0022-202X/ 85/8405- 0374$02.00/0 THE ,)O ltNAI. OF INVESTIGATIV E DEitMATOLOr.Y, 84:374- 377, 1985 Vol. 84, No.5 Copyrighl 1985 by T he Williams & Wilkins Co. Printed in U.S .A. REPORTS

Evaluation of Anchoring Fibrils and Other Components of the Dermal­ Epidermal Junction in Dystrophic Epidermolysis Bullosa by a Quantitative Ultrastructural Technique*

MICHAEL J. TIDMAN, M.B., B .Sc., M.R.C.P.t AND ROBIN A. J. EADY, M.B., F.R.C.P. Department of Electron Microscopy and Ce ll Pathology, Institute of De rmatology, London, U.K.

To examine the possibility that differences in the may be indistinguishable from the dominant condition (DDEB) structure and population density of anchoring fibrils [2]. Severe RDEB, however, is a much more aggressive disease (AF) and other components of the dermal-epidermal characterized by generalized fragility of the skin and wide­ junction might distinguish between genetically and clin­ spread scarring of the skin and mucous membranes, associated ically distinct varieties of dystrophic epidermolysis bul­ with stunted growth , c hronic anemia, esophageal strictures, losa (DEB), a controlled ultrastructural morphometric dental decay, pseudosyndactyly, and fl exion contractures (1] . study of nonseparated keratinocyte-associated dermal­ This typical clinical picture all ows ready distinction from the epidermal junction was undertaken in a total of 17 localized form. patients with DEB. Seven patients had dominant DEB, In all variants of DEB, blisters arise immediately beneath 3 had localized recessive DEB, and 7 had severe, gen­ the lamina densa of t he dermal-epidermal junction (DEJ) in eralized · recessive DEB. Nonlesional, unscarred skin the most superficial region of the dermis [3], where anchoring was obtained from standard body regions. Criteria for fibrils (AF) are normally located. The biochemical nature and the identification of AF were a mandatory union with function of AF are still not known, but the strong union that the lamina densa and the presence of central banding normall y exists between the lamina densa a nd AF (4], which and/or fanning of the extremities. are derived from epidermis and dermis respectiv ely [5], suggests No AF were detected in 9 technically suitable samples that AF serve as attachment devices. In DDEB the AF have from patients with severe recessive DEB. Structurally been reported to be reduced in number (6] and abnormal in normal AF were present, but significantly reduced in structure [7,8]. Furthermore, in the Pasini type of DDEB the number, in both dominant a nd localized recessive DEB, reduction in AF numbers was found in both bli ster predilection compared with site-matched samples from 12 healthy and nonpredilection sites [7], whereas in t he milder Cockayne­ adults. There was no difference in AF characteristics Touraine form, AF numbers were noted to be decreased only between dominant and localized recessive DEB, or be­ in predilection sites (8] . Some workers have found that AF tween sites of predilection and non predilection for blis­ were present, but reduced in number, in both locali zed and ters. The presence or absence of albopapuloid lesions in ge neralized forms of RDEB [9], whereas others have described dominant DEB did not influe nce AF counts. There was a total absence of AF in severe generalized RDEB [10,11]. no difference in numbers of hemidesmosomes, basal cell Quantitative data on the AF in DEB are limited, but a plasmalemma) vesicles, or dermal microfibril bundles in morphometric analysis of the components of normal DEJ has any g roup of DEB patients compared with controls. demonstrated that AF numbers vary widely among individuals, Thus, a lthough severe mutilating DEB can be distin­ between body regions in the same individual, and even within guished by routine transmission electron microscopy, a s ingle s kin sample, and has highlighted the potential pitfalls the dominant and localized recessive forms cannot be in evaluating AF numbers from isolated microscopic fields differentiated on the basis of AF structure or numbers. [1 2]. The aim of t he present investigation was to use the same morphomet ri c technique on skin from patients with DEB to quantify AF a nd other components of the DEJ, and to ascertain Dystrophic epidermolysis bullosa (DEB) may be inherited whether DDEB and RDEB can be distinguished using ultra· autosomall y in either a dominant o r a recessive manner, each structural criteria. This would have o bvious implications for type having a wide spectrum of severi ty [1 ]. This distinction both prognostication and ge netic counseling, and may fa cilitate often proves difficult in sporadic cases because the clinical the prenatal diagnosis of DEB. features of the mild locali zed form of recessive DEB (RDEB)

MATERIALS AND METHODS ManuRcript received July 24 , 1984; accepted for publication Novem· ber 2, 1984. Patients and Co ntrols Thi s work was supported hy grants from the Dystrophic Epidermo· Seventee n patients with DEB were studied, comp ri sing 7 with lysis Bu ll osa Research Association and the Dunhill Trust. DDEB, 3 with mi ld locali zed RDEB , a nd 7 with severe generali zed *Presented in part at the Join t International Meeting of The Society RDEB. In each case the diagnosis was unequivocal and based on clinical for Investigative Dermatology, In c. and t~e European Society for and ge netic features. Electron microsco py was used to establish that Dermatological Research, Washingto n, D.C., Apr il 27- May 1, 1983. blisterin g occurred beneath the lamina densa. All the subjects with Abstract in J Invest Dermatol 80:342, 1983. DDEB had affected fam ily members in 2 or more ge nerations, and each 1' Dr. Tidman is a Medical Research Coun cil Training Fell ow. of th e 3 patients with the mild loca lized form of RDEB had a simi larly Reprint requests to: Robin A. J. Eady, M.S. , Institute of Dermatol­ affected sibling in the abse nce of parental involvement. Five of the ogy, Homerton Grove, London E9 6BX, England. patients with seve re genera lized RDEB were s poradic cases and the Abbreviations: other 2 were siblin gs (CH and SH). AF: anchorin g fibril(s) The DDEB group included a father and daughter (DP and SG) and DDEB: dominant dystrophic epidermolysis bu llosa a grandmother and grandso n (BB and KM) . In 6 of the 7 patients the DEB: dyRtroph ic ep idermolys is bu llosa co ndition was localized, with blistering, sca rring, and mi lia formation DEJ: derm al-epide rmal junction co nfined to bony prominences on the extremiti es, and with typical RDEB: recessive dystrophic epidermo lysis bullosa dystrophic changes of the nai ls. The seventh patient (KB) had a more 374 May 1985 ANCHORING FIBRILS IN DYSTROPHIC EPIDERMOLYSIS BULLOSA 375

widespread invo lvement wi th extensive scarring of his trunk a nd limbs, and basal plasma membrane, excluding DEJ associated with melano­ a nd acLi ve bli stering co ntinuing in t hese a reas despite his advanced cytes. The numerical densities of AF and dermal mi crofibril bundles age. He occasionally s uffered with bli stering of t he buccal mucous wi t h respect t.o lamina den sa, a nd of hemidesmosomes and plasmalem: membranes and episodes of dysphagia had been associated with radio­ mal vesicles, with respect to basal plasma membrane, were determined, logic evidence of an esophageal web. However, t he co ndition could be using the criteria previously defin ed [1 2], by the same observer. An AF t r aced through 3 ge nerations a nd there was no evidence of fin ge r was counted only if it united with the la mina densa and demonstrated webbing. In this group t he presence or absence of albopapul oid lesions, irregular banding of the central portion or fanning of its extremities or which a re thought to be a dist inguishing feature of t he Pasini variant both these characteristics, thus excluding wisp like structures normally of DDEB [1 ,13], was recorded (Table 1). Albopapuloid lesions were present beneath t he la mina densa, and avoiding confusion with other defined as hypopigmented macules or papules, with a wrinkled surface, fibrillar structures such as elastic mi crolibrils a nd co ll agen fibrils. u s ually located over the lower back. The absence of fresh, natural blisters in all 7 subjects wi th DDEB at the t ime of co nsultation and RESULTS t h e inability to experimentally provoke blister formation by vigorous friction with a pencil eraser, as is often the case [1 4], prevented the The c linical details a nd AF counts for the DDEB group are ult rastructural confirmation of a sub-lamina densa separation, al­ recorded in Table I. The presence o f albopapuloid lesions did t h ough clinically the diagnosis in each patient was beyond doubt. not appear to bear a direct relationship to the clinical severity The 3 s ubjects with mi ld locali zed RDEB included 2 sisters (MD of the co ndition, and it is of interest that 1 of the subjects (IH) a nd BB). In both, s kin lesions were confined to bony prominences on with the mild locali zed form of RDEB also had albopapuloid t h e extremities and consisted of old scarring as blisters had occurred lesions over the lumbar r egion, as did his 7-year-old brother. o nly infrequently s ince c hildhood. Both had suffered episodes of dys­ The possession of albopapuloid lesions does, however, seem to phagia a nd mild iron-defi ciency anemia, and an esophageal web h ad be a familial trait. been co nfirmed radiologically in one. The t hird subject (!H) had In subjects with DDEB, compared wi th site- matched samples extremely mild di sease affecting knees a nd ankles without nail dystro ­ phy or gastrointestinal involvement. from healthy adults using an unpaired t-test, AF were signifi­ The 7 patients with severe ge nerali zed RDEB all manifested t he cantly reduced in number in both the lower leg (p < 0.001) and mutilating changes that typify t his condition [1 ,2 ], wi th, in particular, thigh (p < 0.001), with no difference apparent between these 2 extensive c utaneous and oral mucous membrane scarring, pseudosyn­ sites (p > 0.5). Compared with normal skin (Fig 1) , the AF d actyly, a nd fl exion co ntractu res of t he digits. appeared morphologically normal (Fig 2) with no evidence The co nt rol group co nsisted of 7 males and 5 females, aged 20-60 years, who were the subjects in a detail ed morphometri c study of normal DEJ [1 2].

Tissue Samples and Processing Nonlesional and clinically unscarred skin, wi th n o history of having previously blistered, was obtained, after infiltration of the surrounding a rea wit h local anesthetic, from the inner aspects of the lower leg (just above the ankle) and thigh . These 2 s ites represent, respectively, areas of predilection and nonpredilection for blister formation in the milder forms of DEB, and in normal skin AF counts in these 2 regions do not differ [ 12] . In severe RDEB there is a ge neralized blistering tendency, and the skin is so fragile that the trauma of the biopsy may be sufficient to cause dermal-epidermal separation. Samples were processed in a standard fashion for transmission electron microscopy as previously described [1 2]. Briefly, tissue was fixed in 2% formaldehyde and 2.5% glu taraldehyde in 0.04 M (final concentration) cacodylate buffer co ntaining 0.05% calcium chloride a nd 5% sucrose (pH 7.4), and then postfixed in 1.3% os mium tetroxide buffered wi th 0.067 M s-collidine (pH 7.4) on i ce, before dehydration in ethanol and embedding in Epon 812. Ult rathin sections were cut from randomly selected blocks, stained with uranyl acetate and lead cit rate, and viewed with a JEOL 100 CX electron microscope. Electron micrographs of overlapping fields of randomly selected segments of interappendageal keratinocyte-associated DEJ were taken, printed, and assembled into montages wit h a fina l magnification of between 35,000x a nd 43,000X, accurately determined by means of a calibration grating.

Morphometry A M.O.P. semiautomatic im age analysis system (Reichert-Jung, . FIG 1. The normal dermal-epidermal junction, with the a nchoring Vienna, Au stria) was used to measure 40-pm lengths o f lamina densa fibnls arrowed. Calibration. bar= 0.25 pm.

TABLE I. Dominant dystroph ic epidermolysis bullosa: patient d e taiL~ and anchoring fibril counts

Age/sex Albopapuloid Anchoring fibri l count• Subject Clin ica l severity (yrs) lesions Lower leg Thigh Controls (n = 12) 20-60/7M:5F 63.5 ± 12.5 (SEM)b 61.9 ± 12.4 (SEM)b (range 16- 100) (range 19- 120) DP 50/M Localized Absent 28 31 SG 28/F Localized Absent 12 5 BB 73/F Localized Present 0 0 KM 25/M Locali zed Present 0 2 MS 39/F Localized Present 0 25 PE 29/ M Localized Not known ' 2 d KB 60/M Generali zed Absent 0 0 "Number of anchoring fibrils per 40-pm lamina densa. b Corrected t o a llow for individual differences. ' Not recorded at initial consultation and patient died shortly a fterwards from a cerebrovascular accident. d Tissue sample tec hnically unsatisfactory. 376 TIDMAN AND EADY Vol. 84, No. 5 elsewhere of damaged or rudimentary structures. Comparin g DEJ. There was no evidence of structures that could be inter­ patients with and wi thout albopapuloid lesions, AF counts in preted as damaged or abnormal AF. t hi gh skin were not significantly different (p > 0.5). One patient The numerical data on the other major components of the (DP) had AF numbers in both t he lower leg and thigh which DEJ are recorded in Table III. No significant differences from fell within the lower part of the normal range, although hi s daughter (SG) had fewer. The 2 subjects (BB and KB) for whom no AF were recorded, both possessed sparse numbers of morphologically normal AF, which, by chance, were not located within t he measured lengths of DEJ. The AF data from the 3 patients with mild locali zed RDEB showed a similar pattern (Table II). All 3 subjects possessed AF with a normal morphology (Fig 3), but their numbers were significantly reduced in the thigh (p < 0.001) and also reduced in the lower leg, although t he data from t he latter site were too few to enable a proper statistical comparison. No difference was apparent in t he AF counts in t he t high skin of patients with mild RDEB and DDEB (p > 0.5). One subject with locali zed RDEB (MD) had AF counts in both sites which fell wi thin t he lower part of the normal range, and although no AF were recorded in the t high of her sister (BB), scanty AF were present outside the measured lengths of DEJ. In 5 of the 7 subjects with the severe mutilating form of RDEB, the trauma of the biopsy procedure caused dermal­ epidermal separation t hroughout skin samples taken from the FIG 3. The dermal-epidermal junction in mild, localized recessive lower leg (Table II). In the 9 remaining specimens, no AF were dystrophic epidermolysis bullosa (subject MD, lower leg), demonstrat­ seen (Fig 4) along lengths representing up to 130 J.Lm of intact ing structurally normal anchoring fibrils (arrows ) beneath t he lamina densa. E = epidermis, D = dermis. Calibration bar= 0.5 J.lm .

FIG 4. The dermal-epidermal junction in severe, generali zed reces­ FIG 2. T he dermal-epidermal junction in dominant dystrophic epi­ sive dystrophic epidermolysis bullosa (subject VH , t high) , show ing an dermolysis bullosa (subject DP, lowe r leg), demonstrating morpholog­ absence of anchoring fibrils beneath the lamina densa. Among the ically normal anchoring fib ril s (arrows). LD =Lamina densa (obliquely poorly stained coll agen fibrils (C) are fibrils of small diameter which sectioned) . Calibration bar = 0.25 J.lm. are well -stained (arrow). E =epidermis. Calibration bar= 0.5 11 m.

TABLE II. Recessive dystrophic epidermolysis bullosa: patient details and anchoring fibril counts

Age/sex Sublam ina densa Anchoring fibril coun t" Subject (yrs) Clinical severity spli t Lower leg Thigh Controls (n = 12) 20-60/7M:5F 63.5 ± 12.5 (SEM) 61.9 ± 12.4 (SEM) (range 16- 100) (range 19- 120) MD 57/F Locali zed Present 19 21 BB 61/F Localized Present 11 0 IH 12/M Locali zed Not done ND" 27 CH 24/F Generalized Present 0' od SH 22/M Generali zed Present ND' 0 so 15/M Generali zed Not done 0 0 VH 3/M Generalized Present ND' 0 ES 12/F Generalized Present NO' 0 JB :3/M Generalized Present NO' 0 SB 32/M Generalized Not done ND' 0 "Number of anchoring fibrils per 40-llm lamina densa. 1 ' Not done because tissue sample technica lly unsatisfactory. ' On ly 30 11m of unseparated skin avail able fo r evaluation. d Only 24 11m of unseparated skin avail able for evaluation. •· Not done because dermal-epidermal separation throughout t he biopsy specimen prevented evaluation. May 1985 ANCHORING FIBRILS IN DYSTROPHIC EPIDERMOLYSIS BULLOSA 377

TABLE III. Numerical values of h.emidesmosomes, plasmalemma/ vesicles, and dermal microfibril bundles in dystrophic epidermolysis bullosa Co ntrols" DDEB Mi ld RDEB Severe RDEB Lower leg Thigh Lower leg Thi gh Lower leg Th igh Lower leg Th igh (n = 12) (n = 12) (n = 7) (n = 6) (n = 2) (n = 3) (n = 2) (n = 7) Hemidesmosomes• 66.1 ± 5.7 74.0 ± 5.7 59.7 ± 3.0 59.2 ± 4.7 72.0 74.7 ± 3.9 59.5 61.3 ± 5.0 P lasmalemma! vesi- 25.8 ± 5.8 27. 1 ± 5.8 25.0 ± 2.1 24.7 ± 3.4 40.8 30.3 ± 1.5 24.5 30.3 ± 6.7 cles• Dermal microfibril 0.8 ± 0.3 1.0 ± 0.3 1.1 ± 0.4 0.5 ± 0.3 0 1.3 ± 0.9 0.5 1.0 ± 0.4 bundles' Abbreviations: DDEB = dominant dystrophic epidermolysis bullosa. Mild RDEB = localized form of recessive dystrophic epidermolysis bullosa. Severe RDEB = generali zed form of recessive dystrophic epidermolys is bullosa. • Values adjusted for differences among individuals. • Mean ± SEM per 40-l'm basal plasma membrane. 'Mean ± SEM per 40-l'm lam in a densa. con trol values were present for he midesmosome, plasmalemma! We thank Mr. J. E. E. Brassey and Mr. D. B. Gunner for technical vesicle, a nd dermal microfibril bundle counts in a ny of t he 3 assistance, Mrs. L. J. Barducc i for typing the manuscript, and our variants of DEB. clinical colleagues, especially Dr. R. S. Well s, fo r allowi ng us to study their patients. DISCUSSION REFERENCES The present study showed a total absence of AF in t he severe 1. Gedde-Dahl T: Epidermolysis bullosa: A clinical, ge netic and epi­ gene ralized fo rm of RDEB. This confirms previous reports demiological study. Baltimore, Johns Hopkins Press, 1971 based on observations us ing routine transmission electron mi­ 2. Bauer EA, Bri ggaman RA: The mechanobullous diseases (epider­ croscopy [10] a nd a monoclonal a nt ibody to AF [11], and molysis bullosa), Dermatology in General Medicine, 2d ed. Ed­ s u ggests t hat t he presence of AF in skin samples from cases ited by TB Fitzpatrick, AZ Eisen, K Wolff, IM Freedberg, KF causing diagnostic uncertainty excludes t he possibility of severe Austen. New Yo rk, McGraw-Hill, 1979, pp 334-347 3. Pearso n RW: Studies on the pathogenesis of epidermolysis bullosa. mutilating RDEB. The loss of AF is unlikely to be a result of J Invest Dermatol 39 :5~1 -5 75 , 1962 · unrecognized scarring because scar t issue from healt hy subjects 4. Heaphy MR, Winkelmann RK: The human cutaneous basement contains normal AF (Tidman and Eady, unpublished data). membrane-anchoring fibril complex: preparation and ultrastruc­ AF were present, but in reduced numbers, in all subjects with ture. J Invest Dermatol 68:177-178, 1977 DDEB a nd localized RDEB, irrespective of clinical severity, 5. Briggaman RA , Dalldorf FG, Wheeler CE: Formation and origin of basal lamina and anchoring fibrils in adul t human skin. J Cell a lt hough t hey were often very scanty, and examination of long Bioi 51:384- 395, 1971 length s of DEJ was at t imes necessary to find t hem. There was 6. Anton-Lamprecht I, Schnyder UW: Epidermolysis bullosa dystro­ no difference in AF counts between DDEB a nd locali zed RDEB, phica dominans. Ein Defekt der anchoring fibrils? Dermatolo­ and in both conditions the AF appeared morphologically nor­ gica 147:289- 298, 1973 mal, with no evidence of degradation. These results indicate 7. Hashimoto I, Anton-Lamprecht I, Gedde-Dahl T, Schnyder UW: Ultrastructural studies in epidermolysis bullosa hereditaria. I. that DDEB a nd t he localized form of RDEB cannot be distin­ Dominant dystrophic type of Pasini. Arch Dermatol Forsch guis hed by t heir ult rastructural features. 252:167-178, 1975 The distinction between t he Cockayne-Toura ine and Pasini 8. Hashimoto I, Gedde-Dahl T, Schnyder UW, Anton-Lamprecht I: variants of DDEB depends la rgely on t he presence, in t he Ultrastructural studies in epidermolysis bullosa hereditaria. II. latter, of a lbopapuloid lesions [1], the etiology a nd significance Dominant dystrophic type of Cockayne and Touraine. Arch Dermatol Res 255:285- 295, 1976 of which are not known. The clinical features of t hese 2 variants 9. Hashimoto I, Schnyder UW, Anton-Lamprecht I, Geddc-Dahl T, are otherwise simila r, despite the emphasis placed on t he hy­ Ward S: Ultrastructural studies in epidermolysis bullosa here­ perplastic or atrophic nature of the scar tissue [2]. The abnor­ ditaria. III. Recessive dystrophic types with dermolytic bli stering mality of glycosaminoglycan metabolism in the Pasini variant (Hallopeau-Siemens types and in verse type). Arch Dermatol Res [15], however, supports the existence of heterogeneity within 256:137-150, 1976 the DDEB group, although the present study did not show 10. Briggaman RA, Wheeler CE: Epidermolysis bullosa dystrophica­ recessive: a possible role of anchoring fibrils in the pathogenesis. ultrastructural differences between DDEB patients with a nd J Invest Dermatol 65:203- 211, 1975 w it hout albopapuloid lesions. 11. Goldsmith LA, Briggaman RA: Monoclonal antibodies to anchor­ In an earlier study of the Cockayne-Touraine variant of ing fibrils for the diagnosis of epidermolysis bullosa. J Invest DDEB [8] up to 10 times as many AF were found compa red Dermatol 81:464- 466, 1983 with the numbers obtained in t he present investigation. This 12. T idman MJ, Eady RAJ: Ultrastructural morphometry of normal human dermal-epidermal junction: the influence of age, sex, and discrepancy is probably due to t he use of different criteria for body region on laminar and nonlaminar co mponents. J In vest counting AF, and emphasizes t he difficulty in identifying AF Dermatol 83:448-453, 1984 unequivocally by t heir morphologic characteristics. 13. Pasini A: Dystrophie cutanee bulleuse atrophi ante et albo-papu­ Thus, although a ll forms of DEB are associated with a loide. Ann Dermatol Syphiligr {Paris) 9:1044- 1066, 1928 numerical abnormali ty of AF, whether t his is the primary defect 14. Eady RAJ, T idman MJ: Diagnosing epidermolysis bullosa. Br J Dermatol 108:621- 626, 1983 is uncertain. It remains to be determined whether t he increased 15. Bauer EA, Fiehler WK, Esterly NB: Increased glycosaminoglycan production of coll agenase by fibroblasts [16,17] is t he funda­ accumulation as a genetic characteristic in cell cultures of one mental abnorma li ty in RDEB. variety of dominant dystrophic epidermolysis bullosa. J Clin In conclusion, t he examination by routine transmission elec­ Invest 64:32- 39, 1979 tron microscopy of long lengths of DEJ, sectioned perpendicu­ 16. Bauer EA, Eisen AZ: Recessive dystrophic epidermolysis bullosa: evidence for increased collagenase as a genetic characteristic in lar to t he skin surface, is of limited diagnostic a nd t herefore cell culture. J Exp Med 148:1378-1387, 1978 prognostic value in dystrophic forms of epidermolysis bullosa. 17. Va lle K-J , Bauer EA: Enhanced biosynthesis of human skin col­ This technique cannot distinguish between the dominant a nd lagenase in fibroblast cultures from recessive dystrophic epider­ localized recessive forms of DEB. molysis bullosa. J Clin Invest 66:176- 187, 1980