Transforming Chemical Safety Assessment Using Transcriptomics to Define Toxicological Potency and Common Mechanisms of BPA Replacements

Abstract

Humans are exposed to chemicals that disrupt the function of endocrine tissues through the use of everyday consumer products and other environmental exposures. Bisphenol A (BPA) is used to manufacture polycarbonate plastics and epoxies and is a developing agent in thermal paper. BPA is an endocrine disruptor that acts on several nuclear hormone receptors, including the estrogen receptor. Most people have measurable levels of BPA in their urine, and exposure is associated with several negative health outcomes. In the late 2000s, BPA was regulated in Canada and other jurisdictions worldwide. This has led to the proliferation of chemical alternatives to BPA, sometimes in products labelled 'BPA-free'. Thus, humans are exposed to BPA alternatives, often as mixtures. Unfortunately, most BPA alternatives are data-poor, limiting regulatory toxicology assessment. The goal of my thesis was to use high-throughput transcriptomics to evaluate the potency of data-poor BPA alternative chemicals and characterize mechanisms of action in vitro. I produced global gene expression profiles to derive measures of biological activity and potency of 20 individual BPA alternatives and mixtures of these chemicals. Gene set enrichment analysis was used to identify perturbed biological pathways and upstream regulators. Transcriptomic biomarkers were applied to identify chemicals that activated the estrogen receptor alpha (ERα) and cellular stress response pathways. Bisphenol AF (BPAF) was the most potent chemical tested overall, activating the ERα at the lowest concentration, followed by BPA. Most BPA alternatives activated ERα and related pathways and enriched similar gene sets associated with dysregulated cell cycle control and cancer-related outcomes in vitro. Next, I used a concentration response model to test whether mixtures of BPA alternative chemicals have additive effects in breast cancer cells. Both equimolar and complex mixtures were additive, activated the ERα, and had similar gene expression profiles as their individual components. Lastly, I sought to test BPA alternative chemicals in a more physiologically relevant model. Human mammary epithelial cells (HMECs) from multiple donors were exposed to a subset of BPA alternatives chemicals. These chemicals had similar potency rankings in MCF-7 and HMEC cells, with BPAF exhibiting the highest potency. Some of the BPA alternatives activated cancer-related pathways, while others inhibited the same pathways in HMECs. Overall, my thesis demonstrates that many BPA alternatives have comparable toxicological potencies to BPA, behave additively in mixtures, and have similar mechanisms of action in vitro. Importantly, this research identified BPAF as toxicologically potent BPA alternative chemical. Finally, these data demonstrate a pragmatic approach to the implementation of transcriptomics in regulatory toxicology and decision making.