Fig 1.
FTCD SNP rs61735836 is associated with the all three arsenic species measured in urine (iAs%, MMA%, and DMA%).
Quantile-quantile plots (A-C) for all 19,992 post-QC exome chip variants and scatterplots (D-F) depicting the association between rs61735836 (301G>A, Val101Met) and arsenic species percentages (iAs% on left, MMA% center, and DMA% right) among 1,660 HEALS participants.
Fig 2.
Regional association plots for the FTCD region (21q22.3).
The vertical axes show the–log10(P-value) for the association of SNP allele counts with (A) DMA%, (B) MMA%, (C) iAs% (A-C based on 1,660 HEALs participants), and (D) arsenic induced skin lesions (2,401 cases, 2,472 controls).
Table 1.
Odds ratios (ORs) for the association between rs61735836 carrier status and arsenic-induced skin lesions, including exposure-stratified ORs.
Fig 3.
The minor allele of missense variant rs61735836 changes a valine to a methionine.
The minor allele (A, MAF = 0.07) changes a valine (V, circled) codon (GTG, red box) to a methionine (M) codon (ATG/AUG) at a site that is highly conserved across vertebrates. This change introduces a potential start codon in exon 3 which is the first 5´-proximal Kozak consensus sequence ([A/G]xxAUGG) in the FTCD gene.
Fig 4.
The role of FTCD in histidine catabolism and the one-carbon/folate cycle, which provides methyl groups for arsenic methylation (by AS3MT) via the methionine cycle.
The formininotranserase domain of FTCD catalyzes the transfer of a formimino group from N-formimino-L-glutamate (FIGLU) (or a formyl group from N-formyl-L-glutamate) to tetrahydrofolate (THF) producing formimino-THF. The cyclodeaminase domain of FTCD then catalyzes the removal of ammonia from formimino-THF, generating 5,10-methenyl THF, which can then be converted to 5:10 methylene-THF or THF, both key components of the canonical one-carbon/folate cycle (shown in bold). The folate cycle contributes one-carbon groups to the methionine cycle, which in turn supplies these groups to methyltransferases (such as AS3MT) involved in methylation of arsenic, DNA, and other substrates. DHF, dihydrofolate; dUTP, deoxyuridine triphosphate.
Fig 5.
Mendelian randomization supports a causal effect of arsenic metabolism efficiency on arsenic-induced skin lesion risk.
Horizontal and vertical error bars for each SNP correspond to the 95% CI for the beta coefficient for its association with DMA% and skin lesion risk, respectively. The slope of the diagonal line (-0.013) is the inverse-variance-weighted estimate of the causal effect (i.e., the ln(OR), corresponding to OR = 0.89 for a 10% increase in DMA%; P = 6x10-8).