Toxicological Testing
A major bottleneck in the drug development process is toxicological testing. Traditional in vivo tests performed in animals are difficult to automate; one can't put a rat into a 96-well plate! How does one manage this leg of the development process, where not "failing early" can mean millions of dollars in wasted capital?
Technical improvements in tissue culture and advances in molecular biology have increased the understanding of cellular and molecular processes and the differences in these processes between humans and animals. Tools from cellular and molecular biology are being used to develop research strategies for identifying primary target genes. Moreover, the costs of assessing potential health effects of newly identified or synthesized therapeutic compounds (new chemical entities, NCEs) necessitate alternatives to animal testing. In vitro testing provides the researcher with considerably more control of the variables than whole-animal testing; however, an advantage of whole-animal testing allows for any adverse effects of “uncontrolled” variables to demonstrate broader-scale problems. New tools for toxicity testing must be looked on as adjuncts to traditional testing methods. Any testing method has inherent difficulties: when using whole animals, data must be extrapolated from one species to another; when using cell or tissue culture assays, data must be extrapolated to the whole organism. In vitro toxicological methods have allowed an earlier assessment of an NCE’s toxicity. Early determination of pharmaceutical properties can serve as predictors of a compound's possible developmental success — or lack thereof. Therefore, implementation of high-throughput ADME-Tox assays that address absorption, metabolism, and physical-chemical properties of potential therapeutics may serve to minimize discovery to market attrition.
Alternatives to whole-animal testing include endpoint assays, cell and tissue culture assays, the use of tissue slices, toxicokinetic modeling, and structure-activity relationships and databases. One of the difficulties with cell culture has to do with maintaining differentiated cells. Cells in culture tend to become unspecialized after a short time, losing the characteristics of the organ or tissue from which they were taken. Immortalized cells that have been genetically altered could prove useful for toxicity testing, although in in vitro testing, one looks for cells that respond closely to those of the intact human body. Continuous cell lines have undergone extensive selection for the ability to grow in culture, whereas normal cells have complex requirements for growth and differentiation in culture. The importance and great variety of growth factors, cell regulators, and mediators must also be taken into account.
Early drug discovery ADME cell-based assays, such as fast Caco-2 screens, can help rejection of test compounds that lack good efficacy and toxicity profiles. A cost-effective high-throughput method, PAMPA (parallel artificial membrane permeability analysis), which uses an artificial phospholipid membrane that models passive transport of epithelial cells, is becoming increasingly popular. In addition, drug-drug interactions occur when one drug alters the pharmacokinetics of a co-administered drug. This effect is often the result of drug-induced induction or inhibition of cytochrome P450 (CYP) enzymes. Induction of specific P450 enzymes may alter the metabolic profile of a drug by increasing metabolism or by creating an alternate pathway of metabolism. These changes can have profound effects on the pharmacology and toxicology of drugs. The effects of NCEs on the induction of CYP enzymes can be measured, assessed, evaluated, and performed in multiple species as numerous in vitro test systems are available. An understanding of species-specific changes in these important drug-metabolizing enzymes can provide important information for predicting how a drug is handled in animals versus humans. Moreover, these data may also provide an explanation for why a test compound that produces adverse effects in rodent studies may not have a similar effect in humans.
A high-throughput fluorescence assay using cDNAexpressed human CYP isozymes and fluorogenic substrates has been reported for the study of CYP inhibition. In the early stages of drug discovery, the fluorescence assay for CYP inhibition could be used in conjunction with a human liver microsomal assay to identify potential drug interactions between NCEs and established products. Such automated assays can be used for high-throughput ADME-Tox screening in early drug discovery although such screening this early in development is often cost-prohibitive for most companies.
Drug candidates may demonstrate cytotoxicity through apoptosis or necrosis. Different parameters for apoptosis can be measured compared with necrosis. Apoptosis is characterized by early events, such as expression of phosphatidylserine on the cell surface and fragmentation of DNA, followed by loss of membrane integrity and mitochondrial function. Parameters, such as drug concentration, time of exposure, and measurement of DNA fragmentation, can be customized for in vitro cytotoxicity assays. Necrosis occurs through the action of toxic factors that act within the cell, such as irreversible inhibitors of protein, RNA, or DNA synthesis, or mitotic poisons. Protein and nucleic acid synthesis rates are then measured to determine drug toxicity.
In vitro systems have provided information on metabolic pathways and mechanisms of action and have identified appropriate animal models for extrapolating to humans. When human cells are used, species extrapolation is less important and only a minimal amount of animal study is needed to confirm in vitro findings. From a scientific perspective, the FDA and most other regulatory agencies commonly require toxicology testing in two species, one of which is nearly always required to be a primate species.
The main questions concerning the use of in vitro assays are: 1) How do we extrapolate from an in vitro system to an in vivo system (i.e., how do we relate effects in single cells to complex interactions in whole animals); 2) How do we use available in vitro and in vivo data to design better experimental approaches; and 3) How do we predict potential biological effects from the chemical structure of a substance? It is important to note that research using cells, tissue cultures, or non-mammalian systems is conducted not only as an alternative to using mammals but because a given alternative system best answers the question under study. In vitro studies also allow researchers to understand the discrete steps in a specific sequence of events, which is difficult to do in whole animals. Using in vitro data, it is possible not only to extrapolate the human response but often to look at the pharmacokinetics of other species as well.
A wide array of tools is available for toxicity testing that has the capacity to greatly increase knowledge of the complex systems under investigation. Managing this step of the development process occurs simultaneously with pharmacokinetic analyses and, often, with early clinical trials. Toxicology testing development is often hampered by the requirement for validation and regulatory approval. Before non-animal toxicity tests may be officially accepted by regulatory agencies, it is generally agreed that the validity of the new methods must be demonstrated in an independent, scientifically sound validation program. The development of new, faster, more efficient in vitro tests, which are more prognostic and can be extrapolated to whole-animal testing, can result in a marked reduction in the use of in vivo testing, and the conjunction of the two can result in faster testing and lower development costs. It is not practical or realistic to assume that an in vitro test can fully predict a whole-body response, regardless of species used. The FDA requires extensive toxicology testing (both in terms of formal testing and in terms of safety assessments in human subjects) not only to identify “common” risks but also to identify those 1-in-10,000 or 1-in- 100,000 events that are severely debilitating or deadly. The establishment of data-based performance benchmarks will better guide reviewers of a validation study in setting realistic performance expectations given the real-world technical limitations characteristic of the current state of the art. Progress has been and continues to be made at this level, critical for the management of the drug development process, and of toxicological testing specifically, to maintain not only a full spectrum of potential therapeutics in the pipeline but a semblance of control over rising R&D costs.
The author would like to thank Dr. Janice Badger for constructive comments and suggestions.
Barbara VanRenterghem, Ph.D. is Science Editor for Lab Manager.