MTHFR genetic polymorphisms and susceptibility to childhood acute lymphoblastic leukemia

In the letter by Kumar and colleagues in this issue of Blood, the authors failed to reveal any association between MTHFR genetic variants (677C>T and 1298A>C) and the risk of childhood acute lymphoblastic leukemia (ALL) in an association study targeting 460 patients recruited between 1983 to 2003 and 1472 controls of German origin. As mentioned by the authors, factors related to both genetics and nutrition might explain the discrepancies between Kumar et al's data and those reported by other groups.^1-4 

MTHFR is one of the best examples of a polymorphic gene investigated in the context of cancer susceptibility, particularly leukemia (see Robien and Ulrich⁵  for further discussion). This might be explained by the availability of several groups of patients worldwide that have been used to validate genetic epidemiology data.

In a French-Canadian cohort (n = 271 patients), we observed an underrepresentation of CC1298 homozygotes among the ALL patients compared with controls (4.4% vs 10.3%, respectively), suggesting a protective effect of this variant (odds ratio [OR] = 0.4;95% CI, 0.2-0.8; P < .01).⁶  The analysis of both genotypes showed that CC677/AA1298 individuals were at a higher risk of ALL compared with those with other genotypes (OR = 1.8; 95% CI, 1.1-2.8; P = .02). This led to the suggestion that carriers of alleles T677 and C1298 have a decreased susceptibility to ALL, particularly the homozygotes for either one or another of these variants: TT677/AA1298 (OR = 0.4; 95% CI, 0.2-0.9; P = .02) or CC677/CC1298 (OR = 0.3; 95% CI, 0.1-0.6; P < .001). Similar results were obtained with a parental trio design, thus decreasing putative population stratification bias. These data confirm and extend previous findings by Skibola et al,³  who reported reduced frequency of the same MTHFR haplotypes in adult leukemia (n = 71 patients from United Kingdom), as well as results of Wiemels et al,²  who found the protective effect of these variants in infant leukemia with MLL rearrangements and hyperdiploid pediatric leukemia (n = 253 patients from United Kingdom). Furthermore, Franco et al⁴  reported in a group of 71 Brazilian patients a protective effect of the T677 variant. Taken together, these 4 studies including 666 ALL patients from 4 distinct cohorts observed similar protective effects of at least one of the tested MTHFR variants, suggesting that population-to-population differences are a less plausible explanation; although, an apparent European north-to-south range of frequencies has been reported.⁷ 

Considering the reported protective association between folate supplements in pregnancy and the risk of common childhood ALL,⁸  maternal diet might represent a suitable determinant of the disease, particularly for the effects of MTHFR variants that seem to be modified by folate levels. Indeed, we observed a protective effect in TT677/AA1298, CC677/CC1298, and CC677/AC1298 only in children born before Health Canada's recommendation to give folate supplements during pregnancy, thus presumably reflecting the maternal folate insufficiency during pregnancy.⁶  However, Kumar et al did not observe such association. In this regard, it is worth mentioning that a study addressing the prenatal vitamin supplementation issue reported substantial differences in the number of mothers taking vitamins during pregnancy, ranging from 3% in France to 90% in US centers.⁹  We propose that such gene-nutrient interaction might explain, at least in part, the differences observed between genetic association studies.