Potential conflict of interest: Nothing to report.
Hepatic differentiation at the molecular level is poorly understood, mainly because of the lack of a suitable model. Recently, using adherent monoculture conditions, we demonstrated the direct differentiation of hepatocytes from embryonic stem (ES) cells. In this study, we exploited the direct differentiation model to compare the gene expression profiles of ES cell–derived hepatocytes with adult mouse liver using DNA microarray technology. The results showed that the ES cell–derived hepatocyte gene expression pattern is very similar to adult mouse liver. Through further analysis of gene ontology categories for the 232 most radically altered genes, we found that the significant categories related to hepatic function. Furthermore, through the use of small interfering RNA technology in vitro, hepatocyte nuclear factor 3β/FoxA2 was identified as having an essential role in hepatic differentiation. These results demonstrate that ES cell–derived hepatocytes recapitulate the gene expression profile of adult mouse liver to a significant degree and indicate that our direct induction system progresses via endoderm differentiation. In conclusion, our system closely mimics in vivo hepatic differentiation at the transcriptional level and could, therefore, be useful for studying the molecular basis of hepatocyte differentiation per se. (HEPATOLOGY 2005.)
Embryonic stem (ES) cells have the ability to differentiate into a variety of cell lineages.1, 2 ES cells can be propagated indefinitely in an undifferentiated state but, when provided with the appropriate signals, have the capacity to differentiate—presumably via the formation of precursor cells—into almost any mature cell phenotype. The recapitulation of developmentally regulated gene expression patterns enables the analysis of developmental processes on a cellular level in vitro. Therefore, ES cells have been used as a complementary experimental model to analyze early differentiation events in vivo. These regulatory genes are often transcriptional factors that activate or repress patterns of gene expression that create the phenotypic change seen during stem cell differentiation.3, 4
At the molecular level, the induction of liver development and its progress is characterized by the expression of transcription factors, such as hepatocyte nuclear factors (HNFs), CCAAT/enhancer-binding proteins, and GATA-binding proteins. These so-called “liver-enriched” transcription factors show a specific expression pattern during organogenesis with a distinct narrow time interval of transcription initiation.5, 6 Recently, we established the direct differentiation of functional hepatocytes from an adherent monoculture condition of ES cells without forming embryoid bodies and clearly identified the growth factors that direct hepatic fate specification (the hepatic induction factor cocktail [HIFC] differentiation system)7 based on the in vivo induction system.8 These cells expressed several differentiation markers of mature hepatocytes and rescued experimental liver injury when they were transplanted into animals. Our next goal was to clarify the molecular events involved in hepatic differentiation.
In this study, we show that the gene expression pattern of ES cell–derived hepatocytes is similar to that in adult mouse liver. Furthermore, by using small interfering RNA (siRNA) technology, we demonstrate that HNF3β/FoxA2 plays a critical role in early hepatic differentiation from ES cells. These results suggest that the HIFC differentiation system allows hepatic differentiation of ES cells via endoderm differentiation, thus recapitulating the in vivo liver developmental program.
The HIFC Differentiation System and Culture of ES Cells.
pALB-EGFP/ES cells, a J1 cell clone of 129X1/SvJ male origin, were cultured as previously described.8, 9 The HIFC differentiation system was performed as follows: undifferentiated ES cells were treated for 3 days in an ES cell culture medium containing leukemia inhibitory factor (100 U/mL) and 10−8 mol/L all-trans-retinoic acid; for 5 days with a leukemia inhibitory factor (−) culture medium in the presence of a growth factor combination (fibroblast growth factor 1, 100 ng/mL; fibroblast growth factor 4, 20 ng/mL; hepatocyte growth factor, 50 ng/mL [VERITAS, Tokyo, Japan]) in gelatin-coated dishes (Asahi Techno Glass, Chiba, Japan); and for 2 days with a leukemia inhibitory factor (−) culture medium in the presence of oncostatin M (10 ng/mL) (VERITAS) in type I collagen–coated dishes (Asahi Techno Glass). Differentiated hepatocytes were identified as green fluorescent protein (GFP)-positive. GFP gene expression was monitored via fluorescence microscopy (Nikon, Tokyo, Japan).
Isolation of Total RNA.
Total RNA was extracted from undifferentiated ES cells, HIFC-treated cells, and HIFC-untreated cells using ISOGEN solution (Nippon Gene, Tokyo, Japan) according to the manufacturer's protocol and treated with deoxyribonuclease (DNase I, amplification grade; TaKaRa, Kyoto, Japan).
AceGene mouse oligo chip subsets A and B (DNA Chip Research and Hitachi Software, Yokohama, Japan) DNA microarrays were used according to the manufacturer's instructions (http://www.dna-chip.co.jp/thesis/AceGeneProtocol.pdf). MIAME (minimum information about a microarray experiment) guidelines were applied for microarray data analysis.10 RNA was extracted from day 0 (undifferentiated ES cells), day 6 (+) (HIFC-treated sample), day 6 (−) (HIFC-untreated sample), day 10 (+) (HIFC-treated sample), and day 10 (−) (HIFC-untreated sample) cells and 129X1/SvJ mouse liver (8 weeks old). As a total RNA reference, equal amounts of day 0, day 6 (+), day 6 (−), day 10 (+), and day 10 (−) RNA were mixed together. To ensure data reliability, weak signal spots were removed according to the following criteria. For each sample, the experiment was repeated once, wherein the dye was reversed between the experimental and the reference sample to account for dye-incorporation bias. This resulted in a data matrix of 9,172 genes that contained no missing data.
Genes [day 10 (+) sample data] that showed an increase or decrease in their expression levels of more than twofold compared with day 10 (−) sample data were designated as upregulated and downregulated genes, respectively. A hierarchical cluster was produced from upregulated and downregulated gene data using a Euclidean distance calculation based on the Unweighted Pair Group Method with Arithmetic Mean by GenMaths software (Applied Maths). Gene ontology (GO) categories were assigned to genes based on the AceGene microarray database (DNA Chip Research and Hitachi Software, Yokohama, Japan). The significance of GO term appearance in the upregulated and downregulated genes (compared with all genes: 9,172 genes) was calculated as the P value using the software GO Term Finder adapted to the AceGene microarray (http://db.yeastgenome.org/cgi-bin/SGD/GO/goTermFinder). Cutoff points were set at 0.005.11 By comparing the day 10 (+) and day 10 (−) and focusing on genes with twofold or more alteration in expression level, we reasoned that these gene profiles represented the main cell population (GFP-positive cells) and not the minority cell populations (GFP-negative cells).
Reverse-Transcription Polymerase Chain Reaction.
Total RNA (1 μg) was reverse-transcribed using the SuperScript II First-Strand Synthesis System (Invitrogen, Tokyo, Japan) according to the manufacturer's protocol. Polymerase chain reaction (PCR) reactions were performed using the Ampli Taq Gold kit (Roche Diagnostics, Tokyo, Japan). PCR primers and conditions are listed in Table 1.
Table 1. Primers and Conditions for RT-PCR and Real-Time PCR
GeneBank Accession No.
Annealing Temperature (°C)
HNF1α S: 5′-GCCTAATGGCCTTGGAGAAA-3′
HNF1β S: 5′-ACCAAGCCGGTTTTCCATAC-3′
HNF3α S: 5′-GTGAAGATGGAAGGGCATGA-3′
HNF3β S: 5′-CGCCAACATGAACTCGATGA-3′
HNF3γ S: 5′-TGAAGATGGAGGCTCATGAC-3′
HNF4α S: 5′-ATTCATCCAACAGCCTGAGC-3′
HNF6 S: 5′-CAAATGGCTTTGAAGCCCAC-3′
C/EBPα S: 5′-CAACGCCCGCCTTTGGCTTT-3′
C/EBPβ S: 5′-CGGATCAAACGTGGCTGAG-3′
GATA-4 S: 5′-CAAGATGAACGGCATCAACC-3′
GATA-6 S: 5′-CATCACCATCACCCGACC-3′
Trf S: 5′-ACTGCCATTCGGAATCA-3′
Apoa2 S: 5′-CACTGCTGGTAACCATCTG-3′
Cyp2e1 S: 5′-GTCAGAGGCGCATCGT-3′
Pzp S: 5′-TACTTGGCTCACTGCGTTTG-3′
Mup1 S: 5′-GCTCCGAATTATCTATGGTT-3′
Cst3 S: 5′-GTTCCTGCTGGCCGTCCT-3′
Anxa2 S: 5′-CATTGCCTTCGCCTATCAGA-3′
Acvr2b S: 5′-CCGGCATGAAGCACGAAA-3′
βactin S: 5′-AGAGCAAGAGAGGTATCCTG-3′
HNF3β S: 5′-CGAGAACGGCTGCTACCTG-3′
ALB S: 5′-AAGCTGAGACCTTCACCTTC-3′
TTR S: 5′-AAAGACCTCTGAGGGATCCT-3′
G6Pase S: 5′-CGCAGCAGGTGTATACTATG-3′
GAPDH S: 5′-ATCACTGCCACCCAGAAGAC-3′
Transfection With HNF3β/FoxA2 siRNA.
Synthetic 21-nt RNAs were purchased from QIAGEN (Tokyo, Japan) in deprotected, desalted, and annealed form. The sequences of our prepared mouse HNF3β/FoxA2 siRNAs are listed in Table 2. Control luciferase GL2 siRNA was purchased from Dharmacon (Boulder, CO). Twenty-four hours before transfection with the ES cell culture medium, 1 × 104 cells/mL of regularly passaged ESJ1 pALB-EGFP cells were plated. Lipofectamine 2000-mediated transient transfection of siRNA was performed on days 1 and 3 as described by the manufacturer (Invitrogen). GFP-positive cells were monitored on day 7 via fluorescence microscopy (Nikon).
Table 2. Sequences of HNF3β/FoxA2 siRNA
ES cells were transfected with an Alexa-labeled siRNA (QIAGEN, Tokyo, Japan) in the same manner. Five hours after transfection, the cells were washed with phosphate-buffered saline and observed via fluorescence microscopy (Nikon) at ×100 magnification. Alexa-positive cells and total cells were counted in five randomly selected fields. The transfection efficiency was calculated as a percentage of Alexa-positive cells per well.
SYBR Green Real-Time PCR.
Real-time RT-PCR primers are listed in Table 1. The complementary DNAs were used for PCR using Platinum SYBR Green qPCR SuperMix UDG (Invitrogen) in triplicate. Optimization of the real-time PCR reaction was performed according to the manufacturer's instructions (PE Applied Biosystems, Tokyo, Japan). All quantitations were normalized to an endogenous control GAPDH. Relative gene induction values were calculated according to the manufacturer's instructions.
The results are given as the mean ± SD. Statistical analysis was conducted using ANOVA with the Bonferroni correction for multiple comparisons. A P value of .05 or less was considered significant.
Differentiation and Microarray Analysis of ES Cell–Derived Hepatocytes.
We previously established and reported the HIFC differentiation system,7 whereby ES cells could be differentiated in 10 days into hepatocytes without forming embryoid bodies at a rate of 29.6% ± 1.8% (data not shown). Therefore, our novel culture system represents a reproducible model for studying the molecular mechanisms underlying hepatic differentiation per se.7 To analyze genes related to hepatic differentiation, a profile of day 10 cells treated with HIFC [10 (+)] was compared with that of day 10 cells without HIFC [10 (−)]. Of the 9,172 genes analyzed, 232 genes showed a significant twofold or more alteration of the expression level, indicating that expression levels of these genes were altered by HIFC treatment.
Of the 232 genes, 30 of the most highly upregulated and downregulated genes are listed in Tables 3 and 4, respectively. Genes upregulated during the hepatic differentiation of ES cells comprised many metabolic enzymes, such as cytochrome P450 and alcohol dehydrogenase,12 and serum proteins, such as transthyretin, albumin, and major urinary protein 1 (Table 3).13, 14 On the other hand, downregulated genes in the hepatic differentiation of ES cells contained glycolysis-related proteins such as phosphoglycerate kinase 1 and lactate dehydrogenase 1, A chain (Table 4). Additionally, many of the 232 genes in the ES cell–derived hepatocytes are cell growth–, proliferation-, and physiology-associated genes. This includes 51 genes for cellular metabolism, 53 genes related to development, morphogenesis and differentiation, and 14 genes encoding cell–cell signaling and cell adhesion. Furthermore, several ES cell marker genes containing Oct3/4 and Nanog were downregulated as measured via microarray profiles and RT-PCR analysis (data not shown). Collectively, these results suggest that HIFC treatment induced differentiation from ES cells into cells with a gene expression profile typical of differentiated hepatocytes.
Table 3. Upregulated Genes in ES Cell–Derived Hepatocytes
GeneBank Accession No.
NOTE. Two hundred genes show the increase in their expression levels of more than twofold compared with day 10(−). Thirty of the most upregulated genes are listed.
BCL2/adenovirus E1B 19kDa-interacting protein 1, NIP3
Phosphoglycerate kinase 1
Keratin complex 1, acidic, gene 19
Hypothetical protein MGC37636
Heterogeneous nuclear ribonucleoprotein A1
Lactate dehydrogenase 1, A chain
RIKEN cDNA 1500011H22 gene
RIKEN cDNA 0710007A14 gene
Clustering Analysis of ES Cell–Derived Hepatocytes.
Unsupervised hierarchical cluster analysis was performed to sort the 232 altered genes (Fig. 1A). The case cluster analysis of microarray data revealed a striking mirroring of gene clusters between adult liver and day 10 (+) cells, indicating that ES cell–derived hepatocytes were similar to adult mouse liver in another aspect of the gene expression pattern. Interestingly, these analyses showed that significant gene expression alterations occurred during each stage of the process of hepatic differentiation [day 0, day 6 (+), and day 10 (+)], suggesting that ES cells differentiated into mature hepatocytes from nonhepatic precursor cells. The hierarchical clustering method delineated two distinctive major clusters according to the following expression patterns: group A (downregulated genes), consisting of 32 genes, and group B (upregulated genes), consisting of 200 genes (Fig. 1A).
To validate the results of the microarray analysis, a random selection of genes that were upregulated and downregulated in day 10 (+) cells was analyzed. As shown in Fig. 1B, the expression of upregulated genes Trf, Apoa2, and Cyp2e1 and downregulated genes Cst3, Anxa2, and Acvr2b was confirmed via RT-PCR. These results confirm the accuracy of the transcriptional regulation that was acquired from the microarray experiments.
GO Classification of ES Cell–Derived Hepatocytes.
Using the database, the microarray analysis data were integrated to identify the GO biological processes for the upregulated and downregulated genes. This analysis indicated that four GO groups were highly significant for the upregulated and downregulated genes vis-à-vis the parent population (9,172 genes) (Table 5). The probabilities of observing such a number of genes in these categories by chance were extremely small, ranging from 3.7 × 10−4 to 2.5 × 10−3. Interestingly, all of these GO groups are relevant to hepatocyte function, reinforcing the conclusion that hepatocyte-related genes were induced by HIFC treatment. For example, electron transport contains CYP 2e1 and CYP 2d10 genes. Thus, the result of GO analysis suggests that hierarchical clustering analysis of the hepatic differentiation of ES cells reflects a gene expression pattern similar to that of normal mouse liver.
Table 5. Significance of Gene Ontology Appearance in Upregulated and Downregulated Genes
NOTE. The significance of GO term (biological process) appearance in upregulated and downregulated genes was calculated as a P value by the software.
Of 232 genes, GO biological process is known for 183. Others are unknown.
Of 9,172 genes, GO biological process is known for 6,699. Others are unknown.
6 out of 183 genes (3.3%)
17 out of 6,699 genes (0.3%)
Response to pest/pathogen/parasite
13 out of 183 genes (7.1%)
103 out of 6,699 genes (1.5%)
6 out of 183 genes (3.3%)
22 out of 6,699 genes (0.3%)
14 out of 183 genes (7.7%)
140 out of 6,699 genes (2.1%)
Expression Pattern of Liver-Enriched Transcriptional Factors.
To elucidate the gene expression profile of endodermal and hepatic transcription factors in the HIFC differentiation system, RT-PCR analysis was performed. Endoderm-associated transcription factors HNF1β/TCF2, HNF3α/FoxA1, HNF3β/FoxA2, and HNF4α were upregulated from day 1 during HIFC differentiation (Fig. 2A). These genes were not expressed in undifferentiated ES cells (day 0) and were not detected until day 3 in the HIFC-untreated cells (Fig. 2A). In contrast, the hepatocyte-enriched transcription factors HNF6/OC-1 and CCAAT/enhancer-binding proteins α and β were upregulated at the hepatocyte maturation stage (day 6 and thereafter) in HIFC-treated cells (Fig. 2B), whereas control cells did not show any upregulation of these transcription factors at any stage. These results indicated that endoderm-associated transcription factors were expressed at the early stage of the HIFC differentiation system and that, during the late stage of this system, hepatocyte-enriched transcription factors required for the maturation of hepatocytes were induced. These data suggest that HIFC differentiation directs ES cells to generate mature hepatocytes via endoderm differentiation.
HNF3β/FoxA2 Signaling Is Essential for Hepatic Differentiation of ES Cells.
To further confirm that lineage-specific transcription factor networks were regulated in the HIFC differentiation system and that ES cells differentiated into hepatocytes via endoderm differentiation, we disrupted the expression of the endoderm-associated transcription factor HNF3β/FoxA2, which is expressed during the early stage of the HIFC differentiation system (Fig. 2A). First, the transfection efficiency of synthetic siRNA in ES cells was estimated to be approximately 70% using Alexa-labeled siRNA detectable via fluorescence microscopy (Fig. 3A). Second, we chose several siRNAs targeted against several sequences in the coding region of mouse HNF3β/FoxA2. HNF3β/FoxA2-siRNAs were transfected with liposome on days 1 and 3 in the HIFC-differentiation system. Among tested siRNA species, the HNF3β/FoxA2-siRNA1 treatment showed the most significant effect, with over 60% inhibition of HNF3β/FoxA2 expression levels, compared with the luciferase-siRNA control, by real-time PCR analysis (Fig. 3B). Therefore, HNF3β/FoxA2-siRNA1 was selected for further studies.
We investigated whether loss of HNF3β/FoxA2 expression affects the hepatic differentiation of ES cells by examining the effect of HNF3β/FoxA2-siRNA1 on the GFP-positive rate in the siRNA-transfected cells. GFP expression depends on albumin (ALB) expression under the control of the ALB promoter. As shown in Fig. 4A, GFP-positive cells were observed in both HNF3β/FoxA2 siRNA1–transfected and luciferase siRNA–transfected cells. However, the HNF3β/FoxA2 siRNA1–transfected cells showed a greater than 60% decrease in GFP-positive rate compared with luciferase siRNA–transfected cells (Fig. 4B). This finding shows that differentiation of ALB-positive cells is suppressed in HNF3β/FoxA2 siRNA1–transfected cells. Additionally, real-time RT-PCR analysis showed that expression levels of transthyretin and glucose-6-phosphatase, which were typically expressed in mature hepatocytes, were also downregulated as well as albumin in the HNF3β/FoxA2 siRNA1–transfected cells (Fig. 4C). These results suggest that transcription factor networks are regulated precisely in the HIFC differentiation system and that ES cells differentiate into mature hepatocytes via endoderm differentiation.
During mouse development, the first morphological sign of hepatogenesis is at 8.5 days postcoitum, when the visceral floor of the foregut endoderm thickens to form the liver diverticulum.15 Tissue transplantation studies have shown that signals from the cardiac mesoderm induce the hepatic differentiation of the endoderm.16, 17 Cells of the hepatic endoderm then begin to migrate in a cord-like fashion into the surrounding mesenchyme of the septum transversum. There, the cords intermingle with the vitelline veins, which themselves begin to anastomose into a venous bed where the liver begins to form.18 Just before birth and shortly thereafter, a large number of liver metabolic enzymes are induced. After birth, the liver acquires additional metabolism functions and becomes fully mature.19 To understand the molecular mechanism for liver development in vitro, previous studies have attempted to show that ES cells differentiate into hepatocytes by forming embryoid bodies in vitro.20–23 None of the reports, however, elucidated the molecular mechanisms underlying the differentiation of ES cells into hepatocytes, because the formation of embryoid bodies is not easily controlled. In this study, using the HIFC differentiation system, our data clearly demonstrated the applicability of microarray analysis and the controlled and reproducible hepatic differentiation of ES cells. Moreover, analysis of GO groups indicated that significant categories of GO appearance in 232 genes contained categories relevant to hepatic function. This integrative perspective on the gene expression profile validates our HIFC differentiation method for hepatic induction from ES cells. Therefore, the microarray analysis readout provides a potentially valuable resource for further detailed dissection of key molecules and demonstrates that these gene networks require that ES cells differentiate into hepatocytes.
The expression of endoderm-associated transcription factor genes was detected in the early stage of the hepatic differentiation of ES cells. Previous reports showed that these transcription factors were also expressed in the visceral endoderm in vivo and induce transactivation of several endoderm- and hepatocyte-specific factors, including transthyretin, albumin, and L-pyruvate kinase.24–28 Additionally, HNF4α is essential for morphological and functional differentiation of hepatocytes, accumulation of hepatic glycogen stores, and generation of the hepatic epithelium.29 A previous report suggested that the cell maturation effect of oncostatin M is mediated through HNF4α in cultured human fetal hepatocytes.30 RT-PCR analysis revealed that, in our system, HIFC-treated cells showed increased HNF4α expression critically at a late stage in the medium containing oncostatin M. Thus, our experimental data suggest that the HIFC differentiation system induces ES cells to differentiate into mature hepatocytes via the endoderm.
Visceral endoderm is an extraembryonic tissue derived from the inner cell mass of the embryo. It is functionally homologous to hepatocytes involved in lipid and glucose metabolism and serum protein production. Because of the difficulty in distinguishing visceral endoderm from embryonic endoderm, it is theoretically possible that our ES cell–derived hepatocytes might be hepatocyte-like cells from the visceral endoderm. However, a recent study demonstrated that the Cyp7A1 gene is expressed in hepatocytes derived from definitive endoderm, but not visceral endoderm.31 In agreement with these findings, our HIFC-treated ES cell–derived hepatocytes expressed Cyp7A1 (data not shown). Additionally, based on both our previous7 and present data, we have demonstrated via extensive characterization that our system produces the hepatocytes from mouse ES cells. Moreover, recent data from our laboratory has indicated that the HIFC system can be modified to allow differentiation of human mesenchymal stem cells into functional human hepatocytes (unpublished observation).
HNF3β/FoxA2 is known to be important in endoderm specification and subsequent hepatocyte differentiation in vivo.26, 32–34 On the other hand, no reports have shown its necessity in in vitro experiments because of the inability to differentiate ES cells into hepatocytes. In this study, HNF3β/FoxA2 siRNA transfection clearly repressed the differentiation into ALB-expressed cells. This result is in line with previous reports that HNF3β/FoxA2 is involved in early hepatocyte differentiation based on an in vivo animal model, suggesting that our HIFC differentiation system may recapitulate hepatic differentiation in vivo and may be useful for the study of a liver-specific transcriptional network.
Much is known about the molecular control of ectoderm and mesoderm development; however, little is known about endoderm regionalization and subsequent organ formation, which have been difficult to study in vitro because of the lack of a suitable model. According to our data, the HIFC differentiation system allows ES cell differentiation into hepatocytes along a developmentally similar route. Previously, we reported that the HIFC differentiation system was critical for alpha-fetoprotein (AFP) gene expression at an early stage.7 AFP is first detected in the gut endoderm at the four-somite stage of the mouse embryo.35, 36 In addition, AFP expression declines rapidly after birth, and the level of its messenger RNA in adult liver is less than 0.01% of that in fetal liver. During rapid hepatocyte proliferation, such as liver regeneration or tumorigenesis, AFP expression is reactivated. Hence, AFP might be a marker for liver progenitor cells or proliferating liver cells, such as oval cells.37, 38 Therefore, using this system, hepatic stem cells such as oval cells might be differentiated from ES cells and may be acquired abundantly. In support of this presumption, oval-like cell marker genes containing cytokeratin 19,39 Thy-1,40 γ-glutamyl transpeptidase,41 and Dlk42 are detectable using the HIFC differentiation system at early and middle stages (unpublished observations). Furthermore, this system is a valuable production model for the identification and characterization of factors involved in endoderm specification and hepatogenesis from hepatic stem cells. In the future, this system could be applied to the screening of key molecules for hepatic specification through an siRNA library approach.
In conclusion, the HIFC differentiation system closely mimics the actual in vivo hepatic developmental program: in development stage 1, pluripotent ES cells appear; in stage 2, endoderm markers, such as HNF3β/FoxA2, become positive; in stage 3, immature hepatocytes expressing ALB and AFP appear; and in stage 4, mature hepatocytes positive for ALB and tryptophan 2,3-dioxygenase are produced. This system could be a valuable tool for a variety of applications, such as toxicity testing, drug screening, study of the molecular biology of hepatitis B and C viruses, and production of therapeutic hepatocytes from human ES cells suitable for cell transplantation.
We are very grateful to Dr. Fumitaka Takeshita, Ayako Inoue, Kimi Honma, Nachi Namatame, Maho Kodama, Shinobu Ueda, and Akemi Sugai for their excellent technical assistance.