A strategy to validate a selection of human effect biomarkers using adverse outcome pathways: Proof of concept for phthalates and reproductive effects.

Human biomonitoring measures the concentrations of environmental chemicals or their metabolites in body fluids or tissues. Complementing exposure biomarkers with mechanistically based effect biomarkers may further elucidate causal pathways between chemical exposure and adverse health outcomes. We combined information on effect biomarkers previously implemented in human observational studies with mechanisms of action reported in experimental studies and with information from published Adverse Outcome Pathways (AOPs), focusing on adverse reproductive effects of phthalate exposure. Phthalates constitute a group of chemicals that are ubiquitous in consumer products and have been related to a wide range of adverse health effects. As a result of a comprehensive literature search, we present an overview of effect biomarkers for reproductive toxicity that are substantiated by mechanistic information. The activation of several receptors, such as PPARα, PPARγ, and GR, may initiate events leading to impaired male and female fertility as well as other adverse effects of phthalate exposure. Therefore, these receptors appear as promising targets for the development of novel effect biomarkers. The proposed strategy connects the fields of epidemiology and toxicology and may strengthen the weight of evidence in observational studies that link chemical exposures to health outcomes.

Tracing nitrogenous disinfection byproducts after medium pressure UV water treatment by stable isotope labeling and high resolution mass spectrometry.

Advanced oxidation processes are important barriers for organic micropollutants (e.g., pharmaceuticals, pesticides) in (drinking) water treatment. Studies indicate that medium pressure (MP) UV/H2O2 treatment leads to a positive response in Ames mutagenicity tests, which is then removed after granulated activated carbon (GAC) filtration. The formed potentially mutagenic substances were hitherto not identified and may result from the reaction of photolysis products of nitrate with (photolysis products of) natural organic material (NOM). In this study we present an innovative approach to trace the formation of disinfection byproducts (DBPs) of MP UV water treatment, based on stable isotope labeled nitrate combined with high resolution mass spectrometry. It was shown that after MP UV treatment of artificial water containing NOM and nitrate, multiple nitrogen containing substances were formed. In total 84 N-DBPs were detected at individual concentrations between 1 to 135 ng/L bentazon-d6 equivalents, with a summed concentration of 1.2 µg/L bentazon-d6 equivalents. The chemical structures of three byproducts were confirmed. Screening for the 84 N-DBPs in water samples from a full-scale drinking water treatment plant based on MP UV/H2O2 treatment showed that 22 of the N-DBPs found in artificial water were also detected in real water samples.

Assessment of drinking water safety in the Netherlands using nationwide exposure and mortality data.

BACKGROUND: Although drinking water in the Netherlands is generally accepted as safe, public concern about health risks of long-term intake still exist. OBJECTIVE: The aim was to explore associations between drinking water quality for nitrate, water hardness, calcium and magnesium and causes-of-death as related to cardiovascular diseases amongst which coronary heart disease and colorectal cancer. METHODS: We used national administrative databases on cause-specific mortality, personal characteristics, residential history, social economic indicators, air quality and drinking water quality for parameters specified by the EU Drinking Water Directive. We put together a cohort of 6,998,623 persons who were at least 30 years old on January 1, 2008 and lived for at least five years on the same address. The average drinking water concentration over 2000-2010 at the production stations were used as exposure indicators. We applied age stratified Cox proportional hazards models. RESULTS: Magnesium was associated with a reduced risk for mortality due to coronary heart diseases: HR of 0.95 (95% CI: 0.90, 0.99) per 10 mg/L increase. For mortality due to cardiovascular diseases, a 100 mg/L increase in calcium was associated with a HR of 1.08 (95% CI: 1.03, 1.13) and an increase of 2.5 mmol/L of water hardness with a HR of 1.06 (95% CI: 1.01, 1.10). The results show an elevated risk for coronary heart disease mortality at calcium concentrations below 30 mg/L, but over the whole exposure range no exposure response relation was observed. For other combinations of drinking water quality parameters and cause-specific mortality studied, no statistical significant associations were identified. CONCLUSION: We identified in this explorative study a protective effect of magnesium for the risk of mortality to coronary heart disease. Also we found an increased risk of mortality due to cardiovascular disease associated with the concentration of calcium and the water hardness in drinking water.

Comparing conventional and green fracturing fluids by chemical characterisation and effect-based screening.

There is public and scientific concern about air, soil and water contamination and possible adverse environmental and human health effects as a result of hydraulic fracturing activities. The use of greener chemicals in fracturing fluid aims to mitigate these effects. This study compares fracturing fluids marketed as either 'conventional' or 'green', as assessed by their chemical composition and their toxicity in bioassays. Chemical composition was analysed via non-target screening using liquid chromatography - high resolution mass spectrometry, while toxicity was evaluated by the Ames fluctuation test to assess mutagenicity and CALUX reporter gene assays to determine specific toxicity. Overall, the results do not indicate that the 'green' fluids are less harmful than the 'conventional' ones. First, there is no clear indication that the selected green fluids contain chemicals present at lower concentrations than the selected conventional fluids. Second, the predicted environmental fate of the identified compounds does not seem to be clearly distinct between the 'green' and 'conventional' fluids, based on the available data for the top five chemicals based on signal intensity that were tentatively identified. Furthermore, Ames fluctuation test results indicate that the green fluids have a similar genotoxic potential than the conventional fluids. Results of the CALUX reporter gene assays add to the evidence that there is no clear difference between the green and conventional fluids. These results do not support the claim that currently available and tested green-labeled fracturing fluids are environmentally more friendly alternatives to conventional fracturing fluids.

Prioritizing anthropogenic chemicals in drinking water and sources through combined use of mass spectrometry and ToxCast toxicity data.

Advancements in high-resolution mass spectrometry based methods have enabled a shift from pure target analysis to target, suspect and non-target screening analyses to detect chemicals in water samples. The multitude of suspect chemicals thereby detected needs to be prioritized for further identification, prior to health risk assessment and potential inclusion into monitoring programs. Here, we compare prioritization of chemicals in Dutch water samples based on relative intensities only to prioritization including hazard information based on high-throughput in vitro toxicity data. Over 1000 suspects detected in sewage treatment plant effluent, surface water, groundwater and drinking water samples were ranked based on their relative intensities. Toxicity data availability and density in the ToxCast database were determined and visualized for these suspects, also in regard to water relevant mechanisms of toxicity. More than 500 suspects could be ranked using occurrence/hazard ratios based on more than 1000 different assay endpoints. The comparison showed that different prioritization strategies resulted in significantly different ranking, with only 2 suspects prioritized based on occurrence among the top 20 in the hazard ranking. We therefore propose a novel scheme that integrates both exposure and hazard data, and efficiently prioritizes which features need to be confidently identified first.

Risk-based approach in the revised European Union drinking water legislation: Opportunities for bioanalytical tools.

A plethora of in vitro bioassays are developed in the context of chemical risk assessment and clinical diagnostics to test effects on different biological processes. Such assays can also be implemented in effect-based monitoring (EBM) of (drinking) water quality alongside chemical analyses. Effects-based monitoring can provide insight into risks for the environment and human health associated with exposure to (unknown) complex, low-level mixtures of micropollutants, which fits in the risk-based approach that was recently introduced in the European Drinking Water Directive. Some challenges remain, in particular those related to selection and interpretation of bioassays. For water quality assessment, carcinogenesis, adverse effects on reproduction and development, effects on xenobiotic metabolism, modulation of hormone systems, DNA reactivity, and adaptive stress responses are considered the most relevant toxicological endpoints. An evaluation procedure of the applicability and performance of in vitro bioassays for water quality monitoring, based on existing information, has been developed, which can be expanded with guidelines for experimental evaluations. In addition, a methodology for the interpretation of in vitro monitoring data is required, because the sensitivity of specific in vitro bioassays in combination with sample concentration may lead to responses of chemicals (far) below exposure concentrations that are relevant for human health effects. Different approaches are proposed to derive effect-based trigger values (EBTs), including EBTs based on (1) relative ecotoxicity potency, (2) health-based threshold values for chronic exposure in humans and kinetics of reference chemicals, and (3) read-across from (drinking) water guideline values. Effects-based trigger values need to be chosen carefully in order to be sufficiently but not overly conservative to indicate potential health effects. Consensus on the crucial steps in the selection and interpretation of in vitro bioassay data will facilitate implementation and legal embedding in the context of water quality monitoring of such assays in EBM strategies. Integr Environ Assess Manag 2019;15:126-134. © 2018 The Authors. Integrated Environmental Assessment and Management published by Wiley Periodicals, Inc. on behalf of Society of Environmental Toxicology & Chemistry (SETAC).

Chemical and bioassay assessment of waters related to hydraulic fracturing at a tight gas production site.

Publicly available chemical assessments of hydraulic fracturing related waters are generally based on shale gas practices in the U.S. There is a lack of information on hydraulic fracturing related gas development from EU countries and more generally on other types of extractions. This research fills this knowledge gap by presenting chemical and bioassay assessments of hydraulic fracturing related waters from a tight gas development in the Netherlands. Fracturing fluid, flowback water and groundwater from surrounding aquifers before and after the actual fracturing were analysed by means of high resolution liquid chromatography tandem mass spectrometry, the Ames test and three chemical activated luciferase gene expression bioassays aimed at determining genotoxicity, oxidative stress response and polyaromatic hydrocarbon contamination. After sample enrichment a higher number of peaks can be found in both fracturing fluid and flowback samples. No clear differences in chemical composition were shown in the groundwater samples before and after hydraulic fracturing. Preliminary environmental fate data of the tentatively identified chemicals points towards persistence in water. Clear genotoxic and oxidative stress responses were found in the fracturing fluid and flowback samples. A preliminary suspect screening resulted in 25 and 36 matches in positive and negative ionisation respectively with the 338 possible suspect candidates on the list. Extensive measures relating to the handling, transport and treatment of hydraulic fracturing related waters are currently in place within the Dutch context. The results of the present study provide a scientific justification for such measures taken to avoid adverse environmental and human health impacts.

Toxicological risk assessment and prioritization of drinking water relevant contaminants of emerging concern.

Toxicological risk assessment of contaminants of emerging concern (CEC) in (sources of) drinking water is required to identify potential health risks and prioritize chemicals for abatement or monitoring. In such assessments, concentrations of chemicals in drinking water or sources are compared to either (i) health-based (statutory) drinking water guideline values, (ii) provisional guideline values based on recent toxicity data in absence of drinking water guidelines, or (iii) generic drinking water target values in absence of toxicity data. Here, we performed a toxicological risk assessment for 163 CEC that were selected as relevant for drinking water. This relevance was based on their presence in drinking water and/or groundwater and surface water sources in downstream parts of the Rhine and Meuse, in combination with concentration levels and physicochemical properties. Statutory and provisional drinking water guideline values could be derived from publically available toxicological information for 142 of the CEC. Based on measured concentrations it was concluded that the majority of substances do not occur in concentrations which individually pose an appreciable human health risk. A health concern could however not be excluded for vinylchloride, trichloroethene, bromodichloromethane, aniline, phenol, 2-chlorobenzenamine, mevinphos, 1,4-dioxane, and nitrolotriacetic acid. For part of the selected substances, toxicological risk assessment for drinking water could not be performed since either toxicity data (hazard) or drinking water concentrations (exposure) were lacking. In absence of toxicity data, the Threshold of Toxicological Concern (TTC) approach can be applied for screening level risk assessment. The toxicological information on the selected substances was used to evaluate whether drinking water target values based on existing TTC levels are sufficiently protective for drinking water relevant CEC. Generic drinking water target levels of 37 µg/L for Cramer class I substances and 4 µg/L for Cramer class III substances in drinking water were derived based on these CEC. These levels are in line with previously reported generic drinking water target levels based on original TTC values and are shown to be protective for health effects of the majority of contaminants of emerging concern evaluated in the present study. Since the human health impact of many chemicals appearing in the water cycle has been studied insufficiently, generic drinking water target levels are useful for early warning and prioritization of CEC with unknown toxicity in drinking water and its sources for future monitoring.

Exploration of ToxCast/Tox21 bioassays as candidate bioanalytical tools for measuring groups of chemicals in water.

The present study explores the ToxCast/Tox21 database to select candidate bioassays as bioanalytical tools for measuring groups of chemicals in water. To this aim, the ToxCast/Tox21 database was explored for bioassays that detect polycyclic aromatic hydrocarbons (PAHs), aromatic amines (AAs), (chloro)phenols ((C)Ps) and halogenated aliphatic hydrocarbons (HAliHs), which are included in the European and/or Dutch Drinking Water Directives. Based on the analysis of the availability and performance of bioassays included in the database, we concluded that several bioassays are suitable as bioanalytical tools for assessing the presence of PAHs and (C)Ps in drinking water sources. No bioassays were identified for AAs and HAliHs, due to the limited activity of these chemicals and/or the limited amount of data on these chemicals in the database. A series of bioassays was selected that measure molecular or cellular effects that are covered by bioassays currently in use for chemical water quality monitoring. Interestingly, also bioassays were selected that represent molecular or cellular effects that are not covered by bioassays currently applied. The usefulness of these newly identified bioassays as bioanalytical tools should be further evaluated in follow-up studies. Altogether, this study shows how exploration of the ToxCast/Tox21 database provides a series of candidate bioassays as bioanalytical tools for measuring groups of chemicals in water. This assessment can be performed for any group of chemicals of interest (if represented in the database), and may provide candidate bioassays that can be used to complement the currently applied bioassays for chemical water quality assessment.

In vitro immunotoxicity of bis(tri-n-butyltin)oxide (TBTO) studied by toxicogenomics.

The biocide and environmental pollutant bis(tri-n-butyltin)oxide (TBTO) causes thymus atrophy in rodents. Whether the depletion of thymic lymphocytes by tributyltin compounds may be the result of inhibition of cell proliferation or induction of apoptosis is subject of debate. We examined gene expression profiles in primary rat thymocytes exposed to TBTO in vitro at dose levels of 0, 0.1, 0.3, 0.5, and 1.0microM. By measuring cell viability and apoptosis, exposure conditions were selected that would provide information on changes in gene expression preceding or accompanying functional effects of TBTO. Several processes related to TBTO-induced toxicity were detected at the transcriptome level. Effects on lipid metabolisms appeared to be the first indication of disruption of cellular function. Many transcriptional effects of TBTO at higher dose levels were related to apoptotic processes, which corresponded to present or subsequent thymocyte apoptosis observed phenotypically. The gene expression profile was, however, not unambiguous since expression of apoptosis-related genes was both increased and decreased. Stimulation of glucocorticoid receptor signaling appeared to be a relevant underlying mechanism of action. These findings suggest that TBTO exerts its toxic effects on the thymus primarily by affecting apoptotic processes, but the possibility is discussed that this may in fact represent an early effect that precedes inhibition of cell proliferation. At the highest dose level tested, TBTO additionally repressed mitochondrial function and immune cell activation. Our in vitro toxicogenomics approach thus identified several cellular and molecular targets of TBTO that may mediate the toxicity towards thymocytes and thereby its immunosuppressive effects.

Evaluation of immunomodulation by Lactobacillus casei Shirota: immune function, autoimmunity and gene expression.

Lactic acid bacteria are claimed to have immunomodulating effects. Stimulation as well as suppression of T helper (Th)1 mediated immune responses, have been described for various strains. Experiments involving Lactobacillus casei Shirota (LcS) detected mainly enhancement of innate immune responses and promotion of Th1 mediated immune reactivity. To confirm and further investigate modulation of Th1 responses and development of autoimmune disease by LcS, the consequences of oral administration of LcS were assessed in several experiments. The effect of LcS varied between the different models. No modulation was found in the mitogen-induced cell proliferation and cytokine release assays in mesenteric lymph nodes of Wistar rats. LcS inhibited the Th1 mediated immune response in an adapted murine Local Lymph Node Assay (LLNA) in BALB/c mice, whereas experimental autoimmune encephalomyelitis (EAE) in Lewis rats was aggravated. These varying effects on Th1 responses indicate that beneficial as well as harmful effects on immune related disorders could occur after LcS consumption. Since microarray analysis is suggested to be more sensitive and predictive than functional tests, gene expression profiling was included as an alternative endpoint in the testing of immunomodulation. The detected gene expression profiles did not reflect the effects of LcS on the immune system. Microarray analysis may therefore have no more predictive value than immune function assays when investigating immunomodulation by probiotics. To gain further insight into effects of probiotics on immune function, experiments including cytokine assays and gene expression analysis combined with disease models could be useful.

Use of the local lymph node assay in assessment of immune function.

The murine local lymph node assay (LLNA) was originally developed as a predictive test method for the identification of chemicals with sensitizing potential. In this study we demonstrated that an adapted LLNA can also be used as an immune function assay by studying the effects of orally administered immunomodulating compounds on the T-cell-dependent immune response induced by the contact sensitizer 2,4-dinitrochlorobenzene (DNCB). C57Bl/6 mice were treated with the immunotoxic compounds cyclosporin A (CsA), bis(tri-n-butyltin)oxide (TBTO) or benzo[a]pyrene, (B[a]P). Subsequently, cell proliferation and interferon-gamma (IFN-gamma) and interleukin (IL)-4 release were determined in the auricular lymph nodes (LNs) after DNCB application on both ears. Immunosuppression induced by CsA, TBTO and B[a]P was clearly detectable in this application of the LLNA. Cytokine release measurements proved valuable to confirm the results of the cell proliferation assay and to obtain an indication of the effect on Th1/Th2 balance. We believe to have demonstrated the applicability of an adapted LLNA as an immune function assay in the mouse.

Influence of process conditions and water quality on the formation of mutagenic byproducts in UV/H2O2 processes.

UV/H2O2 processes in drinking water treatment may generate byproducts which cause an increased response in Ames fluctuation assays. As this probably involves a mixture of substances in very low concentrations, it is challenging to identify the individual byproducts. Therefore it was studied under which conditions mutagenic byproducts are formed and how this can be prevented. It was found that positive Ames fluctuation test responses only are obtained when Medium Pressure UV lamps are used, and not with Low Pressure lamps. This probably is explained by the photolysis of nitrate, which plays an important role in the formation of mutagenic byproducts. The most important parameters involved in the formation of such byproducts were demonstrated to be the nitrate concentration, the natural organic matter, the UV spectrum of the lamps, and the UV dose applied. These factors explain up to 74-87% of the Ames fluctuation test responses after UV/H2O2 drinking water treatment. By taking this into account, drinking water utilities can estimate whether UV processes applied in their case may cause the formation of mutagenic byproducts, and how to take measures to prevent it.

Keratinocyte gene expression profiles discriminate sensitizing and irritating compounds.

Many chemicals can induce allergic contact dermatitis. Because evaluation of skin sensitizing potential by animal testing is prohibited for cosmetics, and screening of many chemicals is required within Registration, Evaluation, Authorisation and Restriction of Chemicals, urgent need exists for predictive in vitro assays to identify contact allergens. Keratinocytes (KC) are the first cells encountered when chemicals land on the skin. Therefore, KC form an important site of haptenization and their metabolism is likely to be important. Moreover, KC secrete mediators that affect processing and presentation of haptenized proteins by dendritic cells. To develop a KC-based in vitro assay to predict sensitizing potential of chemicals, in vitro exposure effects of eight contact sensitizers and six irritants on the KC cell line HaCaT were examined by gene profiling. Classifiers predictive of the class sensitizers or irritants were calculated, based on support vector machine (SVM) and random forest (RF) algorithms. Classifiers using high-ranking genes were 70% (SVM) and 62% (RF) accurate, based on three (SVM) and two to five (RF) features. Classifiers using oxidative stress pathway gene sets were 68-73% (SVM) and 69-71% (RF) accurate. Cross-validation showed that the top-3 of most discriminating genes added up to 13 genes and included oxidative stress gene HMOX1 irrespective of the chemical left out. Moreover, HMOX1 was the most significantly regulated gene. Gene Set Enrichment Analysis showed upregulation of "Keap1 dependent" and "oxidative stress" gene lists. In conclusion, KC expression profiling can identify contact sensitizers, providing opportunities for nonanimal testing for sensitizing potential. Moreover, our data suggest that contact sensitizers induce the oxidative stress pathway in KC.

Gene expression changes in the mesenteric lymph nodes of rats after oral peanut extract exposure.

New techniques are needed to broaden the understanding of the food allergic response. The capacity of peanut extract to influence gene expression profiles was investigated in a Brown Norway rat model for food allergy. Brown Norway rats were sensitized to peanut extract (0, 1 and 10 mg/rat/d) by daily oral gavage and were dissected after 3, 7 or 14 days of exposure. RNA extracted from mesenteric lymph nodes of individual rats were hybridized against a common reference pool on Agilent whole rat genome (4*44k) arrays. The raw data were normalized and statistically analyzed using the statistical program R. A False Discovery Rate of 10% and a Fold Ratio of - 1.5 < or = Fold Ratio or Fold Ratio > or = 1.5 between the experimental groups and their respective control groups were applied. Differentially expressed genes were clustered into a heatmap. Functional annotation and GeneOntology term enrichment were examined. Furthermore, the involvement of the differentially expressed genes in specific cellular pathways was investigated with MetaCore. Gene expression changes, which were both dose- and time-dependent, were detected after sensitization to peanut. A total of 64 genes were differentially expressed, of which 60 were up-regulated and four were down-regulated. These changes were related to the regulation of immunological processes, most notably increased cell division. The findings indicate that responses to peanut include proliferation of immunologically relevant tissues, which can be identified by analysis of gene expression profiles. This may lay a basis for further research into possibilities for discrimination of allergenic from non-allergenic proteins.

Overlapping gene expression profiles of model compounds provide opportunities for immunotoxicity screening.

In order to investigate immunotoxic effects of a set of model compounds in mice, a toxicogenomics approach was combined with information on macroscopical and histopathological effects on spleens and on modulation of immune function. Bis(tri-n-butyltin)oxide (TBTO), cyclosporin A (CsA), and benzo[a]pyrene (B[a]P) were administered to C57BL/6 mice at immunosuppressive dose levels. Acetaminophen (APAP) was included in the study since indications of immunomodulating properties of this compound have appeared in the literature. TBTO exposure caused the most pronounced effect on gene expression and also resulted in the most severe reduction of body weight gain and induction of splenic irregularities. All compounds caused inhibition of cell division in the spleen as shown by microarray analysis as well as by suppression of lymphocyte proliferation after application of a contact sensitizer as demonstrated in an immune function assay that was adapted from the local lymph node assay. The immunotoxicogenomics approach applied in this study thus pointed to immunosuppression through cell cycle arrest as a common mechanism of action of immunotoxicants, including APAP. Genes related to cell division such as Ccna2, Brca1, Birc5, Incenp, and Cdkn1a (p21) were identified as candidate genes to indicate anti-proliferative effects of xenobiotics in immune cells for future screening assays. The results of our experiments also show the value of group wise pathway analysis for detection of more subtle transcriptional effects and the potency of evaluation of effects in the spleen to demonstrate immunotoxicity.

Toxicogenomics in the assessment of immunotoxicity.

Microarray analysis is used for simultaneous measurement of expression of thousands of genes in a given sample and as such extends and deepens our understanding of biological processes. Application of the technique in toxicology is referred to as toxicogenomics. The examples of assessment of immunotoxicity by gene expression profiling presented and discussed here, show that microarray analysis is able to detect known and novel effects of a wide range of immunomodulating agents. Besides the elucidation of mechanisms of action, toxicogenomics is also applied to predict consequences of exposing biological systems to toxic agents. Successful attempts to classify compounds using signature gene expression profiles have been reported. These did, however, not specifically focus on immunotoxicity. Databases containing expression profiles can facilitate the applications of toxicogenomics. Platforms and methodologies for gene expression profiling may vary, however, hampering data compiling across different laboratories. Therefore, attention is paid to standardization of the generation, reporting, and management of microarray data. Obtained gene expression profiles should be anchored to pathological and functional endpoints for correct interpretation of results. These issues are also important when using toxicogenomics in risk assessment. The application of toxicogenomics in evaluation of immunotoxicity is thus not yet without challenges. It already contributes to the understanding of immunotoxic processes and the development of in vitro screening assays, though, and is therefore expected to be of value for mechanistic insight into immunotoxicity and hazard identification of existing and novel compounds.

Gene expression profiling of Bis(tri-n-butyltin)oxide (TBTO)-induced immunotoxicity in mice and rats.

Bis(tri-n-butyltin)oxide (TBTO) is one of the organotin compounds that have been used as biocides and occur as persistent environmental pollutants. Human exposure to these compounds occurs through consumption of meat and fish products in which they accumulate. The most sensitive endpoint of TBTO exposure is immunotoxicity. TBTO causes thymus atrophy and thereby interferes with T-lymphocyte-mediated immune responses. Tributyltin compounds have been found to adversely affect a wide range of cellular components and processes in many species, organ systems, and cell types. Both inhibition of proliferation and induction of apoptosis have been observed in thymocytes. We conducted microarray experiments in mice and rats in order to investigate if the immunosuppressive actions of TBTO could be detected by gene expression profiling, and if so, to elucidate the mechanisms of action. Gene expression changes that were detected in mouse thymuses after exposure to a maximum tolerable dose of TBTO correlated to previously observed effects. Most notably, reduction of expression of cell surface determinants and T-cell receptor chains, suppression of cell proliferation, and a possible involvement of nuclear receptors in interference with lipid metabolism by TBTO were observed. The TBTO-induced thymus involution may therefore primarily be caused by inhibition of thymocyte proliferation. In contrast, in rats only limited effects of a lower dose of TBTO were found at the gene expression level in the thymus, even though thymus involution was observed. Here, most gene expression regulation by TBTO was detected in the liver. These preliminary results indicate that gene expression analysis is able to reveal effects of TBTO and to gain insight into its molecular mechanism of action. It may even be a suitable tool to investigate immunotoxicology in general. However, dose and inter-species differences are apparently clearly reflected in the gene expression profiles.