The potential mediating role of the gut microbiome and metabolites in the association between PFAS and kidney function in young adults: A proof-of-concept study

Abstract

Background: Chronic kidney disease (CKD) affects over 10 % of the global population and can lead to kidney failure and death. Exposure to per- and polyfluoroalkyl substances (PFAS) is associated with increased risk of CKD, yet studies examining the mechanisms linking PFAS and kidney function are lacking. In this exploratory study, we examined longitudinal associations of PFAS exposure with kidney function, and tested if associations were mediated by altered gut bacterial taxa or plasma metabolites using a multi-omics mediation analysis.

Methods: Seventy-eight young adults from the Children's Health Study were included in this longitudinal cohort study. At baseline, seven plasma PFAS and untargeted plasma metabolomics were measured using liquid chromatography/mass-spectrometry. Baseline gut bacterial abundance was characterized using 16S rRNA sequencing and examined at the genus level. At follow-up, serum creatinine and cystatin-C concentrations were quantified to estimate glomerular filtration rate (eGFR). High-dimensional multi-omics analyses were conducted to assess the association between baseline PFAS exposure with follow-up eGFR, mediated by gut microbiome and circulating metabolite levels.

Results: PFAS burden score, a variable developed to estimate exposure to chemical mixtures, was associated with kidney function. Each standard deviation increase in baseline PFAS burden score was associated with a 2.4 % lower eGFR at follow-up (95 % CI:[0.1 %,4.8 %]). Following high-dimensional mediation analyses with the microbiome and circulating metabolites, a joint component (characterized by reduced Lachnospiraceae and 17b-estradiol and increased succinate, retinoate and dodecanoic acid) and a metabolite component (characterized by increased hypotaurine and decreased D-pinitol and ureidopropionate) mediated 38 % and 50 % of the effect between PFAS burden score and eGFR, respectively.

Conclusion: Our proof-of-concept analysis provides the first evidence that reduced short-chain fatty acid-producing bacteria and anti-inflammatory metabolites may link PFAS exposure with impaired kidney function. This study raises the possibility of future targeted interventions that can alter gut microbiome or circulating metabolite profiles to prevent PFAS induced kidney damage.

Keywords: Gut microbiome; Kidney damage; Mediation analysis; Metabolism; PFAS exposure.

Figure 1.

Coefficient plots for results of multiple linear regression between all baseline PFAS concentrations and PFAS burden score and biomarkers of kidney function at follow-up, including eGFR, creatinine concentration and Cystatin C. All analyses performed with log-transformed and scaled variables outcome variables and scaled exposure variables. Beta coefficients interpreted as a % change in the biomarker for each 1 standard deviation (SD) increase in PFAS concentration or PFAS burden score (PFAS standard deviations shown in Table 2). All analyses adjusted for age, sex, race/ethnicity, parental education and baseline creatinine.

Figure 2.

Results of the JIVE multi-omics dimensionality reduction. A) JIVE variance explained by each component. Joint components explained by the 5 joint factors, individual components each explained by 5 individual components. B) Heatmap of all components with factor loadings for each feature and taxa.

Figure 2.

Results of the JIVE multi-omics dimensionality reduction. A) JIVE variance explained by each component. Joint components explained by the 5 joint factors, individual components each explained by 5 individual components. B) Heatmap of all components with factor loadings for each feature and taxa.

Figure 3.

Results from causal mediation analysis of HIMA selected factor components. Figure shows the JIVE joint and metabolite components that were selected by HIMA as significantly mediating the relationship between PFAS burden score and eGFR. The bar charts within the joint and metabolite components represent the features that characterize each protect component. Causal mediation analysis was performed on the two HIMA-selected JIVE components, and the effect estimates are shown on the figure. Increasing the PFAS burden score by 1 SD was associated with a 20% [−3.2%, 39%] decrease in the protective joint component; while increasing the protective joint component by 1 SD was associated with a 2.6% [0.5%, 4.8%] increase in eGFR. Increasing the PFAS burden score by 1 SD was associated with a 23% [2.5%, 39%] decrease in the protective metabolite component; while increasing the protective metabolite component by 1 SD was associated with a 3.7% [1.4%, 6.1%] increase in eGFR.