PFOS disrupts the development of early germline and surrounding embryonic soma at environmentally relevant exposure levels

STUDY QUESTION Can an embryoid body model of germ cell specification predict the impact of per- and polyfluoroalkyl substances (PFAS) on early embryonic and germline development at exposure levels relevant to human populations? SUMMARY ANSWER Perfluorooctanesulfonic acid (PFOS) impairs early germline specification and surrounding embryonic somatic development in the embryoid body system at exposure levels detected in the serum of pregnant women. WHAT IS KNOWN ALREADY PFAS are highly stable and persistent synthetic chemicals used in industrial processes and consumer products widely detected in humans. Although mounting evidence indicates that chronic PFAS exposure and bioaccumulation can adversely affect human health, their impact on early embryonic and germline development remains largely unknown. STUDY DESIGN, SIZE, DURATION We used a stem cell-based human embryoid body model for early germ cell specification to assess the effects of commonly environmentally found PFAS, perfluorooctanesulfonic acid (PFOS), perfluorooctanoic acid (PFOA), and perfluorohexanesulfonic acid (PFHxS), on human germ cell specification and early embryonic somatic development. PFAS exposure levels were chosen to reflect serum concentrations measured in pregnant women in Sweden, with the 1× nominal concentration defined as 7 nM for PFOS and 4 nM for both PFOA and PFHxS. For each compound control, 1×, 5×, 10×, 25×, or 100× exposures were compared. Four experimental replicates were carried out for each tested compound. PARTICIPANTS/MATERIALS, SETTING, METHODS Human induced pluripotent stem cells (hiPSCs) were exposed to PFOS, PFOA, and PFHxS to assess their effects on cell growth, embryoid body development, and germ cell specification. hiPSC expansion and embryoid body development were monitored by phase-contrast microscopy. Following differentiation, germ cells were quantified by fluorescence-activated cell sorting (FACS) after embryoid body dissociation and staining for the early human germ cell surface markers tissue-nonspecific alkaline phosphatase (TNAP) and podoplanin (PDPN). Bulk RNA sequencing was performed to evaluate the effects of PFAS exposure on the transcriptomic landscape of the in vitro-derived germ cells. MAIN RESULTS AND THE ROLE OF CHANCE When exposing the embryoid bodies to PFAS concentrations resembling the levels measured in the bloodstream of pregnant women, we find that PFOS impaired embryoid body development in a concentration-dependent manner, with effects already observed at the levels detected in that population group. On the other hand, PFOA caused a slight, non-significant downward trend in embryoid body size, while PFHxS induced weaker and non-monotonic effects. Upon PFOS exposure, germ cells were still specified but exhibited marked transcriptomic alterations, characterized by suppression of developmental and signalling pathways and enrichment of RNA-processing and ribosome biogenesis. In contrast, no significantly differentially expressed genes were detected in their precursors following PFOS exposure. LARGE SCALE DATA N/A LIMITATIONS, REASONS FOR CAUTION Although the embryoid body model closely captures the posterior region of the gastrulating post-implantation embryo, it does not fully recapitulate the entire embryo and its implantation environment. WIDER IMPLICATIONS OF THE FINDINGS Collectively, our results identify PFOS as a potent disruptor of embryonic and germline development and establish the embryoid body model for early germ cell specification as a robust and sensitive candidate new approach methodology (NAM) for evaluating developmental and reproductive toxicity (DART) at relevant exposure levels for human populations. FUNDING L.N.A. acknowledges support from the Marie Skłodowska-Curie Actions Individual Fellowship (101278886: GERMFIT) from the European Commission. T.M. acknowledges support from Sällskapet Barnavård (2026-012). K.J. acknowledges support from the Foundation for Pediatric Research, the Finnish Cancer Society, the Swedish Childhood Cancer Foundation (KP2020-0012), and the Birgitta and Carl-Axel Rydbeck’s Research Grant for Paediatric Research (2020-00335, 2021-00079, and 2023-00380). Y.Z. acknowledges support from National Natural Science Foundation of China (Youth Program; grant number: 82505173), and the National Academic Infrastructure for Supercomputing in Sweden (NAISS), partially funded by the Swedish Research Council through grant agreement no. 2022-06725. P.D. acknowledges support from Swedish Research Council FORMAS (2024-01415) and Karolinska Institutet consolidator grant. V.M.L. acknowledges support from the ERC Consolidator Grant 3DMASH (101170408), the European Union's Horizon Europe program NEMESIS (101137405), the Swedish Research Council (2021-02801, 2023-03015 and 2024-03401), the Novo Nordisk Foundation (NNF23OC0085944 and NNF23OC0084420) and the Robert Bosch Foundation, Stuttgart. J.P.A.-L was supported by a Starting Grant in Medicine and Health (2022-01467) from the Swedish Research Council and Birgitta and Carl-Axel Rydbeck Research Grant for Paediatric Research (2024-00208). DISCLOSURES V.M.L. is CEO and shareholder of HepaPredict AB, as well as co-founder and shareholder of Shanghai Hepo Biotechnology Ltd. All other authors declare no conflicts of interest.