Chronic Mercury Exposure and GSTP1 Polymorphism in Munduruku Indigenous from Brazilian Amazon

Genetic polymorphisms may be involved with mercury levels and signs and symptoms of intoxication from this exposure. Therefore, the aims were to describe the frequency of the GSTP1 polymorphism and to evaluate its effects on mercury levels and neurological signs in three Munduruku indigenous villages in the Brazilian Amazon. One-hundred-and-seven indigenous (over 12 years old) were included and genotyped (rs1695) using a TaqMan validated assay. Then, associations were evaluated by binary logistic regression, using odds ratios (OR) and 95% confidence intervals (CI). Mean age was 27.4 ± 13.9 years old, 52.3% were male, mean hair mercury concentration was 8.5 ± 4.3, exceeding the reference limit (≥6.0 µg/g), and were different among the three villages: 13.5 ± 4.6 µg/g in Sawré Aboy, 7.4 ± 2.3 µg/g in Poxo Muybu and 6.9 ± 3.5 µg/g in Sawré Muybu. The minor allele frequency of GSTP1 G was significantly different among the villages: 57% Sawré Muybu, 21% Poxo Muybu and 15% Sawré Aboy. Finally, after adjustment, GSTP1 GG and GA genotypes were associated with lower levels of Hg (OR = 0.13; CI95% = 0.03–0.49) and abnormal somatosensory signs (OR = 3.7; 95%IC = 1.5–9.3), respectively. In conclusion, monitoring this population is imperative to identify individuals at higher risk of developing signs of chronic mercury exposure based on the genetic profile.


Introduction
Mercury is the third most toxic element on the planet and can be found in different forms in the environment, of which organic mercury (methylmercury-MeHg) is the most dangerous chemical form and a global public health concern [1]. Artisanal small-scale gold mining (ASGM) is one of the main sources of human exposure to elemental mercury (Hg(0)), and approximately 3000 tons of Hg were released into the environment between 1975 and 2002, mostly in South America [2]. In the last years, it is estimated that ASGM (also called garimpo) increased over 90% in the Brazilian Amazon [3]. The Amazonian indigenous peoples need natural resources to live and the ASGM aftereffects threaten their livelihoods and cause health risks [4].
(m2). The Hemocue device was used to assess the hemoglobin (Hb) levels and determine the prevalence of anemia (Hb < 11.5 g/dL).
Participants underwent a systematized neurological examination protocol, specially developed for this research, carried out by three of the authors (RAAO, BDP and BHR). For somatosensory signs diagnosis, distal pinprick perception, distal thermal sensitivity, hallux or thumb vibration sensitivity and feet mechanical detection threshold were evaluated. Classical diagnostic research criteria for the diagnosis of peripheral polineuropathy, or neuropathy burden, were used [23] and included the presence of abnormalities in at least one somatic sensory domain and/or alterations in the ankle jerk reflex. Toe amyotrophy and ankle jerk reflex have been observed for the motor function diagnosis, while the brief cognitive screening battery (BCSB), verbal fluency test and stick design test were performed to evaluate the cognitive functions. The neurological assessment of the studied population is described in more detail elsewhere [6].

Hair Mercury Analysis
Mercury exposure levels were determined by the measurements of mercury in the hair of participants. Hair samples were removed with the aid of stainless-steel dissection scissors, close to the scalp in the occipital region and stored in paper envelopes, individually identified. Then, the samples were sent to the Toxicology Laboratory, in the Environment Section of the Evandro Chagas Institute (IEC), in Belém-Pará, Brazil, for an analysis of total mercury levels (THg). The complete applied methodology was described by Basta et al., 2021 [4].
Mercury exposure levels <6.0 µg/g in hair were considered as a safe health limit, following the WHO recommendations [24]. Therefore, levels of hair Hg were categorized in ≤6.0 µg/g or >6.0 µg/g, according to the safety dose recognized by WHO and the median observed in the studied population.

DNA Extraction and GSTP1 Genotyping
Samples from oral mucosa epithelial cells were collected using sterile swabs, stored in a buffered solution, individually identified, and transported to the Laboratory of Pharmaceutical Science-LAPESF of the State University of Rio de Janeiro, West Zone Campus, in Rio de Janeiro-RJ. The access to the genomic DNA was performed using an extraction kit (Qiagen, Hilden, Germany), following the procedures recommended by the manufacturer. Briefly, samples are incubated at 56 • C with 20 µL of proteinase and 400 µL of lysis buffer to release intracellular material. The mixture is then centrifuged and 400 µL of ethanol is added, allowing DNA precipitation. Then, 700 µL of the mixture is transferred to a silica column, which, after centrifugation, retains the DNA. After that, two washing steps are carried out to remove PCR inhibitors, such as divalent cations and proteins, followed by a full-speed spin to remove all traces of wash buffers from the silica column. Finally, we use a low-salt buffer to elute the purified, ready-to-use DNA.
The genotyping analysis of the GSTP1 (chr11:67585218) A>G (rs1695) missense variant was performed using a TaqMan allelic discrimination assay (C_3237198_20) by a 7500 Real-Time System (Applied Biosystems, Foster City, CA, USA), as previously described [14] and the GSTP1 A>G allele frequency and genotype distribution were derived by gene counting.

Data Analysis
Continuous variables were presented as the mean, median, and their respective ranges. Linearity was tested with the Shapiro-Wilk normality test and the differences between groups for variables that did not present a normal distribution were evaluated by the Kruskal-Wallis (KW) nonparametric test. Categorical data were presented as number (n) and frequency (%) and analyzed using the chi-square test or Fisher's exact test, if necessary. Deviations from Hardy-Weinberg equilibrium (HWE) in the GSTP1 A>G polymorphism frequency were assessed by the goodness-of-fit Chi-squared (χ 2 ) test.
The associations between categorical variables and levels of mercury exposure were evaluated by determining the odds ratios (OR) and their respective 95% confidence intervals (95% CI), with adjustment for possible confounding factors, using binary logistic regression models. Age was considered a confounder for the association between the GSTP1 A>G SNP and Hg levels, while age and Hg levels were considered confounders for the association between the GSTP1 A>G SNP and the prevalence of neurological signs. All analyses were performed using IBM SPSS 20.0 Statistics for Windows (SPSS Inc., Chicago, IL, USA) and a significance level of 0.05 was adopted.

Results
The present study was conducted with 107 individuals older than 12 years old, residents from Poxo Muybu (31.8%), Sawré Aboy (21.5%) and Sawré Muybu (46.7%) villages. The majority of the individuals were male in Poxo Muybu and Sawré Aboy villages. The mean age in the overall population was 27.4 ± 13.9 years old (median 24.0, ranging from 12.0 to 72.0), with older individuals living in Sawré Muybu (58%) and younger individuals living in Sawré Aboy (60.8%). Approximately 5% of individuals showed anemia, and it was more common in residents of Sawré Aboy.
The mean level of mercury exposure in the studied population was 8.5 ± 4.3 (median 7.4, ranging from 2.0 to 22.8) and was not normally distributed (Shapiro-Wilk test p-value < 0.001). The Sawré Aboy village presented higher levels of mercury exposure when considering either the WHO cutoff point (6.0 µg/g) and the cutoff determined by the overall median of the studied population (7.4 µg/g). The Sawré Aboy residents also presented higher prevalence of abnormal somatosensory signs and cognitive functions ( Table 1). When considering the prevalence of at least one abnormal neurological sign, 57.9% of individuals were affected in the overall population and a higher prevalence (82.6%) was found in Sawré Aboy residents (p-value = 0.004), comparing the three villages (data not shown). The rate of successful genotyping of the GSTP1 A>G SNP was 99.1%, since only a sample from one individual did not amplify during the PCR experiment. The distribution of the polymorphism was in Hardy-Weinberg equilibrium for the overall population (p-value = 0.79) and for each of the villages Poxo Muybu, Sawré Aboy and Sawré Muybu (p-values = 0.61, 0.39 and 0.19, respectively). The frequency distribution of the GSTP1 A>G polymorphism is presented in Figure 1. The distributions of the genotypes and minor allele frequency (MAF) among the three villages were statistically different (p-value < 0.0001), with Sawré Muybu residents presenting the highest frequency of the GSTP1 GG variant genotype and no individual from Sawré Aboy having that genotype. In addition, the Sawré Aboy village presented the highest frequency of the GSTP1 AA wild-type genotype. The rate of successful genotyping of the GSTP1 A>G SNP was 99.1%, since only a sample from one individual did not amplify during the PCR experiment. The distribution of the polymorphism was in Hardy-Weinberg equilibrium for the overall population (pvalue = 0.79) and for each of the villages Poxo Muybu, Sawré Aboy and Sawré Muybu (pvalues = 0.61, 0.39 and 0.19, respectively). The frequency distribution of the GSTP1 A>G polymorphism is presented in Figure 1. The distributions of the genotypes and minor allele frequency (MAF) among the three villages were statistically different (p-value < 0.0001), with Sawré Muybu residents presenting the highest frequency of the GSTP1 GG variant genotype and no individual from Sawré Aboy having that genotype. In addition, the Sawré Aboy village presented the highest frequency of the GSTP1 AA wild-type genotype.  To investigate the associations between the GSTP1 A>G SNP and Hg exposure (Table 2), we first considered the Hg levels categorized according to the WHO guidelines (≤6.0 µg/g or >6.0 µg/g). It was observed that the GSTP1 GG genotype was associated with lower levels of Hg exposure in both crude and adjusted analyses. Furthermore, this association is slightly stronger in the recessive model (AA + AG versus GG). Interestingly, the Sawré Aboy residents showed significant (p-value < 0.0001) higher levels of Hg (13.5 ± 4.6 µg/g; median = 12.5, range = 4.8-22.8) compared with Poxo Muybu (7.4 ± 2.3 µg/g; median = 7.20, range = 2.8-12.9; p-value < 0.0001) and Sawré Muybu (6.9 ± 3.5 µg/g; median = 6.38, range = 2.0-16.0, p-value < 0.0001), and no individual from Sawré Aboy presented the GSTP1 GG variant genotype (Figure 1), which was associated with protection (lower levels of Hg). In addition, we investigated this association by considering the Hg levels categorized by the overall median (7.4 µg/g) and the results remained the same. Therefore, for the subsequent analysis, we considered only the Hg ≤ 6.0 (µg/g) 2 versus > 6.0 (µg/g) 2 classification.
Finally, the association between the GSTP1 A>G polymorphism and the prevalence of neurological signs was investigated. When considering the presence of any abnormal somatosensory sign, the GSTP1 AG genotype showed a positive association in both crude and adjusted analyses for age and Hg exposure levels (OR = 3.70 and 95% IC = 1.47-9.29; OR = 3.70 and 95% IC = 1.47-9.29, respectively), while the presence of abnormal motor or cognitive signs was not statistically significant (data not shown). In addition, each neuro-logical sign individually and GSTP1 A>G polymorphism also investigated (Tables 3 and 4). For the somatosensory signs, individuals carrying the GSTP1 AG genotype had an approximately 4-fold higher chance of having clinical signs of polyneuropathy, while the other signs were not significant among different GSTP1 genotypes (Table 3). There was a low prevalence of abnormal motor functions in the studied population (4.7% for toe amyotrophy and 16.8% for abnormal ankle jerk reflex) and all individuals who presented toe amyotrophy were carriers of the GSTP1 AG genotype (p-value = 0.04), however, it was not possible to estimate the OR and the respective 95% IC (Table 4). Regarding the cognitive evaluation, no association was observed with GSTP1 genotypes, however, the BCSB was significantly different among the groups, and approximately 42% of the altered results in this battery were from individuals carrying the GSTP1 AA genotype (Table 4), which was positively associated with higher levels of Hg in Table 2.

Discussion
The present study described the association of a single nucleotide polymorphism from a gene involved in the Hg metabolism and the presence of neurological signs of chronic mercury exposure in 107 indigenous adults residing in a region of the Brazilian Amazon affected by illegal mining activities. The Hg levels and GSTP1 A>G polymorphism genotype frequencies were significantly different among the three villages (Poxo Muybu, Sawré Aboy and Sawré Muybu). The GSTP1 AA genotype was associated with higher levels of Hg and the heterozygotic genotype (GSTP1 GA) was associated with a higher chance of having an abnormal somatosensory sign and neuropathy burden. In addition, the Sawré Aboy residents showed: (i) higher levels of Hg and (ii) higher frequency of the GSTP1 AA genotype.
Our findings retake the discussion initiated by Basta et al., 2021, who observed in this same population that Hg exposure levels and impaired neurological functions were higher in the Sawré Aboy village, which is located downstream from the Jamanxin River and is the closest village to the mining activities in comparison to the Sawré Muybu and Poxo Muybu ones [4,6]. In addition, these findings can also be explained by the higher frequency of the The distribution profile of the GSTP1 A>G (rs1695) polymorphism varies according to the population, with the MAF (allele G) ranging from 24.7% to 43% in the Brazilian population [25][26][27][28][29][30][31]. As far as we know, there is no data on GSTP1 polymorphism frequency from other indigenous Brazilian populations focusing on mercury exposure. A recent study investigated the GSTP1 frequency in the nonindigenous Brazilian population and observed that the MAF (ranging from 30.7% to 31.5%) did not differ between the groups based on skin color: black, mulatto, nonwhite and white [32]. Here, the GSTP1 MAF in the overall indigenous population was 36.8%, similar (41.8%) to that described in the indigenous from the Amazon admitted to the Indian House of Health in Boa Vista, Roraima, Brazil (CASAI/RR), aiming to evaluate the genetic profile (TP53 and GSTP1 genes) and prostatic features [26]. However, a significant difference was found among the three villages: 21.2% Poxo Muybu, 15.2% Sawré Aboy and 57.0% Sawré Muybu. The Munduruku are an indigenous community from the Brazilian Amazon that lives isolated from other Amerindian populations and rarely display admixture with nonindigenous. Despite living near each other for centuries, these populations maintain their distinctiveness, according to the substantial differences found in their genetic profiles [4]. For this reason, it is not possible to extrapolate these results, not even for the same ethnic group. Therefore, the implementation of an individual genetic diagnosis is necessary.
The GST superfamily is comprised of 16 genes, including GSTP1, and encodes vital defense enzymes involved in the detoxification of reactive oxygen species and metal biotransformation [33,34]. GSTP1 is a key enzyme because it is the most widely expressed GST in the body. In addition to being frequent, the GSTP1 A>G non-synonymous polymorphism changes significantly the enzyme-substrate affinity [18,19,[35][36][37]. The literature is not conclusive about the relationship between the GSTP1 A>G polymorphism and levels of Hg exposure. Our results contrast with studies where the minor allele GSTP1 G was associated with higher Hg levels from different exposure sources and biological matrices such as hair [9,38] and urine [13], and is agreement with other studies where the GSTP1 G showed a protective effect against elevated Hg levels dosed from hair [22] and blood samples [21,39]. One possible reason for the discrepancy among the studies is the use of different biological matrices to measure the Hg exposure, depending on the route of exposure, which may not present the same association with the polymorphism in the GSTP1 gene, which is mostly involved with the metabolism of MeHg ingested from the diet [9,15,39]. Furthermore, this raises the question of whether Hg dosed from hair samples reflects the circulating levels of the metal or, instead, represents how much is being excreted from the organism. Nevertheless, the measurement of Hg in hair samples is advantageous due to the simple collection and transportation, which are easier than for blood and urine samples [40,41].
There is evidence that the GSTP1 polymorphism is associated with several neurological conditions, such as multiple sclerosis [42] and autism spectrum disorders [43]. To date, however, no other clinical study has found an association between the GSTP1 A>G polymorphism and signs of neurotoxicity induced by MeHg. We believe that the enzymatic alteration caused by this SNP may alter the MeHg exposure levels in the peripheral and central nervous systems. This, in addition to the enhancement of oxidative stress, may play a role in the pathogenesis of the neurological abnormalities found in this series [44]. In the present study, despite the association found for the GSTP1 AG genotype, it is not possible to conclude that the SNP is involved with the onset of neurological signs and symptoms due to the small number of individuals enrolled and the low prevalence of these outcomes in the population. Furthermore, this might be due to the low prevalence of the neurological signs and symptoms of chronic mercury exposure and the low frequency of the GSTP1 AA genotype in Sawré Muybu and the GSTP1 GG genotype in Poxo Muybu and Sawré Aboy, ensuring a higher frequency of the GSTP1 AG genotype in the overall population. In addition, the individuals were young, and this kind of outcome is more prevalent in later ages, which could be better investigated in a longitudinal observation of this population, that also has an elevated frequency of polymorphism associated with mercury exposure levels.
Studies with indigenous communities have multiple ethical aspects that need to be considered, such as the cultural importance of the indigenous body and, therefore, all specimens that can be collected, such as tissue, blood, and hair [45]. In the present study, we were able to count on the engagement of community leaders, who made it possible to collect hair and oral mucosa samples and perform the needed analysis. This access was allowed thanks to the effort and competence of the multidisciplinary team that was assembled to evaluate this population, involving biologists, physicians, nurses, psychologists, and pharmaceutical geneticists. Efforts are being made to create a closer interaction between researchers and indigenous communities, to increase research recruitment and also to provide reports of individual results obtained with the studies [4]. This approach is especially important when genetic studies are carried out because individuals must be aware of the benefits of engaging in the research [4,14,45].
Although some limitations need to be discussed, such as nonmeasured variables that would be important to evaluate, such as occupation, alcohol consumption and smoking status, those can be sources of bias. Furthermore, the small sample size may lead to relatively weak power to detect the real association between the polymorphism and the neurological signs and symptoms. However, the potential benefits that can be provided to the communities must be considered. Therefore, in order to reduce some of these limitations and expand the clinical evaluation carried out among the Munduruku indigenous people, a longitudinal study is being designed to evaluate levels of mercury exposure in different matrices and the chronic effects of mercury exposure, in addition to identifying possible polymorphisms as susceptibility biomarkers of these events. Furthermore, the villages are in a high-risk area for ASGM since levels of Hg are extremely high, which reveals the extreme vulnerability of this population and the need for constant environmental and individual monitoring through a wide genetic evaluation that can easily be carried out with noninvasive biological matrices, such as the oral mucosa cells used in this work.

Conclusions
This is the first study to evaluate the association of the GSTP1 A>G polymorphism with Hg levels and the neurological signs that can be caused by chronic exposure to heavy metals in an extremely vulnerable population from the Brazilian Amazon. The Hg levels were significantly higher in Sawré Aboy residents, followed by those from Poxo Muybu and Sawré Muybu, while the GSTP1 G allele, which is associated with lower levels of Hg, was more frequent in Sawré Aboy residents, followed by Poxo Muybu residents, and finally Sawré Muybu residents. The GSTP1 AA and GA genotypes were associated with higher levels of Hg and a higher chance of having an abnormal somatosensory sign, respectively. Our findings highlight the importance of monitoring the indigenous communities in the Brazilian Amazon, who live in a vulnerable situation caused by the abandonment of the state. In conclusion, it is evident that further wide evaluation is imperative to this population, in order to identify individuals at higher risk of developing signs and symptoms of chronic mercury exposure. Thus, individual genetic diagnosis tests are crucial, since it is not possible to extrapolate data from other populations, even from other Amerindians.  , the study began with a pre-study consultation, carried out in August 2019, during a visit to the villages, in which two authors (SSH and PCB) and local indigenous leaders participated. At the time, the study objectives were presented and discussed (https://youtu. be/oFEYEGxNmns?t=4, accessed on 4 January 2023). After answering questions and receiving approval of the proposal from the communities, we received support from the coordination of the Special Indigenous Sanitary District of the Tapajós River through the multidisciplinary indigenous health team to carry out the study. In addition, the interviews and data collection started only after the participants had their questions answered and given formal consent in the Informed Consent Form (ICF) by the children's guardians.
Informed Consent Statement: Informed consent was obtained from all subjects involved in the study.
Data Availability Statement: Not applicable.