Polymorphisms of GSTP1, ERCC2 and TS-3′UTR are associated with the clinical outcome of mFOLFOX6 in colorectal cancer patients

The aim of the current study was to examine whether polymorphisms in drug metabolism genes have any clinical impact on patients treated with 5-fluorouracil (FU)/oxaliplatin for metastatic colorectal cancer (MCRC). In total, 63 patients with MCRC were recruited and treated with a modified FOLFOX6 (mFOLFOX6) treatment as a first-line chemotherapy. Polymorphisms in five drug metabolism genes and two DNA-repair genes were assessed in these patients using polymerase chain reaction (PCR), a PCR restriction fragment length polymorphism (PCR-RFLP) technique or invader techniques. These included a 28-bp tandem repeat in the 5′-untranslated region (UTR) and 6-bp deletions in the 3′-UTR of thymidylate synthase (TS), methylenetetrahydrofolate reductase (MTHFR; Ala677Val), glutathione S-transferase π (GSTP1; IIe105Val), GST θ1 (GSTT1; deletion) and GST μ1 (GSTM1; deletion) and the two DNA-repair genes, excision repair cross-complementing-1 (ERCC1; Asp118Asn) and ERCC2 (Lys751Gln). The correlation between these polymorphisms and the clinical outcome, including drug response, progression-free survival (PFS), overall survival (OS) and the incidence of peripheral neuropathy, were evaluated. Patients with the GSTP1-105 A/A genotype had poor responses to mFOLFOX6 treatment compared with those with the GSTP1-105 A/G and G/G genotypes (P=0.01). The median PFS of patients with the ERCC2-751 A/A genotype tended to be longer than that of patients with the ERCC2-751 A/C genotype (P=0.05). Patients with the TS-3′-UTR −6/−6 genotype had a significantly longer OS compared with patients with other genotypes (P=0.003). A statistically significant association between the incidence of peripheral neuropathy higher than grade 2 and the GSTP1-105 (P=0.03) and GSTM1 genotypes (P=0.02) was identified by multivariate logistic regression analyses. Results demonstrated that polymorphisms in GSTP1-105, ERCC2-751 and the 3′-UTR of TS may be a statistically significant predictors of clinical outcome. GSTP1-105 and GSTM1 genotypes may be useful markers of severe peripheral neuropathy in MCRC patients treated with 5-FU/oxaliplatin as first-line chemotherapy.


Introduction
Advances in chemotherapeutic regimens for metastatic colorectal cancer (MCRC) patients, including FOLFOX treatment comprising of a combination of 5-f luorouracil (FU)/leucovorin (LV) and oxaliplatin, have improved overall survival (OS) (1,2). It has been reported that the efficacy rate of FOLFOX treatment varies between 20 and 50% in MCRC patients (1,(3)(4)(5). Individuals who receive chemotherapy commonly suffer from side effects, including myelosuppression, nausea, diarrhea and peripheral neuropathy (1)(2)(3). Numerous patients undergoing FOLFOX treatment have complained of oxaliplatin-induced peripheral neuropathy. Therefore, several markers for predicting the efficacy of FOLFOX treatment have been investigated to identify patients with favorable treatment prognoses. Gene expression analysis, associated with the metabolism of 5-FU and oxaliplatin, has been intensively studied (6)(7)(8)(9). It has been reported that thymidylate synthase (TS) and thymidine phosphorylase (TP) mRNA expression levels are useful markers for predicting the efficacy of FOLFOX treatment in CRC patients with liver metastasis (10). In addition, advances in molecular biology indicate that a number of drug metabolism genes have polymorphisms that alter levels of expression. Among these, polymorphisms have been identified in TS, excision repair cross-complementing-1 (ERCC1) and ERCC2, glutathione S-transferase π (GSTP1), GST θ1 (GSTT1), GST µ1 (GSTM1) and methylenetetrahydrofolate reductase (MTHFR), which exert functions in drug metabolism and antidotal effects on the 5-FU and oxaliplatin pathways. Studies have demonstrated that specific polymorphisms of these genes are associated with the efficacy of FOLFOX treatment in MCRC patients (11)(12)(13)(14).
TS and MTHFR are associated with the metabolism of 5-FU, indicating that their altered expression affects the response to 5-FU-based chemotherapy. The enzyme product of TS is critical for catalyzing the methylation of deoxyuridine-5'-monophosphate to deoxythymidine-5'-monophosphate in de novo DNA synthesis. Fluorodeoxyuridine monophosphate (FdUMP), the metabolic product of 5-FU, forms complexes with TS and 5,10-methylenetetrahydrofolate, resulting in the inhibition of DNA synthesis. Two polymorphisms have been identified in TS, a variable length tandem repeat polymorphism in the 5'-untranslated region (UTR) that consists of two or three 28-bp repeated sequences and a 6-bp insertion/deletion (6+/6-) in the 3'-UTR. A number of studies have described correlations between genotype patterns of polymorphisms in TS and the efficacy of FOLFOX treatment for MCRC patients (11)(12)(13)(14). However, current evidence is insufficient to confirm a statistically significant correlation. MTHFR is important for folate metabolism and catalyzing the conversion of 5,10-methylenetetrahydrofolate to 5-methyltetrahydrofolate. Two important polymorphisms in MTHFR, C677T and A1298C, have been studied (12)(13)(14) and have been identified to affect the enzyme activity of MTHFR (15,16), leading to the accumulation of 5,10-methylenetetrahydrofolate and the enhanced sensitivity of 5-FU by forming complexes with TS and FdUMP. Studies have described the correlation among these polymorphisms and the efficacy of FOLFOX treatment for MCRC patients (12,14).
The expression levels of ERCC1, ERCC2, GSTP1, GSTT1 and GSTM1 have been hypothesized to be associated with the efficacy of platinum compounds, including cisplatin and oxaliplatin. ERCC-1 and -2 are involved in DNA repair and tolerance of DNA damage through the nucleotide excision repair pathway. The enhanced expression of these proteins may lead to the resistance to platinum drugs. A common C to T transition at codon 118 of ERCC1 has been shown to increase expression in patients with the T/T genotype compared with patients with the C/T or C/C genotypes (17), despite the T/T polymorphism producing the identical amino acid, asparagine. Studies have revealed that patients with the T/T genotype have a poor outcome compared with patients with the C/T or C/C genotypes (13). During oxaliplatin-based chemotherapy, the prognosis for MCRC patients with the C/C genotype is more encouraging than that of patients with other genotypes (11,18,19). The xeroderma pigmentosum group D (XPD) gene, ERCC2, has three common polymorphisms in codons 156, 312 and 751. The polymorphism at codon 751 (A>C: Lys>Gln) is associated with the clinical outcome of MCRC patients receiving FOLFOX treatment (13,20).
The GST family includes at least five subclasses with major biological roles in the detoxification of genotoxic compounds. GSTP1, GSTT1 and GSTM1 genotypes have been extensively studied for drug response, including oxaliplatin-based treatment (21)(22)(23). A single nucleotide polymorphism at codon 105 (A>G: Ile>Val) of GSTP1 affects enzyme activity (24). Several studies have demonstrated that among MCRC patients receiving oxaliplatin-based treatment, patients with GSTP1-105 A/G and G/G genotypes have a more favorable outcome compared with patients with the GSTP1-105 A/A genotype (21)(22)(23).
In the present study, correlations were identified between the polymorphism patterns of TS, MTHFR, ERCC1, ERCC2, GSTP1, GSTT1 and GSTM1 and the clinical outcome, including the incidence of peripheral neuropathy, in Japanese MCRC patients who were treated with modified FOLFOX6 (mFOLFOX6).

Materials and methods
Patients and clinical procedures. The current study was performed in accordance with the ethical guidelines for clinical studies with approval from the institutional ethics committee.  (85 mg/m 2 ) and LV (200 mg/m 2 ) for 2 h, followed by a rapid intravenous bolus infusion of 5-FU (400 mg/m 2 ) for 5 min and a continuous intravenous infusion of 5-FU (2,400 mg/m 2 ) for 46 h. This regimen was repeated every 2 weeks. Table I  DNA extraction and analysis of polymorphisms. Genomic DNA was extracted from 23 blood samples and 40 normal colonic mucosae from each enrolled patient using the QIAamp DNA Blood and QIAamp DNA Mini kits (Qiagen, Tokyo, Japan). Polymorphisms were analyzed by polymerase chain reaction (PCR), a PCR restriction fragment length polymorphism (PCR-RFLP) technique and a PCR-invader method. Primer sequences and restriction enzymes of all genes examined are presented in Table II.
Statistical analysis. Continuous data are presented as the median and range. Mann-Whitney U, Fisher's exact probability and χ 2 tests were used where applicable. A survival analysis was conducted using the Kaplan-Meier method. The log-rank test was used to determine the significance of the survival curves. The OS period was calculated between the time of surgery and the date of mortality of any cause. OS was censored from the time of the individuals last visit to the hospital or December 2010, depending on which was the first event. Logistic regression was used to determine independent predictors of adverse events. P<0.05 was considered to indicate a statistically significant difference. All statistical analyses were performed using a statistical software package (StatFlex ver.6.0; Artech, Osaka, Japan).

Results
Correlation between polymorphisms in TS, MTHFR, ERCC1, ERCC2, GSTP1, GSTT1 and GSTM1 and the response rate to mFOLFOX6 treatment. Polymorphisms of GSTP1-105 were shown to significantly correlate with the efficacy of mFOLFOX6 treatment (Table III). The frequencies of GSTP1-105 A/A, A/G and G/G genotypes were 70, 25 and 3%, respectively. In the responder group, fewer patients expressed the GSTP1-105 A/A genotype than the GSTP1-105 A/G or G/G genotypes (P=0.01). No significant differences were identified between the polymorphisms of other genes and the efficacy of mFOLFOX6 treatment.   Fig. 1).
Correlation between polymorphisms in TS, MTHFR, ERCC1, ERCC2, GSTP1, GSTT1 and GSTM1 and incidence of peripheral neuropathy in patients treated with mFOLFOX6. The incidence of peripheral neuropathy of grades 2 (n=42) and 3 (n=2) was found to significantly correlate with the GSTP1-105 (P=0.05) and GSTM1 (P= 0.03) genotypes, as identified by univariate regression analyses (Table IV). Peripheral neuropathy occurred in the majority of patients with the GSTP1-105 A/G and G/G genotypes compared with patients with the GSTP1-105 A/A genotype. Individuals who were GSTM1-negative also had peripheral neuropathy, whereas individuals who were GSTM1-positive did not. A statistically significant correlation between the incidence of peripheral   neuropathy higher than grade 2 and the GSTP1-105 (P= 0.03) and GSTM1 (P=0.02) genotypes was determined using multivariate regression analysis.

Discussion
In the present study, specific polymorphisms of genes involved in 5-FU/oxaliplatin metabolism were demonstrated to be significantly associated with the clinical outcome of Japanese MCRC patients who received first-line chemotherapy with mFOLFOX6. The response to mFOLFOX6 treatment in patients with the GSTP1-105 A/G and G/G genotypes was significantly improved compared with that of patients with the GSTP1-105 A/A genotype. In addition, the ERCC2-751 and TS-3'-UTR genotypes were shown to significantly correlate with PFS and OS, respectively. The results indicated that polymorphisms in the oxaliplatin-associated genes, GSTP1-105 and ERCC2-751, were hypothesized to be important for the prediction of primary clinical outcome, including drug responses and PFS, for MCRC patients treated with mFOLFOX6. Second-and third-line chemotherapy regimens also affected OS. Of the 63 patients, 44 were treated with FOLFIRI following the FOLFOX regimen and 5-FU treatment continued throughout. Therefore, it is possible that polymorphisms in the genes involved in 5-FU metabolism contribute to OS in long-term observations. Previous studies have revealed that GSTP1-105 genotypes are associated with the clinical outcome of MCRC patients who receive 5-FU/oxaliplatin as first-line chemotherapy (11)(12)(13)(21)(22)(23). As GSTP1 expression is enhanced in CRC (26) it has been hypothesized to be involved in the resistance to platinum compounds (27). The enzyme activity of the GSTP1-105 A/G and G/G genotypes is lower than that of the GSTP1-105 A/A genotype (24). In addition, clinical assessments of the correlation between the GSTP1 genotype and the clinical outcome in MCRC patients treated with 5-FU/oxaliplatin appears to be consistent with basic studies. In the present study, patients with the GSTP1-105 A/G and G/G genotypes were revealed to have a significantly improved response to mFOLFOX6 treatment when compared with patients with the GSTP1-105 A/A genotype. Previous studies have indicated that the GSTP1-105 A/G and G/G genotypes are significantly more common than the GSTP1-105 A/A genotype among patients who respond to 5-FU/oxaliplatin treatment (22,23). In addition to drug response, several studies have demonstrated that MCRC patients with the GSTP1-105 A/G and G/G genotypes have favorable outcomes following oxaliplatin-based treatment compared with patients with the GSTP1-105 A/A genotype (11,21). This tendency was also recognized in the results of the current study. The frequencies of the GSTP1-105 A/A, A/G and G/G polymorphisms were 70, 25 and 3%, respectively, in the Japanese population, which is similar to frequencies reported in other Asian populations, including Chinese and Taiwanese (11,22,24). In American and European populations, there is an almost equal frequency of GSTP1-105 A/A and A/G carriers, which combine to make a total of ~90% of all patients. By contrast, the frequency of the GSTP1-105 G/G genotype is ~10% in these populations (11)(12)(13)21,24). Regardless of ethnic differences, the association of the GSTP1-105 genotype with the clinical outcome is consistent among all MCRC patients who receive 5-FU/oxaliplatin as first-line chemotherapy.
In addition to GSTP1 and ERCC-1 and -2, members of the nucleotide excision repair pathway are involved in repair and tolerance of DNA damage and also encode key enzymes for oxaliplatin metabolism. Several studies have demonstrated that the ERCC1-118 and ERCC2-751 genotypes are associated with the clinical outcome of MCRC patients receiving oxaliplatin-based treatment (11,13,(18)(19)(20). In the present study, the ERCC2-751 genotypes were significantly associated with PFS, whereas no significant difference was identified between the ERCC1-118 genotype and the clinical outcome. The PFS of the patients with the ERCC2-751 A/A genotype was longer than that of patients with the ERCC2-751 A/C genotype, and this was consistent with previous studies (11,13,20). The distribution of ERCC2-751 polymorphisms clearly differs between Asian and Western individuals. Among Asians, the frequencies of the ERCC2-751 A/A, A/C and C/C genotypes are 84-92, 8-16 and 0%, respectively (11,22), whereas among Americans and Europeans the frequencies are 25-38, 50-61 and 11-15%, respectively (11)(12)(13)20,22,23). There is a high possibility that the majority of Asians carry the ERCC2-751 A/A genotype, leading to promising outcomes of oxaliplatin-based chemotherapy.
A statistically significant association between TS-3'-UTR genotypes and OS was identified in the current study. Treatment with FOLFIRI was also administered as second-line chemotherapy to ~70% of patients receiving mFOLFOX6 treatment, and hence, the patients with increased survival rates were exposed to 5-FU for a long time. Numerous studies have indicated that MCRC patients with lower TS expression have a favorable outcome following 5-FU-based chemotherapy compared with patients with high TS expression (10,28). A previous study revealed that TS mRNA expression in rectal cancer patients with the TS-3'-UTR -6/-6 and -6/+6 genotypes was significantly lower compared with patients with the TS-3'-UTR +6/+6 genotype, resulting in a favorable outcome following neoadjuvant 5-FU-based chemoradiation (29). Among CRC patients receiving 5-FU-based adjuvant treatment, the OS of patients with the TS-3'-UTR -6/-6 genotype was significantly longer compared with that of patients with other genotypes (30). Although there are various types of cancer, an encouraging association between clinical outcome and the TS-3'-UTR -6/-6 genotype has been identified in Asian gastric cancer patients receiving mFOLFOX6 treatment (31). Several studies have reported that there is no correlation between the TS-3'-UTR genotype and the clinical outcome of MCRC patients receiving 5-FU/oxaliplatin treatment (12)(13)(14)23). The frequency distribution of the TS-3'-UTR genotype may lead to discrepancies in the clinical outcome. In the current study, the frequencies of the TS-3'-UTR -6/-6, -6/+6 and +6/+6 genotypes were 38, 46 and 16%, respectively, while in the USA and Europe these genotypes are 10-16, 37-5 and 33-53%, respectively (11)(12)(13)23). Further studies may be required to clarify the association between these differences in ethnicity and the efficacy of anti-cancer drugs.
While mFOLFOX6 treatment improves the survival rate of MCRC patients, adverse events, including myelosuppression, nausea, diarrhea and peripheral neuropathy, are common. In particular, peripheral neuropathy, caused by cumulative administration of oxaliplatin, directly affects the quality of life and is a major reason for the discontinuation of oxaliplatin chemotherapy. Thus, predictive markers of peripheral neuropathy are required for prospective evaluations. In agreement with previous studies, the incidence of peripheral neuropathy higher than grade 2 was identified to significantly correlate with the GSTP1-105 and GSTM1 genotypes (13,22). Notably, peripheral neuropathy in patients with the GSTP1-105 A/G and G/G genotypes was of greater intensity compared with that of patients with the GSTP1-105 A/A genotype. A statistically significant correlation was identified between the GSTP1-105 genotype and the clinical outcome. Therefore, the GSTP1-105 polymorphism may serve as a double-edged marker for predicting response to 5-FU/oxaliplatin treatment and the intensity of oxaliplatin-associated peripheral neuropathy.
In the present study, the association among gene polymorphisms that affect the metabolism of 5-FU oxaliplatin and the clinical outcome in Japanese patients with MCRC was identified. Ethnic differences in the frequency distribution of polymorphisms, which preclude the extrapolation of clinical studies between Western and Asian populations, were also identified. Therefore, the present study is likely to improve chemotherapy for individuals of Asian descent. Consistent with studies in Western patients, the polymorphisms of GSTP1-105, ERCC2-751 and the 3'-UTR of TS were associated with the clinical outcome of FOLFOX treatment in Japanese MCRC patients. Therefore, these polymorphisms may be significant predictors of clinical outcome globally. However, GSTP1-105 and GSTM1 genotypes may be more useful as markers for severe oxaliplatin-induced peripheral neuropathy in Japanese patients compared with Western patients.