Associations of arsenic exposure and folate in maternal leukocyte DNA methylation: a case-control study of mothers with spina-bifida affected children

Abstract Background In Bangladesh, more than a quarter of drinking water tubewells are contaminated with arsenic above the national standard (50 µg/l), while nearly half exceed the World Health Organization guideline (10 µg/l). Among other negative health consequences, arsenic is a suspected environmental risk factor for neural tube defects (NTDs), including spina bifida. Maternal folate status protects against NTDs, though recent evidence suggests arsenic attenuates folate’s protective effects. Arsenic is methylated prior to excretion with methyl groups produced in one-carbon metabolism, for which folate is a cofactor. We thus hypothesized that DNA methylation (DNAme) may provide insight into the interactions between arsenic, maternal folate levels, and offspring spina bifida. Methods Here we analyzed leukocyte DNAme using the Illumina MethylationEPIC v2.0 array in 374 women from Bangladesh, 246 with a previous spina-bifida affected birth and 128 controls. Chronic arsenic exposure was evaluated in maternal toenail; fasting plasma folate was measured at blood draw. Linear models evaluated DNAme associated with offspring spina bifida, arsenic, folate, and their interaction terms. Results Maternal DNAme was associated with spina bifida at 71 CpGs, arsenic at 6 CpGs, and folate at 33 CpGs (all FDR < 0.05). The spina bifida*arsenic and arsenic*folate interactions were associated with 11 and 28 CpGs, respectively, while spina bifida*folate returned no significant associations. We observed lower DNAme in mothers of spina bifida cases at significant loci, including in developmental genes such as HOXB3 and HOXB4. Arsenic’s influence on DNAme was more pronounced in individuals with low folate. Conclusions Our work suggests that postnatal maternal leukocyte DNAme is associated with offspring spina bifida status and is modified by arsenic exposure but not plasma folate.