| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
OPEN ACCESS ARTICLE
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
EMBRYONIC STEM CELLS/INDUCED PLURIPOTENT STEM CELLS |
ari
a
aInstitute of Neurophysiology, Center of Physiology and Pathophysiology, University of Cologne, Cologne, Germany;
bMedical Clinic and Policlinic III, Justus-Liebig University of Giessen and Marburg, Giessen, Germany;
c Axiogenesis AG, Cologne, Germany
Key Words. Murine embryonic stem cells • Transgenic cell clones • Hepatocyte-like cells • In vitro differentiation • Antibiotic selection
Correspondence: Irina Drobinskaya, Ph.D., Institute for Neurophysiology, Center of Physiology and Pathophysiology, University of Cologne, Robert-Koch Str. 39, D-50931 Cologne, Germany. Telephone: + 49-221-478-86772; Fax: + 49-221-478-3834; e-mail: irina.drobinskaya{at}uni-koeln.de; or Eugen Kolossov, Ph.D., Axiogenesis AG, Nattermannallee 1, Building S20, D-50829 Cologne, Germany. Telephone: + 49-221-99-881818; Fax: + 49-221-99-881810; e-mail: eugen.kolossov{at}axiogenesis.com
Received April 22, 2008;
accepted for publication June 2, 2008.
First published online in STEM CELLS EXPRESS June 12, 2008.
 |
ABSTRACT |
|---|
 TopFootnotes Abstract IntroductionMaterials and MethodsResultsDiscussionConclusionDisclosure of Potential...AcknowledgmentsReferences |
|---|
-fetoprotein gene promoter driving a cassette for the enhanced green "live" fluorescent reporter protein (eGFP) and a puromycin resistance gene. Under established culture conditions these clones allowed for both monitoring of differentiation and for puromycin selection of hepatocyte-committed cells in a suspension mass culture of transgenic ESC aggregates ("embryoid bodies" [EBs]). When plated on fibronectin, the selected eGFP-positive cells formed colonies, in which intensely proliferating hepatocyte precursor-like cells gave rise to morphologically differentiated cells expressing -1-antitrypsin, -fetoprotein, and albumin. A number of cells synthesized glycogen and in some of the cells cytokeratin 18 microfilaments were detected. Major hepatocyte marker genes were expressed in the culture, along with the gene and protein expression of stem/progenitor markers, suggesting the features of both hepatocyte precursors and more advanced differentiated cells. When cultured in suspension, the EB-derived puromycin-selected cells formed spheroids capable of outgrowing on an adhesive substrate, resembling the behavior of fetal mouse hepatic progenitor cells. The established system based on the highly efficient selection/purification procedure could be suitable for scalable generation of ESC-derived hepatocyte- and hepatocyte precursor-like cells and offers a potential in vitro source of cells for transplantation therapy of liver diseases, tissue engineering, and drug and toxicology screening. Disclosure of potential conflicts of interest is found at the end of this article.
|
INTRODUCTION |
|---|
|
TopFootnotesAbstractIntroductionMaterials and MethodsResultsDiscussionConclusionDisclosure of Potential...AcknowledgmentsReferences |
|---|
However, there are a number of principal problems, which have to be solved before such an approach could become a clinical reality. One of the most obvious ones is that terminally specified cells of a certain lineage represent only a minor cell fraction within a differentiating ESC culture. Furthermore, persistent undifferentiated ESCs remaining in the culture possess a high risk of teratocarcinoma development, particularly after an autologous transplantation or after transplantation into immunosuppressed recipients [18, 19]. Therefore, a large-scale production of highly purified cell lineages of interest is a paramount task of the ESC-based approach in regenerative medicine.
As previously shown, cells of the outer layer of differentiating ESC aggregates ("embryoid bodies" [EBs]) resemble endodermal cells of the native embryo regarding the expression of endodermal markers including -fetoprotein (AFP) [11, 20–22]. Several groups have demonstrated the hepatocyte-like differentiation of the outer-layer AFP-positive cells after their separation from EBs [11, 12, 23]. Attempts to isolate ESC-derived hepatocyte-like cells were, to date, based on fluorescence-activated sorting (FACS) of cells expressing enhanced green fluorescent protein (eGFP) under control of the AFP gene promoter [12] or on the elimination of undifferentiated cells by magnetic activated cell sorting (MACS) [24]. However, due to its low yield/cost ratio, as well as to the problem of the cell vitality, FACS procedure cannot be considered as a relevant approach for the large-scale production of purified cells. Applicability of MACS is limited by the lack of sufficiently specific membrane receptors in hepatocytes, necessary to ensure their efficient separation from other cell types. By contrast, approaches using genetic selection have been proven to be highly efficient in terms of purity of the resulting cell cultures. They were successfully used for selection of cardiomyocytes, endothelial cells, and neuronal cells [25–29] and have been shown to be suitable for the large-scale production in a bioreactor mode [26].
The goal of the present study was to investigate the development of AFP-expressing cells within EBs, to examine the pattern of hepatocyte-like differentiation of these cells, and to establish a system for efficient generation and antibiotic selection of hepatocyte-like cells and their precursors. For this purpose, we used EB cultures derived from transgenically modified ESCs possessing an expression vector, which contains both antibiotic resistance and "live" fluorescent reporter genes under control of a common tissue-specific gene promoter. Based on the published data on the hepatocyte-like differentiation of ESCs within EBs and taking into account that AFP is highly expressed in fetal hepatocytes from the early stages of liver embryogenesis onward [30], we used the promoter-enhancer element of the AFP gene as a driving element for the reporter- and drug-resistance cassettes. This approach allowed for tracing both differentiation and selection processes in a live mode, as well as for generating highly purified cultures of hepatocyte- and hepatocyte precursor-like cells under mass culture conditions.
|
MATERIALS AND METHODS |
|---|
|
TopFootnotesAbstractIntroductionMaterials and MethodsResultsDiscussionConclusionDisclosure of Potential...AcknowledgmentsReferences |
|---|
Generation of Transgenic ESC Clones
The AFP-PIG vector was linearized by the XhoI restrictase and 30 µg was used for transfection into 5 x 106 mouse D3 ESCs through the standard electroporation procedure [33]. Transfected cells were maintained for 8 days on a feeder layer of mitomycin-treated mouse embryonic fibroblasts in Dulbecco's modified Eagle's medium (DMEM) containing 15% (vol/vol) fetal bovine serum (FBS), 0.1 mM minimum essential medium (MEM) nonessential amino acids, 100 U/ml penicillin, 100 µg/ml streptomycin (all from Invitrogen, Eggenstein, Germany, http://www.invitrogen.com), 0.05 mM β-mercaptoethanol (Sigma-Aldrich, München, Germany, http://www.sigmaaldrich.com), 500 U/ml leukemia inhibitory factor (LIF; ESGRO) (Chemicon International, Hampshire, U.K., http://www.chemicon.com), und 300 µg/ml G418 (Invitrogen). Colonies of G418-resistant ESCs were then picked and propagated on a fibroblast feeder layer in the above-described culture medium.
Mass Culture of Transgenic EBs and Monitoring and Puromycin Selection of eGFP-Expressing Cells
Colonies of transgenic ESCs were dissociated with 0.25% trypsin-EDTA (Invitrogen). To generate EBs in a mass culture mode and to start differentiation, a single-cell suspension (2.5 x 105 cells per 1 ml DMEM-based medium) was prepared and maintained in bacterial Petri dishes (Labomedic, Bonn, Germany, http://www.labomedic.de) at 37°C and 5% CO2 on a shaker (reciprocating/20 mm/50 minutes–1). From this step onward, a medium containing 20% FBS and the above-mentioned supplements (with exception of LIF and G418), indicated in the sections below as "medium" was used. After 2 days (by this time, the EBs became macroscopically visible), the EB culture was diluted to about 30 EBs per 1 ml medium and further incubated under the above-mentioned culture conditions (without shaking) for different time periods. Half of the medium was changed every other day. The EBs were then concentrated to about 150 EBs per 1 ml medium, and 5 µg/ml puromycin (Sigma-Aldrich) was introduced for various time periods. The same puromycin concentration was used in all selection procedures described below, independently of culture conditions. Green fluorescent cell areas within EBs before and during the puromycin application were live monitored under the fluorescent microscope Axiovert 200M (Zeiss, Jena, Germany, http://www.zeiss.com), using the HQ-Longpass Filterset for Enhanced-GFP. This and other filter sets used in this study were from AHF Analysentechnik (Tübingen, Germany, http://www.ahf.de).
Generation and Monitoring of EB-Derived Adherent Cell Cultures
Puromycin-treated EBs were dissociated with collagenase Type II (CellSystems, St. Katharinen, Germany, http://www.cellsystems.de) for 20 minutes at 37°C on the shaker (2 ml of 0.1% collagenase solution was added to portions of approximately 3,000 EBs each). After addition of 8 ml of medium to dilute the enzyme, the resulting cell aggregates were centrifuged (1000 rpm, 5 minutes), washed once with the medium, resuspended in 1 ml medium, and plated onto cell culture dishes (20 µl/cm2 surface area) precoated with fibronectin (CellSystems) (15 µg/ml in phosphate-buffered saline [PBS]). The plated cultures were maintained at 37°C and 5% CO2 in the medium permanently containing puromycin. Adherent cells were monitored using the Filterset for 4',6-diamidino-2-phenylindole (DAPI), Hoechst for transmission light microscopy, and the HQ-Longpass Filterset for Enhanced-GFP for registering eGFP fluorescence.
Suspension Culture Derived from Collagenase-Treated EBs
The 9-day-old suspension EB culture was exposed to puromycin for another 3 days and then dissociated with collagenase Type II as mentioned above. The dissociated material was filtered through 100-µm nylon cell strainers (BD Biosciences, Heidelberg, Germany, http://www.bdbiosciences.com), centrifuged (1000 rpm, 5 minutes), resuspended in the puromycin-containing medium (5 ml per cell pellet derived from approximately 3,000 EBs), and cultured in Petri dishes at 37°C and 5% CO2 for another 5 days. The culture was monitored under the fluorescent microscope, using the HQ-Longpass Filterset for Enhanced-GFP. Estimation of the size of eGFP-fluorescent cell aggregates ("spheroids") (30–60 for each time point) was performed with the help of the Analysis-Pro-3.2 software (Olympus Soft Imaging Solutions GmbH, Münster, Germany, http://www.olympus-sis.com).
Cultivation of Hepatoma Cells
Mouse hepatoma cells of the line HEPA 1–6 (German Collection of Microorganisms and Cell Cultures [DSMZ], Braunschweig, Germany, http://www.dsmz.de) were cultivated in DMEM supplemented with 15% (vol/vol) FBS, 0.1 mM MEM nonessential amino acids, 100 U/ml penicillin, 100 µg/ml streptomycin, and 0.05 mM β-mercaptoethanol. The cells were split every fourth day, according to the manufacturer's protocol.
Propidium Iodide Labeling of EB Cultures for Microscopic Detection of Dead Cells
Intact or puromycin-treated EBs were incubated in the medium in the presence of 1 µg/ml propidium iodide (PI) (Sigma-Aldrich) for 15 minutes at 37°C. The EBs were then monitored under the fluorescent microscope, using the HQ-Longpass Filterset for Enhanced-GFP and the HQ-Filterset for Texas Red.
FACS Analysis of Cell Cultures
Quantity of eGFP-expressing cells was analyzed in intact EB suspension cultures at day 7 or day 9 of the differentiation protocol, as well as in cultures of adherent puromycin-selected cells. The latter were generated from 9-day-old EBs exposed to puromycin for another 3 days and then treated with collagenase, as described in "Generation and Monitoring of EB-Derived Adherent Cell Cultures." For analysis we used cells after 3-day-long growth on fibronectin in the presence of puromycin. The cell cultures were dissociated with 0.25% trypsin-EDTA, washed once with the medium, and then resuspended in CellWash (BD Pharmingen, Heidelberg, Germany, http://www.bdbiosciences.com). The single-cell suspensions (minimum 10,000 cells per sample) were analyzed with the FACScan flow cytometer equipped with a 488-nm argon-ion laser (Becton Dickinson, Höhenkirchen, Germany, http://www.bd.com). To exclude dead cells from calculation, 1 µg/ml of PI was added to cell suspensions prior to analysis. The emitted fluorescence of eGFP and PI was measured at 530 nm (fluorescein isothiocyanate [FITC] band-pass filter) and 585 nm (phycoerythrin band-pass filter), respectively. The PI-negative cell population displaying intermediate and high forward scatter was gated and the percentage of eGFP-positive cells was determined using the CellQuest software (Becton Dickinson). Nontransfected ESCs were used as a negative control.
Reverse-Transcription–Polymerase Chain Reaction Analysis
Total RNA was extracted from ESCs, intact suspension EB cultures, puromycin-selected plated cultures after 3-day-long growth on fibronectin, mouse embryonic fibroblasts, and 129/Sv mouse liver with the High Pure RNA Isolation Kit (Roche Applied Science, Mannheim, Germany, http://www.roche-applied-science.com), including DNase treatment, according to the manufacturer's protocol. Expression of following genes was analyzed: hepatocyte nuclear factor 3β (HNF3β), Sry-box containing gene 17 (Sox17), -fetoprotein (AFP), -1-antitripsin (AAT), albumin, glucose transporter 2 (Glut2), liver-specific organic anion transporter 1 (LST-1), liver glycogen synthase 2 (Gys2), tyrosine aminotransferase (Tat), cytokeratin 8 (CK8), cytokeratin 18 (CK18), cytokeratin 19 (CK19), CD34 antigen (CD34), cell surface protein stem cell antigen 1 (Sca-1), thymus cell surface antigen 1 (Thy-1), receptor tyrosinkinase c-Kit (c-Kit), NK2 transcription factor related, locus 5 (Nkx2.5), paired box gene 6 (Pax6), goosecoid homeobox (Gsc), and glyceraldehyde phosphate dehydrogenase (GAPDH). RNA was reverse-transcripted with the Moloney murine leukemia virus reverse transcriptase (Sigma-Aldrich) and random nanomer primers (MWG, Ebersberg/München, Germany, http://www.mwg-biotech.com). For polymerase chain reaction (PCR), the GoTag Flexi DNA Polymerase (Promega, Mannheim, Germany, http://www.promega.com) was used. cDNA corresponding to 0.2 µg RNA was denatured at 94°C for 2 minutes, followed by 35 cycles of the following: 30-second denaturation at 94°C, 30-second annealing at respective temperatures, and 1-minute elongation at 72°C. Final extension step was at 72°C for 7 minutes. For primer sequences see references [6, 7, 34–44] as well as supplemental online Table 1 also containing information on PCR products and annealing temperatures used.
Fluorescent Immunohistochemistry Assay
Plated cell cultures were fixed with 4% paraformaldehyde (PFA) (Fluka, Neu-Ulm, Germany, http://www.fluka.org) in PBS for 30 minutes at room temperature, washed with PBS, permeabilized with 0.5 M NH4Cl in PBS containing 0.25% Triton X-100 (Sigma-Aldrich), and blocked with 5% bovine serum albumin in PBS for 1 hour. All above-mentioned reagents were from Applichem (Darmstadt, Germany, http://www.applichem.de). Corresponding primary antibodies were then added for overnight incubation at 4°C: rabbit polyclonal anti-AFP antibody (1:100) (DAKO, Hamburg, Germany, http://www.dako.com), goat polyclonal anti-AAT antibody (1:100) (Santa Cruz Biotechnology, Santa Cruz, CA, http://www.scbt.com), goat polyclonal anti-albumin antibody (1:100) (Bethyl Laboratories, Montgomery, TX, http://www.bethyl.com), or mouse monoclonal anti-CK18 (biotin) antibody (Abcam, Cambridge, U.K., http://www.abcam.com). For detection of membrane-associated markers, rabbit polyclonal anti-
-like leucine zipper kinase (Dlk) (1:100), rat monoclonal anti-Thy1 (biotin) (1:100), rat monoclonal anti-CD34 (biotin) (1:100), or rat monoclonal anti-c-Kit (biotin) antibody (1:100) (all from Abcam) were used, whereby the permeabilization step was omitted. As secondary antibodies, the cyanin 3 (Cy3)-conjugated mouse anti-rabbit IgG (1:1000) (Rockland Immunochemicals, Gilbertsville, PA, http://www.rockland-inc.com) or the Cy3-conjugated rabbit anti-goat IgG (1:500) (Jackson ImmunoResearch Europe, Cambridgeshire, U.K., http://www.jacksonimmuno.com) was applied for 1 hour at room temperature. The biotin-conjugated antibodies bound to the cells were detected by treatment with Cy3-conjugated streptavidin (25 µg/ml) (The Jackson Laboratory, Bar Harbor, Maine, http://www.jax.org) for 1 hour at 37°C. As isotype controls, respective rabbit or goat polyclonal IgGs and mouse or rat biotinylated monoclonal IgGs (Abcam) were used. After washing out the secondary antibodies, cell nuclei were stained with DAPI (Vector Laboratories, Burlingame, CA, http://www.vectorlabs.com). Cy3 fluorescence and DAPI staining were evaluated microscopically using the HQ-Filterset for Cy3 and the Filterset for DAPI, Hoechst, respectively. The PFA fixation preserves the eGFP fluorescence so that eGFP-positive cells could also be detected in these stainings.
Labeling and Detection of S-Phase Nuclei in Adherent Cells
To investigate cell proliferation, the Bromo-2'-deoxy-uridine Labeling and Detection Kit I (Roche Applied Science) was used according to the manufacturer's protocol. Briefly, at day 2 after plating the cells, 5-bromo-2'-deoxy-uridine (BrdU) was applied to adherent cell cultures for 48 hours. Thereafter, the cells were fixed with the ethanol-glycine fixative for 20 minutes at –20°C and incubated with antibodies provided in the kit. The secondary antibody was FITC conjugated. In addition, nuclei were stained with DAPI as described above. The HQ-Longpass Filterset for Enhanced-GFP was used for microscopic detection of BrdU-positive FITC-labeled nuclei.
Detection of Glycogen-Synthesizing Cells in Adherent Cultures
Glycogen synthesis in cell cultures growing on fibronectin was analyzed by periodic acid-Schiff (PAS) staining as previously described [14]. The cells were oxidized in 1% periodic acid (Sigma-Aldrich) for 5 minutes at room temperature. After washing three times in distilled H2O they were treated with Schiff's reagent (Sigma-Aldrich) for 15 minutes and washed in distilled H2O for 10 minutes. Microscopic observations were carried out using the Hoechst filter for transmission light microscopy.
Test for the Presence of Undifferentiated ESCs in Adherent and Spheroid Cell Cultures
"Feeder test" for controlling cell cultures for the presence of undifferentiated ESCs was carried out as previously described [27]. Puromycin-selected adherent cultures at day 3 after plating the collagenase-dissociated EBs on fibronectin, spheroids at day 5 of cultivation in the puromycin-containing medium, as well as 9-day-old EBs exposed to puromycin for another 3 days were dissociated with 0.25% trypsin-EDTA, washed once with the medium, and seeded onto a fibroblast feeder layer at a density of 10,000 cells per square centimeter of surface area. The cultures were then maintained for 2 weeks under conditions typically used for expansion of ESCs and monitored microscopically for the appearance of ESC colonies.
|
RESULTS |
|---|
|
TopFootnotesAbstractIntroductionMaterials and MethodsResultsDiscussionConclusionDisclosure of Potential...AcknowledgmentsReferences |
|---|
|
Puromycin Treatment of EB Cultures Allows Selection of Viable eGFP-Expressing Cells
To select eGFP-expressing cells, we applied puromycin to suspension cultures of 9-day-old or 12-day-old AFP-PIG EBs containing green-fluorescent cells. These cells survived during at least 10 days of the puromycin treatment and formed brightly fluorescent areas (Fig. 2A, top panel, exemplified for EBs treated with puromycin from day 9 onward). (See also a general differentiation and selection scheme in supplemental online Fig. 1, based on the results obtained in this study.) A similar pattern of puromycin resistance was observed in eGFP-positive EBs generated from another transgenic clone (AFP-PIG-1). Puromycin selection resulted in death of eGFP-negative but not of eGFP-positive cells, as clearly indicated by PI staining of EBs at day 3 of drug application (Fig. 2Ba). In untreated EBs only scarce dead cells were stained with PI (Fig. 2Bb). These data proved a highly efficient selection of eGFP-positive cells.
|
Adherent Cell Cultures Generated from Puromycin-Treated EBs: Proliferation and Differentiation Within Growing Colonies
To further characterize the puromycin-resistant cells comprising eGFP-expressing clusters within EBs, we examined adherent cultures derived from puromycin-treated and enzymatically dissociated EBs. Puromycin was introduced at different stages of the EB culture protocol for various periods of time. The drug-treated EBs were dissociated with collagenase and plated onto fibronectin-coated dishes in the puromycin-containing medium. Notably, the dissociation by collagenase resulted in a suspension containing mostly eGFP-positive cell aggregates of different size, similar to those shown in Figure 7A (leftmost panel) below. At day 3 after plating, the adherent cells were fixed and analyzed by fluorescent immunostaining.
The outcome and features of adherent cultures where found to be dependent on a combination of two factors: (a) EB age at a starting time point of drug selection and (b) duration of drug treatment. Thus, when 9-day-old EBs were exposed to puromycin for 10 days, only few adherent colonies containing 15–30 cells each, which were positive for both eGFP and AAT, were detected (Fig. 2Ea). When puromycin was first applied to EBs at day 12 of differentiation, and collagenase dissociation was performed 10 days later, the selected cells formed even smaller colonies (Fig. 2Eb) (data of three experiments). Reduction of the duration of drug treatment to 3 days before the collagenase dissociation only slightly increased the efficiency of colony formation in the case of 12-day-old EBs (data of three experiments, not shown). In contrast, the same 3-day-long puromycin treatment resulted in a massive growth of the plated culture (up to confluent) when the drug was applied earlier, namely after 9 days of the EB development. In confluent regions of this culture, bright eGFP-fluorescent zones of a high cell nuclei density, displaying a low level of the AAT expression, as well as strongly AAT-positive areas containing relatively scarce nuclei, were observed (Fig. 3A). Microscopic observations of single colonies at the edge of the confluent culture allowed for explanation of this heterogeneity. The colonies typically featured a heterogeneous morphological structure: in their central parts, small, strongly eGFP-positive cells displaying a minor AAT-bound fluorescence were concentrated, whereas peripheral regions consisted of large, morphologically differentiated cells intensely expressing AAT (Fig. 3B, 3C, see also scheme in supplemental online Fig. 1). BrdU labeling revealed a high proliferative activity in the middle of colonies where the percentage of S-phase nuclei was distinctively higher than at the periphery (Fig. 3D and supplemental online Table 2). The culture also stained positively for AFP and albumin, without any distinctive difference between the colony center and periphery. The overwhelming majority of eGFP-positive cells located at peripheries of the colonies stained positively for AAT (94%), AFP (93%), and albumin (98%) (calculated for five colonies each, 367, 279, and 273 cells total, respectively). This suggests that more than 90% of these cells expressed all these proteins. A number of binucleated cells, as well as some cells expressing CK18 with a typical intermediate microfilament pattern [65–67], were detected (Fig. 4A, 4B). Positive PAS staining of numerous cells (Fig. 4C) testified to their capability to synthesize glycogen, which is one of the functional characteristics of hepatocytes.
|
|
|
|
From day 4 of cultivation on fibronectin onward, the eGFP fluorescence of cells at colony peripheries visually decreased, and the decrease was also seen when the drug pressure was withdrawn on day 3 of the adhesive culture, suggesting down-regulation of the AFP gene expression in the course of differentiation.
Hence, the data described in this section demonstrate that the EB-derived puromycin-selected adherent cell culture features characteristics of both early hepatocyte precursors and advanced differentiated cells. Approximate estimation of efficiency of the above-described protocol revealed an output of about 105 selected cells from 106 original transgenic AFP-PIG ESCs.
Puromycin-Selected Cells Are Able to Proliferate in Suspension
As we detected the fraction of highly proliferating cells in the adherent puromycin-selected cultures, we next examined whether those cells are capable of proliferation in a liquid medium too, which would be indicative of their progenitor character. Nine-day-old EBs exposed to puromycin for another 3 days were treated with collagenase, filtered through 100-µm cell strainers, and maintained in suspension in the permanent presence of puromycin. Under these conditions, the green-fluorescent multicellular aggregates separated from EBs remarkably increased in size, forming spheroids, during 5 days of cultivation (Fig. 7A, 7B). When plated onto fibronectin-coated dishes, the 4-day-old spheroids gave rise to eGFP-positive outgrowths that contained small cells located close to the adherent spheroids, as well as larger, morphologically advanced differentiated cells, some of which synthesized glycogen (Fig. 7C). This behavior of the eGFP-expressing cells selected from suspension EB cultures is similar to that of fetal mouse hepatic stem/progenitor cells forming growing spheroids in a liquid medium, which outgrow to expanding colonies after plating onto an adhesive substrate [84].
|
To determine whether final drug-selected cells still contain undifferentiated ESCs, we carried out the so-called feeder test, in which we seeded trypsin-dissociated 3-day-old adherent cultures or 5-day-old spheroids onto a fibroblast feeder layer. In both cases, no ESC colonies were detected within 2 weeks after seeding. In contrast, when 9-day-old EBs exposed to puromycin for another 3 days were treated with trypsin and seeded in the same way, dozens of ESC colonies were detected already at day 7 after seeding (data obtained on three independent cell cultures). This suggests that after puromycin selection from EBs, eGFP-expressing cell cultures (plated or further maintained in suspension) become completely free from undifferentiated ESCs when exposed to the drug for several more days. This result also supports the suggestion that the minor eGFP/PI-negative cell fraction detected by FACS did not represent undifferentiated ESCs.
|
DISCUSSION |
|---|
|
TopFootnotesAbstractIntroductionMaterials and MethodsResultsDiscussionConclusionDisclosure of Potential...AcknowledgmentsReferences |
|---|
To generate, select, and purify ESC-derived hepatocyte-like cells and their precursors under mass culture conditions, we implicated both the live fluorescent (eGFP) and selection (puromycin resistance) markers under common control of the AFP gene promoter in the designed AFP-PIG expression vector. In our experiments this vector appeared as an effective tool enabling the detection and selection of both hepatocyte- and hepatocyte precursor-like cells. This is in accordance with data that showed that in the early-stage mouse embryo the onset of AFP expression occurs in the endoderm of the embryonic region and AFP is synthesized at a high level in the endoderm-derived embryonic hepatocytes [85].
From the early stage (6 days) of the EB suspension culture onward we already detected the AFP promoter-driven eGFP expression, which temporarily correlated with the profile of the AFP gene expression. Localization of eGFP-positive cells to the outer layer of the AFP-PIG EBs is in line with previous studies, which showed that AFP expression as well as expression of other endodermal markers occur in cells of the outer EB surface [11, 20–22]. In some publications these cells were identified as resembling the extraembryonic endoderm of the yolk sac [20, 21, 84]. On the other hand, it was demonstrated that AFP-expressing cells isolated from the outer rim of EBs were able to differentiate into hepatocyte-like cells [11, 12, 23]. The fact that the eGFP-expressing cells selected in our experiments gave rise to hepatocyte- and hepatocyte precursor-like cells shows that at least a part of the outer-layer cells became hepatocyte committed during the EB differentiation, thus suggesting their origin from the definitive endoderm. Notably, no EB expansion into cystic structures thought to resemble the extraembryonic visceral endoderm [20, 86, 87] was observed in our experiments at least up to day 18 of the EB culture.
Approximately 24 hours after the first detection of eGFP-positive cells (namely, at day 7 of the mass culture), their selection by puromycin was feasible. However, it was impossible to keep them viable longer than 2 days, despite sufficient activity of the AFP promoter driving the Pac cassette, as shown by PI labeling. We speculate that these "early-stage" eGFP-expressing cells might represent an endoderm-like and/or a very early hepatocyte precursor cell population(s), which need(s) crosstalk signaling from surrounding (drug sensitive) cell types (e.g., from mesoderm-like cells [88]) for further differentiation. Loss of such signaling may trigger a cascade of apoptotic events in selected cells that lead to cell death following the principle "differentiate or die." In contrast, from day 9 of the EB development onward, eGFP-expressing hepatocyte-committed cells seemed to be able to carry out the differentiation program in the absence of other cell types, thereby enabling their stable puromycin selection. The long-term viability of eGFP-positive cells and rapid death of the eGFP-negative cell fraction demonstrated both high specificity and efficiency of the selection procedure.
The proliferative capacity of cells selected in the suspension EB culture, reflected in their ability to form colonies after separation with collagenase and plating on fibronectin, was dependent on the "age" of EBs at the starting time point of the 3-day-long drug selection: even its modest shift, namely from day 9 to day 12 of the EB culture, significantly reduced both the amount and the size of adherent colonies. Prolongation of drug treatment to 10 days led to a similar effect. Thus, it can be assumed that eGFP-expressing cells, which feature high proliferative capacity when selected and separated under the optimized temporal condition, display a relatively rapid transition from actively dividing precursors to more differentiated cells. The effect of the developmental stage on the ability of the progenitor cells to proliferate was also evident after their plating onto the adhesive substrate: after the initial intensive colony expansion during the first 3–4 days, the colony growth was arrested in conditions of both continued puromycin application and withdrawal of the drug on day 3 of culturing.
In adherent colonies, the hepatocyte precursors underwent further development toward advanced differentiated cells. The latter was confirmed by the differentiated-like morphology of cells located at the colony peripheries and their strongly positive immunostaining for AAT, as well as by the appearance of binucleated and glycogen-synthesizing cells. Moreover, cells with intermediate microfilament structure were identified by immunostaining for CK18, the main type I cytokeratin in differentiated hepatocytes. The development of hepatocyte-like cells in adherent cultures in the presence of puromycin was also corroborated by the expression of the specific gene marker of mature hepatocytes Tat [37, 63] and of the gene for the liver-specific anion transporter LST-1 [60]. On the other hand, gene and protein expression of progenitor cell markers that, besides hematopoietic and nonhematopoietic bone marrow cells [89–94], salivary gland [75], neurons [95, 96], and endothelial cells [97], are also typical for hepatoblasts (common progenitors for both hepatocytes and biliary epithelial cells) isolated from fetal liver [40, 74, 76, 77, 98], as well as for hepatic oval (progenitor) cells found in adult liver [70–73, 99], suggested the presence of precursor cells in the adherent culture. The profile of the expression of progenitor marker proteins in respect of the colony morphology (center/periphery) is in line with previously published data attributing CD34 and c-Kit to early hepatocyte precursors [70, 71, 73] and ascribing Thy-1 and Dlk to advanced differentiated hepatoblasts and fetal hepatocytes [74, 77]. The absence of transcripts for Nkx2.5 and Pax6, on the one hand, and, on the other hand, expression of Gsc confirm the definitive endoderm origin of the EB-derived puromycin-selected adherent culture.
Interestingly, the observed "differentiation gradient" from center to periphery within growing colonies resembles the polarized organization of the liver acinus where maturation process begins with stem and precursor cells in the periportal zone and ends with the senescing parenchyma close to the central vein [100].
In our experiments, the progenitor character of cells selected in the EB mass culture by puromycin application was also confirmed by their ability to multiply and form spheroids in suspension culture under ongoing pressure of the drug. This growth behavior is in line with the similar feature described for neuronal [101], endothelial [102], dermal [103], mammary [104], and fetal mouse hepatic [84] stem/progenitor cells cultured in liquid or semiliquid medium. Further functional characterization of the highly enriched and purified cell cultures described above is on our agenda for the next future.
Precursor cells committed to a certain cell type are obviously advantageous for cell replacement therapy. Thus, a high level of liver reconstitution after transplantation of fetal liver stem/progenitor cells was recently reported [105]. Implication of the AFP promoter as a driving element for the eGFP and Pac cassettes possesses an additional advantage for future therapeutic applications. As the selected precursor-like cells are capable of further differentiation, and because the activity of the AFP promoter would decline to a low level in adult hepatocytes [106], the immunological response of a recipient against products of the transgenes integrated in transplanted cells would be diminished or prevented. For this reason, investigation of the post-transplantation fate of hepatocyte precursor-like cells selected in our culture system would be of a great interest. On the other hand, the AFP promoter-driven eGFP expression may be not sufficient for the long-term monitoring of transplanted cells just because of the above-mentioned downregulation of the promoter activity. To overcome this problem, the transgenic cells used for experimental transplantations should be labeled with a long-lived fluorescent dye (for example, with help of dye-based C dots [107]) or additionally transfected with a vector containing a reporter gene (e.g., lacZ) under control of a constitutive gene promoter.
The remarkable feature of the described culture system is an opportunity to generate and select the cells of interest in suspension culture, thereby enabling their large-scale production and making the whole processing feasible for a bioreactor.
|
CONCLUSION |
|---|
|
TopFootnotesAbstractIntroductionMaterials and MethodsResultsDiscussionConclusionDisclosure of Potential...AcknowledgmentsReferences |
|---|
|
DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST |
|---|
|
TopFootnotesAbstractIntroductionMaterials and MethodsResultsDiscussionConclusionDisclosure of Potential...AcknowledgmentsReferences |
|---|
|
ACKNOWLEDGMENTS |
|---|
|
TopFootnotesAbstractIntroductionMaterials and MethodsResultsDiscussionConclusionDisclosure of Potential...AcknowledgmentsReferences |
|---|
.). We thank Dr. Silke Schwengberg and Simone Wagner for valuable discussions; Dr. Andreas Ehlich for supplying the GAPDH primer sequences; Niculina Dalibor, Lubov Rochlina, Célestin Ndombasi Tona, Tamara Rotshteyn, Rebecca Dieterich, Peter Metzger, Josef Tenelsen, and Doris Erb for technical assistance; and Susanne Wood for revising the manuscript.
The current addresses of noncorresponding authors are as follows: T.L.: Medical Clinic and Policlinic III, Justus-Liebig University of Giessen and Marburg, Giessen, Germany; T..: Institute for Neurophysiology, Center of Physiology and Pathophysiology, University of Cologne, Cologne, Germany; R.G.B., Medical Clinic and Policlinic III, Justus-Liebig University of Giessen and Marburg, Giessen, Germany; H.B., Axiogenesis AG, Cologne, Germany; J.H.: Institute for Neurophysiology, Center of Physiology and Pathophysiology, University of Cologne, Cologne, Germany.
|
FOOTNOTES |
|---|
.: collection and assembly of data, data analysis and interpretation, manuscript writing; R.G.B.: financial support, administrative support; H.B.: administrative support, provision of study material; J.H.: administrative support, provision of study material, data analysis and interpretation, final approval of manuscript. |
REFERENCES |
|---|
|
TopFootnotesAbstractIntroductionMaterials and MethodsResultsDiscussionConclusionDisclosure of Potential...AcknowledgmentsReferences |
|---|
