Metabolic disorders, including nonalcoholic fatty liver disease (NAFLD), metabolic syndrome, type 2 diabetes, and obesity have reached epidemic proportions. These diseases are all related to dysfunction of the liver, which performs over 500 critical functions including lipid metabolism and blood glucose regulation. Moreover, NAFLD, type 2 diabetes, metabolic syndrome, and obesity are closely connected (Figure 1). For example, NAFLD is a risk factor for the development of type 2 diabetes, which, in turn, is a major contributor to progressive inflammation of the liver (nonalcoholic steatohepatitis [NASH]). Metabolic syndrome—linked with obesity—is associated with the risk of developing type 2 diabetes and cardiovascular diseases. Additionally, inherited metabolic disorders, like familial hypercholesterolemia, alpha-1-antitrypsin deficiency, and glycogen storage diseases, involve defective genes that cause metabolic enzyme deficiencies and liver dysfunction.

Figure 1. The relationships between metabolic diseases and liver dysfunction. Dysfunction in the liver can lead to interrelated metabolic disorders. Metabolic syndrome, defined as abdominal obesity and elevated blood pressure, plasma glucose, serum triglycerides, or low high-density-lipoprotein levels, can be caused by liver dysfunction and can lead to type 2 diabetes. Dysfunction in the liver can also lead to nonalcoholic fatty liver disease (NAFLD), which occurs when fat is deposited (steatosis) in the liver due to metabolic dysfunction rather than alcohol use. NAFLD can progress to nonalcoholic steatohepatitis (NASH) and ultimately cirrhosis. Patients with NAFLD can often develop insulin resistance, and patients with diabetes can develop NASH, demonstrating a distinct role of the liver in mediating complex interactions between these interrelated metabolic diseases (Tilg, Moschen, and Roden 2016).

The molecular mechanisms of metabolic disease progression in the liver are not well understood, and in vitro disease models are crucial to their discovery. Predictive in vitro models should exhibit metabolic characteristics linked to hepatic insulin resistance, lipid metabolism, and accumulation of free fatty acids. Currently, primary hepatocytes are the most common cell model, but their utility is constrained by large variation between donors and a finite number of cells harvestable from each donor.

Eliminating these constraints, hepatocytes differentiated from human induced pluripotent stem (hiPS) cells provide an unlimited supply of functional hepatocytes from each parental cell line, which can be derived noninvasively from a diseased or healthy person. This model is particularly amenable to gene editing technology, which makes it possible to design cells with specific disease-relevant mutations—a boon for those studying the cause and progression of metabolic diseases of the liver. However, hiPS cell-derived hepatocytes are only useful as an in vitro model system and for gene editing applications if they recapitulate critical hepatocyte functions.

Until recently, the functionality of hiPS cell-derived hepatocytes was insufficient for modeling metabolic processes. To address these problems, we developed a robust hepatocyte differentiation protocol—which promotes a highly synchronized differentiation pattern across multiple hiPS cell lines (Ghosheh et al. 2016)—and a novel maintenance medium, which enables the long-term culture of our hiPS cell-derived hepatocytes. To create these cells, we terminally differentiated three hiPS cell lines (ChiPSC12, ChiPSC18, and ChiPSC22; abbreviated C12, C18, and C22) from different individuals into three lines of Cellartis enhanced hiPS-HEP cells; one of these lines is included in each complete kit. With up to three renewable, reliable sources of cells, perform functional studies reflecting interindividual variation and be confident across experiments due to low batch-to-batch variability. Cellartis enhanced hiPS-HEP cells display adult characteristics and have substantial drug-metabolizing functionality that can be maintained for two weeks. Here, we present data demonstrating the suitability of these hiPS cell-derived hepatocytes for modeling the role of hepatocytes in metabolic disorders.

Enhanced hiPS-HEP cells express hepatocyte markers and display functional characteristics of mature hepatocytes

Hepatocyte nuclear factor 4α (HNF4α) is a transcription factor required for liver development and for controlling the expression of liver-specific genes, and it is associated with several critical metabolic pathways (Gonzalez 2008). HNF4α immunostaining showed over 90% of hepatocytes differentiated with our protocol expressed HNF4α (Asplund et al. 2016). Asialoglycoprotein receptor 1 (ASGPR1) is another liver-specific cell surface protein and a marker for mature hepatocytes (Takayama et al. 2014; Peters et al. 2016). In our experiments, both HNF4α and ASGPR1 are well-expressed in enhanced hiPS-HEP cells from all three hiPS cell lines (Figure 2), as shown by homogeneous nuclear staining for HNF4α and homogeneous membrane staining for ASGPR1. Notably, the mRNA expression levels of HNF4α and ASGPR1 in enhanced hiPS-HEP cells on Day 12 post-thawing are comparable to levels in cryopreserved human primary hepatocytes (hphep cells) at Day 1 post-thawing, indicating that the enhanced hiPS-HEP cells express these hepatocyte markers at high levels over an extended culture time.

Figure 2. Enhanced hiPS-HEP cells express mature hepatocyte markers. Panel A. Staining for HNF4α and ASGPR1 was performed on enhanced hiPS-HEP cells from three different hiPS cell lines (C12, C18, and C22) on Day 12 post-thawing and cryopreserved human primary hepatocytes (hphep cells) on Day 1 post-thawing. Panel B. Similarly, staining for ASGPR1 was performed on enhanced hiPS-HEP cells from C18 on Day 12 and hphep cells on Day 1. Scale bar = 50 µm. Panel C. mRNA expression analysis of HNF4α and ASGPR1, as quantified by transcriptome analysis, was performed on enhanced hiPS-HEP cells from C12, C18, and C22 on Day 13 post-thawing (n = 2 batches per cell line), and on hphep cells on Day 1 post-thawing (n = 3 donors). Data are presented as mean values of log₂ intensity +/– standard deviation.

Of the many functions performed by mature hepatocytes in vivo, synthesis and secretion of both albumin and urea are commonly used to evaluate hepatocyte in vitro models. Albumin regulates the oncotic pressure of blood, and low serum albumin levels can indicate liver cirrhosis or chronic hepatitis. Urea is a normal byproduct of protein breakdown by the liver, and impaired urea synthesis can also indicate liver dysfunction. Another characteristic used to evaluate hepatocyte models is the presence of functional alpha-1-antitrypsin (α1AT), a protease inhibitor produced in the liver. α1AT deficiency leads to chronic tissue breakdown and liver damage.

Albumin and α1AT are well-expressed in enhanced hiPS-HEP cells from all three hiPS cell lines (Figure 3, Panels A, B, and C), as shown by immunostaining and mRNA expression analysis. Notably, only a subset of cells is strongly immunopositive for albumin in enhanced hiPS-HEP cells and in hphep cultures, in agreement with the metabolic zonation (Jungermann and Kietzmann 2000) observed in liver slices (Figure 3, Panel A).

Additionally, enhanced hiPS-HEP cells express genes involved in the urea cycle within the same range as hphep cells (Figure 3, Panel D), and the cells secrete albumin and urea (Figure 3, Panels E and F). Albumin secretion in enhanced hiPS-HEP cells on Days 4, 6, 12, and 20 post-thawing occurs at similar or higher levels as in hphep cells on Day 1 post-thawing. Urea secretion on Days 13 and 20 post-thawing is lower than in hphep cells on Day 1 post-thawing, and it increases over time.

Figure 3. Enhanced hiPS-HEP cells display functional characteristics of mature hepatocytes. Staining for albumin (Panel A) and α1AT (Panel B) was performed on cryopreserved human primary hepatocytes (hphep cells) on Day 1 post-thawing, on enhanced hiPS-HEP cells from C18 on Day 12 post-thawing, and on a human liver section (albumin only). Scale bar = 50 µm. Panel C. mRNA expression analysis of ALB and α1AT, as quantified by transcriptome analysis, was performed in enhanced hiPS-HEP cells from C12, C18, and C22 on Day 13 post-thawing (n = 2 batches per cell line) and compared to hphep cells on Day 1 post-thawing (n = 3 donors). Data are presented as mean values of log₂ intensity +/– standard deviation. Panel D. mRNA expression analysis of urea cycle enzymes, as quantified by transcriptome analysis, was performed in enhanced hiPS-HEP cells from C12, C18, and C22 on Day 13 post-thawing (n = 2 batches per cell line) and compared to hphep cells on Day 1 post-thawing (n = 3 donors). Data are presented as mean values of log₂ intensity +/– standard deviation. Abbreviations: CPS1 = carbamoyl-phosphate synthase; OTC = ornithine carbamoyltransferase; ASS1 = argininosuccinate synthase; ASL = argininosuccinate lyase; and ARG1 = arginase-1. Panel E. Albumin secreted into the medium during a 24-hr period was measured by ELISA in enhanced hiPS-HEP cells from C12, C18, and C22 (two technical replicates per cell line) on Days 4, 6, 12, and 20 post-thawing and in hphep cells (n = 3 donors) on Day 1 post-thawing. Data represents normalization of albumin secretion to total protein content. Panel F. Urea secreted into the medium during a 24-hr period was measured by ELISA in enhanced hiPS-HEP cells from C12, C18, and C22 (two technical replicates per cell line) on Days 13 and 20 post-thawing and in hphep cells (n = 3 donors) on Day 1 post-thawing. Data represents normalization of urea secretion to total protein content.

Enhanced hiPS-HEP cells show functional glucose regulation

Hepatocytes carry out energy metabolism by regulating glucose production. When blood glucose levels are high, hepatocytes respond to insulin by increasing glycogen storage, decreasing gluconeogenesis, and decreasing glycogenolysis. Conversely, when blood glucose levels are low, hepatocytes respond to glucagon and glucocorticoids by decreasing glycogen storage and producing glucose via gluconeogenesis and glycogenolysis. In NAFLD, metabolic syndrome, and type 2 diabetes, hepatocytes become insulin resistant, and glucose builds up in the blood.

A relevant hepatocyte model for metabolic diseases should demonstrate normal insulin response and functional glucose regulation. Here, we show that enhanced hiPS-HEP cells respond to insulin with phosphorylation of protein kinase B-α (Akt), even at low insulin concentrations (Figure 4, Panels E and F), and that the genes involved in glycogen metabolism, gluconeogenesis, and insulin signaling are expressed at similar levels as in hphep cells (Figure 4, Panels B, C, and D). Furthermore, we show that enhanced hiPS-HEP cells can store glycogen (Figure 4, Panel A). Notably, both in hphep cells and enhanced hiPS-HEP cultures, only a subset of hepatocytes is strongly stained for glycogen storage—again in agreement with the metabolic zonation observed in the liver lobe (Figure 3, Panel A).

Figure 4. Enhanced hiPS-HEP cells demonstrate normal insulin response and functional glucose regulation. Panel A. Glycogen storage was visualized in enhanced hiPS-HEP cells from C18 fixed on Day 12 post-thawing and hphep cells fixed on Day 1 post-thawing by Periodic acid-Schiff (PAS) staining. Scale bar = 100 µm. Panels B–D. mRNA expression analysis of genes involved in glycogen metabolism (Panel B), gluconeogenesis (Panel C), and insulin signaling (Panel D), as quantified by transcriptome analysis, was performed in enhanced hiPS-HEP cells from C12, C18, and C22 on Day 12 post-thawing (n = 2 batches per cell line) and compared to gene expression measurements performed on hphep cells on Day 1 post-thawing (n = 3 donors). Data are presented as mean values of log₂ intensity +/– standard deviation. Abbreviations: AGL = glycogen debranching enzyme; GSK3A/B = glycogen synthase kinase 3 α/β; GYS2 = glycogen synthase 2; PYGL = glycogen phosphorylase, liver form; PCK1/2 = phosphoenolpyruvate carboxykinase 1/2; FBP1/2 = fructose-1,6-bisphosphatase 1/2; G6PC = glucose-6-phosphatase; G6PD = glucose-6-phosphate 1-dehydrogenase; GLUT-2 = glucose transporter 2; INSR = insulin receptor; IRS1/2 = insulin receptor substrate 1/2; PIK3CA/B/D = PI3-kinase subunit α/β/δ; and AKT1 = protein kinase B-α. Panel E. The effect of insulin on phosphorylation of Akt protein was measured by Western blot in enhanced hiPS-HEP cells from C18 on Day 12 post-thawing (n = 3 for the [+] condition; n = 1 for the [–] control). Panel F. Akt phosphorylation at pS473 was also measured by ELISA in enhanced hiPS-HEP cells from C18 on Day 6 post-thawing (n = 1 for each treatment concentration). The 0-nM and 2-nM insulin treatment conditions were tested, but did not result in measurable Akt phosphorylation. For the experiments represented in Panels B–E, data are presented as mean values +/– standard deviation.

Enhanced hiPS-HEP cells show functional lipid metabolism

In a healthy person, the liver transports fatty acids, sterols, and lipoproteins into and out of the liver in response to feeding and fasting. In various disease states, metabolic processes are dysfunctional and the liver begins to accumulate these substances. Normally, the liver regulates the level of cholesterol, which is carried to and from tissues by lipoproteins in the blood, by taking up low-density lipoproteins (LDL) and secreting very-low-density lipoproteins (VLDL) and high-density lipoproteins (HDL). Proprotein convertase subtilisin/kexin type 9 (PCSK9) is an enzyme that regulates plasma cholesterol homeostasis through its interaction with the LDL receptor (LDLR). When an LDL particle (carrying cholesterol) binds to the LDLR, the particle is trafficked into the hepatocyte, and PCSK9 targets the receptor for lysosomal degradation. If PCSK9 is blocked, the LDL receptor is recycled back into the cell membrane and can remove additional LDL particles from the extracellular fluid. Thus, PCSK9 is an excellent target for clinical inhibitors that lower blood LDL concentration and therefore cholesterol levels; in fact, two different drugs were approved for this purpose by the US Food and Drug Administration in 2015.

Here, we show that enhanced hiPS-HEP cells express principal genes involved in uptake, synthesis, and beta-oxidation of fatty acids at similar levels as hphep cells (Figure 5, Panel C). Furthermore, enhanced hiPS-HEP cells express high levels of the LDL receptor (Figure 5, Panel B) and take up fluorescently labeled LDL (Figure 5, Panel A). Other genes relevant for regulating blood cholesterol levels are well-expressed in enhanced hiPS-HEP cells, e.g., apolipoprotein B (a VLDL), apolipoprotein A1 (an HDL), PCSK9, sterol regulatory element-binding proteins 1 and 2 (SREBP-1 and -2), and lipoprotein lipase (LPL), which hydrolyzes triglycerides in lipoproteins into free fatty acids and glycerol. Taken together, enhanced hiPS-HEP cells express multiple genes that are important for lipid metabolism.

Figure 5. Enhanced hiPS-HEP cells demonstrate important features of lipid metabolism. Panel A. Enhanced hiPS-HEP cells from C12 and C18 were incubated for 3 or 24 hr with LDL-DyLight immediately prior to fixation on Day 6 post-thawing. Scale bar = 25 µm. Other fixation timepoints (Days 4 and 12 post-thawing) showed similar LDL uptake (data not shown). Panels B and C. mRNA expression analysis of genes involved in blood cholesterol regulation and fatty acid metabolism, as quantified by transcriptome analysis, was performed in enhanced hiPS-HEP cells from C12, C18, and C22 on Day 12 post-thawing (n = 2 batches per cell line) and compared to hphep cells on Day 1 post-thawing (n = 3 donors). Data are presented as mean values of log₂ intensity +/– standard deviation. Abbreviations: LDLR = LDL receptor; APOB/A1 = apolipoprotein B-100/A1; PCSK9 = proprotein convertase subtilisin/kexin type 9; SREBP-1/2 = sterol regulatory element-binding protein 1/2; LPL = lipoprotein lipase; FATP2/4/5 = fatty acid transporter protein 2/4/5; FASN = fatty acid synthase; SCD5 = stearoyl-CoA desaturase 5; ACADL = acyl-CoA dehydrogenase; L-FABP = liver fatty acid-binding protein; and CPT1A = carnitine O-palmitoyltransferase 1.

Modeling nonalcoholic fatty liver disease (NAFLD)

NAFLD is an imbalance between the uptake and removal of lipids in hepatocytes, which leads to abnormal triglyceride accumulation, or steatosis. In late-stage NAFLD, lipid accumulation generates chronic ER stress which, in turn, results in inflammation. This is known as nonalcoholic steatohepatitis (NASH) (Zhang et al. 2014, Zhang and Kaufman, 2008). ER stress activators, such as thapsigargin, can be used as a model to see if ER stress results in elevated levels of the inflammation and fibrosis marker TNFα.

Here we show that enhanced hiPS-HEP cells, when treated with high concentrations of oleic acid (OA) or lower concentrations of oleic acid in combination with thapsigargin for 24 hr, accumulate triglycerides in lipid droplets (Figure 6, Panel A), and display elevated expression of TNFα mRNA (Figure 6, Panel B).

Figure 6. Accumulation of triglycerides and elevated levels of the inflammation marker TNFα upon induction of steatosis. Enhanced hiPS-HEP cells from C18 on Day 6 post-thawing were incubated for 24 hr with BSA only as a vehicle control, oleic acid (OA) coupled to BSA at 200 µM—alone or in combination with 1 µM thapsigargin—and oleic acid coupled to BSA at 600 µM. Panel A. Fixed cells were stained with Oil Red O, resulting in red staining of lipid droplets in the cells. Scale bar = 50 µm. Other fixation timepoints (Days 4 and 12 post-thawing) showed similar levels of Oil Red O staining (data not shown). Similar results were also achieved with enhanced hiPS-HEP cells from C12 and C22 (data not shown). Panel B. As measured by qPCR, mRNA expression of TNFα was increased in response to oleic acid with and without thapsigargin as compared to the control. For each condition tested, n = 3, and data are presented as mean values +/– standard deviation.

Human iPS cell-derived hepatocytes differentiated with our robust differentiation protocol and cultured using a novel maintenance medium provide an inexhaustible, consistent supply of functional hepatocytes that can be used to advance the understanding of diseases related to dysfunction in liver metabolism, including NAFLD/NASH, type 2 diabetes, and metabolic syndrome. Like human primary hepatocytes, Cellartis enhanced hiPS-HEP cells show important characteristics of mature hepatocytes such as expression of HNF4α, ASGPR1, and α1AT; albumin and urea secretion; and principal features of functional glucose and lipid regulation. Furthermore, enhanced hiPS-HEP cells respond to steatosis-inducing conditions with an inflammatory response that mimics progressing NAFLD. With full cellular functionality obtainable on Day 5, the v2 kits can achieve a >14-day assay window, typically from Day 5 through at least Day 19. Enhanced hiPS-HEP cells are available from three different iPSC lines derived from healthy donors.

Cell culture

Cryopreserved enhanced hiPS-HEP cells derived from ChiPSC12, ChiPSC18, and ChiPSC22 cell lines (abbreviated C12, C18, and C22) were thawed, plated, and maintained according to the user manual for the Cellartis enhanced hiPS-HEP v2 kits. Cells were maintained for up to 21 days post-thawing with media changes every second or third day. Cryoplateable human primary hepatocytes (BioreclamationIVT) were thawed and plated according to the manufacturer's protocol.

Transcriptome analysis

Total RNA from Cellartis enhanced hiPS-HEP cells from C12, C18, and C22 (n = 2 batches per cell line) was extracted on Day 13 post-thawing as well as from hphep cells (n = 3 donors) Day 1 post-thawing using the GenElute RNA/DNA/Protein Plus Purification Kit (Sigma Aldrich). Transcriptome data was generated using the Affymetrix Gene 2.0 ST Array. Raw microarray data was imported into R software and signal intensities were normalized by the Robust Multichip Average method in the oligo package (Carvalho and Irizarry 2010). Microarray probe IDs were converted to HGNC gene symbols by Ensembl BioMart (release 90, assembly GRCh38) using the biomaRt package (Yates et al. 2016; Durinck et al. 2009).

Immunocytochemistry

Cells were stained as described previously (Ulvestad et al. 2013). Cellartis enhanced hiPS-HEP cells were fixed on Day 12 post-thawing and hphep cells on Day 1 post-thawing, by a 15-min incubation with 4% formaldehyde. Cells were stained with the following primary and secondary antibodies: rabbit anti-α1AT (1:200); rabbit anti-HNF4α (1:300); rabbit anti-albumin (1:1,000); mouse anti-ASGPR (1:50); donkey anti-rabbit IgG, Alexa Fluor 594 (1:1,000); or goat anti-mouse, Alexa Fluor 488 (1:1,000). Representative pictures were merged with the respective DAPI staining using Image J software.

Protein quantification

Cells were washed once with PBS and lysed in 0.02 mM NaOH for 16 hr at 4°C, and then stored at –20°C until analysis. Protein levels were quantified using the Pierce BCA Protein Assay Kit (Thermo Fisher Scientific) according to the manufacturer's protocol.

Periodic acid-Schiff (PAS) staining

Glycogen storage was visualized by PAS staining of fixed enhanced hiPS-HEP cells (C18, Day 12 post-thawing) and hphep cells (Day 1 post-thawing). Cells were stained with the Glycogen Assay Kit (Sigma) according to the manufacturer's protocol.

LDL uptake

To estimate the uptake of low-density lipoproteins (LDL) in enhanced hiPS-HEP cells (from C12, C18, and C22) on Days 4, 6, and 12 post-thawing, cells were incubated for 3 or 24 hr with LDL-DyLight (1:100). Cells were washed once with PBS and immunofluorescence was documented.

Oil Red O staining

Enhanced hiPS-HEP cells were incubated on Day 6 post-thawing for 24 hr with BSA only as a vehicle control, oleic acid coupled to BSA at 200 µM, alone or in combination with thapsigargin, and oleic acid at 600 µM. Fixed cells (see section on immunocytochemistry) were stained with Oil Red O (0.005 g ORO/ml of 2-propanol diluted 3:2 in distilled water) for 30 min. Next, cells were harvested and mRNA expression of TNFα was analyzed using qRT-PCR.

Western blot

Enhanced hiPS-HEP cells (from C18, Day 12 post-thawing) were incubated in insulin-free medium for 3 hr and then treated for 10 min with 0 nM or 100 nM insulin. Phosphorylated AKT (Cell Signaling) and total AKT (Cell Signaling) were quantified by Western blot (NuPAGE 4-12% Bis-Tris Protein Gels, Thermo Fisher Scientific).

Albumin secretion

Albumin secretion from enhanced hiPS-HEP cells (from C12, C18, and C22) was analyzed on Days 4, 6, 12, and 20 post-thawing and from hphep cells 24 hr post-thawing. The culture medium was collected after 24 hr of conditioning and albumin content was analyzed with the Albuwell kit (Exocell) according to the manufacturer's protocol. Albumin content was normalized by the total amount of protein per well.

Urea secretion

On Days 13 and 20 post-thawing, enhanced hiPS-HEP cells were incubated with 5 mM ammonium chloride for 24 hr. After 24 hr, medium was collected and urea secretion was analyzed with the QuantiChrom Urea Assay Kit (BioAssay Systems). Urea content was normalized to the total amount of protein per well.

Insulin response

On Day 12 post-thawing, enhanced hiPS-HEP cells from C18 were incubated in insulin-free medium for 3 hr and then treated for 10 min with 0 nM (–) or 100 nM (+) insulin. Phosphorylated Akt and total Akt in insulin-treated cells (+) and untreated controls (–) were quantified by Western blot analysis. Akt phosphorylation at pS473 was also measured by ELISA (Thermo Fisher Scientific). On Day 6 post-thawing, enhanced hiPS-HEP cells from C18 were incubated in insulin-free medium for 3 hr and then treated for 10 min with an insulin concentration ranging from 0 to 100 nM.

Acknowledgments

Data was kindly provided by Ann Hammarstedt (Department of Molecular and Clinical Medicine, Gothenburg University, Sweden), who performed the Western blot analysis of AKT proteins, incubation of cells in MCD-media, and subsequent qRT-PCR analysis.

Jane Synnergren and Benjamin Ulfenborg (School of Bioscience, System Biology Research Center, University of Skövde, Sweden) kindly performed the bioinformatic analysis of the transcriptome data.  

Asplund, A. et al. One Standardized Differentiation Procedure Robustly Generates Homogenous Hepatocyte Cultures Displaying Metabolic Diversity from a Large Panel of Human Pluripotent Stem Cells. Stem Cell Rev. Reports 12, 90–104 (2016).

Carvalho, B. S. & Irizarry, R. A. A framework for oligonucleotide microarray preprocessing. Bioinformatics 26, 2363–2367 (2010).

Durinck, S., Spellman, P. T., Birney, E. & Huber, W. Mapping identifiers for the integration of genomic datasets with the R/Bioconductor package biomaRt. Nat. Protoc. 4, 1184–91 (2009).

Ghosheh, N. et al. Highly Synchronized Expression of Lineage-Specific Genes during In Vitro Hepatic Differentiation of Human Pluripotent Stem Cell Lines. Stem Cells Int. 2016, 1–22 (2016).

Gonzalez, F. J. Regulation of hepatocyte nuclear factor 4 alpha-mediated transcription. Drug Metab. Pharmacokinet. 23, 2–7 (2008).

Jungermann, K. & Kietzmann, T. Oxygen: Modulator of metabolic zonation and disease of the liver. Hepatology 31, 255–260 (2000).

Peters, D. T. et al. Asialoglycoprotein receptor 1 is a specific cell-surface marker for isolating hepatocytes derived from human pluripotent stem cells. Development 143, 1475–81 (2016).

Takayama, K. et al. Prediction of interindividual differences in hepatic functions and drug sensitivity by using human iPS-derived hepatocytes. Proc. Natl. Acad. Sci. U. S. A. 111, 16772–7 (2014).

Tilg, H., Moschen, A. R. & Roden, M. NAFLD and diabetes mellitus. Nat. Rev. Gastroenterol. Hepatol. 14, 32–42 (2016).

Ulvestad, M. et al. Drug metabolizing enzyme and transporter protein profiles of hepatocytes derived from human embryonic and induced pluripotent stem cells. Biochem. Pharmacol. 86, 691–702 (2013).

Yates, A. et al. Ensembl 2016. Nucleic Acids Res. 44, D710–D716 (2016).

Zhang, K. et al. From endoplasmic-reticulum stress to the inflammatory response. Nature. 454, 455–462 (2008).

Zhang, X-Q. et al. Role of endoplasmic reticulum stress in the pathogenesis of nonalcoholic fatty liver disease. World J. Gastroenterol. 20, 1768–1776 (2014).