Current opinion in toxicology最新文献

The concept of a human body-on-a-chip model emulating entire organismal homeostasis has raised high expectations during the past decade. Microphysiological systems combining all essential organs in a common medium circuit have the potential to revolutionize not only basic research but also the drug development process. Hence, various researchers aimed at developing such organismal systems, but none went beyond initial scientific publication. In the following, we will highlight the current consensus of opinion on how to approach organismal homeostasis on-a-chip and summarize major challenges and how they might be tackled by current developments.

The development of human-induced pluripotent stem cell–derived cardiomyocytes (iPSC-CMs) has opened a new era to address the challenge of improving drug-induced cardiotoxicity prediction. Human iPSC-CMs can be generated from individuals with diverse genetic backgrounds and varying disease status, which provides unprecedented opportunities to assess drug-induced cardiotoxicity at the population level, ultimately, realizing personalized cardiac safety prediction and permitting mechanistic insights into genetic predisposition of drug-induced cardiotoxicity at the molecular and cellular levels. Reviewed herein are successful applications and limitations in using patient-specific iPSC-CMs for cardiac safety evaluation. Future directions for iPSC-CMs are also discussed. The aim of this review is to promote the further development of human iPSC-CM technology to address existing gaps in drug development, improve the prediction of patient susceptibility to therapeutic drugs, and enhance postmarketing surveillance of severe adverse drug reactions.

DILIsym®, a quantitative systems toxicology model developed over the last decade by the drug-induced liver injury (DILI)-sim Initiative, has provided novel insights regarding mechanisms underlying drug-induced liver injury and why animal models sometimes fail to accurately assess the liver safety liability of new drug candidates. For example, DILIsym, but not routine preclinical testing, predicted the human hepatotoxicity of the migraine drugs telcagepant and MK3207 that terminated their clinical development. DILIsym also predicted that the next in-class drug, ubrogepant, would be relatively safe for the liver; this prediction was prospectively confirmed in phase-3 clinical trials leading to FDA approval without liver safety warnings. DILIsym also identifies mechanisms underlying liver toxicity, and this information can identify patient-specific risk factors for drug-induced liver injury including drug–drug interactions and certain disease states, improving risk management and pharmacovigilance. DILIsym provides an example of how increased application of quantitative systems toxicology modeling should lead to more efficient development of new drugs.

Over the past two decades, our understanding of genetic heritability has been derived from candidate gene and genome-wide studies looking at common allelic variant associations. As our access to advanced genomics technologies increases, so too does the availability of pharmacogenomic data for predicting the risk of adverse drug reactions (ADRs). We now have the ability to look at the contribution of rare and even personal genomic variants on ADR risk. However, the increase in data will be accompanied by challenges in interpretation and implementation. This review looks at the current position of drug safety pharmacogenomics and discusses the challenges, as well as some possible future directions.

The science of in vitro studies has advanced dramatically over the last 100 years, particularly in regards to their application in toxicology. Recent developments such as in vitro three-dimensional cultures and microphysiological systems have offered the promise of even greater replication of in vivo function. Nonetheless, the challenge of validating a system's performance and extrapolating it's responses to those of an animal or human remains

Computational pharmacokinetic modeling methods, such as physiologically based pharmacokinetic (PBPK) modeling, have shown great promise for use in translational research as well as regulatory assessments. PBPK models are assumption-based simplifications of the complex biological system modeled and have high data demands for model parameterization and verification. However, unlike empirical models that rely on multiple observations from a single system, PBPK models uniquely allow for data to be obtained from multiple platforms (in silico, in vitro, and in vivo). Furthermore, these data are integrated by the principles of physiology and pharmacology/toxicology to make predictions in domains with sparse observations. Our article provides an overview of scientific utility of PBPK modeling in translational research and regulatory toxicology using some case examples that highlight the important role of PBPK model-based predictions in contributing to regulatory assessments of diverse types of chemicals, ranging from food and environmental chemicals to drugs intended for use in veterinary and human medicine. At present, collective efforts are ongoing for establishing uniformity, consistency, and transparency within many areas of PBPK modeling, and with continuing advances in the field of computational pharmacokinetic, PBPK modeling has the potential to contribute to reliable alternatives to animal testing in the future.

In vivo imaging has a potential to bring innovation to translational toxicology for drug discovery. This includes magnetic resonance imaging, positron emission tomography (PET), single photon emission tomography, and computed tomography. The utility of imaging comes from its capacity to provide minimally invasive biomarkers for safety profiling and decision-making in drug discovery. Nonspecific biomarkers, such as magnetic resonance imaging relaxometry, computed tomography density, or PET ¹⁸F-fluoro-2-deoxy-d-glucose are better suited for preclinical general toxicology and clinical monitoring, while specific ones, such as most of PET and single photon emission tomography ligands, are better suited to clarify mechanisms of toxicity or unwanted target engagement. However, the use of these biomarkers is sporadic and governed by scientific interest and availability rather than its utility. A systematic approach is needed to qualify these biomarkers with regulatory authorities so translational imaging could be incorporated into drug development and its unique potential translated into safer and cheaper medicines.

During the past decade, quantitative structure–activity relationship (QSAR) enjoyed an ever-increasing application in various fields including translational sciences. This review summarizes the progress in data preprocessing, processing, and validation techniques as well as the standardization in reporting of QSARs and the legislative framework promoting the use of computational approaches as viable tools for reducing animal testing. Software products focused on prediction of translational end-points and recently published individual models are discussed briefly. Particular attention is given to challenges springing from the immense complexity of translational QSARs.

The dog is an important species used in preclinical studies in support of human drug product development. Likewise, because of the many active pharmaceutical ingredients with therapeutic relevance to both humans and dogs, extrapolation can also occur in the reverse, from human to dog. In either situation, it is important to appreciate species-specific factors influencing drug pharmacokinetics (absorption, metabolism, disposition, and elimination) and the potential impact of disease on the applicability of these extrapolations. Furthermore, tools such as physiologically based pharmacokinetic models not only enable investigators to extrapolate species-specific data on systemic or organ exposure to the parent compound and metabolite(s) but also facilitate an interrogation of factors that can lead to species-specific differences in drug effectiveness and toxicity. In this review, we explore the factors and tools that comprise our current arsenal for understanding and predicting human–canine comparative toxicity.