|Home | About | Journals | Submit | Contact Us | Français|
Ptch1 is a key regulator of embryonic development, acting through the sonic hedgehog (SHH) signaling pathway. Ptch1 is best known as a tumor suppressor, since germline or somatic mutations in Ptch1 lead to the formation of skin basal cell carcinomas (BCCs). Here, we show that Ptch1 also acts as a lineage-dependent oncogene, as overexpression of Ptch1 in adult skin in K14PtchFVB transgenic mice synergizes with chemically induced Hras mutations to promote squamous carcinoma development. These effects were not due to aberrant activation of SHH signaling by the K14PtchFVB transgene, as developmental defects in the highest expressing transgenic lines were consistent with inhibition of this pathway. Carcinomas from K14PtchFVB transgenic mice had only a small number of non-proliferative Ptch1 transgene positive cells, suggesting that the Ptch1 transgene is not required for tumor maintenance, but may play a critical role in cell fate determination at the initiation stage.
Ptch1 is a key signaling receptor in the sonic hedgehog (SHH) signaling pathway, which is critical in embryonic development, tissue patterning, and cell fate decisions (Chen, 1996; Ingham, 1991; Marigo et al., 1996). Mutations resulting in loss of Ptch1 function or aberrant expression of genes in the SHH pathway lead to developmental defects and a cancer-prone phenotype. Germline mutations in the human PTCH1 gene are responsible for Gorlin syndrome (also called nevoid basal cell carcinoma syndrome, NBCCS), which is characterized by various developmental malformations and a high predisposition to skin basal cell carcinoma (BCC) development (Hahn, 1996; Johnson et al., 1996). Mice partially deficient in functional Ptch1 exhibit developmental abnormalities that are comparable to those of NBCCS patients. Additionally, these mice develop tumors such as medulloblastomas, rhabdomyosarcomas, and BCCs following ultraviolet (UV) or ionizing radiation exposure, demonstrating a critical role for Ptch1 in the regulation of developmental homeostasis and tumor suppression (Aszterbaum, 1999; Goodrich, 1997; Hahn et al., 1998).
We have previously identified a polymorphic variant in the mouse Ptch1 gene (PtchFVB) that conferred susceptibility to early post-natal squamous cell carcinoma (SCC) development in transgenic mice expressing a mutant H-ras gene under the control of the keratin 5 promoter (K5Hras mice) (Wakabayashi et al., 2007). In this model, both Ptch1 and activated Ras are expressed under the control of constitutive keratin promoters during skin development (Byrne et al., 1994), and it remained possible that this paradoxical effect of Ptch1 in promoting cancer may be linked to developmental abnormalities related to this specific model. Therefore, we tested the role of Ptch1 in adult-onset skin tumor formation using the classical DMBA(dimethylbenzanthracene)/TPA(12-O-tetradecanoylphorbol-13-acetate) model involving sporadic mutation of the endogenous H-ras gene, rather than the high levels of mutant H-ras expressed during development in the K5Hras transgenic mice. Our data demonstrate that even low levels of Ptch1 transgene expression that do not perturb expression of Gli transcription factors promote formation of malignant SCCs. We conclude that the SHH pathway can play either positive or negative roles in development of alternative tumor types within the same tissue.
We previously generated four different transgenic lines (lines 6, 7, 8, and 10) expressing PtchFVB under the control of the K14 promoter at variable levels, with line 6 showing the highest Ptch1 transgene expression (Wakabayashi et al., 2007). Initially, these transgenic lines were generated and maintained on the FVB/N background. Continuous breeding to the same background resulted in poor litter sizes, particularly in line 6 with the highest PtchFVB levels. To maintain this line, we crossed the line 6 K14PtchFVB mice to wild-type C57BL/6 mice and obtained litters on this hybrid background. We observed major phenotypic abnormalities in the line 6 K14PtchFVB mice on both FVB/N and hybrid backgrounds. First, when compared to their wild-type littermates, line 6 K14PtchFVB mice exhibited shorter mean body lengths (5.6 ± 0.48 cm vs. 7.9 ± 0.25 cm, P=0.00065 by t-test) and lower mean body weights (5.9 ± 0.74 g vs. 14.7 ± 0.68 g, P<0.00001 by t-test) at 4 weeks of age (Figure 1a). This smaller body size was not visible in the other transgenic lines apart from a slight size difference in line 8 K14PtchFVB mice (data not shown). Furthermore, while the wild-type littermates showed normal eye development, line 6 K14PtchFVB mice clearly displayed ocular defects involving bilateral and unilateral underdevelopment of the ocular structures (Figure 1b). Newborn line 6 K14PtchFVB mice exhibited either aphakia with microphthalmia (Figure 1b, upper middle panel), or anophthalmia (Figure 1b, upper right panel). Similar ocular defects have been described in mice with mutations in Gas1, a membrane-bound glycoprotein that is known to antagonize SHH signaling (Lee et al., 2001a, 2001b; Seppala et al., 2007). Adult line 6 K14PtchFVB mice demonstrated complete and permanent closure of the developmentally abnormal eyes (Figure 1b, lower panel). Finally, just as limb deformities have been linked to defects in both mouse and human genes involved in the SHH pathway, the line 6 K14PtchFVB mice displayed abnormal limb development (Figure 1c). Both forelimbs and hindlimbs of line 6 K14PtchFVB mice showed oligodactyly with fewer digits, consistent with a general inhibition of SHH signaling during mouse development. The defective eyes and limb deformities were not observed in lines 7 and 8 K14PtchFVB mice. These phenotypic abnormalities in Ptch1 transgenic mice are clearly opposite to those observed in Ptch1+/− mice or animals expressing transgenic Shh (Goodrich, 1997; Oro et al., 1997). We conclude that the Ptch1 transgene, although expressed in a subset of tissues under the control of the K14 promoter, is acting in the expected fashion as an inhibitor of SHH signaling during mouse development.
To explore the oncogenic function for Ptch1 in the setting of adult onset skin carcinoma development, we performed two-stage skin carcinogenesis by treating DMBA followed by TPA. Because of the difficulties in breeding the line 6 transgenic mice, we carried out the carcinogenesis studies with lines 7 and 8 transgenic mice on the FVBxC57BL/6 F1 genetic background (Wakabayashi et al., 2007). By 20 weeks post-treatment, both lines of K14PtchFVB transgenic mice and wild-type mice developed a similar number of papillomas, (Figure 2a, P=0.89 by Kruskal-Wallis test). It is known that a subset of benign papillomas induced by DMBA/TPA progress to malignant squamous (SCCs), or spindle cell carcinomas (SpCCs) (Burns, 1991; Oft et al., 2002). Thus, we monitored the mice for carcinoma development for up to 50 weeks post-TPA treatment. At this time point, 71% of line 7 K14PtchFVB mice and 84% of line 8 K14PtchFVB mice had developed carcinomas, compared to only 35% of wild-type mice (Figure 2b) (P=0.023 for line 7, P=0.0011 for line 8 by Fisher’s exact test). The carcinoma latency period was significantly shortened in K14PtchFVB mice compared to wild-type mice, which translated to lower carcinoma-free survival at 50 weeks (Figure 2c, P=0.0027 by Kaplan-Meier method). The skin carcinomas were either SCCs or SpCCs in both K14PtchFVB and wild-type mice, with no difference in terms of the frequency of histological subtypes (Table S1). A hallmark of DMBA-initiated skin tumors is activating H-ras mutations at codon 61(CAA to CTA, Gln to Leu) (Balmain, 1984; Quintanilla et al., 1986). A similar codon 61 CTA mutation frequencies was detected in carcinomas from K14PtchFVB (24/27; 89%) and wild-type (10/11; 91%) mice (Table S1). Taken together, these data support a positive role of Ptch1 expression in promoting sensitivity to carcinoma development initiated by mutation of Ras in adult mice.
We previously showed that the product of the mouse Tid1 tumor suppressor gene showed differential binding to PtchFVB and PtchB6 in 293T cells (Wakabayashi et al., 2007). In order to explore the possible mechanisms by which Ptch1 expression may interact with H-Ras signaling, we investigated the effects of H-RAS transfection or TPA treatment on, Tid1-Ptch1 binding and on activation of the MapK signaling pathway. Figure 3a shows that transfection of mutant HRAS into 293T cells led to accumulation of the GTP-bound form of RAS, but surprisingly had only a minor effect on activation of phospho-Erk (P-Erk) (Figure 3a, lanes 1–2). Co-transfection of Ptch1 (PtchFVB or PtchB6) in this assay led to only a modest but reproducible stimulation of P-Erk, and this was increased slightly by co-transfection of oncogenic V12HRAS. In parallel experiments, oncogenic HRAS had little influence on the binding of Tid1 to PtchFVB or PtchB6 as shown by co-immunoprecipitation assays (Figure 3b). However, TPA treatment that had a very strong effect on the activation of P-Erk (Figure 3c, second to bottom panels) destabilized the strong binding between Tid1 and Ptch1, especially for the PtchB6 variant (Figure 3c, lanes 3 and 4). This reduced level of binding, seen both at 30 min and 1hr after TPA treatment, was more similar to that seen with the FVB variant protein, suggesting that activation of HRAS/Mapk signaling may directly affect the stability of the Ptch1-Tid1 complex.
Loss of the inhibitory function of Ptch1 in the SHH signaling pathway results in up-regulation of Gli1, Gli2, and Ptch1 itself (Aszterbaum, 1999; Dahmane et al., 1997). To investigate whether Ptch1 overexpression in skin and skin SCCs gives rise to perturbations in SHH signaling, we performed qRT-PCR for the major target genes Ptch1, Gli1, Gli2, as well as Ccnb1 (cyclin B1) which was proposed to act together with Ptch1 in regulating the cell cycle (Barnes et al., 2001). While the average transcript level of total Ptch1 was elevated in skin (Figure 4a, P<0.0001 by t-test), and to a lesser extent also in skin cancers from K14PtchFVB mice (Figure 4b, P=0.0014 by t-test), Gli1 and Gli2 levels in skin (Figure 4a) and skin carcinomas (Figure 4b) were comparable between K14PtchFVB and wild-type mice. The average Ccnb1 transcript level was not different between the two genotypes, both in normal skin and in skin tumors (Figure 4a–b). We conclude that all of these known SHH pathway candidate genes are not significantly disrupted in expression levels, at least at the whole tissue level, in Ptch1 transgenic skins.
To investigate other possible consequences of Ptch1 overexpression in K14PtchFVB mice, we chose candidate genes on the basis of correlation in transcript levels with Ptch1 in normal skin from interspecific backcross mice (Quigley et al., 2009; additional data not shown). We performed qRT-PCR analysis for eight candidate genes – Bmp6, Bnc2, Hdgfrp3, Nt5e, Sox4, Gli3, Lphn1, Ncdn, all of which were significantly correlated with Ptch1 expression in normal skin. The expression analysis showed that only the Sox4 transcript level was significantly decreased in adult skins of K14PtchFVB mice compared to controls (Figure 4c, P=0.001 by t-test), but no difference was seen in skin carcinomas (Figure 4b). This reduced level of Sox4 was also seen, but to a lesser extent in, newborn skins from lines 6 and 8 K14PtchFVB mice (Figure 4d, P=0.023 by t-test).
To examine Ptch1 transgene expression at the cellular level in skin and skin carcinomas from transgenic mice, we carried out an immunohistochemical analysis using antibodies against the HA tag present at the C-terminus of the Ptch1 transgene. Cells positive for anti-HA were simultaneously positive for anti-Ptch1 by co-staining. First, we evaluated normal skin from the K14PtchFVB transgenic lines for the presence of HA/Ptch-positive epithelial cells. Among the three K14PtchFVB transgenic lines, a small number of HA/Ptch positive cells was only detectable in skin from line 6 K14PtchFVB mice (Figure 5a). These HA/Ptch positive cells were clustered in small foci rather than being scattered as single cells at the basal layer of the epidermis (Figure 5a). On the other hand, no HA-positive cells were identified in skin from lines 7 and 8 K14PtchFVB mice. It appears that while the Ptch1 mRNA level remains elevated in adult skin from K14PtchFVB mice compared to wild-type mice, significant protein expression is only detectable in clusters of epidermal cells during development.
Next we examined the presence of HA/Ptch positive cells in skin carcinomas from K14PtchFVB mice. Of 24 transgenic mouse carcinomas tested, 20 (83%) showed small clusters of HA/Ptch positive cells, but none were detected in carcinomas from wild-type mice (Table S1). These cells were mainly detected in SCCs, and if present in SpCCs, they were predominantly in areas of the tumor that showed squamous differentiation (Figure 5b–c). In the SCCs, the HA/Ptch positive cells were located at the superficial aspect of the spinous layer, near the interface with the keratinized/cornified cells (Figure 5b), particularly in tumor regions with strong keratin 14 expression (Figure 6a). The cell proliferation marker, Ki67, was frequently expressed in the skin cancer cells, but the HA/Ptch positive cells (78–100%) were largely negative for Ki67 expression (Figure 6b), indicating that these cells are likely non-proliferative. We also performed co-staining of the known skin stem cell markers, Sox2, Pax6, or Cd34, with the HA/Ptch tag (Takahashi, 2008; Li, 2005; Malanchi et al., 2008). Some HA/Ptch positive cells showed co-staining with Sox2 and Pax6, but the pattern was essentially random (Figure 6c), and these cells were largely negative for Cd34 (Figure 6c). We conclude that any HA/Ptch positive cells that are found in skin carcinomas are predominantly terminally differentiated cells with no significant proliferative and stem cell-like activity.
In this study, we have explored the paradoxical role of exogenous Ptch1 expression in promoting Ras-driven SCC development. Line 6 K14PtchFVB mice with the highest Ptch1 overexpression had severe developmental defects including growth retardation, loss of digits in the fore- and hindlimbs, as well as abnormalities of the eyes. The similarity between these phenotypes and those caused by germline mutations leading to loss of SHH pathway function (Chiang, 1996; Chiang, 2001; Zhang et al., 2001) indicates that expression of Ptch1 driven by the K14 promoter during development (Byrne, 1994; Kopan et al., 1989) results in perturbation of Shh signaling by either cell autonomous or non cell-autonomous mechanisms, leading to the observed major developmental abnormalities in developing epithelia.
Skin tumor susceptibility studies were performed on two transgenic lines (lines 7 and 8), in which Ptch1 transgene expression was relatively low. Nevertheless, these lines were found to be highly susceptible to chemically induced carcinomas of the skin in a two-stage carcinogenesis study. Gene expression analysis detected no changes in the major downstream effectors of SHH signaling such as Gli1 or Gli2 in adult skin, although it remains possible that single cells or clusters of Ptch1-positive cells in transgenic skin could express altered levels of these markers. Of the other candidate genes selected based on correlations with Ptch1 expression in heterogeneous mouse populations (Quigley et al, 2009), only Sox4 was expressed at lower level in the skins of K14PtchFVB mice. A previous study suggested that PTCH1 is one of the SOX4 transcriptional target genes (Scharer et al., 2009), and coexpression of Sox4 and Shh in the developing hair germ was observed (Kobielak et al., 2007). It is therefore possible that one consequence of Ptch1 overexpression is deregulation of stem cell dynamics through a feedback leading to altered Sox4 expression in K14PtchFVB mice. At the molecular level, in vitro studies using human 293T cells indicated that one possible mechanism by which Ptch1 and the Ras pathway interact is through modulation of Map kinase signaling. Strong activation of P-Erk by TPA led to disruption of the previously reported strong binding of the Ptch1 PtchB6 variant to mTid1 (Wakabayashi et al, 2007), a tumor suppressor protein originally discovered in Drosophila (Canamasas et al., 2003). It is possible that activation of Ras/MapK signaling alleviates the tumor suppressor functions of Tid1 through reduction of its interaction with Ptch. Further studies using conditional knock-out alleles will be required to determine whether similar mechanisms play a role in promoting the growth of Ras-initiated cells in mouse skin.
We also tested the possibility that the subpopulation of cells within carcinomas that continued to express the Ptch1 transgene might be involved in tumor maintenance, or express markers of skin stem cells. However, these cells were only found in very well differentiated tumors, or in highly differentiated cells within more aggressive tumors. They were negative for expression of cell cycle marker, suggesting that they do not display proliferative activity, and were also negative for markers such as Cd34, Pax6 and Sox2 that have been implicated in control of skin stem cells. These data are compatible with the interpretation that the Ptch1 transgene expression, although clearly acting to promote SCC development, is no longer required for the maintenance of these tumors and is silenced during tumor progression.
Our data are compatible with the hypothesis that the SHH pathway, when activated through loss of Ptch1 or activation of downstream effectors such as Smoothened (Smo) or Gli1/2, leads to increased susceptibility to BCC formation by promoting the cell fate decision leading to the appropriate cell of origin for these tumors. Overexpression of Ptch1 however promotes an alternative epidermal cell fate decision leading to increased SCC formation. The concept that developmental regulators that promote one particular lineage can also simultaneously suppress an alternative lineage is well known (Davidson 2010). Such a mechanism may have an important influence on susceptibility to different types of tumors in the same tissue, conferring susceptibility to one type but resistance to an alternative tumor arising from a different lineage. Polymorphisms in genes that influence alternative cell fate decisions may therefore play an important role in determining individual tumor susceptibility at an early stage of carcinogenesis.
All animal experiments were performed under the UCSF Institutional Animal Care And Use Committee (IACUC) approval. The original K14PtchFVB mice from lines 6, 7, and 8 were previously described (Wakabayashi et al., 2007). Lines 7 and 8 were maintained on the FVB/N background and crossed with C57BL/6 mice to generate F1 wild-type and K14PtchFVB mice for two stage chemical carcinogenesis. 23 line 7 K14PtchFVB and 26 line 8 K14PtchFVB mice along with 35 wild-type mice (combined from both lines) were used to induce chemical carcinogenesis. Mice were genotyped using PCR primers previously described (Wakabayashi et al., 2007). To induce skin carcinogenesis, a single dose of DMBA was topically applied at 8 weeks of age followed by twice-weekly application of TPA for 20 weeks as described (Balmain, 1984; Quintanilla et al., 1986). Papilloma number was recorded from 10 weeks post-DMBA up to 20 weeks and carcinoma development was monitored up to 50 weeks post-TPA treatment. Mice were sacrificed if they were moribund, if there were excessive tumor loads, if any single tumor exceeded 1.5 cm in diameter, or at the termination of the experiment. Tumors removed by surgical dissection were immediately snap-frozen in liquid nitrogen or fixed in 10% buffered formalin solution for further analysis. Paraffin-embedded tumor sections were stained with H&E for histopathological analyses.
Frozen skin and tumor specimens were powdered in liquid nitrogen with pestle and mortar and homogenized for DNA and RNA preparation. For RNA extraction, specimens were homogenized and dissolved in TRIzol Reagent (Invitrogen, Carlsbad, CA) and purified for RNA isolation according to manufacturer’s instruction. Genomic DNA was isolated using standard phenol/chloroform extraction following an overnight incubation at 55°C with Proteinase K (Sigma, St. Louis, MO) in a lysis buffer (50 mM Tris at pH 8.0, 100 mM EDTA, 100mM NaCl, 1% SDS).
To determine a specific H-ras mutation, CAA to CTA at codon 61, exon 2 of the H-ras gene was amplified using the primers previously described (Nagase et al., 2003). Amplified PCR fragments was digested with restriction enzyme XbaI (NEB, Ipswich, MA) at 37°C for 2 hours and electrophoresed in a 4% Nusieve 3:1 agarose gel (Lonza, Walkersville, MD).
293T cells were grown at 37°C in DMEM supplemented with 10% FBS, penicillin, streptomycin, and glutamine and maintained in a humidified atmosphere of 5% CO2. HA-tagged PtchFVB and PtchB6 constructs in pcDNA3.1 and pLXSP3 retrovirus expressing V12HRAS were previously described (Wakabayashi et al., 2007). For transient transfection or co-transfection, Lipofectamine 2000 was used according to the manufacturer’s instruction (Invitrogen, Carlsbad, CA)
Western blotting was performed as previously described (Wakabayashi et al., 2007). Antibodies ERK1/2 (137F5), P-ERK1/2 (20G11), Tid1 (RS13 and RS-11), H-Ras (C-20, sc-520), HA-tag (mouse monoclonal 6E2 and goat polyclonal ab9134), β-actin (from Sigma) were used. For TPA treatment, transfected 293T cells were incubated with 500nM of TPA for 30 min and 1 hour. For immunoprecipitation, 500ug of lysates were processed with HA Tag IP/Co-IP kit (Thermo Scientific, Waltham, MA) and Dynabeads Protein G IP kit (Invitrogen, Carlsbad, CA) according to manufacturers’ instructions. For Ras-GTP assay, cells were washed with ice-cold PBS and lysed in 1X MLB (Magnesium-containing Lysis Buffer) with protease and phosphatase inhibitor cocktails. Raf-RBD pull down was performed with Ras Activation Assay kit containing Raf-1 RBD agarose beads according to manufacturer’s manual (EMD Millipore, Billerica, MA).
Single strand cDNAs were synthesized from 1ug of DNase-treated total RNA from skins and skin carcinomas using SuperScript® III First-Strand Synthesis System (Invitrogen, Carlsbad, CA) according to manufacturer’s manual. Pre-designed TaqMan probes and primers for Ptch1 (Mm01306905), Gli1 (Mm00494654), Gli2 (Mm01293111), Ccnb1 (Mm00838401), Sox4 (Mm00486317), Bmp6 (Mm01332882), Bnc2 (Mm01266537), Nt5e (Mm00501917), Hdgfrp3 (Mm01324333), Gli3 (Mm00492345), Lphn1 (Mm00492345), Ncdn (Mm00449529), Actb (beta-actin, 4352933E), and Gapdh (Mm99999915) were used for qRT-PCR analysis (ABI, Foster City, CA). Amplification of cDNAs was carried out in triplicate for each sample in the ABI 7900HT system according to the manufacturer’s protocol. The normalized transcript level was determined by calculating dCt (delta Ct) values by subtracting mean Ct values of target genes with mean Ct values of housekeeping genes. Then, 1/dCt values were used to obtain representative values indicating higher values to higher transcript levels.
Skin and carcinomas specimens were fixed in buffered formalin solution and processed for paraffin embedding. Sections (5μm in thickness) were de-paraffinized with xylene two times for 10 minutes and processed in a pressure cooker for antigen retrieval using Triology solution (Cell Marques, Rocklin, CA). The sections were blocked in 10% donkey serum supplemented with 0.3% Triton X-100 for 1 hour and incubated with primary antibody (HA tags, goat, Abcam, Cambridge, MA) for 1 hour at room temperature. After rinse with PBS, the sections were incubated with FITC-conjugated secondary antibody for 1 hour at room temperature. For double staining, HA tag incorporated sections were further blocked in 10% goat serum supplemented with 0.3% Triton X-100 for 1 hour and incubated with rabbit Ptch1 (Abcam, Cambridge, MA), Ki67 (Neomarkers, Fremont, CA), K14 (Covance, Princeton, NJ), Sox2 (Abcam, Cambridge, MA), Pax6 (Abcam, Cambridge, MA), or Cd34 (rat, BD Pharmingen, San Diego, CA) antibodies. After rinse with PBS, the sections were incubated with Alexa 555-conjugated anti-rabbit or anti-rat secondary antibodies followed by DAPI staining. The section images were examined under the fluorescent microscope (Olympus BX60, Center Valley, PA).
These studies were supported by NCI MMHCC (National Cancer Institute Mouse Models of Human Cancer Consortium) grant 2U01 CA08422-06 to AB and Program Project Grant (PPG) 5P01AR050440 from the US NIAMS (National Institute Of Arthritis And Musculoskeletal And Skin Diseases), National Institutes of Health (NIH). AB acknowledges support from the Barbara Bass Bakar Chair of Cancer Genetics.
Conflict of Interest
The authors state no conflict of interest.