Identifying molecular markers for tumor recurrence is critical in successfully selecting patients with colon cancer who are more likely to benefit from adjuvant chemotherapy. We investigated the effect of single-nucleotide polymorphisms (SNP) within genes involved in oxaliplatin and fluoropyrimidines metabolism, DNA repair mechanisms, drug transport, or angiogenesis pathways on outcome for patients with stage II and III colon cancer treated with adjuvant chemotherapy. Genomic DNA was extracted from formalin-fixed paraffin-embedded samples of 202 patients with stage II and III colon cancer receiving oxaliplatin-based adjuvant chemotherapy from January 2004 to December 2009. Genotyping was performed for 67 SNPs in 32 genes using the MassARRAY (SEQUENOM) technology. Our results were validated in an independent cohort of 177 patients treated with the same chemotherapy regimens. The combination of the selectin E (SELE) rs3917412 G>A G/G and the methylentetrahydrofolate reductase (MTHFR) rs1801133 T/T genotypes was associated with a significantly increased risk for recurrence in both the training [RR = 4.103; 95% confidence interval (CI), 1.803–9.334; P = 0.001] and the validation cohorts (RR = 3.567; 95% CI, 1.253–10.151; P = 0.017) in the multiple regression analysis considering the stage, lymphovascular invasion, and bowel perforation as covariates. The combined analysis of these polymorphisms was also significantly associated with overall survival in both cohorts (RR = 3.388; 95% CI, 0.988–11.623; P = 0.052, and RR = 3.929; 95% CI, 1.144–13.485; P = 0.020, respectively). Our findings suggest that the SELE rs3917412 and MTHFR rs1801133 SNPs could serve as pharmacogenetic predictors of tumor recurrence in patients with early-stage colon cancer treated with oxaliplatin-based adjuvant chemotherapy, thus allowing personalized selection of treatment to optimize clinical outcomes. Mol Cancer Ther; 13(9); 2226–37. ©2014 AACR.

Oxaliplatin- and fluoropyrimidine-based chemotherapy is currently recommended by the main clinical practice guidelines as the standard adjuvant treatment for completely resected stage III and selected high-risk stage II colon cancer. Its benefits have been most clearly demonstrated in stage III disease, with an approximately 30% reduction in the risk of recurrence and a 22% to 32% reduction in mortality (1–3). The value of adjuvant chemotherapy in stage II disease remains controversial and data from randomized trials and meta-analyses indicate that, if there is a benefit for fluoropyrimidine-based chemotherapy in those patients, it does not exceed an absolute improvement in 5-year survival of 5% (4, 5).

Conventional clinicopathological risk factors such as staging, grade, or tumor histology have been used as the standard measure of prognosis for many years. However, these factors do not clearly distinguish between patients who have a high or low risk of disease recurrence, and do not predict which patients are likely to benefit from chemotherapy. In current practice, the majority of patients with stage II and III colon cancer receive treatment unnecessarily, either because they were cured with surgery alone or because they will relapse despite adjuvant treatment. It is therefore essential to improve prognostication beyond that achieved by the current tumor–node–metastasis (TNM) staging system and to develop reliable molecular markers to stratify patients with early-stage colon cancer for risk of relapse, which in turn would allow treatment options to be tailored to the individual, spare other needless toxicity and the financial burden of chemotherapy that will not work.

Several potential molecular predictors of recurrence risk and chemotherapy benefit have been investigated (microsatellite instability, KRAS expression, p53 expression, gene, or microRNA signatures), but they still require validation and are not part of standard clinical practice (6–9). More recently, the necessary interactions between tumor and host in the oncological process have raised questions about how the individual genomic biology influences cancer progression and response to chemotherapy, consequently generating a great amount of critical data. In the last few years, several genetic polymorphisms within genes involved in the metabolism, transport, and detoxification of fluoropyrimidines and oxaliplatin have been considered in an attempt to optimize treatment in patients with colon cancer (10–18), but, up to now, their clinical applications remain limited. Furthermore, numerous variants have been identified in genes related to the DNA-repair complex, angiogenesis, and other relevant pathways that may alter their expression and/or activity, thereby causing interindividual differences in tumor recurrence capacity and chemoresistance (18–22).

Many of the recently published studies have used a methodology based on one or a few single-nucleotide polymorphisms (SNP). Although this approach could be useful, it is necessary to previously know the SNPs and the function of the genes involved, which has often generated contradictory outcomes. In addition, because chemotherapy affects multiple targets, there is little likelihood that a single SNP could account for the variability of the treatment outcomes. Conversely, with the recent advance in molecular genotyping technology, the genotyping of a large number of SNPs involved in the different stages that take part in the triggering of drug activation, metabolisms, and target expression is now possible. This new approach seems to have a greater capability in examining genetic differences in individual patients with regard to drug response or side effects (18, 23).

In this study, we performed a comprehensive analysis to evaluate associations between a panel of SNPs within candidate genes known or suspected to be involved, based on previously described associations or putative functional effects, in oxaliplatin and fluoropyrimidine pathways with clinical outcome in a cohort of patients with early-stage colon cancer treated with adjuvant chemotherapy. In addition, we have explored whether SNPs with significant prognostic value showed stage- or gender-specific differences and if such differences correlated with clinical outcomes.

Study protocol and design, treatment, and evaluations

Patients with high-risk stage II and III colon cancer treated at 2 Spanish institutions in Madrid (La Paz University Hospital and Puerta de Hierro Majadahonda University Hospital) between January 2004 and December 2009 were eligible for our study. The study protocol specified the inclusion criteria as follows: (i) completely resected colon adenocarcinoma located at ≥15 cm of the anal verge as determined by endoscopy or above the peritoneal reflection in the surgical resection without any evidence of metastatic disease; (ii) stage III or high-risk stage II colon cancer with at least one of the following clinicopathological features: T4 primary, inadequate lymph node sampling (less than 14 lymph nodes at the time of surgery), tumor presentation with bowel obstruction or perforation, lymphovascular or perineural invasion, poorly differentiated histology, or a high preoperative serum carcinoembryonic antigen (CEA) level; (iii) oxaliplatin- and fluoropyrimidine-based adjuvant chemotherapy regimen within 6 to 8 weeks of surgery. Main exclusion criteria were as follows: (i) macroscopic or microscopic evidence of residual tumor in the surgical specimen; (ii) any prior adjuvant or neoadjuvant radiotherapy; (iii) severe renal or hepatic disorder, bone marrow suppression, or disabling peripheral neuropathy; (iv) history of other malignancies within the previous 5 years, except curatively treated cervix carcinoma in situ or basal cell carcinoma of the skin. To validate our results, we used an independent cohort of patients with high-risk stage II and III colon cancer treated with the same chemotherapy regimens at 4 different Institutions in Valencia (La Fe University Hospital, Valencian Oncology Institute), Alicante (Elche General University Hospital), and Barcelona (Clinic University Hospital) at the same time period.

All patients were treated with oxaliplatin-based adjuvant chemotherapy (12 cycles of FOLFOX or 8 cycles of CAPOX regimens) as determined by each responsible physician according to the patient characteristics or local practice guidelines.

Before adjuvant treatment, patients were medically assessed and baseline demographic and pathologic variables (age, sex, site of disease, tumor grade, T-stage, N-stage, number of resected nodes, lymphovascular or perineural invasion, bowel obstruction or perforation, preoperative serum CEA level, and chemotherapy regimen) were recorded by reviewing medical histories and pathologic tumor samples. The staging of colorectal cancer was assessed according to the TNM classifications set out by the American Joint Committee on Cancer (AJCC). All patients were included in the colon cancer surveillance program of each attending institution providing the following information: history, physical examination, and CEA determination every 3 months for 2 years and every 6 months at years 3, 4, and 5 after surgery; computed tomography scans of chest and abdomen every 6 months; and colonoscopy at year 1 and thereafter every 3 to 5 years.

The protocol was approved by the local Ethic Committees and all participants signed informed consent for the analysis of molecular correlates.

Selection of candidate polymorphisms

Common and putatively functional SNPs within genes involved in the metabolism and detoxification of oxaliplatin and fluoropyrimidines, DNA repair mechanisms, drug transport, epidermal growth factor receptor (EGFR), and vascular endothelial growth factor receptor (VEGF)–dependent and –independent angiogenesis pathways have been selected on the basis of the public literature resources and databases. Haplotype-tagging SNP and additional informative SNP were selected from the International Haplotype Mapping (HapMap; www.hapmap.org) and from the National Center for Biotechnology Information SNP database (http://www.ncbi.nlm.nih.gov/SNP), respectively.

We used the following criteria to select genes for study: (i) credible scientific basis to support a gene's involvement in the previously referred pathways; (ii) genes with an established, well-documented genetic polymorphism that could alter its function in a biologically relevant manner [either published data or predicted function using Functional SNP (F-SNP) database, http://compbio.cs.queensu.ca/F-SNP]; and (iii) minor allelic frequency of 5% or more in Caucasians (for relative allelic frequencies of the polymorphisms in different ethnicities, we refer to the population genetic section in the Ensembl Genome Browser: http://uswest.ensembl.org/index.html). Overall, 67 SNPs in 32 genes were analyzed (Table 1).

Table 1.

Analyzed SNPs

CategoryGeneSNPs
Metabolisms/detoxification of fluoropyrimidines Thymidylate synthase (TYMSrs34743033, rs2244500 
 Dihidropyrimidine dehydrogenase (DPYDrs3918290, rs6663828, rs970337 
 Methylentetrahydrofolate reductase (MTHFRrs1801133, rs1801131, rs6541003 
 Orotate phosphoribosyltransferase (OPRTrs1801019, rs3772807 
Metabolisms/detoxification of oxaliplatin and DNA repair genes Glutatione S-transferase pi 1 (GSTP1rs1695, rs749174 
 Excision repair cross-complementing 1 (ERCC1rs11615, rs 3212964 
 Excision repair cross-complementing 2 (ERCC2/XPDrs1799793, rs 238404, rs13181 
 Excision repair cross-complementing 5 (ERCC5/XPGrs4150279, rs4150360 
 Excision repair cross-complementing 6 (ERCC6rs2228527, rs7907557 
 X-ray repair cross-complementing group 1 (XRCC 1rs25487, rs25489, rs12611088, rs3213255 
 X-ray repair cross-complementing group 2 (XRCC 2rs3218536,rs 3218408, rs3111417 
 RAD23 homolog B (RAD23Brs2147072, rs10759225, rs1805329 
 Methylguanine-DNA methyltransferase (MGMTrs1803965, rs656639 
 Xeroderma pigmentosum, complementation group A (XPArs3176751, rs3176639 
 Xeroderma pigmentosum, complementation group C (XPCrs2733534 
EGFR pathway Epidermal growth factor (EGFrs6824594, rs929446 
 Epidermal growth factor receptor (EGFRrs11543848 
 Akt murine thymoma viral oncogene homolog 1 (AKT1rs3803304, rs10142069, rs3001371 
 Akt murine thymoma viral oncogene homolog 2 (AKT2rs2304186, rs7260517 
Angiogenesis Vascular endothelial growth factor (VEGFrs833070, rs3025039 
 Interleukin 8 (IL8rs4073, rs2227306 
Others ATP-binding cassette, subfamily B, member 1 (ABCB1rs1045642 
 ATP-binding cassette, subfamily G, member 2 (ABCG2rs2231142, rs2728124, rs3114018 
 Sulfotransferase (SULT1A1rs1968752 
 Selectin E (SELErs5361, rs3917412, rs3917436 
 Matrix metalloproteinase 1 (MMP1rs498186 
 Matrix metalloproteinase 3 (MMP3rs602128 
 Cyclin H (CCNHrs2230641, rs3093816 
 Intercellular adhesion molecule 1 (ICAM1rs3093030 
 Insulin-like growth factor 1 receptor (IGF1Rrs2229765, rs939626 
 Mechanistic target of rapamycin (MTORrs2295080, rs357278, rs6895953 
 LSM3 homolog, U6 small nuclear RNA associated (LSM3rs2607739 
CategoryGeneSNPs
Metabolisms/detoxification of fluoropyrimidines Thymidylate synthase (TYMSrs34743033, rs2244500 
 Dihidropyrimidine dehydrogenase (DPYDrs3918290, rs6663828, rs970337 
 Methylentetrahydrofolate reductase (MTHFRrs1801133, rs1801131, rs6541003 
 Orotate phosphoribosyltransferase (OPRTrs1801019, rs3772807 
Metabolisms/detoxification of oxaliplatin and DNA repair genes Glutatione S-transferase pi 1 (GSTP1rs1695, rs749174 
 Excision repair cross-complementing 1 (ERCC1rs11615, rs 3212964 
 Excision repair cross-complementing 2 (ERCC2/XPDrs1799793, rs 238404, rs13181 
 Excision repair cross-complementing 5 (ERCC5/XPGrs4150279, rs4150360 
 Excision repair cross-complementing 6 (ERCC6rs2228527, rs7907557 
 X-ray repair cross-complementing group 1 (XRCC 1rs25487, rs25489, rs12611088, rs3213255 
 X-ray repair cross-complementing group 2 (XRCC 2rs3218536,rs 3218408, rs3111417 
 RAD23 homolog B (RAD23Brs2147072, rs10759225, rs1805329 
 Methylguanine-DNA methyltransferase (MGMTrs1803965, rs656639 
 Xeroderma pigmentosum, complementation group A (XPArs3176751, rs3176639 
 Xeroderma pigmentosum, complementation group C (XPCrs2733534 
EGFR pathway Epidermal growth factor (EGFrs6824594, rs929446 
 Epidermal growth factor receptor (EGFRrs11543848 
 Akt murine thymoma viral oncogene homolog 1 (AKT1rs3803304, rs10142069, rs3001371 
 Akt murine thymoma viral oncogene homolog 2 (AKT2rs2304186, rs7260517 
Angiogenesis Vascular endothelial growth factor (VEGFrs833070, rs3025039 
 Interleukin 8 (IL8rs4073, rs2227306 
Others ATP-binding cassette, subfamily B, member 1 (ABCB1rs1045642 
 ATP-binding cassette, subfamily G, member 2 (ABCG2rs2231142, rs2728124, rs3114018 
 Sulfotransferase (SULT1A1rs1968752 
 Selectin E (SELErs5361, rs3917412, rs3917436 
 Matrix metalloproteinase 1 (MMP1rs498186 
 Matrix metalloproteinase 3 (MMP3rs602128 
 Cyclin H (CCNHrs2230641, rs3093816 
 Intercellular adhesion molecule 1 (ICAM1rs3093030 
 Insulin-like growth factor 1 receptor (IGF1Rrs2229765, rs939626 
 Mechanistic target of rapamycin (MTORrs2295080, rs357278, rs6895953 
 LSM3 homolog, U6 small nuclear RNA associated (LSM3rs2607739 

Polymorphisms genotyping

Tumor samples and corresponding adjacent normal mucosa tissues from the study patients were obtained at the time of surgery. For the SNP analysis, tumor sections were selected from formalin-fixed paraffin-embedded (FFPE) tumor tissue blocks containing at least 80% of tumor tissue. Genomic DNA was extracted using the “MasterPure Complete DNA Purification Kit” (Epicentre Biotechnologies) according to manufacturer's instructions.

Genotyping analysis was performed using the SEQUENOM MassARRAY platform [National Genotyping Center (CeGen), Santiago de Compostela, Spain]. Genotyping process consists of 2 reactions. In the first one, DNA fragments containing the SNP of interest are amplified using a multiplex polymerase chain reaction restriction fragment length polymorphism (PCR-RFLP) technique. In the second one, the allelic discrimination of single base extension (SBE) reaction is performed. The primer hybridizes near the polymorphic site being genotyped and extends a base according to the sequence of the variant site. Through the use of matrix-assisted laser desorption ionization time-of-flight (MALDI-TOF) mass spectrometry, the mass of the extended primer is determined. The end result is a mass spectrum of the multiple products obtained using extension primers of different lengths in a single reaction, which generates sufficiently spaced groups of peaks. Then, the analysis software assigns genotypes to SNPs in each trial based on the masses obtained. The peak corresponding to the primers that have not extended is used as a control to determine the effectiveness of each extension reaction. The calculation of the ratios of the peaks of each allele generates a report with the cluster analysis (24, 25).

Statistical analysis

The primary endpoint of this study was the relationship between SNPs and disease-free survival (DFS). The secondary endpoint was the relationship between SNPs and overall survival (OS). DFS was computed from the date of intended curative resection to the date of the first documented tumor recurrence. DFS was censored at the date of death or at the last follow-up for disease-free patients. OS was defined as the time from the surgery until death from any cause.

The genotypes for each SNP were analyzed as a 3-group categorical variable (reference model). Deviation from Hardy–Weinberg equilibrium (HWE) for allelic distributions of all SNPs was tested using a χ2 test. We removed SNPs that deviated from HWE in the validation analysis. The distribution of SNPs across baseline demographic, clinical, and pathologic characteristics was examined using the Mann–Whitney test for continuous variables and the χ2 test or Fisher exact test for categorical variables.

The DFS and OS estimates were calculated using Kaplan–Meier curves. The associations between clinicopathological characteristics and SNPs with DFS and OS were assessed using a univariate survival analysis based on the log-rank test. The Cox proportional hazards regression model was used to evaluate the independent effects of the SNPs on DFS and OS using the clinicopathological data and the chemotherapy regimen as covariates. Genotypes were grouped after trying to seek the most significant association with phenotypes. Two-sided P values <0.05 were considered statistically significant. All statistical analyses were carried out using the SAS statistical package software version 9.2 (SAS Institute Inc.).

Patient characteristics and survival

A total of 405 patients were registered onto the study between January 2004 and December 2009. The genotyping analysis was possible in tumor samples from 202 of the 221 (91.4%) patients included as the training cohort and in 177 of the 184 (96.19%) patients included as the validation cohort. The median follow-up was 51.4 months (range, 7–96) in the training set and 49.7 months (range, 9–82) in the validation set. Detailed clinicopathological characteristics of the patients included in the analysis are listed in Table 2. The incidence of grade 2 tumors, perineural invasion, and bowel perforation was significantly higher in the training cohort than in the validation cohort. The 74.8% of patients in the training set were treated with the CAPOX regimen and the 25.2% with FOLFOX, whereas most patients (87.6%) received FOLFOX adjuvant chemotherapy in the validation cohort. No statistically significant differences with respect to other variables were found between both cohorts.

Table 2.

Clinical and pathologic characteristics in the training and validation cohorts

CharacteristicsAll patients (n = 379)Training set (n = 202)Validation set (n = 177)P
Median age (range) 61.94 (23–85) 63.82 (23–85) 59.8 (23–76) 0.231 
Sex 
 Men 194 (51.18%) 115 (56.93%) 79 (44.63%) 0.068 
 Women 185 (48.81%) 87 (43.06%) 98 (55.36%)  
Tumor site 
 Cecum-right colon 115 (30.34%) 60 (29.7%) 55 (33.07%) 0.235 
 Transverse colon 23 (6.06%) 15 (7.42%) 8 (4.51%)  
 Left colon 52 (13.72%) 28 (13.86%) 24 (13.55%)  
 Sigma 189 (49.86%) 99 (49%) 90 (50.84%)  
Differentiation 
 Grade 1 60 (15.83%) 19 (9.4%) 41 (23.16%) 0.001 
 Grade 2 274 (72.29%) 160 (79.2%) 114 (64.4%)  
 Grade 3 45 (11.87%) 23 (11.4%) 22 (12.42%)  
Stage 
 II 105 (27.7%) 60 (29.7%) 45 (25.42%) 0.208 
 III 274 (72.3%) 142 (70.3%) 132 (74.57%)  
Lymphovascular invasion 
 Yes 164 (43.27%) 88 (43.56%) 76 (42.93%) 0.311 
 No 215 (56.73%) 114 (56.43%) 101 (57.06%)  
Perineural invasion 
 Yes 119 (31.39%) 73 (36.13%) 46 (25.98%) 0.025 
 No 260 (68.61%) 129 (63.86%) 131 (74.01%)  
Bowel obstruction 
 Yes 89 (23.48%) 53 (26.23%) 36 (20.33%) 0.184 
 No 290 (76.51%) 149 (73.76%) 141 (79.66%)  
Bowel perforation 
 Yes 61 (16.1%) 41 (20.3%) 20 (11.29%) 0.012 
 No 318 (83.9%) 161 (79.7%) 157 (88.71%)  
Type of adjuvant CT 
 FOLFOX 206 (54.4%) 51 (25.24%) 155 (87.57%) 0.001 
 CAPOX 173 (45.6%) 151 (74.75%) 22 (12.43%)  
CharacteristicsAll patients (n = 379)Training set (n = 202)Validation set (n = 177)P
Median age (range) 61.94 (23–85) 63.82 (23–85) 59.8 (23–76) 0.231 
Sex 
 Men 194 (51.18%) 115 (56.93%) 79 (44.63%) 0.068 
 Women 185 (48.81%) 87 (43.06%) 98 (55.36%)  
Tumor site 
 Cecum-right colon 115 (30.34%) 60 (29.7%) 55 (33.07%) 0.235 
 Transverse colon 23 (6.06%) 15 (7.42%) 8 (4.51%)  
 Left colon 52 (13.72%) 28 (13.86%) 24 (13.55%)  
 Sigma 189 (49.86%) 99 (49%) 90 (50.84%)  
Differentiation 
 Grade 1 60 (15.83%) 19 (9.4%) 41 (23.16%) 0.001 
 Grade 2 274 (72.29%) 160 (79.2%) 114 (64.4%)  
 Grade 3 45 (11.87%) 23 (11.4%) 22 (12.42%)  
Stage 
 II 105 (27.7%) 60 (29.7%) 45 (25.42%) 0.208 
 III 274 (72.3%) 142 (70.3%) 132 (74.57%)  
Lymphovascular invasion 
 Yes 164 (43.27%) 88 (43.56%) 76 (42.93%) 0.311 
 No 215 (56.73%) 114 (56.43%) 101 (57.06%)  
Perineural invasion 
 Yes 119 (31.39%) 73 (36.13%) 46 (25.98%) 0.025 
 No 260 (68.61%) 129 (63.86%) 131 (74.01%)  
Bowel obstruction 
 Yes 89 (23.48%) 53 (26.23%) 36 (20.33%) 0.184 
 No 290 (76.51%) 149 (73.76%) 141 (79.66%)  
Bowel perforation 
 Yes 61 (16.1%) 41 (20.3%) 20 (11.29%) 0.012 
 No 318 (83.9%) 161 (79.7%) 157 (88.71%)  
Type of adjuvant CT 
 FOLFOX 206 (54.4%) 51 (25.24%) 155 (87.57%) 0.001 
 CAPOX 173 (45.6%) 151 (74.75%) 22 (12.43%)  

In the training set, 63 (31.18%) patients had experienced tumor recurrence and the 3-year DFS was 72.2% [95% confidence interval (CI), 69–75.4]. Thirty-one patients (15.34%) died and 3-year OS was 89.1% (95% CI, 86.9–91.3). Patients with stage III [hazard ratio (HR), 3.86; 95% CI, 1.78–8.37; P = 0.001], lymphovascular invasion (HR, 2.15; 95% CI, 1.28–3.59; P = 0.003) and bowel perforation (HR, 1.89; 95% CI, 1.034–3.46; P = 0.039) were more likely to develop tumor recurrence in the multivariate analysis considering the most relevant demographic and clinicopathological characteristics as covariates. We did not observe any significant associations between other variables and DFS. In the validation set, 49 (27.68%) patients had a tumor recurrence, and the 3-year DFS was 75.6% (95% CI, 70.1–82.9). Twenty-seven patients (15.25%) died and 3-year OS was 87.4% (95% CI, 85–89.8).

SNP associated with survival

Genotyping was successful in at least 90% of cases for each SNP analyzed in tumor tissue, with the exception of X-ray cross-complementing gene-1 (XRCC1) rs25487 (82.06%) and thymidylate synthase (TYMS) rs34743033 (81.54%). In failed cases, genotyping was not successful because of limited quantity of extracted genomic DNA. The allelic frequencies for all polymorphisms were within the probability limits of HWE, with the exception of RAD23B rs1805329 (data not shown). Concordance between genotyping in tumor samples and adjacent normal mucosa tissue was found in the 95% to 99% of analyzed SNPs. There were no significant associations between the SNPs and baseline demographic, clinical, or pathologic characteristics (data not shown).

In the training cohort, selectin E (SELE) rs3917412 G>A and methylenetetrahydrofolate reductase (MTHFR) rs1801133 C>T SNPs were significantly associated with DFS in the univariate analysis. A significant association was also found in the logistic regression analysis using the stage, lymphovascular invasion, and bowel perforation as covariates. A total of 111 patients (56.06%) were homozygous for the G allele (G/G) of the SELE rs3917412 G>A SNP, 77 (38.89%) were heterozygous (G/A) and 10 (5.05%) were homozygous for the A allele (A/A; Supplementary Table S1). Patients with the G/G homozygous genotype have a 3-year DFS of 67.6% (95% CI, 63.2%–72%) compared with 80.4% (95% CI, 76.1%–84.7%) in heterozygous or homozygous carriers of the A allele (HR, 0.524; 95% CI, 0.310–0.884; P = 0.016; Fig. 1A). In the case of the MTHFR rs1801133 C>T SNP, 86 patients (42.57%) were homozygous for the C allele (C/C), 77 (38.12%) were heterozygous (C/T) and 26 (12.87%) were homozygous for the T allele (T/T; Supplementary Table S1). Patients with the T/T homozygous genotype have a 3-year DFS of 61.5% (95% CI, 52%–71%) compared with 74.2% (95% CI, 70.8%–77.4%) in heterozygous or homozygous carriers of the C allele (HR, 1.924; 95% CI, 1.019–3.634; P = 0.044; Fig. 1B).

Figure 1.

DFS by SELE rs3917412 (A) and MTHFR rs1801133 (B) SNPs in the training cohort.

Figure 1.

DFS by SELE rs3917412 (A) and MTHFR rs1801133 (B) SNPs in the training cohort.

Close modal

Furthermore, we examined the possible interaction between the SELE rs3917412 G>A and MTHFR rs1801133 C>T SNPs and outcome. In the combined genotype analysis, patients harboring the SELE rs3917412 G/G and MTHFR rs1801133 T/T genotypes were more likely to develop tumor recurrence than those displaying the SELE rs3917412 any A and/or MTHFR rs1801133 any C genotypes (HR, 3.28; 95% CI, 1.46–7.37; P = 0.004). The ability to predict tumor recurrence of this combined analysis was independently validated in the second cohort (HR, 3.85; 95% CI, 1.35–10.95; P = 0.011). In addition, the presence of the unfavorable haplotype was also significantly associated with shorter DFS in the multivariate analysis using the disease stage, lymphovascular invasion, and bowel perforation as covariates, in the discovery (HR, 4.103; 95% CI, 1.803–9.334; P = 0.001) and validation sets (HR, 3.567; 95% CI, 1.253–10.151; P = 0.017; Table 3 and Fig. 2). About OS, the combination of the SELE rs3917412 G/G and MTHFR rs1801133 T/T genotypes was associated with worse outcome in both the training (HR, 3.388; 95% CI, 0.988–11.623; P = 0.052) and the validation cohorts (HR, 3.929; 95% CI, 1.144–13.485; P = 0.028; Supplementary Table S2 and Fig. 3).

Figure 2.

DFS by combination of SELE rs3917412 and MTHFR rs1801133 SNPs in the training (A) and validation cohorts (B).

Figure 2.

DFS by combination of SELE rs3917412 and MTHFR rs1801133 SNPs in the training (A) and validation cohorts (B).

Close modal
Figure 3.

OS by combination of SELE rs3917412 and MTHFR rs1801133 SNPs in the training (A) and validation cohorts (B).

Figure 3.

OS by combination of SELE rs3917412 and MTHFR rs1801133 SNPs in the training (A) and validation cohorts (B).

Close modal
Table 3.

Cox regression multivariate analysis for DFS in the training and validation cohorts

Training cohortValidation cohort
nHR (95% CI)PnHR (95% CI)P
Stage   0.001   0.343 
 Stage II 60 (29.7%) 1 (reference)  45 (25.42%) 1 (reference)  
 Stage III 142 (70.3%) 3.03 (1.43–6.43)  132 (74.57%) 1.53 (0.63–3.67)  
Lymphovascular invasion   0.003   <0.001 
 No 114 (56.43%) 1 (reference)  101 (57.06%) 1 (reference)  
 Yes 88 (43.56%) 2.26 (1.31–3.87)  76 (43.94%) 4.67 (2.28–9.55)  
Bowel perforation   0.349   <0.001 
 No 161 (79.7%) 1 (reference)  157 (88.71%) 1 (reference)  
 Yes 41 (20.3%) 1.32 (0.74–2.35)  20 (11.29%) 3.56 (1.88–6.77)  
SELE rs3917412 + MTHFR rs1801133   0.003   0.035 
 2 favorable (SELE G/A-A/A and MTHFR C/T-C/C) 71 (38.17%) 1 (reference)  63 (38.41%) 1 (reference)  
 1 favorable (SELE G/A-A/A or MTHFR C/T-C/C) 100 (53.76%) 1.68 (0.93–3.03) 0.087 92 (56.09%) 1.99 (1.013–3.909) 0.046 
 2 unfavorable (SELE G/G and MTHFR T/T) 15 (8.06%) 4.103 (1.803–9.33) 0.001 9 (5.48%) 3.56 (1.25–10.15) 0.017 
Training cohortValidation cohort
nHR (95% CI)PnHR (95% CI)P
Stage   0.001   0.343 
 Stage II 60 (29.7%) 1 (reference)  45 (25.42%) 1 (reference)  
 Stage III 142 (70.3%) 3.03 (1.43–6.43)  132 (74.57%) 1.53 (0.63–3.67)  
Lymphovascular invasion   0.003   <0.001 
 No 114 (56.43%) 1 (reference)  101 (57.06%) 1 (reference)  
 Yes 88 (43.56%) 2.26 (1.31–3.87)  76 (43.94%) 4.67 (2.28–9.55)  
Bowel perforation   0.349   <0.001 
 No 161 (79.7%) 1 (reference)  157 (88.71%) 1 (reference)  
 Yes 41 (20.3%) 1.32 (0.74–2.35)  20 (11.29%) 3.56 (1.88–6.77)  
SELE rs3917412 + MTHFR rs1801133   0.003   0.035 
 2 favorable (SELE G/A-A/A and MTHFR C/T-C/C) 71 (38.17%) 1 (reference)  63 (38.41%) 1 (reference)  
 1 favorable (SELE G/A-A/A or MTHFR C/T-C/C) 100 (53.76%) 1.68 (0.93–3.03) 0.087 92 (56.09%) 1.99 (1.013–3.909) 0.046 
 2 unfavorable (SELE G/G and MTHFR T/T) 15 (8.06%) 4.103 (1.803–9.33) 0.001 9 (5.48%) 3.56 (1.25–10.15) 0.017 

Finally, we have tested whether the SELE rs3917412 and MTHFR rs1801133 SNPs could be associated with stage- or gender-specific outcome. The allelic distribution of these 2 polymorphisms did not vary significantly between males and females or stage II and III colon cancer (data not shown). When the patient population was separated by stage in the multivariate model, we found that the combination of the SELE rs3917412 G/G and MTHFR rs1801133 T/T genotypes was significantly associated with DFS in stage III patients in both cohorts. In the training cohort, patients with the SELE G/G and MTHFR T/T genotype had a shorter 3-year DFS of 50.2% (95% CI, 44%–56.4%) compared with those with the SELE any A and MTHFR any C (71.9%; 95% CI, 66.2%–77.6%) or SELE any A or MTHFR any C genotypes (68.3%; 95% CI, 61.5%–75.1%; HR, 2.657; 95% CI, 1.303–5.418; P = 0.027). In the validation cohort, patients with the SELE G/G and MTHFR T/T genotype also had a higher risk of recurrence (3-year DFS, 49.9%; 95% CI, 44%–56.1%) than SELE any A and MTHFR any C (3-year DFS, 77.8%; 95% CI, 71.2%–84.1%) or SELE any A or MTHFR any C carriers (3-year DFS, 69.6%; 95% CI, 63.9–75.8; HR, 2.591; 95% CI, 1.023–6.566; P = 0.045; Supplementary Table S3). In the case of stage II disease, the SELE G/G and MTHFR T/T genotype was associated with a nonsignificant trend toward decreased DFS compared with patients with the SELE any A and/or MTHFR any C genotypes in both the training (HR, 3.93; 95% CI, 0.994–16.478; P = 0.051) and the validation cohorts (HR, 4.174; 95% CI, 0.880–21.082; P = 0.106). In the sex-specific analysis, female patients harboring the combination of the SELE rs3917412 G/G and MTHFR rs1801133 T/T genotypes were more likely to develop recurrence than those displaying the SELE rs3917412 any A and/or MTHFR rs1801133 any C genotypes in the training set (3-year DFS, 72.6% SELE any A and MTHFR any C, 69.4% SELE any A or MTHFR any C, 48.9% SELE G/G and MTHFR T/T; HR, 4.119; 95% CI, 1.466–11.577; P = 0.007) and the validation set (3-year DFS, 79.4% SELE any A and MTHFR any C, 70.2% SELE any A or MTHFR any C, 49.2% SELE G/G and MTHFR T/T; HR, 4.322; 95% CI, 0.973–12.32; P = 0.051; Supplementary Table S4). However, there were no statistically significant survival differences according to these polymorphisms in male patients.

Individualization of adjuvant therapy in colon cancer using robust, evidence-based prediction of efficacy is a highly attractive strategy. To the best of our knowledge, this is the largest pharmacogenomic study performed in a Caucasian cohort of patients with high-risk stage II and III colon cancer who had been homogenously treated with oxaliplatin- and fluoropyrimidine-based adjuvant chemotherapy. We were able to demonstrate that the SELE rs3917412 and the MTHFR rs1801133 germline SNPs are independently associated with tumor recurrence and could serve as gender- and stage-specific predictors of outcome. In addition, this is the first study in which the association between significant SNPs and patients' outcome has been validated in an independent cohort of patients with stage II and III colon cancer treated with the same chemotherapy regimens.

The prognostic and predictive value of SNPs in genes encoding key enzymes for the metabolism of fluoropyrimidines, such as TYMS, dihydropyrimidine dehydrogenase (DPYD), or MTHFR, has been assessed in patients with colon cancer, but the results are often contradictory and inconclusive, particularly in patients treated adjuvantly (10–13, 17, 18, 26). Functional polymorphisms within genes involved in DNA repair pathways or in the metabolism and detoxification of oxaliplatin have also been associated with the efficacy of this agent. Most analysis have been conducted in patients with metastatic disease (12, 15, 16, 18), whereas the pharmacogenetic of oxaliplatin-based adjuvant chemotherapy has not been widely studied (14, 17, 27). In a previous European report including 98 patients with stage III colon cancer, none of the tested SNPs in glutathione S-transferase pi 1 (GSTP1), excision repair cross-complementing 1 (ERCC1), and excision repair cross-complementing 2 (ERCC2) were reliable markers of response to oxaliplatin therapy (14). Furthermore, in contrast to a pharmacogenetic study of adjuvant FOLFOX for Korean patients in which polymorphisms in XRCC1 and TYMS were related to DFS (17), no significant associations between all these SNPs and outcome have been found in our report.

This study, however, is the largest to show that one of the SELE SNPs may be an important prognostic marker for colon cancer relapse after oxaliplatin-based adjuvant chemotherapy. E-selectin is a specific cell adhesion molecule expressed on endothelial cells, which is activated by cytokines released during the inflammatory process and involved in adhesion and extravasation of leucocytes carrying the E-selectin ligands (28). Intriguingly, cancer cells hijack the inflammatory system and interact with the vascular endothelium via E-selectin. This interaction confers metastatic properties to colon cancer cells by promoting its extravasation (29), inducing chemotaxis (30, 31) and triggering the activation of the promigratory p38 and prosurvival ERK and PI3K/NF-κB pathways (32). The correlation between the expression of E-selectin and the E-selectin ligands with the metastatic potential of malignant cells has been well documented for colon cancer (20, 30, 33, 34). Several SNPs have been identified within the SELE gene. The most common one, SELE rs5361, is present in the 10% to 15% of the Caucasian population. It is located in exon 4 and results in the substitution of a serine by an arginine (S128R) within the extracellular domain of the receptor, increasing the adhesion of cancer cells to the endothelium (20, 30). Few studies to date have evaluated the potential prognostic role of SELE SNPs in patients with colon cancer. A French report has found that the variant allele of the SELE rs5361 SNP was associated with a higher risk of relapse and death in patients with stage II and III colon cancer (20). Although these results could not be confirmed in our series, we have demonstrated, however, the prognostic value of the SELE rs3917412 G>A SNP. Curiously, this mutation occurs in nonexpressed intronic regions of the SELE gene and its functional consequences are not definitively established. Although the detailed molecular mechanisms involved in how the SELE rs3917412 G>A SNP exerts effects on colon cancer cells are unknown and the probability of inducing biological effects on gene expression is higher for polymorphisms located in exonic regions, it is possible that those located in nonexpressed regions could affect the protein function through enhanced mutability because of altered DNA sequence context, increased RNA splicing events, altered transcript stability or translation, or interfered gene expression (35). A possible explanation for the worse outcome in patients homozygous for the G allele in our analysis might be the increased affinity and adhesiveness of E-selectin for its ligands in tumor cells in this subgroup. In addition, the ancestral allele may increase the release of E-selectin from the vascular endothelium, providing a soluble stimulant of motility to direct cancer cells to the target organ, thus leading to a higher risk of metastatic seeding.

MTHFR is a key enzyme in the regulation of the intracellular folate levels, which mediates the irreversible conversion of 5,10-methylentetrahydrofolate (CH2FH4) to 5-methyltetrahidrofolate (CH3FH4). Experimental and clinical studies (36) have demonstrated that the balance of CH2FH4 and CH3FH4 is a crucial factor in fluoropyrimidines-induced cytotoxicity, as these drugs exert their effect by inhibiting TYMS through the formation of a ternary complex, involving 5-fluorouracil (5-FU), TYMS, and CH2FH4. Two common MTHFR SNPs, the rs1801133 677 C>T and the rs1801131 1298 A>C, are associated with decreased MTHFR enzymatic activity and thus increasing the intracellular levels of CH2FH4 available for the complex (36). Regarding to the MTHFR rs1801133 SNP, chemotherapy heterozygotes have a 65% lower enzymatic activity compared with wild-type colon cancer, whereas homozygous variants TT have a 30% lower activity. Furthermore, the TT genotype is associated with lower plasma folate, higher homocysteine levels, and lower DNA methylation levels in vivo (36). Therefore, it can be hypothesized that patients with the variant MTHFR alleles (677T or 1298C) will have increased sensitivity to fluoropyrimidine-based treatment than those bearing the common variants (677C or 1298A) and, possibly, an elevated risk of toxicity, by impairment of DNA synthesis and repair. Several studies, the majority conducted on patients with advanced colon cancer, have tried to clarify the possible predictive role of MTHFR SNPs on fluoropyrimidines activity and clinical outcome showing conflicting results (11–13, 26, 36–44). In our analysis, patients harboring the MTHFR rs1801133 T/T genotype were at higher risk to develop tumor recurrence. In accordance with our data, some studies show a negative effect on survival or response for the 677T allele (13, 37). However, other authors show no influence of MTHFR rs1801133 C>T genotype (12, 26, 38, 39) and others again show a positive effect of the T allele (11, 37, 40). Our results are not consistent with in vitro findings that suggest that patients with the T/T genotype would have reduced MTHFR activity compared with those with C/C and C/T genotypes, potentiating the antitumoral activity of fluoropyrimidines. Although the results from in vitro and clinical studies should not be directly compared, if we assume the functional effect of the MTHFR rs1801133 C>T allele, a possible explanation for our divergent findings may be that the cellular availability of 5,10-MTHF may depend not only on MTHFR genotype but also on other cofactors such as the folate dietary intake (45). Furthermore, population-based studies report an increased risk of malignancy and worse survival outcome in patients with the T/T genotype. It is possible therefore that the T/T genotype is an independent adverse prognostic factor regardless of fluoropyrimidine-based therapy (42).

In addition, we have shown that the combination of SELE rs3917412 and MTHFR rs1801133 SNPs could serve as a gender- and stage-specific genomic predictor of tumor recurrence in patients with stage II and III colon cancer. These results are consistent with previous reports which have found that polymorphisms in the MTHFR gene or other genes involved in relevant colon cancer pathways could predict gender- and stage-specific colon cancer outcome both in early-stage and advanced disease (44, 46–48). We have demonstrated that the SELE rs3917412 G/G and MTHFR rs1801133 T/T genotypes are significantly associated with a higher risk or recurrence in stage III patients, with a similar but nonsignificant trend in high-risk stage II patients. These data support the role of the SELE rs3917412 and MTHFR rs1801133 SNPs as molecular predictors of recurrence in both stages, but the limited sample size in the case of stage II disease makes it difficult to draw statistically significant conclusions. About the role of MTHFR SNPs as gender-specific prognostic markers, 2 previous reports have shown a significant association between the rs1801131 A1298C SNP and OS in female patients with metastatic colon cancer on 5-FU–based chemotherapy, but not in male patients (44, 46). However, in contrast to our findings, no differences in survival in men or women groups according to the MTHFR rs1801133 C677T SNP have been found. The underlying biologic mechanism of why sex may affect the association of MTHFR SNPs and clinical outcome is not clear. It has been postulated that female patients' estrogen status may interact with MTHFR genotype through decreasing homocysteine levels, but this impact is diminished in male patients (49). A second possible explanation is that sex is a strong predictor of methylation levels in the MTHFR gene, with males showing higher relative methylation levels (50).

Discrepancy between published data and our results could be explained by a number of factors, including differences in chemotherapy regimens, limited samples sizes, different technological platforms and sources of clinical material, and the absence of a validation cohort in most reports. Differences in ethnicity between populations may translate into contrasting pharmacogenomic interactions or activation of alternate pathways that affect tumor biology and behavior. Furthermore, it should be considered that the clinical significance of genotyping a single polymorphism is limited and may be inadequate to predict outcome in the adjuvant setting. Variations in genetic combinations and possible interdependence with other predictive enzymes are related to outcome and could also explain some of the difficulties in obtaining reproducible and uniform results when using single enzyme polymorphisms as predictive markers (13). Therefore, an approach that combines relevant SNPs of the genes involved in chemotherapy metabolism will yield a better, more logical model for explaining individual variations in treatment efficacy than single SNPs, as suggested in our study. This aspect is even more relevant if the pharmacogenetic analysis is performed in patients receiving combination chemotherapy with 2 or more drugs targeting independent biologic pathways (i.e., FOLFOX or CAPOX in our analysis).

Our study has a number of limitations such as its retrospective nature, although this point does not seem to affect the assessment of the primary endpoint (DFS). In addition, we carried out SNP testing in archival paraffin-embedded tumor samples. This could mean that we were unable to correct the effects of loss of heterozygosity (LOH), which would produce fewer heterozygous individuals than expected. However, concordance was found in the 95% to 99% of patients when genotypes for each SNP in tumor samples and adjacent nontumor tissue were compared. In addition, the fact that our population genotype distribution was in HWE with genotype frequencies in accordance with previous reports (28) suggests that LOH is of minor importance in our cohort. Finally, we have used a candidate polymorphism approach, which allowed us to focus on potentially functional SNPs reported in the literature but did not comprehensively cover all SNPs in the entire genome. Some important SNPs may have been missed or the observed associations may result from genetic linkages with other untyped SNPs. Nonetheless, despite its limitations, our analysis includes one of the largest populations studied with a long postoperative follow-up. Moreover, an almost unique advantage is that all patients were treated with oxaliplatin-based regimens along the same guidelines, presenting an exceptional opportunity for determining the effect of polymorphisms on oxaliplatin and fluoropyrimidines effect, avoiding bias introduced by varying treatment regimens and effects of prior therapies. Although imbalances in tumor grade, perineural invasion, bowel perforation, or adjuvant chemotherapy regimens between cohorts have been found, there are not statistically significant differences in the distribution of relevant SNPs genotypes according to these parameters neither in the training nor in the validation series. About this point, it should be emphasized that the ability to predict tumor recurrence and survival of our pharmacogenetic approach in 2 independent cohorts, despite some differences in baseline characteristics and adjuvant regimens, would reinforce the validity of our findings and the possibility to extrapolate them to populations other than the ones here studied.

In conclusion, our results highlight new ways to explore in order to improve the prognostic stratification of early-stage colon cancer. Screening for SELE or MTHFR SNPs could effectively classify patients treated with oxaliplatin-based adjuvant chemotherapy into groups at low and high risk of disease recurrence, thereby adding prognostic value to significant clinicopathological risk factors used to assess these patients' prognosis. Therefore, new treatment options and more intensive strategies of follow-up may be explored after surgery in patients carrying these sequence variants. Although our findings require confirmation in large-scale prospective trials, they reinforce the concept that a pharmacogenetic approach potentially offers clinical value in directing personalized selection of adjuvant chemotherapy in patients with colon cancer in order to optimize clinical outcomes.

No potential conflicts of interest were disclosed.

Conception and design: A. Custodio, J. Moreno-Rubio, A. Sánchez, J. Feliu

Development of methodology: A. Custodio, J. Moreno-Rubio, E. Díaz-López, A. Sánchez, P. Cejas, J. Feliu

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): A. Custodio, J. Moreno-Rubio, J. Aparicio, J. Gallego-Plazas, R. Yaya, J. Maurel, E. Burgos, A. Calatrava, E. Andrada, E. Díaz-López, A. Sánchez, J. Feliu

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): A. Custodio, J. Moreno-Rubio, E. Burgos, A. Sánchez, R. Madero, J. Feliu

Writing, review, and/or revision of the manuscript: A. Custodio, J. Moreno-Rubio, J. Aparicio, R. Yaya, J. Maurel, N. Rodríguez-Salas, A. Sánchez, J. Feliu

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): A. Custodio, J. Moreno-Rubio, A. Sánchez

Study supervision: A. Custodio, J. Moreno-Rubio, E. Burgos, A. Sánchez, J. Feliu

Other (histopathological evaluation of specimens, determination of histological subtypes, and determination of tumor stage): D. Ramos

This study was supported by the “Sociedad Española de Oncología Médica (SEOM)-Roche 2008-2009” and “conSEPOC” grants to J. Feliu.

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.

1.
André
T
,
Boni
C
,
Navarro
M
,
Tabernero
J
,
Hickish
T
,
Topham
C
, et al
Improved overall survival with oxaliplatin, fluorouracil, and leucovorin as adjuvant treatment in stage II or III colon cancer in the MOSAIC trial
.
J Clin Oncol
2009
;
27
:
3109
16
.
2.
Yothers
G
,
O'Connell
MJ
,
Allegra
CJ
,
Kuebler
JP
,
Colangelo
LH
,
Petrelli
NJ
, et al
Oxaliplatin as adjuvant therapy for colon cancer: updated results of NSABP C-07 trial, including survival and subset analysis
.
J Clin Oncol
2011
;
29
:
3768
74
.
3.
Haller
DG
,
Tabernero
J
,
Maroun
J
,
de Braud
F
,
Price
T
,
Van Cutsem
E
, et al
Capecitabine plus oxaliplatin compared with fluorouracil and folinic acid as adjuvant therapy for stage III colon cancer
.
J Clin Oncol
2011
;
29
:
1465
71
.
4.
Quasar Collaborative Group
Gray
R
,
Barnwell
J
,
McConkey
C
,
Hills
RK
,
Williams
NS
, et al
Adjuvant chemotherapy in patients with colorectal cancer: a randomised study
.
Lancet
2007
;
370
:
2020
9
.
5.
Figueredo
A
,
Coombes
ME
,
Mukherjee
S
. 
Adjuvant therapy for completely resected stage II colon cancer
.
Cochrane Database Syst Rev
2008
;
16
:
CD005390
.
6.
Sinicrope
FA
,
Mahoney
MR
,
Smyrk
TC
,
Thibodeau
SN
,
Roepman
P
,
Bertagnolli
MM
, et al
Prognostic impact of deficient DNA mismatch repair in patients with stage III colon cancer from a randomized trial of FOLFOX-based adjuvant chemotherapy
.
J Clin Oncol
2013
;
31
:
3664
72
.
7.
Salazar
R
,
Roepman
P
,
Capella
G
,
Moreno
V
,
Simon
I
,
Dreezen
C
, et al
Gene expression signature to improve prognosis prediction of stage II and III colorectal cancer
.
J Clin Oncol
2011
;
29
:
17
24
.
8.
Tejpar
S
,
Bertagnolli
M
,
Bosman
F
,
Lenz
HJ
,
Garraway
L
,
Waldman
F
, et al
Prognostic and predictive biomarkers in resected colon cancer: current status and future perspectives for integrating genomics into biomarker training
.
Oncologist
2010
;
15
:
390
404
.
9.
Zhang
JX
,
Song
W
,
Chen
ZH
,
Wei
JH
,
Liao
YJ
,
Lei
J
, et al
Prognostic and predictive value of a microRNA signature in stage II colon cancer: a microRNA expression analysis
.
Lancet Oncol
2013
;
14
:
1295
306
.
10.
Lurje
G
,
Zhang
W
,
Yang
D
,
Groshen
S
,
Hendifar
AE
,
Husain
H
, et al
Thymidylate synthase haplotype is associated with tumour recurrence in stage II and stage III colon cancer
.
Pharmacogenetic Genomics
2008
;
18
:
161
8
.
11.
Delgado-Plasencia
L
,
Medina-Arana
V
,
Bravo-Gutiérrez
A
,
Pérez-Palma
J
,
Álvarez-Argüelles
H
,
Salido-Ruiz
E
, et al
Impact of the MTHFR C677T polymorphism on colorectal cancer in a population with low genetic variability
.
Int J Colorectal Dis
2013
;
28
:
1187
93
.
12.
Negandhi
AA
,
Hyde
A
,
Dicks
E
,
Pollet
W
,
Younghusband
BH
,
Parfrey
P
, et al
MTHFR Glu429Ala and ERCC5 His46His polymorphisms are associated with prognosis in colorectal cancer patients: analysis of two independent cohorts from Newfoundland
.
PLoS ONE
2013
;
8
:
e61469
.
13.
Taflin
H
,
Wettergren
Y
,
Odin
E
,
Carisson
G
,
Derwinger
K
. 
Gene polymorphisms MTHFRC677T and MTRA2756G as predictive factors in adjuvant chemotherapy for stage III colorectal cancer
.
Anticancer Res
2011
;
31
:
3057
62
.
14.
Fariña Sarasqueta
A
,
van Lijnschoten
G
,
Lemmens
VE
,
Rutten
HJ
,
van den Brule
AJ
. 
Pharmacogenetics of oxaliplatin as adjuvant treatment in colon carcinoma: are single-nucleotide polymorphisms in GSTP1, ERCC1, and ERCC2 good predictive markers?
Mol Diagn Ther
2011
;
15
:
277
83
.
15.
Yin
M
,
Yan
J
,
Martínez-Balibrea
E
,
Graziano
F
,
Lenz
HJ
,
Kim
HJ
, et al
ERCC1 and ERCC2 polymorphisms predict clinical outcomes of oxaliplatin-based chemotherapies in gastric and colorectal cancer: a systemic review and meta- analysis
.
Clin Cancer Res
2011
;
17
:
1632
40
.
16.
Ye
F
,
Liu
Z
,
Tan
A
,
Liao
M
,
Mo
Z
,
Yang
X
. 
XRCC1 and GSTP1 polymorphisms and prognosis of oxaliplatin-based chemotherapy in colorectal cancer: a meta-analysis
.
Cancer Chemother Pharmacol
2013
;
71
:
733
40
.
17.
Lee
KH
,
Chang
HJ
,
Han
SW
,
Oh
DY
,
Im
SA
,
Bang
YJ
, et al
Pharmacogenetic analysis of adjuvant FOLFOX for Korean patients with colon cancer
.
Cancer Chemother Pharmacol
2013
;
71
:
843
51
.
18.
Madi
A
,
Fisher
D
,
Maughan
T
,
James
Colley
,
Angela
M
. 
Meade, Julie Maynard, et al. Comprehensive pharmacogenetic profiling of advanced colorectal cancer
.
J Clin Oncol
2013
;
31
(
suppl; abstr 3509
).
19.
Lurje
G
,
Zhang
W
,
Schultheis
AM
,
Yang
D
,
Groshen
S
,
Hendifar
AE
, et al
Polymorphisms in VEGF and IL-8 predict tumour recurrence in stage III colon cancer
.
Ann Oncol
2008
;
19
:
1734
41
.
20.
Hebbar
M
,
Adenis
A
,
Révillion
F
,
Duhamel
A
,
Romano
O
,
Truant
S
, et al
E-selectin gene S128R polymorphisms is associated with poor prognosis in patients with stage II or III colorectal cancer
.
Eur J Cancer
2009
;
45
:
1871
6
.
21.
Lin
M
,
Gu
J
,
Eng
C
,
Ellis
LM
,
Hildebrandt
MA
,
Lin
J
, et al
Genetic polymorphisms in microRNA-related genes as predictors of clinical outcomes in colorectal adenocarcinoma patients
.
Clin Cancer Res
2012
;
18
:
3982
91
.
22.
Absenger
G
,
Benhaim
L
,
Szkandera
J
,
Zhang
W
,
Yang
D
,
Labonte
MJ
, et al
The cyclin D1 (CCND1) rs9344 G>A polymorphism predicts clinical outcome in colon cancer patients treated with adjuvant 5-FU-based chemotherapy
.
Pharmacogenomics J
2014
;
14
:
130
4
.
23.
Lee
S
,
Kwon
MS
,
Park
T
. 
Network graph analysis of gene-gene interactions in genome-wide association study data
.
Genomics Inform
2012
;
10
:
256
62
.
24.
Oeth
P
,
del Mistro
G
,
Marnellos
G
,
Shi
T
,
van den Boom
D
. 
Qualitative and quantitative genotyping using single base primer extension coupled with matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MassARRAY)
.
Methods Mol Biol
2009
;
578
:
307
43
.
25.
Meyer
K
,
Ueland
PM
. 
Use of matrix-assisted laser desorption/ionization time-of-flight mass spectrometry for multiplex genotyping
.
Adv Clin Chem
2011
;
53
:
1
29
.
26.
Afzal
S
,
Jensen
SA
,
Vainer
B
,
Dvorak
A
,
Roy
S
,
Alyasiri
A
, et al
MTHFR polymorphisms and 5-FU-based adjuvant chemotherapy in colorectal cancer
.
Ann Oncol
2009
;
20
:
1660
6
.
27.
Huang
MY
,
Fang
WY
,
Lee
SC
,
Cheng
TL
,
Wang
JY
,
Lin
SR
. 
ERCC2 2251A>C genetic polymorphisms was highly correlated with early relapse in high-risk stage II and stage III colorectal cancer patients: a preliminary study
.
BMC Cancer
2008
;
8
:
50
.
28.
Wiese
G
,
Barthel
SR
,
Dimitroff
CJ
. 
Analysis of physiologic E-selectin-mediated leukocyte rolling on microvascular endothelium
.
J Vir Exp
2009
;
11
:
pii:1009
.
29.
Gout
S
,
Tremblay
PL
,
Huot
J
. 
Selectins and selectins ligands and extravasation of cancer cells and organ selectivity of metastasis
.
Clin Exp Metastasis
2008
;
25
:
335
44
.
30.
Alessandro
R
,
Seidita
G
,
Flugy
AM
,
Damiani
F
,
Russo
A
,
Corrado
C
, et al
Role of S128R polymorphism of E-selectin in colon metastasis formation
.
Int J Cancer
2007
;
121
:
528
35
.
31.
Tremblay
PL
,
Huot
J
,
Auger
FA
. 
Mechanisms by which E-selectin regulates diapedesis of colon cancer cells under flow conditions
.
Cancer Res
2008
;
68
:
5167
76
.
32.
Porquet
N
,
Poirier
A
,
Houle
F
,
Pin
AL
,
Gout
S
,
Tremblay
PL
, et al
Survival advantages conferred to colon cancer cells by E-selectin-induced activation of the PI3K-NFκB survival axis downstream of death receptor-3
.
BMC Cancer
2011
;
11
:
285
.
33.
Köhler
S
,
Ullrich
S
,
Ritcher
U
,
Schumacher
U
. 
E-/P- selectins and colon carcinoma metastasis: first in vivo evidence for their crucial role in a clinically relevant model of spontaneous metastasis formation in the lung
.
Br J Cancer
2010
;
102
:
602
9
.
34.
St Hill
CA
,
Baharo-Hassan
D
,
Farooqui
M
. 
C2-O-sLeX glycoproteins are E-selectin ligands that regulate invasion of human colon and hepatic carcinoma cells
.
PLoS ONE
2011
;
6
:
e16281
.
35.
Cooper
DN
. 
Functional intronic polymorphisms: buried treasure awaiting discovery within our genes
.
Hum Genomics
2010
;
4
:
284
8
.
36.
De Mattia
E
,
Toffoli
G
. 
C677T and A1298C MTHFR polymorphisms, a challenge for antifolate and fluoropyrimidine-based therapy personalization
.
Eur J Cancer
2009
;
45
:
1333
51
.
37.
Derwinger
K
,
Wettergren
Y
,
Odin
E
,
Carlsson
G
,
Gustavson
B
. 
A study of the MTHFR gene polymorphism C677T in colorectal cancer
.
Clin Colorectal Cancer
2009
;
8
:
43
8
.
38.
Marcuello
E
,
Altés
A
,
Menoyo
A
,
Rio
ED
,
Baiget
M
. 
Methylenetetrahydrofolate reductase gene polymorphisms: genomic predictors of clinical response to fluoropyrimidine-based chemotherapy
.
Cancer Chemother Pharmacol
2006
;
57
:
835
40
.
39.
Ruzzo
A
,
Graziano
F
,
Loupakis
F
,
Rulli
E
,
Canestrari
E
,
Santini
D
, et al
Pharmacogenetic profiling in patients with advanced colorectal cancer treated with first-line FOLFOX-4-chemotherapy
.
J Clin Oncol
2007
;
25
:
1247
54
.
40.
Etienne-Grimaldi
MC
,
Milano
G
,
Maindrault-Goebel
F
,
Chibaudel
B
,
Formento
JL
,
Francoual
M
, et al
Methylenetetrahydrofolate reductase (MTHFR) gene polymorphisms and FOLFOX response in colorectal cancer patients
.
Br J Clin Pharmacol
2010
;
69
:
58
66
.
41.
Cohen
V
,
Panet-Raymond
V
,
Sabbaghian
N
,
Morin
I
,
Batist
G
,
Rozen
R
. 
Methylenetetrahydrofolate reductase polymorphism in advanced colorectal cancer: a novel genomic predictor of clinical response to fluoropyrimidine-based chemotherapy
.
Clin Cancer Res
2003
;
9
:
1611
5
.
42.
Ettiene
MC
,
Formentto
JL
,
Chazal
M
,
Francoual
M
,
Magné
N
,
Formento
P
, et al
Methylenetetrahydrofolate reductase polymorphisms and response to fluorouracil-based treatment in advanced colorectal cancer patients
.
Pharmacogenetics
2004
;
14
:
785
92
.
43.
Jakobsen
A
,
Nielsen
JN
,
Gyldenkerne
N
,
Lindeberg
J
. 
Thymidylate synthase and methylenetetrahydrofolate reductase gene polymorphism in normal tissue as predictors of fluorouracil sensitivity
.
J Clin Oncol
2005
;
23
:
1365
9
.
44.
Zhang
W
,
Press
OA
,
Haiman
CA
,
Yang
DY
,
Gordon
MA
,
Fazzone
W
, et al
Association of methylenetetrahydrofolate reductase gene polymorphisms and sex-specific survival in patients with metastatic colon cancer
.
J Clin Oncol
2007
;
25
:
3726
31
.
45.
Kawakami
K
,
Ruszkiewicz
A
,
Bennett
G
,
Moore
J
,
Watanabe
G
,
Iacopetta
B
. 
The folate pool in colorectal cancers is associated with DNA hypermethylation and with a polymorphism in methylenetrahydrofolate reductase
.
Clin Cancer Res
2003
;
9
:
5860
5
.
46.
Paré
L
,
Salazar
J
,
del Río
E
,
Baiget
M
,
Altés
A
,
Marcuello
E
, et al
Methylenetetrahydrofolate reductase gene polymorphisms: genomic predictors of clinical response to fluoropyrimidine-based chemotherapy in females
.
Clin Oncol
2008
;
26
:
3468
.
47.
Benhaim
L
,
Gerger
A
,
Bohanes
P
,
Páez
D
,
Wakatsuki
T
,
Yang
D
, et al
Gender-specific profiling in SCN1A polymorphisms and time-to-recurrence in patients with stage II/III colorectal cancer treated with adjuvant 5-fluorouracil chemotherapy
.
Pharmacogenomics J
2014
;
14
:
135
41
.
48.
Ning
Y
,
Gerger
A
,
Zhang
W
,
Hanna
DL
,
Yang
D
,
Winder
T
, et al
Plastin polymorphisms predict gender and stage-specific colon cancer recurrence after adjuvant chemotherapy
.
Mol Cancer Ther
2014
;
13
:
528
39
.
49.
Castro
R
,
Rivera
I
,
Ravasco
P
,
Jakobs
C
,
Blom
HJ
,
Camilo
ME
, et al
5,10-methylenetetrahydrofolate reductase 677C→T and 1298A→C mutations are genetic determinants of elevated homocysteine
.
QJM
2003
;
96
:
297
303
.
50.
Sarter
B
,
Long
TI
,
Tsong
WH
,
Koh
WP
,
Yu
MC
,
Laird
PW
, et al
Sex differential in methylation patterns of selected genes in Singapore Chinese
.
Hum Genet
2005
;
117
:
402
3
.