Abstract
Purpose: Gene silencing by hypermethylation plays an important role in proximal colon carcinogenesis. Conversely, DNA hypomethylation has been associated with distal colon cancer (CLC). Methylenetetrahydrofolate reductase (MTHFR) catalyzes the conversion of 5′,10′-methylenetetrahydrofolate to 5′-methyl tetrahydrofolate, which serves as methyl donor in the remethylation of homocysteine to methionine. A common MTHFR 677 C→T polymorphism is characterized by reduced catalytic activity, which affects methionine synthesis and DNA methylation.
The aim of the study was to investigate the role of MTHFR 677 C→T gene polymorphism in the tumorigenesis of proximal and distal CLC in a monoinstitutional group of patients in North Italy.
Experimental Design: One-hundred thirty four consecutive proximal and 142 consecutive distal CLC patients, and 279 control subjects without cancer were genotyped for MTHFR using PCR-restriction fragment-length polymorphism analysis.
Results: The prevalence of the 677 TT genotype was significantly (P = 0.005) lower in patients with proximal tumors (10 of 134, 7%) than in subjects with distal tumors (28 of 142, 20%). Case/control approach indicated that individuals homozygous for the 677 TT allele had a significantly reduced risk (2.8-fold) (adjusted odds ratio, 0.36; 95% confidence intervals, 0.14–0.91) of developing proximal CLC compared with those harboring the wild-type or heterozygous genotype (677 CC or 677 CT). No significant association between CLC risk and TT genotype was observed in patients with distal tumors (odds ratio, 1.01; 95% confidence interval, 0.48–2.14).
Conclusions: Our findings support a role for MTHFR 677 TT genotype in reducing proximal CLC risk in North Italy.
INTRODUCTION
Genetic and epigenetic factors affecting DNA methylation and gene expression are involved in the development of CLC3 (reviewed in Ref. 1 ), and it is debated whether a diet deficient in methyl groups, particularly folate, and vitamins B6, B12, and methionine predisposes to CLC (2) . Deficiency in methyl groups lowers the concentration of SAM, possibly reducing DNA methylation and decreasing the synthesis of thymidine from uracil, resulting in misincorporation of uracil in place of thymidine, leading to DNA strand breaks (3) . MTHFR plays a central role in folate metabolism, and is an important factor in DNA methylation (4) , synthesis, and repair (5) . MTHFR catalyzes the reduction of 5,10-methylenetetrahydrofolate to 5-methyltetrahydrofolate, which converts homocysteine to methionine, the precursor of SAM, which acts as a methyl donor for intracellular methylation reactions, particularly DNA methylation. Two common germ-line mutations have been identified in the MTHFR locus at nucleotides 677 (677 C→T, alanine→valine) and 1298 (1298 A→C, glutamate→alanine). Both variants are associated with an increased thermolability and diminished specific activity of the enzyme, but the 677 C→T variant has a major role in affecting MTHFR activity (6 , 7) . Whereas the clinical significance of the MTHFR 1298 A→C mutation needs to be clarified, the clinical implications of the MTHFR 677 C→T mutation are better understood. Patients with the 677 C→T variant have impaired remethylation of homocysteine to methionine, and subsequent hyperhomocysteinaemia and diminished genomic DNA methylation, particularly when folate intake is inadequate (4) . The MTHFR 677 TT genotype is associated with cardiovascular disease, congenital abnormalities, and pregnancy outcome. An increased risk associated with the TT genotype for acute leukemia, esophageal squamous cell and gastric carcinoma, and other cancers has been reported (5 , 8, 9, 10) , but the role of the MTHFR 677 C→T variant in the risk for CLC remains controversial (11, 12, 13, 14, 15, 16) . Previous studies have suggested that the MTHFR 677 TT genotype is slightly inversely associated with CLC. However, this association needs additional explanations because distal and proximal CLC tumors seem to have different tumorigenesis mechanisms (17) , which could be affected by the MTHFR variant in a different manner. In particular, associations have been reported between proximal CLC location and hypermethylation of genes involved in mismatch repair leading to MSI or genes involved in cell cycle control (18 , 19) . Conversely, hypomethylation has been associated with the chromosomal instability that characterizes distal CLCs (1 , 17) .
The aim of the study was to investigate whether the MTHFR 677 C→T gene polymorphism has a different role in the tumorigenesis of proximal and distal CLC in a monoinstitutional group of patients in North Italy. The prevalence of MTHFR 1298 A→C mutation in proximal and distal CLC was also investigated.
MATERIALS AND METHODS
Sample Collection.
The study included 276 patients with CLC and 279 healthy controls. One hundred thirty-four consecutive patients with disease located in the proximal colon (i.e., from the caecum to the splenic flexure) and 142 consecutive cases with distal colon (i.e., descending and sigmoid colon) underwent primary surgery at the Arcispedale S. Anna, Ferrara, Italy between January 1999 and December 2000. Selection was performed to have ∼50% proximal CLCs and ∼50% distal CLCs. All of the cases had conventional histopathological parameters analysis, including American Joint Committee on Cancer/Unio Internationale Contra Cancrum Tumor-Node-Metastasis stage, tumor type, and grade of differentiation. Tumors were typed as adenocarcinomas or mucinous adenocarcinomas according to the criteria of the WHO (20) . Grade of differentiation (well/moderate or poor) was defined following the WHO guidelines (20) . One-hundred ninety six patients received adjuvant/cytoriductive chemotherapy after primary surgery. All of the cases were from the same geographic area as the 279 consecutive control individuals, who were blood donors and provided peripheral blood. All of the patients and donors gave informed consent. The study was approved by the Review Board of the Institution. Of the total 276 cases, 35 subjects were excluded from MSI analysis, because tumor DNA was not available.
Analysis of the MTHFR Genotype.
Normal (nontumoral) DNA was used for MTHFR analysis. The source for normal DNA was peripheral blood lymphocytes for healthy controls and peripheral blood lymphocytes or normal colon mucosa obtained during primary surgery for patients. Genomic DNA was extracted with conventional methods from frozen (−80°C) or formalin-fixed, paraffin-embedded tissues. MTHFR 677 genotyping was performed using a 100 Programmable Thermal Controller (MJ Research, Inc., Genenco, Florence, Italy). Two-hundred ng of human genomic DNA were amplified with 100 pmol each of forward primer 5′-GCACTTGAAGGAGAAGGTGTC-3′ and reverse primer 5′-AGGACGGTGCGGTGAGAGTG-3′, 1.5 mm MgCl2, 5% formamide, 0.2 μm each deoxynucleotide triphosphate and 1 unit Taq Polymerase (Polimed, Florence, Italy) in a total volume of 100 μl. PCR conditions were as follows: denaturation at 94°C for 5 min, followed by 35 cycles at 94°C for 30 s, 51°C for 30 s, and 72°C for 30 s; the terminal elongation was performed at 72°C for 5 min. The PCR products were ethanol precipitated, digested with TaqI and run on a 4% agarose gel. The variant allele is created by a C to T change at nucleotide 677, which introduces a TaqI site; homozygous individuals (TT) show two fragments of 173 and 30 bp, heterozygous individuals (CT) show three fragments of 203, 173, and 30 bp, and wild-type individuals (CC) show only one band of 203 bp. DNA from three lymphoblastoid cell lines with the 677 TT, 677 CT, and 677 CC genotypes was used as quality control. Data obtained with TaqI were also validated by comparing results obtained with HinfI digestion according to the methods described by Frosst et al. (6) . Laboratory investigators were blinded to the colon localization.
The presence of the MTHFR 1298 polymorphism was analyzed using the method of Skibola et al. (8) with slight modifications. DNA was amplified with the forward primer 5′-CTTTGGGGAGCTGAAGGACTACTAC-3′ and the reverse primer 5′-CACTTTGTGACCATTCCGGTTTG-3′. Amplification and PCR conditions were the same as for MTHFR 677. The variant allele is created by an A to C change at nucleotide 1298, which abolishes a MboII site; wild-type individuals (AA) show five bands of 56, 31, 30, 28, and 18 bp, whereas homozygous variant individuals (CC) show four bands of 84, 31, 30, and 18 bp. The restricted product was analyzed by electrophoresis in a 4% agarose gel stained with ethidium bromide.
Microsatellite Analysis.
DNA was extracted from formalin-fixed, paraffin-embedded or frozen tissue. Tumor specimens were chosen by microscopic examination, carefully avoiding areas enriched in nonneoplastic cells and necrotic debris. Tumor and corresponding normal DNA samples (normal mucosa) were analyzed for MSI with the Bethesda panel of five microsatellite markers (BAT26, BAT25, D2S123, D5S346, and D17S250; Ref. 21 ) and a fluorescence-based multiplex PCR method, using the hereditary nonpolyposis colorectal carcinoma Microsatellite Instability Test kit (Boehringer Mannheim, Mannheim, Germany) according to manufacturer’s instructions. PCR products were run in an ABI PRISM 377 DNA sequencer (Perkin-Elmer Applied Biosystems Division, Foster City, CA) and analyzed by the GeneScan 3.1 software (Perkin-Elmer). Five additional microsatellites: BAT40, D10S197, D18S58, D18S69, and l-myc were analyzed when only one locus from the first panel showed MSI or when one or more loci were not evaluable. Tumors were scored has having MSI-H when additional bands were present in the PCR products from at least two loci of the first panel or from at least 30% of all of the examined loci, if additional microsatellites were analyzed. Tumors were scored as having MSS/MSI-L when any (MSS) or <30% (MSI-L) of the examined loci showed additional bands (alleles or MSI).
Statistical Analysis.
The distribution of individual characteristics was evaluated by simple descriptive statistics. Differences among distributions of selected variables were evaluated by use of the Fisher exact tests for categorical data. The Kruskal-Wallis test was performed for statistical evaluation of the significant difference of the distributions of continuous variables across the MTHFR genotype.
The Hardy-Weinberg equilibrium assumption was assessed by the standard method of matching the observed numbers of individuals in the different genotype categories with those expected under Hardy-Weinberg equilibrium for the estimated allele frequency and comparing the Pearson goodness-of-fit statistic with the χ2 distribution with 1 degree of freedom. Genotype distributions were compared with the use of contingency table analysis.
OR, and corresponding 95% CI, were computed using multiple logistic regression models using as dependent variable case/control status.
A multinomial logistic regression model was used to permit, simultaneously, the analysis of two subgroups of CLC cases classified according to MSI status and/or site and the series of population controls. All of the statistical tests and Ps were two-tailed.
RESULTS
677 MTHFR C→T Polymorphism.
Two hundred and seventy-six CLC patients entered the study; 134 (48.6%) consecutive patients had proximal CLC and 142 (51.4%) consecutive patients had distal CLC. Patient characteristics in relation to the 677 MTHFR genotype are summarized in Table 1⇓ . Data were compared with those obtained in a series of 279 consecutive blood donors used as control. Of all of the cases, 93 (33.7%) had the CC, 145 (52.5%) the CT, and 38 (13.8%) the TT genotype; allele frequencies were T = 40% and C = 60%. The distribution among controls was 83 CC (29.8%), 140 CT (50.2%), and 56 TT (20.1%); allele frequencies were T = 45.2% and C = 54.8%. These genotype frequencies were not different from those expected from the Hardy-Weinberg equilibrium (30.1% CC, 49.5% CT, and 20.4% TT (exact P estimated by using conventional Monte Carlo method = 0.9).
Patient characteristics at baseline according to MTHFR 677 C→T polymorphism
The prevalence of the MTHFR 677 C→T mutation among cases (38 of 276, 13.8%) was significantly lower (P = 0.05) than controls (56 of 279, 20.1%). However, considering the patients according to their tumor localization, such differences remained significant in patients with proximal colon tumor (P = 0.0009) but not in patients with distal CLC. The distribution of the MTHFR 677 TT variant in the patients with proximal CLC was significantly lower than that in the patients with distal tumor (P = 0.005). Only 10 of the 134 patients (7.5%) with proximal CLC were homozygous for the T allele, whereas the TT mutation was observed in 28 of the 142 patients (19.7%) with distal CLC. The frequency of the TT genotype in distal colon patients was superimposable to that of controls (Table 2)⇓ .
Distribution of MTHFR 677 C→T polymorphism, among controls and CLC cases according to tumor localization and for MSI status
We observed a significantly decreased risk associated with the TT genotype in patients with proximal cancer compared with controls. According to logistic regression model, when the 677 MTHFR CC and CT genotypes were combined and used as the reference group, the patients with the TT genotype had ∼65% risk reduction of developing proximal CLC (OR adjusted for sex and age, 0.36; 95% CI, 0.14–0.91). Conversely, we did not observe any differences in risk associated with the MTHFR TT genotype in patients with distal cancer (OR, 1.01; 95% CI, 0.48–2.14). Considering the entire group of patients, the OR for the TT genotype was 0.72 (95% CI, 0.35–1.49; Table 2⇓ ). Finally, the relative risk in developing CLC was not in an allele-dose relationship because no significant risk differences between wild-type homozygous MTHFR 677 CC and heterozygous MTHFR 677 CT genotypes was observed when either controls were compared with proximal (OR, 0.92; 95% CI, 0.43–1.98) or distal (OR, 0.81; 95% CI, 0.37–1.80) tumors and using wild-type (CC) as reference category.
MTHFR 1298 A→C Polymorphism.
Table 3⇓ shows patient characteristics in relation to the 1298 MTHFR genotype. The distribution of MTHFR 1298 AA, AC, and CC genotypes was 122 (44.2%), 129 (46.7%), and 25 (9.1%), respectively, among cases and: AA = 133(47.6%), AC = 121 (43.4%), and CC = 25 (9.0%), respectively, among controls. The latter were also in accordance with the Hardy-Weinberg equilibrium (48.1% AA, 41.5% AC, and 9.4% CC; P = 0.7). The allele frequency for A and C was 68% and 32%, respectively, among cases, and 69% and 31%, respectively, among controls. The frequency of mutated MTHFR 1298 CC genotype was lower among proximal (7 of 134, 5.2%) than distal tumors (18 of 142, 12.7%; P = 0.04); when the MTHFR 1298 AA+AC genotype was defined as reference, the adjusted OR for MTHFR 1298 CC genotype was 0.39 (95% CI, 0.16–0.98). However, such differences among cases with proximal and distal tumor localization did not determine any statistically significant differences in CLC risk when proximal tumors were compared with controls. In the case-control statistical analysis, the adjusted OR for MTHFR 1298 CC genotype was 0.72 (95% CI, 0.23–2.73; P = 0.24) for proximal CLC and OR = 2.02 (95% CI, 0.72–5.76; P = 0.24) for distal tumors. Analysis of the joint effect between MTHFR 677 and 1298 genotypes was performed. Only 1 case (distal CLC) and 2 controls were homozygous for both MTHFR 677 TT and MTHFR 1298 CC, whereas 77 cases [42 (31.3%) proximal CLC and 35 (24.6%) distal CLC] and 72 (25.8%) controls exhibited both the heterozygous genotypes MTHFR 677 CT and MTHFR 1298 AC. No significant joint effect in the development of CLC was observed (data not shown).
Patient characteristics at baseline according to MTHFR 1298 A→C polymorphism
MSI Status and MTHFR Polymorphisms.
MSI analysis was performed in CLC from 241 patients (128 proximal and 113 distal CLC). Fifty of the 128 (39.1%) patients with proximal CLC showed MSI-H tumors, and 78 (60.9.7%) had MSS tumors (8 of these patients had MSI-L). Only 1 of the 113 patients with distal CLC showed MSI-H. The incidence of the MTHFR 677 mutation was higher in MSS (31 of 190, 16%) than in MSI tumors (3 of 51, 5.9%), even if such differences did not reach a statistical significance (P = 0.07). Comparison between MSI status and 677 MTHFR polymorphism by CLC subsets showed that the frequency of the MTHFR 677 TT genotype was not statistically different (P, not significant) in MSI (3 of 50 = 6.0%) compared with MSS tumors (7 of 78 = 9.0%) in the proximal CLC patients (Table 2)⇓ . Finally, no significant associations were observed between MSI status and MTHFR 1298 genotype. Overall, the incidence of MTHFR 1298 CC mutation was 19 of 190 (10%) among MSS, and 2 of 51 (4%) among MSI tumors (P = 0.3, NS). Frequency of MTHFR 1298 CC genotype among proximal MSS and MSI CLC was 7 of 78 (9%) and 3 of 50 (6%), respectively (P = 0.7, not significant).
DISCUSSION
Several data provide evidence that epigenetic mechanisms of gene silencing by hypermethylation play a differential role in proximal versus distal colon carcinogenesis. Hypermethylation of genes involved in mismatch repair, such as hMLH1, or in cell cycle control, such as p16INK4a and p14 (ARF) is, in fact, preferentially found in proximal rather than in distal CLC (1 , 17, 18, 19 , 22) . The transcriptional silencing of tumor suppressor or mismatch repair genes has been attributed to the aberrant cytosine methylation of promoter-region CpG islands.
In this study, we asked whether availability of methyl donors, primarily SAM, could affect the tumorigenesis of proximal and distal CLC. To this purpose, we genotyped colon patients for the C→T variants at the MTHFR 677 locus, because individuals homozygous for the MTHFR 677 C→T variant have impaired remethylation of homocysteine to methionine, the precursor of SAM (5) .
We observed a significant decrease in the distribution of MTHFR 677 TT genotype frequencies among patients with proximal CLC compared with those with distal CLC. Subjects homozygous for the MTHFR 677 TT had about a 65% risk reduction of developing proximal CLC compared with controls. However, we found no significant association between the risk of developing distal CLC and homozygosity for the MTHFR 677 T allele. These data strongly suggest a different role for MTHFR 677 TT genotype in the tumorigenesis of proximal and distal CLC. Probably, hypermethylation of CpC island, possibly involved in proximal CLC, could occur to a lesser extent in individuals with the MTHFR mutation compared with those harboring the wild-type (CC) and/or heterozygous (CT) genotype. In our study patients with CT and CC genotype were combined, because previous reports indicated only slight reduction of MTHFR activity because of the CT genotype compared with wild-type (5 , 6) . In addition, an allele-dose relationship was not observed in our study. This confirms most of the reported associations of this polymorphism with diseases or biomarkers, suggesting that disease susceptibility is not determined in a gene-dose manner (11 , 12 , 14 , 23) .
The aim of this study was to investigate the role of MTHFR 677 C→T polymorphism in the risk of developing proximal and distal CLC, respectively. For this purpose we enrolled ∼50% consecutive patients with proximal CLC and 50% with distal CLC that do not reflect the normal distribution of CLC localization (proximal:distal = 1:4). Therefore, the percentage of patients with MTHFR 677 C→T mutation in the entire group of our patients (proximal and distal) could be lower than that expected in a population with a normal distribution. In some published studies (11 , 12 , 16) CLCs were considered as a whole. This could explain controversial results reported previously in the MTHFR 677 TT genotype risk. In our study, allele and genotype frequencies among controls were consistent with those derived from the Hardy-Weinberg equilibrium, and the observed effect was not affected by other potential predictors of CLC risk such as age, sex, Dukes’ stage, and histological grade. Therefore, is unlikely that subject selection or unknown confounding factors could have biased our result in this study.
Another common polymorphism has been described in the MTHFR locus at nucleotide 1298, in which adenine is replaced by cytosine (MTHFR 1298 A→C). The frequency of the homozygous MTHFR 1298 CC genotype in proximal CLC was significantly (P = 0.04) lower than in distal CLC (OR, 0.39; 95% CI, 0.16–0.98). The risk in developing proximal tumors compared with controls was decreased ∼30%, but such differences were not statistically significant (OR, 0.72; 95% CI, 0.23–2.73). Finally, we did not find any statistically significant joint effect between MTHFR 677 and 1298 genotypes on risk of proximal CLC. These results are in agreement with previous reports indicating that the MTHFR 1298 A→C polymorphism has a lower effect on MTHFR activity than the MTHFR 677 C→T variant, which marginally affects remethylation of homocysteine (24) .
In our study we noted a trend, although not statistically significant (P = 0.07), in the association between low frequency of MTHFR 677 TT genotype and MSI tumors compared with MSS CLC. Moreover, we observed a relatively similar incidence of MTHFR TT genotype among MSI and MSS proximal tumors. This last finding is in accordance with previous reports indicating that only 45% of CLC with the type C (cancer-related) CpG-island methylator phenotype show MSI (25) . Possibly, silencing by hypermethylation of mismatch repair genes, such as hMLH1, originates proximal MSI tumors, whereas silencing by hypermethylation of cycle-related genes, such as p16INK4a or p14 (ARF), leads to the development of proximal MSS tumors. On this ground MTHFR 677 TT genotype could affect the CpG-island methylator phenotype more than MSI status.
It was thought that MTHFR 677C→T polymorphism could have alternative biochemical mechanisms other than DNA hypomethylation in the tumorigenesis of CLC. In the presence of an adequate folate and B12 intake, MTHFR 677 TT individual may be at reduced risk of CLC possibly because of the intracellular increasing of 5,10-methylenetetrahydrofolate, a cofactor for the novo synthesis of nucleotides and DNA repair (5) . In our study, we have no data on folate intake; therefore, we cannot draw any conclusion about the role of dietary folates and intracellular 5,10-methylenetetrahydrofolate levels in our patients. Nevertheless, we think that with regard to proximal tumors, folate intake could have some preventive effect only in patients with the 677 MTHFR TT mutation as also suggested by Slattery et al. (14) . Conversely, folate could be detrimental or have no effect in developing proximal CLC in subjects with no mutated 677 MTHFR genotypes. Probably, in this case, a chemopreventive therapy with drugs inducing DNA hypomethylation could be more indicated.
In conclusion, our data demonstrate that MTHFR 677 TT genotype is a determinant factor in lowering the risk of developing proximal CLC in North Italy, without affecting distal colon carcinogenesis, and confirm a different role of DNA methylation in the developing of proximal and distal CLC tumors. We are confident that the present data will stimulate biological and clinical studies aimed at better understanding CLC tumorigenesis and developing new chemoprevention strategies based on MTHFR genotype and CLC localization.
Footnotes
-
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 Supported in part by a grant of Ricerca Finalizzata, Ministry of Health, Italy (N°118-2001).
-
↵2 To whom requests for reprints should be addressed, at Experimental and Clinical Pharmacology Unit, CRO-National Cancer Institute, Via Pedemontana Occidentale 12, 33081 Aviano (PN), Italy. Phone: 39-0434-659612; Fax: 39-0434-659659; E-mail: gtoffoli{at}cro.it
-
↵3 The abbreviations used are: CLC, colon cancer; MTHFR, methylentetrahydrofolate reductase; SAM, S-adenosylmethionine; MSI, microsatellite instability; OR, odds ratio; MSS, microsatellite stable; MSI-H, microsatellite instability high; MSI-L, microsatellite instability low; CI, confidence interval.
- Received April 29, 2002.
- Revision received September 10, 2002.
- Accepted September 11, 2002.
References
Volume 9, Issue 2
