Maternal Polychlorinated biphenyl 126 (PCB 126) exposure modulates offspring gut microbiota irrespective of diet and exercise

Abstract

The gut microbiota plays an important role throughout the lifespan in maintaining host health, and several factors can modulate microbiota composition including diet, exercise, and environmental exposures. Maternal microbiota is transferred to offspring during early life; thus, environmental exposures before gestation may also modulate offspring microbiota. Here we aimed to investigate the effects of maternal exposure to dioxin-like polychlorinated biphenyls (PCBs) on the microbiota of aged offspring and to determine if lifestyle factors, including maternal exercise or offspring high-fat feeding alter these associations. To test this, dams were exposed to PCB 126 (0.5 μmole/kg body weight) or vehicle oil by oral gavage during preconception, gestation, and during lactation. Half of each group was allowed access to running wheels for ≥7 days before and during pregnancy and up through day 14 of lactation. Female offspring born from the 4 maternal groups (PCB exposure or not, with/without exercise) were subsequently placed either on regular diet or switched to a high-fat diet during adulthood. Microbiota composition was quantified in female offspring at 49 weeks of age by 16S rRNA sequencing. Maternal exposure to PCB 126 resulted in significantly reduced richness and diversity in offspring microbiota regardless of diet or exercise. Overall compositional differences were largely driven by offspring diet, but alterations in specific taxa due to maternal PCB 126 exposure, included the depletion of Verrucomicrobiaceae and Akkermansia muciniphila, and an increase in Anaeroplasma. Perturbation of microbiota due to PCB 126 may predispose offspring to a variety of chronic diseases later in adulthood.

Introduction

Gut microbiota refers to the set of microorganisms such as bacteria, viruses, protozoa, and fungi that coexist in the digestive tract and make up an integral part of the body [1]. Numerous factors impact the form and function of gut microbiota, with dysbiosis and modulation of microbiota impacting human health [2]. A host’s environment and lifestyle have long been shown to modulate microbiota taxa and diversity with much emphasis focusing on diet and personal exposures including cigarette smoking. Recently, exposures to ubiquitous environmental chemicals including persistent organic pollutants such as polychlorinated biphenyls (PCBs) have also been shown to directly or indirectly impact gut health [3-7]. Although PCBs are no longer manufactured in the United States since they were banned in 1979 due to adverse health effects on multiple organ systems [8-11], they continue to contaminate our environment through multiple routes including improper disposal or incineration of contaminated waste and leaks from old transformers and capacitors[12, 13]. Currently, the major source of human exposure to PCBs is through consumption of contaminated food especially fish from contaminated lakes and rivers, meat, and dairy products [12].

Microbiota is vertically transferred from mothers to offspring, thus exposures that impact the microbiota of mothers prior to or during conception—as well as during pregnancy—may also impact their offspring’s microbiota [14, 15]. As modulation of normal microbiota homeostasis is associated with many chronic diseases, this paradigm can be seen as a growing research area related to the Developmental Origins of Health and Disease (DOHaD) hypothesis [16, 17]. Studies in adult rodents have shown that PCBs and dioxin-like pollutant mixtures can significantly impact gut microbiota diversity and taxa [3, 7, 18-20]. We and others have shown that PCB exposure significantly decreases the relative abundance of Proteobacteria, Bacteroidales family S7-25, and Alistipes compared with unexposed controls [6]. A decrease in diversity and an increase in the Firmicutes to Bacteroidetes ratio (F/B ratio) has also been demonstrated [2, 21]. These dioxin-like pollutants have been linked to metabolic dysfunction, reproductive disorders, neurobehavioral impairments, and immune dysfunction [22], which is potentially mediated by alterations to microbiota.

Several studies have implicated maternal diet and exercise with modulation of offspring gut microbiota [19, 23-27], but to our knowledge, no group has focused on other environmental exposures such as dioxin-like pollutants. In addition, only a few studies in adult rodents have investigated if other factors known to impact the microbiota, such as diet and exercise, may mask or modulate toxicant-induced changes in the gut microbiota [6, 18]. Our present study seeks to examine the impact of maternal exposure to the model dioxin-like pollutant, PCB 126, on offspring gut microbiota, and if maternal exercise or offspring diet can mask or exacerbate these changes. We hypothesized that perinatal exposure to PCBs will result in decreased alpha and compositional differences in the gut microbiota of the offspring, despite the offspring never being directly exposed to PCBs via oral gavage. We further hypothesized that these changes in offspring may be attenuated by maternal exercise, and that feeding offspring with a high-fat diet will exacerbate the effects on the gut microbiota.

Materials and Methods

In vivo study design

These studies were carried out at the University of Kentucky according to an approved Institutional Animal Care and Use Committee (IACUC # 2009-0503) protocol. Two to four-month-old proven ICR breeders from Envigo Harlan were shipped to the University of Kentucky and housed four per cage for one week to acclimate. During this period, they were on regular mouse diet 2018 (Harlan). The following week, they were individually housed in circadian lightbox cabinets (Phenome Technologies, Lincolnshire, IL) for habituation and placed on maternal breeder diet 5008 (Lab Diet, St. Louis, MO). A week after being in the light box, dams were weighed and placed in the following groups: Group 1- Sedentary/Vehicle (Sed/Veh) exposed to safflower oil with no access to a running wheel, Group 2- Exercise/Vehicle (Ex/Veh) exposed to safflower oil with voluntary access to a running wheel, Group 3-Sedentary/PCB (Sed/PCB) exposed to PCB126 (0.5 μmole/kg body weight PCB126) with no access to a running wheel, and Group 4- Exercise/PCB exposed to PCB126 with voluntary access to a running wheel. All mice were housed in an environmentally controlled vivarium between 20-22°C with unlimited access to food and water under a controlled photoperiod (14:10-h light-dark). If applicable, mice had open access to the running wheels that were mounted within each cage (Lincolnshire, IL). A mechanical counter was used to record wheel rotations to a desktop computer via ClockLab software (Actimetrics, Wilmette, IL). Sedentary female mice were housed in nearly identical cages that did not contain running wheels as we did in our previous study[28]. Dams were allowed to exercise ≥7 days before and during pregnancy and up through day 14 of lactation. The exercise was completely voluntary.

Dams were exposed 3 times to PCB 126: The first exposure was administered 72 hours before mating. The second exposure was administered on approximate gestational day 7 and the last exposure was administered on a postnatal day 7 (PND7) during lactation. The current study utilizes a similar dosing paradigm in previous studies [29, 30]. The pregnant dams in the exercise cohort had continual access to running wheels throughout pregnancy until 14 days after giving birth, at which point wheels were removed to prevent possible harm to growing pups Litters were culled to eight or nine pups ~48 h after birth. Pups were cross-fostered from other litters from the same group if they did not have at least eight pups.

Female offspring were weaned at 21d (3 weeks) and placed on Teklad Global 18% Protein Rodent Diet No. 2018 (regular mouse diet 2018). The housing density was five per cage and each weanling cage had a (mix of litter)-one litter from a different dam within the same treatment group. At 39 weeks of age, half of the female offspring were placed on high-fat feeding D12492 with 60 kcal % fat (Research Diets; New Brunswick, NJ) as a means to induce a secondary metabolic hit in adulthood in an attempt to further elucidate PCB 126 induced impacts. The other half of the offspring remained on regular diet for the duration of the study. The offspring themselves did not have access to running wheels during any portion of the study. At the end of the study, offspring were fasted for 3 hours before euthanasia via carbon dioxide, and cecum samples were collected at 49 weeks of age, snap-frozen in liquid nitrogen, and stored at −80 degrees C until analysis (offspring n=5 from each group). As is typical of DOHaD studies, offspring sample size indicates the number of dams/litters (only one offspring per litter were included in each of the 4 maternal exposure/exercise groups; though siblings were included in the offspring dietary feeding portion of the experiment). Male offspring were used for different experiments, so were not included in the microbiota studies.

DNA extraction and 16S rRNA sequencing

DNA extraction and 16S rRNA gene sequencing of the V4-V5 region was conducted by the Environmental Sample Preparation and Sequencing Facility (ESPSF) at Argonne National Laboratory as discussed previously[2]. Data processing was conducted using Quantitative Insights into Microbial Ecology (QIIME version 1.9) by ESPSF. Reads were joined using the command join_paired_ends.py and demultiplexed using the command “split_libaries_fastq.py” with the default parameters of 1.5 mismatches. To pick de novo operational taxonomic units (OTUs), the command “pick_de_novo_otus.py” was used and observations that were present less than twice were filtered from the OTU table before further analysis. Read depths in the filtered, unrarefied table ranged from 56 to 174,324 (Median [Q1; Q3]=85,836 [72,674; 102,775]). The sample with the minimum depth of 56 was considered defective (2^nd smallest depth=56,285) and was excluded from analyses. A rarefied OTU table was calculated for alpha and beta diversity analyses, which was derived using a multiple rarefaction algorithm [31] at a depth of 56,285 (retaining all non-defective samples). The unrarefied OTU table was using for taxonomic testing, described in detail below.

Statistical analyses

A power calculation was performed to determine the detectable effect size for the association between alpha diversity and PCB exposure by diet/exercise using two-way ANOVA with interaction terms. Assuming a type I error rate of 5%, we are able to detect an effect size of (quantified using Cohen’s f) of 0.46 with 80% power. According to Cohen [32], 0.10 is considered a small effect size, 0.30 is considered a medium effect size, and 0.50 is considered a large effect size. Therefore, we are only able to detect interactions with large effect sizes given our current sample size.

Alpha diversity metrics (richness, Pielou’s evenness, and Faith’s phylogenetic diversity) were calculated using the R packages picante [33] and vegan[34]. Unweighted and weighted UniFrac beta diversity metrics were calculated using the R package phyloseq [35] Weighted Unifrac accounts for the abundance of the observed organisms, whereas unweighted Unifrac only considers the presence or absence of organisms [36]. Differences in alpha diversity metrics were tested using linear regression models, first examining the 3-way interaction (PCB*exercise*diet), followed by 2-way interactions and main effects. Compositional difference testing was performed with PERMANOVA using the R package vegan, testing 3-way and 2-way interactions, and main effects. Principal coordinate analysis (PCoA) was utilized to visually examine compositional distinction across experimental conditions. While examining compositional shifts in the microbiota is helpful to observe broad changes, assessing specific taxonomic differences, specifically at the family, genus, and OTU level, allows for the assessment of individual groups of microbes impacted and provides a basis for an inference of microbial functionality. The R package DESeq2 [37] was used for taxonomic testing. Multiple testing correction was performed using Benjamini and Hochberg false discovery rate (FDR) correction with p_FDR<0.05 considered significant. For the dam and terminal body weight measurements, and F/B ratio, the statistical significance for all (p<0.05) was determined by Two-way or Three-way ANOVA analysis and posthoc comparisons by the Holm-Sidak method.

Results

Anthropometrics

This dose of PCB 126 or exercise paradigm did not impact body weights of the dams in any group during the duration of the study (Supplemental Figure 1). Neither maternal exercise nor PCB exposure significantly affected offspring body weight, but high-fat-fed offspring weighed significantly more than regular diet -fed mice (p=0.009) (Supplemental Figure 2). PCB exposure did not impact on running distance of dams for any group (data not shown).

Gut microbiota differences in alpha diversity by PCB exposure, diet, and exercise

When alpha diversity was assessed, no significant 3-way or 2-way interactions were noted for all three metrics (Table 1). After adjusting for both diet and exercise, significant main effects were observed for PCB 126 exposure for both richness (p=0.045) and Faith’s diversity (p=0.039), in which PCB exposure resulted in lower levels of both metrics. Additionally, significant main effects of high-fat diet consumption were observed for richness (p<0.001), evenness (p<0.001), and diversity (p<0.001), in which high-fat diet consumption yielded lower levels compared to the regular fat diet. No significant effects of maternal exercise were observed.

Table 1:

Differences in alpha diversity by PCB exposure, diet, and exercise.

	Richness	Pielou’s Evenness	Faith’s Diversity
		β (95% CI); p-value
3-Way Interactions a
PCB*Exercise*Diet	−69 (−646, 508); p=0.81	−0.03 (−0.12, 0.06); p=0.48	−17.2 (−50.1, 15.8); p=0.30
2-Way Interactions b
PCB* Exercise	−18 (−302, 265); p=0.90	0.02 (−0.03, 0.06); p=0.46	1.6 (−14.8, 18.1); p=0.84
PCB*Diet	141 (−143, 425); p=0.32	0.02 (−0.02, 0.06); p=0.33	4.0 (−12.4, 20.5); p=0.62
Exercise*Diet	−167 (−451, 116); p=0.24	0.001 (−0.04, 0.04); p=0.98	−13.1 (−29.5, 3.4); p=0.12
Main Effects c
PCB (Exposed vs. Unexposed)	−144 (−284, −3); p=0.045	−0.01 (−0.03, 0.01); p=0.28	−8.6 (−16.8, −0.5); p=0.039
Exercise (Yes vs. No)	33 (−108, 173); p=0.64	0.01 (−0.01, 0.03); p=0.40	2.4 (−5.8, 10.6); p=0.56
Diet (High Fat vs. Regular)	−325 (−465, −185); p<0.001	−0.07 (−0.09, −0.04); p<0.001	−28.2 (−36.4, −20.0); p<0.001

^aIn a linear model including all main effects, 2-way interactions, and the 3-way interaction

^bIn a linear model including all main effects and 2-way interactions

^cIn a linear model including all main effects

Gut microbiota compositional differences by PCB exposure, diet, and exercise

When compositional differences were visually assessed, a large effect of diet most clearly separated microbiota compositions (Supplemental Figure 3 and Supplemental Figure 4). When associations were formally tested in PERMANOVA models (Table 2), the 3-way interaction for Weighted UniFrac distance had a p-value of 0.056; thus, exploratory sub-group analyses was completed (Table 3). Similar significant interactions between PCB exposure and exercise were observed for mice on a high-fat diet (R²=0.098; p=0.042) and a regular diet (R²=0.094; p=0.030). In weighted UniFrac subgroup analyses, a PCB exposure effect was only significant among offspring who were from exercised mothers and fed a regular diet (R²=0.175; p=0.033). Similar results were seen for Unweighted UniFrac distances (Tables 2 and 3). If interaction effects are ignored, significant main effects were observed for diet only (Table 2), using both unweighted and weighted UniFrac (diet explains 13% and 32% of the variability in unweighted and weighted UniFrac distances, respectively (both p<0.001)).

Table 2:

Compositional differences by PCB exposure, diet, and exercise.

	Unweighted UniFrac		Weighted UniFrac
	R2	p-value	R2	p-value
3-Way Interaction a
PCB*Exercise*Diet	0.028	0.092	0.034	0.056
2-Way Interactions b
Exercise*Diet	0.028	0.091	0.020	0.26
Diet*PCB	0.025	0.18	0.017	0.39
Exercise*PCB	0.028	0.084	0.030	0.090
Main Effects c
PCB	0.027	0.14	0.021	0.26
Diet	0.133	<0.001	0.321	<0.001
Exercise	0.027	0.12	0.033	0.072

^aIn a PERMANOVA model including all main effects, 2-way interactions, and the 3-way interaction

^bIn a PERMANOVA model including all main effects and 2-way interactions

^cIn a PERMANOVA model including all main effects

Table 3:

Compositional differences by PCB exposure – subgroup specific effects.

	Unweighted UniFrac		Weighted UniFrac
	R2	p-value	R2	p-value
2-Way Interactions within Subgroups
PCB*Diet among Exercised	0.056	0.094	0.067	0.069
PCB*Diet among Non-Exercised	0.050	0.21	0.039	0.29
PCB*Exercise among High Fat Diet	0.062	0.040	0.098	0.042
PCB*Exercise among Regular Diet	0.068	<0.001	0.094	0.030
PCB Effect within Subgroups
PCB among Exercised, High Fat Diet	0.138	0.061	0.189	0.077
PCB among Exercised, Regular Diet	0.144	0.009	0.175	0.033
PCB among Non-Exercised, High Fat Diet	0.116	0.23	0.175	0.11
PCB among Non-Exercised, Regular Diet	0.134	0.016	0.131	0.25

Examination of the F/B ratio was employed to assess any interactions in broad compositional changes, as Firmicutes and Bacteroidetes are the two predominant phyla in the mammalian gut microbiota (Supplemental Figure 5). No significant 3-way or 2-way interactions were noted for the F/B ratio. However, a significant main effect of exercise was noted, with a lower F/B ratio observed in the offspring of exercised mothers (p=0.034). Additionally, a significant main effect of diet was noted, with a higher F/B ratio observed in high-fat-fed mice (p=0.015).

Taxonomic Differences Associated with PCB Exposure

For family-level abundances, after adjusting for diet and exercise, PCB exposure resulted in a significant increase in Anaeroplasmataceae abundance (log₂ fold change=3.07; p_FDR=0.009), as well as a significant reduction in Verrucomicrobiaceae abundance (log₂ fold change=−2.69; p_FDR<0.001), after adjusting for diet and exercise (Figure 1). At the genus level, PCB exposure resulted in a significant reduction in Akkermansia abundance compared to the vehicle-treated group (log₂FC=−2.78; p_FDR<0.001), as well as a significant increase in Anaeroplasma (log₂FC=2.98; p_FDR=0.014) and Prevotella (log₂FC=2.84; p_FDR=0.033) (Figure 2). At the OTU level, a total of 15 OTUs were significantly associated with PCB exposure after diet and exercise adjustment, the majority of which (12/15) were associated with decreased abundance following PCB exposure (Figure 3). Specifically, PCB exposure significantly decreased the abundance of OTUs in the families Alcaligenaceae (Sutterella sp.), Desulfovibrionaceae, Erysipelotrichaceae, Ruminococcaceae, and Verrucomicrobiaceae (Akkermansia muciniphila), while it significantly increased the abundance of an OTU in the family Anaeroplasmataceae (Anaeroplasma sp.)

Figure 1:

Families significantly associated with PCB exposure (p_FDR<0.05), after adjusting for diet and exercise. Log₂ fold change estimates compare PCB exposed to unexposed (positive: higher abundance in PCB exposed; negative: lower abundance in PCB exposed).

Figure 2:

Genera significantly associated with PCB exposure (p_FDR<0.05), after adjusting for diet and exercise. Log₂ fold change estimates compare PCB exposed to unexposed (positive: higher abundance in PCB exposed; negative: lower abundance in PCB exposed).

Figure 3:

OTUs significantly associated with PCB exposure (p_FDR<0.05), after adjusting for diet and exercise. Log₂ fold change estimates compare PCB exposed to unexposed (positive: higher abundance in PCB exposed; negative: lower abundance in PCB exposed).

Discussion

The gut microbiota is a sensitive marker of lifestyle factors including diet and exercise but also may be a sensitive marker of environmental pollutant exposure [38, 39]. Each of these factors during the perinatal period has been demonstrated to be involved in shaping the long-term health and disease risk of the offspring. Additionally, the perinatal period is important in shaping host microbial composition, which is believed to be a component contributing to offspring health and chronic disease risk [40, 41]. Thus, in the present study, we sought to examine the impact of maternal toxicant exposure on offspring gut microbial composition and to determine if other factors known to impact gut health, specifically diet and exercise, may modulate PCB-induced changes. Our results showed that maternal exposure to PCB 126 resulted in significantly reduced alpha diversity in offspring gut microbiota at during adulthood and changes in specific taxa, but significant interactions with diet or exercise were not observed.

Exposure to PCBs is commonly through dietary means thus PCB exposure could detrimentally impact the gut microbial composition and the gastrointestinal environment. A previous 14-week study, that used similar PCB 126 concentrations, also found that PCB exposure resulted in a reduction in alpha diversity measures [2], but this was completed in non-pregnant mice. It was observed in this study that PCB exposure induces shifts in specific microbial populations including lower levels of Clostridiales, Ruminococci, and Lactobacilli, and increased levels of Akkermansia [2]. These pollutant-induced microbial alterations are associated with systemic complications including toxicant-induced steatohepatitis [4] and neurotoxicity [42]. Notably, the present study differs from most of the previous research in that it examines the impact of maternal PCB exposure on offspring microbiota. While research is largely limited, Rude and colleagues found that exposure via the maternal diet resulted in a dysbiotic microbial signature [6].Specifically, they observed alterations in offspring beta diversity as a result of maternal PCB exposure. While we did not observe a significant impact of PCB exposure on overall beta diversity measures in the present study, we did find a significant interaction between PCB exposure and exercise in subgroup analysis, and we also observed that PCB exposure effects were masked in the high-fat diet in subgroup analysis (Table 3). However, we did observe that PCB exposure significantly decreased richness and phylogenetic diversity overall (Table 1), and also observed several taxonomic differences by PCB exposure overall. When comparing the present study to Rude et. al, there were no similarities in specific taxa altered as a result of maternal PCB exposure. The differences in our findings compared to Rude et. al. may be due to their use of a different mouse model, our exclusive use of female offspring, use of a different PCB mixture and/or differences in timing of PCB exposure [6]. Nevertheless, this growing body of evidence supports that there is a potential for maternal PCB exposure to impact offspring microbial composition, and thus should be further explored.

In the present study, because some of the 2-way and 3-way interactions for beta diversity measures had a p-value less than 0.10, we conducted exploratory analyses to further examine subgroup-specific effects (Table 3). We observed an interaction between PCB exposure and exercise in mice that were on a high-fat diet for both unweighted and weighted UniFrac distances. We also observed similar interactions between PCB exposure and exercise in mice that were on a regular diet. While exploratory, these findings suggest that there may be important interactions between maternal PCB exposure and exercise on offspring microbial composition that should be examined further in future studies utilizing a larger sample size. In a similar analysis examining the impact of PCB exposure in the four diets + exercise sub-groups, we observed a significant effect of PCB exposure on the microbial composition of mice that were exercised and on a regular diet (weighted and unweighted Unifrac distances). While preliminary, these findings suggest that the effect of maternal PCB exposure may be influenced by the combination of diet and exercise variables, and highlight a need for further exploration.

In the present study we observed several taxa that were significantly different between maternally PCB-exposed or non-exposed offspring. Firstly, Akkermansia muciniphila was significantly reduced as a result of maternal PCB exposure (Figure 3). It should be noted that Verrucomirobiacae and Akkermansia statistically exhibit the same magnitude of response since A. muciniphila is the sole microbe currently classified under that family and genus (Figures 1 and 2). A reduction in A. muciniphila has also been observed by others as a result of direct PCB exposure. For example, Tian and colleagues found that early-life PCB exposure significantly reduced A. muciniphila at 13 weeks after exposure. Interestingly, this group also found that this alteration persisted into adulthood [5]. Of note, A. muciniphila plays a role in the maintenance of the intestinal barrier, has been consistently inversely related to chronic disease risk and inflammation, and may play a role in the maintenance of metabolic health [43-45]. Desulfovibrionaceae spp. were also decreased in the present study as a result of maternal PCB exposure (Figure 3). While little has been observed with this family as a result of persistent organic pollutant exposure, several species within this family have the sulfate-reducing capacity with hydrogen sulfide as a product, which may increase intestinal inflammation and disrupt the intestinal barrier [46, 47]. However, other studies have demonstrated that sulfate-reducing bacteria like Desulfovibrio spp. do not always elicit detrimental impacts and may even have protective effects [48, 49]. In the present study, we also observed that Sutterella sp. was decreased as a result of PCB exposure. This alteration in Sutterella sp. has been observed with pollutant exposure previously. Johanson et al. observed that in a study examining maternal exposure to a mixture of persistent organic pollutants, an unclassified Sutterella sp. was reduced with pollutant exposure [50]. While most microbial groups were found to be reduced with maternal PCB exposure, one group that was higher with PCB exposure was Anaeroplasmataceae, including Anaeroplasma sp. (Figures 1 and 3). Although research is largely lacking, some studies have found an Anaeroplasma sp. to be increased in mice with a diabetic phenotype [51]. More research is needed to determine the main functions of these Anaeroplasma sp. before a clear hypothesis can be made as to why we observed an increase in offspring of maternally PCB-exposed mice.

In the present study, exercise was introduced as a potential modifier of pollutant-induced impact on the microbiome, as our prior research clearly demonstrates the effects of prenatal exercise on matured offspring insulin sensitivity and improves adult offspring glucose homeostasis [28, 52, 53]. Additionally, others have demonstrated that exercise is a probable modulator of the gut microbiome [54, 55] and recent studies have confirmed the link of exercise with the intestinal microbial community [56, 57]. Despite this growing body of research into the developing gut microbiome and dysbiosis, and the well-established benefits of exercise on maternal-child health and perinatal outcomes [58-60], few studies have investigated the impact of maternal exercise before, during and post-pregnancy on the gut microbiome of offspring. Notably, recent studies have found that maternal rats who had greater physical activity during their pregnancy transferred microbes beneficial to gut health to their offspring [19, 27]. In the current study, we did not observe exercise to be protective against pollutant-induced effects on the microbiome, as there were no significant 2-way or 3-way interactions noted for measures of microbial diversity or composition. Yet, from our compositional subgroup analysis we did observe a significant interaction between PCB exposure and exercise irrespective of diet. Future research should explore whether type, during, or timing of exercise intervention impacts any effects that may be exerted on the microbiome.

In addition to maternal exercise, we were also interested if high-fat feeding would exacerbate any observed PCB 126 effects in offspring. High-fat diets have been associated with alterations in alpha diversity, specifically, reductions in richness and evenness [61, 62], and modulate the toxicity of environmental pollutants in multiple mouse models [12, 47-49]. For example, Chi and colleagues showed that mice gut microbiota was significantly impacted by PCB exposure following high-fat diet feeding [63-65]. Additionally, high-fat feeding has been observed to result in an increased F/B ratio [2, 5]. We did not observe any significant interactions between diet and pollutant exposure in the present study, but we observed significant main effects due to diet. The microbial changes observed as a result of diet in the present study were consistent with the known effects of high-fat feeding, including an overall effect on alpha diversity and F/B ratio, with high-fat feeding being associated with lower measures of microbial richness, evenness, and diversity, and a higher F/B ratio [61, 62].

This study had multiple limitations including relatively small sample sizes for the individual groups and the exclusive examination of female offspring. Some studies have demonstrated that female mice appear to be more sensitive to PCB-induced toxicity, but we were unfortunately not able to investigate sex differences herein [66, 67]. Though we detected an overall effect of PCB exposure on gut microbiota, our study was only powered to detect large effect sizes for interactions between PCB and diet or exercise, and these largely failed to be detected. Future studies may consider larger sample sizes to better investigate attenuation by maternal exercise and exacerbation by high-fat diet. Additionally, it is important to note that human exposure to toxicants, like PCB 126, occurs at usually low levels over the life course. Our mouse model of relatively high acute exposures attempts to increase body burdens quickly during the short time of gestation in mice. Unfortunately, blood levels of PCB 126 were not measured in Dams from this current study. Additionally, we acknowledge the limitations of 16S rRNA sequencing for microbial analysis, as it only provides a snapshot of the microbial community and does not allow for a direct understanding of functionality. Similarly, here we could not determine if the microbiota changes are caused by maternal microbiota changes or if the developmental exposures to the PCBs themselves that alter the offspring microbiota. Future studies should focus on utilization of shotgun metagenomics approaches to more accurately investigate species and strain level changes as well as assess functional capacity of the microbial communities. Another limitation of our study is that we cannot determine if effects of maternal PCB exposure in the offspring is due to preconception exposure, in utero exposure, or to exposure during lactation, as there is evidence that dioxin-like PCBs can cross the placental barrier during gestation and excrete in milk during lactation [68, 69]. Future studies utilizing more rigorous cross-fostering strategies will be required to tease out these differences in exposure timing. Nevertheless, the findings of this study contribute to the growing body of evidence demonstrating the detrimental impact of dioxin-like pollutant exposure during critical windows of exposure and development and indicate a need for future research to examine this relationship and the role of lifestyle modifications in the exacerbation or mitigation of PCB-induced effects.

Highlights.

• Maternal exposure to PCB 126 decreased microbiome diversity in aged offspring

• Maternal exposure to PCB 126 impacted depleted offspring Akkermansia muciniphila

• Maternal exercise or offspring high-fat feeding did not modulate PCB effects