Mode Of Toxic Action Research Articles

Development and validation of theoretical linear solvation energy relationship models for toxicity prediction to fathead minnow (pimephales promelas)

The acute toxicity predictive models are vitally important for the toxicological information used in the ecological risk assessments. In this study, we used Verhaar classification scheme to group compounds into five modes of toxic action. The quantum chemical descriptors that characterize the electron donor–acceptor property of the compounds were introduced into the theoretical linear solvation energy relationship (TLSER) models. The predictive models have relatively larger data sets, which imply that they cover a wide applicability domain (AD). All models were developed following the Organization for Economic Co-operation and Development (OECD) QSAR models development and validation guidelines. The adjusted determination coefficient (Radj2) and external explained variance (QEXT2) of the models were ranging from 0.707 to 0.903 and 0.660 to 0.858, respectively, indicating high goodness-of-fit, robustness and predictive capacity. The cavity term (McGowans volume) was the most significant descriptor in the models. Moreover, the electron donor–acceptor (E-TLSER) models are comparable to the TLSER models for the toxicity prediction to fathead minnow. Thus, the E-TLSER models developed can be used to predict acute toxicity of new compounds within the AD.

Understanding bioavailability and toxicity of sediment‐associated contaminants by combining passive sampling with in vitro bioassays in an urban river catchment

Bioavailable and bioaccessible fractions of sediment-associated contaminants are considered as better dose metrics for sediment-quality assessment than total concentrations. The authors applied exhaustive solvent extraction and nondepletive equilibrium sampling techniques to sediment samples collected along the Brisbane River in South East Queensland, Australia, which range from pristine environments to urban and industry-impacted areas. The wide range of chemicals expected prevents comprehensive chemical analysis, but a battery of cell-based bioassays sheds light on mixture effects of chemicals in relation to various modes of toxic action. Toxic effects were expressed as bioanalytical equivalent concentrations (BEQs) normalized to the organic carbon content of each sediment sample. Bioanalytical equivalent concentrations from exhaustive extraction agreed fairly well with values estimated from polydimethylsiloxane passive sampling extracts via the constant organic carbon to polydimethylsiloxane partition coefficient. Agreement was best for bioassays indicative of photosynthesis inhibition and oxidative stress response and discrepancy within a factor of 3 for the induction of the aryl hydrocarbon receptor. For nonspecific cytotoxicity, BEQ from exhaustive extraction were 1 order of magnitude higher than values from equilibrium sampling, possibly because of coextraction of bioactive natural organic matter that led to an overestimation of toxicity in the exhaustive extracts, which suggests that passive sampling is better suited in combination with bioanalytical assessment than exhaustive extraction.

Evaluation of critical body residue data for acute narcosis in aquatic organisms

The Environmental Residue Effects Database was evaluated to identify critical body residues of organic chemicals causing acute baseline neutral narcosis in aquatic organisms. Over 15 000 records for >400 chemicals were evaluated. Mean molar critical body residues in the final data set of 161 records for 29 chemicals were within published ranges but varied within and among chemicals and species (~3 orders of magnitude), and lipid normalization did not consistently decrease variability. All 29 chemicals can act as baseline neutral narcotics, but chemicals and/or their metabolites may also act by nonnarcotic modes of action. Specifically, nonnarcotic toxicity of polycyclic aromatic hydrocarbons and/or their biotransformation derivatives may be a significant source of variability. Complete testing of the narcosis-critical body residue hypothesis was confounded by data gaps for key toxicity modifying factors such as metabolite formation/toxicity, lipid content/composition, other modes of toxic action, and lack of steady-state status. Such problems impede determination of the precise, accurate toxicity estimates necessary for sound toxicological comparisons. Thus, neither the data nor the chemicals in the final data set should be considered definitive. Changes to testing designs and methods are necessary to improve data collection and critical body residue interpretation for hazard and risk assessment. Each of the toxicity metrics discussed-wet weight and lipid weight critical body residues, volume fraction in organism lipid, and chemical activity-has advantages, but all are subject to the same toxicity modifying factors.

Prediction of Aquatic Toxicity Mode of Action Using Linear Discriminant and Random Forest Models

The ability to determine the mode of action (MOA) for a diverse group of chemicals is a critical part of ecological risk assessment and chemical regulation. However, existing MOA assignment approaches in ecotoxicology have been limited to a relatively few MOAs, have high uncertainty, or rely on professional judgment. In this study, machine based learning algorithms (linear discriminant analysis and random forest) were used to develop models for assigning aquatic toxicity MOA. These methods were selected since they have been shown to be able to correlate diverse data sets and provide an indication of the most important descriptors. A data set of MOA assignments for 924 chemicals was developed using a combination of high confidence assignments, international consensus classifications, ASTER (ASessment Tools for the Evaluation of Risk) predictions, and weight of evidence professional judgment based an assessment of structure and literature information. The overall data set was randomly divided into a training set (75%) and a validation set (25%) and then used to develop linear discriminant analysis (LDA) and random forest (RF) MOA assignment models. The LDA and RF models had high internal concordance and specificity and were able to produce overall prediction accuracies ranging from 84.5 to 87.7% for the validation set. These results demonstrate that computational chemistry approaches can be used to determine the acute toxicity MOAs across a large range of structures and mechanisms.

Development of QSAR-based two-stage prediction model for estimating mixture toxicity

Conventionally, concentration addition (CA) and independent action (IA) models based on additive toxicity are often used to estimate the mixture toxicity of similarly- and dissimilarly-acting chemicals, respectively. A two-stage prediction (TSP) model has been developed as an integrated addition model that can perform the CA and IA calculations stage by stage. However, the use of the conventional TSP model is limited if the mode of toxic action (MoA) for every mixture component is not readily known. The aim of this study was to develop and evaluate a quantitative structure–activity relationship-based TSP (QSAR-TSP) model for estimating mixture toxicity in the absence of knowledge on the MoAs of the constituents. For this purpose, different clustering methods of mixture constituents using computerized analysis based on the structural similarity between chemicals were applied as a part of the predictions of mixture toxicity. The relative importance of molecular descriptors was additionally determined by Random Forest analysis. This study highlights the prediction power of the QSAR-TSP model and its potential to overcome the limitations of the conventional TSP model, and how clustering methods of mixture components that employ chemical structural information to categorize might be applied to predict mixture toxicity effectively.

Data mining toxicity effects through an ontology approach to investigate toxicity mode of action

Data mining toxicity effects through an ontology approach to investigate toxicity mode of action

Simplifying complexity: Mixture toxicity assessment in the last 20 years

Simplifying complexity: Mixture toxicity assessment in the last 20 years

Catalase has a key role in protecting cells from the genotoxic effects of monomethylarsonous acid: A highly active metabolite of arsenic

Although it is widely known that arsenic-contaminated drinking water causes many diseases, arsenic's exact mode of action (MOA) is not fully understood. Induction of oxidative stress has been proposed as an important key event in the toxic MOA of arsenic. The authors' studies are centered on identifying a reactive species involved in the genotoxicity of arsenic using a catalase (CAT) knockout mouse model that is impaired in its ability to breakdown hydrogen peroxide (H2 O2 ). The authors assessed the induction of DNA damage using the Comet assay following exposure of mouse Cat(+/) (+) and Cat(-) (/) (-) primary splenic lymphocytes to monomethylarsonous acid (MMA(III) ) to identify the potential role of H2 O2 in mediating cellular effects of this metalloid. The results showed that the Cat(-) (/) (-) lymphocytes are more susceptible to MMA(III) than the Cat(+/) (+) lymphocytes by a small (1.5-fold) but statistically significant difference. CAT activity assays demonstrated that liver tissue has approximately three times more CAT activity than lymphocytes. Therefore, Comet assays were performed on primary Cat(+/) (+) , Cat(+/) (-) , and Cat(-) (/) (-) hepatocytes to determine if the Cat(-) (/) (-) cells were more susceptible to MMA(III) than lymphocytes. The results showed that the Cat(-) (/) (-) hepatocytes exhibit higher levels of DNA strand breakage than the Cat(+/) (+) (approximately fivefold) and Cat(+/) (-) (approximately twofold) hepatocytes exposed to MMA(III) . Electron spin resonance using 5,5-dimethyl-1-pyrroline-N-oxide as the spin-trap agent detected the generation of ·OH via MMA(III) when H2 O2 was present. These experiments suggest that CAT is involved in protecting cells against the genotoxic effects of the ·OH generated by MMA(III) .

Comparative study to predict toxic modes of action of phenols from molecular structures

Quantitative structure–activity relationship models for the prediction of mode of toxic action (MOA) of 221 phenols to the ciliated protozoan Tetrahymena pyriformis using atom-based quadratic indices are reported. The phenols represent a variety of MOAs including polar narcotics, weak acid respiratory uncouplers, pro-electrophiles and soft electrophiles. Linear discriminant analysis (LDA), and four machine learning techniques (ML), namely k-nearest neighbours (k-NN), support vector machine (SVM), classification trees (CTs) and artificial neural networks (ANNs), have been used to develop several models with higher accuracies and predictive capabilities for distinguishing between four MOAs. Most of them showed global accuracy of over 90%, and false alarm rate values were below 2.9% for the training set. Cross-validation, complementary subsets and external test set were performed, with good behaviour in all cases. Our models compare favourably with other previously published models, and in general the models obtained with ML techniques show better results than those developed with linear techniques. We developed unsupervised and supervised consensus, and these results were better than our ML models, the results of rule-based approach and other ensemble models previously published. This investigation highlights the merits of ML-based techniques as an alternative to other more traditional methods for modelling MOA.

Current developments in toxicological research on arsenic.

There is a plethora of recent publications on all aspects relevant to the toxicology of arsenic (As). Over centuries exposures to arsenic continue to be a major public health problem in many countries. In particular, the occurrence of high As concentrations in groundwater of Southeast Asia receives now much attention. Therefore, arsenic is a high-priority matter for toxicological research. Key exposure to As are (traditional) medicines, combustion of As-rich coal, presence of As in groundwater, and pollution due to mining activities. As-induced cardiovascular disorders and carcinogenesis present themselves as a major research focus. The high priority of this issue is now recognized politically in a number of countries, research funds have been made available. Also experimental research on toxicokinetics and toxicodynamics and on modes of toxic action is moving very rapidly. The matter is of high regulatory concern, and effective preventive measures are required in a number of countries.

Mortality, bioaccumulation and physiological responses in juvenile freshwater mussels (Lampsilis siliquoidea) chronically exposed to copper

Several studies have indicated that the early life stages of freshwater mussels are among the most sensitive aquatic organisms to inorganic chemicals, including copper. However, little is known about the toxic mode of action and sub-lethal effects of copper exposure in this group of imperiled animals. In this study, the physiological effects of long-term copper exposure (survival, growth, copper bioaccumulation, whole-body ion content, oxygen consumption, filtration rate, ATPase activities, and biomarkers of oxidative stress) were evaluated in juvenile (6 month old) mussels (Lampsilis siliquoidea). The mussels’ recovery capacity and their ability to withstand further acute copper challenge were also evaluated in secondary experiments following the 28 day exposure by assessing survival, copper bioaccumulation and whole-body ion content. Mussels chronically exposed to 2 and 12μg Cu/L showed significantly higher mortality than those held under control conditions (mortality 20.9, 69.9 and 12.5%, respectively), indicating that juvenile L. siliquoidea is underprotected by the U.S. Environmental Protection Agency (USEPA) biotic ligand model (BLM)-derived chronic water quality criteria (WQC) (2.18μg Cu/L) and the hardness-derived USEPA WQC (12.16μg Cu/L). Soft tissue copper burden increased equally for both copper exposures, suggesting that chronic toxicity is not associated with copper bioaccumulation. Several physiological disturbances were also observed during chronic copper exposure. Most relevant was a decrease in whole-body sodium content paralleled by an inhibition of Na+ K+-ATPase activity, indicating a metal-induced ionoregulatory disturbance. Filtration and oxygen consumption rates were also affected. Redox parameters (reactive oxygen production, antioxidant capacity against peroxyl radicals, glutathione-S-transferase (GST) activity, and glutathione (GSH) concentration) did not show clear responses, but membrane damage as lipid peroxidation (LPO) was observed in both copper exposures. Mussels previously held in control conditions or pre-exposed to 2μg dissolved Cu/L were able to maintain their ionic homeostasis and did not experience mortality after the 4-d recovery period. In contrast, those previously exposed to 12μg dissolved Cu/L exhibited 50% mortality indicating that they had already reached a ‘point of no return’. Pre-exposure to copper did not influence mussel response to the copper challenge test. As observed for the chronic exposure, mortality of mussels held in the absence of copper and submitted to the challenge test was also associated with an ionoregulatory disturbance. These results indicate that ionoregulatory disruption in freshwater mussels chronically exposed to copper is the main mechanism of toxicity and that redox parameters do not appear to be useful as indicators of sub-lethal copper toxicity in these animals.

Comparison of toxicity and transcriptomic profiles in a diatom exposed to oil, dispersants, dispersed oil

Dispersants are commonly used to mitigate the impact of oil spills, however, the ecological cost associated with their use is uncertain. The toxicity of weathered oil, dispersed weathered oil, and the hydrocarbon-based dispersant Slickgone NS®, to the diatom Phaeodactylum tricornutum has been examined using standardized toxicity tests. The assumption that most toxicity occurs via narcosis was tested by measuring membrane damage in diatoms after exposure to one of the petroleum products. The mode of toxic action was determined using microarray-based gene expression profiling in diatoms after exposure to one of the petroleum products. The diatoms were found to be much more sensitive to dispersants than to the water accommodated fraction (WAF), and more sensitive to the chemically enhanced WAF (CEWAF) than to either the WAF itself or the dispersants. Exposure to dispersants and CEWAF caused membrane damage, while exposure to WAF did not. The gene expression profiles resulting from exposure to all three petroleum mixtures were highly similar, suggesting a similar mode of action for these compounds. The observed toxicity bore no relationship to PAH concentrations in the water column or to total petroleum hydrocarbon (TPH), suggesting that an undescribed component of the oil was causing toxicity. Taken together, these results suggest that the use of dispersants to clean up oil spills will dramatically increase the oil toxicity to diatoms, and may have implications for ecological processes such as the timing of blooms necessary for recruitment.

Predicting Adult Fish Acute Lethality with the Zebrafish Embryo: Relevance of Test Duration, Endpoints, Compound Properties, and Exposure Concentration Analysis

The zebrafish embryo toxicity test has been proposed as an alternative for the acute fish toxicity test, which is required by various regulations for environmental risk assessment of chemicals. We investigated the reliability of the embryo test by probing organic industrial chemicals with a wide range of physicochemical properties, toxicities, and modes of toxic action. Moreover, the relevance of using measured versus nominal (intended) exposure concentrations, inclusion of sublethal endpoints, and different exposure durations for the comparability with reported fish acute toxicity was explored. Our results confirm a very strong correlation of zebrafish embryo to fish acute toxicity. When toxicity values were calculated based on measured exposure concentrations, the slope of the type II regression line was 1 and nearly passed through the origin (1 to 1 correlation). Measured concentrations also explained several apparent outliers. Neither prolonged exposure (up to 120 h) nor consideration of sublethal effects led to a reduced number of outliers. Yet, two types of compounds were less lethal to embryos than to adult fish: a neurotoxic compound acting via sodium channels (permethrin) and a compound requiring metabolic activation (allyl alcohol).

Oxidative stress and detoxification biomarker responses in aquatic freshwater vertebrates exposed to microcystins and cyanobacterial biomass

Cyanobacterial blooms represent a serious threat to the aquatic environment. Among other effects, biochemical markers have been studied in aquatic vertebrates after exposures to toxic cyanobacteria. Some parameters such as protein phosphatases may serve as selective markers of exposure to microcystins, but under natural conditions, fish are exposed to complex mixtures, which affect the overall biomarker response. This review aims to provide a critical summary of biomarker responses in aquatic vertebrates (mostly fish) to toxic cyanobacteria with a special focus on detoxification and oxidative stress. Detoxification biomarkers such as glutathione (GSH) and glutathione-S-transferase (GST) showed very high variability with poor general trends. Often, stimulations and/or inhibitions and/or no effects at GSH or GST have been reported, even within a single study, depending on many variables, including time, dose, tissue, species, etc. Most of the oxidative stress biomarkers (e.g., superoxide dismutase, catalase, glutathione peroxidase, and glutathione reductase) provided more consistent responses, but only lipid peroxidation (LPO) seemed to fulfill the criteria needed for biomarkers, i.e., a sufficiently long half-life and systematic response. Indeed, reviewed papers demonstrated that toxic cyanobacteria systematically elevate levels of LPO, which indicates the important role of oxidative damage in cyanobacterial toxicity. In summary, the measurement of biochemical changes under laboratory conditions may provide information on the mode of toxic action. However, comparison of different studies is very difficult, and the practical use of detoxification or oxidative stress biomarkers as diagnostic tools or early warnings of cyanobacterial toxicity is questionable.

Quantifying intrinsic chemical reactivity of molecular structural features for protein binding and reactive toxicity, using the MOSES chemoinformatics system

Covalent binding of xenobiotic compounds to endogenous biomolecular sites, e.g. protein residues, leads to potentially irreversible toxic effects such as enhanced acute toxicity or skin sensitization [1]. This mechanistic knowledge provides the basis for the in silico prediction of these toxicities, as required by the EU REACH legislation and the EU Cosmetics Directive. A general toxicity prediction can be based on three consecutive steps [2]: (1.) Identification of a potential reactive protein binding mechanism via a set of molecular structural features. Those structural features can be encoded by the Chemical Subgraph Representation Markup Language (CSRML), that supports a flexible annotation of meta information, including physicochemical properties as annotations. (2.) Confirmation and quantification of (bio)chemical reactivity. The potential for a chemical to be reactive can be captured by mechanistically based QSAR models. This intrinsic reactivity is calculated rapidly by descriptors of the chemoinformatics platform Molecular Structure Encoding System (MOSES) [3]. It represents electronic, steric and resonance effects in chemical structures. The performances obtained by these reactivity models are close to time-consuming quantum chemical reactivity calculations, e.g., se = 0.44 versus 0.40 for glutathione adduct formation via Michael addition, comparing predicted values to an experimental reactivity dataset [1]. (3.) Establishing a relationship between calculated reactivity and toxicity. The predicted intrinsic reactivity values were linked to the computational prediction for different modes of toxic action, with good correlations between predicted and in vitro toxicity (up to r2=0.86). The combined use of structural information and computed reactivity could assist in the non-animal based risk assessment of chemicals for regulatory purposes and in the application of integrated testing strategies (ITS). The research has received funding from the EU FP7 COSMOS Project (grant agreement n° 266835) and financing from COLIPA.

TRENDS OF DITHIOCARBAMATE RESIDUES IN RASPBERRIES IN THE REPUBLIC OF SERBIA OVER THE PERIOD 2007/2010

Dithiocarbamates (DTC) are among the most commonly used in the world. In the Republic of Serbia almost 300 active substances are registered including five dithiocarbamate fungicides (ziram, mancozeb, metiram, propineb and tiram). For raspberry protection from fungal diseases, only mancozeb combined with metalaxyl-M (Ridomil Gold MZ 68-WG; mankozeb 640 g/kg + metalaxyl-M 40 g/kg), of all dithiocarbamates, is authorized for use in the Republic of Serbia. Dithiocarbamates present in food can be of health concern. They have at least three modes of toxic action. These include: (a) the capacity to generate carbon disulfide (CS2), (b) biotransformation into ethylenethiourea (ETU), (c) the ability to chelate physiologically important ions (e.g., copper). The Serbian government authorities adapted the maximum residue levels (MRLs) to the MRLs set by the European Union in 2010 (The Official gazette 27/2010) but dithiocarbamate levels were not set for all agricultural products. The MRLs for dithiocarbamates in raspberries (dithiocarbamates expressed as CS2, including maneb, mancozeb, metiram, propineb, thiram and ziram) in the EU and the Republic of Serbia is 0.05 mg/kg. The paper presents the method validation results and the results of dithiocarbamate monitoring study. Our study comprised the determination of the dithiocarbamate residues content in 103 raspberry samples from the market of the Republic of Serbia taken during 2007/2010. Over 20% of samples were above the MRLs (0.05 mg/kg), with the highest detected concentration of 0.8 mg/kg. The headspace GC/MS multi-residue method, based on the analysis of carbon disulphide obtained after the treatment of analytes with hydrochloric acid in the presence of tin (II)chloride was used. The averaged recovery from raspberries was spiked in the range of 0.05 - 1.25 mg/kg, is 91.35%. The limits of detection (LOD = 3 x δ baseline noise) and quantification (LOQ = 10 x δ baseline noise) were approximately 0.01 and 0.03 mg/kg, respectively. The linearity obtained without matrix was R2=0.992, and with matrix was R2=0.999. The matrices did not have interfering peaks in the time of CS2 elution.

The role of surface charge in cellular uptake and cytotoxicity of medical nanoparticles

Many types of nanoparticles (NPs) are tested for use in medical products, particularly in imaging and gene and drug delivery. For these applications, cellular uptake is usually a prerequisite and is governed in addition to size by surface characteristics such as hydrophobicity and charge. Although positive charge appears to improve the efficacy of imaging, gene transfer, and drug delivery, a higher cytotoxicity of such constructs has been reported. This review summarizes findings on the role of surface charge on cytotoxicity in general, action on specific cellular targets, modes of toxic action, cellular uptake, and intracellular localization of NPs. Effects of serum and intercell type differences are addressed. Cationic NPs cause more pronounced disruption of plasma-membrane integrity, stronger mitochondrial and lysosomal damage, and a higher number of autophagosomes than anionic NPs. In general, nonphagocytic cells ingest cationic NPs to a higher extent, but charge density and hydrophobicity are equally important; phagocytic cells preferentially take up anionic NPs. Cells do not use different uptake routes for cationic and anionic NPs, but high uptake rates are usually linked to greater biological effects. The different uptake preferences of phagocytic and nonphagocytic cells for cationic and anionic NPs may influence the efficacy and selectivity of NPs for drug delivery and imaging.

R(+)-, Rac-, and S(–)-Bupivacaine Stereostructure-Specifically Interact with Membrane Lipids at Cardiotoxically Relevant Concentrations

It remains questionable whether local anesthetics can interact with membrane lipids at clinically relevant concentrations to show the difference between enantiomers. We compared the effects of bupivacaine stereoisomers on biomimetic membranes containing cardiolipin and cholesterol. Bupivacaine interacted with the membranes at cardiotoxic 5 μM with the potency being S(-)-enantiomer < racemate < R(+)-enantiomer, which agreed with the rank order of their cardiotoxicity. Such differences became greater with decreasing drug concentrations, possibly explaining the inconsistent cardiotoxic potencies of bupivacaine stereoisomers reported previously. The interactivity with biomembranes may in part contribute to the mode of toxic action of local anesthetics.

Inhibition of Snowshoe Hare Succinate Dehydrogenase Activity as a Mechanism of Deterrence for Papyriferic Acid in Birch

The plant secondary metabolite papyriferic acid (PA) deters browsing by snowshoe hares (Lepus americanus) on the juvenile developmental stage of the Alaska paper birch (Betula neoalaskana). However, the physiological mechanism that reduces browsing remains unknown. We used pharmacological assays and molecular modeling to test the hypothesis that inhibition of succinate dehydrogenase (SDH) is a mode of action (MOA) of toxicity of PA in snowshoe hares. We tested this hypothesis by measuring the effect of PA on the activity of SDH in liver mitochondria isolated from wild hares. In addition, we used molecular modeling to determine the specific binding site of PA on SDH. We found that PA inhibits SDH from hares by an uncompetitive mechanism in a dose-dependent manner. Molecular modeling suggests that inhibition of SDH is a result of binding of PA at the ubiquinone binding sites in complex II. Our results provide a MOA for toxicity that may be responsible for the concentration-dependent anti-feedant effects of PA. We propose that snowshoe hares reduce the dose-dependent toxic consequences of PA by relying on efflux transporters and metabolizing enzymes that lower systemic exposure to dietary PA.

Metabolic responses of Eisenia fetida after sub-lethal exposure to organic contaminants with different toxic modes of action

Nuclear magnetic resonance (NMR) – based metabolomics has the potential to identify toxic responses of contaminants within a mixture in contaminated soil. This study evaluated the metabolic response of Eisenia fetida after exposure to an array of organic compounds to determine whether contaminant-specific responses could be identified. The compounds investigated in contact tests included: two pesticides (carbaryl and chlorpyrifos), three pharmaceuticals (carbamazephine, estrone and caffeine), two persistent organohalogens (Aroclor 1254 and PBDE 209) and two industrial compounds (nonylphenol and dimethyl phthalate). Control and contaminant-exposed metabolic profiles were distinguished using principal component analysis and potential contaminant-specific biomarkers of exposure were found for several contaminants. These results suggest that NMR-based metabolomics offers considerable promise for differentiating between the different toxic modes of action (MOA) associated with sub-lethal toxicity to earthworms.