Chemical hazard characterisation of PFAS : current approaches in dose-response assessments and the lipidomic disruption in human cell models

Per- and polyfluoroalkyl substances (PFAS) are a group of synthetic chemicals, abundantly produced and with a wide range of applications. The partial or full fluorination of alkyl substances with varying functional groups, leads to physicochemical properties that are mainly characterised by persistence and amphiphilicity. While these properties are highly valuable for industrial- and consumer products, they pose major challenges in environmental pollution and health implications. Especially influences on the lipid metabolism have been frequently connected to PFAS exposures, yet lack deeper mechanistic understanding, and were, hence, the main investigated experimental subject of this thesis.

Regulatory efforts for PFAS have been made in the past and are currently underway. A harmonised approach for the assessment of human health effects, following PFAS exposure is, however, still needed. Dose-response assessment, through benchmark dose (BMD) modelling is the state of the science for hazard characterisation for risk assessment purposes. Specifically, the selection of a critical effect size (CES) - a threshold for adversity - of the assessed effect (response), is important in this context.

This thesis set out to tackle the outlined challenges in toxicological assessments of PFAS in three main approaches.

1) Testing contrasting ways to select the CESs in different BMD modelling approaches and assess the outcome metrics for select PFAS case-studies (Study I).

2) Utilise human cell models to gain insights into PFAS-effects on the human lipid metabolism. This was done by investigating PFAS (single and mixture) exposures and PPARa receptor activation, lipid metabolism-related gene expression and lipid accumulation in Study III. Further, PFAS (single) exposure was investigated with regards to affecting hepatocyte lipidomic profiles in Study II.

3) Assessing the meaningfulness of comparative relative potency factors (RPFs) for PFAS, through characterisation of RPFs, based on the endpoints in Study III.

The findings of Study I reveal that the CES choice alters the results of a dose- response assessment significantly for both, frequentist and Bayesian BMD modelling. Further, it became apparent that for the selected case-studies, the Bayesian BMD modelling - paired with flexible, effect-specific CES selection - led to more stable and biologically relevant results, supporting their use in regulatory decision-making.

Study II unveiled wide-range influences of PFAS exposures on the intracellular (hepatocyte) lipid profile. Legacy PFAS (e.g., PFOS and PFOA), as well as substitute PFAS (e.g., HFPO-DA and ADONA), were shown to have lipid profile-altering properties, with PFOS and PFOA having displayed the largest effects.

The cell-based assays in Study III confirmed PFAS influences in multiple mechanistic steps of the lipid metabolism. This further underscores the body of evidence for the investigated PFAS and PFAS mixtures to be involved in alterations of important lipid metabolic pathways with likely relevance to cardiovascular diseases and other metabolic conditions. With regards to their relative potencies, PFAS appeared to be endpoint-specific, with no unambiguous pattern of potency.

This thesis offers an assessment of the BMD methodology in chemical hazard characterisation within the context of assessing the risks from PFAS. Further, multi-endpoint assessment and exposures to single PFAS and PFAS mixtures of human relevance, highlighted the importance of the lipid metabolism as a major target system for PFAS toxicity. The use of RPFs to compare PFAS effects in an endpoint-specific manner, however, needs to be further investigated and a universal approach across endpoints appears to be very challenging.

List of scientific papers

I. BRUNKEN, L., Vieira Silva, A., & Öberg, M. (2025). Selection of the critical effect size alters hazard characterization - a retrospective analysis of key studies used for risk assessments of PFAS. Frontiers in Toxicology. 7. https://doi.org/10.3389/ftox.2025.1525089

II. Kashobwe, L.#, Sadrabadi, F.#, BRUNKEN, L.#, Coelho, A. C. M. F., Sandanger, T. M., Braeuning, A., Buhrke, T., Öberg, M., Hamers, T., & Leonards, P. E. G. (2024). Legacy and alternative per- and polyfluoroalkyl substances (PFAS) alter the lipid profile of HepaRG cells. Toxicology. 506, 153862. # Equally contributed. https://doi.org/10.1016/j.tox.2024.153862

III. BRUNKEN, L., Vieira Silva, A., & Öberg, M. Relative Potency of PFAS in Human Cell Models: Linking PPARa Activation, Gene Regulation, and Lipid Accumulation. [Manuscript]