Profiling the endocrine disrupting properties of triazines, triazoles and short-chain PFAS.

Persistent, mobile and toxic (PMT) compounds released to the environment are likely to pollute drinking water sources due to their slow environmental degradation (persistency) and high water solubility (mobility). The aim of the present study was to create in vitro hazard profiles for sixteen triazoles, nine triazines and eleven PFAS based on their agonistic and antagonistic effects in estrogen receptor (ER), androgen receptor (AR) and thyroid hormone receptor (TR) reporter gene assays, their ability to bind human transthyretin (TTR), and their effects on steroidogenesis. The triazole fungicides tetraconazole, bitertanol, fenbuconazole, tebuconazole, cyproconazole, difenoconazole, propiconazole, paclobutrazol and triadimenol had agonistic or antagonistic effects on the ER and AR. Difenoconazole, propiconazole and triadimenol were also found to be TR antagonists. The triazine herbicide ametryn was an ER, AR and TR antagonist. The same nine triazole fungicides and the triazines atrazine, deethyl-atrazine and ametryn affected the secretion of steroid hormones. Furthermore, PFAS compounds PFBS, PFHxS, PFHxA, PFOS, PFOA and GenX and the triazoles bitertanol, difenoconazole and 4-methyl benzotriazole were found to displace T4 from TTR. These results are in line with earlier in vitro and in vivo studies on the endocrine disrupting properties of triazines, triazoles and PFAS. The present study demonstrates that this battery of in vitro bioassays can be used to profile compounds from different classes based on their endocrine disrupting properties as a first step to prioritize them for further research, emission reduction, environmental remediation and regulatory purposes.

A comparison of the TempO-Seq and Affymetrix microarray platform using RTqPCR validation.

Next-generation risk assessment relies on mechanistic data from new approach methods, including transcriptome data. Various technologies, such as high-throughput targeted sequencing methods and microarray technologies based on hybridization with complementary probes, are used to determine differentially expressed genes (DEGs). The integration of data from different technologies requires a good understanding of the differences arising from the use of various technologies.To better understand the differences between the TempO-Seq platform and Affymetrix chip technology, whole-genome data for the volatile compound dimethylamine were compared. Selected DEGs were also confirmed using RTqPCR validation. Although the overlap of DEGs between TempO-Seq and Affymetrix was no higher than 37%, a comparison of the gene regulation in terms of log2fold changes revealed a very high concordance. RTqPCR confirmed the majority of DEGs from either platform in the examined dataset. Only a few conflicts were found (11%), while 22% were not confirmed, and 3% were not detected.Despite the observed differences between the two platforms, both can be validated using RTqPCR. Here we highlight some of the differences between the two platforms and discuss their applications in toxicology.

Uncovering the toxicity mechanisms of a series of carboxylic acids in liver cells through computational and experimental approaches.

Hepatotoxicity poses a significant concern in drug design due to the potential liver damage that can be caused by new drugs. Among common manifestations of hepatotoxic damage is lipid accumulation in hepatic tissue, resulting in liver steatosis or phospholipidosis. Carboxylic derivatives are prone to interfere with fatty acid metabolism and cause lipid accumulation in hepatocytes. This study investigates the toxic behaviour of 24 structurally related carboxylic acids in hepatocytes, specifically their ability to cause accumulation of fatty acids and phospholipids. Using high-content screening (HCS) assays, we identified two distinct lipid accumulation patterns. Subsequently, we developed structure-activity relationship (SAR) and quantitative structure-activity relationship (QSAR) models to determine relevant molecular substructures and descriptors contributing to these adverse effects. Additionally, we calculated physicochemical properties associated with lipid accumulation in hepatocytes and examined their correlation with our chemical structure characteristics. To assess the applicability of our findings to a wide range of chemical compounds, we employed two external datasets to evaluate the distribution of our QSAR descriptors. Our study highlights the significance of subtle molecular structural variations in triggering hepatotoxicity, such as the presence of nitrogen or the specific arrangement of substitutions within the carbon chain. By employing our comprehensive approach, we pinpointed specific molecules and elucidated their mechanisms of toxicity, thus offering valuable insights to guide future toxicology investigations.

Protectiveness of NAM-based hazard assessment - which testing scope is required?

Hazard assessment (HA) requires toxicity tests to allow deriving protective points of departure (PoDs) for risk assessment irrespective of a compound's mode of action (MoA). The scope of in vitro test batteries (ivTB) thereby necessitated for systemic toxicity is still unclear. We explored the protectiveness regarding systemic toxicity of an ivTB with a scope, which was guided by previous findings from rodent studies, where examining six main targets, including liver and kidney, was sufficient to predict the guideline scope-based PoD with high probability. The ivTB comprises human in vitro models representing liver, kidney, lung and the neuronal system covering transcriptome, mitochondrial dysfunction and neuronal outgrowth. Additionally, 32 CALUX®- and 10 HepG2 BAC-GFP reporters cover a broad range of disturbance mechanisms. Eight compounds were chosen for causing adverse effects such as immunotoxicity or anemia in vivo, i.e., effects not directly covered by assays in the ivTB. PoDs derived from the ivTB and from oral repeated dose studies in rodents were extrapolated to maximum unbound plasma concentrations for comparison. The ivTB-based PoDs were one to five orders of magnitude lower than in vivo PoDs for six of eight compounds, implying that they were protective. The extent of in vitro response varied across test compounds. Especially for hematotoxic substances, the ivTB showed either no response or only cytotoxicity. Assays better capturing this type of hazard would be needed to complement the ivTB. This study highlights the potentially broad applicability of ivTBs for deriving protective PoDs of compounds with unknown MoA.

Animal tests are used to determine which amount of a chemical is toxic ('threshold of toxicity') and which organs are affected. In principle, the threshold can also be derived solely from tests with cultured cells. However, only a limited number of cell types can practically be tested, so one challenge is to determine how many and which types shall be tested. In animal studies, only few organs including liver and kidney are regularly among those most sensitively affected. We explored whether a cell-based test battery representing these sensitive organs and covering important mechanisms of toxicity can be used to derive protective human thresholds. To challenge this approach, eight chemicals were tested that primarily cause effects in organs not directly represented in our test battery. Results provided protective thresholds for most of the investigated compounds and gave indications how to further improve the approach towards a full-fledged replacement for animal tests.

Toward implementing virtual control groups in nonclinical safety studies.

Historical data from control groups in animal toxicity studies is currently mainly used for comparative purposes to assess validity and robustness of study results. Due to the highly controlled environment in which the studies are performed and the homogeneity of the animal collectives it has been proposed to use the historical data for building so-called virtual control groups, which could replace partly or entirely the concurrent control. This would constitute a substantial contribution to the reduction of animal use in safety studies. Before the concept can be implemented, the prerequisites regarding data collection, curation and statistical evaluation together with a validation strategy need to be identified to avoid any impairment of the study outcome and subsequent consequences for human risk assessment. To further assess and develop the concept of virtual control groups the transatlantic think tank for toxicology (t4) sponsored a workshop with stakeholders from the pharmaceutical and chemical industry, academia, FDA, pharmaceutical, contract research organizations (CROs), and non-governmental organizations in Washington, which took place in March 2023. This report summarizes the current efforts of a European initiative to share, collect and curate animal control data in a centralized database and the first approaches to identify optimal matching criteria between virtual controls and the treatment arms of a study as well as first reflections about strategies for a qualification procedure and potential pitfalls of the concept.

Animal safety studies are usually performed with three groups of animals where increasing amounts of the test chemical are given to the animals and one control group where the animals do not receive the test chemical. The design of such studies, the characteristics of the animals, and the measured parameters are often very similar from study to study. Therefore, it has been suggested that measurement data from the control groups could be reused from study to study to lower the total number of animals per study. This could reduce animal use by up to 25% for such standardized studies. A workshop was held to discuss the pros and cons of such a concept and what would have to be done to implement it without threatening the reliability of the study outcome or the resulting human risk assessment.

Update of the Cancer Potency Database (CPDB) to enable derivations of Thresholds of Toxicological Concern (TTC) for cancer potency.

The purpose of this study was to update the existing Cancer Potency Database (CPDB) in order to support the development of a dataset of compounds, with associated points of departure (PoDs), to enable a review and update of currently applied values for the Threshold of Toxicological Concern (TTC) for cancer endpoints. This update of the current CPDB, last reviewed in 2012, includes the addition of new data (44 compounds and 158 studies leading to additional 359 dose-response curves). Strict inclusion criteria were established and applied to select compounds and studies with relevant cancer potency data. PoDs were calculated from dose-response modeling, including the benchmark dose (BMD) and the lower 90% confidence limits (BMDL) at a specified benchmark response (BMR) of 10%. The updated full CPDB database resulted in a total of 421 chemicals which had dose-response data that could be used to calculate PoDs. This candidate dataset for cancer TTC is provided in a transparent and adaptable format for further analysis of TTC to derive cancer potency thresholds.

Exposure considerations in human safety assessment: Report from an EPAA Partners' Forum.

Understanding and estimating the exposure to a substance is one of the fundamental requirements for safe manufacture and use. Many approaches are taken to determine exposure to substances, mainly driven by potential use and regulatory need. There are many opportunities to improve and optimise the use of exposure information for chemical safety. The European Partnership for Alternative Approaches to Animal Testing (EPAA) therefore convened a Partners' Forum (PF) to explore exposure considerations in human safety assessment of industrial products to agree key conclusions for the regulatory acceptance of exposure assessment approaches and priority areas for further research investment. The PF recognised the widescale use of exposure information across industrial sectors with the possibilities of creating synergies between different sectors. Further, the PF acknowledged that the EPAA could make a significant contribution to promote the use of exposure data in human safety assessment, with an aim to address specific regulatory needs. To achieve this, research needs, as well as synergies and areas for potential collaboration across sectors, were identified.

Workshop report: Challenges faced in developing inhalation thresholds of Toxicological Concern (TTC) - State of the science and next steps.

A challenging step in human risk assessment of chemicals is the derivation of safe thresholds. The Threshold of Toxicological Concern (TTC) concept is one option which can be used for the safety evaluation of substances with a limited toxicity dataset, but for which exposure is sufficiently low. The application of the TTC is generally accepted for orally or dermally exposed cosmetic ingredients; however, these values cannot directly be applied to the inhalation route because of differences in exposure route versus oral and dermal. Various approaches of an inhalation TTC concept have been developed over recent years to address this. A virtual workshop organized by Cosmetics Europe, held in November 2020, shared the current state of the science regarding the applicability of existing inhalation TTC approaches to cosmetic ingredients. Key discussion points included the need for an inhalation TTC for local respiratory tract effects in addition to a systemic inhalation TTC, dose metrics, database building and quality of studies, definition of the chemical space and applicability domain, and classification of chemicals with different potencies. The progress made to date in deriving inhalation TTCs was highlighted, as well as the next steps envisaged to develop them further for regulatory acceptance and use.

Substantiate a read-across hypothesis by using transcriptome data-A case study on volatile diketones.

This case study explores the applicability of transcriptome data to characterize a common mechanism of action within groups of short-chain aliphatic α-, ß-, and γ-diketones. Human reference in vivo data indicate that the α-diketone diacetyl induces bronchiolitis obliterans in workers involved in the preparation of microwave popcorn. The other three α-diketones induced inflammatory responses in preclinical in vivo animal studies, whereas beta and gamma diketones in addition caused neuronal effects. We investigated early transcriptional responses in primary human bronchiolar (PBEC) cell cultures after 24 h and 72 h of air-liquid exposure. Differentially expressed genes (DEGs) were assessed based on transcriptome data generated with the EUToxRisk gene panel of Temp-O-Seq®. For each individual substance, genes were identified displaying a consistent differential expression across dose and exposure duration. The log fold change values of the DEG profiles indicate that α- and ß-diketones are more active compared to γ-diketones. α-diketones in particular showed a highly concordant expression pattern, which may serve as a first indication of the shared mode of action. In order to gain a better mechanistic understanding, the resultant DEGs were submitted to a pathway analysis using ConsensusPathDB. The four α-diketones showed very similar results with regard to the number of activated and shared pathways. Overall, the number of signaling pathways decreased from α-to ß-to γ-diketones. Additionally, we reconstructed networks of genes that interact with one another and are associated with different adverse outcomes such as fibrosis, inflammation or apoptosis using the TRANSPATH-database. Transcription factor enrichment and upstream analyses with the geneXplain platform revealed highly interacting gene products (called master regulators, MRs) per case study compound. The mapping of the resultant MRs on the reconstructed networks, visualized similar gene regulation with regard to fibrosis, inflammation and apoptosis. This analysis showed that transcriptome data can strengthen the similarity assessment of compounds, which is of particular importance, e.g., in read-across approaches. It is one important step towards grouping of compounds based on biological profiles.

Potential of the zebrafish (Danio rerio) embryo test to discriminate between chemicals of similar molecular structure-a study with valproic acid and 14 of its analogues.

Valproic acid is a frequently used antiepileptic drug and known pediatric hepatotoxic agent. In search of pharmaceuticals with increased effectiveness and reduced toxicity, analogue chemicals came into focus. So far, toxicity and teratogenicity data of drugs and metabolites have usually been collected from mammalian model systems such as mice and rats. However, in an attempt to reduce mammalian testing while maintaining the reliability of toxicity testing of new industrial chemicals and drugs, alternative test methods are being developed. To this end, the potential of the zebrafish (Danio rerio) embryo to discriminate between valproic acid and 14 analogues was investigated by exposing zebrafish embryos for 120 h post fertilization in the extended version of the fish embryo acute toxicity test (FET; OECD TG 236), and analyzing liver histology to evaluate the correlation of liver effects and the molecular structure of each compound. Although histological evaluation of zebrafish liver did not identify steatosis as the prominent adverse effect typical in human and mice, the structure-activity relationship (SAR) derived was comparable not only to human HepG2 cells, but also to available in vivo mouse and rat data. Thus, there is evidence that zebrafish embryos might serve as a tool to bridge the gap between subcellular, cell-based systems and vertebrate models.

Application of high-throughput transcriptomics for mechanism-based biological read-across of short-chain carboxylic acid analogues of valproic acid.

Chemical read-across is commonly evaluated without specific knowledge of the biological mechanisms leading to observed adverse outcomes in vivo. Integrating data that indicate shared modes of action in humans will strengthen read-across cases. Here we studied transcriptomic responses of primary human hepatocytes (PHH) to a large panel of carboxylic acids to include detailed mode-of-action data as a proof-of-concept for read-across in risk assessment. In rodents, some carboxylic acids, including valproic acid (VPA), are known to cause hepatic steatosis, whereas others do not. We investigated transcriptomics responses of PHHs exposed for 24 h to 18 structurally different VPA analogues in a concentration range to determine biological similarity in relation to in vivo steatotic potential. Using a targeted high-throughput screening assay, we assessed the differential expression of ~3,000 genes covering relevant biological pathways. Differentially expressed gene analysis revealed differences in potency of carboxylic acids, and expression patterns were highly similar for structurally similar compounds. Strong clustering occurred for steatosis-positive versus steatosis-negative carboxylic acids. To quantitatively define biological read-across, we combined pathway analysis and weighted gene co-expression network analysis. Active carboxylic acids displayed high similarity in gene network modulation. Importantly, free fatty acid synthesis modulation and stress pathway responses are affected by active car­boxylic acids, providing coherent mechanistic underpinning for our findings. Our work shows that transcriptomic analysis of cultured human hepatocytes can reinforce the prediction of liver injury outcome based on quantitative and mechanistic biological data and support its application in read-across.

RespiraTox - Development of a QSAR model to predict human respiratory irritants.

Respiratory irritation is an important human health endpoint in chemical risk assessment. There are two established modes of action of respiratory irritation, 1) sensory irritation mediated by the interaction with sensory neurons, potentially stimulating trigeminal nerve, and 2) direct tissue irritation. The aim of our research was to, develop a QSAR method to predict human respiratory irritants, and to potentially reduce the reliance on animal testing for the identification of respiratory irritants. Compounds are classified as irritating based on combined evidence from different types of toxicological data, including inhalation studies with acute and repeated exposure. The curated project database comprised 1997 organic substances, 1553 being classified as irritating and 444 as non-irritating. A comparison of machine learning approaches, including Logistic Regression (LR), Random Forests (RFs), and Gradient Boosted Decision Trees (GBTs), showed, the best classification was obtained by GBTs. The LR model resulted in an area under the curve (AUC) of 0.65, while the optimal performance for both RFs and GBTs gives an AUC of 0.71. In addition to the classification and the information on the applicability domain, the web-based tool provides a list of structurally similar analogues together with their experimental data to facilitate expert review for read-across purposes.

Integrate mechanistic evidence from new approach methodologies (NAMs) into a read-across assessment to characterise trends in shared mode of action.

Read-across approaches often remain inconclusive as they do not provide sufficient evidence on a common mode of action across the category members. This read-across case study on thirteen, structurally similar, branched aliphatic carboxylic acids investigates the concept of using human-based new approach methods, such as in vitro and in silico models, to demonstrate biological similarity. Five out of the thirteen analogues have preclinical in vivo studies. Three out of them induced lipid accumulation or hypertrophy in preclinical studies with repeated exposure, which leads to the read-across hypothesis that the analogues can potentially induce hepatic steatosis. To confirm the selection of analogues, the expression patterns of the induced differentially expressed genes (DEGs) were analysed in a human liver model. With increasing dose, the expression pattern within the tested analogues got more similar, which serves as a first indication of a common mode of action and suggests differences in the potency of the analogues. Hepatic steatosis is a well-known adverse outcome, for which over 55 adverse outcome pathways have been identified. The resulting adverse outcome pathway (AOP) network, comprised a total 43 MIEs/KEs and enabled the design of an in vitro testing battery. From the AOP network, ten MIEs, early and late KEs were tested to systematically investigate a common mode of action among the grouped compounds. The targeted testing of AOP specific MIE/KEs shows that biological activity in the category decreases with side chain length. A similar trend was evident in measuring liver alterations in zebra fish embryos. However, activation of single MIEs or early KEs at in vivo relevant doses did not necessarily progress to the late KE "lipid accumulation". KEs not related to the read-across hypothesis, testing for example general mitochondrial stress responses in liver cells, showed no trend or biological similarity. Testing scope is a key issue in the design of in vitro test batteries. The Dempster-Shafer decision theory predicted those analogues with in vivo reference data correctly using one human liver model or the CALUX reporter assays. The case study shows that the read-across hypothesis is the key element to designing the testing strategy. In the case of a good mechanistic understanding, an AOP facilitates the selection of reliable human in vitro models to demonstrate a common mode of action. Testing DEGs, MIEs and early KEs served to show biological similarity, whereas the late KEs become important for confirmation, as progression from MIEs to AO is not always guaranteed.

The in vitro assessment of the toxicity of volatile, oxidisable, redox-cycling compounds: phenols as an example.

Phenols are regarded as highly toxic chemicals. Their effects are difficult to study in in vitro systems because of their ambiguous fate (degradation, auto-oxidation and volatility). In the course of in vitro studies of a series of redox-cycling phenols, we found evidences of cross-contamination in several in vitro high-throughput test systems, in particular by trimethylbenzene-1, 4-diol/trimethylhydroquinone (TMHQ) and 2,6-di-tertbutyl-4-ethylphenol (DTBEP), and investigated in detail the physicochemical basis for such phenomenon and how to prevent it. TMHQ has fast degradation kinetics followed by significant diffusion rates of the resulting quinone to adjacent wells, other degradation products being able to air-diffuse as well. DTBEP showed lower degradation kinetics, but a higher diffusion rate. In both cases the in vitro toxicity was underestimated because of a decrease in concentration, in addition to cross-contamination to neighbouring wells. We identified four degradation products for TMHQ and five for DTBEP indicating that the current effects measured on cells are not only attributable to the parent phenolic compound. To overcome these drawbacks, we investigated in detail the physicochemical changes occurring in the course of the incubation and made use of gas-permeable and non-permeable plastic seals to prevent it. Diffusion was greatly prevented by the use of both plastic seals, as revealed by GC-MS analysis. Gas non-permeable plastic seals, reduced to a minimum compounds diffusion as well oxidation and did not affect the biological performance of cultured cells. Hence, no toxicological cross-contamination was observed in neighbouring wells, thus allowing a more reliable in vitro assessment of phenol-induced toxicity.

NAM-supported read-across: From case studies to regulatory guidance in safety assessment.

The use of new approach methodologies (NAMs) in support of read-across (RAx) approaches for regulatory purposes is a main goal of the EU-ToxRisk project. To bring this forward, EU-ToxRisk partners convened a workshop in close collaboration with regulatory representatives from key organizations including European regulatory agencies, such as the European Chemicals Agency (ECHA) and the European Food Safety Authority (EFSA), as well as the Scientific Committee on Consumer Safety (SCCS), national agencies from several European countries, Japan, Canada and the USA, as well as the Organisation for Economic Cooperation and Development (OECD). More than a hundred people actively participated in the discussions, bringing together diverse viewpoints across academia, regulators and industry. The discussion was organized starting from five practical cases of RAx applied to specific problems that offered the oppor-tunity to consider real examples. There was general consensus that NAMs can improve confidence in RAx, in particular in defining category boundaries as well as characterizing the similarities/dissimilarities between source and target substances. In addition to describing dynamics, NAMs can be helpful in terms of kinetics and metabolism that may play an important role in the demonstration of similarity or dissimilarity among the members of a category. NAMs were also noted as effective in providing quanti-tative data correlated with traditional no observed adverse effect levels (NOAELs) used in risk assessment, while reducing the uncertainty on the final conclusion. An interesting point of view was the advice on calibrating the number of new tests that should be carefully selected, avoiding the allure of "the more, the better". Unfortunately, yet unsurprisingly, there was no single approach befitting every case, requiring careful analysis delineating the optimal approach. Expert analysis and assessment of each specific case is still an important step in the process.

Threshold of Toxicological Concern-An Update for Non-Genotoxic Carcinogens.

The Threshold of Toxicological Concern (TTC) concept can be applied to organic compounds with the known chemical structure to derive a threshold for exposure, below which a toxic effect on human health by the compound is not expected. The TTC concept distinguishes between carcinogens that may act as genotoxic and non-genotoxic compounds. A positive prediction of a genotoxic mode of action, either by structural alerts or experimental data, leads to the application of the threshold value for genotoxic compounds. Non-genotoxic substances are assigned to the TTC value of their respective Cramer class, even though it is recognized that they could test positive in a rodent cancer bioassay. This study investigated the applicability of the Cramer classes specifically to provide adequate protection for non-genotoxic carcinogens. For this purpose, benchmark dose levels based on tumor incidence were compared with no observed effect levels (NOELs) derived from non-, pre- or neoplastic lesions. One key aspect was the categorization of compounds as non-genotoxic carcinogens. The recently finished CEFIC LRI project B18 classified the carcinogens of the Carcinogenicity Potency DataBase (CPDB) as either non-genotoxic or genotoxic compounds based on experimental or in silico data. A detailed consistency check resulted in a dataset of 137 non-genotoxic organic compounds. For these 137 compounds, NOEL values were derived from high quality animal studies with oral exposure and chronic duration using well-known repositories, such as RepDose, ToxRef, and COSMOS DB. Further, an effective tumor dose (ETD10) was calculated and compared with the lower confidence limit on benchmark dose levels (BMDL10) derived by model averaging. Comparative analysis of NOEL/EDT10/BMDL10 values showed that potentially bioaccumulative compounds in humans, as well as steroids, which both belong to the exclusion categories, occur predominantly in the region of the fifth percentiles of the distributions. Excluding these 25 compounds resulted in significantly higher but comparable fifth percentile chronic NOEL and BMDL10 values, while the fifth percentile EDT10 value was slightly higher but not statistically significant. The comparison of the obtained distributions of NOELs with the existing Cramer classes and their derived TTC values supports the application of Cramer class thresholds to all non-genotoxic compounds, such as non-genotoxic carcinogens.

Risk assessment for irritating chemicals - Derivation of extrapolation factors.

Irritation of the eyes and the upper respiratory tract are important endpoints for setting guide values for chemicals. To optimize the use of the often-limited data, we analysed controlled human exposure studies (CHS) with 1-4 h inhalation of the test substance, repeated dose inhalation studies in rodents, and Alarie-Tests and derived extrapolation factors (EF) for exposure duration, inter- and intraspecies differences. For the endpoint irritating effects in the respiratory tract in rodents, geometric mean (GM) values of 1.9 were obtained for the EF for subacute→subchronic (n = 16), 2.1 for subchronic→chronic (n = 40), and 2.9 for subacute→chronic (n = 10) extrapolation. Based on these data we suggest an EF of 2 for subchronic→chronic and of 4 for subacute→chronic extrapolation. In CHS, exposure concentration determines the effects rather than exposure duration. Slight reversible effects during 4 h exposure indicate that an EF of 1 can be considered for assessing chronic exposures. To assess species extrapolation, 10 chemicals were identified with both, reliable rat inhalation studies and CHS. The GM of the ratio between the No Observed Adverse Effect Concentration (NOAEC) in rats and humans was 2.3 and increased to 3.6 when expanding the dataset to all available EF (n = 25). Based on these analyses, an EF of 3 is suggested to extrapolate from a NOAEC in a chronic rat study to a NOAEC in a CHS. The analysis of EFs for the extrapolation from a 50% decrease in respiratory frequency in the Alarie test in mice (RD50) to a NOAEC in a CHS resulted in a GM of 40, for both, the reliable (n = 11) and the overall dataset (n = 19). We propose to use the RD50 from the Alarie test for setting guide values and to use 40 as EF. Efs for intraspecies differences in the human population must account for susceptible persons, most importantly for persons with chemical intolerance (CI), who show subjective signs of irritation at low concentrations. The limited data available do not justify to deviate from an EF of 10 - 20 as currently used in different regulatory settings.

Combining In Vivo Data with In Silico Predictions for Modeling Hepatic Steatosis by Using Stratified Bagging and Conformal Prediction.

Hepatic steatosis (fatty liver) is a severe liver disease induced by the excessive accumulation of fatty acids in hepatocytes. In this study, we developed reliable in silico models for predicting hepatic steatosis on the basis of an in vivo data set of 1041 compounds measured in rodent studies with repeated oral exposure. The imbalanced nature of the data set (1:8, with the "steatotic" compounds belonging to the minority class) required the use of meta-classifiers-bagging with stratified under-sampling and Mondrian conformal prediction-on top of the base classifier random forest. One major goal was the investigation of the influence of different descriptor combinations on model performance (tested by predicting an external validation set): physicochemical descriptors (RDKit), ToxPrint features, as well as predictions from in silico nuclear receptor and transporter models. All models based upon descriptor combinations including physicochemical features led to reasonable balanced accuracies (BAs between 0.65 and 0.69 for the respective models). Combining physicochemical features with transporter predictions and further with ToxPrint features gave the best performing model (BAs up to 0.7 and efficiencies of 0.82). Whereas both meta-classifiers proved useful for this highly imbalanced toxicity data set, the conformal prediction framework also guarantees the error level and thus might be favored for future studies in the field of predictive toxicology.

Integrated In Silico Models for the Prediction of No-Observed-(Adverse)-Effect Levels and Lowest-Observed-(Adverse)-Effect Levels in Rats for Sub-chronic Repeated-Dose Toxicity.

Repeated-dose toxicity (RDT) is a critical endpoint for hazard characterization of chemicals and is assessed to derive safe levels of exposure for human health. Here we present the first attempt to model simultaneously no-observed-(adverse)-effect level (NO(A)EL) and lowest-observed-(adverse)-effect level (LO(A)EL). Classification and regression models were derived based on rat sub-chronic repeated dose toxicity data for 327 compounds from the Fraunhofer RepDose database. Multi-category classification models were built for both NO(A)EL and LO(A)EL though a consensus of statistics- and fragment-based algorithms, while regression models were based on quantitative relationships between the endpoints and SMILES-based attributes. NO(A)EL and LO(A)EL models were integrated, and predictions were compared to exclude inconsistent values. This strategy improved the performance of single models, leading to R2 greater than 0.70, root-mean-square error (RMSE) lower than 0.60 (for regression models), and accuracy of 0.61-0.73 (for classification models) on the validation set, based on the endpoint and the threshold applied for selecting predictions. This study confirms the effectiveness of the modeling strategy presented here for assessing RDT of chemicals using in silico models.