Development of algal interspecies correlation estimation models for chemical hazard assessment.

Web-based Interspecies Correlation Estimation (ICE) is an application developed to predict the acute toxicity of a chemical from 1 species to another taxon. Web-ICE models use the acute toxicity value for a surrogate species to predict effect values for other species, thus potentially filling in data gaps for a variety of environmental assessment purposes. Web-ICE has historically been dominated by aquatic and terrestrial animal prediction models. Web-ICE models for algal species were essentially absent and are addressed in the present study. A compilation of public and private sector-held algal toxicity data were compiled and reviewed for quality based on relevant aspects of individual studies. Interspecies correlations were constructed from the most commonly tested algal genera for a broad spectrum of chemicals. The ICE regressions were developed based on acute 72-h and 96-h endpoint values involving 1647 unique studies on 476 unique chemicals encompassing 40 genera and 70 species of green, blue-green, and diatom algae. Acceptance criteria for algal ICE models were established prior to evaluation of individual models and included a minimum sample size of 3, a statistically significant regression slope, and a slope estimation parameter ≥0.65. A total of 186 ICE models were possible at the genus level, with 21 meeting quality criteria; and 264 ICE models were developed at the species level, with 32 meeting quality criteria. Algal ICE models will have broad utility in screening environmental hazard assessments, data gap filling in certain regulatory scenarios, and as supplemental information to derive species sensitivity distributions. Environ Toxicol Chem 2016;35:2368-2378. Published 2016 Wiley Periodicals Inc. on behalf of SETAC. This article is a US government work and, as such, is in the public domain in the United States of America.

Comparison of species sensitivity distributions derived from interspecies correlation models to distributions used to derive water quality criteria.

Species sensitivity distributions (SSD) require a large number of measured toxicity values to define a hazard level protective of multiple species. This investigation comprehensively evaluated the accuracy of SSDs generated from toxicity values predicted from interspecies correlation estimation (ICE) models. ICE models are log-log correlations of multiple chemical toxicity values for a pair of species that allow the toxicity of multiple species to be predicted from a single measured acute toxicity value for a surrogate species. ICE SSDs were generated using four surrogate species (fathead minnow, Pimephales promelas; rainbow trout, Oncorhynchus mykiss; sheepshead minnow, Cyprinodon varigatus; and water flea, Daphnia magna). ICE-based hazard concentrations (HC5s) from the 5th percentile of the log-logistic distribution of toxicity values were compared to HC5s determined from the acute toxicity of 55 chemicals from the United States Environmental Protection Agency Ambient Water Quality Criteria (AWQC). Measured fish and invertebrate acute toxicity data and HC5s from the AWQC data sets were compared to ICE-based HC5s. Surrogate species choice was found to be an important consideration in developing predictive HC5s. These results illustrated that fish predict fish betterthan invertebrates and D. magna predicted invertebrates better than most fish. For example, a mixed model of predicted fish and invertebrates from fathead minnow and D. magna as surrogate species provided predictive relationships with an average factor of 3.0 (+/- 6.7) over 7 orders of toxic magnitude and several chemical classes (HC5(predicted)/HC5(measured)). The application of ICE models is recommended as a valid approach for generating SSDs and hazard concentrations for chemicals with limited toxicity data.

The reduced local lymph node assay: the impact of group size.

The local lymph node assay (LLNA) is a skin sensitization test that provides animal welfare benefits. To reduce animal usage further, a modified version (rLLNA) was proposed. Conducting the rLLNA as a screening test with a single high dose group and vehicle control differentiated accurately between skin sensitizers and non-sensitizers. This study examined whether a reduction in animal number/group is feasible. Historical data were utilized to examine the impact of conducting the rLLNA with two mice/group. To assess the effect on the stimulation index (SI) 41 datasets with individual animal data derived using five mice/group were analysed. SIs were calculated on all possible combinations of two control and two high dose group disintegrations per minute (dpm) values. For 25 of 33 sensitizer datasets, > 96% of possible dpm combinations resulted in a calculated SI > 3. The lowest percentages of positive SIs were observed with weak allergens when, in the standard LLNA, the mean SIs would have been nearer to the threshold value of 3. The results indicate that moderate, strong and extreme allergens are more likely than weak allergens to be identified as sensitizers when group sizes of two mice are used within the rLLNA. It is concluded that a rLLNA with two mice/group would display decreased sensitivity and is inappropriate for use in hazard identification.

Extrapolating local lymph node assay EC3 values to estimate relative sensitizing potency.

The assessment of the potency of a skin sensitizing chemical is a key starting point for its subsequent risk assessment/management. The Local Lymph Node Assay can provide information on the relative skin sensitizing potency of contact allergens by interpolation from the dose response curve the concentration of a chemical required to elicit a threshold positive response (EC3 value). However, interpolation requires that the dose response curve have at least one stimulation index (SI) value above and one SI value below the threshold value of 3. For instances where all test concentrations result in SI values above 3, there was a need to develop a method that would permit estimation of EC3 values. This has been achieved by log-linear extrapolation using the two lowest test concentrations from the dose response curve. Before applying this approach, it is important that data quality is assessed. The dose response must include concentrations on the linear portion of the curve and, ideally, the SI induced by the lowest dose should approach 3. Judicious use of this approach for extrapolating EC3 values can provide information on a likely potency classification for use in risk assessment and may avoid the need for repeat animal testing.

Quantification of chemical peptide reactivity for screening contact allergens: a classification tree model approach.

In the interest of reducing animal use, in vitro alternatives for skin sensitization testing are under development. One unifying characteristic of chemical allergens is the requirement that they react with proteins for the effective induction of skin sensitization. The majority of chemical allergens are electrophilic and react with nucleophilic amino acids. To determine whether and to what extent reactivity correlates with skin sensitization potential, 82 chemicals comprising allergens of different potencies and nonallergenic chemicals were evaluated for their ability to react with reduced glutathione (GSH) or with two synthetic peptides containing either a single cysteine or lysine. Following a 15-min reaction time with GSH, or a 24-h reaction time with the two synthetic peptides, the samples were analyzed by high-performance liquid chromatography. UV detection was used to monitor the depletion of GSH or the peptides. The peptide reactivity data were compared with existing local lymph node assay data using recursive partitioning methodology to build a classification tree that allowed a ranking of reactivity as minimal, low, moderate, and high. Generally, nonallergens and weak allergens demonstrated minimal to low peptide reactivity, whereas moderate to extremely potent allergens displayed moderate to high peptide reactivity. Classifying minimal reactivity as nonsensitizers and low, moderate, and high reactivity as sensitizers, it was determined that a model based on cysteine and lysine gave a prediction accuracy of 89%. The results of these investigations reveal that measurement of peptide reactivity has considerable potential utility as a screening approach for skin sensitization testing, and thereby for reducing reliance on animal-based test methods.

Interspecies correlation estimates predict protective environmental concentrations.

Environmental risk assessments often use multiple single species toxicity test results and species sensitivity distributions (SSDs) to derive a predicted no-effect concentration in the environment, typically the 5th percentile of the SSD, termed the HC5. The shape and location of the distribution are best known when populated with numerous toxicity values. To help overcome the cost of multiple toxicity tests, we explored the potential of the U.S. EPA's Interspecies Correlation Estimation (ICE) program to predict single species toxicity values from a single known toxicity value. ICE uses the initial toxicity estimate for one species to produce correlation toxicity values for multiple species, which can be used to develop SSD and HC5. To test this approach to deriving HC5, we generated toxicity values based on measured toxicity values for three surrogate species Pimephales promelas (Fathead minnow), Onchorynchus mykiss (Rainbow trout), and Daphnia magna (water flea). Algal taxa were not used due to the paucity of high quality algal-aquatic invertebrate and algal-fish correlations. The compounds used (dodecyl linear alkylbenzenesulfonate (LAS), nonylphenol, fenvalerate, atrazine, and copper) have multiple measured toxicity values and diverse modes of action and toxicities. Distribution parameters and HC5 values from the measured toxicity values were compared with ICE predicted distributions and HC5 values. While distributional parameters (scale and intercept) differed between measured and predicted distributions, in general, the ICE-based SSDs had HC5 values that were within an order of magnitude of the measured HC5 values. Examination of species placements within the SSDs indicated that the most sensitive species were coldwater species (e.g., salmonids and Gammarus pseudolimnaeus). These results raise the potential of using quantitative structure activity models to estimate HC5s.

Development of a peptide reactivity assay for screening contact allergens.

Allergic contact dermatitis resulting from skin sensitization is a common occupational and environmental health problem. In recent years, the local lymph node assay (LLNA) has emerged as a practical option for assessing the skin sensitization potential of chemicals. In addition to accurate identification of skin sensitizers, the LLNA can also provide a reliable measure of relative sensitization potency; information that is pivotal in successful management of human health risks. However, even with the significant animal welfare benefits provided by the LLNA, there is still interest in the development of nonanimal test methods for skin sensitization testing. One characteristic of a chemical allergen is its ability to react with proteins prior to the induction of skin sensitization. The majority of chemical allergens is electrophilic and as such reacts with nucleophilic amino acids like cysteine or lysine. In order to determine if reactivity correlates with sensitization potential, 38 chemicals representing allergens of different potencies (weak to extreme) and nonsensitizers were evaluated for their ability to react with glutathione or three synthetic peptides containing either cysteine, lysine, or histidine. Following a 15-min reaction time for glutathione or a 24 h reaction period for the three synthetic peptides, the samples were analyzed by HPLC. UV detection was used to monitor the depletion of glutathione or the peptide following reaction. The results demonstrate that a significant correlation (Spearman correlation) exists between allergen potency and the depletion of glutathione (p = 0.001), lysine (p = 0.025), and cysteine (p = 0.020), but not histidine. The peptide with the highest sensitivity was cysteine (80.8%) whereas histidine was the least sensitive (11.5%). The data presented show that measuring peptide reactivity has utility for screening chemicals for their skin sensitization potency and thus potential for reducing our reliance on animal test methods.

Identification of a gene expression profile that discriminates indirect-acting genotoxins from direct-acting genotoxins.

During the safety evaluation process of new drugs and chemicals, a battery of genotoxicity tests is conducted starting with in vitro genotoxicity assays. Obtaining positive results in in vitro genotoxicity tests is not uncommon. Follow-up studies to determine the biological relevance of positive genotoxicity results are costly, time consuming, and utilize animals. More efficient methods, especially for identifying a putative mode of action like an indirect mechanism of genotoxicity (where DNA molecules are not the initial primary targets), would greatly improve the risk assessment for genotoxins. To this end, we are participating in an International Life Sciences Institute (ILSI) project involving studies of gene expression changes caused by model genotoxins. The purpose of the work is to evaluate gene expression tools in general, and specifically for discriminating genotoxins that are direct-acting from indirect-acting. Our lab has evaluated gene expression changes as well as micronuclei (MN) in L5178Y TK(+/-) mouse lymphoma cells treated with six compounds. Direct-acting genotoxins (where DNA is the initial primary target) that were evaluated included the DNA crosslinking agents, mitomycin C (MMC) and cisplatin (CIS), and an alkylating agent, methyl methanesulfonate (MMS). Indirect-acting genotoxins included hydroxyurea (HU), a ribonucleotide reductase inhibitor, taxol (TXL), a microtubule inhibitor, and etoposide (ETOP), a DNA topoisomerase II inhibitor. Microarray gene expression analysis was conducted using Affymetrix mouse oligonucleotide arrays on RNA samples derived from cells which were harvested immediately after the 4 h chemical treatment, and 20 h after the 4 h chemical treatment. The evaluation of these experimental results yields evidence of differentially regulated genes at both 4 and 24 h time points that appear to have discriminating power for direct versus indirect genotoxins, and therefore may serve as a fingerprint for classifying chemicals when their mechanism of action is unknown.

Use of a B cell marker (B220) to discriminate between allergens and irritants in the local lymph node assay.

It has been shown that exposure of mice to contact allergens induces B cell activation in the draining lymph nodes (DLN), as seen by an increase in the percentage of B220+ or IgG/IgM+ cells. We have now examined whether the measurement of the percentage of B220+ cells could be used as an alternative or supplementary endpoint for the local lymph node assay (LLNA) to differentiate between allergenic responses and those few irritants that induce low-level proliferation in the DLN. Mice were treated on the ears, daily for 3 consecutive days, with various allergens (1-chloro-2,4-dinitrobenzene, alpha-hexylcinnamaldehyde, trinitrochlorobenzene, isoeugenol, and eugenol) or irritants (benzalkonium chloride, methyl salicylate, salicylic acid, and sodium lauryl sulfate). The DLN were excised 72 h following the final topical treatment, and the cells were prepared for B220 analysis using flow cytometry. The percentage of B220+ cells in lymph nodes derived from test and vehicle-treated animals was determined for 5 allergens and 4 irritants tested in multiple experiments (n = 3 to 17). As expected, the percentage of B220+ B cells was increased with each of the allergens tested, whereas irritant treatment did not cause similar increases. Moreover, the method was reproducible. For example, the strong allergen, 1-chloro-2,4-dinitrobenzene and the weak allergen, alpha-hexylcinnamaldehyde were identified as allergens in 17 of 17 and in 12 of 13 experiments, respectively. The percentage of B220 values for each chemical treatment (41 observations for allergens; 28 observations for irritants) versus the percentage of B220 values for the concurrent vehicle controls were plotted, and a classification tree model was developed that defined a B220 test:vehicle ratio cutoff of 1.25 for discriminating between allergens (>1.25) and irritants (<1.25). Using this B220 test:vehicle ratio of 1.25 in 93% of the 69 independent observations made, the allergens and irritants tested were identified correctly. Finally, to evaluate the performance of this model in a second independent laboratory, 3 allergens and 2 irritants were tested. Each of the allergens and irritants were classified correctly using the B220 test:vehicle ratio cutoff of 1.25. These data demonstrate that analysis of B220 expression in DLN may be useful in differentiating between allergen and irritant responses induced in chemically treated mice.