Identification of polychlorinated biphenyls (PCBs) and PCB metabolites associated with changes in the gut microbiome of female mice exposed to an environmental PCB mixture.

Identification of polychlorinated biphenyls (PCBs) and PCB metabolites associated with changes in the gut microbiome of female mice exposed to an environmental PCB mixture - PMC Skip to main content An official website of the United States government Here's how you know Here's how you know Search Log in Primary site navigation Search Log in Search PMC Full-Text Archive Search in PMC Journal List User Guide As a library, NLM provides access to scientific literature. Inclusion in an NLM database does not imply endorsement of, or agreement with, the contents by NLM or the National Institutes of Health. Learn more: PMC Disclaimer | PMC Copyright Notice J Hazard Mater . Author manuscript; available in PMC: 2026 Jun 5. Published in final edited form as: J Hazard Mater. 2025 Feb 20;489:137688. doi: 10.1016/j.jhazmat.2025.137688 Identification of polychlorinated biphenyls (PCBs) and PCB metabolites associated with changes in the gut microbiome of female mice exposed to an environmental PCB mixture Laura E Dean 1, † , Hui Wang 1, † , Xueshu Li 1 , Rachel L Fitzjerrells 2, 3 , Anthony E Valenzuela 4 , Kari Neier 5 , Janine M LaSalle 5 , Ashutosh Mangalam 6 , Pamela J Lein 4 , Hans-Joachim Lehmler 1, * Author information Article notes Copyright and License information 1 Department of Occupational and Environmental Health, University of Iowa, Iowa City, IA 52242, USA 2 Interdisciplinary Graduate Program in Informatics, University of Iowa, Iowa City, IA 52242, USA 3 College of Dentistry, University of Iowa, Iowa City, IA 52242, USA 4 Department of Molecular Biosciences, University of California, Davis, CA 95616, USA 5 Department of Medical Microbiology and Immunology, University of California, Davis, CA 95616, USA 6 Department of Pathology, University of Iowa, Iowa City, IA 52242, USA * Corresponding Author: Dr. Hans-Joachim Lehmler, The University of Iowa, Department of Occupational and Environmental Health, University of Iowa Research Park, #221 IREH, Iowa City, IA 52242-5000, Fax: (319) 335-4290, hans-joachim-lehmler@uiowa.edu † Both authors contributed equally to this manuscript. Issue date 2025 Jun 5. PMC Copyright notice PMCID: PMC12002644  NIHMSID: NIHMS2061747  PMID: 40020572 The publisher's version of this article is available at J Hazard Mater Abstract Polychlorinated biphenyls (PCBs) are neurotoxic hazardous materials that may cause toxicity via the gut-liver-brain axis. This study investigated PCB × microbiome interactions in adult female mice exposed orally to an environmental PCB mixture. Female mice (6-week-old) were exposed daily for 7 weeks to peanut butter containing 0, 0.1, 1, or 6 mg/kg/day of PCBs. Twenty hours after the final exposure, the cecal content was collected to characterize the microbiome composition and predicted function. PCB and its metabolites in feces were analyzed using gas chromatography-tandem mass spectrometry (GC-MS/MS), while cecal content was assessed with liquid chromatography-high resolution mass spectrometry (LC-HRMS). MARBLES exposure influenced the abundance of microbial taxa and predicted functions within the cecal content. Complex PCB and metabolite mixtures were detected in the gastrointestinal tract. Network analysis revealed associations between specific parent PCBs and metabolites with changes in the abundance of bacteria in the gastrointestinal tract. These findings demonstrate that individual PCBs and their metabolites significantly influence the abundance of specific bacteria in the gastrointestinal tract following oral PCB exposure. These findings inform further research targeting the microbiome to attenuate the adverse health outcomes of PCB exposure. Keywords: MARBLES mix, Polychlorinated Biphenyls, Metabolites, Microbiome, multi-omics, integration Graphical Abstract 1. Introduction Polychlorinated biphenyls (PCBs) are a group of synthetic organic chemicals. Due to their chemical stability, non-flammability, and insulating properties, PCBs were widely used in industrial applications, including coolants and lubricants in electrical equipment, hydraulic systems and sealants [ 1 - 3 ]. Their production was banned in the United States in the late 1970s, and by most other countries by the early 2000s due to environmental and health concerns [ 4 , 5 ]. However, many PCBs are still produced inadvertently [ 6 ]. PCBs persist in the environment because of their resistance to degradation [ 2 , 7 ]. Their accumulation in aquatic and terrestrial food chains leads to widespread and ongoing human exposures [ 8 - 10 ]. PCB concentrations have been measured in many foods, with the highest PCB concentrations found in salmon, canned tuna, and beef steak [ 8 , 10 ]. Dietary PCB exposures can vary significantly between individuals depending on the type of foods they consume; those who consume a higher-fat diet have a higher exposure to PCBs [ 8 , 10 ]. Cytochrome P450 enzymes metabolize PCBs to hydroxylated metabolites (OH-PCBs). The formation of OH-PCBs has been studied extensively in various species, for example, guinea pigs, hamsters, mice, rabbits, rats, and pigeons [ 11 - 17 ]. The rates of PCB metabolism vary across species, with different cytochrome P450 isoforms contributing to the formation of OH-PCBs, potentially resulting in complex mixtures of different OH-PCB congeners. OH-PCBs can undergo further oxidative metabolism or conjugation reactions to form complex mixtures of sulfated and other metabolites that are typically eliminated with the feces [ 18 - 20 ]. Across species, PCBs are metabolized not only by the liver but also within the intestine, where PCB sulfates can be deconjugated via sulfatases to form OH-PCBs [ 21 , 22 ]. PCBs and their metabolites are associated with a range of adverse health effects, including neurotoxicity, endocrine disruption, and immune system impairment [ 19 , 23 , 24 ]. The gut microbiome, the complex community of trillions of microorganisms that live in the gastrointestinal tract [ 25 ], is a possible target of PCBs. It plays a pivotal role in maintaining the host’s health by contributing to essential physiological processes, including digestion, nutrient absorption, immune system regulation, and protection against pathogens [ 26 , 27 ]. Alteration in microbial communities (dysbiosis), has been linked to various diseases, including diabetes and cardiovascular diseases [ 26 ]. Growing evidence demonstrates the existence of the gut-liver-brain axis, a complex and bidirectional communication network linking the gastrointestinal system, liver, and brain [ 28 , 29 ]. Microbial imbalances have been shown to contribute to the pathophysiology of psychiatric disorders, including neurodevelopmental disorders [ 30 ]. Several factors, such as antibiotic use, diet, stress, and exposure to environmental pollutants, including PCBs, have been linked to dysbiosis [ 25 , 31 ]. Emerging research demonstrates that oral PCB exposure significantly alters the gut microbiome in animal models, leading to changes in microbial diversity and abundance [ 32 - 35 ]. These alterations can impact the ability of the gut to maintain homeostasis and perform essential functions, resulting in a disruption of the gut barrier, enhanced inflammatory responses, and altered immune system function [ 36 , 37 ]. For instance, changes in microbial abundances were observed alongside an impairment of the intestinal mucosal barrier due to the induction of inflammation, initiation of autophagy, and activation of apoptosis after exposure to PCB52 [ 32 ]. Furthermore, PCB-mediated effects on the gut microbiome can affect other host organ systems, including the liver, potentially affecting neurotoxic outcomes via the gut-liver-brain axis [ 38 - 40 ]. Characterizing the interactions between PCB exposure and the gut microbiome is essential for determining the mechanisms underlying PCB-induced neurodevelopmental and other psychiatric disorders. However, critical knowledge gaps remain regarding the specific PCB congeners and metabolites that alter the intestinal microbiome. Our study leverages samples generated as part of the Markers of Autism Risk in Babies-Learning Early Signs (MARBLES) study investigating PCB developmental neurotoxicity [ 41 - 46 ] to identify individual PCB congeners or metabolites associated with alterations in the abundance of specific bacteria and their predicted functions in female mice following oral exposure to a human-relevant PCB mixture. This work lays the groundwork for future mechanistic studies, with the long-term goal of targeting the microbiome to prevent adverse effects on the brain caused by exposure to complex PCB mixtures. 2. Materials and Methods 2.1. Chemicals and Analytical Standards CAUTION: For safety reasons, PCBs and dichloromethane, group 1 human carcinogens, and diazomethane, a toxic and explosive derivatization reagent, must be handled in a chemical fume hood with appropriate personal protective equipment. The PCB nomenclature used in this manuscript is based on the US EPA Table of PCB congeners [ 47 ]. The full names of OH-PCBs used in this manuscript is adapted from the nomenclature proposed by Maervoet and co-workers [ 48 ]. An abbreviated version is also provided, where the first number indicates the position of the hydroxy group on the biphenyl moiety, and the second number reflects the corresponding PCB congener (e.g. 4-52 indicates 4-OH-PCB52). The abbreviations and unique identifiers for the OH-PCBs are detailed in Table S1 . A detailed description of the sources of key chemicals and analytical standards ( Table S2 and Table S3 ) is provided in the Supplementary Material . 2.2. Preparation of the MARBLES mix The MARBLES mix (Sethi et al., 2019) used for dosing was composed of 12 different congeners (% of total mass): 28 (48.2%), 11 (24.3%), 118 (4.9%), 101 (4.5%), 52 (4.5%), 153 (3.1%), 180 (2.8%), 149 (2%), 138 (1.7%), 84 (1.5%), 135 (1.3%), and 95 (1.2%). The PCB congeners used to make the MARBLES mix were synthesized and authenticated as previously described [ 46 , 49 ]. The MARBLES PCBs were dissolved in peanut oil. Aliquots of this PCB solution were homogeneously mixed with peanut butter using a bullet blender (Next Advance, Troy, NY, USA) to obtain the PCB dosing formulations [ 50 ]. Control peanut butter was prepared by mixing peanut oil with peanut butter. 2.3. Animals and Exposure All procedures involving animals were conducted as part of a larger PCB developmental neurotoxicity study in accordance with the NIH Guide for the Care and Use of Laboratory Animals and approved by the University of California Davis Animal Care and Use Committee. Details regarding the animal study have been reported in a series of publications following the ARRIVE guidelines [ 41 , 42 , 45 , 50 , 51 ]. Briefly, approximately 6-week-old female mice that were hybrid 75%C57Bl/6J/ 25% SVJ129, as determined by single nucleotide polymorphism analysis, were group housed in clear plastic cages with corncob bedding. This wildtype line matched the genetic background of the transgenic mice used in the earlier studies. Mice were maintained on a 12 h light/dark cycle at 22 ± 2°C and 40-50% humidity with food (Diet 5058, LabDiet, Saint Louis, MO, USA) and water available ad libitum . Mice were exposed orally every day for 7 weeks to the MARBLES mix at low (0.1 mg/kg/d), medium (1 mg/kg/d), or high (6 mg/kg/d) doses, or a sham exposure with peanut butter only (0 mg/kg/d). Approximately 20 hours after the final PCB exposure, mice were euthanized via CO2 asphyxiation, followed by blood collection via cardiac puncture and transcardial perfusion with cold phosphate buffered saline (PBS). Cecal content was collected and stored at −80 °C for DNA isolation for microbiome analyses and PCB metabolite extraction for subject-screening liquid chromatography-high resolution mass spectrometry (LC-HRMS) analyses. Fresh feces pellets were collected from the bedding and stored at −80°C for the PCB and OH-PCB analysis using gas chromatography-tandem mass spectrometry (GC-MS/MS). 2.4. Microbiome Analysis DNA was isolated from cecal content utilizing the DNeasy PowerLyzer PowerSoil Kit (Qiagen, Germantown, MD, USA) following the manufacturer's instructions. A Nanodrop (Thermo Scientific, Waltham, MA, USA) was used to quantify DNA, which was kept at −80°C until further analysis. 16S rRNA sequencing was performed with the MiSeq platform by amplifying the V3-V4 region. Fastq files were used for downstream analysis, conducted using dada2 script in R [ 52 , 53 ] to generate amplicon sequence variants (ASVs). The ASV table was then further analyzed through MicrobiomeAnalyst [ 54 , 55 ]. Features were filtered based on the inter-quantile range with a minimum count of 200, a prevalence in the sample equal to 20%, and a percentage to remove of 10%. Data were transformed by the relative log expression before further analyses. Alpha diversity was measured by Shannon Diversity at the genus level. Principal Coordinate Analysis (PCoA) and Jensen-Shannon Divergence at the feature level were used to measure beta diversity. Random forest analyses (5000 trees, 7 predictors, randomness on) and linear discriminant analysis effect size (LEfSe) were performed to identify differentially abundant taxa [ 56 ]. In addition, Phylogenetic Investigation of Communities by Reconstruction of Unobserved States (PICRUSt) analysis was used to predict the functional profile of the microbial communities [ 57 ]. 2.5. Topic Modeling Topic modeling was performed in R (version 4.3.1.) [ 58 ] as previously described [ 59 ]. Briefly, a phyloseq (version 1.44.0) [ 60 ] object was built using microbiome data and metadata. There were initially 283 features and 25 remained for analysis post filtering (prevalence < 1×10 −5 ). Each of the three exposure groups (low, medium, and high) underwent the following assessment: The ideal topic number for each group was assessed based on various metrics described previously [ 61 , 62 ] using the ldatuning (version 1.0.2) [ 63 ] package. To build the topics, the method "VEM" was chosen to perform variational inference. Next, the ideal number of topics was derived for each exposure group with latent Dirichlet allocation using the topicmodels citation package [ 64 , 65 ]. A document-term matrix was created with this topic information. A new phyloseq object was built utilizing this table to assess sample assignment to topics. Lastly, to assess which topics were more often assigned to samples from exposure groups, LinDA was utilized from the MicrobiomeStat package [ 66 , 67 ]. The parameters were feature.dat.type = "count", is.winsor = F, p.adj.method = “BH”. Topics were significant if they reached a p-value of ≤ 0.05 and an FDR of ≤ 0.25. 2.6. Targeted GC-MS/MS analysis of PCBs and OH-PCBs from Feces 2.6.1. Extraction of PCBs and OH-PCBs Aliquots of feces (40±6 mg) were homogenized in 3.0 mL of 0.2 M sodium acetate buffer (pH=5) using a TissueRuptor (Qiagen, Hilden, Germany) in a glass tube. For the method blank samples, 50 μL of Milli-Q water was added. Each sample was spiked 60 ng of 3-F,4'-PCB3 sulfate (in dimethyl sulfoxide (DMSO), serving as a surrogate standard for PCB sulfate) and 50 μL of sulfatase (type H-2 from Helix pomatia , Sigma-Aldrich, Burlington, MA, USA). The mixtures were incubated at 37 °C for 16 hours in a shaking water bath to deconjugate PCB sulfate and glucuronide metabolites. Samples were subsequently kept at 110 °C for 10 min to denature protein. Surrogate standards for PCBs (PCB15 and PCB117, 10 ng each in isooctane) and OH-PCBs (4'-9, 4-91, and 4'-159, 10 ng each in methanol) were spiked into all samples, followed by the addition of 1 mL KCl (1%, w/w), 1 mL of 6 M HCl, 4 mL of 2-propanol, and 4 mL of hexane:methyl tert-butyl ether (MTBE) mixture (1:1, v/v). The samples were inverted for 5 min and centrifuged at 1,811 g for 5 min to facilitate phase separation. The upper organic phase was transferred to a new glass vial containing 4 mL of 1% (w/w) KCl. The lower aqueous phase was re-extracted with 3 mL of hexane and the organic layer was combined with the previous organic layer. The combined mixture was then invertedfor 5 min and centrifuged at 1,811 g for another 5 min. The upper organic phase was transferred to a new glass vial, and the lower aqueous phase was re-extracted with 3 mL of hexane. The combined organic phase was evaporated under a gentle stream of nitrogen to about 0.5 mL. OH-PCBs in the extracts were derivatized to the corresponding methoxy (MeO)-PCBs by adding 5 drops of methanol and 0.5 mL of diazomethane (about 5 mmol) in diethyl ether [ 68 ] at 4 °C for approximately 16 h. The solvent and excess diazomethane were evaporated to near dryness under a gentle stream of nitrogen in a chemical fume hood. After further clean-up using 2 mL of 2-propanol and 2 mL of 0.1 M tetrabutylammonium hydrogen sulfate (TBA), the extracts were subjected to a solid phase extraction (SPE) clean-up with acidified silica gel (sulfuric acid:silica gel = 1:5, w/w). Once the extracts were added to the cartridge, the PCB and metabolites were eluted with 14 mL of dichloromethane (DCM). The sample eluents were then concentrated to near dryness, and 3 mL of hexane was added to reconstitute until concentrated to a final volume of 50 μL. The extracts were spiked with the internal standard (d-PCB30 and PCB204) and transferred into a GC vial with an insert for GC-MS/MS analysis. 2.6.2. GC-MS/MS Instrumental Analysis of PCBs and OH-PCBs PCBs and OH-PCBs (as methylated derivatives) were quantified simultaneously using a GC-MS/MS system (Agilent 7890B GC system, Agilent 7000D Triple Quad, and Agilent 7693 autosampler; Agilent Technologies, Santa Clara, CA, USA) equipped with a Supelco SPB-Octyl capillary column (50% n-octyl/50% methyl siloxane, 30 m length, 0.25 mm inner diameter, and 0.25 μm film thickness; Supelco, Bellefonte, PA, USA). Helium was used as the carrier gas (flow rate of 0.8 mL/min) and nitrogen as the collision gas. The gas chromatograph was operated in the solvent vent injection mode with the following setup: an initial temperature of 45 °C, for 0.06 min, a ramp of 600 °C/min to reach an inlet temperature of 325 °C at 5 psi. The GC oven temperature program was set to 45 °C for 2 min, then increased from 45 °C to 75 °C at a rate of 100 °C/min, hold for 5 min, increased from 75 to 150 °C at 15 °C/min, hold for 1 min, and finally increased from 150 °C to 280 °C at 2.5 °C/min, and hold for 5 min. The triple quadrupole electron ionization source temperature was set to 280 °C. A list of precursor ions, product ions, and collision energies for each analyte is provided in Table S4 . The levels of PCBs and OH-PCBs were determined by the relative chromatographic peaks area to internal standard and corrected by surrogate recoveries. 2.6.3. Quality Assurance and Quality Control for GC-MS/MS Analyses Solvent blanks, method blanks, and ongoing precision and recovery (OPR) standards were extracted in parallel with all samples to assess the reproducibility and precision of the analyses. PCB and OH-PCB levels were adjusted for the recoveries of the respective surrogate standard. The method detection limit (MDL) was calculated using the following formula: MDL = mean blank + t (0.01, n-1) * SD blank , where mean blank represents the average of the method blanks, t (0.01, n-1) is the Student's t-value for n–1 degrees of freedom at a confidence level of 99%, and SD blank is the standard deviation of the method blanks. A similar formula was used to calculate the limits of detection (LOD): LOD = meancontrol + t (0.01, n-1) * SD control , where the mean control and SD control represent the average and standard deviation, respectively, using control (or sham) tissue samples. The OPR recoveries, surrogate standard recoveries, MDLs, and LODs are reported in Table S5 - S7 . 2.7. Analysis of PCB Metabolites in Cecal Content Using Subject-Screening LC-HRMS 2.7.1. Extraction of PCB Metabolites from Cecal Content An aliquot of cecal content (30±7 mg, n = 25) was extracted and analyzed using LC-HRMS to identify PCB metabolites as described previously [ 69 , 70 ]. In brief, 30 mg of cecal content was mixed with 2 mL of Milli-Q water in glass tubes. Surrogate standards, 3-F,4'-OH-PCB3 and 3-F,4'-PCB3 sulfate (10 ng each), were spiked into the mixtures. Then, 4 mL of acetonitrile with 1 % (v/v) formic acid was added and vortexed. Two hundred mg of NaCl and 800 mg of MgSO 4 were added into the tubes, followed by vigorous shaking. The tubes were centrifuged at 1,811 g for 5 min. The organic phases were passed through commercial SPE cartridges (3 mL, HybridSPE, Sigma-Aldrich, St. Louis, Missouri, USA) filled with a 3 g mixture of MgSO 4 and Na 2 SO 4 (1:1, w/w). The extracts were concentrated to dryness using a Savant SpeedVac SPD vacuum concentrator (Thermo Scientific, Waltham, Massachusetts, USA) at 35 °C. Three hundred μL acetonitrile was added to reconstitute the extracts. Extracts were centrifuged at 1,811 g for 5 min, and the supernatants were transferred to a microcentrifuge tube. The solvents were evaporated to dryness again using SpeedVac at 35°C. The resulting extracts were reconstituted with 200 μL of water:acetonitrile (1:1, v/v). Samples were spiked with perfluorooctanesulfonic acid potassium salt (PFOS, 10 ng) as volume corrector. Finally, the samples were kept at −20 °C for at least 30 min, centrifuged repeatedly at 4 °C for 10 min at 16,000 g, and stored the supernatant at −80 °C before LC-HRMS analysis. 2.7.2. LC-HRMS Analysis of Extracts from Cecal Content PCB metabolites from cecal content extracts were analyzed at the University of Iowa High-Resolution Mass Spectrometry Facility on a Q-Exactive Orbitrap mass spectrometer (ThermoFisher Scientific, Waltham, MA, USA) with a Vanquish Flex ultra-high-performance liquid chromatograph (UHPLC, ThermoFisher Scientific) equipped with an Acquity UPLC-C18 column (1.7 μM particle size, 2.1 mm inner diameter, 100 mm length, Waters, Milford, MA, USA). The mobile phase used a mixture of water (A) and acetonitrile (B), with a flow rate of 0.3 mL/min. The LC system operated at a pressure range of 4000 to 8000 psi and used the following gradient program to separate the analytes: starting at 5% B (acetonitrile), hold for 1 min, increase linearly to 95% B, hold for 3 min, and returning to 5% B, with a hold for 4 min before the next injection. The sample injection volume was 2 μL. The Orbitrap was operated in the negative electrospray ionization mode, with a current of 18.2 μA and a spray voltage of 2,472 V. The auxiliary gas flow rate was 2 mL/min, and the sheath gas flow rate was 48 mL/min. The auxiliary and capillary temperatures were 413 °C and 256 °C, respectively. Analyses were performed in the full scan mode with a range of 85 to 1000 m/z . The maximum interval time was 200 ms, the full resolution was set to 70,000, and the autogain control target was 1×10 6 . 2.7.3. Processing of LC-HRMS data The chromatographic peaks of suspected PCB metabolites were extracted from the acquired data files (.raw) using Thermo Xcalibur (version 4.1, Thermo Fisher Scientific). The extraction was conducted using the following criteria: an m/z tolerance of 5 ppm, five decimal places for mass precision, and a smoothing factor of 7. The isotopic pattern of chlorine was utilized to confirm the metabolite identification as a PCB metabolite [ 69 ]. The chromatographic and mass spectrometric data were then exported. The relative PCB metabolite levels are expressed as the peak area normalized to the peak area of PFOS [ 71 ]. Only metabolites with more than 50% detection rate in each group are reported. 2.7.4. QA/QC for LC-HRMS analyses. Solvent (50% water/50% acetonitrile) and method blanks were used in the LC-HRMS to monitor carryover and assess background contamination. No PCB metabolites were detected in these blank samples. 3-F,4'-OH-PCB3 and 3-F,4'-PCB3 sulfate were added to all samples to assess the recoveries, reproducibility and precision of the analyses. The recoveries of both surrogate standards were 53±11% (range 31% to 70%, n=30) for 3-F,4'-OH-PCB3 and 60±12% (range 34% to 95%, n=30) for 3-F,4'-PCB3 sulfate across all samples. 2.8. Statistical Analyses Log fold change and p-values based on a Mann-Whitney U test with Holm-Sidak multiple comparisons were calculated based on the microbial ASV data of the features identified by random forest analysis (MDA > 0.01) and PICRUSt analyses. PCB and OH-PCB levels are expressed as mean ± standard deviation. PCB and OH-PCB mass profiles were compared using the similarity coefficient θ, where cos θ = 1 indicates that profiles are identical and cos θ = 0 indicates that profiles are entirely different [ 72 , 73 ]. A Pearson correlation analysis was conducted to compare the correlation of microbial abundance and PCB and metabolites levels (both targeted and nontargeted results; Table S8 ). 2.9. Network analysis of PCB and metabolite levels and microbe abundance A comprehensive network analysis was performed using xMWAS (v. 0.552) [ 74 ] to explore the interactions between PCB and PCB metabolite levels in the cecum or feces and microbial abundance in the cecum [ 75 ]. xMWAS calculates pairwise association scores, which approximate correlation coefficients, between each PCB or PCB metabolite and individual microbes. This is achieved through partial least-squares (PLS) regression analysis. A multi-data integrative network is then generated [ 76 ], incorporating only associations with an |association score| > 0.4 and a p-value < 0.05 (determined by Student's t-test). A multilevel community detection algorithm [ 77 ] is applied to refine the network further, identifying clusters of PCBs and PCB metabolites and microbes that showed strong internal connections but sparse links to other network components. Finally, Cytoscape (v. 3.10.1) was used to visualize the resulting correlation networks [ 78 , 79 ]. 3. Results and Discussion 3.1. Microbial diversity within or between exposure groups No significant changes in alpha diversity were observed based on Shannon Diversity ( Figure S1A ). The beta diversity analysis also indicated that microbial populations did not differ between exposure groups ( Figure S1B ). Similarly, the microbiota populations showed no difference in beta diversity across exposure groups in wild-type mice exposed to the MARBLES mix via the maternal diet [ 50 ] or in wild-type rats exposed to PCB52 via nose-only inhalation [ 75 ]. However, oral PCB exposure to various PCBs has been shown previously to alter alpha and beta diversity in both mice and rats [ 33 , 80 - 84 ]. 3.2. Effect of the MARBLES mix on microbial abundance in the cecum Random forest revealed two phyla, three classes, three orders, three families, three genera, and two species as important in distinguishing exposure groups ( Figure 1A ). These taxa were then compared using Mann-Whitney U test to determine significant changes based on dose. Figure 1. Open in a new tab Exposure to different doses of the MARBLES mix affects the gut microbiome of mice, as assessed by (A) random forest analysis, (B) KEGG pathway analysis based on predicted functional changes, and (C) topic modeling. (A) Taxa at the phylum, class, order, family, genus, and species were significantly altered by exposure to the MARBLES mix at all PCB doses. The abundance data were analyzed by random forest followed by the Mann-Whitney U test and are plotted as the fold change compared to controls. (B) Nineteen different pathways were predicted to be significantly altered compared to controls across all three exposure groups based on PICRUSt analysis followed by Kegg pathway analyses. (C) Topic modeling revealed a microbiome community significantly more often found in the control group compared to mice with high exposure to PCBs. The x-axis represents the probability of bacterial assignment to community. The y-axis depicts the top 20 assigned bacteria. P, phylum; C, class; O, order; F, family; G, genus; S, species. The order Lactobacillales was significantly decreased after exposure to the high dose compared to the vehicle (p = 0.0025) ( Figure 2A ). At the family and genera levels, Lactobacillaceae and Lactobacillus were significantly decreased at the medium and high concentrations ( Figure 2B and Figure 2C ). LEfSe analysis also confirmed significant changes in the Lactobacillaceae family ( Figure 2B ). Lactobacillales are called lactic acid bacteria since they mainly produce lactic acid from carbohydrate fermentation [ 85 , 86 ]. An increased abundance of Lactobacillales can inhibit pathogens, protect against disease, enhance the immune response of the host, and reduce cholesterol due to the colonization of the epithelium [ 85 , 86 ]. Therefore, a decrease in Lactobacillales abundance in the intestine could indicate increased inflammation and inflammatory markers. Figure 2. Open in a new tab Exposure of female mice to the MARBLES mix significantly decreased the abundance of (A) Lactobacillales at the order level, (B) Lactobacillaceae at the family level, and (C) Lactobacillus at the genus level in a dose-dependent manner. Statistically significant changes in bacterial abundance were assessed with the Mann-Whitney U test with Holm-Sidak multiple comparisons, with p<0.05 considered significant. At the genus level, random forest analysis revealed no significant difference in the abundance of Bacteroidales S7-24 and Alistipes between sham and MARBLES-exposed mice. In contrast, in wild-type mice exposed to a medium dose of the MARBLES mix through the maternal diet during gestation and lactation, Bacteroidales S7-24 and Alistipes abundances were significantly decreased compared to controls [ 50 ]. At the same time, Lactobacillus abundance remained unchanged following PCB exposure in the earlier study. At the phylum level, exposure to a low dose of the MARBLES mix increased the abundance of Verrucomicrobiota compared to the sham control ( Figure 1A ). In contrast, the medium and high PCB doses compared to sham did not significantly affect this phylum. The abundance of Verrucomicrobiales increased with low and medium doses of the MARBLES mix but decreased at the high dose. Inhalation of PCB 52 also affected the abundance of Verrucomicrobiota and Verrucomicrobiales in adolescent rats, with a decrease observed in the cecum following PCB 52 inhalation [ 75 ] and in mice exposed to oral PCB 126 [ 39 ]. In contrast, several studies have indicated that Verrucomicrobiota abundance increased following exposure to oral PCBs [ 80 , 81 , 83 , 87 ]. Our results in female mice exposed to the MARBLES mix differ distinctively from the earlier study investigating the microbiome in mice exposed to the MARBLES mix via the maternal diet [ 50 ]. Both animal studies were performed in animal facilities at the University of California (UC) Davis as part of a larger study investigating the developmental neurotoxicity of the MARBLES mix [4146, 88 ]. The differences in microbiome composition observed between both studies may reflect the differences in the exposure timeline and age differences in the microbiome. Additionally, seasonal, environmental, and other factors specific to the UC Davis facility may explain these differences [ 89 , 90 ]. The comparison between the two studies highlights the well-documented sensitivity of the microbiome to environmental factors [ 91 ] and the challenges of the reproducibility and generalizability of studies investigating the role of the microbiome in toxic outcomes [ 92 ]. 3.3. Effect of the MARBLES mix on the predicted functional composition of the microbiota PICRUSt, a bioinformatics software designed to predict the functional composition of metagenomes using marker gene data, such as 16S rRNA gene sequences was used to explore the potential functional effects of the microbial alterations caused by exposure to the MARBLES mix. This analysis predicted 965, 965, and 642 enzymes to be altered by the taxon changes caused by exposure of mice to low, medium, or high doses of the MARBLES mix via the diet. Among the predicted enzyme alterations, six were significantly affected by the low dose, two by the medium dose, and none by the high dose. These enzyme alterations are associated with predicted alterations in several pathways ( Table S9 ). Briefly, oral exposure to the MARBLES mix was predicted to significantly alter 41 pathways (10 de-enriched and 31 enriched pathways) in the low-dose group, 31 pathways (7 de-enriched and 24 enriched pathways) in the medium-dose group, and 39 pathways (18 de-enriched and 11 enriched pathways) in the high-dose group. The top ten pathways altered following PCB exposure were analyzed using KEGG enrichment pathway analysis ( Figure 1B ). Eight pathways were predicted to be significantly altered after exposure to the MARBLES mix across all exposure groups. These pathways showed either enrichment or suppression in both the medium and high-dose groups, while the opposite effect was observed in the low-dose group. For example, the "ribosome" pathway was significantly enriched in the medium and high-dose groups but was predicted to be significantly de-enriched in the low-dose group compared to controls. An earlier study also reported enrichment of the "ribosome" pathway after exposure to stigma maydis polysaccharide in an autism spectrum disorder rat model [ 30 ]. These results indicate that the “ribosome" pathway may play a role in psychiatric disorders; however, more studies are needed to confirm this hypothesis. In comparison to the control group, the "COVID-19" pathway was significantly enriched in mice exposed to medium and high PCB doses, but it was suppressed in those exposed to the low PCB dose. The "COVID-19" pathway is enriched when there is an increase in inflammatory markers [ 93 ], indicating that inflammation is prevalent in the gut lumen at higher PCB doses. Another inflammatory pathway, "monobactam biosynthesis," was significantly suppressed after exposure to the MARBLES mix at the medium and high dose group but enriched significantly in the low dose group. Monobactam is a bioactive compound that can lead to the synthesis of antibiotics [ 94 ]. The class Actinobacteria has been associated with the enrichment of monobactam biosynthesis [ 94 ]. As shown in Figure 1A , levels of Actinobacteria were decreased after exposure to the medium and high doses of MARBLES mix. As described above, the reduction of Lactobacillales abundance could be another factor in the predicted enrichment of the "COVID-19" pathway and suppression of "monobactam biosynthesis" pathway at higher PCB exposures [ 85 , 86 ]. 3.4. Topic modeling of microbial communities affected by PCB exposure Topic modeling [ 59 ], a machine learning approach to identify hidden patterns or "topics" was used to determine microbiome communities altered following exposure to the MARBLES mix. Using methods specified by CaoJuan2009 and Arun2010, it was observed that the ideal topic numbers were 6 for the low exposure group, 9 for the medium exposure group, and 6 for the high exposure group ( Figure S2A , C , and E ). Of these topics, only the high vs. sham exposure group had one topic out of six identified as significant ( Figure S2B , D , and F ). The top 20 bacteria found within this community are shown in Figure 1C . Of note, Lactobacillaceae was one of the most often assigned bacteria. This community has a lower probability of being assigned to the high exposure than the control group. Consistent with this observation, Lactobacillaceae abundance significantly decreased at high concentrations ( Figure 2B and Figure 2C ). No community types significantly differed from controls in the low and medium exposure groups following PCB exposure. 3.5. PCB and metabolite levels in feces and cecal content Excretion of unresorbed parent PCBs and biliary or intestinal excretion of the parent PCB after absorption from the gastrointestinal tract are minor routes of PCB elimination [ 95 ]. Most PCBs are excreted as the more polar metabolites via the feces, which, as assessed using the network analysis described below, can affect microbiome composition. For example, complex PCB metabolite mixtures have been detected in the gastrointestinal content and the feces of rodents following oral exposure to a complex, environmental PCB mixture [ 70 ]. We used GC-MS/MS and LC-HRMS analyses to characterize the complex PCB and metabolite mixtures present in the gastrointestinal tract of female mice exposed to the MARBLES mix. Subject-screening LC-MS/MS analyses were used to characterize the polar metabolites present in the cecum. Because of the limited quantity of cecal content, GC-MS/MS analysis was used to measure PCB and total OH-PCB levels in fecal samples collected as part of the animal study. These fecal measurements likely reflect the gastrointestinal PCB and OH-PCB profiles, as daily exposure to the MARBLES mix is expected to achieve steady-state PCB and metabolite levels within the gastrointestinal tract. 3.5.1. GC-MS/MS analysis of PCBs in feces All PCB congeners present in the MARBLES mix were detected in the feces, except for PCB84 and PCB95 ( Figure 3 ). The levels of both congeners were below the detection limit in the low-dose group. PCB levels typically increased dose-dependently in the feces ( Figures 3C - F , Table S10 ). A comparison of the PCB profiles showed that the low-dose group differed from the profile of the MARBLES mix (cos θ = 0.81); see Figure 4 . In contrast, the medium (cos θ = 0.95) and high dose (cos θ = 0.95) groups were more like the MARBLES mix. The dose-dependent difference in the fecal PCB residue profiles may be because more of the MARBLES mix in the medium and high dose groups were not fully absorbed in the gut. As a result, the fecal PCB profiles are more comparable to the original MARBLES mix. Analogously, we reported that the fecal PCB residues in mice exposed to a chiral PCB congener displayed a less pronounced enantiomeric enrichment due to unresorbed PCB in the gastrointestinal tract [ 96 ]. Figure 3. Open in a new tab PCBs and OH-PCBs concentration (ng/g feces) detected in feces from MARBLES-exposed mice. (A) PCB and (B) OH-PCB levels in feces were depicted in heatmaps and representative PCB (C-F) and OH-PCB (G-J) levels were shown in bar plots. (K-N) Simple linear regression models were fitted for representative PCBs and their OH-PCB metabolites, with their R 2 and 95% confidence limits (dashed line) of the best-fit line (solid line) shown in the figure. Data were shown as mean ± standard deviation (n = 6). Values used in the heatmaps are log-transformed. The y-axis in all the bar plots (C-J) is in log scale. Values below LOD are not plotted in the figures. ND: not detected (values are below LOD). Figure 4. Open in a new tab The profile of (A) PCB and (B) OH-PCB in the feces of MARBLES-exposed mice are shown as their mass percentage. The similarity coefficients cos θ are summarized for the (C) PCB and (D) OH-PCB profiles. The PCB profile was also compared to the MARBLES mix (panels A and C). When comparing fecal PCB profiles across exposure groups, the profiles of PCBs in medium and high dose groups were identical, with a similarity coefficient cos θ > 0.99 ( Figure 4A , 4B ). The profiles in both medium- and high-exposure groups were dominated by PCB 28, followed by PCB11 > PCB118 > PCB101 ≈ PCB52 ≈ PCB153 ≈ PCB180 ( Table S10 ). A different ranking was observed for PCB congeners in the low-dose group. PCB28 remained the predominant congener, but its mass percentage was not as high as in the medium and high dose groups, with PCB levels following the rank order PCB28 > PCB52 ≈ PCB101 > PCB11 ≈ PCB153 ≈ PCB118 ( Table S10 ). The tissue disposition of PCBs in the brain, liver, and serum from the same animals investigated in this study reported PCB tissue profiles comparable to the profiles observed in feces [ 88 ]. For example, PCB28 was the major PCB congener in tissues, followed by PCB118, PCB180, PCB153, PCB52, and PCB138, whereas PCB11 is barely detectable in serum and liver and below the limit of detection in the brain. 3.5.2. GC-MS/MS analysis of OH-PCBs in feces The total levels of each OH-PCB congener were measured in parallel with the parent PCB congeners after deconjugation with sulfatase from Helix pomatia [ 97 ]. Twenty-three OH-PCBs were detected in the feces from the medium and high-dose group ( Figure 3B , Table S10 ). Only 17 OH-PCBs were identified in the low-dose group. In comparison, 21 OH-PCBs were detected in serum, 15 in the liver, and only 8 in the brain from the same animals investigated in this study [ 88 ]. Fecal OH-PCB levels increased in a dose-dependent manner ( Figures 3G - 3J ). 4,5-11 and 5,6-11, two minor metabolites of PCB11, were an exception, likely because of the low detection frequency of 33% in the medium dose group. Moreover, the fecal levels of OH-PCBs, such as 4-11, 3'-28, 4-52, and 3-118, correlated with the levels of the corresponding parent PCB ( Figures 3K - 3N ; for a summary of correlations between PCB and OH-PCB, see Table S11 ). The OH-PCBs profiles in all three exposure groups were identical, with cos θ > 0.99 ( Figure 4C - 4D ). Two PCB28 metabolites, 5-28 and 3'-28, accounted for more than 80% of the total OH-PCB across all three exposure groups, followed by 4-52 > 4-11 > 5-95 ≈ 4-95 ( Figure 4C ). Several of the OH-PCBs detected in the feces are human-relevant PCB metabolites. For example, 3'-28 has been reported in plasma samples from PCB occupationally exposed individuals from Germany [ 98 ]. 5-28 and 3'-28 have been detected in PCB28 metabolism studies with recombinant human CYP1A2 and transgenic HEK293 cells expressing CYP1A2 [ 99 ]. 4-52 has been identified as a major metabolite of PCB52 in rodents and recombinant human CYP2A6 [ 18 , 71 ]; however, this metabolite has not been reported in humans to-date. Finally, 4-11 has been detected as the major metabolite of PCB11 in rodents and human populations [ 100 , 101 ]. The overall profile of OH-PCB in feces is comparable to that in liver and serum reported in our earlier study [ 88 ]; however, the total OH-PCB levels in feces are much higher than those observed in tissues. For example, in the high-dose group, the average concentration of 5-28 in feces was 15000 ng/g, while 2380 ng/g in the liver, 81.1 ng/g in the serum, and 0.7 ng/g in the brain, indicating feces is a primary route of excretion for OH-PCB or their conjugates. Consistent with this interpretation, a significant percentage of the total dose of PCB91 and PCB95, two congeners present in the MARBLES mix, were eliminated as OH-PCB metabolites or the corresponding conjugates in mice following oral exposure [ 102 , 103 ]. 3.5.3. LC-HRMS analysis of PCB metabolites in cecal content The PCB metabolites in the cecal content were analyzed using a subject-screening LC-HRMS method [ 69 , 70 ]. Seven classes of metabolites were detected in the feces using LC-HRMS analysis, including OH-PCBs (C 12 H 9-n Cl n O, n=1-7; 12 total analytes), PCB sulfates (C 12 H 9-n Cl n SO 4 , n=3-7; 6 total analytes), hydroxylated PCB sulfates (OH-PCB sulfates, C 12 H 9-n Cl n /SO 5 , n=1-6; 15 total analytes), methoxylated OH-PCBs (MeO-OH-PCBs, C 13 H 11-n Cl n O 2 , n=5-6; 3 total analytes), dihydroxylated (diOH-PCBs, C 12 H 9-n Cl n O 2 , n=2 or 4 or 6; 7 total analytes), PCB sulfonates (C 12 H 9-n Cl n SO 3 , n=1-2 and 4-6; 5 total analytes), and methoxylated PCB sulfates (MeO-PCB sulfates, C 13 H 6 Cl 5 SO 5 , n=5; 1 analyte). Example chromatograms and the accurate masses and chlorine isotope patterns of the molecular ions of selected metabolites are shown in Figure 5 . Figure 5. Open in a new tab Non-target analysis of PCB metabolites using LC-HRMS reveals the presence of different homologs of (A) OH-PCBs, (B) PCB sulfates, (C) OH-PCB sulfates, (D) diOH-PCBs, (E) MeO-OH-PCBs, (F) PCB sulfonates, and (G) MeO-PCB sulfates in the cecal content of female mice exposed to different doses of the MARBLES mix. The bar charts show the relative abundance of the homologs of each metabolite. Values are the relative abundance relative to internal standard PFOS. Only metabolites with more than 50% detection frequency were included in the figure. ND: not detected. The relative levels of all PCB metabolites increased dose-dependently across all homolog groups ( Figure 6 ). Based on the relative peak area, the PCB metabolite levels followed the descending order: OH-PCB > PCB sulfate > OH-PCB sulfate > MeO-OH-PCB > diOH-PCB > PCB sulfonate > MeO-PCB sulfate ( Table S12 ). Mono- to hepta-chlorinated OH-PCBs, which are formed by oxidation of the parent PCBs by cytochrome P450 enzymes [ 19 ], accounted for > 80% of the total metabolites (86% for L, 87% for M, and 89% for H) detected by LC-HRMS. Tri-, penta-, and hexachlorinated OH-PCBs were the major OH-PCBs ( Figure 6A ). This finding is consistent with the GC-MS/MS analysis results, which indicate that 5-28, a tri-chlorinated metabolite, is the most abundant metabolite detected in the feces. Interestingly, a mono-chlorinated OH-PCB was detected in the LC-HRMS analysis in the high-dose group. This metabolite is likely the result of a dechlorination reaction. Several studies have reported that mammalian cytochrome P450 enzymes can form various dechlorinated PCB metabolites [ 99 , 104 , 105 ]. Figure 6. Open in a new tab The chromatogram (A1-F1) and mass spectra of the molecular ion (A2-F2) for representative metabolites from the non-target analysis, including (A1,A2) pentachlorinated OH-PCBs, (B1,B2) pentachlorinated PCB sulfates, (C1,C2) tetrachlorinated OH-PCB sulfates, (D1,D2) tetrachlorinated diOH-PCBs, (E1,E2) pentachlorinated MeO-OH-PCBs, and (F1,F2) pentachlorinated PCB sulfonates. Panels A1-F1 show the chromatogram of the three most abundant isotopes with their molecular formulas labeled. The measured accurate mass, shown in panels A2-F2, matched the theoretical accurate mass and the chlorine isotopic pattern. PCB sulfates were the second most abundant PCB metabolites detected in the cecum (5-10% of total). These metabolites are formed by the sulfation of OH-PCBs by sulfotransferases [ 106 ]. This process promotes the detoxification and excretion of PCBs. The further oxidation of PCB sulfate leads to the formation of OH-PCB sulfate. As illustrated in the PCB metabolic pathway in Figure 7 , additional PCB metabolites present in the gastrointestinal tract include diOH-PCBs, MeO-OH-PCBs, PCB sulfonates, and MeO-PCB sulfates as minor metabolites. These metabolites were mainly detected in the high-dose group ( Figure 6 ). Similar metabolites were observed in the intestinal content or feces of mice exposed orally to another environmental PCB mixture [ 70 ]. Figure 7. Open in a new tab Seven classes of PCB metabolites were detected in the cecum or feces of female mice exposed to the MARBLES mix. Representative structures are shown, but the position of the substituents is uncertain. Metabolites that were detected in our study are labeled in blue. A: Cytochrome P450; B: Sulfotransferase; C: Epoxide hydrolase; D: Dihydrodiol dehydrogenase; E: Glutathione-S-transferase; F: γ-Glutamyl transpeptidase; G: Cysteinylglycinase; H: Cysteine S-conjugate β-lyase; I: Sulfatase. 3.6. Network Analysis of PCB and microbiome data Network analysis was used to integrate the PCB and PCB metabolite data with the results from the microbiome analysis to assess if specific PCB congeners or PCB metabolites are associated with abundance changes of specific bacteria due to PCB exposure ( Figure 8A ). Pearson correlation analyses were performed to confirm the significance of the correlation observed in the network analysis ( Figures 8C , 8D ). Figure 8. Open in a new tab Integration of PCB and microbiome data allows the identification of classes of PCBs or their metabolites associated with specific bacteria in female mice exposed to different doses of the MARBLES mix. Networks of (A) PCBs and OH-PCBs, determined in the feces by GC-MS/MS, and (B) other polar PCB metabolites, determined by LC-HRMS in the cecal content, and the cecal microbiome. Rectangles represent PCBs and their metabolites, while microbes are represented in ovals. The color coding of PCBs or PCB metabolites and microbes represents clusters of correlations identified by xMWAS. Line colors represent negative (blue) and positive (red) correlations. Pearson correlation analyses confirm the results from the network analysis for (C) PCBs and OH-PCBs from the GC-MS/MS analyses and (D) PCB metabolites from the LC-HRMS analysis and the cecal microbiome. The network of PCB and OH-PCB metabolites, determined in the feces by GC-MS/MS, and the microbiome data are depicted in Figure 8A (R >0.4, p-value < 0.05). Four clusters containing 29 PCB metabolites and 7 microbes showed 53 correlations (25 positive and 28 negative). One order (Oscillospirales), one family ( Peptococcaceae ), and five genera ( UCG-1 , Roseburia , Lachnoclostridium , Lactobacillus, and Lactococcus ) were correlated with PCBs and PCB metabolites. One cluster included Oscillospirales , which was positively correlated with several parent PCBs (28, 52, 95, 101, 118, 136, and 180) but no PCB metabolites, and Lactobacillus , which was negatively correlated with several of the same parent PCBs. The cluster also included two OH-PCB95 metabolites. A second cluster included Lactococcus , which was negatively correlated with seven, mostly higher chlorinated OH-PCBs. In contrast, Lachnoclostridium was positively correlated with three OH-PCBs in the third cluster. The fourth cluster included Roseburia and Peptococcaceae , which were positively correlated positively with mostly lower chlorinated OH-PCB metabolites, and UCG-1, which was negatively correlated with several of the same OH-PCB metabolites. The network of the polar PCB metabolite, determined by LC-HRMS in the cecal content, and the microbiome data are depicted in Figure 8B (R > 0.4, p-value < 0.05). Four clusters containing 50 PCB metabolites, and seven microbes showed 168 correlations (36 positive and 132 negative). One large cluster contained two families ( UCG-1 , Lachnospiraceae ) and one genus ( Lachnospiraceae UCG-4 ) were negatively associated with diverse PCB metabolites, including diOH-PCBs and their conjugates. Because diOH-PCBs are reactive and toxic PCB metabolites [ 107 ], the presence of these metabolites in the gastrointestinal tract may be acutely toxic to the bacteria included in this cluster. In the second large cluster, one phylum (Firmicutes) and one genus ( Lactococcus ) correlated positively or negatively with diverse PCB metabolites. One smaller cluster included Lactobacillus , which was negatively associated with a PCB sulfonate metabolite. In the network analysis shown in Figure 8A , Lactobacillus was also negatively related to several PCB metabolites. Another cluster contained Roseburia , which was negatively correlated with an OH-PCB metabolite. In contrast, Roseburia showed positive correlations with OH-PCB metabolites in the network, as shown in Figure 8A . Conclusions This study provides novel insights into the interactions between PCBs and the gut microbiome. Consistent with earlier work, this study demonstrated that oral exposure to a human-relevant PCB mixture, the MARBLES mix, impacts microbial abundance and function in female mice. While no significant differences in alpha or beta diversity were observed across exposure groups, PCB exposure resulted in dose-dependent alterations in specific microbial taxa, for example, a reduction in Lactobacillus abundance at medium and high doses. The study also revealed the presence of complex mixtures of the parent PCBs, OH-PCBs, and other PCB metabolites in the gastrointestinal tract following oral PCB exposure. Utilizing both datasets, the network analysis identified specific parent PCB congeners, lower or higher chlorinated OH-PCBs, and diOH-PCBs and their conjugates that affected the abundance of specific bacteria in the gastrointestinal tract. These associations may reflect the direct toxicity of specific PCB or metabolite congeners to particular bacteria. This work highlights the utility of integrating PCB and microbiome data to uncover PCB x microbiome interactions that may play a role in PCB-induced toxicity through the gut-liver-brain axis. Supplementary Material 1 NIHMS2061747-supplement-1.docx (483.1KB, docx) Environmental Implication. Polychlorinated biphenyls (PCBs), hazardous environmental contaminants, linked with neurological disorders, are hypothesized to cause neurotoxicity via mechanisms involving the gut-liver-brain axis. We examined changes in the gut microbiome and PCB metabolism following exposure to a complex PCB mixture, the MARBLES mix. We identified individual PCBs and metabolites associated with cecal microbial composition and predicted functional changes. These findings demonstrate that PCBs alter the gut microbial composition and possibly function. Highlights. PCB exposure dose-dependently altered cecal bacterial levels, including Lactobacillales Lactobacillaceae was among the most frequently identified bacteria in topic modeling PCB exposure altered predicted functional pathways relevant to inflammation Complex PCB and metabolite mixtures were detected in gut content and feces Network analysis linked altered gut microbiota to specific PCB levels in GI content Acknowledgments: The authors would like to thank Rachel F. Marek and Keri C. Hornbuckle from the Analytical Core of Iowa Superfund Research Program for supporting the GC-MS/MS analyses, the staff of the University of Iowa's Microbiome Core for providing 16S rRNA sequencing services, Dr. Lynn M. Teesch and Vic R. Parcell from the High-Resolution Mass Spectrometry Facility at the University of Iowa for help with the LC-HRMS analysis. Funding Sources: This work was supported by grants R01 ES014901, R01 ES031098, R01 ES034691 and R01 ES029213 from the National Institute of Environmental Health Sciences, National Institutes of Health. In addition, the Environmental Health Sciences Research Center (P30 ES005605 and P30 ES023513) and the Iowa Superfund Research Program (P43 ES013661) supported the authentication of the test compounds and the analytical work. RLF was supported by the National Institutes of Health under award number F31 DE033564. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institute of Environmental Health Sciences or the National Institutes of Health. Footnotes Research Data The sequence data have been uploaded to the Sequence Read Archive ( https://www.ncbi.nlm.nih.gov/sra ; last accessed December 11, 2024) under the BioProject ID PRJNA1196926 and are freely accessible. All other data reported in this manuscript are publicly available on Iowa Research Online and can be accessed at DOI: 10.25820/data.007461 . Declaration of Generative AI and AI-assisted technologies in the writing process The author(s) used ChatGPT ( chat.openai.com ) and Grammarly ( https://www.grammarly.com/ ) to edit sections of the paper. After using these tools/services, the author(s) reviewed and edited the content as needed and take full responsibility for the content of the publication. Declaration of Competing Interest The authors declare the following financial interests/personal relationships which may be considered as potential competing interests: Author AM is the inventor of a technology claiming the use of Prevotella histicola for treating autoimmune diseases. The patent for the technology is owned by Mayo Clinic, which has given Evelo Biosciences an exclusive license. AM received royalties from Mayo Clinic (paid by Evelo Biosciences). However, no funds or products from the patent were used in the present study. Author PJL was hired as an expert witness by lawyers representing a group of plaintiffs alleging they were harmed by exposure to PCBs in school air to testify on PCB neurotoxicity. The defendant was Pharmacia, a successor company to Monsanto. All other authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. References 1. Bannavti MK, Jahnke JC, Marek RF, Just CL, & Hornbuckle KC (2021). Room-to-room variability of airborne polychlorinated biphenyls in schools and the application of air sampling for targeted source evaluation. Environ Sci Technol, 55(14), 9460–9468. 10.1021/acs.est.0c08149 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 2. IARC. (2016). Polychlorinated biphenyls and polybrominated biphenyls. IARC Monogr Eval Carcinog Risks Hum, 107, 9–500. [ PMC free article ] [ PubMed ] [ Google Scholar ] 3. Marek RF, Thorne PS, Herkert NJ, Awad AM, & Hornbuckle KC (2017). Airborne pcbs and oh-pcbs inside and outside urban and rural u.S. Schools. Environ Sci Technol, 51(14), 7853–7860. 10.1021/acs.est.7b01910 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 4. Erickson MD, & Kaley RG 2nd. (2011). Applications of polychlorinated biphenyls. Environ Sci Pollut Res Int, 18(2), 135–151. 10.1007/s11356-010-0392-1 [ DOI ] [ PubMed ] [ Google Scholar ] 5. Grossman E (2013). Nonlegacy pcbs: Pigment manufacturing by-products get a second look. Environ Health Perspect, 121(3), A86–93. 10.1289/ehp.121-a86 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 6. EPA. (2024). Inadvertent pcbs. https://www.epa.gov/pcbs/inadvertent-pcbs [ Google Scholar ] 7. ATSDR. (2000). Toxicological profile for polychlorinated biphenyls (pcbs). U.S. Dept. of Health and Human Services, Public Health Service, Agency for Toxic Substances and Disease Registry,. http://purl.fdlp.gov/GPO/gpo31597 [ PubMed ] [ Google Scholar ] 8. Saktrakulkla P, Lan T, Hua J, Marek RF, Thorne PS, & Hornbuckle KC (2020). Polychlorinated biphenyls in food. Environ Sci Technol, 54(18), 11443–11452. 10.1021/acs.est.0c03632 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 9. Vorkamp K (2016). An overlooked environmental issue? A review of the inadvertent formation of pcb-11 and other pcb congeners and their occurrence in consumer products and in the environment. Sci Total Environ, 541, 1463–1476. [ DOI ] [ PubMed ] [ Google Scholar ] 10. Ampleman MD, Martinez A, DeWall J, Rawn DF, Hornbuckle KC, & Thorne PS (2015). Inhalation and dietary exposure to pcbs in urban and rural cohorts via congenerspecific measurements. Environ Sci Technol, 49(2), 1156–1164. 10.1021/es5048039 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 11. Borlakoglu JT, Haegele KD, Reich HJ, Dils RR, & Wilkins JP (1991). In vitro metabolism of [14c]4-chlorobiphenyl and [14c]2,2',5,5'-tetrachlorobiphenyl by hepatic microsomes from rats and pigeons. Evidence against an obligatory arene oxide in aromatic hydroxylation reactions. Int J Biochem, 23(12), 1427–1437. 10.1016/0020-711x(91)90286-v [ DOI ] [ PubMed ] [ Google Scholar ] 12. Borlakoglu JT, & Wilkins JPG (1993). Metabolism of di-, tri-, tetra-, penta- and hexachlorobiphenyls by hepatic microsomes isolated from control animals and animals treated with aroclor 1254, a commercial mixture of polychlorinated biphenyls (pcbs). Comp Biochem Physiol C Toxicol Pharmacol, 105(1), 95–106. https://doi.org/ 10.1016/0742-8413(93)90064-R [ DOI ] [ PubMed ] [ Google Scholar ] 13. Haraguchi K, Kato Y, Koga N, & Degawa M (2004). Metabolism of polychlorinated biphenyls by gunn rats: Identification and serum retention of catechol metabolites. Chem Res Toxicol, 17(12), 1684–1691. 10.1021/tx0498096 [ DOI ] [ PubMed ] [ Google Scholar ] 14. Haraguchi K, Kato Y, Koga N, & Degawa M (2005). Species differences in the tissue distribution of catechol and methylsulphonyl metabolites of 2,4,5,2',5'-penta- and 2,3,4,2',3',6'-hexachlorobiphenyls in rats, mice, hamsters and guinea pigs. Xenobiotica, 35(1), 85–96. 10.1080/00498250400026456 [ DOI ] [ PubMed ] [ Google Scholar ] 15. Ohta C, Haraguchi K, Kato Y, Endo T, Kimura O, & Koga N (2015). Distribution and excretion of 2,2′,3,4′,5,5′,6-heptachlorobiphenyl (cb187) and its metabolites in rats and guinea pigs. Chemosphere, 118, 5–11. https://doi.org/ 10.1016/j.chemosphere.2014.05.019 [ DOI ] [ PubMed ] [ Google Scholar ] 16. Ohta C, Haraguchi K, Kato Y, Endo T, & Koga N (2013). [species difference in the metabolism of 2, 2', 3, 4', 5, 5'-hexachlorobiphenyl (cb146) by animal and human liver microsomes]. Fukuoka Igaku Zasshi, 104(4), 161–169. [ PubMed ] [ Google Scholar ] 17. Ohta C, Haraguchi K, Kato Y, & Koga N (2005). In vitro metabolism of 2,2′,3,4′,5,5 ′,6-heptachlorobiphenyl (cb187) by liver microsomes from rats, hamsters and guinea pigs. Xenobiotica, 35(4), 319–330. 10.1080/00498250500087507 [ DOI ] [ PubMed ] [ Google Scholar ] 18. Shimada T, Kakimoto K, Takenaka S, Koga N, Uehara S, Murayama N, Yamazaki H, Kim D, Guengerich FP, & Komori M (2016). Roles of human cyp2a6 and monkey cyp2a24 and 2a26 cytochrome p450 enzymes in the oxidation of 2,5,2',5'-tetrachlorobiphenyl. Drug Metab Dispos, 44(12), 1899–1909. 10.1124/dmd.116.072991 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 19. Grimm FA, Hu D, Kania-Korwel I, Lehmler HJ, Ludewig G, Hornbuckle KC, Duffel MW, Bergman A, & Robertson LW (2015). Metabolism and metabolites of polychlorinated biphenyls. Crit Rev Toxicol, 45(3), 245–272. 10.3109/10408444.2014.999365 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 20. Bullert A, Li X, Chunyun Z, Lee K, Pulliam CF, Cagle BS, Doorn JA, Klingelhutz AJ, Robertson LW, & Lehmler HJ (2023). Disposition and metabolomic effects of 2,2',5,5'-tetrachlorobiphenyl in female rats following intraperitoneal exposure. Environ Toxicol Pharmacol, 102, 104245. 10.1016/j.etap.2023.104245 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 21. Grimm FA, He X, Teesch LM, Lehmler HJ, Robertson LW, & Duffel MW (2015). Tissue distribution, metabolism, and excretion of 3,3'-dichloro-4'-sulfooxy-biphenyl in the rat. Environ Sci Technol, 49(13), 8087–8095. 10.1021/acs.est.5b01499 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 22. Zhang CY, Flor S, Ruiz P, Ludewig G, & Lehmler HJ (2021). Characterization of the metabolic pathways of 4-chlorobiphenyl (pcb3) in hepg2 cells using the metabolite profiles of its hydroxylated metabolites. Environ Sci Technol, 55(13), 9052–9062. 10.1021/acs.est.1c01076 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 23. Granillo L, Sethi S, Keil KP, Lin Y, Ozonoff S, Iosif AM, Puschner B, & Schmidt RJ (2019). Polychlorinated biphenyls influence on autism spectrum disorder risk in the marbles cohort. Environ Res, 171, 177–184. 10.1016/j.envres.2018.12.061 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 24. Paranjape N, Strack S, Lehmler HJ, & Doorn JA (2024). Astrocyte mitochondria are a sensitive target of pcb52 and its human-relevant metabolites. ACS Chem Neurosci, 15(15). 10.1021/acschemneuro.4c00116 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 25. Freedman SN, Shahi SK, & Mangalam AK (2018). The "gut feeling": Breaking down the role of gut microbiome in multiple sclerosis. Neurotherapeutics, 15(1), 109–125. 10.1007/s13311-017-0588-x [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 26. Dupont HL, Jiang ZD, Dupont AW, & Utay NS (2020). The intestinal microbiome in human health and disease. Trans Am Clin Climatol Assoc, 131, 178–197. https://www.ncbi.nlm.nih.gov/pubmed/32675857 [ PMC free article ] [ PubMed ] [ Google Scholar ] 27. Mayer EA (2011). Gut feelings: The emerging biology of gut-brain communication. Nat Rev Neurosci, 12(8), 453–466. 10.1038/nrn3071 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 28. Giovannini MG, Lana D, Traini C, & Vannucchi MG (2021). The microbiota–gut–brain axis and alzheimer disease. From dysbiosis to neurodegeneration: Focus on the central nervous system glial cells. J Clin Med, 10(11), 2358. https://www.mdpi.com/2077-0383/10/11/2358 https://mdpi-res.com/d_attachment/jcm/jcm-10-02358/article_deploy/jcm-10-02358-v2.pdf?version=1623318031 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 29. Wu H, Zhang Y, Yu J, & Shi M (2023). Editorial: Gut-liver-brain axis: A complex network influences human health and diseases. Front Neurosci, 17, 1241069. 10.3389/fnins.2023.1241069 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 30. Yang X, Li J, Zhou Y, Zhang N, & Liu J (2022). Effect of stigma maydis polysaccharide on the gut microbiota and transcriptome of vpa induced autism model rats. Front Microbiol, 13, 1009502. 10.3389/fmicb.2022.1009502 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 31. Wen L, & Duffy A (2017). Factors influencing the gut microbiota, inflammation, and type 2 diabetes. J Nutr, 147(7), 1468S–1475S. 10.3945/jn.116.240754 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 32. Xu LL, Zhang QY, Chen YK, Chen LJ, Zhang KK, Wang Q, & Xie XL (2022). Gestational pcb52 exposure induces hepatotoxicity and intestinal injury by activating inflammation in dam and offspring mice: A maternal and progeny study. Environ Pollut, 313, 120186. 10.1016/j.envpol.2022.120186 [ DOI ] [ PubMed ] [ Google Scholar ] 33. Chi Y, Wang H, Lin Y, Lu Y, Huang Q, Ye G, & Dong S (2019). Gut microbiota characterization and lipid metabolism disorder found in pcb77-treated female mice. Toxicology, 420, 11–20. 10.1016/j.tox.2019.03.011 [ DOI ] [ PubMed ] [ Google Scholar ] 34. Agarwal M, Hoffman J, Ngo Tenlep SY, Santarossa S, Pearson KJ, Sitarik AR, Cassidy-Bushrow AE, & Petriello MC (2023). Maternal polychlorinated biphenyl 126 (pcb 126) exposure modulates offspring gut microbiota irrespective of diet and exercise. Reprod Toxicol, 118, 108384. 10.1016/j.reprotox.2023.108384 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 35. Lim JJ, Li X, Lehmler HJ, Wang D, Gu H, & Cui JY (2020). Gut microbiome critically impacts pcb-induced changes in metabolic fingerprints and the hepatic transcriptome in mice. Toxicol Sci, 177(1), 168–187. 10.1093/toxsci/kfaa090 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 36. Di Vincenzo F, Del Gaudio A, Petito V, Lopetuso LR, & Scaldaferri F (2024). Gut microbiota, intestinal permeability, and systemic inflammation: A narrative review. Intern Emerg Med, 19(2), 275–293. 10.1007/s11739-023-03374-w [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 37. Mostafavi Abdolmaleky H, & Zhou J-R (2024). Gut microbiota dysbiosis, oxidative stress, inflammation, and epigenetic alterations in metabolic diseases. Antioxidants, 13(8), 985. https://www.mdpi.com/2076-3921/13/8/985 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 38. Li T, Tian D, Lu M, Wang B, Li J, Xu B, Chen H, & Wu S (2022). Gut microbiota dysbiosis induced by polychlorinated biphenyl 126 contributes to increased brain proinflammatory cytokines: Landscapes from the gut-brain axis and fecal microbiota transplantation. Ecotoxicol Environ Saf, 241, 113726. 10.1016/j.ecoenv.2022.113726 [ DOI ] [ PubMed ] [ Google Scholar ] 39. Tian Y, Rimal B, Gui W, Koo I, Yokoyama S, Perdew GH, & Patterson AD (2022). Early life short-term exposure to polychlorinated biphenyl 126 in mice leads to metabolic dysfunction and microbiota changes in adulthood. Int J Mol Sci, 23(15), 1–13. 10.3390/ijms23158220 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 40. Tian Y, Rimal B, Gui W, Koo I, Smith PB, Yokoyama S, & Patterson AD (2022). Early life polychlorinated biphenyl 126 exposure disrupts gut microbiota and metabolic homeostasis in mice fed with high-fat diet in adulthood. Metabolites, 12(10), 1–14. 10.3390/metabo12100894 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 41. Keil Stietz KP, Sethi S, Klocke CR, de Ruyter TE, Wilson MD, Pessah IN, & Lein PJ (2021). Sex and genotype modulate the dendritic effects of developmental exposure to a human-relevant polychlorinated biphenyls mixture in the juvenile mouse. Front Neurosci, 15, 766802. 10.3389/fnins.2021.766802 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 42. Matelski L, Keil Stietz KP, Sethi S, Taylor SL, Van de Water J, & Lein PJ (2020). The influence of sex, genotype, and dose on serum and hippocampal cytokine levels in juvenile mice developmentally exposed to a human-relevant mixture of polychlorinated biphenyls. Curr Res Toxicol, 1, 85–103. 10.1016/j.crtox.2020.09.001 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 43. Mouat JS, Li X, Neier K, Zhu Y, Mordaunt CE, La Merrill MA, Lehmler HJ, Jones MP, Lein PJ, Schmidt RJ, & LaSalle JM (2023). Networks of placental DNA methylation correlate with maternal serum pcb concentrations and child neurodevelopment. Environ Res, 220, 115227. 10.1016/j.envres.2023.115227 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 44. Sethi S, Keil KP, Chen H, Hayakawa K, Li X, Lin Y, Lehmler HJ, Puschner B, & Lein PJ (2017). Detection of 3,3'-dichlorobiphenyl in human maternal plasma and its effects on axonal and dendritic growth in primary rat neurons. Toxicol Sci, 158(2), 401–411. 10.1093/toxsci/kfx100 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 45. Sethi S, Keil Stietz KP, Valenzuela AE, Klocke CR, Silverman JL, Puschner B, Pessah IN, & Lein PJ (2021). Developmental exposure to a human-relevant polychlorinated biphenyl mixture causes behavioral phenotypes that vary by sex and genotype in juvenile mice expressing human mutations that modulate neuronal calcium. Front Neurosci, 15, 766826. 10.3389/fnins.2021.766826 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 46. Sethi S, Morgan RK, Feng W, Lin Y, Li X, Luna C, Koch M, Bansal R, Duffel MW, Puschner B, Zoeller RT, Lehmler HJ, Pessah IN, & Lein PJ (2019). Comparative analyses of the 12 most abundant pcb congeners detected in human maternal serum for activity at the thyroid hormone receptor and ryanodine receptor. Environ Sci Technol, 53(7), 3948–3958. 10.1021/acs.est.9b00535 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 47. EPA. (2023). Table of polychlorinated biphenyl (pcb) congeners. https://www.epa.gov/pcbs/table-polychlorinated-biphenyl-pcb-congeners 48. Maervoet J, Covaci A, Schepens P, Sandau CD, & Letcher RJ (2004). A reassessment of the nomenclature of polychlorinated biphenyl (pcb) metabolites. Environ Health Perspect, 112(3), 291–294. 10.1289/ehp.6409 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 49. Aghajani M, Graham S, McCafferty C, Shaheed CA, Roberts T, DeSouza P, Yang T, & Niles N (2018). Clinicopathologic and prognostic significance of programmed cell death ligand 1 expression in patients with non-medullary thyroid cancer: A systematic review and meta-analysis. Thyroid, 28(3), 349–361. 10.1089/thy.2017.0441 [ DOI ] [ PubMed ] [ Google Scholar ] 50. Rude KM, Pusceddu MM, Keogh CE, Sladek JA, Rabasa G, Miller EN, Sethi S, Keil KP, Pessah IN, Lein PJ, & Gareau MG (2019). Developmental exposure to polychlorinated biphenyls (pcbs) in the maternal diet causes host-microbe defects in weanling offspring mice. Environ Pollut, 253, 708–721. 10.1016/j.envpol.2019.07.066 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 51. Keil Stietz KP, Kennedy CL, Sethi S, Valenzuela A, Nunez A, Wang K, Wang Z, Wang P, Spiegelhoff A, Puschner B, Bjorling DE, & Lein PJ (2021). In utero and lactational pcb exposure drives anatomic changes in the juvenile mouse bladder. Curr Res Toxicol, 2, 1–18. 10.1016/j.crtox.2021.01.002 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 52. Callahan BJ, McMurdie PJ, Rosen MJ, Han AW, Johnson AJ, & Holmes SP (2016). Dada2: High-resolution sample inference from illumina amplicon data. Nat Methods, 13(7), 581–583. 10.1038/nmeth.3869 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 53. Shahi SK, Zarei K, Guseva NV, & Mangalam AK (2020). Microbiota analysis using two-step pcr and next-generation 16s rrna gene sequencing. J Vis Exp, 1(152), 1–21. 10.3791/59980 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 54. Chong J, Liu P, Zhou G, & Xia J (2020). Using microbiomeanalyst for comprehensive statistical, functional, and meta-analysis of microbiome data. Nat Protoc, 15(3), 799–821. 10.1038/s41596-019-0264-1 [ DOI ] [ PubMed ] [ Google Scholar ] 55. Dhariwal A, Chong J, Habib S, King IL, Agellon LB, & Xia J (2017). Microbiomeanalyst: A web-based tool for comprehensive statistical, visual and metaanalysis of microbiome data. Nucleic Acids Res, 45(W1), W180–W188. 10.1093/nar/gkx295 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 56. Segata N, Izard J, Waldron L, Gevers D, Miropolsky L, Garrett WS, & Huttenhower C (2011). Metagenomic biomarker discovery and explanation. Genome Biol, 12(6), R60. 10.1186/gb-2011-12-6-r60 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 57. Langille MG, Zaneveld J, Caporaso JG, McDonald D, Knights D, Reyes JA, Clemente JC, Burkepile DE, Vega Thurber RL, Knight R, Beiko RG, & Huttenhower C (2013). Predictive functional profiling of microbial communities using 16s rrna marker gene sequences. Nat Biotechnol, 31(9), 814–821. 10.1038/nbt.2676 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 58. R Core Team. (2021). R: A language and environment for statistical computing https://www.R-project.org/ 59. Fitzjerrells RL, Ollberding NJ, & Mangalam AK (2024). Looking at the full picture, using topic modeling to observe microbiome communities associated with disease. Gut Microbes Reports, 1(1), 1–11. 10.1080/29933935.2024.2378067 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 60. McMurdie PJ, & Holmes S (2013). Phyloseq: An r package for reproducible interactive analysis and graphics of microbiome census data. PLoS One, 8(4), e61217. 10.1371/journal.pone.0061217 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 61. Cao J, Xia T, Li J, Zhang Y, & Tang S (2009). A density-based method for adaptive lda model selection. Neurocomputing, 72(7), 1775–1781. https://doi.org/ 10.1016/j.neucom.2008.06.011 [ DOI ] [ Google Scholar ] 62. Arun R, Suresh V, Veni Madhavan CE, & Narasimha Murthy MN (2010). On finding the natural number of topics with latent dirichlet allocation: Some observations. Advances in Knowledge Discovery and Data Mining, Berlin, Heidelberg. [ Google Scholar ] 63. Nikita M, & Chaney N (2020). Ldatuning: Tuning of the latent dirichlet allocation models parameters. https://github.com/nikita-moor/ldatuning [ Google Scholar ] 64. Grün B, Hornik K, Blei DM, Lafferty JD, Phan X-H, Matsumoto M, Nishimura T, & Cokus S (2024). Topicmodels: Topic models. https://cran.r-project.org/web/packages/topicmodels/index.html [ Google Scholar ] 65. Grün B, & Hornik K (2011). Topicmodels: An r package for fitting topic models. J Stat Softw, 40(13), 1–30. 10.18637/jss.v040.i13 [ DOI ] [ Google Scholar ] 66. Zhou H, He K, Chen J, & Zhang X (2022). Linda: Linear models for differential abundance analysis of microbiome compositional data. Genome Biol, 23(1), 95. 10.1186/s13059-022-02655-5 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 67. Xhang X, Chen J, & Zhou H (2024). Microbiomestat: Statistical methods for microbiome compositional data. https://cran.r-project.org/web/packages/MicrobiomeStat/index.html [ Google Scholar ] 68. Black H (1983). The preparation and reactions of diazomethane. Aldrichimica acta, 16, 3. [ Google Scholar ] 69. Li X, Hefti MM, Marek RF, Hornbuckle KC, Wang K, & Lehmler H-J (2022). Assessment of polychlorinated biphenyls and their hydroxylated metabolites in postmortem human brain samples: Age and brain region differences. Environ Sci Technol, 56(13), 9515–9526. 10.1021/acs.est.2c00581 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 70. Li X, Liu Y, Martin JW, Cui JY, & Lehmler HJ (2021). Nontarget analysis reveals gut microbiome-dependent differences in the fecal pcb metabolite profiles of germ-free and conventional mice. Environ Pollut, 268(Pt A), 115726. 10.1016/j.envpol.2020.115726 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 71. Bullert AJ, Li X, Gautam B, Wang H, Adamcakova-Dodd A, Wang K, Thorne PS, & Lehmler H-J (2024). Distribution of 2,2′,5,5′-tetrachlorobiphenyl (pcb52) metabolites in adolescent rats after acute nose-only inhalation exposure. Environ Sci Technol, 58(14), 6105–6116. 10.1021/acs.est.3c09527 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 72. Li X, Zhang C, Wang K, & Lehmler H-J (2020). Fatty liver and impaired hepatic metabolism alter the congener-specific distribution of polychlorinated biphenyls (pcbs) in mice with a liver-specific deletion of cytochrome p450 reductase. Environ Pollut, 266, 115233. https://doi.org/ 10.1016/j.envpol.2020.115233 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 73. Davis J (2002). Statistics and data analysis in geology (3 ed.). John Wiley. [ Google Scholar ] 74. Uppal K, Ma C, Go YM, Jones DP, & Wren J (2018). Xmwas: A data-driven integration and differential network analysis tool. Bioinformatics, 34(4), 701–702. 10.1093/bioinformatics/btx656 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 75. Dean LE, Wang H, Bullert AJ, Wang H, Adamcakova-Dodd A, Mangalam AK, Thorne PS, Ankrum JA, Klingelhutz AJ, & Lehmler H-J (2024). Inhalation of 2,2′,5,5′-tetrachlorobiphenyl (pcb52) causes changes to the gut microbiome throughout the gastrointestinal tract. J Hazard Mater, 480, 135999. https://doi.org/ 10.1016/j.jhazmat.2024.135999 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 76. Cao K-AL, Rossouw D, Robert-Granié C, & Besse P (2008). A sparse pls for variable selection when integrating omics data. Stat Appl Genet Mol Biol, 7(1), Article 35. https://doi.org/doi: 10.2202/1544-6115.1390 [ DOI ] [ PubMed ] [ Google Scholar ] 77. Blondel VD, Guillaume J-L, Lambiotte R, & Lefebvre E (2008). Fast unfolding of communities in large networks. J Stat Mech: Theory Exp, 2008(10), P10008. 10.1088/1742-5468/2008/10/P10008 [ DOI ] [ Google Scholar ] 78. Shannon P, Markiel A, Ozier O, Baliga NS, Wang JT, Ramage D, Amin N, Schwikowski B, & Ideker T (2003). Cytoscape: A software environment for integrated models of biomolecular interaction networks. Genome Res, 13(11), 2498–2504. 10.1101/gr.1239303 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 79. Otasek D, Morris JH, Bouças J, Pico AR, & Demchak B (2019). Cytoscape automation: Empowering workflow-based network analysis. Genome Biol, 20(1), 185. 10.1186/s13059-019-1758-4 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 80. Petriello MC, Hoffman JB, Vsevolozhskaya O, Morris AJ, & Hennig B (2018). Dioxin-like pcb 126 increases intestinal inflammation and disrupts gut microbiota and metabolic homeostasis. Environ Pollut, 242(Pt A), 1022–1032. 10.1016/j.envpol.2018.07.039 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 81. Wahlang B, Alexander NC 2nd, Li X, Rouchka EC, Kirpich IA, & Cave MC (2021). Polychlorinated biphenyls altered gut microbiome in car and pxr knockout mice exhibiting toxicant-associated steatohepatitis. Toxicol Rep, 8, 536–547. 10.1016/j.toxrep.2021.03.010 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 82. Wang H, Adamcakova-Dodd A, Flor S, Gosse L, Klenov VE, Stolwijk JM, Lehmler HJ, Hornbuckle KC, Ludewig G, Robertson LW, & Thorne PS (2020). Comprehensive subchronic inhalation toxicity assessment of an indoor school air mixture of pcbs. Environ Sci Technol, 54(24), 15976–15985. 10.1021/acs.est.0c04470 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 83. Cheng SL, Li X, Lehmler HJ, Phillips B, Shen D, & Cui JY (2018). Gut microbiota modulates interactions between polychlorinated biphenyls and bile acid homeostasis. Toxicol Sci, 166(2), 269–287. 10.1093/toxsci/kfy208 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 84. Choi JJ, Eum SY, Rampersaud E, Daunert S, Abreu MT, & Toborek M (2013). Exercise attenuates pcb-induced changes in the mouse gut microbiome. Environ Health Perspect, 121(6), 725–730. 10.1289/ehp.1306534 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 85. Zhao X, Zhong X, Liu X, Wang X, & Gao X (2021). Therapeutic and improving function of lactobacilli in the prevention and treatment of cardiovascular-related diseases: A novel perspective from gut microbiota. Front Nutr, 8, 693412. 10.3389/fnut.2021.693412 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 86. Ilinskaya ON, Ulyanova VV, Yarullina DR, & Gataullin IG (2017). Secretome of intestinal bacilli: A natural guard against pathologies. Front Microbiol, 8, 1666. 10.3389/fmicb.2017.01666 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 87. Rude KM, Keogh CE, & Gareau MG (2019). The role of the gut microbiome in mediating neurotoxic outcomes to pcb exposure. Neurotoxicology, 75, 30–40. 10.1016/j.neuro.2019.08.010 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 88. Bullert AJ, Wang H, Valenzuela AE, Neier K, Wilson RJ, Badley JR, LaSalle JM, Hu X, Lein PJ, & Lehmler H-J (2024). Interactions of polychlorinated biphenyls and their metabolites with the brain and liver transcriptome of female mice. ACS Chem Neurosci, 15(21), 3991–4009. 10.1021/acschemneuro.4c00367 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 89. Franklin CL, & Ericsson AC (2017). Microbiota and reproducibility of rodent models. Lab Anim (NY), 46(4), 114–122. 10.1038/laban.1222 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 90. Hasan N, & Yang H (2019). Factors affecting the composition of the gut microbiota, and its modulation. PeerJ, 7, e7502. 10.7717/peerj.7502 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 91. Rothschild D, Weissbrod O, Barkan E, Kurilshikov A, Korem T, Zeevi D, Costea PI, Godneva A, Kalka IN, Bar N, Shilo S, Lador D, Vila AV, Zmora N, Pevsner-Fischer M, Israeli D, Kosower N, Malka G, Wolf BC, Avnit-Sagi T, Lotan-Pompan M, Weinberger A, Halpern Z, Carmi S, Fu J, Wijmenga C, Zhernakova A, Elinav E, & Segal E (2018). Environment dominates over host genetics in shaping human gut microbiota. Nature, 555(7695), 210–215. 10.1038/nature25973 [ DOI ] [ PubMed ] [ Google Scholar ] 92. Knight R, Vrbanac A, Taylor BC, Aksenov A, Callewaert C, Debelius J, Gonzalez A, Kosciolek T, McCall L-I, McDonald D, Melnik AV, Morton JT, Navas J, Quinn RA, Sanders JG, Swafford AD, Thompson LR, Tripathi A, Xu ZZ, Zaneveld JR, Zhu Q, Caporaso JG, & Dorrestein PC (2018). Best practices for analysing microbiomes. Nat Rev Microbiol, 16(7), 410–422. 10.1038/s41579-018-0029-9 [ DOI ] [ PubMed ] [ Google Scholar ] 93. He S, Zhao C, Guo Y, Zhao J, Xu X, Hu Y, Lian B, Ye H, Wang N, Luo L, & Liu Q (2023). Alterations in the gut microbiome and metabolome profiles of septic mice treated with shen fuhuang formula. Front Microbiol, 14, 1111962. 10.3389/fmicb.2023.1111962 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 94. Zada S, Xie J, Yang M, Yang X, Sajjad W, Rafiq M, Hasan F, Hu Z, & Wang H (2021). Composition and functional profiles of microbial communities in two geochemically and mineralogically different caves. Appl Microbiol Biotechnol, 105(23), 8921–8936. 10.1007/s00253-021-11658-4 [ DOI ] [ PubMed ] [ Google Scholar ] 95. Birnbaum LS (1985). The role of structure in the disposition of halogenated aromatic xenobiotics. Environ Health Perspect, 61, 11–20. 10.1289/ehp.856111 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 96. Kania-Korwel I, Zhao H, Norstrom K, Li X, Hornbuckle KC, & Lehmler HJ (2008). Simultaneous extraction and clean-up of polychlorinated biphenyls and their metabolites from small tissue samples using pressurized liquid extraction. J Chromatogr A, 1214(1–2), 37–46. 10.1016/j.chroma.2008.10.089 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 97. Dhakal K, He X, Lehmler HJ, Teesch LM, Duffel MW, & Robertson LW (2012). Identification of sulfated metabolites of 4-chlorobiphenyl (pcb3) in the serum and urine of male rats. Chem Res Toxicol, 25(12), 2796–8204. 10.1021/tx300416v [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 98. Quinete N, Esser A, Kraus T, & Schettgen T (2017). Pcb 28 metabolites elimination kinetics in human plasma on a real case scenario: Study of hydroxylated polychlorinated biphenyl (oh-pcb) metabolites of pcb 28 in a highly exposed german cohort. Toxicol Lett, 276, 100–107. https://doi.org/ 10.1016/j.toxlet.2017.05.025 [ DOI ] [ PubMed ] [ Google Scholar ] 99. Randerath I, Quinete N, Müller JP, Stingl J, Bertram J, Schettgen T, Kraus T, & Ziegler P (2024). Partial dechlorination of 2,4,4′-trichlorobiphenyl (pcb 28) mediated by recombinant human cyp1a2. Arch Toxicol, 98(1), 159–163. 10.1007/s00204-023-03621-1 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 100. Hu X, Lehmler H-J, Adamcakova-Dodd A, & Thorne PS (2013). Elimination of inhaled 3,3′-dichlorobiphenyl and the formation of the 4-hydroxylated metabolite. Environ Sci Technol, 47(9), 4743–4751. 10.1021/es3049114 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 101. Zhu Y, Mapuskar KA, Marek RF, Xu W, Lehmler HJ, Robertson LW, Hornbuckle KC, Spitz DR, & Aykin-Burns N (2013). A new player in environmentally induced oxidative stress: Polychlorinated biphenyl congener, 3,3'-dichlorobiphenyl (pcb11). Toxicol Sci, 136(1), 39–50. 10.1093/toxsci/kft186 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 102. Wu X, Barnhart C, Lein PJ, & Lehmler HJ (2015). Hepatic metabolism affects the atropselective disposition of 2,2',3,3',6,6'-hexachlorobiphenyl (pcb 136) in mice. Environ Sci Technol, 49(1), 616–625. 10.1021/es504766p [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 103. Wu X, Zhai G, Schnoor JL, & Lehmler H-J (2020). Atropselective disposition of 2,2′,3,4′,6-pentachlorobiphenyl (pcb 91) and identification of its metabolites in mice with liver-specific deletion of cytochrome p450 reductase. Chem Res Toxicol, 33(6), 1328–1338. 10.1021/acs.chemrestox.9b00255 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 104. Zhang C-Y, Li X, Flor S, Ruiz P, Kruve A, Ludewig G, & Lehmler H-J (2022). Metabolism of 3-chlorobiphenyl (pcb 2) in a human-relevant cell line: Evidence of dechlorinated metabolites. Environ Sci Technol, 56(17), 12460–12472. 10.1021/acs.est.2c03687 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 105. Tulp MTM, Bruggeman WA, & Hutzinger O (1977). Reductive dechlorination of chlorobiphenylols by rats. Experientia, 33(9), 1134–1136. 10.1007/BF01922284 [ DOI ] [ PubMed ] [ Google Scholar ] 106. Duffel MW, & Lehmler HJ (2024). Complex roles for sulfation in the toxicities of polychlorinated biphenyls. Crit Rev Toxicol, 54(2), 92–122. 10.1080/10408444.2024.2311270 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 107. Liu J, Tan Y, Song E, & Song Y (2020). A critical review of polychlorinated biphenyls metabolism, metabolites, and their correlation with oxidative stress. Chem Res Toxicol, 33(8), 2022–2042. 10.1021/acs.chemrestox.0c00078 [ DOI ] [ PubMed ] [ Google Scholar ] Associated Data This section collects any data citations, data availability statements, or supplementary materials included in this article. Supplementary Materials 1 NIHMS2061747-supplement-1.docx (483.1KB, docx) ACTIONS View on publisher site PDF (2.8 MB) Cite Collections Permalink RESOURCES Similar articles Cited by other articles Links to NCBI Databases Follow NCBI NCBI on X (formerly known as Twitter) NCBI on Facebook NCBI on LinkedIn NCBI on GitHub NCBI RSS feed Connect with NLM NLM on X (formerly known as Twitter) NLM on Facebook NLM on YouTube National Library of Medicine 8600 Rockville Pike Bethesda, MD 20894 Web Policies FOIA HHS Vulnerability Disclosure Help Accessibility Careers NLM NIH HHS USA.gov Back to Top

Related documents

• Effects of Endocrine Disruptors on the Gut Microbiota and Assessment of Their Impact on Colorectal Cancer Development (PERMICA)#1008341
• The French Gut-colo#906071
• Li, F.; Wu, H.; Li, L.; Li, X.; Zhao, J.; Peijnenburg, W. J. G. M. Docking and QSAR study on the binding interactions between polycyclic aromatic hydrocarbons and estrogen receptor. Ecotoxicology and Environmental Safety 2012, 80, 273 - 279.#933219
• Spatial transcriptomic profiling uncovers the molecular effects of the neurotoxicant polychlorinated biphenyls (PCBs) in the brains of adult mice#224199
• Nationwide cohort study reveals low bifidobacteria and distinct microbiota composition and function in Swedish newborns.#610382