Computational analysis for hepatic safety signals of constituents present in botanical extracts widely used by women in the United States for treatment of menopausal symptoms

Black cohosh, red clover, hops, and chasteberry are botanicals commonly used to alleviate menopausal symptoms in the US, and are examined in this study as part of a FDA Office of Women's Health research collaboration to expand knowledge on the safety of these botanical products. Computational approaches using classic (quantitative) structure-activity relationships ((Q)SAR), probabilistic reasoning, machine learning methods, and human expert rule-based systems were employed to deliver human hepatobiliary adverse effect predictions. The objective is to profile and analyze constituents that are alerting for the human hepatobiliary adverse effects. Computational analysis of positively predicted constituents showed that common structural features contributing to the hepatobiliary adverse effect predictions contain phenolic, flavone, isoflavone, glucoside conjugated flavone and isoflavone, and 4-hydroxyacetophenone structures. Specifically, protocatechuic acid from black cohosh, benzofuran and 4-vinylphenol from chasteberry, and xanthohumol I from hops were botanical constituents predicted positive for liver toxicity endpoints and were also confirmed with literature findings. However, comparison between the estimated human exposure to these botanical constituents and the LOAEL and NOAEL in published animal liver toxicology studies for these constituents demonstrated varying margins of safety. This study will serve as regulatory decision support information for regulators at the FDA to help with the process of prioritizing chemicals for testing.