When analyzing the genetic selleck impact on U-Cd by B-Cd tertiles instead, rs11076161 was significantly associated with U-Cd in the highest tertile: for increasing U-Cd with increasing number of variant alleles the p-value for trend was 0.01 in the unadjusted model and p = 0.001 for the model adjusted for age, sex and smoking. There was no interaction (p > 0.05) between exposure and genotype for B-Cd and U-Cd: the mean differences between the

genotypes were similar in each exposure group (Fig. 1, Supplementary Figs. 1–2). First, genetic effect modification on Cd-related excretion of low molecular weight proteins was evaluated in different exposure groups. For rs11076161, the genotype was significantly (p-value = 0.045, adjusted for age, sex and smoking) associated with UNAG in the highly polluted group: the variant homozygotes demonstrated the highest levels of see more UNAG (Fig. 2A). The same pattern was seen for UB2M (p = 0.052; Fig. 2B). The same pattern, but non-significant, was seen for rs10636 (Supplementary

Figs. 3a and b; UB2M p = 0.28 in the high exposure group; UNAG p = 0.13), but no effect of rs28366003 was found. In the alternative analyses by B-Cd tertiles, the genetic influence of rs11076161 became more obvious in the highest tertile on UNAG (p-value = 0.01) and UB2M (p-value = 0.002; both p-values for trends in models adjusted for age, sex and smoking). In the models stratifying for B-Cd tertiles instead of exposure groups, associations of the other two SNPs with UNAG and UB2M disappeared. Secondly, genetic effect modification on Cd-related levels of UNAG and UB2M was evaluated by using Cd in blood or in urine as exposure markers. The four exposure–response marker combinations that had significant or nearly significant interaction p-values in Table 3 were selected for calculation of genotype-specific association coefficients

which are presented in Table 4. Coefficients of non-significant associations are presented in Supplementary Table S1. There was a significant interaction of MT1A rs11076161 with B-Cd (adjusted p = 0.001), as well as weakly with U-Cd aminophylline (adjusted p = 0.062, unadjusted p = 0.053) for concentrations of UB2M ( Table 3). Carriers of the variant genotype AA demonstrated a steeper slope for the association between B-Cd/U-Cd and UB2M compared to carriers of genotype GG ( Table 4). A significant interaction with rs11076161 and B-Cd, but not with U-Cd was found for UNAG concentrations. Carriers of the variant genotype AA were associated with a steeper slope for the association between B-Cd and UNAG compared to carriers for genotypes AG or GG ( Table 4; Fig. 3). Rs28366003 modified the association between U-Cd and UNAG, where individuals carrying the variant genotype demonstrated a shallower slope compared to the common genotype. Although there were only 10 carriers of the GG genotype, we analyzed whether they had even higher UNAG levels in relation to U-Cd levels compared to the heterozygotes.