Mice with a Targeted Mutation of Patched2 Are Viable but Develop Alopecia and Epidermal Hyperplasia

Abstract

Hedgehog (Hh) signaling plays pivotal roles in tissue patterning and development in Drosophila melanogaster and vertebrates. The Patched1 (Ptc1) gene, encoding the Hh receptor, is mutated in nevoid basal cell carcinoma syndrome, a human genetic disorder associated with developmental abnormalities and increased incidences of basal cell carcinoma (BCC) and medulloblastoma (MB). Ptc1 mutations also occur in sporadic forms of BCC and MB. Mutational studies with mice have verified that Ptc1 is a tumor suppressor. We previously identified a second mammalian Patched gene, Ptc2, and demonstrated its distinct expression pattern during embryogenesis, suggesting a unique role in development. Most notably, Ptc2 is expressed in an overlapping pattern with Shh in the epidermal compartment of developing hair follicles and is highly expressed in the developing limb bud, cerebellum, and testis. Here, we describe the generation and phenotypic analysis of Ptc2^tm1/tm1 mice. Our molecular analysis suggests that Ptc2^tm1 likely represents a hypomorphic allele. Despite the dynamic expression of Ptc2 during embryogenesis, Ptc2^tm1/tm1 mice are viable, fertile, and apparently normal. Interestingly, adult Ptc2^tm1/tm1 male animals develop skin lesions consisting of alopecia, ulceration, and epidermal hyperplasia. While functional compensation by Ptc1 might account for the lack of a strong mutant phenotype in Ptc2-deficient mice, our results suggest that normal Ptc2 function is required for adult skin homeostasis.

The Hedgehog (Hh) family of secreted signaling molecules plays critical roles in embryonic patterning and organ development in Drosophila melanogaster and vertebrate species. Many components of the Hh signaling pathway were first identified in Drosophila, and recent studies have revealed both evolutionarily conserved and divergent aspects of the pathway (17, 31, 48, 49). In both Drosophila and vertebrate cells, two transmembrane proteins, Patched (Ptc) and Smoothened (Smo), are involved in the reception of Hh signals. Ptc inhibits the signaling activity of Smo when Hh is absent, and binding of Hh to Ptc relieves the inhibition on Smo, resulting in the transcriptional activation of Hh target genes. Vertebrates possess two Patched genes, Ptc1 and Ptc2. In humans, Ptc1 is a tumor suppressor gene mutated in nevoid basal cell carcinoma syndrome (NBCCS), which is characterized by various developmental anomalies and a predisposition to develop basal cell carcinoma (BCC) and medulloblastoma (MB) (25). Furthermore, Ptc1 mutations are also found in sporadic forms of BCC and MB. Ptc1^+/− mice display a mutant phenotype similar to those observed in NBCCS patients and develop tumors, including MB and skin tumors (5, 24, 27, 28).

Ptc encodes a receptor protein containing 12 hydrophobic membrane-spanning domains, intracellular amino- and carboxy-terminal regions, and two large hydrophilic extracellular loops, where Hh ligand binding occurs (29, 40, 60). Responsiveness of cells to Hh can be abolished by mutating or deleting the extracellular domains of Ptc (11, 47). The Ptc receptor belongs to a family of integral-membrane proteins that characteristically possess a sterol-sensing domain (SSD), which is implicated in vesicle trafficking (35); however, the role of the SSD of Ptc in Hh signaling remains unclear. In Drosophila, Ptc and Hh colocalize to intracellular vesicles in Hh-responding cells, suggesting that the Hh-Ptc complex is internalized upon binding (10, 12, 41, 61). Similarly, in mammalian cells, Sonic hedgehog (Shh) can be internalized by cells transfected with Ptc1, followed by targeting of both proteins to the lysosome (42). These findings support a model of Ptc-mediated endocytosis for regulating the availability of the ligand. Internalization of Ptc and Ptc1 can occur via dynamin-dependent endocytosis mediated by clathrin-coated pits (13, 30). Caveolin, a major coat protein of caveolae, has also been identified as a Ptc-binding partner, suggesting that non-clathrin-coated plasma membrane invaginations (caveolae) may be involved in the endocytosis and trafficking of Ptc (32).

The precise mechanism by which Ptc regulates Smo is unclear. A catalytic model for Ptc function has been proposed (62); the levels of free Ptc (unbound by Hh) determine the degree of pathway activity as well as the amount of Hh ligand required for stimulation of the pathway (62). Ptc shows similarity to the resistance, nodulation, division (RND) family of bacterial proton gradient-driven transmembrane molecular transporters. In bacteria, RND proteins are responsible for removing antibiotics, toxic organic compounds, and metal ions from cells (64). Therefore, like other RND proteins, Ptc might function as a molecular transporter. In addition, the intracellular loop of Ptc has been shown to interact with cyclin B1, resulting in inhibition of cell proliferation by mediating the localization of phosphorylated cyclin B1 (9).

Most studies on Ptc function focus on Ptc1, and very little is known about the role of Ptc2 in development. We have previously identified the mouse Ptc2 gene and shown that it displays overlapping expression with Shh during epidermal development in mouse embryos (46). Mouse Ptc2 is located at mouse chromosome 4C7-D1 (23), which is syntenic to human chromosome 1p36-31.3, where multiple tumor suppressor genes have been mapped (45). Reverse transcription-PCR (RT-PCR) analysis revealed Ptc2 expression in the adult mouse brain, stomach, intestine, kidney, heart, lung, and liver. RNA in situ hybridization analysis also detected Ptc2 expression in primary and secondary spermatocytes of adult mouse testes (14). Similar to Ptc1, Ptc2 also contains 12 transmembrane domains and two large extracellular loops, but, different from Ptc1, Ptc2 possesses much shorter amino- and carboxy-terminal regions.

Both PTCH1 (57, 58) and PTCH2 (52) were shown to generate several splice variants. In PTCH2 alternative splicing appears to affect mostly the C-terminal region and the sterol-sensing domain (52).

Biochemical studies revealed that human PTCH1 and PTCH2 bind to all Hh family members (Sonic hedgehog, Desert hedgehog [Dhh], and Indian hedgehog [Ihh]) with similar affinities (14). Since Ptc2 is highly expressed in the testis, it has been suggested that it acts as the receptor for Dhh, which is required for germ cell development (14). Recently, it was suggested that the Dhh-Ptc2 signaling pathway is involved in the maintenance of adult nerves where Ptc2, but not Ptc1, is expressed in peripheral nerve cells. Interestingly, Ptc2 and Dhh expression was upregulated in regenerating nerves, suggesting that Ptc2 is the major receptor of Dhh in adult nerves (8). In addition, PTCH2 splice variants were able to reconstitute a Dhh-dependent transcriptional response in cells lacking Ptc1 function (52). It was reported that Ptc2 and Ihh are highly expressed in equine osteochondrosis-affected cartilage and repair tissue, suggesting that Ihh-Ptc2 signaling might play a role in diseased adult cartilage (56).

PTCH2 mutations have been detected in some cases of sporadic BCC and MB, suggesting that it might play a role in tumorigenesis (58). Interestingly, like PTCH1, PTCH2 is highly expressed in both familial and sporadic BCC (66) and MB (36), indicating that Ptc2 is a target gene of Shh signaling in the skin and that Ptc2 may be under negative regulation by Ptc1.

Murine hair follicle development begins at embryonic day 14 (E14), when the mesenchyme instructs the overlaying ectoderm to form an epidermal placode. In response the placode signals to mesenchymal fibroblasts to form a dermal condensate. The dermal condensate responds by instructing follicular keratinocytes to proliferate and differentiate into the mature follicle. During follicular development, Shh is required for proliferation: in Shh^−/− embryos hair follicle development is arrested (59), while overexpression of Shh (K14-Shh) results in hyperproliferative basaloid lesions developing at the expense of hair follicles (1, 50). Postnatal follicles cycle through successive phases of anagen (active growth), catagen (regression), and telogen (resting), and Shh acts as a biological switch instructing hair follicles to enter anagen (55).

To determine the role of Ptc2 in mammalian development, we generated a targeted mutant allele, Ptc2^tm1, by homologous recombination in embryonic stem (ES) cells. Our analysis suggests that Ptc2^tm1 is likely a hypomorphic allele of Ptc2. We present here a phenotypic analysis of Ptc2^tm1/tm1 mice. Our results indicate that Ptc2^tm1/tm1 mice develop normally, are viable and fertile, and do not display any obvious defects in hair follicle, limb, neural, or testis development. However, Ptc2^tm1/tm1 male mice develop skin lesions with progressing age consisting of alopecia (hair loss) and epidermal hyperplasia, suggesting a role for Ptc2 in adult epidermal homeostasis.

MATERIALS AND METHODS

Gene targeting of Ptc2.

A Ptc2 clone was isolated from an Sv/129 mouse genomic library for the construction of the targeting vector, which contains a 2.0-kb XhoI-XhoI fragment and a 4.5-kb XhoI-BamHI fragment flanking the phosphoglycerol kinase-neo cassette (Fig. 1A). A herpes simplex virus thymidine kinase gene was inserted next to the longer arm and used for negative selection. The targeting vector was confirmed by restriction analysis and DNA sequencing. Mouse R1 ES cells were electroporated with the vector and subjected to positive selection with G418 and negative selection with ganciclovir. Targeted ES cell clones were identified by Southern blot analysis, and two independent Ptc2 mutant clones were used to generate germ line chimeras by embryo aggregation. F1 animals were verified by Southern blot hybridization using a 700-bp DNA fragment as indicated in Fig. 1A. Mice heterozygous for the Ptc2 mutant allele (Ptc2^tm1/+ mice) were interbred to generate Ptc2^tm1/tm1 mice. PCR genotyping was performed on ear punch specimens using a combination of three primers, WT1 (5′-CCTGGCCCACCTGTGCCTTA-3′), WT2 (5′-GACCTTCACTAACCTCAGGC-3′), and Neo (5′-TTCCATTTGTCACGTCCTGCACG-3′), to amplify a 150-bp wild-type fragment and a 350-bp mutant fragment (Fig. 1B).

FIG. 1.

Disruption of Ptc2 by gene targeting. (A) Targeting strategy illustrating the genomic organization, a restriction map of the wild-type (Wt) locus, the targeting vector, and the targeted allele (Mt). Ptc2 is located in chromosome 4 and consists of 22 predicted exons. The black bar indicates the position of the probe used for Southern hybridization. Arrows indicate positions of genotyping primers. X, XbaI; Xh, XhoI; B, BamHI; E, EcoRI. (B) Genotyping of progeny by PCR analysis. PCR amplification generated wild-type (150-bp) and mutant (350-bp) bands. (C) Northern blot analysis. The major transcript (a) encoded by Ptc2 is absent in Ptc2^tm1/tm1 testis RNA. Four transcripts (b to e), ranging from 2.4 to 7.5 kb, are present in Ptc2^tm1/tm1 testis RNA. (D) RT-PCR analysis of Ptc2 expression in skin, cerebellum, and testis of wild-type and mutant mice. The diagram indicates the locations of primer pairs Ptc2-A (325 bp), Ptc2-B (600 bp), and Ptc2-C (350 bp) relative to the insertion (arrowhead). Similar results were obtained for all transgenic tissues analyzed. Ptc2-A and Ptc2-C amplified fragments of the expected sizes in wild-type and Ptc2 mutant samples. Ptc2-B amplified expected transcript 1 in the wild type only, transcripts 2 and 3 in the mutant only, and transcript 4 present in the wild type and mutant. (E) Schematic representation of sequencing results obtained for wild-type and mutant transcripts amplified by Ptc2-B primers. Arrows indicate locations of forward (BF) and reverse (BR) primers relative to the genomic sequence. Three alternative splice forms of Ptc2 were isolated, with deletions of exon 6, exons 5 and 6, and exons 6 and 7. (F) Schematic depicting the location of alternative splice forms present in Ptc2 mutants. The translational start ATG codon resides in exon 1. The putative first extracellular loop, implicated in interacting with Shh, is encoded by exons 2 to 9, while the putative sterol-sensing domain is encoded by exons 9 to 13. Our targeting strategy aimed at disrupting the first large extracellular loop of Ptc2, thereby abolishing its interaction with Shh. Ptc2-Δ6,7 represents an alternative splice form of Ptc2 in which exons 6 and 7 are deleted, resulting in an in-frame deletion. In the Ptc2 mutant, two additional splice forms are generated, Ptc2-Δ6 and Ptc2-Δ5,6.

Northern analysis.

Poly(A)⁺ RNA from mouse testis was prepared using an Oligotex kit (QIAGEN) according to the manufacturer's instructions. Three micrograms of RNA was separated on an agarose formaldehyde gel and transferred onto a nitrocellulose membrane (Genescreen). A Ptc2 cDNA probe was labeled with [α-³²P]dCTP using Ready-To-Go DNA-labeling beads (Amersham).

RT-PCR analysis.

Total RNA was extracted from mouse testis, skin, and cerebellum using Trizol (Invitrogen) according to the manufacturer's instructions. RT-PCRs were performed using the Superscript one-step RT-PCR kit (Invitrogen) and the following primers: Ptc2-AF (5′-GTCTCCGAGTGGCTG TAA-3′), Ptc2-AR (5′-TTCTCAATCATCCGTTCG-3′), Ptc2-BF (5′-CACCC CCGAGGCACTTGA-3′), Ptc2-BR (5′-GCCCCGGAAGTGCTCGTA-3′), Ptc2-CF (5′-GTGGCTCCCCCTTCCTCTTCT-3′), Ptc2-CR (5′-AGGGGCAAAGGTCTGTTCC-3′), GAPDH-F (5′-GTGGCAAAGTGGAGATTGTTGCC-3′), GAPDH-R (5′-GATGATGACCCGTTTGGCTCC-3′), Ptc1-F (5′-AACAAAAATTCAACCAAACCTC-3′), Ptc1-R (5′-TGTCTTCATTCCAGTTGATGTG-3′), Gli1-F (5′-TTCGTGTGCCATTGGGGAGG-3′), and Gli1-R (5′-CTTGGGCTCCACTGTGGAGA-3′). Reaction conditions are available upon request. The intensity of Ptc1 and Gli1 amplification products was analyzed and compared to the internal control (GAPDH), using ImageJ software. Ptc2 transcripts were subcloned into a TA vector (Invitrogen), and the products were examined by DNA sequencing.

Generation of myc-tagged Ptc2 expression constructs.

Template cDNA was generated from wild-type testes RNA (Superscript first-strand synthesis system for RT-PCR; Invitrogen). Overlapping primers were designed to amplify Ptc2 cDNA lacking the start codon (AccuPrime Pfx DNA polymerase; Invitrogen). Based on sequencing information obtained regarding Ptc2 mutant transcripts (see above) primers were designed to amplify the affected regions (exons 1 to 4, 1 to 5, 7 to 12, and 8 to 12). Products were verified by sequencing and assembled in a pcDNA3.1 vector (Invitrogen) modified by in-frame insertion of an N-terminal myc cassette (ATGCAAAAACTCATCTCAGAAAGAGGATCTG). The following constructs were generated: Ptc2-WT, Ptc2-Δ5,6, Ptc2-Δ6, and Ptc2-Δ6,7. Primers and reaction conditions are available upon request.

Confocal microscopy and immunofluorescence.

CH310T1/2 cells were maintained in Dulbecco's modified Eagle medium supplemented with 10% fetal bovine serum. Cells were seeded on gelatin-treated coverslips and transfected using Fugene6 (Roche). At 48 h posttransfection, cells were fixed, followed by incubation with anti-c-myc (Santa Cruz) or antihemagglutinin (HA; Santa Cruz). Slides were processed for indirect fluorescence (fluorescein isothiocyanate-conjugated secondary antibody; Jackson Laboratories) visualization.

Luciferase assays.

Ptc1^−/− embryonic fibroblasts were maintained as previously described (6). 8xGli-BS-luc (54) reporter assays were performed as previously described (18) following transfection of Ptc1-HA (6), Ptc2-myc, Ptc2-Δ5,6-myc, Ptc2-Δ6-myc, or Ptc2-Δ6,7-myc expression constructs. Gli-dependent transcription was measured and normalized using a dual-luciferase reporter assay (Promega). Data were obtained from three independent experiments, each performed in triplicate, for Fig. 2C and two independent experiments, each performed in triplicate, for Fig. 2D.

FIG. 2.

In vitro analysis of Ptc2 mutant forms. (A) Schematic indicating myc-tagged expression constructs for wild-type and mutant Ptc2. Deletion of exon 6 in Ptc2-Δ6 likely results in a premature truncation of the Ptc2 protein (gray shading). (B) Subcellular localization of Ptc1, Ptc2 wild-type, and Ptc2 mutant constructs. Ptc2-Δ5,6 is present in the cytoplasmic and nuclear compartments, while Ptc2-Δ6 and Ptc2-Δ6,7 are localized similarly to Ptc1 and Ptc2. (C) Luciferase assay showing that Ptc2 can function similarly to Ptc1 as a negative regulator of Shh signaling. The data shown are the averages from three independent experiments performed in triplicate. (D) Graph showing the effect of Ptc2 mutant forms on Gli-dependent transcription represented as relative luciferase activity and relative % inhibition. When compared to wild-type Ptc2, each of the mutants showed a statistically significant increase in Gli-dependent transcription (**, P < 0.05). Ptc2-Δ6 and Ptc2-Δ6,7 behaved similarly, and the differences in their relative luciferase activities were not statistically significant. Ptc2-Δ5,6 had the greatest decrease in its ability to inhibit Gli-dependent transcription compared to Ptc2. The results are from two independent experiments performed in triplicate.

Histological and skeletal analyses.

Embryos at E15.5 and E18.5 from heterozygous intercrosses were collected and fixed in 4% (wt/vol) paraformaldehyde (PFA) in phosphate-buffered saline (PBS) overnight at 4°C, processed, and embedded in paraffin following standard procedures. Dorsal and front paw skin samples from adult wild-type, heterozygous, and homozygous mutant animals were also processed. For histological analysis, 5-μm sections were prepared and stained with hematoxylin and eosin. Skeletal analysis of E18.5 embryos was performed using alcian blue/alizarin red bone staining (44).

In situ hybridization.

Tissues and embryos were fixed in 4% PFA in PBS for 16 h at 4°C. Paraffin sections (5 μm) or whole-mount embryos were subjected to in situ hybridization with digoxigenin-dUTP-labeled riboprobes as described previously (19). Plasmids used for generating riboprobes were those carrying Ptc1, Gli1, Shh (46), Keratin-15 (B. Morgan), Keratin-17 (P. Coulombe), and Ptc2-N (generated by RT-PCR; nucleotides 504 to 827).

Immunohistochemistry.

Immunohistochemistry was carried out as described previously (16) on 5 μm paraffin sections. Dilutions and other details concerning antibodies are available upon request. The following primary antibodies were used: keratin-5, keratin-14, keratin-10, loricrin (Covance), cyclin D1/D2, PCNA, phospho-histone H3 (Ser10) (Cell Signaling Technology), and GATA-3 (Santa Cruz). Fluorescein isothiocyanate- or tetramethyl rhodamine isocyanate-conjugated (Jackson Laboratories), as well as biotinylated (Vector Labs), secondary antibodies were used, followed by visualization with the ABC Vectastain and VIP or NovaRed color substrate kits (Vector Labs). Endogenous alkaline phosphatase activity was detected using rehydrated paraffin sections. Sections were equilibrated in APB buffer (0.1 M NaCl, 0.05 M MgCl, 0.1 M Tris, pH 9.5, 0.1% Tween 20) for 45 min, followed by a 2-hour incubation with BM Purple (Roche) at room temperature in the dark. Following immunohistochemistry, sections were dehydrated and mounted using Permount. Sections were examined and photographed using an Axioskop microscope (Carl Zeiss).

RESULTS

Targeted disruption of Ptc2.

To investigate the role of Ptc2 in mammalian development, we generated Ptc2^tm1/tm1 mice. The Ptc2 gene was disrupted in ES cells by targeted insertion of a neomycin resistance cassette (pGKNeopA) in exon 6 (Fig. 1A), which encodes part of the large extracellular loop of the Ptc2 protein (Fig. 1F), implicated in direct interaction with Shh (29, 40). This modification was expected to result in a loss-of-function mutant allele due to several stop codons created 3′ to the insertion. Germ line chimeras were obtained from two independently targeted ES cell lines, and both lines gave rise to fertile heterozygous mutant mice. Intercrosses of heterozygous mutant mice were used to generate homozygous mutant mice. Southern blot hybridization confirmed the disruption of the Ptc2 gene (data not shown). Northern blot analysis revealed that the 5-kb Ptc2 transcript is absent in the testes of Ptc2^tm1/tm1 animals (Fig. 1C). Instead, several mutant transcripts ranging from 2.4 to 7.5 kb were detected. We performed RT-PCR on skin, cerebellum, and testis, followed by sequence analyses to further characterize the Ptc2 mutant transcripts (Fig. 1D and E; see Fig. S1 in the supplemental material). Regions 5′ and 3′ to the insertion were still transcribed from the mutant allele (products Ptc2-A and Ptc2-C in Fig. 1D). However, using Ptc2-B primers, the 600-bp wild-type transcript was found to be absent in the Ptc2^tm1/tm1 sample (Fig. 1D, #1). Instead, three mutant transcripts were detected. In wild-type samples, the intensity of the 600-bp transcript was greater than the intensities of any of the transcripts in the mutant. All the mutant transcripts were sequenced. Transcripts 2 and 3 have deletions of exon 6 and exons 5 and 6, respectively. Transcript 4 lacks exons 6 and 7 (Fig. 1E; see Fig. S1 in the supplemental material) and appears to be a minor splice product of Ptc2 present in both wild-type and Ptc2^tm1/tm1 RNA. A summary of Ptc2 products generated is presented in Fig. 1E and F; see also Fig. S1 in the supplemental material.

We were unable to determine whether Ptc2^tm1 is a protein-null allele since attempts to generate a Ptc2-specific antibody were unsuccessful. Instead we generated myc-tagged expression constructs for wild-type and mutant Ptc2 (summarized in Fig. 2A) and performed an in vitro analysis. Ptc1, Ptc2, Ptc2-Δ6, and Ptc2-Δ6,7 appeared to be localized perinuclearly, while Ptc2-Δ5,6 showed diffuse cytoplasmic and nuclear staining (Fig. 2B). Luciferase assays were performed to determine the effect of Ptc2 mutant forms on Gli-dependent transcription (Fig. 2C and D). First we showed that Ptc2 acts similarly to Ptc1 to decrease Gli-dependent transcription (Fig. 2C) when expressed in Ptc1^−/− embryonic fibroblast cells. Having established the assay, we examined the mutant forms of Ptc2. We found that, when compared with wild-type Ptc2, Ptc2-Δ5,6 had greatly reduced inhibitory activity, whereas the activities of Ptc2-Δ6 and Ptc2-Δ6,7 were compromised to a lesser extent (Fig. 2D). We consistently observed great variability in the results obtained for Ptc2-Δ6. This mutant form is predicted to be a protein truncated at amino acid 206. Interestingly, Ptc2-Δ5,6, in addition to having lost its ability to negatively regulate Shh signaling, also has altered subcellular localization compared to Ptc2, Ptc2-Δ6, and Ptc2-Δ6,7. Western blot analysis suggested that Ptc2-Δ5,6 is deficient in protein maturation, indicating that the deleted exons might be encoding domains necessary for early folding or glycosylation of Ptc2 (data not shown). Ptc1 missense mutations in regions encoding the large extracellular loop have previously been shown to affect glycosylation of Ptc1 when overexpressed in Ptc1^−/− cells (7). Together, these results suggest that Ptc2^tm1 likely represents a hypomorphic allele of Ptc2.

Normal Ptc2 function is not required for viability.

To determine whether Ptc2 is required during development, we analyzed offspring from Ptc2^tm1/+ intercrosses. PCR genotyping at 3 weeks of age (n = 158) revealed normal Mendelian ratios of Ptc2^+/+ (22%, n = 35), Ptc2^tm1/+ (54%, n = 86), and Ptc2^tm1/tm1 (23%, n = 37) mice, indicating that normal Ptc2 function is not essential for embryogenesis or postnatal survival. Breeding of male and female Ptc2^tm1/tm1 mice resulted in healthy offspring, suggesting that Ptc2 is not required for reproduction. Furthermore, Ptc2^tm1/tm1 mice did not exhibit any differences in mortality when compared with Ptc2^tm1/+ or Ptc2^+/+ control littermates. Thus, normal Ptc2 function is not required for embryonic development, viability, and fertility.

Normal limb development in Ptc2 mutant mice.

Ptc2 is highly expressed in the posterior mesenchyme of the developing limb buds and can be induced by Shh (37, 51). To investigate whether Ptc2 plays a role in Shh signaling during limb development, we examined the expression of two Shh target genes, Ptc1 and Gli1, in Ptc2^tm1/tm1 mice by whole-mount RNA in situ hybridization (Fig. 3A to H). Consistent with the notion that Ptc2 may act negatively in Shh signaling, E11.5 Ptc2^tm1/tm1 mutant limb buds showed a slight anterior expansion of Ptc1 and Gli1 expression. Despite the expansion of Shh target gene expression, alcian blue and alizarin red staining of E18.5 skeletons did not reveal any abnormalities in Ptc2^tm1/tm1 mice (Fig. 3I to L), suggesting that reduced Ptc2 function does not perturb limb patterning and development.

FIG. 3.

Normal limb development in Ptc2-deficient mice. (A to H) Whole-mount in situ hybridization of E11.5 limb buds. Expression of Ptc1 (A, B, E, and F) and Gli1 (C, D, G, and H) in forelimb (A to D) and hind limb (E to H) buds of wild type (A, C, E, and G) and Ptc2^tm1/tm1 (B, D, F, and H) embryos is shown. Bars containing graded shading indicate increased expression of Ptc1 and Gli1 in Ptc2^tm1/tm1 forelimb and hind limb buds. (I to L) Alcian blue and alizarin red staining of E18.5 wild-type and Ptc2^tm1/tm1 limbs revealed no significant difference in skeletal patterning and development.

Ptc2 mutants have normal cerebellum and testis development.

Both Ptc1 and Ptc2 are expressed in the developing cerebellum, but Ptc2 expression is more restricted and is confined to the proliferative region of the external germinal layer (EGL) at postnatal day 7. This suggests a selective involvement of Ptc2 in proliferative populations of the EGL (36). However, examination of gross morphology as well as histological analysis did not reveal any abnormalities in the adult Ptc2 mutant cerebella (Fig. 4A to D). In situ hybridization analysis indicated that Ptc1 and Gli1 are not upregulated in Ptc2^tm1/tm1 cerebella (Fig. 4I to L). Another major site of Ptc2 expression is the primary and secondary spermatocytes in the seminiferous tubules of the adult testes (14). Since primary and secondary spermatocytes are in close contact with supporting Sertoli cells, which express Dhh, it is possible that Ptc2 may act as a receptor for Dhh in the testes. However, Ptc2 mutant males are fertile and histological analysis of Ptc2^tm1/tm1 testis did not reveal any abnormalities (Fig. 4E to H). Similar to what was found for cerebella, expression of Ptc1 and Gli1 was unaffected in the testes of Ptc2 mutants (data not shown). Semiquantitative RT-PCR analysis of Ptc1 and Gli1 mRNA levels (Fig. 4M) confirmed results obtained by in situ hybridization. Together, these results indicate that normal Ptc2 function is not required for the development of the cerebellum or testes.

FIG. 4.

Ptc2-deficient mice develop normal cerebellums and testes, and Ptc1 and Gli1 are not upregulated. Histological staining of cerebellums (A to D) and testes (E to H) of adult wild-type (A, C, E, and G) and Ptc2^tm1/tm1 (B, D, F, and H) mice is shown. (A, B, E, and F) Low-magnification views; C, D, G, H, and I to L, high-magnification views. (I to L) In situ hybridization of Ptc1 (I and J) and Gli1 (K and L) showing normal expression in Ptc2 mutants. IGL, Internal germinal layer; Pu, Purkinje cell layer. (M) Semiquantitative RT-PCR analysis of Ptc1 and Gli1 expression levels in wild-type (wt) and Ptc2^tm1/tm1 (mt) cerebellum and testis. Results obtained from RT-PCR analysis were normalized to GAPDH (internal control) and are represented as expression detected in wild-type samples compared to mutant samples.

Ptc2 is not required for embryonic hair follicle development.

Shh signaling is essential for the development of epidermal structures such as hair follicles, and hair follicle development is arrested in Shh^−/− mice (16, 59). We have previously shown that Ptc2 and Shh are coexpressed in the developing hair follicles during mouse embryogenesis (45). To determine whether Ptc2 plays a role in hair follicle development, we analyzed E15.5 and E18.5 Ptc2^tm1/tm1 skin. Histological analysis showed that Ptc2 mutants develop hair follicles with normal morphology and density (Fig. 5A, G, M, and S). Marker gene analysis indicated that the differentiation of epidermal cells appears grossly normal. Similar to wild-type skin (Fig. 5B to E and N to Q), Ptc2^tm1/tm1 skin showed normal expression of Keratin-17 (Fig. 5H and T) and keratin-14 (Fig. 5I and U) in the basal layer, keratin-10 in the suprabasal layer (Fig. 5J and V), and loricrin in the cornified layer (Fig. 5K and W) of the epidermis. Furthermore, epidermal proliferation of Ptc2^tm1/tm1 skin was also similar to that of wild-type skin, as revealed by the expression pattern (Fig. 5F, L, R, and X) and the number of phospho-histone H3-positive cells (Fig. 5Y). These results indicate that normal Ptc2 function is not required for embryonic hair follicle development.

FIG. 5.

Ptc2 is not required for embryonic hair follicle development. Histological analysis of wild-type (A and M) and Ptc2^tm1/tm1 (G and S) hair follicle development at E15.5 (A and G) and E18.5 (M and S) revealed normal differentiation and proliferation. In situ hybridization and immunohistochemistry for markers of hair follicle development at E15.5 and E18.5 are shown. Expression levels of Keratin-17 (B, H, N, and T), keratin-14 (C, I, O, and U), keratin-10 (D, J, P, and V), loricrin (E, K, Q, and W), and phospho-histone H3 (PH3) (F, L, R, and X) appear to be normal in the Ptc2^tm1/tm1 epidermis. (Y) Cell proliferation is not affected in Ptc2^tm1/tm1 skin at E15.5 or E18.5. Data represent the average numbers of PH3-positive cells (arrowheads in panels F, L, R, and X) obtained from counting 40 contiguous microscopic fields per sample. The error bars indicate standard deviations, and a t test revealed that the differences are not statistically significant. Dashed lines indicate epidermis-dermis boundaries. ep, epidermis; de, dermis; bl, basal layer; sbl, suprabasal layer; cl, cornified layer. Scale bar: 50 μM.

We next analyzed whether reduction of Ptc2 function affects the Shh response in developing skin. In situ hybridization analysis indicated that the expression of Gli1 (Fig. 6H and K) was not affected, while Ptc1 (Fig. 6G and J) appeared to be ectopically activated in the interfollicular epidermis of Ptc2^tm1/tm1 skin at E18.5. Ptc1 also appeared slightly elevated in E18.5 Ptc2^tm1/tm1 hair follicles, despite normal follicular expression of Shh (Fig. 6C, F, I, and L). Interestingly, hair follicles did not exhibit morphological defects despite ectopic Ptc1 expression, suggesting that Ptc1 might be upregulated as a mechanism to compensate for reduced Ptc2 function.

FIG. 6.

Ptc1 is upregulated in Ptc2^tm1/tm1 embryonic skin. Shown is expression of Ptc1 (A, D, G, and J), Gli1 (B, E, H, and K), and Shh (C, F, I, and L) in wild-type (A to C and G to I) and Ptc2^tm1/tm1 (D to F and J to L) embryos at E15.5 (A to F) and E18.5 (G to L). Ptc1 is slightly upregulated in E18.5 hair follicles (black arrowhead) and ectopically expressed in the interfollicular epidermis (IFE; red arrowhead) in Ptc2^tm1/tm1 skin. Dashed lines indicate epidermis-dermis boundaries. ep, epidermis; de, dermis; hf, hair follicle. Scale bar: 50 μM.

Ptc2 mutant males develop alopecia and skin lesions.

Despite the lack of an embryonic skin phenotype, 48% of Ptc2^tm1/tm1 male animals (and Ptc2^tm1/+ male animals to a lesser extent) develop alopecia (hair loss) and 15% developed spontaneous skin lesions (Fig. 7; Table 1). Ptc2^tm1/tm1 male animals exhibit an earlier onset of the phenotype (6 to 7 months of age) compared to Ptc2^tm1/+ male animals (9 to 10 months of age). At 15 months of age, 80% of Ptc2^tm1/tm1 male animals developed the full spectrum of lesions while 74% of Ptc2^tm1/+ male animals had lesions (Table 1). Histological analysis of a total of 187 wild-type, Ptc2^tm1/+, and Ptc2^tm1/tm1 adult male and female animals (Table 1) revealed that lesions, including epidermal hyperplasia, dermal hyperplasia, hair follicle loss, and ulceration, are observed only in skin samples of Ptc2^tm1/+ and Ptc2^tm1/tm1 male animals (Fig. 7E, F, H, and I), but not in those of Ptc2^tm1/+ and Ptc2^tm1/tm1 female animals (data not shown). The lesions consistently developed on the dorsal and front paw skin (Fig. 7B and C). We excluded the possibility that these lesions resulted from fighting by housing one male per cage and closely observing them for abnormal grooming behavior.

FIG. 7.

Ptc2 mutant males develop alopecia and skin lesions. Ptc2^tm1/tm1 males with ulceration and alopecia on back skin (B) and front paw skin (C) (arrowheads) are shown compared to a normal wild-type male (A). (D and G) Histological staining of wild-type back skin (D) and front paw skin (G). (E and F) Epidermal hyperplasia mutant back skin (E) and front paw skin (F). (H and I) Ulceration shown in mutant back skin (H) and mutant front paw skin (I). ep, epidermis; de, dermis; hf, hair follicle; hy; hypodermis; mu; muscle; ul, ulcer. Scale bar: 50 μM.

TABLE 1.

Phenotypic analysis of adult Ptc2 mutant mice

Phenotype	% of mice
	Male			Female
	Wild type (n = 13)	Ptc2tm1/+ (n = 38)	Ptc2tm1/tm1 (n = 40)	Wild type (n = 12)	Ptc2tm1/+ (n = 34)	Ptc2tm1/tm1 (n = 50)
Epidermal hyperplasia	0	29	60	0	3	0
Dermal proliferation	0	61	70	0	0	0
Hair follicle loss	0	21	48	0	0	0
Ulcers	0	3	15	0	0	0
Normal	100	26	20	100	97	100

Since Ptc2^tm1/tm1 mice developed hair loss, we next wanted to determine whether mutant mice fail to form the appropriate layers of the hair follicle. We performed this study on postnatal day 12 skin since all hair follicles are synchronized in the first anagen phase of the hair cycle. Histological (Fig. 8A and C) and marker analyses (Fig. 8B and D to P) revealed that the epidermis is normal at this stage. Keratin-15 expression (Fig. 8B and D) in Henle's layer of the inner root sheath (IRS) was normal. GATA-3 (Fig. 8E and G), a marker for Huxley's layer and the cuticle of the IRS (33), was also expressed in a normal pattern. Keratin-5 (Fig. 8F and H), expressed in the outer root sheath, and keratin-6 (Fig. 8M and O), expressed in the companion cell layer, were expressed similarly in wild-type and mutant samples. We concluded that all the layers of the hair follicles expressed the appropriate markers, suggesting that hair loss in Ptc2 mutants is not due to failures in forming the inner or outer root sheaths.

FIG. 8.

Hair loss in Ptc2 mutants is not due to aberrant hair follicle development. Shown are histological and immunohistochemical analyses of development of the differentiated layers of the hair follicle and epidermis. (A and C) Normal morphology of hair follicles. All markers analyzed are expressed in the normal pattern. (B and D) Keratin-15 expression in Henle's layer of the IRS; (E and G) GATA-3 is expressed in Huxley's layer of the IRS; (F and G) keratin-5 in the outer root sheath (ORS) and alkaline phosphatase (AP) in the dermal papillae; (M and O) keratin-6 in the companion cell layer. Epidermal development was also normal. (F, H, I, and K) Keratin-5 and keratin-14 in the basal cell layer; (J and L) keratin-10 in the suprabasal layer; (N and P), loricrin (Lor) in the cornified layer. Scale bars: 50 μM. DAPI, 4′,6′diamidino-2-phenylindole; HE, hematoxylin and eosin.

To examine the nature of the skin lesions in Ptc2^tm1/tm1 mice, we analyzed the expression of differentiation and proliferation markers of the epidermis. The analysis revealed that the expression of loricrin, which is a marker of the cornified layer, is grossly normal in the Ptc2 mutant lesions (Fig. 9A and B). High levels of keratin-14 expression in areas of hyperplasia suggest that the hyperplasia is due to an expanded proliferative basal layer (Fig. 9C and D). As expected for adult skin during the repair process, the domain of Keratin-17 expression is expanded in the suprabasal cells (Fig. 9E and F) and, consistent with epidermal hyperplasia, the proliferative marker PCNA is also expressed at high levels in the suprabasal cells (Fig. 9G and H).

FIG. 9.

Marker analysis of skin lesions in Ptc2 mutant males. Shown are immunohistochemistry and in situ hybridization of epidermal compartments. Loricrin (Lor) expression appears to be normal (A and B). Keratin-14 expression is expanded suprabasally in ulcerated Ptc2^tm1/tm1 skin (D) compared to wild-type epidermis, where it is expressed only in the basal layer (C). Keratin-17 (E and F) expression indicates expansion of the basal layer in areas of epidermal hyperplasia in affected Ptc2 mutant skin (F). Similar expression of PCNA was observed in wild-type (G) and Ptc2^tm1/tm1 skin (H). Multiple layers of PCNA-expressing cells could be detected in tissue underlying ulcerated lesions (H, inset). ep, epidermis; de, dermis; hf, hair follicle; ul, ulcer. Scale bars: 50 μM.

Shh signaling is activated in regions of epidermal hyperplasia in affected skin of Ptc2 mutants.

Ptc1, Gli1, and Shh are normally expressed at low levels in adult skin (Fig. 10A, C, and E) but are upregulated in regions of epidermal hyperplasia in Ptc2 mutants (Fig. 10B, D, and F). Consistent with increased proliferative capacities and Shh signaling, cyclin D1/D2 protein levels are highly elevated in areas of epidermal hyperplasia (Fig. 10G and H). Previous reports indicated that D-type cyclins are involved in Hh-mediated growth control (20, 34, 38, 43, 65), and their upregulation contributes to tumor growth associated with Hedgehog pathway activation (26). These results suggest that the Shh signaling pathway is activated in epidermal hyperplasia associated with the Ptc2 mutation.

FIG. 10.

Sonic hedgehog signaling is activated in the Ptc2^tm1/tm1 epidermis. Shown is marker gene analysis by in situ hybridization (A to F) and immunohistochemistry (G to H). Ptc1 (A and B), Gli1 (C and D), Shh (E and F), and cyclin D1/D2 (G to H) expression is increased in areas of epidermal hyperplasia in affected Ptc2^tm1/tm1 skin. Arrowheads indicate low levels of Gli1, Shh, and cyclin D1/D2 expression in hair follicles of wild-type skin. ep, epidermis; ul, ulcer. Scale bar: 50 μM.

DISCUSSION

Ptc2^tm1 is likely a hypomorphic mutant allele.

The objective of the present study was to elucidate the role of Ptc2 in mammalian development and Shh signaling. Through targeted insertion of a neomycin cassette in exon 6, we have generated Ptc2^tm1 mice. Northern blot and RT-PCR analyses indicated that Ptc2^tm1/tm1 mice lack the major Ptc2 transcript but produce several minor mutant transcripts. We identified a minor splice variant of Ptc2, Ptc2Δ6,7, present in both wild-type and mutant samples. Ptc2Δ6,7 displayed similar subcellular localization as wild-type Ptc2 and possessed only a slightly reduced ability to inhibit Gli-dependent transcription. Previous studies revealed several splice variants for human PTCH2 involving the SSD (52). In our study, alternative splicing appears to affect only regions 5′ to the SSD, revealing a possible difference in the function of mouse and human Ptc2/PTCH2 proteins. Overexpression studies of Ptc2 splice variants resulting from the Ptc2^tm1 mutation suggested that they have a compromised ability to negatively regulate Shh signaling. Taken together, our molecular characterizations suggest that Ptc2^tm1 is likely a hypomorphic mutant allele of Ptc2.

Normal Ptc2 function is not essential for embryogenesis.

We show here that Ptc2^tm1/tm1 mice are viable and fertile and do not display any obvious defects. Although Ptc2 is expressed in the skin, limb, testes, and cerebellum, our results clearly indicate that reduced Ptc2 function does not impede embryogenesis, viability, and reproduction. Ptc2 displays overlapping expression with Shh in the epidermal compartment of the developing hair follicles, where Ptc1 and Ptc2 show a mostly nonoverlapping expression pattern, suggesting that Ptc2 might play a unique role during embryonic hair follicle development. Interestingly, in Ptc2^tm1/tm1 mutants hair follicle development is not perturbed. There are at least two explanations for the lack of a strong mutant phenotype in Ptc2^tm1/tm1 mice. First, Ptc2^tm1/tm1 mice might still possess residual Ptc2 function that is sufficient for programming normal embryonic development. Second, Ptc1 might compensate for the loss of Ptc2 function in Ptc2^tm1/tm1 mice. As Ptc1^−/− mice die at E9, well before the onset of the major Ptc2 expression, the latter hypothesis awaits the conditional knockout of Ptc1 (22) in a Ptc2 mutant background.

Upregulation of Shh target genes suggests that Ptc2 acts as a negative regulator of Shh signaling.

Despite the lack of an apparent mutant phenotype, we can detect a slight upregulation of Ptc1 and Gli1 expression in the developing limb buds and hair follicles of Ptc2^tm1/tm1 mice, suggesting that, like Ptc1, Ptc2 acts as a negative regulator of Hh signaling. It remains to be determined whether the upregulation of Ptc1 in Ptc2^tm1/tm1 mice contributes to functional compensation. It has been shown that Ptc2 expression is upregulated in Ptc1^−/− cells and that Ptc2 expression apparently does not compensate for the lack of Ptc1 function (6). Interestingly, we also detected upregulation of Ptc1 expression in the interfollicular epidermis of E18.5 Ptc2 mutant skin. This finding might suggest that Ptc2 has a specific function in preventing ectopic expression of Ptc1 in the epidermis.

Ptc2 plays an important role in epidermal homeostasis in mature skin.

Interestingly, although adult Ptc2 mutant mice appear grossly normal, we find that mutant male animals develop skin lesions consisting of alopecia and ulceration with progressing age. These observations indicate that normal Ptc2 function is required for skin homeostasis and that the phenotype is sex dependent. Histological and marker analysis shows that Ptc2-deficient mice have severe epidermal hyperplasia and that the Shh signaling cascade is activated in these lesions. In contrast, cell proliferation and the Shh signaling cascade are not affected in the normal skin of Ptc2-deficient mice. Hh pathway activation has been linked to transcriptional activation of cell cycle genes, thus predisposing cells to higher rates of proliferation and hyperplasia. Interestingly, although we have observed Hh pathway activation and upregulation of cyclin D1/D2 expression in the epidermal hyperplasia of Ptc2-deficient mice, these lesions never develop into BCC or other skin tumors.

In humans, the etiology and genetic basis of male pattern baldness (androgenetic alopecia) are unclear. Androgenetic alopecia appears to be caused by a combination of genetic predisposition and elevated levels of circulating androgen (21). Normal hair follicles undergo a continuous process of renewal throughout life. In androgenetic alopecia, there is a shortening of the anagen (growth) phase of the hair follicles. Since Shh signaling is required for the initiation of anagen (55), it is possible that epidermal cells lacking normal Ptc2 function may exhibit a defect in responding to Shh during the adult hair cycle. It is currently unclear why the Ptc2^tm1/tm1 phenotype affects only adult males. Intriguingly, it has been reported that, in humans, males heal acute cutaneous wounds more slowly than females and have an altered inflammatory response (2, 3, 63). It has been also demonstrated that the male genotype is a strong positive risk factor for impaired wound healing in the elderly. The mechanisms underlying such sex differences have not been elucidated, although endogenous testosterone has been found to inhibit cutaneous wound healing (4). Interestingly, male gender is also a strong predisposing factor for BCCs in humans and is associated with more BCCs developing per year compared to females (53). In a study of radiation-induced BCCs in Ptc1 heterozygous mice, only males developed infiltrative BCCs (39). It is possible that BCC development is affected by hormone status and that Ptc genes are involved in responses to steroid hormones or their precursors (15).

Is Ptc2 a tumor suppressor?

In humans, disease-associated Ptc1 mutations are dispersed over the entire coding sequence. About three-fourths of them are predicted to result in truncated proteins, while the rest are in-frame insertions and deletions. Studies aimed at investigating the effect of Ptc1 missense mutations (6, 7), identified in humans with BCC or NBCCS, suggested that Ptc1 mutants that had greatly reduced or absent activity might be deficient in maturation and/or endocytosis. No correlation was found between the position of the mutation in the protein sequence and the degree to which Ptc1 function was disrupted, but a mutation in the large extracellular loop blocked Ptc1 maturation and suggested that this loop might be involved in folding or glycosylation. Similarly Ptc2 mutant forms generated in our study also disrupt the large extracellular loop, and from Western blot analysis we predict that Ptc2-Δ5,6 mice have similar defects in maturation. Interestingly, while Ptc2 deletions might affect Ptc2 similarly to missense mutations in Ptc1, Ptc2^tm1/tm1 mice did not develop tumors.

In this study, we have shown that normal Ptc2 function is not required for embryogenesis, viability, and reproduction. We speculate that Ptc1 may be able to compensate for the lack of Ptc2 function during development; however, Ptc2 plays an indispensable role in Shh signaling in the adult male mouse skin and is required for skin homeostasis.