Human Three-Dimensional Hepatic Models: Cell Type Variety and Corresponding Applications

Abstract

Owing to retained hepatic phenotypes and functions, human three-dimensional (3D) hepatic models established with diverse hepatic cell types are thought to recoup the gaps in drug development and disease modeling limited by a conventional two-dimensional (2D) cell culture system and species-specific variability in drug metabolizing enzymes and transporters. Primary human hepatocytes, human hepatic cancer cell lines, and human stem cell-derived hepatocyte-like cells are three main hepatic cell types used in current models and exhibit divergent hepatic phenotypes. Primary human hepatocytes derived from healthy hepatic parenchyma resemble in vivo-like genetic and metabolic profiling. Human hepatic cancer cell lines are unlimitedly reproducible and tumorigenic. Stem cell-derived hepatocyte-like cells derived from patients are promising to retain the donor's genetic background. It has been suggested in some studies that unique properties of cell types endue them with benefits in different research fields of in vitro 3D modeling paradigm. For instance, the primary human hepatocyte was thought to be the gold standard for hepatotoxicity study, and stem cell-derived hepatocyte-like cells have taken a main role in personalized medicine and regenerative medicine. However, the comprehensive review focuses on the hepatic cell type variety, and corresponding applications in 3D models are sparse. Therefore, this review summarizes the characteristics of different cell types and discusses opportunities of different cell types in drug development, liver disease modeling, and liver transplantation.

Keywords: drug development; hepatic cell types; hepatocyte transplantation; in vitro 3D model; liver disease.

FIGURE 1

Cellular composition of the liver. (A) Gross structure and blood supplies of the liver. The liver is a dark reddish-brown organ supplied by two distinct blood sources: oxygenated blood from the hepatic artery (HA) and nutrient-rich blood from the hepatic portal vein (PV). (B) Hepatic lobules are composed of hepatocytes arranged in linear cords radiating out from the central vein (CV) and portal triads including the bile duct (BD), HA, and PV. (C) The representative hepatic functional unit in hexagonal hepatic lobules is composed of diverse cell types. Besides parenchymal cells, non-parenchymal cells also support the hepatic structure and function. Hepatic blood vessels are lined by specialized fenestrated liver sinusoidal endothelial cells (LSECs). Kupffer cells (KCs) are macrophages found in the sinusoids. Hepatic stellate cells (HSCs) locate in the space of Disse. Cholangiocytes line the bile ducts. (D) Hepatic cell types used in in vitro 3D models can be obtained from healthy hepatic tissue, hepatocellular carcinoma, induced somatic cells, and human embryos. (E) Advanced 3D culture techniques. Scaffold-free, scaffold-based, and 3D bioprinting techniques are applied to build 3D models. (F) Current 3D hepatic models include spheroid, liver-on-a-chip, micropatterned co-culture (MPCC), and organoids models. (G) Corresponding applications of 3D hepatic models derived by various cell types. Developed diseased or healthy 3D models are promising in the areas of hepatotoxin screening, idiosyncratic drug reaction study, diverse disease modeling, and hepatocyte transplantation.

FIGURE 2

3D cell culture paradigms support disease modeling of NAFLD. (A) Representative pictures of lipid accumulated in PHH spheroids upon exposure to cyclosporine A (30 μM) for 48 h; scale bars, 100 μm. (B) Illustration of the PHH liver-on-a-chip model. (C) Intracellular fat accumulation in the liver-on-a-chip model under fat condition (Lower) was noted after Oil Red O staining. (D) Schematic demonstration of stiffness measurement by AFM. The top region of each single organoid (14 × 14 matrix in a 25 × 25 μm square) was scanned with an AFM cantilever. Scale bar, 100 μm. Young’s modulus on each scanned spot is shown in the heatmap. (A) Modified with permission from Bell et al. (2016). (B,C) Modified with permission from Kostrzewski et al. (2017). (D) Modified with permission from Ouchi et al. (2019).

FIGURE 3

Selected representative 3D models of viral hepatitis B. (A) Schematic representation of the HBV life cycle in infected hepatocytes. (B) Illustration of the PHH liver-on-a-chip model. Microfluidic recirculation was driven by a micro-pump and went through collagen-coated polystyrene scaffolds seeded with PHHs. (C) Immunofluorescence of HBV viral antigens (HBsAg and HBcAg) 10 days following infection of the PHH liver-on-a-chip model with patient-derived HBV (100 GE/cell). (D) (Left) Schematic representation of the protocol for liver organoid generation and differentiation from single donor umbilical cord (UC)-derived hiPSCs, HUVECs, and mesenchymal cells (MCs). (Right) Morphology of single donor cell–derived liver organoids (SDC-LOs) on days 1 and 15. Scale bars, 100 μm. (E) Immunofluorescence staining of ALB and NTCP in differentiated SDC-LOs. Green, ALB; red, NTCP. Scale bars, 50 μm. (A) Modified with permission from Yan et al. (2012). (B,C) Modified with permission from Ortega-Prieto et al. (2018). (D,E) Modified with permission from Nie et al. (2018).

FIGURE 4

Selected representative 3D models of monogenic diseases. (A) Representative bright-field and lipid-fluorescent images of organoids from Wolman disease patients and non-Wolman donors after oleic acid treatment. (B) (Left) Comparison of the P3NP secretion level of non-Wolman and Wolman human liver organoids (HLOs). (Right) Comparison of Young’s modulus of non-Wolman and Wolman organoids, n = 7. (C) Differentiated liver organoids from MM, MZ, and ZZ patients. Specific detection of total AAT protein with anti-AAT-B9 and AAT polymers with anti-AAT-D11 is shown in green fluorescence. (D) ELISA measurement of A1AT secretion in supernatants from healthy donor and patient organoids after 11 days of differentiation. (E–H) Immunohistochemistry for A1AT on liver tissue (E,G) and liver-derived organoids from a healthy donor (F) and a representative A1AT deficiency patient (H). Black arrows indicate A1AT proteins aggregated in patient-derived liver tissue (G) and organoids (H). Scale bars, 20 mm. All results are represented as mean ± SEM. Figure modified with permission from Ouchi et al. (2019), Gomez-Mariano et al. (2020), and Huch et al. (2015).