The human chemical blood exposome reflects a lifetime of exposure to environmental chemicals from different sources, like water, food, air, and consumer products. Many of these compounds are persistent organic pollutants (POPs), including perfluoroalkyl substances (PFAS), polychlorinated biphenyls (PCBs), organochlorine pesticides (OCPs) and polybrominated diphenyl ethers (PBDEs). However, their blood concentrations and relative profiles vary between individuals. Current risk assessments, typically based on studies of single chemicals, do not reflect the actual exposure to complex mixtures and may underestimate the health impacts of environmental contaminants. It is therefore important to study the effects of real-world POP mixtures, particularly for sensitive toxicological endpoints like hormone signaling, which regulates vital processes such as growth, reproduction, and metabolism.

This thesis aims to bridge exposomics and in vitro toxicology through a novel proof-of-principle approach. Mixtures of POPs detected in the blood of different individuals were reconstructed using non-contact acoustic liquid dispensing. These mixtures were tested using optimized in vitro OECD assays, enabling medium- to high-throughput screening to assess effects on cell viability, testosterone and estradiol synthesis, and estrogen receptor activity for insights into endocrine disruptive potential.

The findings demonstrate that the reconstructed personalized mixtures from unique individuals induced various effects, including decreased cell viability and endocrine disruption, at concentrations found in human blood. Notably, these effects were not simply dependent on the total concentration or number of POPs in the mixture. Furthermore, population-based mixtures did not capture the diversity of effects observed in the reconstructed mixtures, underscoring limitations in generalized mixture testing and risk assessments. Testing personalized mixtures, divided into sub-mixtures by chemical class, revealed effects on testosterone synthesis that could explain the bioactivity of some but not all whole mixtures. The results also revealed effects from sub-mixtures not apparent in the whole mixture tests, highlighting the complexity of mixture toxicology, which warrant further studies into underlying mechanisms.

Overall, the results in thesis demonstrate the importance of personalized toxicology in assessing the effects of real-world chemical mixtures. The established approach is adaptable to a range of in vitro models and techniques for studying various endpoints and chemicals. The thesis underscores the need to consider population variability and interactive effects in chemical mixtures within toxicology studies. By supporting the implementation of specific and generic mixture allocation factors, this novel mixture testing strategy can improve risk assessments, protect sensitive subpopulations, and promote comprehensive public health measures.