Oral Presentation SETAC Asia-Pacific Virtual Conference 2022

Ecotoxicity characterization of pharmaceuticals in relation to physiological responses of algae for sufficiently conservative environmental risk assessment using in silico tools (#72)

Yoshikazu NINOMIYA 1 , Haruna Watanabe 1 2 , Takahiro Yamagishi 1 2 , Hiroshi Yamamoto 1 2
  1. The University of Tokyo, Tsukuba, Ibaraki, JAPAN, Japan
  2. National Institute for Environmental Studies, Tsukuba, Ibaraki, JAPAN, 日本

While current major in silico tools for ecotoxicity prediction such as KAshinhou Tool for Ecotoxicity (KATE) (https://kate.nies.go.jp/index-e.html) of Japan Ministry of the Environment and ECOSAR of USEPA are developed for industrial chemicals, it would be preferable to further incorporate these to improve the efficiency of environmental risk assessment of pharmaceuticals, which are considered as emerging contaminants of concern in the aquatic environment. The biggest challenge in the application of these in silico tools to pharmaceuticals is that the models are mostly trained by industrial chemicals and only a few pesticides. It may be necessary to create an algorithm to identify those pharmaceuticals not suitable for sufficiently conservative risk assessment by preventing the severe underestimation of their ecotoxicity. In this study, we predicted the toxicity values of algal growth inhibition tests (72 h-EC50 or NOEC) for approximately 300 pharmaceuticals using the latest version of KATE (KATE2020 ver. 3.0), and classified based on their (sub)structure(s), mechanisms of action, target molecules, and physicochemical properties, especially for those with high algal toxicity. As a result, less than 50% of the selected pharmaceuticals were successfully predicted their toxicity to algae using KATE while Predicted/Measured ratios for about 40% of the predictable pharmaceuticals ranged from 0.1 to 10. 30 highly underestimated pharmaceuticals with large difference from baseline toxicity (based on the QSAR model for narcotic group or neutral organics) were selected and further analyzed the reasons of the discrepancy. The pharmaceuticals that could be unsuitable for QSAR can be characterized by linking their characteristics and potential effects on the physiological responses of algae. We also found similarities between the substructures of pharmaceuticals and those found in the Herbicide Resistance Action Committee (HRAC) and Fungicide Resistance Action Committee (FRAC) classifications. With careful use, current in silico tools could be one of options for environmental risk assessment.