BP180 Gene Delivery in Junctional Epidermolysis Bullosa

Home , Hemidesmosome

CS Seitz1, GJ Giudice2, SD Balding2, MP Marinkovich1 and PA Khavari1 1VA Palo Alto Health Care System, Palo Alto, CA and Department of Dermatology, Stanford University, Stanford, CA; and 2Department of Dermatology, Medical College of Wisconsin, Milwaukee, WI, USA

Epidermolysis bullosa (EB) comprises a family of inherited patients with generalized atrophic JEB. Restoration of full- blistering skin diseases for which current therapy is only length BP180 protein expression was associated with palliative. Junctional EB (JEB) involves dissociation of the adhesion parameter normalization of primary JEB kera- dermal–epidermal junction and results from mutations in tinocytes in vitro. These cells were then used to regenerate a number of genes that encode vital structural proteins, human skin on immune-deﬁcient mice. BP180 gene- including BP180 (type XVII collagen/BPAG2). In order to transduced tissue demonstrated restoration of BP180 gene develop a model of corrective gene delivery for JEB, we expression at the dermal–epidermal junction in vivo while produced a retroviral expression vector for wild-type untransduced regenerated JEB skin entirely lacked BP180 human BP180 and used it to restore BP180 protein expression. These ﬁndings provide a basis for future efforts expression to primary keratinocytes from BP180-negative to acheive gene delivery in human EB skin tissue.

Keywords: epidermolysis bullosa; BP180/type XVII collagen/BPAG2; gene therapy; skin disease

Introduction With the molecular characterization of the underlying genetic defects responsible for these blistering disorders, Epidermolysis bullosa (EB) is the term utilized to charac- the opportunity for the development of rational genetic 1 terize a family of inherited blistering diseases. EB sub- therapies has arrived. COL17A1 encodes BP180 and is groups are divided based on the level of blister formation mutated in a subtype of JEB characterized by a generally within the skin, with location of the split found to corre- non-lethal clinical course with blisters that heal leaving late with the genetic lesion responsible for the underlying atrophic skin changes.9 BP180 is also involved in the 2–5 disorders. In the case of epidermolysis bullosa simplex, pathogenesis of the autoimmune blistering disease, bull- the plane of blister formation occurs in the most super- ous pemphigoid,17 in which patient autoantibodies are ficial location of EB subgroups, namely within the lower generated against BP180. BP180 is a collagenous trans- epidermis, and has been shown due to mutations in membrane type II protein component of the hemidesmo- 6,7 genes encoding basal keratins 5 and 14 and more some18–20 involved in adhesion of the epidermis to the 8 recently, plectin. Abnormalities of protein components underlying dermis.21–23 The size and multiple domains of of the basement membrane zone are associated with a this protein pose major barriers to developing effective deeper split at the dermal–epidermal junction, producing conventional pharmacologic therapy for BP180-deficient the clinical disorders collectively known as junctional epi- JEB. Therapeutic gene delivery is a general approach dermolysis bullosa (JEB). Proteins encoded by genes being developed in a wide range of tissues and the skin mutated in JEB include BP180 (type XVII collagen, is an attractive tissue for this technology by virtue of its 9 ␣ ␤ ␥ 10,11 BPAG2 or HD-4), Laminin 5 3, 3 and 2 chains accessibility and by the existence of such well-charac- ␣ ␤ 12–14 and integrins 6 and 4. Deeper blisters occurring terized monogenic recessive diseases to serve as human below the level of those seen in JEB characterize a group disease model prototypes. Development of gene therapy of patients afflicted with dystrophic EB have been found for EB requires establishment of models of restorative 15,16 to be due to mutations in the type VII collagen gene. gene delivery. In order to achieve this, we transduced Because the proteins affected in EB are generally large early passage keratinocytes from BP180-deficient JEB structural proteins with multiple functional domains, patients with a retroviral expression vector for wild-type small molecule pharmacotherapy has thus far been of human BP180. BP180 gene transfer results in improve- little benefit. In contrast to the major advances in under- ment of functional defects in JEB cell adhesion in vitro, standing the pathogenesis of these disorders, progress in as well as restores expression of BP180 protein at the the treatment of EB patients has been minimal and basement membrane zone in JEB skin regenerated in vivo. current therapy is only supportive. Results BP180 gene transfer to BP180-negative JEB Correspondence: PA Khavari, VA Palo Alto Health Care System, Derma- keratinocytes tology Service, 3801 Miranda Ave, Palo Alto, CA 94304, USA In order to develop a model of gene delivery in EB we Received 7 June 1998; accepted 24 August 1998 produced an amphotropic retroviral expression vector for BP180 gene delivery CS Seitz et al 43 human BP180 (Figure 1a). This vector was used to trans- duce primary keratinocytes from two patients with generalized atrophic benign EB. The patients in this study lack BP180 protein expression but express normal levels of other cutaneous basement membrane zone proteins, including laminin 5 and ␣6␤4 integrin. The patient whose cells were used for in vivo experiments has also been confirmed to have mutations in both BP180 alleles (Marinkovich MP et al, unpublished). Untransduced JEB cells demonstrated no full-length BP180 protein expression as detected by a panel of monoclonal and polyclonal antibodies to human BP180. JEB cells transduced with this vector, however, demonstrated full-length BP180 in amounts comparable to that seen in normal cells (Figure 1b), with similar results obtained in cells from both patients. Figure 2 Effect of BP180 gene transfer on cellular adhesion in vitro. Forty-eight hours after gene transfer, cells were treated with 0.05% tryp- Effect of BP180 gene transfer on cellular adhesion in sin in phosphate-buffered saline and the percentage dissociation determ- vitro ined in triplicate from triplicate transductions at the range of timepoints To determine if BP180 gene transfer exerts a functional shown. NL, normal control primary keratinocytes from individuals without skin disease; EB[−], untransduced primary BP180-negative JEB effect on JEB cells, we performed an in vitro assay of the patient keratinocytes; EB[+], primary BP180-negative JEB patient kera- central function in which these cells are defective, tinocytes transduced with retroviral expression vector for BP180. adhesion. Defective adhesive properties, believed responsible for the clinical blistering seen in JEB patients, can be readily measured in JEB cells in vitro.24 We performed a trypsinization assay to determine the kinetics of cell dis- Regeneration of JEB skin tissue in vivo on immune- sociation from a tissue culture substrate. BP180 gene deficient mice transfer is associated with a restoration of dissociation After restoring BP180 expression to JEB cells and ident- kinetics to normal levels (Figure 2a), suggesting a ifying evidence of a corrective functional effect in vitro, functional correction of this abnormality. we wished to determine if such genetically engineered early passage cells could be useful in BP180 gene delivery in vivo. To do this, we established an in vivo human tissue model of BP180-negative JEB. Transduced and untreated JEB patient cells, along with normal controls, were used to regenerate human skin tissue on immunodeficient mice.25,26 Untreated JEB cell grafts did not demonstrate frank evidence of clinical blistering. Four weeks after the operation, skin tissue was excised and analyzed. Both transduced and untransduced JEB cells regenerated skin with a histologic appearance similar to that produced by normal controls (Figure 3a and b). However, a subepidermal blister was detected histologically in one skin tissue graft regenerated from untransduced JEB cells (Figure 3c) in a pattern similar to that seen commonly in regenerat- ing subepidermal blisters.27 Such histologic evidence of blistering was not observed in any of the stepwise tissue sections from normal controls or from JEB skin regenerated after BP180 gene transfer in this study and has also not been observed in analysis of normal control human grafts generated by similar methods in previous studies.28 The human species origin, as well as the expression of basement membrane components other than BP180, was confirmed by immunostaining with species-specific antibodies for human laminin 5 ␤3 chain (Figure 3d–f).29

Restoration of BP180 expression to regenerated JEB Figure 1 BP180 gene transfer to BP180-negative JEB keratinocytes. skin in vivo (a) Diagram of amphotropic retroviral expression vector for human To determine if BP180 gene transfer in vitro before graft- BP180. The basepair (bp) positions of the 7311 bp vector are noted below ing could lead to restored BP180 expression in vivo, the schematic. (b) Restoration of BP180 expression in JEB cells. Forty- immunostaining was performed with antibodies to BP180 eight hours after gene transfer, cell extracts were prepared, electrophoresed on regenerated human skin tissue specimens. At 4 weeks via 8% SDS-PAGE and subjected to immunoblotting with antibody to after grafting, untransduced cells regenerated human epi- BP180. NL, primary keratinocytes from patients with normal skin; EB[−], untransduced primary BP180-negative JEB patient keratinocytes; EB[+], dermis that entirely lacked BP180 protein expression, as primary BP180-negative JEB patient keratinocytes transduced with expected. BP180 vector-transduced JEB cells, however, retroviral expression vector for BP180. demonstrated restoration of BP180 expression in the cor- BP180 gene delivery CS Seitz et al 44

Figure 3 Regeneration of JEB skin tissue in vivo on immune-deficient mice. JEB keratinocytes and normal controls were used to regenerate human skin on immune deficient mice in vivo. (a–c) Histologic appearance of regenerated JEB skin. BP180[−], tissue regenerated from primary JEB keratinocytes lacking BP1N80 expression; BP180[+], tissue regenerated from primary JEB keratinocytes from the same patient after BP180 gene transfer. Scale bars (a,b) 25 ␮m and (c) 100 ␮m. (d–f) Confirmation of species origin and basement membrane zone component production in regenerated skin. Immunofluo- rescence staining of regenerated JEB and control skin grafts was performed with antibodies specific for human laminin. Scale bars (d–f) 25 ␮m.

rect basement membrane zone location, in a pattern Among the first of these is re-introduction of the relevant similar to normal control (Figure 4a–f). wild-type gene to patient cells in vitro, with transformed cell lines usually used initially, as has previously been Discussion achieved in cells immortalized from a patient with another EB subtype deficient in the laminin 5 ␥2 chain.24 Defects in the basal layer program of epidermal gene When effective gene transfer approaches have been expression lead to a number of inherited blistering dis- developed, this work is extended to non-immortalized eases, including JEB. Following identification of relevant primary cells of the target tissue as we have done here. disease genes, the development of rational genetic ther- A related next step is demonstration of restored gene apies may require progression through a series of steps. expression and correction of a measurable disease charac-

Figure 4 Restoration of BP180 expression to regenerated JEB skin in vivo. Immunofluorescence staining of regenerated JEB and control skin grafts was performed with antibodies specific for human BP180. (a,d) BP180[+], tissue regenerated from BP180-deficient primary JEB keratinocytes after BP180 gene transfer in vitro. (b,e) BP180[−], tissue regenerated from primary JEB keratinocytes from the same patient without previous BP180 gene transfer. (c,f) NL, normal control regenerated tissue from keratinocytes derived from patients without skin disease. Scale bars, 100 ␮m for (a–c); and 25 ␮m for (d–f). BP180 gene delivery CS Seitz et al 45 teristic in primary patient cells in vitro. In the case of skin, protein expression to JEB skin tissue, restoration of gene such a demonstration is best followed by establishment expression via the gene transfer approach here has been of an in vivo model of restored gene expression for a observed to be lost over the course of a month.26,32,33 This given disease using the non-transformed cells and gen- loss appears to be due to promoter inactivation rather eral approaches that are planned for eventual use in than death of genetically engineered cells.30 This work, humans. These last two steps are the focus of the by demonstrating the first restoration of gene expression current work. in non-immortalized JEB skin tissue in vivo, thus rep- Here we have used gene transfer to JEB cells in order resents a discrete advance in efforts to develop ex vivo to restore expression of BP180, a protein component of gene therapy for JEB. Before application to humans, how- hemidesmosomes necessary for normal dermal–epider- ever, extensive efforts will be required to develop a mal cohesion. Immunostaining performed with human refined model of corrective gene delivery meeting key species-specific antibody to BP180 on adjacent mouse additional criteria, such as sustainability. New vectors skin demonstrated no specific staining, therefore exclud- designed to overcome this problem show promise in this ing the possibility that selective murine tissue ingrowth regard and may support gene delivery in human skin could account for the restoration of BP180 in JEB skin. To through multiple epidermal turnover cycles.30 These vec- verify that other basement membrane zone components tors offer hope for future efforts to achieve long-lasting were present in unengineered BP180-negative JEB skin therapeutic gene delivery in JEB gene transfer models as regenerated in vivo, we performed immunostaining with a step before human trials. Generation of this model of antibody to laminin 5, another important basement mem- gene transfer in human JEB offers a next step in the path- brane zone protein that is affected in other types of JEB. way to developing rational corrective genetic therapies Patient and control specimens all demonstrated a normal for EB and other human genetic skin disorders. laminin 5 expression pattern, indicating that the lack of BP180 expression in unengineered JEB skin recapitulates the disease picture seen in humans and is not due to a Materials and methods global failure of basement membrane zone component formation by BP180-negative cells in this model. These JEB patients, cells and cell culture human species-specific antibodies to laminin 5 also pro- JEB patients were characterized as having junctional EB vided further confirmation of the human origin of the on the basis of previously published criteria.1 These analyzed tissue. included the clinical presence of a congenital mechano- The JEB patient keratinocytes studied here regenerated bullous eruption in a generalized distribution that led to human skin tissue with no clear clinical evidence of a atrophic scars, a lack of internal organ involvement and blistering phenotype, however, subepidermal bulla for- evidence of a subepidermal split on histologic examin- mation was observed at the histologic level. The reason ation. In addition, JEB patient skin in the two unrelated for a lack of clinical bullae is unclear, however, it may Caucasian patients studied here, was characterized as be explained by the fact that generalized atrophic JEB is lacking BP180 expression but demonstrated normal levels characterized by significant areas of uninvolved skin and of expression of other basement membrane zone compo- by a milder blistering phenotype than that seen in the nents implicated in EB, including laminin 5 chains, ␣6␤4 1 severe Herlitz form. In addition, our xenograft model integrins and type VII collagen. Normal control kera- may lack elements, such as adequate surface area or tinocytes were obtained from unrelated individuals unaf- appropriate types of mechanical trauma, needed to recap- fected by EB and without evidence of any other skin dis- itulate formation of clinically apparent blisters. BP180 ease. Biopsies were obtained in accordance with Stanford protein is normally only expressed in the basement mem- IRB approved protocols and early passage keratinocytes brane zone by cells in the basal layer of stratified epi- were grown as previously described.34,35 thelia. In these studies, we used a retroviral promoter which is active in all layers of stratified epithelium,30 however, we found this basal pattern of protein Amphotropic retroviral expression vector for human expression was maintained. This retention of basement BP180 A 4751 bp HindIII/XhoI fragment corresponding to the membrane zone expression and the lack of evidence of 20 suprabasal protein expression could point to the exist- full length human BP180 cDNA minus 3’ polyadenyl- ence of mechanisms constraining BP180 protein ation sequences was subcloned into the HindIII and SalI expression to the basal surface of cells of the basal layer, sites of the MFG based episomal expression vector LZRS36 and amphotropic retrovirus produced as pre- perhaps via complex formation with integrins associated 35 with the hemidesmosome.22,23 While retroviral LTR viously described. elements can express certain target genes in all layers of stratified epithelium,30 however, in the absence of in situ BP180 gene transfer to JEB cells hybridization data we cannot conclude that BP180 Retroviral transduction of BP180 cells was performed as mRNA is actually stably expressed throughout the epi- previously described.35 Briefly, 5 × 104 primary kera- dermis. These data do suggest, however, that well charac- tinocytes were seeded per 35-mm well and incubated for terized non-targeted promoters such as the retroviral LTR 12–24 h. Cells were pre-incubated in 2.5 ml of SFM/154 or endogenous housekeeping gene promoters could be media with 5 ␮g/ml polybrene for 5 min, followed by useful in efforts to restore BP180 expression in JEB. addition of 2.5 ml thawed viral supernatant sup- The restoration achieved here and in work with other plemented to 5 ␮g/ml polybrene. Plates were centrifuged genodermatoses26,31 defines a gene transfer approach that at 300 g for 1 h at 32°C. The medium was then aspirated is generally not durable enough to be immediately from wells, after which 3 ml fresh SFM/154 media was applied to humans. Although we have restored BP180 added for cell recovery and further growth in vitro. BP180 gene delivery CS Seitz et al 46 BP180 gene expression analysis linked to mutations in the gene (LAMC2) for the gamma 2 sub- For immunoblotting, keratinocyte cell extracts were pre- unit of nicein/kalinin (LAMININ-5). Nat Genet 1994; 6: 299–304. pared35 and 20 ␮g of extract protein electrophoresed on 11 Pulkkinen L et al. Mutations in the gamma 2 chain gene an 8% denaturing polyacrylamide gel before gene trans- (LAMC2) of kalinin/laminin 5 in the junctional forms of epider- fer and blotting with antibodies to BP180. Both mono- molysis bullosa. Nat Genet 1994; 6: 293–297. 12 Vidal F et al. Integrin beta 4 mutations associated with junctional clonal37 and rabbit polyclonal38 anti-BP180 antibodies epidermolysis bullosa with pyloric atresia. Nat Genet 1995; 10: were used. Antibodies to BRG1, a constitutively 39 229–234. expressed protein, was used as an internal control to 13 Ruzzi L et al. A homozygous mutation in the integrin alpha6 verify equivalence of extract quality and protein transfer gene in junctional epidermolysis bullosa with pyloric atresia. J between samples. Biopsy specimens were obtained from Clin Invest 1997; 99: 2826–2831. regenerated skin xenografts at 4 weeks after the oper- 14 Pulkkinen L et al. Homozygous alpha6 integrin mutation in ation. Frozen tissue skin sections were obtained in step- junctional epidermolysis bullosa with congenital duodenal atre- wise fashion through each tissue block. Sections were sia. Hum Mol Genet 1997; 6: 669–674. then fixed in acetone at room temperature and incubated 15 Christiano AM et al. A missense mutation in type VII collagen with antibody to BP180 noted above37 or to a monoclonal in two affected siblings with recessive dystrophic epidermolysis antibody to laminin 5␤3 chain40 followed by a fluor- bullosa. Nat Genet 1993; 4: 62–66. escein-labeled secondary antibody (Sigma, St Louis, MO, 16 Hilal L et al. A homozygous insertion-deletion in the type VII USA) before analysis. Immunofluorescence and immuno- collagen gene (COL7A1) in Hallopeau-Siemens dystrophic epid- 40 ermolysis bullosa. Nat Genet 1993; 5: 287–293. blot analysis was performed as previously described. 17 Liu Z et al. A passive transfer model of the organ-specific autoimmune disease, bullous pemphigoid, using antibodies gener- Production of engineered JEB skin in vivo ated against the hemidesmosomal antigen, BP180. J Clin Invest Human JEB and control epidermis 1.5 × 1.5 cm in size 1993; 92: 2480–2488. was regenerated on immune-deficient SCID mice from 18 Diaz LA et al. Isolation of a human epidermal cDNA corre- cells grown in vitro as previously described.25,26 Five mice sponding to the 180-kD autoantigen recognized by bullous pem- were grafted for each of the following: normal kera- phigoid and herpes gestationis sera. Immunolocalization of this tinocytes, untransduced JEB keratinocytes and BP180- protein to the hemidesmosome. J Clin Invest 1990; 86: 1088–1094. transduced JEB keratinocytes. 19 Giudice GJ, Squiquera HL, Elias PM, Diaz LA. Identification of two collagen domains within the bullous pemphigoid autoantigen, BP180. J Clin Invest 1991; 87: 734–738. Acknowledgements 20 Giudice GJ, Emery DJ, Diaz LA. Cloning and primary structural analysis of the bullous pemphigoid autoantigen BP180. J Invest This work was supported by the Office of Research and Dermatol 1992; 99: 243–250. Development, Department of Veterans Affairs (VA) and 21 Masunaga T et al. The extracellular domain of BPAG2 localizes by NIH AR44012 to PAK and MPM, AR40410 to GJG and to anchoring filaments and its carboxyl terminus extends to the AR43371 to PAK. CSS is the recipient of a post-doctoral lamina densa of normal human epidermal basement membrane. fellowship award from Deutsche Forschungsgemeinsch- J Invest Dermatol 1997; 109: 200–206. aft. We thank Q Lin, H Tran, N Griffiths and M Janoff 22 Hopkinson SB, Baker SE, Jones JC. Molecular genetic studies of a for assistance and A Lane, E Bauer and L Diaz for help- human epidermal autoantigen (the 180-kD bullous pemphigoid ful advice. antigen/BP180): identification of functionally important sequences within the BP180 molecule and evidence for an inter- action between BP180 and alpha 6 integrin. J Cell Biol 1995; 130: References 117–125. 23 Borradori L et al. The localization of bullous pemphigoid antigen 1 Lin AN, Carter DM. Epidermolysis bullosa. Annu Rev Med 1993; 180 (BP180) in hemidesmosomes is mediated by its cytoplasmic 44: 189–199. domain and seems to be regulated by the beta4 integrin subunit. 2 Uitto J, Pulkkinen L, McLean WH. Epidermolysis bullosa: a J Cell Biol 1997; 136: 1333–1347. spectrum of clinical phenotypes explained by molecular hetero- 24 Gagnoux-Palacios L et al. Functional re-expression of laminin-5 geneity. Mol Med Today 1997; 3: 457–465. in laminin-gamma2-deficient human keratinocytes modifies cell 3 Paller AS. The genetic basis of hereditary blistering disorders. morphology, motility, and adhesion. J Biol Chem 1996; 271: Curr Opin Pediatr 1996; 8: 367–371. 18437–18444. 4 Uitto J, Pulkkinen L, Christiano AM. Molecular basis of the dys- 25 Medalie DA et al. Evaluation of human skin reconstituted from trophic and junctional forms of epidermolysis bullosa: composite grafts of cultured keratinocytes and human acellular mutations in the type VII collagen and kalinin (laminin 5) genes. dermis transplanted to athymic mice. J Invest Dermatol 1996; 107: J Invest Dermatol 1994; 103: 39S–46S. 121–127. 5 Marinkovich MP. The molecular genetics of basement membrane diseases. Arch Dermatol 1993; 129: 1557–1565. 26 Choate KA, Medalie DA, Morgan JR, Khavari PA. Corrective 6 Coulombe PA et al. Point mutations in human keratin 14 genes gene transfer in the human skin disorder lamellar ichthyosis. of epidermolysis bullosa simplex patients: genetic and func- Nat Med 1996; 2: 1263–1267. tional analyses. Cell 1991; 66: 1301–1311. 27 Lever WF, Schaumberg-Lever G. Noninfectious vesicular and 7 Bonifas JM, Rothman AL, Epstein EH Jr. Epidermolysis bullosa bullous diseases. In: Histopathology of the Skin. Lippincott: simplex: evidence in two families for keratin gene abnormalities Philadelphia, 1990, pp 622–634. (see comments). Science 1991; 254: 1202–1205. 28 Fan H, Oro AE, Scott MP, Khavari PA. Induction of basal cell 8 Smith FJ et al. Plectin deficiency results in muscular dystrophy carcinoma features in transgenic human skin expressing sonic with epidermolysis bullosa. Nat Genet 1996; 13: 450–457. hedgehog. Nature Med 1997; 3: 788–792. 9 McGrath JA et al. Mutations in the 180-kD bullous pemphigoid 29 Marinkovich MP et al. Basement membrane proteins kalinin and antigen (BPAG2), a hemidesmosomal transmembrane collagen nicein are structurally and immunologically identical. Lab Invest (COL17A1), in generalized atrophic benign epidermolysis bul- 1993; 69: 295–299. losa. Nat Genet 1995; 11: 83–86. 30 Deng H, Lin Q, Khavari PA. Sustainable cutaneous gene deliv- 10 Aberdam D et al. Herlitz’s junctional epidermolysis bullosa is ery. Nat Biotechnol 1997; 15: 1388–1391. BP180 gene delivery CS Seitz et al 47 31 Freiberg RA et al. A model of corrective gene transfer in X-linked 37 Pohla-Gubo G et al. Diminished expression of the extracellular ichthyosis. Hum Molec Genet 1997; 6: 937–943. domain of bullous pemphigoid antigen 2 (BPAG2) in the epider- 32 Fenjves ES, Yao SN, Kurachi K, Taichman LB. Loss of expression mal basement membrane of patients with generalized atrophic of a retrovirus-transduced gene in human keratinocytes. J Invest benign epidermolysis bullosa. Exp Dermatol 1995; 4: 199–206. Dermatol 1996; 106: 576–578. 38 Fairley JA et al. Expression pattern of the bullous pemphigoid- 33 Gerrard AJ, Hudson DL, Brownlee GG, Watt FM. Towards gene 180 antigen in normal and neoplastic epithelia. Br J Dermatol therapy for haemophilia B using primary human keratinocytes. 1995; 133: 385–391. Nat Genet 1993; 3: 180–183. 39 Khavari PA et al. BRG1 contains a conserved domain of the 34 Rheinwald JG, Green H. Serial cultivation of strains of human SWI2/SNF2 family necessary for normal mitotic growth and epidermal keratinocytes: the formation of keratinizing colonies transcription. Nature 1993; 366: 170–174. from single cells. Cell 1975; 6: 331–343. 40 Marinkovich MP, Lunstrum GP, Burgeson RE. The anchoring 35 Choate KA et al. Transglutaminase 1 delivery to lamellar ichthy- filament protein kalinin is synthesized and secreted as a high osis keratinocytes. Hum Gene Ther 1996; 7: 2247–2253. molecular weight precursor. J Biol Chem 1992; 267: 17900–17906. 36 Kinsella TM, Nolan GP. Episomal vectors rapidly and stably produce high-titer recombinant retrovirus. Hum Gene Ther 1996; 7: 1405–1413.