Prenatal PFAS exposure and outcomes related to maternal gut microbiome composition in later pregnancy

Abstract

The composition of the gut microbiome is dependent on factors including diet, lifestyle, and exposure to environmental chemicals, and has implications for human health. Per-and polyfluoroalkyl substances (PFAS), a class of man-made chemicals that have nonstick and flame-retardant properties may impact on gut microbiome composition. Our objective was to elucidate links between PFAS and maternal gut microbiome composition in two geographically diverse sites of the Environmental Influences on Child Health Outcomes program. The present analysis includes participants in the Atlanta African American Maternal Child Cohort; ATL AA and a predominately non-Hispanic White subsample of the Michigan Archive for Research on Child Health Cohort; MARCH with serum or plasma PFAS concentrations measured in early or late pregnancy and 16s rRNA sequencing from maternal gut microbiome samples available primarily in later pregnancy (2nd-3rd trimester). Linear regression models tested associations between prenatal PFAS levels (separately for the 1st/3rd trimesters) and measures of alpha diversity, bacterial composition differences, and differential taxonomic abundance. Bayesian Kernel Machine Regression and Elastic net regression mixture modeling were also incorporated. In both cohorts, multiple PFAS were significantly associated with the relative abundance of specific microbiome taxa even after adjustment for covariates including maternal diet, age, race, BMI, and smoking; A total of 16 significant family-level associations were identified for ATL AA (e.g., PFOA with Clostridiaceae; natural log fold change = 0.94) and 13 significant family-level associations identified for MARCH e.g., PFOS with Desulfovibrionaceae; natural log fold change =-1.53 (pFDR<0.05), but similarities between cohorts were lacking. Mixture analyses did not identify interactive or combined effects but did provide modest evidence of inclusion of individual PFAS in beta diversity models in both cohorts. In 2 distinct cohorts, there were significant associations between prenatal PFAS and the relative abundance of several bacterial taxa, but these differences were cohort-specific. This work suggests that PFAS may modulate the gut microbiome during pregnancy.