Drug-induced liver injury (DILI) is a leading cause of drug attrition. Significant and well-documented differences between animals and humans in liver pathways now necessitate the use of human-relevant in vitro liver models for testing new chemical entities during preclinical drug development. Consequently, several human liver models with various levels of in vivo-like complexity have been developed for assessment of drug metabolism, toxicity, and efficacy on liver diseases. Recent trends leverage engineering tools, such as those adapted from the semiconductor industry, to enable precise control over the microenvironment of liver cells and to allow for miniaturization into formats amenable for higher throughput drug screening. Integration of liver models into organs-on-a-chip devices, permitting crosstalk between tissue types, is actively being pursued to obtain a systems-level understanding of drug effects. Here, we review the major trends, challenges, and opportunities associated with development and implementation of engineered liver models created from primary cells, cell lines, and stem cell-derived hepatocyte-like cells. We also present key applications where such models are currently making an impact and highlight areas for improvement. In the future, engineered liver models will prove useful for selecting drugs that are efficacious, safer, and, in some cases, personalized for specific patient populations.

Hyperglycemia in type 2 diabetes mellitus has been linked to non-alcoholic fatty liver disease, which can progress to inflammation, fibrosis/cirrhosis, and hepatocellular carcinoma. Understanding how chronic hyperglycemia affects primary human hepatocytes (PHHs) can facilitate the development of therapeutics for these diseases. Conversely, elucidating the effects of hypoglycemia on PHHs may provide insights into how the liver adapts to fasting, adverse diabetes drug reactions, and cancer. In contrast to declining PHH monocultures, micropatterned co-cultures (MPCCs) of PHHs and 3T3-J2 murine embryonic fibroblasts maintain insulin-sensitive glucose metabolism for several weeks. Here, we exposed MPCCs to hypo-, normo- and hyperglycemic culture media for ~3 weeks. While albumin and urea secretion were not affected by glucose level, hypoglycemic MPCCs upregulated CYP3A4 enzyme activity as compared to other glycemic states. In contrast, hyperglycemic MPCCs displayed significant hepatic lipid accumulation in the presence of insulin, while also showing decreased sensitivity to insulin-mediated inhibition of glucose output relative to a normoglycemic control. In conclusion, we show for the first time that PHHs exposed to hypo- and hyperglycemia can remain highly functional, but display increased CYP3A4 activity and selective insulin resistance, respectively. In the future, MPCCs under glycemic states can aid in novel drug discovery and mechanistic investigations.

Engineered liver models with varying levels of in vivo-like complexities provide investigators with the opportunity to develop assays with sufficient complexity and required throughput. Control over cell-cell interactions and co-culture with stromal cells in both two dimension and three dimension are critical for enabling stable liver models. The validation of liver models with diverse sets of compounds for different applications, coupled with an analysis of cost:benefit ratio, is important for model adoption for routine screening. Ultimately, engineered liver models could significantly reduce drug development costs and enable the development of more efficacious and safer therapeutics for liver diseases.

Hepatitis C virus (HCV) infection leads to persistence in the majority of cases despite triggering complex innate immune responses within the liver. Although hepatocytes are the preferred site for HCV replication, nonparenchymal cells (NPCs) can also contribute to antiviral immunity. Recent innovations involving single-genome amplification (SGA), direct amplicon sequencing, and phylogenetic inference have identified full-length transmitted/founder (T/F) viruses. Here, we tested the effect of HCV T/F viral RNA (vRNA) on innate immune signaling within hepatocytes and NPCs, including the HepG2 and Huh 7.5.1 cell lines, a human liver endothelial cell line (TMNK-1), a plasmacytoid dendritic cell line (GEN2.2), and a monocytic cell line (THP-1). Transfection with hepatitis C T/F vRNA induced robust transcriptional upregulation of type I and III interferons (IFNs) within HepG2 and TMNK-1 cells. Both the THP-1 and GEN2.2 lines demonstrated higher type I and III IFN transcription with genotype 3a compared to genotype 1a or 1b. Supernatants from HCV T/F vRNA-transfected TMNK-1 cells demonstrated superior viral control. Primary human hepatocytes (PHH) transfected with genotype 3a induced canonical pathways that included chemokine and IFN genes, as well as overrepresentation of RIG-I (DDX58), STAT1, and a Toll-like receptor 3 (TLR3) network. Full-length molecular clones of HCV induce broad IFN responses within hepatocytes and NPCs, highlighting that signals imparted by the various cell types within the liver may lead to divergent outcomes of infection. In particular, the finding that HCV genotypes differentially induce antiviral responses in NPCs and PHH might account for relevant clinical-epidemiological observations (higher clearance but greater necroinflammation in persistence with genotype 3).

Human liver models are useful for assessing compound metabolism/toxicity; however, primary human hepatocyte (PHH) lots are limited and highly variable in quality/viability. In contrast, cell lines, such as HepaRG, are cheaper and more reproducible surrogates for initial compound screening; however, hepatic functions and sensitivity for drug outcomes need improvement. Here, we show that HepaRGs cocultured with murine embryonic 3T3-J2 fibroblasts, previously shown to induce PHH functions, could address such limitations. We either micropatterned HepaRGs or seeded them ‘randomly’ onto collagen-coated plates before 3T3-J2 coculture. Micropatterned cocultures (HepaRG-MPCCs) secreted ∼2- to 4-fold more albumin and displayed more stable cytochrome-P450 activities than HepaRG conventional confluent monocultures (HepaRG-CCs) and micropatterned HepaRG monocultures (HepaRG-MPHs) for 4 weeks, even when excluding dimethyl sulfoxide from the medium. Furthermore, HepaRG-MPCCs had the most albumin-only positive cells (hepatic), lowest cytokeratin 19 (CK19)-only positive cells (cholangiocytic), and highest mean albumin intensity per cell than HepaRG random cocultures (HepaRG-RCCs) and monocultures; however, 80-84% of HepaRGs remained bipotential (albumin+/CK19+) across all models. The 3T3-J2s also induced higher albumin in HepaRG spheroids than HepaRG-only spheroids. Additionally, while rifampin induced CYP3A4 in HepaRG-MPCCs and HepaRG-CCs, only HepaRG-MPCCs showed the dual omeprazole-mediated CYP1A2/3A4 induction as with PHHs. Lastly, when treated for 6 days with 47 drugs and evaluated for albumin and ATP to make binary hepatotoxicity calls, HepaRG-MPCCs displayed a sensitivity of 54% and specificity of 100% (70%/100% in PHH-MPCCs), whereas HepaRG-CCs misclassified several hepatotoxins. Ultimately, HepaRG-MPCCs could be a more cost-effective and reproducible model than PHHs for executing a tier 1 compound screen.