Portal de Periódicos da CAPES

Mouse Models in Preclinical Studies for Pachyonychia Congenita

2005; Elsevier BV; Volume: 10; Issue: 1 Linguagem: Inglês

10.1111/j.1087-0024.2005.10206.x

1529-1774

Chen Jiang, Dennis R. Roop,

Wnt/β-catenin signaling in development and cancer

The similarities between the human and mouse genomes often allow researchers to make accurate predictions about the roles of their human counterparts. Because of the similar physiology between these two mammals, mice are used extensively in the laboratory to investigate the mechanisms of human diseases. Furthermore, mice provide us with the option of testing the toxicity of drugs and the safety of therapeutic approaches prior to human application. Here, we review the existing mouse models involving the keratin genes (K6a, K6b, K16, and K17) that cause the human genetic disorder pachyonychia congenita (PC). We also suggest methods to more accurately model this autosomal dominant skin condition in the mouse in order to better understand the pathophysiological processes underlying PC and importantly, provide a test-bed for testing emerging therapies in vivo. The similarities between the human and mouse genomes often allow researchers to make accurate predictions about the roles of their human counterparts. Because of the similar physiology between these two mammals, mice are used extensively in the laboratory to investigate the mechanisms of human diseases. Furthermore, mice provide us with the option of testing the toxicity of drugs and the safety of therapeutic approaches prior to human application. Here, we review the existing mouse models involving the keratin genes (K6a, K6b, K16, and K17) that cause the human genetic disorder pachyonychia congenita (PC). We also suggest methods to more accurately model this autosomal dominant skin condition in the mouse in order to better understand the pathophysiological processes underlying PC and importantly, provide a test-bed for testing emerging therapies in vivo. epidermolysis bullosa simplex epidermolytic hyperkeratosis pachyonychia congenita Although their exact physiology may differ, mice are still believed to be one of the most reliable models for mimicking human diseases. Genetically engineered mouse models mimic most human disease phenotypes faithfully. In some instances, however, they may reproduce only certain aspects of the human disease, may be more severely affected than the human cases, or may have no clinical phenotype at all. Nevertheless, in many cases, mouse models of human diseases have allowed researchers to uncover the progression and underlying mechanisms of the disease process. Genetically engineered mice include transgenic mice, knockout mice, and knock-in mice. Figure 1 illustrates how these mouse models are generated. Genetically engineered mouse models have provided researchers with valuable in vivo tools to study skin cancer, skin development, and inherited skin diseases. During the last decade, substantial progress has been made in determining the genetic basis of several inherited skin diseases (Lane and McLean, 2004Lane E.B. McLean W.H.I. Keratins and skin disorders.J Pathol. 2004; 204: 355-366Crossref PubMed Scopus (162) Google Scholar). The success of these efforts now allows investigators to rationally design therapeutic strategies for these diseases. In this regard, advances in vector design and the development of efficient, safe, and specific delivery systems, have made it possible to contemplate the use of gene therapy approaches for correcting inherited skin diseases. Prior to testing gene therapy approaches in human patients, one would ideally like to test the safety and efficacy of these approaches in an animal model that mimics the human disease both genetically and clinically. This review will briefly describe all existing mouse models that have been generated with genetic alterations in genes that have been shown to cause pachyonychia congenita (PC) in humans. Currently, there are no mouse models that accurately reproduce the clinical phenotypes observed in PC patients. We will attempt to explain why previous mouse models have failed to exhibit PC phenotypes. Finally, we will outline a strategy for generating what we believe would be an ideal mouse model for PC. We anticipate that this mouse model will be a valuable resource to other investigators in their attempts to develop and test therapeutic approaches for PC. As described byLeachman et al., 2005Leachman S.A. Kaspar R.L. Flechman P. et al.Clinical and pathological features of pachyonychia congenita.J Investig Dermatol. 2005; 10: 3-17Abstract Full Text Full Text PDF Scopus (139) Google Scholaret al (in this issue), the clinical features of PC include various ectodermal abnormalities characterized most distinctly by the presence of hypertrophic nail dystrophy (nail bed hyperkeratosis with nail plate thickening and deformation) and focal keratoderma. Briefly, dominant-negative mutations in keratins 6a or 16 (K6a, K16) cause PC-1, whereas similar mutations in K6b or K17 are associated with the PC-2 phenotype. Ideally, a PC mouse model should be able to represent the clinical manifestation and genetic changes in PC. To date, there are seven mouse models carrying genetic modifications of PC-related genes. The characteristics of these mouse models are summarized in Table I.Table ICurrent mouse models with genetic alterations in genes implicated in causing PCGeneType of mouse modelGenetic modificationPhenotypeReferencesK6aTransgenicDeletion of 52 amino acids (residue 125–176) between head and 1A helix domainInterepidermal blisterTakahashi and Coulombe, 1996Takahashi K. Coulombe P.A. A transgenic mouse model with an inducible skin blistering disease phenotype.Proc Natl Acad Sci USA. 1996; 93: 14776-14781Crossref PubMed Scopus (33) Google ScholarK6aTransgenicTruncation deleting the 2B regionLethal blister or alopeciaWojcik et al., 1999Wojcik S.M. Imakado S. Seki T. et al.Expression of MK6a dominant-negative and C-terminal mutant transgenes in mice has distinct phenotypic consequences in the epidermis and hair follicle.Differentiation. 1999; 65: 97-112Crossref PubMed Scopus (22) Google ScholarK6aTransgenicReplacement of E2 by HK1-tagHyperkeratosis and late-onset alopeciaWojcik et al., 1999Wojcik S.M. Imakado S. Seki T. et al.Expression of MK6a dominant-negative and C-terminal mutant transgenes in mice has distinct phenotypic consequences in the epidermis and hair follicle.Differentiation. 1999; 65: 97-112Crossref PubMed Scopus (22) Google ScholarK6aKnockoutDeletion of K6a locusDelay of reepithelizationWojcik et al., 2000Wojcik S.M. Bundman D.S. Roop D.R. Delayed wound healing in keratin 6a knockout mice.Mol Cell Biol. 2000; 20: 5248-5255Crossref PubMed Scopus (116) Google ScholarK6a/bKnockoutDeletion of K6a and K6b locusBlistering in oral mucosaWojcik et al., 2000Wojcik S.M. Bundman D.S. Roop D.R. Delayed wound healing in keratin 6a knockout mice.Mol Cell Biol. 2000; 20: 5248-5255Crossref PubMed Scopus (116) Google ScholarK6a/bKnockoutDeletion of K6a and K6b locusHyperkeratotic plaque formation on the tongueWojcik et al., 2001Wojcik S.M. Longley M.A. Roop D.R. Discovery of a novel murine keratin 6 (K6) isoform explains the absence of hair and nail defects in mice deficient for K6a and K6b.J Cell Biol. 2001; 154: 619-630Crossref PubMed Scopus (82) Google ScholarK6hfKnock-inPoint mutation of codon N158 (corresponding to mutation of N171 in PC patient)Defects in hair shaft and fragile nailsUnpublished dataK17KnockoutDeletion of K17 locusAge and strain specific alopeciaMcGowan et al., 2002McGowan K.M. Tong X. Colucci-Guyon E. Langa F. Babinet C. Coulombe P.A. Keratin 17 null mice exhibit age- and strain-dependent alopecia.Genes Dev. 2002; 16: 1412-1422Crossref PubMed Scopus (93) Google ScholarPC, pachyonychia congenita Open table in a new tab PC, pachyonychia congenita The first mouse model engineered to express a mutant form of K6a was established by the laboratory of Pierre Coulombe (Takahashi and Coulombe, 1996Takahashi K. Coulombe P.A. A transgenic mouse model with an inducible skin blistering disease phenotype.Proc Natl Acad Sci USA. 1996; 93: 14776-14781Crossref PubMed Scopus (33) Google Scholar). The Coulombe laboratory created a transgenic mouse with an internal deletion of 52 amino acids between the head and 1A helix domain of human K6a (for details of the functional domains of keratins, seeSmith et al., 2005Smith F.J.D. Liao H. Cassidy A.J. et al.The genetic basis of pachyonychia congenita.J Invest Dermatol. 2005; 10: 21-30Abstract Full Text Full Text PDF Scopus (83) Google Scholaret al, in this issue) followed by a c-myc tag, designated as ΔK6a-myc. The deleted region includes N171, a mutational “hot spot” for PC-1 (Lin et al., 1999Lin M.T. Levy M.L. Bowden P.E. Magro C. Baden L. Baden H.P. Roop D.R. Identification of sporadic mutations in the helix initiation motif of keratin 6 in two pachyonychia congenita patients: Further evidence for a mutational hot spot.Exp Dermatol. 1999; 8: 115-119Crossref PubMed Scopus (15) Google Scholar). Expression of the transgene was achieved by topical application of the phorbol ester, TPA. Although the mutant K6a protein produced from the ΔK6a-myc transgene acts in a dominant-negative fashion and interferes with the formation of 10 nm (intermediate) filaments, it does not resemble any known mutation found in PC patients, as it is a major deletion (52 amino acids). In addition, the expression characteristics of the transgene were not comparable with that of the endogenous MK6a gene because the latter is artificially modified to be overexpressed. As a consequence, this mouse model did not exhibit a naturally occurring PC phenotype. Instead, intraepidermal blisters were only observed after TPA-treated skin was subjected to tape stripping. The lack of a characteristic PC phenotype in this mouse model was most likely because of the fact that a transgenic approach was used in its generation, not a knock-in approach. Our laboratory also generated transgenic mice expressing a mutant form of MK6a that was truncated at the beginning of the 2B region of the rod domain (Δ2B-P) (Wojcik et al., 1999Wojcik S.M. Imakado S. Seki T. et al.Expression of MK6a dominant-negative and C-terminal mutant transgenes in mice has distinct phenotypic consequences in the epidermis and hair follicle.Differentiation. 1999; 65: 97-112Crossref PubMed Scopus (22) Google Scholar). Some of these mice exhibited a severe and lethal skin blistering phenotype. Others, expressing low levels of the truncated MK6a transgene, survived to adulthood without blisters but developed severe alopecia. Similar hair loss was also seen in animals that expressed another mutant K6a transgene in which the E2 domain of MK6a was replaced by the C-terminus of human K1, which encodes an epitope tag (ΔE2-HK1). Mice expressing this transgene developed hyperkeratotic skin lesions and alopecia beginning around 8–10 wk of age (Wojcik et al., 1999Wojcik S.M. Imakado S. Seki T. et al.Expression of MK6a dominant-negative and C-terminal mutant transgenes in mice has distinct phenotypic consequences in the epidermis and hair follicle.Differentiation. 1999; 65: 97-112Crossref PubMed Scopus (22) Google Scholar). Despite the fact that both of these transgenic lines expressing different mutant forms of MK6a developed alopecia and hyperkeratosis in some regions of the skin, neither developed hyperkeratosis of the paws or the oral cavity, or hypertrophic nails as would be expected for PC-1. This is most likely because of the fact that MK6a transgenes used in these experiments were not efficiently expressed in the oral cavity, foot pads, or nail beds. To further investigate the function of K6a, our laboratory also deleted this gene from the mouse genome (Wojcik et al., 2000Wojcik S.M. Bundman D.S. Roop D.R. Delayed wound healing in keratin 6a knockout mice.Mol Cell Biol. 2000; 20: 5248-5255Crossref PubMed Scopus (116) Google Scholar). Surprisingly, the K6a-null mice did not exhibit apparent structural defects in the skin and its appendages. Also, the healing of full thickness skin wounds in these mice was normal compared with WT littermates. However, a delay in re-epithelialization following tape stripping to remove all layers of the epidermis, was observed. As the phenotypic consequences of complete lack of expression of MK6a are surprisingly minimal in the mouse, these observations are potentially relevant to therapeutic strategies that might be contemplated for PC patients. It may be difficult to develop strategies that will specifically eliminate expression of the mutant K6a allele. Alternative strategies designed to eliminate expression of both the wild-type and mutant K6a alleles are much more feasible. Importantly, the subtle phenotype of the MK6a knockout mice suggests that this strategy may have minimal side effects if attempted in humans, i.e., if it were possible to specifically ablate or substantially reduce K6a expression using small interfering RNA (siRNA) or a similar method, without affecting K6b expression, this might well ameliorate the phenotype of PC patients carrying K6a mutations. To determine the consequences of loss of expression of both the K6a and K6b genes, the Coulombe laboratory deleted both genes from the mouse germline. They discovered that K6a/b-null mice have changes in the oral mucosa resembling those of PC-1 patients (Wong et al., 2000Wong P. Colucci-Guyon E. Takahashi K. Gu C. Babinet C. Coulombe P.A. Introducing a null mutation in the mouse K6alpha and K6beta genes reveals their essential structural role in the oral mucosa.J Cell Biol. 2000; 150: 921-928Crossref PubMed Scopus (75) Google Scholar). This mouse model, however, lacked obvious alterations in nail morphology, a conspicuous feature in PC, as determined by histological analysis of cross-sections of the nail tissues of 5–10 d old pups (Wong et al., 2000Wong P. Colucci-Guyon E. Takahashi K. Gu C. Babinet C. Coulombe P.A. Introducing a null mutation in the mouse K6alpha and K6beta genes reveals their essential structural role in the oral mucosa.J Cell Biol. 2000; 150: 921-928Crossref PubMed Scopus (75) Google Scholar). This was most likely because of the fact that the K6a/b-null mice are representative of a recessive disorder; i.e., the mice do not express a dominant-negative interfering mutant keratin protein as found in all of the human forms of PC. Our laboratory also produced K6a/b-null mice. However, unlike the K6a/b-null mice, however, generated by Wong et al which died postnatally, approximately 25% of our K6a/b-null mice survived to adulthood (Wojcik et al., 2001Wojcik S.M. Longley M.A. Roop D.R. Discovery of a novel murine keratin 6 (K6) isoform explains the absence of hair and nail defects in mice deficient for K6a and K6b.J Cell Biol. 2001; 154: 619-630Crossref PubMed Scopus (82) Google Scholar). Mice which failed to survive exhibited oral lesions at the back of the tongue, similar to those observed by Wong et al. Surprisingly, the surviving double knockout mice did not show any abnormalities of the hair follicles or nails. This prompted the search for other K6 genes that might be expressed in the nail bed and hair follicles, and compensated for the loss of K6a/b. Immunofluorescent staining of surviving K6a/b-null mice with an antibody raised against the C-terminus of MK6b showed cross-reactivity in the hair follicle and nail bed. Therefore, the K6a/b-null mice led to the discovery of a third isoform of the MK6 gene family, MK6hf (Wojcik et al., 2001Wojcik S.M. Longley M.A. Roop D.R. Discovery of a novel murine keratin 6 (K6) isoform explains the absence of hair and nail defects in mice deficient for K6a and K6b.J Cell Biol. 2001; 154: 619-630Crossref PubMed Scopus (82) Google Scholar). MK6hf was named because of its high homology and similar expression patterns to that of HK6hf. Based on the knowledge gained from previous studies on K6a-null and K6a/b-null mice, we hypothesized that the loss of function of the MK6hf by deleting the gene from the mouse genome may not exhibit phenotypes because of possible compensation by K6a and/or K6b. Among the known mutations in the human K6 genes which have been reported to cause PC, deletion of codon 171 of K6a, encoding Asparagine (N), is considered to be a mutational “hot spot.” Therefore, in an attempt to generate a PC phenotype, we introduced this mutation into the corresponding residue of the MK6hf gene by deleting its codon 158. Transfection of this dominant mutant complementary DNA (cDNA) into PtK2 cells showed that the endogenous keratin filament network was disrupted, suggesting a dominant-negative effect of the mutation (unpublished data). Therefore, we proceeded to develop a mouse model in which one wild-type K6hf allele was replaced with the K6hf Δ158 allele. These mutant mice displayed a thin coat of fur compared with WT littermates. The hair phenotype is more evident on the ventral side of the body. Microscopically, blebs were observed along the hair shafts of a small percentage of the hair. Some hair shafts have more than one bleb. This focal keratinization defect (blebs) along the hair shaft may cause hair fragility and explain the sparse coat of the mutant mice. In addition to the hair phenotype, nails of the mutant mice also appear to be fragile and are easily lost, probably because of fragility of keratinocytes in the nail bed (unpublished data). The nail phenotype is frequently observed in mutant adult male mice which are prone to fighting when housed together. To date, this is the only genetically engineered mouse model to exhibit clinical phenotypes similar to those observed in PC patients. The restriction of the phenotypes to the hair and nails is consistent with the expression pattern of the endogenous K6hf gene. Thus, this mouse model does not actually mimic PC-1 or PC-2. Instead, these mice suggest that PC patients who do not exhibit classical features of PC-1 or PC-2, i.e., only their nails and hair follicles show clinical involvement, may actually have mutations in HK6hf. Several K17 mutations have been reported in PC-2 and recent data suggest a link between the late-onset of PC and K17 (seeSmith et al., 2005Smith F.J.D. Liao H. Cassidy A.J. et al.The genetic basis of pachyonychia congenita.J Invest Dermatol. 2005; 10: 21-30Abstract Full Text Full Text PDF Scopus (83) Google Scholaret al, in this issue). With respect to mouse models that have been generated with genetic alterations in K17, K17-null mice exhibit an age and strain specific alopecia (McGowan et al., 2002McGowan K.M. Tong X. Colucci-Guyon E. Langa F. Babinet C. Coulombe P.A. Keratin 17 null mice exhibit age- and strain-dependent alopecia.Genes Dev. 2002; 16: 1412-1422Crossref PubMed Scopus (93) Google Scholar). But, no obvious abnormalities were observed in the nails, as observed for PC-2 patients. The lack of the classical phenotypic features of PC-2 in this mouse model is most easily explained by the fact that these mice were not engineered to express a dominant-negative mutant form of K17 as observed in human patients. The recently identified K17n, whose expression is highly restricted in the nails of mice, may also be a contributing factor in the absence of PC phenotypes (Tong and Coulombe, 2004Tong X. Coulombe P.A. A novel mouse type I intermediate filament gene, keratin 17n (K17n), exhibits preferred expression in nail tissue.J Invest Dermatol. 2004; 122: 965-970Abstract Full Text Full Text PDF PubMed Scopus (24) Google Scholar). The spatial expression of keratin genes is tightly regulated. These genes are constitutively expressed or induced in designated layers of the epidermis. Therefore, studies on other keratin genes that do not necessarily cause PC may provide meaningful insights into the pathological course of PC. In addition, advances in embryonic stem (ES) cell technology have allowed the generation of mouse models that mimic many human diseases. But, it is not possible to establish viable mouse models for all inherited diseases. For example, newborn mice genetically engineered to reproduce inherited forms of skin fragility syndromes often die a few days after birth because of uncontrolled water loss and/or severe infections. We think that there are two other mouse models of inherited skin diseases that are relevant to PC. To overcome the problem associated with the lethality caused by skin fragility syndromes in establishing mouse models, our laboratory has recently utilized an inducible system that allows the activation of mutations in a restricted area of the skin. Using this system, we were able to develop viable, inducible mouse models that mimic epidermolysis bullosa simplex (EBS) (Cao et al., 2001Cao T. Longley M.A. Wang X.J. Roop D.R. An inducible mouse model for epidermolysis bullosa simplex: Implications for gene therapy.J Cell Biol. 2001; 152: 651-656Crossref PubMed Scopus (122) Google Scholar) and epidermolytic hyperkeratosis (EHK) (Arin et al., 2001Arin M.J. Longley M.A. Wang X.J. Roop D.R. Focal activation of a mutant allele defines the role of stem cells in mosaic skin disorders.J Cell Biol. 2001; 152: 645-649Crossref PubMed Scopus (64) Google Scholar) at both the genetic and phenotypic levels. These mouse models have provided new insights into the molecular and cellular basis of EHK and EBS and have important implications for gene therapy approaches for these diseases. As PC exhibits clinical features of both EHK and EBS, we provide a brief description of these mouse models and explain how they might be relevant to PC. The clinical features of PC-1 are in many respects similar to EHK, because under constitutive conditions, both K10, whose mutation is responsible for EHK, and K6a are expressed primarily in the differentiated layers of the affected epithelia. Therefore, it is beneficial to describe the transgenic mouse model that mimics EHK (Arin et al., 2001Arin M.J. Longley M.A. Wang X.J. Roop D.R. Focal activation of a mutant allele defines the role of stem cells in mosaic skin disorders.J Cell Biol. 2001; 152: 645-649Crossref PubMed Scopus (64) Google Scholar). This mouse model differs from the human disease in that expression of the mutant K10 allele, which contains an Arg 154 Cys mutation equivalent to the Arg 156 Cys mutation found in ∼50% of patients with the severe form of EHK (Arin et al., 2001Arin M.J. Longley M.A. Wang X.J. Roop D.R. Focal activation of a mutant allele defines the role of stem cells in mosaic skin disorders.J Cell Biol. 2001; 152: 645-649Crossref PubMed Scopus (64) Google Scholar), can be restricted to a small area of the skin. To develop an inducible model for EHK, a recently developed transgenic mouse model that allows the focal deletion of genomic sequences upon topical application of an inducer was exploited (Berton et al., 2000Berton T.R. Wang X.J. Zhou Z. Kellendonk C. Schutz G. Tsai S. Roop D.R. Characterization of an inducible, epidermal-specific knockout system: Differential expression of lacZ in different Cre reporter mouse strains.Genesis. 2000; 26: 160-161Crossref PubMed Scopus (43) Google Scholar). This system utilizes tissue-specific expression of the Cre recombinase fused to a truncated progesterone receptor (PR1) under the control of a K14 promoter (K14-CrePR1) Figure 2. The Cre recombinase activity in this fusion protein is sequestered in the cytoplasm until the progesterone antagonist, RU486, induces translocation of the fusion protein into the nucleus (Kellendonk et al., 1996Kellendonk C. Tronche F. Monaghan A.P. Angrand P.O. Stewart F. Schutz G. Regulation of Cre recombinase activity by the synthetic steroid RU 486.Nucleic Acids Res. 1996; 24: 1404-1411Crossref PubMed Scopus (210) Google Scholar). Mice carrying the K14-CrePR1 transgene were bred with heterozygous mtK10neo mice. RU486 was topically applied once daily to the forelimbs and chest. After 4 d of RU486 treatment, blisters were observed on the forelimbs and chest of bigenic pups, whereas they were never observed on the untreated areas or on treated monogenic or wild-type pups Figure 3. Although treatment with RU486 was terminated after blister formation was observed, persistent focal blistering and hyperkeratosis was observed throughout the lifespan of treated mice. The hyperkeratosis of the paw of the mouse resembles palmoplantar hyperkeratosis observed in PC-1 patients.Figure 3Focal induction of blisters in the inducible epidermolytic hyperkeratosis (EHK) mouse model. The left panel shows blister formation on the forelimbs and chest 4 d after treatment. Unlike epidermolysis bullosa simplex (EBS)-induced blisters, EHK blisters persist after treatment with the activator is stopped. Shown in the right panel is a mouse at 3 mo following blister induction. But, persistent blistering has now been observed for over 1 y.View Large Image Figure ViewerDownload (PPT) The data summarized above suggests that once the mutant K10 allele is activated in epidermal stem cells, these mutant stem cells persist for the life of the mouse. Further support for this hypothesis and for the fact that EBS stem cells exhibit very different properties from EHK stem cells was obtained when the same inducible system was used to develop a mouse model for EBS (Cao et al., 2001Cao T. Longley M.A. Wang X.J. Roop D.R. An inducible mouse model for epidermolysis bullosa simplex: Implications for gene therapy.J Cell Biol. 2001; 152: 651-656Crossref PubMed Scopus (122) Google Scholar). EBS is very similar to EHK, except that blisters develop in the basal keratinocytes of the epidermis because of a dominant mutation in either K5 or K14. In this respect, PC-1 could also be similar to EBS, if the HK6a gene is induced in keratinocytes in the basal layer in response to injury or blistering. Unlike EHK lesions, which persist for the life of the mouse, EBS blisters healed within 2 wk after induction and never reappeared Figure 4 (Cao et al., 2001Cao T. Longley M.A. Wang X.J. Roop D.R. An inducible mouse model for epidermolysis bullosa simplex: Implications for gene therapy.J Cell Biol. 2001; 152: 651-656Crossref PubMed Scopus (122) Google Scholar). Thus, activation of the mutant K14 allele in epidermal stem cells results in their rapid displacement whereas EHK stem cells fail to exhibit a growth disadvantage. The phenotypes exhibited by these inducible models provide an explanation for the existence of mosaic forms of EHK, but not EBS. The schematic shown in Figure 5 illustrates what has been observed in these inducible models. In the inducible EBS model, focal activation of the mutant K14 allele occurs in stem cells, resulting in fragility. With time, neighboring stem cells (where the neomycin-resistance (neo)-cassette is still in place) migrate into the area and displace the defective EBS stem cells. In the inducible EHK model, removal of the neo-cassette from the mutant K10 gene also occurs in stem cells. The mutant K10 protein, however, is not expressed in stem cells, but only in the progeny of stem cells after they have differentiated and moved into the suprabasal layers. Therefore, there is no adverse selection against epidermal stem cells with K10 mutations, and these stem cells continue to give rise to defective differentiated progeny for the life of the mouse. Our models suggest that gene therapy approaches will be very different for EHK and EBS. Whereas repopulation of a phenotypic area by corrected stem cells is predicted to occur in the case of EBS (Cao et al., 2001Cao T. Longley M.A. Wang X.J. Roop D.R. An inducible mouse model for epidermolysis bullosa simplex: Implications for gene therapy.J Cell Biol. 2001; 152: 651-656Crossref PubMed Scopus (122) Google Scholar), successful approaches for EHK will have to include a strategy that allows selection of genetically corrected epidermal stem cells and ablation of defective EHK stem cells. In humans, PC resembles the persistent hyperkeratosis phenotype as found in EHK. Therefore, the EHK mouse model suggested that selective ablation of defective PC stem cells may be indispensable to reverse the phenotype. During the development of these mouse models, we made an unexpected observation that has additional implications for the development of therapeutic approaches for these diseases. As illustrated in Figure 2, the targeting constructs used to make the inducible mouse models for EBS and EHK contain a neo-cassette, flanked by loxP sites, in the first intron. We discovered that heterozygous mice which retain the neo-cassette, either mtK10neo or mtK14neo express these mutant alleles at levels that are ∼35%–40% (mtK10) (Arin et al., 2001Arin M.J. Longley M.A. Wang X.J. Roop D.R. Focal activation of a mutant allele defines the role of stem cells in mosaic skin disorders.J Cell Biol. 2001; 152: 645-649Crossref PubMed Scopus (64) Google Scholar) or 50% (mtK14) (Cao et al., 2001Cao T. Longley M.A. Wang X.J. Roop D.R. An inducible mouse model for epidermolysis bullosa simplex: Implications for gene therapy.J Cell Biol. 2001; 152: 651-656Crossref PubMed Scopus (122) Google Scholar) of the wild-type alleles. As mtK10neo and mtK14neo mice exhibited either very subtle or no detectable phenotypes, respectively, therapeutic strategies designed to increase the ratio of wild-type K10/K14 to mutant K10/K14 to approximately 2–1 might result in sub-clinical disease. Thus, in contrast to the general assumption that gene therapy approaches for dominant diseases must either correct the mutant allele or completely inhibit its expression, our mouse models predict that amelioration of the EHK and EBS phenotypes may be achieved by partial suppression of the mutant keratin alleles or over-expression of the wild-type alleles. This may also be true for PC. Therefore, establishing a PC mouse model is essential to prove the usefulness of hypotheses drawn from the above relevant mouse models. In addition to providing insight into the pathophysiological mechanisms of PC, it is anticipated that the PC mouse model will provide investigators with an animal model suitable for performing preclinical studies to address issues such as safety and efficacy that will be required to obtain approval for clinical trials in humans. Therefore, the ideal mouse model should mimic PC at both the genetic and phenotypic levels. Furthermore, as certain therapeutic approaches for treating PC-1 patients may utilize strategies that are very selective for nucleotide sequences in the human K6a gene, the PC mouse model should be engineered to contain a mutant human K6a allele, i.e., a “humanized” mouse PC model. As discussed above, it is also imperative that a knock-in strategy be used in the development of the PC mouse model, as this will ensure that the mutant allele is expressed at the same level as the wild-type allele and in the appropriate tissues. As summarized in Figure 1, the use of traditional transgenic approaches results in random integration of the transgene into the mouse genome and increases the likelihood that the transgene will not be able to accurately mimic the expression levels and patterns of endogenous genes. An additional consideration in designing the PC mouse model is the potential deleterious or even lethal effect of constitutive expression of a mutant K6a allele. As discussed above, previous studies demonstrated that constitutive overexpression of a dominant-negative mutant form of K6a resulted in severe hair loss and skin fragility that ultimately resulted in death. Therefore, we believe that it would be prudent to utilize the same inducible system described above for the EBS and EHK mouse models in the generation of the PC mouse model. This would permit the induced expression of the mutant K6a allele in restricted focal areas of the skin and ensure that viability of the mouse model is maintained. An inducible mouse model for PC-1 that contains sequences identical to human K6a around the N171K mutation. As discussed above, this proposed mouse model should be viable, and mimic PC-1 at both genetic and phenotypic levels. To accomplish this aim, an inducible mouse model should be established in which the mutational region of the MK6a is replaced with the corresponding sequences found in mutant HK6a. The strategy to accomplish this goal is outlined in Figure 6. Based on a blast search of homologous regions between HK6a and MK6a, it would be technically easiest and scientifically rational to replace a region between codon 120 and 168 in exon 1 of MK6a with the corresponding sequence of HK6a. As outlined in Figure 6, an 8 kb genomic sequence of MK6a is cloned into a plasmid that has a PGK-TK cassette. The partial replacement of the MK6a gene and introduction of the mutation can be engineered simultaneously. Ideally, prior to insertion of the neo-cassette in intron 1 of MK6a Figure 6, the “humanized” mutant MK6a constructs can be transfected into cell lines to confirm that the mutant K6a protein is expressed and that it does cause collapse of the endogenous K18 network. The final step is to introduce the neo-cassette (Figure 6b). The ideal neo-cassette for inducible gene targeting has been engineered in this laboratory. It is flanked by loxP sites. It also contains multiple polyadenylation sequences (polyA signals). The neo-cassette not only facilitates the identification of successfully targeted ES cell clones, but it also prevents transcriptional read-through, i.e., it acts as a STOP sequence to prevent expression of the mutant allele. When the neo-cassette is removed by Cre-mediated recombination via the loxP sites, normal transcription is resumed. Thus, induction of the mutant allele is achieved Figure 7. After linearization, the targeting vector will be electroporated into ES cells. FIAU and G418 resistant ES cell colonies will be screened by Southern blot with the designated 5′ probe and 3′ probe, as well as the neo-cassette probe, as outlined in Figure 7. Successfully targeted ES cell clones will be injected into C57BL/6 blastocysts to obtain chimeric pups. Once chimeric pups are born, they will be backcrossed into the C57BL/6 background to confirm germline transmission. When germline transmission is confirmed, heterozygous mice will be mated with mice expressing the inducible Cre recombinase. Heterozygous mice will be mated with our K5-CrePR1 mice (Zhou et al., 2002Zhou Z. Wang D. Wang X.J. Roop D.R. In utero activation of K5: CrePR1 induces gene deletion.Genesis. 2002; 32: 191-192Crossref PubMed Scopus (30) Google Scholar). Double transgenic mice will be treated topically with RU486 to confirm focal activation of the mutant “humanized” K6a allele and induction of PC-1 phenotypes. Although the inducible mouse model gives us the advantage of producing viable lines, it is also important to observe the constitutive effects of the dominant-negative mutation in mice. Therefore, the neo-cassette from successfully targeted ES cell clones should be excised by transfecting a Cre-expression plasmid. Cre-mediated recombination will excise the neo-cassette located between the two loxP sites. These ES cell clones will be transferred into blastocysts to produce a constitutive PC mouse model. The animal models described above should mimic PC-1 at both genetic and phenotypic levels. The inducible PC-1 mouse model also ensures that the line will be viable, in case constitutive expression of the mutant K6a causes death. The “humanized” mutant sequence also provides researchers with an ideal preclinical model to assess therapeutic strategies that may eventually be used in humans. An inducible mouse model for PC-1 in which the entire coding region is replaced with HK6a sequence containing the N171K mutation. It may be possible to selectively inhibit expression of mutant K6a transcripts through the use of siRNAs or ribozymes that recognize sequences around the site of the mutation (seeLewin et al., 2005Lewin A.S. Glazer P.M. Milstone L.M. Gene therapy for autosomal dominant disorders of Keratins.J Invest Dermatol. 2005; 10: 47-61Abstract Full Text Full Text PDF Scopus (31) Google Scholaret al, in this issue). As an alternative approach, however, it may be feasible to completely eliminate expression of both wild-type and mutant K6a alleles. We previously showed that ablation of the MK6a gene from the mouse germline had only minor phenotypic consequences (Wojcik et al., 2000Wojcik S.M. Bundman D.S. Roop D.R. Delayed wound healing in keratin 6a knockout mice.Mol Cell Biol. 2000; 20: 5248-5255Crossref PubMed Scopus (116) Google Scholar). To develop a mouse model to assess this strategy, it is necessary to completely replace the MK6a sequence with that of HK6a. This model will provide numerous target sites to ablate expression of K6, rather than just those sequences around the site of the mutation. A potential limitation of such a mouse model expressing the entire human gene, however, is that the human gene may not function or it may not function in the same way as it does in humans. For these reasons, it is best to keep as many of the potential regulatory regions of the mouse K6a unaltered as possible, such as the 5′ or 3′ untranslated region (UTR) in exon 1 and exon 9. Given the sensitivity and specificity of the above-mentioned gene targeting methods, the coding region of HK6a should provide an ample number of sites that can be tested for specific targeting of HK6a. Strategies used in construction of this targeting vector are essentially the same as described for PC mouse model 1, except that all exons and introns of mouse K6a will be replaced with human exons and introns (excluding the 5′ or 3′UTR in exon 1 and exon 9) Figure 8. As described for PC Mouse Model 1, the targeting vector containing the entire HK6a sequence and N171K mutation will be electroporated in ES cells. Multiple probes (not shown), however, are necessary to identify correctly targeted ES cell clones. This is because of the highly homologous sequence between MK6a and HK6a. Correctly targeted ES cell clones will be injected to blastocysts to obtain chimeras. Also as described in PC Mouse Model 1, chimeras will be backcrossed to obtain heterozygous mice. Inducible and constitutive mouse models that express the mutant HK6a gene can be established as described for PC Mouse Model 1 Figure 8. Replacement of the entire MK6a gene locus is straightforward, however, screening for targeted ES cells may be difficult, as exons 1–6 of HK6a and MK6a are highly homologous. Alignment of these two genes reveals that there is a 72 bp consecutively identical sequence in exon 6. Undesired homologous recombination may happen between these regions during ES cell targeting. Therefore, as mentioned above, additional probes are needed to screen for successfully targeted ES cell clones. Also, the HK6a introns may result in undesired gene splicing and altered regulation of expression. To circumvent this problem, an alternative strategy to target the ES cells may be necessary. The alternative strategy is illustrated in Figure 9. In this strategy, exon 1 (excluding the 5′ UTR) and intron 1 of MK6a will be replaced with that of HK6a, including the N171K mutation. Then, exon 2–9 (including all introns in between but excluding 3′ UTR of exon 9) of MK6a will be replaced with the corresponding cDNA of HK6a Figure 9. This alternative approach will exchange all of the coding sequence of MK6a with the mutant human sequence, but the inclusion of cDNA encoding exons 2–9 will greatly simplify ES cell targeting. The ultimate role of the PC mouse model is to provide investigators with a relevant and reliable source of cells and tissues, as well as whole animals to assess the efficacy of various therapeutic approaches. For example, keratinocytes can be isolated from PC mouse skin and cultured in vitro. These cells can be used to rapidly screen for siRNAs designed to specifically degrade mutant K6a transcripts. Likewise, primary keratinocytes from the PC mouse model can be used to assess the efficacy of other inhibitory molecules as discussed byLewin et al., 2005Lewin A.S. Glazer P.M. Milstone L.M. Gene therapy for autosomal dominant disorders of Keratins.J Invest Dermatol. 2005; 10: 47-61Abstract Full Text Full Text PDF Scopus (31) Google Scholaret al (in this issue). Once effective inhibitory molecules are identified using the in vitro assays, whole animals can be used to assess their effectiveness in vivo. Because of the barrier functions of the skin, delivery of inhibitory molecules into intact skin may prove to be a formidable challenge. In both cases, the mouse model will be invaluable for the in vivo assessment of different routes of delivery of inhibitory molecules. Keratinocytes isolated from the PC mouse model could also be used to screen libraries of small molecules that might up-regulate other keratins that would effectively suppress the mutant K6a phenotype by diluting its concentration relative to the wild-type protein. Again, once small molecules are identified in vitro, their effectiveness and safety can be assessed in vivo in the PC mouse model. We anticipate that a PC mouse model that accurately re-capitulates the human PC phenotype will be a very valuable tool in initial attempts to develop therapeutic approaches for this genetic disorder. We are aware that there are major differences between mouse and human skin, however, such as hair density and skin thickness. Methods of delivering therapeutic molecules into mouse skin may not prove effective for human skin. Therefore, an alternative strategy would be the use of xenografts of skin from PC patients onto immuno-compromised mice. The availability of PC skin biopsies for grafting may limit the use of this model. Therefore, the combined use of the PC mouse model as a first screen to identify promising candidate molecules and the xenograft model as a final proof for efficacy may provide the ideal combination of preclinical models for the assessment of therapeutic approaches for PC.