Bisphenol A exposure affects specific gut taxa and drives microbiota dynamics in childhood obesity

Abstract

Cumulative xenobiotic exposure has an environmental and human health impact which is currently assessed under the One Health approach. Bisphenol A (BPA) exposure and its potential link with childhood obesity that has parallelly increased during the last decades deserve special attention. It stands during prenatal or early life and could trigger comorbidities and non-communicable diseases along life. Accumulation in the nature of synthetic chemicals supports the "environmental obesogen" hypothesis, such as BPA. This estrogen-mimicking xenobiotic has shown endocrine disruptive and obesogenic effects accompanied by gut microbiota misbalance that is not yet well elucidated. This study aimed to investigate specific microbiota taxa isolated and selected by direct BPA exposure and reveal its role on the overall children microbiota community and dynamics, driving toward specific obesity dysbiosis. A total of 333 BPA-resistant isolated species obtained through culturing after several exposure conditions were evaluated for their role and interplay with the global microbial community. The selected BPA-cultured taxa biomarkers showed a significant impact on alpha diversity. Specifically, Clostridium and Romboutsia were positively associated promoting the richness of microbiota communities, while Intestinibacter, Escherichia-Shigella, Bifidobacterium, and Lactobacillus were negatively associated. Microbial community dynamics and networks analyses showed differences according to the study groups. The normal-weight children group exhibited a more enriched, structured, and connected taxa network compared to overweight and obese groups, which could represent a more resilient community to xenobiotic substances. In this sense, subnetwork analysis generated with the BPA-cultured genera showed a correlation between taxa connectivity and more diverse potential enzymatic BPA degradation capacities.IMPORTANCEOur findings indicate how gut microbiota taxa with the capacity to grow in BPA were differentially represented within differential body mass index children study groups and how these taxa affected the overall dynamics toward patterns of diversity generally recognized in dysbiosis. Community network and subnetwork analyses corroborated the better connectedness and stability profiles for normal-weight group compared to the overweight and obese groups.

Keywords: BPA; amplicon-sequencing; culturomics; microbiota; obesity; xenobiotics.

Fig 1

Description of the children’s gut microbiota in NW, OW, and O groups assessed by culturomics. (A) Cumulative bar plot showing the phylogeny of isolated microorganisms at the genus level. (B and C) Box plots of α-diversity indices of the microbiota across the three groups. Medians and interquartile ranges are shown. Group differences in the observed number of amplicon sequence variants and Shannon α-diversity were calculated using multivariable linear models on the log-transformed data and corrected for age. P values shown in the figure correspond to pairwise comparisons assessed with the multcomp package. (D) Nonmetric multidimensional scaling (NMDS) plot based on Bray-Curtis dissimilarity between samples, with data points and ellipses colored by study groups. Results of the permutational multivariate analysis of variance (PERMANOVA) test (R² and P value corrected for age) based on Bray-Curtis dissimilarity index ordination are indicated within the plot. Ellipses represent the standard deviation of all points within a study group.

Fig 2

(A) Tile map showing frequency of species isolation at 20 and 50 ppm of BPA in NW, O, and OW groups. The species are ordered according to their counts. (B) Venn diagram analysis shows common, shared, and unique species in the study groups. (C) Fold change (log₁₀) plots of BPA-tolerant genera (NW vs O; NW vs OW; O vs OW).

Fig 3

Description of the children’s gut microbiota according to the study groups assessed by 16S rRNA gene amplicon sequencing. (A) The figure displays the mean relative abundance of bacterial genera in the study groups. (B and C) Box plots of the α-diversity indices of the microbiota across the three groups. Medians and interquartile ranges are shown. Group differences in observed number of ASVs and Shannon α-diversity were calculated using multivariable linear models on the raw data and log transformed, respectively, and corrected for age. P values shown in the figure correspond to pairwise comparisons assessed with the multcomp package. (D) NMDS plot based on Bray-Curtis dissimilarity between samples, with data points and ellipses colored by study groups. Results of the PERMANOVA test (R² and P value corrected for age) are indicated within the plot. Ellipses represent the standard deviation of all points within a study group.

Fig 4

Volcano plot depicting differential abundant ASVs related to the three study groups. (A–C) Shade color corresponds to the study group where the specific ASVs were enriched. Gray dots represent species that were not differentially abundant; green dots represent ASVs that had a log2(FC) >1 but were not significantly differentially abundant after correcting for multiple testing; red dots represent ASVs that were significantly differentially abundant and had a log2(FC) >1. All these tests were corrected for age. Abbreviations: FC, fold change; ns, non-significant.

Fig 5

Microbial networks and network metrics for each study group. (A–C) Overall networks of the gut microbial community for each study group. Nodes represent individual ASVs. Shaded areas around groups of nodes represent clusters defined by cluster fast greedy community analysis. (D–G) Degree, betweenness, harmonic centrality, and hub score, respectively, across the three study groups. (H) Main clusters (clusters with more than 10 ASVs) and number of ASVs within each cluster for the three study groups networks. (I) Venn diagram analysis of the hub taxa shared within the study groups.

Fig 6

Cultured taxa biomarkers’ influence on the α-diversity indices of the microbial community. Spearman’s rank test shows the correlation between discriminant taxa of BPA tolerance with α-diversity metrics.

Fig 7

Correlation between biomarker’s genera and community structure based on β-diversity. NMDS plot based on Bray-Curtis dissimilarity between samples, with data points shaped by study groups and colored by relative abundance of the biomarker’s genera. (A) Bifidobacterium; (B) Romboutsia; (C) Intestinibacter; (D) Clostridium sensu stricto 1; (E) Turicibacter; (F) Lactobacillus; (G) Bacillus; results of the PERMANOVA test (R² and P value) are indicated within the plots.

Fig 8

Microbial subnetworks of biomarkers ASVs for each study group. (A–C) Subnetworks of the BPA-tolerant gut microbial ASVs for each study group. Nodes represent individual ASVs. Shaded areas around groups of nodes represent clusters defined by cluster fast greedy community analysis. BPA metabolic classification: B, biodegrader; T, tolerant BPA; R, resistant.