Transcriptional, Functional, and Mechanistic Comparisons of Stem Cell-Derived Hepatocytes, HepaRG Cells, and Three-Dimensional Human Hepatocyte Spheroids as Predictive In Vitro Systems for Drug-Induced Liver Injury

Abstract

Reliable and versatile hepatic in vitro systems for the prediction of drug pharmacokinetics and toxicity are essential constituents of preclinical safety assessment pipelines for new medicines. Here, we compared three emerging cell systems-hepatocytes derived from induced pluripotent stem cells, HepaRG cells, and three-dimensional primary human hepatocyte (PHH) spheroids-at transcriptional and functional levels in a multicenter study to evaluate their potential as predictive models for drug-induced hepatotoxicity. Transcriptomic analyses revealed widespread gene expression differences between the three cell models, with 8148 of 17,462 analyzed genes (47%) being differentially expressed. Expression levels of genes involved in the metabolism of endogenous as well as xenobiotic compounds were significantly elevated in PHH spheroids, whereas genes involved in cell division and endocytosis were significantly upregulated in HepaRG cells and hepatocytes derived from induced pluripotent stem cells, respectively. Consequently, PHH spheroids were more sensitive to a panel of drugs with distinctly different toxicity mechanisms, an effect that was amplified by long-term exposure using repeated treatments. Importantly, toxicogenomic analyses revealed that transcriptomic changes in PHH spheroids were in compliance with cholestatic, carcinogenic, or steatogenic in vivo toxicity mechanisms at clinically relevant drug concentrations. Combined, the data reveal important phenotypic differences between the three cell systems and suggest that PHH spheroids can be used for functional investigations of drug-induced liver injury in vivo in humans.

Fig. 1.

PHHs cultured in 3D spheroids resemble freshly isolated cells regarding expression patterns of drug-metabolizing enzymes, drug transporters, and hepatic markers. (A) Expression of phase I (CYP2C8, CYP2C9, CYP2D6, and CYP3A4) and phase II (GSTT1 and UGTA1) metabolic enzymes, drug transporters (SLCO1B1 and ABCB11), ligand-activated nuclear receptors (CAR, PXR, and PPARA) as well as the critical hepatic transcription factor HNF4A and the main hepatocyte secretory product, albumin (ALB), were quantified in PHH spheroids by quantitative polymerase chain reaction and normalized to expression in freshly isolated cells of the same donors (n = 3 to 4 donors; donor demographics are shown in Table 1). Importantly, with the exception of CYP2C8 (33% of expression of freshly isolated cells, P = 0.001) and CYP2C9 (40%, P = 0.004), no significant differences in expression levels between freshly isolated cells and PHH spheroids were detected. Error bars indicate S.E.M. **P < 0.01 (heteroscedastic two-tailed t test). (B) Expression levels of genes analyzed in (A) were elevated up to 1834-fold in the 3D spheroids compared with 2D cultured PHHs from the same donors after 7 days in culture. FC, Fold change; n.s., not significant (P > 0.05).

Fig. 2.

Transcriptomic profiling of hepatic in vitro models reveals wide-scale differences in global gene expression. (A) Heat map depicting differentially expressed genes in PHH spheroids (donor 1; blue), HepaRG cells (red), and hiPS-Hep cells (green) at 2, 7, and 14 days. Overall, 8148 of 17,462 genes analyzed were found to be differentially expressed after multiple testing correction (Benjamini–Hochberg FDR < 0.05). PHH spheroids showed elevated expression of genes involved in endogenous and xenobiotic metabolism (P_adjusted = 3 × 10⁻³³), whereas HepaRG and hiPS-Hep cells exhibited, among others, elevated transcript levels of genes involved in proliferation (P_adjusted = 0.0083) and ribosomes (P_adjusted = 0.0034). Average values of three technical triplicates are presented as mean centered and σ normalized. (B) Principal component analysis revealed clear separation of the three cell models, which even increased over time (time progression is indicated as shades of purple). Notably, temporal changes of the transcriptomic signatures were more evident for HepaRG and hiPS-Hep cells during the culture period (indicated by arrows), whereas the transcriptomes of PHH spheroids remained temporally stable. PC, principal component.

Fig. 3.

Expression levels of important ADME genes differ substantially between the three hepatic in vitro models. PHH spheroids, HepaRG cells, and hiPS-Hep cells showed pronounced expression differences in phase I enzymes (A), phase II enzymes (B), and drug transporters (C). Median expression values of three technical replicate microarray measurements are shown for each cell system and time point. Data are presented as mean centered and σ normalized. CES, Carboxylesterase; EPHX, Epoxide Hydrolase; FMO, Flavin Containing Monooxygenase; GPX, Glutathione Peroxidase; GSR, Glutathione-Disulfide Reductase; MGST, Microsomal Glutathione S-Transferase; NAT, N-Acetyltransferase; SULT, Sulfotransferase; TAP, ATP-Binding Cassette Transporter.

Fig. 4.

The sensitivity to model DILI compounds differs drastically between hiPS-Hep, HepaRG, and PHH cell models. (A) PHH spheroids, HepaRG cells, and hiPS-Hep cells were treated with APAP, aflatoxin B1, amiodarone, chlorpromazine, troglitazone, and ximelagatran in single-dose (48 hours, black) or repeated-exposure (7 days, brown; and 14 days, orange) experiments. Data are presented as the percentage relative to the viability of vehicle-treated controls at the same time point. For PHHs, two replicate experiments (both from donor 1) with six replicate measurements per concentration and time point are shown. For HepaRG cells, three replicate experiments with three replicate measurements per concentration and time point are shown. For hiPS-Hep cells, two replicate experiments with three replicate measurements per concentration and time point are shown. Error bars indicate S.E.M. (B) Semi-log plot showing the temporal evolution of sensitivity in PHHs (blue), HepaRG cells (red), and hiPS-Hep cells (green). Dashed lines indicate therapeutic exposure levels (human cmax). Note that long-term exposure resulted in increased sensitivity toward the hepatotoxins used in all cell systems. PHH spheroids detected toxicity at clinically relevant exposure levels for all compounds, with the exception of ximelagatran.

Fig. 5.

PHH spheroids constitute the most sensitive in vitro cell culture system tested. Heat map summarizing the sensitivities of the three cell systems to cytotoxicity, as shown in Fig. 4. Data are presented as mean centered and σ normalized and are related to therapeutic (ximelagatran and troglitazone) or toxic (APAP, aflatoxin B1, amiodarone, chlorpromazine) exposure values. Single, double, and triple asterisks indicate sensitivity < 30× C_max, < 10× C_max, and < 1× C_max, respectively. C_max or exposure values were obtained from the following references: APAP, 700 µM (Vale and Proudfoot, 1995); aflatoxin B1, 0.03 µM (Hassan et al., 2006); amiodarone, 3.9 µM (Regenthal et al., 1999); chlorpromazine, 1.6 µM, (Regenthal et al., 1999); troglitazone, 2.82 µM (Loi et al., 1999); and ximelagatran, 0.3 µM, (Schützer et al., 2004).

Fig. 6.

PHH spheroids faithfully mimic compound-specific transcriptional toxicity effects observed in vivo. Transcriptomic analyses of PHH spheroids treated chronically (14 days) with subtoxic concentrations (IC₁₀) of aflatoxin B1, amiodarone, and chlorpromazine. (A) Venn diagram showing significantly dysregulated genes compared with DMSO controls (Benjamini–Hochberg multiple testing correction, FDR < 0.05). Gene set enrichment analysis revealed that compound-specific toxicity responses (e.g., DNA damage-related pathways, perturbations of bile acid metabolism, and PPAR signaling) were detected in aflatoxin B1–, chlorpromazine-, and amiodarone-treated spheroids. (B–D) Targeted analysis of genes implicated in aflatoxin B1 (B), amiodarone (C), and chlorpromazine (D) toxicity in vivo. Genes whose expression was up- or downregulated in vivo are shown in shades of red and blue, respectively. (E) Expression of cellular ABC and SLC transporters was broadly inhibited upon chlorpromazine treatment. *P < 0.05; **P < 0.01; ***P < 0.001 (heteroscedastic two-tailed t test compared with DMSO control at the same time point). ELOVL = Elongation Of Very Long Chain Fatty Acids Protein, FC = Fold change.