p53 Mutation Arising in Arg72 Allele in the Tumorigenesis and Development of Carcinoma of the Urinary Tract¹

The tumor-suppressive functions of p53 stem, in part, from its capacity 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 increased risk of tumorigenesis (1). Oncogenic studies have revealed that missense mutations of the p53 gene lead to critical structural changes in the p53 product and are among the most common genetic alterations in human cancers.

It is known that a p53 sequence polymorphism found in the general population that results in either Arg³ or Pro, or at residue 72 produces marked change in the structure of p53 protein and thereby affects migration on SDS-PAGE (2). Several reports have noted epidemiological differences in the prevalence or prognostic significance of this p53 polymorphism in certain cancer types (3, 4, 5, 6, 7). Recent reports suggested that the allele encoding Arg72 is a significant risk factor for the development of cervical and esophageal cancers associated with human papillomavirus (8, 9, 10). Two recent reports additionally indicated that the Arg72 allele is preferentially mutated and retained in various human tumors (11, 12). However, the biopathological significance of these variants of p53 polymorphism is unclear. To determine whether a common polymorphic residue of Arg72 within p53 could represent a risk factor for mutations in p53, we analyzed the relationship between p53 mutations and genotypes of its codon 72 in human TCCs of the urinary tract. We next determined the identity of the deleted allele in the TCCs arising in individuals with Pro/Arg heterozygotes that had undergone LOH at the TP53 locus. From these data, the relationship with various clinical and pathological factors were also analyzed.

Tumor tissue specimens were obtained from 141 cases of human TCCs of the renal pelvis, ureter, and urinary bladder, which were obtained by nephroureterectomy (58 cases), radical or partial cystectomy (51 cases), and transurethral resection (49 cases) between 1998 and 2002 at the Department of Urology of Kochi Medical School and the Divisions of Urology of both Kochi Takasu Hospital and Fujisaki Hospital. Cases with preoperative radiation or chemotherapy were not selected. The patients included 106 men and 35 women. The median age of the patients was 69.4 years (range 55–86). The Institutional Review Committee for clinical investigation reviewed and approved the collection of pairs of tumor and corresponding normal mucosa or peripheral blood tissue samples for genetic analysis. Paired normal and tumor tissues were processed in a similar fashion at all three of the institutions from which genomic DNA was isolated.

Genomic DNAs isolated from TCCs have been analyzed by direct DNA sequencing concentrating on exons 4–9 of the p53 gene according to our procedure described previously (13). Primers were prepared to amplify exons 4–9, including each splicing site, of the p53 gene for direct DNA sequencing analysis. The sequence of each set of primers for PCR and the conditions of sequencing analysis of p53 were as described previously (13). Using genomic DNAs from each normal tissue, the identity of the amino acid at codon 72 in exon 4 of p53, Pro (CCC) or Arg (CGC), was also determined by the sequencing. In the determination of genotype of codon 72 polymorphism, the Pro or Arg type was scored if the area under the guanine peak was reduced to <50% of its cytosine allele (this indicate Pro type) or if the area under the cytosine peak was reduced to <50% of its guanine area (this indicate Arg type). For each primer pair, sense primer only was fluorescently labeled at the 5′ end. Template genomic DNAs were amplified by PCR using a thermal cycler (Takara Co Ltd., Ohtsu, Japan) with Taq polymerase (Perkin-Elmer, Norwalk, CT). Direct genomic DNA sequencing of PCR products was performed using Automated Laser Fluorescent ALF sequencer (Pharmacia LKB Biotechnology AB, Uppsala, Sweden) on a 6% polyacrylamide denaturing gel. In the cases that detected p53 mutation showing heterozygosity (Pro/Arg) at codon 72, genomic DNA fragments from the tumors containing exon 4 to each mutated exon detected were amplified by PCR, successively subcloned into pGEM-T easy plasmids (Promega, Madison, WI), and transfected into JH109 cells to identify the codon 72 genotype of the mutant allele of p53 (14). At least five JH109 clones of each case were expanded and subjected to sequence analysis.

LOH was determined using a p53 microsatellite marker (TP53) amplified by PCR as described previously (15). LOH was defined as a reproducible reduction of ≥30% of either the smaller or larger allele. Samples that were homozygous for the marker were defined as noninformative and excluded from calculations for LOH frequency.

The χ² test at the 5% level was used to determine statistical associations between p53 mutations and the status of polymorphisms of codon 72. Analysis of associations with clinicopathological factors was also performed using the χ² test.

We analyzed genomic DNA from 141 individuals with TCCs for mutations in p53 by DNA sequencing and identified p53 mutations in 49 TCCs examined (34.8%; 49 of 141). Of 49 mutations of p53, 4 were nonsense mutations (3 at codon 192: C/T-Gln/stop; 1 at codon 213: C/T-Arg/stop), 1 had insertion of 7 bp at codon 299, 1 had deletion of 2 bp at codon 209, and 1 had deletion of 1bp at codon 279; the remaining 42 were missense point mutations leading to amino acid substitutions. Most of the mutations identified in this study would be predicted to alter the conformation of p53.

Direct sequencing of p53 at exon 4 also yielded the frequencies of Pro/Pro, Pro/Arg, and Arg/Arg alleles at codon 72. As shown in Table 1, the allele frequencies in our 141 TCC group were 24 (17.0%) with homozygosity for Pro, 53 (37.6%) with heterozygosity for Pro/Arg, and 64 (45.4%) with homozygosity for Arg nearly equal to those in a group of 110 healthy Japanese women (16) and 58 normal esophageal mucosa (9) reported previously. The frequencies of the two alleles in a series of 49 TCCs with p53 mutation were compared with those in a series of 92 TCCs with wild-type p53. There were no significant differences in p53 codon 72 genotypes between the two groups (χ² = 0.678, P = 0.713).

Of 49 TCCs with p53 mutations, 40 cases were performed PCR genotyping for LOH status of the tumor DNAs using a p53 microsatellite marker (TP53). Table 2 summarizes the correlation between mutant p53 status and codon 72 polymorphism or LOH detected in 40 TCCs. Among 40 TCCs, 11 tumors mutant for p53 arose in individuals with heterozygosity (Pro/Arg). Additional sequencing of pGEM-T easy plasmid clones with genomic DNAs of tumor tissues containing each p53 mutated allele revealed that of the 11 TCCs, the mutant allele was Arg in 8 cases and Pro in 3 cases (Fig. 1, A and B). There was a bias to mutate and express the Arg allele in TCCs examined. We next determined the identity of the deleted allele in the 11 TCCs arising in heterozygosity (Pro/Arg) individuals that had undergone LOH at the p53 locus. LOH at TP53 was detected in 6 of the 11 TCCs. The lost allele was Pro in 5 of these TCCs and Arg in 1. Thus, in most TCCs in which p53 was mutated, there was a bias to mutate and express the Arg allele arising in Pro/Arg heterozygotes, although the number of samples is too small to reach statistical significance (selected allele in p53 mutation, Arg72:Pro72 = 8:3; lost allele, Arg72:Pro72 = 1:5).

Of the p53 -mutated 40 TCCs examined, the mutant allele was Arg in 29 cases and Pro in 11 cases. Table 3 summarizes the relationship between the p53 mutations arising in Arg72- or Pro72-containing alleles and clinical and pathological features of the 40 TCCs examined. A significant association was detected between p53 mutations within Arg72-containing allele and tumor stage (χ² = 5.320, P = 0.021). In contrast, no significant correlation was detected between p53 mutations within Arg72-containing allele and any other clinical or pathological parameters, including age, sex, site of tumor, and tumor grade.

The exon 4 is one of the largest exons of p53, lies close to the central region, and is necessary for DNA-specific binding (17). A common p53 polymorphism, encoding either Pro or Arg at residue 72, produces marked change in the structure of p53. This polymorphism also lies in a region of p53 that is involved in the induction of apoptosis (18). Two recent reports indicated that the Arg72 allele is preferentially mutated and retained in various human tumors arising in Pro/Arg heterozygotes, and that the p53 mutant plays as a more potent inhibitor of p73 activity when p53 has Arg72 rather than Pro72 (11, 12). These findings suggest that this polymorphism acts as an intragenic modifier of mutant p53 behavior and has an effect on the biological activity of p53. In the present study, we investigated the hypothesis that the p53 Arg72 allele is preferentially mutated in tumors by screening the p53 genotype and mutations. In our series of TCCs, most patients were Arg/Arg or Pro/Arg, but the distribution of these genotypes was typical of a Japanese population (9, 16). We demonstrated that the structural change conferred on p53 by Arg72 polymorphism was associated with p53 mutations and also fund that the Arg72-containing allele was preferentially retained in p53 -mutated TCCs arising in individuals with germ-line heterozygosity (Pro/Arg). Although the sample size used was small and failed to reach statistical significance, the results of this study confirm those of the previous studies of tumors (11, 12), suggesting that mutant alleles containing Arg72 may be preferentially selected during tumorigenesis and affect mutant behavior in urothelial TCCs.

At present, the significance of the p53 codon 72 polymorphism remains obscure, both in terms of cancer epidemiology and pathobiology. Several studies have reported an association of p53 codon 72 variants (Pro/Pro, Pro/Arg, and Arg/Arg) with tumor susceptibility (3, 4, 5, 6, 7, 8, 9, 10). Patients, especially smokers, with the Pro/Pro genotype are more likely to develop lung cancer (5). In contrast, nonsmokers with lung cancer have an increased frequency of the Arg/Arg genotype (3). The genotype of the codon 72 polymorphic site varied with race in patients with gastric cancer (7). Recently, Storey et al. (8) reported that a majority of women affected by human papillomavirus-induced cervical cancer are homozygous for the Arg allele compared with unaffected women. Thus, p53 codon 72 variants may serve as risk factors for human tumors. In the present study, we speculate on the potential impact of the codon 72 polymorphic site in regard to the tumor behavior of urothelial TCCs and revealed that tumors with p53 mutation arising in Arg72-containing allele became progressively more abundant with increase in tumor stage. Our in vivo findings may provide potential evidence accounting for the more aggressive behavior of this type of p53 -mutated TCCs. Additional comprehensive studies using a spectrum of excised human carcinoma tissue samples from greater numbers of tumors will be needed to elucidate the association between polymorphic residue within p53 and mutant behavior of p53 in human carcinogenesis.