Maternal PBDE exposure disrupts gut microbiome and promotes hepatic proinflammatory signaling in humanized PXR-transgenic mouse offspring over time

Abstract

Developmental exposure to the persistent environmental pollutant, polybrominated diphenyl ethers (PBDEs), is associated with increased diabetes prevalence. The microbial tryptophan metabolite, indole-3-propionic acid (IPA), is associated with reduced risk of type 2 diabetes and lower-grade inflammation and is a pregnane X receptor (PXR) activator. To explore the role of IPA in modifying the PBDE developmental toxicity, we orally exposed humanized PXR-transgenic (hPXR-TG) mouse dams to vehicle, 0.1 mg/kg/day DE-71 (an industrial PBDE mixture), DE-71+IPA (20 mg/kg/day), or IPA, from 4 weeks preconception to the end of lactation. Pups were weaned at 21 days of age and IPA supplementation continued in the corresponding treatment groups. Tissues were collected at various ages until 6 months of age (n = 5 per group). In general, the effect of maternal DE-71 exposure on the gut microbiome of pups was amplified over time. The regulation of hepatic cytokines and prototypical xenobiotic-sensing transcription factor target genes by DE-71 and IPA was age- and sex-dependent, where DE-71-mediated mRNA increased selected cytokines (Il10, Il12p40, Il1β [both sexes], and [males]). The hepatic mRNA of the aryl hydrocarbon receptor (AhR) target gene Cyp1a2 was increased by maternal DE-71 and DE-71+IPA exposure at postnatal day 21 but intestinal Cyp1a1 was not altered by any of the exposures and ages. Maternal DE-71 exposure persistently increased serum indole, a known AhR ligand, in age- and sex-dependent manner. In conclusion, maternal DE-71 exposure produced a proinflammatory signature along the gut-liver axis, including gut dysbiosis, dysregulated tryptophan microbial metabolism, attenuated PXR signaling, and elevated AhR signaling in postweaned hPXR-TG pups over time, which was partially corrected by IPA supplementation.

Keywords: AhR; PBDE; PXR; diabetes; inflammation; metabolomics; microbiome.

Figure 1.

Experimental design of the study. Humanized PXR-transgenic (hPXR) dams were exposed to one of the following: (1) standard rodent chow, (2) DE-71 (an industrial PBDE mixture) via diet, (3) indole propionic acid (IPA) supplementation via drinking water; or (4) DE-71 via diet + IPA via drinking water, from 4 weeks preconception until lactation. The pups were weaned at postnatal day (PND) 21, when the DE-71 exposure stopped; whereas the IPA supplementation continued in the 2 corresponding groups (3 and 4). Tissues were collected at PND 21 as well as 3 months and 6 months of age. Intestinal contents were subjected to metagenomic shotgun sequencing. Serum tryptophan metabolites were quantified using LC-MS. Untargeted metabolomics was also performed in serum. RT-qPCR was performed to quantify the mRNA expression of distinct liver genes in these pups. Three-way analysis of variance (ANOVA) was first performed to determine the effect of sex, age, and chemicals, as well as their potential interactions. Because the primary focus of the study is the effect of chemicals, at each sex and age, 1-way ANOVA was then performed followed by Duncan’s post hoc test. Statistically significance differences were considered at p < .05.

Figure 2.

Differentially regulated intestinal bacteria at the species level in 3 ages of the pups (PND 21, 3 months, and 6 months of age; n = 5 per group). A, Number of species significantly changed by DE-71 exposure over time points. B, Examples of differentially regulated microbial species (top row), as well as the differentially regulated microbial DNA encoding swarming-related enzymes (bottom row) at PND 21. C, Bacteria species level at 3 months of age where maternal DE-71 exposure reduced the antiobesity and anti-inflammatory species in both sexes (top row) and the differentially regulated microbial DNA encoding swarming-related enzymes (bottom row) at 3 months of age. D, Bacteria species level (top row) and the differentially regulated microbial DNA encoding swarming-related enzymes (bottom row) at 6 months of age. To determine the chemical effect, statistical analysis was done by 1-way ANOVA followed by Duncan’s post hoc test with p-value < .05 being statistically significant. Groups a, b, and ab represent different post hoc groups.

Figure 3.

Differentially regulated intestinal bacterial KEGG pathway related to obesity and/or diabetes at PND 21 (A), 3 months of age (B), and 6 months of age (C). To determine the chemical effect, at each age and sex, statistical analysis was done by 1-way ANOVA followed by Duncan’s post hoc test, p-value < .05. Groups a, b, ab, bc, and c represent different post hoc groups.

Figure 4.

Tryptophan metabolites in serum and livers of mouse pups of the 4 exposure groups at various postnatal ages (PND 21, 3 months, and 6 months; n = 5 per group). A, Diagram of microbial tryptophan metabolism. B, Statistically significant tryptophan metabolites in serum at PND 21 (top row) and 3 months of age (bottom row). C, Statistically significant tryptophan metabolite in the liver of pups at 3 months of age. Statistical analysis was done by 1-way ANOVA followed by Duncan’s post hoc test, p-value < .05. Groups a, b, and ab represent different post hoc groups. Tryptophan metabolites in serum and livers of pups at 6 months of age were not included because they were not statistically significant.

Figure 5.

Two-way hierarchical clustering of differentially regulated untargeted metabolites in serums of PND 21 pups (n = 5 per group). A, Differentially regulated untargeted metabolites in the male serums of PND 21 pups. B, Differentially regulated untargeted metabolites in the female serums of PND 21 pups. C, Bar plots of statistically significant untargeted metabolites related to obesity and diabetes of 21 days old pups. Statistical analysis was done by 1-way ANOVA followed by Duncan’s post hoc test; asterisks on Figures 7A and 7B: p-value < .05. Groups a, b, and ab represent different post hoc groups.

Figure 6.

Two-way hierarchical clustering of differentially regulated untargeted metabolites in serums of 3 months old pups of the 4 exposure groups (n = 5 per group). A, Differentially regulated untargeted metabolites in the male serums of 3 months old pups. B, Differentially regulated untargeted metabolites in the female serums of 3 months old pups. C, Bar plot of untargeted metabolite related to obesity of 3 months old pups. Statistical analysis was done by 1-way ANOVA followed by Duncan’s post hoc test; asterisks on Figures 8A and 8B: p-value < .05. Groups a, b, and ab represent different post hoc groups.

Figure 7.

Two-way hierarchical clustering of differentially regulated untargeted metabolites in serums of 6 months old pups of the 4 exposure groups (n = 5 per group). A, Differentially regulated untargeted metabolites in the male serums of 6 months old pups. B, Differentially regulated untargeted metabolites in the female serums of 6 months old pups. C, Bar plots of statistically significant untargeted metabolites related to obesity and diabetes of 6 months old pups. Statistical analysis was done by 1-way ANOVA followed by Duncan’s post hoc test; asterisks on Figures 8A and 8B: p-value < .05. Groups a, b, and ab represent different post hoc groups.

Figure 8.

Summary diagram of the major findings of this study. Overall, the effect of maternal DE-71 exposure on hPXR-TG mice produced a proinflammatory signature within the gut-liver axis associated with dysregulated tryptophan microbial metabolism, attenuated PXR signaling, and elevated AhR signaling in an age- and sex-dependent, and some but not all of these effects can be at least partially corrected by IPA supplementation.