J7Med Genet 1998;35:513-518

New lethal disease involving type I and IIIcollagen defect resembling gerodermaosteodysplastica, De Barsy syndrome, andEhlers-Danlos syndrome IV

Arja Jukkola, Saila Kauppila, Leila Risteli, Katri Vuopala, Juha Risteli, Jaakko Leisti,Leila Pajunen

AbstractWe describe the clinical findings andbiochemical features of a male childsuffering from a so far undescribed lethalconnective tissue disorder characterisedby extreme hypermobility ofthe joints, laxskin, cataracts, severe growth retardation,and insufficient production of type I andtype III procollagens. His features arecompared with Ehlers-Danlos type IV, DeBarsy syndrome, and geroderma osteo-dysplastica, as these disorders show somesymptoms and signs shared with ourpatient.The child died because of failure of the

connective tissue structures joining theskull and the spine, leading to progressivespinal stenosis. The aortic valve wastranslucent and insufficient. The clinicalsymptoms and signs, together with histo-logical findings, suggested a collagen de-fect. Studies on both skin fibroblastcultures and the patient's serum showedreduced synthesis of collagen types I andIII at the protein and RNA levels. The sizesof the mRNAs and newly synthesised pro-teins were normal, excluding gross struc-tural abnormalities. These findings arenot in accordance with any other collagendefect characterised so far.(7Med Genet 1998;35:513-518)

Keywords: types I and III collagen defect; gerodermaosteodysplastica; De Barsy syndrome; Ehlers-Danlos IV

The heritable disorders of connective tissue area large, heterogeneous group of disorders,some of which can be traced back to specificdefects in the genes of connective tissueproteins, whereas the molecular background ofothers still remains unknown.' In the presentstudy, we describe a child with findingssuggesting a generalised disorder of the skeletaland locomotive system. His signs includedextreme hypermobility of the joints, inability tomove, laxity of the skin, severe growth retarda-tion, osteoporosis, and progressive spinal ste-nosis, which proved fatal. We have comparedhis signs and biochemical findings with threepreviously characterised heritable disorders ofconnective tissue, geroderma osteodysplastica,De Barsy syndrome, and type IV Ehlers-Danlos syndrome, and suggest that he suffered

from an unknown disease entity combiningsome characteristics of all of them.

Case reportThe patient was born at term after an uncom-plicated delivery as the first child of healthy,non-consanguineous, 27 year old, Finnish par-ents. The pregnancy had been followed be-cause of mild hypertonia, oligohydramnios,and poor fetal growth. His birth weight was2180 g, length 43 cm, head circumference 32.5cm, and Apgar scores 9/9/9. He was small fordates with thin and translucent skin, increasedvisibility of superficial veins, reduced subcuta-neous fat tissue, and extremely loose joints andbilateral inguinal hernia, suggesting a connec-tive tissue disorder. Severe growth retardation(below -6 SD) was obvious. The infant alsosuffered from bilateral cataracts and disloca-tion of the left hip and had a club foot on theright side.The bilateral cataracts were operated on at

the age of 6 months and since then he success-fully wore contact lenses. There was noresponse to the attempts to correct the hip dis-location and it later became bilateral. Theinguinal hernias were operated on with appar-ently normal wound healing afterwards.The patient's motor development was se-

verely delayed, as he dislocated both large andsmall joints several times a day. He never learntto crawl, sit, or use his hands freely for playing.He was unable to hold his head in an uprightposition for any length of time, and mostlyspent his days sitting in a baby seat or on hisparents' lap. Muscular hypotonia and briskreflexes were observed right from the begin-ning, but significant clonus only became obvi-ous with increasing age and other symptoms.The boy spoke his first words at the age of 16

months but his speech development did notproceed further. He could understand simple,every day talk and used fewer than 10 shortwords, mimicking, and a couple of signs tocommunicate. Eye contact often made himproduce peculiar grimacing faces.By the age of 2 years 10 months he no longer

had the strength to turn from back to side, andcould lift his arm at maximum to a horizontallevel and hold only very light objects in hishands. By that time it also became evident thateven slightly upright positions were verypainful for him. CT scanning of the head andcervical vertebrae showed this to be because ofan anomalous atlas and dens filling the

Department ofOncology, Universityof Oulu, FinlandA Jukkola

Department ofMedical Biochemistry,University of Oulu,FinlandS KauppilaL Risteli

Department ofPathology, Universityof Oulu, FinlandK Vuopala

Department of ClinicalChemistry, Universityof Oulu, FinlandJ Risteli

Department of ClinicalGenetics, University ofOulu, FinlandJ LeistiL Pajunen

Correspondence to:Dr Pajunen, TahkokangasHospital, Kiilakiventie 5,FIN-90250 Oulu, Finland

Received 8 July 1997 Revisedversion accepted forpublication21 November 1997

513

on April 24, 2021 by guest. P

rotected by copyright.http://jm

g.bmj.com

/J M

ed Genet: first published as 10.1136/jm

g.35.6.513 on 1 June 1998. Dow

nloaded from

4Jukkola, Kauppila, Risteli, et al

4'

.,-

Figure 1 Pictures of the patient at 2 years of age (reproduced with permission). Anteroposterior and lateral profiles (A, B) showing sagging cheeks, olderappearance,frontalflatness, prominent nose, maxillary recession, mandibular prognathism, and visible superficial veins. Wrinkled skin, ankle deformities,and hyperextensibility of the joints (C), particularly in the fingers (D), are clearly visible. The radiographs of the forearm (E, 1 day old) and the knees (F,2years 4 months old) showed no obvious bone dysplasia, but showed the development of osteoporosis with increasing age and immobilisation.

foramen magnum so that only a 5 mm canalwas left for the spinal cord.As the patient's symptoms of spinal stenosis

were progressive, surgical decompression wasperformed. He recovered well from the opera-tion, but became increasingly restless and inpain during the next two weeks at home. Thecast for neck support was changed withoutcomplications. However, within 24 hours hedeveloped an unresponsive raised temperature,a heart rate over 200/min, and became uncon-scious. His blood pressure was extremely lowand did not react to medication. He died at theage of 2 years 11 months.

Table 1 Concentration ofamino terminal propeptide of type III procollagen (PIIINP) andthe carboxy terminal propeptide of type I procollagen (PICP) in serum samples. (Age andsex matched reference intervals in parentheses)

Patient Mother Father

PIIINP (tg/1) 9.6 (23-43) 3.6 (1.7-4.2) 3.2 (1.7-4.2)PICP (tgI1) 380 (800-2800) 190 (50-170) 124 (38-202)

Table 2 Concentration of the amino terminal propeptide of type III procollagen (PIIINP)and the carboxy terminal propeptide of type I procollagen (PICP) in the medium ofculturedfibroblasts. The results are means of duplicate determinations

PIIINP(in the medium pgll) PICP

Control 160 (100%) 4400 (100%)Patient 46 (29%) 700 (16%)

Clinical findings and necropsyThe boy was examined by a clinical geneticistimmediately after birth and regularly severaltimes a year until his death. His head was largewith frontal bossing and the fontanelles andcranial sutures were wide open. The mouthand nose were small, the neck was short, theleft ear was slightly dysmorphic, and his smalleyes had corneal clouding with bilateralcataracts, the fundus being normal. He kept hiswrists in volar flexion and his long fingersshowed hyperextensibility of the joints andloose, aged skin. The thumb was dug into thepalm and there were bilateral simian creases(fig 1A-D).The radiographs showed hip luxation and

pes calcaneovalgus on the left and club foot onthe right. The neurocranium was asymmetricaland flattened anteroposteriorly, the cranialbones were thin, and there were multiplewormian bones. The mandible was slightlyhypoplastic. The right side of the atlas wasattached to the base of the skull leading to mildscoliosis. No clearly recognised bone dysplasiawas evident and long tubular bones wereproportionally formed; osteoporosis was notprominent during the first months. However,with increasing age osteoporosis could be seenboth in the tubular bones and spine (fig 1 E, F).

Ultrasound examination of the abdomen wasunremarkable apart from agenesis of the rightkidney. He had a systolic murmur consideredto be harmless, but the echocardiogram

514

....p

on April 24, 2021 by guest. P

rotected by copyright.http://jm

g.bmj.com

/J M

ed Genet: first published as 10.1136/jm

g.35.6.513 on 1 June 1998. Dow

nloaded from

New lethal disease involving type I and III collagen defect

phoresis according to Laemmli,5 using 6.5%acrylamide in the separating gels. The radioac-tive bands were visualised by fluorography,using Kodak X-Omat autoradiographic films.

RADIOIMMUNOASSAYSRadioimmunoassays of the equilibrium satura-tion type were used to determine the concen-trations of the amino-terminal propeptide ofhuman type III procollagen (PIIINP) and thecarboxy-terminal propeptide of type I procolla-gen (PICP) in the culture medium, cellfractions, and serum samples, as measures oftype III and type I collagen synthesis, respec-tively. The radioimmunoassays used are spe-cific and showed no cross reactions with eachother.6 7

Figure 2 Sodium dodecylsulphate-polyacrylamide gel electrophoresis analysis of (H)proline containing macromolecules in the medium and in the cellfraction of culturedfibroblasts after 20 hours of continuous labelling. All samples were reduced beforeelectrophoresis. The samples were medium of control cell (lane 1), medium ofpatient cells(lane 2), cellfraction of control cells (lane 3), cell fraction ofpatient cells (lane 4), andimmunoprecipitates obtained with antibodies against the amino-terminal propeptide of typeIII procollagen (PIIINP) in the medium of control cells (lane 5), medium ofpatient cels(lane 6), cellfraction of control cells (lane 7), and cellfraction ofpatient cells (7ane 8). Themolecular mass markers, indicated by arrows, were reduced proal (III) or proal (I) (1),pNal (I) orpNal (III) (2), al (I) or al (III) (3), and a2(I) (4).

showed normal intracardiac anatomy andfunction. Blood lymphocyte karyotype, routineblood studies, and screening tests for metabolicdisorders showed no pathological findings. Inthe CT scan taken some weeks before hisdeath, frontotemporal cortical atrophy wasseen, in addition to the spinal stenosis.At necropsy, the aortic valve was thin, trans-

lucent, and insufficient, which was surprisingas the echocardiogram a few days before deathhad been normal. The left side of the myocar-dium was hypertrophic. There were congestivechanges in the lungs and liver. The centralnervous system was macroscopically normallyformed, but the gyri of the cerebrum were flat-tened. There was oedema and ischaemicchanges in the frontal cortex. These findingswere confirmed by a neuropathological study,in which no additional findings were found.

Materials and methodsCELL CULTURES AND PROCOLLAGEN LABELLINGA skin biopsy was taken from the patient at theage of 7 months for obtaining fibroblastcultures. These cells and primary skin fibro-blasts from a healthy adult control were main-tained in Dulbecco's modification of Eagle'smedium as described previously in detail.2Confluent cultures were labelled overnightwith L-(2,3,4,5-3H) proline, and the mediumand the cell layer were treated and immunopre-cipitated according to published methods.'

IMMUNOPRECIPITATION, GEL ELECTROPHORESIS,AND FLUOROGRAPHYImmunoprecipitation was carried out essen-tially as described by Cooper et al.4 Thelabelled proteins were analysed by sodiumdodecylsulphate-polyacrylamide gel electro-

RNA ANALYSISTotal RNA was isolated from confluent fibro-blast cultures of a healthy control and thepatient's cell line.8 For northern blot hybridisa-tions, cDNA probes for the genes of the proa 1chain of type III procollagen,9 the proa2 chainof type I procollagen,'0 elastin," and ,B actin(Clontech Laboratories, CA) were used sepa-rately. The filter was exposed overnight to theradiographic film after being hybridised withthe probe for 3 actin and for seven days afterthe hybridisations with the probes for type Iand III collagen and elastin.

HISTOLOGICAL STUDIESSkin biopsies taken from the volar surface ofthe left forearm at the age of 2 years 4 monthsas well as the tissue samples collected at

Figure 3 Northern blot analysis ofRNAs. Lanes 1 and 3contain 27 ug of total RNA isolatedfrom the patient'scultured skin fibroblasts; 30 ,ug of total RNA isolatedfromcontrolfibroblasts were run in lanes 2 and 4. Lanes 1 and 2were hybridised with a probe corresponding to humanproal (III) collagen mRNAs, and lanes 3 and 4 with aprobe corresponding to human proa2(I) collagen mRNAs.The bands below represent control hybridisations with aprobe corresponding to fi actin mRNAs.

0-.

515

on April 24, 2021 by guest. P

rotected by copyright.http://jm

g.bmj.com

/J M

ed Genet: first published as 10.1136/jm

g.35.6.513 on 1 June 1998. Dow

nloaded from

56ukkola, Kauppila, Risteli, et al

d. >'it ' , ' X '

p.~~~ ~ ~

xA.. A S , s

.9....

Figure 4 Electron micrograph of the dermis offorearm skin. (A) The collagen fibres are accompanied by thin,disorganisedfibres (arrowhead) without adjacentfibroblasts. (B) Elasticfibres (star) seem to be composed of tubularfibrils,but the core of amorphous material seems to be lacking.

necropsy were fixed in buffered 4% formalin,embedded in paraffin, and stained withhaematoxylin-eosin and Alcian blue-periodicacid-Schiff methods. Elastic fibres were visual-ised with the Verhoff stain.For electron microscopy, tissue samples

from skin and aorta were first fixed in 2.5%glutaraldehyde, then postfixed in 1% OSO4 in0.1 mol/l phosphate buffer, dehydrated inacetone, and embedded in Epon LX 1 12.Ultrathin sections were cut with a ReichertUltracut E-ultramicrotome (Reichert-Jung,Wien, Austria) and examined in a 410 LStransmission electron microscope (PhillipsExport BV, Eindhoven, The Netherlands).

ResultsPRODUCTION OF TYPES I AND III PROCOLLAGEN INVIVO AND IN VITROThe circulating concentrations of two propep-tides set free during the synthesis of type I andtype III collagens, abbreviated PICP andPIIINP, respectively, were within the referenceintervals for age in the mother and father of thepatient. However, in the repeated analyses ofthe serum of the patient both concentrationswere clearly below the reference intervals forage (table 1), suggesting a defect that affectsboth these major collagens.The same methods were used for quantifying

the secretion of types I and III procollagen inconfluent cultures of fibroblasts. In the pa-tient's cultures the concentration of PIIINP inthe culture medium was less than a third andthat ofPICP less than a fifth of the correspond-ing concentrations in the control culture (table2). To characterise the procollagens producedby the cells further, confluent cultures werelabelled with radioactive proline and the newlyformed radioactive proteins studied by gel

electrophoresis and immunoprecipitation withspecific antibody against type III procollagen(fig 2). The medium and cell layer fractions ofa control culture of adult fibroblasts containedthe characteristic bands of type I and type IIIprocollagens and collagens,' whereas thepolypeptide patterns of the patient's culturewere clearly different. In the control, the bandsrelated to type III procollagen could be easilyvisualised by immunoprecipitation with anti-PIIINP antibodies (fig 2, lane 5), whereasnothing could be seen in the patient's mediumtreated identically (fig 2, lane 6). All theseresults agree with an almost total lack of typeIII procollagen production, but type I procolla-gen production is also defective.

MESSENGER RNA ANALYSISNorthern blot hybridisation of the RNAproduced by the patient's fibroblasts showedthe two transcripts corresponding to the proa2chain of type I procollagen in amounts remark-ably lower than those produced by controlcells. The two type III procollagen mRNAbands were barely visible in the patient'ssamples. Hybridisation with the elastin probeshowed that the amount of elastin mRNA wasalso decreased in the patient's fibroblasts (datanot shown). f actin mRNA was analysed as aninternal control and was clearly visible both inthe patient and control samples. The sizes of allthe mRNA species studied appeared normal(fig 3).

HISTOLOGICAL FINDINGSIn skin biopsies, the dermis appeared thinnerand the elastic fibres fragmented and lessnumerous than in age matched control skin.Ultrastructurally, the elastic fibres seemed tolack the amorphous body normally observed,

516

on April 24, 2021 by guest. P

rotected by copyright.http://jm

g.bmj.com

/J M

ed Genet: first published as 10.1136/jm

g.35.6.513 on 1 June 1998. Dow

nloaded from

New lethal disease involving type I and III collagen defect

Table 3 Comparison of the patient with Ehlers-Danlos syndrome type IVgerodermaosteodysplastica, and De Barsy syndrome

Patient EDSIV GO De Barsy

Severe growth retardation + - + +Mental retardation - - - +Progressive microcephaly - - - +Cataracts/corneal opacities + - - +Hypermobility of small joints + + + +Hypermobility of large joints +Generalised osteoporosis - - + +Hernias + +Thin translucent skin/visible veins + + + +Loose skin + - + +Vascular manifestations + +Inheritance ? AD AR ARPremature ageing + - + +Abnormal elastic fibres + - +/- +Cartoon dwarf-like facial appearance + - +Verified type I collagen defect + -

Verified type III collagen defect + +

AR=autosomal recessive inheritance, AD=autosomal dominant inheritance. + indicates positivefindings and - indicates both negative findings and findings not known.

and appeared fragmented and overstained (fig4A). In some locations, the collagen fibres werein close contact with a delicate meshwork ofapparently disorganised fibres (fig 4B). A strik-ing feature of the dermis was the sparseness, oreven lack, of fibroblast cells.

In the biopsy taken from the ascending aortaat necropsy, collagen fibres were of normalappearance, but there was more material stain-ing with Alcian blue, suggesting increasedamounts of proteoglycan, between the collagenfibres. The elastic fibres appeared fragmentedbut evenly stained.

DiscussionOur patient suffered from a disorder character-ised by extreme hypermobility of the joints,inability to move, laxity of the skin, severegrowth retardation, cataracts, local osteoporo-sis, and progressive spinal stenosis. He hadseveral features in common with three inher-ited diseases affecting connective tissue,growth, and ageing (table 3): subtype IV ofEhlers-Danlos syndrome (EDS), gerodermaosteodysplastica (GO), and De Barsy syn-drome. A verified type III collagen defect ischaracteristic of EDS IV,` but our patient hadno bruising or scarring tendencies. His growthretardation and poor development were also fartoo severe for a typical EDS IV patient, and thetype I collagen defect observed is definitely notpresent in this disease (table 3).Some of the findings are similar to cardinal

features described in patients with GO (table3)." In particular, his facial appearance withthe small, slightly dysmorphic left ear and hisclose resemblance to the dwarfs of WaltDisney's Snow White are remarkably similar tothe proband (III.3) of the second family in thereport ofHunter et al'4 on GO. However, severegeneralised osteoporosis has been one of themain features in other published GO patients.In our patient, osteoporosis became evidentonly later and increased with age and longimmobilisation. Also, the facial skin was notwrinkled, but rather had a tight appearance. Asfar as we know, all geroderma patients haveeventually been ambulatory and had normalspeech development, unlike our patient.

Both abnormal1" 15 and normal'6 findings inskin biopsies have been reported in GO. Thefragmentation of the elastic fibres seen in skinin GO as well as in our patient is not necessar-ily a specific finding. To our knowledge, noabnormal biochemical collagen findings havebeen reported in this disorder.

Degeneration of the elastic tissue in skin andcornea, together with retarded growth andmental retardation, are features of De Barsysyndrome.'7 Mental retardation is severe inmost patients and some patients never learn tospeak, but intelligence can also be only mildlyimpaired,'8 or even normal. The patients showathetoid and choreiform movements and pecu-liar grimacing. With the exception of the reportby Hoefnagel et al,'9 degeneration of collagen-ous or elastic fibres has usually been observedwhen a skin biopsy has been performed. Theposmatal development of microcephaly, typicalof De Barsy patients, was not found in ourpatient, but occasional peculiar grimacing wasseen.Our patient's fibroblasts showed clearly

diminished production ofmRNA for both typeI and III procollagen (fig 3). Also theproduction of the corresponding proteins wasdecreased, suggesting possible defects at thegene level. Practically no newly synthesisedtype III procollagen could be detected in thecell cultures, indicating minimal production ofthis protein (fig 2). The control samples werefrom adult skin, but this should not affect theresults as collagen synthesis is at its most activein infancy and early childhood.20The serum findings agree with the in vitro

results on type III procollagen production(table 1). The circulating concentration of PII-INP, a propeptide released from type IIIprocollagen, was 50 to 80% lower than normalfor his age group. A similar decrease was seenin the concentration of PICP, a propeptidederived from type I collagen synthesis. Theparents showed concentrations within the adultreference intervals in both analyses (table 1).The similarly decreased production rate of

two different collagen types was a surprisingfinding. The three genes for type I (COLlAland COL1A2) and type III (COL3A1) procol-lagens are all located on differentchromosomes.21 However, it is highly improb-able that there would be mutations in all oreven two of these genes in the same patient. Wedid not detect any gross structural defectseither in the proteins or in the correspondingmRNAs. In fibroblast cells, the expression oftype I and type III procollagen is tightlycoordinated," whereas in mature osteoblastsonly type I collagen is produced. Neither ofthese regulatory mechanisms is completelyknown at the molecular level.Our patient could in principle have a defec-

tive gene for some regulatory protein involvedin the coordinated regulation of type I and typeIII collagen expression. As a genetic defect thisshould have led to a combined clinical pictureof severe osteoporosis or osteogenesis imper-fecta and EDS IV, being most probably earlylethal. However, our patient does not fulfil thecriteria of osteogenesis imperfecta or severe

517

on April 24, 2021 by guest. P

rotected by copyright.http://jm

g.bmj.com

/J M

ed Genet: first published as 10.1136/jm

g.35.6.513 on 1 June 1998. Dow

nloaded from

58ukkola, Kauppila, Risteli, et al

osteoporosis. Radiographic findings in thepatient did not show the changes of osteogen-esis imperfecta. The immunological and bio-chemical methods showed reduced synthesis oftype I procollagen, but also an extensive lack oftype III procollagen, which is not typical ofosteogenesis imperfecta. In osteogenesis im-perfecta, procollagen type III synthesis is notaffected.The clinical picture and the collagen defect

of the patient resemble the process of acceler-ated senescence of the fibroblasts and othercells of mesenchymal origin, as the capacity forcollagen turnover tends to become lower. Thisis indirectly suggested in our patient by analmost normal collagen content of the skintogether with a minimal amount of newly syn-thesised collagen.

Genetic counselling for this family is difficultas both autosomal recessive and dominantinheritance are found in the disorders dis-cussed here. Being the offspring of clinicallyhealthy parents with biochemically normal col-lagen findings, our patient's mutation is mostprobably new if dominantly inherited. How-ever, the other modes of inheritance cannot beexcluded either. The radioimmunoassays usedin this study were clinically useful and easy fordiagnosing the major changes involving thetype I and III collagen metabolism of ourpatient. If the collagen defects detected in thischild could be measured directly in amnioticfluid as in osteogenesis imperfecta (Kauppila etal, unpublished data), prenatal diagnosis usingspecific radioimmunoassays could be offered tothis family in the future.

We gratefully acknowledge the expert technical assistance ofMsKristiina Apajalahti and Ms Outi Laurila.

1 Beighton P. McKusick's heritable disorders of connective tissue.St Louis: Mosby, 1993.

2 Jukkola A, Risteli J, Niemela 0, Risteli L. Incorporation ofsulphate into type III procollagen by cultured humanfibroblasts. Identification of tyrosine 0-sulphate. EurJ Bio-chem 1986;154:219-24.

3 Jukkola A, Risteli J, Risteli L. Effect of dextran on synthesis,secretion and deposition of type III procollagen in culturedhuman fibroblasts. Biochem J 199 1;279:49-54.

4 Cooper AR, Kurkinen M, Taylor A, Hogan BLM. Studieson the biosynthesis of laminin by murine parietalendoderm cells. EurJ_ Biochem 198 1;119: 189-97.

5 Laemmli UK. Cleavage of structural proteins during theassembly of the head of bacteriophage T4. Nature1970;277:680-5.

6 Risteli J, Niemi S, Trivedi P, Maentausta 0, Mowat AP, Ris-teli L. Rapid equilibrium radioimmunoassay for theaminoterminal propeptide of human type III procollagen.Clin Chem 1988;34:715-18.

7 Melkko J, Niemi S, Risteli L, Risteli J. Radioimmunoassay ofthe carboxyterminal propeptide of human type I procolla-gen. Clin Chem 1990;36:1328-32.

8 Chirgwin IM, Przybyla AE, MacDonald RJ, Rutter WJ. Iso-lation of biologically active ribonucleic acid from sourcesenriched in ribonuclease. Biochemistry 1979;18:5294-9.

9 Loidl HR, Brinker JM, May M, et al. Molecular cloning andcarboxylpropeptide analysis ofhuman type III procollagen.Nucleic Acids Res 1984;12:9383-94.

10 Bernard MP, Myers JC, Chu ML, Ramirez F, EikenberryEF, Prockop DJ. Structure of a cDNA for the proa2 chainof human type I procollagen. Comparison with chickcDNA for proa2(I) identifies structurally conservedfeatures of the protein and the gene. Biochemistry 1983;22:1139-44.

11 Fazio MJ, Olsen DR, Eunkyung AK, et al. Cloning offull-length elastin cDNA library: further elucidation ofalternative splicing utilizing exon-specific oligonucleotides.Invest Dermatol 1988; 91:458-64.

12 Pope FM, Nicholls AC, Narcisi P, et al. Type III collagenmutations in Ehlers Danlos syndrome type IV and otherrelated disorders. Clin Exp Dermatol 1988;13:285-302.

13 Lisker R, Hernandez A, Martinez-Lavin M, et al. Geroder-mia osteodysplastica hereditaria: report of three affectedbrothers and literature review. Am J Med Genet 1979;3:389-95.

14 Hunter AGW, Martsolf JT, Baker CG, Reed MH.Geroderma osteodysplastica. A report of two affectedfamilies. Hum Genet 1978;40:311-24.

15 Klein D, Bamatter F, Franceschetti A, Boreaux G, BrocherJEW, Holenstein P. Une affection liee au sexe: lagerodermie osteodysplasique hereditaire (20 ansd'observation). Rev Oto-Neuro-Ophtalmol 1968;40:415-21.

16 Suter H, Tonz 0, Scharli A. Geroderma osteodysplasticaheriditaria (GOH) in a girl. In: Papadatos CJ, BartsocasCS, eds. Skeletal dysplasias. New York: Alan R Liss,1982:327-9.

17 De Barsy AM, Moens E, Dierckx L. Dwarfism, oligophreniaand degeneration of the elastic tissue in skin and cornea. Anew syndrome? Helv PaediatrActa 1968;23:305-13.

18 Kunze J, Majewski F, Montgomery P, Hockey A, Karkut I,Riebel T. De Barsy syndrome-an autosomal recessive,progeroid syndrome. EurJ Pediatr 1985;144:348-54.

19 Hoefnagel D, Pomeroy J, Wurster D, Saxon A. Congenitalathetosis, mental deficiency, dwarfism and laxity of skinand ligaments. Helv PaediatrActa 197 1;26:397-402.

20 Trivedi P, Risteli J, Risteli L, Hindmarsh PC, Brook CGD,Mowat AP. Serum concentrations of type I and III procol-lagen propeptides as biochemical markers of growth ininfants and children with growth disorders. Pediatr Res199 1;30:276-80.

21 Vuorio E, de Crombrugghe D. The family of collagen genes.Annu Rev Biochem 1990;59:837-72.

22 Vuorio T, Makela JK, Kiihari VM, Vuorio E. Coordinatedregulation of type I and type III collagen production andmRNA levels ofproal (I) and proa2(I) collagen in culturedmorphea fibroblasts. Arch Dermatol Res 1987;279:154-60.

518

on April 24, 2021 by guest. P

rotected by copyright.http://jm

g.bmj.com

/J M

ed Genet: first published as 10.1136/jm

g.35.6.513 on 1 June 1998. Dow

nloaded from