Effects of per- and poly-fluorinated alkyl substances on pancreatic and endocrine differentiation of human pluripotent stem cells

Abstract

Perfluorooctanoic acid (PFOA) and perfluorooctanesulfonic acid (PFOS) are typical per- and poly-fluorinated alkyl substances (PFASs) that epidemiological studies have already associated with diabetes. However, insufficient data on their toxicity have been reported to explain any mechanism of action, which could justify such an association. Meanwhile, short-chain PFASs designed to substitute PFOA and PFOS, have already raised increasing concerns for their biosafety. Here, we evaluated whether common PFASs affected pancreatic and endocrine cell development using a human pluripotent stem cell pancreatic induction model and human pancreatic progenitor cell (hPP) endocrine induction model. The short-chain PFASs, pentafluorobenzoic acid, perfluorohexanoic acid, perfluorobutanesulfonic acid, and perfluorohexanesulfonic acid, homologous to PFOA or PFOS, did not significantly disrupt hPP generation, unlike PFOA and PFOS, based on pancreatic and duodenal homeobox 1 (PDX1) expression. However, SRY box 9 (SOX9) expression was suppressed in PDX1+ hPPs. All six PFASs did not disrupt SOX9 expression or hPP proliferation. However, endocrine differentiation of hPPs was affected, as indicated by neurogenin-3 (NGN3) downregulation, owing to abnormal increases in SOX9 and hairy and enhancer of split-1 (HES1) expressions. Thus, hyperactivation of NOTCH signaling was repressed after hPPs committed to the endocrine lineage. In conclusion, our study demonstrates how powerful human pluripotent stem cell-based pancreatic differentiation models can be in developmental toxicity evaluations, compared to traditional toxicity assays, mostly based on live animals. Moreover, our findings suggest that PFASs may alter pancreatic development after the pancreatic domain emerges from the gut tube, and provide insights into their toxicity mechanisms.