Multiple studies demonstrate mixture effects arising from the interactive toxicity of environmental chemicals in human blood, but identifying the main toxic drivers remains challenging. In a recent proof-of-principle in vitro study, we showed that personalized mixtures (PMs), reconstructed from 24 persistent organic pollutant (POPs) concentrations measured in individual blood samples from Swedish adults, induced unique interindividual effects on H295R cell viability and steroidogenesis. Here, we followed up by testing submixtures of four PMs (PM#3, PM#4, PC1-OC-Mix, and PC2-PFAS-Mix), separated by the chemical classes perfluoroalkyl substances (PFASs), organochlorine pesticides (OCPs), polychlorinated biphenyls (PCBs), and polybrominated diphenyl ethers (PBDEs). Submixtures of PFAS and OCPs induced significant effects on testosterone synthesis at low (1×) and medium (10×) concentrations, consistent with effects observed in the corresponding whole PMs, and were therefore likely the primary drivers of the whole-mixture effects on testosterone. Notably, some submixtures altered estradiol and testosterone levels in ways not observed in full PMs, suggesting antagonistic interactions across chemical classes when combined. Potential antagonistic interaction in more complex mixtures, independent of concentration, was also observed within OCP submixtures, as only the less complex OCP mixtures lacking DDE or DDT induced testosterone synthesis. For additional mechanistic insight, we expanded the H295R assay to include oxidative stress analyses, which revealed no effects from the PMs. RT-qPCR analysis showed downregulation of CYP11A1 after exposure to PM#3 and PM#4 at high concentrations (100×), suggesting a feedback mechanism contributing to suppressed testosterone synthesis.

Toxicological studies of single chemicals overlook the real-world complexity of human exposure, where multiple compounds may interact to disrupt biological processes such as endocrine signaling. Moreover, the chemical exposome, the sum of an individual's chemical burden, varies markedly between people, yet its biological implications remain unclear. To address this gap, we reconstructed individualized human chemical exposomes to assess their effects on estrogen receptor (ER) activity. Sixteen exposomes comprising 24 persistent organic pollutant (POP) were derived from blood profiles of participants in the Swedish Västerbotten Intervention Programme. Using automated, non-contact acoustic liquid dispensing, we reconstructed 14 personalized mixtures (PMs) reflecting individual blood compositions and two formulated mixtures (FM) representing the cohort's median and maximum population exposure levels. ER activity was assessed in VM7Luc4E2 cells using a high-throughput 384-well adaption of the OECD No. 455 assay at 1×, 10× and 100× blood concentrations. While most individual POPs showed no or weak ER activity, three mixtures induced ER agonism. The PM-High, corresponding to the individual with the highest total POP levels, and the FM-Median activated the ER at 100×, while the FM-Maximum induced activation at 10× and 100×. Removing β-HCH and trans-nonachlor from the active mixtures abolished or reduced ER activity. Co-treatment with physiological estradiol levels increased ER responses in six mixtures, PM#1 (1× and 10×), PM#4 (100×), PM#8 (10×), PM#9 (100×), PM#10 (1×) and the FM-Median (1×), indicating potentiation of endogenous hormonal signaling. Overall, this study reveals endocrine activity in real-world POP mixtures and advances high-throughput screening as a scalable approach for individualized exposome-based health risk evaluation.