Monika Nendza1, Andrea Wenzel. 1. Analytisches Laboratorium, Luhnstedt, Germany. AL-Luhnstedt@t-online.de
Abstract
BACKGROUND AND SCOPE: Predictive toxicology, particularly quantitative structure-activity relationships (QSARs), require classification of chemicals by mode of action (MOA). MOA is, however, not a constant property of a compound but it varies between species and may change with concentration and duration of exposure. A battery of MOA-specific in-vitro and low-complexity assays, featuring biomolecular targets for major classes of environmental pollutants, provides characteristic responses for (1.) classification of chemicals by MOA, (2.) identification of (eco)toxicity profiles of chemicals, (3.) identification of chemicals with specific MOAs, (4.) indication of most sensitive species, (5.) identification of chemicals that are outliers in QSARs and (6.) selection of appropriate QSARs for predictions. METHODS: Chemicals covering nine distinct modes of toxic action (non-polar non-specific toxicants (n=14), polar non-specific toxicants (n=18), uncouplers of oxidative phosphorylation (n=25), inhibitors of photosynthesis (n=15), inhibitors of acetylcholinesterase (n=14), inhibitors of respiration (n=3), thiol-alkylating agents (n=9), reactives (irritants) (n=8), estrogen receptor agonists (n=9)) were tested for cytotoxicity in the neutralred assay, oxygen consumption in isolated mitochondria, oxygen production in algae, inhibition of AChE, reaction with GSH and activity in the yeast estrogen receptor assay. Data on in-vivo aquatic toxicity (LC50, EC50) towards fish, daphnids, algae and bacteria were collected from the literature for reasons of comparison and reference scaling. RESULTS AND DISCUSSION: In the MOA-specific in-vitro test battery, most test chemicals are specifically active at low concentrations, though multiple effects do occur. Graphical and statistical evaluation of the individual classes versus MOA 1 (non-polar non-specific toxicants) identifies interactions related to predominant MOA. Discriminant analyses (DA) on subsets of the data revealed correct classifications between 70% (in-vivo data) and >90% (in-vitro data). Functional similarity of chemical substances is defined in terms of their (eco)toxicity profiles. Within each MOA class, the compounds share some properties related to the rate-limiting interactions, e.g., steric fit to the target site and/or reactivity with target biomolecules, revealing a specific pattern (fingerprint) of characteristic effects. CONCLUSION: The successful discrimination of toxicant classes by MOA is based on comprehensive characterization of test chemicals' properties related to interactions with target sites. The suite of aquatic in-vivo tests using fish, daphnids, algae and bacteria covers most acute effects, whilst long-term (latent) impacts are generally neglected. With the MOA-specific in-vitro test battery such distinctions are futile, because it focuses on isolated targets, i.e. it indicates the possible targets of a chemical regardless of the timescale of effects. The data analysis indicates that the in-vitro battery covers most effects in vivo and moreover provides additional aspects of the compounds' MOA. RECOMMENDATION AND PERSPECTIVE: Translating in-vitro effects to in-vivo toxicity requires combining physiological and chemical knowledge about underlying processes. Comparison of the specific in-vitro effects of a compound with the respective sensitivities of aquatic organisms indicates particularly sensitive species. Classifications of toxicants by MOA based on physicochemical descriptors provides insight to interactions and directs to mechanistic QSARs.
BACKGROUND AND SCOPE: Predictive toxicology, particularly quantitative structure-activity relationships (QSARs), require classification of chemicals by mode of action (MOA). MOA is, however, not a constant property of a compound but it varies between species and may change with concentration and duration of exposure. A battery of MOA-specific in-vitro and low-complexity assays, featuring biomolecular targets for major classes of environmental pollutants, provides characteristic responses for (1.) classification of chemicals by MOA, (2.) identification of (eco)toxicity profiles of chemicals, (3.) identification of chemicals with specific MOAs, (4.) indication of most sensitive species, (5.) identification of chemicals that are outliers in QSARs and (6.) selection of appropriate QSARs for predictions. METHODS: Chemicals covering nine distinct modes of toxic action (non-polar non-specific toxicants (n=14), polar non-specific toxicants (n=18), uncouplers of oxidative phosphorylation (n=25), inhibitors of photosynthesis (n=15), inhibitors of acetylcholinesterase (n=14), inhibitors of respiration (n=3), thiol-alkylating agents (n=9), reactives (irritants) (n=8), estrogen receptor agonists (n=9)) were tested for cytotoxicity in the neutralred assay, oxygen consumption in isolated mitochondria, oxygen production in algae, inhibition of AChE, reaction with GSH and activity in the yeast estrogen receptor assay. Data on in-vivo aquatic toxicity (LC50, EC50) towards fish, daphnids, algae and bacteria were collected from the literature for reasons of comparison and reference scaling. RESULTS AND DISCUSSION: In the MOA-specific in-vitro test battery, most test chemicals are specifically active at low concentrations, though multiple effects do occur. Graphical and statistical evaluation of the individual classes versus MOA 1 (non-polar non-specific toxicants) identifies interactions related to predominant MOA. Discriminant analyses (DA) on subsets of the data revealed correct classifications between 70% (in-vivo data) and >90% (in-vitro data). Functional similarity of chemical substances is defined in terms of their (eco)toxicity profiles. Within each MOA class, the compounds share some properties related to the rate-limiting interactions, e.g., steric fit to the target site and/or reactivity with target biomolecules, revealing a specific pattern (fingerprint) of characteristic effects. CONCLUSION: The successful discrimination of toxicant classes by MOA is based on comprehensive characterization of test chemicals' properties related to interactions with target sites. The suite of aquatic in-vivo tests using fish, daphnids, algae and bacteria covers most acute effects, whilst long-term (latent) impacts are generally neglected. With the MOA-specific in-vitro test battery such distinctions are futile, because it focuses on isolated targets, i.e. it indicates the possible targets of a chemical regardless of the timescale of effects. The data analysis indicates that the in-vitro battery covers most effects in vivo and moreover provides additional aspects of the compounds' MOA. RECOMMENDATION AND PERSPECTIVE: Translating in-vitro effects to in-vivo toxicity requires combining physiological and chemical knowledge about underlying processes. Comparison of the specific in-vitro effects of a compound with the respective sensitivities of aquatic organisms indicates particularly sensitive species. Classifications of toxicants by MOA based on physicochemical descriptors provides insight to interactions and directs to mechanistic QSARs.
Authors: Mariann Rand-Weaver; Luigi Margiotta-Casaluci; Alpa Patel; Grace H Panter; Stewart F Owen; John P Sumpter Journal: Environ Sci Technol Date: 2013-09-30 Impact factor: 9.028