Bas J. Blaauboer

A roadmap for the development of alternative (non-animal) methods for systemic toxicity testing - t4 report

Systemic toxicity testing forms the cornerstone for the safety evaluation of substances. Pressures to move from traditional animal models to novel technologies arise from various concerns, including: the need to evaluate large numbers of previously untested chemicals and new products (such as nanoparticles or cell therapies), the limited predictivity of traditional tests for human health effects, duration and costs of current approaches, and animal welfare considerations. The latter holds especially true in the context of the scheduled 2013 marketing ban on cosmetic ingredients tested for systemic toxicity. Based on a major analysis of the status of alternative methods (Adler et al., 2011) and its independent review (Hartung et al., 2011), the present report proposes a roadmap for how to overcome the acknowledged scientific gaps for the full replacement of systemic toxicity testing using animals. Five whitepapers were commissioned addressing toxicokinetics, skin sensitization, repeated-dose toxicity, carcinogenicity, and reproductive toxicity testing. An expert workshop of 35 participants from Europe and the US discussed and refined these whitepapers, which were subsequently compiled to form the present report. By prioritizing the many options to move the field forward, the expert group hopes to advance regulatory science.

Methods of in vitro toxicology

In vitro methods are common and widely used for screening and ranking chemicals, and have also been taken into account sporadically for risk assessment purposes in the case of food additives. However, the range of food-associated compounds amenable to in vitro toxicology is considered much broader, comprising not only natural ingredients, including those from food preparation, but also compounds formed endogenously after exposure, permissible/authorised chemicals including additives, residues, supplements, chemicals from processing and packaging and contaminants. A major promise of in vitro systems is to obtain mechanism-derived information that is considered pivotal for adequate risk assessment. This paper critically reviews the entire process of risk assessment by in vitro toxicology, encompassing ongoing and future developments, with major emphasis on cytotoxicity, cellular responses, toxicokinetics, modelling, metabolism, cancer-related endpoints, developmental toxicity, prediction of allergenicity, and finally, development and application of biomarkers. It describes in depth the use of in vitro methods in strategies for characterising and predicting hazards to the human. Major weaknesses and strengths of these assay systems are addressed, together with some key issues concerning major research priorities to improve hazard identification and characterisation of food-associated chemicals.

Interlaboratory comparison of total cytochrome P-450 and protein determinations in rat liver microsomes

Assay conditions in determining total cytochrome P-450 in four laboratories were compared. Although the determination was derived from the original Omura and Sato method in each laboratory, the four standard protocols differed slightly, resulting in considerable differences in the results. Since the cytochrome P-450 content is usually expressed per mg protein, the protein assay conditions were evaluated as well. Furthermore, we compared the cytochrome P-450 values obtained by the CO- and the dithionite (DT)-difference methods. The effect of a number of variables in the assay was investigated. The influence of the storage temperature of the microsomes was ascertained as well as effects of the gassing time with CO and the time between addition of dithionite, CO-gassing and the recording of the difference spectra. After evaluating these variables a standard operation procedure was established. Using this procedure the interlaboratory coefficient of variation for total cytochrome P-450 was 4.8%, a value which was comparable to the intralaboratory coefficients of variation. The final results also show that the millimolar extinction coefficient for the DT-difference method is higher than for the CO-difference method.

The isoenzyme pattern of cytochrome P450 in rat hepatocytes in primary culture, comparing different enzyme activities in microsomal incubations and in intact monolayers

Changes in the isoenzyme pattern of cytochrome P450 during culture were investigated in primary cultures of rat hepatocytes, measuring specific enzyme activities in microsomes prepared from cultured cells as well as in intact monolayers. Assays of 7-ethoxyresorufin O-deethylation (EROD), 7-pentoxyresorufin O-depentylation (PROD), aniline 4-hydroxylation (AH) and the specific regioselective hydroxylation of testosterone were used as representatives of the activities of seven isoenzymes of cytochrome P450. The isoenzyme profile expressed as catalytic activities was qualitatively and quantitatively similar in microsomes obtained from freshly isolated hepatocytes in comparison with microsomes obtained from whole livers of untreated rats. There was a relatively high activity in EROD, AH and the oxidation of testosterone at the 7 alpha, 2 alpha, 6 beta, 16 alpha and 17 sites (androstenedione). During culture, these microsomal enzyme activities declined at a similar rate to ca. 50% of the activities of microsomes prepared from freshly isolated hepatocytes after 24 hr and to 15% after 96 hr. The overall decline of cytochrome P450-dependent activities during culture was not accompanied with gross changes in catalytic profile. Determining the same drug-metabolizing activities directly in intact hepatocyte monolayers revealed a much higher metabolic rate for all measured P450-dependent activities. The profile of the catalytic activities was essentially the same as measured in microsomes prepared from cultured hepatocytes. The relatively low activity towards the 7 alpha site of testosterone measured in intact hepatocytes, however, remained constant during culture. Determination of enzyme activities directly in intact hepatocytes is a convenient way of studying changes in monooxygenase activities of different P450 isoenzymes in vitro.

Comparison of cytochrome P450 isoenzyme profiles in rat liver and hepatocyte cultures. The effects of model inducers on apoproteins and biotransformation activities.

The metabolic profile of seven subfamilies of cytochrome P450 (P450IA, IIA, IIB, IIC, IIE, IIIA, IVA) was studied in rat liver (in vivo) and in primary hepatocyte cultures (in vitro) after treatment with various inducers. The dealkylation of 7-ethoxyresorufin (EROD) and 7-pentoxyresorufin (PROD), aniline 4-hydroxylation and the regio- and stereoselective hydroxylation of testosterone were measured to characterize the isoenzyme pattern in intact hepatocytes and in liver microsomes. Occurrence of isoenzyme apoproteins was determined using Western blotting. Primary cultures of rat hepatocytes retain the capacity to respond to inducers of isoenzymes belonging to six different subfamilies (P450IA, IIA, IIB, IIC, IIIA and IVA). Treatment of cells with beta-naphthoflavone revealed a P450-activity profile similar to in vivo, namely a highly induced EROD (P450IA1), a small enhancement of testosterone 7 alpha-hydroxylation (P450IIA) and a marked reduction in 2 alpha- and 16 alpha-hydroxylation (P450IIC11). Exposure of cultured cells to phenobarbital resulted in a higher testosterone 16 beta-hydroxylation (reflecting P450IIB), though to a lesser extent than in vivo. The induction of P450IIIA due to both phenobarbital and dexamethasone, as mirrored by 6 beta- and 15 beta-hydroxylation of testosterone, was the same in cultured hepatocytes and in vivo. Treatment of cells with clofibric acid resulted in an induction profile similar to the one observed in liver microsomes from clofibrate-treated rats: the apoprotein P450IVA as well as the apoprotein P450IIB1/2 and its associated activities (PROD and testosterone 16 beta-hydroxylation) were induced. Isoniazid, a known in vivo inducer of P450IIE1 and aniline 4-hydroxylation, did not change any of the determined P450-dependent activities in vitro.

Biokinetic Modeling and in Vitro–in Vivo Extrapolations

The introduction of in vitro methodologies in the toxicological risk assessment process requires a number of prerequisites regarding both the toxicodynamics and the biokinetics of the compounds under study. In vitro systems will need to be relevant for measuring those structural and physiological changes that are good indicators for adverse effects. Furthermore, the dose metric found to have an effect in the in vitro system should be relevant. One element in defining the appropriate dose metric is related to the kinetic behavior of the compound in the in vitro system: binding to proteins, binding to plastic, evaporation, and the interaction between the culture medium and the cells. Ways to measure and model “in vitro biokinetics” are described. Second, the appropriate dose metric in vitro, e.g., the effective concentration, will need to be extrapolated to relevant in vivo exposure scenarios. The application of physiologically based biokinetic modelling is essential in such extrapolations. The parameters needed to build these models often can be estimated based on nonanimal data, namely chemical properties (QSARs) and in vitro experiments.

Application of integrated transcriptomic, proteomic and metabolomic profiling for the delineation of mechanisms of drug induced cell stress.

High content omic techniques in combination with stable human in vitro cell culture systems have the potential to improve on current pre-clinical safety regimes by providing detailed mechanistic information of altered cellular processes. Here we investigated the added benefit of integrating transcriptomics, proteomics and metabolomics together with pharmacokinetics for drug testing regimes. Cultured human renal epithelial cells (RPTEC/TERT1) were exposed to the nephrotoxin Cyclosporine A (CsA) at therapeutic and supratherapeutic concentrations for 14days. CsA was quantified in supernatants and cellular lysates by LC-MS/MS for kinetic modeling. There was a rapid cellular uptake and accumulation of CsA, with a non-linear relationship between intracellular and applied concentrations. CsA at 15μM induced mitochondrial disturbances and activation of the Nrf2-oxidative-damage and the unfolded protein-response pathways. All three omic streams provided complementary information, especially pertaining to Nrf2 and ATF4 activation. No stress induction was detected with 5μM CsA; however, both concentrations resulted in a maximal secretion of cyclophilin B. The study demonstrates for the first time that CsA-induced stress is not directly linked to its primary pharmacology. In addition we demonstrate the power of integrated omics for the elucidation of signaling cascades brought about by compound induced cell stress.

The effect of beclobric acid and clofibric acid on peroxisomal β-oxidation and peroxisome proliferation in primary cultures of rat, monkey and human hepatocytes

The peroxisome-proliferating effects of clofibric acid and beclobric acid were studied in primary cultures of hepatocytes derived from rat, monkey (Macaca fascicularis) and human liver. Determination of peroxisomal fatty acid beta-oxidation and morphometrical analysis of the peroxisomal compartment were performed after incubation of 1-day-old hepatocyte cultures for 3 days with either compound. In rat liver cell cultures both compounds gave a 10-fold increase in peroxisomal beta-oxidation, a 3-fold increase in the relative number of peroxisomes and a 1.5-fold increase in the mean size of peroxisomes. Beclobric acid gave its maximal effect at a concentration of 10 microM, which is at least one order of magnitude lower than the maximum-effect concentration of clofibric acid. At concentrations greater than 300 microM beclobric acid was cytotoxic. No stimulation of peroxisomal fatty acid beta-oxidation was found in either monkey or human hepatocyte cultures. Morphometrical analysis also showed no increase in the peroxisomal compartment in cultures derived from these species, as indicated by the lack of increase in both relative number and size of peroxisomes. In all three species tested beclobric acid was equally cytotoxic for hepatocytes in vitro. These results are of relevance for the interpretation of the peroxisome-proliferating effects of clofibrate and similar compounds in rats. Since peroxisome proliferation may be correlated to increased hepatic tumour incidences in the rat, the absence of peroxisome proliferation in primates suggests the absence of tumourogenic activity by hypolipidemic compounds in these species.

The use of in vitro toxicity data and physiologically based kinetic modeling to predict dose-response curves for in vivo developmental toxicity of glycol ethers in rat and man.

At present, regulatory assessment of systemic toxicity is almost solely carried out using animal models. The European Commissions REACH legislation stimulates the use of animal-free approaches to obtain information on the toxicity of chemicals. In vitro toxicity tests provide in vitro concentration-response curves for specific target cells, whereas in vivo dose-response curves are regularly used for human risk assessment. The present study shows an approach to predict in vivo dose-response curves for developmental toxicity by combining in vitro toxicity data and in silico kinetic modeling. A physiologically based kinetic (PBK) model was developed, describing the kinetics of four glycol ethers and their embryotoxic alkoxyacetic acid metabolites in rat and man. In vitro toxicity data of these metabolites derived in the embryonic stem cell test were used as input in the PBK model to extrapolate in vitro concentration-response curves to predicted in vivo dose-response curves for developmental toxicity of the parent glycol ethers in rat and man. The predicted dose-response curves for rat were found to be in concordance with the embryotoxic dose levels measured in reported in vivo rat studies. Therefore, predicted dose-response curves for rat could be used to set a point of departure for deriving safe exposure limits in human risk assessment. Combining the in vitro toxicity data with a human PBK model allows the prediction of dose-response curves for human developmental toxicity. This approach could therefore provide a means to reduce the need for animal testing in human risk assessment practices.

Dose metric considerations in in vitro assays to improve quantitative in vitro-in vivo dose extrapolations

Challenges to improve toxicological risk assessment to meet the demands of the EU chemicals legislation, REACH, and the EU 7th Amendment of the Cosmetics Directive have accelerated the development of non-animal based methods. Unfortunately, uncertainties remain surrounding the power of alternative methods such as in vitro assays to predict in vivo dose-response relationships, which impedes their use in regulatory toxicology. One issue reviewed here, is the lack of a well-defined dose metric for use in concentration-effect relationships obtained from in vitro cell assays. Traditionally, the nominal concentration has been used to define in vitro concentration-effect relationships. However, chemicals may differentially and non-specifically bind to medium constituents, well plate plastic and cells. They may also evaporate, degrade or be metabolized over the exposure period at different rates. Studies have shown that these processes may reduce the bioavailable and biologically effective dose of test chemicals in in vitro assays to levels far below their nominal concentration. This subsequently hampers the interpretation of in vitro data to predict and compare the true toxic potency of test chemicals. Therefore, this review discusses a number of dose metrics and their dependency on in vitro assay setup. Recommendations are given on when to consider alternative dose metrics instead of nominal concentrations, in order to reduce effect concentration variability between in vitro assays and between in vitro and in vivo assays in toxicology.