,12 antibodies against

anti-EC5 gave negative staining by indirect immunofluores-

cence due to the cryptic nature of the EC5 domain. We

can hypothesize that anti-Dsg1 antibodies detected in our

parasitic patients might be directed against cryptic, conform-

ational epitopes on the EC5 domain. Further studies using

an immunoprecipitation assay with recombinant proteins are

needed to confirm this point.

The second finding of this study is that pemphigus

patients had negative serological tests for hydatidosis and

leishmaniasis. Similar findings were also observed in Fogo

selvagem sera, which did not recognize parasitic antigens.

Diaz et al. hypothesized that saliva of the haematophagous

vector containing cadherin proteins could be the sensitizing

antigens and not the parasite itself.7 Although there is no

haematophagous vector for hydatidosis, this parasite could

use ‘cadherin-like’ molecules in the initial phase of invasion

Fig. 1. Analysis of sera from patients with pemphigus foliaceus (n),

hydatidosis (h),visceral leishmaniasis (e) and normal subjects (s) in

ELISA using the recombinant ectodomain of desmoglein (Dsg) 1 (a)

and of Dsg3 (b). The red line shows the cut-off value. The mean

index value of each group is indicated.

Table 1 Reactivity of IgG subclasses with desmoglein 1 and 3 in

patients with hydatidosis and visceral leishmaniasis

IgG subclasses

Hydatidosis

(n ¼ 35)

Visceral leishmaniasis

(n ¼ 23)

Dsg1 ELISAIgG 14 5

IgG1 1 (7%) 1 (20%)IgG2 14 (100%) 1 (20%)

IgG3 1 (7%) 2 (40%)IgG4 0 (0%) 0 (0%)

None 0 (0%) 1 (20%)Dsg3 ELISA

IgG 3 4IgG1 1 (33%) 0 (0%)

IgG2 3 (100%) 1 (25%)IgG3 0 (0%) 3 (75%)

IgG4 0 (0%) 0 (0%)None 0 (0%) 0 (0%)



592 Correspondence

(oncosphere stage) to cross the intestinal barrier. It is of note

that many arguments are in line with a potential role of

hydatidosis in the aetiopathogenesis of pemphigus. Firstly,

these two diseases are endemic in rural areas and occur more

frequently in women. Interestingly, contact with ruminants,

which are hosts for Echinococcus granulosis, was found to be a

risk factor for Tunisian endemic pemphigus.6 Secondly,

hydatidosis seems to offer an adequate microenvironment for

IgG4 isotype switching.13

In conclusion, our results confirm that anti-Dsg1 antibodies

are prevalent in patients with parasitic diseases, suggesting a

possible link of these environmental factors to Tunisian endemic

pemphigus. Additional experiments and follow-up of subjects

involved in the present study are needed to confirm a direct link

between these parasitoses and Tunisian endemic pemphigus.

Acknowledgments

This work was supported by a grant from Le Comite Mixte

Pour La Cooperation Universitaire Franco-Tunisenne. We thank

Pr Louzir H for kindly providing leishmania extract.

M. KAL L E L S E L LAM I*

M. Z I TOUN I*

W. TOMB AR I*

M. BEN AYED�O. AB IDA�

L . LAADHAR*

M. MOKN I�B. FEZ ZA§

H. TURK I–

I . MOKHTAR**

A. BEN OSMAN�R. KAMOUN MOHAMED§

P . JO LY��F . TRON��

D. G I L B E R T��H. MASMOUD I�

S . MAKN I*

(MEMBER S O F TH E ‘FRANCO-TUN I S I AN GROUP OF SURV EY AND

RE S E ARCH ON TUN I S I AN ENDEM I C P EMPH IGU S ’ )

*Laboratoire d’Immunologie and

�Service de Dermatologie,

Hopital La Rabta 1007, Tunis, Tunisia

�Laboratoire d’Immunologie,

Hopital Habib Bourguiba, Sfax, Tunisia

§Service de Dermatologie,

Hopital Charles Nicolle, Tunis, Tunisia

–Service de Dermatologie,

Hopital Hedi Chaker, Sfax, Tunisia

**Service de Dermatologie,

Hopital Habib Thameur, Tunis, Tunisia

��INSERM U519, Institut Federatif

de Recherche Multidisciplinaire

sur les Peptides (IFR23), Faculte Mixte

de Medecine et de Pharmacie, Rouen, France

Correspondence: M. Kallel Sellami


References

1 Koulu L, Kusumi A, Steinberg MS et al. Human autoantibodies

against a desmosomal core protein in pemphigus foliaceus. J ExpMed 1984; 160:1509–18.

2 Ogawa MM, Hashimoto T, Konohana A et al. Immunoblot analysisof Brazilian pemphigus foliaceus antigen using different antigen

sources. Arch Dermatol Res 1990; 282:84–8.3 Morini JP, Jomaa B, Gorgi Y et al. Pemphigus foliaceus in young

women. An endemic focus in the Sousse area of Tunisia. ArchDermatol 1993; 129:69–73.

4 Bastuji Garin S, Souissi R, Blum L et al. Comparative epidemiologyof pemphigus in Tunisia and France: unusual incidence of pemphi-

gus foliaceus in young Tunisian women. J Invest Dermatol 1995;104:302–5.

5 Kallel Sellami M, Ben Ayed M, Mouquet H et al. Anti-desmoglein 1antibodies in Tunisian healthy subjects: arguments for the role of

environmental factors in the occurrence of Tunisian pemphigusfoliaceus. Clin Exp Immunol 2004; 137:195–200.

6 Bastuji Garin S, Turki H, Mokhtar I et al. Possible relation ofTunisian pemphigus with traditional cosmetics: a multicenter case–

control study. Am J Epidemiol 2002; 155:249–56.7 Diaz LA, Arteaga LA, Hilario-Vargas J et al. Anti-desmoglein-1 anti-

bodies in onchocerciasis, leishmaniasis, and Chagas disease suggest

a possible etiological link to Fogo selvagem. J Invest Dermatol 2004;123:1045–51.

8 Ben Rachid MS, Ben Ammar R, Redissi T et al. Geography of majorparasitosis in Tunisia. Arch Inst Pasteur Tunis 1984; 61:17–41.

9 Hashimoto T, Ogawa MM, Konohana A, Nishikawa T. Detection ofpemphigus vulgaris and pemphigus foliaceus antigens by immuno-

blot analysis using different antigen sources. J Invest Dermatol 1990;94:327–31.

10 Maalej IA, Chenik M, Louzir H et al. Comparative evaluation ofELISAs based on ten recombinant or purified Leishmania antigens

for the serodiagnosis of Mediterranean visceral leishmaniasis. Am JTrop Med Hyg 2003; 68:312–20.

11 Rickard MD, Honey RD, Brumley JL, Mitchell GF. Serological diag-nosis and post-operative surveillance of human hydatid disease.

The enzyme-linked immunosorbent assay (ELISA) using variousantigens. Pathology 1984; 16:211–15.

12 Li N, Aoki V, Hans-Filho G et al. The role of intramolecular epitopespreading in the pathogenesis of endemic pemphigus foliaceus

(Fogo selvagem). J Exp Med 2003; 197:1501–10.13 Aceti A, Pennica A, Teggi A et al. IgG subclasses in human hydatid

disease: prominence of the IgG4 response. Int Arch Allergy Immunol1993; 102:347–51.


Needle-free anaesthesia prior to botulinumtoxin type A injection treatment of palmarand plantar hyperhidrosis

DOI: 10.1111/j.1365-2133.2007.07691.x

SIR, Botulinum toxin type A (BTX-A) has emerged as a new

technique to treat intractable cases of focal hyperhidrosis

(HH).1 Injections on the palms and soles, although effective,

are very painful and need some form of anaesthesia. Peripheral

nerve blockade, the standard method of anaesthesia used, may

have serious drawbacks,2 particularly when used repeatedly.

Needle-free anaesthesia may offer an alternative solution.3 The

technique involves the use of needle-free injection of 0Æ02–

0Æ04 mL of 2% lidocaine without adrenaline prior to BTX-A

injection with needle. The mixture of lidocaine and BTX-A

does not jeopardize toxin potency.4

The device used to achieve needle-free anaesthesia is

the MED-JET� MBX (Medical International Technologies,



Correspondence 593

Montreal, QC, Canada; Fig. 1), which has the advantage over

similar devices such as the Dermojet� (Robbins Instruments,

Chatham, NJ, U.S.A.), of providing an adjustable range of vol-

umes and pressures necessary to control the penetration depth.

For palmar HH, the volume is adjusted to 0Æ02 mL per spurt

(range 0Æ01–0Æ1 mL) and the pressure system set to 120 p.s.i.

A first shot of lidocaine is discharged at the centre of the palm

and if a superficial skin weal fails to appear, the pressure is

steadily raised until a weal is obtained. Once the weal is

formed, injections are evenly distributed at 1Æ5–2-cm intervals

over the whole palmar surface including the fingers. For plan-

tar HH, higher volumes (0Æ04 mL) and pressure may be

required to produce an anaesthetic weal. On average, 50 sites

per hand or 30 sites per foot are injected in this manner.

BTX-A is injected through a 31 gauge needle preferably after

each sequence of four or five lidocaine weals (Fig. 2). Injec-

tions are carried out intradermally rather than subcutaneously

in an attempt to reduce the incidence of significant muscular

weakness.5 Universal precautions including full-face shield or

face mask and protective eyeglasses are taken to avoid contam-

ination caused by potential blood splatter. The device has

recently been approved by Health Canada, and U.S. Food and

Drug Administration approval is pending.

Between August 2004 and July 2005, 16 patients have been

treated with this technique. Clinical data were collected

through a questionnaire and are summarized in Table 1. All

Fig 2. Botulinum toxin type A (BTX-A)

injection with 31 gauge needle follows each

sequence of formation of four or five

lidocaine weals with the needle-free device.

Fig 1. The MED-JET� MBX model by Medical

International Technologies Inc. (Montreal,

Canada), a subsidiary of Medical International

Technology U.S.A.



594 Correspondence

patients found the needle-free anaesthesia tolerable and accep-

ted to be reinjected if their treatment was effective.

The injected dose varied between 50 and 200 units of

BTX-A (Botox�; Allergan Inc., Irvine, CA, U.S.A.) recon-

stituted in preserved saline (3–5 mL per vial of 100 units).

Initially 50 or 100 units of BTX-A were injected according

to the size of the injected site and later increased up to a

total of 200 units per site if a satisfactory result was not

obtained. All patients gave written informed consent prior to

commencement of the BTX-A injections. All patients were eli-

gible candidates to receive BTX-A injections after failing to

respond to topical preparations6 and/or iontophoresis.

Patient 1 experienced an unexpected remission of her axil-

lary and plantar HH, following her palmar treatment with

BTX-A. Patients 4 and 6 complained of severe muscle weak-

ness of their hands that persisted for 26 and 12 weeks,

respectively. The time lapse between the injection of BTX-A

and its onset of action was 6 weeks in patients 11 and 15.

Patient 16 was needle phobic: he requested to be treated

with BTX-A directly on his left hand. BTX-A was reconstituted

in 1Æ5 mL preserved saline and the volume per spurt was

adjusted to 0Æ03 mL. This setting provided 2 units of BTX-A

per spurt for a total dose of 100 units. After failing to respond

to this dosage, he was given an additional 50 units of BTX-A

on his left hand, and 150 units on his right hand, this time

with needle-free anaesthesia prior to BTX-A injection with

needle. His treatment remained effective even 6 months after

his initial injections.

The average duration of the effect of BTX-A was 8 months

in this group. When the effect of BTX-A started to fade away,

topical aluminium chloride preparations were used to try to

extend the interval between BTX-A injections. Other side-

effects were limited to mild ecchymoses that faded away in a

week or two.

The majority of recent research has demonstrated less

painful injection with needle-free jet devices compared with

traditional needle delivery.7 Needle-free devices are used to

inject all types of liquid: anaesthetics, corticosteroids and even

BTX-A.

Dermojet� has been successfully used to inject BTX-A for

plantar HH8 but the authors found the method inadvisable for

palmar HH because of possible injury to the superficial palmar

nerves or vessels. Using smaller volumes and lower pressure

would certainly control the desired penetration depth and mini-

mize this type of injury. Further studies using objective evalu-

ation methods are warranted on a larger number of patients.

A . BENOHAN I ANDermatology Division, Department of Medicine,

University of Montreal,

Montreal, Quebec, Canada


References

1 Lowe N, Campanati A, Bodokh I et al. The place of botulinum toxin

type A in the treatment of focal hyperhidrosis. Br J Dermatol 2004;151:1115–22.

2 Hayton MJ, Stanley JK, Lowe NJ. A review of peripheral nerveblockade as local anaesthesia in the treatment of palmar hyperhidro-

sis. Br J Dermatol 2003; 149:447–51.3 Benohanian A. Surgical pearl: use of needle-free anesthesia in the

treatment of palmar hyperhidrosis with botulinum A toxin. J Am

Acad Dermatol 2005; 52:1073–4.4 Gassner HG, Sherris DA. Addition of an anesthetic agent to

enhance the predictability of the effects of botulinum toxin type Ainjections: a randomized controlled study. Mayo Clin Proc 2000;

75:701–4.5 Heckmann M, Schaller M, Plewig G, Ceballos-Baumann A. Optimi-

zing botulinum toxin therapy for hyperhidrosis. Br J Dermatol 1998;138:553–4.

Table 1 Demographic details and results of 16 patients treated with botulinum toxin type A for palmar or plantar hyperhidrosis

Patient

Sex/age

(years)

Injection

date Site

Units

per side

Efficacy

in months

Muscle weakness

duration (weeks)

Willingness to

be reinjecteddespite muscle

weakness

1 F/46 6 Aug 04 Hands 200 12 2 Yes2 M/44 10 Aug 04 Hands 200 15 0

3 M/31 10 Aug 04 Hands 200 15 04 F/45 10 Aug 04 Hands 160 15 26 Yes

5 F/24 3 Sep 04 Hands 100 Failure 06 F/45 28 Sep 04 Hands 150 4 12 No

7 M/33 28 Sep 04 Hands 150 3 2 Yes8 F/27 5 Oct 04 Hands 100 13 0

9 M/24 5 Oct 04 Hands 100 5 010 M/30 23 Nov 04 Hands 50 12 0

11 F/32 14 Dec 04 Hands 50 11 012 F/53 1 Feb 05 Feet 100 8 0

13 M/33 1 Feb 05 Feet 100 7 0

14 F/42 1 Mar 05 Hands 100 3 12 Yes15 F/44 5 Apr 05 Feet 100 4 0

16 M/32 14 Jul 05 Hands 150 > 5 2 Yes



Correspondence 595

6 Benohanian A. Antiperspirants and deodorants. Clin Dermatol 2001;19:398–405.

7 Ellis GL, Owens A. The efficacy and acceptability of using a jetinjector in performing digital blocks. Am J Emerg Med 1993; 11:648–

50.8 Vadoud-Seyedi J. Treatment of plantar hyperhidrosis with botulinum

toxin type A. Int J Dermatol 2004; 43:969–71.


Increased human papillomavirus type 31 DNAload in a verrucous high-grade intraepithelialneoplasia of a human immunodeficiencyvirus-infected patient with extensive bowenoidpapulosis

DOI: 10.1111/j.1365-2133.2007.07690.x

SIR, Bowenoid papulosis (BP) represents a multifocal intraepi-

thelial neoplasia primarily located in the anogenital region of

predominantly young adults.1 Although BP shows clinically

benign-looking papular lesions, histopathological findings

reveal features of a squamous cell carcinoma (SCC) in situ, and

high-risk human papillomaviruses (HPVs), mainly HPV 16,

are regularly found in the lesions. BP usually has a benign

course and tends to regress spontaneously, in contrast to

Bowen’s disease (BD), which develops into SCC after a certain

duration of time. Malignant transformation has so far only

sporadically been reported in BP. Immunosuppression, espe-

cially human immunodeficiency virus (HIV) infection, greatly

increases the risk for premalignant and invasive HPV-related

conditions.2

We report a 57-year-old HIV-infected heterosexual man

with extensive genital BP. He was diagnosed as HIV-1 positive

in May 2005, presenting with a severely compromised immuno-

logical and virological HIV status (CD4+ cells 67 lL)1; HIV

RNA 244 000 copies mL)1). At first physical examination, he

revealed widespread brownish papules with a scaly or smooth

surface located on the mons pubis, genitals, inguinal folds,

buttocks, and the inner aspects of the thighs (Fig. 1a). Con-

comitant Candida albicans infection of the intertriginous areas

was present. Furthermore, a wart-like, verrucous nodule at the

perineal area, 8 · 8 mm in diameter, was detected that dif-

fered clinically from the other BP lesions (Fig. 1b). Histopatho-

logical examination of the nodule revealed acanthosis,

parakeratosis, and atypical cells within the entire epidermis

consistent with SCC in situ. Immunohistochemical staining of

the nodule using anti-p16 monoclonal antibody demonstrated

strong and diffuse immunoreactivity in both the nuclei and

the cytoplasm. p16 staining of the nodule reached the upper

layers of the epidermis (Fig. 2) and exceeded that of the BP

lesions (Table 1). HPV typing for 37 different low- and high-

risk HPVs from the two different types of lesions (the verru-

cous nodule shown in Fig. 1b and the widespread brownish

papules shown in Fig. 1a) showed identical high-risk HPV

types in both types of lesions (HPV 31 and 52). The low-risk

HPV 84 was present only in the brownish papules (Table 1).

Further genital low- or high-risk HPV types were not detected.

Interestingly, HPV 31 loads (quantitated by real-time polymer-

ase chain reaction with type-specific primers and probes and

expressed as HPV DNA copies per b-globin gene copy3) sig-

nificantly differed in the two types of lesions: HPV 31 load of

the verrucous nodule was two orders of magnitude higher

than that in the brownish papules (Table 1). Semiquantitative

viral load determination for HPV 52 and HPV 84 revealed

similar loads in the two brownish papules, and a lower HPV

52 load in the warty nodule compared with the brownish

papules (Table 1). The nodule was surgically removed and all

other lesions were treated with cautery fulguration. Imiqui-

mod 5% cream therapy was offered to prevent recurrences,

but was declined by the patient. He finally refrained from

therapy and was lost to follow up.

Fig 1. (a) Widespread brownish papules with

a scaly or smooth surface located on the mons

pubis, genitals, inguinal folds, and inner

aspects of the thighs. (b) Singular, wart-like

nodule located on the perineum.



596 Correspondence

In the past, BP was mainly considered to be a benign

condition despite distinctive histopathological findings of

intraepithelial neoplasia, and no malignant transformation

was observed in the long-term follow up of a large collective

of patients with BP.4 However, development of SCC from BP

was reported recently, and most of these patients had signs

of immunosuppression. Although HPV 16 is found in the

majority of patients with BP, five cases of HPV 31-associated

BP have been reported in the literature so far, and two of

them transformed into SCC.5,6 Interestingly, elevated amounts

of HPV 31 were detected in the SCC as compared with other

BP lesions in the case of a woman with pancytopenia.6 This

finding is in line with our present patient, in whom an

exceedingly high HPV 31 DNA load was found in a singular

warty lesion as compared with the surrounding BP lesions.

Furthermore, immunohistochemical analysis of p16, a marker

that demonstrates HPV oncogene expression independent of

the infecting HPV type,7 revealed clearly more p16 positivity

within the epidermis of the warty lesion (Table 1). Besides

integration (which could occur alongside the high HPV 31

load), increased HPV oncogene expression could be

explained by mutations in the E6/E7 promotor or simply by

the large numbers of viral genomes present (elevated gene

dose). It is tempting to speculate that SCC transformation

could have occurred without treatment in this patient, and

that the elevated HPV 31 load is related to the advanced

grade of intraepithelial neoplasia. For cervical intraepithelial

neoplasia (CIN) an association between increased high-risk

HPV DNA loads and advanced CIN is established.8 Moreover,

our virological findings could corroborate the possibility that

both BP and BD are part of a spectrum of the same disease.

Considering the clinical features of the verrucous nodule one

might speculate that a common genital wart developed

within an area of dysplasia. However, no HPV types associ-

ated with genital warts were detectable within the lesion

(Table 1). We recently found elevated high-risk HPV DNA

loads in cervical high-grade squamous intraepithelial lesions

of HIV-positive women, compared with HIV-negative

women.3 Moreover, we observed high incidences of anal

intraepithelial neoplasia in HIV-positive homosexual men.2

Accordingly, all HIV-infected patients should be carefully

Fig 2. Immunohistochemical analysis of the squamous cell carcinoma

in situ (warty nodule) using anti-p16 monoclonal antibody

demonstrates a strong and diffuse immunoreactivity for p16 in both

the nuclei and cytoplasm, including also the upper layers of the

epidermis (original magnification · 40).

Table 1 Patient’s virological and histological features

Site of the lesion Clinical feature HPV typesaHPV 31

DNA loadb,cHPV 52

DNA loadcHPV 84

DNA loadc Histologydp16

staininge

Mons pubis Brownish papule 31, 52, 84 26b

1 : 10 000c1 : 10 000c 1 : 10c II 50%

Thigh (left) Brownish papule 31, 52, 84 51b

1 : 1000c1 : 10 000c 1 : 10c I 20%

Perineal area Warty nodule 31, 52 5735b

1 : 1 000 000c1 : 1000c – III 70%f

HPV, human papillomavirus. aHPV typing was performed as described previously.2,9 Probes used for enzyme immunoassay (EIA) hybridiza-

tion were described previously.10 bQuantitative HPV DNA load determination was performed by real-time polymerase chain reaction (PCR)with type-specific primers and probes as described previously.3 HPV 31 DNA load is expressed as HPV 31 DNA copies per b-globin gene

copy. cSemiquantitative HPV DNA load determination was performed by serial dilution (from 1 : 10 to 1 : 100 000 000) of extractedsample DNA in Tris–ethylenediamine tetraacetic acid buffer containing human placental DNA (40 ng lL)1). HPV group-specific PCR2 of the

diluted samples was performed and PCR products were hybridized to type-specific probes.9,10 The last dilution that gave a positive signal inthe EIA-based hybridization9 with type-specific probes10 for HPV 31, 52 and 84, respectively, is stated. dGrade of intraepithelial neoplasia.ep16 staining results are expressed as the percentage of positive cells per field. fReaching the upper third of the epithelium.



Correspondence 597

screened for HPV-related disorders, including BP, and should

be treated as early as possible, as they have a higher risk of

malignant transformation, widespread disease, and frequent

recurrences.

Acknowledgments

We thank Monika Junk, Barbara Panz and Sabine Richter

for excellent technical assistance. This study was supported

by the Federal Ministry of Education and Research, German

Competence Network HIV/AIDS, Grant No. 01 KI 0501.

A . KREUT ER*

N.H. BROCKMEY ER*

H. P F I S T E R�P . ALTMEY ER*

U. WIE LAND�FOR THE GERMAN COMPE T ENC E

NETWORK HIV/AIDS

*Department of Dermatology and Allergology,

Ruhr-University Bochum, Gudrunstr.

56, 44791 Bochum, Germany

�Institute of Virology, University of Cologne,

Cologne, Germany


References

1 Wade TR, Kopf AW, Ackerman AB. Bowenoid papulosis of the

genitalia. Arch Dermatol 1979; 115:306–8.2 Kreuter A, Brockmeyer NH, Hochdorfer B et al. Clinical spectrum

and virologic characteristics of anal intraepithelial neoplasia in HIVinfection. J Am Acad Dermatol 2005; 52:603–8.

3 Weissenborn SJ, Funke AM, Hellmich M et al. Oncogenic humanpapillomavirus DNA loads in human immunodeficiency virus-posi-

tive women with high-grade cervical lesions are strongly elevated.J Clin Microbiol 2003; 41:2763–7.

4 Patterson JW, Kao GF, Graham JH, Helwig EB. Bowenoid papulo-sis: a clinicopathologic study with ultrastrucural observations. Cancer

1986; 57:823–36.5 Yoneta A, Yamashita T, Jin HY et al. Development of squamous cell

carcinoma by two high-risk human papillomaviruses (HPVs),a novel HPV-67 and HPV-31 from bowenoid papulosis. Br J Dermatol

2000; 143:604–8.6 Hama N, Ohtsuka T, Yamazaki S. Elevated amount of human papil-

lomavirus 31 DNA in a squamous cell carcinoma developed frombowenoid papulosis. Dermatology 2004; 209:329–32.

7 Sano T, Oyama T, Kashiwabara K et al. Expression status ofp16 protein is associated with human papillomavirus oncogenic

potential in cervical and genital lesions. Am J Pathol 1998; 153:1741–8.

8 Snijders PJ, Hogewoning CJ, Hesselink AT et al. Determination ofviral load thresholds in cervical scrapings to rule out CIN 3 in

HPV 16, 18, 31 and 33-positive women with normal cytology.

Int J Cancer 2006; 119:1102–7.9 Jacobs MV, Snijders PJ, van den Brule AJ et al. A general primer

GP5+/GP6(+)-mediated PCR-enzyme immunoassay method forrapid detection of 14 high-risk and 6 low-risk human papillo-

mavirus genotypes in cervical scrapings. J Clin Microbiol 1997; 35:791–5.

10 van den Brule AJ, Pol R, Fransen-Daalmeijer N et al. GP5+/6+ PCRfollowed by reverse line blot analysis enables rapid and high-

throughput identification of human papillomavirus genotypes.J Clin Microbiol 2002; 40:779–87.


Telangiectatic cutaneous metastasis fromcarcinoma of the prostate

DOI: 10.1111/j.1365-2133.2006.07696.x

SIR, Carcinoma of the prostate is the most common cancer in

men, but is only rarely associated with metastases to the skin.1

Cutaneous metastases, when they occur, are usually nodular.2

Telangiectatic cutaneous metastases are very rare and are usu-

ally described in patients with breast cancer. There is a single

report in the literature of a telangiectatic cutaneous metastasis

from prostate cancer.3 We describe a second patient develop-

ing a telangiectatic cutaneous metastasis 12 years after diagno-

sis with prostate cancer.

An 83-year-old man with a 12-year history of prostate can-

cer metastatic to bone presented with a 3-month history of a

slowly enlarging patch on his left chest wall. The lesion was

mildly pruritic and occasionally associated with a burning sen-

sation. He denied any trauma or application of topical agents

to the area and had not had any similar lesions in the past. He

had received various treatments for his prostate cancer, includ-

ing goserelin acetate, bicalutamide, flutamide and radiation

therapy to the pelvis. At presentation, he was receiving palli-

ative radiation therapy and pamidronate.

The patient appeared weak and cachectic. On physical

examination he had a nontender, nonblanching, violaceous

and telangiectatic patch located anterior to the left axilla on

the chest wall (Fig. 1a). There were no other skin findings.

Pertinent laboratory studies included a prostate-specific anti-

gen (PSA) level of 227 ng L)1 (normal range, 0–4 ng mL)1),

which had risen from 95Æ8 ng mL)1 in the 2 months pre-

ceding his presentation with the skin lesion. A full blood

count and chemistry profile were within normal limits. A skin

biopsy from the patch showed intravascular aggregates of PSA-

positive adenocarcinoma within ectatic papillary dermal blood

vessels (Fig. 1b, c). The patient died from disease 6 months

after presenting with his cutaneous metastasis.

In the U.S.A., carcinoma of the prostate is the most com-

mon form of cancer in men and the second most common

cause of cancer-related mortality. It is estimated that 300 000

new cases are detected each year, of which roughly 41 000

(13Æ7%) will ultimately prove to be fatal.1 Carcinoma of the

prostate preferentially metastasizes to bone, lung, liver and

adrenal glands, whereas skin metastasis is considered excep-

tional and accounts for fewer than 1% of cutaneous metasta-

ses.2 In a retrospective study of 4020 patients with metastatic

disease, Lookingbill et al.4 found that 420 (10%) had cutane-

ous metastases. Most arose from primary tumours in the lung,

large intestine and oral cavity. None of the cutaneous metasta-

ses was of prostatic origin.4

When prostate carcinoma does metastasize to the skin, the

metastases usually appear as multiple or solitary nodules on

the abdominal skin, in the suprapubic area, and on the



598 Correspondence

anterior aspect of the thighs.2,5,6 Less frequently, the metasta-

ses involve other sites and may not have the typical nodular

morphology (Table 1).3,5,7–23 A telangiectatic cutaneous meta-

stasis from prostatic carcinoma has been described only once

in the literature. More frequently, metastases with this mor-

phology arise from the breast, the parotid gland, or in associ-

ation with apocrine carcinoma.3,24–27

Clinically, telangiectatic cutaneous metastases may demon-

strate telangiectasia, or erythema, or show purpuric papules,

or lymphangioma circumscriptum-like pseudovesicles.24–27

The single reported case of telangiectatic metastatic prostate

cancer presented as a plaque resembling angiosarcoma of the

scalp.3 Histological examination demonstrates ectatic, often

engorged, and predominantly superficial dermal blood vessels

that contain erythrocytes and aggregates of neoplastic cells

that frequently appear ‘free-floating’ within the vascular

lumina.24,26,28 Metastases from a prostatic primary carcinoma

have the typical histological features of prostatic adenocarci-

noma and are immunoreactive for both PSA and prostatic acid

phosphatase.2

Table 1 Atypical clinical presentations of cutaneous metastatic prostatic carcinoma

First author, year, reference Morphology Site Atypical features

Azana, 19937 Nodule Neck

Bluefarb, 19578 Nodules Abdomen Zosteriform distributionBoswell, 20053 Violaceous plaque Forehead Resembling angiosarcoma

Cox, 19949 Inflammatory papules, plaques Thigh, inguinal Resembling cellulitis; histopathology:lymphatic infiltration

Landow, 198010 Nodules Cutaneous lip, scalpLandow, 198010 Nodule Supraclavicular

Marcoval, 199811 Papulonodules Nipple and periareolar Resembling mammary Paget’s diseaseMueller, 20045 Nodular plaques Chest

Ng, 200012 Inflammatory plaque Thigh Resembling cellulitis; histopathology:lymphatic infiltration

Offidani, 199713 Nodules Scalp, chestPeison, 197114 Nodule Scalp Resembling sebaceous cyst

Pieslor, 198615 Nodule Umbilicus Resembling Sister Joseph’s noduleRazvi, 197516 Nodules Nose, abdomen

Reingold, 196617 Nodule ScalpRonchese, 194018 Nodules Scalp Resembling turban tumours

Rossetti, 199119 Nodule Chest Resembling basal cell carcinomaSchellhammer, 197320 Nodules Scalp

Sharma, 200521 Papulonodules Face, neck, chest Multiple nodules resemblingtrichoepitheliomas

Stahl, 198022 Ulcerated nodules Dorsal hands, forearms Resembling pyodermaWhitmore, 198623 Indurated, erythematous plaque Chest Resembling morphoea

(a) (b) (c)

Fig 1. (a) Clinical appearance of a nontender, nonblanching, violaceous and telangiectatic patch located anterior to the left axilla on the chest

wall. (b) Prostate-specific antigen-positive tumour masses within dilated blood vessels (original magnification · 200). (c) Tumour and thrombus

within dilated blood vessel (haematoxylin and eosin; original magnification · 200).



Correspondence 599

Prostate carcinoma may metastasize via four mechanisms:

local extension from an underlying tumour, implantation

within a surgical scar, lymphatic spread and haematogenous

spread.29 Cutaneous metastases at sites distant from the pri-

mary tumour are thought to result from haematogenous

dissemination.29 Our case, with intravascular aggregates of

PSA-positive tumour cells, seems to corroborate this hypoth-

esis. It is possible that the patch of telangiectatic carcinoma

represents in-transit metastasis that would have preceded

eventual involvement of more distant sites such as the face

or scalp.

Cutaneous metastasis of prostate carcinoma, regardless of

how it manifests, is a marker of advanced disease and por-

tends a grave prognosis, with most patients dying from their

disease within 1 year of presentation of cutaneous findings.29

Acknowledgments

We are grateful to Dr Beverly Faulkner-Jones for proofreading.

S . REDDY

R .H. BANG*

M.E . CONTR E RA S*

Department of Dermatology, University of Texas

Health Science Center, San Antonio, TX, U.S.A.

*Department of Dermatology, University of

New Mexico, Albuquerque, NM, U.S.A.

Correspondence: R.H. Bang.


References

1 Boring CC, Squires TS, Tong T et al. Cancer statistics, 1994. CA

Cancer J Clin 1994; 44:7–26.2 Duran EP, Paradela A, Farina MC. Cutaneous metastases from pro-

static carcinoma. J Surg Oncol 1996; 62:144–7.3 Boswell JS, Davis MDP. Violaceous plaque on the forehead clinically

resembling angiosarcoma: cutaneous metastasis in a patient withprostatic adenocarcinoma. J Am Acad Dermatol 2005; 53:744–5.

4 Lookingbill DP, Spangler N, Helm KF. Cutaneous metastases inpatients with metastatic carcinoma: a retrospective study of 4020

patients. J Am Acad Dermatol 1993; 29:228–36.5 Mueller TJ, Wu H, Greenberg RE et al. Cutaneous metastases from

genitourinary malignancies. Urology 2004; 63:1021–6.6 Steinkraus V, Lange T, Abeck D et al. Cutaneous metastases from

carcinoma of the prostate. J Am Acad Dermatol 1994; 31:665–6.7 Azana JM, de Misa RF, Gomez MI et al. Cutaneous metastases from

prostatic cancer. J Dermatol 1993; 20:786–8.

8 Bluefarb SM, Wallk S, Gecht M. Carcinoma of the prostate withzosteriform cutaneous metastases. Arch Dermatol 1957; 76:402–7.

9 Cox SE, Ponciano D, Cruz JR. A spectrum of inflammatory metasta-sis to skin via lymphatics: three cases of carcinoma erysipeloides.

J Am Acad Dermatol 1994; 30:304–7.10 Landow RK, Rhodes DW, Bauer M. Cutaneous metastases: report of

two cases of prostate cancer. Cutis 1980; 26:399–401, 409.11 Marcoval J, Moreno A, Jucgla A et al. Prostatic adenocarcinoma with

cutaneous metastases overlying oestrogen-induced gynaecomastia.Clin Exp Dermatol 1998; 23:119–20.

12 Ng CS. Carcinoma erysipeloides from prostate cancer presenting ascellulitis. Cutis 2000; 65:215–16.

13 Offidani A, Simonetti O, Cellini A et al. Skin metastases from prostatecancer associated with malignant melanoma. Cutis 1997; 59:278–80.

14 Peison B. Metastasis of carcinoma of the prostate to the scalp. Simula-tion of a large sebaceous cyst. Arch Dermatol 1971; 104:301–3.

15 Pieslor PC, Hefter LG. Umbilical metastases from prostatic carci-noma – Sister Joseph nodule. Urology 1986; 27:558–9.

16 Razvi M, Firfer R, Berskson B. Occult transitional cell carcinoma ofthe prostate presenting as skin metastasis. J Urol 1975; 113:734–5.

17 Reingold IM. Cutaneous metastases from internal carcinoma. Cancer1966; 19:162–8.

18 Ronchese F. Metastases of the scalp mimicking turban tumours.Arch Dermatol Syph 1940; 41:639–48.

19 Rossetti RB, Villaca Neto CM, Paschoal LH, Burnier Junior M.

Cutaneous metastasis originating from prostate adenocarcinoma.Int J Dermatol 1991; 30:363.

20 Schellhammer PF, Milsten R, Bunts RC. Prostatic carcinoma withcutaneous metastases. Br J Urol 1973; 45:169–72.

21 Sharma R, Chandra M. Cutaneous metastases from carcinoma ofthe prostate: a case report. Dermatol Online J 2005; 11:24.

22 Stahl D, Veien NK. Cutaneous metastases simulating other derma-toses. Cutis 1980; 26:282–4.

23 Whitmore WF Jr. Stage A prostate cancer. J Urol 1986; 136:883.24 Lin JH, Lee JY, Chao SC, Tsao CJ. Telangiectatic metastatic breast

carcinoma preceded by en cuirasse metastatic breast carcinoma. Br JDermatol 2004; 150:523–4.

25 Dobson CM, Tagore V, Myint AS, Memon A. Telangiectatic meta-static breast carcinoma in face and scalp mimicking cutaneous

angiosarcoma. J Am Acad Dermatol 2003; 48:635–6.26 Zanca A, Ferracini U, Bertazzoni MG. Telangiectatic metastasis from

ductal carcinoma of the parotid gland. J Am Acad Dermatol 1993;28:113–14.

27 Kiyohara T, Kumakiri M, Kawami K et al. Apocrine carcinoma ofthe vulva in a band-like arrangement with inflammatory and tel-

angiectatic metastasis via local lymphatic channels. Int J Dermatol2003; 42:71–4.

28 Pickard C, Callen JP, Blumenreich M. Metastatic carcinoma of thebreast: an unusual presentation mimicking cutaneous vasculitis.

Cancer 1987; 59:1184–6.29 Powell FC, Venencie PY, Winkelman RK. Metastatic prostate

carcinoma manifesting as penile nodules. Arch Dermatol 1984;120:1604–6.


Isolated recurrence of vesicobullousincontinentia pigmenti in a schoolgirl

DOI: 10.1111/j.1365-2133.2006.07700.x

SIR, The inflammatory vesicobullous first stage of incontinentia

pigmenti (IP) can recur, albeit rarely. It has previously been

described only in infants and young children under the age of

18 months and it has usually been preceded by a systemic

infection.

A 5-year-old schoolgirl presented with a 2-day history of a

linear rash on her right leg extending from the medial knee to

the buttock and consisting of vesicobullous lesions arising on

inflamed skin (Fig. 1). There was no history of anteceding



600 Correspondence

systemic illness or trauma. In addition, there were symmetrical

linear white atrophic streaks on her trunk and limbs, as a con-

sequence of her neonatal involvement with the same rash.

Swabs for bacterial and viral culture were negative and

C-reactive protein was normal. She was known to have

maternally inherited IP and had first presented 2 days after

birth with a whorled vesicobullous erythematous eruption on

both legs and the trunk. She had subsequently remained well,

with no complications.

A diagnosis of isolated recurrence of vesicobullous IP was

made. The rash was treated with once daily application of a

potent steroid and settled after a fortnight, in keeping with

the usual course in which the acute skin lesions last between

2 weeks and 4 months.1 Unilateral presentation of cutaneous

IP is rare and recurrence in such a distribution has not been

reported. The vesicobullous nature of the lesions and their

rapid resolution distinguished the rash from the differential of

lichen striatus, in which lesions are usually papular, reach

their maximal extent in 2 weeks and tend to resolve after

3–6 months.

In IP the mutated nuclear essential modulator (NEMO)

gene encodes a nonfunctional protein that is unable to acti-

vate nuclear factor-jB, which protects against tumour necro-

sis factor (TNF)-induced apoptosis. Most cases of IP have a

single acute episode in which most NEMO mutated cells are

destroyed. Subsequent infections, associated with the produc-

tion of TNF-a, can trigger apoptosis in residual NEMO

mutated cell lines.1 Reactivation can also occur after a phys-

ical insult such as laser2 or liquid nitrogen therapy.3 Previ-

ously reported cases of recurrence can be divided into two

groups: (i) infants and children under the age of 18 months,

with vesicobullous lesions in previously spared skin, with

five of the eight reported cases associated with a preceding

trigger; and (ii) school-age children or adults, with areas of

whorled erythema, without vesiculation, localized to previ-

ously affected areas, less predictably associated with an ante-

ceding trigger (Table 1).

Our patient uniquely re-presented with asymmetrical

vesicobullous lesions of IP at the age of 5 years in the absence

of a systemic illness. It is difficult to understand how her sig-

nificant number of residual NEMO mutated cells had been

Table 1 Reported cases of reactivation of

incontinentia pigmenti

Study Age (years)

Number of

recurrences

Preceding systemic

trigger

Presence of

vesicles

Bodak et al.1 < 1 One Yes YesBodak et al.1 < 1 Two Yes Noa

Bodak et al.1 < 1 Several Yes (all) YesBodak et al.1 < 1 Several Some Yes

van Leeuwen et al.4 < 1 One No YesSahn and Davidson5 < 1 Several No Yes

Llombart et al.6 1Æ5 Several Yes Noa

Patrizi et al.7 1Æ5 One No Yes

This case 5 One No YesBodak et al.1 7 Several Some Noa

Pfau and Landthaler8 7 Several Yes (all) Noa

Barnes9 Adult One No Noa

Bessems et al.10 Adolescence Several No Noa

aWhorled demarcated erythema only.

Fig 1. Vesicobullous eruption arising on inflamed skin of right

medial thigh.



Correspondence 601

spared apoptosis for so long, and the mechanism responsible

for triggering the vesicobullous eruption at this late age.

We report this case to highlight the possibility of unilateral

reactivation of the vesicobullous stage of IP several years after

the initial presentation, in the absence of a systemic illness,

and to emphasize the importance of obtaining a neonatal

and family history to explain a vesicobullous eruption in

childhood.

S . DARN E

A . J . CARM I CHA E L

Department of Dermatology, The James Cook

University Hospital, Middlesbrough TS4 3BW,

U.K.


References

1 Bodak N, Hadj-Rabia S, Hamel-Teillac D et al. Late recurrenceof inflammatory first-stage lesions in incontinentia pigmenti.

Arch Dermatol 2003; 139:201–4.2 Nagase T, Takanashi M, Takada H, Ohmore K. Extensive vesiculo-

bullous eruption following limited ruby laser treatment for incon-tinentia pigmenti: a case report. Australas J Dermatol 1997; 58:155–7.

3 Gatalano RA, Lopatynsky M, Tasman WS. Treatment of proliferativeretinopathy associated with incontinentia pigmenti. Am J Ophthalmol

1990; 110:701–2.

4 van Leeuwen RI, Wintzen M, van Praag MCG. Incontinentia pig-menti: an extensive second episode of a ‘first-stage’ vesicobullous

eruption. Pediatr Dermatol 2000; 17:70.5 Sahn EE, Davidson LS. Incontinentia pigmenti: three cases with

unusual features. J Am Acad Dermatol 1994; 31:852–7.6 Llombart B, Garcia L, Monteagudo C et al. Incontinentia pigmenti:

a case with an unusual course. J Eur Acad Dermatol 2005; 19:394–6.

7 Patrizi A, Neri I, Guareschi E, Cocchi G. Bullous recurrent eruptionof incontinentia pigmenti. Pediatr Dermatol 2004; 21:613–14.

8 Pfau A, Landthaler M. Recurrent inflammation in incontinentiapigmenti of a seven-year-old child. Dermatology 1995; 191:161–3.

9 Barnes CM. Incontinentia pigmenti. Report of a case with persistentactivity into adult life. Cutis 1978; 22:621–4.

10 Bessems PJM, Jagtman BE, van de Staak WJB et al. Progressivepersistent hyperkeratotic lesions in incontinentia pigmenti. Arch

Dermatol 1998; 124:29–30.


Infantile systemic hyalinosis: a case reportand mutation analysis in a Chinese infant

DOI: 10.1111/j.1365-2133.2006.07701.x

SIR, Infantile systemic hyalinosis (ISH; MIM 236490) is a rare

autosomal recessively inherited disease characterized by the

deposition of amorphous hyaline material in various tissues.

In 2003, Hanks et al.1 and Dowling et al.2 first reported muta-

tions in the gene encoding capillary morphogenesis protein 2

(CMG2) to be the cause of ISH. We present a Chinese patient

with ISH with two heterozygous mutations in CMG2.

This patient was the first child of healthy nonconsanguine-

ous Chinese parents. He was born at term by caesarean section

with a body weight of 2650 g. One month after birth he

developed violaceous papules on his face and refractory diar-

rhoea. Flexion contracture of the large joints gradually

appeared, and as a result he was lying in a leech-like position.

The bone of his left leg fractured after intensive physiotherapy

at 6 months of age. Violaceous plaques and hypertrophy of

the skin in his anogenital area spread slowly after 10 months

of age. Gingival hypertrophy began from 11 months. There

was no similar disease in his parents’ families.

Physical examination showed papules and nodular rashes

without clear demarcation on his face, back and anogenital area.

The consistency of the skin lesions varied from soft to hard.

There were also livid-red macules over bony prominences such

as the elbow and spine. His skin was generally oedematous and

thick. Pigmentation over the ankles and the metacarpal joints of

both hands, flexion contracture of large joints and hypertrophic

gingivae were quite characteristic (Fig. 1). His intelligence was

within normal range.

Screening for inheritable metabolic diseases including

amino acids in urine and plasma as well as serum immuno-

globulins was normal. Other blood biochemistry examinations

including serum vitamin A and 1,25-(OH)2 vitamin D3 were

within normal limits. His hands showed diffuse osteoporosis,

thin cortices and soft tissue swelling on X-ray examination.

(b)

(c)

Fig 1. The 20-month-old boy with infantile systemic hyalinosis

showing (a) violaceous nodules over his spinal prominences,

violaceous plaques in the napkin area, joint contracture of extremities,

and a leech-like body position; (b) hypertrophic gingivae; and

(c) a swollen and erythematous appearance of the metacarpo-

phalangeal and interphalangeal joints.



602 Correspondence

A skin biopsy was taken from a lesion on his back. In the

dermis, we found plenty of amorphous hyaline matrix with

scattered fibroblasts and telangiectasia on haematoxylin and

eosin staining. On periodic acid–Schiff staining, the dermis

showed many oval or polygonal cells of a chondroid appear-

ance embedded in an abundant, amorphous and eosinophilic

matrix (Fig. 2).

Electron microscopic examination of the skin biopsy speci-

men showed fibroblasts with enlarged rough endoplasmic

reticulum containing dense particles, and intracytoplasmic and

membrane-bound vesicles of various sizes. These vesicles also

contained fine granular and fibrillar materials. In some places,

the basement membrane zone was discontinuous. Much

fibrillogranular material was located among collagen fibres.

The 17 coding exons of CMG2 were amplified from geno-

mic DNA samples by polymerase chain reaction, and directly

sequenced.1 A heterozygous 304–305insA mutation in exon 4

of CMG2 was found in the patient and his mother. A heterozy-

gous 1074delT mutation in exon 13 of CMG2 was found in

the patient and his father (cDNA sequence NM_058172 in

GenBank was used for nucleotide numbering, counting from

the A of the translational initiation codon ATG). These muta-

tions were not detected in 50 normal controls. The 304–

305insA allele results in a truncated protein of only 103

amino acid residues including two abnormal amino acid resi-

dues at the carboxyl terminus, and the 1074delT allele codes a

truncated protein of 407 amino acid residues including 49

abnormal amino acid residues at the carboxyl terminus. Alter-

natively, mutant CMG2 transcripts were absent in our patient’s

tissues due to nonsense-mediated mRNA decay. The two

mutations are of recessive inheritance as his parents, each car-

rying one mutant allele, were normal.

Two distinct syndromes due to mutations in CMG2 have

been described to date: ISH and juvenile hyaline fibromatosis.

ISH has an earlier onset, a more severe course and the wide-

spread deposition of hyaline material throughout the skin,

gastrointestinal tract, endocrine glands and muscles.3 Diagnosis

is made based on the clinical features such as cutaneous

lesions, gingival hyperplasia, osteoporosis and joint contrac-

ture as well as the findings of amorphous hyaline material in

the dermis.3,4

The protein encoded by CMG2 is a transmembrane protein

with a von Willebrand factor type A (vWA) domain in its

extracellular region.5 Genotype–phenotype analyses suggested

that in-frame and missense mutations within the cytoplasmic

domain result in a mild phenotype of juvenile hyaline fibro-

matosis, whereas mutations located in the vWA domain cause

a more severe form of the disease. In this case, the hetero-

zygous 304–305insA mutation causes premature translation

termination located in the vWA domain, whereas the hetero-

zygous 1074delT mutation results in abnormal amino acid

residues and premature translation termination confined to the

cytoplasmic domain. A homozygous 1074delT mutation was

previously found in two families.1 Therefore, the 1074 site

may be a mutation hotspot in the gene. However, this is the

first case of ISH caused by compound heterozygous mutations

in CMG2, and the 304–305insA mutation is a new finding not

reported in previous papers.

Y-C . HUANG

Y-Y. X IAO

Y-H. ZHENG

W. JANG

Y-L . YANG*

X- J . ZHU

Departments of Dermatology and

*Pediatrics, Peking University First Hospital,

Beijing 100034, China

Correspondence: Xue-Jun Zhu.


Y-C.H. and Y-Y.X. contributed equally to this work.

References

1 Hanks S, Adams S, Douglas J et al. Mutations in the gene encoding

capillary morphogenesis protein 2 cause juvenile hyaline fibro-matosis and infantile systemic hyalinosis. Am J Hum Genet 2003;

73:791–800.2 Dowling O, Difeo A, Ramirez MC et al. Mutations in capillary mor-

phogenesis gene-2 result in the allelic disorders juvenile hyalinefibromatosis and infantile systemic hyalinosis. Am J Hum Genet 2003;

73:957–66.

Fig 2. (a) Biopsy of a skin lesion reveals an

amorphous hyaline matrix in the dermis with

telangiectasia (haematoxylin and eosin;

original magnification · 100); (b) oval or

polygonal cells of a chondroid appearance in

the dermis (periodic acid–Schiff; original

magnification · 400).



Correspondence 603

3 Landing BH, Nadorra R. Infantile systemic hyalinosis: report of fourcases of a disease, fatal in infancy, apparently different from juvenile

systemic hyalinosis. Pediatr Pathol 1986; 6:55–79.4 Keser G, Karabulut B, Oksel F et al. Two siblings with juvenile

hyaline fibromatosis: case reports and review of the literature.Clin Rheumatol 1999; 18:248–52.

5 Bell SE, Mavilla A, Salazar R et al. Differential gene expression duringcapillary morphogenesis in 3D collagen matrices: regulated expres-

sion of genes involved in basement membrane matrix assembly, cellcycle progression, cellular differentiation and G-protein signaling.

J Cell Sci 2001; 114:2755–73.


Scytalidium infection associated withdyskeratosis congenita

DOI: 10.1111/j.1365-2133.2006.07702.x

SIR, Scytalidium dimidiatum (the anamorphic form of Nattrassia man-

giferae, formally known as Hendersonula toruloidea), a saprobic

mould, was first recognized as a human pathogen in 1970,1

with clinical manifestations similar to those of dermatophyto-

sis.2 These nondermatophytes are infrequently diagnosed

because the routine mycological medium used for isolation of

dermatophytes, Sabouraud’s dextrose agar (SDA), contains

cycloheximide, which inhibits their growth.1 Several cases of

S. dimidiatum have been reported in connexion with coexistent

systemic diseases such as chronic active hepatitis, systemic

lupus erythematosus and diabetes mellitus, as well as with var-

ious skin diseases such as eczema, pruritus, lichen simplex,

lichen planus, psoriasis, leprosy and urticaria.2,3 This is the

first report of onychomycosis caused by S. dimidiatum in a

patient with dyskeratosis congenita (DC).

A 16-year-old Indian boy, a field worker by profession,

presented to the dermatology outpatient clinic in July 2005

with a 6-year history of toenail and fingernail dystrophy.

Yellow–brown discoloration of the toenails was noticed ini-

tially, followed by laterodistal dystrophy of the big toe and

subsequently other toenails (Fig. 1). Microscopic examination

and fungal culture of nail scrapings performed elsewhere were

negative for dermatophyte and nondermatophyte moulds. He

was treated unsuccessfully with oral griseofulvin and oral fluc-

onazole but, by September 2001, all his fingernails had also

become similarly affected (Fig. 1).

Direct light microscopy of nail scrapings using 20%

potassium hydroxide solution revealed fungal hyphae, brown in

colour, a rare finding, with little variation in hyphal width.

Culture of nail scrapings was performed on SDA contain-

ing 0Æ005% chloramphenicol with and without 0Æ05%

cycloheximide. Scytalidium dimidiatum was isolated after 2 weeks

of incubation at 25 �C on three occasions on cycloheximide-

free SDA. Positive culture was confirmed by gross colony mor-

phology and colour: pale-olivaceous grey with scant aerial

mycelium and a pale mousy grey undersurface (Fig. 2). Micro-

scopic observations of a lactophenol–cotton blue mount pre-

pared from colonies revealed abundant hyphal loops with little

variation in hyphal width. There were scanty, long cylindrical to

barrel shaped, light brown arthrospores (3 · 2 lm) (Fig. 2).

Further examination of the patient revealed reticulate

pigmentation covering his entire body present since child-

hood, nail dystrophy (Fig. 1), and marked leucoplakia on the

tongue and oral mucosa. He had sparse hair on the lateral half

Fig 1. Clinical appearance: (a) reticulate

hyperpigmentation; (b) palmar

hyperkeratosis; (c,d) onychodystrophy.



604 Correspondence

of the eyebrows, scarring alopecia, and hyperkeratosis of both

palms (Fig. 1). The presence of other features such as dental

caries, marked epiphora and subnormal intelligence further

supported the diagnosis of DC. Respiratory, cardiovascular and

abdominal examinations were normal. Blood counts were

within normal limits. Chest X-ray and abdominal ultrasound

revealed no abnormality. Enzyme-linked immunosorbent assay

for human immunodeficiency virus types 1 and 2 was nega-

tive. Skin biopsy revealed a sparse superficial perivascular

lymphocytic infiltrate with numerous melanophages within

the papillary dermis. The papillary dermis was slightly thick-

ened and showed delicate fibroplasia and mucin. The over-

lying epidermis showed focal vacuolar change in the basal

layer and infiltration of the interface by lymphocytes. The epi-

dermis was flattened in places (Fig. 3).

Our patient was given oral isotretinoin for DC. He was

commenced on oral itraconazole 200 mg twice daily for

7 days each month and topical terbinafine cream, without any

improvement in 6 months.

Scytalidium dimidiatum infection has been reported from South

Asia, Africa, the West Indies and South America.2 Our patient

was infected by a strain resembling the slow-growing type

(form 3), which occurs predominantly in individuals from the

Indian subcontinent.2

DC is a rare inherited disease characterized by a triad of

clinical manifestations including abnormal skin pigmentation,

nail dystrophy and mucosal leucoplakia.4 Along with the

above diagnostic triad our patient also presented with palmar

hyperkeratosis, which was negative for fungus on microscopy

and culture, and dental caries, both of which have been repor-

ted with high frequency in patients with DC.4 Bone marrow

suppression, which presents as anaemia in half of patients

with DC in their second or third decade,4 was not observed in

Fig 3. Photomicrograph showing superficial perivascular lymphocytic

infiltrate with numerous melanophages within papillary dermis

and focal vacuolar changes in the basal layer of the epidermis

(haematoxylin and eosin).

Fig 2. (a) Surface and (b) undersurface of

Scytalidium dimidiatum culture on Sabouraud’s

dextrose agar. (c) Lactophenol–cotton blue

mount showing abundant hyphal loops.

(d) Arthroconidia of S. dimidiatum (Indian

subcontinental strain).



Correspondence 605

our patient. Nail dystrophy in DC involves the fingernails first,

with longitudinal fissuring, thinning, atrophy, distortion, and

rarely nail loss.5 In onychomycosis due to S. dimidiatum the

changes are thickening, discoloration and laterodistal inva-

sion.3 The nail changes in our patient may have resulted from

the combined effect of both Scytalidium infection and DC.

Among the antifungals, griseofulvin, ketoconazole and fluc-

onazole are ineffective against S. dimidiatum, while there are

isolated reports of efficacy of oral itraconazole, intravenous

amphotericin B, topical amorolfine and topical ciclopiroxol-

amine.6 Terbinafine has in vitro activity but no clinical effect.7

It is important to study the efficacy of new antifungal drugs

(voriconazole, ravuconazole, posaconazole and caspofungin)8

in Scytalidium infection and to develop specific therapy for this

orphan disease. Correct diagnosis of Scytalidium infection is

necessary to avoid lengthy and potentially toxic antifungal

treatment. Furthermore, in a patient with DC it may be worth-

while to evaluate the nail changes with regard to the possibil-

ity of associated fungal infection.

Acknowledgments

We thank Dr Uday Khopkar, Dermatologist, K.E.M Hospital,

Mumbai for his expert opinion on histopathology.

J . GARG

R . T I L AK

A.K . GULAT I

S . S INGH*

P . PRAKA SH

A. GARG

Departments of Microbiology and

*Dermatology, Institute of Medical Sciences,

Banaras Hindu University,

Varanasi, UP, India

Correspondence: Atul Garg.


References

1 Gentles JC, Evans EGV. Infections of the feet and nails with Henderson-ula toruloidea. Sabouraudia 1970; 8:72–5.

2 Sigler L, Summerbell RC, Poole L et al. Invasive Nattrassia mangiferaeinfections: case report, literature review, and therapeutic and

taxonomic appraisal. J Clin Microbiol 1997; 35:433–40.3 Hay RJ, Moore MK. Clinical features of superficial fungal infections

caused by Hendersonula toruloidea and Scytalidium hyalinum. Br J Dermatol1984; 110:677–83.

4 Kraemer KH. Heritable diseases with increased sensitivity to cellularinjury. In: Fitzpatrick’s Dermatology in General Medicine (Freedberg IM,

Eisen AZ, Wolff K, Austen KF, Goldsmith LA, Katz SI, FitzpatrickTB, eds), 5th edn, Vol. 2. New York: McGraw-Hill, 1999; 1858–

9.5 Bagby GC, Lipton JM, Sloand EM, Schiffer CA. Marrow failure.

Hematology 2004; 1:318–36.6 Goon ATJ, Seow CS. Three cases of Nattrassia mangiferae (Scytalidium

dimidiatum) infection in Singapore. Int J Dermatol 2002; 41:53–5.7 Clayton YM. Relevance of broad-spectrum and fungicidal activity of

antifungals in the treatment of dermatomycoses. Br J Dermatol 1994;130 (Suppl. 43):7–8.

8 Lecha M, Effendy I, Chauvin MF et al. Treatment options – develop-ment of consensus guidelines. J Eur Acad Dermatol Venereol 2005; 19

(Suppl. 1):25–33.


Pityriasis rubra pilaris and hypothyroidism.Efficacy of thyroid hormone replacementtherapy in skin recovery

DOI: 10.1111/j.1365-2133.2006.07697.x

SIR, A 24-year-old man, without any relevant medical history,

presented with progressive erythroderma that had begun

3 months previously. His skin lesions were initially localized

on the forearms but a rapid diffuse extension had occurred.

Psoriasis had been suspected, and he had undergone psoralen

plus ultraviolet A therapy, in association with topical steroids,

without any improvement. At admission to our department,

he presented with erythroderma with rare areas of normal

skin on the limbs. Follicular keratotic papules and orange

palmoplantar keratoderma were observed and the face and

scalp were severely affected. We noticed eyebrow depilation

and marked alopecia associated with apathy of the patient

(Fig. 1a). Laboratory studies showed an inflammatory syn-

drome. Blood count and serum IgE level were normal. Human

immunodeficiency virus (HIV), hepatitis B virus and hepatitis

C virus serologies were negative. Lipid determination showed

a borderline high level of cholesterol at 5Æ87 mmol L)1 (desir-

able < 5Æ3; high > 6), without hypertriglyceridaemia. Skin

biopsy showed an alternation of parakeratosis and ortho-

hyperkeratosis especially in follicular orifices. In the dermis

there was a moderate lymphocytic infiltrate. The histopatholo-

gical and clinical evidence led to the diagnosis of pityriasis

rubra pilaris (PRP).

In view of the eyebrow depilation, diffuse alopecia and

dyslipidaemia, hypothyroidism was suspected. This was con-

firmed by high serum levels of thyroid-stimulating hormone

(TSH; 25 lU mL)1; normal 0Æ35–3Æ5), T3 (2 pmol L)1; nor-

mal 2Æ5–5) and T4 (8Æ6 pmol L)1; normal 10–20). The hypo-

thyroidism was considered to be autoimmune, with elevation

of antithyroglobulin antibodies. Vitamin A and serum carotene

levels were normal.

Oral thyroid hormone replacement therapy was begun

(L-thyroxine with progressive dose increases), associated with

topical corticosteroids and emollients. This led to dramatic

improvement of skin lesions after only 1 month of treatment

(Fig. 1b), while the patient had previously been erythroder-

mic for 6 months without any improvement. In parallel, we

noted hair regrowth and recovery of functions, while his thy-

roid status remained elevated (TSH 8Æ6 lU mL)1). Topical

corticosteroids were stopped in 1 month. The patient was still

in complete remission after more than 1 year of follow up

and his serum TSH level was normal.

Our patient showed severe classical adult PRP (type I)1 and

concomitant autoimmune hypothyroidism. Thyroid hormone

replacement allowed a rapid and complete remission of PRP

without any additional treatment. This unusual association has

been reported only twice but is probably not fortuitous as it

may be explained by the physiopathology of PRP. The first



606 Correspondence

previously reported case was mentioned with neither details

nor reference in a literature review.2 The second case reported

was of erythrodermic PRP in a 4-year-old child, associated

with severe nonautoimmune hypothyroidism.3 As in our case,

complete remission of PRP was obtained after 3 months of

thyroid hormone replacement.

This dramatic skin improvement under thyroid hormone

replacement therapy is remarkable and argues against a fortuit-

ous association. While spontaneous remission of PRP is fre-

quent (80%), it occurred after several months or even after

1–3 years, especially in erythrodermic types.4 Systemic treat-

ments (retinoids, methotrexate) are usually necessary to accel-

erate the remission, and topical treatments such as topical

corticosteroids are generally less effective.5 In our patient,

improvement was rapidly obtained after a course of L-thyrox-

ine, without addition of any other systemic drug, suggesting

the direct effect of this treatment on PRP.

The pathogenesis of PRP is not clear, although a dis-

order of vitamin A metabolism has been suggested.6

Firstly, the histological appearance of PRP is quite similar

to that of hypovitaminosis A. Secondly, PRP has been

successfully treated in the past by high doses of vitamin

A.7 As thyroid hormone deficiency inhibits the transform-

ation of carotene to vitamin A, it may be responsible for

the occurrence of PRP in this patient. Our patient had

normal levels of carotene and vitamin A, but these serum

levels do not strictly reflect their reserves, especially in

the liver. Thus, an abnormality of vitamin A metabolism,

induced by hypothyroidism, could explain the occurrence

of PRP in association with hypothyroidism, and its recov-

ery after thyroid hormone replacement.

PRP associated with hypothyroidism may be a particular

subtype of PRP (such as the previously described association

between HIV and PRP8). This association, although unusual,

has to be made known because of the spectacular effect of

thyroid hormone replacement. Hypothyroidism must be sus-

pected when PRP occurs in association with asthenia and

apathy, as in our patient.

V . ORLA ND IN I

O. COGRE L

M.S . DOUTRE

C . BEY LOT

M. BEY LOT-BARRY


Hopital Haut-Leveque, avenue de Magellan,

33604 Pessac, France


References

1 Griffiths WAD. Pityriasis rubra pilaris. Clin Exp Dermatol 1980; 5:105–12.

2 Gross DA. Pityriasis rubra pilaris. Report of a case and analysis ofthe literature. Arch Dermatol 1969; 99:710–16.

3 Tunnessen WW, Nieburg PI, Voorhess ML. Hypothyroidism and pit-yriasis rubra pilaris. Response to thyroid hormone. J Pediatr 1976;

88:456–8.4 Griffiths WAD. Pityriasis rubra pilaris: the problem of its classifica-

tion. J Am Acad Dermatol 1992; 26:140–2.5 Chaabouni M. Treatment of pityriasis rubra pilaris. Ann Dermatol

Venereol 1999; 126:836.6 Finzi AF, Altomare G, Bergamaschini L, Tucci A. Pityriasis rubra

pilaris and retinol-binding protein. Br J Dermatol 1981; 104:253–6.7 Griffiths WAD. Vitamin A and pityriasis rubra pilaris. J Am Acad

Dermatol 1982; 7:555.8 Sanchez-Regana M, Fuentes CG, Creus I et al. Pityriasis rubra pilaris

and HIV infection: a part of the spectrum of HIV-associated follicu-lar syndrome. Br J Dermatol 1995; 133:818–19.


(a) (b)

Fig 1. (a) Severe erythema of the face and

eyebrow depilation. (b) Comparative

photograph after 1 month of thyroid

hormone replacement therapy.



Correspondence 607

Nicorandil-induced peristomal ulcers:is nicorandil also associated withgastrointestinal fistula formation?

DOI: 10.1111/j.1365-2133.2006.07698.x

SIR, Further to the recently published report by Claeys et al.1 of

two patients with perivulval and perianal ulceration induced

by nicorandil, we present a series of four cases of nicorandil-

induced peristomal ulceration in patients with a previous his-

tory of diverticular disease requiring gastrointestinal surgery.

To our knowledge this is a previously unreported association.

Patient 1 was an 83-year-old woman who presented with

painful vulval and peristomal ulceration. She had a defunction-

ing colostomy following the development of a colovaginal fis-

tula in 2004; a barium enema performed prior to the surgery

demonstrated diverticular disease. In 2005 she presented with

vaginal discharge and was found to have several vesicovaginal

fistulae; however, these were managed conservatively because

of medical comorbidity. She had a history of ischaemic heart

disease and had been taking nicorandil since 2001. Examina-

tion revealed three vulval ulcers and a well-defined sloughy

peristomal ulcer. There were no clinical features of pyoderma

gangrenosum (Fig. 1). Nicorandil was discontinued and alter-

native antianginal therapy (isosorbide mononitrate) com-

menced. At 2 months, the ulcers had healed without any

additional treatment.

Patient 2 was an 82-year-old woman seen with a 2-week

history of painful peristomal ulceration. She had a history of a

colovaginal fistula secondary to diverticular disease that had

been treated 1 year previously with a Hartman’s procedure

(resection of part of the colon with formation of a colos-

tomy). At the time of surgery she also had drainage of an

abscess in the left groin that had since failed to heal. She was

on multiple medications for ischaemic heart disease, hyperten-

sion and also osteoarthritis, including nicorandil 20 mg twice

daily. On examination she had a sloughy peristomal ulcer and

an area of ulceration in the left groin. After consultation with

her general practitioner the nicorandil was discontinued and

the ulcers were treated topically with sulcrafate powder. At

1 month review there was considerable improvement and by

10 months the ulcers had healed.

Patient 3 was a 59-year-old man who 3 months prior to pre-

sentation had had an ileostomy performed for presumed fistulat-

ing diverticular disease. Two months postsurgery he had

developed painful peristomal ulceration that had not responded

to treatment with topical triamcinolone (Adcortyl in Orabase�;

Bristol-Myers Squibb, Uxbridge, U.K.). He had a previous his-

tory of rheumatoid arthritis, ischaemic heart disease and asthma,

and had been taking nicorandil 20 mg twice daily. He had two

shallow peristomal ulcers. The ulcers were treated topically with

sulcrafate powder and the nicorandil was discontinued. At

3 month review the ulcers had completely healed.

Patient 4 was a 63-year-old woman with ischaemic heart

disease on nicorandil 20 mg daily who developed perianal

ulcers and then a colovesical fistula in 2005 necessitating a

defunctioning colostomy. She was seen in 2006 with peristo-

mal and perianal ulcers that healed within 3 months after

stopping nicorandil and changing to isosorbide mononitrate.

The initial healing of the peristomal ulcer was complicated by

prominent overgranulation treated with corticosteroid (Syna-

lar� gel; GP Pharma, Lidlington, U.K.).

The association between nicorandil and oral or anal ulceration

is well known.2,3 There are also a few reports of gastrointestinal

ulceration linked to nicorandil therapy, including a report of a

patient with multiple ulcers of the small and large intestine that

resolved on stopping nicorandil and a report of ulceration and

perforation of the terminal ileum possibly due to nicorandil.4,5

Interestingly, two of our patients had cutaneous ulceration prior

to the development of fistulae requiring stoma formation. The

theoretical link between fistulating bowel disease and nicorandil

therapy is currently being investigated.

The mechanism of nicorandil-induced mucosal ulceration is

as yet unknown; however, various hypotheses have been sug-

gested including a direct effect of the drug or its metabolites

on the mucosa, and a vascular steal phenomenon.6 In some

patients dose reduction may be sufficient to promote ulcer

healing; however, some will require cessation of the drug and

institution of alternative antianginal therapy. Recognition of

the potential association of nicorandil and peristomal ulcer-

ation and fistulating gastrointestinal disease is important, as

these conditions have high morbidity.

S . OGDEN

Y . MUKA SA

C .C . LYON*

I .H. COUL SON

Department of Dermatology, Burnley General

Hospital, Burnley BB10 2PQ, U.K.

*Department of Dermatology, York District

Hospital, York YO31 8HE, U.K.


References

1 Claeys A, Weber-Muller F, Trechot P et al. Cutaneous, perivulvar andperianal ulcerations induced by nicorandil. Br J Dermatol 2006;

155:494–6.2 Healy CM, Smyth Y, Flint SR. Persistent nicorandil induced oral

ulceration. Heart 2004; 90:e38.Fig 1. Peristomal ulcer (patient 1).



608 Correspondence

3 Katory M, Davies B, Kelty C et al. Nicorandil and idiopathic analulceration. Dis Colon Rectum 2005; 48:1442–6.

4 Egred M, Andron M, Morrison WL. Nicorandil may be associatedwith gastrointestinal ulceration. BMJ 2006; 332:889.

5 King PM, Suttie SA, Jansen JO, Watson AJM. Perforation of the ter-minal ileum: a possible complication of nicorandil therapy. Surgeon

2004; 2:56–7.6 Watson A, Al-Ozairi O, Fraser A et al. Nicorandil associated anal

ulceration. Lancet 2002; 360:546–7.


Variability in the clinical pattern of cutaneousside-effects of drugs with systemic symptoms:does a DRESS syndrome really exist?

DOI: 10.1111/j.1365-2133.2006.07704.x

SIR, In a recent study Peyriere et al.1 stated that the existence of a

clinical entity, known under various names including HSS (anti-

convulsant hypersensitivity syndrome), DRESS (drug reaction

with eosinophilia and systemic symptoms), DIDMOHS (drug-

induced delayed multiorgan hypersensitivity syndrome) and

DIHS (drug-induced hypersensitivity syndrome) cannot be

denied but that its definition, clinical and biological pattern, and

limits must be more accurately reappraised. We can fully

endorse that a gold standard is lacking, as is also repeatedly sta-

ted in the literature. Also the lack of consensus on nosology is

obvious, but this is minor if there is agreement on what are the

main characteristics of the ‘syndrome’. Although it is generally

accepted that a syndrome by its nature comprises a variable

combination of symptoms, the acronym DRESS is questioned as

eosinophilia need not necessarily be present in this syndrome.

The diversity of cutaneous adverse drug reactions is nearly

infinite. It makes sense to isolate syndromes, rather than to

consider the whole as a continuum, if it helps in finding ori-

ginal clinical patterns, courses, causes, mechanisms and treat-

ment. From long discussions between experts from different

countries in recent medical meetings on drug hypersensitivity

it appears that whatever the denomination, HSS/DRESS is

characterized by a variable combination of: (i) drug-induced

immunological background; (ii) later onset than other drug

reactions; (iii) longer duration than common ‘drug rashes’;

(iv) multiorgan involvement; (v) lymphocyte activation (node

enlargement, lymphocytosis, atypical lymphocytes); (vi) eosino-

philia; and (vii) frequent virus reactivation.

HSS/DRESS is specifically complicated because besides its

rather variable presentation it is a diagnosis by exclusion. Its

main features such as rash, fever and organ involvement can also

be attributed to a wide range of other causes such as infections,

and to concomitant and pre-existing diseases. Hence each symp-

tom should always be thoroughly investigated for its relation to

the syndrome. Not all symptoms and signs are always recog-

nized, and asymptomatic systemic involvement such as eosino-

philia and atypical lymphocytes are often not determined or are

determined too late, leading to their under-reporting. In add-

ition, partly due to the relatively long latency after initiation and

the long duration after cessation of the culprit drug, the symp-

toms are often not recognized as drug related. General aware-

ness of HSS/DRESS is very important due to the severity and

life-threatening potential of this type of drug reaction.

The RegiSCAR study group (as its predecessors EuroSCAR

and SCAR) is performing a prospective study of severe cutan-

eous adverse reactions (SCAR) in Austria, France, Germany,

Israel, Italy and the Netherlands, in order to investigate their

risk factors and mechanisms based on a large multinational

registry. Former projects of the group dealt with the spec-

trum of Stevens–Johnson syndrome/toxic epidermal necrolysis

(SJS/TEN)2 and acute generalized exanthematous pustulosis

(AGEP).3 Crucial to these studies has been a clear case defini-

tion. The combination of a scoring system and judgement of

cases by a review committee (blinded for possible risk factors)

has proven effective for validation in SJS/TEN and AGEP. The

group’s intention to extend investigations to HSS/DRESS raised

the need for an equally reliable approach for those cases.

RegiSCAR has collected cases of HSS/DRESS since 2002.

Patients are actively detected through a hospital network cov-

ering about 170 million inhabitants, using selected inclusion

criteria (Table 1). Information on reported cases of HSS/

DRESS is obtained by trained local interviewers using standard-

ized questionnaires, comprising elaborate questions on drug

use, morphology and extent of the rash, involvement of

lymph nodes and other organs, laboratory and clinical param-

eters to judge organ involvement as attributable to HSS/

DRESS, and course of the disease. Where possible, clinical pic-

tures and results of histological examination in the active

phase of the eruption are collected. Interviews take place at

the acute stage of the disease with a follow up at 8 ± 2 weeks

and 1 year, if the patient’s informed consent for participation

in a cohort is obtained. In addition, blood samples are taken

for immunological and genetic research.

Due to the complexity and variability of HSS/DRESS, inter-

pretation of clinical findings and laboratory data by an unorgan-

ized reviewing process would not have produced consistent

Table 1 Inclusion criteria for potential case of HSS/DRESS inRegiSCAR

HospitalizationReaction suspected to be drug related

Acute skin rasha

Fever above 38 �Ca

Enlarged lymph nodes at at least two sitesa

Involvement of at least one internal organa

Blood count abnormalitiesLymphocytes above or below the laboratory limitsa

Eosinophils above the laboratory limits (in percentage or

absolute count)a

Platelets below the laboratory limitsa

aThree or more required.



Correspondence 609

results. Based on information from the literature and clinical ex-

perience of the review committee we reached consensus on a

scoring system for our study which would allow for reproduci-

bly classifying cases as definite, probable, possible or no case.

Thorough case assessment was done on the basis of clinical pic-

tures and analysis of the collected data by experienced clinicians,

as this cannot be replaced by a scoring system alone, but will

always need professional judgement. An overview of the scoring

system is given in Table 2. To prevent bias, the review commit-

tee was blinded to the suspected drugs.

Although we are aware that virus reactivation may play a

role in the syndrome, we do not count the related organ in

case of a positive virus serology. However, it is still a matter

of debate whether reactivation of several herpesviruses in the

course of the disease is part of the syndrome or should be

interpreted as a complication, resulting in a more protracted

and relapsing disease.4,5

In pharmacovigilance, data are often only received retro-

spectively, whereas we most often see the patient in the acute

stage of the disease and systematically collect far more detailed

data, permitting us better to judge the presented symptoms.

Moreover, the advantages of the scale of our multinational

study over a national one, as proposed by Peyriere et al.,1 in a

rare syndrome such as HSS/DRESS, are obvious.

We anticipate our case definition and system of validation

will lead to a reliable identification of cases of HSS/DRESS for

further studies of pharmacoepidemiological and genetic risk

factors, as well as the immunological background. We expect

to be able to answer several long-standing questions after fur-

ther case enrolment in 12–18 months.

S .H. KARDAUN

A. S I DOROF F*

L . VAL EYR I E -AL LANORE�S . HAL E VY�

B.B . DA V IDOV I C I�M. MOCKENHAUPT§

J .C . ROU J E AU�

Center for Blistering Diseases,


University Medical Center Groningen,

University of Groningen, Hanzeplein 1,

9713 GZ Groningen, the Netherlands

*Department of Dermatology and Venereology,

Medical University of Innsbruck,

Innsbruck, Austria

�Department of Dermatology, Hopital

Henri Mondor, University of Paris XII,

Creteil, France

�Department of Dermatology,

Soroka University Medical Center,

Ben-Gurion University of the Negev,

Beer-Sheva, Israel

§Dokumentationszentrum schwerer

Hautreaktionen (dZh),

Department of Dermatology, University Medical Center,

Freiburg, Germany


Table 2 Scoring system for classifying HSS/DRESS cases as definite, probable, possible or no case

Score )1 0 1 2 Min. Max.

Fever ‡ 38Æ5 �C No/U Yes )1 0

Enlarged lymph nodes No/U Yes 0 1Eosinophilia No/U 0 2

Eosinophils 0Æ7–1Æ499 · 109 L)1 ‡ 1Æ5 · 109 L)1

Eosinophils, if leucocytes < 4Æ0 · 109 L)1 10–19Æ9% ‡ 20%

Atypical lymphocytes No/U Yes 0 1Skin involvement )2 2

Skin rash extent (% body surface area) No/U > 50%Skin rash suggesting DRESS No U Yes

Biopsy suggesting DRESS No Yes/UOrgan involvementa 0 2

Liver No/U YesKidney No/U Yes

Lung No/U YesMuscle/heart No/U Yes

Pancreas No/U Yes

Other organ No/U YesResolution ‡ 15 days No/U Yes )1 0

Evaluation of other potential causesAntinuclear antibody

Blood cultureSerology for HAV/HBV/HCV

Chlamydia/mycoplasmaIf none positive and ‡ 3 of above negative Yes 0 1

Total score )4 9

U, unknown/unclassifiable; HAV, hepatitis A virus; HBV, hepatitis B virus; HCV, hepatitis C virus. aAfter exclusion of other explanations:

1, one organ; 2, two or more organs. Final score < 2, no case; final score 2–3, possible case; final score 4–5, probable case; final score > 5,definite case.



610 Correspondence

References

1 Peyriere H, Dereure O, Breton H et al. Variability in the clinical pat-tern of cutaneous side-effects of drugs with systemic symptoms:

does a DRESS syndrome really exist? Br J Dermatol 2006; 155:422–8.2 Auquier-Dunant A, Mockenhaupt M, Naldi L et al. Correlations

between clinical patterns and causes of erythema multiforme majus,Stevens–Johnson syndrome, and toxic epidermal necrolysis: results of

an international prospective study. Arch Dermatol 2002; 138:1019–24.3 Sidoroff A, Halevy S, Bavinck JN et al. Acute generalized exanthema-

tous pustulosis (AGEP) – a clinical reaction pattern. J Cutan Pathol2001; 28:113–19.

4 Kano Y, Hiraharas K, Sakuma K, Shiohara T. Several herpesvirusescan reactivate in a severe drug-induced multiorgan reaction in the

same sequential order as in graft-versus-host disease. Br J Dermatol2006; 155:301–6.

5 Seishima A, Yamanaka S, Fujisawa T et al. Reactivation of humanherpesvirus (HHV) family members other than HHV-6 in drug-

induced hypersensitivity syndrome. Br J Dermatol 2006; 155:344–9.


Book review

DOI: 10.1111/j.1365-2133.2006.07722.x

Textbook of Pediatric Dermatology, 2nd edn. J. Harper, A. Oranje

& N. Prose (editors). Oxford: Blackwell Publishing, 2005;

2272 pp. ISBN: 9781405110464. Price £350.00.

This extensive, two-volume text represents an international

collaboration of the world leaders in the field of paediatric

dermatology. It is meticulously and systematically organized

in sections and subsections based around disorders with com-

mon or related pathogenesis, modes of clinical presentation

and diseases presenting within specific body sites. As such it is

highly user friendly, allowing the reader not only to find

comprehensive details of specific disorders but also to retrieve

information to aid in the differential diagnosis and investiga-

tion of more ill-defined clinical presentations.

There have been many advances in the field of paediatric

dermatology since the first edition was published in 2000, par-

ticularly in the field of molecular genetics and in the under-

standing of the pathomechanisms of disorders. The second

edition has been extensively updated to reflect this knowledge

and the impact it has had on disease classification. There are

nine additional chapters and many new contributing authors.

While common clinical problems such as eczema and skin

infections are given significant coverage, much rarer disorders

are not short changed and as such this is an excellent refer-

ence text. Indeed, the largest section is devoted to the full

breadth of genetic skin disease and is a particular strength of

this book. There is extensive use of colour photography which

is generally of high quality. There has been a much greater

emphasis in the second edition on the practical aspects of dis-

ease management with useful algorithms and protocols in

many chapters which will undoubtedly be of great benefit to

the practising dermatologist.

This is a state-of-the-art textbook which no dermatology

department should be without and I suspect that most dermato-

logists with any paediatric practice will want to own a copy.

D.K .B . ARMSTRONGRoyal Hospital, Belfast BT12 6BA, U.K.

News and Notices

DOI: 10.1111/j.1365-2133.2007.07788.x

Congress: 2nd International Congress on Psoriasis

Dates: JUNE 21–24, 2007

Venue: Paris, Palais des congres, FRANCE

Web site: http://www.pso2007.com

Contact:

PSO 2007 c/o MCI

24 rue Chauchat

75009 Paris

FRANCE

Phone: +33 (1) 53 85 82 59

Fax: +33 (1) 53 85 82 83

Email: [email protected]

Main topics:

– Scoring and monitoring the severity of the disease

– Management of the severe clinical manifestations

– Psoriasis in children and pregnancy

– Difficult to treat localisations

– Topical Treatment: what to choose and how to use

– Risk management and treatment optimisation: combination

and rational strategies

– Phototherapies: what to choose and how to use

– Alternative treatment

– Biologics

– Extracutaneous manifestations in Psoriasis



Book review 611

North West Vascular Anomalies Consensus Workshop

Wednesday 23rd May 2007

Call for Abstracts

A Scientific workshop to develop a consensus in the treatment

of vascular malformations throughout the United Kingdom.

Venue – Crowne Plaza Hotel Liverpool.

For further information abstract forms or registration forms,

please visit http://www.vascularanomalies.org.uk or contact

[email protected]; [email protected]

Royal Liverpool Children’s NHS Trust Eaton Road Liverpool

L12-2AP.

Corrigenda

DOI: 10.1111/j.1365-2133.2007.07789.x

Familial cutaneous collagenomas resulting from a novelmutation in LEMD3. Br J Dermatol 2007, Hershkovitzet al.

In the above mentioned article1 there was an error in one

of the author names. D.B. Amitai should have appeared as

B. Amitai.

The authors apologise for these errors.

References

1 Hershkovitz D, Amitai B, Sprecher E. Familial cutaneous collageno-mas resulting from a novel mutation in LEMD3. Br J Dermatol 2007;

156:375–7.

DOI: 10.1111/j.1365-2133.2007.07790.x

Validation of the U.K. Working Party diagnostic cri-teria for atopic eczema in a Xhosa-speaking Africanpopulation. Br J Dermatol 2007, Chalmers et al.

In the above mentioned article1 there were some errors in

the reference list.

Reference 2 should have appeared as follows and not as

was published:

2 Williams HC, Burney PG, Hay RJ et al. The UK Working

Party’s Diagnostic Criteria for Atopic Dermatitis. I. Deriva-

tion of a minimum set of discriminators for atopic derma-

titis. Br J Dermatol 1994;131:383–96.

Reference 15 should have appeared as follows and not as

was published:

15 Williams HC, Burney PG, Pembroke AC, Hay RJ. The

UK Working Party’s Diagnostic Criteria for Atopic Derma-

titis. III. Independent hospital validation. Br J Dermatol

1994;131:406–16.

The authors apologise for these errors.

References

1 Chalmers DA, Todd G, Saxe N, Milne JT, Tolosana S, Ngcelwane PN,

Hlaba BN, Mngomeni LN, Nonxuba TC, Williams HC. Validation ofthe U.K. Working Party diagnostic criteria for atopic eczema in a

Xhosa-speaking African population. Br J Dermatol 2007; 156:111–16.

DOI: 10.1111/j.1365-2133.2007.07791.x

Ciclosporin use in toxic epidermal necrolysis. Br JDermatol 2007, Danby.

In the above mentioned article1 there was an error in one

of the corresponding email addresses.

[email protected] should have appeared as fwdljm@

tds.net.

The author apologises for this error.

References

1 Danby FW. Ciclosporin use in toxic epidermal necrolysis. Br J Dermatol

2007; 156:390–1.



612 Corrigenda

British Journal of Dermatologylib.ajaums.ac.ir/booklist/339455.pdf · C. Mitteldorf R. Stadler H.P....

Documents

Transcript of British Journal of Dermatologylib.ajaums.ac.ir/booklist/339455.pdf · C. Mitteldorf R. Stadler H.P....