J7 Clin Pathol: Mol Pathol 1 996;49:M225-M23 1

Lymphocytes, cytokines and adhesion moleculesin chronic graft versus host disease

Selim Aractingi, Eliane Gluckman, Caroline Le Goue, Louis Dubertret,Edgardo D Carosella

AbstractAims-To determine which inflammatoryand immune pathways are implicated inthe development of chronic graft versushost disease (GvHD) and whether differ-ences between these pathways are respon-sible for the different presentations ofchronic GvHD.Methods-Biopsy specimens of diseasedand normal skin were obtained frompatients presenting with lichen planus-like and sclerodermatous type chronicGvHD. Expression of epidermal cyto-kines, adhesion molecules and lymphoidsurface markers was analysed by means ofimmunohistochemistry. Apoptosis wasdetected using the in situ nick end-labelling method.Results-In both GvHD lesion types,CD8+ cells predominated in the epider-mis, whereas CD4+ cells were the mostprevalentinthedermis.Apoptotickeratino-cytes were found in diseased skin only andFas antibodies labelled a considerablenumber of keratinocytes. The epidermisin both types oflesions expressed interleu-kin (IL) la, tumour necrosis factor (TNF)a and intercellular adhesion molecule(ICAM)-1, but dermal vascular cell adhe-sion molecule (VCAM)-1 expression wasrestricted to specimens of lichen planus-like GvHD. ILla and E-selectin wereexpressed in normal looking skin of 55%

Service de Rechercheen Hemato-Immunologie(DRM-DSV, CEA),Hopital St Louis,Centre Hayem,1 Avenue ClaudeVellefaux, 75475 Paris,FranceS AractingiE GluckmanC Le GoueL DubertretE D Carosella

Service de Greffe deMoelleE Gluckman

Correspondence to:Dr S Aractingi.

Accepted for publication16 April 1996

and 80%, respectively, of patients withlichen planus-like GvHD.Conclusion-The similarity between ex-pression ofepidermal cytokines and adhe-sion molecules (with the exception ofVCAM-1) and lymphocyte phenotype inlichen planus-like and sclerodermatousGvHD strongly suggests that the latteroccurs as a consequence of the healingprocess. VCAM-1 distinguishes betweenlichen planus-like and sclerodermatouslesions. ILla and E-selectin are potentialearly markers of chronic GvHD.(7 Clin Pathol: Mol Pathol 1996;49:M225-M23 1)

Keywords: graft versus host disease, skin, lymphocytescytokines, adhesion molecules.

Chronic graft versus host disease (GvHD) isone of the most disabling complications ofbone marrow transplantation (BMT). About30-50% of transplant recipients still livingthree months after transplantation will presentwith features of this disease.' 2In contrast toacute GvHD, where a cytotoxic reaction drivenby donor T lymphocytes against host antigenshas been demonstrated,3 4the pathogenesis ofchronic GvHD is poorly understood. Evidencefor an interaction between anti-host and auto-immune phenomena is described fre-quently,5-'0 but the precise pathways leading tothe development of the lesions are still unclear.

Table 1 Antibodies used

Antibody Reactivity Source

Lymphocyte surface antigensCD3 T lymphocytes Dako

(CD3 complex associated with alx or y6TCR)CD4 Helper/inducer T lymphocytes DakoCD8 Suppressor/cytotoxic lymphocytes DakoCD16 Natural killer cells DakoCD25 IL-2R Dako

(Tac; IL-2Ra chain)CD30 Activated lymphoid cells DakoCD45RO Memory T cells DakoCD56 Natural killer cells Dako6TCR yS T cells receptor Immunotecha,TCR ac T cell receptor ImmunotechCD20 B cells DakoPerforin Perforin expression PharmacellCD 95 (Fas) Apo-Fas protein on target cells ImmunotechCytokinesILlc ILI protein R&D SystemsIL8 IL8 protein R&D SystemsTNFa TNF protein R&D SystemsAdhesion moleculesCD106 VCAM-1: ligand for VLA-4 cells NovocastraE-selectin Endothelial adhesion molecule GenzymeCD54 ICAM 1: ligand for LFA-1 and Mac-l Novocastra

TNF = tumour necrosis factor.

M225

on June 9, 2021 by guest. Protected by copyright.

http://mp.bm

j.com/

Clin M

ol Pathol: first published as 10.1136/m

p.49.4.M225 on 1 A

ugust 1996. Dow

nloaded from

http://mp.bmj.com/

Aractingi, Gluckman, Le Goui, Dubertret, Carosella

4.4,~ ~ ~ ; 4..

*4 ~~~

a-~~~~~,:I ' _ -,

t , ~;F't\' -+

E '' S' Sw''.-.

., *at g < Ssr

\ ^ 's e\' io;

va.>9 h*\ .9 .1

_. O

\.

A Ju. > .. .v

Figure 1 (A) CD8 + infiltrate in superficial dermis and epidermis in a patient with lichen planus-like GvHD (originalmagnification x100). (B) CD8 + cells in sclerodermatous GvHD (original magnification x100).

The skin is the majiGvHD and patientsplanus-like or scleboth. The importubetween keratinocytphocytes in triggeririnflammation has bskin diseases.'1-'6 Asterised by a cutaneoithese interactions ndevelopment. In oipathways are implicEwhether the differexGvHD could be sethese pathways, the cmolecules and cytcphenotype of the 1

* 1-

6"'

1., -,.~~~~~~~A1

or target organ in chronics may present with lichen~rodermatous lesions, ornnri- nf tht- inteXrnrtionn

and the mechanisms of epidermal damagewere also investigated.

*lui;IIt;4A Methodstes, endothelium and lym-Leg endor lifyn caneousm- Transplant recipients who presented betweeneg or amplinyeg cutaneous October 1993 and October 1994 with cutane-chrounicrlGvHd inscharac- ous lichen planus-like or de novo scleroderma-s yhroncyt infilratio tous lesions were examined by one of us (SA).uslymphocytic infiltrat Patients were considered to have lichen planus-nay also play a rolewn its like GvHD if they presented with the typical

rted in chteroiniceGvD a physical findings and if examination of tissuesections showed epidermal hyperplasia, eosi-rices in chronic cutaneous nophilic body formation and a lymphocytic

conpareyi to skinferancJes in infiltrate in the epidermis and dermis. Patients)npresow skanalysed.Thesn were considered to have sclerodermatouskines was analysed. Thewymphoid infiltrating cells GvHD if they presented with typical symp-toms, if there was no history of lichen planus

. lesions at the sclerotic sites and if histopathol-ogy demonstrated epidermal atrophy, base-

4- ' ment membrane flattening and fibrosis of thez papillary dermis.

c CONTROLSBiopsy specimens were taken of the diseasedand healthy skin. Each specimen was divided intwo: one half was embedded in OCT and snapfrozen in liquid nitrogen; the other was fixed informalin, embedded in paraffin wax andstained with haematoxylin and eosin.

Biopsy specimens were obtained from pa-tients who received grafts more than threemonths previously, who had other skin diseases(one case of late porphyria, two cases ofadverse reaction to drugs, one case of solarerythema) or healed lichen planus-like GvHD(one case), and from patients who had

riginal undergone total body irradiation and autolo-gous BMT but who had no skin abnormalities

Figure 2 Immunohistochemical labelling of a hair follicle with CD8 antibody (omagnification x250).

B

4A~*'*9

49 ,.zw

M226

,ll.s

on June 9, 2021 by guest. Protected by copyright.

http://mp.bm

j.com/

Clin M

ol Pathol: first published as 10.1136/m

p.49.4.M225 on 1 A

ugust 1996. Dow

nloaded from

http://mp.bmj.com/

Lymphocytes, cytokines and adhesion molecules in chronic graft versus host disease

Table 2 Lymphocyte surface antigens in chronic GvHD

Histological diagnosis Lymphocytic T cells TCR Activation NK B cellsinfiltration CD3 CD25 + CD16 CD20

CD4 CD30 + CD56CD8 CD45RO +

CEO XyLchen planus-like + + + + + + + - + - -Sclerodermatous + + + + + - + - -Normal skin - - - - - - -

(four cases). Normal skin removed cbreast plastic surgery was also analysedcases).

IMMUNOHISTOCHEMISTRYCryostat sections (4-6 gm) were exaximmunohistochemically using an indireotin antibody conjugated technique. Sewere fixed in cold acetone and incubatecnormal serum for 30 minutes at roomperature. Sections were then incubatedeach of the specific antibodies directed athe different antigens listed in table 1 fminutes. The sections were washed inphate buffered saline (PBS) and thenbated with a biotin conjugated antdirected against the primary antibody.was used as the chromogen. Lymphocytic

5i J

sV'';--''.'tv_,.-r

tration was scored semiquantitatively as fol-lows: -, no labelled cells present; +, scatteredlabelled cells; ++, moderate infiltration; and+++, dense infiltration.

Negative controls included substitution ofthe primary antibody by an antibody of anirrelevant specificity but of the same IgGisotype. Positive controls included samples ofpsoriasis, tuberculoid leprosy, T and B celllymphoma, and adverse drug reactions.

gainst LABELLING OF APOPTOTIC CELLSFor 30 Apoptosis was detected in diseased and normalphos- skin by end labelling fragmented DNA usingincu- the ApopTag kit (Oncor, Maryland, USA).'7ibody Briefly, paraffin wax sections were first deparaf-AEC finised and digested with proteinase K. Cryo-inifi- stat sections were fixed in neutral buffered for-

malin and then post-fixed in ethanol:aceticacid (2:1) for five minutes at room tempera-ture. Slides were incubated in a humidifiedchamber at 37°C for one hour in a solutioncontaining terminal deoxynucleotidyl trans-ferase (TDTase), digoxigenin labelled dUTPand dATP. The sections were then washed and

-. incubated with the peroxidase labelled anti-digoxigenin antibody for 30 minutes. Thereaction products were developed with AEC,washed again and the slides counterstainedwith haematoxylin and eosin. The sectionswere then examined under a Leitz orthoplanmicroscope.

4%i 1.I.-

11 ..'

Figure 3 Diffuse labelling of keratinocytes with anti-Fas antigen in chronic lichenplanus-like GvHD (original magnification x250).

Figure 4 Expression ofILla in epidermal keratinocytes (original magnification xlOO)

ResultsHISTOLOGYOn histological examination 13 patients hadtypical features of lichen planus-like GvHDand five of sclerodermatous GvHD. Lympho-cytic infiltration was present in both forms.Epidermal abnormalities were not present andlymphocytes were observed infrequently in thesuperficial dermis in normal skin of the samepatients. The diagnoses of skin diseases in con-trol patients were also confirmed histologically.

IMMUNOHISTOCHEMISTRYScattered CD3+ cells were present in the der-mis of normal skin in all patients with chonicGvHD. A CD3+ infiltrate was observed in sec-tions of diseased skin in both patient groups.CD4+ and CD8+ subsets were found. How-ever, on semiquantitative analysis, theCD4+ CD8+ infiltrate was more dense inpatients with lichen planus-like GvHD (table2). In both types of GvHD, CD8+ lympho-cytes predominated in the epidermis whereasCD4+ cells were more prevalent in the dermis(table 2; figs 1A and 1B). CD8+ cells werealso found also in follicular epithelium in thosewith lichen planus-like GvHD (fig 2). T cells

M227

0II-0-I

I f.-II

. a.1

I

0I e

-1

7 ;.,I t -.1I

on June 9, 2021 by guest. Protected by copyright.

http://mp.bm

j.com/

Clin M

ol Pathol: first published as 10.1136/m

p.49.4.M225 on 1 A

ugust 1996. Dow

nloaded from

http://mp.bmj.com/

Aractingi, Gluckman, Le Goue, Dubertret, Carosella

Table 3 Cytokines and adhesion molecules expressed in chronic GvHD

Lichen planus-like GvHD Sclerodermatous GvHD

Diseased Normal Diseased Normal

TNFa 13/13 3/9 4/5 0/4ILla 13/13 7/9* 5/5 0/5*IL8 13/13 9/9 5/5 4/5VCAM-1 12/13** 1/10 0/5** 0/5E-selectin 12/13 6/9t 3/5 0/5tCD54 13/13 0/9 4/5 0/5

*p < 0.05; **p < 0.01; tp = 0.062.

had the following phenotype: T cell receptor(TCR) ax+, TCRy6-, CD25+, CD30+,CD45RO+ . B lymphocytes and natural killercells were not detected. Perforin expressionwas not seen. Patients who experienced anadverse reaction to drugs (CD16+, CD56+),those with B cell lymphoma (CD20+) orleprosy (TCRy6+, perforin+) served as con-trols. In lichen planus-like GvHD diffuse stain-ing of keratinocyte membranes was noted withFas antibody (fig 3). This expression wasrestricted to basal layer in sclerodermatousGvHD. Fas expression was not observed innormal skin.

Table 3 and figs 4-7 summarise cytokine andadhesion molecule expression. Interleukin (IL)

.... 8~~~

Figure 5 Membrane expression ofCD54 in keratinocytes from diseased skin (originalmagnification x250)

J, ~ ~

tZ Kc- .

/~~~~z

t\S , t . - ,X,,

W~~~~~~~~~~~~~~~~~~~~~~~~.'1.X*SFigure 6 Regular distribution of VCAM-J in the dermis ofpatients with lichenplanus-like GvHD (original magnification xlOO)

la and tumour necrosis factor (TNF) a werenearly always expressed in diseased epidermisof both types of GvHD. IL8 was expressed inthe cytoplasm of all keratinocytes in diseasedand healthy skin of patients in a pattern identi-cal with that in normal controls; the intercellu-lar spaces were unstained (data not shown).expression of vascular cell adhesion molecule(VCAM)-1 in lichen planus-like and scleroder-matous GvHD differed significantly (p < 0.01;Fisher's exact test). ILla was expressed inseven of nine biopsy specimens of normal skinfrom patients with lichen planus-like GvHD,but not in normal skin of patients with sclero-dermatous GvHD (p < 0.05; Fisher's exacttest). Expression of E-selectin was observed insix of nine biopsy specimens of normal skinfrom patients with lichen planus-like GvHDbut not in normal skin of patients with sclero-dermatous GvHD (p = 0.062; Fisher's exacttest).

CONTROLSVCAM-1 expression was not observed in anyof the allogeneic BMT recipients with otherskin conditions. IL1a and E-selectin wereexpressed in two of five of these patients. Alymphocytic infiltrate was not observed inpatients who underwent total body irradiationfollowed by autologous BMT. ILla andE-selectin were expressed in two of fourpatients who had received radiotherapy for lessthan one year.

APOPTOSISNumerous keratinocytic nuclei were labelled inboth GvHD types (fig 8A). Labelled kerati-nocytes were present mainly in the basal layerand in the suprabasal to a lesser degree,especially in patients with lichen planus-likeGvHD. Follicular keratinocytes were alsolabelled (data not shown). Staining was notobserved in controls and in normal skin frompatients with GvHD (fig 8B).

DiscussionChronic GvHD is a multisystemic diseasewhich mainly involves the skin, mouth andliver. Its pathogenesis seems to be the conse-quence of a complex interaction between anti-host and auto-immune phenomena.5'0 Anti-host reactions are demonstrated by thefrequent isolation, from peripheral blood ofpatients with chronic GvHD, of CD4+ cloneswhich proliferate when cultured with cryopre-served host lymphocytes.5 6Furthermore, cyto-toxic clones directed against host lymphocyteshave been isolated from specimens of skin frompatients with chronic GvHD.' However,CD4+ clones recognising the donor Ia grouphave been isolated from mice with chronicGvHD, suggesting the presence of an auto-immune effect.8 Patients with chronic GvHDmay develop lymphoid atrophy and thrombo-cytopenia, indicating that donor cells are alsotargeted.9 CD8+ cells predominate in involvedtissues" and electron microscopy studies showthat some mononuclear cells are in direct con-tact with keratinocytes,'9 implying direct cyto-toxicity. Our results demonstrate that infiltrat-

M228

on June 9, 2021 by guest. Protected by copyright.

http://mp.bm

j.com/

Clin M

ol Pathol: first published as 10.1136/m

p.49.4.M225 on 1 A

ugust 1996. Dow

nloaded from

http://mp.bmj.com/

Lymphocytes, cytokines and adhesion molecules in chronic graft versus host disease

Figure 7 E-selectin expression in dermal vessels (original magnification x250)

ing lymphocytes in lichen planus-like andsclerodermatous GvHD have a similar pheno-type and distribution pattern. Indeed, CD8+cells predominated in the epidermis andCD4+ cells in the dermis. The presence ofCD8+ cells at the site of cellular damagesuggests strongly that these are the effectorcells. Invasion of follicle walls by CD8+ cells inlichen planus-like lesions, very similar to thechanges described in acute GvHD,20 is inaccordance with this hypothesis. The possibil-ity that these observations are the consequenceof reactions directed against epidermal stemcells, which are more numerous in lowerfollicles,2' could not, however, be demonstratedin this study.

Perforin, part of the major cytotoxicity path-way,22 does not seem to play a role in chronic

K.eee.. A

,-'lC/,.o ,8A'"

.62)

J,S.i;tXjo

:,-'1e)

Aractingi, Gluckman, Le Goue, Dubertret, Carosella

phocytes expressing VLA-4 to endothelium."TNFa induces the expression of bothE-selectin and VCAM-1 molecules34 on skinendothelial cells, whereas IL4 triggers selectiveexpression of VCAM-1.35 36 It has recentlybeen shown that administration of antibodiesdirected against IL4 inhibits features of GvHDin mice."7 As VCAM-1 was expressed in lichenplanus-like GvHD only, where infiltrating Tlymphocytes were present, and never in healthyskin (although ILla and TNFac were some-times expressed by keratinocytes), VCAM-1expression could therefore occur secondary tothe lymphocytic infiltration and possibly to IL4expression. Interestingly, Norton and Sloanereported that VCAM-1 was expressed only inpatients who developed acute GvHD at a laterdate."

Sclerodermatous GvHD, therefore, is char-acterised by the production of adhesionmolecules and cytokines by keratinocytes andby epidermal apoptosis and lymphocytic infil-tration. These features are similar, althoughmilder, than those found in lichen planus-likeGvHD. The nature of the T cell infiltrate isalso very similar in both GvHD types. Theseresults suggest strongly that de novo scleroder-matous GvHD may represent healing ofdiscrete epidermal events. This hypothesis is inaccordance with previous studies showing thatthe development of sclerotic lesions in a lateevent in chronic GvHD'0 3"A0 and with clinicalreports of sclerodermatous GvHD developingat sites of lichen planus lesions.4' The localisa-tion of fibrosis in superficial dermis 4' and thesequential deposition of collagen III and I4' insclerodermatous GvHD, very similar to theprocesses occurring during tissue repair,43 alsosupport this hypothesis.

E-selectin and ILl x were expressed in unin-volved skin of 55% and 80% of patients,respectively, with lichen planus-like lesionselsewhere. This was not the case in patientswith sclerodermatous GvHD. The significanceof the expression of these molecules in the nor-mal skin of patients with lichen planus-likeGvHD is difficult to explain. It is possible thatan epidermal injury was responsible for theproduction of ILla, which in turn inducedexpression of E-selectin.'4 In patients withacute GvHD TNFa mRNA expression super-sedes the appearance of a lymphocytic infil-trate.44 4 A better understanding of the role ofE-selectin and ILI x in the pathogenesis ofchronic GvHD may be forthcoming on analy-sis of these molecules at day 100 after BMT.

In conclusion, chronic cutaneous GvHDseems to occur secondary to the infiltration byT cells which in turn induce apoptosis of kera-tinocytes, possibly via the production of Fasligand. T cell infiltration may be triggered bythe epidermal expression of ILla and may inturn amplify inflammation by inducingVCAM-1 expression on dermal vessels. Insome patients, this results in the formation ofsclerotic lesions. Further studies are requiredto elucidate the pathogenesis of this debilitat-ing chronic disease.This work was supported in part by Ligue contre lecancer-Comite de Paris and ARTM.

1 Parkman R. Graft-versus-host disease. Ann Rev Med 1991;42:189-97.

2 Atkison K, Horowitz M, Deeg J, Gale RP, van Bekkun DW,Gluckman E, et al. Risk factors for chronic graft versushost disease. Blood 1990;175:2459-64.

3 Gaschet J, Mahe B, Milpied N, Devilder MC, Dreno B,Bignon JD, et al. Specificity of T-cell invading the skin dur-ing acute graft-versus-host disease after semi-allogeneicbone-marrow transplantation. J Clin Invest 1993;91:12-20.

4 Kasten Sportes CM, Masset M, Varrin F, Devergie A,Gluckman E. Phenotype and function ofT-lymphocyte theskin during graft-versus-host disease following allogenicbone-marrow transplantation Transplantation 1989;47:621-4.

5 Goulmy E, Gratama JW, Blokland E, Zwaan FE, van RoodJJ. Minor transplantation antigen detected by MHCrestricted cytotoxic T lymphocytes during graft versushost disease. Nature 1983;302: 159-61.

6 Tsoi MS, Storb R, Dobbs S, Medill L, Thomas ED. Cellmediated immunity to non-HLA antigens of the host bydonor lymphocytes in patients with chronic graft-versus-host disease. J Immunol 1980;125:2258-62.

7 Atkinson K. Pathogenesis of human chronic graft-vs-hostdisease. In: Burakoff SJ, Deeg HJ, Ferrara J, Atkinson K,eds. Graft-vs-Host-Disease. Hematology. Vol 12. New York:Marcel Dekker, 1990:615-23.

8 Parkman R. Clonal analysis of murine graft versus host dis-ease. JImmunol 1986;136:3543-8.

9 Anasetti C, Rybka W, Sullivan KM, Banaji M, Slichter SJ.Graft-v-host disease is associated with autoimmune-likethrombocytopenia. Blood 1989;73: 1054-8.

10 Shulman HM, Sullivan KM, Weiden PL MacDonald GB,Striker G, Sale GE, et al. Chronic graft-versus-hostsyndrome in man. A long-term clinicopathologic study of20 Seattle patients. AmJMed 1980;69:204-17.

11 Barker JNWN, Mitra RS, Griffiths CEM, Dixit VM, Nick-oloff BJ. Keratinocytes as initiators of inflammation. Lan-cet 1991;337:211-14.

12 Bos JD, Kapsenberg ML. The skin immune system:progress in cutaneous biology. Immunol Today 1993;14:75-8.

13 Norris DA. Cytokine modulation of adhesion molecules inthe regulation of immunologic cytotoxicity of epidermaltargets. Invest Dermatol 1990;95:1 1 1S-20S.

14 Dustin ML, Singer KH, Tuck DT, Springer TA. Adhesionof T lymphoblasts to epidermal keratinocytes is regulatedby interferon and is mediated by ICAML. J Exp Med1988;167: 1323-6

15 Griffiths CEM, Voorhees JJ, Nickoloff BJ. Characterizationof intercellular adhesion molecule- I and HLA-DR expres-sion in normal and inflammed skin: Modulation byrecombinant gamma-interferon and tumor necrosis factor.J Am Acad Dermatol 1989;20:17-29.

16 Norton J, Sloane JP. ICAM-1 on epidermal keratinocytes incutaneous graft-versus-host disease. Transplantation 1991;51:1203-6.

17 Gavrieli Y, Shemen Y, BenSasson SA. Identification of pro-grammed cell death in situ via specific labelling of nuclearDNA fragmentation. J Cell Biol 1992;1 19:493-501.

18 Atkinson K, Munro V, Vasak E, Biggs JC. Mononuclear cellsubpopulation in the skin defined by monoclonal antibod-ies after HLA-identical sibling marrow transplantation. BrJ Dermatol 1986;114:145-60.

19 Fuji H, Oshashi M, Nagura H. Immuno-histochemicalanalysis of oral lichen planus-like eruption in graft-versus-host disease in man.lImmunol 1978;120:1485-92.

20 Murphy GF, Lavker RM, Whitaker D, Korngold R.Cytotoxic folliculitis in GVHD. Evidence for stem cellinjury. J Cutan Pathol 1991;18:309-14.

21 Rochat A, Kobayashi K, Barrandon Y Location of stem cellsof human hair follicles by clonal analysis. Cell 1994;76:1063-73.

22 Kagi D, Ledermann B, Burki K, Seiler P, Odermatt B, OlsenKJ, et al. Cytotoxicity mediated by T cells and naturalkiller cells is greatly impaired in perforin deficient mice.Nature 1994;369:31-7.

23 Takata M. Immunohistochemical identification of perforinpositive cytotoxic lymphocytes in graft versus host disease.Am J Clin Pathol 1995;103:324-9.

24 Rouvier E, Luciani MF, Golstein P. Fas involvement in Caindependent T cell mediated cytotoxicity. J Exp Med1993;177: 195-200.

25 Sayama K, Yonehara S, Watanabe Y, Miki Y. Expression ofFas antigen on keratinocytes in vivo and induction of apop-tosis in cultured keratinocytes. J Invest Dermatol 1994;103:330-4.

26 Acevedo A, Aramburu J, Lopez J, fernandez Herrera J,Fernandez-Ranada JM, Lopez-Botet M. Identification ofnatural killer cells in lesions of human cutaneous GVHD:expression of a novel NK associated surface antigen(Kp43) in mononuclear infiltrates. J Invest Dermatol 199 1;97:659-66.

27 Schattenfroh NC, Hoffman RA, McCarthy SA, SimmonsRL. Phenotypic analysis of donor cells infiltrating thesmall intestinal epithelium and spleen during graft versushost disease. Tranplantation 1995;59:268-73.

28 Koch AE, Kronfeld-Harrington LB, Szeckanez Z, ChoMM, Haines GK, Harlow LA, et al. In situ expression ofcytokines and cellular adhesion molecules in the skin ofpatients with systemic sclerosis. Their role in early and latedisease. Pathobiology 1993;61:239-46.

29 Sollberg S, Peltonen J, Uitto J, Jimenez SA. Elevated expres-sion of beta 1 and beta 2 integrins, intercellular adhesionmolecule-I and endothelial leukocyte adhesion molecule-i

M230

on June 9, 2021 by guest. Protected by copyright.

http://mp.bm

j.com/

Clin M

ol Pathol: first published as 10.1136/m

p.49.4.M225 on 1 A

ugust 1996. Dow

nloaded from

http://mp.bmj.com/

Lymphocytes, cytokines and adhesion molecules in chronic graft versus host disease

in the skin of patients with systemic sclerosis of recentonset. Arthritis Rheum 1992;35:290-8.

30 Groves RW, Ross EL, Barker JNWN, MacDonald DM. Vas-cular adhesion molecule-1: expression in normal and dis-eased skin and regulation in vivo by interferon gamma.Am Acad Dermatol 1993;29:67-72.

31 Norton J, Sloane JP, Al-Saffar N, Haskard DO. Vessel asso-ciated adhesion molecules in normal skin and acute graft-versus-host disease. Clin Pathol 1991 ;44:586-91

32 Norton J, Sloane JP. A prospective study of cellular andimmunologic changes in skin of allogeneic bone marrowrecipients. Am Clin Pathol 1994;101:597-602.

33 Adams DH. Leucocyte endothelial interactions and regula-tion of leucocyte migration. Lancet 1994;ii:831-6.

34 Bevilacqua MP. Endothelial leukocyte adhesion molecules.Annu Rev Immunol 1993;11:767-84.

35 Thornhill MH, Kyan Aung U, Haskard DO. IL4 increaseshuman endothelial cell adhesiveness for T cells but notneutrophils. I Immunol 1990;144:3060-4.

36 Swerlick RA, Lee KH, Li L, Sepp NT, Caughman SW, Law-ley TJ. regulation of vascular cell adhesion molecule-i onhuman dermal microvascular endothelial cells. 7 Immunol1992;149:698-705.

37 Ushiyama C, Hirano T, Miyajima H, Okumura K, Ovary Z,Hashimoto H. Anti-IL 4 antibody prevents graft versushost disease in mice after bone marrow transplantation. 7Immunol 1995;154:2687-92.

38 Shulman H, Sale G, Lerner K, Barker A, Weiden P, Sullivan

K, et al. Chronic cutaneous graft versus host disease inman. Am 7 Pathol 1978;92:545.

39 Sullivan KM, Shulman HM, Storb R, Weiden PL,Witherspoon RP, MacDonald GB, et al. Chronic graft ver-sus host disease in 52 patients: adverse natural course andsuccessful treatments with combination immunosuppres-sion. Blood 1981;57:267-76.

40 Volc-Platzer B, Stingl G. Cutaneous graft-vs-host disease.In: Burakoff SJ, Deeg HJ, Ferrara J, Atkinson K, eds.Graft-vs-Host-Disease. Hematology. Vol 12. New York: Mar-cel Dekker, 1990:245-54.

41 Chosidow 0, Bagot M, Vernant JP, Roujeau JC, CordonnierC, Kuentz M, et al. Sclerodermatous chronic graft versushost disease. JAm Acad Dermatol 1992;26:49-55.

42 Janin-Mercier A, Devergie A, VanCauwenberge D, SauratJH, Bourges M, Lapiere CM, et al. Immunohistologic andultrastructural study of the sclerotic skin in chronic graftversus host disease in man. Am Pathol 1984;115:296-306.

43 Bailey AJ, Bazin S, Delaunay A. Changes in the nature of thecollagen during development and resorption of granula-tion tissue. Biochem Biophys Acta 1973;328:383-90.

44 Rowbottom AW, Norton J, Riches PG, Hobbs JR, PowlesRL Sloane JP. Cytokine gene expression in skin andlymphoid organs in human graft versus host disease.I ClinPathol 1993;46:341-4.

45 Piguet PF, Grau GE, Allet B, Vassalli P. Necrosisfactor-cachectin is an effector of skin and gut lesions of theacute phase of graft-versus-host disease. I Exp Med 1987;166:1280-9.

M231

on June 9, 2021 by guest. Protected by copyright.

http://mp.bm

j.com/

Clin M

ol Pathol: first published as 10.1136/m

p.49.4.M225 on 1 A

ugust 1996. Dow

nloaded from

http://mp.bmj.com/