Several single nucleotide polymorphisms SNPs in
Several single nucleotide polymorphisms (SNPs) in humans affect CYP3A expression and activity (Lamba et al., 2002). The CYP3A4 SNP rs2740574 (also referred to as CYP3A4*1B, by The Human Cytochrome P450 (CYP) Allele Nomenclature Database, http://www.cypalleles.ki.se/) is a change in the promoter region that can potentially give higher expression levels of CYP3A4 (Lamba et al., 2002). CYP3A5 rs776746 SNP (CYP3A5*3A) encodes a truncated mRNA transcript, resulting in no protein expression (Knops et al., 2015). CYP3A7 rs2257401 (CYP3A7*2) is a non-synonymous coding SNP that yields an enzyme with increased catalytic efficiency (Rodriguez-Antona et al., 2005). In this study we examined whether these polymorphic variants of CYP3A genes, which predict different levels of CYP activity, modify the association between prenatal exposure to MeHg and neurodevelopmental outcomes in children. We studied three birth cohorts from coastal populations; the Seychelles Child Development Study (SCDS), INMA – Environment and Childhood in Spain, and PHIME (Public Health Impact of long-term, low-level Mixed Element Exposure in susceptible population strata) in Italy and Greece. A lack of adverse neurodevelopmental effects during infancy associated with prenatal MeHg in these birth cohorts has been previously reported (Davidson et al., 2008, Llop et al., 2012, Strain et al., 2015, Valent et al., 2013).
Results Maternal and child characteristics are presented in Table 1. Mothers in NC1 and NC2 (Seychelles) were slightly younger than mothers in INMA-Spain and PHIME-Italy and Greece. The percentage of mothers with university degree was the lowest in PHIME-Greece, followed by INMA-Spain, PHIME-Italy, NC1 and NC2-Seychelles. Nearly half of the INMA-Spain mothers belonged to the lowest social class, followed by women in PHIME-Italy, PHIME-Greece, NC1 and NC2-Seychelles. The proportion of smokers was higher in the Mediterranean cohorts than in Seychelles. THg concentrations in maternal hair and cord blood were the highest in Seychelles (5.8 and 3.9μg/g in NC1 & NC2 maternal hair respectively and 39.3μg/L in NC1 cord blood). Among the Mediterranean cohorts, the highest cord blood THg concentrations were found in INMA-Spain (11.3μg/L), followed by PHIME-Greece (7.5μg/L) and finally PHIME-Italy (5.6μg/L). The maternal fish consumption was higher in Seychelles (9.1 meals/week in NC1 and 8.5 meals/week in NC2), followed by INMA-Spain (6.2 meals/week), PHIME-Greece (3.2 meals/week) and Italy (2.5 meals/week). In all cohorts, the minor Cilomilast of each CYP3A SNP correspond to the higher-activity allele with respect to predicted CYP enzyme activity (i.e., G in rs2257401 (CYP3A7), A in rs776746 (CYP3A5), and G in rs2740574 (CYP3A4)), with the exception of SCDS where higher-activity alleles A in rs776746 and G in rs2740574 were the major alleles (Table 2). Across all Mediterranean cohorts, the MAFs for all three CYP3A genes were relatively low (range from 1.4% to 12.4%), with minor cohort-differences in MAF for each SNP, and similar to reported MAFs for European populations (Supplementary Table S2). In contrast, in Seychelles (NC1 and NC2) the MAFs were much higher than the Mediterranean cohorts across all three CYP3A genes (range 44.8–55.1%, Table 2). These frequencies are similar to those previously reported for African populations and are somewhat higher than frequencies observed in Asian populations (Supplementary Table S2). There was evidence of linkage disequilibrium (LD) between some of the SNPs: the LD (r2) between rs776746 and rs2257401 were 0.39 in PHIME-Greece, 0.56 in PHIME-Italy, 0.60 in SCDS and 0.73 in INMA. The LD between all other SNP pairs was below 0.35. With respect to THg concentrations in cord blood, no significant differences were seen as a function of CYP3A genotype (Table 2). Effect modification of CYP3A genotype on the association between prenatal exposure to MeHg and child neurodevelopment was evaluated. For CYP3A7, the overall coefficient of cord blood THg on MDI was positive (i.e. improving MDI score with increasing cord blood THg concentrations) and statistically significant (β=2.90; 95% CI 1.53, 4.27) for children carrying GG or GC (Fig. 1A). For children carrying CC, this coefficient became near null (β=0.20; 95% CI −0.49, 0.89) (Fig. 1B). The interaction between CYP3A7 genotype and cord blood THg was statistically significant (p<0.05) for the Mediterranean cohorts. No effect modification of CYP3A7 was evident for PDI (Fig. 1C–D). For CYP3A5, the overall coefficient for the association between cord blood THg and MDI was positive for children carrying AA or AG (β=2.51; 95% CI 1.04, 3.98) (Fig. 2A) and near the null for children carrying GG (β=0.40; 95% CI −2.27, 1.06) (Fig. 2B). The interaction between cord blood THg and CYP3A5 genotype was statistically significant (p<0.05) for INMA-Spain and PHIME-Greece. For PDI, the overall coefficient was positive but not statistically significant and the interaction was significant only for PHIME (Fig. 2C–D). Results from the analysis of CYP3A4 rs2740574 were similar to the CYP3A7 and CYP3A5 SNPs, but the interaction p-value in the MDI models was <0.05 for INMA only (Fig. 3). Coefficients were virtually the same when we included maternal fish intake in the minimally adjusted models (Supplemental material Tables 2-4). Similar results were observed when we used the common variables adjusted models (Supplemental material Tables 5-7).