Application of emerging toxicity screens in drug discovery: challenges and implications

Abstract

Future Medicinal ChemistryVol. 1, No. 7 EditorialFree AccessApplication of emerging toxicity screens in drug discovery: challenges and implicationsAlan GE Wilson, Xuemei Liu and Jeffrey A KramerAlan GE Wilson† Author for correspondenceVice President, Drug Metabolism, Pharmacokinetics, Toxicology and Pathology, Lexicon Pharmaceuticals Inc., 8800 Technology Forest Place, The Woodlands, TX 77381-1160, USA. Search for more papers by this authorEmail the corresponding author at awilson@lexpharma.com, Xuemei LiuDrug Metabolism, Pharmacokinetics, Toxicology and Pathology, Lexicon Pharmaceuticals Inc., 8800 Technology Forest Place, The Woodlands, TX 77381, USA. Search for more papers by this authorEmail the corresponding author at mliu@lexpharma.com and Jeffrey A KramerDrug Metabolism, Pharmacokinetics, Toxicology and Pathology, Lexicon Pharmaceuticals Inc., 8800 Technology Forest Place, The Woodlands, TX 77381, USA. Search for more papers by this authorEmail the corresponding author at jkramer@lexpharma.comPublished Online:21 Oct 2009https://doi.org/10.4155/fmc.09.108AboutSectionsPDF/EPUB ToolsAdd to favoritesDownload CitationsTrack CitationsPermissionsReprints ShareShare onFacebookTwitterLinkedInReddit The discovery and development of drugs remain very challenging, time consuming, expensive and hampered by low success rates. It can take up to 12–15 years to bring a drug to market with an estimated cost in excess of US$800 million [1–3]. Safety issues remain a major cause of the late-stage attrition and low success rate. Discovering safety issues late in development is potentially financially devastating for a company. It is therefore imperative that potential safety issues are identified early. In this article, we highlight some selected in silico, in vitro and in vivo approaches that are being used to predict potential toxicity early in the drug-discovery phase with the hope of reducing late-stage attrition and improving success rates.Classical toxicity testing relies primarily on observing adverse biological responses in homogeneous groups of animals exposed to high doses of a test agent. However, these tests provide little information on modes and mechanisms of action, information critical for understanding interspecies differences in toxicity and the translation of preclinical data to humans. Moreover, these preclinical studies are expensive, time consuming and require large numbers of animals. The current focus and urgency in the pharmaceutical industry is to shorten the timelines for all aspects of drug discovery and to improve the ability to filter out early potential safety issues. We are therefore seeing an increasing interest and focus of developing high-throughput screening (HTS) approaches. These early HTS approaches that are completed in the early drug-discovery phase even before the efficacy screenings and not necessarily linked to the mechanistic-driven strategy will gain increasing importance in the coming years in allowing early identification of potential toxicity issues. In addition, the continuing understanding and advancement of ‘omics’ (i.e., genomics, proteomics and metabolonomics) technologies will elucidate patterns of biomarkers changes that aid the prediction of different toxicities. These approaches include promoter-reporter assays, gene expression arrays and bioanalytical methods. In silico and in vitro tests have the advantage of being able to rapidly evaluate large numbers of chemicals, greatly reducing animal use and accelerating lead identification and optimization. The potential of in silico approaches was heralded with great fanfare with the promise of rapid impact in decreasing attrition and increasing productivity. Not surprisingly, in recent years, we have seen a more reasoned and rational expectation of the potential of in silico predictive models and the increasing development and application of both in silico models and toxicity databases for a number of toxicity end points (e.g., mutagenicity, carcinogenicity, skin sensitization, reproductive and cardiac toxicity). In addition, pharma companies are now increasingly utilizing their in-house data for the development of both generic and chemotype-specific in silico models. Examples of some of the commercially available in silico models for toxicity prediction have been discussed recently [1,2,4–6]. Regulatory agencies are also providing leadership in the improvement of our toxicity prediction approaches and aiding in the development and validation of computational in silico approaches. For example, in 2007, the USEPA launched ToxCast™ [7], which is designed to use HTS data to build computational models to forecast the potential human toxicity of chemicals. These hazard predictions will provide USEPA regulatory programs with science-based information helpful in prioritizing chemicals for more detailed toxicological evaluations and lead to more efficient use of animal testing. In addition, the US FDA has a long interest and experience in working with industry on the evaluation of in silico approaches of toxicity prediction and have a number of programs focused on developing in silico models for various toxicities. Finally, the implementation of REACH in the EU will provide important data sources for the development of QSAR and other models to aid toxicity prediction.The development of predictive in vitro toxicity assays is conducted primarily with cells or cell lines, optimally from humans. The use of a mammalian cell system is preferable due to substantial differences in cellular outer surfaces and the fact that chemical interactions via transcription factors can be monitored only in mammalian cells, unless the relevant transcription factors are transfected into prokaryote cells [1,2]. Some examples of toxicity pathways and mechanisms that could be evaluated with high-throughput methods will now be discussed.Oxidative stressOxidative stress triggers cell death or apoptotic pathways by several distinct mechanisms. Among these, the mitochondrial death pathway induced by oxidative stress has lately drawn considerable attention [8–10]. The cell death is either apoptotic or necrotic depending on the cellular redox status. The activation of anti-oxidant response element signaling occurs through the oxidation of sentinel sulfhydryls on the protein Keap1 [8]. Some agents, such as chlorine, activate Nrf2 signaling in vitro and the oxidative stress is likely the cause of irritation and toxicity in the respiratory tract.Mitochondrial toxicityMitochondrial function plays a central role in various cellular activities, producing 95% of cellular ATP requirements and participating in a range of physiological processes [11,12]. A variety of drugs have been shown to affect the electron transport chain, coupling of oxidative phosphorylation, β-oxidation or other mitochondrial functions [12,13]. Such events may lead to the opening of a large pore across the mitochondrial membranes – the membrane permeability transition pore – eventually leading to the apoptosis or necrosis of cells, depending on the cellular ATP content. These drugs may, therefore, lead to organ damage, particularly in the liver, kidney, heart or skeletal muscle [11].Endoplasmic reticulum stress & heat-shock responseThe endoplasmic reticulum (ER) is central in eukaryotic cells for lipid synthesis, protein folding and protein maturation [14,15]. The ER is the major signal-transducing organelle that senses and responds to changes in cellular homeostasis. ER stress can be induced by an accumulation of unfolded protein aggregates (unfolded protein response) or by excessive protein traffic such as that caused by viral infection (ER overload response) [14–16]. If the stress cannot be resolved, apoptosis may be triggered, resulting in cell death [14,15].Apoptosis, cell growth & proliferationA universal cellular stress effect is the impairment of growth and proliferation, which represents an adaptive and integrated part of the stress response [17]. It allows the preservation of energy and reducing equivalents and redirects the utilization of these important metabolites toward macromolecular stabilization and repair [17]. In eukaryotic cells, the activation of cell cycle checkpoints is a key aspect of the cellular stress response [7]. Cell cycle checkpoints monitor macromolecular integrity and the successful completion of cellular processes prior to initiating the next phase in the cell cycle [18]. The eukaryotic cell cycle proteins that control checkpoints maintain the fidelity of DNA replication, repair and cell division in normal as well as stressed cells [19].DNA damage & repairDamage to DNA induces several cellular responses that enable the cell either to eliminate or cope with the damage, or to activate a programmed cell death process, presumably to eliminate cells with potentially catastrophic mutations [20]. The DNA damage-response reactions include: ▪ Repair damage and restoration of the DNA duplex;▪ Activation checkpoint to allow the repair and prevention of the transmission of damaged chromosomes;▪ Transcriptional response, which may be beneficial to the cell;▪ Apoptosis, which eliminates seriously deregulated cells [21–23].DNA repair mechanisms include direct repair, base-excision repair, nucleotide-excision repair, double-strand break repair and cross-link repair [21]. Damage to DNA structures induces repair enzymes that act through GADD45 and other proteins [22,24]. Unrepaired damage increases the risk of mutation during cell division and increases the risk of cancer [22,24].hERG channel interactionDrug-induced prolongation of the rate-corrected QT interval (QT[C]I) occurs as an unwanted effect of diverse clinical and investigational drugs and carries a risk of potentially fatal cardiac arrhythmias [25]. hERG is the gene encoding the α-subunit of channels that mediate the rapid delayed rectifier K+ current, which plays a vital role in depolarizing the ventricles of the heart [26]. Most QT(C)I-prolonging drugs can inhibit the function of recombinant hERG K+ channels and, consequently, in vitro hERG assays are used widely as front-line screens in cardiac safety testing of novel chemical entities [25–27].PXR, CAR, PPAR & AhR response pathwaysMany exogenous compounds alter gene expression through the orphan nuclear receptors, such as the aryl hydrocarbon receptor (AhR), constitutive androstane receptor (CAR), pregnane X receptor (PXR) and peroxisome proliferator activated receptor-α (PPARα) [28,29]. Even though activation of the xenobiotic metabolizing pathways might reduce the active xenobiotics by eliminating them from the body as metabolites [30], it may increase exposure to potentially more toxic metabolites. Additionally, induction of some oxidative enzymes may cause rodent-specific effects [4,31], for which a mechanistic understanding can provide an opportunity to manage a preclinical finding and advance an otherwise promising molecule.Endogenous hormone response pathwaysOne of the difficult challenges in toxicology is to devise reliable assays for the prediction of nongenotoxic rodent carcinogens [31–33]. Hormone-related cancers, namely breast, endometrium, ovary, prostate, testis, thyroid and osteosarcoma, share a unique mechanism of carcinogenesis. The changes of transcriptionally active hormone receptors, including estrogen, androgen, thyroid and progesterone receptors, drive cell proliferation and thus the opportunity for the accumulation of random genetic errors [34].Metabolite-mediated toxicityMirroring the metabolism of the whole animal by in silico and in vitro systems is a continuing and difficult challenge [35]. In silico models are available that provide a prediction of in vivo metabolism, and various in vitro approaches, including cells from a number of tissues and different preparations, are used to assess metabolism-mediated toxicity. For the foreseeable future, any strategy should also include the assessment of metabolism in the whole animal. Just as important is the identification of species differences in metabolism that may lead to toxic metabolites that are unique either to humans or one of the common preclinical species.In vivo safety assessmentAlthough human-specific toxicities are a significant cause of safety-related attrition and the primary reason for postapproval withdrawal, unacceptable safety profiles in regulated nonclinical studies continue to be a significant contributor to compound failures. Whereas in silico and in vitro approaches, such as those detailed above, may facilitate lead prioritization and provide invaluable mechanistic details and vital early safety signals, a key to the decision to advance or deprioritize a lead compound is a suitable safety margin. Accurate estimates of an in vivo safety margin can be difficult to determine from in vitro assays, so the inclusion of early in vivo toxicity signal-generation studies is a key piece of any nonclinical safety-assessment strategy [4]. However, the predictivity of the common preclinical species for human toxicity is incomplete [36], necessitating new approaches both in vitro and in vivo. The use of in vivo genetic models has been described both for predicting clinical efficacy [37] and carcinogenicity [38]; these technologies hold promise for the generation of mechanistic information vital to issue management and risk assessment. Finally, the evaluation of alternative in vivo models may also allow the earlier identification of human-specific toxicities [39,40].The goal of toxicity testing is to develop data that can ensure appropriate protection of public health from the potential adverse effects of exposure to xenobiotics. This goal provides a clear standard for judging the performance of prospective screening assays. In general, early higher-throughput approaches might be expected to have less predictive accuracy than later, lower-throughput screens. One strategy would be to err on the side of ‘false-positives’, which would eliminate compounds that would turn out negative in regulatory assays [41]. An alternate but less-preferable strategy would be to err on the side of ‘false-negatives’, which would increase the probability of a compound passing into development that would ultimately turn out positive in in vitro toxicity assays [41,42]. Another key potential problem with early predictive toxicity screens is compound impurities related to synthetic catalysts and source materials in the early drug-discovery phase. According to current understanding, virtually all of the early in silico, in vitro and in vivo screening tests are of limited use in directly establishing an estimation of the risk for humans. Improving the sensitivity of these tests through the better understanding of the basic cellular processes and alterations that can affect the integrity of the genetic material will require new mechanistic insights and better cell culture models. An obvious approach for the early identification of toxicity among discovery compounds is to combine predictions from different high-throughput assays and integration of bioassay results with in silico models, such that more accurate assessment may be generated. However, it cannot be stressed strongly enough that, for any predictive technology to add value by reducing cycle time and attrition rate, it will be critical that the relevance of these technologies and translatability to humans is fully understood. While no preclinical approach is expected to be totally predictive, the strategic application of a variety and combination of early in silico, in vitro and in vivo animal studies heralds the possibility and opportunity for transforming the success and productivity of drug discovery and development.AcknowledgementsOur thanks to Melinda M Albright for her assistance in the preparation of this article.Financial & competing interests disclosureThe authors have no relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties.No writing assistance was utilized in the production of this manuscript.Bibliography1 Wiebel FJ, Andersson TB, Casciano DA et al. Genetically engineered cell lines: characterisation and applications in toxicity testing. The report and recommendations of ECVAM workshop 26. Altern. Lab. Anim.25,625–639 (1997).Crossref, Google Scholar2 Hasspieler BM, Haffner GD, Adeli K. In vitro toxicological methods for environmental health testing. Rev. Environ. Health.11,213–227 (1996).Crossref, Medline, CAS, Google Scholar3 Adams CP, Brantner VV. Estimating the cost of new drug development: is it really 802 million dollars? Health Aff.25,420–428 (2006).Crossref, Medline, Google Scholar4 Helma C, Cramer T, Kramer S, De Raedt L. Data mining and machine learning techniques for the identification of mutagenicity inducing substructures and structure activity relationships of noncongeneric compounds. J. Chem. Inf. Comput. Sci.44,1402–1411 (2004).Crossref, Medline, CAS, Google Scholar5 Helma C. Data mining and knowledge discovery in predictive toxicology. SAR QSAR Environ. Res.15,367–383 (2004).Crossref, Medline, CAS, Google Scholar6 Dearden JC. In silico prediction of drug toxicity. J. Comput. Aided Mol. Des.17,119–127 (2003).Crossref, Medline, CAS, Google Scholar7 Judson R, Richard A, Dix DJ et al. The toxicity data landscape for environmental chemicals. Environ Health Perspect.117,685–695 (2009).Crossref, Medline, CAS, Google Scholar8 Kohen R, Nyska A. Oxidation of biological systems: oxidative stress phenomena, antioxidants, redox reactions, and methods for their quantification. Toxicol. Pathol.30,620–650 (2002).Crossref, Medline, CAS, Google Scholar9 Apel K, Hirt H. Reactive oxygen species: metabolism, oxidative stress, and signal transduction. Annu. Rev. Plant Biol.55,373–399 (2004).Crossref, Medline, CAS, Google Scholar10 Spector A. Review: Oxidative stress and disease. J. Ocul. Pharmacol. Ther.16,193–201 (2000).Crossref, Medline, CAS, Google Scholar11 Will Y, Hynes J, Ogurtsov VI, Papkovsky DB. Analysis of mitochondrial function using phosphorescent oxygen-sensitive probes. Nat. Protoc.1,2563–2572 (2006).Crossref, Medline, CAS, Google Scholar12 Neupert W, Herrmann JM. Translocation of proteins into mitochondria. Annu. Rev. Biochem.76,723–749 (2007).Crossref, Medline, CAS, Google Scholar13 Ernster L, Schatz G. Mitochondria: a historical review. J. Cell Biol.91(Suppl.),S227–S255 (1981).Crossref, Medline, CAS, Google Scholar14 Schroder M. Endoplasmic reticulum stress responses. Cell Mol. Life Sci.65,862–894 (2008).Crossref, Medline, CAS, Google Scholar15 Marciniak SJ, Ron D. Endoplasmic reticulum stress signaling in disease. Physiol. Rev.86,1133–1149 (2006).Crossref, Medline, CAS, Google Scholar16 Xu C, Bailly-Maitre B, Reed JC. Endoplasmic reticulum stress: cell life and death decisions. J. Clin. Invest.115,2656–2664 (2005).Crossref, Medline, CAS, Google Scholar17 Kultz D. Molecular and evolutionary basis of the cellular stress response. Annu. Rev. Physiol.67,225–257 (2005).Crossref, Medline, Google Scholar18 Hartwell LH, Weinert TA. Checkpoints: controls that ensure the order of cell cycle events. Science246,629–634 (1989).Crossref, Medline, CAS, Google Scholar19 Hartwell LH, Kastan MB. Cell cycle control and cancer. Science266,1821–1828 (1994).Crossref, Medline, CAS, Google Scholar20 Kaufmann WK, Paules RS. DNA damage and cell cycle checkpoints. FASEB J.10,238–247 (1996).Crossref, Medline, CAS, Google Scholar21 Hiom K. DNA repair: a riddle at a double-strand break. Curr. Biol.19,R331–R333 (2009).Crossref, Medline, CAS, Google Scholar22 Cohn MA, D’Andrea AD. Chromatin recruitment of DNA repair proteins: lessons from the fanconi anemia and double-strand break repair pathways. Mol. Cell.32,306–312 (2008).Crossref, Medline, CAS, Google Scholar23 Altieri F, Grillo C, Maceroni M, Chichiarelli S. DNA damage and repair: from molecular mechanisms to health implications. Antioxid. Redox Signal.10,891–937 (2008).Crossref, Medline, CAS, Google Scholar24 Sheikh MS, Hollander MC, Fornance AJ Jr. Role of Gadd45 in apoptosis. Biochem. Pharmacol.59,43–45 (2000).Crossref, Medline, CAS, Google Scholar25 Wible BA, Hawryluk P, Ficker E, Kuryshev YA, Kirsch G, Brown AM. HERG-Lite: a novel comprehensive high-throughput screen for drug-induced hERG risk. J. Pharmacol. Toxicol. Methods.52,136–145 (2005).Crossref, Medline, CAS, Google Scholar26 Priest BT, Bell IM, Garcia ML. Role of hERG potassium channel assays in drug development. Channels2(2) (2008).Crossref, Medline, Google Scholar27 Yao X, Anderson DL, Ross SA et al. Predicting QT prolongation in humans during early drug development using hERG inhibition and an anaesthetized guinea-pig model. Br. J. Pharmacol.154,1446–1456 (2008).Crossref, Medline, CAS, Google Scholar28 Giguere V. Orphan nuclear receptors: from gene to function, Endocr. Rev.20,689–725 (1999).Medline, CAS, Google Scholar29 Kliewer SA, Willson TM. Regulation of xenobiotic and bile acid metabolism by the nuclear pregnane X receptor, J. Lipid Res.43,359–364 (2002).Crossref, Medline, CAS, Google Scholar30 Nebert DW. Drug-metabolizing enzymes in ligand-modulated transcription. Biochem. Pharmacol.47,25–37 (1994).Crossref, Medline, CAS, Google Scholar31 Elcombe CR, Odum J, Foster JR et al. Prediction of rodent nongenotoxic carcinogenesis: evaluation of biochemical and tissue changes in rodents following exposure to nine nongenotoxic NTP carcinogens. Environ. Health Perspect.110,363–375 (2002).Crossref, Medline, CAS, Google Scholar32 Hoel DG, Haseman JK, Hogan MD, Huff J, McConnell EE. The impact of toxicity on carcinogenicity studies: implications for risk assessment. Carcinogenesis9,2045–2052 (1988).Crossref, Medline, CAS, Google Scholar33 Tennant RW, Elwell MR, Spalding JW, Griesemer RA. Evidence that toxic injury is not always associated with induction of chemical carcinogenesis. Mol. Carcinog.4,420–440 (1991).Crossref, Medline, CAS, Google Scholar34 Aranda A, Pascual A. Nuclear hormone receptors and gene expression. Physiol. Rev.81,1269–1304 (2001).Crossref, Medline, CAS, Google Scholar35 Coecke S, Ahr H, Blaauboer BJ et al. Metabolism: a bottleneck in in vitro toxicological test development. The report and recommendations of ECVAM workshop 54. Altern. Lab Anim.34,49–84 (2006).Crossref, Medline, CAS, Google Scholar36 Olson H, Betton G, Robinson D et al. Concordance of the toxicity of pharmaceuticals in humans and in animals. Regul. Toxicol. Pharmacol.32,56–67 (2000).Crossref, Medline, CAS, Google Scholar37 Zambrowicz BP, Sands AT. Knockouts model the 100 best-selling drugs – will they model the next 100? Nat. Rev. Drug Discov.2,38–51 (2003).Crossref, Medline, CAS, Google Scholar38 MacDonald J, French JE, Gerson RJ et al. The utility of genetically modified mouse assays for identifying human carcinogens: a basic understanding and path forward. The Alternatives to Carcinogenicity Testing Committee ILSI HESI. Toxicol. Sci.77,188–194 (2004).Crossref, Medline, CAS, Google Scholar39 Baryla UM, Fleming JS, Stanton HC. The anesthetized ferret, an in vivo model for evaluating inotropic activity: effects of milrinone and anagrelide. J. Pharmacol. Methods.20,299–306 (1988).Crossref, Medline, CAS, Google Scholar40 Faiola B, Falls JG, Peterson RA et al. PPAR α, more than PPAR δ, mediates the hepatic and skeletal muscle alterations induced by the PPAR agonist GW0742. Toxicol. Sci.105,384–394 (2008).Crossref, Medline, CAS, Google Scholar41 Mannisto PT, Kirkland D, Viluksela M, Tikkanen L. Toxicological studies with dithranol and its 10-acyl analogues. Arch. Toxicol.59,180–185 (1986).Crossref, Medline, CAS, Google Scholar42 Preston J, Hoffmann G. Casarett and Doull’s Toxicology. McGraw Hill, Berkshire, UK (2001).Google ScholarFiguresReferencesRelatedDetailsCited ByPredicting in vivo safety characteristics using physiochemical properties and in vitro assaysNigel Greene & Minghu Song1 September 2011 | Future Medicinal Chemistry, Vol. 3, No. 12 Vol. 1, No. 7 Follow us on social media for the latest updates Metrics History Published online 21 October 2009 Published in print October 2009 Information© Future Science LtdAcknowledgementsOur thanks to Melinda M Albright for her assistance in the preparation of this article.Financial & competing interests disclosureThe authors have no relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties.No writing assistance was utilized in the production of this manuscript.PDF download