QSAR models and their corresponding toxicity mechanism for mixture toxicity of organic pollutants

Abstract

Environmental contaminants are frequently encountered as mixtures rather than single chemicals. However, current environmental quality standards and ecological risk assessments are mainly based on single toxicities, and pay little attention to the potential hazards of mixture exposure. The study of mixture toxicity is therefore not only of theoretical value, but also of deep practical significance. This review mainly focuses on quantitative structure-activity relationship (QSAR) models for mixture toxicities of organic pollutants and the corresponding toxicity mechanisms. The organic pollutants are classified as non-polar narcosis chemicals, polar narcosis chemicals, and reactive chemicals, according to their modes of action. QSAR models of the three types of chemical are described based on their toxicity mechanisms respectively. QSAR models of non-polar narcosis chemicals are described. The QSAR models of the mixture toxicities of non-polar narcosis chemicals are the simplest. The toxicity only depends on the total hydrophobicity of the mixture and can be predicted using a uniform model. QSAR models for polar narcosis chemicals are summarized. The toxicities of narcosis chemicals are determined not only by the total hydrophobicity but also by hydrogen-bonding effects. QSAR models for reactive chemicals are described in detail. There is no uniform model, because the toxicity mechanisms of the reactive chemicals are complicated. However, three aspects of the latest research on quantitative prediction methods for toxicities of mixtures containing reactive chemicals are summarized: the factors affecting multicomponent mixtures containing reactive chemicals; the interactions among the single chemicals in the mixture; and the interactions between the chemicals and the target protein. It is pointed out that identifying the molecular biological mechanisms of chemicals using advanced experimental methods (such as genomics and metabolomics) and formulating a common prediction model based on these mechanisms should be the direction of future developments.