Effects of Early Life Exposures to the Aryl Hydrocarbon Receptor Ligand TCDF on Gut Microbiota and Host Metabolic Homeostasis in C57BL/6J Mice

Background: Exposure to persistent organic pollutants (POPs) and disruptions in the gastrointestinal microbiota have been positively correlated with a predisposition to factors such as obesity, metabolic syndrome, and type 2 diabetes; however, it is unclear how the microbiome contributes to this relationship. Objective: This study aimed to explore the association between early life exposure to a potent aryl hydrocarbon receptor (AHR) agonist and persistent disruptions in the microbiota, leading to impaired metabolic homeostasis later in life. Methods: This study used metagenomics, nuclear magnetic resonance (NMR)– and mass spectrometry (MS)–based metabolomics, and biochemical assays to analyze the gut microbiome composition and function, as well as the physiological and metabolic effects of early life exposure to 2,3,7,8-tetrachlorodibenzofuran (TCDF) in conventional, germ-free (GF), and Ahr-null mice. The impact of TCDF on Akkermansia muciniphila (A. muciniphila) in vitro was assessed using optical density (OD 600), flow cytometry, transcriptomics, and MS-based metabolomics. Results: TCDF-exposed mice exhibited lower abundances of A. muciniphila, lower levels of cecal short-chain fatty acids (SCFAs) and indole-3-lactic acid (ILA), as well as lower levels of the gut hormones glucagon-like peptide 1 (GLP-1) and peptide YY (PYY), findings suggestive of disruption in the gut microbiome community structure and function. Importantly, microbial and metabolic phenotypes associated with early life POP exposure were transferable to GF recipients in the absence of POP carry-over. In addition, AHR-independent interactions between POPs and the microbiota were observed, and they were significantly associated with growth, physiology, gene expression, and metabolic activity outcomes of A. muciniphila, supporting suppressed activity along the ILA pathway. Conclusions: These data obtained in a mouse model point to the complex effects of POPs on the host and microbiota, providing strong evidence that early life, short-term, and self-limiting POP exposure can adversely impact the microbiome, with effects persisting into later life with associated health implications. https://doi.org/10.1289/EHP13356


Introduction
Persistent organic pollutants (POPs) are a serious and global threat to human health owing to their persistence, prevalence, and bioaccumulative nature.Exposure to POPs is associated with various health problems, including cancer, 1,2 neurological disorders, 3 immune suppression, 4 and reproductive disorders, 5 as well as metabolic disease, such as obesity 6,7 and diabetes. 8,9Accumulating evidence suggests that exposure to POPs during early life may increase the risk of developing diabetes or obesity later in life in humans; 10,11 however, the mechanisms underlying this connection remain poorly defined.
2,3,7,8-Tetrachlorodibenzofuran (TCDF), a common environmental pollutant and POP, is a high-affinity aryl hydrocarbon receptor (AHR) ligand.The AHR is the sole ligand-activated receptor in the Per-ARNT-Sim family of transcription factors that mediates the effects of a variety of exogenous and endogenous small molecules. 12,13In terms of structure and acute toxicity, TCDF is closely related to 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD), whose toxicity is mediated through the AHR. 14Unlike TCDD, TCDF has a short half-life (2-4 d in mouse and 1-3 y in human) and is quickly eliminated from the body compared with the more persistent TCDD (8-12 d in mouse 15 and 7-10 y in human). 16umans are exposed to POPs primarily through high-fat foods, such as meat, meat products, dairy products, and some fish. 17In spite of the importance of the potential adverse health risks derived from human exposure to these POPs, which is mainly via dietary intake, the current available scientific information on this topic is limited to a few countries. 179][20] A recent study suggested that long-term exposure (continuous 14 wk) to TCDF in mice induced hepatic lipogenesis, an early hallmark of metabolic dysfunction-associated steatotic liver disease (formerly known as nonalcoholic fatty liver disease). 21Although many studies have reported that POPs quickly and profoundly affect host health, 19,20 it is still unclear what the long-lasting consequences are after those with shorter half-lives are eliminated from the body.
The gut microbiota plays crucial roles in maintaining human health. 22Mounting evidence suggests that the gut microbiota affects glucose and metabolic homeostasis. 23,24Possible mechanisms linking the gut microbiota to glucose homeostasis may include increased intestinal permeability, low-grade endotoxemia, changes in the production of short-chain fatty acids (SCFAs) or branched-chain amino acids, alterations in bile acid metabolism, and effects on the secretion of gut hormones. 239][30] Our previous study reported that POPs not only altered bacterial physiology but also significantly affected the metabolism and gene expression of microbial communities in vitro. 31Recently, we demonstrated in mice that early life exposure to 3,3 0 ,4,4 0 ,5-pentacholorobiphenyl (PCB 126), one of the most acutely toxic polychlorinated biphenyl (PCB) congeners with a long half-life (17 d in mouse and 4.5 y in human), 32 had a substantial impact on bacteria in adulthood at the community structure, metabolic, and functional levels, independent of diet. 33,34These findings raise the possibility that microbial toxicity could be a key target of early life exposure to environmental pollutants, potentially contributing to an increased risk of metabolic disorders later in life.
Here, we investigated the gut microbiome composition and function, as well as the physiological and metabolic effects of early life exposure to TCDF, a representative of a large class of POPs in conventional, germ-free (GF), and AHR knockout [Ahrnull (Ahr −= − )] mice using metagenomics, nuclear magnetic resonance (NMR)-and mass spectrometry (MS)-based metabolomics, and biochemical assays.The effects of TCDF on Akkermansia muciniphila (A.muciniphila) in vitro were analyzed using optical density (OD 600), flow cytometry, transcriptomics, and MS-based metabolomics.

In Vitro Bacterial Culture and Flow Cytometry
A. muciniphila DSM 22959 (Leibniz Institute DSMZ) was cultured using BHI CHV media.Overnight cultures were then diluted 1:200 into fresh BHI CHV media and then immediately treated with DMSO (1%, vol/vol) as the vehicle or with two doses of TCDF (high: 6 lM and low: 0:6 lM, n = 6) at 37°C, according to our previous study. 31he growth rate of A. muciniphila was measured by optical density (OD 600) using a Multiskan Sky Microplate spectrophotometer (Thermo Fisher Scientific).Flow cytometry, metabolomics, and transcriptomics analysis were performed on the treated cultures after 36 h of incubation.The in vitro experiment was performed in a monitored anaerobic chamber [Coy Laboratory Products; 20% carbon dioxide (CO 2 ), 5% hydrogen (H 2 )] and repeated at least three times.
After incubation, A. muciniphila was washed and resuspended with 1 × reduced phosphate-buffered saline containing 1 mg=mL L-cysteine.For physiology assessments, bacteria samples were stained by four fluorescent dyes, including SybrGreen, Pi, CFDA, and DiBAC 4 . 31,35All cytometric analyses were analyzed on a BD Accuri C6 plus flow cytometer (BD Biosciences), and data were analyzed with FlowJo software (version 10; FlowJo, LLC).The percentages of Pi-, CFDA-, and DiBAC 4 -stained cells were normalized to the total bacterial counts obtained by SybrGreen staining.

Mice
Male C57BL/6J mice were used in the four series of experiments (Table S1).The experiments were performed using protocols approved by the institutional animal care and use committees of The Pennsylvania State University (Penn State; PROTO202001416) and Montana State University (2022-54-117).
Experiment 1: short-and long-duration experiment.Threeweek-old male C57BL/6J wild-type (WT) mice were obtained from Jackson Laboratories.After acclimatization, the mice were trained to eat bacon-flavored dough pills that were prepared using a tablet mold (Total Pharmacy Supply) for 5 d.After training, the mice were fed pills containing TCDF (dissolved in acetone) or acetone alone as the vehicle, and one pill uniformly contained 0:46 lg TCDF (24 lg=kg as the final dose 19,20 ).The pills were dried at room temperature to ensure complete acetone evaporation.This dose is higher than those of polychlorinated dibenzo-p-dioxins and dibenzofurans typically found in the diet of adults [0.3-3.0 pg toxic equivalents (TEQs)/kg body weight]. 36The mice were housed singly in an empty cage and monitored to ensure the pill was consumed in the morning for each of the 5 d.Two models were used to evaluate the effect of early life TCDF exposure on mice (Figure 1A).For the acute, short-duration model, the mice (n = 6 per group) were sacrificed on the day after the last TCDF exposure.For the long-duration model, the mice (n = 6 per group) were sacrificed at 3 months after TCDF exposure.In these two models, urine and feces were collected using separate cages before sacrifice.Blood collection via cardiac puncture was performed post-sacrifice, and serum samples were obtained by centrifugation (5,000 × g, 4°C) for 5 min using BD Microtainer Blood Collection Tubes.Liver, cecal content, adipose, and intestinal tissue samples were collected immediately after sacrifice by CO 2 asphyxiation and kept at −80 C for subsequent analysis.Two independent longduration experiments were conducted in the animal facility at Penn State for replication purposes.
Experiment 2: GF and Ahr À=À mice with TCDF experiment.GF C57BL/6J mice were from the Penn State Gnotobiotic Facility.Male Ahr −= − mice (C57BL/6J congenic mice) were obtained from C. Bradfield (University of Wisconsin) and inbred to maintain the line, which is used as a powerful model to understand the physiological roles of the AHR and toxicology of halogenated aromatic pollutants. 37TCDF (24 lg=kg) or corn oil as the vehicle were administered to age-matched male GF and Ahr −= − mice by oral gavage once daily for 5 d (n = 4 per group) (Figure 2F).After 5 d, all mice were sacrificed by CO 2 asphyxiation and liver and intestine tissue samples were collected.
Experiment 3: GF mice with cecal microbiota transplantation experiment.GF mice were from the gnotobiotic facilities at Penn State and Montana State University.Cecal microbiota transplantation was performed as follows (Figure 3G): Four-week-old male GF C57BL/6J mice (n = 5 per group) were orally gavaged with 100 lL of cecal suspension (100 mg in Experiment 4: GF mice with A. muciniphila and antibiotic cocktail experiment.GF mice were from the Montana State University Gnotobiotic Facility.A. muciniphila was administered to GF mice by oral gavage at one dose of 10 7 colony forming units (CFUs)/0:1 mL suspended in sterile BHI CHV media containing an end concentration of glycerol (15% vol/vol) (Figure 4D).The viability of A. muciniphila in feces was confirmed by bacterial Gram stain (Fisher Scientific) and A. muciniphila-specific polymerase chain reaction (PCR) primers (Table S2).After 1 wk colonization, the mice (n = 3-5) were treated for 2 wk with an antibiotic cocktail solution composed of 2 mg=mL of neomycin, 2 mg=mL of streptomycin, and 2 mg=mL of bacitracin in sterile drinking water (Akk + ABX) or sterile water (Akk).Control groups were administered the antibiotic cocktail solution without A. muciniphila (ABX) for 2 wk.Blood collection was performed via cardiac puncture post-sacrifice, and serum samples were obtained by centrifugation (5,000 × g, 4°C) for 5 min using BD Microtainer Blood Collection Tubes.Liver, cecal content, and intestinal tissue samples were harvested after sacrifice by CO 2 asphyxiation and kept at −80 C for subsequent analysis.
Experiment 5: GF mice with cecal microbiota transplantation along with A. muciniphila supplementation experiment.GF mice were from the Penn State Gnotobiotic Facility.Cecal microbiota transplantation along with A. muciniphila supplementation was performed as follows (Figure 4I): 6-wk-old male GF male C57BL/6J mice (n = 5 per group) were orally gavaged with 100 lL of cecal suspension (100 mg in 1 mL of sterile BHI CHV media) from vehicle-or TCDF-treated mice in the long-duration model along with or without A. muciniphila (10 7 CFU/0:1 mL) supplementation.At 1-wk posttransplantation, the mice were sacrificed by CO 2 asphyxiation and blood, liver, cecal content, and intestinal tissue samples were harvested and kept at −80 C for subsequent analysis.The viability of A. muciniphila in the cecal contents was confirmed by A. muciniphila-specific PCR primers (Table S2).

Blood Clinical Biochemistry and Cytokine Analysis
Serum alkaline phosphatase (ALP) was measured using the VetScan VS2 and the Mammalian Liver Profile rotor (Abaxis Inc.) according to the manufacturer's instructions.Serum cytokine levels were measured with a BioPlex 200 mouse cytokine array/chemokine array 32-Plex by Eve Technologies.

Liver Triglyceride and Glutathione Quantification
Liver triglyceride (TG) was measured using the Triglyceride Colorimetric Assay Kit according to the manufacturer's protocol (Cayman Chemical).The ratios of reduced glutathione (GSH) and

Fecal Mucin and Serum Lipopolysaccharide Quantification
Fecal mucin levels were measured using the Fecal Mucin Assay Kit according to the manufacturer's protocol (Cosmo Bio USA).Serum lipopolysaccharide (LPS) levels were measured using Pierce LAL Chromogenic Endotoxin Quantitation Kit (Thermo Fisher Scientific).

Glucose Tolerance Test
The glucose tolerance test was done as previously described 38 and followed a 6 h fast.In brief, blood glucose was measured following intraperitoneal injection of 40% glucose (2 g=kg body weight).The blood glucose concentrations were measured using a OneTouch Ultra 2 meter (LifeScan).
After centrifugation (12,000 × g, 4°C) for 10 min, the top layer was dried in a vacuum and reconstituted in 2 mL of hexane, followed by cleaning using a Florisil column as per the US Environmental Protection Agency 1613 method. 39Column packing materials for the cleanup procedure included anhydrous sodium sulfate and Florisil 60/100 mesh from Restek.Liver TCDF levels were measured by an Agilent Technologies 7890A-5975C gas chromatography-MS (GC-MS) system equipped with a Rxi-5ms (Restek) capillary column [30 m, 0:25-mm internal diameter (ID), 0:25-lm film thickness].The results were quantified by a standard curve.The limit of detection was 10 pg=L.

H NMR-Based Metabolomics Experiments
Hydrophilic metabolites from cecal content, liver, and urine samples for NMR analysis were extracted as previously described. 40Liver lipid quantification was done using published methods, 41 with minor modifications.Briefly, 50 mg of liver tissue was homogenized in 1 mL of precooled chloroform/methanol mix [1:1 (vol/vol)], followed by adding 296 lL of water.After centrifugation (10,000 × g, 4°C) for 10 min, the lower phase was dried in a vacuum and reconstituted in 600 lL of deuterated chloroform (CDCl 3 ) containing 0.03% (vol/vol) tetramethylsilane (TMS). 1 H NMR spectra were acquired using a Bruker Avance NEO 600 MHz spectrometer equipped with an inverse cryogenic probe (Bruker Biospin) at 298 K. NMR spectra of the liver, cecal content, and urine samples were recorded with a typical 1D NMR spectrum named NOESYPR1D.The metabolites were assigned on the basis of a set of 2D NMR spectra and published results. 41,42All 1 H NMR spectra were corrected for phase and baseline distortions and quantified for hydrophilic metabolites using Chenomx NMR Suite 8.4 (Chenomx Inc.).The absolute quantification for lipid classed in liver were calculated according to the following expression, as previously reported. 19,41 = where Cx is the lipid concentration, Ix is the integral of the lipid proton peak, Nx is the number of protons in the lipid proton peak, Cs is the concentration of TMS as the internal standard, Is is the integral of TMS proton peak, Ns is the number of protons in TMS, V is the volume of the analyzed extract, and M is the weight of liver tissue analyzed.

LC-MS-Based Metabolomics Experiments
Lipid analysis of liver and bacteria samples was performed using a Vanquish UHPLC system (Thermo Fisher Scientific) connected to an Orbitrap Fusion Lumos Tribrid MS using a heated electrospray ionization (H-ESI) ion source (all Thermo Fisher Scientific) with a Waters charged surface hybrid (CSH) C18 column (1:0 × 150 mm, 1:7-lm particle size).Hydrophilic metabolites of mouse cecal and bacteria were performed with a Dionex Ultimate 3000 quaternary HPLC system connected to Exactive Plus Orbitrap MS (Thermo Fisher Scientific) with a Waters XSelect high strength silica (HSS) T3 column (2:1 × 100 mm, 2:5-lm particle size).The liver samples (20 mg) were extracted twice with 0:5 mL of precool isopropanol/water/ethyl acetate (30:10:60, vol/vol/vol) containing 1:1,000 EquiSPLASH (Avanti Polar Lipid).After homogenization and centrifugation (12,000 × g, 4°C) for 10 min, the combined supernatants were dried in a vacuum and reconstituted in 200 lL of isopropanol/acetonitrile/water (45:35:20, vol/vol/vol).After a centrifugation (12,000 × g, 4°C) for 10 min, 40 lL of supernatants were transferred to autosampler vials for LC-MS analysis.The hydrophilic metabolites and lipid of bacteria were extracted as previously described. 31Briefly, bacteria samples (1 mL) were extracted with 1 mL of precool chloroform/methanol (2:1, vol/vol).After homogenization and freeze-thaw, 250 lL of HPLC water was added to the samples and vortexed, followed by centrifugation (12,000 × g, 4°C for 10 min).The top phase was collected for hydrophilic metabolites and the bottom phase was collected for lipid analysis.Two phases were dried down in a vacuum and reconstituted in 60 lL of 3% methanol containing 1 lM chlorpropamide (hydrophilic metabolites) or 100 lL of isopropanol/acetonitrile/water (50:25:25, vol/vol/vol) containing 1:1,000 EquiSPLASH (lipids).LC-MS data were analyzed by the open-source software MS-DIAL. 43MetaMapp network analysis for bacteria hydrophilic metabolites was performed to visualize and integrate biochemical pathways and chemical similarities. 35,44eatmaps were plotted using R software package "pheatmap."

Fatty Acid Analysis by GC-MS
Quantitative measurements of liver fatty acids were performed as previously reported. 45,46Briefly, liver (10 mg) samples were homogenized in 500 lL of methanol containing fatty acid internal standards.After acetyl chloride-catalyzed methyl esterification, the identification and quantification of methylated fatty acids was performed on an Agilent Technologies 7890A-5975C GC-MS system equipped with a Rxi-5ms (Restek) capillary column (30 m, 0:25-mm ID, 0:25-lm film thickness).The fatty acid standards including the Suplelco 37 Component FAME Mix (Sigma-Aldrich) and two internal standards (C17:0 and C23:0) were used in this analysis.The relative amount of fatty acids was determined relative to internal standards.

Bacteria DNA Isolation and qPCR
The DNA of A. muciniphila and mouse cecal content samples was extracted using E.Z.N.A. stool DNA kit (Omega Bio-Tek Inc.) according to the manufacturer's protocol.Total bacteria and A. muciniphila quantification were calculated using a standard curve with the threshold (C T ) value vs. the number of bacterial CFUs. 47qPCR reactions were carried out with PowerUp SYBR Green Master Mix (Applied Biosystems) on a QuantStudio 3 real-time PCR system (Thermo Fisher Scientific).Primers can be found in Table S2.Reactions were performed using the cycle conditions (50°C, 2 min, 95°C, 2 min; 95°C, 1 s, 60°C, 30 s, for 45 cycles; 95°C, 15 s, 60°C, 1 min, 95°C, 15 s).

Metagenomic Analysis
Mouse cecal DNA samples were sequenced using NextSeq Mid-Output 150 × 150 paired-end sequencing.The resulting reads were processed using fastp (version 0.23.2) 48 with automatic adapter detection, poly G removal, and sliding window quality filtering.After quality control, the median sequencing depth was 2.69 Gb (range: 2.08-3.8).To obtain the highest possible assignment rates of mice taxa, taxonomic characterization was performed with MetaPhlAn4 (version 4.0.6) 49by incorporating species-level genome bins 50 with additional estimation of the unassigned fraction.The median unassigned percentage was 20.03 (range: 9.30-59.33);however, this did not vary significantly between treatment groups (p > 0:05 , Mann-Whitney U-test).To negate any potential effects of variable unassigned taxa, data were transformed by centered logratio (CLR) with zeros replaced with two-thirds of the lowest detected relative abundance on a per-sample basis, and ordinations performed using the Aitchison/CLR Euclidean distance. 51Statistical analysis of distance matrixes was performed using ADONIS2/ PERMANOVA function of vegan (version 2.6-4, R software package) with both treatment group and cage as terms. 52Statistical analysis of differential species abundance was performed using a Mann-Whitney U-test followed by a Benjamini-Hochberg false discovery rate (FDR) correction with significance determined as FDR <0:2.Gene family abundances were calculated using HUMAnN (version 3.5) against the UniRef90 Metacyc 25 databases. 53Gene family abundances were normalized to reads per thousand bases per genome equivalent (RPKG) by estimating genome equivalents using MicrobeCensus. 54Statistical testing of gene family abundances was performed using Welch's t-test with a Benjamini-Hochberg FDR cutoff of 0.2.

Transcriptomic Analysis
Total RNA was isolated from 7 mL of A. muciniphila culture by the addition of 1 mL of TRIzol, and the DNA was removed by RNase-Free DNase Set (QIAGEN).After purification with PureLink RNA Mini Kit (Invitrogen), total RNA was measured and checked by an Agilent Bioanalyzer.The 16S and 23S rRNA fractions were removed from total RNA using the RiboMinus Bacteria 2.0 Transcriptome Isolation kit (Invitrogen).The deletion of 16S and 23S rRNA was checked again with the Agilent Bioanalyzer.Depleted RNA samples were submitted to the Microbial Genome Sequencing Center for RNA sequencing.Library prep was performed using an Illumina Stranded RNA library kit and sequenced on a MiSeq to generate ∼ 12 million 50-bp paired-end reads per sample.
Obtained demultiplexed reads were checked for quality using Fastp 48 to remove bad quality reads, low complexity regions and adapters.Kraken2 was used to check for contamination of reads by other bacteria. 55Filtered reads were then aligned to A. muciniphila ATCC BAA-835 reference assembly (GCF_000020225.1_ASM2022v1)using bowtie2. 56Alignment qualities were checked using Qualimap. 57eatureCounts function of Rsubread package in R was used to count the reads in terms of coding sequences. 58Only uniquely mapped reads were considered for counting.Raw counts were then normalized and then analyzed for differential gene expression using the DESeq2 package. 59Further, visualization and analysis were performed in R (version 4.3.3,R Development Core Team) and GraphPad Prism (version 6.0; GraphPad).

Transcriptomic and Metabolomics Data Integration and Visualization
Metabolomics data and transcriptomics data were visualized using MetaboMAPS. 60

Sanger Sequencing
The DNA from A. muciniphila was confirmed by Sanger sequencing, as previously described. 35The result was analyzed with Nucleotide BLAST (BLAST +2.15.0: 31 October 2023) (Table S4).

Statistics
All data values are expressed as the mean ± standard deviation (SD) or median and interquartile range.Graphical illustrations and statistical analyses were performed using GraphPad Prism 6.0 (GraphPad).Two-way analysis of variance (ANOVA) with repeated measures was used for body weight gain.Two-tailed unpaired t-test analyses or Mann-Whitney U-test in vivo study and one-way ANOVA followed by Dunnett's multiple comparisons in vitro study were used.

Effects of TCDF Exposure in Early Life on Body Weight and Glucose Tolerance
Given the considerably shorter half-life of TCDF in rodents relative to TCDD, 15,16 TCDF levels in the liver of mice from long duration (3 months after final exposure) were below the limit of detection via GC-MS (Figure 1B).This observation was supported by data showing no significant different in the expression of liver and ileal AHR target genes Cyp1a1 and Cyp1a2, which have long been used as sensitive indicators of AHR activation for hazard identification and risk assessment of environmental pollutants, 61 at 3 months after exposure compared with significantly higher gene expression of those genes at short duration (the day after final exposure) (Figure 1C,D).The significantly higher ratios of GSSG to GSH and serum ALP levels were observed with TCDF exposure at the short-duration time point (Figure S1A,B).No overt toxicity was observed at the long-duration time point, supported by no significant differences in serum ALP and cytokines, expression of intestinal cytokine mRNA, and liver histopathology (Figure S1A-D and Table S5).No significant differences in food intake were observed in mice after TCDF exposure (Figure S2A).Interestingly, early life 5 d of TCDF exposure resulted in significantly higher levels of body weight and epididymal white adipose tissue (eWAT) later in life (Figure 1E,F; Figures S2B and S3).Impaired glucose tolerance was observed at 3 months after TCDF exposure (Figure 1G), which was associated with significantly higher levels of liver glycogen and uridine diphosphate glucose (UDP-glucose) (Figure 1H).

Effects of TCDF Exposure in Early Life on Liver Lipogenesis
Previous research in mice has found that TCDF exposure induces hepatic lipid accumulation, thus leading to metabolic-associated fatty liver disease (formerly called nonalcoholic fatty liver disease). 21To investigate the effects of early life TCDF exposure on liver lipid metabolism, we used 1 H NMR-and MS-based metabolomics along with conventional, GF, and Ahr −= − mice.The TG assay and quantitative 1 H NMR analysis showed significantly higher levels of liver lipids with TCDF exposure at the shortduration time point (Figure 2A,B).UHPLC-MS/MS global analysis also confirmed that 5 d of TCDF exposure immediately resulted in significant differences in liver lipid (Figure 2C).Subtle differences in liver profiles were observed with TCDF exposure at 3 months after exposure (Figure 2A-C), which was consistent with no obvious differences in hepatic fat accumulation (Oil Red O staining) or macrophages (F4/80-positive cells) from TCDF-exposed mice from the long-duration time point (Figure S4).To further explore the effect on lipid metabolism, fatty acid compositional analysis was performed using 1 H NMR and GC-MS.Consistently, quantitative 1 H NMR and targeted GC-MS analysis revealed significantly higher levels of hepatic fatty acids from TCDF-exposed mice at the short-duration time point (Figure 2D; Figure S5).TCDF did not result in significant differences in fatty acid profiling in liver at 3 months after exposure (Figure 2D; Figure S5).Consistent with the metabolomics data, mRNA expression of genes involved in de novo fatty acid biosynthesis was also significantly higher in the liver at the shortduration time point, but we noted few differences at the longduration time point after TCDF exposure (Figure 2E).To test whether these lipid changes by TCDF exposure could be specific to AHR activation or microbiota changes, we measured the liver lipid profile in Ahr −= − and GF mice with 5 d of TCDF exposure (Figure 2F-H; Figure S6).No significant differences in the lipid profile or mRNA expression of liver AHR target genes were observed in Ahr −= − mice with 5 d of TCDF exposure (Figure 2F-H).However, we observed significantly higher levels of liver lipids in addition to the induction of mRNA expression of AHR target genes in GF mice with 5 d of TCDF exposure (Figure S6).

Effects of TCDF Exposure in Early Life on the Gut Microbiome Composition and Function
To further explore the long-term influence of early life TCDF exposure on the gut microbiota, metagenomics and metabolomics analysis were performed.We assessed the influence of caging on microbiota composition in both short and long-duration experiments.The TCDF effect exhibited a stronger impact on microbiota composition than the cage effect in the long-duration model (long-duration: treatment p = 0:002, R 2 = 0:17; cage p = 0:165, R 2 = 0:19; short-duration: treatment p = 0:110, R 2 = 0:11; cage p = 0:044, R 2 = 0:25).Neither group had a difference in alpha diversity (Figure S7A).Principal coordinate analysis of cecal species-level abundances revealed significant effects of TCDF exposure on the microbiome composition only later in life (Figure 3A), whereas no significantly different species were observed in the short duration (FDR-corrected Mann-Whitney p < 0:2).Five days of TCDF exposure resulted in significantly lower levels in relative abundances of species A. muciniphila, Parasutterella excrementihominis, and Bifidobacterium pseudolongum and higher levels in relative abundance of species Methanomethylovorans SGB40959 at 3 months after exposure (Figure 3B).Analysis of differentially abundant gene families identified genes corresponding to amino acid, nucleotide, and carbohydrate metabolism that were significantly regulated by TCDF exposure later in life (Figure 3C).
We then sought to investigate the influence of early life TCDF exposure on microbial function (Figure 3D-F; Figure S7B,C).Significant differences in urinary bacterial metabolites and circulating LPS, the primary component of the outer membrane of Gram-negative bacteria, 18 were observed with TCDF exposure later in life (Figure S7B,C).Significantly lower levels of the cecal tryptophan metabolite indole-3-lactic acid (ILA) that is associated with intestinal inflammation 62 were observed with TCDF exposure later in life (Figure 3D; Table S6).TCDF exposure also resulted in significantly lower levels of SCFAs later in life and significantly lower mRNA expression for G proteincoupled receptors (GPCRs) including GPR41, GPR43, and GPR119 in colon tissues (Figure 3E SCFAs. 63,64The mRNA expression for GLP1 and PYY also showed significantly lower levels in colon tissues from mice with TCDF exposure later in life (Figure 3F).Next, we investigated whether this effect could be transferred to GF mice via cecal microbiota transplantation (Figure 3G-K; Figure S8).Two weeks after conventionalization with either cecal contents from control mice or those treated with TCDF, mice exhibited a stable gut microbiome but no significant differences in the body weight (Figure S8A,B).GF recipients from TCDFexposed mice exhibited significantly lower levels of cecal SCFAs and mRNA expression of GPCRs and gut hormones GLP1 and PYY in the colon after 4 wk of conventionalization (Figure 3H,I; Figure S9).Moreover, we observed significantly higher levels of liver UDP-glucose as well as mildly higher liver glucose (p = 0:06) and fasting blood glucose levels (p = 0:06) in GF recipients from TCDF-exposed mice after 4 wk conventionalization (Figure 3J,K).No significant differences in expression of liver AHR target gene or lipid profiling were observed in TCDF-transplantation mice after 4 wk conventionalization (Figure S8C-E).

Effects of TCDF Exposure on A. muciniphila and Its Production of Healthy Promoting Microbial Metabolites
Our findings revealed that early life TCDF exposure resulted in significantly lower abundances of cecal A. muciniphila at 3 months after exposure (Figure 4A,B; Figure S10), which was consistent with the significantly higher levels of fecal mucin (Figure S11).GF recipients from TCDF-exposed mice also exhibited significantly lower abundances of cecal A. muciniphila (Figure 4C).
Next, to examine the functional role of A. muciniphila in vivo, we colonized antibiotic-treated or nontreated GF mice with A. muciniphila (Figure 4D-H).Interestingly, GF recipients colonized with A. muciniphila activated AHR target gene expression in the colon (Figure 4E), but there were no significant differences in the liver and ileum (Figure S12).This is consistent with significantly higher levels of cecal tryptophan metabolites, including ILA, which are considered reliable sources of AHR ligands 13 (Figure 4F; Table S7).Furthermore, GF mice colonized with A. muciniphila exhibited significantly higher levels of cecal SCFAs and mRNA expression of GPCRs and gut hormones in the colon (Figure 4G,H).
To further investigate whether A. muciniphila could reverse the effects from TCDF exposure, we colonized GF mice with cecal transplants from mice treated with TCDF in the longduration model along with A. muciniphila supplementation (Figure 4I).The administration of A. muciniphila was effective in ameliorating the effects of TCDF, supported by significantly higher levels of cecal ILA and SCFAs, along with higher mRNA expression of the AHR target gene, GPCRs, and gut hormones in the colon (Figure 4J-M; Figure S13 and Table S8).

Effects of TCDF Exposure on the Metabolism and Gene Expression of A. muciniphila in Vitro
Next, we focused on the effect of TCDF on A. muciniphila in vitro.Consistent with the in vivo study, significantly slower growth of A. muciniphila was observed with two doses of TCDF (0:6 lM and 6 lM) exposure in vitro (Figure 5A).The high dose of TCDF (6 lM) resulted in significantly higher proportions of damaged bacteria (Pi-stained) and lower CFDA-stained cells (Figure 5B).No significant effects were observed in the proportions of SybrGreen-and DiBAC 4 -stained bacteria with the two doses of TCDF exposure (Figure S14).
To further explore the influence of TCDF on A. muciniphila, transcriptomics [RNA sequencing (RNA-Seq)], coupled with metabolomic analysis, were performed (Figure 5C-H).Principal component analysis of the RNA-seq data showed distinct separation between A. muciniphila with vehicle and two doses of TCDF (0:6 lM and 6 lM) (Figure 5C).Among the 615 significantly different genes, 35 genes were down-regulated, and 12 genes were up-regulated with fold changes >1:5 in A. muciniphila with the higher dose of TCDF (6 lM) compared with vehicle (Figure 5D).Gene orthologs involved in amino acid metabolism, carbohydrate metabolism, nucleotide metabolism, membrane integration, vitamin metabolism, and translation pathways were most significantly lower in A. muciniphila with the two doses of TCDF exposure (0:6 lM and 6 lM) (Figure 5E).Of particular note, the significant down-regulation of the A. muciniphila-derived ILA pathway was observed with the higher dose of TCDF (6 lM) compared with vehicle (Figure 5F; Table S9).UHPLC-MS-based global metabolomic analysis revealed markedly lower levels in carbohydrate, amino acid, and nucleotide metabolism with the two doses of TCDF exposure (Figure 5G).Moreover, significantly higher levels in bacterial membrane lipids, including carnitine (CAR), ceramide (Cer), sphingolipid (SL), lysophosphatidylcholine (LPC), sphingomyelin (SM), diglycerol (DG), and TG, were observed with the two doses of TCDF exposure (Figure 5H).

Discussion
This study provides evidence suggesting that early life disruptions in the gut microbiome due to environmental pollutants are crucial mechanisms linking the association between environmental pollutant exposure and metabolic disorders later in life.Infants and children may experience increased exposure to environmental chemicals compared with adults who may have been exposed in utero and via breast milk, [65][66][67][68] but only limited attention has been given to the subsequent outcomes. 69Our work in mice demonstrates that early life exposure to an environmental pollutant may increase the risk of metabolic disorders later in life through disruption of the gut microbiome (Figure S15).
The establishment of the microbiome in early life is gaining appreciation as a major influencer in human development and long-term health outcomes. 70,71Accumulating evidence suggests that microbiota disruptions caused by antibiotic exposure in early life can have long-lasting effects on host health and significantly increase the risk of developing obesity and associated metabolic disorders later in life. 30,72Currently, POPs are increasingly recognized as influential modulators of the gut microbiota. 18,20,31e did not observe significant differences in food intake or overt toxicity in TCDF-exposed during the long duration, factors that could potentially impact the gut microbiota. 73,74We observed a long-lasting impact on the gut microbiota due to early life TCDF exposure.For example, the lower mRNA expression for GPCRs and gut hormones, even though there were no significantly different species in the short-duration experiment, suggests impaired gut microbiota function after 5 d of early life TCDF exposure.A similar observation has been reported in antibiotic studies, demonstrating long-term effects of early life antibiotic use on gut microbiota in adulthood even in the absence of drug exposure in adulthood. 30,757][78] Increased levels of circulating LPS and decreased levels of cecal ILA later in life were observed with early life TCDF exposure, suggesting intestinal inflammation and immune dysregulation. 6,18ur data indicate that the disruption of the gut microbiome by early life TCDF exposure is associated with impaired glucose homeostasis later in life.Altered gut microbiota composition and function have been associated with obesity and obesity-related metabolic disorders in both animals and humans. 79,80Recent in vivo and in vitro studies suggest that SCFAs, the primary Environmental Health Perspectives 087005-10 132(8) August 2024 metabolites of the gut microbiota, may regulate host energy metabolism by increasing gut hormones, such as PYY and GLP-1, via the activation of SCFA receptors. 81,82Changes in the microbiome community, reductions in colonic SCFAs levels, inhibition of GLP-1 signaling, and reduced insulin sensitivity have been reported with antibiotic treatment. 83Similarly, we observed decreased levels of cecal SCFAs, as well as lower expression of SCFA receptors and gut hormones GLP-1 and PYY later in life with early life TCDF exposure, likely contributing to impaired glucose homeostasis and the development of obesity. 84Remarkably, we observed that this effect could be transferred to GF mouse recipients through microbiota transplantation from TCDF-exposed mice.Our study provides evidence in a mouse model linking disturbances in the gut microbiota with early life environmental pollutant exposure and the subsequent development of metabolic disease.This study emphasizes the importance of the microbiome as a key target of early life environmental pollutant exposure.TCDF serves as a model for a variety of persistent AHR ligands, and assessing the risk of early life exposure needs to consider all of these compounds in future studies.
There is increasing discussion regarding A. muciniphila as a beneficial contributor to gut health and host glucose homoestasis. 77,78,85Supplementation with A. muciniphila in overweight and obese humans has shown improvement in several metabolic parameters in clinical trials. 85,86In this work, we observed significantly lower levels of A. muciniphila associated with TCDF exposure in both in vivo and in vitro models.Consistently, reductions in carbohydrate, amino acid, and nucleotide metabolism (at the transcriptional and metabolic levels in vitro) were also observed with A. muciniphila with TCDF exposure, even at a low concentration (0:6 lM).Profound changes in the lipid profile of A. muciniphila were observed with TCDF exposure, indicating disturbances in bacterial membrane function. 31Specifically, TCDF exposure significantly down-regulated the A. muciniphila-derived ILA pathway in vitro, which is consistent with in vivo findings of significantly lower levels of cecal ILA with early life TCDF exposure.The presence of ILA has recently been reported in breastfed infants colonized with Bifidobacterium infantis, and this microbial tryptophan metabolite has been shown to decrease enteric inflammation through the activation of AHR. 62,87However, relatively little is known about which and how other microbiota regulate this antiinflammatory metabolite.We have demonstrated that A. muciniphila is capable of producing ILA, which is negatively correlated with TCDF exposure both in vivo and in vitro.A more comprehensive understanding of the dynamics of microbial tryptophan catabolites and their functional implication during early life environmental exposure is needed.Our data suggests that the significant impact of TCDF exposure to A. muciniphila could be a contributing factor that mediates the relationship between early life TCDF exposure and the potential increased risk of metabolic disorder later in life.
The present study has several limitations.TCDF serves as a model for various AHR ligands, and future studies assessing the risk of early life exposure need to consider the combined influence of chemical mixtures.Moreover, additional studies using Ahr-null mice and humanized AHR mice may be needed to determine whether microbial tryptophan catabolites, such as ILA, exert their effects through AHR during early life environmental exposure.In addition, more studies should consider sex-specific differences in the composition of gut microbiota, as well as the possible mechanisms involved.Last, we recognize, however, that sample size and the dosing regimen could be modified to capture subtle changes in the gut microbiome community structure and more closely resemble typical human doses, respectively.
Our work demonstrates that early life exposure to TCDF in mice leads to a persistent disruption in the gut microbiome community structure and, importantly, its function, which was associated with the impaired metabolic homeostasis later in life.We find TCDF exposure reduced a potentially beneficial microbe, A. muciniphila, and that reduction was observed in in vivo and in vitro models and that the effects of TCDF could be reversed with A muciniphila supplementation.This study provides new insight into microbiota-derived cues that drive the development of metabolic diseases later in life.

Figure 1 .
Figure 1.Impacts of early life TCDF exposure on body weight and glucose tolerance.(A) Experimental schematic for determining the short-and long-duration effect of early life TCDF exposure.(B) TCDF levels in the liver of mice after exposure were detected via quantitative GC-MS analysis (n = 6).(C,D) qPCR analysis of AHR targeted genes in the ileum (C) and liver (D) after TCDF exposure (n = 6).(E,F) Body weight (E) and epididymal white adipose tissue (eWAT) weight (F) after TCDF exposure (n = 6).(G) Blood glucose levels following intraperitoneal injection of glucose and area under the curve (AUC) at 3 months after TCDF exposure (n = 9 ).(H) Targeted 1 H NMR analysis of liver glucose, glycogen, and UDP-glucose with TCDF exposure (n = 6).p-Values were calculated using two-way ANOVA with repeated measures (E), two-tailed unpaired t-test or Mann-Whitney U-test.Data in all graphs are presented as means ± SDs.The numeric data are shown in Excel TablesS1-S6.Note: AHR, aryl hydrocarbon receptor; ANOVA, analysis of variance; GC, gas chromatography; LC, liquid chromatography; MS, mass spectrometry; NMR, nuclear magnetic resonance; qPCR, quantitative polymerase chain reaction; SD, standard deviation; TCDF, 2,3,7,8-tetrachlorodibenzofuran; UDP-glucose, uridine diphosphate glucose.
1 mL of sterile BHI Pathways were condensed into a custom figure.The enzymes were mapped to genes.The concentration of