Introduction

Breast cancer, a malignant proliferation of the epithelial cells that line the ducts or lobules of the breast, is the most common malignancy in women [1], accounting for approximately one-third of all cancers in women worldwide [2]. Although many risk factors for the development of breast cancer have been identified, such as the inherited genetic predisposition, the molecular mechanisms related to breast carcinogenesis remain under investigation [3, 4]. The disease seems to be the result of cumulative alterations of oncogenes and tumor suppressor genes that lead to clonal outgrowth of progressively malignant cells [5, 6].

Tumor suppressor gene p53 which located on 17p13 is one of the major markers of human tumor and one of the most commonly mutated genes in human cancer [7]. The p53 protein has a very important function in many physiological processes, such as cell cycle arrest, DNA repair, apoptosis, and gene transcription [8]. In addition to acquired mutations that alter its function in p53, there are many studies which have been identified in the p53 gene, the p53 codon 72 (rs1042522) polymorphism of exon 4 is a common single nucleotide polymorphism (SNP), where the variant encodes a proline (CCC) rather than an arginine (CGC) residue [9], it can affect p53 function. The two polymorphic variants have been indicated that their structure and biological properties were not the same [10].

Many studies have reported the role of p53 polymorphisms at codon 72 (rs1042522), IVS3 16 bp (rs17878362) and IVS6+62A>G (rs1625895) with breast cancer risk [15–68], but the results were inconclusive, some original studies thought that these polymorphisms were association with breast cancer risk, but others had different opinions. In addition, previous meta-analysis on p53 showed conflicting results. Hence, the correlation of this polymorphic gene remains unknown. In order to explore the association between p53 codon 72 (rs1042522), IVS3 16 bp (rs17878362) and IVS6+62A>G (rs1625895) polymorphisms with breast cancer risk, a Meta-analysis was conducted to summarize the data. Meta-analysis is a powerful tool for summarizing the different studies. It can not only overcome the problem of small size and inadequate statistical power of genetic studies of complex traits, but also provide more reliable results than a single case–control study.

Materials and methods

Search strategy and selection criteria

All the case–control studies were identified by a computerized literature search of the PubMed, EBSCO, and CGEMS database (prior to September 2010) using the following words and terms: “p53”, “TP53”, “polymorphism,” and “breast,” as well as their combinations. Only research articles were included and the language was not limited. The included studies have to meet the following criteria: (1) only the case–control studies and cohort studies were considered; (2) they were designed to evaluate the p53 codon 72, IVS3 16 bp (rs17878362) and IVS6+62A>G (rs1625895) polymorphisms and breast cancer risk, (3) the amount of published data was sufficient to allow estimation of an odds ratio (OR) with 95% confidence interval (CI); and (4) the distribution of genotypes among controls are in Hardy–Weinberg equilibrium (P < 0.01).

Data extraction

Information was carefully extracted from all eligible studies independently by two authors (He and Su) according to the inclusion criteria listed above. The following data were collected from each study: first author, year of publication, original country and ethnicity of the sample, source of controls, and genotype distribution. Disagreement was resolved by discussion between the two authors. If they could not reach a consensus, another author was consulted to resolve the dispute, and a final decision was made by two of this group of three authors. When a study reported results on different sub-populations according to ethnicity, we considered each sub-population as a separate study in the meta-analysis.

Statistical analysis

The strength of association between p53 polymorphisms and breast cancer risk was assessed by Crude ORs with the corresponding 95% CIs. The pooled ORs were performed for an additive model (CC vs. YY), recessive model (CC vs. CY+YY) and a dominant model (CC+CY vs. YY). Heterogeneity among studies was checked by the Q test; the P value of more than 0.1 for the Q test indicates a lack of heterogeneity among studies, so the pooled OR was calculated by the fixed-effects model [69]. Otherwise, a random-effects model was used [70]. If heterogeneity was present we might use meta-regression analysis in exploring sources of heterogeneity [71]. In addition, subgroup analyses were conducted by ethnicity and resource of controls. Sensitivity analyses were performed to estimate the robustness of the summary estimate of alteration in breast cancer risk conferred by p53 codon 72 (rs1042522), IVS3 16 bp (rs17878362), and IVS6+62A>G (rs1625895) polymorphisms. Begg’s funnel plots [72] and Egger’s linear regression test [73] were used to assess publication bias. In the control group, Hardy–Weinberg equilibrium (HWE) was tested for using a goodness-of-fit chi-square test. All of the calculations were performed using STATA version 10.0 (STATA Corporation, College Station, TX).

Results

Study characteristics

Table 1 listed the main characteristics and genotype distribution of codon 72 polymorphism (rs1042522), with a total of 56 eligible studies met the inclusion criteria, including first author, published year, ethnicity, original country, source of controls, and genotype distribution. However, the study of Pharoah et al. [50] and the study of Samson et al. [51] were excluded because subjects had been included by Baynes et al. [38] and Rajkumar et al. [44]. The distribution of genotypes in the controls was consistent with Hardy–Weinberg equilibrium in all studies except for two studies (P < 0.01) [21, 65], these studies were excluded in this meta-analysis. Hence, leaving 52 eligible studies (27,046 cases and 30,998 controls) that had assessed the association between the codon 72 polymorphism and breast cancer risk.

Table 1 Main characteristics and genotype distribution of codon 72 polymorphism (rs1042522)
Full size table

Table 2 listed the main characteristics and genotype distribution of IVS3 16 bp polymorphism (rs17878362), with a total of 15 eligible studies (3,332 cases and 3,700 controls) for investigating IVS3 16 bp polymorphism and breast cancer risk.

Table 2 Main characteristics and genotype distribution of IVS3 16 bp polymorphism (rs17878362)
Full size table

Table 3 listed the main characteristics and genotype distribution of IVS6+62A>G polymorphism (rs1625895), with a total of 14 eligible studies (8,787 cases and 9,869 controls) for investigating VS6+62A>G polymorphism and breast cancer risk.

Table 3 Main characteristics and genotype distribution of IVS6+62A>G polymorphism (rs1625895)
Full size table

Meta-analysis results

Codon 72 polymorphism

Table 4 listed the main results of the meta-analysis of codon 72 polymorphism and breast cancer risk. When all the eligible studies were pooled into this meta-analysis of codon 72, there was no evidence of significant association between breast cancer risk and p53 codon 72 polymorphism in any genetic model (dominant model: odds ratio [OR] = 0.97, 95% confidence interval [CI] = 0.90–1.05, P value of heterogeneity test [P h] < 0.001; recessive model: OR = 0.96, 95% CI = 0.88–1.06, P h = 0.009; additive model: OR = 0.95, 95% CI = 0.85–1.07, P h < 0.001). Significant between-study heterogeneity was detected in any genetic model. Hence, we performed the stratified analysis according to ethnicity, source of controls, and sample size. In the stratified analysis for India population, significantly decreased risk of breast cancer was observed in additive model (OR = 0.62, 95% CI = 0.46–0.82, P = 0.001, P h = 0.153, Fig. 1) and recessive model (OR = 0.70, 95% CI = 0.50–0.92, P h = 0.463, Fig. 2).

Table 4 Description of current and published meta-analysis for select codon 72 polymorphism and breast cancer
Full size table
Fig. 1
figure 1

Meta-analysis of OR for TP53 codon 72 polymorphism associated with breast cancer risk (additive model)

Full size image
Fig. 2
figure 2

Meta-analysis of OR for TP53 codon 72 polymorphism associated with breast cancer risk (recessive model)

Full size image

Previous codon 72 polymorphism

Three meta-analyses have been previously published for codon 72 polymorphism and breast cancer risk [11–14]. Table 4 listed the main results of meta-analysis of previous codon 72 polymorphism and breast cancer risk.

The study of Zhang et al. [11] had 39 studies, when all the eligible studies were pooled into the meta-analysis, significantly decreased risk of breast cancer was observed in dominant model (OR = 0.90, 95% CI: 0.82–0.99). In the stratified analysis by ethnicity, significantly decreased risk was also observed in European populations (dominant model: OR = 0.88, 95% CI: 0.80–0.98). In the stratified analysis by source of controls, they found that the variant genotypes were associated with a significantly decreased breast cancer risk in dominant model and additive model (dominant model: OR = 0.87. 95% CI: 0.78–0.97; homozygote comparison: OR = 0.88, 95% CI: 0.78–1.00).

The study of Hu et al. [12] had 37 studies, significantly decreased risk of breast cancer was found between Mediterranean and Northern European populations (recessive model: OR = 0.32, 95% CI: 0.24–0.44; additive model: OR = 0.35, 95% CI: 0.21–0.60). The data for this meta-analysis only included 375 cases and 389 controls from 6 studies between Mediterranean and Northern European populations.

The study of Ma et al. [13] had 21 studies, when all the eligible studies were pooled into the meta-analysis, significantly increased risk of breast cancer was observed in dominant model (OR = 1.179, 95% CI = 1.020–1.362). In the stratified analysis by source of controls, significantly increased risk was also observed by population-based study (dominant model: OR = 1.23, 95% CI = 1.05–1.43; recessive model: OR = 1.16, 95% CI = 1.01–1.33; additive model: OR = 1.28, 95% CI = 1.04–1.59).

The study of Francisco et al. [14] had 42 case–control studies reporting an association between the p53 codon 72 polymorphism and breast cancer. When all the eligible studies were pooled into the meta-analysis, no significant association of breast cancer risk was found in any genetic model. In the stratified analysis by source of country, significantly decreased risk was observed in Indian population (dominant model: OR = 0.75, 95% CI = 0.61–0.93; Arg/Arg vs. Pro/Pro: OR = 0.70, 95% CI = 0.53–0.91; recessive model: OR = 0.77, 95% CI = 0.61–0.97) in Indian population.

IVS3 16 bp polymorphism

Table 5 listed the main results of the meta-analysis of the IVS3 16 bp polymorphism and breast cancer risk. When all the eligible studies were pooled into the meta-analysis, significantly increased risks of breast cancer were observed in any genetic model (dominant model: OR = 1.14, 95% CI = 1.02–1.27, P = 0.017, P value of heterogeneity test [P h] = 0.30, Fig. 3; recessive model: OR = 1.61, 95% CI = 1.21–2.25, P = 0.001, P h = 0.25, Fig. 4; additive model: OR = 1.66, 95% CI = 1.24–2.21, P = 0.001, P h = 0.28, Fig. 5). Moreover, significant between-study heterogeneity was not detected in the meta-analysis of the IVS3 16 bp polymorphism and breast cancer under any genetic model.

Table 5 Results of the meta-analysis for IVS3 16 bp and breast cancer risk
Full size table
Fig. 3
figure 3

Meta-analysis of Ors for TP53 16 bp polymorphism in intro 3 associated with breast cancer risk (dominant model)

Full size image
Fig. 4
figure 4

Meta-analysis of Ors for TP53 16 bp polymorphism in intro 3 associated with breast cancer risk (recessive model)

Full size image
Fig. 5
figure 5

Meta-analysis of OR for TP53 16 bp polymorphism in intro 3 associated with breast cancer risk (additive model)

Full size image

IVS6+62A>G polymorphism

Table 6 listed the main results of the meta-analysis of the IVS6+62A>G polymorphism and breast cancer risk. When all the eligible studies were pooled into the meta-analysis, no significant association of breast cancer risk was found in any genetic model (dominant model: OR = 1.03, 95% CI = 0.91–1.18, P = 0.82, P value of heterogeneity test [P h] = 0.009; recessive model: OR = 0.93, 95% CI = 0.76–1.14, P = 0.49, P h = 0.85; additive model: OR = 0.93, 95% CI = 0.76–1.14, P = 0.51, P h = 0.80). Moreover, significant between-study heterogeneity was not detected in the meta-analysis of the IVS6+62A>G polymorphism and breast cancer under any genetic model, except for dominant model (P = 0.009 for heterogeneity).

Table 6 Results of the meta-analysis for IVS6+62A>G and breast cancer risk
Full size table

Next, we performed stratified analysis by source of controls and ethnicity, in stratified subgroup meta-analysis, the IVS6+62A>G polymorphism was not found to be associated with breast cancer risk too.

Sensitive analysis

We tested the inclusion criteria of this meta-analysis by a sensitivity analysis. Sensitivity analysis were conducted to determine whether modification of the inclusion criteria of this meta-analysis affected the results, A single study involved in the meta–analysis was deleted each time to reflect the influence of individual data set to the pooled ORs, and the corresponding pooled ORs were not essentially altered (data not shown), indicating that our results were statistically robust.

Publication bias

Both Begg’s funnel plot and Egger’s test were performed to access the publication bias of this meta-analysis. Begg’s funnel plots did not reveal any evidence of obvious asymmetry in any genetic model in the overall meta-analysis (Figures not shown). The Egger’s test results suggested no evidence of publication bias in the meta-analysis of codon 72 (P = 0.259 for dominant model, P = 0.514 for recessive model, P = 0.328 for additive model); IVS3 16 bp (P = 0.869 for dominant model, P = 0.694 for recessive model, P = 0.744 for additive model) and IVS6+62A>G (P = 0.663 for dominant model, P = 0.566 for recessive model, P = 0.426 for additive model), indicating that our results were statistically robust.

Discussion

Many studies have reported the role of p53 polymorphisms at codon 72 (rs1042522), IVS3 16 bp (rs17878362) and IVS6+62A>G (rs1625895) with breast cancer risk [15–68], but the results were inconclusive, some original studies thought that these polymorphisms were associated with breast cancer risk, but other original studies thought no association with breast cancer risk. In addition, previous meta-analysis on codon 72 polymorphism showed conflicting results. Hence, a meta-analysis was conducted to explore the association between p53 codon 72, IVS3 16 bp and IVS6+62A>G polymorphisms and breast cancer risk.

Our present meta-analysis, which included 27,046 cases and 30,998 cases from 52 studies, explored the association between the p53 codon 72 polymorphism and breast cancer risk. The results indicated that codon 72 polymorphism may be not associated with breast cancer risk in Caucasian population. In the stratified analysis for Indian population, significantly decreased risk of breast cancer was observed in additive model (OR = 0.62, 95% CI = 0.46–0.82, P = 0.001, P h = 0.153) and recessive model (OR = 0.70, 95% CI = 0.50–0.92, P h = 0.463). The result indicated that codon 72 polymorphism may be associated with breast cancer risk, but there are only four studies in Indian population, to determine whether codon 72 polymorphism be applied to clinical genotyping for risk assessment still require large scale breast cancer case–controls researches in Indian population.

In this meta-analysis, significant association of the IVS3 16 bp polymorphism and breast cancer risk was found (dominant model: OR = 1.14, 95% CI = 1.02–1.27, P = 0.017, P value of heterogeneity test [P h] = 0.30; recessive model: OR = 1.61, 95% CI = 1.21–2.25, P = 0.001, P h = 0.25; additive model: OR = 1.66, 95% CI = 1.24–2.21, P = 0.001, P h = 0.28). The result indicated that IVS3 16 bp polymorphism may increase risk of developing breast cancer. To determine whether this marker should be applied to clinical genotyping for risk assessment still require large scale breast cancer case–control researches.

Meanwhile, no significant association of the IVS6+62A>G polymorphism and breast cancer risk was found. Hence, IVS6+62A>G may have no strong association with breast cancer risk, at least in our meta-analysis.

Previous meta-analysis on p53 codon 72 showed conflicting results. We have read with great interest the article “No significant association between the p53 codon 72 polymorphism and breast cancer risk: a meta-analysis of 21 studies involving 24,063 subjects” Published online on May 23, 2010 issue of “Breast Cancer Research and Treatment” [13]. The study of Ma [13] have 21 case–control studies, his conclusion indicate that it provided strong evidence that the P53 codon 72 polymorphism is not association with the risk of developing breast cancer. Ma et al. [13] concluded that no significant association was found between the P53 codon 72 polymorphism and breast cancer risk when all the eligible studies were pooled into the meta-analysis, but significant risk of breast cancer was observed in dominant model (OR = 1.179, 95% CI = 1.020–1.362). Similarly, Ma et al. [13] demonstrated that no significant association was observed for any of the genetic models in the stratified analysis by source of controls. But in the stratified analysis by source of controls, significant increased risks were observed by source of controls (dominant model: OR = 1.23, 95% CI = 1.05–1.43; recessive model: OR = 1.16, 95% CI = 1.01–1.33; additive model: OR = 1.28, 95% CI = 1.04–1.59). Hence, the ongoing uncertainty still existed and the conclusion by Ma et al. [13] was not entirely credible. In addition, several sizeable eligible studies have not been included in this meta-analysis, we thought that these studies satisfied the search criteria. Importantly, the data reported by Ma et al. [13] for the study by Schimit et al. [41] do not seem in line with the data provided by Schimit et al. [41] in their original publication. The numbers reported by Ma et al. [13] for Arg/Arg, Arg/Pro, Pro/Pro, in cases and controls, are 2797-2008-386 and 2024-1523-287, respectively. Interestingly enough, after carefully studying the data presented by Schimit et al. [41], the frequencies that we have retrieved on the 8,345 cases and 6,849 controls were 4499-3228-618 and 3661-2677-511, respectively. The data reported by Ma et al. [13] for the study by Sjalander et al. [16] do not seem in line with the data provided by Sjalander et al. [16] in their original publication too. The numbers reported by Ma et al. [13] for Arg/Arg, Arg/Pro, Pro/Pro, in cases and controls, are 24-93-95 and 61-253-375, respectively. Interestingly enough, after carefully studying the data presented by Sjalander et al. [16], the frequencies that we have retrieved on the 212 cases and 689 controls were 95-93-24 and 375-253-61, respectively. The data reported by Ma et al. [13] for the study by Sprague et al. [42] do not seem in line with the data provided by Sprague et al. [42] in their original publication too. The numbers reported by Ma et al. [13] for Arg/Arg, Arg/Pro, Pro/Pro, in cases and controls, are 823-570-89 and 705-490-83, respectively. Interestingly enough, after carefully studying the data presented by Sprague et al. [42], the frequencies that we have retrieved on the 1,653 cases and 1,854 controls were 909-644-100 and 1021-704-129, respectively. The data reported by Ma et al. [13] for the study by Weston et al. [17] do not seem in line with the data provided by Weston et al. [17] in their original publication too. The numbers reported by Ma et al. [13] for Arg/Arg, Arg/Pro, Pro/Pro, in cases and controls, are 6-27-32 and 3-42-72, respectively, in Caucasian. Interestingly enough, after carefully studying the data presented by Weston et al. [17], the frequencies that we have retrieved on the 65 cases and 117 controls were 32-27-6 and 72-42-3 in Caucasian, respectively.

Secondly, we have also read great interest the recent meta-analysis by Zhang et al. [11], the study of Zhang [11] have 39 case–control studies, the results suggested that p53 codon 72 polymorphism may contribute to susceptibility to breast cancer, especially in Europeans. Zhang et al. [11] concluded that significant association was found between the TP53 codon 72 polymorphism and breast cancer risk in the stratified analysis by ethnicity (Arg/pro vs. Arg/Arg: OR 0.89, 95% CI 0.80–0.99; dominant model: OR 0.88, 95% CI 0.80–0.98) and source of controls (Arg/pro vs. Arg/Arg: OR 0.88, 95% CI 0.78–0.98; dominant model: OR 0.87, 95% CI 0.78–0.97). But P value of Q test for heterogeneity test <0.001, when heterogeneity was very big, the results cannot be concluded that p53 codon 72 polymorphism may contribute to susceptibility to breast cancer, especially in Europeans. Hence, the ongoing uncertainty still existed and the conclusion by Zhang et al. [11] was not entirely credible. In addition, the study of by Baynes et al. [38] and the study by Pharoah et al. [50] essentially represent the same study, two studies by Buyru et al. [24, 74] have been included in this meta-analysis; however, careful inspection of both studies reveals that the same cases hace been included in them. Hence, incorporating one of the two studies by Buyru et al. might seem more appropriate. Importantly, several sizeable eligible studies have not been included in Zhang et al. [11], we thought that these studies satisfied the search criteria.

Thirdly, we have also read with great interest the recent meta-analysis by Hu et al. [12], the study of Hu et al. [12] had 37 case–control studies, the results suggest that codon 72 had a potential role in association with breast cancer risk within certain populations or regions. Significantly decreased risk was observed by source of Ethnicity (dominant model: OR = 0.32, 95% CI = 0.24–0.44; Pro/Pro vs. Arg/Arg: OR = 0.35, 95% CI = 0.21–0.60) in the Mediterranean studies. In the Mediterranean was Caucasian, in addition, all eligible study was small sample in the Mediterranean. Hence, the ongoing uncertainty still existed and the conclusion by Hu et al. [12] was not entirely credible.

Last, we have also read with great interest the recent meta-analysis by Francisco et al. [14], the study of Francisco et al. [14] had 42 case–control studies reporting an association between the p53 codon 72 polymorphism and breast cancer. Significantly decreased risk was observed in Indian population (dominant model: OR = 0.75, 95% CI = 0.61–0.93; Arg/Arg vs. Pro/Pro: OR = 0.70, 95% CI = 0.53–0.91; recessive model: OR = 0.77, 95% CI = 0.61–0.97) in Indian population. The study of Francisco et al. [14] had only five case–control studies in Indian population, which include 715 cases and 1,668 controls. However, in our present meta-analysis, which including four case–control studies in Indian population, significantly decreased risk was only observed in additive model and recessive model. Sample size was not large in our present meta-analysis and Francisco et al. [14], hence, the results should be interpreted with caution. To determine whether codon 72 polymorphism be applied to clinical genotyping for risk assessment still require large scale breast cancer case–controls researches in Indian population.

However, there are several limitations in this meta-analysis. Our results should be interpreted with caution. First, the controls were not uniformly defined. Although all the controls were healthy populations, most of them were common populations, some controls were Population-based; other controls were hospital-based. Hence, non–differential misclassification bias is possible. Second, in the subgroup analysis may have had insufficient statistical power to check an association, Third, we were also unable to examine the interactions among gene–environment, lacking of the original data of the included studies limited our further evaluation of potential interactions, which may be an important component of the association between p53 codon 72 polymorphism and environment and breast cancer risk. Four, it was much difficult to get the all articles published in various language. We only included the studies published in English and Chinese. Last, our results were based on unadjusted published estimates. Because of data limitations, we were unable to adjust them such as age, smoking, alcohol consumption et al.

Overall, our results indicated that IVS3 16 bp polymorphism may increase risk of developing breast cancer; and codon 72 homozygous mutants may be associated with decreased breast cancer risk in India population.