The MTHFR 1298A>C Polymorphism and Genomic DNA Methylation in Human Lymphocytes

The altered properties of lymphoid neoplastic cells have been traditionally attributed to changes in the gene sequence or structure such as point mutations, gene deletions/rearrangements, or double strand breaks with subsequent increased risk of chromosomal translocations. Epigenetic abnormalities, widely described in several solid tumor malignancies, have been similarly found in hematopoietic cancer cells, suggesting that epigenetic phenomena can play a substantial role also in the carcinogenetic process affecting this particular tissue (1). DNA methylation is the most important epigenetic feature of eukaryotic DNA for the regulation of gene expression and the maintenance of genome integrity (2, 3). Preserving the patterns of DNA methylation is a major feature of DNA for the epigenetic control of genome stability and the regulation of gene transcription (4). An altered profile of DNA methylation is a frequent finding in cancer diseases, indicating that an aberrant regulation of the mechanisms of DNA methylation is crucial for the understanding of mechanisms of carcinogenesis (1, 5) and may constitute a good target for future therapies (5, 6). Abnormal genomewide DNA methylation is a phenomenon known to lead to chromosome instability and damage (7), genetic abnormalities that are frequently described in malignancies, mostly those of the lymphoid tissue.

A crucial role for the control of DNA methylation as well as DNA synthesis and repair mechanisms is played by the provision of adequate pools of methyl groups through the functional activity of methylenetetrahydrofolate reductase (MTHFR) in one-carbon metabolism. The availability of methyl groups for DNA methylation reactions is mainly regulated by the activity of MTHFR, which is indeed a fundamental enzyme in one-carbon metabolism. Its substrate, 5,10-methylenetrahydrofolate, is necessary for thymidylate and purine synthesis and its enzymatic product regulates the folate-dependent de novo synthesis of 5-methyltetrahydrofolate, the prominent circulatory form of folate and the precursor of S-adenosylmethionine, universal methyl donor for several biological methylation reactions, including that of DNA (8). We recently described that the different availability of methylated forms of folate in homozygous mutant carriers of the MTHFR 677C>T polymorphism (9) affects genomic DNA methylation through gene-nutrient interaction (10, 11). We observed that carriers of the 677TT genotype have a diminished level of DNA methylation compared with those with the 677CC wild-type genotype but only the 677TT subjects with low levels of folate accounted for the diminished DNA methylation, thus indicating that the MTHFR 677C>T polymorphism influences DNA methylation level through an interaction with folate status (11, 12).

A second common mutation in the MTHFR gene, 1298A>C, was reported. First described in association with ovarian carcinoma (13), it was subsequently investigated for its association with risk of neural tube defects, a group of diseases closely linked to one-carbon metabolism functions (14). However, whether or not these associations are ultimately due to a reduced activity of the mutant MTHFR 1298A>C enzyme is yet controversial (14-18). Most recently, polymorphic sites in folate metabolizing genes, particularly those affecting MTHFR activity, have been associated with a reduced risk of adult (19-22) and childhood (23-25) acute lymphoblastic leukemia development as well as with modulation of the response to chemotherapy (26). Furthermore, epidemiologic studies have shown an association between MTHFR polymorphisms and susceptibility to cancer affecting also other tissues (27-30). Because of the fundamental role of folate in nucleic acid synthesis and repair systems (31) as well as DNA methylation (12), alterations in this metabolic pathway seem an elective candidate target for the understanding of folate-related carcinogenesis with potential diagnostic and therapeutic implications (25, 26). Furthermore, it is known that MTHFR 677C>T and folate, independently or interactively, are important determinants of genome stability (7), cell proliferation, and chromosomal breakage, particularly in lymphoid cells (32-35).

Recent reports indicate that the MTHFR 1298A>C polymorphism is associated with a decreased risk of acute lymphoblastic leukemia in adults and children carrying the mutant allele either in heterozygosity (1298AC) or homozygosity (1298CC; refs. 19, 24). Furthermore, it has been reported that individuals with the combined 1298AA/677TT or 1298CC/677CC genotype show respectively a 2.5- and 3.3-fold reduction of risk for childhood acute lymphoblastic leukemia (25). Although several hypotheses have been proposed, there is yet no clear evidence for a link between this polymorphism and DNA methylation status.

We designed the present study precisely with the aim of evaluating the interaction between the MTHFR 1298A>C mutation and folate in lymphocyte DNA methylation status as a potential mechanism by which this polymorphism exerts a protective effect, particularly on the risk of lymphoid cell malignancies, either alone or in combination with the MTHFR 677C>T.

We studied 198 unrelated, age- and sex-matched subjects from a single geographic area and recruited those referred to the Section of Internal Medicine of the Clinical and Experimental Medicine Department and to the Institute of Cardiology and Cardiovascular Surgery of the University of Verona (Verona, Italy), as previously described (36). Subjects for whom a complete set of biochemical data pertaining to folate-homocysteine metabolism (plasma folate levels, vitamin B12, and total plasma homocysteine concentrations) and a complete clinical history were available, were selected for the present study on gene-nutrient interaction between folate and MTHFR polymorphisms. Subjects were considered as a whole with no distinction referred to their pathologic status, and excluded from the study were those with conditions known to influence homocysteine levels such as those subjects taking folate or vitamin B₁₂ or any multivitamin supplementation, those with any major acute illness, and those using drugs interfering with folate metabolism (i.e., penicillamine, anticonvulsants, and methotrexate). Informed consent was obtained from all the subjects and the study received the approval of our Institutional Review Board.

Samples of venous blood were drawn from each subject after an overnight fast. DNA was extracted from peripheral lymphocytes using a phenol/chloroform/isoamyl alcohol protocol. All the laboratory methods for biochemical analyses have been done as previously described (36). Levels of total homocysteine (tHcy) were measured by high-performance liquid chromatography with fluorescent detection according to Araki and Sako (37). Plasma folate was measured by an automated chemiluminescence method (Chiron Diagnostics, East Walpole, MA).

For the MTHFR 677C>T single nucleotide polymorphism, we used primer sets, PCR-RFLP method, and nomenclature as previously described by Frosst et al. (38), after optimizing the conditions in our laboratory. The 1298A>C mutation analysis was done as described by Weisberg et al. (ref. 15; i.e., by PCR followed by Fnu4HI endonuclease restriction). Positive and negative mutation controls were always included in each gel during the study for the genotyping of both the 677C>T and 1298A>C mutations (15).

To measure genomic DNA methylation, we used a recently developed liquid chromatography electrospray ionization mass spectrometry method that allows to define DNA methylation status as the absolute amount (in nanograms) of 5-methyl-2′-deoxycytidine (mCyt) per microgram of DNA (11, 39). Briefly, 1 μg of DNA was hydrolyzed by sequential digestion with three nucleases. DNA hydrolyzates were then injected onto an analytic column and subsequently separated by reversed-phase high-performance liquid chromatography in isocratic mode. The four major DNA bases and 5-methyl-2′-deoxycytidine were resolved and eluted in a short time run. The identification of 5-methyl-2′-deoxycytidine was obtained by spectral analysis of chromatographic peaks by isotope-labeled compounds as internal standard (39).

All computations were done by using the Systat software 10.0 program for Windows (2000, SPSS, Inc., Chicago, IL). The distribution of continuous variables in groups is expressed as mean ± SD. Analysis was done with log-transformed data for all skewed variables including DNA methylation. Geometric means (antilogarithms of the transformed means) are presented, and 95% confidence intervals are calculated using the transformed values and displayed as the antilogarithms of the transformed data. Statistical significance for differences in continuous variables was tested by Student's unpaired t test or by AN(C)OVA with Tukey's post hoc comparison of the means when appropriate. Categorical variables were analyzed using a χ² test or linear regression analysis when appropriate. Statistical significance refers to a two-tailed analysis where P < 0.05.

Each end point was evaluated for the three MTHFR 1298A>C polymorphism genotypes (1298AA, 1298AC, and 1298CC; Table 1; Fig. 1) as well as for the three different combined genotype groups for 1298A>C and 677C>T polymorphisms (1298AA/677TT, 1298AA/677CC, and 1298CC/677CC; Table 2; Fig. 2).

MTHFR 1298A>C Polymorphism. The overall allele frequency of the 1298C allele in the studied population was 0.33 and the genotype frequency was as follows: 1298AA, 46% (n = 91); 1298AC, 33% (n = 65); and 1298CC, 21% (n = 42).

Combination of MTHFR 1298A>C and 677C>T Polymorphisms. Among 91 1298AA subjects, 19 (21%) carried a 677CC genotype and 72 (79%) carried a 677TT genotype. The 677T and 1298C alleles only occurred in trans; no individual presented a 1298CC/677TT genotype configuration.

MTHFR 1298A>C Polymorphism. As shown in Table 1, plasma folate concentrations did not differ among the three genotype groups. However, when the subjects carrying the 1298C allele either in hetero- or homozygosity were considered together, plasma folate concentration within the 1298AA group was significantly lower than that found in those carrying at least one 1298C allele (10.87 versus 12.22 nmol/L, P < 0.05).

Combined MTHFR 1298A>C and 677C>T Polymorphisms. Plasma folate concentration was lower in the 1298AA/677TT group compared with the 1298CC/677CC group (10.73 versus 12.62 nmol/L, P = 0.043; Table 2). Comparisons of 1298AA/677TT with 1298AA/677CC and 1298CC/677CC groups did not show any significant difference (Table 2).

MTHFR 1298A>C Polymorphism. As shown in Table 1, the concentration of tHcy in the 1298AA group was significantly higher than that in either 1298AC or 1298CC group (18.71 versus 13.97 μmol/L and 18.71 versus 14.45 μmol/L, P < 0.0001 and P = 0.001, respectively). The difference was significant even when the carriers of the 1298C allele were considered together (1298AA versus 1298AC/1298CC; 18.71 versus 14.16 μmol/L, P < 0.0001). When folate levels were taken into consideration, the 1298AA subjects with low folate concentrations (i.e., below the median value, 12 nmol/L) showed a significantly higher tHcy concentration compared with that of 1298AA individuals with adequate folate levels (i.e., above the median value, 12 nmol/L; 22.08 versus 13.18 μmol/L, P < 0.0001). Within the 1298CC genotype carrier group, those with low plasma folate levels also had higher tHcy concentrations compared with those with high folate concentrations (19.46 versus 13.93 μmol/L, P < 0.0001). However, there was a significant difference in tHcy concentrations between 1298AA and 1298CC genotypes (19.29 versus 15.89 μmol/L, P = 0.002) at the low level of folate, but no significant difference between 1298AA and 1298AC (19.29 versus 16.27, P = 0.11). Such difference was no longer present if the comparison was made between subjects with levels of folate above the median (13.63 versus 12.62 μmol/L, P = 0.321, and 13.63 versus 13.53 μmol/L, P = 0.891, for the comparison between 1298AA versus 1298CC and 1298AA versus 1298AC, respectively).

Combination of MTHFR 1298A>C and 677C>T Polymorphisms. Total plasma homocysteine concentration in the 1298AA/677TT group was significantly higher than that in the 1298AA/677CC and 1298CC/677CC group (19.57 versus 15.79 μmol/L, P = 0.045, and 19.57 versus 14.41 μmol/L, P = 0.001, respectively; Table 2). Even when the 1298AA/677TT was compared with the 1298AA/677CC and 1298CC/677CC groups taken together, the difference remained statistically significant (19.57 versus 14.86 μmol/L, P = 0.001). However, no difference was detected for the comparison of the 1298AA/677CC with the 1298CC/677CC group (15.79 versus 14.41 μmol/L, P = 0.358; Table 2). When we analyzed the data according to plasma folate concentrations, the 1298AA/677TT group showed a significant difference between low- and high-folate groups (23.87 versus 13.09 μmol/L, P < 0.0001). A difference according to low or high plasma folate levels was also detected within the 1298CC/677CC group (16.87 versus 12.55 μmol/L, P = 0.012). However, no differences were found within the 1298AA/677CC group (16.03 versus 13.49 μmol/L, P = 0.257).

MTHFR 1298A>C Polymorphism. DNA methylation status within MTHFR 1298AA group was significantly decreased compared with that of either 1298AC (3.72 versus 8.59 ng mCyt/μg DNA, P < 0.0001) or 1298CC (3.72 versus 6.79 ng mCyt/μg DNA, P = 0.007; Table 1). DNA methylation was significantly decreased in the MTHFR 1298AA group compared with the 1298AC and 1298CC genotypes considered together (3.72 versus 7.82 ng mCyt/μg DNA, P < 0.0001). There was no significant difference between the 1298AC and 1298CC groups for DNA methylation status. However, when plasma folate levels were taken into account, the reduced DNA methylation in the 1298AA group was evident only for those subjects with low plasma folate concentrations (below the median value, i.e., 12 nmol/L), whereas under conditions of adequate plasma folate, the levels of genomic DNA methylation were similar to that of the other two MTHFR genotypes (Fig. 1).

To further evaluate the relationship between folate and MTHFR genotypes, we did additional analyses. After dividing the population sample into tertiles of plasma folate according to the three different 1298A>C genotypes, we observed that only among the individuals carrying the 1298AA genotype was there a graded effect of folate in determining levels of DNA methylation (2.39 ng mCyt/μg DNA for the lowest tertile, 5.33 ng mCyt/μg for the intermediate tertile, and 7.54 ng mCyt/μg DNA for the highest tertile, P for trend = 0.005). Among the carriers of the 1298AC or 1298CC genotype, however, DNA methylation status was not different across the increasing tertiles of plasma folate.

Combination of MTHFR 1298A>C and 677C>T Polymorphisms. Genomic DNA methylation status of the MTHFR 1298AA/677TT group was significantly decreased compared with that of 1298AA/677CC (3.11 versus 7.29 ng mCyt/μg DNA, P = 0.001) and 1298CC/677CC groups (3.11 versus 7.14 ng mCyt/μg DNA, P = 0.0001; Table 2). There was no significant difference between the 1298AA/677CC and 1298CC/677CC groups (Table 2). However, when considering plasma folate levels, the reduced DNA methylation was evident only for the 1298AA/677TT subjects with low plasma folate concentrations (1298AA/677TT, 2.44; 1298AA/677CC, 5.57; and 1298CC/677CC, 7.53 ng mCyt/μg DNA), whereas under conditions of adequate plasma folate, genomic DNA methylation status was similar to the other two MTHFR genotype groups (1298AA/677TT, 8.31 ng mCyt/μg DNA; 1298AA/677CC, 9.49 ng mCyt/μg DNA; and 1298CC/677CC, 6.69 ng mCyt/μg DNA; Fig. 2). No statistically significant difference for DNA methylation status was observed within the 1298AA/677CC and 1298CC/677CC groups between low and high levels of folate (5.57 versus 9.49 ng mCyt/;g DNA and 7.53 versus 6.69 ng mCyt/μg DNA, P = 0.67 and P = 0.69, respectively; Fig. 2).

Epidemiologic studies recently observed an association of polymorphic variants of the MTHFR gene with reduced risk of adult and childhood acute lymphoblastic leukemia (19, 22-25) and with the modulation of risk of other forms of cancer (27-30, 40, 41). The MTHFR 677TT homozygote mutant genotype has been consistently reported to have a protective effect against the risk of cancer of the colorectum (27, 42) and of the lymphoid tissue in several studies (19, 20, 22-25). The role of MTHFR polymorphisms in modulating cancer risk, however, is strongly associated with folate status, so that the protective effect of the 677TT genotype, under adequate folate conditions, turns to a situation of elevated risk of cancer of the colorectum among MTHFR 677TT carriers with low folate status or intake (27, 43). On the other hand, although some studies observed a protective effect also for the MTHFR 1298CC mutant genotype, results in this regard are conflicting (22, 23). Because the function of MTHFR is essential for providing methyl groups for DNA methylation (27), it is highly plausible that enzymatic variants due to functional polymorphisms may alter this major epigenetic feature of DNA, patterns of which greatly affect carcinogenesis (1, 3, 5). Different from 677C>T, the biochemical effect of the 1298A>C mutation on MTHFR function is still controversial (14-18), and its role in the risk of cancer is therefore more questionable (40, 41).

In the present study, we sought to evaluate the MTHFR 1298A>C polymorphism interaction with folate status and the subsequent effects on genomic methylation in human lymphocyte DNA as a potential mechanism underlying the association between the MTHFR 1298A>C polymorphism and the risk of cancer. We observed that lymphocyte DNA from subjects with the MTHFR 1298AA wild-type genotype has reduced genomic DNA methylation compared with that of carriers of either the 1298AC or 1298CC genotype. This finding referred only to those subjects whose plasma folate concentrations are low with no difference in DNA methylation among the three 1298A>C genotypes when stratified for high levels of folate. A further evidence of such gene-nutrient interaction was the detection of a graded effect of folate in determining DNA methylation, as shown by the increasing DNA methylation levels across increasing tertiles of plasma folate concentrations. This gene-nutrient interaction phenomenon affecting DNA methylation is supported also by the observation of a similar effect of the 1298AA genotype on tHcy levels, as shown by the higher tHcy in the 1298AA group compared with carriers of at least one 1298C allele, as similarly described for the 677TT genotypes (36). Because tHcy is a product of the intracellular hydrolysis of S-adenosylhomocysteine, which is known to be an inhibitor of DNA methylation, tHcy levels could be also considered as a marker of tissue DNA methylation status, as reported by Yi et al. (44).

The most crucial conflict in the present observation, however, is that the MTHFR 1298AA wild-type genotype for the MTHFR 1298A>C mutation shows a model of interaction with folate similar to that of the mutant MTHFR 677TT genotype. When considered by combined polymorphic genotypes, the 1298AA/677TT group showed decreased plasma folate concentration, increased tHcy, and decreased DNA methylation status compared with the 1298AA/677CC and 1298CC/677CC groups, and even when folate concentration was taken into account, only the 1298AA/677TT genotype group showed a decreased genomic DNA methylation when plasma folate levels are low. Furthermore, when we compared the 1298AA/677CC and 1298CC/677CC groups, to avoid any possible influence of the mutant 677TT genotype, there was no difference in plasma folate, tHcy, and lymphocyte genomic DNA methylation. These data suggest that the MTHFR 1298A>C polymorphism does not impair the enzyme function significantly enough to affect the pathway of homocysteine remethylation and biological methylation functions, thus substantiating previous findings showing no effect of the 1298A>C polymorphism per se in determining tHcy (14, 18), and leading to the understanding of a minor effect of the 1298A>C mutation on this pathway of one-carbon metabolism as a consequence of reduced MTHFR enzyme function.

Plausible explanations for the similar features observed for the 1298AA and 677TT arise from the analysis of the distribution of the 677C>T polymorphism among carriers of the 1298A>C polymorphism. Noteworthy, we observed that among 1298AA carriers, 79% had a 677TT genotype, and all of the 1298AC and 1298CC carriers had a 677CC genotype. Furthermore, the 1298C and 677T alleles only occurred in trans in our population, consistent with several previous reports on the very rare occurrence of the two polymorphisms on the same allele (14, 15, 18). This combined genotype setting, therefore, may explain the similar pattern of DNA methylation for the 1298AA and 677TT genotypes, and that of the 1298CC with the coexistence of the 677CC genotype. Indeed, such observations confirm as well the population specificity of these genotypes and could give reason for the controversial association and potentially ethnic-related association of MTHFR genotypes with acute lymphoblastic leukemia in some reports (22, 45).

Consistent with the results of the present study were the reports on the structural and biochemical characteristics of the two different enzyme variants, MTHFR 677C>T and 1298A>C (46, 47). The 677C>T polymorphism lies at the base of binding site for the MTHFR cofactor, flavin adenine dinucleotide, and has been shown to affect the MTHFR enzyme activity more significantly than the 1298A>C variant (14, 46). Individuals carrying the 677TT genotype show a 30% in vitro enzyme activity compared with the 677CC wild-type (38, 46). The MTHFR 1298A>C variant, instead, has been shown to have biochemical properties that are not markedly distinguishable from those of the 1298AA wild-type genotype (47). Further reports on the biochemical relevance of the 1298A>C mutation showed also that lymphocytes from individuals carrying the mutant 1298CC genotype have ∼60% specific wild-type MTHFR activity in vitro compared with the 1298AA wild-type (14), although such a decrease did not seem sufficient to affect tHcy levels (14, 15, 18). Thus, considering the strong effect of the MTHFR 677C>T variant on determining the availability of methylated forms of folate and on remethylation pathway (9, 11), we can speculate that the genomic DNA methylation pattern in 1298AA genotypes is mostly driven by the concomitant presence of the 677TT genotype. The effect of the 1298AA/677TT genotype also is likely due primarily to the high portion of 677TT in the combined genotype group, and the effect of both 1298AA/677CC and 1298CC/677CC combined genotypes is mainly related to the simultaneous presence of the 677CC genotype.

Therefore, it may be of crucial importance for future epidemiologic studies to evaluate both MTHFR single polymorphisms as well as combined genotype configurations in the MTHFR gene (48) and to simultaneously consider environmental factors such as folate status (49). Indeed, whereas a strong gene-nutrient interaction affects the MTHFR 677C>T polymorphism in determining DNA methylation, the mildly decreased activity due to MTHFR 1298A>C mutation may affect a different pathway in one-carbon metabolism, such as that of DNA synthesis, and therefore accounts for a different mechanism for the reduced risk of leukemia in carriers of the 1298C allele. Specific mechanistic studies are needed to clarify the molecular bases of such associations.

The evaluation of MTHFR common mutations and folate status should be considered in the view of an individually tailored assessment of cancer risk because it may as well lead to the ultimate goal of providing, in the near future, a better clinical and preventive approach (50).

We are indebted to all subjects who consented to participate to this study. We are especially thankful to Diego Minguzzi and Patrizia Pattini from the Department of Clinical and Experimental Medicine at the University of Verona School of Medicine, Italy, for their excellent technical assistance and valuable contribution in data collection.