Research Article: Pluripotent Stem Cell-Derived Human Tissue: Platforms to Evaluate Drug Metabolism and Safety

Date Published: February 3, 2018


Author(s): Jose Meseguer-Ripolles, Salman R. Khetani, Javier G. Blanco, Mairi Iredale, David C. Hay.


Despite the improvements in drug screening, high levels of drug attrition persist. Although high-throughput screening platforms permit the testing of compound libraries, poor compound efficacy or unexpected organ toxicity are major causes of attrition. Part of the reason for drug failure resides in the models employed, most of which are not representative of normal organ biology. This same problem affects all the major organs during drug development. Hepatotoxicity and cardiotoxicity are two interesting examples of organ disease and can present in the late stages of drug development, resulting in major cost and increased risk to the patient. Currently, cell-based systems used within industry rely on immortalized or primary cell lines from donated tissue. These models possess significant advantages and disadvantages, but in general display limited relevance to the organ of interest. Recently, stem cell technology has shown promise in drug development and has been proposed as an alternative to current industrial systems. These offerings will provide the field with exciting new models to study human organ biology at scale and in detail. We believe that the recent advances in production of stem cell-derived hepatocytes and cardiomyocytes combined with cutting-edge engineering technologies make them an attractive alternative to current screening models for drug discovery. This will lead to fast failing of poor drugs earlier in the process, delivering safer and more efficacious medicines for the patient.

Partial Text

Despite improvements in drug screening, there is still a high percentage of drug attrition during development. This presents in either in pre-clinical modeling, clinical trials or after drug approval, with greater expense incurred the further along the pipeline the compounds are removed. Therefore, fast failing is key to improving the success and the cost of human drug development. The percentage of drug failure at phase II and phase III is high and the main reasons for failure are the lack of efficacy, 48% in phase II and 55% in phase III, and safety, 25% in phase II and 14% in phase III (1). A recent study analyzed the main reasons for a drug withdrawn from the market because of adverse effects from 1950 to 2014. Hepatotoxicity (18%) represented the first reason for drug withdrawal followed by immune-related reactions (17%) and with cardiotoxicity third (14%). Hepatotoxicity and cardiotoxicity represent serious concerns in drug development. Side toxic effects are often detected at later stages of the development or even after the drug approval. Because of that, there is a need to improve current screening models to improve the early detection of hepatotoxic and cardiotoxic drugs (2). Although high-throughput screening platforms permit the testing of large compound libraries during drug development, the high attrition rates demonstrate the need for improved screening platforms and more reliable pre-clinical models. An essential component of this is to improve model fidelity (for a detailed review see (3)). Key to this is our ability to recapitulate organ physiology ‘in the dish’. Improvements in this space will likely lead to improved safety, efficacy and reduced development costs (3).

The liver plays a central role in drug disposition; it is responsible for drug uptake, metabolism and excretion. Several factors are involved in the off-target effects of drugs with differential metabolism playing a key role. Cytochrome P450 enzymes play an important roles during drug metabolism, with five family members (CYP1A2, CYP2C9, CYP2C19, CYP2D6, and CYP3A4) responsible for the metabolism of approximately 90% of marketed drugs. Genetic polymorphisms in the CYP450 family members affects drug metabolism, efficacy and safety (13,14). Polymorphisms in phase II enzymes, such as UDP-glucuronosyltransferases, N-acetyltransferases, and glutathione S-transferases, and ABC transporters are also known to influence metabolism and drug exposure (for a review please see (14)). Following drug metabolism, the metabolites are an important concern. Those can be active and provide patient benefit; however, they also expose the organ to adverse events, including endoplasmic reticulum stress (19,20). This can lead to alterations in cell signaling pathways that can alter the cell fate upon toxic insult such as NF-κB and Nrf-2 (15,16). Mitochondrial stress is also evident, altering cellular ATP and reactive oxygen species levels triggering cell death pathways (17,18). Drug metabolism and cell stress are therefore key concerns during the drug development process to reduce possible side effects of new drugs.

The gold standard model for the study of drug metabolism during drug development is the primary hepatocyte. The main disadvantages of primary hepatocytes are their rapid loss of phenotype post isolation and isolation costs (423). Therefore, researchers have searched for more accessible and cheaper alternatives. Cancer-derived cell lines, such as HepG2, HuH-7, Hep3B, or Fa2N-4, and HepaRG have been used to characterize some determinants of dug metabolism (24,27,28). Pluripotent cell-derived models have been proposed as an alternative cell source for screening (6,29,30). In many cases, pluripotent cell-derived models exhibit drug sensitivities patterns similar to primary cells (5,7,31–33). Moreover, the use of pluripotent stem cells allows the user to derive somatic cells from defined background, thereby offering insight into idiosyncratic DILI (34,35). To date, most of the work has focused on monolayer hepatocyte systems derived from induced pluripotent stem cells, through defined and reproducible differentiation protocols. Despite these advances, monolayer cultures of hepatocytes face significant limitations and do not emulate the complexity of the liver in terms of tissue organization, blood flow, and different cell-cell interactions. To overcome these limitations, organoid or spheroid models have been developed showing promising results (36,37). Although they require more complex differentiation protocols, organoids/ spheroids better recapitulate human tissue structure and display more mature and functional phenotype such as improved cytochrome P450 3A4 activity, greater expression of Phase II and III enzymes, combined with reduced fetal gene expression and longer lifespan (38). While promising the current challenges that face the B3D field^ is cost-effective manufacture, experimental reproducibility and automated scale-up for application.

Despite the advances, further refinement is required to better model the physiology of the organ of interest. Doing so will increase specificity and sensitivity of the screening models, thereby reducing the potential for off target drug events and failure. The ability to move beyond the current limitations requires interdisciplinary collaboration. By combining the best stem cell models with chemistry, physics, and engineering, new automated screening assays with improved function and physiology can be developed. Current areas of promise are discussed in the final section.

We believe that with the recent advances in stem cell differentiation, scale up, and performance, it is possible to create more accurate human tissue models for drug development. Advances in stem cell biology over the last decade has provided the field with more accurate human cell-based models that recapitulate key aspects of human drug metabolism, with better precision than cancer lines. These systems also provide comparable activity to primary cells. While this is encouraging, further improvements are necessary to improve predictive power. Tissue engineering has already played an important role in this space, with organ-on-chip devices now available via several commercial sources. Future efforts in the field should focus on developing high-throughput multi-organ systems, capable of real-time monitoring and multiplexing to reduce costs and improve the quality of data output.




Leave a Reply

Your email address will not be published.