Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2023 Feb 10.
Published in final edited form as: Expo Health. 2022 Feb 3;14(4):941–949. doi: 10.1007/s12403-022-00468-2

Alterations in Microbial-Associated Fecal Metabolites in Relation to Arsenic Exposure Among Infants

Margaret R Karagas 1, Susan McRitchie 2, Anne G Hoen 1, Cindy Takigawa 1, Brian Jackson 3, Emily R Baker 4, Juliette Madan 1,5, Susan J Sumner 2, Wimal Pathmasiri 2
PMCID: PMC9918239  NIHMSID: NIHMS1818701  PMID: 36776720

Abstract

In utero and early life exposure to inorganic arsenic (iAs) alters immune response in experimental animals and is associated with an increased risk of infant infections. iAs exposure is related to differences in the gut microbiota diversity, community structure, and the relative abundance of individual microbial taxa both in laboratory and human studies. Metabolomics permits a direct measure of molecular products of microbial and host metabolic processes. We conducted NMR metabolomics analysis on infant stool samples and quantified the relative concentrations of 34 known microbial-related metabolites. We examined these metabolites in relation to both in utero and infant log2 urinary total arsenic concentrations (utAs, the sum of iAs and iAs metabolites) collected at approximately 6 weeks of age using linear regression models, adjusted for infant sex, age at sample collection, type of delivery (vaginal vs. cesarean section), feeding mode (breast milk vs. any formula), and specific gravity. Increased fecal butyrate (b = 214.24), propionate (b = 518.33), cholate (b = 8.79), tryptophan (b= 14.23), asparagine (b = 28.80), isoleucine (b = 65.58), leucine (b = 95.91), malonate (b = 50.43), and uracil (b = 36.13), concentrations were associated with a doubling of infant utAs concentrations (p< 0.05). These associations were largely among infants who were formula fed. No clear associations were observed with maternal utAs and infant fecal metabolites. Metabolomic analyses of infant stool samples lend further evidence that the infant gut microbiota is sensitive to As exposure, and these effects may have functional consequences.

Keywords: Arsenic, Metabolomics, In utero exposure, Postnatal exposure

Introduction

Arsenic (As) is naturally present in the earth’s crust and has been used for a broad range of industrial purposes, including as antibacterial and immune modulating medicines in the nineteenth and early twentieth centuries (Kapp 2018). Anthropogenic and natural sources of As contribute to widespread exposure primarily from the food and drinking water systems. The World Health Organization estimates that at least 140 million people worldwide across 50 countries are exposed to drinking water exceeding the provisional guideline of 10 μg/L As. Populations relying on private, unregulated water systems remain vulnerable to As concentrations above regulatory limits. Fish and seafood contain As in forms that are not metabolized by humans and therefore considered non-toxic such as arsenobetaine. However, other dietary sources, especially rice and rice products, fruits and fruit juices, and seaweeds, may contain appreciable concentrations of inorganic As (iAs) and other potentially detrimental forms of As (EFSA Panel on Contaminants in the Food Chain 2009). Diet is the primary source of As exposure for the majority of people and contributes to disproportionately higher intakes of As among young children for their body weight (EFSA Panel on Contaminants in the Food Chain 2009). This poses a health concern, especially for infants consuming formula mixed with potentially contaminated water and when transitioning to solid foods such as infant rice cereal (EFSA Panel on Contaminants in the Food Chain 2009; Carignan et al. 2016; Signes-Pastor et al. 2018).

Evidence both from laboratory experiments and epidemiologic investigations support immunotoxic effects of As on both innate and adaptive immunity (Dangleben et al. 2013). In our earlier work, we observed an increased risk of infant infection associated with higher in utero As concentrations, and specifically lower respiratory infections Farzan et al. 2016. Immunity develops in utero and evolves during the first years of life (Dietert et al. 2010). The gut microbiota plays an essential, bidirectional role in this process—gut microbiota stimulate the maturation of the neonatal immune system, and in turn, an infant’s immune response helps to shape the composition of microbes inhabiting the gut (Madan et al. 2012). We previously reported on the relation between infants’ urinary As concentrations and gut microbiota composition in a pregnancy cohort of maternal–child dyads from New Hampshire (Hoen et al. 2018; Laue et al. 2020). Notably, we observed decreased relative abundance of keystone taxa in the genera Bacteroides and Bifidobacterium involved in immune maturation (Hoen et al. 2018; Laue et al. 2020). Analysis of fecal samples using next-generation sequence-based methods allowed us to observe the genetic make-up of microbes in the gut and estimate its diversity, community structure, and composition. However, it only can be used to make inferences about the collective function of the gut microflora. Metabolomics complements genetic sequence-based profiling by providing a critical bridge from microbiota composition to the complex array of small molecules that directly influence biologic activities. As-induced functional changes in the gut microbiota as reflected in the metabolome have been characterized from mouse experiments (Li et al. 2019). By comprehensively profiling the fecal metabolome in the context of a longitudinal pregnancy cohort study, we sought to gain a clearer window into the phenotype of the human infant gut microbiota related to As exposure.

Materials and Methods

Study Population

Our study is based on the ongoing New Hampshire Birth Cohort Study of women recruited during pregnancy and whose offspring are followed to update exposure and health information (Hoen et al. 2018). Reproductive and medical history, health, diet, and lifestyle factors were ascertained from questionnaires and medical record review during pregnancy, and a maternal urine sample was collected at approximately 24 to 28 weeks gestation. Newborn infant characteristics were documented from the delivery medical records, and both infant urine and stool samples were collected at approximately 6 weeks of life.

Assessment of Maternal Pregnancy (In Utero) and Infant Postnatal As Exposure

iAs undergoes a series of reduction and oxidative methyl processes from iAsIV to iAsIII to MMAV to monmethyl-arsonus acid (MMAIII) to DMAV and is excreted in these forms (Abuawad et al. 2021). Therefore, to estimate As exposure, we analyzed urine samples collected during pregnancy at 24–28 weeks gestation and during infancy at approximately 6 weeks of life for As species [arsenite (iAsIII), arsenate (iAsV), monomethylarsonic acid (MMA), dimethylarsinic acid (DMA) and arsenobetaine (AsB)] using high-performance liquid chromatography (HPLC)–inductively coupled plasma mass spectrometry, ICPMS (Hoen et al. 2018). To compute total urinary As we summed the individual fractions of iAs (iAsIII and iAsIV) and the metabolites of iAs, MMA, and DMA (utAs = iAs + MMA + DMA), excluding arsenobetaine found in fish and seafood, which is unmetabolized and considered non-toxic. Although MMAIII is excreted it is rapidly converted to DMA and therefore typically undetectable using standard approaches. Specific gravity of urine was determined by a handheld refractometer with automatic temperature compensation (PAL-10S; ATAGO Co. Ltd.) to adjust for urinary dilution. We further divided MMA by iAs and DMA by MMA to calculate the primary and secondary methylation indices (PMI and SMI), respectively, as indicators of iAs metabolic capacity (Shen et al. 2016). Detection limits for iAs, MMA, and DMA were 0.1, 0.02 and 0.02 μg/L. The spiked recovery rates averaged 82% for iAs, 91% for MMA, 89% for DMA and 94% for AsB.

Stool Collection

Infant diapers containing stool were collected at home by caregivers, stored in a home freezer, and then brought frozen to the 6-week postpartum visit in thermal transport bags on ice packs or transported directly on ice packs. Upon receipt, diapers were stored at − 80 °C until processing. Following an overnight thaw at 4 °C, the stool was aliquoted and frozen at − 80 °C. Stool samples aliquoted into tubes certified as trace element free were used for metabolomic analysis.

Sample Preparation and Data Acquisition

De-identified aliquots of infant stool samples, along with replicates for quality control (QC), were shipped to the NIH Eastern Regional Comprehensive Metabolomics Research Core on dry ice and immediately stored at − 80 °C after being logged in for metabolomics analysis. The metabolomics analysis were adapted from previously described procedures (Brim et al. 2012, 2017; Livanos et al. 2016). Briefly, samples were randomized into batches. In each batch, samples were thawed, and ~150 mg of stool was transferred to MagNA Lyser tubes after recording the weight; samples were then homogenized with 50% acetonitrile in water by using a bead homogenizer (100 mg fecal mass/mL). Homogenized samples were centrifuged at 16,000 rcf, and the supernatant was separated into another tube. An aliquot (1000 μL, 100 mg equivalent of fecal mass) was transferred into an Eppendorf tube and lyophilized overnight. The dried extract was reconstituted in 700 μL of NMR master mix (containing 0.2 M phosphate, 0.5 mM DSS-d6, and 0.2% sodium azide), vortexed on a multitube vortexer at speed 5 for 2 min, and centrifuged at 16,000 rcf for 5 min. A 600 μL aliquot of the supernatant was transferred into a pre-labeled 5 mm NMR tube for data acquisition on a 700 MHz spectrometer. Additionally, pooled QC samples (study pools created from randomly selected study samples and batch pools) were generated from supernatants of study samples, and aliquots of pooled QC samples were dried and reconstituted similar to study samples described above and used for further QC purposes. 1H NMR spectra of feces samples were acquired on a Bruker 700 MHz NMR spectrometer using a 5 mm cryogenically cooled ATMA inverse probe and ambient temperature of 25 °C. A 1D NOESY presaturation pulse sequence (noesygppr1d, [recycle delay, RD]-90°-f1-90°-fm-90°-acquire free induction decay (FID)]) was used for data acquisition (Beckonert et al. 2007; Dona et al. 2014). For each sample, 64 transients were collected into 64k data points using a spectral width of 12.02 ppm, 2 s relaxation delay, 10 ms mixing time, and an acquisition time of 3.899 s per FID. The water resonance was suppressed using resonance irradiation during the relaxation delay and mixing time. NMR spectra were processed using TopSpin 3.5 software (Bruker-Biospin, Germany). Spectra were zero-filled, and Fourier transformed after exponential multiplication with a line broadening factor of 0.5. Phase and baseline of the spectra were manually corrected for each spectrum. Spectra were referenced internally to the DSS-d6 signal (d=0 ppm). The quality of each NMR spectrum was assessed for the level of noise and alignment of identified markers. Spectra were assessed for missing data and underwent quality checks. NMR bins (0.5–9.0 ppm) were created excluding water (4.73–4.85 ppm) using intelligent bucket integration of 0.04 ppm bucket width with 50% looseness using ACD Spectrus Processor (ACD Labs, Inc., Toronto, Canada). Integrals of each of the bins were normalized to the total integral of each of the spectra. Chenomx NMR Suite 8.1 Professional (Chenomx, Inc., Edmonton, AB, Canada) (Weljie et al. 2006) was used to determine the relative concentrations of library-matched metabolites previously identified as associating with host and gut microbes co-metabolism (Li et al. 2008; Zheng et al. 2011; Nicholson et al. 2012).

Statistical Analyses

Statistical analyses were conducted using data from infants at approximately 6 weeks who had fecal metabolomics data and complete covariate data. Descriptive statistics were calculated for the participant characteristics, As concentrations, and relative concentration of metabolites. The masked QC replicates were used to calculate the intra-class correlation (ICC) for each metabolite with a detectable relative concentration. Any metabolite with an ICC < 0.2 was excluded from the analysis.

Normalized binned NMR data were Pareto-scaled and centered prior to multivariate analysis using SIMCA 14 (Sartorius Data Analytics, Umeå, Sweden). The scores plot from the principal component analysis (PCA) was inspected to ensure that the laboratory QC pool samples were clustered in the center of study samples used to create the pools, a method widely applied to metabolomic studies (Chan et al. 2011; Masson et al. 2011; Broadhurst et al. 2018).

Spearman correlations were calculated for each metabolite’s relative concentration with urinary total As (utAs). Metabolites with p-value < 0.1 were used as the outcome in multivariable linear regression models to determine the associations between log2 transform of urinary As concentration (maternal and infant utAs separately). Models with infant urinary As concentrations were adjusted for infant age, sex, feeding method (exclusive breastfeeding or formula/mixed feeding at 6 weeks of age), urine-specific gravity, and delivery mode (vaginal or cesarean section). Models with maternal urinary As concentrations were adjusted for urine-specific gravity, infant age, sex, feeding method, and delivery mode. Continuous variables were centered prior to modeling. We also performed the analysis stratified by below or equal to or above the median infant urinary PMI (0.35) and SMI (8.06).

SAS 9.4 (SAS Institute, Inc., Cary, NC) was used for calculating descriptive statistics, hypothesis tests, correlations, and linear regression, and ICC coefficients of replicate samples. For this exploratory study, p-values < 0.05 were considered to be statistically significant and were not adjusted for multiple testing (Bender and Lange 2001; Xi et al. 2014).

Pathway Enrichment Analysis

GeneGo MetaCore (Clarivate Analytics, PA) was used to assess the enrichment of perturbed metabolic pathways derived from the concentration data. MetaCore uses the hypergeometric test, which represents the enrichment of certain metabolites in a pathway, together with the false discovery rate (FDR). A q-value < 0.05 is considered indicative of significant enrichment in pathways.

Results

Study Population

A total of 83 infants with NMR metabolomics data acquired from infant stool samples also had maternal urinary As species measured at approximately 24 to 28 weeks gestation and complete data for model covariates. Eighty-one infants with NMR metabolomics data also had an analyzed postnatal, approximately 6-week urine sample for As species and complete data for model covariates. The mean age of the 81 infants at stool and urine sample collection was 46 days, 62% were boys, 26% were delivered by cesarean section, and 42% were exclusively breast-fed (Table 1). The utAs concentration during pregnancy was on average 4.1 mg/L, with a range of 0.2 to 21.0 mg/L (Table 1). Among infants at approximately 6 weeks of age, the average utAs concentration was 0.6 μg/L with a range of 0.1 to 5.2 μg/L (Table 1). The mean (range) of the individual As species were 0.1 μg/L (undetectable to 1.0 μg/L) for iAs, 0.1 μg/L (undetectable to 0.4 μg/L) for MMA, 0.4 μg/L for DMA (undetectable to 4.5 μg/L) and 0.1 μg/L for AsB (undetectable to 1.2 μg/L). Maternal utAs concentrations averaged 4.1 μg/L, with a range of 0.2 to 21.0 μg/L (Table 1). Urinary specific gravity was within a narrow range in both maternal and infant samples (mean = 1.01; range 1.00, 1.03 and mean 1.00; range 1.00, 1.02, respectively).

Table 1.

Selected characteristics of infants from the New Hampshire Birth Cohort Study

Maternal utAs analysis (n = 83) Infant utAs analysis (n = 81)
Characteristic Mean [range] or N (%) Mean [range] or N (%)
Infant age at sample collection (days) 45.5 [28.0, 115.0] 45.6 [28.0, 115.0]
Sex
 Male 53 (63.9%) 50(61.7%)
 Female 30 (36.1%) 31 (38.3%)
Delivery mode
 C-section 22 (26.5%) 21 (25.9%)
 Vaginal 61 (73.5%) 60(74.1%)
Feeding at ~ 6 weeks
 Exclusively breast fed 35 (42.2%) 34 (42.0%)
 Formula/mixed fed 48 (57.8%) 47 (58.0%)
 Missing 0 0
Maternal age at enrollment (years) 31.5 [21.1,43.7] N/A*
Maternal urine, total As excluding AsB (mg/L) 4.1 [0.2,21.0] N/A*
Infant urine, total As excluding AsB (mg/L) N/A* 0.6 [0.1, 5.2]
iAs N/A* 0.1 [0, 1.01
MMA N/A* 0.4 [0, 41
DMA N/A* 0.1 [0, 4.5]
Open in a new tab
*

NIA not applicable

Quality Control

ICCs for formate and fumarate fell below 0.2 and thus were excluded from further analyses. The average ICC coefficients for the remaining 34 metabolites ranged from 0.2 for isobutyrate to 0.99 for succinate (Supplemental Table 1) and averaged 0.75 for those metabolites used in the analysis. Laboratory QC pools were centered in the PCA plots of the samples from which the pools were created, further indicating that the NMR data were of high quality (data not shown).

Linear Regression

Eighteen of the 34 concentration fitted metabolites with ICC ≥ 0.2 had p-values < 0.1 for Spearman correlations with utAs (Supplemental Table 2). These 18 metabolites measured from infant fecal samples were used as the dependent variables in multivariable linear regression models to examine the associations between maternal and infant urinary utAs (independent variables) after adjusting for covariates. A doubling of infant utAs concentrations was associated with statistically significant increases (p < 0.05) in the relative concentrations of infant fecal short-chain fatty acids (SCFAs) butyrate (b = 214.24) and propionate (b = 518.33); the bile acid cholate (b = 8.79); the amino acids asparagine (b = 28.80), isoleucine (b = 65.58), leucine (b = 95.91); and tryptophan (b= 14.23), the pyrimidine uracil (b = 36.13), and the organic acid malonate (b = 50.43, Table 2). Positive associations tended to occur among infants fed formula, with negative associations for certain metabolites among exclusively breast-fed infants (Supplemental Table 3). Additional interactions were observed with phenylalanine and proline (Supplemental Table 3). Maternal urinary As was related only to infant fecal acetate concentration in unadjusted models (rs= − 0.24, p = 0.031) but was no longer statistically significant after adjustment (b= − 847.51, p = 0.115). Associations with infant urinary As and tryptophan were largely among those with high PMI (p for interaction = 0.0228, Supplemental Table 4), and those for cholate were largely among those with low SMI (p for interaction = 0.0037, Supplemental Table 5). However, most interaction terms did not reach statistical significance.

Table 2.

Changes in microbe related metabolite concentrations with a doubling (log2) of infant urinary arsenic concentrations (n = 81) from multivariable linear regression models

Metabolite β (95% Q) p-value
Short-chain fatty acids
 Butyrate 214.24 (83.78, 344.70) 0.0016
 Propionate 518.33 (94.39,942.28) 0.0172
 Isobutyrate 2.61 (−0.86,6.07) 0.1378
Lipids
 Glycerol −67.89 (−139.16, 3.38) 0.0616
Bile acid
 Cholate 8.79(4.21, 13.36) 0.0003
Amino acids
 Tryptophan 14.23 (3.71,24.74) 0.0087
 Lysine 63.29 (−42.79, 169.36) 0.2383
 Asparagine 28.80(10.27,47.33) 0.0028
 Methionine 16.91 (−5.32, 39.14) 0.1338
 Proline 17.09 (−28.28, 62.46) 0.4553
 Isoleucine 65.58 (10.56, 120.60) 0.0201
 Leucine 95.91 (8.88, 182.95) 0.0312
 Glutamate 119.69 (−22.02,261.40) 0.0966
 Phenylalanine 18.38 (−12.46,49.23) 0.2388
Sugars
 Fucose −126.31 (−307.21,54.59) 0.1683
Organic acid
 Malonate 50.43 (3.14,97.72) 0.0369
Pyrimidine
Uracil 36.13 (2.74,69.52) 0.0343
Other
 Propylene glycol −62.58 (−128.62, 3.46) 0.0629
Open in a new tab

Adjusted for infant sex, age, type of delivery (vaginal vs. C-section), feeding mode (breast milk vs. any formula), and specific gravity

Pathway Analysis of Concentration Data

MetaCore pathway enrichment analysis using the metabolites associated with utAs in infant urine by Spearman correlation (p < 0.1, Supplemental Table 6). Top enriched pathways identified included aminoacyl-t-RNA biosynthesis, amino acid dependent mTORC1 activation (signal transduction), amino acid metabolism (lysine, branched chain amino acids, BCAAs (isoleucine, leucine, valine), and methionine), saturated fatty acid synthesis to hexadecenoic acid, regulation of lipid metabolism (by niacin and isoprenaline), immune responses [through myeloid-derived suppressor cells (MDSC), M2 macrophages, and Treg cell-mediated modulation of antigen-presenting cell (APC) functions] were among the top enriched pathways (Fig. 1).

Fig. 1.

Fig. 1

Open in a new tab

Top 10 enriched pathways in the MetaCore pathway enrichment analysis for the metabolites (concentration data) associated with infant utAs. See Supplementary Table 6 or the full list of enriched pathways

Discussion

As exposure remains a major public health concern problem worldwide. The gut microbiota biochemically transforms As compounds (Coryell et al. 2019; McDermott et al. 2020) and at the same time may be affected by As exposure (Chi et al. 2018; McDermott et al. 2020). In our prospective pregnancy cohort study, we observed positive associations between infant urinary As concentrations and the relative concentration of nine infant fecal metabolites (asparagine, butyrate, cholate, isoleucine, leucine, malonate, propionate, tryptophan, and uracil) in multivariable regression analyses. Maternal urinary As concentrations during pregnancy were unrelated to infant fecal metabolites overall. Only a weakly negative association (adjusted p = 0.115) was observed between maternal urinary As concentrations and infant fecal acetate concentrations. However, as As concentrations change over the course of pregnancy (Hopenhayn et al. 2003; Tseng 2009; Gardner et al. 2011; Gao et al. 2019a, b; Gao et al. 2019a, b), it is conceivable that our sampling of ~ 24–28 weeks gestation did not capture the relevant time window to influence maternal–fetal transfer of the microbiome during delivery or otherwise influence either microbe or host metabolic products. We found concentrations of SCFAs, bile acids, amino acids, organic acids, and pyrimidines associated with infant urinary As concentrations. These changes may reflect impacts on the microbial composition of the gut or activation or detoxification pathways of either the host or microbes, epigenetic effects or other mechanisms. Thus, alterations in the metabolic pathways associated with these metabolites by As may provide insights on the mechanistic interplay between As and the gut microbiota with functional health consequences.

We previously identified associations between infant urinary As concentration and the gut microbiota at about 6 weeks of age (Hoen et al. 2018). Findings were especially evident among those receiving formula, as in the current study. Based on earlier work of our study and others (Carignan et al. 2016), formula results in higher As exposure due to both the formula powder and the water used to mix the formula. Household tap water can contain high levels of As in our rural cohort which whose households were served by private unregulated water systems such as bedrock wells as an eligibility criterion (Hoen et al. 2018). Eight genera, six within the phylum Firmicutes, were enriched with higher As exposure. This is consistent with our current results of a positive relationship with fecal butyrate and propionate concentrations which are produced by anaerobic fermentation of dietary carbohydrates by Firmicutes (Louis and Flint 2017; Appert et al. 2020).

SCFAs, succinate, formate, acetate, butyrate, and propionate. are important energy sources for intestinal epithelial cells, have diverse regulatory functions, and impact host physiology and immunity (Louis and Flint 2017; Appert et al. 2020). Of these, acetate has the highest systemic concentrations, as they can be produced by most gut anaerobes, whereas propionate and butyrate are products of only a select group of gut microbiota (Louis and Flint 2017). Butyrate, a SCFA produced by bacterial fermentation of dietary fiber, is considered a key metabolite during infant gut development in part for its immunoregulatory effects (Roduit et al. 2019). Microbial metabolites, including butyrate, also are hypothesized to play an important role in the gut-brain axis by modulating the functional and signaling activity of brains cells and the blood–brain barrier and influence risk of neurodevelopmental outcomes such as autism (Smith 2015; Liu et al. 2019; Silva et al. 2020). There is evidence of functional redundancy of butyrate producers with co-occurrence of Clostridiaceae, Ruminococcaceae, and Lachnospiraceae, dominated by the endospore-forming Clostridiaceae (Appert et al. 2020). Further, proteolytic microbiota in the gut produce butyrate and propionate from peptide and amino acid fermentation (Louis and Flint 2017).

Propionate is a metabolite produced by genera in both Firmicutes and Bacteroidetes, including Bacteroides and Clostridium G2 (Gonzalez-Garcia et al. 2017; Louis and Flint 2017). Like propionate, malonate, which also was increased in relation to higher As exposures, plays a role in tricarboxylic acid cycle and other bacterial metabolic processes (Suvorova et al. 2012). In addition, malonate competitive inhibits succinate dehydrogenase and is involved in the metabolism of propionate (Suvorova et al. 2012). Thus, perturbations in either gut microbes that produce propionate, or in Krebs cycle metabolism that produces malonate, could in part explain the observed differences in propionate and malonate metabolism. However, our results need to be replicated in further studies.

In pathway analyses, we found alterations in metabolites enriched in MDSCs and M2 macrophages in cancer, and immune response, Treg cell-mediated modulation of APC functions) in relation to infant urinary As concentrations. MDSCs inhibit T cell function. T cell alterations occur with leukemia treatment with As (Gao et al. 2017) and have been observed both in highly drinking water-exposed populations (Burchiel et al. 2020), and in our own US cohort of infants exposed prenatally (Nygaard et al. 2017), we observed decreased cord blood naïve T-cells in relation to utAs concentrations during pregnancy.

Enriched pathways identified in our analyses included aminoacyl-tRNA biosynthesis and amino acid metabolism [lysine, BCAAs (isoleucine, leucine, valine), and methionine]. Martin and colleagues likewise found urinary As concentrations associated with differences in aminoacyl-tRNA-biosynthesis in plasma of adult diabetics from the Chihuahua cohort (Martin et al. 2015). In an As-exposed pregnancy cohort from Durango, Mexico, Laine, and colleagues found that maternal urinary iAs% and MMA% related to newborn cord blood aminoacyl-tRNA biosynthesis. Thus, our findings based on fecal metabolites likely reflect host as well as microbial metabolism.

In humans, As undergoes methyl metabolism, and methionine is involved in producing S-adenosylmethionine, the major methyl donor. Interestingly, methionine metabolism was among the top pathways enriched in our study. Both laboratory and population-based studies report changes in DNA methylation as well as H3 lysine 9 dimethylation (Howe and Gamble 2016) in relation to As exposure, which in turn could influence gene expression and cell fate (Tollervey and Lunyak 2012). Seventeen different types of PTMs on more than 30 amino acids have been identified for human H3 alone including the acetylation (ac) and methylation (me) of lysine residues (K) (Howe et al. 2017). Rats deprived of methionine lived longer and had attenuated age-related T cell changes (Miller et al. 2005). Thus, higher fecal methionine concentrations could affect immune function and downstream gut microbiota composition. We observed few differences in our associations by primary or secondary methylation status; and not for methionine as found in a prior study of prenatal urinary As in relation to cord blood metabolites (Laine et al. 2017). Further, in our study, higher infant urinary As associated with higher fecal uracil. Arsenic trioxide was found to perturb SUMO- and folate-dependent nuclear de novo thymidylate (dTMP) biosynthesis, which can lead to misincorporation of uracil into DNA and genome instability (Kamynina et al. 2017).

We also found increases in isoleucine and leucine concentrations in relation to an infant’s urinary As concentrations. In a Caenorhabditis elegans model, response to As toxicity was in part driven by genetic variation in dbt-1 (Zdraljevic et al. 2019), which encodes the E2 subunit of the branched-chain keto acid dehydrogenase (BCKDH) complex involved in BCAA metabolism. BCAA changes following As treatment further suggested BCAA metabolism as a target of As toxicity. This is consistent with our findings that isoleucine and leucine may be altered in infants with higher As exposure, although additional mechanistic and epidemiologic studies are needed.

We found that tryptophan, which is converted by gut bacteria to indole (Lu et al. 2014) was positively associated with As concentrations, especially among those with higher PMIs; this suggests a possible reduction in gut microbial conversion of tryptophan into indole containing metabolites. Cholate is a primary bile acid synthesized in the liver from the oxidation of cholesterol. Bile acids further undergo deconjugation and dihydroxylation by gut microbes (Tian et al. 2020). In our study, the positive association with fecal cholate was stronger among those with lower SMIs. Metabolism of iAs, and accumulation of MMA, in particular, has been associated with a myriad of health outcomes (Abuawad et al. 2021). Whether fecal metabolic differences exist by As metabolic capacity (measured by urinary metabolites or genetic characterization) and whether these differences influence children’s later risk of disease merits further investigation. Thus, while preliminary, our findings align with known processes and may inform new avenues of mechanistic exploration.

In summary, our exploratory study indicates that As exposure in infants may have functional perturbations in the infant gut microbiota–host interactions consistent with our previous microbiota analysis. These perturbations could be attributed primarily to bile acid metabolism, SCFA and organic acid metabolism, amino acid metabolism, and pyrimidine metabolism. Our findings further support earlier findings that As exposure in infants affects the developing infant gut microbiota.

Supplementary Material

Supplementary Material

Funding

This study was supported by National Institute of Environmental Health Sciences (Grant No. P01ES022832), National Library of Medicine (Grant No. R01LM012723) and National Institute of Diabetes and Digestive and Kidney Diseases (Grant No. U24DK097193).

Footnotes

Declarations We declare this manuscript to be original, has not yet been published, and includes information relevant to your journal. This manuscript is not being considered for publication elsewhere. Data can be made available on reasonable request.

Supplementary Information The online version of this article (https://doi.org/10.1007/s12403-022-00468-2) contains supplementary material, which is available to authorized users.

References

  1. Abuawad A, Bozack AK, Saxena R, Gamble MV (2021) Nutrition, one-carbon metabolism and arsenic methylation. Toxicology 457:152803. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Appert O, Garcia AR, Frei R, Roduit C, Constancias F, Neuzil-Bunesova V, Ferstl R, Zhang J, Akdis C, Lauener R, Lacroix C, Schwab C (2020) Initial butyrate producers during infant gut microbiota development are endospore formers. Environ Microbiol 22(9):3909–3921 [DOI] [PubMed] [Google Scholar]
  3. Beckonert O, Keun HC, Ebbels TM, Bundy J, Holmes E, Lindon JC, Nicholson JK (2007) Metabolic profiling, metabolomic and metabonomic procedures for NMR spectroscopy of urine, plasma, serum and tissue extracts. Nat Protoc 2(11):2692–2703 [DOI] [PubMed] [Google Scholar]
  4. Bender R, Lange S (2001) Adjusting for multiple testing—when and how? J Clin Epidemiol 54(4):343–349 [DOI] [PubMed] [Google Scholar]
  5. Brim H, Lee EL, Nelson KE, Smoot DT, Sears CL, Hassanzadeh H, Pathmasiri W, Sumner SC, Ashktorab H (2012) Mol671 a comprehensive taxonomic, metagenomic and metabolomic gut flora analysis reveals distinct profiles in healthy and colon adenoma African Americans. Gastroenterology 142(5):S655 [Google Scholar]
  6. Brim H, Yooseph S, Lee E, Sherif ZA, Abbas M, Laiyemo AO, Varma S, Torralba M, Dowd SE, Nelson KE, Pathmasiri W, Sumner S, de Vos W, Liang Q, Yu J, Zoetendal E, Ashktorab H (2017) A microbiomic analysis in African Americans with colonic lesions reveals Streptococcus sp. VT162 as a marker of neoplastic transformation. Genes (Basel) 8(11):314. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Broadhurst D, Goodacre R, Reinke SN, Kuligowski J, Wilson ID, Lewis MR, Dunn WB (2018) Guidelines and considerations for the use of system suitability and quality control samples in mass spectrometry assays applied in untargeted clinical metabolomic studies. Metabolomics 14(6):72. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Burchiel SW, Lauer FT, Factor-Litvak P, Liu X, Islam T, Eunus M, Abu Horayara M, Islam MT, Rahman M, Ahmed A, Cremers S, Nanda-kumar R, Ahsan H, Olopade C, Graziano J, Parvez F (2020) Arsenic exposure associated T cell proliferation, smoking, and vitamin D in Bangladeshi men and women. PLoS ONE 15(6):e0234965. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Carignan CC, Punshon T, Karagas MR, Cottingham KL (2016) Potential exposure to arsenic from infant rice cereal. Ann Glob Health 82(1):221–224 [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Chan EC, Pasikanti KK, Nicholson JK (2011) Global urinary metabolic profiling procedures using gas chromatography-mass spectrometry. Nat Protoc 6(10): 1483–1499 [DOI] [PubMed] [Google Scholar]
  11. Chi L, Gao B, Tu P, Liu CW, Xue J, Lai Y, Ru H, Lu K (2018) Individual susceptibility to arsenic-induced diseases: the role of host genetics, nutritional status, and the gut microbiome. Mamm Genome 29(l-2):63–79 [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Coryell M, Roggenbeck BA, Walk ST (2019) The human gut microbiome’s influence on arsenic toxicity. Curr Pharmacol Rep 5(6):491–504 [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Dangleben NL, Skibola CF, Smith MT (2013) Arsenic immunotoxicity: a review. Environ Health 12(1):73. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Dietert RR, DeWitt JC, Germolec DR, Zelikoff JT (2010) Breaking patterns of environmentally influenced disease for health risk reduction: immune perspectives. Environ Health Perspect 118(8): 1091–1099 [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Dona AC, Jimenez B, Schafer H, Humpfer E, Spraul M, Lewis MR, Pearce JT, Holmes E, Lindon JC, Nicholson JK (2014) Precision high-throughput proton NMR spectroscopy of human urine, serum, and plasma for large-scale metabolic phenotyping. Anal Chem 86(19):9887–9894 [DOI] [PubMed] [Google Scholar]
  16. EFSA Panel on Contaminants in the Food Chain (2009) Scientific opinion on arsenic in food. EFSA J 7(10): 1351 [Google Scholar]
  17. Farzan SF, Li Z, Korrick SA, Spiegelman D, Enelow R, Nadeau K, Baker E, Karagas MR (2016) Infant infections and respiratory symptoms in relation to in utero arsenic exposure in a U.S. Cohort: Environ Health Perspect 124(6):840–847 [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Gao Q, Jiang J, Chu Z, Lin H, Zhou X, Liang X (2017) Arsenic trioxide inhibits tumor-induced myeloid-derived suppressor cells and enhances T-cell activity. Oncol Lett 13(4):2141–2150 [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Gao S, Lin P-I, Mostofa G, Quamruzzaman Q, Rahman M, Rahman ML, Su L, Hsueh Y-M, Weisskopf M, Coull B, Christiani DC (2019a) Determinants of arsenic methylation efficiency and urinary arsenic level in pregnant women in Bangladesh. Environ Health 18(1):94. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Gao S, Mostofa MG, Quamruzzaman Q, Rahman M, Rahman M, Su Hsueh YM, Weisskopf M, Coull B, Christiani DC (2019b) Gene–environment interaction and maternal arsenic methylation efficiency during pregnancy. Environ Int 125:43–50 [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Gardner RM, Nermell B, Kippler M, Grandér M, Li L, Ekström EC, Rahman A, Lönnerdal B, Hoque AM, Vahter M (2011) Arsenic methylation efficiency increases during the first trimester of pregnancy independent of folate status. Reprod Toxicol 31 (2):210–218 [DOI] [PubMed] [Google Scholar]
  22. Gonzalez-Garcia RA, McCubbin T, Navone L, Stowers C, Nielsen LK, Marcellin E (2017) Microbial propionic acid production. Fermentation 3(2):21 [Google Scholar]
  23. Hoen AG, Madan JC, Li Z, Coker M, Lundgren SN, Morrison HG, Palys T, Jackson BP, Sogin ML, Cottingham KL, Karagas MR (2018) Sex-specific associations of infants’ gut microbiome with arsenic exposure in a US population. Sci Rep 8(1): 12627. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Hopenhayn C, Huang B, Christian J, Peralta C, Ferreccio C, Atallah R, Kalman D (2003) Profile of urinary arsenic metabolites during pregnancy. Environ Health Perspect 111(16): 1888–1891 [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Howe CG, Gamble MV (2016) Influence of arsenic on global levels of histone posttranslational modifications: a review of the literature and challenges in the field. Curr Environ Health Rep 3(3):225–237 [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Howe CG, Liu X, Hall MN, Ilievski V, Caudill MA, Malysheva O, Lomax-Luu AM, Parvez F, Siddique AB, Shahriar H, Uddin MN, Islam T, Graziano JH, Costa M, Gamble MV (2017) Sex-specific associations between one-carbon metabolism indices and post-translational histone modifications in arsenic-exposed Bangladeshi adults. Cancer Epidemiol Biomark Prev 26(2):261–269 [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Kamynina E, Lachenauer ER, DiRisio AC, Liebenthal RP, Field MS, Stover PJ (2017) Arsenic trioxide targets MTHFD1 and SUMO-dependent nuclear de novo thymidylate biosynthesis. Proc Natl Acad Sci USA 114(12):E2319–E2326 [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Kapp RW (2018) Arsenic toxicology ☆. In: Reference module in biomedical sciences. Elsevier, Amsterdam [Google Scholar]
  29. Laine JE, Bailey KA, Olshan AF, Smeester L, Drobná Z, Stýblo M, Douillet C, García-Vargas G, Rubio-Andrade M, Pathmasiri W, McRitchie S, Sumner SJ, Fry RC (2017) Neonatal metabolomic profiles related to prenatal arsenic exposure. Environ Sci Technol 51(1):625–633 [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Laue HE, Moroishi Y, Jackson BP, Palys TJ, Madan JC, Karagas MR (2020) Nutrient-toxic element mixtures and the early postnatal gut microbiome in a United States Longitudinal Birth Cohort. Environ Int 138:105613. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Li M, Wang B, Zhang M, Rantalainen M, Wang S, Zhou H, Zhang Y, Shen J, Pang X, Zhang M, Wei H, Chen Y, Lu H, Zuo J, Su M, Qiu Y, Jia W, Xiao C, Smith LM, Yang S, Holmes E, Tang H, Zhao G, Nicholson JK, Li L, Zhao L (2008) Symbiotic gut microbes modulate human metabolic phenotypes. Proc Natl Acad Sci USA 105(6):2117–2122 [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Li X, Brejnrod AD, Ernst M, Rykaer M, Herschend J, Olsen NMC, Dorrestein PC, Rensing C, Sorensen SJ (2019) Heavy metal exposure causes changes in the metabolic health-associated gut microbiome and metabolites. Environ Int 126:454–467 [DOI] [PubMed] [Google Scholar]
  33. Liu S, Li E, Sun Z, Fu D, Duan G, Jiang M, Yu Y, Mei L, Yang P, Tang Y, Zheng P (2019) Altered gut microbiota and short-chain fatty acids in Chinese children with autism spectrum disorder. Sci Rep 9(1):287. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Livanos AE, Greiner TU, Vangay P, Pathmasiri W, Stewart D, McRitchie S, Li H, Chung J, Sohn J, Kim S, Gao Z, Barber C, Kim J, Ng S, Rogers AB, Sumner S, Zhang XS, Cadwell K, Knights D, Alekseyenko A, Backhed F, Blaser MJ (2016) Antibiotic-mediated gut microbiome perturbation accelerates development of type 1 diabetes in mice. Nat Microbiol 1(11): 16140. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Louis P, Flint HJ (2017) Formation of propionate and butyrate by the human colonic microbiota. Environ Microbiol 19(1):29–41 [DOI] [PubMed] [Google Scholar]
  36. Lu K, Abo RP, Schlieper KA, Graffam ME, Levine S, Wishnok JS, Swenberg JA, Tannenbaum SR, Fox JG (2014) Arsenic exposure perturbs the gut microbiome and its metabolic profile in mice: an integrated metagenomics and metabolomics analysis. Environ Health Perspect 122(3):284–291 [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Madan JC, Farzan SF, Hibberd PL, Karagas MR (2012) Normal neonatal microbiome variation in relation to environmental factors, infection and allergy. Curr Opin Pediatr 24(6):753–759 [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Martin E, Gonzalez-Horta C, Rager J, Bailey KA, Sanchez-Ramirez B, Ballinas-Casarrubias L, Ishida MC, Gutierrez-Torres DS, Hernandez Ceron R, Viniegra Morales D, Baeza Terrazas FA, Saunders RJ, Drobna Z, Mendez MA, Buse JB, Loomis D, Jia W, Garcia-Vargas GG, Del Razo LM, Styblo M, Fry R (2015) Metabolomic characteristics of arsenic-associated diabetes in a prospective cohort in Chihuahua, Mexico. Toxicol Sci 144(2):338–346 [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Masson P, Spagou K, Nicholson JK, Want EJ (2011) Technical and biological variation in UPLC–MS-based untargeted metabolic profiling of liver extracts: application in an experimental toxicity study on galactosamine. Anal Chem 83(3): 1116–1123 [DOI] [PubMed] [Google Scholar]
  40. McDermott TR, Stolz JF, Oremland RS (2020) Arsenic and the gastrointestinal tract microbiome. Environ Microbiol Rep 12(2): 136–159 [DOI] [PubMed] [Google Scholar]
  41. Miller RA, Buehner G, Chang Y, Harper JM, Sigler S, Smith-Wheelock M (2005) Methionine-deficient diet extends mouse lifespan, slows immune and lens aging, alters glucose, T4, IGF-I and insulin levels, and increases hepatocyte MIF levels and stress resistance. Aging Cell 4(3): 119–125 [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Nicholson JK, Holmes E, Kinross J, Burcelin R, Gibson G, Jia W, Pettersson S (2012) Host-gut microbiota metabolic interactions. Science 336(6086): 1262–1267 [DOI] [PubMed] [Google Scholar]
  43. Nygaard UC, Li Z, Palys T, Jackson B, Subbiah M, Malipatlolla M, Sampath V, Maecker H, Karagas MR, Nadeau KC (2017) Cord blood T cell subpopulations and associations with maternal cadmium and arsenic exposures. PLoS ONE 12(6):e0179606. [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Roduit C, Frei R, Ferstl R, Loeliger S, Westermann P, Rhyner C, Schiavi E, Barcik W, Rodriguez-Perez N, Wawrzyniak M, Chassard C, Lacroix C, Schmausser-Hechfellner E, Depner M, von Mutius E, Braun-Fahrlander C, Karvonen AM, Kitjavainen PV, Pekkanen J, Dalphin JC, Riedler J, Akdis C, Lauener R, O’Mahony L, PASTURE/EFRAIM Study Group (2019) High levels of butyrate and propionate in early life are associated with protection against atopy. Allergy 74(4):799–809 [DOI] [PubMed] [Google Scholar]
  45. Shen H, Niu Q, Xu M, Rui D, Xu S, Feng G, Ding Y, Li S, Jing M (2016) Factors affecting arsenic methylation in arsenic-exposed humans: a systematic review and meta-analysis. Int J Environ Res Public Health 13(2):205. [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Signes-Pastor AJ, Cottingham KL, Carey M, Sayarath V, Palys T, Meharg AA, Folt CL, Karagas MR (2018) Infants’ dietary arsenic exposure during transition to solid food. Sci Rep 8(1):7114. [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Silva YP, Bernardi A, Frozza RL (2020) The role of short-chain fatty acids from gut microbiota in gut-brain communication. Front Endocrinol 11:25. [DOI] [PMC free article] [PubMed] [Google Scholar]
  48. Smith PA (2015) The tantalizing links between gut microbes and the brain. Nature 526(7573):312–314 [DOI] [PubMed] [Google Scholar]
  49. Suvorova IA, Ravcheev DA, Gelfand MS (2012) Regulation and evolution of malonate and propionate catabolism in proteobacteria. J Bacteriol 194(12):3234–3240 [DOI] [PMC free article] [PubMed] [Google Scholar]
  50. Tian Y, Gui W, Koo I, Smith PB, Allman EL, Nichols RG, Rimal B, Cai J, Liu Q, Patterson AD (2020) The microbiome modulating activity of bile acids. Gut Microbes 11(4):979–996 [DOI] [PMC free article] [PubMed] [Google Scholar]
  51. Tollervey JR, Lunyak VV (2012) Epigenetics. Epigenetics 7(8): 823–840 [DOI] [PMC free article] [PubMed] [Google Scholar]
  52. Tseng CH (2009) A review on environmental factors regulating arsenic methylation in humans. Toxicol Appl Pharmacol 235(3):338–350 [DOI] [PubMed] [Google Scholar]
  53. Weljie AM, Newton J, Mercier P, Carlson E, Slupsky CM (2006) Targeted profiling: quantitative analysis of 1H NMR metabolomics data. Anal Chem 78(13):4430–4442 [DOI] [PubMed] [Google Scholar]
  54. Xi B, Gu H, Baniasadi H, Raftery D (2014) Statistical analysis and modeling of mass spectrometry-based metabolomics data. Methods Mol Biol (Clifton NJ) 1198:333–353 [DOI] [PMC free article] [PubMed] [Google Scholar]
  55. Zdraljevic S, Fox BW, Strand C, Panda O, Tenjo FJ, Brady SC, Crombie TA, Doench JG, Schroeder FC, Andersen EC (2019) Natural variation in C. elegans arsenic toxicity is explained by differences in branched-chain amino acid metabolism. eLife 8:e40260. [DOI] [PMC free article] [PubMed] [Google Scholar]
  56. Zheng X, Xie G, Zhao A, Zhao L, Yao C, Chiu NH, Zhou Z, Bao Y, Jia W, Nicholson JK, Jia W (2011) The footprints of gut microbial-mammalian co-metabolism. J Proteome Res 10(12):5512–5522 [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary Material
NIHMS1818701-supplement-Supplementary_Material.docx (34.8KB, docx)

RESOURCES