Chapter 31 - Deploying QSAR to discriminate excess toxicity and identify the toxic mode of action of organic pollutants to aquatic organisms

Abstract

In aquatic toxicology, determining the toxic mode of action (MOA) for diverse chemicals is the basis of ecological risk assessment and plays a driving role in chemical regulations. Toxic MOA identification can contribute to quantitative structure–activity relationship (QSAR) model development and be leveraged to evaluate the toxic effects of organic pollutants. Beyond this, MOA identification improves understanding of toxic mechanisms and provides scientific evidence for ecological risk assessment. To identify the MOA of organic compounds, specifically reactive compounds, the concept of excess toxicity has been employed to discriminate the elevated toxic responses- from baseline narcotic effects. Toxicity above that associated with narcosis is defined as “excess toxicity.” Interspecies variation in sensitivity to toxicants can be substantial, with the most sensitive species being of utmost concern for risk management. These differences in sensitivity between species may result from a number of factors including physiology, exposure time for testing, parameters (growth, reproduction, survival) used, and life stage. The acute toxicity of class-based organic pollutants to fish, Daphnia magna (DM), Tetrahymena pyriformis (TP), and Vibrio fischeri (VF), has been investigated. The results showed that baseline compounds may share the same MOAs between species. Moreover, the toxicity sensitivity of baseline compounds compared across DM, fish, and VF is similar. Therefore, the threshold log(TR)=1, which is based on the distribution of toxicity data to fish, can be used to discriminate reactive compounds from baseline narcotics for DM and VF. On the other hand, the toxicity sensitivity in TP is very different from the other three species for baseline compounds. Importantly, this suggests that the log(TR)=1 obtained from fish toxicity is not an ideal threshold of excess toxicity for TP. The threshold log(TR)=0.8 was determined by investigating the relationship between the species sensitivity, the absolute averaged residuals between the predicted baseline toxicity, and experimental toxicity. The toxic MOA of baseline and less inert and reactive compounds on fish, DM, TP, and VF were identified based on the species-specific log(TR)=1 and log(TR)=0.8. The results indicated that less inert compounds exhibit higher toxicity on TP and VF than the high-level species fish and DM. This striking finding may be attributed to the fact that hydrophilic compounds more easily pass through the cell membrane than the skin of organisms and have higher bioconcentrations in TP and VF, leading to higher toxic effects. Most compounds exhibit excess toxicity in multiple species. However, few compounds exhibit excess toxicity to only one species. Bioconcentration factors (BCFs) calculated by the discrimination of excess toxicity from baseline effect have been investigated. Based on the analysis of 951 acute toxicity data in fish (LC50) and 1088 BCFs, some compounds were identified as reactive compounds that exhibit excess toxicity. BCF can significantly affect the TR value. The excess toxicity of reactive chemicals from baseline should be determined by the TR of internal effect concentrations but not external effect concentrations. The foundation of the discrimination of excess toxicity from baseline effect is the linear relationship between log(BCF) and hydrophobicity expressed as log KOW. However, log(BCF) is not linearly related with log KOW for all the observed compounds. The BCFs with log KOW >7 or <0 are either overestimated or underestimated by the linear baseline BCF model. Parallel lines are observed from calculated log critical body residual (CBR) values for baseline and less inert compounds. The log(BCF) values overestimated or underestimated by log KOW from the baseline BCF model can result in misclassification on baseline, less inert, and reactive compounds. DM toxicity data determined at three pH values for 61 ionic compounds were used to investigate the effect of pH on the discrimination of excess toxicity. The resulting apparent toxicities are significantly less than the baseline levels. Analysis of the effect of pH on BCF showed that the log(BCF) values are significantly overestimated for strongly ionizable compounds, leading to apparent toxicities substantially less than the baseline toxicities, and the TRs are substantially less than zero. A theoretical equation between the apparent toxicities and pH values has been developed based on CBRs. The apparent toxicities are linearly related to fractions of the unionized form but not to the pH values. The determined apparent toxicities fit well with the toxicities predicted by the equation. The toxicities in the unionized form calculated from the equation are close to or greater than the baseline level for almost all the strongly ionizable compounds, which are very different from the apparent toxicities. The studied ionizable compounds can be classified as baseline, less inert, and reactive compounds in toxicity. Due to ionization in water rather than poor target reactivity, some ionizable compounds do not exhibit excess toxicity at a specific pH value. This chapter reviews work on MOA identification and QSAR modeling of excess toxicity of organic chemicals on aquatic organisms.