p63, a homologue of the tumor suppressor p53, is critical for the development and maintenance of squamous epithelia. p63 is specifically expressed in the basal layers of stratified epithelial tissues and is considered a specific marker for cells of this type. The role of p63 in tumorigenesis remains poorly defined. Numerous studies have highlighted the oncogenic potential of the predominant p63 isoform ΔNp63α; however, data suggest that other p63 proteins can act as tumor suppressors or alter the metastatic potential of tumors. ΔNp63α can act as a transcriptional repressor, but the link between the transcriptional functions of p63 and its biological role is still unclear. In this study, we used a loss-of-function approach to investigate the transcriptional programs controlled by p63. Disruption of p63 in squamous cell lines resulted in down-regulation of transcripts specifically expressed in squamous tissues and a significant alteration of keratinocyte differentiation. Interestingly, we found that disruption of p63 led to up-regulation of markers of nonepithelial tissues (mesenchyme and neural tissue) in both primary and immortalized squamous cells. Many of these up-regulated genes are associated with increased capacity for invasion and metastasis in tumors. Furthermore, loss of p63 expression was accompanied by a shift toward mesenchymal morphology and an increase in motility in primary keratinocytes and squamous cell lines. We conclude that loss of endogenous p63 expression results in up-regulation of genes associated with invasion and metastasis, and predisposes to a loss of epithelial and acquisition of mesenchymal characteristics. These findings have implications for the role of p63 in both development and tumorigenesis. (Cancer Res 2006; 66(15): 7589-97)
p63 is a homologue of the tumor suppressor p53 ( 1– 4). Unlike p53, which is inactivated in a majority of human cancers but is largely dispensable for normal development, p63 is critical for the development and maintenance of stratified epithelial tissues. p63−/− mice display gross developmental abnormalities; the most striking of these is a complete lack of all stratified epithelia and their derivatives, including epidermis, mammary glands, prostate, and other tissues ( 5, 6). This phenotype is recapitulated in zebrafish, in which disruption of p63 results in a lack of epidermal morphogenesis ( 7, 8). p63 is immunolocalized primarily in the basal compartment of epithelia, including epidermis, oral mucosa, cervix, vaginal epithelium, urothelium, prostate, and breast ( 1, 9, 10); it has been postulated that epithelial stem cells reside in these basal layers ( 11). Taken together, these data suggest that p63 is crucial for the development and maintenance of stem cells in stratified epithelia.
The p63 gene can be expressed as six different transcripts; all of these maintain a DNA-binding domain with significant homology but somewhat altered target specificity to that of p53 ( 12, 13). Due to alternate promoter usage, isoforms of p63 have different NH2 termini with either transactivating (TA) or dominant-negative (ΔN) activities; however, in epithelial cells, ΔNp63α is the predominant form expressed ( 9, 10, 12, 14, 15). The ΔNp63α protein can act as a transcriptional repressor in vitro and in animal models and can strongly oppose p53- or TAp63-mediated transactivation ( 1, 7, 8, 12).
The role of p63 in tumorigenesis is complex and remains poorly defined. Many studies have highlighted the oncogenic potential of ΔNp63α ( 16– 23). However, other data suggest that the p63 gene acts as a tumor suppressor ( 24, 25), although p63 is rarely mutated in human cancers like classic tumor suppressor genes. Even more recent data show that TAp63α, a transactivating isoform that induces apoptosis in many assays, can promote tumor progression and metastasis in mice ( 26). Clearly, further investigation is required to elucidate the role of p63 in tumor development and progression and to link the transcriptional functions of p63 proteins to these biological effects.
In this study, we used a loss-of-function approach to investigate the transcriptional programs controlled by p63. Disruption of p63 expression in squamous cell lines led to down-regulation of markers of squamous epithelium, and keratinocytes with reduced p63 expression failed to properly differentiate in vitro. Interestingly, loss of p63 expression in primary, immortalized, and transformed cells resulted in up-regulation of genes expressed primarily in nonepithelial tissues (mesenchymal and neural). Many of these up-regulated genes are associated with an increased propensity for tumor invasion and metastasis. These changes in gene expression were accompanied by a shift toward mesenchymal morphology and an increase in motility in cells with disruption of p63. These data suggest that ΔNp63α acts to maintain epithelial character in cells which express it, and that loss of ΔNp63α in tumors may predispose to invasion and metastasis.
Materials and Methods
Cell culture. The human keratinocyte cell lines HaCaT and HaCaT-RG were generously provided by P. Boukamp (German Cancer Research Center, Heidelberg, Germany); the squamous carcinoma cell lines SCC-1, SCC-6, SCC-17B, and SCC-74B were gifts from T. Carey (University of Michigan, Ann Arbor, MI); the squamous carcinoma cell lines SCC-012 and SCC-028 were kindly provided by D. Sidransky (Johns Hopkins University, Baltimore, MD). Human epidermal keratinocytes (HEK) and normal human dermal fibroblasts (NHDF) were obtained from the Vanderbilt Skin Disease Research core. HaCaT, HaCaT-RG, SCC-1, SCC-6, SCC-17B, SCC-74B, HeLa [American Type Culture Collection (ATCC), Manassas, VA], A-431 (ATCC), and NHDF were cultured in DMEM supplemented with 10% FCS and 1% penicillin-streptomycin. SCC-012 and SCC-028 cells were cultured in RPMI 1640 supplemented with 10% FCS and 1% penicillin-streptomycin. The A-549 human lung adenocarcinoma cell line (ATCC) was cultured in DMEM supplemented with 10% FCS, 10 μg/mL insulin, and 1% penicillin-streptomycin. All cells were cultured at 37°C with 5% CO2.
Cell transfection/infection and small interfering RNA. The following targeting sequences were used for small interfering RNA (siRNA): p63, 5′-AACAGCCATGCCCAGTATGTA-3′; p63-2, 5′-AAAGCAGCAAGTTTCGGACAG-3′; green fluorescent protein (GFP), 5′-AAGCTGACCCTGAAGTTCATC-3′. Targeting oligonucleotides for p63 and GFP were designed as previously described ( 27). pCEP-H1 ϕ, pCEP-H1 GFP, and pCEP-H1 p63 expression vectors were generated as previously described ( 28). HaCaT, HaCaT-RG, SCC-1, SCC-6, SCC-012, and A-549 cells were transfected using Fugene 6 (Roche, Indianapolis, IN). Cells were selected with hygromycin B 48 hours after transfection and harvested for RNA isolation as described below. pRetroSUPER-GFP and pRetroSUPER-p63-2 were constructed by phosphorylating targeting oligonucleotides by kinase treatment; complementary oligos were annealed together and subcloned into the BglII and HindIII sites of pRetroSUPER. Phoenix cells were transfected with pRetroSUPER plasmids, and retroviral supernatants were collected and used to infect HEKs 48 hours after transfection. HEKs were selected with puromycin after 48 hours of infection and were harvested for RNA isolation or used in migration assays as described below.
Western analyses. Cells were harvested by trypsinization and lysed as previously described ( 28). Western analysis was done as previously described ( 28) with the following primary antibodies: α-p63 monoclonal antibody Ab-1 (Oncogene Research Products, Calbiochem, La Jolla, CA) and α-β-actin polyclonal antibody I-19 (Santa Cruz Biotechnology, Santa Cruz, CA).
RNA isolation and microarray experiments. HEKs were harvested, and total RNA was isolated using the Aurum Total RNA Mini kit (Bio-Rad, Richmond, CA). HaCaT, HaCaT-RG, SCC-1, SCC-6, SCC-012, and A-549 cells were harvested, and mRNA was isolated as previously described ( 29). Each RNA sample was assayed for integrity using Agilent's Bioanalyzer microfluidic assay (Agilent Technologies, Palo Alto CA). Spectophotometric and fluorometric detection of protein and DNA contamination, respectively, were also used in the quality control process. Following quality control, the RNA was prepared for microarray analysis using the standard Affymetrix protocol (Affymetrix, Inc., Santa Clara, CA). Briefly, a total of 300 ng of mRNA were reverse transcribed to double-stranded cDNA using an oligo-dT primer coupled to a T7 promoter. In vitro transcription from the double-stranded cDNA was carried out using T7 polymerase and incorporating biotin-modified CTP and UTP ribonucleotides. The biotinylated cRNA (15 μg) was fragmented and hybridized to the Affymetrix GeneChips U133A and U133B (HaCaT-RG and SCC-1) and U133Plus2.0 (HaCaT, SCC-012, and SCC-6). Following hybridization for 16 hours at 45°C, hybridized cRNAs were washed and detected via streptavidin coupled to phycoerythrin using the Affymetrix 450 Fluidics Station and recommended protocols. Results were visualized by laser scanner (Affymetrix GeneChip Scanner 3000), and the image data were quantified to generate gene expression values and ratios of gene expression between the hybridized samples. Microarray data analyses were done using the GeneSpring software platform (Silicon Genetics, Redwood City, CA). Data were normalized on a per chip basis to the 50th percentile, then normalized on a per gene basis to the median signal, and finally normalized on a per gene basis to generate ratios of p63 siRNA to control signal for each cell line. As a starting point for analysis, genes up-regulated or down-regulated 2-fold in at least three of the five cell lines were selected for further investigation. A condition tree was generated from the cell line comparisons using the gene expression ratios of probes present on all arrays used and the standard correlation function. Gene trees were generated from up-regulated and down-regulated gene lists using the standard correlation function.
Quantitative reverse transcription-PCR. cDNA was made from the purified mRNA using Taqman Reverse Transcription Reagents (Applied Biosystems, Foster City, CA) according to the manufacturer's protocol; 100 ng of mRNA were used for each reaction. A 1:5 dilution of the resulting cDNA was used for all real-time reverse transcription-PCR (RT-PCR) experiments. Real-time PCR was conducted using an icycler (Bio-Rad) with the iQ SYBR Green Supermix (Bio-Rad) according to the manufacturer's instructions. Primers specific for each gene were designed using the Beacon Designer 3 software (Premier Biosoft, Palo Alto, CA). Primer sequences are available upon request. All primers were used at a concentration of 200 nmol/L. The cycling conditions for all genes were 95°C for 3 minutes followed by 40 cycles of 95°C for 10 seconds and annealing at 55°C to 61°C for 45 seconds, with data acquisition during each cycle. Melting curve analysis was conducted following PCR cycling to verify purity and quality of the PCR product.
Keratinocyte differentiation. HaCaT cells were transfected with plasmids encoding GFP or p63 siRNA and selected with hygromycin B as described above. Cells were grown to confluence and switched into differentiation media (serum-free DMEM with 1.4 mmol/L CaCl2) for 21 days. Differentiation medium was changed every 48 hours. Control (day 0) and differentiated cells were photographed and harvested for cornified envelope assays as described below.
Cornified cell envelope assay. HaCaT cornified cell envelopes were counted as previously described ( 30). Briefly, cells were trypsinized and resuspended in 1 mL of PBS plus 2 mmol/L EDTA. Ten-microliter aliquots were removed to count total cells. The remainder of the cells were centrifuged, resuspended in 1 mL of cell envelope dissociation buffer [2% SDS, 20 mmol/L DTT, 5 mmol/L EDTA, 0.1 mol/L Tris-HCl (pH 8.5)], and boiled for 5 minutes. Detergent-insoluble cell envelopes were cooled, centrifuged, and resuspended in 50 μL PBS. Cell envelopes were visualized by phase-contrast microscopy and counted with a hemacytometer.
Migration assays. HaCaT-RG and SCC-6 cells were transfected with plasmids encoding GFP siRNA or p63 siRNA and selected with hygromycin B as described above. HEKs were infected with pRetroSUPER expressing either GFP siRNA or p63 siRNA and selected with puromycin as described above. Cells were trypsinized and counted, and 2 × 105 HaCaT-RG and SCC-6 cells or 5 × 105 HEKs were plated onto 6.5-mm-diameter, 8.0-μm-pore size polycarbonate transwell filters in a 24-well plate (Corning Costar Corp., Cambridge, MA). Twenty-four hours after plating, cells on top of the membrane were wiped off with a cotton swab, and cells on the bottom of the membrane were fixed in methanol and stained with crystal violet as previously described ( 31). Cells in five, 200 × magnification, fields were counted per membrane.
Disruption of p63 expression leads to down-regulation of markers of squamous epithelium. To determine if p63 is necessary for the expression of markers of squamous epithelium, we disrupted p63 expression in a panel of transformed and nontransformed squamous cell lines: HaCaT, HaCaT-RG, SCC-012, SCC-6, and SCC-1. HaCaT is an immortalized keratinocyte cell line that is able to differentiate ( 32), whereas HaCaT-RG cells are a rapidly growing variant of the HaCaT cells that have lost the ability to differentiate; SCC-012, SCC-6, and SCC-1 cells are head and neck squamous carcinoma cell lines. These cell lines were transfected with a hygromycin-selectable vector expressing p63 siRNA or a control vector, then selected with hygromycin B, and harvested. Quantitative RT-PCR analysis showed that p63 expression was reduced by 50% to 85% by p63 siRNA transfection in a cell line–dependent manner (data not shown); these results were confirmed by Western blot ( Fig. 1A ). As previously reported, ΔNp63α is the predominant p63 protein expressed in these cell lines ( 9, 10, 12, 14, 15), and TAp63 isoform expression is not evident by Western analysis ( Fig. 1B). TAp63 could be detected by quantitative RT-PCR in these cell lines; however, the expression level was below the limit of quantitation for the assay. Semiquantitative calculations indicated that the ratio of ΔNp63 to TAp63 transcripts was >100:1 (data not shown). Therefore, although we cannot rule out TAp63 involvement in the processes described herein, it is likely that knockdown of ΔNp63α is critical to the transcriptional and biological effects observed in response to p63 siRNA. In addition, none of these cell lines examined express functional p53 (HaCaT and HaCaT-RG cells express mutant p53, whereas SCC-1 SCC-6, and SCC-012 cells do not express detectable p53); therefore, the effects of p63 siRNA can be considered p53 independent. The majority of these cell lines do not express detectable levels of p73, the third member of the p53 family. SCC-012 cells do express p73; however, when these cells were transfected with p63 siRNA, there was not a reduction in p73 expression (data not shown).
Control and p63 siRNA expressing cells were analyzed for global gene expression using the Affymetrix microarray platform. All gene expression data from microarray experiments have been deposited in the National Center for Biotechnology Information Gene Expression Omnibus (GEO, http://www.ncbi.nlm.nih.gov/geo/) and are accessible through GEO Series accession no. GSE4975. Data were normalized to the control sample for each of the five cell lines, and changes in gene expression in response to p63 disruption were examined across the panel of cell lines. Many of the transcripts down-regulated after p63 disruption represent genes that are markers of squamous epithelia ( Fig. 2 ). These include components of the basement membrane, keratinocyte structural proteins, skin-derived antimicrobial peptides, and components of the cornified cell envelope. Changes in a subset of these genes were validated by quantitative RT-PCR ( Table 1 ). Consistent with the microarray results, the genes shown in Table 1 were significantly down-regulated by disruption of p63 in multiple squamous cell lines. Cystatin A, keratin 6A, S100 proteins, and SPRR3 are keratinocyte structural proteins and components of the cornified cell envelope ( 33); hyaluronan synthase 3 is highly expressed in basal keratinocytes and is important for extracellular matrix formation in epidermis ( 34). We conclude from these data that p63 controls the expression of multiple markers of squamous epithelium, although the specific isoforms of p63 executing this function cannot be conclusively determined from these results.
Disruption of p63 expression prevents keratinocyte differentiation. One of the major physiologic roles of squamous cells is to form cornified envelopes and serve a barrier function upon terminal differentiation. To determine if the changes in gene expression we observed had biological implications for the function of squamous cells, we examined the ability of keratinocytes to differentiate after disruption of p63 expression. HaCaT keratinocytes were transfected with vectors expressing GFP or p63 siRNA, selected, grown to confluence, and induced to differentiate. Cells expressing GFP siRNA began to stratify and formed opaque, proteinaceous layers on top of the monolayer of cells, consistent with keratinocyte differentiation in vivo. ( Fig. 3A ). In contrast, HaCaT cells expressing p63 siRNA failed to show any visible indication of differentiation ( Fig. 3A). To quantify this observation, we counted the cornified envelopes formed by differentiating keratinocytes expressing either GFP or p63 siRNA. Induction of differentiation induced a robust increase in the number of detergent-insoluble cornified envelopes in cells expressing GFP siRNA; expression of p63 siRNA abrogated this increase in cornified envelope formation ( Fig. 3B). In addition, we observed up-regulation of loricrin and filaggrin, markers of keratinocyte differentiation, in control HaCaT cells but not cells in which p63 expression had been disrupted by siRNA (data not shown). We conclude from these data that p63 is required for both the expression of transcripts associated with stratified squamous epithelia and for the ability of squamous cells to execute a functional program of terminal differentiation. It has been previously reported that overexpression of ΔNp63α inhibits the terminal differentiation of keratinocytes ( 35). This finding can be explained if commitment to a squamous lineage and subsequent terminal differentiation are considered distinct cellular processes; down-regulation of ΔNp63α may be important for terminal differentiation, whereas p63 is required for the upstream events defining the overall squamous epithelial phenotype. These results are consistent with the reported phenotypes of p63−/− animal models ( 6, 7), and taken together, these studies suggest that p63 is necessary for the induction and maintenance of a squamous epithelial cell fate.
Disruption of p63 leads to up-regulation of mesenchymal and neural genes. Because ΔNp63α can act as a transcriptional repressor in vitro and in vivo ( 7, 8, 12), we next examined genes that were up-regulated by disruption of p63 expression in our microarray analyses. We found that many of the genes up-regulated by loss of p63 could be characterized as mesenchymal or neural-specific genes ( Fig. 4 ). These include structural components of connective tissue and muscle, nonepithelial adhesion molecules, neurotransmitter precursors, transporters, and receptors, and synaptic vesicle-trafficking components ( Fig. 4). A subset of these up-regulated genes was validated by quantitative RT-PCR ( Table 2 ). These genes are involved in a variety of biological processes, including adhesion, extracellular matrix formation, and tissue morphogenesis and development; a common trait, however, is that all these genes are expressed primarily in nonepithelial cells and tissues. Many of these genes are also associated with increased propensity for tumor cell invasion and metastasis. Some of the genes most consistently up-regulated by loss of p63 expression (N-cadherin, L1 cell adhesion molecule, periostin, and Wnt-5A) have been well characterized as promoting cell motility and tumor progression and metastasis ( 36– 41).
Loss of p63 expression leads to acquisition of mesenchymal traits in squamous cells. The genes differentially expressed in response to disruption of p63 expression suggested that loss of p63 in these cells had shifted the transcriptional program from that typically seen in squamous epithelial tissue to one more characteristic of a mesenchymal phenotype. Several of the genes up-regulated in response to disruption of p63 expression ( Table 2) are associated with mesenchymal phenotype and increased metastatic potential in cancer cells ( 37, 42). In addition, we found that reduced p63 expression led to a decrease in Wnt-4 expression and a concomitant increase in Wnt-5A expression in the majority of the squamous cell lines tested; this specific switch in expression of Wnt family members has been reported to be associated with epithelial to mesenchymal transition in squamous cell carcinomas of the head and neck ( 43). To further examine if p63 expression was associated with an epithelial phenotype, we compared relative levels of p63 protein and cell morphology in a panel of squamous cell lines ( Fig. 5A and B ). The origin of these cell lines were keratinocytes (HaCaT and HaCaT-RG), epidermoid carcinoma (A-431), and squamous cell carcinomas of the head and neck (SCC-1, SCC-6, SCC-012, SCC-17B, SCC-028, and SCC-74B) and cervix (HeLa). HEKs and primary cultures of dermal fibroblasts were used as references for epithelial and mesenchymal morphology, respectively. Squamous cell lines expressing ΔNp63α generally had morphologic features similar to primary keratinocytes and characteristic of epithelial cells, such as regular polygonal shape and well-defined cell-cell contacts ( Fig. 5A and B). In contrast, squamous cell lines lacking ΔNp63α expression seemed more spindle shaped and elongated, with increased numbers of processes and poorly defined cell-cell contacts; these are characteristics of mesenchymal cells, such as fibroblasts ( Fig. 5A and B). Interestingly, SCC-6 cells display an intermediate morphology with both regular, polygonal cells and numerous cells with elongated processes; this cell line expresses very low levels of ΔNp63α compared with other p63-expressing lines ( Fig. 5A and B).
Because cell motility is a hallmark of a mesenchymal phenotype, we determined if p63 expression affected the motility of transformed and nontransformed squamous epithelial cells. The motility of HaCaT, SCC-17B, SCC-6, and SCC-74B cells was measured using transwell migrations assays. Cells were plated into the top chambers of transwell membrane dishes; after 24 hours, cells on the bottom of the membranes were counted. Replicate platings of the cells in monolayer culture were also counted at 24 hours to control for cell proliferation. The motility of these squamous cell lines inversely correlated with expression of ΔNp63α; SCC-74B cells, which lack ΔNp63α expression, displayed the highest level of migration, whereas HaCaT and SCC-17B cells, which strongly express ΔNp63α, displayed low levels of migration ( Fig. 5C). SCC-6 cells, which express very low levels of ΔNp63α, had increased motility compared with the HaCaT and SCC-17B cells but significantly less than the SCC-74B cells. To test more directly if p63 expression affected cell motility, the migration of HaCaT-RG and SCC-6 cells transfected with vectors expressing GFP siRNA or p63 siRNA and HEKs infected with a retrovirus expressing GFP siRNA or p63 siRNA was investigated using the transwell assay. Knockdown of p63 with siRNA led to an ∼2.5- to 3-fold increase in migration in the HaCaT-RG and SCC-6 cell lines as well as the primary HEKs ( Fig. 5D). GFP and p63 siRNA expressing cells plated in parallel in monolayer cultures showed no significant difference in cell number 24 hours after plating, suggesting that differences in migration were not due to altered proliferation or plating efficiency (data not shown). These data show that loss of p63 expression in squamous cells leads to increased cell motility, a characteristic of cells of mesenchymal origin and tumors with high metastatic potential. These findings are consistent with the changes in gene expression observed in squamous cells with disrupted p63 expression and suggest that loss of p63 permits a shift from epithelial to mesenchymal character.
Finally, to extend observations concerning gene expression made in immortalized and transformed cell lines to p63 function in normal cells, we disrupted p63 expression in primary HEKs using a second, unique p63 siRNA sequence ( Fig. 5E). We found that loss of p63 in HEKs results in many of the same changes in gene expression observed in other squamous cell lines ( Fig. 5F). Not every transcriptional change seen in the cell line panel ( Tables 1 and 2) was observed in primary cells; perhaps, loss of p63 is insufficient to affect expression of these genes without the additional genetic and epigenetic alterations present in immortalized and transformed cells.
The definitive role of the p63 gene in tumor formation and progression is controversial. p63 is a putative oncogene; ΔNp63α is overexpressed in a number of epithelial cancers, often as a result of genomic amplification of the p63 locus, and numerous studies have highlighted the oncogenic potential of ΔNp63α ( 16– 23). A number of biochemical functions of ΔNp63α have been reported that could mediate ΔNp63α-mediated oncogenesis; these include activation of β-catenin signaling ( 22), up-regulation of Hsp-70 ( 44), repression of the proapoptotic protein insulin-like growth factor binding protein-3 ( 45), and suppression of p73-dependent apoptosis ( 46). In contrast to a role in promoting tumorigenesis, p63 expression is implicated as a positive prognostic factor in cancer progression and outcome. p63 expression is associated with favorable prognosis in patients with lung cancer ( 23); furthermore, loss of p63 expression in bladder cancer is associated with progression to more invasive and metastatic tumors ( 47, 48).
In this study, we show that loss of p63 expression in squamous cells results in an up-regulation of genes associated with tumor invasion and metastasis, including N-cadherin; we postulate that much of this up-regulation is due to loss of ΔNp63α-mediated repression. Consistent with this conjecture, it was recently shown that ectopic expression of the transactivating TAp63α in mice predisposes to up-regulation of N-cadherin expression, epithelial-to-mesenchymal transition, and metastasis of chemically induced tumors ( 26). Our findings that endogenous p63 suppresses expression of genes and phenotypic characteristics associated with mesenchymal cells and metastatic tumors provides a potential mechanism explaining the association between loss of p63 expression and poor prognosis in human cancers. It is possible that ΔNp63α acts to promote early steps in tumorigenesis by protecting cells from growth arrest and apoptosis, while at the same time acting as a metastasis suppressor by maintaining the epithelial character of cancer cells.
p63 is critical for the development of stratified squamous epithelia and their derivatives, such as mammary and prostate glands. However, the mechanism by which p63 executes this biological function is still in question. Here, we present evidence that p63 controls transcriptional programs that specify a squamous epithelial phenotype. Disruption of p63 expression in squamous cells led to down-regulation of genes expressed specifically in squamous epithelia and abrogated the ability of keratinocytes to form cornified envelopes and differentiate. A role for the p63 gene in definition of a squamous epithelial phenotype is consistent with the phenotype of p63−/− mice, as well as a recent report that ectopic expression of p63 in single layered epithelium can drive a transition to a stratified squamous phenotype ( 49). Strikingly, we found that loss of p63 expression in squamous cells also led to the up-regulation of genes normally associated with mesenchymal and neural tissues and an increase in mesenchymal appearance and motility. p63 deficiency can also lead to cellular senescence; Keyes et al. used both germ line and conditional p63 knockout mouse models to show that loss of p63 results in the activation of programs of senescence ( 50). It is reasonable to posit that loss of p63 in the context of intact checkpoint signaling in mouse epidermis promotes irreversible arrest and senescence due to alterations in the transcriptional programs we describe in combination with potential reactivation of inappropriate developmental programs.
In adult tissues, p63 is immunolocalized to sites of epithelial-mesenchymal apposition: these include the basal layer of epidermis and other stratified squamous tissues and the basal cells of complex glandular structures, such as the breast, prostate, lacrimal, and salivary glands ( 10). In addition, the phenotype of p63-deficient animals indicates that p63 plays a role in developmental processes that rely heavily on proper epithelial-mesenchymal interactions ( 5, 6). p63 is necessary for limb formation in both mice and zebrafish, presumably due to its role in maintenance of the apical ectodermal ridge, a site of complex epithelial-mesenchymal crosstalk ( 5– 8). Furthermore, both mice and humans with defects in p63 function have craniofacial abnormalities, including clefting of the lip and palate ( 5, 6, 51). Morphogenesis of the palate and facial structure involves coordinated communication between epithelial and mesenchymal components during the process. The discovery that p63 promotes an epithelial phenotype and suppresses mesenchymal specification in squamous cells provides insight to its role in epithelial-mesenchymal interactions. It is possible that p63 helps to define the line of demarcation between epithelium and stroma by specifying cells in which it is expressed as epithelial, while still allowing these cells the plasticity necessary for physiologic and pathologic epithelial to mesenchymal transitions. Continued identification of the target genes and protein interactions of ΔNp63α will provide insight to the complex role of p63 in development and tumorigenesis.
Grant support: NIH grants CA70856 and CA105436 (J.A. Pietenpol), NIH training grants GM073407 and CA009385 (C.E. Barbieri), grants and ES00267 and CA68485 (Core services).
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
We thank Dr. Matthew Westfall for construction of the ΔNp63α adenovirus, the members of the Pietenpol laboratory for critical reading of the article and helpful discussions, and the members of Resource for technical assistance and helpful discussions.
Note: All microarray experiments were done in the Vanderbilt Microarray Shared Resource.
- Received June 2, 2006.
- Accepted June 5, 2006.
- ©2006 American Association for Cancer Research.