- Email: [email protected]
Hepatocyte-like cells derived from human embryonic stem cells specifically via definitive endoderm and a progenitor stage
Journal of Biotechnology 145 (2010) 284–294
Contents lists available at ScienceDirect
Journal of Biotechnology journal homepage: www.elsevier.com/locate/jbiotec
Hepatocyte-like cells derived from human embryonic stem cells specifically via definitive endoderm and a progenitor stage Gabriella Brolén a , Louise Sivertsson c , Petter Björquist a , Gustav Eriksson a , Monica Ek c , Henrik Semb d , Inger Johansson c , Tommy B. Andersson b , Magnus Ingelman-Sundberg c , Nico Heins a,∗ a
Cellartis AB, Arvid Wallgrens Backe 20, 41346 Göteborg, Sweden Development DMPK & Bioanalysis, AstraZeneca R&D Mölndal, 431 83 Mölndal, Sweden Karolinska Institute, Section of Pharmacogenetics, Department of Physiology and Pharmacology, Nanna Svartz väg 2, 17177 Stockholm, Sweden d Stem Cell Center, Lund University, BMC B10, Lund 22184, Sweden b c
a r t i c l e
i n f o
Article history: Received 24 March 2009 Received in revised form 9 November 2009 Accepted 12 November 2009
Keywords: Human embryonic stem cell Definitive endoderm Hepatocyte Liver progenitor Activin
a b s t r a c t Human embryonic stem cells offer a potential unlimited supply for functional hepatocytes, since they can differentiate into hepatocyte-like cells displaying a characteristic hepatic morphology and expressing various hepatic markers. These cells could be used in various applications such as studies of drug metabolism and hepatotoxicity, which however, would require a significant expression of drug metabolizing enzymes. To derive these cells we use a stepwise differentiation protocol where growth- and maturation factors are added. The first phase involves the formation of definitive endoderm. Next, these cells are treated with factors known to promote the induction and proliferation towards hepatic progenitor cell types. In the last phase the cells are terminally differentiated and maturated into functional hepatocyte-like cells. The cultures were characterized by analysis of endodermal or hepatic markers and compared to cultures derived without induction via definitive endoderm. Hepatic functions such as urea secretion, glycogen storage, indocyanine green uptake and secretion, and cytochrome P450-expression and activity were evaluated. The DE-Hep showed a hepatocyte morphology with sub-organized cells and exhibited many liverfunctions including transporter activity and capacity to metabolize drugs specific for important cytochrome P450 sub-families. This represents an important step in differentiation of hESC into functional hepatocytes. © 2009 Elsevier B.V. All rights reserved.
1. Introduction Pluripotent human stem cells are expected to change the accessibility to a variety of human cell types. The possibility to propagate pluripotent human embryonic stem cells (hESC) and subsequently differentiate them into the desired target cell types will provide a stable and virtually unlimited supply of cells for a range of applications in vivo and in vitro. The liver is the major organ for metabolism and detoxification in the human body, and therefore huge efforts have been undertaken in order to identify a reliable source of functional hepatocyte-like cell types for in vitro testing. Up to now the complexity and function of the liver are not reflected by any available in vitro cell type or cellular system. The availability of human primary liver cells is limited and furthermore these cells lose their normal phenotype and functional properties (typically within 24 h) when used in
∗ Corresponding author. Fax: +46 31 758 09 10. E-mail address: [email protected] (N. Heins). 0168-1656/$ – see front matter © 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.jbiotec.2009.11.007
vitro. Transformed hepatic cell lines contain very low levels of drug metabolizing enzymes and have expression profiles of other important proteins substantially different from the native hepatocyte in vivo. Thus, many tests e.g. drug metabolism and toxicity are still performed using animal material, even though liver metabolism is known to be species-specific. Due to this, difficulties to predict liver toxicity in another species than the one tested remain present. Therefore early prediction of human liver toxicity liabilities is of principal importance when selecting compounds to enter clinical trials. Accordingly there is an urgent need for a model system that mimics human liver cells and that is able to predict effects of candidate molecules in the development of new drugs or chemicals. Regarding both availability and physiological relevance human pluripotent stem cells may serve as an ideal renewable source of functional human hepatocytes. When hESC have been placed in a proper environment, certain hepatic characteristics have been observed 2–4 weeks after differentiation. The present investigation is based on the fact that definitive endoderm (DE) cells give rise to endodermal organs and consequently to e.g. hepatic cell types. Early endoderm development is not well understood. Fate
G. Brolén et al. / Journal of Biotechnology 145 (2010) 284–294
Fig. 1. Schematic overview of the differentiation strategy into definitive endoderm derived hepatocytes (DE-Hep). The differentiation protocol is divided into three main phases. The starting material consists of undifferentiated hESC cultured on MEF. In phase I the hESC are induced into definitive endoderm (DE). In phase II the DE is induced to early liver endoderm or liver progenitor cells. During phase III the DE-Hep cultures are maturated.
mapping studies of cultured mouse embryos (Lawson et al., 1986, 1991; Lawson and Pedersen, 1987) have revealed that DE begins to form at the embryonic days 6–6.5 (E6–6.5) and that by the end of gastrulation (E7.5), some labeled cells only give rise to endodermal derivatives. Other fate mapping studies (Lawson et al., 1991; Tremblay and Zaret, 2005) suggested that the first endodermal cells that migrate through the primitive streak (PS) at E6.5 are fated to become liver, ventral pancreas, lungs and stomach. A complication in the study of endoderm is that mammals generate extraembryonic endoderm (ExE). ExE arises at the blastocyst stage and eventually forms two subpopulations: visceral endoderm and parietal endoderm. These cells share the expression of many genes with DE (cells that give rise to the endodermal organs), including the transcription factors Sox17 (Kanai-Azuma et al., 2002), HNF3 and HNF4␣ (Belo et al., 1997; Sasaki and Hogan, 1993). D’Amour et al. (2005, 2006) have developed a protocol for deriving DE from hESC. Previous studies have identified cells with some hepatocyte-like characteristics, e.g. cytochrome P450 (CYP) activity and ability to store glycogen, in differentiated hESC cultures (Rambhatla et al., 2003; Cai et al., 2007; Duan et al., 2007; Agarwal et al., 2008). So far the cells generated have not shown the metabolic qualities necessary for potentially replacing traditional liver systems in terms of drug transporter expression and specific CYP expression patterns needed for industrial applications. In this investigation several cell lines have been induced by growth factors to generate hepatocyte-like cells via definitive endoderm (DE-Hep). The cells were subsequently cultured under conditions known to support the development of liver cell lineages (Hamazaki et al., 2001; Hamazaki and Terada, 2003). Our strategy involves three distinct important steps in liver development. The study shows that the hepatocyte-like cell population obtained via DE has many similarities with human adult hepatocytes, including capacity to metabolize several specific pharmaceutical compounds.
285
ation into DE recombinant human Activin A (100 ng/ml, R&D) was supplemented to the medium, the following days of DE differentiation (phase I) Activin A (100 ng/ml) and sometimes FGF2 (4 ng/ml) or Wnt3a (50 ng/ml) was added to the medium. Phase II was carried out either in RPMI or DMEM (both from Invitrogen/Gibco). Different growth factors were added to the media. In phase II, RPMI advanced medium supplemented with PEST (1%) (Invitrogen/Gibco) and Glutamax (1%) (Invitrogen–Sigma) was used as a basic medium. BMP4 (100 ng/ml, PromoKine) and FGF2 (4 ng/ml) were added to the medium to differentiate the DE into hepatic progenitors. In some experiments in phase II the medium was supplemented with FGF1 (100 ng/ml, PromoKine), FGF2 (5 ng/ml, R&D), BMP2 (50 ng/ml, PromoKine), BMP4 (200 ng/ml) and 0.2% FBS (Gibco), with medium change every second day. At phase III hepatocyte maturation was carried out in HCM (Cambrex) with dexamethasone (Dex) (0.1 M, Calbiochem), oncostatin M (OSM) (10 ng/ml, PromoKine), HGF (20 ng/ml, PromoKine) and SingleQuots (Lonza). The medium was changed every second day. 2.3. Culture of controls and tissue material As controls we used intrinsic differentiated hESC as well as mouse embryonic feeder cells (MEF), HepG2 cells, human primary hepatocytes, and human liver samples. Intrinsic differentiated hepatocyte-like cells that we use as control in this study were derived according to Söderdahl et al. (2007) and Ek et al. (2007). hESC were allowed to differentiate for 18–45 days in VitroHESTM supplemented with 4 ng/ml FGF2. MEFs were derived and cultured as previously described (Hogan et al., 1994). HepG2 cells (HB-8065, American Type Culture Collection) were cultured as previously described (Butura et al., 2004). HepG2 cells (HB-8065, American Type Culture Collection, Manassas, VA) were cultured in minimum essential medium (MEM) supplemented with 10% FBS, 1% penicillin–streptomycin, 1% sodium pyruvate and 1% non-essential amino acids (all from Invitrogen) and collected by trypsinization when 80–90% confluence were reached. Platable cryopreserved human primary hepatocytes (In Vitro Technologies) were cultured according to the manufacturer instructions. In short, hepatocytes were thawed at 37 ◦ C and propagated at 0.3 million cells per ml and well in InVitroGRO CP medium (In Vitro Technologies) at collagen I-coated 24-well plates. After 24–48 h in vitro the hepatocytes were fixed and subjected to immunofluorescence stainings. After 24–120 h in vitro the hepatocytes were harvested for activity assays and mRNA. All cultures were kept at 37 ◦ C, 5% CO2 and 90–95% humidity. Human liver samples were obtained from Sahlgrenska Hospital (Göteborg, Sweden) and originated from patients undergoing liver resection. All tissues were obtained through qualified medical staff, with donor consent and with the approval of the Local Ethics Committee at Sahlgrenska Hospital.
2. Methods
2.4. Immunocytochemical methods and antibodies
2.1. Culture of hESC
Cells in culture were fixed in 4% (w/v) paraformaldehyde (PFA) for 15 min, and washed in PBS. For visualizing extracellular proteins cells were blocked for 30 min in 5% FBS in PBS, for intracellular protein staining cells were blocked and permeabilized for 30 min in 5% FBS in 0.1% TritonX-100 in PBS. The primary antibodies were incubated in 1% FBS in PBS overnight at 4 ◦ C and the secondary antibodies were diluted in 1% FBS in PBS for 1 h at room temperature (RT). For CYP-stainings the cells were incubated with secondary antibodies in PBS for 3 h at RT. All washes were performed in PBS. To visualize the nucleus cells were incubated with DAPI at 0.05 mg/ml for 5 min at RT and mounted in DAKO Cytomation mounting medium. Primary antibodies used were rabbit anti-albumin (1:500, DAKO Cytomation), rabbit anti-␣1 antitrypsin (1:200, DAKO Cytomation), mouse IgG2a ␣FP (1:500,
The hESC lines SA001, SA002, SA002.5 and SA167 (Cellartis AB, Göteborg, Sweden, http://www.cellartis.com) were derived, cultured and characterized as previously described (Heins et al., 2004, 2006). 2.2. Culture of DE-Hep Differentiation of hESC into DE-Hep cells was carried out as described in Fig. 1. Before initiating differentiation, hESC were washed in PBS+/+ (Invitrogen/Gibco). During differentiation into DE (phase I) the fetal calf serum concentration was 0% the first day and 0.2% for the following days. During the first day of differenti-
286
G. Brolén et al. / Journal of Biotechnology 145 (2010) 284–294
Sigma), rabbit anti-CK7 (1:200, Novocastra), mouse IgG1 anti-CK8 (1:200, Santa Cruz), mouse anti-CK18 (1:200, DAKO Cytomation), mouse IgG1 CK19 (1:200, Novocastra), mouse IgG2b anti-CXCR4 (1:250, R&D Systems), rabbit anti-CYP1A2 (1:100, Biomol), sheep anti-CYP3A4/7 (1:100, Biomol), mouse IgG1 anti-EpCAM (1:50, GeneTex), goat anti-LFABP (1:500, Santa Cruz), goat anti-HNF3 (1:400, Santa Cruz Biotechnology), rabbit anti-HNF1␣ (1:400, Santa Cruz), rabbit anti-HNF4␣ (1:300, AH Diagnostics), mouse antiICAM-1 (CD54) (1:500, BD Pharmingen), rabbit anti-MRP2 (1:50, Santa Cruz), goat anti-Sox7 (R&D), rabbit anti-Sox17 (1:2000, a kind gift from E. Baetge). Secondary antibodies used were Alexa Fluor 488 conjugated donkey anti-goat IgG (1:500, Molecular Probes), Alexa Fluor 488 donkey anti-rabbit IgG (1:1000, Molecular Probes), Alexa Fluor 488 donkey anti-sheep IgG (1:1000, Molecular Probes), Alexa Fluor 594 donkey anti-rabbit IgG (1:1000, Molecular Probes), Cy2 donkey anti-mouse IgG (1:100, Jackson Immuno Research), Cy2 donkey anti-rat IgG (1:100, Jackson Immuno Research), Cy3 donkey anti-mouse IgG (1:1000, Jackson Immuno Research), Cy3 donkey anti-rat IgG (1:500, Jackson Immuno Research), fluorescein isothiocyanate (FITC) anti-mouse IgG (1:500, Jackson Immuno Research). Stained samples were examined using a Nikon Eclipse TE2000-U Fluorescence microscope and Nikon Act-1C for DXM1 200C software.
a culture was treated with human saliva for 20 min at RT and subsequently washed in PBS. The human saliva contains ␣-amylase which digests glycogen. Periodic acid was added to the treated and untreated cultures for 5 min at RT followed by repeated washing in PBS. Subsequently, cultures were incubated in Schiff’s reagent for 15 min at RT. After washing in PBS cells were rinsed in H2 O prior to mounting in mounting media.
2.5. Western blot
Gene-specific primer pairs were designed and evaluated for an annealing temperature of 60 ◦ C using freely available Web-based software (Primer3, Netprimer, Beacon Designer 2.1, mFold, and Oligonucleotide Properties Calculator). Primers were designed for the following genes: Sox17, HNF3, CXCR4, ␣FP and ␣1 -antitrypsin. The optimized assays, including reference material, are available from TATAA Biocenter (Göteborg, Sweden, http://www.tataa.com). All Q-PCRs in Fig. 2 were performed with SYBR Green I chemistry in a Rotorgene 3000. The authenticity of the PCR products was verified by melt-curve analysis and agarose-gel electrophoresis. For Q-PCRs, 1× Jump Start Buffer 10× (Sigma–Aldrich), 3 mM MgCl2 (Sigma–Aldrich), 0.3 mM dNTP mix (Sigma–Aldrich), 0.4× SYBR Green (Molecular Probes, Inc., Eugene, OR, http://probes.invitrogen.com), 0.4 M forward primer (MWG Biotech, Ebersberg, Germany, http://www.mwgbiotech.com), 0.4 M reverse primer (MWG Biotech), 0.04 U/l Jump Start taq polymerase (Sigma–Aldrich), and 2 l cDNA template were used in a final volume of 20 l. After an initial denaturation/activation step of 3 min at 95 ◦ C followed 45 cycles of 20 s at 95 ◦ C, 20 s at 60 ◦ C, and 20 s at 72 ◦ C. The detection of fluorescent signal was performed at 72 ◦ C in each cycle. Ct (threshold cycle) values were calculated using the Rotorgene software. For some genes (albumin, ␣1 -antitrypsin, ␣FP, FABP1, HNF3, HNF4␣, CYP3A4, CYP1A2, CYP2C9, OATP2, MRP2, CDX2 and Vilin) TaqMan-based Q-PCR was used, as previously described by Ek et al., 2007.
DE-Hep progenitor cells, DE-Hep cells and its control cells (intrinsically differentiated hESC, HepG2 and human primary hepatocytes) were lysed in TritonX buffer (1% TritonX-100, 1 mM EDTA and 1:100 protease inhibitor cocktail (Roche, Germany)) and centrifuged at 800 × g and 4 ◦ C for 10 min. Protein determination was measured according to the method of Bradford (1976). The following primary antibodies were used: human albumin (Bethyl Laboratories Inc., TX, USA), ␣1 -antitrypsin (Dako Cytomation, Glostrup, Denmark), ␣-fetoprotein (Cell Signaling Technology Inc., MA, USA), CYP3A (BD Biosciences, CA, USA), MRP2, OATP2 (Santa Cruz Biotechnology, CA, USA). Horseradish peroxidase-conjugated secondary antibodies (Dako Cytomation, Glostrup, Denmark) were visualized with Pierce supersignal chemiluminescense substrate (Pierce, IL, USA) using the LAS-1000 + gel documentation system, and analyzed using Image Gauge software (Fujifilm, CT, USA). 2.6. Indocyanine green (ICG) Cardiogreen (Sigma, l2633) was dissolved in 5 ml of solvent in a sterile vial and then added to 20 ml of DMEM containing 10% FBS. The final concentration of the resulting indocyanine green (ICG) solution was 1 mg/ml. The ICG solution was added to the cell culture dish and incubated at 37 ◦ C for 60 min. After the cells were rinsed three times with PBS, the cellular uptake of ICG was examined with a stereomicroscope. After the analysis the dish was refilled with DMEM containing 10% FBS. ICG was eliminated from the cells overnight. 2.7. ELISA analysis of urea in the medium For urea excretion the Roche UREA/BUN assay was used. Urea secretion was analyzed using a kit for kinetic UV assay for urea/urea nitrogen (Roche/Hitachi) by enzyme-linked immunosorbent assay (ELISA) at Klinisk Kemi, C-lab, Sahlgrenska University Hospital, Göteborg. 2.8. Glycogen storage detection Cells were fixed in 100% methanol for 15 min at RT and subsequently washed three times in PBS. As technical negative control,
2.9. RNA extraction and reverse transcription Extraction of total RNA from the cells was performed using RNeasy Mini Kit (Qiagen, Hilden, Germany, http://www.qiagen.com) according to the manufacturer’s instructions. DNase treatment was performed on-column using RNase-free DNase Kit (Qiagen). Reverse transcription was performed using 1 g of total RNA in a final volume of 20 l, using iScript First Strand Synthesis Kit (Bio-Rad Laboratories, Hercules, CA, http://www.bio-rad.com) and a Rotorgene 3000 (Corbett Research, Sydney, Australia, http://www.corbettresearch.com). Each RNA sample was reverse-transcribed in duplicate, and appropriate negative controls were included in each run. 2.10. Quantitative PCR
2.11. Metabolizing capacity of the cells hESC derived hepatocyte-like cells obtained by different differentiation protocols were tested for their ability to metabolize Phenacetin (Aldrich), Midazolam (Sigma) and Diclophenac (Sigma) via the phase I cytochrome P450 enzymes, CYP1A2, CYP3A4/7 and CYP2C9, respectively. In short, released metabolites of respective substances into the medium were measured by LC–MS. Cultures containing hepatocyte-like cells from cell lines SA002, SA002.5 and SA167 were tested at an age of 21–45 days. The substances were incubated as a cocktail in phenol red free medium; 26 M Phenacetin, 9 M Diclophenac and 3 M Midazolam for 6 h, 12 h and 24 h at 37 ◦ C and 5% CO2 . Samples from each culture and time point were collected and centrifuged for 5 min at a speed of 500 × g to get rid of any cell debris. 100 l of the cleared medium sam-
G. Brolén et al. / Journal of Biotechnology 145 (2010) 284–294
287
Fig. 2. Characterization of the cultures in phase I that has been induced to DE by Activin A: (A) phase-contrast image of an Activin A treated culture at day 5, (B) Sox17 (green) and Sox7 (red) immunofluorescence merged micrograph of an Activin A treated culture at day 4 and (C) gene expression of Sox17, HNF3, CXCR4, HNF4␣, ␣1 -antitrypsin and ␣FP analyzed by RT-PCR of growth factor induced and control cultures (intrinsically differentiated cultures without growth factors) at day 5. A Student’s t-test was performed testing the Activin A + FGF2 data or the Activin A vs the control data. One star represents a p-value ≤ 0.05 and two stars ≤0.01. N = 6. Scale bars are 250 m (A) and 50 m (B).
ple was transferred to a 96-well plate and 15 l of acetonitril was added to each well. The samples were frozen until measured and analyzed by LC–MS.
3. Results 3.1. Phase I—inducing undifferentiated hESC into definitive endoderm (DE) We used Activin A or a combination of Activin A and FGF2 to generate functional liver cells via DE in three hESC lines (Fig. 1). Genetic markers of DE induction were examined using quantitative PCR after 3–5 days of hESC colony differentiation in both control and growth factor induced conditions. Activin A treatment leads to morphological changes and distinct differentiation pattern of the colonies. The cells gathered together and formed a ring of epithelial cells around the colony. These homogenous growing cells expanded over the MEF layer coated on the plate. The morphological changes were associated with the expression of endoderm specific genes. Immunohistochemical analysis of Sox17/HNF3, Sox7/HNF3, HNF3/Oct4 and Sox17/Sox7 showed that after 4 days in Activin A (phase I) more than 70% of the remaining cells
were Sox17+ /Sox7− and only few were found to be Sox7+ or HNF3+ /Oct4+ (Fig. 2B and data not shown). Interestingly, the combination of Activin A and FGF2 yielded a higher expression of Sox17, HNF3 and CXCR4 compared to Activin A alone, whereas the control cultures, intrinsically differentiated hESC had significantly less Sox17, HNF3 and CXCR4 expression (Fig. 2). The control cultures revealed significantly less Sox17, HNF3 or CXCR4 expression while there were elevated levels of early liver markers like ␣FP or ␣1 -antitrypsin. At this early stage of development this indicates the presence of ExE in the control cultures (Fig. 2C). Together these findings demonstrated that all of our three analyzed cell lines are capable of differentiating towards DE.
3.2. Phase II—hepatic induction by BMP4 and FGF2 Growth factor combinations were stepwise added that are known to be important for in vivo development of an embryonic liver. At phase II (see Section 2) the cells were exposed to a combination of BMP2/4 and FGF1/2/4 to induce liver development. After some days a heterogeneous population of cells developed, among which many cell types were grouped together
288
G. Brolén et al. / Journal of Biotechnology 145 (2010) 284–294
Fig. 3. Morphology of hESC derived DE-Hep progenitors. The DE cultures strongly responded to the growth factors applied in phase II and the former homogenous epithelial cells changed morphology into polygonal shaped cells (white arrows in A): (A) phase-contrast image of a growth factor treated culture at day 17, (B) a corresponding region labeled with an EpCAM (green) antibody and in (C) co-stained for CK19 (red), (D) overlay of EpCAM and CK19 (white arrows in B, C and D show some of the double positive cells), (E) DE-Hep cells at day 14 are positive for CD54 and in (F) for CK7. Scale bars = 250 m.
in clusters, which grew homogenously (Fig. 3). This early cell population contained many areas of EpCAM expressing cells. Next to these cells many CK7, CK19 and some CD54 positive areas were detected indicating that an early liver cell type has developed (Fig. 3). The growth factor treated cultures expressed early sets of liver genes as ␣1 -antitrypsin, ␣FP, FABP1, HNF3, HNF4␣, CK8 and CK18 (Fig. 4). In the control cultures the expression of EpCAM, ␣1 -antitrypsin, ␣FP, FABP1, CK7/8/18/19 and CD54 expression was detected; however, the morphology of the cells differed considerably between the two culture protocols (data not shown). Quantification by Q-PCR showed significantly elevated levels of ␣1 -antitrypsin, ␣FP and FABP1 expression in the cytokine treated cultures compared to the intrinsically differentiated control (Fig. 4). Western blot showed as well upregulation of the ␣FP gene expression in the DE-Hep progenitor culture. Taken together, phase II cultures consisted mostly of cells expressing early liver markers whereas in the control cultures only few areas of such cells were found. 3.3. Phase III—hepatic maturation In phase III, the cultures were subsequently changed to hepatocyte differentiation medium containing EGF, HGF, insulin, transferrin, dexamethasone, hydrocortisone, ascorbic acid and
oncostatin M in the presence of serum (Fig. 1). At this stage many hepatocyte-like cells emerged and exhibited typical hepatocyte morphology with a polygonal shape, containing distinctive nuclei with a nucleoli. Interestingly, they were arranged in small islet-like clusters mid to late phase III and differed from the morphology of the hepatocyte-like cells derived by the intrinsic method that were located only at the periphery of the colonies (Fig. 5 and Söderdahl et al., 2007). All cytokine induced cell cultures had elevated urea secretion and the expression of CYP3A4, CYP1A2, CYP2C9, ␣1 -antitrypsin, ␣FP, HNF4␣, MRP2 and OATP2 was significantly upregulated compared to the intrinsically differentiated control (Fig. 6). Western blot analysis supported our findings by Q-PCR that CYP3A4, CYP1A2, CYP2C9, ␣1 -antitrypsin, ␣FP, HNF4␣, MRP2 and OATP2 were upregulated in DE-Hep cultures. Interestingly expression levels of CYP3A4 have been found to be dependent on the cell line (data not shown). The immunocytochemical analysis pointed out that many DE-Hep cells were positive for CYP1A, CYP3A, and CK18 (Fig. 7). Interestingly, the CYP3A signal intensity on single cell level was comparable to the expression in primary hepatocyte cultures (data not shown). Transport across hepatocyte plasma membranes is a key parameter in hepatic clearance and usually occurs through different carrier-mediated systems. Biliary elimination of anionic compounds is mediated by the multidrug resistance associated protein
G. Brolén et al. / Journal of Biotechnology 145 (2010) 284–294
289
Fig. 4. DE-Hep progenitor cultures in phase II were characterized by (A) immunocytochemistry (ICC), (B) RT-PCR and (C) Western blot. ICC results are summarized in a table (A). RT-PCR shows that DE-Hep progenitors express higher mRNA levels of ␣FP, ␣1 -antitrypsin and FABP1 than the control (intrinsic differentiated hESC) (B). Western blot analysis shows elevated expression of ␣FP compared to the control (C). All the data presented in this figure were generated at stage II day 15–17 at different passages and were compared to the corresponding control (intrinsic differentiated hESC). A Student’s t-test was performed testing the DE-Hep vs the control data. One star represents a p-value ≤ 0.05, two stars ≤0.01, and three stars ≤0.001. N ≥ 3.
2 (MRP2). At this stage also the expression of MRP2 could be detected in the DE-Hep cultures but not in the control cultures (Figs. 6 and 7). 3.4. Activity of DE-Hep cells In order to assess whether the DE-Hep are functional, they were tested for indocyanine green (ICG) uptake and secretion, ammonia metabolism, glycogen storage and metabolic capability of distinct drugs. The uptake of ICG by DE-Hep was analyzed and indeed showed that many DE-Hep in culture exhibited an ICG uptake (Fig. 7E). The uptake and excretion appeared only in a few cells of the control cultures, indicating that intrinsically differentiating cultures are also able to generate functional hepatocyte-like cells but significantly less in proportion compared to growth factor induced DE-Hep. Since human hepatocytes can make and store glycogen, we analyzed glycogen levels by periodic acid–Schiff staining (Fig. 6F and Figs. 6F and 7). Positive staining was found in both DE-Hep as well as in the intrinsically differentiated control cultures. Albumin, a serum protein produced only in the liver, is the major plasma protein that
circulates in the blood. The cells were evaluated for albumin protein expression. As a final degradation product of protein and amino acid metabolism ammonia is converted in the liver to urea and constitutes the most important catabolic pathway for eliminating excess nitrogen in the human body. We could indeed detect elevated levels of urea in the medium of the phase III cells compared to the intrinsically differentiated control cultures (Fig. 6). 3.5. Metabolism of Phenacetin, Midazolam and Diclophenac by the DE-Hep cells The DE-Hep obtained in phase III were tested for their ability to metabolize Phenacetin, Midazolam and Diclophenac via the phase I cytochrome P450 enzymes, CYP1A, CYP3A and CYP2C, respectively. Metabolites of respective substance released into the medium were determined by liquid chromatography–mass spectrometry (LC–MS). The control cultures and DE-Hep cultures were tested at an age of 18–45 days. The analysis of the metabolites in the samples was performed by LC–MS and revealed that both the DE-Hep and the
290
G. Brolén et al. / Journal of Biotechnology 145 (2010) 284–294
Fig. 5. DE-Hep and control cultures in phase III. Phase-contrast images of DE-Hep cultures at day 17 (A) and day 20 (B). Note, that the cells are starting to from sub-structures, are frequently bi-nucleated and exhibit a hepatocyte-like morphology (black arrows, B). (C) Phase-contrast image of hepatocyte-like cells at day 23 (white arrows), that were generated by intrinsic differentiation and used as controls. These cells are located only at the periphery of the growing colony and show different differentiation pattern and no clear sub-structures. (D and E) Human primary hepatocytes in culture. Scale bars: 50 m (A), 100 m (B), 40 m (C and D), and 100 m (E).
intrinsically differentiated control contained cells that were able to metabolize Phenacetin, Midazolam and Diclophenac. However, the DE-Hep cultures showed higher detectable signs of metabolism of Phenacetin than the intrinsically differentiated control (Fig. 8) and were comparable to the HepG2 control. In contrast to HepG2 could both our intrinsically differentiated control as well as our DE-Hep metabolize Diclophenac. However, the substrates are not specific for 1A2 and 3A4 and could reflect even the activity of other members of the CYP family. Therefore it is important to relate the activity data to the CYP immunostainings of the cultures (Fig. 6A). Activity of the control cultures could reflect therefore the presence of other members of the CYP family since there were found only few CYP1A and CYP3A positive cells. 4. Discussion In vitro the differentiation of human embryonic stem cells (hESC) to a specific germ layer or cell type has proven to be difficult, most likely because cells do not receive the same coordinated developmental cues/signals as they do in vivo. This has particularly affected the differentiation of hESC towards definitive endoderm (DE) and its derivatives, mostly due to the overlap of gene expression patterns between DE and extraembryonic endoderm and the following
lack of specific markers for early definitive endoderm (Pera et al., 2004). In this study, we demonstrate that hESC can be differentiated via DE to functional hepatocyte-like cells (DE-Hep), a terminally differentiated cell type of endoderm. We have used a three-stage protocol for differentiating hESC through a series of endodermal intermediates resembling those that occur during hepatic development in vivo. We characterized the cells at the three differentiation phases at morphological, RNA and protein level using quantitative PCR, immunocytochemistry, Western blot, ELISA and various functionality tests in vitro. Phase I involves the formation of DE. The importance of growth in the presence of Activin A to the differentiation process of hESC was demonstrated previously (D’Amour et al., 2005, 2006; Agarwal et al., 2008). Although methods for hepatic differentiation of hESC have been described, many of them did not exclude the extraembryonic differentiation of hESC (Rambhatla et al., 2003; Hay et al., 2007). This is of critical importance since large sets of genes are co-expressed between liver and yolk sac, such as ␣1-antitrypsin, ␣FP, albumin and even CYPs (Meehan et al., 1984; Pedersen et al., 1985). To overcome this problem, we induced hESC into DE by culturing them for several days in the presence of Activin A/FGF2. Our data in Fig. 2 confirmed that an early ␣FP and ␣1-antitrypsin
G. Brolén et al. / Journal of Biotechnology 145 (2010) 284–294
291
Fig. 6. DE-Hep cultures in phase III characterized by (A) immunocytochemistry (ICC) and ELISA, (B) RT-PCR, and (C) Western blot. The ICC and ELISA results are summarized in a table (A). ICC shows that DE-Hep cells have a marker expression comparable to immature hepatocytes. (B) RT-PCR shows that DE-Hep express much higher mRNA levels of all markers tested than the control culture, and much higher mRNA levels of CYP1A2, CYP3A4, CYP2C9, OATP2 mRNA levels than HepG2. (C) Western blot analysis shows expression of albumin, ␣FP, ␣1 -antitrypsin, CYP1A2, CYP3A4, CYP2C9, MRP2 and OATP2 for DE-Hep cultures in phase III day 29 that was barely or not found in the control. All data presented in this figure were generated at stage III day 24–45. The data were compared to the corresponding control (intrinsic differentiated hESC). Student’s test was performed testing the DE-Hep vs the control data. One star represents a p-value ≤ 0.05, two stars ≤0.01, and three stars ≤0.001. N ≥ 3.
expression, as we regarded as a sign of extraembryonic lineages, decreases significantly when cells have been treated with Activin A. On the other hand could we detect high expression of Sox17, CXCR4 and HNF3 in growth factor treated cultures whereas the intrinsically differentiated control cultures showed high levels of ␣1-antitrypsin and ␣FP (Fig. 2). This effect was even amplified by a combination of Activin A and FGF2. These findings are supported by previous studies with mouse embryonic stem cells where it was shown that interplay between FGFs and Activin A participates in ES cell differentiation to definitive endoderm (Funa et al., 2007). The next stage, phase II, involved the induction towards hepatic progenitor cell types. Combinations of BMPs and FGFs efficiently initiated hepatic differentiation from the growth factor treated cul-
tures as previously reported with mouse ES cell differentiation to hepatic derivates (Gouon-Evans et al., 2006). The combination BMP2/4 and FGF1/2/4 initiated efficiently hepatic differentiation in phase I cultures. At this stage numerous cells expressed early liver genes (Figs. 3 and 4). The DE-Hep cultures expressed significantly more ␣FP, ␣1-antitrypsin and FABP1 than the intrinsically differentiated cultures (Fig. 4). It is known that during liver development in vivo, cholangiocytes and hepatocytes are derived from the common progenitor called hepatoblasts. In our study, EpCAM as well as E-cadherin, was found at this early phase in many of the proliferating cells. The vast majority of EpCAM expressing cells co-expressed CK19 and many of the cells had a typical hepatic morphology indicating the early hepatic cell fate. Furthermore, next to EpCAM and CK19 express-
292
G. Brolén et al. / Journal of Biotechnology 145 (2010) 284–294
Fig. 7. Characterization of DE-Hep in phase III after 21 days in culture. Immunofluorescent labeling of CYP1A (green) (A), CYP3A (green) expression (B), the organic anion transporter MRP2 (red) (C), and CK18 (D). Indocyanine green (ICG) uptake after 1 h incubation (E) and glycogen storage (pink) (F). Note that some bi-nucleated cells store glycogen (black arrow). Scale bars: (A–H) 50 m.
ing regions many CD54 expressing cells were detected supporting a hepatic fate. In phase III the cells maturated in hepatocyte differentiation medium. This phase is characterized by cells of typical hepatocyte
Fig. 8. Functional drug metabolism test of DE-Hep cultures in phase III (day 29) measured by LC–MS. DE-Hep, control (intrinsic differentiation) and HepG2 cultures were tested for their ability to metabolize Phenacetin, Midazolam and Diclophenac via the phase I cytochrome P450 enzymes CYP1A, CYP3A, CYP2C, respectively. The CYP1A activity in DE-Hep cultures was significantly higher compared with control. HepG2 did not show any CYP2C activity. Student’s t-test was performed (DE-Hep vs control). One star represents a p-value ≤ 0.05, two stars ≤0.01, and three stars ≤0.001. N = 3.
morphology with many bi-nuclear cells. Most strikingly, many cells arranged in islands (Fig. 5) of CYP3A4/7 positive cells (Fig. 7). These sub-structures have never been found in the control cultures nor they have been mentioned in previously published studies (Cai et al., 2007; Hay et al., 2007, 2008; Duan et al., 2007; Agarwal et al., 2008). As in most of the published reports our cells also express large sets of liver associated genes and proteins (Figs. 6 and 7). MRP2 and CYPs are indicators of mature hepatocytes and play important functions in the liver. The majority of oxidative metabolic reactions are mediated by the CYP superfamily of enzymes (Sheweita, 2000). For example, CYP1A2, CYP3A4 and CYP2C9 are expressed in human liver microsomes (Stearns et al., 1995) and they are considered relatively hepatocyte specific (Venkatakrishnan et al., 2001), in particular CYP1A2 (McKinnon et al., 1991). The presence of CYP gene expression and especially their metabolic capacity highlights the maturation grade of the DE-Hep in phase III. So far only a few studies have shown constitutive activity of distinct CYPs in hESC derived hepatocyte-like cells (Basma et al., 2009). CYP3A is of special interest, since it plays a central role in the metabolism of aromatic amines, estrogen compounds, and certain drugs (Ingelman-Sundberg, 2004). Our results demonstrate that our guided differentiation of hESC generates a cell population that express liver-specific functions and can be found even in human primary hepatocytes. Additionally, in phase III DE-Hep accumulate glycogen as shown by glycogen staining, take up ICG and secrete urea (Fig. 6). Furthermore we show functionality in drug metabolism in the DE-Hep cultures (Fig. 8). However, we could not
G. Brolén et al. / Journal of Biotechnology 145 (2010) 284–294
see a significant difference between our intrinsically differentiated cells and DE-Hep (Fig. 8). One explanation could be the presence of more intestinal and adipocyte cells in the intrinsically differentiated cells which is supported by the gene array (Synnergren et al., 2009) and by the expression of CYP3A in the gut epithelium (Kolars et al., 1994). Together these findings indicate that DE-Hep have the transcriptional machinery for producing many of the functional proteins present in the liver. Under preparation of this paper an alternative strategy has been published using hepatic induction without biologically relevant factors like BMPs and FGFs. Hay et al. demonstrated recently that after DE induction synthetic low molecular weight compounds such as DMSO initiate liver development. To find out whether this non-biological way of steering cell differentiation, or the strategy using naturally occurring growth factors used in this study, gives rise to the most mature hepatocyte-like cells has to be investigated in comparative studies in the future. In conclusion, gene- and protein expression as well as the functional data clearly demonstrated that differentiation of hESC with this growth factor guided protocol generated hepatocyte-like cells with many liver-like properties. We could show that hESC derived hepatocyte-like cells can organize and form sub-structures in vitro, but most importantly, we demonstrate the presence and functionality of CYP1A, CYP3A and CYP2C, which are shown to metabolize pharmaceutical compounds specific for these three key-enzymes. Our data suggest that the DE-Hep in general exhibit not only higher protein- and gene expression levels than the control cultures, but also express much higher amounts of drug metabolizing enzymes. This indicates a great potential of these cells to be further developed by suitable protocols into hepatocyte-like cells with phenotype mature enough for future use in studies of drug metabolism and drug-induced hepatotoxicity. Acknowledgments We acknowledge Karin Noaksson, Jörg Benecke and Fredrik Wessberg for the hESC propagation; Gunilla Caisander for the primary hepatocyte culture and ELISA analysis. The work was supported by the 6th EU framework program (LSHB-CT-2006-018940) Vitrocellomics, the 6th EU framework program (LSHB-CT-2005-512145) BetaCelltherapy and the Carcinogenomics framework (CARCINO; LSHB-CT-2006-037712). References Agarwal, S., Holton, K.L., Lanza, R., 2008. Efficient differentiation of functional hepatocytes from human embryonic stem cells. Stem Cells 26 (5), 1117–1127. Basma, H., Soto-Gutiérrez, A., Yannam, G.R., Liu, L., Ito, R., Yamamoto, T., Ellis, E., Carson, S.D., Sato, S., Chen, Y., Muirhead, D., Navarro-Alvarez, N., Wong, R.J., Roy-Chowdhury, J., Platt, J.L., Mercer, D.F., Miller, J.D., Strom, S.C., Kobayashi, N., Fox, I.J., 2009. Differentiation and transplantation of human embryonic stem cell-derived hepatocytes. Gastroenterology 136 (3), 990–999. Belo, J.A., Bouwmeester, T., Leyns, L., Kertesz, N., Gallo, M., Follettie, M., De Robertis, E.M., 1997. Cerberus-like is a secreted factor with neutralizing activity expressed in the anterior primitive endoderm of the mouse gastrula. Mechanism of Development 68, 45–57. Bradford, M.M., 1976. A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein–dye binding. Analytical Biochemistry 72, 248–254. Butura, A., Johansson, I., Nilsson, K., Wärngård, L., Ingelman-Sundberg, M., SchuppeKoistinen, I., 2004. Differentiation of human hepatoma cells during confluence as revealed by gene expression profiling. Biochemistry and Pharmacology 67 (7), 1249–1258. Cai, J., Zhao, Y., Liu, Y., Ye, F., Song, Z., Qin, H., Meng, S., Chen, Y., Zhou, R., Song, X., Guo, Y., Ding, M., Deng, H., 2007. Directed differentiation of human embryonic stem cells into functional hepatic cells. Hepatology 45 (5), 1229–1239. D’Amour, K.A., Agulnick, A.D., Eliazer, S., Kelly, O.G., Kroon, E., Baetge, E.E., 2005. Efficient differentiation of human embryonic stem cells to definitive endoderm. Nature Biotechnology 23 (12), 1534–1541. D’Amour, K.A., Bang, A.G., Eliazer, S., Kelly, O.G., Agulnick, A.D., Smart, N.G., Moorman, M.A., Kroon, E., Carpenter, M.K., Baetge, E.E., 2006. Production of pancreatic hormone-expressing endocrine cells from human embryonic stem cells. Nature Biotechnology 24 (11), 1392–1401.
293
Duan, Y., Catana, A., Meng, Y., Yamamoto, N., He, S., Gupta, S., Gambhir, S.S., Zern, M.A., 2007. Differentiation and enrichment of hepatocyte-like cells from human embryonic stem cells in vitro and in vivo. Stem Cells 12, 3058–3068. Ek, M., Söderdahl, T., Küppers-Munther, B., Edsbagge, J., Andersson, T.B., Björquist, P., Cotgreave, I., Jernström, B., Ingelman-Sundberg, M., Johansson, I., 2007. Expression of drug metabolizing enzymes in hepatocyte-like cells derived from human embryonic stem cells. Biochemistry and Pharmacology 74 (3), 496– 503. Funa, N.S., Saldeen, J., Akerblom, B., Welsh, M., 2007. Interdependent fibroblast growth factor and activin A signaling promote the expression of endodermal genes in differentiating mouse embryonic stem cells expressing Src Homology 2-domain inactive Shb. Differentiation 22, 367–376. Gouon-Evans, V., Boussemart, L., Gadue, P., Nierhoff, D., Koehler, C.I., Kubo, A., Shafritz, D.A., Keller, G., 2006. BMP-4 is required for hepatic specification of mouse embryonic stem cell-derived definitive endoderm. Nature Biotechnology 24 (11), 1402–1411. Hamazaki, T., Iiboshi, Y., Oka, M., Papst, P.J., Meacham, A.M., Zon, L.I., Terada, N., 2001. Hepatic maturation in differentiating embryonic stem cells in vitro. FEBS Letters 497 (1), 15–19. Hamazaki, T., Terada, N., 2003. In vitro differentiation of embryonic stem cells into hepatocytes. Methods in Enzymology 365, 277–287. Hay, D.C., Zhao, D., Ross, A., Mandalam, R., Lebkowski, J., Cui, W., 2007. Direct differentiation of human embryonic stem cells to hepatocyte-like cells exhibiting functional activities. Cloning Stem Cells 9 (1), 51–62. Hay, D.C., Zhao, D., Fletcher, J., Hewitt, Z.A., McLean, D., Urruticoechea-Uriguen, A., Black, J.R., Elcombe, C., Ross, J.A., Wolf, R., Cui, W., 2008. Efficient differentiation of hepatocytes from human embryonic stem cells exhibiting markers recapitulating liver development in vivo. Stem Cells, Epub ahead of print. Heins, N., Englund, M.C., Sjöblom, C., Dahl, U., Tonning, A., Bergh, C., Lindahl, A., Hanson, C., Semb, H., 2004. Derivation, characterization, and differentiation of human embryonic stem cells. Stem Cells 22 (3), 367–376. Heins, N., Lindahl, A., Karlsson, U., Rehnström, M., Caisander, G., Emanuelsson, K., Hanson, C., Semb, H., Björquist, P., Sartipy, P., Hyllner, J., 2006. Clonal derivation and characterization of human embryonic stem cell lines. Journal of Biotechnology 122 (4), 511–520. Hogan, B., Beddington, R., Costantini, F., Lacy, E., 1994. Manipulating the Mouse Embryo: A Laboratory Manual. Cold Spring Harbor Laboratory Press, Plainview, NY, pp. 260–261. Ingelman-Sundberg, M., 2004. Human drug metabolising cytochrome P450 enzymes: properties and polymorphisms. Naunyn Schmiedebergs Archives of Pharmacology 369 (1), 89–104. Kanai-Azuma, M., Kanai, Y., Gad, J.M., Tajima, Y., Taya, C., Kurohmaru, M., Sanai, Y., Yonekawa, H., Yazaki, K., Tam, P.P., Hayashi, Y., 2002. Depletion of definitive gut endoderm in Sox17-null mutant mice. Development 129 (10), 2367– 2379. Kolars, J.C., Lown, K.S., Schmiedlin-Ren, P., Ghosh, M., Fang, C., Wrighton, S.A., Merion, R.M., Watkins, P.B., 1994. CYP3A expression in human gut epithelium. Pharmacogenetics 4 (5), 247–259. Lawson, K.A., Meneses, J.J., Pedersen, R.A., 1986. Cell fate and cell lineage in the endoderm of the presomite mouse embryo, studied with an intracellular tracer. Developmental Biology 115 (2), 325–339. Lawson, K.A., Pedersen, R.A., 1987. Cell fate, morphogenetic movement and population kinetics of embryonic endoderm at the time of germ layer formation in the mouse. Development 101 (3), 627–652. Lawson, K.A., Meneses, J.J., Pedersen, R.A., 1991. Clonal analysis of epiblast fate during germ layer formation in the mouse embryo. Development 113 (3), 891– 911. McKinnon, R.A., Hall, P.D., Quattrochi, L.C., Tukey, R.H., McManus, M.E., 1991. Localization of CYP1A1 and CYP1A2 messenger RNA in normal human liver and in hepatocellular carcinoma by in situ hybridization. Hepatology 14 (5), 848– 856. Meehan, R.R., Barlow, D.P., Hill, R.E., Hogan, B.L., Hastie, N.D., 1984. Pattern of serum protein gene expression in mouse visceral yolk sac and fetal liver. EMBO Journal 3 (8), 1881–1885. Pedersen, R.A., Meneses, J., Spindle, A., Wu, K., Galloway, S.M., 1985. Cytochrome P-450 metabolic activity in embryonic and extraembryonic tissue lineages of mouse embryos. Proceedings of the National Academy of Sciences 82 (10), 3311–3315. Pera, M.F., Andrade, J., Houssami, S., Reubinoff, B., Trounson, A., Stanley, E.G., Wardvan Oostwaard, D., Mummery, C., 2004. Regulation of human embryonic stem cell differentiation by BMP-2 and its antagonist noggin. Journal of Cell Science 117, 1269–1280. Rambhatla, L., Chiu, C.P., Kundu, P., Peng, Y., Carpenter, M.K., 2003. Generation of hepatocyte-like cells from human embryonic stem cells. Cell Transplantation 12 (1), 1–11. Sheweita, S.A., 2000. Drug-metabolizing enzymes: mechanisms and functions. Current Drug Metabolism 1 (2), 107–132. Sasaki, H., Hogan, B.L., 1993. Differential expression of multiple fork head related genes during gastrulation and axial pattern formation in the mouse embryo. Development 118, 47–59. Söderdahl, T., Küppers-Munther, B., Heins, N., Edsbagge, J., Björquist, P., Cotgreave, I., Jernström, B., 2007. Glutathione transferases in hepatocyte-like cells derived from human embryonic stem cells. Toxicology In Vitro 21 (5), 929– 937. Stearns, R.A., Chakravarty, P.K., Chen, R., Chiu, S.H., 1995. Biotransformation of losartan to its active carboxylic acid metabolite in human liver microsomes. Role of
294
G. Brolén et al. / Journal of Biotechnology 145 (2010) 284–294
cytochrome P4502C and 3A subfamily members. Drug Metabolism and Disposition 23 (2), 207–215. Synnergren, J., Heins, N., Brolén, G., Eriksson, G., Lindahl, A., Hyllner, J., Olsson, B., Sartipy, P., Björquist, P., 2009. Transcriptional profiling of human embryonic stem cells differentiating to definitive and primitive endoderm and further towards the hepatic lineage. Stem Cells and Development, Epub ahead of print.
Tremblay, K.D., Zaret, K.S., 2005. Distinct populations of endoderm cells converge to generate the embryonic liver bud and ventral foregut tissues. Developmental Biology 280 (1), 87–99. Venkatakrishnan, K., Von Moltke, L.L., Greenblatt, D.J., 2001. Human drug metabolism and the cytochromes P450: application and relevance of in vitro models. Journal of Clinical Pharmacology 41 (11), 1149–1179.
Contents lists available at ScienceDirect
Journal of Biotechnology journal homepage: www.elsevier.com/locate/jbiotec
Hepatocyte-like cells derived from human embryonic stem cells specifically via definitive endoderm and a progenitor stage Gabriella Brolén a , Louise Sivertsson c , Petter Björquist a , Gustav Eriksson a , Monica Ek c , Henrik Semb d , Inger Johansson c , Tommy B. Andersson b , Magnus Ingelman-Sundberg c , Nico Heins a,∗ a
Cellartis AB, Arvid Wallgrens Backe 20, 41346 Göteborg, Sweden Development DMPK & Bioanalysis, AstraZeneca R&D Mölndal, 431 83 Mölndal, Sweden Karolinska Institute, Section of Pharmacogenetics, Department of Physiology and Pharmacology, Nanna Svartz väg 2, 17177 Stockholm, Sweden d Stem Cell Center, Lund University, BMC B10, Lund 22184, Sweden b c
a r t i c l e
i n f o
Article history: Received 24 March 2009 Received in revised form 9 November 2009 Accepted 12 November 2009
Keywords: Human embryonic stem cell Definitive endoderm Hepatocyte Liver progenitor Activin
a b s t r a c t Human embryonic stem cells offer a potential unlimited supply for functional hepatocytes, since they can differentiate into hepatocyte-like cells displaying a characteristic hepatic morphology and expressing various hepatic markers. These cells could be used in various applications such as studies of drug metabolism and hepatotoxicity, which however, would require a significant expression of drug metabolizing enzymes. To derive these cells we use a stepwise differentiation protocol where growth- and maturation factors are added. The first phase involves the formation of definitive endoderm. Next, these cells are treated with factors known to promote the induction and proliferation towards hepatic progenitor cell types. In the last phase the cells are terminally differentiated and maturated into functional hepatocyte-like cells. The cultures were characterized by analysis of endodermal or hepatic markers and compared to cultures derived without induction via definitive endoderm. Hepatic functions such as urea secretion, glycogen storage, indocyanine green uptake and secretion, and cytochrome P450-expression and activity were evaluated. The DE-Hep showed a hepatocyte morphology with sub-organized cells and exhibited many liverfunctions including transporter activity and capacity to metabolize drugs specific for important cytochrome P450 sub-families. This represents an important step in differentiation of hESC into functional hepatocytes. © 2009 Elsevier B.V. All rights reserved.
1. Introduction Pluripotent human stem cells are expected to change the accessibility to a variety of human cell types. The possibility to propagate pluripotent human embryonic stem cells (hESC) and subsequently differentiate them into the desired target cell types will provide a stable and virtually unlimited supply of cells for a range of applications in vivo and in vitro. The liver is the major organ for metabolism and detoxification in the human body, and therefore huge efforts have been undertaken in order to identify a reliable source of functional hepatocyte-like cell types for in vitro testing. Up to now the complexity and function of the liver are not reflected by any available in vitro cell type or cellular system. The availability of human primary liver cells is limited and furthermore these cells lose their normal phenotype and functional properties (typically within 24 h) when used in
∗ Corresponding author. Fax: +46 31 758 09 10. E-mail address: [email protected] (N. Heins). 0168-1656/$ – see front matter © 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.jbiotec.2009.11.007
vitro. Transformed hepatic cell lines contain very low levels of drug metabolizing enzymes and have expression profiles of other important proteins substantially different from the native hepatocyte in vivo. Thus, many tests e.g. drug metabolism and toxicity are still performed using animal material, even though liver metabolism is known to be species-specific. Due to this, difficulties to predict liver toxicity in another species than the one tested remain present. Therefore early prediction of human liver toxicity liabilities is of principal importance when selecting compounds to enter clinical trials. Accordingly there is an urgent need for a model system that mimics human liver cells and that is able to predict effects of candidate molecules in the development of new drugs or chemicals. Regarding both availability and physiological relevance human pluripotent stem cells may serve as an ideal renewable source of functional human hepatocytes. When hESC have been placed in a proper environment, certain hepatic characteristics have been observed 2–4 weeks after differentiation. The present investigation is based on the fact that definitive endoderm (DE) cells give rise to endodermal organs and consequently to e.g. hepatic cell types. Early endoderm development is not well understood. Fate
G. Brolén et al. / Journal of Biotechnology 145 (2010) 284–294
Fig. 1. Schematic overview of the differentiation strategy into definitive endoderm derived hepatocytes (DE-Hep). The differentiation protocol is divided into three main phases. The starting material consists of undifferentiated hESC cultured on MEF. In phase I the hESC are induced into definitive endoderm (DE). In phase II the DE is induced to early liver endoderm or liver progenitor cells. During phase III the DE-Hep cultures are maturated.
mapping studies of cultured mouse embryos (Lawson et al., 1986, 1991; Lawson and Pedersen, 1987) have revealed that DE begins to form at the embryonic days 6–6.5 (E6–6.5) and that by the end of gastrulation (E7.5), some labeled cells only give rise to endodermal derivatives. Other fate mapping studies (Lawson et al., 1991; Tremblay and Zaret, 2005) suggested that the first endodermal cells that migrate through the primitive streak (PS) at E6.5 are fated to become liver, ventral pancreas, lungs and stomach. A complication in the study of endoderm is that mammals generate extraembryonic endoderm (ExE). ExE arises at the blastocyst stage and eventually forms two subpopulations: visceral endoderm and parietal endoderm. These cells share the expression of many genes with DE (cells that give rise to the endodermal organs), including the transcription factors Sox17 (Kanai-Azuma et al., 2002), HNF3 and HNF4␣ (Belo et al., 1997; Sasaki and Hogan, 1993). D’Amour et al. (2005, 2006) have developed a protocol for deriving DE from hESC. Previous studies have identified cells with some hepatocyte-like characteristics, e.g. cytochrome P450 (CYP) activity and ability to store glycogen, in differentiated hESC cultures (Rambhatla et al., 2003; Cai et al., 2007; Duan et al., 2007; Agarwal et al., 2008). So far the cells generated have not shown the metabolic qualities necessary for potentially replacing traditional liver systems in terms of drug transporter expression and specific CYP expression patterns needed for industrial applications. In this investigation several cell lines have been induced by growth factors to generate hepatocyte-like cells via definitive endoderm (DE-Hep). The cells were subsequently cultured under conditions known to support the development of liver cell lineages (Hamazaki et al., 2001; Hamazaki and Terada, 2003). Our strategy involves three distinct important steps in liver development. The study shows that the hepatocyte-like cell population obtained via DE has many similarities with human adult hepatocytes, including capacity to metabolize several specific pharmaceutical compounds.
285
ation into DE recombinant human Activin A (100 ng/ml, R&D) was supplemented to the medium, the following days of DE differentiation (phase I) Activin A (100 ng/ml) and sometimes FGF2 (4 ng/ml) or Wnt3a (50 ng/ml) was added to the medium. Phase II was carried out either in RPMI or DMEM (both from Invitrogen/Gibco). Different growth factors were added to the media. In phase II, RPMI advanced medium supplemented with PEST (1%) (Invitrogen/Gibco) and Glutamax (1%) (Invitrogen–Sigma) was used as a basic medium. BMP4 (100 ng/ml, PromoKine) and FGF2 (4 ng/ml) were added to the medium to differentiate the DE into hepatic progenitors. In some experiments in phase II the medium was supplemented with FGF1 (100 ng/ml, PromoKine), FGF2 (5 ng/ml, R&D), BMP2 (50 ng/ml, PromoKine), BMP4 (200 ng/ml) and 0.2% FBS (Gibco), with medium change every second day. At phase III hepatocyte maturation was carried out in HCM (Cambrex) with dexamethasone (Dex) (0.1 M, Calbiochem), oncostatin M (OSM) (10 ng/ml, PromoKine), HGF (20 ng/ml, PromoKine) and SingleQuots (Lonza). The medium was changed every second day. 2.3. Culture of controls and tissue material As controls we used intrinsic differentiated hESC as well as mouse embryonic feeder cells (MEF), HepG2 cells, human primary hepatocytes, and human liver samples. Intrinsic differentiated hepatocyte-like cells that we use as control in this study were derived according to Söderdahl et al. (2007) and Ek et al. (2007). hESC were allowed to differentiate for 18–45 days in VitroHESTM supplemented with 4 ng/ml FGF2. MEFs were derived and cultured as previously described (Hogan et al., 1994). HepG2 cells (HB-8065, American Type Culture Collection) were cultured as previously described (Butura et al., 2004). HepG2 cells (HB-8065, American Type Culture Collection, Manassas, VA) were cultured in minimum essential medium (MEM) supplemented with 10% FBS, 1% penicillin–streptomycin, 1% sodium pyruvate and 1% non-essential amino acids (all from Invitrogen) and collected by trypsinization when 80–90% confluence were reached. Platable cryopreserved human primary hepatocytes (In Vitro Technologies) were cultured according to the manufacturer instructions. In short, hepatocytes were thawed at 37 ◦ C and propagated at 0.3 million cells per ml and well in InVitroGRO CP medium (In Vitro Technologies) at collagen I-coated 24-well plates. After 24–48 h in vitro the hepatocytes were fixed and subjected to immunofluorescence stainings. After 24–120 h in vitro the hepatocytes were harvested for activity assays and mRNA. All cultures were kept at 37 ◦ C, 5% CO2 and 90–95% humidity. Human liver samples were obtained from Sahlgrenska Hospital (Göteborg, Sweden) and originated from patients undergoing liver resection. All tissues were obtained through qualified medical staff, with donor consent and with the approval of the Local Ethics Committee at Sahlgrenska Hospital.
2. Methods
2.4. Immunocytochemical methods and antibodies
2.1. Culture of hESC
Cells in culture were fixed in 4% (w/v) paraformaldehyde (PFA) for 15 min, and washed in PBS. For visualizing extracellular proteins cells were blocked for 30 min in 5% FBS in PBS, for intracellular protein staining cells were blocked and permeabilized for 30 min in 5% FBS in 0.1% TritonX-100 in PBS. The primary antibodies were incubated in 1% FBS in PBS overnight at 4 ◦ C and the secondary antibodies were diluted in 1% FBS in PBS for 1 h at room temperature (RT). For CYP-stainings the cells were incubated with secondary antibodies in PBS for 3 h at RT. All washes were performed in PBS. To visualize the nucleus cells were incubated with DAPI at 0.05 mg/ml for 5 min at RT and mounted in DAKO Cytomation mounting medium. Primary antibodies used were rabbit anti-albumin (1:500, DAKO Cytomation), rabbit anti-␣1 antitrypsin (1:200, DAKO Cytomation), mouse IgG2a ␣FP (1:500,
The hESC lines SA001, SA002, SA002.5 and SA167 (Cellartis AB, Göteborg, Sweden, http://www.cellartis.com) were derived, cultured and characterized as previously described (Heins et al., 2004, 2006). 2.2. Culture of DE-Hep Differentiation of hESC into DE-Hep cells was carried out as described in Fig. 1. Before initiating differentiation, hESC were washed in PBS+/+ (Invitrogen/Gibco). During differentiation into DE (phase I) the fetal calf serum concentration was 0% the first day and 0.2% for the following days. During the first day of differenti-
286
G. Brolén et al. / Journal of Biotechnology 145 (2010) 284–294
Sigma), rabbit anti-CK7 (1:200, Novocastra), mouse IgG1 anti-CK8 (1:200, Santa Cruz), mouse anti-CK18 (1:200, DAKO Cytomation), mouse IgG1 CK19 (1:200, Novocastra), mouse IgG2b anti-CXCR4 (1:250, R&D Systems), rabbit anti-CYP1A2 (1:100, Biomol), sheep anti-CYP3A4/7 (1:100, Biomol), mouse IgG1 anti-EpCAM (1:50, GeneTex), goat anti-LFABP (1:500, Santa Cruz), goat anti-HNF3 (1:400, Santa Cruz Biotechnology), rabbit anti-HNF1␣ (1:400, Santa Cruz), rabbit anti-HNF4␣ (1:300, AH Diagnostics), mouse antiICAM-1 (CD54) (1:500, BD Pharmingen), rabbit anti-MRP2 (1:50, Santa Cruz), goat anti-Sox7 (R&D), rabbit anti-Sox17 (1:2000, a kind gift from E. Baetge). Secondary antibodies used were Alexa Fluor 488 conjugated donkey anti-goat IgG (1:500, Molecular Probes), Alexa Fluor 488 donkey anti-rabbit IgG (1:1000, Molecular Probes), Alexa Fluor 488 donkey anti-sheep IgG (1:1000, Molecular Probes), Alexa Fluor 594 donkey anti-rabbit IgG (1:1000, Molecular Probes), Cy2 donkey anti-mouse IgG (1:100, Jackson Immuno Research), Cy2 donkey anti-rat IgG (1:100, Jackson Immuno Research), Cy3 donkey anti-mouse IgG (1:1000, Jackson Immuno Research), Cy3 donkey anti-rat IgG (1:500, Jackson Immuno Research), fluorescein isothiocyanate (FITC) anti-mouse IgG (1:500, Jackson Immuno Research). Stained samples were examined using a Nikon Eclipse TE2000-U Fluorescence microscope and Nikon Act-1C for DXM1 200C software.
a culture was treated with human saliva for 20 min at RT and subsequently washed in PBS. The human saliva contains ␣-amylase which digests glycogen. Periodic acid was added to the treated and untreated cultures for 5 min at RT followed by repeated washing in PBS. Subsequently, cultures were incubated in Schiff’s reagent for 15 min at RT. After washing in PBS cells were rinsed in H2 O prior to mounting in mounting media.
2.5. Western blot
Gene-specific primer pairs were designed and evaluated for an annealing temperature of 60 ◦ C using freely available Web-based software (Primer3, Netprimer, Beacon Designer 2.1, mFold, and Oligonucleotide Properties Calculator). Primers were designed for the following genes: Sox17, HNF3, CXCR4, ␣FP and ␣1 -antitrypsin. The optimized assays, including reference material, are available from TATAA Biocenter (Göteborg, Sweden, http://www.tataa.com). All Q-PCRs in Fig. 2 were performed with SYBR Green I chemistry in a Rotorgene 3000. The authenticity of the PCR products was verified by melt-curve analysis and agarose-gel electrophoresis. For Q-PCRs, 1× Jump Start Buffer 10× (Sigma–Aldrich), 3 mM MgCl2 (Sigma–Aldrich), 0.3 mM dNTP mix (Sigma–Aldrich), 0.4× SYBR Green (Molecular Probes, Inc., Eugene, OR, http://probes.invitrogen.com), 0.4 M forward primer (MWG Biotech, Ebersberg, Germany, http://www.mwgbiotech.com), 0.4 M reverse primer (MWG Biotech), 0.04 U/l Jump Start taq polymerase (Sigma–Aldrich), and 2 l cDNA template were used in a final volume of 20 l. After an initial denaturation/activation step of 3 min at 95 ◦ C followed 45 cycles of 20 s at 95 ◦ C, 20 s at 60 ◦ C, and 20 s at 72 ◦ C. The detection of fluorescent signal was performed at 72 ◦ C in each cycle. Ct (threshold cycle) values were calculated using the Rotorgene software. For some genes (albumin, ␣1 -antitrypsin, ␣FP, FABP1, HNF3, HNF4␣, CYP3A4, CYP1A2, CYP2C9, OATP2, MRP2, CDX2 and Vilin) TaqMan-based Q-PCR was used, as previously described by Ek et al., 2007.
DE-Hep progenitor cells, DE-Hep cells and its control cells (intrinsically differentiated hESC, HepG2 and human primary hepatocytes) were lysed in TritonX buffer (1% TritonX-100, 1 mM EDTA and 1:100 protease inhibitor cocktail (Roche, Germany)) and centrifuged at 800 × g and 4 ◦ C for 10 min. Protein determination was measured according to the method of Bradford (1976). The following primary antibodies were used: human albumin (Bethyl Laboratories Inc., TX, USA), ␣1 -antitrypsin (Dako Cytomation, Glostrup, Denmark), ␣-fetoprotein (Cell Signaling Technology Inc., MA, USA), CYP3A (BD Biosciences, CA, USA), MRP2, OATP2 (Santa Cruz Biotechnology, CA, USA). Horseradish peroxidase-conjugated secondary antibodies (Dako Cytomation, Glostrup, Denmark) were visualized with Pierce supersignal chemiluminescense substrate (Pierce, IL, USA) using the LAS-1000 + gel documentation system, and analyzed using Image Gauge software (Fujifilm, CT, USA). 2.6. Indocyanine green (ICG) Cardiogreen (Sigma, l2633) was dissolved in 5 ml of solvent in a sterile vial and then added to 20 ml of DMEM containing 10% FBS. The final concentration of the resulting indocyanine green (ICG) solution was 1 mg/ml. The ICG solution was added to the cell culture dish and incubated at 37 ◦ C for 60 min. After the cells were rinsed three times with PBS, the cellular uptake of ICG was examined with a stereomicroscope. After the analysis the dish was refilled with DMEM containing 10% FBS. ICG was eliminated from the cells overnight. 2.7. ELISA analysis of urea in the medium For urea excretion the Roche UREA/BUN assay was used. Urea secretion was analyzed using a kit for kinetic UV assay for urea/urea nitrogen (Roche/Hitachi) by enzyme-linked immunosorbent assay (ELISA) at Klinisk Kemi, C-lab, Sahlgrenska University Hospital, Göteborg. 2.8. Glycogen storage detection Cells were fixed in 100% methanol for 15 min at RT and subsequently washed three times in PBS. As technical negative control,
2.9. RNA extraction and reverse transcription Extraction of total RNA from the cells was performed using RNeasy Mini Kit (Qiagen, Hilden, Germany, http://www.qiagen.com) according to the manufacturer’s instructions. DNase treatment was performed on-column using RNase-free DNase Kit (Qiagen). Reverse transcription was performed using 1 g of total RNA in a final volume of 20 l, using iScript First Strand Synthesis Kit (Bio-Rad Laboratories, Hercules, CA, http://www.bio-rad.com) and a Rotorgene 3000 (Corbett Research, Sydney, Australia, http://www.corbettresearch.com). Each RNA sample was reverse-transcribed in duplicate, and appropriate negative controls were included in each run. 2.10. Quantitative PCR
2.11. Metabolizing capacity of the cells hESC derived hepatocyte-like cells obtained by different differentiation protocols were tested for their ability to metabolize Phenacetin (Aldrich), Midazolam (Sigma) and Diclophenac (Sigma) via the phase I cytochrome P450 enzymes, CYP1A2, CYP3A4/7 and CYP2C9, respectively. In short, released metabolites of respective substances into the medium were measured by LC–MS. Cultures containing hepatocyte-like cells from cell lines SA002, SA002.5 and SA167 were tested at an age of 21–45 days. The substances were incubated as a cocktail in phenol red free medium; 26 M Phenacetin, 9 M Diclophenac and 3 M Midazolam for 6 h, 12 h and 24 h at 37 ◦ C and 5% CO2 . Samples from each culture and time point were collected and centrifuged for 5 min at a speed of 500 × g to get rid of any cell debris. 100 l of the cleared medium sam-
G. Brolén et al. / Journal of Biotechnology 145 (2010) 284–294
287
Fig. 2. Characterization of the cultures in phase I that has been induced to DE by Activin A: (A) phase-contrast image of an Activin A treated culture at day 5, (B) Sox17 (green) and Sox7 (red) immunofluorescence merged micrograph of an Activin A treated culture at day 4 and (C) gene expression of Sox17, HNF3, CXCR4, HNF4␣, ␣1 -antitrypsin and ␣FP analyzed by RT-PCR of growth factor induced and control cultures (intrinsically differentiated cultures without growth factors) at day 5. A Student’s t-test was performed testing the Activin A + FGF2 data or the Activin A vs the control data. One star represents a p-value ≤ 0.05 and two stars ≤0.01. N = 6. Scale bars are 250 m (A) and 50 m (B).
ple was transferred to a 96-well plate and 15 l of acetonitril was added to each well. The samples were frozen until measured and analyzed by LC–MS.
3. Results 3.1. Phase I—inducing undifferentiated hESC into definitive endoderm (DE) We used Activin A or a combination of Activin A and FGF2 to generate functional liver cells via DE in three hESC lines (Fig. 1). Genetic markers of DE induction were examined using quantitative PCR after 3–5 days of hESC colony differentiation in both control and growth factor induced conditions. Activin A treatment leads to morphological changes and distinct differentiation pattern of the colonies. The cells gathered together and formed a ring of epithelial cells around the colony. These homogenous growing cells expanded over the MEF layer coated on the plate. The morphological changes were associated with the expression of endoderm specific genes. Immunohistochemical analysis of Sox17/HNF3, Sox7/HNF3, HNF3/Oct4 and Sox17/Sox7 showed that after 4 days in Activin A (phase I) more than 70% of the remaining cells
were Sox17+ /Sox7− and only few were found to be Sox7+ or HNF3+ /Oct4+ (Fig. 2B and data not shown). Interestingly, the combination of Activin A and FGF2 yielded a higher expression of Sox17, HNF3 and CXCR4 compared to Activin A alone, whereas the control cultures, intrinsically differentiated hESC had significantly less Sox17, HNF3 and CXCR4 expression (Fig. 2). The control cultures revealed significantly less Sox17, HNF3 or CXCR4 expression while there were elevated levels of early liver markers like ␣FP or ␣1 -antitrypsin. At this early stage of development this indicates the presence of ExE in the control cultures (Fig. 2C). Together these findings demonstrated that all of our three analyzed cell lines are capable of differentiating towards DE.
3.2. Phase II—hepatic induction by BMP4 and FGF2 Growth factor combinations were stepwise added that are known to be important for in vivo development of an embryonic liver. At phase II (see Section 2) the cells were exposed to a combination of BMP2/4 and FGF1/2/4 to induce liver development. After some days a heterogeneous population of cells developed, among which many cell types were grouped together
288
G. Brolén et al. / Journal of Biotechnology 145 (2010) 284–294
Fig. 3. Morphology of hESC derived DE-Hep progenitors. The DE cultures strongly responded to the growth factors applied in phase II and the former homogenous epithelial cells changed morphology into polygonal shaped cells (white arrows in A): (A) phase-contrast image of a growth factor treated culture at day 17, (B) a corresponding region labeled with an EpCAM (green) antibody and in (C) co-stained for CK19 (red), (D) overlay of EpCAM and CK19 (white arrows in B, C and D show some of the double positive cells), (E) DE-Hep cells at day 14 are positive for CD54 and in (F) for CK7. Scale bars = 250 m.
in clusters, which grew homogenously (Fig. 3). This early cell population contained many areas of EpCAM expressing cells. Next to these cells many CK7, CK19 and some CD54 positive areas were detected indicating that an early liver cell type has developed (Fig. 3). The growth factor treated cultures expressed early sets of liver genes as ␣1 -antitrypsin, ␣FP, FABP1, HNF3, HNF4␣, CK8 and CK18 (Fig. 4). In the control cultures the expression of EpCAM, ␣1 -antitrypsin, ␣FP, FABP1, CK7/8/18/19 and CD54 expression was detected; however, the morphology of the cells differed considerably between the two culture protocols (data not shown). Quantification by Q-PCR showed significantly elevated levels of ␣1 -antitrypsin, ␣FP and FABP1 expression in the cytokine treated cultures compared to the intrinsically differentiated control (Fig. 4). Western blot showed as well upregulation of the ␣FP gene expression in the DE-Hep progenitor culture. Taken together, phase II cultures consisted mostly of cells expressing early liver markers whereas in the control cultures only few areas of such cells were found. 3.3. Phase III—hepatic maturation In phase III, the cultures were subsequently changed to hepatocyte differentiation medium containing EGF, HGF, insulin, transferrin, dexamethasone, hydrocortisone, ascorbic acid and
oncostatin M in the presence of serum (Fig. 1). At this stage many hepatocyte-like cells emerged and exhibited typical hepatocyte morphology with a polygonal shape, containing distinctive nuclei with a nucleoli. Interestingly, they were arranged in small islet-like clusters mid to late phase III and differed from the morphology of the hepatocyte-like cells derived by the intrinsic method that were located only at the periphery of the colonies (Fig. 5 and Söderdahl et al., 2007). All cytokine induced cell cultures had elevated urea secretion and the expression of CYP3A4, CYP1A2, CYP2C9, ␣1 -antitrypsin, ␣FP, HNF4␣, MRP2 and OATP2 was significantly upregulated compared to the intrinsically differentiated control (Fig. 6). Western blot analysis supported our findings by Q-PCR that CYP3A4, CYP1A2, CYP2C9, ␣1 -antitrypsin, ␣FP, HNF4␣, MRP2 and OATP2 were upregulated in DE-Hep cultures. Interestingly expression levels of CYP3A4 have been found to be dependent on the cell line (data not shown). The immunocytochemical analysis pointed out that many DE-Hep cells were positive for CYP1A, CYP3A, and CK18 (Fig. 7). Interestingly, the CYP3A signal intensity on single cell level was comparable to the expression in primary hepatocyte cultures (data not shown). Transport across hepatocyte plasma membranes is a key parameter in hepatic clearance and usually occurs through different carrier-mediated systems. Biliary elimination of anionic compounds is mediated by the multidrug resistance associated protein
G. Brolén et al. / Journal of Biotechnology 145 (2010) 284–294
289
Fig. 4. DE-Hep progenitor cultures in phase II were characterized by (A) immunocytochemistry (ICC), (B) RT-PCR and (C) Western blot. ICC results are summarized in a table (A). RT-PCR shows that DE-Hep progenitors express higher mRNA levels of ␣FP, ␣1 -antitrypsin and FABP1 than the control (intrinsic differentiated hESC) (B). Western blot analysis shows elevated expression of ␣FP compared to the control (C). All the data presented in this figure were generated at stage II day 15–17 at different passages and were compared to the corresponding control (intrinsic differentiated hESC). A Student’s t-test was performed testing the DE-Hep vs the control data. One star represents a p-value ≤ 0.05, two stars ≤0.01, and three stars ≤0.001. N ≥ 3.
2 (MRP2). At this stage also the expression of MRP2 could be detected in the DE-Hep cultures but not in the control cultures (Figs. 6 and 7). 3.4. Activity of DE-Hep cells In order to assess whether the DE-Hep are functional, they were tested for indocyanine green (ICG) uptake and secretion, ammonia metabolism, glycogen storage and metabolic capability of distinct drugs. The uptake of ICG by DE-Hep was analyzed and indeed showed that many DE-Hep in culture exhibited an ICG uptake (Fig. 7E). The uptake and excretion appeared only in a few cells of the control cultures, indicating that intrinsically differentiating cultures are also able to generate functional hepatocyte-like cells but significantly less in proportion compared to growth factor induced DE-Hep. Since human hepatocytes can make and store glycogen, we analyzed glycogen levels by periodic acid–Schiff staining (Fig. 6F and Figs. 6F and 7). Positive staining was found in both DE-Hep as well as in the intrinsically differentiated control cultures. Albumin, a serum protein produced only in the liver, is the major plasma protein that
circulates in the blood. The cells were evaluated for albumin protein expression. As a final degradation product of protein and amino acid metabolism ammonia is converted in the liver to urea and constitutes the most important catabolic pathway for eliminating excess nitrogen in the human body. We could indeed detect elevated levels of urea in the medium of the phase III cells compared to the intrinsically differentiated control cultures (Fig. 6). 3.5. Metabolism of Phenacetin, Midazolam and Diclophenac by the DE-Hep cells The DE-Hep obtained in phase III were tested for their ability to metabolize Phenacetin, Midazolam and Diclophenac via the phase I cytochrome P450 enzymes, CYP1A, CYP3A and CYP2C, respectively. Metabolites of respective substance released into the medium were determined by liquid chromatography–mass spectrometry (LC–MS). The control cultures and DE-Hep cultures were tested at an age of 18–45 days. The analysis of the metabolites in the samples was performed by LC–MS and revealed that both the DE-Hep and the
290
G. Brolén et al. / Journal of Biotechnology 145 (2010) 284–294
Fig. 5. DE-Hep and control cultures in phase III. Phase-contrast images of DE-Hep cultures at day 17 (A) and day 20 (B). Note, that the cells are starting to from sub-structures, are frequently bi-nucleated and exhibit a hepatocyte-like morphology (black arrows, B). (C) Phase-contrast image of hepatocyte-like cells at day 23 (white arrows), that were generated by intrinsic differentiation and used as controls. These cells are located only at the periphery of the growing colony and show different differentiation pattern and no clear sub-structures. (D and E) Human primary hepatocytes in culture. Scale bars: 50 m (A), 100 m (B), 40 m (C and D), and 100 m (E).
intrinsically differentiated control contained cells that were able to metabolize Phenacetin, Midazolam and Diclophenac. However, the DE-Hep cultures showed higher detectable signs of metabolism of Phenacetin than the intrinsically differentiated control (Fig. 8) and were comparable to the HepG2 control. In contrast to HepG2 could both our intrinsically differentiated control as well as our DE-Hep metabolize Diclophenac. However, the substrates are not specific for 1A2 and 3A4 and could reflect even the activity of other members of the CYP family. Therefore it is important to relate the activity data to the CYP immunostainings of the cultures (Fig. 6A). Activity of the control cultures could reflect therefore the presence of other members of the CYP family since there were found only few CYP1A and CYP3A positive cells. 4. Discussion In vitro the differentiation of human embryonic stem cells (hESC) to a specific germ layer or cell type has proven to be difficult, most likely because cells do not receive the same coordinated developmental cues/signals as they do in vivo. This has particularly affected the differentiation of hESC towards definitive endoderm (DE) and its derivatives, mostly due to the overlap of gene expression patterns between DE and extraembryonic endoderm and the following
lack of specific markers for early definitive endoderm (Pera et al., 2004). In this study, we demonstrate that hESC can be differentiated via DE to functional hepatocyte-like cells (DE-Hep), a terminally differentiated cell type of endoderm. We have used a three-stage protocol for differentiating hESC through a series of endodermal intermediates resembling those that occur during hepatic development in vivo. We characterized the cells at the three differentiation phases at morphological, RNA and protein level using quantitative PCR, immunocytochemistry, Western blot, ELISA and various functionality tests in vitro. Phase I involves the formation of DE. The importance of growth in the presence of Activin A to the differentiation process of hESC was demonstrated previously (D’Amour et al., 2005, 2006; Agarwal et al., 2008). Although methods for hepatic differentiation of hESC have been described, many of them did not exclude the extraembryonic differentiation of hESC (Rambhatla et al., 2003; Hay et al., 2007). This is of critical importance since large sets of genes are co-expressed between liver and yolk sac, such as ␣1-antitrypsin, ␣FP, albumin and even CYPs (Meehan et al., 1984; Pedersen et al., 1985). To overcome this problem, we induced hESC into DE by culturing them for several days in the presence of Activin A/FGF2. Our data in Fig. 2 confirmed that an early ␣FP and ␣1-antitrypsin
G. Brolén et al. / Journal of Biotechnology 145 (2010) 284–294
291
Fig. 6. DE-Hep cultures in phase III characterized by (A) immunocytochemistry (ICC) and ELISA, (B) RT-PCR, and (C) Western blot. The ICC and ELISA results are summarized in a table (A). ICC shows that DE-Hep cells have a marker expression comparable to immature hepatocytes. (B) RT-PCR shows that DE-Hep express much higher mRNA levels of all markers tested than the control culture, and much higher mRNA levels of CYP1A2, CYP3A4, CYP2C9, OATP2 mRNA levels than HepG2. (C) Western blot analysis shows expression of albumin, ␣FP, ␣1 -antitrypsin, CYP1A2, CYP3A4, CYP2C9, MRP2 and OATP2 for DE-Hep cultures in phase III day 29 that was barely or not found in the control. All data presented in this figure were generated at stage III day 24–45. The data were compared to the corresponding control (intrinsic differentiated hESC). Student’s test was performed testing the DE-Hep vs the control data. One star represents a p-value ≤ 0.05, two stars ≤0.01, and three stars ≤0.001. N ≥ 3.
expression, as we regarded as a sign of extraembryonic lineages, decreases significantly when cells have been treated with Activin A. On the other hand could we detect high expression of Sox17, CXCR4 and HNF3 in growth factor treated cultures whereas the intrinsically differentiated control cultures showed high levels of ␣1-antitrypsin and ␣FP (Fig. 2). This effect was even amplified by a combination of Activin A and FGF2. These findings are supported by previous studies with mouse embryonic stem cells where it was shown that interplay between FGFs and Activin A participates in ES cell differentiation to definitive endoderm (Funa et al., 2007). The next stage, phase II, involved the induction towards hepatic progenitor cell types. Combinations of BMPs and FGFs efficiently initiated hepatic differentiation from the growth factor treated cul-
tures as previously reported with mouse ES cell differentiation to hepatic derivates (Gouon-Evans et al., 2006). The combination BMP2/4 and FGF1/2/4 initiated efficiently hepatic differentiation in phase I cultures. At this stage numerous cells expressed early liver genes (Figs. 3 and 4). The DE-Hep cultures expressed significantly more ␣FP, ␣1-antitrypsin and FABP1 than the intrinsically differentiated cultures (Fig. 4). It is known that during liver development in vivo, cholangiocytes and hepatocytes are derived from the common progenitor called hepatoblasts. In our study, EpCAM as well as E-cadherin, was found at this early phase in many of the proliferating cells. The vast majority of EpCAM expressing cells co-expressed CK19 and many of the cells had a typical hepatic morphology indicating the early hepatic cell fate. Furthermore, next to EpCAM and CK19 express-
292
G. Brolén et al. / Journal of Biotechnology 145 (2010) 284–294
Fig. 7. Characterization of DE-Hep in phase III after 21 days in culture. Immunofluorescent labeling of CYP1A (green) (A), CYP3A (green) expression (B), the organic anion transporter MRP2 (red) (C), and CK18 (D). Indocyanine green (ICG) uptake after 1 h incubation (E) and glycogen storage (pink) (F). Note that some bi-nucleated cells store glycogen (black arrow). Scale bars: (A–H) 50 m.
ing regions many CD54 expressing cells were detected supporting a hepatic fate. In phase III the cells maturated in hepatocyte differentiation medium. This phase is characterized by cells of typical hepatocyte
Fig. 8. Functional drug metabolism test of DE-Hep cultures in phase III (day 29) measured by LC–MS. DE-Hep, control (intrinsic differentiation) and HepG2 cultures were tested for their ability to metabolize Phenacetin, Midazolam and Diclophenac via the phase I cytochrome P450 enzymes CYP1A, CYP3A, CYP2C, respectively. The CYP1A activity in DE-Hep cultures was significantly higher compared with control. HepG2 did not show any CYP2C activity. Student’s t-test was performed (DE-Hep vs control). One star represents a p-value ≤ 0.05, two stars ≤0.01, and three stars ≤0.001. N = 3.
morphology with many bi-nuclear cells. Most strikingly, many cells arranged in islands (Fig. 5) of CYP3A4/7 positive cells (Fig. 7). These sub-structures have never been found in the control cultures nor they have been mentioned in previously published studies (Cai et al., 2007; Hay et al., 2007, 2008; Duan et al., 2007; Agarwal et al., 2008). As in most of the published reports our cells also express large sets of liver associated genes and proteins (Figs. 6 and 7). MRP2 and CYPs are indicators of mature hepatocytes and play important functions in the liver. The majority of oxidative metabolic reactions are mediated by the CYP superfamily of enzymes (Sheweita, 2000). For example, CYP1A2, CYP3A4 and CYP2C9 are expressed in human liver microsomes (Stearns et al., 1995) and they are considered relatively hepatocyte specific (Venkatakrishnan et al., 2001), in particular CYP1A2 (McKinnon et al., 1991). The presence of CYP gene expression and especially their metabolic capacity highlights the maturation grade of the DE-Hep in phase III. So far only a few studies have shown constitutive activity of distinct CYPs in hESC derived hepatocyte-like cells (Basma et al., 2009). CYP3A is of special interest, since it plays a central role in the metabolism of aromatic amines, estrogen compounds, and certain drugs (Ingelman-Sundberg, 2004). Our results demonstrate that our guided differentiation of hESC generates a cell population that express liver-specific functions and can be found even in human primary hepatocytes. Additionally, in phase III DE-Hep accumulate glycogen as shown by glycogen staining, take up ICG and secrete urea (Fig. 6). Furthermore we show functionality in drug metabolism in the DE-Hep cultures (Fig. 8). However, we could not
G. Brolén et al. / Journal of Biotechnology 145 (2010) 284–294
see a significant difference between our intrinsically differentiated cells and DE-Hep (Fig. 8). One explanation could be the presence of more intestinal and adipocyte cells in the intrinsically differentiated cells which is supported by the gene array (Synnergren et al., 2009) and by the expression of CYP3A in the gut epithelium (Kolars et al., 1994). Together these findings indicate that DE-Hep have the transcriptional machinery for producing many of the functional proteins present in the liver. Under preparation of this paper an alternative strategy has been published using hepatic induction without biologically relevant factors like BMPs and FGFs. Hay et al. demonstrated recently that after DE induction synthetic low molecular weight compounds such as DMSO initiate liver development. To find out whether this non-biological way of steering cell differentiation, or the strategy using naturally occurring growth factors used in this study, gives rise to the most mature hepatocyte-like cells has to be investigated in comparative studies in the future. In conclusion, gene- and protein expression as well as the functional data clearly demonstrated that differentiation of hESC with this growth factor guided protocol generated hepatocyte-like cells with many liver-like properties. We could show that hESC derived hepatocyte-like cells can organize and form sub-structures in vitro, but most importantly, we demonstrate the presence and functionality of CYP1A, CYP3A and CYP2C, which are shown to metabolize pharmaceutical compounds specific for these three key-enzymes. Our data suggest that the DE-Hep in general exhibit not only higher protein- and gene expression levels than the control cultures, but also express much higher amounts of drug metabolizing enzymes. This indicates a great potential of these cells to be further developed by suitable protocols into hepatocyte-like cells with phenotype mature enough for future use in studies of drug metabolism and drug-induced hepatotoxicity. Acknowledgments We acknowledge Karin Noaksson, Jörg Benecke and Fredrik Wessberg for the hESC propagation; Gunilla Caisander for the primary hepatocyte culture and ELISA analysis. The work was supported by the 6th EU framework program (LSHB-CT-2006-018940) Vitrocellomics, the 6th EU framework program (LSHB-CT-2005-512145) BetaCelltherapy and the Carcinogenomics framework (CARCINO; LSHB-CT-2006-037712). References Agarwal, S., Holton, K.L., Lanza, R., 2008. Efficient differentiation of functional hepatocytes from human embryonic stem cells. Stem Cells 26 (5), 1117–1127. Basma, H., Soto-Gutiérrez, A., Yannam, G.R., Liu, L., Ito, R., Yamamoto, T., Ellis, E., Carson, S.D., Sato, S., Chen, Y., Muirhead, D., Navarro-Alvarez, N., Wong, R.J., Roy-Chowdhury, J., Platt, J.L., Mercer, D.F., Miller, J.D., Strom, S.C., Kobayashi, N., Fox, I.J., 2009. Differentiation and transplantation of human embryonic stem cell-derived hepatocytes. Gastroenterology 136 (3), 990–999. Belo, J.A., Bouwmeester, T., Leyns, L., Kertesz, N., Gallo, M., Follettie, M., De Robertis, E.M., 1997. Cerberus-like is a secreted factor with neutralizing activity expressed in the anterior primitive endoderm of the mouse gastrula. Mechanism of Development 68, 45–57. Bradford, M.M., 1976. A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein–dye binding. Analytical Biochemistry 72, 248–254. Butura, A., Johansson, I., Nilsson, K., Wärngård, L., Ingelman-Sundberg, M., SchuppeKoistinen, I., 2004. Differentiation of human hepatoma cells during confluence as revealed by gene expression profiling. Biochemistry and Pharmacology 67 (7), 1249–1258. Cai, J., Zhao, Y., Liu, Y., Ye, F., Song, Z., Qin, H., Meng, S., Chen, Y., Zhou, R., Song, X., Guo, Y., Ding, M., Deng, H., 2007. Directed differentiation of human embryonic stem cells into functional hepatic cells. Hepatology 45 (5), 1229–1239. D’Amour, K.A., Agulnick, A.D., Eliazer, S., Kelly, O.G., Kroon, E., Baetge, E.E., 2005. Efficient differentiation of human embryonic stem cells to definitive endoderm. Nature Biotechnology 23 (12), 1534–1541. D’Amour, K.A., Bang, A.G., Eliazer, S., Kelly, O.G., Agulnick, A.D., Smart, N.G., Moorman, M.A., Kroon, E., Carpenter, M.K., Baetge, E.E., 2006. Production of pancreatic hormone-expressing endocrine cells from human embryonic stem cells. Nature Biotechnology 24 (11), 1392–1401.
293
Duan, Y., Catana, A., Meng, Y., Yamamoto, N., He, S., Gupta, S., Gambhir, S.S., Zern, M.A., 2007. Differentiation and enrichment of hepatocyte-like cells from human embryonic stem cells in vitro and in vivo. Stem Cells 12, 3058–3068. Ek, M., Söderdahl, T., Küppers-Munther, B., Edsbagge, J., Andersson, T.B., Björquist, P., Cotgreave, I., Jernström, B., Ingelman-Sundberg, M., Johansson, I., 2007. Expression of drug metabolizing enzymes in hepatocyte-like cells derived from human embryonic stem cells. Biochemistry and Pharmacology 74 (3), 496– 503. Funa, N.S., Saldeen, J., Akerblom, B., Welsh, M., 2007. Interdependent fibroblast growth factor and activin A signaling promote the expression of endodermal genes in differentiating mouse embryonic stem cells expressing Src Homology 2-domain inactive Shb. Differentiation 22, 367–376. Gouon-Evans, V., Boussemart, L., Gadue, P., Nierhoff, D., Koehler, C.I., Kubo, A., Shafritz, D.A., Keller, G., 2006. BMP-4 is required for hepatic specification of mouse embryonic stem cell-derived definitive endoderm. Nature Biotechnology 24 (11), 1402–1411. Hamazaki, T., Iiboshi, Y., Oka, M., Papst, P.J., Meacham, A.M., Zon, L.I., Terada, N., 2001. Hepatic maturation in differentiating embryonic stem cells in vitro. FEBS Letters 497 (1), 15–19. Hamazaki, T., Terada, N., 2003. In vitro differentiation of embryonic stem cells into hepatocytes. Methods in Enzymology 365, 277–287. Hay, D.C., Zhao, D., Ross, A., Mandalam, R., Lebkowski, J., Cui, W., 2007. Direct differentiation of human embryonic stem cells to hepatocyte-like cells exhibiting functional activities. Cloning Stem Cells 9 (1), 51–62. Hay, D.C., Zhao, D., Fletcher, J., Hewitt, Z.A., McLean, D., Urruticoechea-Uriguen, A., Black, J.R., Elcombe, C., Ross, J.A., Wolf, R., Cui, W., 2008. Efficient differentiation of hepatocytes from human embryonic stem cells exhibiting markers recapitulating liver development in vivo. Stem Cells, Epub ahead of print. Heins, N., Englund, M.C., Sjöblom, C., Dahl, U., Tonning, A., Bergh, C., Lindahl, A., Hanson, C., Semb, H., 2004. Derivation, characterization, and differentiation of human embryonic stem cells. Stem Cells 22 (3), 367–376. Heins, N., Lindahl, A., Karlsson, U., Rehnström, M., Caisander, G., Emanuelsson, K., Hanson, C., Semb, H., Björquist, P., Sartipy, P., Hyllner, J., 2006. Clonal derivation and characterization of human embryonic stem cell lines. Journal of Biotechnology 122 (4), 511–520. Hogan, B., Beddington, R., Costantini, F., Lacy, E., 1994. Manipulating the Mouse Embryo: A Laboratory Manual. Cold Spring Harbor Laboratory Press, Plainview, NY, pp. 260–261. Ingelman-Sundberg, M., 2004. Human drug metabolising cytochrome P450 enzymes: properties and polymorphisms. Naunyn Schmiedebergs Archives of Pharmacology 369 (1), 89–104. Kanai-Azuma, M., Kanai, Y., Gad, J.M., Tajima, Y., Taya, C., Kurohmaru, M., Sanai, Y., Yonekawa, H., Yazaki, K., Tam, P.P., Hayashi, Y., 2002. Depletion of definitive gut endoderm in Sox17-null mutant mice. Development 129 (10), 2367– 2379. Kolars, J.C., Lown, K.S., Schmiedlin-Ren, P., Ghosh, M., Fang, C., Wrighton, S.A., Merion, R.M., Watkins, P.B., 1994. CYP3A expression in human gut epithelium. Pharmacogenetics 4 (5), 247–259. Lawson, K.A., Meneses, J.J., Pedersen, R.A., 1986. Cell fate and cell lineage in the endoderm of the presomite mouse embryo, studied with an intracellular tracer. Developmental Biology 115 (2), 325–339. Lawson, K.A., Pedersen, R.A., 1987. Cell fate, morphogenetic movement and population kinetics of embryonic endoderm at the time of germ layer formation in the mouse. Development 101 (3), 627–652. Lawson, K.A., Meneses, J.J., Pedersen, R.A., 1991. Clonal analysis of epiblast fate during germ layer formation in the mouse embryo. Development 113 (3), 891– 911. McKinnon, R.A., Hall, P.D., Quattrochi, L.C., Tukey, R.H., McManus, M.E., 1991. Localization of CYP1A1 and CYP1A2 messenger RNA in normal human liver and in hepatocellular carcinoma by in situ hybridization. Hepatology 14 (5), 848– 856. Meehan, R.R., Barlow, D.P., Hill, R.E., Hogan, B.L., Hastie, N.D., 1984. Pattern of serum protein gene expression in mouse visceral yolk sac and fetal liver. EMBO Journal 3 (8), 1881–1885. Pedersen, R.A., Meneses, J., Spindle, A., Wu, K., Galloway, S.M., 1985. Cytochrome P-450 metabolic activity in embryonic and extraembryonic tissue lineages of mouse embryos. Proceedings of the National Academy of Sciences 82 (10), 3311–3315. Pera, M.F., Andrade, J., Houssami, S., Reubinoff, B., Trounson, A., Stanley, E.G., Wardvan Oostwaard, D., Mummery, C., 2004. Regulation of human embryonic stem cell differentiation by BMP-2 and its antagonist noggin. Journal of Cell Science 117, 1269–1280. Rambhatla, L., Chiu, C.P., Kundu, P., Peng, Y., Carpenter, M.K., 2003. Generation of hepatocyte-like cells from human embryonic stem cells. Cell Transplantation 12 (1), 1–11. Sheweita, S.A., 2000. Drug-metabolizing enzymes: mechanisms and functions. Current Drug Metabolism 1 (2), 107–132. Sasaki, H., Hogan, B.L., 1993. Differential expression of multiple fork head related genes during gastrulation and axial pattern formation in the mouse embryo. Development 118, 47–59. Söderdahl, T., Küppers-Munther, B., Heins, N., Edsbagge, J., Björquist, P., Cotgreave, I., Jernström, B., 2007. Glutathione transferases in hepatocyte-like cells derived from human embryonic stem cells. Toxicology In Vitro 21 (5), 929– 937. Stearns, R.A., Chakravarty, P.K., Chen, R., Chiu, S.H., 1995. Biotransformation of losartan to its active carboxylic acid metabolite in human liver microsomes. Role of
294
G. Brolén et al. / Journal of Biotechnology 145 (2010) 284–294
cytochrome P4502C and 3A subfamily members. Drug Metabolism and Disposition 23 (2), 207–215. Synnergren, J., Heins, N., Brolén, G., Eriksson, G., Lindahl, A., Hyllner, J., Olsson, B., Sartipy, P., Björquist, P., 2009. Transcriptional profiling of human embryonic stem cells differentiating to definitive and primitive endoderm and further towards the hepatic lineage. Stem Cells and Development, Epub ahead of print.
Tremblay, K.D., Zaret, K.S., 2005. Distinct populations of endoderm cells converge to generate the embryonic liver bud and ventral foregut tissues. Developmental Biology 280 (1), 87–99. Venkatakrishnan, K., Von Moltke, L.L., Greenblatt, D.J., 2001. Human drug metabolism and the cytochromes P450: application and relevance of in vitro models. Journal of Clinical Pharmacology 41 (11), 1149–1179.