Metabolic Alteration of Gut Microbiota Is Associated With Host Non-alcoholic Fatty Liver Disease Induced by Pyrethroid in Xenopus Laevis

Background: There is increasing awareness of the signicance of the gut microbiome to host health, and a clear relationship has been established between the perturbed gut microbiome and multiple diseases. Cis-bifenthrin, a widely used agricultural pyrethroid insecticide, has been implicated as a cause of hepatotoxicity due to the oxidative stress produced during its metabolism by the liver. Studies have demonstrated the role of gut microbiota in gut-liver axis, it is possible that the perturbation of gut microbiota may also contribute to the toxicity of cis-bifenthrin on the liver. Results: 16S rRNA gene sequencing suggested that cis-bifenthrin exposure signicantly perturbed the gut microbiota composition, and metabolomics analysis showed signature metabolic shifts arising from exposure. Moreover, we also found altered functional regulation of lipids in the liver after cis-bifenthrin exposure, and the accumulation of lipid droplets in hepatocytes was observed. Conclusions: Our results suggested cis-bifenthrin exposure disturbed the gut microbiota community and metabolite prole in frogs. Specically, changes in bile acid metabolites altered bile acid hepatoenteral circulation, which affected lipid metabolism in the liver and ultimately caused the development of fatty liver disease. Our ndings reveal novel insights into gut microbiota-host axis in frogs, and the perturbed microbial function provides novel mechanism contributing to cis-bifenthrin-induced toxicity. Hepatic gene expression prole in X. laevis after exposure to 0.3 μg/L cis-BF. a, Signicant enriched pathways obtained by KEGG pathway analysis. b, Network-based analysis of the differentially expressed genes involved in the fatty acid metabolism pathway, PPAR signaling pathway and fatty acid degradation pathway. c, Heatmap of differently expressed genes of PPAR signaling pathway. d, Heatmap of differently expressed genes of fatty acid metabolism pathway. e, Signicant enriched pathways for both differentially expressed genes and differentially changed metabolites. Data are presented as mean ± SEM, NS means there is no signicant difference compared with the control group.


Background
A growing body of research has demonstrated the essential role of the gut microbiome in the normal development and overall health of the host organism [1,2]. The gut microbiota control physiology beyond the intestinal lumen as they can signal to cells throughout the host to regulate their activity through the production of metabolically active microbial metabolites [3,4]. Intestinal dysbiosis can contribute to hepatic metabolic diseases via various obesity-associated mechanisms, including abnormalities in lipid metabolism [5]. Current reports on the chemical disruption of host-associated microbiota have generally been concerned with persistent organic pollutants (POPs) or antibiotics [6,7], and fewer studies have well characterized the microbiome-associated challenges to healthy hosts from pesticide exposure.
Pesticides are indispensable for modern agricultural production. In recent decades, synthetic pyrethroid (SP) use has increased continuously due to its insecticidal potency and shorter breakdown time than organophosphate pesticides [8]. Today, SPs have extensive indoor and outdoor applications, including in agriculture, public health, and residential usage [9]. SP contamination is extensively observed in soil, water, streambed sediments and indoor dust. For instance, a maximum value of 901 ng/L of SP has been detected in the surface water samples of California coastal watersheds, with bifenthrin being the most commonly detected SP [10,11]. The maximum concentration of SP pesticides found in the surface waters of the Pearl River Basin (China) was 29.72 g/L [12]. Additionally, SPs were detected in wild sh from Iberian river basins (Spain), with the highest concentration detected being 4938 ng/g lipid weight [13]. Moreover, it is worth noting that a considerable amount of research has shown that SPs ultimately pose a health threat to wildlife and humans. Mounting laboratory studies have reported the toxic effects of SPs, including developmental toxicity, neurotoxicity, and reproductive toxicity, in different nontarget organisms [14][15][16][17]. In addition, in recent years, considering the indispensable and crucial role of the gut microbiome, speci c attention has also been given to the in uence of SPs on the abundance, composition and metabolic function of the gut microbiome. A previous study indicated that lowconcentration exposure to permethrin signi cantly changed the abundance of the gut microbiota in rats but also altered the metabolic pro les, including the acetic and propionic acid contents [18]. However, the adverse in uences of SPs on gut microbiota community structures and metabolite-mediated host metabolism are largely unknown. Mechanistic insight into the interplay of gut microbiota, host metabolism, and metabolic health has gained signi cant attention over the past decade [19]. The gut microbiota has been considered a "microbial organ" within the intestine that modulates the health phenotype of the host via a range of microbial metabolites. Bile acids are classical examples of metabolites that are metabolized in the intestine by the gut microbiota into secondary bile acids and therefore regulate host metabolic pathways [19].
Since the gut microbiota have been shown to play indispensable roles in the maintenance of normal hepatic functions and the pathogenesis of disease, greater insights into the mechanisms driving the interplay between SPs and the gut microbiome are needed to reveal SP-induced adverse effects and related susceptibilities. Therefore, the present study was conducted to address whether SP exposure would alter the gut microbiota community structure as well as the gut microbiota-related metabolic functions, and whether SP-induced intestinal dysbiosis would further impact the metabolic process of the host liver. We applied 16S rRNA gene sequencing, liquid chromatography-mass spectrometry (LC-MS) metabolomics pro ling, and transcriptomics to probe functional interactions between liver function and the gut microbiome after exposure to pyrethroids. Cis-bifenthrin (cis-BF) is among the most commonly used SPs and is detected the most frequently [10,11]. Amphibians are suitable for studying the effects of stressors from environmental toxicants since they can provide more comprehensive toxicological endpoints [20]. The African clawed frog (Xenopus laevis, X. laevis) is a widely used amphibian model organism and is regarded as an invaluable tool for ecotoxicological assessment [21,22]. Therefore, cisbifenthrin was chosen as the representative SP, and African clawed frogs were used as the bio model. Our study provides the rst evidence of the functional interactions between liver function and the gut microbiome after pyrethroid exposure.

Methods
Animals and exposure X. laevis adults were obtained from Nasco (Fort Atkinson, WI, USA). To alleviate suffering, animals were maintained humanely in accordance with the guidelines of the American Society for Testing and Materials [23]. Adult frogs (2 years old) were exposed to the environmentally relevant concentrations of 0.06 and 0.3 μg/L cis-BF (Sigma, St. Louis, MO, USA). Stock solutions were prepared with alcohol, and the alcohol concentrations were identical for all treatment groups. Each treatment consisted of 27 male frogs that were randomly assigned to three replicates. Experiments were carried out at the same photoperiod (12 h light: 12 h dark) and temperature (22°C). During the 3 months of exposure, the exposure solutions were replaced daily to maintain the concentration of cis-BF. After exposure, the frogs were anesthetized in MS-222 and frozen in liquid nitrogen for subsequent analysis. All studies were approved by the Experimental Animal Ethics Committee of Zhejiang University.

Extraction and quanti cation of cis-BF
The cis-BF in the exposure solutions and intestines was extracted and detected according to a previous study with slight modi cations [24]. For exposure solutions, 200 mL of exposure solution was mixed with sodium chloride (5.0 g) and dichloromethane (200 mL) in a 500 mL separatory funnel. The mixture was shaken for 5 min and left at room temperature for 5 min to separate the organic and aqueous phases. Then, the organic phase was collected, and the extraction procedure was repeated once again. The combined organic phase was evaporated to near dryness on a rotatory evaporator at 40°C. Frog intestine contents (1 g) were homogenized and transferred into a 50 mL polypropylene tube preloaded with 0.25 g NaCl and 1 g MgSO 4 . Then, 5 mL acetonitrile was added to extract cis-BF, followed by sonication for 20 min. The mixture was centrifuged at 4500 rpm for 10 min, and the organic phase was transferred into a glass ask. The extraction procedure was repeated once again, the organic phase was combined and subsequently evaporated almost to dryness on a rotatory evaporator at 40°C. The residues from the exposure solution and tissues were reconstituted with 1 mL acetonitrile and transferred into a 1.5 mL polypropylene tube that contained 200 mg PSA. After vortexing and centrifugation (12000 g, 10 min), the supernatant was ltered with a 0.22-μm polytetra uoroethylene lter and transferred into vials for quantitative analysis.

Oil red O staining
Fresh hepatic tissues were xed in 4% (v/v) formaldehyde in phosphate buffer for 24 hours. The hepatic tissues were transferred into 15% and 30% sucrose solutions in turn at 4°C for dehydration. After dehydration, the hepatic tissues were embedded in optimal cutting temperature medium and subsequently cut into 8-μm-thick sections. The sections were stained with Oil Red O solution (Servicebio, Wuhan, China) for 10 min in the dark. The background differentiation procedure was performed in 60% isopropanol (Servicebio, Wuhan, China). Then, the sections were washed in water for 10 s and stained with hematoxylin. Glycerin gelatin (Servicebio, Wuhan, China) was used to seal the slices. The slices were observed and photographed on a light microscope (Olympus, Tokyo, Japan).

16S rRNA gene sequencing
Total genomic DNA was extracted from the intestinal contents of frogs using the PowerSoil® DNA isolation kit according to the manufacturer's instructions. Extracted DNA was ampli ed to target the V3-V4 regions of the 16S rRNA gene of bacteria using the primer pair 341F (CCTAYGGGRBGCASCAG) and 806R (GGACTACNNGGGTATCTAAT). The amplicons were assessed on the Qubit@ 2.0 Fluorometer (Thermo Scienti c) and sequenced on an Ion S5TM XL platform. To obtain high-quality clean reads, the raw reads were ltered on Cutadapt (Martin M., 2011). In addition, chimera sequences were removed with the UCHIME algorithm (UCHIME Algorithm). Uparse software (Uparse v7.0.1001) was used to choose operational taxonomic units (OTUs) with a threshold of 97% sequence similarity. A representative sequence from each OTU was selected for taxonomic annotation using the Silva Database (Version 132). The taxonomic information of the 16S rRNA gene sequences was obtained at different classi cation levels, including the phylum, class, order, family and genus levels. Alpha diversity (observed species, Chao1, Shannon, Simpson, ACE, and good coverage indices) and beta diversity were calculated with QIIME (Version 1.7.0). Linear discriminant analysis (LDA) effect size (LEfSe) was performed using the LEfSe tool.

Metabolomics analysis
The intestinal contents (100 mg) were ground with liquid nitrogen, and the homogenate was resuspended in prechilled 80% methanol (-20°C). After vortexing, the samples were incubated for 1 h at -20°C and then centrifuged at 14000 g for 20 min at 4°C. The supernatants were transferred to a new tube, dried in a vacuum concentrator, and then reconstituted in 60% methanol for LC-MS/MS analysis. LC-MS/MS analysis was performed on a Vanquish UHPLC system (Thermo Fisher) coupled with an Orbitrap Q Exactive HF-X mass spectrometer (Thermo Fisher). Chromatographic separation was performed using a Hyperil Gold column (100×2.1 mm, 1.9 μm). For the positive polarity mode (pos), the eluents were 0.1% FA in water (A) and methanol (B). For the negative polarity mode (neg), the eluents were 5 mM ammonium acetate (A, pH 9.0) and methanol (B). The solvent gradient was programmed as follows: 98% A, 1.5 min; 98-0% A, 12.0 min; 0% A, 14.0 min; 0-98% A, 14.1 min; 98% A, 17 min. The ow rate was 0.2 mL/min. The other optimized parameters were as follows: spray voltage, 3.2 kV; capillary temperature, 320°C; sheath gas ow rate, 35 arb; and aux gas ow rate, 10 arb. The raw data were processed with Compound Discoverer 3.0 (CD 3.0, Thermo Fisher) for peak alignment, peak picking, and quantitation for each metabolite. Then, the molecular features were matched with the mzCloud (https://www.mzcloud.org/) and ChemSpider (http://www.chemspider.com/) databases.

Transcriptomics analysis
The mRNA-Seq experiment was completed by Novogene Bioinformatics Technology Co., Ltd. (Beijing, China). Total RNA was isolated from frog liver using TRIzol reagent (Takara Bio Inc., Kusatsu, Japan). The RNA purity and integrity were assessed by using the A260/A280 ratio and agarose-formaldehyde gel electrophoresis, and the RNA concentration was also measured. A total amount of 3 μg of RNA was used for RNA sample preparation, and the sequencing libraries were generated using the NEBNext® UltraTM RNA Library Prep Kit for Illumina® (NEB, USA). The quality of the sequencing library was assessed on the Agilent Bioanalyzer 2100 system. Then, the library was sequenced on an Illumina HiSeq platform to generate 125 bp/150 bp paired-end reads. Reads containing adapters and poly-N and low-quality reads were removed to obtain clean reads. FeatureCounts v1.5.0-p3 was employed to calculate the clean read numbers mapped to each gene, and the FPKM value of each gene was also calculated based on the length of the gene and the read counts mapped to this gene. The DESeq2 R package (1.16.1) was used to assess the differential expression analysis results, and genes with an adjusted p-value < 0.05 were considered differentially expressed. The statistical enrichment of differentially expressed genes in KEGG pathways was performed using the clusterPro ler R package.

Quantitative real-time PCR (qPCR) validation
Frog liver tissues (n = 3 replicates; each replicate contained three tissues) were homogenized with RNAiso plus (Takara Bio Inc, Kusatsu, Japan) to extract the total RNA. The total RNA contents were measured on an Epoch Microplate Spectrophotometer (BioTek Instruments, Inc., Winooski, VT, USA). The quality of the extracted total RNA was also examined via the 260/280 nm ratios and 1% agarose gel electrophoresis.
The Prime Script TM RT reagent Kit (Takara Bio Inc, Kusatsu, Japan) was used to synthesize the cDNA. The real-time PCR was carried out on an ABI 7300 system (Applied Biosystems, CA, USA) using the SYBR Green kits (Takara Bio Inc, Kusatsu, Japan). Speci c primers of the target genes were identi ed by the NCBI Primer-BLAST and showed in the SI (Table S1). The transcriptional stability of the reference gene was accessed using the GeNorm, and nally the 18s was chosen as the housekeeping gene. The gene transcription levels were calculated using 2 −ΔΔCt method.

Statistics
Data were presented as the mean ± standard error (SEM). SPSS 20.0 (SPSS, Chicago, IL, USA) was used to perform statistical analyses, and the comparison between the exposure groups and control group was performed using one-way analysis of variance (ANOVA) followed by Tukey's test. A p < 0.05 was were considered to be signi cant.

Cis-BF contents
The cis-BF concentrations were measured at 0 h and 24 h after renewing the exposure solutions, and the measured cis-BF concentrations were similar to the nominal exposure concentrations (Table S2). The cis-BF concentrations in the intestinal contents of X. laevis are shown in Fig. 1a. In the 0.06 and 0.3 μg/L exposure groups, the cis-BF concentrations in the intestinal contents were 170.35 ± 10.95 and 266.60 ± 30.15 μg/kg, respectively (Fig. 1a). Cis-BF was not detected in the intestinal contents of the control group.
Lipid accumulation in the liver Exposure to 0.3 μg/L cis-BF increased the hepatosomatic index (HSI) of X. laevis (Fig. 1b). Compared with the control group (Fig. 1c), the hepatic lipid droplets in the liver of X. laevis in the 0.06 (Fig. 1d) and 0.3 μg/L (Fig. 1e) exposure groups were obviously increased (Fig. 1f).

Gut microbiome changes
Cis-BF exposure signi cantly altered the composition, abundance and diversity of the gut microbiome community. In terms of the gut microbial structure, obvious differences were observed between the control group and all cis-BF exposure groups. The 0.06 and 0.3 μg/L cis-BF exposure groups were more similar (ANOSIM; R = 0.07, p = 0.17), with all cis-BF exposure groups showing obvious differences from the control group (0.06: R = 0.18, p = 0.018; 0.3: R = 0.25, p = 0.013; Table 1). Compared with the control group, 0.3 μg/L cis-BF exposure signi cantly decreased the diversity of the microbiome (Simpson: p = 0.0379 and Shannon: p = 0.0312; Fig. 2a, b and Table 2). A network based on the top 50 OTUs was also constructed to visualize the different bacteria in the intestinal contents of X. laevis in the control and cis-BF exposure groups. The result showed that 11 and 20 OTUs exclusively emerged in 0.06 and 0.3 μg/L exposure groups, respectively (Fig. 2c). Fig. 2d illustrates the relative abundance of the top 10 microbiota at the phylum level in X. laevis. At the phylum level, Firmicutes and Fusobacteria were the most abundant phyla, followed by Proteobacteria, Bacteroidetes, Actinobacteria, Cyanobacteria, Synergistetes, Chloro exi, Tenericutes and Armatimonadetes. Obvious increases in the abundances of Chloro exi and Armatimonadetes were observed in all cis-BF exposure groups. However, 0.3 μg/L cis-BF exposure signi cantly decreased the abundance of Bacteroidetes. At the genus level, several signi cantly changed genera are shown in Fig. 3a. Exposure to cis-BF (0.06 and 0.3 μg/L) also increased the abundances of the pathogenic bacteria Plesiomonas (Fig. 3b) and Aeromonas (Fig. 3c). LEfSe was performed to identify the speci c bacteria that served as biomarkers in the control group and cis-BF treatment groups. The LEfSe results illustrated that 14 biomarkers with an LDA score > 4 were differentially abundant between the control group and the cis-BF treatment groups (Fig. 3d).
The metabolic pro le of the gut microbiome Three months of cis-BF exposure signi cantly disturbed the metabolic pro les of the X. laevis intestinal contents. Speci cally, 60 (pos:36 and neg:24) and 73 (pos:36 and neg:37) metabolites with a fold change > 1.5 and p < 0.05 were observed in X. laevis treated with 0.06 ( Figure S1a; Table S3) and 0.3 μg/L ( Fig.  4a; Table S3) cis-BF, respectively.
In the 0.3 μg/L cis-BF treatment group, the contents of taurocholic acid (TCA), glycochenodeoxycholic acid (GCDCA) and glycocholic acid (GCA) were lower than those in the control group. However, the deoxycholic acid (DCA), taurochenodeoxycholic acid (TCDCA) and cholic acid (CA) levels were unaffected by cis-BF (Fig. 4b). In the 0.06 μg/L cis-BF treatment group, there were no signi cant changes in terms of bile acids (Fig. 4b).
All metabolites that changed signi cantly upon cis-BF exposure were subjected to KEGG pathway enrichment analysis, and 7 pathways were signi cantly enriched in the 0.3 μg/L cis-BF exposure group (Fig. 4c). The signi cantly enriched metabolic pathways included primary bile acid biosynthesis, secondary bile acid biosynthesis, bile secretion and cholesterol metabolism. In the 0.06 μg/L cis-BF treatment group, pyrimidine metabolism pathway was signi cantly enriched ( Figure S1b; Table S4). Correlation analysis based on the top 20 OTUs and bile acids (Fig. 4d) indicated that g_Clostridium_sensu_stricto_1, f_Lachnospiraceae, f_Porphyromonadaceae, g_Desulfovibrio, f_Erysipelotrichaceae, f_Peptostreptococcaceae, p_Firmicutes, f_Family_XIII, g_Lachnoclostridium and s_Fusobacterium_varium were obviously positively correlated with bile acids. However, g_Romboutsia showed a negative correlation with taurochenodeoxycholic acid.

Differentially expressed genes
We identi ed 5055 (2620 downregulated and 2435 upregulated genes; Figure S2a) and 5188 (2570 downregulated and 2618 upregulated genes; Figure S2b) differentially expressed genes in the 0.06 and 0.3 μg/L cis-BF treatment groups, respectively. All differentially expressed genes were assigned for KEGG pathway analysis, with the altered pathways shown in Table S5. The KEGG results showed that 34 and 17 pathways were obviously enriched in the 0.06 (Table S5) and 0.3 μg/L (Fig. 5a) cis-BF treatment groups, respectively. The enriched signaling pathways associated with metabolic signaling pathways, including fatty acid degradation, fatty acid metabolism and the PPAR signaling pathway, indicated that cis-BF exposure could in uence fatty acid metabolism in the livers of frogs (Fig. 5b). Heatmaps of the differentially expressed genes involved in PPAR signaling pathway and fatty acid metabolism pathway are shown in Fig. 5c and 5d, respectively. The pathways of primary bile acid biosynthesis and fatty acid degradation were signi cant enriched for both differentially expressed genes and differentially changed metabolites (Fig. 5e). The gene expression levels obtained from the transcriptomics was validated by using qPCR, and the fold-changes of selected genes were in agreement with transcriptomic results (Table  S6).

Discussion
In this study, we clearly showed that environmentally relevant levels of cis-BF exposure induced a signi cant change in the frog gut microbiome and its metabolic pro les, leading to disorders of lipid metabolism in the liver. In addition, lipid metabolism disorders were strongly related to changes in bile acid production in the gut. Our results suggested that gut microbiota dysbiosis is involved in the abnormal lipid metabolism caused by SP.
High deposition of cis-BF in the gut contents of frogs indicates bioaccumulation and bioavailability of these compounds in the intestinal system, which suggests that cis-BF may interact with interstitial cells and the intestinal microbiome. The microbiome is relatively plastic and can be rapidly altered through factors such as diet, drugs, probiotics, and xenobiotic exposure [25]. Our 16S rRNA gene sequencing results (ANOSIM) showed signi cantly different community structures between groups, which indicates that cis-BF exposure signi cantly altered the community structures of the frog gut microbiome. A reduction in bacterial diversity was observed in cis-BF-treated X. laevis, suggesting that cis-BF altered the patterns of the gut microbial ecosystems, which may alter the mutually bene cial interaction toward another stable but harmful balance [26]. We identi ed a number of bacterial genera perturbed in X. laevis after cis-BF exposure. It is remarkable that several pathogenic bacteria of the Plesiomonas and Aeromonas genera were only observed in treated frogs. Plesiomonas is a microorganism of aquatic origin and plays a role as a pathogen of both humans and animals causing occasional opportunistic infections and diarrhea [27]. Aeromonas species that are known to be pathogenic in humans are also pathogenic in sh [28]. Thus, our results showed that SP exposure induced the prevalence of several potentially pathogenic bacteria, which could increase the risk of disease development. In addition, several genera of the families Erysipelotrichaceae, Ruminococcaceae and Lachnospiraceae, which belong to the Firmicutes phylum, decreased in the intestinal tract of frogs. Previous studies have con rmed that the members of the family Erysipelotrichaceae are associated with lipidemic pro les within the host [29,30]. The families Lachnospiraceae and Ruminococcaceae, which belong to Clostridiales within Firmicutes, have been associated with the maintenance of gut health [31]. A few members of the family Lachnospiraceae or Ruminococcaceae are known to convert conjugated primary bile acids to secondary bile acids. The loss of Lachnospiraceae and Ruminococcaceae in acute and chronic intestinal diseases has been linked to bile acid metabolic changes [32]. For instance, a previous study reported that in pouch patients who experience ulcerative colitis, members of the family Ruminococcaceae were signi cantly depleted, and this condition was associated with secondary bile acid de ciency [33]. Mice fed lard and bile acids showed changes in the dominant intestinal bacterial community, including a decline in the relative abundance of Lachnospiraceae linked to hepatic lipid rearrangements [34]. Therefore, our results suggested that the decrease in certain genera in the intestinal tracts of treated frogs may induce changes in the contents of microbial metabolites that can act both locally and systemically.
Consistent with the microbiome compositional changes, the metabolite pro les were also signi cantly altered, with a number of key pathways being disrupted by cis-BF exposure. In particular, a perturbation was found in cholesterol metabolism, primary bile acid biosynthesis, secondary bile acid biosynthesis and bile secretion in the intestinal microbiota of frogs treated at higher doses. Bile acids are metabolites of cholesterol that are synthesized in the liver and secreted into the intestine, and a portion of the total bile acids are dehydroxylated by intestinal bacteria to hydrophobic secondary bile acids [35]. Our results indicated that cis-BF possesses biological activities on the composition of the gut microbiota, which may cause changes in microbial metabolites. Speci cally, decreased levels were found for bile acids, including glycocholic acid (GCA), glycochenodeoxycholic acid (GCDCA), taurocholic acid (TCA), and sulfoglycolithocholic acid. An important nding in this study is that changes in bile acids were strongly positively associated with the abundance of the Clostridium_sensu_stricto_1 bacterial family (p 0.01). Linear discriminant analysis effect size also suggested that Clostridium_sensu_stricto_1 was a typical biomarker found in the control group compared with the cis-BF exposure groups. In fact, a previous study reported that a small population of intestinal species in the genus Clostridium spp. are capable of deconjugating conjugated BAs and metabolizing bile acids [36]. Therefore, Clostridium_sensu_stricto_1 is probably one of the most important genera that mediates BA pro le changes induced by cis-BF. Obviously, future studies are needed to address the role of the gut microbiome in the changes in its metabolites after exposure to environmental toxicants. As a class of host-derived metabolites, BAs are signaling molecules through a variety of receptors to maintain metabolic homeostasis, such as hepatic lipid and energy homeostasis [37]. Disorders in bile acid metabolism could cause cholestatic liver diseases, dyslipidemia, fatty liver diseases, diabetes, etc [37]. In addition, a defect in the synthesis of the primary bile acids (such as cholic acid or chenodeoxycholic acid) in the liver would cause cholestasis and the malabsorption of fat and fat-soluble vitamin [38]. It was also observed that bacterial dysbiosis observed in cirrhosis is linked to low bile acid levels entering the intestine [36]. Microbe-derived metabolites and the signaling pathways they affect might play important roles in the development of nonalcoholic fatty liver disease [39]. Therefore, the modi cation of BA metabolites after cis-BF exposure may perturb the normal physiological process of hepatic metabolism in frogs.
Transcriptomic analysis of the liver was performed in parallel, and differentially expressed genes were identi ed. KEGG analysis revealed that fatty acid metabolism-related terms were signi cantly enriched among these differentially expressed genes. For example, cis-BF treatment in frogs altered the activation of fatty acid metabolism, fatty acid degradation and the peroxisome proliferator activated receptor (PPAR) signaling pathway. PPARs are nuclear hormone receptors that modulate the expression of genes and are involved in adipogenesis and maintenance of metabolic homeostasis [40]. In particular, xPPAR alpha and PPARG coactivator 1 alpha (ppara and ppargc1a) gene were downregulated signi cantly, meanwhile, the mRNA of genes associated with fatty acid transport processes were signi cantly decrease. The inhibition of the expression levels of the receptors and these signaling pathway-related genes may lead to a decrease in disorders of fatty acid oxidation and lipid catabolism, resulting in increased lipid deposition in the liver. The association between the hepatic gene enriched pathways and the gut metabolic pro le was evaluated, and signi cant enrichment was found in the primary bile acid biosynthesis pathway. Bile acids circulate between the liver and the intestine and back to the liver, which called enterohepatic circulation of the bile, play an essential role in the liver, biliary, and intestinal function and disease [37]. Clinical studies indicate that dysregulation of bile acid homeostasis and its related signaling pathways is associated with the occurrence of liver disease [39]. Our results further con rmed the important role of bile acids in liver metabolism, especially when challenged by environmental pollutants, and bile acid disorders have a part in inducing liver metabolism disorder. Although the regulatory mechanism remains unclear, as previous reports have suggested that bile acids can induce the expression of the human PPAR alpha gene and contribute to liver disease by altering nuclear bile acid receptor farnesoid X receptor (FXR) signaling [41,42], indicating that the suppressed BA-FXR signaling pathway might participate in the regulation of fatty acid metabolism in liver. Further studies are needed to reveal the role of FXR in the regulation of lipid homeostasis and metabolic pathways upon cis-BF exposure. Our results suggest that changes in the metabolic capacities of intestinal microbes reduces the lipid metabolism function of the liver.
Lipids are stored within lipid droplets in cells [43]. To further investigate whether the hepatic metabolic pathway changes that were induced by cis-BF exposure led to alterations in lipid storage, we measured the formation of lipid droplets in hepatocytes. We observed that liver sections from exposed frogs showed increased lipid droplets by using oil red O staining. A previous study reported that lipid droplet accumulation is the hallmark of nonalcoholic fatty liver disease [43]. Our results indicated that the inhibition of metabolism-related molecular signaling pathways in the liver induces the accumulation of lipids, which leads to cellular dysfunction and liver disease, such as hepatic steatosis.

Conclusion
In summary, the present study demonstrates that environmentally relevant levels of cis-BF exposure signi cantly altered the structure and composition of intestinal microorganisms, resulting in changes in the intestinal metabolites involved in multiple metabolic pathways, especially bile acid metabolism, and ultimately interfered with lipid metabolism in the X. laevis liver. Our results indicated that gut microbes play a crucial role in regulating the toxicity of environmental toxicants. In particular, the interplay between gut microbes and xenobiotics may promote the unexpected development of toxicity, which is easily overlooked due to the lack of information on the direct-action sites of xenobiotic compounds in organisms. Therefore, our results highlight the capacity for physiological regulation by intestinal microorganisms, particularly in the potential ampli cation of chemical toxicity resulting from environmental toxicant exposure, which is important for the identi cation of potential ecological risks of pesticide use.

Declarations
Ethics approval and consent to participate Not applicable.

Consent for publication
Not applicable.

Availability of data and materials
The raw datasets supporting the results of this article has been deposited in the NCBI and the project numbers are PRJNA725147 and PRJNA725067.

Competing interests
The authors declare that they have no competing interests.  Magni cation, ×200. f, Relative Oil Red O area in the livers of X. laevis in the control and exposure groups (n = 3). Data are presented as mean ± SEM, *p < 0.05 and **p < 0.01 indicate signi cant differences compared with the control group.

Figure 2
Cis-BF exposure changed the gut microbiota community structure of X. laevis. Simpson (a) and Shannon (b) diversity of gut microbiota in X. laevis (n = 8). c, Network-based analysis of top 50 OTUs of gut microbiota in X. laevis. d, Bacterial community compositions at phylum level (top 10) of gut microbiota in X. laevis (n = 8). Data are presented as mean ± SEM, *p < 0.05 indicates a signi cant difference compared with the control group. Effects of cis-BF exposure on the gut microbiota composition at the genus level. a, The signi cant changed genera of gut microbiota in X. laevis after exposure to cis-BF (n = 8). The relative abundances of the pathogenic bacteria Plesiomonas (b) and Aeromonas (c) of gut microbiota in X. laevis (n = 8). d, The most differentially abundant taxa between control and cis-BF exposure groups. Data are presented as mean ± SEM, **p < 0.01 indicates a signi cant difference compared with the control group. NS means there is no signi cant difference compared with the control group.  6). b, The bile acids concentrations of the intestinal contents in X. laevis (n = 6). Data are presented as mean ± SEM, *p < 0.05 and **p < 0.01 indicate signi cant differences compared with the control group. c, Enrichment ratio of pathways by KEGG pathway analysis in 0.3 μg/L cis-BF exposure group. d, Correlations between bile acids concentrations and the relative abundance of top 20 genera.