From the AGA Journals

Bioengineered liver models screen drugs and study liver injury

Thirty to 50 new drugs are approved in the United States annually, which costs approximately $2.5 billion/drug in drug development costs. Nine out of 10 drugs never make it to market, and of those that do, adverse events affect their longevity. Hepatotoxicity is the most frequent adverse drug reaction, and drug-induced liver injury, which can lead to acute liver failure, occurs in a subset of affected patients. Understanding a drug’s risk of hepatotoxicity before patients start using it can not only save lives but also conceivably reduce the costs incurred by pharmaceutical companies, which are passed on to consumers.

Dr. Rotonya Carr Assistant professor of medicine, division of gastroenterology, University of Pennsylvania

Dr. Rotonya Carr

In Cellular and Molecular Gastroenterology and Hepatology, Underhill and Khetani summarize available and emerging cell-based, high-throughput systems that can be used to predict hepatotoxicity. These modalities include cellular microarrays of single cells; cocultures of liver parenchymal and nonparenchymal cells; organoids (3-D organ-like structures); and liver-on-a-chip devices (complex perfusion bioreactors that allow for modulation of the cellular micro-environment). These in vitro systems have not only enabled investigators to screen multiple drugs at the same time but also have informed the clinical translation of these technologies. For example, the extracorporeal liver assist device – essentially, a liver bypass – and similar bioartificial liver devices can in principal temporarily perform some of the major liver functions while a patient’s native liver heals from drug-induced liver injury or other hepatic injury.

However, just as we have seen with the limitations of the in vitro systems, bioartificial livers are unlikely to be successful unless they integrate the liver’s complex functions of protein synthesis, immune surveillance, energy homeostasis, and nutrient sensing. The future is bright, though, as biomedical scientists and bioengineers continue to push the envelope by advancing both in vitro and bioartificial technologies.

Rotonya Carr, MD, is an assistant professor of medicine in the division of gastroenterology at the University of Pennsylvania, Philadelphia. She receives research support from Intercept Pharmaceuticals.



Bioengineered liver models have enabled recapitulation of liver architecture with precise control over cellular microenvironments, resulting in stabilized liver functions for several weeks in vitro. Studies have focused on using these models to investigate cell responses to drugs and other stimuli (for example, viruses and cell differentiation cues) to predict clinical outcomes. Gregory H. Underhill, PhD, from the department of bioengineering at the University of Illinois at Urbana-Champaign and Salman R. Khetani, PhD, from the department of bioengineering at the University of Illinois in Chicago presented a comprehensive review of the these advances in bioengineered liver models in Cellular and Molecular Gastroenterology and Hepatology (doi: 10.1016/j.jcmgh.2017.11.012).

Drug-induced liver injury (DILI) is a leading cause of drug attrition in the United States, with some marketed drugs causing cell necrosis, hepatitis, cholestasis, fibrosis, or a mixture of injury types. Although the Food and Drug Administration requires preclinical drug testing in animal models, differences in species-specific drug metabolism pathways and human genetics may result in inadequate identification of potential for human DILI. Some bioengineered liver models for in vitro studies are based on tissue engineering using high-throughput microarrays, protein micropatterning, microfluidics, specialized plates, biomaterial scaffolds, and bioprinting.

High-throughput cell microarrays enable systematic analysis of a large number of drugs or compounds at a relatively low cost. Several culture platforms have been developed using multiple sources of liver cells, including cancerous and immortalized cell lines. These platforms show enhanced capabilities to evaluate combinatorial effects of multiple signals with independent control of biochemical and biomechanical cues. For instance, a microchip platform for transducing 3-D liver cell cultures with genes for drug metabolism enzymes featuring 532 reaction vessels (micropillars and corresponding microwells) was able to provide information about certain enzyme combinations that led to drug toxicity in cells. The high-throughput cell microarrays are, however, primarily dependent on imaging-based readouts and have a limited ability to investigate cell responses to gradients of microenvironmental signals.

Liver development, physiology, and pathophysiology are dependent on homotypic and heterotypic interactions between parenchymal and nonparenchymal cells (NPCs). Cocultures with both liver- and nonliver-derived NPC types, in vitro, can induce liver functions transiently and have proven useful for investigating host responses to sepsis, mutagenesis, xenobiotic metabolism and toxicity, response to oxidative stress, lipid metabolism, and induction of the acute-phase response. Micropatterned cocultures (MPCCs) are designed to allow the use of different NPC types without significantly altering hepatocyte homotypic interactions. Cell-cell interactions can be precisely controlled to allow for stable functions for up to 4-6 weeks, whereas more randomly distributed cocultures have limited stability. Unlike randomly distributed cocultures, MPCCs can be infected with HBV, HCV, and malaria. Potential limitations of MPCCs include the requirement for specialized equipment and devices for patterning collagen for hepatocyte attachment.

Next Article: