- Top of page
- Material and methods
Hereditary nonpolyposis colorectal cancer (HNPCC) is an autosomal dominantly inherited cancer syndrome associated with germline mutations in DNA mismatch repair (MMR) genes. Recently a polymorphism at codon 72 (R72P) in the tumour suppressor gene TP53 has been implicated in the age of disease onset in HNPCC. In this report we have studied a large cohort of HNPCC patients to assess the impact of this polymorphism on disease expression and age of diagnosis of colorectal cancer (CRC). DNA samples from 218 HNPCC mutation positive patients from Australia and Poland were genotyped for the arginine to proline change at codon 72 in the TP53 gene. The association between the polymorphism and disease characteristics (mutation status, disease expression and age of diagnosis of CRC) was tested using Pearson's Chi-square and Kaplan–Meier survival analysis. Our study of Australian and Polish HNPCC patients does not provide evidence for an association between the Arg/Pro (GC) genotype of the R72P polymorphism and age of diagnosis of CRC. The R72P polymorphism was examined in HNPCC patients and found to be not associated with disease development in either the Australian or Polish populations. When gene mutation status (hMLH1 or hMSH2) was included in the analysis some evidence of an affect was observed. The genotyping revealed in the Australian population that the R72P polymorphism was under-represented in the hMSH2 group whereas it was over-represented in the Polish hMSH2 group. A similar trend was observed for hMLH1 in both groups but was not significant. Age of diagnosis of CRC in HNPCC patients is therefore more complex than that predicted by the R72P TP53 polymorphism alone, suggesting an inter-relationship with other genetic and/or environmental factors. © 2005 Wiley-Liss, Inc.
Hereditary nonpolyposis colorectal cancer (HNPCC) is an autosomal dominantly inherited cancer syndrome that is characterized by early-onset epithelial cancers. Patients with HNPCC have an 80% risk of developing colorectal cancer (CRC) during their lifetime, at a mean age of disease onset that is much lower than that of the general population.1 They are also at risk of developing other epithelial malignancies in a variety of organs that include the endometrium (40–60%2), small bowel, stomach, ovary, bladder, pancreas and the urinary tract.1, 3 HNPCC is associated with germline mutations in DNA mismatch repair (MMR) genes, with the majority affecting hMLH1 and hMSH2.4, 5, 6, 7, 8 The primary function of MMR genes is to eliminate base–base mismatches and insertion–deletion loops which arise as a consequence of DNA polymerase slippage during DNA replication.7, 9
Recently, it has been reported that HNPCC patients heterozygous for the arginine (CGC) to proline (CCC) polymorphism at codon 72 (R72P) in TP53, developed colorectal cancer at an average age of 13 years younger than those who were homozygote wildtype for R72P.10TP53 is a tumour suppressor gene, which regulates the transcription of genes necessary for the maintenance of genomic integrity; blocking cell proliferation after DNA damage and initiating apoptosis if DNA damage is too extensive.11, 12, 13TP53 acts to reduce the incidence of cancers by mediating apoptosis in cells with activated oncogenes. Mutations in TP53 are found in 50–55% of all human cancers.14 DNA damage and genotoxic stress can cause induction of TP53 leading to the biological outcome of growth arrest or apoptosis.15 The structural difference between the 2 variants at codon 72 are a result of a polymorphism that was identified in 1987.16 In 1999 it was shown that the 2 forms of the protein were not functionally equivalent and that the Arg72 variant induces apoptosis faster and suppresses tumour growth more efficiently than the Pro72 variant.17 A recent study provides evidence that the Arg72 variant of TP53 is at least 5 times more efficient at inducing apoptosis than the Pro72 variant.18 Furthermore, Pim and Banks19 have demonstrated that the Pro72 variant appears to induce higher levels of G1 arrest than the Arg72 variant. Another study suggests that the polymorphism contributes to a genetically determined variability in apoptotic susceptibility among old people,20 which could be relevant in the context of an age-related pathological condition.
The frequency of the TP53 polymorphism at codon 72 is known to differ by geographic region and Storey et al.21 raised the possibility of major effects of population stratification on allele frequency, which was supported by Bereir et al.22 According to Pegoraro et al.23 the most marked difference in allele frequency distribution can be seen between the Caucasian and Japanese populations and Africans. The Pro72 allele frequency shows a significantly linear correlation with geographical latitude in the northern hemisphere with 17% in Swedish Saamis and 63% in native Nigerians.24 A highly significant association between the TP53 polymorphism and ethnicity was also observed in a large study (n = 3,371) in North America.25
The TP53 polymorphism at codon 72 has been widely studied in a variety of malignancies. Some cervical tumours are HPV associated, and individuals with the Arg/Arg (GG) genotype was suggested to be more susceptible to HPV-associated tumourigenesis via TP53 inactivation.21 Subsequent reports, with only a few exceptions,23, 26, 27 failed to confirm this hypothesis.25, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38 Similarly, 18–36% of head and neck tumours are HPV-associated but no association between the TP53 polymorphism and head and neck tumours was observed in these studies.39, 40, 41 However, the Pro/Pro (CC) genotype of TP53 has been shown to be more frequent in patients with chronic myeloid leukaemia (CML) than in controls42 and to cause an increased risk of lung cancer among Taiwanese females.43TP53 mutations in non-small cell lung cancer (NSCLC) are increased in carriers of the Pro/Pro (CC) genotype44 and a higher frequency of the Arg/Arg (GG) genotype has been linked to increased risk of breast cancer45, 46 and bladder cancer.47
HNPCC patients are already at high risk of developing cancer by virtue of their decreased capacity to repair DNA mismatches. The presence of a polymorphism in TP53, which affects its function, may be observed as a change in disease phenotype. To confirm the results of Jones et al.,10 we investigated a larger cohort of HNPCC patients to determine the effect of the R72P change, and any influence on disease expression or age of diagnosis of CRC in MMR gene mutation carriers.
- Top of page
- Material and methods
The role of modifier genes in otherwise single gene disorders is becoming recognised as an important adjunct in understanding the differences that can be observed in individuals who harbour mutations in the same gene. The focus of this study has been on a gene involved in cell cycle control, regulation of DNA repair and apoptosis. Inheriting a deficiency in 2 genes associated with DNA repair could affect the accumulation of mutations in cells and thereby alter the risk of tumour development. Functional studies indicate that the Pro/Pro (CC) genotype is less efficient in inducing apoptosis, and induces higher levels of G1 arrest.17, 18, 19 The presence of this change in association with a DNA MMR mutation may be expected to segregate with particular disease characteristics such as a reduced age of disease expression. In support of this notion Jones et al.10 reported that individuals harbouring the Arg/Pro (GC) genotype developed their cancers 13 years younger on average than individuals with the Arg/Arg (GG) genotype. Our study of 218 HNPCC patients (84 Australian and 134 Polish) does not provide the same evidence for the association between the heterozygous (GC) genotype of the R72P polymorphism of TP53 and age of diagnosis of CRC in HNPCC patients.
Potential limitations of our study include comparing all unaffected patients (age 9–93 years) with all patients affected with CRC. This could result in an underestimation of the effect of the TP53 polymorphism, given that the youngest individuals have not lived through much risk of developing disease. A group of patients unaffected with CRC over the age of 45 years was therefore studied, and we were able to show that the allele frequency distribution among an older unaffected population was the same as in the other groups. Our subject group consists of probands and some relatives of probands (26% in Australian samples and 58% in the Polish sample), giving an uneven proportion of CRC cancer patients in the 2 populations (p = 0.01). This may have possibly influenced the difference seen between hMLH1 and hMSH2 mutation carriers. The allele frequency distribution of the Pro/Pro (CC) genotype is low with only 12 patients harbouring this genotype, which does not lend itself to any effective evaluation.
Within the context of the Caucasian population the median age of diagnosis of CRC is the same for both the Australian and Polish populations (42 and 44 years, respectively). This is in accordance with median age of probands (44 years) in a study of Caucasian HNPCC patients in USA.49 Mutation positive HNPCC patients have an estimated 80% lifetime risk of developing cancer50 and our study population include both cancer patients and unaffected MMR gene mutation carriers. Therefore, Kaplan–Meier survival analysis was used to determine the median age of diagnosis of CRC, which is the age at which 50% of the population is cancer free and includes unaffected participants in the calculation. According to Jones et al.10 Kaplan–Meier analysis is the most powerful analysis in detecting the modifying effects of SNPs.
The frequency of the 3 genotypes; Arg/Arg, Arg/Pro and Pro/Pro in our subject groups was 61, 36 and 4% in the Australian group and 58, 35 and 7% in the Polish group, respectively. This is in accordance with frequencies of control groups from Germany (59, 34 and 7%),41 The Netherlands (57, 37 and 7%),30 northern Italy (61, 34 and 5%)38 and control groups from other studies undertaken in Caucasians.31, 32, 33, 51, 52, 53 This provides evidence that there is no difference in the TP53 allele frequencies between HNPCC patients in our study and other European populations. As shown in Figures 1 and 2 there is a marked difference between the Australian and Polish hMLH1 and hMSH2 mutation carriers, where only the hMSH2 group reached significance. The proportion of hMLH1 and hMSH2 mutations is the same in both populations, so are the proportion of affected patients within the hMLH1 and hMSH2 carriers (46% in both populations). The allele frequency distribution in the Polish group is similar in both hMLH1 and hMSH2 carriers (see Fig. 2), while they are not in the Australian group (see Fig. 1). The hMSH2 carriers in Australia have ∼40% more patients harbouring the Arg/Arg (GG) genotype than the hMLH1 carriers. This difference is most likely accounted for by the fact that this population is not random and do contain relatives of probands, and it is likely that if a much larger population was studied that this difference would disappear.
Zhang et al.54 suggest that Arg/Arg (GG) genotype may be associated with later onset of gastric cancer or prolonged survival of patients with gastric adenocarcinoma, which is in accordance with Jones et al.'s10 results (GG genotype developed CRC 13 years later than GC genotype). However, the number of cases used for this inference was small and consequently lacked statistical power. Of the few studies that managed to confirm the original research by Storey et al.21 on cervical cancer, one was among a small group (n = 23) of Israeli Jewish woman (who have a low incidence of cervical cancer) compared to controls (from different ethnic groups).26 Pegoraro et al.23 claims that their results can clarify some of these conflicting results, since most studies have not made a distinction in HPV-status; in the absence of HPV-16 or -18 infections, the Arg72 allele may be a risk factor in patients at low to intermediate risk of uterine cervical carcinoma. A higher frequency of the Arg/Arg (GG) genotype has been linked to an increased risk of bladder cancer47 and breast cancer46 when compared to a control population in Greece and among Jewish woman in Israel, who have a very low frequency of the Arg/Arg (GG) genotype (∼20%). In the literature there is controversy with respect to the role of the R72P polymorphism and disease. In our study we do not see any obvious relationship between this polymorphism and disease expression or age of diagnosis of CRC.
The diverse results for the R72P polymorphism may possibly be related to a polymorphism in MDM2, a protein that binds directly to and inhibits TP53. The SNP309 of MDM2 has recently been associated with a lower age of onset of tumour development (on average 9 years) and an increased occurrence of multiple primary tumours in Li-Fraumeni Syndrome patients.55 This study has been able to support the hypothesis that heightened levels of MDM2 (caused by the mutant genotype of SNP309) results in the inability to properly stabilise TP53 response to cellular stress, e.g. DNA damage. If this is the case, TP53 could indirectly affect the age of diagnosis of CRC in HNPCC patients because of the heightened levels of MDM2.
Overall the R72P allele frequencies were the same in both populations and the differences observed in the hMLH1 and hMSH2 carriers does not appear to be a determinant for the age of CRC diagnosis (as revealed by Kaplan–Meier analysis). These results indicate that there is no association between the TP53 codon 72 polymorphism and HNPCC patients. To confirm these results, a much larger sample population needs to be analysed.
In conclusion, MMR gene mutation carriers in HNPCC families have considerable variation in the age of diagnosis of CRC and disease expression. We have demonstrated that age of diagnosis of CRC in HNPCC patients are more complex than that predicted by the R72P TP53 polymorphism. It is likely that other modifying factors, both genetic and environmental, play a role in the variation in disease expression observed in HNPCC.