Individual Variations in Inorganic Arsenic Metabolism Associated with AS3MT Genetic Polymorphisms

Individual variations in inorganic arsenic metabolism may influence the toxic effects. Arsenic (+3 oxidation state) methyltransferase (AS3MT) that can catalyze the transfer of a methyl group from S-adenosyl-l-methionine (AdoMet) to trivalent arsenical, may play a role in arsenic metabolism in humans. Since the genetic polymorphisms of AS3MT gene may be associated with the susceptibility to inorganic arsenic toxicity, relationships of several single nucleotide polymorphisms (SNPs) in AS3MT with inorganic arsenic metabolism have been investigated. Here, we summarize our recent findings and other previous studies on the inorganic arsenic metabolism and AS3MT genetic polymorphisms in humans. Results of genotype dependent differences in arsenic metabolism for most of SNPs in AS3MT were Inconsistent throughout the studies. Nevertheless, two SNPs, AS3MT 12390 (rs3740393) and 14458 (rs11191439) were consistently related to arsenic methylation regardless of the populations examined for the analysis. Thus, these SNPs may be useful indicators to predict the arsenic metabolism via methylation pathways.


Introduction
Groundwater pollution by inorganic arsenic (IA) derived from natural origin is a serious issue of public health in the world, especially in developing countries where clean and safe surface water is not available in many places. In these areas, high arsenic concentrations exceeding the WHO guideline (10 μg/L) for drinking [1] have been reported in the groundwater [2][3][4]. It is well known that IA is one of the carcinogenic chemicals for the human. Chronic arsenic exposure causes skin pigmentation, hyperkeratosis, and cancers in the skin, lung, bladder, liver, and kidney resulting in high mortalities [1,5,6].
In general, humans can metabolize ingested IA to monomethylated arsenic (MMA) and then dimethylated arsenic (DMA), through arsenic (+3 oxidation state) methyltransferase (AS3MT). The pathways are discussed in detail in a later section of this review. Urinary arsenical profile has been used to assess arsenic exposure and its metabolism in individuals. In general, concentration ratios of MMA/IA and DMA/MMA in the urine indicate first and second methylation capacity, respectively. The metabolic action of AS3MT is considered to have an association with accumulation profile of various arsenic compounds in the body and consequently with susceptibility to toxic effects of arsenic. Several previous epidemiological studies showed that people who had high percentage of DMA (%DMA) in the urine could excrete more arsenic from the body because of high arsenic methylation capacity [7,8]. In other epidemiological studies, high concentration or percentage of MMA in the urine of populations with arsenic-related diseases were reported when compared with that of healthy people [9][10][11][12].
Vahter [13] showed that relative distributions of IA, MMA, and DMA in the urine of various populations are generally 10-30%, 10-20%, and 60-70%, respectively. On the other hand, there are large variations in the arsenic metabolism at individual and population levels [7]. It is known that biological and environmental factors including age, sex, pregnancy, arsenic exposure level, smoking habits, nutritional status and diet, potentially relate to the inter-individual variation (reviewed by Tseng [14]). With regard to ethnic related variation, it was shown that Andes women in arsenic contaminated area had only a low percentage of MMA[V] (0-11%) in the urine [15], whereas residents ingesting arsenic through drinking water in the northeast Taiwan had an extraordinarily high %MMA[V] (26.9%) in the urine [16]. These results indicate that genetic polymorphisms of certain enzymes that can metabolize arsenic through methylation pathways may be one of factors associated with the variation of arsenic profiles.
Genetic analyses of SNPs in the enzymes that may be involved in arsenic metabolism have been recently developed to evaluate the variation in the metabolism or toxic impacts of arsenic at individual or population level. We have also reported the association of arsenic metabolism-SNPs in AS3MT in Vietnamese [17][18][19][20][21]. However, a limited number of publications regarding the association of SNPs in AS3MT with arsenic methlyation are available. Moreover, the published results are based on fragmentary investigations and thus consistent conclusions have not yet been drawn. Hernandez and Marcos [22] have recently reviewed the role of genetic polymorphisms in enzymes related to arsenic metabolism including AS3MT. Thereafter, the number of publications regarding genetic polymorphisms dependent metabolism and toxicity of arsenic is increasing.

Association of Arsenic Methylation with Genotypes in Human AS3MT
Notation of SNPs in human AS3MT is individually defined among studies. In this review, we use the notation by SNP ID which indicates the polymorphism identification number relative to the location in the consensus sequence (AY817668) with the first base of the consensus number 1 and dbSNP rs# cluster id which is available on Search for SNP on NCBI Reference Assembly [33]. Chromosome positions of genetic polymorphisms in AS3MT are indicated in Figure 2.

In Vitro Studies
Drobna et al. [34] reported the association of genotypes in AS3MT with arsenic methylation in human hepatocytes. They treated primary hepatocytes from eight human donors with graded concentrations of As[III] (0.330 nmol/mg protein) for 24 h to evaluate the variation in arsenic metabolism. One of eight donor's cells showed a higher methylation rate than other cells at a high exposure level (30 nmol/mg protein). The methylation rate was consistent with heterozygote of Met287Thr at amino acid base (14458 at nucleotide base, rs11191439) mutation of AS3MT, suggesting a linkage of this genotype with arsenic methylation.

Reduction
Reductive methylation  In vitro functional analysis of human AS3MT was performed by Wood et al. [30]. The authors initially sequenced DNA samples from African-American (n = 60) and Caucasian-American  (n = 60), and identified 26 SNPs and one VNTR. Among the SNPs, three SNPs, Arg173Trp, Met287Thr, and Thr306Ile in the exon region were non-synonymous. In the AS3MT-expressed COS-1 cells, which was treated with 12.5 nM As[III], the enzyme activity and immunoreactive protein expression of 287Thr variant of AS3MT were significantly higher than those of the wild type, whereas 173Trp and 306Ile variants showed opposite trends. Reporter gene assays using reporter constructs containing different length 5'-UTR VNTR sequences of AS3MT gene demonstrated that the transcriptional expression of this gene depends on the number of the repeat sequence. Given these results, the authors suggested genetic polymorphisms in AS3MT may contribute to individual differences in the expression and function of AS3MT, and, consequently, to variation in the risk of arsenic-dependent carcinogenesis.
Following the above initial study, the same group [36] found a clear association of percentage of (%) MMA[V] with genetic polymorphism in AS3MT 35587 (rs11191453) only for children. However, given that the participant selection may be biased; children of the above study belong to more indigenous American ancestry than adults [33,34], these age-dependent associations of genetic polymorphisms in AS3MT with arsenic metabolism need reconsideration.
According to Valenzuela et al. [37], AS3MT polymorphisms may influence skin lesions caused by arsenic exposure through drinking water. This group investigated three SNPs, AS3MT 4602 (rs7085104), 14458 (rs11191439, Met287Thr), and 35587 (rs11191453) in 71 individuals with skin lesions and 51 controls without the lesions from Zimapan, Hidalgo in Mexico. In this study area, medians of arsenic concentrations in wells and human urine samples were 84 μg/L and 109 μg/g creatinine, respectively. No significant association of AS3MT 35587 (rs11191439) was found with arsenic metabolism and skin disease. Among the subjects, AS3MT 4602 (rs7085104) GG  (rs11191439) CT + CC carriers were significantly higher than those in AS3MT 14458 (rs11191439) TT carriers. This suggests that AS3MT 14458 (rs11191439) CT + CC carriers may have a poor methylation capacity. Interestingly, the frequency of this SNP carriers in the skin lesion group was marginally (p = 0.055) higher than the control subjects, suggesting that AS3MT 14458 (rs11191439) CT + CC may be involved in skin lesions. Although the way in which genetic polymorphism of AS3MT 14458 (rs11191439) affects skin lesions is not clear, it can be presumed that the skin disease may be related to the metabolic pathway of IA.
Recently, from a large sample size (n = 405), Gomez-Rubio et al. [38]  was higher in CC genotype of AS3MT 30312 (rs10883795), followed by TC and TT genotypes. On the contrary, exonic AS3MT 14458 (rs11191439) showed no significant association with arsenic methylation, probably due to the small number of variants for this SNP. As described earlier, a previous study by this group revealed a children specific association of AS3MT genotypes with arsenic metabolism [35,36], but the association was also observed in adults when analyses were performed using a large number of subjects [38].
Sampayo-Reyes et al. [39] evaluated DNA damages by a comet assay in residents who were exposed to IA from drinking water (1-187 μg/L) in Torreon, Coahuila, Mexico. The results revealed a significant positive correlation between DNA damages and urinary arsenic concentrations, suggesting that this population is under threat by exposure to arsenic. Furthermore, AS3MT 14458 (rs11191439, Met287Thr) influenced DNA damages in children; DNA damages of subjects having AS3MT 14458 (rs11191439, Met287Thr) C allele were higher than those having T allele (p = 0.034). Hence, the authors suggested that the arsenic exposure through drinking water induces DNA damage and this effect may become more profound by AS3MT 14458 (rs11191439, Met287Thr).

Argentina
Significant relationships between genetic polymorphisms in AS3MT and arsenic metabolism in Argentina were reported by Schläwicke Engström et al. [40]. This group analyzed arsenic compounds in the urine, whole blood, and buccal cells and polymorphisms in AS3MT of 147 indigenous women from a village in the northern Argentinean Andes. The residents have been exposed to high levels (about 200 μg/L) of arsenic through drinking water. For AS3MT, nine polymorphisms (AS3MT 5019 (rs17880342), 5194 (rs3740400), 8788 (rs17881367), 12390 (rs3740393), 12590 (rs3740392), 14215 (rs3740390), 14458 (rs11191439), and 35991 (rs10748835)) were identified. Among these SNPs, intronic AS3MT 12390 (rs3740393), 14215 (rs3740390), and 35991 (rs10748835), which had a strong LD with each other (r 2 = 0.74-0.95), were significantly associated with the secondary methylation step of arsenic. For example, AA type of AS3MT 35991 (rs10748835) had a lower %MMA[V] and higher %DMA[V] and DMA[V]/MMA[V] ratios in the urine than the GG and GA types. They also found that the genotype frequencies in three SNPs in AS3MT, which had correlations with the secondary methylation of arsenic in a population of Argentina, were higher than those in other populations. This may be due to a positive selection of AS3MT genetic polymorphisms that can efficiently metabolize arsenic for detoxification, because these populations have suffered from arsenic toxicity for thousands of years [41,42].
Recently, Schläwicke Engström et al. [43] genotyped another two SNPs, AS3MT 4602 (rs7085104) in 5' terminal and AS3MT 5194 (rs3740400) in intron 1, and evaluated the influences on arsenic metabolism in the same subjects as in the previous cases. These SNPs comprised a strong LD (r 2 = 0.97) with each other and also with AS3MT 12390 (rs3740393), 14215 (rs3740390), and 35991 (rs10748835) which were reported in the previous study [40]. This LD cluster was found only in this Argentinean population, but not in other HapMap populations. The SNPs, AS3MT 4602 (rs7085104) and 5194 (rs3740400) were clearly associated with %MMA[V] and %DMA[V].

Chile
Hernández et al. [44] investigated 50 workers who have been occupationally exposed to high levels of arsenic in a copper smelting plant in Chuquicamata, Chile. Mean values of exposure periods and total arsenic concentration in the urine of the workers were 15.55 years and 152.8 ppb, respectively. In this study, nine polymorphisms were found from the sequencing of exons and flanking regions of AS3MT gene. Three polymorphisms, AS3MT 4965 (rs17881215), exon 1: 5'-UTR (inside VNTR), and 14458 (rs11191439) in AS3MT, which comprised one LD cluster, were significantly associated with the variation in %MMA [V].
To confirm the genetic polymorphism-dependent metabolism of arsenic, Hernández et al. [45] re-investigated the relationships of the arsenic profile and AS3MT SNPs in larger population (n = 207), in which the urine had 91.26 ppb of the arsenic level on average. The results revealed that exonic AS3MT 14458 (rs11191439) TC + CC showed higher %MMA[V] in the urine than the wild type as observed in the previous study [44]. In summary, the authors concluded that AS3MT 14458 (rs11191439) TC + CC genotypes increase the methylation capacity from IA to MMA.

Central Europe
Lindberg et al. [46] investigated the association between the genetic polymorphism in AS3MT and arsenic metabolism in populations from Hungary, Romania and Slovakia (n = 415) that have been exposed to low arsenic levels. This investigation is distinct from others, and is meaningful because variation of the arsenic metabolism depending on the exposure could be negligible. Even though the median urinary arsenic concentration of the individuals was low (8.0 μg/L), percentages of arsenic compounds in the urine varied widely. Through the multivariate analysis, they found that AS3MT 14458 (rs11191439) in exon was one of the major factors that can influence arsenic metabolism; the TC + CC type had higher %MMA[V] and lower %DMA[V] than the TT type. This trend was much more pronounced in men than in women.

India
West Bengal in India is one of the most severe arsenic endemic areas in the world. It has been estimated that about 6,000,000 people are exposed to arsenic through the consumption of groundwater [3]. However, only 300,000 individuals (5% of total exposed population) within this population had arsenic-related disease [47]. De Chaudhuri et al. [48] hypothesized that low occurrence of the disease may be related to the genetic variability. Hence, the association of exonic genetic polymorphisms in AS3MT with the occurrence of arsenic-induced skin lesions was investigated. They first analyzed 13 genotypes in AS3MT in 25 cases with skin lesions and 25 controls from North 24 Parganas, Nadia, and Murshidabad Districts in West Bengal, and identified eight common genotypes. Mean concentrations of arsenic in drinking water and human urine were 163.16 μg/L and 301.17 μg/L, respectively, in the cases with skin lesions, and 161.71 μg/L and 283.28 μg/L, respectively, in controls. For these samples, there were no significant differences in the arsenic concentrations between case and control groups. On the other hand, arsenic concentrations in nail and hair of case subjects (5.39 μg/g dry wt and 3.13 μg/g dry wt, respectively) were significantly higher than those of control group (2.39 μg/g dry wt and 1.61 μg/g dry wt, respectively). The authors did not carry out urinary arsenic speciation and thus they did not mention arsenic methylation. Among the identified eight genotypes, AS3MT 14458 (rs11191439) was genotyped in 229 cases and 199 controls, and the distribution of genotype frequencies was compared between cases and controls. However, no significant difference was detected between the two groups, and thus the exonic genetic polymorphisms in AS3MT may not be linked to the occurrence of arsenic-induced skin lesions.

Vietnam
Arsenic pollution in groundwater in Vietnam was reported by Berg et al [50]. The authors detected more than 3,000 μg/L of As concentration in groundwater from around Hanoi. Following this initial attempt, many studies have reported the concentration and distribution patterns of As in groundwater and sediment samples in northern and southern Vietnam [51][52][53][54][55][56].
While intensive studies have focused on the levels of arsenic in environmental samples, there is little information on arsenic exposure to, and toxic impacts on, local residents [18,53,[57][58][59][60][61]. Furthermore, frequency distribution of AS3MT polymorphisms and the association with arsenic methylation ability have not yet been clarified. Hence, our group has investigated arsenic exposure, methylation capacity, and SNPs in AS3MT in Vietnamese [17][18][19][20][21].
To clarify the relationship between arsenic methylation and SNPs in AS3MT, we collected groundwater as well as human hair, urine, and blood samples from Hoa Hau and Liem Thuan in Ha Nam Province, which are located in the Red River Delta. For the human samples, each arsenic compound was quantified and 17 SNPs in AS3MT were genotyped [17,18,20]. The arsenic level in filtered water that some families consumed in Hoa Hau drastically (around 93%) decreased when compared with that in raw groundwater. Geometric means (GMs) of arsenic concentration in drinking water (50.1 μg/L) and human hair (0.351 μg/g dry wt) from Hoa Hau were significantly higher than those from Liem Thuan (1.7 μg/L in the water and 0.232 μg/g dry wt in the hair). On the other hand, there was no significant difference in urinary arsenic level and composition of arsenicals between Hoa Hau (GM, 92.6 μg/g creatinine) and Liem Thuan (GM, 97.9 μg/g creatinine), suggesting that the residents in Hoa Hau have not recently been exposed to arsenic through drinking water and/or there are arsenic sources other than drinking water in Liem Thuan. Among the 17 SNPs in AS3MT investigated, 14 unexonic SNPs were categorized into four LD clusters; AS3MT 3963 (rs7098825), 6144 (rs17878846), 12390 (rs3740393), 14215 (rs3740390), 35587 (rs11191453), and 37950 (rs17879819) as cluster 1, AS3MT 4602 (rs7085104), 35991 (rs10748835), and 37853 (rs11191459) as cluster 2, AS3MT 4740 (rs12416687) and 12590 (rs3740392) as cluster 3, and AS3MT 5913 (rs4917986), 9749 (rs17881367), and 27215 (rs11191446) as cluster 4. We also found that sex, age, and body mass index (BMI) are correlated with urinary arsenic profile in this Vietnamese population. Thus, to understand the SNP-dependent association with arsenic metabolism without co-effects, we further conducted an analysis that was adjusted for sex, age, and BMI and detected significant associations for 12 SNPs. In cluster 1, homozygotes in AS3MT 12390 (rs3740393) GG and 35587 (rs11191453)

Allele Frequency of SNPs in AS3MT
As we have mentioned above, several SNPs in AS3MT may influence arsenic metabolism. However, only limited information on frequencies of genetic polymorphisms in AS3MT was available for the number of the SNPs and the association with ethnicity. We have reported distribution of SNPs in AS3MT in some Asian NP has no C allele of AS3MT 4740 (rs12416687). Relatively high frequencies of AS3MT 7395 (rs12767543) A allele and 8979 (rs72920657) A allele were found in CH and VN, respectively. AS3MT 6144 (rs17878846) A allele, AS3MT 37616 (rs4568943) A allele, and AS3MT 37950 (rs17879819) C allele frequencies in OVA and GHA populations were higher than Asians, but comparable with those in MN. Frequencies of AS3MT 10209 (rs3740394) A allele, AS3MT 26790 (rs11191445) C allele, and AS3MT 35991 (rs10748835) A allele in Asians (except for AS3MT 35991 (rs10748835) A allele in OVA) were higher than those in other ethnic groups, while the opposite trend was observed for AS3MT (rs11191439) T allele. For AS3MT 14215 (rs3740390), the allele distribution was different even within Chinese (CH vs CHB(H) and CHD(D)) and African groups (OVA and GH vs LWK(L), MKK(K), YRI(Y), and ASW(A)). TIB also showed a different pattern of AS3MT 14215 (rs3740390) allele from other Asian populations. Relatively lower C allele frequency of AS3MT 25986 (rs7085854) was detected in JP, JPT, SKR, and MN. TIB showed different allele distribution of AS3MT 35587 (rs11191446). To understand these variations, further research is necessary.

Conclusions and Future Perspectives
In this review, we summarized recent results on association of AS3MT genetic polymorphism with arsenic methylation. Although there are some inconsistencies between the genotypes and metabolism in human case studies, we found consistent results on two SNPs, AS3MT 12390 (rs3740393) in intron and 14458 (rs11191439, Met287Thr) in exon, in all the nations (Figure 4 and Table 2). This indicates that these SNPs may be ethnically independent polymorphisms, but may affect arsenic methylation.
AS3MT 12390 (rs3740393) is an intronic polymorphism, but the genotype with C allele in this SNP may have a higher second methylation capacity as predicted from the consistent results of high DMA[V]/MMA [V] in Mexicans, Argentines, Taiwanese, and Vietnamese ( Table 2). Although AS3MT 12390 (rs3740393) belongs to a LD cluster in these populations, other SNPs within the cluster are different among populations (Table 2). Therefore, AS3MT 12390 (rs3740393) may be involved in an independent reaction from other SNPs in the same LD on a worldwide population basis. The frequency of AS3MT 12390 (rs3740393) CC or CG types was much higher than the GG type in Argentina's population; the allele frequencies of C and G were 72% and 28%, respectively [40]. This allele frequency distribution was drastically different from other populations (Table 1 and Figure 3). This unique distribution may contribute to the results indicating that Argentina's population has a higher %DMA and a lower %MMA in the urine, when compared with yet another study [15]. It is noteworthy to study whether this specific genotype distribution of AS3MT 12390 (rs3740393) is observed only in this group (Argentinean Andes) and how this unique SNP selection took place.      NS   5913  rs4917986  Intron 3  C/T  T  6144  rs17878846  Intron 3  A/T  A  7395  rs12767543  Intron 3  A/G  G  8788  rs17881367  Intron 5  -/TTT  NAV  8979  rs7920657  Intron 5  A/T  T  9749  rs17881367  Intron 5  A/G  A  10209  rs3740394  Intron 6  C/T  T  12390  rs3740393  Intron 6  C/G  C  12590  rs3740392  Intron 7  C/T  T  13599  rs10883797  Intron 7  C/G  C  14215 rs3740390 Intron 8 C/T C    AS3MT 14458 (rs11191439), locates in the exon region, is synonymous substitution (Met to Thr substitution) at the amino acid base 287, and its functional difference has been investigated in some studies. By reviewing the in vitro [30,34] and human case studies [17][18][19][20][21][35][36][37][38][39][40][43][44][45][46][47][48][49], we found that C allele containing genotype in this SNP consistently showed a higher first methylation capacity than the T allele carrier. Interestingly, the effect of this SNP has been observed even in subjects exposed to low levels of arsenic [46]. Hence, it seems that the association of AS3MT 14458 polymorphism with arsenic methylation may occur irrespective of arsenic exposure status. This allele frequency was relatively low in Asian populations (Table 1 and Figure 3), indicating that among different global populations, Asians may be genetically less sensitive to arsenic methylation.
Increased %MMA in the urine linked to cancer risk has been suggested in previous epidemiological studies [9][10][11][12]. Thus, AS3MT 14458 (rs11191439) C allele carriers, who have a higher methylation capacity from IA to MMA, may be at higher risk of arsenic related cancers. The association of this SNP with human health effect was investigated in Mexicans [37] and Indians [48]. According to the study by De Chaudhuri et al. [48], there was no significant interaction between this SNP and skin lesions. On the other hand, Valenzuela et al. [37] found higher frequency of C allele type of AS3MT 14458 (rs11191439) TC + CC in people with cancers, although the level of significance was not high (p = 0.055). Hence, at present, no rigid evidence has been provided to link the AS3MT 14458 (rs11191439) dependent variation in arsenic methylation to arsenic-induced carcinogenesis. Apart from an evidence of cancer, Sampayo-Reyes et al. [39] found that AS3MT 14458 (rs11191439) is linked with DNA damages among children living in arsenic contaminated sites in Mexico (p = 0.034). Although these authors did not analyze arsenic compounds in urine and thus did not take into consideration the methylation capacity, they suggested that increased production of MMA[III] during methylation processes in AS3MT 14458 (rs11191439, Met287Thr) C allele carrier might induce genotoxicity.
Up to now, it still remains unknown how AS3MT 12390 (rs3740393) and 14458 (rs11191439) can influence the structure and function of this enzyme. Also, several results of AS3MT genotype-dependent arsenic methylation are interesting, because some non-exonic (intron, UTR, or 5' and 3' gene) SNPs may play important roles during methylation. The study on splicing variants of AS3MT gene and their function is one such issue to be pursued.
There are several tasks for future study regarding the role of AS3MT polymorphisms on arsenic methylation. As shown in this review, most studies encompassed only a small number of subjects and SNPs in AS3MT and thus, they merely provide fragmental and weak significant results at particular level(s). As shown in Table 2, results on the genotype association of AS3MT 4602 (rs7085104), 5194 (rs3740400), 7395 (rs12767543), 12590 (rs3740393), 14215 (rs3740390), 35587 (rs11191453), and 35991 (rs10748835) with arsenic metabolism were not consistent among different country groups. Larger scale studies would be required to gain enough statistical power and strengthen currently proposed assumptions on the linkages of AS3MT SNPs with arsenic methylation and cancer risk.
SNPs in AS3MT can provide useful information on the medical treatment of some diseases or risk assessment. For example, arsenic trioxide has been used for the treatment of acute promyelocytic leukemia, although the systematic mechanism is not fully understood [66]. The information on gene-arsenic interactions presented here can be applied to a personalized medicine for leukemia or other cancers. To better understand the variations in arsenic metabolism and to assess the potential susceptibility to its toxic and carcinogenic effects at individual and population levels, more comprehensive studies would be necessary.