Purpose: A common polymorphism of the tumor suppressor gene TP53 at codon 72 has been associated with human cancer susceptibility. The prognostic role of the polymorphism was assessed in 102 patients with advanced gastric adenocarcinoma.
Experimental Design: We followed up 102 consecutive Caucasian patients with advanced gastric adenocarcinoma for >5 years and determined the status of the TP53 codon 72 polymorphism in DNA samples extracted from archived gastric tissues.
Results: The frequency of the arginine homozygous allele was positively correlated to patient age at baseline (P = 0.002). However, the age-related increase in the percentage of codon 72 arginine p53 was not correlated to the prognosis for gastric cancer patients. Multivariable analysis in patients who had surgery showed that baseline age may be inversely associated with patient survival (odds ratio, 1.1; 95% confidence interval, 1.0–1.2; P = 0.02). Furthermore, alcohol consumption may be associated with reduced survival (P = 0.06).
Conclusions: These findings indicate that codon 72 arginine p53 may not be associated with a prolonged survival in patients with advanced gastric adenocarcinoma, but further study is needed to assess whether this polymorphism is associated with a late onset or slow progress of early gastric adenocarcinoma.
The TP53 gene, located on chromosome 17p13, is one of the important tumor suppressor genes in controlling DNA transcription, cell cycle regulation, and tumor suppression. The p53 protein induces apoptosis or cell cycle arrest in response to DNA damage, allowing the injured cells to be destroyed or repaired before reinitiating replicative DNA synthesis (1) . TP53-null mice are susceptible to an array of spontaneous tumors in early adult life (2) . The importance of p53 as a tumor suppressor has also been demonstrated in humans with the rare autosomal-dominant Li-Fraumeni syndrome, who carry heterozygous TP53 mutations in the germline. On loss of the wild-type TP53 allele, these individuals develop a variety of mesenchymal and epithelial tumors at an early age (3) . Mutations in TP53 are found in >50% of all human cancers (4) , comprising >50 different cell and tissue types, indicating that there is a powerful selection for loss of p53 activity during tumor development. A common polymorphism of TP53 at codon 72 results in either a variant protein with a proline (Pro) residue (CCC) or an arginine (Arg) residue (CGC; Refs. 5 , 6 ). There are several differences among the p53 variants in their ability to bind components of the transcriptional machinery, to activate transcription, to induce apoptosis, and to repress the transformation of primary cells (7) . This polymorphism has been associated with the susceptibility of several human cancers (8, 9, 10, 11, 12) and prognosis (13) . Patients with the Pro/Pro genotype are more likely to develop lung cancer than those with other genotypes (14, 15, 16) , particularly patients who are smokers. In contrast, nonsmokers with lung cancer have an increased frequency of the homozygous arginine genotype (8 , 11) . Increased frequency of the proline allele (Pro/Pro or Arg/Pro genotypes) was also found in patients with breast cancer (12) . Although controversial findings are reported on the relationship between the polymorphism and cervical cancer (17 , 18) , an in vitro study suggested that the arginine form of p53 was significantly more susceptible to E6 oncoprotein (derived from tumor-associated human papillomaviruses)-mediated degradation than the proline form (17) . Thus, the p53 variants may serve as risk factors for major human neoplasms and may play a role in modulating environmental risk factors for cancer.
Several studies have suggested that the codon 72 polymorphism may be implicated in the development of gastric adenocarcinoma. Kawajiri et al. (16) studied 347 healthy controls and 140 gastric cancer patients in a Japanese population and found that the genotype distribution among gastric cancer patients was significantly different from that of healthy controls, with 48.6% Arg/Arg and 3.6% Pro/Pro in gastric cancer patients compared with 41.5% and 10.9% in healthy controls (χ2 = 7.3; df, 2; P = 0.03). Recently however, another Japanese study showed that gastric cancer patients had a codon 72 genotype distribution pattern similar to that of controls, although they found that the Pro/Pro genotype was more frequent in patients with diffuse-type gastric cancer (19) . Shepherd et al. (20) examined the relationship between the codon 72 polymorphism and susceptibility to gastric cancer in a group of American gastric cancer patients. They noticed a higher frequency of arginine homozygotes for patients with gastric cardia cancer compared with those with cancer in the antrum or corpus. Recently we examined the relationship between codon 72 genotypes and gastric adenocarcinoma in a group of Caucasian patients. An age-related increase in the frequency of homozygous arginine allele in gastric cancer patients was observed, but not in noncancer controls (21) . These data may suggest that codon 72 arginine p53 is associated with a later onset of gastric cancer or prolonged survival of patients who have had gastric adenocarcinoma. To examine whether the age-related changes in the distribution of codon 72 genotypes is associated with patient survival, we further studied the prognosis of 102 Caucasian gastric cancer patients in relation to their codon 72 polymorphism status.
MATERIALS AND METHODS
The study recruited 118 consecutive Caucasian patients who underwent gastroscopy or gastrectomy between January 1997 and October 2000 at Bristol Royal Infirmary (Bristol, United Kingdom). The prognostic details were available for 102 of them (63 males; median age at diagnosis, 74 years; range, 33–99 years). Clinical data, including diagnosis, race, age, and gender, for each patient were obtained from patient records and endoscopy and histology reports. All patients were staged to identify those who had operable disease. This was done through routine examination, computed tomography scans, staging laparoscopy, and endoscopic ultrasound. The classification of cardia and noncardia gastric cancer was based on International Classification of Diseases for Oncology code, 2nd edition (22) . Patients were excluded from the study if they had severe systemic diseases or cancers of other systems. All of the histological diagnoses were reassessed on original H&E slides by two pathologists (R. F. and M. M.), who were blinded to the clinical details and research outcome. The predominant WHO histological type, the degree of differentiation (well, moderate, or poor), and the Lauren’s histological classification (intestinal or diffuse type) were recorded. Heterogeneous tumors were graded according to the worst differentiated areas. International Union Against Cancer tumor stage was determined by reviewing pathological and clinical records and histological material. The study was approved by the research ethics committee of The United Bristol Healthcare National Health Services Trust. The clinical outcome, including length of survival, was obtained from patient records and by contacting each patient’s general practitioner. The median follow-up period for all patients was 5 months (range, 0–64 months).
Analysis of TP53 Codon 72 Polymorphism.
The preparation of genomic DNA and subsequent analysis of the codon 72 polymorphism were as described previously (21) . Briefly, gastric tissue blocks from gastric tumor tissue and some adjacent well-separated normal tissue were used for genomic DNA preparation. All of the samples were obtained before patients received any medical treatment for their cancer. A QIAamp DNA mini kit (Qiagen, Crawley, United Kingdom) was used for DNA extraction. PCR amplification was performed with two sets of primers: the outer oligonucleotide primers (forward, 5′-GCT CTT TTC ACC CAT CTA CAG-3′; reverse, 5′-TGA AGT CTC ATG GAA GCC AGC-3′) and the inner oligonucleotide primers (forward, 5′-TCC CCC TTG CCG TCC CAA-3′; reverse, CGT GCA AGT CAC AGA CTT-3′). The restriction enzyme cuts within the sequence corresponding to the arginine codon (CGC) at position 72 to generate two visible fragments of 160 and 119 bp, leaving the proline allele uncut. The positive control was a DNA sample with Pro/Arg genotype to confirm digestion of the PCR products. The negative control was a complete PCR reaction mixture with template DNA omitted to exclude the possibility of cross-contamination during sample preparation. The genotypes of >10% of the samples were reassessed to confirm the results. DNA sequencing was performed on selected PCR products to further confirm the authenticity of the genotype analysis. We also analyzed DNA samples from well-separated normal tissues and from gastric tumor tissues of 36 cancer patients to compare their codon 72 polymorphic status and found that the genotypes from the two sites for each patient were all identical.
Statistical analyses were performed with the SPSS for Windows (Standard version, Nov 1999) statistical package. Kaplan-Meier curves were constructed to assess survival, and differences in survival among groups were analyzed by the log-rank test. The Cox proportional hazard model was used to determine whether factors such as age, gender, TNM pathological stage, histological type, treatment, lifestyle (cigarette smoking and alcohol consumption), and TP53 codon 72 polymorphism were associated independently with survival.
Of 102 patients for whom prognostic data were available, 45 had surgery and 57 had received palliative therapy (13 of them had radio- or chemotherapy). Their details are presented in Table 1⇓ . The overall genotype frequencies for Pro/Pro, Arg/Pro, and Arg/Arg were 4%, 43%, and 53%, respectively. The genotype distribution was similar in men and women (χ2 = 1.8; df, 2; P = 0.4). The presence of the arginine homozygous allele was positively associated with age, whereas the proline homozygous and heterozygous genotypes showed no such correlation (P = 0.006; Fig. 1⇓ ).
There were 73 (90%) patients who died of gastric cancer and 21 patients who were still alive in the end of the study. Eight patients died of diseases other than gastric cancer and were excluded from subsequent survival analysis. The median follow-up time was 38 months (range, 22–64 months) for the survivors and 3 months (range, 0–55 months) for patients who died of gastric cancer. Patients with the arginine homozygous allele and proline alleles (Pro/Arg and Pro/Pro) had a median survival of 5 and 8 months, respectively. Initial analyses suggested that patient age at baseline, tumor tissue differentiation, TNM pathological stage, and treatment received (surgery or palliative therapy) significantly influenced patient survival (Table 2)⇓ . Other factors, such as patient age in the end of the study or at death, codon 72 genotypes, and tumor site, showed no significant correlation with patient survival. To further examine the possible role of the codon 72 polymorphism, we subsequently assessed the prognosis values of codon 72 genotypes by stratifying various clinicopathological parameters of the patients. Among the parameters analyzed, cigarette consumption was an important patient survival modifier in relation to codon 72 genotypes. Smokers who harbored the homozygous arginine allele had significantly longer survival than those smokers possessing other genotypes (27 versus 5 months; P = 0.02). However, this was not seen in nonsmokers (Table 3)⇓ .
The median survival for patients who received surgical treatment was 31 months (range, 1–64 months), which was significantly longer than those patients in the palliative treatment group (2 months; range, 1–39 months; P < 0.0001). Because of the very short survival period in the patient group receiving palliative treatment, subsequent multivariable analysis was performed only on patients who had surgery. Patient age at baseline showed a significant association with the survival of this patient group, with an odds ratio of 1.1 (95% confidence interval, 1.0–1.2; P = 0.02). No other factor showed links to survival except that alcohol consumption was marginally associated with reduced survival (P = 0.06; Table 4⇓ ).
This study examined the prognostic values of the TP53 codon 72 polymorphism and other clinicopathological parameters in 102 Caucasian patients with gastric adenocarcinoma, with a follow-up period of >5 years. Consistent with our previous observation, there was a baseline age-related increase in the frequency of the arginine homozygous allele for the codon 72 polymorphism (21) . The most obvious potential explanation could have been that this genotype conferred a survival advantage and therefore was more frequent in the older patients. However, the data in the study do not support this hypothesis, and the genotype was not correlated with the survival of gastric cancer patients. Multivariable analysis of prognostic factors showed that the baseline age of postsurgery patients may be inversely associated with the survival of cancer patients. Furthermore, the data also showed that patients who consume alcohol might have a worse prognosis than those who drink no alcohol.
The development of gastric adenocarcinoma may involve several different stages. The early stage has been attributed to alterations in DNA caused by chronic inflammation, an imbalance between epithelial-cell proliferation and apoptosis, and gastric colonization by enteric bacteria with nitrate reductase activity, which facilitates the formation of carcinogenic nitrosamines in a milieu of atrophy and achlorhydria. It may take many decades for normal gastric mucosa to progress to chronic gastritis, gastric atrophy, intestinal metaplasia, and finally to early gastric cancer (23 , 24) , and it may take an additional 8 years for the early gastric cancer to become advanced gastric cancer (25 , 26) . One of the functions of the p53 protein in a cell exposed to DNA-damaging agents is to induce G1 arrest to allow the cell machinery to repair DNA damage or, in case of failure, to induce apoptosis. It has been shown that the p53 Arg/Arg variant induces apoptosis with faster kinetics and suppresses transformation more efficiently than the p53 Pro/Pro variant. The p53 Arg/Arg variant is a better inducer of transcription because it has a stronger affinity for the TAFII32 and TAFII70 transcription factors (7) . This could confer a survival advantage, and the Arg/Arg genotype could then become enriched in older patients. However, the protective role of p53 protein may be more important in the early development of the cancer, rather than in the late stage, when the DNA damage may be not obvious and ∼50% of the TP53 gene is mutated in the tumor tissues (27) . This may explain why we could not find a relationship between the polymorphism and survival of patients with advanced gastric adenocarcinoma.
Recent literature suggests that age, lesion sites, histology type, and tumor stage are linked to the prognosis of advanced gastric cancer (28 , 29) . Our data suggest that resection with curative intention can markedly improve patient survival. Unfortunately, many patients in our study presented with a very advanced stage of cancer or were unfit for operation. Our multivariate analysis showed that baseline age may be inversely associated with the survival of postoperative patients. The data did not clearly show why age was relevant to survival, but advanced age may be related to more advanced disease and poorer health status. Alcohol consumption and cigarette smoking have been linked to the development of gastric cancer (30) , but their roles in the prognosis of patients with this cancer remain controversial (30) . Chao et al. (31) recently showed, in a large cohort study with a 14-year follow-up, that prolonged use of tobacco products is associated with increased stomach cancer mortality in men and women. Although in the initial analysis of our data there was a marked difference in the survival of smokers between patients with the homozygous arginine allele and those who possessed proline alleles, it disappeared after subsequent multivariate analysis. This finding may indicate that the interaction between smoking and the codon 72 polymorphism does not play an important role in the prognosis of advanced gastric cancer. The relationship between alcohol consumption and prognosis in gastric cancer is still not clear (32) . Our data showed a marginal reduction in survival in patients who consumed alcohol compared with those who drank no alcohol. Clearly, further study with increased sample size is needed to examine these findings.
In summary, the present study did not find a clear association between the TP53 codon 72 polymorphism and the prognosis of patients with advanced gastric adenocarcinoma, with a follow-up period >5 years. This finding suggests that the baseline-age-related increase in the frequency of the arginine homozygous allele of the codon 72 polymorphism is not related to a better prognosis for patients with advanced gastric cancer. However, this study could not rule out the possibility that this polymorphism plays a role in the early development of gastric cancer. Furthermore, we showed that baseline age of postsurgery patients may be inversely associated with patient survival. Patients who consume alcohol might have a worse survival than those who drink no alcohol. Clearly, a well-designed population-based study is needed to examine the role of the codon 72 polymorphism in the early development of gastric cancer.
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.
Requests for reprints: Zun Wu Zhang, Division of Surgery, Level 7, Bristol Royal Infirmary, University of Bristol, Bristol BS2 8HW, United Kingdom. Phone: 44-117-928-3282; Fax: 44-117-925-2736; E-mail:
- Received June 10, 2003.
- Revision received September 15, 2003.
- Accepted September 26, 2003.