Hepatocyte-like cells differentiated from human induced pluripotent stem cells: Relevance to cellular therapies

Abstract

Maturation of induced pluripotent stem cells (hiPSCs) to hepatocyte-like cells (HLCs) has been proposed to address the shortage of human hepatocytes for therapeutic applications. The purpose of this study was to evaluate hiPSCs, HLCs and hepatocytes, all of human origin, in terms of performance metrics of relevance to cell therapies. hiPSCs were differentiated to HLCs in vitro using an established four-stage approach. We observed that hiPSCs had low oxygen consumption and possessed small, immature mitochondria located around the nucleus. With maturation to HLCs, mitochondria showed characteristic changes in morphology, ultrastructure, and gene expression. These changes in mitochondria included elongated morphology, swollen cristae, dense matrices, cytoplasmic migration, increased expression of mitochondrial DNA transcription and replication-related genes, and increased oxygen consumption. Following differentiation, HLCs expressed characteristic hepatocyte proteins including albumin and hepatocyte nuclear factor 4-alpha, and intrinsic functions including cytochrome P450 metabolism. But HLCs also expressed high levels of alpha fetoprotein, suggesting a persistent immature phenotype or inability to turn off early stage genes. Furthermore, the levels of albumin production, urea production, cytochrome P450 activity, and mitochondrial function of HLCs were significantly lower than primary human hepatocytes.

Conclusion

hiPSCs offer an unlimited source of human HLCs. However, reduced functionality of HLCs compared to primary human hepatocytes limits their usefulness in clinical practice. Novel techniques are needed to complete differentiation of hiPSCs to mature hepatocytes.

Introduction

Liver transplantation is a successful treatment for patients with end-stage liver disease. However, transplantable donor livers are in short supply. Hepatocyte transplantation and bioartificial liver (BAL) devices have been proposed as therapeutic alternatives to the shortage of transplantable livers. BAL is an extracorporeal supportive therapy developed to bridge patients with liver failure to liver transplantation or to recovery of the native liver. Hepatocyte transplantation is best suited for patients with metabolic liver disease for which smaller number of cells (<10% of liver mass) may be curative. Both BAL and hepatocyte transplantation are cellular therapies that avoid use of a whole liver.

Though once controversial, it is now well accepted that hepatocytes can be derived from progenitor cells which include pluripotent stem cells, either embryonic or native to the liver or in blood (Basma et al., 2009; Wang et al., 2011). Furthermore, techniques now exist for production of human induced pluripotent stem cells (hiPSCs) from somatic cells (Yu et al., 2007). Therefore, in theory, hiPSCs could provide an unlimited source of human hepatocytes for BAL therapy and cell transplantation. Preliminary reports indicate that human hepatocyte-like cells (HLCs) can be derived from hiPSCs under in vitro conditions (Si-Tayeb et al., 2010). These findings are exciting since they suggest the possibility of producing HLCs from the patient's own cells and cell transplantation without immunosuppression.

The HLCs derived from hiPSCs express characteristic hepatocyte proteins including alpha-1-antitrypsin, albumin (Alb), and hepatocyte nuclear factor 4-alpha (HNF4α). They also display intrinsic hepatocyte functions including cytochrome P450 (CYP) metabolism. The efficiency of induced pluripotent stem cells (iPSCs) directed-differentiation into HLCs is variable. Some protocols describe over 80% differentiation efficiency, but none yet achieve complete differentiation of hiPSCs into hepatocytes. Transplantation of undifferentiated iPSCs in immunodeficient recipients results in the formation of teratomas. However, the risks and benefits of transplantation of iPSCs into immunocompetent recipients are poorly studied. Reports of BAL therapy using HLCs derived from hiPSCs do not yet exist.

The present study was designed to assess the differentiation status and functionality of three human cell types (hiPSCs, HLCs and primary hepatocytes) under in vitro conditions with regards to their usefulness in human cell therapy. Functionality and differentiation of these three cell types were judged by liver-specific biochemical activities including the urea cycle, mitochondrial maturation including ultrastructure, and liver-related gene expression.

Experimental procedures

Feeder-free iPSCs culture conditions

Human iPSCs were generated from human cardiac fibroblasts (HCFs) as previously described (Thatava et al., 2011). Briefly, 1×10⁵ HCFs (ScienCell Research Laboratories, Carlsbad, CA, USA; no. 6300) were seeded 1 day before infection in each well of six-well plates with Dulbecco's modified Eagle's medium (DMEM) containing 10% FBS, penicillin (100 U/mL) and streptomycin (100 mg/mL) (complete DMEM), and transduced with pluripotency factor-expressing lentiviral vectors (octamer-binding transcription factor 4 (OCT4), SOX2, KLF4, c-MYC). Putative hiPSCs colonies were observed 1 to 2 weeks after vector transduction. Clones of hiPSCs were picked based on morphology and size. A suitable clone (HCF#1 hiPSCs) was selected. This clone was maintained on tissue culture plates coated with Matrigel (BD Biosciences, San Jose, CA, USA; no. 354277) in feeder-free medium (HEScGRO [Millipore, no SCM020] supplemented with 20% mTeSR1 complete medium [STEMCELL Technologies Inc., Vancouver, BC, Canada; no. 05850]) under a mix of 5% CO₂/ambient air.

In vitro differentiation of hiPSCs to HLCs

HLCs were produced by differentiation of HCF#1 hiPSCs in vitro using an established 4-stage approach (Si-Tayeb et al., 2010) with minor modifications: Stage 1 (S1)—endoderm induction, Stage 2 (S2)—hepatic specification, Stage 3 (S3)— hepatoblast expansion and Stage 4 (S4)—hepatic maturation. Briefly, endoderm differentiation was initiated by removing feeder-free medium and exposing 70% confluent cultures of HCF#1 hiPSCs to RPMI1640 media containing B27 Supplements Minus Insulin (Invitrogen, 0050129-SA), 100 ng/mL Activin A (R&D Systems, 338-AC-050/CF), 10 ng/ml bone morphogenetic protein 4 (BMP4) (Peprotech) and 20 ng/ml fibroblast growth factor 2 (FGF2) (Invitrogen, PHG0024). After 8 days of culture in 5% CO₂ with ambient oxygen, culture dishes containing induced definitive endoderm were moved to 5%CO₂/4%O₂ in RPMI/B27 Supplements (Invitrogen, 0080085-SA) medium supplemented with 20 ng/mL BMP4 and 10 ng/mL FGF2 for 5 days. We then cultured the specified hepatic cells in RPMI/B27 supplemented with 20 ng/mL hepatocyte growth factor (HGF, Peprotech), under 5% CO₂/4%O₂ for 5 days. For the final stage of differentiation, cultures were transferred to 5% CO₂/ambient O₂, and the media were replaced with HCM™ hepatocyte culture medium (Lonza, Walkersville, MD USA, CC-3198 but omitting the EGF) supplemented with 20 ng/mL Oncostatin M (R&D Systems, 295-OM-010/CF) for an additional 5 days. In summary, cell-types were selected as follows: hiPSCs (day 0), S1—endoderm (day 8), S2—specified hepatic (day 13), S3—hepatoblast (day 18), and S4—HLCs (day 23).

Primary hepatocytes

The human hepatocytes were purchased from Yecuris Corporation. These primary human hepatocytes were produced by robust in vivo expansion in genetically engineered mice (Azuma et al., 2010). Human hepatocytes were cultured in collagen IV coated plates with William's E medium.

Immunofluorescence

For immunocytochemistry, cultured cells were fixed with 4% paraformaldehyde for 30 min at room temperature, washed with phosphate-buffered saline (PBS) three times for 5 min each, and permeabilized with 0.5% Triton X-100 (Sigma) in PBS for 15 min, and blocked with 3% BSA in PBS for 15 min. Cells were incubated overnight at 4 °C with primary antibodies diluted in 1% BSA in PBS. Primary antibodies included OCT4 (Santa Cruz Biotechnology Inc., CA, #sc-9081), GATA4 (Santa Cruz, #sc-1237), HNF4α (Santa Cruz, #sc-6556), alpha fetoprotein (AFP) (Sigma-Aldrich, #A8452), and albumin (Bethy Laboratories Inc. A80-129A). Antigens were visualized using secondary antibodies conjugated with either Alexa 488 or Alexa 568 (Invitrogen, #A11057, A11055, A21202, A21206).

Periodic acid-schiff (PAS) staining

PAS staining was performed on the last day of S4, using a commercial staining kit (Sigma-Aldrich, #395) according to the manufacturer's instructions.

Quantitative reverse transcription polymerase chain reaction (QRT-PCR)

Total RNA was extracted from cultured cells at each time point using an RNeasy Plus MiniKit (Qiagen, Valencia, CA) according to the manufacturer's instructions. All PCR were performed in 96-well optical reaction plates (Applied BioSystems, Foster City, CA) in 20 μL reaction volume using SYBR Green PCR Master Mix (Applied Biosystems, No. 4389986). Reactions were carried out on the ABI PRISM® 7900HT Sequence Detection System (SDS) (Applied Biosys-tems) and data were analyzed with the SDS Version 2.4 software. Triplicate amplifications were carried out for each target gene. Quantification was performed using the comparative Ct method normalized with the housekeeping gene glyceraldehyde 3-phosphate dehydrogenase (GAPDH). Expression of each target gene at each stage of differentiation was normalized to its level in hiPSCs (day 0). For expression of genes that could not be detected in hiPSCs, we used expression levels on day 8 for normalization. Primer sequences are listed in Table 1.

Table 1.

Primer sequences for RT-PCR.

Genes	Abbr.	5′→3′	Product size (bp)
Alpha fetoprotein	AFP	Forward: AGCTTGGTGGTGGATGAAAC	248
		Reverse: CCCTCTTCAGCAAAGCAGAC
Albumin	Alb	Forward: GCACAGAATCCTTGGTGAACAG	101
		Reverse: ATGGAAGGTGAATGTTTCAGCA
Arginase 1	Arg1	Forward: GCAGAAGTCAAGAAGAACGG	153
		Reverse: GGTTGTCAGTGGAGTGTTG
Arginosuccinate lyase	Asl	Forward : GAGGTGCGGAAGCGGATCAATGT	233
		Reverse: TTGGTGCAGTAGAGGATGAGGTCC
Arginosuccinate synthetase	Ass	Forward: TCGTGCATCCTCGTGTGGCTGAA	157
		Reverse: CCACAAACTCCCTGCTGACATCC
ATP synthase 5 subunit e	Atp5g1	Forward: TTCCAGACCAGTGTTGTCTCC	146
		Reverse: GACGGGTTCCTGGCATAGC
ATP synthase 8	ATP8	Forward: AACAAGTGCTCCATCAATGGC	146
		Reverse: AGGCTGGGGTCCCAAAATAAG
Carbamoyl-phosphate synthetase 1	Cps1	Forward: GGCCATCCATCCTCTGTTGC	171
		Reverse: GCTAAGTCCCAGTTCATCCA
C-X-C chemokine receptor type 4	Cxcr4	Forward: CCTATGCAAGGCAGTCCATGT	86
		Reverse: GGTAGCGGTCCAGACTGATGA
Cytochrome P450, family 2, subfamily C, polypeptide 19	Cyp2C19	Forward: TCCAGATATGCAATAATTTTCCCAC	164
		Reverse: GAAGCAATCAATAAAGTCCCGA
Cytochrome P450, family 3, subfamily A, polypeptide 4	Cyp3A4	Forward: CAGGAGGAAATTGATGCAGTTTT	78
		Reverse: GTCAAGATACTCCATCTGTAGCACAGT
Forkhead box A2	FoxA2	Forward: ATTGCTGGTCGTTTGTTGTG	169
		Reverse: TACGTGTTCATGCCGTTCAT
Glyceraldehyde 3-phosphate dehydrogenase	GAPDH	Forward: CAGAACATCATCCCTGCCTCTAC	251
		Reverse: TTGAAGTCAGAGGAGACCACCTG
GATA binding protein 4	Gata4	Forward: CTAGACCGTGGGTTTTGCAT	275
Hepatocyte nuclear factor 4α	HNF4α	Forward: TGTCCCGACAGATCACCTC	121
		Reverse: CACTCAACGAGAACCAGCAG
N-acetylglutamate synthase	Nags	Forward: CAGCAGGGCGTATCCAGTC	180
		Reverse: GTTGTGTCGAAGCGCGTCTA
NADH dehydrogenase subunit 1	ND1	Forward: TTCTAATCGCAATGGCATTCCT	109
		Reverse: AAGGGTTGTAGTAGCCCGTAG
NADH dehydrogenase subunit 5	ND5	Forward: TTCATCCCTGTAGCATTGTTCG	154
		Reverse: GTTGGAATAGGTTGTTAGCGGTA
Octamer-binding transcription factor 4	Oct4	Forward: AGCGAACCAGTATCGAGAAC	142
		Reverse: TTACAGAACCACACTCGGAC
Ornithine carbamoyltransferase	Otc	Forward: AATCTGAGGATCCTGTTAAACAATG	106
		Reverse: CCTTCAGCTGCACTTTATTTTGTAG
Mitochondrial-specific DNA polymerase gamma	Polg	Forward: GCTGCACGAGCAAATCTTCG	163
		Reverse: GTCCAGGTTGTCCCCGTAGA
Mitochondrial-specific DNA polymerase gamma 2	Polg2	Forward: GCTTAGCCGGGATTCTCTTCT	98
		Reverse: CCGAGGTCCACCATTCTGC
Sex determining region Y box 17	Sox17	Forward: GTGGACCGCACGGAATTTG	94
		Reverse: GGAGATTCACACCGGAGTCA
Mitochondrial transcription factor A	Tfam	Forward: ACCGAGGTGGTTTTCATCTG	218
		Reverse: TATATACCTGCCACTCCGCC
Mitochondrial uncoupling protein 2	Ucp2	Forward: ATGTTGCTCGTAATGCCATTGT	104
		Reverse: GCAAGGGAGGTCATCTGTCA

Analysis of mitochondria

Mitochondrial staining

Cells for mitochondria labeling were washed twice with PBS and incubated at 37 °C for 20 min in 50 nM mitochondrion-selective dye MitoTracker Red CMXRos (Molecular Probes Inc., #M7512). Nuclei were stained with DAPI. Staining solution was then replaced with fresh pre-warmed media, and cells were visualized using a confocal microscope LSM 510 (Zeiss, Germany, Axiovert 100M).

Count of mitochondria number

Transmission electron microscopy (TEM) was used to investigate the morphological features of the mitochondrial network within differentiated and undifferentiated cells. Cells were fixed in Trumps solution, sectioned, and imaged by TEM in the Microscopy Core Facility (Mayo Clinic, Rochester, MN). Mitochondria numbers were quantified by manually counting the number of mitochondria per cell. A minimum of 12 cells were examined in 4–6 representative TEM images (80,000-fold magnification).

Mitochondrial DNA copy number by quantitative polymerase chain reaction (QPCR)

Copy number of mitochondrial DNA (mtDNA) was estimated by QPCR analysis using the mitochondrial gene NADH dehydrogenase subunit 1 (ND1) and ND5. Total cellular DNA was collected and amplified. ND1 and ND5 levels were normalized to half the level of GAPDH since each cell contains two copies of genomic DNA compared to a single copy of DNA per chromosome. Each sample was run in triplicate, and QPCR analysis was performed as described above. Primer sequences are reported in Table 1.

Quantification of albumin, AFP, ammonia, ureagenesis, diazepam metabolism

On days 0, 8, 13, 18, and 23 of differentiation culture and day 1 of human hepatocyte culture, old medium was exchanged with fresh William's E medium supplemented with 10% fetal bovine serum (FBS; Mediatech, Inc., Herndon, VA), 20 μM diazepam and 5% v/v heavy deuterium-enriched ammonia gas (¹⁵ND₃) (2.23 mM; Cambridge Isotope Laboratories, Inc, Andover, MA). Cells were incubated for an additional 24 h. Supernatant medium was collected and stored at −20 °C prior to analysis. The concentration of human albumin in the sampled medium was determined by enzyme-linked immunosorbent assay (ELISA) kit (Bethyl, Montgomery, TX, #E80-129). The concentration of human AFP was determined in the Automated ImmunoAssay Laboratory of the Mayo Division of Clinical Biochemistry and Immunology. The concentration of total urea was performed using a QuantiChromTM DIUR-500 Urea Assay Kit (BioAssay System, Hayward, CA) according to the manufacturer's instructions. Production of urea through the complete urea cycle was determined by the conversion of heavy ammonia (¹⁵ND₃) to native and deuterium-enriched urea. Isotopes of urea were quantified by capillary gas chromatography/mass spectrometry as previously reported (Rinaldo, 2008). Concentrations of diazepam and its three major metabolites (nordiazepam, temazepam, oxazepam) were determined by high performance liquid chromatography (HPLC) with mass spectrometry detection in the CTSA Metabolomics Core lab at Mayo Clinic (Baskin-Bey et al., 2005; Mandrioli et al., 2008). Samples were deconjugated in β-glucuronidase/arylsulfatase (600 Fishman units and 4800 Roy units/mL, respectively; Roche, Indianapolis, IN) and sodium acetate buffer (0.1 M) for 2 h. Reactions were quenched with ethanol. Data acquisition was completed by Chrom Perfect® software (Denville, NJ). All assays were performed in triplicate.

Oxygen consumption

Rates of oxygen consumption were determined by syringe technique as previously reported (Lillegard et al., 2011). Briefly, cells were resuspended in 3 mL of culture medium and oxygen tensions of the medium were determined at 2 min intervals over 6 min by a GEM Premier 3000 gas analyzer. The rate of oxygen consumption was determined using the equation:

$$\mathit{\text{Oxygen consumption}}(\text{mol}O2/\text{min})=(\mathit{\text{pO}}2.\text{initial mmHg}-\mathit{\text{pO}}2.\text{final mmHg})\times (1.29\times {10}^9\text{molO}2/\text{mL}/\text{mmHg})\times (\text{volumeml})/(\text{timemin})$$

Statistical analysis

Data are expressed as mean ± standard deviation of independent samples. Comparisons between two groups were performed by unpaired Student's t-test. p-Values≤0.01 were considered significant. Data were analyzed by SPSS 16 and Excel 2003, Microsoft, Redmond, WA.

Results

Generation HLCs from hiPSCs

The morphology of hiPSCs changed significantly from day 0 to day 23 during four stages of differentiation. The light microscopic appearance of representative cultures of hiPSCs (day 0), endoderm (day 8), hepatic specific (day 13), hepatoblast (day 18), HLCs (day 23), and primary human hepatocytes are shown in Fig. 1A. hiPSCs gradually transformed into less dense, flatter cells with prominent nuclei and spiky shape (day 8) before eventually forming an epithelial monolayer on day 23. Other features of HLCs included a large cytoplasmic-to-nuclear ratio, numerous and prominent nucleoli, and occasional binucleated cells. These features were also characteristic of primary hepatocytes (Fig. 1A).

Figure 1.

Generation of HLCs from hiPSCs. A. Cell morphology by light phase microscopy. The cells gradually transformed into less dense, flatter layers with prominent nuclei (S1), a spiky shape, and then formed an epithelial monolayer (S4). HLCs with large cytoplasmic-to-nuclear ratio, numerous, and prominent nucleoli were observed on day 23, which were of similar morphology and character to primary hepatocytes. B. Relative expression of early stage genes of differentiation. The early stage of hiPSCs differentiation was characterized by the loss of pluripotency marker OCT4. All of the other genes were expressed significantly higher at day 23 (HLCs) than at day 0 (hiPSCs). The special markers for different stages of differentiation (GATA4 for S1, HNF4α for S2, AFP for S3 and Alb for S4) confirmed the successful differentiation of hiPSCs to HLCs. C. Immunofluorescence staining of human cells during 4 stages (23 days) of differentiation. Human hepatocytes are included for comparison (far right column).

To confirm the differentiation status of cells during the four stages of maturation we examined the relative expression of 22 genes using QRT-PCR analyses. These genes included the pluripotency marker (OCT4) and four endoderm markers (FOXA2, GATA4, CXCR4, SOX17) shown in Fig. 1B. Expression of three liver-related genes (HNF4α, AFP, Alb) are shown in Fig. 2A. Expression of the six urea cycle genes (Nags, Cps1, Otc, Ass, Asl, Arg1) were included in Fig. 4A because of their importance to ammonia detoxification and their therapeutic role in cell therapy of liver failure. Expression of CYP3A4 and CYP2C19, whose protein products are important phase 1 enzymes, are shown in Fig. 5A. Expression of Atp5g1, Atp8, Ucp2, Tfam, Polg, Polg2 (Fig. 6A) were used to assess mitochondrial differentiation.

Figure 2.

Characteristics of the HLCs. A. Relative expression of three hepatocyte-related genes using QRT-PCR analyses. Day 23 HLCs expressed higher levels of AFP RNA and lower levels of Alb mRNA than the primary hepatocytes. B. Accumulation of albumin and AFP in the culture supernatant were consistent with expression of these genes in iPS cells (day 0), HLCs (day 23) and hepatocytes. No human albumin was detected in medium of day 0 cultures, which confirmed specificity of our human albumin ELISA. Data are represented as mean ± standard deviation. C. Immunofluorescence staining for AFP and Alb. On day 23 most HLCs stained positively for AFP and Alb, while human hepatocytes only stained positively for Alb.

Figure 4.

Expression of urea cycle genes and ureagenesis. A. The QRT-PCR assay of urea cycle genes. HLCs expressed lower mRNA level of urea cycle genes than the primary hepatocytes. B. Hepatoblasts (day 18) and HLCs (day 23) showed the ability to produce urea. However, the concentration of urea in these cultures was significantly lower than cultures of primary hepatocytes (p<0.01). Furthermore, the formation of heavy urea from heavy ammonia was only demonstrated by primary hepatocytes, indicating that only primary hepatocytes were capable of normal ureagenesis. Data are represented as mean ± standard deviation.

Figure 5.

CYP3A4 activity and diazepam metabolism. A. Expression of CYP3A4 and CYP2C19 increased significantly during the 4 stages of differentiation and was greatest in primary human hepatocytes. ND—expression of CYP3A4 was not detectable in hiPSCs. B. Clearance of diazepam and formation of its three major metabolites (nordiazepam, temazepam, oxazepam) was highest in cultures of primary hepatocytes (p<0.01), indicating greater phase 1 metabolism by primary hepatocyte than HLCs (day 23) or iPSCs (day 0). Data are represented as mean ± standard deviation.

Figure 6.

Mitochondrial features during differentiation of hiPSCs to hepatocytes. A. Expression of Polg, Polg2, Ucp2, Atp5g1 and Atp8 role steadily during stages of differentiation, while expression of Tfam declined over this time frame. B. The ratio of mitochondrial DNA to nuclear DNA rose significantly (p<0.01) during differentiation reflecting the robust expansion of mitochondria in differentiating cells. ND1 and ND5 are genes encoded on mitochondrial DNA. Data are represented as mean ± standard deviation. C. Top row—cells were examined by confocal microscopy using MitoTracker Red to assess their expanding density of functional mitochondria (red stain) during stages of differentiation. Nuclei were stained blue by DAPI. Middle row (Low power TEM) and Lower Row (high power TEM)—mitochondria were observed to elongate, develop swollen cristae and dense matrices, and migrate deeper into the cytoplasm during stages of differentiation. The mitochondria of HLCs (day 23) appeared less mature than the mitochondria in primary hepatocytes. D. Mitochondria were counted from TEM images. The number of mitochondria rose steadily through stages of differentiation and reached a maximum in primary hepatocytes. The number of mitochondria in HLCs (day 23) was significantly less than in primary hepatocytes (p<0.01). Data are represented as mean ± standard deviation. E. The level of oxygen consumption by HLCs (day 23) was significantly lower than primary hepatocytes (p<0.01). Data are represented as mean ± standard deviation.

As expected, loss of the pluripotency marker OCT4 was observed during the 4 stages of differentiation as shown in Fig. 1B. Expression of the four endoderm genes shown in Fig. 1B was significantly higher on day 8 and at all later time points compared to hiPSCs (day 0).

To further assess the differentiation of hiPSCs, we performed immunofluorescence staining (Fig. 1C). All of the undifferentiated hiPSCs expressed the pluripotency marker OCT4 on day 0. By day 8, end of the S1, a sharp decline in OCT4 staining and a rapid rise in the number of cells staining positive for the definitive endoderm marker GATA4 was observed. This pattern of staining suggested efficient differentiation of iPSCs to endoderm. By day 13, end of S2, HNF4α stained cells were observed. HNF4α is believed to be essential for specification of hepatic progenitors from human pluripotent stem cells (DeLaForest et al., 2011). By day 18, end of S3, the addition of hepatocyte growth factor (HGF) to the culture conditions resulted in high levels of expression of AFP, and indicated that the specified cells had committed to a hepatoblast fate. By day 23, end of S4, remaining cells stained brightly for albumin that could be quantified in the media by ELISA (Fig. 2B). On average, 80% of the day 23 cells were Alb-positive (Fig. 2C).

On day 23, after completion of the differentiation protocol, the cells were found to display several known hepatic functions. PAS staining (Fig. 3) indicated glycogen synthesis by the differentiated cells. Analyses of the culture media revealed the ability of day 23 cells to secrete albumin (Fig. 2B), produce urea (Fig. 4B), and metabolize diazepam (CYP3A4 and CYP2C19 enzyme activities) (Fig. 5B).

Figure 3.

PAS staining for glycogen synthesis. Left—hiPSCs stained very weakly positive for PAS on day 0. Middle—HLCs stained positively for PAS on day 23. A focus of less differentiated cells that do not stain for PAS is shown in the upper center of the image. Primary hepatocytes uniformly stain positive for PAS.

We also examined the mitochondria to assess differentiation of hiPSCs to HLCs (Fig. 6). Cell ultrastructure by TEM demonstrated that undifferentiated hiPSCs on day 0 possessed the fewest number and smallest size of mitochondria (Fig. 6C). On day 0, mitochondria were round-shaped, had poorly developed crista, and were located near the nucleus. As hiPSCs lost pluripotency and committed to a hepatic differentiation fate, the expression of mtDNA transcription and replication factors was upregulated (Fig. 6A). The number of mitochondria/cell (Fig. 6D) and mtDNA copies/cell (Fig. 6B) also increased as cells lost pluripotency. With differentiation, both cells and their mitochondria increased in size. The ultrastructural features of mitochondria also became more mature (Fig. 6C). Maturing mitochondria acquired an elongated morphology with swollen cristae and dense matrices as they migrated into wider cytoplasmic areas. Concomitantly, the rate of oxygen consumption by differentiated cells increased (Fig. 6E) as aerobic metabolism became active and more efficient. Collectively, these observations were consistent with successful production of HLCs from hiPSCs.

Human HLCs vs. primary human hepatocytes

Higher AFP and lower albumin by HLCs

To test whether HLCs can provide functionality to ex vivo cell therapies, such as a BAL, we compared the extent of differentiation of HLCs to primary human hepatocytes. We first compared the relative gene expression of Alb and AFP in HLCs and human hepatocytes using QRT-PCR. The results reveal that HLCs expressed a high level of albumin mRNA, but this level was still lower than the level in primary hepatocytes (Fig. 2A). Consistent with this difference in gene transcription, we observed significantly higher concentrations of albumin in the supernatant of cultures of primary hepatocytes compared to HLCs (Fig. 2B). In contrast, primary hepatocytes expressed low levels of AFP mRNA compared to HLCs (Fig. 2A). No AFP was detected in supernatant of primary hepatocytes (Fig. 2B) and primary human hepatocytes did not stain for AFP by immunocytochemistry (Fig. 2C). Cultures of HLCs expressed high levels of AFP mRNA (Fig. 2A) and secreted high levels of AFP into the medium (Fig. 2B). Interestingly, most HLCs stained positively for AFP by immunocytochemistry (Fig. 2C). An inverse relationship was observed between the expression pattern of AFP and Alb in HLCs. These findings suggest differences between HLCs and primary hepatocytes, consistent with a less mature phenotype of HLCs than primary hepatocytes.

Incomplete urea cycle activity by HLCs

Compared to primary hepatocytes, day 23 HLCs expressed lower levels of mRNA of the urea cycle genes (Fig. 4A). The HLCs showed an ability to produce urea (Fig. 4B). However, this level of production was significantly lower than primary hepatocytes. More importantly, neither hiPSCs nor HLCs were able to produce heavy urea when heavy ammonia (¹⁵ND₃) was supplemented to their culture (Fig. 4B). In contrast, over 40% of urea contained heavy isotopes after the culture medium of primary hepatocytes was supplemented with heavy ammonia (Fig. 4B). These data indicated urea cycle activity is complete in primary hepatocytes, but incomplete in less mature cells such as HLCs and iPSCs. The urea formed in cultures of HLCs and iPSCs likely represented the activity of its final step (i.e., arginase), rather than activity of the complete urea cycle.

Lower CYP activity by HLCs

The CYP enzymes are critical enzymes of phase 1 metabolism. In particular, CYP3A4 is normally abundant in human liver. Of little surprise, expression of CYP3A4 gene was not detectable in hiPSCs, and rose steadily during the four stages of differentiation (Fig. 5A). However, CYP3A4 expression did not reach the level of primary hepatocytes. Functionality of the CYP enzymes was assessed by the elimination of the classical substrate, diazepam, and appearance of its major metabolites (temazepam, nordiazepam, and oxazepam). Temazepam is a marker of 3-hydroxylation by CYP3A4; nordiazepam is a marker of N-dealkylation by CYP2C19; and oxazepam is the product of both CYP activities. Therefore, Fig. 5B demonstrates significantly higher CYP3A4 and CYP2C19 activity by primary hepatocytes than HLCs. The higher concentrations of nordiazepam in cultures of HLCs and primary hepatocytes suggest that CYP2C19 is the major pathway of diazepam elimination by both cell types. The low level of diazepam elimination in cultures of hiPSCs suggests non-specific loss of drug since no major metabolites were detected in its supernatant.

Immature mitochondria and lower oxygen consumption by HLCs

Because of their role in energy production and therefore functionality of hepatocytes, mitochondria were a focus of our analysis. Mitochondria were assessed in terms of gene expression (Fig. 6A), replication (Figs. 6B and D), ultrastructure (Fig. 6C), and respiration (oxygen consumption, Fig. 6E). Replication of mtDNA is regulated by the nuclear-encoded mitochondrial transcription factor A (TFAM) and the mitochondrial-specific DNA polymerase gamma (POLG), which consists of a catalytic (POLG1) and an accessory (POLG2) subunit. Our QRT-PCR showed that both Polg and Tfam were expressed at baseline (day 0) and during the steps of differentiation. Expression of Tfam declined during differentiation, while expression of the five other mitochondrial genes (Polg, Polg2, Ucp2, Atp5g, Atp8) was increased (Fig. 6A). Consistent with the rise in mitochondrial gene expression during differentiation, mtDNA copy number increased 50-fold from hiPSCs to HLCs (Fig. 6B). Mitochondrial number per cell and mtDNA copy number were highest in primary hepatocytes (Figs. 6B and D). The difference in maturity of HLCs vs. primary hepatocytes was further uncovered by the ultrastructural features and redox status of their mitochondria (Fig. 6C). Consequently, oxygen consumption by primary human hepato-cytes was significantly higher than human HLCs (Fig. 6E).

Discussion

Access to an abundant, high quality supply of hepatocytes with therapeutic potential for cell transplantation and extracorporeal support of patients in liver failure has been the objective of several research programs. Though hepatic differentiation of hiPSCs has been described (Chen et al., 2012), the functionality of HLCs derived from hiPSCs remains poorly studied in the context of primary human hepatocytes or their application in cell based therapies. The Fox group has reported an implantable BAL device containing HLCs derived from embryonic stem cells (ESCs) in a murine model of liver failure (Soto-Gutierrez et al., 2006). However, no reports have addressed functionality of HLCs derived from hiPSCs in a cell-based BAL. Our results show a characteristic polygonal epithelial morphology and large cytoplasm of HLCs derived from hiPSCs. These HLCs contained cytoplasmic granules of hepatocytes and expressed hepatocyte-specific genes. Additionally, HLCs exhibited known functions of mature hepatocytes, such as albumin production and the appearance of urea in culture medium. We also observed CYP activity and PAS staining consistent with glycogen storage by HLCs. Collectively, these observations confirm the generation of HLCs from hiPSCs.

To address the question of whether these HLCs could serve in cellular therapies, a variety of markers of hepatocyte differentiation were considered. For example, production of albumin, a classic marker of hepatocyte differentiation, was observed in HLCs (Fig. 2). However, production of albumin has been shown to be non-predictive of response to cell-based therapy (Dunn et al., 1991; Koike et al., 1996; Nahmias et al., 2007). Alternatively, mitochondria play a crucial role in many cellular functions relevant to cell based therapy including energy production, differentiation, apoptosis, and ammonia detoxification. Mitochondria are the site of several enzymes of the urea cycle and thus essential to ammonia detoxification and prevention of hepatic encephalopathy. Our data indicate significant maturation of mitochondria in HLCs compared to their hiPSCs precursors. However the mitochondria of HLCs remain noticeable less mature than the gold standard - primary hepatocytes.

Our experience with hiPSCs and HLCs is analogous to the maturation of ESCs. Reviewed evidence suggests that respiratory activity in mitochondria of ESCs is low to minimize generation of reactive oxygen species which may damage and mutate DNA during early development (Parker et al., 2009). As stem cells of the early embryo commit to differentiation, their mitochondria undergo dramatic functional maturation. Successful differentiation of mammalian ESCs involves transcription and replication of the mtDNA genome, replication of mitochondria, and up-regulation of enzymes required for aerobic metabolism (Chen et al., 2008; Cho et al., 2006; Chung et al., 2007; Schieke et al., 2008; Suhr et al., 2010). These changes are needed to fulfill the elevated ATP requirements of fully differentiated cells. In the case of primary hepatocytes, they are rich in mitochondria to fulfill their many vital roles including metabolism, synthesis of blood proteins and biotransformation. Our results suggest that HLCs derived from hiPSCs are capable of many of these functions, but they do not appear to detoxify ammonia to urea under ex vivo conditions.

Our results also demonstrate that cultured HLCs express higher levels of AFP and lower levels of albumin than human hepatocytes produced by in vivo expansion in genetically engineered mice (Azuma et al., 2007). Our observations are similar to other reports of significant but lower than normal production of albumin by HLCs (Cai et al., 2007; Sancho-Bru et al., 2011; Si-Tayeb et al., 2010; Song et al., 2009). The high levels of AFP production in HLCs are unexplained, but suggest that HLCs exhibit an inability to turn off early stage gene(s) as the mechanism of persistent immature phenotype.

As mentioned above, the liver is a major site of detoxification, and ammonia is arguably the most important endogenous toxin which requires clearance by the liver (Denis et al., 1983). Ammonia accumulation causes cerebral edema, the most feared complication of acute liver failure (Haussinger et al., 1992). A functioning urea cycle must be present for ammonia to be effectively detoxified and eliminated from the body (Ytrebo et al., 2009). Therefore, ammonia removal and urea production are important considerations for ex vivo cell therapies. Our data indicate that HLCs are inferior to primary hepatocytes at both ammonia detoxification and ureagenesis. The CYP enzymes, found in high levels in the human liver, are also important to detoxification of endogenous and exogenous wastes. Our HLCs expressed high levels of mRNA from CYP3A4 and CYP2C19 and detectable CYP3A4 and CYP2C19 enzyme activity. However, the CYP activities of HLCs were significantly below that of primary human hepatocytes.

Mitochondria play a vital role in the balance between pluripotency, cellular differentiation, and replication (Armstrong et al., 2010). TFAM, POLG and POLG2 are the key genes in regulating mtDNA replication. All three genes are nuclear-encoded, but produce transcription factors which are translocated to the mitochondria. We observed that the expression of POLG and POLG2 increased progressively in hiPSCs, HLCs, and primary hepatocytes; however, expression of TFAM declined with maturity. Our data is consistent with reports that an increase in the number of TFAM molecules above the optimal mtDNA:TFAM stoichiometry has inhibitory effects on mtDNA transcription and replication (Garstka et al., 2003; Webb and Smith, 1977). Therefore, a reduction in levels of TFAM protein may explain, in part, the increased transcription of mtDNA through the 4 stages of differentiation (Kanki et al., 2004).

Mammalian mitochondria are composed of approximately 1000 different proteins, but only 13 proteins are encoded by the mitochondrial genome. We selected Atp5g, Atp8 and Ucp2 as protein markers of the mitochondrial genome. We observed progressively higher expression of Ucp2, Atp5g and Atp8 in hiPSCs, HLCs, and primary hepatocytes. It has hitherto been believed that the ratio of mtDNA to nuclear DNA reflects the tissue concentration of mtDNA per cell. Control of mtDNA copy number is crucial for successful differentiation of ESCs (Facucho-Oliveira et al., 2007), and presumably also of hiPSCs. With the hepatic differentiation of hiPSCs, we observed an increased mtDNA copy number. MitoTracker staining also rose to reflect the expanded number of mitochondria in the cytoplasm during hepatic differentiation.

Our hiPSCs were derived from human cardiac fibroblasts. These parental fibroblasts possessed a large supply of mitochondria. However, reduction and dedifferentiation of mitochondria was essential for proper energy production and for prevention of DNA damage by oxidative stress in their conversion to hiPSCs. In turn, re-differentiation of hiPSCs to HLCs involved resupplying mitochondrial mass to meet the elevated ATP requirements of differentiated hepatocytes (Brown, 1992). It is possible that the difference in maturity profile and functionality between our HLCs and primary hepatocytes can be explained by our original source of somatic cells, the cardiac fibroblasts. Prior work with hiPSCs derived from fibroblasts suggests that mitochondrial numbers returned to their pre-reprogrammed state or lower levels after re-differentiation (Armstrong et al., 2010; Suhr et al., 2010). Whether the lower mitochondrial functionality of our HLCs was due to their cell of origin is unknown. Alternatively, the mitochondrial complement of HLCs may require more time and more replicative cycles to realize its full potential (Suhr et al., 2010). Further work is needed to clarify the role of original somatic cell type and other factors in re-programming iPSCs to new mature cell types.

Conclusion

In conclusion, our results confirm that functional HLCs can be generated from hiPSCs. However, our studies suggest that HLCs produced in vitro are not sufficiently mature to serve as ex vivo cell therapies, such as an extracorporeal BAL. Our results support those of Takayama et al. (2012), and suggest that a more native milieu is needed to complete hepatic differentiation of HLCs. Other reports suggest that in vivo transplantation of HLCs derived from hiPSCs can rescue rodents from lethal drug-induced acute liver failure (Chen et al., 2012), reduce liver fibrosis in a mouse model (Asgari et al., in press), enhance liver regeneration in mice (Espejel et al., 2010), and stabilize chronic liver disease (Choi et al., 2011). These in vivo studies suggest that novel techniques are still needed to complete hepatic differentiation of HLCs and expand them in numbers sufficient for human therapy. One possibility for expansion and maturation of HLCs is a genetically engineered large animal, similar but larger than the source of human hepatocytes in our study (Azuma et al., 2007), to serve as an in vivo hepatocyte incubator (Hickey et al., 2011). The future success of ex vivo cell therapies depends on novel techniques to provide an abundant, high quality supply of functionally normal hepatocytes.

Abbreviations

AFP

alpha fetoprotein

Alb

albumin

ATP5g1

mitochondrial F0 complex, subunit C1 (subunit 9), ATP synthase 5 subunit e

ATP8

ATP synthase 8

BAL

bioartificial liver

BMP4

bone morphogenetic protein 4

CYP

cytochrome P450

CYP2C19

CYP Family 2, Subfamily C, polypeptide 19

CYP3A4

CYP Family 3, Subfamily A, polypeptide 4

ESC

embryonic stem cell

FGF2

fibroblast growth factor 2

GATA4

GATA-binding protein 4

HCF

human cardiac fibroblast

HGF

hepatocyte growth factor

hiPSC

human induced pluripotent stem cell

HLC

hepatocyte-like cell

HNF4α

hepatocyte nuclear factor 4-alpha

iPSC

induced pluripotent stem cell

MtDNA

mitochondrial DNA

ND1

NADH dehydrogenase subunit 1

ND5

NADH dehydrogenase subunit 5

OCT4

octamer-binding transcription factor 4

PAS

Periodic Acid-Schiff

QRT-PCR

Quantitative Reverse Transcription Polymerase Chain Reaction

QPCR

Quantitative Polymerase Chain Reaction

TEM

transmission electron microscopy