Application of Medium and High Throughput Screening for In Vitro Toxicology in the Pharmaceutical Industry

Home / New Perspectives / Emerging Technologies / Application of Medium and High Throughput Screening for In Vitro Toxicology in the Pharmaceutical Industry

Emerging Technologies

Application of Medium and High Throughput Screening for In Vitro Toxicology in the Pharmaceutical Industry

By Willem G.E.J. Schoonen, Walter M.A. Westerink & G. Jean Horbach Merck Sharp & Dohme Corp. (Merck & Co., Inc.) Published: October 28, 2010
About the Author(s)
Willem Schoonen studied biology at the University of Utrecht, where he graduated for his PhD in 1987 on steroid pheromones in fish. From 1988-1990 he worked at the Free University of Amsterdam on hyperoxic cellular toxicity. From 1990-2007 he worked at N.V. Organon, which merged in 2008 into Schering-Plough and in 2010 into MSD. He worked in the field of nuclear receptors, endocrinology, pharmacology and in vitro toxicology and toxicogenomics. He is group leader of In Vitro Toxicology and Toxicogenomics within the department of Toxicology and Drug Disposition.

Willem Schoonen
Department of Toxicology and Drug Disposition Merck Sharp & Dohme
P.O. Box 20 5340 BH Oss The Netherlands

Walter Westerink studied medical biochemistry at the Hogeschool Enschede and acquired his master degree in molecular life sciences in 2007. From 1997-2007 he was employee at N.V. Organon, which merged in 2008 into Schering-Plough and in 2010 into MSD. He worked in the field of inflammation, in vitro toxicology and toxicogenomics. He is research scientist in the group of In Vitro Toxicology and Toxicogenomics within the department of Toxicology and Drug Disposition.

Walter Westerink
Department of Toxicology and Drug Disposition Merck Sharp & Dohme
P.O. Box 20 5340 BH Oss The Netherlands

G. Jean Horbach studied biochemistry at the University of Leiden and did his PhD at the University of Maastricht, the Netherlands. From 1996-2007 he was employee at N.V. Organon, which merged in 2008 into Schering-Plough and in 2010 into MSD. He worked in the field of drug metabolism and genotoxicity. He is section head of the Mechanistic and Investigative Toxicology within the department of Toxicology and Drug Disposition.

G. Jean Horbach
Department of Toxicology and Drug Disposition Merck Sharp & Dohme
P.O. Box 20 5340 BH Oss The Netherlands

The attrition rate in the pharmaceutical industry due to toxicity in preclinical and clinical studies is between 30 to 40% (Kola & Landis, 2004; Guengerich & MacDonald, 2007). This implies that within the current strategy of predicting (pre-)clinical safety the challenge remains to identify (pre-)clinical safety liabilities early in the drug development process, which will lead to the better selection of drug candidates. The process of drug development is depicted in Figure 1.

Figure 1. The process of drug development within a pharmaceutical company, from target to hit, to lead, to development candidate (DC), statement of no objection (SNOB), final development candidate (FDC) and product.
An optimal strategy could be to study toxicity early during the lead discovery and lead optimisation phase with an in vitro screening battery covering a number of critical failure factors e.g. cytotoxicity, genotoxicity, non-genotoxic carcinogenicity, carcinogenicity, nuclear receptor activation, inflammation, enzyme induction, enzyme competition, enzyme polymorphisms and drug metabolism.

Starting at these early phases within the development process supposes the following requirements in the test battery:

  1. the amounts of compound needed should be low, 10 to 20 mg.
  2. the turn-around time for these assays should be short, between 0.5 – 5 days.
  3. the throughput of these assays should be 40 up to 80 compounds per assay.
  4. these assays should identify organ, cell and species specific differences.

It is expected that the implementation of such a complete test battery with parallel screening will considerably reduce the number of animal studies and their associated costs conducted on development candidates that are going to fail later in the drug development process. In addition, the attrition rate of newly developed drugs may be reduced in the final pre-clinical and clinical research phases.

Current toxicological research is focussed on in vivo studies, in which mouse, rat, rabbit, monkey and dog are the most commonly used animal models. Based on the high attrition rate of 30 to 40% in clinical studies, the prediction of in vivo toxicity in animal models is still not adequate enough for the solid and correct prediction of human toxicity. Nevertheless the validation of new in vitro models is mainly based on the knowledge of these ‘semi-predictive’ in vivo animal models. Therefore, if the goal of the development of novel predictive in vitro tools is the prediction of human toxicity, an emphasis will need to be put on clinical data as a reference point and not on pre-clinical data. To support these statements some examples will be given in this paper.

For the measurement of in vitro and in vivo genotoxicity several assay methods are commonly used. Very well known are the Ames mutagenicity test using Salmonella typhimurium (Ames, et al., 1973; McCann, et al., 1975) and the in vitro clastogenicity tests with Chinese hamster ovary and lung cells, as well as in vitro, ex vivo and in vivo clastogenicity tests in both rats and humans with lymphoblasts and/or erythrocytes (Miller, et al., 1998; McGregor, et al., 2000). Also for these assays a similar concern for the proper predictions among tests and across species exists.

First, there is always the fear for too many false positive and false negative scores, which is more or less a numbers game. Scientists within drug discovery want to discriminate the true positives from the true negatives as early as possible in a set of compounds from a particular chemical series. Therefore the true positive and true negative scores must be as high as possible. On the other hand, the chance that a complete series of structurally related compounds will show positive results in several of these genotoxicity tests can not be called coincidence. Neither can the chance that one of these compounds is positive in many tests with the same biological paradigm be called coincidence.

Next, of course, you may also have species specific tests, in which only one test might give a positive result. For in vitro clastogenicity assays, the large number of false positives with Chinese hamster ovary and/or lung cells appear to be due to a specific p53 mutation (Chaung, et al., 1997; Hu, et al, 1999; Tzang, et al., 1999; Kirkland, et al., 2007). In literature it has been shown, that a wild-type p53 gene is essential for the initiation of repair of DNA damage. Therefore, a mutation in such a gene will definitely lead to overestimation of the incidence on DNA damage. Similarly, humans and mice with particular gene mutations in DNA repair genes, like Xerodermata pigmentosa gene variants A and C, RAD51 genes and analogues thereof, as well as of Fanconi anemia, are more vulnerable to externally induced changes in the DNA, resulting in higher incidence of tumors.

On the other hand, if compounds are used in tests in which cells or animals contain these specific mutations, it does give you the wrong prediction on the potency or severity of genotoxicity of a compound, but the right prediction on the potency or severity of the particular repair mechanism for genotoxicity. Always the main question should be whether the correct reference points are used for obtaining the most objective answers.

Second, species differences may be responsible for incorrect human risk prediction as the metabolic activity or capacity may differ in Salmonella, yeast, hamsters, rats and humans. Moreover, in all in vitro assays with cells of these species, a cytochrome P450 enzyme enriched liver homogenate of rats is used for the introduction of metabolic activity. This enrichment of enzymes can also lead to an overestimation of the risk, since in humans the metabolic enzyme capacity may be lower or different from that of rats. The species specific differences in phase I, II and III metabolism may lead to a different kind of metabolites and thus species specific toxic liabilities. Several examples exist where compounds were shown to be completely safe in rats and dogs, but showed specific human toxicities (Smith and Schmid, 2006), like mitochondrial disorders (Finsterer & Segall, 2010), myopathies (mitochondrial damage: clofibrate, fenofibrate (Brunmair, et al., 2004), cardiomyopathies (mitochondrial damage: doxorubicin (Olson, et al, 1988; Mordente, et al., 2003), mitochondrial liver toxicities by reverse transcriptase inhibitors, like zalcitabine, stavudine and zidovudine (Spengler, et al., 2002), hepatotoxicities, like liver cholestasis by anabolic androgens, estrogens, and chlorpromazine (Accatino, et al., 1995, 1996; Frezza, et al., 1988; Ishak & Zimmerman, 1987; Kreek, et al., 1967; Reyes, et al., 1981; Sanchez Pozzi, et al., 2003; Simon, et al., 1996; Welder, et al., 1995; Westaby, et al., 1977) or liver necrosis in human or liver hypertrophy in rat and mice by troglitazone versus pioglitazone and rosiglitazone (Scheen, 2001; Smith, 2003).

Other examples are digoxin, which appears to be a very specific human potassium channel inhibitor (Jover, et al., 1992; Okey, et al., 1986), being completely inactive in the rodent and dog, and taxol, which is a mammalian specific genotoxic or cytostatic drug that is not active in the cellular assays with Salmonella or yeast (Kirkland, et al., 2008; Birrell, et al, 2010; Westerink, et al., 2010).

To cope with this inconsistency between species, there is an urgent need for a good characterisation of large compound sets with particular toxic liabilities across species using species-specific precision cut tissue slices and/or species- specific primary or permanent cell lines, such as hepatocytes, lymphocytes, lymphoblasts, cardiomyocytes, myocytes, fibroblasts and/or kidney, neuronal or glioma cells. Initiatives as AcuteTox (Clemedson et al., 2006, 2007), supported by ECVAM, and ToxCast (Dix, et al., 2007; Andersen & Krewski, 2009), supported by the US government, are very valuable in this respect as many in vitro tests and in silico models are already performed and analyzed in these initiatives.

Although the proof of concept was already given in the past with Multicenter Evaluation of In vitro Cytotoxicity (MEIC) studies on these technologies (Clemedson & Ekwall, 1999), these new techniques are not yet routinely employed in the pharmaceutical industry. However, we need to be careful that the drugs we are developing are not becoming the best in class for our test animals, yet retain a risk for humans. A combination of several informative in vitro tests can be of additive value for the identification of the mechanism of action and its severity in advance of the current gold standard of a toxicological animal read-out. In this respect, toxicogenomics and metabonomics may become very useful additional techniques in the ultimate identification of the modes and/or mechanisms of action of a drug with in vitro assays.

Within our company we embraced the in vitro toxicological strategy and built, in line with earlier publications, a panel of assays. Although not yet complete, this allows early identification of several intermediate and strong toxic liabilities of particular substitutes in a compound. With trial and error this even resulted in the adaptation and/or omission of certain chemical (sub-)structures with the development of new molecules by an iterative process (Fig. 2).

The assay battery was developed to identify the toxic liabilities in the past and present (Schoonen, et al., 2008), in which specific focus was given towards genotoxicity and reproductive toxicity. The complete test battery is shown in Table 1. The focus on genotoxicity and reproductive toxicity was the result of the portfolio of compounds prepared within legacy Organon (a pharmaceutical company, which merged into Schering-Plough in 2007 and later into MSD in 2009) in the area of male and female contraception, hormone replacement therapy and fertility (Schoonen, et al., 2009). Steroid hormones, like estrogens, androgens and progestagens, their non-steroidal hormone counterparts as well as their antihormones were the main chemicals used in this research area.

Figure 2. View on the lead optimisation process of new drugs
Many of these reproductive (anti-)hormones did result in increased genotoxicity and reproductive toxicity at very high dose levels, as can be expected due to the influence of (anti)-estrogens, (anti)-androgens, and (anti)-progestagens on the development and/or growth of breast, prostate and endometrium tumors, respectively. Moreover, estrogens and progestagens can block separately or in combination the oestrus cycle and estrogens as well as antiprogestagens can induce abortions. In addition, effects on the gender of the foetus with respect to androgenic compounds and 19-nor-progestagenic compounds given at high dosages to the mother during foetal development are well-known. Also diethylstilbestrol (DES), used for pregnancy maintenance in the mother in the past, caused in the offspring complex disorders in the vagina and clitoris of females and the genitals and/or penis formation in males.

For in silico analysis the DEREK program is used in house as a pre-screening tool for the identification of genotoxic compounds. Since the identification is based on chemical entities in a molecule and not on the whole molecular structure as such, a lot of genotoxic compounds are not or wrongly identified by this program. Mutalert was built in house to improve these shortcomings (Kazius, et al., 2005). This program resulted in better predictions, but still the outcome of the biological tests showed a better total performance.
Table 1. Scheme of the test battery for the in silico and in vitro toxicity screening procedures used or to be used within our company.

Test BatteryTest TypeToxicity Screening Procedures
Program modelsin silicoDEREK
Genotoxicity in bacteria, yeast and human cellsin vitroVitotox: SOS repair system
RadarScreen: RAD54 promoter
human HepG2 cells: RAD51C promoter
Cystatin A promoter
p53 Responsive Element
Cytotoxicity in human and rat cellsin vitroGlutathione depletion with monochlorobimane
Membrane damage via calcein uptake
Mitochondrial damage (Alamar Blue, CytoLyte, ATP-Lyte, LuxCell)
Cellular proliferation with Hoechst 33342
Nrf2-induced luminescent oxidative stress assay
Non-genotoxic carcinogenicity in human and rat cellsin vitroArylhydrocarbon Receptor (AhR) activation
Peroxisomal Proliferative Activation Receptor α (PPARα)
Constitutive Androstane Receptor (CAR)
Nuclear receptor activation with human receptorsin vitroPregnane X receptor (PXR)
Cytochrome P450 enzyme activation measured with Promega specific luciferin substrate assaysin vitroCYP 1A1, 1A2, 2C8, 2C9 and 3A4
Cytochrome P450 enzyme fluorescent competition assaysin vitroCYP 1A1, 1A2, 2A6, 2C8, 2C9, 2D6, 2E1 and 3A4
Reproductive hormonesin vitroEstrogen Receptor α and β (ERα and ERβ)
Androgen Receptor (AR)
Progesterone Receptor A and B (PR-A and PR-B)
Mineralocorticoid Receptor (MR)
Glucocorticoid Receptor (GR)
Follicle Stimulating Hormone receptor (FSH receptor)
Luteinizing Hormone receptor (LH receptor)
Thyroid Stimulating Hormone receptor (TSH receptor)
Embryotoxicityin vitroPeroxisomal Proliferative Activation Receptor α and γ (PPARα, γ)
Retinoic Acid Receptors α, β, and γ (RARα, β, γ)
Retinoic Acid X Receptors α, β, and γ (RXRα, β, γ)

For screening of genotoxicity and/or carcinogenicity a panel of five sensitive luminescent promoter based assays have been developed for three different species. A bacterial assay with Salmonella tiphimurium strain TA104 is carried out with the SOS repair system, which activates a luciferase expression system via a cascade of reactions (van der Lelie, et al., 1997; Verschaeve, et al., 1999). This so-called Vitotox assay (Gentaur, Brussels, Belgium), correlates very well with the full Ames test results (Westerink, et al., 2009). This test identifies mutagenic compounds, but besides that also some of the clastogenic compounds can be recognized. For fast identification of clastogenicity, i.e. chromosomal aberrations and micronucleus induction, the yeast RadarScreen assay (reMYND, Leuven, Belgium) with the RAD54 promotor in combination with a β-galactosidase read-out and the addition of a specific 6-O-β-galactopyranosyl-luciferin, ATP and luciferase cocktail (Promega) has been used. The correlation with the in vitro micronucleus and chromosomal aberration test data with Chinese hamster ovary and lung cells was very good (Westerink, et al., 2009).

Unfortunately, also here too many false positives were recognized when compared to the in vivo situation in rats and humans. Good sensitivity, selectivity and predictivity for the clastogenicity tests was obtained for the human hepatocyte permanent liver cell line clones of HepG2 with either the promoter of RAD51C, Cystatin A or the responsive element of the p53 protein interaction. The RAD51C gene is activated if chromosomal aberrations are induced. The Cystatin A gene is a sort of anti-apoptotic gene, allowing the cell to repair DNA damage before apoptotic cell death will be activated. In a similar way, the p53 pathway is activated to restore the DNA damage before initiation of apoptotic cell death. These human HepG2 cells are active metabolizers according to Knasmüller, et al. (2004) and Mersch-Sundermann, et al. (2004), which was confirmed by us, in the absence of the rat S9 metabolic activation system, for the following drugs benzo[a]pyrene, aflatoxin B1, dimethylbenzanthracene (DMBA), nitrofurantoin, 2-acetylaminofluorene (AAF) and etoposide (Westerink, et al., 2010), implying sufficient capacity of CYP1A1, CYP1A2 and CYP3A4. However, cyclophosphamide and dimethylnitrosamine are not metabolized by these cells, due to the lack of sufficient amounts of the enzymes of CYP2B6 and CYP2E1, respectively.

Addition of the rat S9 fraction to HepG2 cells for 3 hours can overcome this shortcoming in enzyme activity (Susan Aldridge, 2010; Boehme, et al, 2010; see presentation of Kathleen Böhme, Merck Serono, at the human predictive ADME/TOX meeting Brussels, 2010). This may also be the case for the activity of CYP2C8 and CYP2C9, which in HepG2 cells was below the detection level in our studies and those of Kathleen Böhme.

For the prediction of human genotoxicity, these cells were tested against 62 reference compounds of ECVAM (Kirkland, et al., 2008) and a large number of genotoxic and non-genotoxic compounds (190) with diverse mechanisms of action, including direct acting genotoxicants, topoisomerase inhibitors, nucleotide/DNA synthesis inhibitors, reactive oxygen species generators, and aneugenes (Westerink, et al., 2009, 2010). The sensitivity, specificity and predictivity for the ECVAM list for these three HepG2 assays – if combined – was 85% (17/20), 81% (34/42) and 82% (51/62), respectively. For the second set of 190 compounds, for which in vivo animal data are available for only 70 compounds for comparison, sensitivity, specificity and predictivity show levels of 74% (28/38), 81% (26/32) and 77% (54/70), respectively.

For cytotoxicity there are various possibilities to build a smart panel of assays, in which spectrophotometric, fluorometric or luminometric assays can be used. We are currently using five of these assays on a routine basis, i.e. for glutathione depletion, membrane damage, mitochondrial activity (Alamar Blue & Luxcell) and cellular proliferation (Schoonen, et al., 2005a, 2005b). These assays give a relatively good view on acute toxic effects, like necrosis and mitochondrial impairment. Of course many other good assays are also available for the mitochondrial activity evaluation, such as the tetrazolium salts MTT, NBT, XXT or the luminescent CytoLyte assay. For membrane damage, the enzyme activity measurements of lactate dehydrogenase, aspartate aminotransferase, alanine aminotransferase or glutathione-S-transferase can also be used, while for the energy status ATP-Lyte or Bright-Glo tests can be applied. Recently the oxidative stress assay with the Nrf2 responsive element promoter in combination with luciferase has been added to our panel (Westerink, et al., 2010). This assay is informative for the kind of stress induced by both pharmaceutical and genotoxic compounds.

A relative new approach, High Content Screening (HCS), is taking the lead for cellular toxicity analysis lately. Peter O’Brien and his colleagues (2006) started with this technique within Pfizer with HepG2 cells as well as with primary rat hepatocytes. In principle, they studied the cells for toxicity on an individual basis by microscopy and staining with different fluorophores. Most common is coloration of the nucleus (e.g. DAPI or Hoechst 33342) and cytosol (DRAQ5™), or on the levels of intracellular calcium (Fluo-4), membrane leakage (TOTO-3), and mitochondrial membrane potential (TMRM). The sensitivity (97%) and specificity (87%) for cellular toxicity within HepG2 cells compared with human in vivo data were both astonishing high when compared with those of the primary rat and human hepatocytes, being only in the order of sensitivity of 70% and 60%, respectively (Xu, 2007; Xu, et al., 2008).

These good predictions for human HepG2 cells were confirmed by Johnson & Johnson, Belgium. This company identified 8 out of 10 compounds, being non-toxic in the rat, but toxic in the human clinical setting, as being toxic in HepG2 cells with HCS (Vanparys, 2007). These positive findings show the capacity of this technique to improve the quality of the selection within the drug screening process. Implementation of HCS in our department is ongoing with the focus on genotoxicity by analysis of in vitro micronuclei and on general toxicity for phospholipidosis.

For carcinogenicity and non-genotoxic carcinogenicity normally two-year rat and mouse studies are needed for risk assessment. New transgenic animal models with a higher sensitivity were developed, in which certain DNA repair mechanisms were made deficient. The treatment time in these animal studies is decreased from two years to nine months. With the in vitro assays for genotoxicity mentioned above also carcinogenic effects can be identified, but now within three days. On the other hand, specific non-genotoxic carcinogens must be identified with other in vitro methods of short duration. In rats and mice it is known that nuclear receptors like Peroxisomal Proliferation Activating Receptor α (PPARα), Aryl hydrocarbon Receptor (AhR), or Constitutive Androstane Receptor (CAR) can activate carcinogenesis via their cognate ligands (Fingerhut, et al., 1991; Manz, et al., 1991; Phillips & Goodman, 2009; Van den Heuvel, 1999; Zober, et al., 1990). In these species, WY 14.643, troglitazone and fibrates are well characterized PPARα inducers, TCDD (dioxine) and specific polychlorinated biphenyls are well defined AhR inducers, while phenobarbital is a potent CAR inducer (Oliver & Roberts, 2002; Lake, 2009). These ligands all induce rat and mouse specific non-genotoxic carcinogenicity.

Replacement of the specific mouse PPARα and CAR by their human counterpart leads with these specific ligands to silencing of these effects (Lin, 2008; Morimura, et al., 2005). Moreover species specific interactions can also cause different species specific risk factors as shown for several compounds for the human and rat AhR by Westerink, et al. (2008). Thus species specific activation processes are very relevant and identification of the specific rat and/or mouse PPARα, AhR or CAR activation, may help in the segregation of the risk for humans. Thus early screening for one week by in vitro research on these receptors can be very beneficial to prevent unnecessary animal studies for 2 years for carcinogenicity for compounds that activate the PPARα, AhR or CAR in rats. Such compounds can or may already be replaced by modified synthetic compounds lacking such receptor interactions.

Prediction of nuclear receptor activation of the pregnane X receptor and activation of rat CYP3A1 or human CYP3A4 can be done by both primary and permanent cell lines. In both primary as well as permanent liver hepatocytes the receptor or enzyme induction can be measured with Q-PCR, with direct LC-MS methods measuring the conversion of testosterone into 6β-hydroxytestosterone (CYP3A1/3A4) or by the indirect conversion of a specific luciferin-substrate (Promega) for CYP3A4 into free luciferin, which quantity can be measured by the addition of ATP and luciferase. Such specific luciferin substrates are also available for CYP1A1, 1A2, 1B1, 2C8, 2C9, 2C19 and 2D6. An alternative for studying the influence of a compound on the interaction with the cytochrome P450 enzymes is by competition analysis for each enzyme with enzyme specific fluorometric substrates and inhibitors as a reference control. In this respect, assays have been set up for CYP1A1, 1A2, 2A6, 2C8, 2C9, 2D6, 2E1 and 3A4 (Schoonen, et al, 2009).

Within the field of reproductive toxicity, embryotoxicity and teratogenicity the role of reproductive hormones with respect to timing and dosing is very important for the different species examined. For instance, estrogens can disturb the reproductive cycle and induce embryotoxicity. Progestagens produced within the corpora lutea maintain the endometrium and/or the placenta in its proper condition. Glucocorticoids given at high concentrations during embryonic development are known to induce malformations like cleft palate formation. Moreover, an enhanced exposure to follicle stimulating hormone (FSH) and a sharp peak of luteinising hormone (LH) at the end of the follicular maturation will lead to the induction of the maturation and ovulation of the oocyte within the mature Graafian follicles. The timing of this process is very delicate and needs to be carried out very accurately.

After coping, this will then proceed in the fertilisation of the egg with sperm in the oviduct, and later on in the implementation of the egg or blastocyst into the endometrium. If no rejection of the blastocyst occurs, embryonic development can progress and the placenta may be formed. Androgens are very important in the late phase of the gender development, as they are essential for the regression of the Müllerian duct and the maintenance of the Wolffian duct. Therefore androgens are well-known as reprotoxic hormones. For the measurement of estrogenic, progestagenic, androgenic, glucocorticoid and mineralocorticoid activity and their antihormonal activities, their cognate receptors are available in Chinese hamster ovary cells with a specific promoter and luciferase based read-out measuring the potency of the compound or its metabolites (Schoonen, et al., 2009). Also for FSH and LH receptors, luciferase based promoter assays are available.

In a similar fashion, nuclear receptor assays for retinoic acid receptors (RAR) α, β and γ, retinoic acid X receptors (RXR) α, β and γ, and thyroid hormone receptor (TR) α and β were developed for the area of embryotoxicity. For human RAR and RXR the assay has been developed in HepG2 cells, while for TR this process is ongoing. This research was set-up as especially retinoic acid (Vitamin A), 9-cis retinol, all trans-retinoic acid, and thyroid hormones are known to be embryotoxic compounds. Both high as well as low concentrations of vitamin A and thyroid hormone analogues are inducing embryotoxicity. Moreover, inhibitors of the enzymes involved in the synthesis or breakdown of retinoic acid or thyroid hormone, like retinoic acid dehydrogenase, retinaldehyde dehydrogenase or thyroxin deiodase, can induce embryotoxicity. Therefore compounds that influence the activity of (RAR) α, β and γ or retinoic acid X receptors (RXR) α, β and γ, or thyroid hormone receptor (TR) α and β may be of importance in this respect. Other well-known compounds that induce embryotoxicity are the cytostatics 5-fluorouracil and methotrexate. Both of these compounds are involved in the synthesis of thymidine and a blockade in this synthesis also leads to cytotoxicity.

Besides the above mentioned fast screening methods, the possibility exists to perform more in depth analysis. With toxicogenomics we try to make progress in the analysis of hepatotoxic, non-genotoxic as well as genotoxic compounds. For this purpose studies are carried out in vivo with rats, in vitro with HepG2 cells and with precision cut liver slices, in which all liver cell types are present in their natural architecture. This latter is important since drug-induced toxicity often is a multi-cellular process involving not only hepatocytes but also other cell types such as Kupffer and stellate cells. The results of the microarray studies show that the in vitro profiles of gene expression in rat liver slices cluster per compound and incubation time, and when analyzed in a commercial gene expression database (ToxShield™), can predict the in vivo toxicity and pathology (Elferink, et al., 2008). The prediction of human specific toxicity using human liver slices is currently under investigation. In HepG2 cells, necrotic and cholestatic compounds can be distinguished from genotoxicants. Data show that the rat liver slice system and HepG2 cells represent appropriate tools for the prediction of liver toxicity.

In this review the importance of in vitro toxicity analysis on rat and human tissues in lead discovery and lead optimisation has been discussed. An appropriate and well-considered in vitro screening strategy may significantly reduce the high attrition rate in (pre-)clinical studies. Another advantage of in vitro studies is that they often can go into more detail in the mechanisms of action of the drug interactions. A scheme of potential assays for innovative drug development has been described to identify different kinds of toxicity with respect to species differences, the cellular distribution or failure sites, DNA or receptor interaction, enzyme competition or inhibition level, or the receptor mediated activation processes. Although not yet complete, these mechanisms can help to identify the early onset of toxicological effects. Further improvement will be needed to incorporate the influence of phase II and III metabolism, of idiosyncratic toxicity failure and immunotoxicity. Further exploration of new techniques in the in vitro toxicity area will help to enlarge the battery of tests and to make the proper adaptations by including new tests to optimize the prediction of human toxicity.
©2010 Willem G.E.J. Schoonen, Walter M.A. Westerink & G. Jean Horbach

Accanito, L., Figuerora, C., Pirazzo, M. & Solis, N. (1995). Enhanced biliary excretion of canalicular membrane enzymes in estradiol-induced and obstructive cholestasis, and effects of different bile acids in the isolated perfused rat liver. J. Hepatol. 22, 658-670.

Accanito, L., Pirazzo, M., Solis, N., Koenig, C.S., Vollrath, V. & Chianale, J. (1996). Modulation of hepatic content and biliary excretion of P-glycoproteins in hepatocellular and obstructive cholestasis in the rat. J. Hepatol. 25, 349-361

Aldridge, S. (2010). New predictive ADME/Tox strategies reduce attrition. Genetic Engineering & Biotechnology News. 30, 1-3.

Ames, B.N., Lee, F.D. & Durston, W.E. (1973). An improved bacterial test system for the detection and classification of mutagens and carcinogens. Proc. Nat. Acad. Sci. USA. 70, 782-786.

Andersen, M.E. & Krewski, D. (2009). Toxicity testing in the 21st century: Bringing the vision to life. Toxicol. Sci. 107, 324-330.

Birrell, L., Cahill, P., Huhges, C., Tate, M. & Walmsley, R.M. (2010). GADD45a-GFP Greenscreen HC assay results for the ECVAM recommended lists of genotoxic and non-genotoxic chemicals for assessment of new genotoxicity tests. Mutat. Res. 695, 87-95.

Böhme, K., Dietz, Y., Hewitt, P. & Mueller, S.O. (2010). Activation of p53 in hepG2 cells as surrogate to detect mutagens and promutagens in vitro. Toxicol. Lett. 198, 272-281.

Brunmair, B., Lest, A., Staniek, K., Gras, F., Scharf, N., Roden, M., et al. (2004).Fenofibrate impairs rat mitochondrial function by inhibition of respiratory complex I. J. Pharm. Exp. Ther. 311, 109-114.

Chaung, W., Mi, L.J. & Boorstein, R.J. (1997). The p53 status of Chinese hamster V79 cells frequently used for studies on DNA damage and DNA repair. Nucleic Acids Res. 25, 992-994.

Clemedson, C. & Ekwall, B. (1999). Overview of the MEIC results: I the in vitroin vitro evaluation. Toxicol. In Vitro. 13, 657-663.

Clemedson, C., Blaauboer, B., Castell, J., Prieto, P., Ristyelli, L., Vericat, J.A., et al. (2006). ACuteTox – Optimation and pre-validation of an invitro test strategy for predicting human acute toxicity. ALTEX. 23, Suppl, 254-258.

Clemedson, C., Kolman, A. & Forsby, A. (2007). The integrated acute systemic toxicity project (ACuteTox). for the optimisation and validation of alternative in vitro tests. Altern. Lab. Anim. 35, 33-38.

Dix, D.J., Houck, K.A., Martin, M.T., Richard, A.M., Setzer, R.W. & Kavlock, R.J. (2007). The ToxCast program for prioritizing toxicity testing of environmental chemicals. Toxicol. Sci. 95, 5-12.

Elferink, M.G.L., Olinga, P., Draaisma, A.L., Merema, M.T., Bauerschmidt, S., Polman, J., et al. (2008). Microarray analysis in rat liver slices correctly predicts in vivo hepatotoxicity. Toxicol. Appl. Pharmacol. 229, 300-309.

Fingerhut, M.A., Halperin, W.E., Marlow, D.A., Piacitelli, L.A, Honchar, P.A., Sweeney, M.H., et al. (1991). Cancer mortality in workers exposed to 2,3,7,8-tetrachlorodibenzo-p-dioxin. N. Engl. J. Med. 324, 212-218.

Finsterer, J. & Segall, L. (2010). Drugs interfering with mitochondrial disorders. Drug Chem. Toxicol. 33, 138-151.

Frezza, M., Tritapepe, R., Pozzato, G. & Di Padova, C. (1988). Prevention of S-adenosylmethionine of estrogen-induced hepatobiliary toxicity in susceptible women. Am. J. Gastroenterol. 83, 1098-1102.

Guengerich, F.P. & MacDonald, J.S. (2007). Applying mechanisms of chemical toxicity to predict drug safety. Chem. Res. Toxicol. 20, 344-369.

Hu, T., Miller, C.M., Ridder, G.M. & Aardema, M.J. (1999). Characterisation of p53 in Chinese hamster cell lines CHO-K1, CHO-WBL, and CHL: Implications for genotoxicity testing. Mutat. Res. 426, 51-62.

Ishak, K.G. & Zimmerman, H.J. (1987). Hepatotoxic effects of the anabolic/androgenic steroids. Semin. Liver Dis. 7, 230-236.

Jover, R., Pondosa, X., Castell, J.V., Gómez-Lechón, M.J. (1992). Evaluation of the cytotoxicity of ten chemicals on human cultured hepatocytes: Predictability of human toxicity and comparison with rodent cell culture systems. Toxicol. In Vitro. 6, 47-52.

Kazius, J., McGuire, R. & Bursi, R. (2005). Derivation and validation of toxicophores for mutagenicity prediction. J. Med. Chem. 48, 312–320.

Kirkland, D., Kasper, P., Muller, L., Corvi, R. & Speit, G. (2008). Recommended lists of genotoxic and non-genotoxic chemicals for assessment of the performance of new or improved genotoxicity tests: A follow-up to an ECVAM workshop. Mutat. Res. 653, 99-108.

Kirkland, D., Pfuhler, S., Tweats, D., Aardema, M., Corvi, R., Darroudi, F., et al. (2007). How to reduce false positive results when undertaking in vitro genotoxicity testing and thus avoid unnecessary follow-up animal tests: Report of an ECVAM Workshop. Mutat. Res. 628, 31-55.

Kola, I. & Landis, J. (2004). Can the pharmaceutical industry reduce attrition rates? Nat. Rev. Drug Discov. 3, 711-715.

Knasmüller, S., Mersch-Sundermann, V., Kevekordes, S., Darroudi, F., Huber, W.W., Hoelzl, C., et al. (2004). Use of human-derived liver cell lines for the detection of environmental and dietary genotoxicants; current status of knowledge. Toxicology. 198, 315-328.

Kreek, M.J., Weser, E., Sleisenger, M.H. & Jeffries, G.H. (1967). Idiopathic chloestasis of pregnancy. The response to challenge with the synthetic estrogen, ethinyl estradiol. N. Engl. J. Med. 277, 1391-1395.

Lake, B.G. (2009). Species differences in the hepatic effects of inducers of CYP2B and CYP4A subfamily forms: Relationship to rodent liver tumour formation. Xenobiotica. 39, 582-596.

Lin, J.H. (2008). Applications and limitations of genetically modified mouse models in drug discovery and development. Curr. Drug Metab. 9, 419-438.

Manz, A., Berger, J., Dwyer, J.H., Flesch-Janys, D., Nagel, S. & Waltsgott, H. (1991). Cancer mortality among workers in chemical plant contaminated with dioxin. Lancet. 338, 959-964.

McCann, J., Choi, E., Yamasaki, E. & Ames, B.N. (1975). Detection of carcinogens as mutagens in the Salmonella/microsome test: Assay of 300 chemicals. Proc. Nat. Acad. Sci. USA. 72, 5135-5139.

McGregor, J.T., Casciano, D. & Müller, L. (2000). Strategies and testing methods for identifying mutagenic risks. Mutat. Res. 455, 3-20.

Mersch-Sundermann, V., Knasmüller, S., Wu, X.J., Darroudi, F. & Kassie, F. (2004). Use of a human-derived liver cell line for the detection of cytoprotective, antigenotoxic and cogenotoxic agents. Toxicology. 198, 329-340.

Miller, B., Pötter-Locher, F., Seelbach, A., Stopper, H., Utesch, D. & Madle S. (1998). Evaluation of the in vitro micronuclei test as an alternative to the in vitro chromosomal aberration assay: Position of the GUM working group on the in vitro micronucleus test. Mutat. Res. 410, 81-116.

Mordente, A., Minotti, G., Martorana, G.E., Silvestrini, A., Giardina, B. & Meucci, E. (2003). Anthracycline secondary alcohol metabolite formation in human or rabbit heart: Biochemical aspects and pharmacologic implications. Biochem. Pharmacol. 66, 989-998.

Morimura, K., Cheung, C., Ward, J.M., Reddy, J.K. & Gonzalez, F.J. (2006). Differential susceptibility of mice humanized for peroxisome proliferator-activated receptor α to WY-14,643-induced liver tumorigenesis. Carcinogenesis. 27, 1074-1080.

O’Brien, P.J., Irwin, W., Diaz, D., Howard-Cofield, E., Krejsa, C.M., Slaughter, M.R., et al. (2006). High concordance of drug-induced human hepatotoxicity with in vitro cytotoxicity measured in a novel cell-based model using high content screening. Arch. Toxicol. 80, 580-604.

Okey, A.B., Roberts, E.A., Harper, P.A. & Denison, M.S. (1986). Induction of drug-metabolizing enzymes: Mechanisms and consequences. Clin. Biochem. 19, 132-141.

Oliver, J.D. & Roberts, R.A. (2002). Receptor-mediated hepatocarcinogenesis: Role of hepatocytes proliferation and apoptosis. Pharmacol. Toxicol. 91, 1-7.

Olson, R.D., Mushlin, P.S., Brenner, D.E., Fleischer, S., Cusak, B.J., Chang, B.K. & Boucek, R.J. (1988). Doxorubicin cardiotoxicity may be caused by its metabolite doxorubicinol. Proc. Nat. Acad. Sci. USA. 85, 3585-3589.

Phillips, J.M. & Goodman, J.I. (2009). Multiple genes exhibit phenobarbital-induced constitutive active/androstane receptor-mediated DNA methylation changes during liver tumorigenesis and in liver tumors. Toxicol. Sci. 108, 273-289.

Reyes, H., Ribalta, J., Gonzalez, M.C., Segovia, N. & Oberhauser, E. (1981). Sulfobromophtalein clearance tests before and after ethinyl estradiol administration, in women and men with familial history of intrahepatic cholestasis of pregnancy. Gastroenterology. 81, 226-231.

Sánchez Pozzi, E.J., Crozenci, F.A., Pellegrino, J.M., Catania, V.A., Luquita, M.G., Roma, M.G., et al. (2003). Ursodeoxycholate reduces ethinylestradiol glucuronidationin the rat: Role of prevention in estrogen-induced cholestasis. J. Pharmacol. Exp. Ther. 306, 279-286.

Scheen, A.J. (2001). Thiazolidinediones and liver toxicity. Diabetes Metab. 27, 305-313.

Schoonen, W.G.E.J., de Roos, J.A.D.M., Westerink, W.M.A. & Débiton, E. (2005a). Cytotoxic effects of 110 reference compounds on HepG2 cells and for 60 compounds on HeLa, ECC-1 and CHO cells. II Mechanistic assays on NAD(P)H, ATP and DNA contents. Toxicol. In Vitro. 19, 491-503.

Schoonen, W.G.E.J., de Roos, J.A.D.M., Westerink, W.M.A. & Débiton, E. (2005b). Cytotoxic effects of 100 reference compounds on Hep G2 and HeLa cells and of 60 compounds on ECC-1 and CHO cells. I Mechanistic assays on ROS, glutathione depletion and calcein uptake. Toxicol. In Vitro. 19, 505-516.

Schoonen, W.G.E.J., Westerink, W.M.A. & Horbach, G.J. (2008). In vitro toxicity screening as pre-selection tool in pre-clinical research. Eur. Pharm. Rev. 3, 40-44.

Schoonen, W.G.E.J., Westerink, W.M.A. & Horbach, G.J. (2009). High-throughput screening for the analysis of in vitro toxicity. Molecular, Clinical and Environmental Toxicology (ed. Andreas Luch). Birkhäuser Publishing, Basel, Switzerland. Vol. 1, chp 14, 401-439.

Simon, F.R., Fortune, J., Iwahashi M., Gartung, C., Wolkoff, A. & Sutherland, E. (1996). Ethinyl estradiol cholestasis involves alterations in expression of liver sinusoidal transporters. Am. J. Physiol. 271G, 1043-1052.

Smith, M.T. (2003). Mechanisms of troglitazone hepatotoxicity. Chem. Res. Toxicol. 16, 679-687.

Smith, D.A. & Schmid, E.F. (2006). Drug withdrawals and the lessons within. Curr. Opin. Drug Disc. Dev. 9, 38-46.

Spengler, U., Lichterfeld, M. & Rockstroh, J.K. (2002). Antiretroviral drug toxicity – a challenge for the hepatologist. J. Hepatol. 36, 283-294.

Tzang, B.S., Lai, Y.C., Hsu, M., Chang, H.W., Chang, C.C., Huang, P.C., et al. (1999). Function and sequence analysis of tumor suppressor gene p53 of CHO.K1 cells. DNA Cell Biol. 18, 315-321.

Van den Heuvel, P.J.V. (1999). Peroxisome proliferator-activated receptors (PPARS) and carcinogenesis. Toxicol. Sci. 47, 1-8.

van der Lelie, D., Regniers, L., Borremans, B., Provoost, A. & Verschaeve, L. (1997). The VITOTOX test, an SOS bioluminescence Salmonella typhimurium test to measure genotoxicity kinetics. Mutat. Res. 389, 279-290.

Vanparys, P. (2007). New models for toxicity testing. Lecture on 2nd Conference of predictive human toxicity and ADME toxicity studies. 25th-26th January, Brussels, Belgium.

Verschaeve, L., Van Gompel, J., Thilemans, L., Regniers, L., Vanparys, P. & van der Lelie, D. (1999). VITOTOX bacterial genotoxicity and toxicity test for the rapid screening of chemicals. Environ. Mol. Mutagen. 33, 240-248.

Welder, A.A., Robertson, J.W. & Melchert, R.B. (1995). Toxic effects of anabolic-androgenic steroids in primary rat hepatic cell cultures. J. Pharmacol. Toxicol. Methods. 33, 187-195.

Westaby, D., Ogle, S.J., Paradinas, F.J., Randell, J.B. & Murray-Lyon, I.M. (1977). Liver damage from long-term methyltestosterone. Lancet. 2, 262-263.

Westerink, W.M.A., Stevenson, J.C.R. & Schoonen, W.G.E.J. (2008). Pharmacologic profiling of human and rat cytochrome P450 1A1 and 1A2 induction and competition. Arch. Toxicol. 82, 909-921.

Westerink, W.M.A., Stevenson, J.C.R., Lauwers, A., Griffioen, G., Horbach, G.J. & Schoonen, W.G.E.J. (2009). Evaluation of the Vitotox™ and RadarScreen assays for the rapid assessment of genotoxicity in the early research phase of drug development. Mutat. Res. 676, 113-130.

Westerink, W.M.A., Stevenson, J.C.R., Horbach, G.J. & Schoonen, W.G.E.J. (2010). The development of RAD51C, Cystatin A, p53, and Nrf2 luciferase-reporter assays in metabolically competent HepG2 cells for the assessment of mechanism-based genotoxicity and of oxidative stress in the early research phase of drug development. Mutat. Res. 696, 21-40.

Xu, J.J. (2007). High content screening with primary rat hepatocytes. Lecture Symposium on: High Content Screening for toxicity (Informa). Le Meridian, Vienna, Austria, 28 June 2007.

Xu, J.J., Henstock, P.V., Dunn, M.C., Smith, A.R., Chabot, J.R. & de Graaf, D. (2008). Cellular imaging predictions of clinical drug-induced liver injury. Toxicol. Sci. 105, 97-105.

Zober, A., Messerer, P. & Huber, P. (1990). Thirty-four-year mortality follow-up of BASF employees exposed to 2,3,7,8-TCDD after the 1953 accident. Int. Arch. Occup. Environ. Health. 62, 139-157.