Research Studies
Study Tracker
Chronic arsenic-fluoride co-exposure impairs spermatogenesis with gut microbiota-associated bile acid and lipid metabolic remodeling.
| Environment International | Chen J, Cui Y, Wang Y, Kong R, Liu P, Su Q, Li Y, Guo H, Fu Y, Zhao Q, Li B, Tian X, Yan X. | 216:110520.
Arsenic Chemical Co-Exposures Reproductive Toxicity Sperm Effects Testes Urine-F Animal Study
View PDF

Abstract

Original full-text study online at
https://www.sciencedirect.com/science/article/pii/S0160412026004782?via%3Dihub

Arsenic and fluoride frequently co-occur in groundwater, but the gut-associated metabolic changes accompanying male reproductive toxicity from their co-exposure remain unclear. We established a chronic developmental drinking water co-exposure model in Sprague–Dawley rats using 100 mg/L NaF and 75 mg/L NaAsO2. Reproductive outcomes were evaluated at postnatal day (PND) 60 and PND120, and reciprocal fecal transfer was performed between control and arsenic–fluoride (AsF) backgrounds. Developmental AsF exposure impaired sperm production and quality, disrupted reproductive hormone homeostasis, altered spermatogenic-stage markers, and induced testicular and epididymal injury. Repeated administration of AsF-donor fecal material to control recipients was accompanied by selected adverse reproductive changes, whereas control-donor material was accompanied by partial changes in the opposite direction in AsF-exposed recipients. Endpoint 16S rRNA sequencing demonstrated exposure-associated microbial community differences, and donor-material administration was accompanied by changes in selected host phenotypes. Untargeted serum metabolomics revealed prominent remodeling of bile acid- and lipid-related signals, including reciprocal Fecal microbiota transplantation-associated patterns and exploratory relationships between bile-acid-related features and reproductive hormones. These metabolic findings were accompanied by differences in colonic junction- and inflammation-related markers, intestinal–hepatic FXR–FGF15-related proteins, hepatic bile acid and lipid endpoints, and testicular autophagy-related markers. Collectively, the findings characterize reproductive hazards at the tested concentrations and support a nonexclusive gut–liver–testis framework in which donor-material-associated microbial and metabolic changes accompany partial modification of recipient phenotypes.

EXCERPTS
1. Introduction

Inorganic arsenic (As) and fluoride (F) are widespread groundwater contaminants whose occurrence and co-enrichment are governed by aquifer mineralogy, pH, redox conditions, and water–rock interactions. Global models indicate that large populations may be exposed to elevated concentrations of either contaminant, while hydrogeochemical investigations have documented As–F co-occurrence in endemic regions and groundwater basins in northern China (He et al., 2020, Kumar et al., 2020, Podgorski and Berg, 2020, Podgorski and Berg, 2022; Z. Wang et al., 2024a, Cao et al., 2024). Mining, coal combustion, industrial discharge, phosphate-related activities, and other human activities may further release or redistribute these elements in soil–water systems (Patel et al., 2023, Shaji et al., 2024). The World Health Organization guideline values for drinking water are 0.01 mg/L for As and 1.5 mg/L for F (World Health Organization, 2022). Their environmental co-occurrence creates a multiple-contaminant context in which biological responses cannot be inferred solely from single-contaminant studies (Mondal and Chattopadhyay, 2020). Therefore, elucidating the health effects and biological pathways associated with chronic As–F co-exposure is important for environmental hazard assessment and evidence-based risk management in co-contaminated regions.

Male reproductive function is sensitive to environmental and metabolic stress because it depends on coordinated germ-cell differentiation, blood–testis barrier integrity, steroidogenesis, and energy metabolism (Rato et al., 2012, Eisenberg et al., 2023, Kimmins et al., 2024). Environmental pollutants have been implicated in altered semen quality and reproductive endocrine function, although the strength of evidence varies across contaminants and study designs (Skakkebæk et al., 2022, Xu et al., 2022). Experimental studies indicate that arsenic can disturb reproductive hormones, damage testicular structure and the blood–testis barrier, and impair spermatogenesis and sperm maturation (Machado-Neves, 2022, Mahboob et al., 2025; F.-W. Niu et al., 2025, Song et al., 2026). Fluoride exposure has likewise been associated in animal models with oxidative injury, altered steroidogenesis, germ-cell damage, and reduced sperm quality (Radovanovi? et al., 2022, Li et al., 2023, Ommati et al., 2025, Talebi et al., 2025). Evidence concerning combined As–F reproductive toxicity is more limited (Mondal and Chattopadhyay, 2020). Our previous study showed that developmental As–F co-exposure was accompanied by testicular injury, changes in autophagy-related markers, and disruption of intestinal microbial balance (Liu et al., 2021). However, whether exposure-associated intestinal and metabolic changes contribute to the reproductive phenotype, and how joint effects vary among endpoints, remain insufficiently defined.

The gut microbiota may modify host responses to environmental chemicals by influencing xenobiotic biotransformation, intestinal barrier function, immune and endocrine signaling, and systemic metabolism (Claus et al., 2016, Licht and Bahl, 2019, Fan and Pedersen, 2021). Arsenic illustrates this bidirectional relationship: gut microorganisms participate in redox and methylation reactions that can modify arsenic metabolism and excretion, whereas arsenic exposure can alter microbial composition and metabolic output (Lu et al., 2014, Coryell et al., 2019, Yang et al., 2021; L. Chen et al., 2023). Fluoride exposure has also been associated with dose-dependent changes in microbial community structure and predicted function, although findings vary according to dose, duration, and host species (Mo et al., 2023, Yasin et al., 2025). These interactions may be especially relevant when exposure spans preconception, gestation, lactation, puberty, and adulthood because early-life microbial assembly and the reproductive endocrine system both undergo substantial developmental remodeling (Qi et al., 2021, Santos-Marcos et al., 2023; S. Wang et al., 2024).

Growing evidence suggests that the gut microbial ecosystem may influence male reproductive function through metabolic, immune, and endocrine pathways relevant to spermatogenesis (Ding et al., 2020; H. Cai et al., 2022; M. Wang et al., 2024a, Wang et al., 2024b, Wang et al., 2024c; W. Chen et al., 2024). Experimental studies have further linked microbial disturbances to coordinated intestinal and testicular responses, while fecal transfer has been shown to modify reproductive phenotypes and systemic and testicular metabolic profiles (Hao et al., 2022; S. Niu et al., 2026). These observations support the gut microbiota as a potential contributor to environmentally induced reproductive injury.

Reciprocal fecal transfer between exposed and control animals provides an experimental approach for testing whether donor-derived fecal material can modify recipient phenotypes and therefore offers stronger evidence than cross-sectional microbial associations alone. Nevertheless, fecal transfer does not by itself establish stable microbial engraftment or microbiota-specific causation, because recipient conditioning, microbial viability, transferred microbial products, soluble metabolites, and chemical carryover may influence the observed effects (Gheorghe et al., 2021). Integrating reciprocal transfer with microbial, intestinal, metabolic, and reproductive assessments may therefore provide a broader basis for investigating gut-associated contributions to toxicant-induced male reproductive injury.

Bile acids provide a biologically plausible interface among intestinal microorganisms, hepatic metabolism, and distal-organ physiology. Primary bile acids synthesized in the liver are deconjugated and further transformed by intestinal microorganisms, generating a bile acid pool shaped jointly by host and microbial processes (Winston and Theriot, 2020, Guzior and Quinn, 2021, Collins et al., 2023). Beyond facilitating lipid absorption, bile acids signal through receptors including the farnesoid X receptor (FXR) and G-protein-coupled bile acid receptor 1. Ileal FXR induces fibroblast growth factor 15 (FGF15), which signals to the liver to regulate bile acid synthesis and broader metabolic processes (Inagaki et al., 2005, Chiang and Ferrell, 2020, Collins et al., 2023). Lipid uptake, synthesis, and remodeling are integral to germ-cell development and steroid hormone production (Miller and Auchus, 2011, Rato et al., 2012), and experimental studies indicate that bile acid receptors can influence testicular physiology and male reproductive function (Baptissart et al., 2014, Martinot et al., 2017). Microbiota-associated bile acid and lipid remodeling could therefore accompany coordinated intestinal, hepatic, and testicular responses.

Autophagy is a highly conserved cellular quality-control process that contributes to organelle turnover, steroidogenesis, germ-cell development, and testicular homeostasis (Gao et al., 2018, Samare-Najaf et al., 2023). Previous studies of combined As–F exposure and fluoride exposure alone have documented changes in testicular autophagy-related proteins and autophagosome-like structures (Liu et al., 2021, Li et al., 2023, Talebi et al., 2025). Because static protein and ultrastructural measurements cannot determine autophagic-flux direction or functional consequence, such findings are appropriately interpreted as accompanying autophagy-related alterations rather than evidence of pathway activation or autophagy-mediated injury (Klionsky et al., 2021).

In this study, we used a relatively high-dose chronic drinking-water exposure model to characterize the reproductive hazards of combined As–F exposure. Sprague–Dawley rats were exposed to sodium arsenite and/or sodium fluoride from parental pre-mating through gestation and lactation, with male offspring continuing their assigned exposures until postnatal day (PND) 60 or PND120. The direct-exposure experiment assessed duration-associated reproductive changes, whereas reciprocal fecal transfer between control and combined-exposure backgrounds was used to evaluate FMT-associated differences in recipient phenotypes under their original drinking-water regimens. Reproductive outcomes were evaluated alongside intestinal, microbial, serum metabolomic, hepatic, and testicular endpoints. Untargeted metabolomic data were further interrogated to identify putatively annotated bile acid signals, while static autophagy-related endpoints were used to characterize accompanying testicular alterations. By integrating these assessments, this study provides a multilevel characterization of reproductive toxicity at the tested concentrations and the associated intestinal, microbial, metabolic, hepatic, and testicular responses.

2.2. Animals, exposure, and experimental allocation

Specific-pathogen-free Sprague–Dawley rats were housed under controlled environmental conditions with food and water provided ad libitum. All experimental procedures were approved by the Institutional Animal Ethics Committee of Shanxi Medical University.

Twenty-four female and 12 male rats were allocated to four drinking-water groups: control water (Con), 75 mg/L NaAsO2 (As), 100 mg/L NaF (F), or the combination of 75 mg/L NaAsO2 and 100 mg/L NaF (AsF). Parental exposure began 10 days before mating. Females and males were paired at a ratio of 2:1, and exposure continued throughout gestation and lactation. After weaning, male offspring remained on the drinking-water regimens assigned to their parental groups until terminal sampling at PND60 or PND120 (n=8 per group).

The exposure concentrations were selected based on our previous experimental model and related reproductive-toxicity literature (Liu et al., 2021, Mahboob et al., 2025, Ommati et al., 2025). Group-level water and food consumption was recorded weekly.

2.3. Reciprocal fecal microbiota transplantation (FMT)

Rats from the PND120 Con and AsF groups served as the respective donor sources. Sixteen male offspring maintained on control water were randomized to receive sterile saline (Con-Sal, n = 8) or fecal suspension derived from AsF donors (Con-AsF, n = 8). Sixteen male offspring maintained under combined AsF exposure were randomized to receive sterile saline (AsF-Sal, n = 8) or fecal suspension derived from control donors (AsF-Con, n = 8). All 32 recipients completed the experimental protocol and remained on their original drinking-water regimens throughout the fecal-transfer period.

On each administration day, freshly voided donor feces were collected and homogenized in approximately five volumes of sterile physiological saline. The resulting suspension was filtered to remove particulate material and centrifuged at 1,200 rpm for 3 min. After removal of the supernatant, the pellet was resuspended in sterile physiological saline to the original suspension volume. The centrifugation–resuspension washing procedure was repeated three times, after which the final pellet was resuspended to the original volume.

Freshly prepared fecal suspension or sterile physiological saline was administered by enema at 1 mL/100 g body weight twice weekly for 7 weeks, for a total of 15 administrations per recipient. Administration ended at PND110, followed by a 10-day post-administration interval before terminal sampling at PND120.

2.4. Exposure biomarker

Urinary As concentrations were measured by atomic fluorescence spectrometry according to the Chinese health industry standard WS/T 474-2015, Determination of arsenic in urine—Hydride generation atomic fluorescence spectrometry, after mixed-acid digestion and prereduction with thiourea and ascorbic acid. Bone F concentrations were measured using a fluoride ion-selective electrode, with the procedure adapted from WS/T 89-2015, Determination of fluoride in urine—Ion-selective electrode method. Bone samples were dried, ashed, acid-extracted, pH-adjusted, and mixed with total ionic-strength adjustment buffer before analysis. Procedural blanks and calibration standards were included, and four animals per group were analyzed.

2.15 Statistical analysis

Data are presented as the mean ± standard deviation (SD). Statistical analyses were primarily performed using R software (version 4.5.2). Comparisons between two independent groups were conducted using Student’s t-test or the Wilcoxon rank-sum test, depending on the distributional characteristics of the data. Multiple-group comparisons were performed using one-way analysis of variance (ANOVA), followed by Tukey’s multiple-comparison test. Prespecified reciprocal FMT comparisons were conducted within each recipient background, with multiplicity adjustment applied where indicated.

For the PND120 Con, As, F, and AsF groups, each endpoint with verified individual-animal data was additionally analyzed using a 2 × 2 factorial model, with arsenic and fluoride exposure entered as binary factors. Type III tests with sum-to-zero contrasts were used to estimate the main effects of arsenic and fluoride and their interaction. The arsenic × fluoride interaction was evaluated before interpreting the main effects. When evidence of interaction was detected, interpretation focused on model-estimated simple effects and contrasts rather than the marginal main effects. The direction of interaction was determined from these estimated contrasts rather than by visual inspection of group means. Heteroskedasticity-consistent type 3 (HC3) standard errors were used in sensitivity analyses (Long and Ervin, 2000).

Western blot band intensities were quantified using ImageJ software (2015 version; National Institutes of Health, Bethesda, MD, USA), and statistical graphs were generated using GraphPad Prism 9.5.0 (GraphPad Software, San Diego, CA, USA). Exact sample sizes, statistical tests, post hoc procedures, and multiplicity adjustments for individual assays are reported in the corresponding figure legends and supplementary tables. All tests were two-sided, and P < 0.05 was considered statistically significant.

3. Results

3.1. Chronic arsenic–fluoride exposure impairs spermatogenesis, with reproductive phenotypes differing following FMT

The experimental design and exposure timeline are shown in Fig. 1a. Weekly food- and water-intake trajectories were summarized descriptively in Fig. S1 and Table S1. At PND120, mean water consumption was 40.88, 19.64, 41.20, and 22.69 mL/rat/day in the Con, As, F, and AsF groups, respectively. Based on nominal elemental concentrations and group-level water-consumption summaries, the estimated daily arsenic administration was 0.849 and 0.981 mg/rat/day in the As and AsF groups, respectively, whereas the estimated fluoride administration was 1.864 and 1.027 mg/rat/day in the F and AsF groups, respectively…

3.2. Direct exposure alters gut microbial community structure, whereas endpoint donor-associated FMT metrics show limited convergence

Bray–Curtis ordinations were analyzed separately for the direct-exposure and FMT recipient models (Fig. 3a–b). In the direct-exposure model, microbial community composition differed significantly among Con60, AsF60, Con120, and AsF120 groups (PERMANOVA R2=0.336, P<0.001), without evidence of unequal multivariate dispersion (PERMDISP P = 0.270). The prespecified Con60 versus AsF60 and Con120 versus AsF120 comparisons were both significant after multiple-testing correction (adjusted P =0.0046 and 0.0138, respectively)…

3.3. Colonic junction- and inflammation-related markers

PND120 intestinal immunofluorescence showed reciprocal marker changes across the two recipient backgrounds (Fig. 4). Administration of AsF-donor fecal material to control recipients was accompanied by reduced ZO-1 and Occludin staining and increased TNF-? expression. Administration of control-donor material to AsF-exposed recipients was accompanied by changes in the opposite direction. Both prespecified comparisons remained significant after Holm correction for ZO-1, Occludin, and TNF-? (adjusted P = 0.002, 0.008, and 0.013, respectively)…

3.4. Untargeted serum metabolomics identifies exposure- and FMT-associated metabolic remodeling

Untargeted serum metabolomics showed broad metabolic differences following arsenic–fluoride exposure and reciprocal FMT. The two prespecified PLS-DA score plots separated the Con120, AsF120, Con-Sal, and Con-AsF groups in Model A and the Con120, AsF120, AsF-Sal, and AsF-Con groups in Model B (Fig. 5a–b)…

3.5. Bile-acid-related features show reciprocal FMT-associated patterns and structured endocrine relationships

A bile-acid-focused analysis of the untargeted metabolomics dataset identified relative bile-acid-related features that differed across the AsF120 versus Con120, Con-AsF versus Con-Sal, and AsF-Con versus AsF-Sal comparisons (Fig. 6a–c). Among the predefined 12-feature panel, TUDCA-, glycodeoxycholic acid-, 7-ketodeoxycholic acid-, GCDCA-3G-, and 7-SCA-related signals showed significant overall group differences after multiple-testing correction (Fig. 6d)…

3.6. Reciprocal FMT is associated with enterohepatic signaling and hepatic metabolic responses at PND120

PND120 liver histology and biochemical measurements differed across the FMT groups (Fig. 7a). Con-Sal recipients showed relatively preserved hepatic architecture, whereas Con-AsF recipients displayed hepatocellular disorganization, lipid vacuolation, and localized inflammatory-cell infiltration. These alterations were less pronounced in AsF-Con than in AsF-Sal recipients…

3.7. Reciprocal FMT is associated with testicular autophagy-related ultrastructure and protein expression

PND120 testicular TEM showed group-dependent ultrastructural changes in Leydig cells, Sertoli cells, spermatogonia, and spermatocytes (Fig. 8a). Con-Sal recipients showed relatively preserved cellular morphology without prominent autophagosome-like structures. In Con-AsF recipients, autophagosome-like structures were observed in Leydig and Sertoli cells. In AsF-Sal recipients, these structures were most apparent in Leydig cells and were also observed in Sertoli cells and spermatogonia. Their distribution and abundance differed after administration of control-donor fecal material to AsF-exposed recipients. Direct-exposure ultrastructural images at PND60 and PND120 are presented in Fig. S16…

3.8. Exploratory toxicogenomic analysis identifies overlapping cellular-stress pathways

The original database-derived analysis was retained in Fig. S17. UpSet analysis identified 51 genes shared among seven male reproductive disease-related categories: azoospermia, hypogonadism, male infertility, spermatogenesis, spermatogenic failure, teratozoospermia, and testicular diseases. Intersection with arsenic- and fluoride-associated genes from the Comparative Toxicogenomics Database identified 64 genes shared across the disease, arsenic, and fluoride domains.

GO and KEGG enrichment analyses identified biological processes and pathways involving oxidative stress, reactive-oxygen-species responses, inflammation, apoptosis regulation, lipid and atherosclerosis, AGE–RAGE signaling, TNF signaling, p53 signaling, cellular senescence, NF-?B signaling, FoxO signaling, mTOR signaling, and autophagy-related pathways.

Protein–protein interaction, MCODE module, and cytoHubba analyses identified highly connected candidates including IL1B, CASP3, TP53, AKT1, PTEN, BCL2, NFE2L2, RELA, MAPK3, CASP8, CASP9, CDKN1A, CTNNB1, CCND1, and ANXA5 (Fig. S17).

4. Discussion

The present study characterized reproductive and multi-organ responses to chronic, relatively high-dose As–F co-exposure across developmental and postnatal periods. At the tested concentrations, exposure was associated with impaired sperm quality, altered reproductive hormones and organ coefficients, testicular and epididymal lesions, changes in spermatogenic-stage markers, intestinal and microbial alterations, systemic metabolic remodeling, hepatic responses, and testicular autophagy-related changes. Factorial analyses further showed that the joint effects of arsenic and fluoride differed among endpoints and were frequently less than additive on the analyzed scale rather than uniformly synergistic. Repeated administration of donor-derived fecal material was accompanied by partially reciprocal differences in selected reproductive, colonic, microbial, metabolic, hepatic, and testicular endpoints. Collectively, these findings support coordinated responses across the gut, liver, and testis, but do not establish a single serial pathway or exclude direct toxicant effects and complementary gut–testis, endocrine, inflammatory, or liver–testis routes.

The reproductive findings are consistent with previous experimental evidence that arsenic and fluoride can interfere with spermatogenesis, steroidogenesis, sperm maturation, redox homeostasis, and testicular structure (Machado-Neves, 2022, Li et al., 2023; F.-W. Niu et al., 2025, Talebi et al., 2025, Song et al., 2026). The exposure design extended from parental premating through gestation and lactation and then continued in male offspring to PND60 or PND120, covering several sensitive periods of reproductive development. The observed changes in sperm density, motility, abnormality, reproductive hormones, and testicular and epididymal morphology therefore reflect the combined influence of developmental exposure history and continued postnatal exposure. Comparisons between PND60 and PND120 broaden the temporal characterization of the phenotype, although these terminal cohorts do not independently separate chronological age from cumulative exposure duration.

The ACRV1 and SYCP3 findings add stage-associated information to the conventional sperm and histopathological endpoints. ACRV1, also known as SP-10, is an acrosomal protein predominantly expressed during the round-spermatid stage and is therefore a marker of postmeiotic spermatid differentiation (Reddi, 2023). In contrast, SYCP3 is a core component of the chromosome axial and lateral elements of the synaptonemal complex and is associated with spermatocytes undergoing meiotic prophase (Guo et al., 2025, Yang and Wang, 2026). The lower ACRV1- and SYCP3-positive areas observed in exposed animals are therefore consistent with disturbances involving postmeiotic spermatids and meiotic spermatocytes, respectively. Partially reciprocal descriptive patterns were also observed among FMT recipients. Nevertheless, positive-area measurements cannot distinguish altered protein abundance from changes in the number or relative representation of marker-positive germ cells, do not constitute stereological quantification of germ-cell populations, and do not evaluate functional acrosome integrity or sperm DNA fragmentation. Because mating success, pregnancy rate, and litter outcomes were not assessed, the reproductive conclusions should remain restricted to impaired spermatogenesis, sperm quality, and testicular injury rather than infertility.

The endocrine findings similarly require context-dependent interpretation. Altered serum and intratesticular testosterone, estradiol, FSH, and LH indicate disruption of reproductive endocrine homeostasis, which may involve Leydig-cell steroidogenesis, substrate availability, hepatic steroid handling, and hypothalamic–pituitary–gonadal feedback (Miller and Auchus, 2011, Rato et al., 2012, Santos-Marcos et al., 2023). However, estradiol did not respond in the same direction in the direct-exposure and reciprocal FMT experiments. This divergence argues against treating estradiol as a linear marker of microbiota-associated reproductive injury or a simple indicator of damage and rescue. Exposure history, aromatization, enterohepatic recycling, microbial deconjugation, hepatic clearance, and endocrine feedback could all contribute, but aromatase activity, estrogen conjugates, microbial ?-glucuronidase activity, hepatic steroid-metabolizing enzymes, and longitudinal hormone profiles were not measured (Ridlon and Bajaj, 2015, Santos-Marcos et al., 2023). The hormone data therefore indicate context-dependent endocrine modulation without identifying the responsible regulatory level.

The microbial findings provide evidence of exposure- and FMT-associated community restructuring. The significant overall PERMANOVA results, together with nonsignificant PERMDISP tests, indicate group-associated differences in community composition that were not explained solely by unequal multivariate dispersion. In contrast, none of the alpha-diversity indices remained significant after adjustment, suggesting that compositional restructuring was more apparent than changes in within-sample richness or evenness. These results are biologically plausible because the gut microbiota can influence xenobiotic biotransformation, intestinal homeostasis, immune signaling, bile acid transformation, and host metabolism (Fan and Pedersen, 2021, Collins et al., 2023). The reciprocal changes in colonic ZO-1, Occludin, and TNF-? staining further show that donor-material administration was accompanied by differences in the colonic junction- and inflammation-related milieu. However, these tissue markers are not direct measurements of intestinal permeability and do not demonstrate systemic or testicular inflammation.

The reciprocal FMT design provides useful but qualified evidence concerning an intestinal contribution to the phenotype. Within each recipient background, animals remained on the same assigned drinking-water regimen, allowing saline- and donor-material-treated groups to be compared without changing the underlying exposure assignment. Directionally reciprocal differences across selected spermatogenic, endocrine, colonic, metabolic, and hepatic endpoints strengthen the internal coherence of the experiment. However, within-background microbial comparisons were not significant after adjustment, recipient communities remained distinguishable from the reciprocal donor references, and recipient-to-donor distance, donor proximity, reciprocal donor-associated ASV abundance, and detected-ASV metrics did not demonstrate stable donor-community convergence. The observed changes are therefore best interpreted as FMT-associated modification of selected recipient phenotypes rather than proof of stable engraftment or microbiota-specific causation.

Nonmicrobial transfer also remains a relevant alternative explanation. The original inocula could not be retrospectively quantified for arsenic, fluoride, or soluble metabolites, and transient transfer cannot be excluded, particularly in the Con-AsF comparison. The absence of terminal differences in urinary arsenic or bone fluoride within the corresponding recipient backgrounds reduces concern about persistent separation in terminal internal burden but does not exclude transient exposure earlier during the seven-week administration period. In addition, recipients were sampled 10 days after the final administration, so the onset, stability, and longer-term persistence of the observed changes cannot be determined. The findings are compatible with partial involvement of the intestinal microbial ecosystem or other transferred fecal components, but they do not establish complete phenotypic rescue or reversibility of the original exposure effects (Gheorghe et al., 2021).

Untargeted metabolomics showed that reciprocal donor-material administration was accompanied by extensive systemic metabolic remodeling, with lipid-related features prominent among the annotated differential signals. However, the metabolite-class plots describe the composition of retained features rather than the total abundance or activity of an entire lipid class. Cross-validated discrimination was also stronger in the AsF-background model than in the control-background model, indicating that the robustness of metabolomic separation differed between comparisons. The PLS-DA, VIP, and pathway-enrichment results should therefore be used to prioritize candidate metabolic processes rather than as independent evidence of pathway activation.

Bile acids provide a biologically plausible interface between microbial and host metabolism because intestinal microorganisms mediate bile acid deconjugation, oxidation, epimerization, and secondary transformation, while bile acids reciprocally regulate microbial ecology and host signaling (Guzior and Quinn, 2021, Collins et al., 2023, Wise and Cummings, 2023). TUDCA and GCDCA-3G were among the bile-acid-related features showing reciprocal FMT-associated patterns, whereas 7-SCA showed recipient-group differences without a marked Con120–AsF120 difference. This distinction indicates that the direct-exposure and FMT contrasts are not interchangeable and that recipient background may shape the detected metabolic response. TUDCA, GUDCA, GCDCA-3G, and 7-SCA also appeared as bridge features in the exploratory genus–bile-acid–hormone network.

These bile acid results require several analytical qualifications. They were obtained through a post hoc, bile-acid-focused interrogation of untargeted LC–MS/MS data, were expressed as relative signal intensities, and were not confirmed using authentic standards or a targeted quantitative assay. The displayed names therefore represent putative annotations rather than definitive molecular identifications. Moreover, genus–bile-acid associations were correlation-based, and bile-acid–hormone relationships used group-level summaries rather than animal-level paired observations. These analyses identify coordinated variation and candidate relationships but cannot establish directionality or mediation.

The intestinal–hepatic FXR–FGF15 pathway offers a plausible physiological context for the coordinated bile acid and lipid findings. Intestinal FXR-induced FGF15 is an established endocrine signal that regulates hepatic CYP7A1-dependent bile acid synthesis, while hepatic FXR-, FGFR4-, SHP-, CYP8B1-, BSEP-, PPAR-, and PGC-1?-related processes participate in bile acid transport and lipid homeostasis (Inagaki et al., 2005, Chiang and Ferrell, 2020). The observed differences in intestinal and hepatic proteins, liver histology, serum total bile acids, and hepatic lipid indices are therefore compatible with altered enterohepatic and hepatic metabolic regulation. Experimental studies also indicate that bile acid receptors can influence testicular steroidogenesis, spermatogenesis, and male fertility, and other FMT or pollutant models have reported coordinated hepatic, circulating, and testicular metabolic changes (Baptissart et al., 2014, Martinot et al., 2017, Hao et al., 2022, Yang et al., 2026).

The present data, however, do not demonstrate that hepatic changes mediated the testicular phenotype. Targeted bile acids and related metabolites were not measured simultaneously in portal blood, systemic circulation, liver, and testis, and no isotope tracing, bile acid supplementation, or tissue-specific FXR/TGR5 intervention was performed. The observed multi-organ coordination is therefore consistent with several nonexclusive routes: gut-derived metabolites or inflammatory signals may act directly on the testis; hepatic processing may modify circulating metabolic signals; endocrine changes may influence both liver and testis; and arsenic and fluoride may act directly on either organ. The gut–liver–testis concept should consequently be presented as an integrative framework supported by coordinated compartmental changes, rather than as a demonstrated unidirectional gut-to-liver-to-testis mechanism.

Lipid remodeling may be relevant to reproductive injury because cholesterol availability, mitochondrial lipid handling, and fatty acid oxidation contribute to Leydig-cell steroidogenesis, Sertoli-cell support, germ-cell development, spermiogenesis, and sperm membrane maturation (Miller and Auchus, 2011, Rato et al., 2012, Gao et al., 2018). Phosphatidylethanolamine, which contributed to the autophagy-related pathway enrichment, also participates in LC3 lipidation during autophagosome formation (Tanida et al., 2004). Nevertheless, pathway enrichment driven by an annotated circulating lipid cannot demonstrate LC3 lipidation or autophagic activity in testicular cells. It provides biological plausibility for a relationship between lipid remodeling and autophagy-related processes but not a mechanistic bridge.

The testicular TEM and Western blot findings should therefore be interpreted as accompanying autophagy-related alterations. Static measurements of LC3-II/LC3-I, p62, Beclin1, ATG5, and ATG7, together with autophagosome-like structures, cannot distinguish increased autophagosome formation from impaired lysosomal degradation or determine whether the response is protective, injury-promoting, or secondary to metabolic stress (Klionsky et al., 2021). Canonical apoptotic markers and testicular inflammatory cytokines were not measured, so apoptotic and inflammatory injury remain possible parallel or interacting mechanisms. The colonic cytokine measurements do not provide evidence of testicular inflammation. Likewise, the database-derived human gene analysis identifies convergence on oxidative-stress, inflammatory, apoptotic, lipid, and autophagy-related pathways but remains exploratory public-data context rather than validation in the experimental rats or in human biological samples.

The study has several strengths, including an exposure design spanning developmental and postnatal periods, factorial analysis of verified PND120 endpoints, reciprocal within-background donor-material comparisons, donor-reference analyses in a common ASV space, animal-level treatment of histological measurements, and integration of reproductive, intestinal, microbial, metabolic, hepatic, and testicular observations. Its main environmental limitation is the exposure intensity. The nominal 75 mg/L NaAsO? and 100 mg/L NaF exposure concentrations were above the corresponding WHO drinking-water guideline values; accordingly, this model is best interpreted as an experimental hazard model rather than a direct simulation of environmentally representative human exposure. Group-level water-consumption data and terminal urinary arsenic and bone fluoride confirm exposure separation but cannot provide individual body-weight-normalized, cumulative, or absorbed doses. The results do not estimate human risk magnitude, establish safe exposure limits, or demonstrate effects at environmentally representative concentrations.

Taken together, the strongest inference is that chronic As–F co-exposure impaired spermatogenesis at the tested concentrations and produced endpoint-specific joint effects, while reciprocal donor-material administration was accompanied by partial modification of selected recipient phenotypes. Coordinated changes in microbial composition, the colonic microenvironment, bile-acid- and lipid-related signals, enterohepatic and hepatic markers, reproductive hormones, and testicular endpoints support a nonexclusive multi-organ framework for further testing. Lower-dose and dose–response experiments, chemically characterized and strain-resolved FMT, longitudinal donor-restructuring analyses, targeted bile acid quantification, tissue-specific metabolomics and receptor intervention, fertility testing, and dynamic autophagic-flux assays will be required to determine which components are causal and whether the findings extend to environmentally relevant exposures.

CRediT authorship contribution statement

Jinyao Chen: Writing – review & editing, Writing – original draft, Visualization, Formal analysis, Data curation, Conceptualization. Yufen Cui: Validation, Investigation, Data curation. Yingda Wang: Validation, Investigation, Data curation. Ruipeng Kong: Validation, Investigation, Data curation. Penghui Liu: Methodology, Investigation, Conceptualization. Qiang Su: Methodology, Investigation, Data curation. Yidi Li: Writing – review & editing, Formal analysis. Hengyi Guo: Validation, Methodology. Yu Fu: Writing – review & editing, Validation. Qian Zhao: Supervision, Resources. Ben Li: Supervision, Resources. Xiaolin Tian: Methodology, Funding acquisition. Xiaoyan Yan: Writing – review & editing, Supervision, Resources, Project administration, Funding acquisition, Conceptualization.

Declaration of competing interest

The authors declare the following financial interests/personal relationships which may be considered as potential competing interests: Xiaoyan Yan reports financial support was provided by National Natural Science Foundation of China. Xiaoyan Yan reports financial support was provided by Department of Science and Technology of Shanxi Province. Xiaolin Tian reports financial support was provided by Department of Science and Technology of Shanxi Province. Xiaoyan Yan reports financial support was provided by Shanxi Provincial Department of Education. Jinyao Chen, Qiang Su, Yidi Li, Zhichao Jia, Shaoping Pan, Xiaolin Tian, Xiaoyan Yan has patent #CN202610156223.X pending to Shanxi Medical University. If there are other authors, they declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgements

This work was supported by the National Natural Science Foundation of China (No. 82173644), the Shanxi Province Central-Guided Local Science and Technology Development Fund Project (No. YDZJSX2024D064), the Shanxi Province Basic Research Program (No. 202403021212261), and the Science and Technology Guidance Project of the Shanxi Province Higher Education “Billion Project”. The authors gratefully acknowledge the Key Laboratory of Pathogenic Mechanisms and Prevention in Coal-Related Environments, Ministry of Education, and the Shanxi Key Laboratory of Environmental Pollution Pathogenicity and Prevention for their support. The authors also sincerely thank all members of the research team for their valuable assistance, insightful discussions, and technical support throughout this study.

Appendix A. Supplementary data

What’s this?

The following are the Supplementary data to this article:Thumbnail

Supplementary Fig. S1.

Thumbnail

Supplementary Fig. S2.

Thumbnail

Supplementary Fig. S3.

Thumbnail

Supplementary Fig. S4.

Thumbnail

Supplementary Fig. S5.

Thumbnail

Supplementary Fig. S6.

Thumbnail

Supplementary Fig. S7.

Thumbnail

Supplementary Fig. S8.

Thumbnail

Supplementary Fig. S9.

Thumbnail

Supplementary Fig. S10.

Thumbnail

Supplementary Fig. S11.

Thumbnail

Supplementary Fig. S12.

Thumbnail

Supplementary Fig. S13.

Thumbnail

Supplementary Fig. S14.

Thumbnail

Supplementary Fig. S15.

Thumbnail

Supplementary Fig. S16.

Thumbnail

Supplementary Fig. S17.

Download: Download spreadsheet (31KB)

Supplementary Table S1.

Download: Download spreadsheet (917KB)

Supplementary Table S2.

Download: Download spreadsheet (583KB)

Supplementary Table S3.

Download: Download spreadsheet (2MB)

Supplementary Table S4.

Download: Download spreadsheet (40KB)

Supplementary Table S5.

Data availability

The data supporting the findings are available from the corresponding author on reasonable request. Supplementary Tables S1–S5 provide the statistical and annotation records described in the text.

References

1 Present address: School of Public Health, Huazhong University of Science and Technology, Wuhan, Hubei 430030, China.
2 Present address: Ordos Center for Disease Control and Prevention, Ordos 017000, Inner Mongolia Autonomous Region, China.