British Journal of Dermatology (1992) 127, 312-317.

Intracellular expression of type VII collagen during wound healing in severe recessive dystrophic epidermolysis bullosa and normal human skin J.A.McGRATH. I.M.LEIGH* AND R.A.J.EADY. St John's Institute of Dennatology. UMDS (St Thomas's campus), St Thomas's Hospital, Iondon, U.K. ' I'kpartment of Experimental Dermatology, The Royal Iondon Hospital. London. U.K. Accepted for puhiication I i May 1992

Summary

The ability of keratinocytes to synthesize basement membrane components in vivo during wound healing in normal human skin and in severe recessive dystrophic epidermolysis bullosa (RDEB) was investigated. Indirect immunofluorescence using anti-type VII collagen (VIIc. recognizing the globular non-helical component of the molecule), anti-type iV collagen, anti-laminin and bullous pemphigoid antisera. was performed on biopsies of intact skin and of healing skin taken between 7 and 14 days after dermatome injury (upper to mid-dermal wounding) in eight patients with severe RDEB and in seven normal subjects. Baseline anti-type VIIc immunofluorescence showed completely absent staining ofthe epidermis, dermis and dermo-epidermal junction in severe RDEB samples, and bright linear dermo-epidermal junction fluorescence in normal human skin. In 5/5 normal human skin samples taken 9-12 days post-wounding, some type VIIc expression was noted within basal cells as well as in a continuous or interrupted linear distribution at the basement membrane zone. In all the severe RUEB biopsies sampled between days 10 and 1 5 (5/5). anti-VIIc fluorescence was also seen with varying intensity within basal and lowermost suprabasal cefls, and in one day 14 sample at the dermo-epidermal junction. Low levels of intracellular type IVc were seen in both groups, but only in those samples taken 7-9 days after injury: later biopsies showed only continuous dermo-epidermal junction staining. Linear basement membrane zone labelling with laminin and bullous pemphigoid antisera was seen in all samples in both sets of subjects, even at day 7. but there was no detectable intracellular antisera staining. The finding of anti-type VII collagen immunofluorescence staining within the lower epidermis in severe RDEB during wound healing implies that these keratinocytes, under certain circumstances, have an inherent capacity to synthesize type VIIc. or at least part ofthe non-collagenous globular region ofthe molecule, and that a complete absence of type VIIc is not the primary abnormality in the most severe form of RDEB.

Recessive dystrophic epidermolysis bullosa (RDEB) is an inherited mechano-bullous disorder characterized clinically by fragile skin and mucous membranes, blisters, erosions, and chronic scarring. The most severe form of the condition is typified morphologically by an absence of recognizable anchoring fibrils at the dermo-epidermal junction and blister formation below the lamina densa.'"' Anchoring fibrils are mainly composed of type VII collagen (VIIc),"* and an absence of immunofluorescence staining with anti-type VIIc antibodies has been widely used in the diagnosis of severe RDEB.^"" It is not clear, however, whether the abnormality of type VIIc in Correspondence: Professor R.A.J.Eady. St John's Institute of I3ermatology, St Thomas's Hospital, Lambeth Palace Road, London SKI 7EH. U.K.

312

severe RDEB represents complete absence, altered synthesis, abnormal polymerization or excessive breakdown of the type VIIc molecule. In normal human skin, type VIIc is synthesized by both keratinocytes and fibroblasts.'*'' although the slow turnover of basement membrane components in intact skin is such that detection of type VIIc by immunofluorescent anti-type VIIc labelling is usually limited to linear staining at the dermo-epidermal junction. Studies of single biopsies of normal or EB skin are only able to give limited information on the dynamics of type VIIc synthesis. However, a process such as wound healing, associated with an increased synthesis of basement membrane components, may provide a usefui model for studying the synthesis and transport of type VIIc and other macro-

TYPE VII COLLAGEN IN WOUND HEALING IN RDEB

molecules of the basement membrane, especially if analysis of sequential biopsies is possible. Such an approach might enable one of the potential abnormalities of type VIIc in severe RDEB to be investigated.

Methods Subjects Eight patients (five male, three female: age range 18-44 years, mean age 27-9 years) with severe RDEB were studied. The diagnosis was based both on characteristic clinical features of severe, generalized, mutilating skin blistering and scarring, and on skin biopsy evidence of a sublamina densa level of blister formation and unrecognizable anchoring fibrils on electron microscopy. Each patient was undergoing plastic surgery to release chronic scarring-associated hand contractures and having a split-thickness skin graft taken from the anterior aspect ofthe thigh with a lubricated dermatome. Seven normal control subjects (six males, one female; age range 28-68 years, mean age 48-9 years) having a similar split-thickness skin graft procedure (e.g. for tattoo removal), using the thigh as donor site, were also studied.

Table I. Number of sequential skin samples from patients with recessive dystrophic epidermolysis bullosa IRDKBI and nurmal controls (NH.SI after

RDFB NHS

7

H

8

I I

t

Skin samples Skin specimens for baseline assessment (day 0) were taken from the edge of the split-thickness skin graft specimen immediately after dermatome removal from the thigh. Following informed consent, each subject then had one further biopsy (6-mm punch biopsy in controls, shave biopsy in severe RDEB group) taken from the healing donor site on the thigh 7-14 days later (see Table 1 for timing of biopsies).

Immunojluorescence Immediately after collection, skin samples were embedded in OCT compound (Miles laboratories. Elkhart. U.S.A.) and snap-frozen in liquid nitrogen. Cryostat sections (5 /im) were then cut, air-dried, pre-incubated with either normal rabbit or normal goat serum at room temperature for 1 5 min to reduce non-specific secondary antibody binding, and then incubated in primary antibody for 1 h at 37°C (Table 2). After washing twice in phosphate-buffered saline (PBS) for 30 min, the sections were incubated in fluorescein-conjugated anti-IgG for 1 h at 37°C, washed in PBS and distilled water, and mounted in glycerol/PBS containing ()• 1 % p-phenylenediamine. Sections incubated in negative control primary sera (rabbit, goat or normal human serum) followed by fluorescent secondary antibody, were also examined.

Results Normal humar} skin

Time of biopsies (day after wounding)

0

313

9

H)

11

12

1i

14

__

1

2

1

— —

2 1

1

—

1 1

Daij 0, Baseline immunofluorescence showed continuous linear dermo-epidermal junction (DEJ) labelling with anti-VIIc (Eig. la). anti-IVc (Fig. lb), anti-laminin, and bullous pemphigoid (Fig. lc} antisera. Fluorescent staining around blood vesseis. nerves and epidermal

Table 2, Antibodies used in Immunofiuorescence study

Primary antibody I,H 7:2 Anti-type IV collagen' Anti-laminin" Rullous pemphigoid antisera^

Comments

Dilution

Secondary antibody

Dilution

Mouse monoclonal (globular region of type VII eoilagen) Rabbit polyclonal Rabbit polyclonal Human polyclonal

1:2 1:100 1:100 1:10

Rabbit anti-mouse' Goat anti-rabbit' Goat anti-rabbit' Goat anti-human'

1:50 1:50 1:50 1:50

"ICN Immunobiologicais, U.K. {"Recognizes 2 30-kDa protein by immunoblotting.

314

J.A.McGRATH et al

Figure 1. Baseline immunofluorescence tindings in normal human skin (NHS, top row) imd severe recessive dystrophic epidermolysis bullosii IRDKB, bottom row), (LI, d| Anti-type VII collagen labelling. There is linear staining iilong the dermo-epidermal junction (Dl-;)| in NHS (a), but completely negative labelling of the DEJ (arrowed) in severe RDF.B (d). lb. e) Anti-type IV collagen labelling. Both groups have continuous fluorescence along the DE) iind around dermal vessels and nerves. In RDF.B |e|, the antibody stains the blister roof. lc, fl Ruilous pemphigoid antiserum labelling. Both sets of patients hiive linear staining of the DE). which localizes to the blister roof in ROKB if). (Magnilications: a x 9 5 , b x 6 5 . c x 100. d x 6 5 , e X 135, f x 100),

appendages was also seen with type IVc and laminin antisera. No intraepidermal labelling was seen. Day 7-14. Anti-VIIc labelling was not noted before day 9. Thereafter it was seen in a continuous or interrupted linear piittern along the DKJ as weii as in granular intracellular deposits within basal keratinocytes (Fig. 2a). although by day 14 no intraepidermal staining was detected and only continuous linear DKJ fluorescence was found. In those sections with continuous DEJ staining, the labelling was most intense around hair follicles, when compared with adjacent interfoilicular epidermis (Fig. 2c). Some intraepidermal labelling with anti-type IVc antiserum was seen on the day 7 sections (Fig. 2f) in addition to interrupted linear DEJ staining and pronounced upper dermal fluorescence in association with new blood vessel formation. By day 10 both the epidermis and DEJ had similar appearances to the baseline state. Laminin and bullous pemphigoid antisera fluorescence was linear and continuous at the dermoepidermal junction in all wound healing specimens examined, with no intracellular labelling seen at any stage. None of the control samples with fluorescent secondary antibody alone had any intraepidermal, DEJ. or dermal staining. There was some slight variability in the intensity of the labelling within samples from different individuals taken at similar times after injury. This might be explained by differences in the original depth of wounding by the dermatome giving rise to variations in the pattern or rate of keratinocyte migration and differentiation. Histologically. there was con-

siderable epidermal hyperplasia at the time of intraepidermal collagen labelling during wound healing. Severe recessive dystrophic epitiennolysis hullosa

Day 0. Baseline anti-VIIc labelling studies showed a complete absence of staining in all eight severe RDEB subjects (Fig. Id). Type IVc (Fig. le). laminin. and bullous pemphigoid (Fig. If) antisera had a similar labelling pattern to normal human skin, although in areas of blistering or dermo-epidermal separation all antibodies stained the roof of the specimen. Day 7-14. At day 7, no anti-VIIc labelling was noted (as in normal human skin). However, in all five biopsy samples taken 10-13 days after wounding some antiV^IIc fluorescence within the basal keratinocytes and lowermost suprabasal cells (Fig. 2b) was seen, albeit less bright than in normal human skin. In the single day 14 biopsy, faint, discontinuous anti-VIIc fluorescence was present at the DEJ (Fig. 2d). Slightly increased dermal fluorescence, possibly in association with flbroblasts, was also found in afl the 10-14 day biopsies. Type IVc (Fig. 2g and 2h), laminin and bullous pemphigoid (Fig. 2e) antisera labelling showed similar characteristics to the normal controls.

Discussion The LH 7:2 monoclonal antibody'" recognizes an epitope on the non-collagenous globular region of type

TYPE VII COLLAGEN IN WOUND HEALING IN RDEB

US

Figure 2. (a) Anti-type VII collagen NHS day 10, Discrete granular staining is seen within basal and lowermost suprabasal cells in addition to continuous linear DEJ fluorescence 1 X 260). (b) Anti-type VII collagen RDEB day 12. Intracellular fluorescence is present within basal and some suprabasal cells (cf, day 0 appearances in Fig. Id). No delinite Dt) staining is seen ( x 14()|. (c) Anti-type VII collagen NHS day 12, Tbere is conlinuous linear staining ofthe DE), but no intraepidermal fluorescence. The foHicular epithelium has brighter labelling than the interfoilicular epidermis (arrowed, x45), (d) Anti-type VII collagen RDEB day 14, No keratinocyte labelling is seen, bul scanty, discontinuous fluorescence is noted beneath the byperplasHc epidermis at (he DE) (cf, day 0 appearances in Fig. Id, x 90|. (e) Bullous pemphigoid antiserum RDF.B day 10. There is continuous linear DFJ fluorescence which localizes to the roof of blistered skin ( x 1 55). (f) Anti-type IV collagen NHS day 7. Granular fluorescence is seen intracellularly within the lower epidennis (arrows). Some dermoepidermal separation is noted even in NHS at this stage in wound healing ( x 1 50), (g) Anlitype IV collagen RDEB day 7, Some discontinuous labelling (arrows) is observed both at the DHJ and within some basal cells. Some dermo-epidermat separation/early blister formation is also present ( x 1 35), (b) Anti-type IV collagen RDEB day 14, Near-continuous linear DEJ lluorescence is seen, with the antibody staining the roof of separated skin ( X 140).

VIIc. a major component of anchoring fibrils.•*'-^' It has been widely used in the diagnosis and classification of dystrophic EB.'' '^ In nearly all cases of severe, mutilating, generalized RDEB there is absent immunofluorescence staining ofthe dermo-epidermal junction with this antibody, as found in the present study. However, similar clinical cases may have positive (bnt reduced) labelling ofthe dermo-epidermal junction using a different, purified, concentrated monoclonal antibody to the globular part of type VIIc.'' suggesting that there may be some basement membrane zone expression of type VIIc, even in severe RDEB. In vitro studies, primarily examining various epithelial cell lines, have suggested that type VIIc is mainly a product of epithelial cells, with a lesser additional contribution fromfibroblasts.•***There has been little in vivo evidence of type VIIc synthesis by keratinocytes apart from studies on grafted cultured epithelium.' ^ The fine granular LH 7:2 fiuorescence staining seen within the basal cells of normal human skin during wound

healing in this study provides IM vivo confirmation of keratinocyte type Vile synthesis. In the severe RDEB skin samples, intraepidermal type VIIc fluorescence during wound healing is a new finding. This suggests that the keratinocytes in severe RDEB do have the capacity to synthesize or upregulate production of at least part of the type VIIc molecule, although exact quantification of synthesis is difficult to assess solely by immunofluorescence. In future, studies using antibodies to other parts of the collagen molecule will be necessary to assess further the completeness of type VIIc synthesis. Additionally, immunoblotting techniques could be used to assess the presence of type VIIc,'^ and the application of in situ hybridization using recently described c-DNA probes for type VIIc mRNA'' would help to confirm or refute the immunofluorescence findings. Skin samples taken after the 2-week study period would also be necessary to comment more fully on the secretion of type VIIc. The finding of intraepidermal type VIIc in severe RDEB during wound healing is somewhat different from

J.A.McGRATH et al

observations noted in recent studies suggesting retention of type VIIc within the epidermis in a small number of dystrophic EB cases.^^"-' In those cases, the type VIIc abnormality was mostly seen as a transient phenomenon representing a temporary defect in intracelluiar processing, possibly calcium dependent."' or impaired secretion of the collagen molecule. In severe RDEB, a more permanent abnormality of type VIIc persists, which may be due to incorrect synthesis (molecular mutations of type VIIc). defective association with accessory macromolecules,-' or excessive molecular breakdown (proteolytic enzyme activity)^*^^ or a combination of one or more of these potential defects. Intracellular type VIIc has also been observed in a proportion of well-differentiated squamous cell carcinomas arising in normal human skin,'*" suggesting that changes in keratinocyte differentiation (as occurs in a regulated fashion in wound healing) have a direct influence on the synthesis and expression of basement membrane components. Further work to determine the cellular growth factors and cytokines involved in promoting type VIIc synthesis (e.g. transforming growth factor-^)''-''•^'* under such circumstances may have some potential therapeutic relevance to dystrophic EB. particularly if a relative deficiency of type VIIc is the essential underlying abnormality. However, if a primary molecular type VIIc defect or abnormalities in secretion or polymerization of type VIIc underlie severe RDEB. sustaining growth factor expression may be inappropriate in attempting to modify this genodermatosis.

Acknowledgments The assistance of Mr Mayou (Consultant Plastic Surgeon. St Thomas's Hospital) in providing skin samples is gratefully acknowledged. Bullous pemphigoid antiserum was kindly provided by Mr Bhogal (St John's Institute of Dermatology). This work was supported by grants from the Dunhill Trust and from the Dystrophic Epidermolysis Bullosa Research Association (DEBRA, U.K.), Dr McGrath is in receipt of a DEBRA research fellowship.

References 1 Tidman MJ, Eady RAJ. Evaluation of anchoring fibrils and other components ofthe dermal-epidermal junction in dystrophic epidermolysis buliosa by a quantitative ultrastructurai technique. / Invest l-)ermatol 19S^:84: J74-7, 2 Briggaman RA, Wheeler CE, Epidermolysis bullosa dystrophicarecessive: a possible role of anchoring fibrils in the pathogenesis. / Invest Dermatol 1975: 65: 203-11. 3 Briggaman RA. Is there any specificity to defects of anchoring

fibrils in epidermolysis bullosa dystrophica, and what does Ihis mean in terms of pytliogenesis.- / Invest Dertnatol 1985:84: J7 I-J. 4 Sakai I.Y. Keene DR. Morris NF, Burgeson RE. Type Vil collagen is a major structural component of anchoring fibrils. / Cell Bio! 1986; 106: lS77-8f>. 5 l£igh IM. Eady RAJ, Heagerty AHM et al. Type VII coHagen is a normai component of epidermal basement membrane, whicii shows aitered expression in recessive dystrophic epidermolysis butlosa. / Invest Dermatol 1988: 90: fi 19-42. b Heagerty AHM, Kennedy AR. Leigh IM et al. Identification of an epidermal basement membrane defect in recessive forms of dystrophic epidermolysis bullosa by LH 7:2 monoclonal antibody: use in diagnosis, Br } IMmatol 1986; 115: 12 S- i 1. 7 Bruckner-Tuderman L, Mitsubashi Y, Schnyder UW. Bruckner P. Anchoring Hhrils and type VII collagen are absent from skin in severe recessive dystrophic epldermolysis bullosa. / Invest lkrmatol 1989; 93: i-9. 8 Stanley JR, Rubinstein N, Klaus-Kovtun V. Kpidermolysis bullosa acquisila antigen is synthesized by both human keratinocytes and butnan dermal libroblasts. / Invest Dermatol 1985; 85: 542-5. 9 Konig A, Bruckner-Tuderman L, Epitheiial-mesenchymal interactions enbance expression of collagen VH in vitro. / Invest Lkrmatol 1991:96:805-8. 10 Leigb IM. Furkis PE, Bruckner-Tuderman L. I,H 7;2 monoclonal antibody detects type VII collagen in the basement membrane of ectodermally derived epithelia including skin. Epithelia 1988; 1: 17-29, 1 1 Sbimizu H. McDonald |N, (lunner DB t'l td. Epidermolysis bullosa acquisita antigen and tbe carboxy terminus of type VII collagen have a c-ommon immunolocalization to anchoring fibrils and lamina densa of basement membrane, Br / Dermatol 1990: 122: 577-85. 12 Heagerty AHM, Kennedy AR, (Gunner DR, Eady RAJ. Rapid prenatal diagnosis and exclusion of epidermolysis bullosa using novel aniibody probes. / Invest liermatol 198fi: 86: 603-5. I i Rusenko KW, Gammon WR, I'ine J-D, Briggaman RA. Tbe carhoxyl-terminal domain of type Vll collagen is present at the basement membrane in recessive dystrophic epidermolysis bullosa. / Invest Dermalol 1989; 92: 62 J-7. 14 Regauer S, Sciler GR, Barrandon Y fl al. Epitbelial origin of cutaneous anchoring fibrils. / Cell Hiol 1990; 111: 2109-15. 1 5 ParenteMG, Chung LC, Ryynanen \elal. Human type VII collagen: cDNA cloning and chromosomal mapping ofthe gene, Proc Matl AcadSci USA 1991; 88: 6951-5. 16 Fine J-D, Horiguchi Y, Stein DH et al. Inlraepidermal type VII collagen. Evidence for abnormal intracytoplasmic processing of a major basement membrane protein in rare patients with dominant and possibly localized recessive forms of dystrophic epidermolysis buliosa.//iHi Acad Dermatol 1990; 22: 188-95. 1 7 Smith LT, Sybert VP, Intra-epidermal retention of type VII collagen in a patient with recessive dystrophic epidermolysis bullosa. / Invest Dermatol 199(1; 94: 261-4, 18 Schofield OMV, Yamamoto A, MclJrath ) ct al. Transient hullous dermolysis of the newhorn; a disorder of type VII collagen secretion.^ Br j Dertnatol 1991; I25|Suppl, 38): 89. 19 Hashimoto K, Matsumoto M, Iacobelli D, Transient bullous dermolysis ofthe newborn. Arch Dermalol 1985; 121: 1429-38. 20 Hasbimoto K, Burk JD, Bale GE el al. Tran.sient bullous dermolysis of the newborn; two additional cases. / Am Acad Dermalol 1989; 21: 7 0 8 - n . 21 Eng AM, Keegan CA, Hashimoto K et al. Transient bullous dermolysis ofthe newborn. Case report and review of patbogenesis. ICutPathol 1991; 18: 328-32.

T Y P E VII COLLAGEN IN W O U N D HEALING IN RDEB

22 Goldsmith LA, McCoon P, Partridge A, Lane AT. Intra-epitbelial anchoring fihril components. Anh Dermatol 1991; 127: 5 J-6. 2 i Sawamura D. Sugawara T, Hashimoto I et al. Increased gene expression of matrix metalloproteinase-J (stromelysin) in skin tibroblasts from palients wilh severe recessive dystrophic epidermolysis bullosa. Biochem Biophys Res Comnnin 1991; 174: 100 i-8, 24 Winberg )(), Gedde-Dabl T |r, Bauer EA. Collagenase expression in skin librobtasts from f;imilies vvilb recessive dystrophic epidermolysis bullosa. / Invesl Dermalol 1989; 92: S2-5, 2 5 Bauer t-lA. Recessive dystrophic epidermolysis bullosa: evidence for an altered coliagenase in fibrobliist cultures. Proc Natl Acad Sci USA 1977;74:4f)4f)-50,

il7

26 Wetzels RHW. Robben HCM. Leigh IM et al. Distribution patterns of type VII collagen in normal and malignant human tissues. Am J Pathol 1991: l 5 9 : 4 S l - 9 . 27 Konig A, Bruckner-Tuderman L. Transforming growtb faclor-beta stimulates collagen Vll expression in keratinocytes and libroblasts in vitro. / Invest Dermatol 1991: 96: 1029. 28 Quaglino D )r, Nanney LB. Ditesbeim |A, Davidson ).M. Transforming growtb factor-beta stimulates wound bealing and modulates extracellular matrix gene expression in pig skin: incisional wound model, / Invest Dermatol 1991: 97: 34-42.