Frequent Alteration of p63 Expression in Human Primary Bladder Carcinomas

p53 is the most frequently mutated tumor suppressor gene identified in human cancers (1). Tumor suppression functions of p53 stem, in part, from its capabilities to induce cell cycle arrest in late G₁ and/or apoptosis in response to genotoxic stress and hypoxia, and mutational inactivation of p53 is associated with an increased risk of tumorigenesis(2). Recently, two members of the p53 family,termed p73 and p63/p40/p51/p73L/KET (hereafter referred to as p63) have been identified at 1p36.3 and 3q27–29, respectively (3, 4, 5, 6, 7, 8). p73 and p63 share remarkable sequence identity to the DNA-binding, transactivation, and oligomerization domains of p53, but contain variable NH₂- and COOH-terminal extensions. When overproduced, both p73 and p63 can induce G₁ cell cycle arrest and/or apoptosis in a p53-like manner and activate the transcription of p53-responsive genes, such as p21^Waf1, suggesting that they might also be tumor suppressors (3, 4, 6, 9).

However, there are currently no genetic evidences that inactivation of p73 is required for transformation or malignant progression of human tumors, except recent reports of epigenetic silencing of p73 by aberrant 5′CpG island methylation in specific types of hematological malignancies such as acute lymphoblastic leukemia,lymphoblastic lymphomas, and Burkitt’s lymphoma (10, 11). Several studies also showed more intense and/or biallelic expression of wild-type p73 in various types of tumors than in normal tissues, which suggested that overexpression of p73 rather than as tumor suppressor may contribute to the tumorigenesis (12, 13). Recently, we also reported an elevated and biallelic expression of p73 in bladder carcinomas, which argues that p73 does not play a role as a tumor suppressor in bladder carcinogenesis (14).

p63 encodes multiple products with transactivating,death-inducing, and dominant-negative activities, which are derived from a single gene with two promoters (TAp63 and ΔNp63) and at least three alternative splicing of the transcripts (α, β, and γ; Refs. 4, 5, 6, 7). TAp63 isotypes with the acidic NH₂-terminal transactivating domain can activate transcription of p53 target genes such as p21^Waf1, whereas ΔNp63 isoforms without the transactivating domain can act as dominant-negative factors toward transactivation by p53 and p63(4, 6). p63 is highly expressed in proliferating basal cells of epithelial layers, including epidermis, cervix, urothelium,and prostate, and the major p63 isoforms in these basal cells lack the transactivating domain (4, 6). Recent studies revealed that mutational alteration of p63 is uncommon in human cancer cell lines and tumors (15, 16). However, it was demonstrated that expression of p63 is low or absent in a subset of lung cancer and ΔNp63 transcript is dominantly expressed in cell lines with high levels of p63expression (17). Expression of TAp63γ was also found to associate with tumor growth in cervical carcinogenesis,and numbers of the cells expressing ΔNp63 and their distribution showed a correlation with anaplasia in squamous cell carcinoma (18, 19).

Although the genomic imbalance at 3q27–29 has not been directly implicated in human cancers, several genetic studies using microsatellites and comparative genomic hybridization demonstrated frequent loss of heterozygosity (3) or amplifications at several regions of chromosome 3 in bladder tumors (20, 21). In the present study, we performed expression and mutation analyses of p63 in 63 bladder specimens, including 47 primary carcinomas and four cell lines, to investigate the potential involvement of p63 alteration in the pathogenesis of bladder cancer. Here, we show that genomic deletion or mutations of p63 is uncommon in primary bladder carcinoma, but its altered expression might contribute to the progression of bladder tumors.

Forty-seven carcinoma and 12 noncancerous bladder tissue specimens,including six matched sets, were obtained from 47 bladder cancer patients and 6 noncancer patients by surgical resection in the Kyung Hee University Medical Center (Seoul, Korea). Tissue specimens were snap-frozen in liquid N₂ and stored at −70°C until used. Four human bladder carcinoma cell lines (J82, T24, HT1197,and HT1376) were obtained from Korea Cell Line Bank (Seoul National University, Seoul, Korea) or American Type Culture Collection(Manassas, VA). Extraction of total cellular RNA and cDNA synthesis were performed as described previously (22). Genomic DNA was extracted from the same cells of the tissues from the DNA phase after RNA was extracted.

For quantitative evaluation by PCR, we initially performed the PCR reaction over a range of cycles (24, 27, 30, 33, 36, 39, and 42 cycles). Diluted cDNA (1:4; 12.5 ng/50 μl_iPCR reaction) undergoing 27–36 cycles was observed to be within the logarithmic phase of amplification and yielded reproducible results with the primers used for p63, p40–1 (sense,5′-CCTGGACGTATTCCACTGAACT-3′) and p40–2 (antisense,5′-CGCTTCGTACCATCACCGTTCT-3′), and an endogenous expression standard gene, GAPDH (14). For isoform-specific quantitation of TAp63 and ΔNp63,primers TAp63–2 (sense, 5′-GACCTGAGTGACCCCATGTG-3′) and p63–2(antisense, 5′-TCTGGATGGGGCATGTCTTTGC-3′) and primers ΔNp63–1 (sense, 5′-TGCCCAGACTCAATTTAGTGAG-3′) and p63–2(see above) were used, respectively. PCR was performed for 36 cycles at 95°C (1 min), 60–63°C (0.5 min), and 72°C (1 min) in 1.5 mm MgCl₂-containing reaction buffer (PCR buffer II; Perkin-Elmer Corp.). Ten microliters of RT-PCR³products were resolved on 2% agarose gels. Quantitation of expression levels was achieved by densitometric scanning of the ethidium bromide-stained gels. Absolute area integrations of the curves representing each specimen were then compared after adjustment for GAPDH expression. For quantitative DNA/PCR analysis of the p63 gene, 200 ng of genomic DNA were used for amplification of exon 5 of the gene with an intron-specific primers p63-E5S (sense,5′-TCTCCTTCCTTTCTCCACTGGC-3′) and p63-E5AS (antisense,5′-TGCCCACAGAATCTTGACCTTC-3′). The GAPDH gene was used for an endogenous control for quantitative DNA/PCR. Integration and analysis were performed using Molecular Analyst software program(Bio-Rad, Hercules, CA).

Nonisotopic RT-PCR-SSCP analysis was performed as described previously(23). The p63 transcript was amplified with six sets of primers that were designed to cover the entire coding region of the gene. Sequences of the primers used for our PCR-SSCP analysis will be obtained on request. The PCR products of over 300 bp in length were digested with endonuclease(s) to increase the sensitivity of SSCP analysis. Twenty microliters of PCR products were mixed with 5 μl of 0.5 N NaOH, 10 mm EDTA, 10 μl of denaturing loading buffer (95% formamide, 20 mm EDTA,0.05% bromphenol blue, and 0.05% xylene cyanol), and 15 μl of ddH₂O. After heating at 95°C for 5 min, samples were loaded in wells precooled to 4°C. SSCP was performed using 8%nondenaturating acrylamide gels containing 10% glycerol at 4–8°C or 18–22°C.

To assess activation of p63 expression, four bladder carcinoma cell lines were plated in 6-well tissue plates 24 h before treatment. 5-Aza-2′-deoxycytidine (Sigma Chemical Co., St. Louis, MO) was added to the fresh medium at concentrations of 1.0 and 2.0 μm in duplicate, and cells were harvested after 3 and 5 days.

To explore the candidacy of p63 as a suppressor in bladder carcinogenesis, we initially evaluated mRNA expression levels of two major isoforms of p63, TAp63, and ΔNp63 in 12 noncancerous bladder tissues by quantitative RT-PCR using isoform-specific primers. As shown in Fig. 1, expression of TAp63 mRNA was easily detectable in all noncancerous tissues we examined, whereas expression of ΔNp63 was extremely low or not detected. No significant variation in expression levels of p63transcripts was recognized among the specimens (TAp63/GAPDH,0.83–1.14; ΔNp63/GAPDH, 0.00–0.28). On the basis of this observation, we arbitrarily classified expression levels less than a half (<0.49; levels 0–2 for TAp63) and more than 2-fold (>0.38; levels 3–5 for ΔNp63) of noncancerous means as abnormal expression. Of four bladder carcinoma cell lines examined, three (J82, T24, and HT1197) and two (HT1197 and HT1376) were identified to express abnormally low TAp63 and high ΔNp63 mRNA,respectively (Table 1).

We next evaluated expression levels of TAp63 mRNA in 47 primary carcinomas. Quantitative RT-PCR analysis revealed that 25(53.2%) carcinomas expressed abnormally low level of TAp63,and in 5 of these TAp63 transcripts were nearly undetectable(Fig. 2,A and Table 1). In addition, tumor-specific reduction of TAp63 was found in two of six matched sets (Fig. 2,B). Interestingly, abnormally low expression (levels 0–2)of TAp63 showed a correlation with tumor stage and grade(Table 1). Whereas abnormal reduction of TAp63 was observed in 28.0% (7 of 25) of superficial tumors (Ta-T1), 81.8% (18 of 22) of invasive tumors (T2-T4) were identified as low or no TAp63expressors. Low expression of TAp63 was also found in 38.9%(7 of 18) of grade I, 43.8% (7 of 16) of grade II, and 84.6% (11 of 13) of grade III tumors. Thus, these observations demonstrate that altered expression of TAp63 mRNA is a frequent event in bladder carcinogenesis and suggest that inactivation of TAp63 might contribute to the malignant progression of primary bladder tumors.

Previous studies using ectopic induction of p63 suggested that ΔNp63 could act as a dominant-negative factor toward the G₁ cell cycle arrest and apoptosis induction by TAp63 or p53 (4, 6). In this context, we investigated the possible involvement of abnormal ΔNp63elevation in bladder tumorigenesis. As shown in Fig. 2, abnormally high expression (levels 3–5) of ΔNp63 was detected in 63.8% (30 of 47) of carcinomas and tumor-specific increase of ΔNp63 was observed in three of six matched sets. However, unlike TAp63, abnormal overexpression of ΔNp63 was not associated with tumor stage and grade (Table 1).

If ΔNp63 acts as a dominant-negative factor toward the growth inhibition or apoptosis by TAp63, alteration of TAp63 and ΔNp63 would be expected to be mutually exclusive in cancer cells. Whereas 19 (86.4%) of 22 normal expressors of TAp63 showed abnormally high ΔNp63, 11 (44.0%) of 25 tumors with low TAp63 expression were identified as high ΔNp63 expressors (Table 2). Likewise, although 14 (82.4%) of 17 normal expressors of ΔNp63 showed abnormally low TAp63,only 11 (36.7%) of 30 tumors with high ΔNp63expression were identified as low TAp63 expressors. Similarly, two (J82 and T24) of the three cell lines with low TAp63 expressed normal levels of ΔNp63, whereas the HT1376 cell line with a normal level of TAp63 expressed abnormally high ΔNp63 (Fig. 1). However, simultaneous alteration of TAp63 and ΔNp63 was also found in 11 (23.4%) of the 47 carcinomas and in one (25.0%) of the four cell lines. Thus, a mutually exclusive expression pattern of TAp63 and ΔNp63 was recognized in general, whereas a subset of tumors showed altered expression of both isoforms.

To evaluate expression of alternatively spliced variants p63α, p63β, and p63γ, we performed PCR amplification of the COOH-terminal portion of p63 transcripts using variant-specific primer sets. p63α and p63γ transcripts were easily detectable in all p63-positive tissues, and no significant difference in expression levels of these two variants was recognized. In contrast, expression of p63β transcripts was not detected under our experiment conditions (data not shown).

To investigate the allelic deletion or mutational alteration of the p63 gene, we performed quantitative DNA/PCR and RT-PCR-SSCP analyses of p63 for 47 primary carcinomas, four cell lines,and 10 noncancerous tissues. Compared with normal tissues, no significant difference was detected in p63 gene levels in tumors, indicating that abnormal expression of p63 mRNA is not associated with allelic alteration of the gene (Figs. 1 and 2). For SSCP analysis, the entire coding region of the transcripts was amplified using six different sets of primers, digested with several different restriction endonucleases, and subjected to electrophoresis under two different running conditions. However, we failed to detect any types of mutation leading to amino acid substitutions or frameshifts, whereas 36.2% (17 of 47) of the same set of primary carcinomas was identified to carry p53 mutations. Thus, this result indicates that, unlike p53, mutational alteration of p63 is not a main genetic event in the bladder carcinogenesis. In addition, no correlation was identified between altered expression of p63 isoforms and p53 status in tumors we analyzed (Table 2). To further define the possible association of p63 with the p53 pathway, we examined expression of p53 target genes such as p21^Waf1, MDM2, and 14–3-3ς. Whereas low expression of p21^Waf1 mRNA was more frequently observed in tumors with p53 mutation (11 of 17, 64.7%) than tumors with wild-type p53 (4 of 30, 13.3%), alteration of TAp63 or ΔNp63 showed no association with mRNA expression of p53 target genes (Table 2). Taken together, these data demonstrate that alteration of p63expression does not correlate with the mutational status of p53 and its target gene expression in bladder carcinomas.

To explore whether abnormal methylation is associated with the altered expression of p63, we treated the four cell lines with a demethylating agent, 5-Aza-2′-deoxycytidine, and analyzed expression levels of TAp63 and ΔNp63. Whereas expression of TAp63 mRNA was induced in T24 by 5-Aza-2′-deoxycytidine treatment, up- and down-regulation of both TAp63 and ΔNp63 transcriptions were observed in HT1376 and HT1197, respectively, and no change in p63 levels was detected in J82 (Fig. 3). These results suggest that abnormal hypermethylation would be one of the causes for the altered expression of p63, but other factors might be implicated in the reciprocal regulation of TAp63 and ΔNp63 in bladder carcinoma.

To further characterize the possible relationship of p63alteration with disease progression, we performed Kaplan-Meier survival analysis for 36 patients whose follow-up history was available. As shown in Fig. 4, expression of TAp63 (P = 0.06)but not ΔNp63 (P = 0.88) was recognized to correlate with cumulative survival of the patients after operation.

Despite its high degree of structural similarity to p53 and a growth inhibitory role by epigenetic control of expression, there are currently few evidences that inactivation of p63 is required for transformation or malignant progression of human tumors. In this study, we first demonstrate that altered expression of TAp63and ΔNp63 is a frequent event and might contribute to the progression of bladder tumors. Our study also suggests that alteration of p63 expression might be caused by more complicated mechanisms, including epigenetic factors rather than allelic or mutational alteration of the gene. It has been previously reported that expression of p63, predominantly ΔNp63, is highly expressed in basal cells with high proliferative potential and is absent from the cells that are undergoing terminal differentiation (4, 6). A recent study also showed that p63 expression is absent or low in a considerable proportion of lung cancers and ΔNp63 transcript is dominantly expressed in cell lines with high levels of p63 expression, whereas only 1 of 44 cell lines but none of 45 primary tumors has p63mutation (15). In squamous cell carcinoma of the skin, the number and distribution of cells expressing ΔNp63 was found to correlate with anaplasia(19). In this context, it is noteworthy that a mutually exclusive alteration of TAp63 and ΔNp63 was recognized in a substantial fraction of bladder tumors we analyzed, and low TAp63 rather than high ΔNp63 expression showed a correlation with tumor stage and grade. However, our finding of simultaneous alteration of TAp63 and ΔNp63 expression in a subset of tumors also raises the possibility that two isoforms of p63 might carry their specific roles in bladder carcinogenesis. This hypothesis is partly supported by our observation that altered expression of TAp63 is correlated with tumor progression and patient survival but no further increase of ΔNp63 alteration is observed in advanced tumors, suggesting its possible contribution to an initial step of bladder tumorigenesis. Further study will be required to gain understanding for the biological significance of isoform-specific roles of p63 in human tumorigenesis.

It has been hypothesized that disruption of normal p53 function results in compensatory or deleterious up-regulation of other members of the p53 gene family or that overexpressed ΔNp63 may bind p53 DNA target sites in a competitive manner or mimic mutant p53, thus act as a dominant-negative factor in wild-type p53-carrying tumor cells (3, 4, 5, 6). However, recent studies demonstrated that p63 transactivates the p53 target genes but the degree of the transactivation by p63 differed from that by p53, and the tumor-derived p63 missense mutations were found to retain their ability to transactivate the MDM2 and/or the Bax promoter but not the p21^Waf1 promoter, indicating that the cellular signal on p63 cross-talks partially, but not completely, with that of the p53 pathway (24, 25). It has been also observed that p53 in cancer cells was not able to interact with endogenous or exogenous p63 or p73 via their respective oligomerization domains while the multiple isoforms of p63, as well as those of p73, are capable of interacting via their common oligomerization domain (26). Moreover, recent studies showed that p53-inactivating viral oncoproteins such as SV40 T antigen,human papilloma virus E6, and adenovirus E1B do not directly interact with p63 and do not inhibit p63-mediated transcription, suggesting that unlike p53, p63 does not seem to be a necessary target in virus-induced cell transformation (27, 28). Consistent with these observations, we identified no association of TAp63 or ΔNp63 expression with the mutational status of p53 and the expressions of p53 target genes such as p21^Waf1, MDM2, and 14–3-3ς in primary bladder tumors and cell lines. Our preliminary work also showed that transient overexpression of wild-type p53 or p73, treatment with a DNA-damaging agent(etoposide), or cell growth under growth factor-deprived culture condition did not significantly affect the expression levels of both TAp63 and ΔNp63 mRNA in bladder carcinoma cell lines (data not shown). Taken together, these results suggest that p63 may not exert a role comparable with p53 and be involved in an unknown tumor suppressor pathway distinct from that of p53.

Recent studies demonstrated that mutations in the p63 gene are rare in human cell lines and tumors. Hagiwara et al.(15) reported that only 2 of 54 human cell lines have either heterozygous mutations or polymorphisms in the putative DNA binding domain of p63. Other investigators identified only four distinct missense mutations of p63 after screening>200 tumors and cell lines (5, 16). In the present study,we also failed to detect allelic deletion or any types of mutation leading to amino acid change of p63 in bladder cancers whereas 36.2% of the same tumors were identified to carry p53 mutations. This result indicates that mutational alteration of p63 may be not a main genetic event in the bladder carcinogenesis and suggests that p63 is unlikely to be a tumor suppressor gene that conforms to a two-hit model of tumorigenesis.

In conclusion, the evidences we obtained here clearly demonstrate that p63 is not a target of sequence alterations in bladder carcinogenesis, but abnormal reduction of TAp63 and/or overexpression of ΔNp63 are frequent and might contribute to the progression of bladder cancer, although the functional significance of p63 alteration was not defined in the present work. Additional studies will be required to elucidate the roles of altered p63 expression in growth and apoptosis of bladder epithelial cells.