Epidermolysis bullosa: Recent advances in understanding pathogenetic mechanisms

EPIDERMOLYSIS BULLOSA: RECENT ADVANCES IN UNDERSTANDING PATHOGENETIC MECHANISMS

ABSTRACT.-Epidermolysis bullosa (EB) is a group of mechanobullous diseases of the skin and mucous membranes characterized by the development of blisters and erosions follotig minor trauma. Hereditary (1) intraepidermal blistering in autoEB exists in four major patterns: (2) junctional blisterhxg in autosomal resomal domhnm t EB simplex; cessive junctional EB; (3) dermal blistering in autosomal dominant dystrophic EB; and (4) dermal blistering in autosomal recessive dystrophic EB. Electron microscopic e xamination of the skin is the current standard for diagnosis, although a number of immunologic reagents have been developed that are diagnostic tools and probes to pathogenesis. Particularly important are the antibody probes that have been developed for use in junctional and dystrophic forms of EB, since several of the monoclonal antiiy preparations also confer at least some information about the likely genetic pattern and prognosis. Antibody preparations have also been used to screen complementary DNA (cDNA) expression libraries for clones representing candidate genes in EB. In EB simplex, mutations in the keratin 14 gene have been discov14 combines with keratin 5 to form a heterodimer, ered. Since keratin the ezistence of these mutations results in poor keratin filament formation in vitro. In recessive junctional EB, antibody stahting patterns suggest involvement of the proteins of the anchoring filaments, such as BM-WO/niceht. A partial cDNA clone representing one of the subunits of BM-BOO/nicein has been developed, and this protein represents an isoform of laminin , suggesting the existence of a family of laminin-like molecules that are important for adhesion of epidermis in the basement membrane zone of the skin. In dominant dystrophic and recessive dystrophic EB, often there are decreased numbers and abnormal appearance of the anchoring: fibrils suggesting involvement of type VII collagen. Genetic linkage has been established to the type VII collagen gene in dominant dystrophic EB families. In recessive dystrophic EB, abnormalities in anchoring fibrils may be due to mutation(s) in the type VII collagen gene and/or due to excessive degradation by increased amounts of collagenase. Spontaneous, transgenic, and xenograft models of EB exist. Among the spontaneous models is recessive dystrophic EB in white alpine sheep, which show no staining for type VII collagen in the skin. A new xenograft model of recessive dystrophic EB has also been developed using EB skin grafted onto severe combined hnmunodeficient mice. A transgenic model has been developed for the herpetiform type of EB 104

Cm-r Probi

Dermatol,

May/June 1992

simplex resulting 14 gene.

Corn the insertion

of a truncated

form

of the keratin

IN BRIEF Epidermolysis bullosa (EB) is a group of mechanobullous diseases of the skin and mucous membranes characterized by the development of blisters and erosions following minor trauma. Hereditary EB can be conveniently EB and simply divided into four major patterns: (1) autosomal dominant simplex in which blister formation occurs primarily within the basal keratinocytes of the epidermis; (2) autosomal recessive junctional EB in which blister formation occurs in the lamina lucida of the epidermal-dermal junction, and the hemidesmosomes are often hypoplastic and reduced in number or absent; (3) autosomal dominant dystrophic EB in which blister formation occurs in the sublamina densa region of the dermis, and there are abnormal anchoring fibrils that may be reduced in number; and (4) autosomal recessive dystrophic EB in which blister formation is also in the sublamina densa region of the dermis, and the anchoring fibrils are quite abnormal in appearance and markedly reduced in number or absent. Although electron microscopic examination of the skin is the current “gold standard” for diagnosis, a number of new immunologic reagents have been developed, mainly with monoclonal antibodies, that are useful not only as diagnostic tools but also as probes to the pathogenesis of various types of EB. Particularly important are the antibody probes that have been developed for use in junctional and dystrophic forms*of EB. Among the antibody preparations that display reduced or absent staining in dystrophic EB are AFl, AFZ, KF-1, LH7:2, and L3d, all directed against the anchoring fibrils, the lamina densa, or type VII collagen. Absent staining with an antibody to chondroitin-6-sulfate proteoglycan has also been observed in recessive dystrophic EB. Two antibody preparations, GB3 and 19-DFJ-1, show reduced or absent staining in recessive junctional EB. Some of these antibody preparations have been important for their use in screening cDNA expression libraries for clones representing some of the candidate genes in EB. The candidate gene approach has been useful in developing molecular biologic tools for exploring the pathogenesis of some genetic varieties of each major group of EB. In one form of dominant EB simplex, the herpetiform type, blister formation is associated with clumping of the tonofilaments of the basal keratinocytes. In this type of EB simplex, mutations in the keratin 14 gene have been discovered. Since keratin 14 combines with keratin 5 to form a heterodimer, the existence of these mutations results in poor keratin filament formation in vitro. In recessive junctional EB, antibody staining patterns suggest involvement of the proteins of the anchoring filaments, such as BM-GOO/nicein. A partial cDNA clone representing one of the subunits of BM-GOO/nicein has been developed, and this protein represents an isoform of laminin. Laminin itself is important for adhesion of epithelia to subjacent stroma. Thus it is now likely that there is a family of laminin-like molecules that are important for adhesion of epidermis in the basement membrane zone of the skin. In dominant dystrophic EB, the decreased numbers and abnormal appearance of the anchoring fibrils suggest involvement of type VII collagen, the major structural protein of the anchoring fib&,, in the pathogenesis. Development of a partial cDNA clone to type VII collagen has established genetic linkage to the type VII collagen gene in four dominant dystrophic EB families. Similarly, in recessive dystrophic EB, there are profound abnormalities in morphology and number of anchoring fibrils. To explore if these abnormalities are due to mutation(s) in the type VII collagen Cur-r Probl Dermatol, May/June 1992

106

gene and/or due to excessive degradation by increased amounts of collagenase, cDNA clones to collagenase and to type VII collagen are now being used to determine if there is linkage to either major candidate gene. Several animal models of EB exist: spontaneous, transgenic, and xenograft. There are spontaneous animal models for EB simplex, junctional EB, and dystrophic EB. Among the best characterized of the spontaneous models is recessive dystrophic EB in the white alpine sheep, which shows no staining of type VII collagen in the skin and thus mimics the human disease immunologically. A transgenic model has been developed for the herpetiform type of EB simplex. Animals that show intraepidermal blistering at birth and associated clumping of the tonofilaments are the result of the insertion of a truncated form of the keratin 14 gene. This finding recapitulates the natural human form of herpetiform EB simplex and will be important for exploring possible means to correct the functional defect in vivo. There is also a new xenograft model of recessive dystrophic EB. EB skin grafted onto severe combined immunodeficient mice maintains complete fidelity to the human disease both morphologically and immunologically. This model should permit exploration of both pathogenetic mechanisms and various potential therapeutic modalities.

Curr

Probl

Dermatol,

May/June 1992

Eugene A. Bauer, M.D., is a dermatologist and investigator who has worked in connective tissue biology for over two decades. Dr. Bauer attended medical school at Northwestern University, graduating in 1967. He served a straight medicine internship, dermatology residency, and fellowship at Washington University and he remained on the faculty there from 1974 through 1963, progressing to the rank of professor. In 1968 he became professor and chairman of the Department of Dermatology at Stanford University School of Medicine. Dr. Bauer’s work, which is supported by the National Institute of Arthritis and Musculoskeletal and Skin Diseases, is centered on the biochemistry and molecular biology of human skin collagenase. His first work on epidermolysis bullosa revealed that skin from recessive dystrophic epidermolysis bullosa patients had elevated levels of collagenase that were also expressed by fibroblasts in vitro. His current work also involves molecular cloning and investigation of structural proteins of the cutaneous basement membrane zone and of the proteases and inhibitors that regulate metabolism of the structural proteins of the skin. Dr. Bauer has received the Montagna Award from the Society of Investigative DermatoloSy and the Sulzberger Award from the American Academy of Dermatology.

Youn H. Kim, MD., is a dermatologist and investigator whose early work focused on the use of magnetic resonance imaging in dermatology with special reference to mechanisms of carcinogenesis. She was graduated from Stanford Vniversity School of Medicine in 1964 after which she was research fellow, resident, and chief resident in dermatology at Stanford. Recently she has developed an animal model of recessive dystrophic epidermolysis bullosa in which skin is placed as a xenograft onto severe combined immunodejicient mice. The blistering phenotype and human antigens are maintained in this model. This model has also served as the basis for the transplantation of the three-dimensional in vitro models of epidermolysis bullosa onto living animals. Dr. Rim’s research is supported by the National Institute of Arthritis and Musculoskeletal and Skin Diseases.

David T. Woodley, M.D., is a dermatologist and investigator who graduated from the University of Missouri Medical School in 1973. Following an internship and residency in internal medicine, Dr. Wood@ was a resident in dermatology at the University of North Carolina at Chapel Hill. He served a two-year research position at the National Institutes of Health, followed by an additional year at the University of Paris. In 1962 he returned to the University of North Carolina as assistant professor and was promoted to associate professor in 2965. He joined the faculty of the Department of Dermatology at Stanford in 1969 as professor and associate chairman. Dr. Woodley has spent most of his research career working on the proteins of the basement membrane zone of the skin and is supported by the National Institute of Arthritis and Musculoskeletal and Skin Diseases. He initially described the autoantigen in acquired epidermolysis bulloss and the existence of circulating autoantibodies in that disease. His work on acquired epidermolysis bullosa includes identification of the antigen as a portion of the type VlZ collagen molecule. Cum Probl

Dematol,

May/June 1992

107

Jouni Uitto, M.D., Ph.D., is a dermatologist and biochemist-molecular biologist who has been an investigator in connective tissue biology for over two decades. Dr. Uitto received both the M.D. and PhD. degrees from the University of Helsinki in 1979. After serving as assistant professor of biochemistry at Rutgers Medical School from 1972-1975, he entered a residency in dermatology at and later joined the faculty of that Washington University School of Medicine institution. Dr. Uitto was named professor of medicine (dermatology) at HarborUCLA Medical Center in 1983 and in 1986 became professor of dermatology, biochemistry, and molecular biology and chairman of the Department of Dermatology of Jeflerson Medical College. Dr. Uittos current work focuses on the molecular and cell biology of proteins of the e,xtracellular matri,x of the skin, particularly those of the cutaneous basement membrane zone that are involved in the pathogenesis of various forms of epidermolysis bullosa. He is the recipient of research funding from the National Institute of Arthritis and Musculoskeletal and Skin Diseases. Dr. Uitto has been honored by many universities and societies, including receiving the Montagna Award from the Society of Investigative Dermatology.

Patrick Verrando, Ph.D., is a biochemist and cell biologist who received his Ph.D. degree in cellular and molecular pharmacology from the University of Nice, France, in 1963. Following a postdoctoral fellowship at the International Center of Dermatological Research, Sophia Antipolis, Dr. Verrando joined the Department of Dermatology of the University of Nice in the Laboratory of Dermatologic Research. Since 1966, he has pursued research on a major component of the cutaneous membrane zone, BM-GtWnicein, a protein involved in the pathogenesis of recessive junctional epidermolysis bullosa. In 1990, Dr. Verrando was a French National Institute of Medical Research eFchange scientist in the Department of Dermatology at Stanford University School of Medicine.

Jean-Paul Ortonne, M.D., is a dermatologist and investigator who received his medical degree f-am the University of Lyon, France, after which he served both residency and research fellowships in the Department of Dermatology under the direction of Professor Jean Thivolet. He further pursued research training in the Department of Dermatology at Harvard Medical School-Massachusetts General Hospital where he investigated pigment cell biology with Dr. Thomas Fitzpatrick. In 1981 Dr. Ortonne became professor and chairman of the Department of Dermatology and director of the Dermatologic Research Laboratory at the University of Nice-Sophia Antipolis Medical School in Nice. In addition to pursuing research in pigment cell biology, Dr. Ortonne’s group has made seminal contributions to the understanding of pathophysiologic mechanisms in recessive junctional epidermolysis bullosa. 108

Curr

Probl

Dermatol, May/June 1992

EPIDERMOLYSIS BULLOSA: RECENT ADVANCES IN UNDERSTANDING PATHOGENETIC MECHANISMS

Epidermolysis bullosa (EB) is the term applied to a group of disorders whose common primary feature is the formation of cutaneous and mucosal blisters and erosions following trivial trauma. The inheritance patterns in EB are autosomal, both recessive and dominant. The major varieties are presented in Table 1, which is a highly simplified classification of EB based on the site of the blister cleavage. This classification tends to deemphasize the clinical- and genetic aspects of any given patient and depends on the current “gold standard” for diagnosis, electron microscopy. As will be apparent from the following discussion, however, the development of new antibody and DNA probes has begun to revolutionize concepts about the pathogenesis of various forms of EB. Table 1 also presents a somewhat more detailed characterization of the pathomorphologic patterns using criteria for types of EB as recently reported

in the United States by the National Epidermolysis Bullosa Registry.’ A crucial feature in understanding the molecular basis for each type of EB is the clinical and genetic heterogeneity that exists within the quite limited number of cleavage planes involved in blister formation as observed histologically. This clinical heterogeneity is important especially for genetic linkage studies. For example, in the case of dominant EB simplex, it is likely that the severity of disease will be correlated with the type of mutations that are ultimately defined. In a second example-recessive dystrophic EB-it is possible that mutations in the collagenous domains of the type VII collagen gene could lead to altered synthesis, secretion, deposition, and/or degradation of this protein with resultant aberrations in the anchoring fibrils in vivo. It is possible to begin to address the molecular

TABLE 1. Classification

of EB

Disease Group

Inheritance

EB simplex Junctional EB

Autosomal Autosomal

Dystrophic

Blister Cleavage dominant recessive

EB

Autosomal dominant, autosomal recessive EB = epidermolysis bullosa. Cur-r Probl Dermatol,

May/June 1992

Intraepidermal with clumped tonofilaments in EB herpetiformis Intralamina lucida with hypoplastic, reduced, or absent hemidesmosomes and absent anchoring filaments in the “lethal” form Sublamina densa with reduced anchoring fibrils in dominant dystrophic EB and reduced or absent anchoring fibrils in recessive dystrophic EB

109

TABLE 2. Immunodiagnostic Antibody

Probes in EB*

Preparation

AFl, AF2 KF-1 LH72 L3d CSPG GB3 19-DES1 EB = epidermolysis bullosa; *Adapted from Fine et al.’ tStaining patterns are those

Antigen

or Epitope

Commentt

Anchoring fibrils Absent in RDEB Lamina densa Reduced in DDEB, absent in RDEB Type VII collagen Reduced or absent in RDEB Type VII collagen Reduced in RDEB Chondmitin-B-sulfate pmteoglycan Absent in RDEB Hemidesmosomes BM-GOO/nicein Usually absent in lethal RJEB, reduced in benign RJEB Anchoring filament Absent in RJEB RDEB = recessive dystmphic EB: DDEB = dominant dystmphic EB; BJEB = recessive junctional EB. most commonly seen.

pathology of each major type of EB largely because of two research appmaches: the development of specific antibodies to various components of the cutaneous basement membrane zone (BMZ) and the implementation of the so-called candidate gene approach. Table 2 presents some of the most important immunologic probes that have been developed and employed in the diagnosis and investigation of EB. The staining patterns of these antibody pmbes are correlated with specific types of EB. Although the current diagnostic standard is ultrastructural analysis, the use of specific antibody preparations has become an essential adjunct to accurate diagnosis and assignment of prognosis. For example, GB3 antibody defines the BM-GOO/nicein protein complex in the epidermal-dermal junction and localizes it to the lower portion of the lamina lucida with immunoelectmn micmscopy. Staining with the antibody is markedly diminished or absent in the so-called lethal form of recessive junctional EB.’ In another example, the LH72 monoclonal antibody defines anchoring fibrils and also stains the epidermal-den& junction histochemically. It resides at the lower partion of the lamina densa and on the ends of the

anchoring fibrils with immunoelectmn micmscopy. Staining with this monoclonal antibody is markedly diminished or absent in recessive dystmphic EB. The candidate gene approach is based on the notion that proteins or antigens that are either in physical proximity to the pathology or, preferably, display abnormal features (e.g., absent staining of the epidermal-dermal junction with antibodies to type VII collagen in recessive dystmphic EB) represent candidate proteins of importance in the pathogenesis of a given type of EB. Table 3 depicts several candidate genes that are currently being examined for abnormal structure, expression, and/or linkage patterns in various forms of EB. Although the BMZ of the skin consists of a large number of distinct structural macromolecules that form a functional matrix at the epidermal-dermal interface, it is obvious that the presence of sufficient quantities of these macromolecules and of their interactions depends not only on adequate synthesis of a structurally normal protein but also on intact cellular machinery for posttranslational modification, and on appropriately expressed and regulated mechanisms of degradation. The re-

TABLE 3. Candidate Type of EB

Genes in Various Forms of EB* Candidate

Genes

Keratins and other cytoskeletal proteins, linking proteins Cell-cell adhesion molecules, beta-l integrins Hemidesmosomal proteins, alpha-6/bets-4 integrin Junctional Anchoring filament proteins (nicein, kalinin, epiligrin) Anchoring fibril proteins (type VII collagen) Dystmphic Enzymes that degrade BMZ structural proteins (type I collagenase, stmmelysin, type IV collagenase) EB = epidermolysis bullosa; BMZ = basement membrane zone. *Adapted from Uitto et al? Simplex

110

Curr

Probl

Dermatol,

May/June 1992

mainder of this review will examine some of the newer concepts emerging from the use of molecular and cell biologic tools to address the problems posed by EB.

DOMINANT EPIDEBMOLYSIS SIMPLEX (DEBS)

BULLOSA

DEBS is defined by the occurrence of intraepidermal blister formation. Several varieties of DEBS exist, including generalized DEBS (Koebner), localized DEBS (Weber-Cockaynel, (Fig 1) DEBS of the Ogna variety described in Norwegian kindreds, and the herpetiform variety of DEBS (DowlingMeara). Common to all these varieties of DEBS are serous or serosanguinous, nonscarring blisters, often being localized to the extremities but being more widespread especially in the generalized (Koebnerl and herpetiform (Dowling-Meara) varieties5 (Figs 2 and 3). Inheritance in each variety is autosomal dominant. Sporadic cases of EB simplex are encountered rather frequently and probably result from gene mutations. The pathogenic mechanisms involved in blister formation are currently unknown, although warming the skin prior to producing blisters experimen-

tally enhances their induction. Conversely, cooling the skin makes blister induction more diRicult.6 It has been suggested that activation of cytolytic enzymes’ or the existence of a mutant, temperaturesensitive structural protein may ultimately be found in this group of disorders. Recent genetic linkage analysis provides support for such a hypothesis. Among the most important observations in DEBS are the recent descriptions of mutations in certain keratin genes. Three groups, using molecular biologic approaches, have demonstrated that the primary defects within certain kindreds are associated with mutations in the keratin genes.‘-” Specifically, the disease in some of the kindreds has been linked to chromosome 12q adjacent to the keratin II gene cluster, while other kindreds demonstrate linkage to chromosome 17q near the keratin I gene cluster. Initial clues to the possible existence of mutations in one or more keratin genes derived from the early observation that clumping of the tonofilaments occurred in association with early blister formation in the herpetiform variev of DEBS. To approach this problem, Fuchs et al created a series of transgenic mice into which had been integrated a truncated portion of the keratin 14 gene.

FIG 1. EB simplex. Blister over the top of the great toe. No evidence Curr

Probl

Dermatol,

May/June 1992

of scarring from previous blisters 111

FIG 2. E B simplex (Weber-Cockayne equipment.

type). Erosions over the palms of a child after climbing on playground

The resulting animals and their offspring developed a clinical disease characterized by severe intraepidermal blistering at birth, and ultrastructural evidence for clumping of the tonofilaments in association with blister formation. The severity of the

clinical disease appeared to correlate with the degree of disruption of filament formation in vitro. By determining the specific mutations in two EB herpetiformis patients, this same group has been able to elucidate the resulting defect in keratin filament

FIG 3. E B simplex (Dowling-Meara 112

type). Groups of small blisters and crusted blisters on the leg Curr

Probl

Dermatol,

May/June 1992

Although the precise pathogenesis of RJEB has yet to be defined, of particular interest is BM-6001 nicein, which was initially recognized by monoclonal antibody GB3.l’ BM-GOO/nicein consists of three disulfide-linked subunit polypeptides that constitute the -600~kDa protein. This protein has been shown by immunoelectron microscopy to localize to anchoring filaments that traverse the lamina lucida and perhaps connect the hemidesmosomal structures to the lamina densa. Two other proteins, kalinin and epiligrin, with similar biochemical features have been described.‘3, l4 It is likely that these three proteins are identical or are closely related members of a gene family. BM-GOO/nicein is the principal candidate gene for BJEB, since immunostaining of skin from paBBCESSIVB JUNCTIONAL EPIDEBMOLYSIS tients with this form of EB shows attenuation or BULLOSA WEB) absence of immunoreactivity (Table 2). It has also Blister formation in RJEB occurs within the lam- been demonstrated that incubation of epidermal ina lucida of the BMZ. There are two main vari- keratinocytes with antibodies to kalinin, perhaps eties: “lethal” BJEB, and a benign type often called an identical protein, results in dissociation of the cells from substrata, suggesting that kalinin serves generalized atrophic benign EB’ (Fig 4). Current research efforts have focused on the more severe as an attachment protein for these cells. In this re(“lethal”) form of the disease, which is character- gard, it is important that a partial (1404-bp) cDNA ized by the onset at birth of severe generalized clone has been developed for the 100~kDa subunit blistering that usually heals without scarring. of BM-600/nicein.15 The clone represents a unique Mild atrophy may result. The nails often show par- protein, a portion of which bears approximately 70% homology to the B2 chain of laminin. In rotary onychia, and chronic plaques of nonhealing granulation tissue are found in the perioral region. shadowing analysis, the protein is an asymmetric The teeth are dysplastic because of an enamel de- 90-nm rod with a single globule at one end and one or two globules at the other. The gene is physfect.

formation. In each patient a point mutation was found in codon 125, normally an arginine, but having a cysteine or histidine in the patients.’ These mutations result in poor keratin filament formation. Because keratin 14 (encoded by a gene that resides on chromosome 17) combines with keratin 5 (a gene on chromosome 12) to form a heterodimeric filament, one would predict that a mutation in keratin 5 would be found as another cause of EBS. Importantly, two groups have now demonstrated strong genetic linkage in different kindreds to chromosome 12q, the physical location of the keratin 5 gene.l’,l’

FIG 4. Junctional Curr

Pmbl

Dennatol,

EB, Tense, nonhemorrhagic

May/June 1992

blister on the scalp 113

ically located on chromosome lq. It seems likely that BM-GOO/nicein is an isoform of laminin and thus functionally may participate in adhesion of basal keratinocytes to the subjacent macromolecular matrix. Possible mechanisms that may be involved in cellular adhesions include a role for the integrins, a family of cell surface receptors that mediate cell-cell and cell-matrix interactions. The cy6B4 integrin, which is expressed by epithelial cells, has been shown immunohistochemically to be present in the cutaneous BMZ. Immunoelectron microscopy localized this integrin to hemidesmosomal complexes.16, I7 It is possible that or6j34 integrins mediate the attachment of hemidesmosomes to the anchoring filament proteins, although immunohistochemical analysis of patients with EB has thus far revealed normal staining for integrin epitopes.

CLINICAL FEATURES

Clinically, the findings of scarring, contractures, and milia formation are frequently useful in categorizing the patient with sporadic EB as having a dystrophic form of the disease. DDEB almost always remains confined to the hands, feet, elbows, and knees (Fig 51. The clinical severity in patients with localized RDEB makes them indistinguishable from those having DDEB. The more severe variety of RDEB begins at birth with cutaneous and mucosal involvement. There is severe generalized blistering, resulting in nail dystrophy; repeated cycles of blistering, scarring and mittenlike encasements of the digits; and flexural contractures of the major joints (Fig 6). Mucosal involvement of the mouth, pharynx, esophagus, and anus is common (Fig 7). Esophageal erosions cause strictures. Conjunctival and corneal erosions may be seen. DOMINANT

DYSTHOPHIC EPIDEHMOLYSIS BULLOSA-DOMINANT AND RECESSIVE (DDEB, BDEB)

Dystrophic EB is defined by the existence of blistering beneath the lamina densa of the cutaneous BMZ. There are two varieties of DDEB and one major, but certainly genetically heterogeneous, form of RDEB.

DYSTROPHIC EB

As was the case with RJEB, the exploration of the structural features of the cutaneous BMZ has been facilitated by molecular cloning. Again, the candidate gene approach has been important. For DDEB, in addition to the sublamina densa blistering, there are also decreased numbers of rudimentary anchoring fibrils that are found only in blisterprone areas in the milder form of DDEB (hyperplastic) but appear to be abnormal and/or absent

FIG 5. Dominant dystrophic 114

EB. Hemorrhagic

blister on back of hand and loss of fingernails

in a child

Curr Probl Dermatol,

May/June

FIG 6. Recessive dystrophic EB. Hemorrhagic

blisters and scarring on a child’s hands.

FIG 7. Recessive dystrophic fant.

EB. Severe erosions of the mouth of an in-

Curr Probl Dermatol, May/June 1992

at all sites in the more severe albopapuloid form of DDEB.’ The observation that anchoring fibrils are composed of type VII collagen1801ghas permitted a more detailed exploration of their possible role in various forms of dystrophic EB. By combining biochemical and molecular biologic data from the recent cloning of a portion of the type VII collagen gene,” the characteristics of type VII collagen have emerged. Type VII collagen consists of a large, -145~kDa collagenous domain, which is interrupted by a noncollagenous segment in the m iddle of the triple-helical portion. The collagenous domain is flanked by two globular noncollagenous segments. The larger one (--145~kDa1was thought to be at the carboxy-terminal end of the protein but has now clearly been shown to be at the amino-terminal end by cDNA sequence analysis and by comparison with peptide sequenceszo The large amino-terminal domains of type VII collagen associate with other macromolecules within the BMZ at the epidermal-dermal junction at one end and with basement membrane-like structures, termed anchoring plaques, at the other end. This association presumably secures the attachment of the basement membrane to the subjacent derm is . In the case of DDEB, strong genetic linkage to the type VII collagen gene locus has been demonstrated.‘l’ ” A polymorphic marker in the type VII collagen gene, together with other restriction fragment-length polymorphisms on chromosome 115

Sp, allowed linkage to be established to type VII collagen in four families with DDEB (LOD scores

and possible heterogeneity within BDEB must be elucidated by further linkage analysis.

2.1-8.8).

RECESSIVE

DYSTROPHZC

EB

The pathogenesis of BDEB has not been defined, although consistent ultrastructural features include a profound decrease or absence of anchoring fibrils and collagenolysis. Two possible mechanisms for blister formation have been proposed: (1) a defective structural protein in the dermis that is responsible for the integrity of the epidermaldermal junction (e.g., type VII collagen of the anchoring fibrils) and (21 destruction of dermal connective tissue by excessive proteolytic action (e.g., collagenase) . Several possible mechanisms might be envisioned as a role for type VII collagen, a major candidate gene for BDEB. Mutations in the triplehelical portion of type VII collagen could prevent or delay the secretion of this collagen from basal keratinocytes. Indeed, recent immunolocalization studies show intracellular accumulation of type VII collagen epitopes within basal keratinocytes in some patients with BDEB, suggesting a secretory defect .23824 A structurally defective form of type VII collagen might also be preferentially, or more rapidly, degraded by proteases. Furthermore, mutations in the noncollagenous portion of type VII collagen, which has putative interactions with other components of the cutaneous BMZ, might result in impairment of anchoring fibril functions. A substantial body of investigation from a number of groups also suggests that dysregulation of collagenase synthesis and/or activity is important in the pathogenesis of blistering BDEB. In vivo quantitation of immunoreactive collagenase has revealed increased levels of the enzyme.25 Skin fibroblasts derived from patients with BDEB demonstrate enhanced synthesis of collagenase in cell culture,26 although this is present in cells from only approximately one fourth of the patients examined.z7 Since it has been shown that anchoring fibrils are comprised of type VII collagen,““’ it has been possible to determine if this collagen can serve as a substrate for collagenase. Seltzer et al” have shown cleavage of type VII collagen by both interstitial and type IV collagenases, thus lending some credibility to the degradation hypothesis. Despite these functional and biochemical data, it has been shown in one family that collagenase could be excluded as the candidate gene on the basis of linkage analysis.” However, this approach was not informative in several additional kindreds, 116

ANIMAL MODELS EPIDERMOLYSIS

OF BULLOSA

Three major types of animal model systems have been described for EB: spontaneous, transgenic, and xenograft models. There are unique advantages and disadvantages to each type of model system. However different, these animal models can serve as important experimental in vivo counterparts for the human diseases. SPONTANEOUS

MODELS

Skin lesions resembling those of the human diseases have been described in dogs, bulls, foals, calves, and sheep. Comprehensive biochemical, immunochemical, and ultrastructural characterization is lacking in most of these animal models of EB. A cutaneous disease similar to EBS has been reported in collie dogs?’ Skin lesions consist of patchy areas of erosions, ulcers, crusts, hypopigmentation, and alopecia. They occur most frequently over exposed bony prominences or joints of the limbs subjected to repetitive trauma. The oral cavity and nails are not affected. Scarring is uncommon, and when present, is composed of areas of hyperpigmentation or hypopigmentation with cutaneous atrophy and parchment-like skin. In most animals, the skin lesions are present before 6 months of age. As in humans, the skin lesions worsen during warmer weather. Although all dogs experience periodic exacerbations and remissions, their general health remains excellent. Histopathologically, the bullae form as a result of cytolysis of basal keratinocytes. Belgian foals that have the clinical and histologic features of junctional EB in humans have been described.31-33 The foals show multiple erosions on glabrous skin and oral, anal, and vaginal mucocutaneous junctions. The erosions and ulcerations result in sloughing of the hooves. All animals have oral lesions and dystrophic teeth, and corneal involvement occurs. Scarring is not a prominent feature. Skin lesions are present either at birth or within the first 2 weeks of age. Histologically, tissue sections from both foals reveal an epidermaldermal separation without basal cell cytolysis. Material of the BMZ, which is positive to periodic acid-Schitf (PAS) staining, can be found on the floor of the blisters. Electron microscopic examination reveals that the separation is in the lamina luCur-r Probl

Dermatol,

May/June 1992

.

cida. Hemidesmosomes appeared to have ruptured in the process of cleft formation. A toy poodle with skin lesions and clinical course resembling the lethal form of junctional EB has also been reported.34 Shortly after birth, the dog developed multiple vesicles and bullae involving the glabrous skin and the mucous membranes. Abnormalities of the teeth or nails were not observed. Microscopically, there was epidermaldermal cleavage separation, and the cleft was localized in the lamina lucida. The hemidesmosomes appeared rudimentary when compared with those of a normal dog. The anchoring fibrils appeared normal. The very first report of animal disease resembling dystrophic EB in humans was in New examiZealand sheep,35 although ultrastructural nation was not performed. Subsequently, a mechanobullous disease clinically similar to -that of the New Zealand sheep was observed in Brangus calves.36 The calves have extensive skin ulcerations on the distal limbs and over pressure points with sloughing of all hooves. Extensive mucous membrane involvement is also present. Histologically, there is an epidermal-dermal separation with the PAS-positive material remaining in the roof of the blister. Ultrastructurally, the cleft is located below the lamina densa, and there is loss of anchoring fibrils. The suggestion is that the Brangus calves clinically resemble the most severe (i.e., the Hallopeau-Siemens) subtype of RDEB. The best characterized spontaneous animal model of EB is the Weisses Alpenschaf sheep, which have clinical manifestations similar to the severe, mutilating subtype (Hallopeau-Siemens) of RDEB.37 The skin lesions appear within a few days after birth and consist of flaccid, serous, and hemorrhagic blisters. Oral mucous membranes and the esophagus are involved, leading to difficulty in feeding and poor growth. Sloughing of the hooves occurs 1 to 3 weeks after birth. Histologically, there is an epidermal-dermal separation with minimal inflammation. Ultrastructural analysis demonstrates that the site of the separation is in the sublamina densa zone. Anchoring fibrils are either rudimentary or absent when compared with those of healthy skin. Immunofluorescent staining for type VII collagen with antibodies to both the triple helical domain and the C-terminus of the type VII collagen molecule shows complete absence of staining. Similarly, immunoblot analysis of skin extracts for type VII collagen reveals no immunoreactive type VII collagen. The pedigree of these animals is consistent with an autosomal recessive inheritance pattern. The advantages of these spontaneous animal Curr Probl Dermatol, May/June 1992

models of EB are (1) availability of tissues that cannot be easily biopsied in humans, such as mucous membranes and extracutaneous sites, and (2) larger amounts of skin that can be obtained for various analyses and tissue culture. However, there are major disadvantages for utilizing the spontaneous models described. First, the spontaneous occurrences are few in number and access to the animals is difficult. Secondly, the affected animals are often too large in size or their fecundity too low for cost-effective breeding for research purposes. Finally, although the diseases resemble those in humans, the exact biochemical and molecular relevance to the human disease must be proven in each case.

TRANSGENIC MODELS The technology for producing useful transgenic animal models has been available since the early 1980s; however, there has been a revolution in this field over the last five years?8-40 With the development of more efficient germ-line transformation techniques and expansion of tissue-specific regulators, there has been a great advance in understanding the mechanism of various human diseases using transgenic models. Although producing transgenic models is costly, the potential attainment of significant knowledge and the ability to address specific questions at the molecular and genetic level are driving forces in the increased utility of this technology. The most important implementation of this technology for EB research has been in the creation of transgenic animals containing a truncated form of the keratin 14 gene.’ The introduction of this human gene to create transgenic mice produced animals with a disease clinically, histologically, and ultrastructurally faithful to the herpetiform type of EB simplex (see above)

XENOGRAFT MODELS An alternative to spontaneous and transgenic models is the use of xenografted animal models of EB. A xenograft model of EB has been established using full-thickness skin tissues from the severe mutilating subtype of RDEB by grafting onto severe combined immunodeficient @CID) mice.41 Immunofluorescence studies were performed up to 24 weeks after transplantation using species-specific antibodies to human class I antigen, mouse class I antigen, and human type IV and VII collagens and with cross-reacting antibodies against laminin and bullous pemphigoid antigen. Staining with the an117

tibody to human class I antigen and with the antibody to mouse class I antigen confirmed the species-specific results obtained with the type IV and type VII collagen antibodies. The RDEB grafts showed essentially no staining with the type VII collagen antibody. Antibodies against laminin and bullous pemphigoid antigen showed normal staining patterns in the RDEB grafts. The model was also ultrastructurally faithful to the human disease. There was an overall paucity of anchoring fibrils in the grafts. Blisters could be induced in these grafts with minor trauma and showed a sublamina densa separation by immunomapping and electron microscopy. As late as 24 weeks after transplantation, the RDEB grafts remained human, were not significantly replaced by mouse cells, and retained the RDEB disease phenotype. The advantage of a xenograft model system is that the tissue investigated is actually of human origin and thus may most closely represent human disease, provided that the grafted human EB skin remains human and retains the EB phenotype. A xenograft model is also far less costly in time and money and easily available. Because only a small part of the animal is involved with human disease, the animals remain healthy for longer duration than spontaneous or transgenic models, and thus long-term experiments can be performed. The difficulties are the small sample sizes that are available for analysis and tissue culture and the concern for the potential gradual loss of the EB phenotype after long periods of time. Thus, it will be necessary to establish that there is longterm fidelity to the human disease characteristics and phenotype before this model can be widely exploited. ACKNOWLEDGMENTS

Supported in part by the following grants from the National Institutes of Health: AR19537, AR41551, AR33625, KllAR01850, POlAR41045, POlAR38923, AR62273, T32AR07411, and RR00070. REFERENCES 1. Fine J-D, Bauer FA, Briggaman RA, et al: Revised clinical and laboratory criteria for subtypes in inherited epidermolysis bullosa: A consensus report by the subcommittee on Diagnosis and Classification of the National Epidermolysis Bullosa Registry. J Am Acad Dermatol 1991; 24:119-135. 2. Verrando P, Blanchet-Bardon C, Pisani A, et al: Monoclonal antibody GB3 defines a widespread defect of several basement membranes and a keratinocyte dysfunction in patients with lethal junctional epidermolysis bullosa. Lab Invest 1991; 64:85-92. 118

3. Heagerty AHM, Kennedy AR, Leigh IM, et al: Identification of an epidermal basement membrane defect in recessive forms of dystrophic epidermolysis bullosa by LH7.2 monoclonal antibody: Use in diagnosis. Br J Dermatof 1986; 15X5 - 131. 4. Uitto J, Bauer EA, Moshell AN: Symposium on epidermolysis bullosa: Molecular biology and pathology of the cutaneous basement membrane zone. J Invest Dermatol 1992; 98:391-400. 5. Bauer EA, Briggaman RA: Hereditary epidermolysis bullosa, in Fitzpatrick TB, Eisen AZ, Wolff K, Freedberg IM, Austen KF teds): Dermatology in General Medicine, 4th ed. New York, McGraw-Hill. In press. 6. Pearson RW: Studies on the pathogenesis of epidermolysis bullosa. J Invest Dermatol 1962; 39:551-575. I: Epidermolysis bul7 Gedde-Dahl T, Anton-Lamprecht losa, in Emory AEH, Rimoin DL (eds): Principles and Practice in Medical Genetics. New York, Livingstone, 1981. 8. Coulombe PA, Hutton ME, Letai A, et al: Point mutations in human keratin 14 genes of epidermolysis bullosa simplex patients: Genetic and functional analyses. Cell 1991; 66:1301-1311. 9. Vassar R, Coulombe PA, Degenstein L, et al: Mutant keratin expression in transgenic mice causes marked abnormalities resembling a human genetic skin disease. Cell 1991; 64:365-380. 10. Bonifas JM, Rothman A-L, Epstein E: Epidermolysis bullosa simplex: Evidence in two families for keratin gene abnormalities. Science 1991; 254:1202-1205. 11. Ryynanen M, Knowlton RG, Uitto J: Mapping of epidermolysis bullosa simplex mutation to chromosome 12. Am J Hum Genet 1991; 459:978-983. 12. Verrando P, Pisani A, Ortonne J-P: The new basement membrane antigen recognized by the monoclonal antibody GB3 is a large size weight glycoprotein: Modulation of its expression by retinoic acid. Biochim Biophys Acta 1988; 942:45-56. 13. Rouselle P, Lunstrum GP, Keene DR, et al: Kalinin: An epithelium-specific basement membrane adhesion molecule that is a component of anchoring filaments. J Ceff Biol 1991; 114:567-576. 14. Carter WG, Ryan MC, Gahr PJ: Epiligrin, a new cell adhesion ligand for integrin cc3pl in epithelial basement membranes. Cell 1991; 65:559-610. 15. Verrando P, Vailly J, Baudoin C, et al: Partial cDNA sequence of the new basement membrane protein nicein (BM-600) reveals discrete homologies with human laminin. Submitted. 16. Stepp MA, Spurr-Michaud S, Tisdale A, et al: a6P4 integrin heterodimer is a component of hemidesmosomes. Proc Nat1 Acad Sci USA 1990; 87:8970-8974. 17. Ryynanen J, Jaakkola S, Engvall E, et al: Expression of B4 integrins in human skin: Comparison of epidermal distribution with Bl integrin epitopes, and modulation by calcium and vitamin D3 in cultured keratinocytes. J Invest Dermatol 1991; 97:562-567. 18. Sakai LY, Keene DR, Morris NP, et al: Type VII collagen is a major structural component of anchoring fibrils. J Cell Biol 1986; 103:1577-1586. 19. Keene DR, Sakai LY, Lunstrum GP, et al: me VII collagen forms an extended network of anchoring fibrils. J Cell Biol 1987; 104:611-621. 20. Parente MG, Chung LC, Ryynanen J, et al: Human type Curr Probl Dermatol, May/June 1992

21.

22.

23. 24.

25. 26.

27.

28.

29.

VII collagen: cDNA cloning and chromosomal mapping of the gene. Proc Nat1 Acad Sci USA 1991; 88:6931-6935. Ryynanen M, Knowlton RG, Parente MG, et al: Human type VII collagen: Genetic linkage of the gene (COL7Al) on chromosome 3 to dominant dystrophic epidermolysis bullosa. &I J Hum Genet 1991; 49:797-803. Ryynanen M, Ryynanen J, Sollberg S, et al: Genetic linkage of type VII collagen (COL7Al) to dominant dystrophic epidermolysis bullosa in families with abnormal anchoring fibrils. J Clin Invest, in press. Smith LT, Sybert VP: Irma-epidermal retention of type VII collagen in a patient with recessive dystrophic epidermolysis bullosa. J Invest Dermatol 1990; 94:261-264. Fine J-D, Horiguichi Y, Stein DH, et al: Intraepidermal type VII collagen: Evidence for abnormal intracytoplasmic processing of a major basement membrane protein in rare patients with dominant and possible localized recessive forms of dystrophic epidermolysis bullosa. J Am Acad Dermatol 1990; 22:188-195. Bauer EA, Gedde-Dahl T, Eisen AZ: The rule of human skin collagenase in epidermolysis bullosa. J Invest Dermatol 1977; 68:119-124. Valle K-J, Bauer EA: Enhanced biosynthesis of human skin collagenase in fibrublast cultures from recessive dystrophic epidennolysis bullosa. J Clin Invest 1980; 66:176- 187. Winberg J-O, Gedde-Dahl T, Bauer FA: Collagenase expression in skin fibroblasts frum families with recessive dystrophic epidermolysis bullosa. J Invest Dermatol 1989; 92:82-85. Seltzer JL, Eisen AZ, Bauer EA, et al: Cleavage of type VII collagen by interstitial collagenase and by type IV collagenase (gelatinase) derived from human skin. J Biol Chem 1989; 26413822-3826. Hovnanian A, Duquesnoy P, Amselem S, et al: Exclusion of linkage between the collagenase gene and generalized recessive dystruphic epidermolysis bullosa phenotype. J C/in Invest 1991; 88:1716-1721.

Curr Probl Dermatol, May/June 1992

30. Scott D, Schultz R: Epidermolysis bullosa simplex in the collie dog. J Am Vet Med Assoc 1977; 171:721-727. 31. Johnson G, Kohn C, Johnson C, et al: Ultrastructure of junctional epidermolysis bullosa in Belgian foals. J Comp Path01 1988; 98:329-336. 32. Frame S, Harrington D, Fessler J, et al: Hereditary junctional mechanobullous disease in a foal. J Am Vet Med Assoc 1988; 193:1420-1424. 33. Kohn C, Johnson G, Garry F, et al: Mechanobullous disease in two Belgian foals. Equine Vet J 1989; 21297-301. 34. Dunstan R, Sills R, Wilkinson J, et al: A disease resembling junctional epidermolysis builosa in a toy poodle. Am J Dermatopathol 1988; 10:442-447. 35. Alley M, O’Hara P, Middelberg A: An epidermolysis bullosa of sheep. N Z Vet J 1974; 22:55-59. 36. Thompson K, Crandell R, Rugeley W, et al: A mechanobullous disease with sub-basilar separation in Bran@IS cakes. Vet Path01 1985; 22283-285. 37. Bruckner-Tuderman L, Guscetti F, Ehrensperger F: Animal model for dermolytic mechanobullous disease: Sheep with recessive dystrophic epidermolysis bullosa lack collagen VII. J Invest Dermatol 1991; 96:452-458. 38. Palmiter R, Brinster R: Germ-line transformation of mice. Ann Rev Genet 1986; 20:465-499. 39. Bailleul B, Surani M, White S, et al: Skin hyperkeratosis and papilloma formation in transgenic mice expressing a ras oncogene from a suprabasal keratin promoter. Cell 1990; 62:697- 708. 40. Sandgren E, Luetteke N, Palmiter R, et al: Overexpreasion of TGF-alpha in transgenic mice: Induction of epi: thelial hyperplasia, pancreatic metaplasia, and carcinoma of the breast. Cell 1990; 61:1121-1135. 41. Kim YH, Woodley D, Wynn K, et al: Recessive dystruphic epidermolysis bullosa phenotype is preserved in xenografts using SCID mice: Development of an experimental in vivo model. J Znvest Dermalol 1992; 98:191-197.

119