Introduction
Human embryonic stem cells (hESCs) have the potential to undergo long term self-renewal
in vitro and to produce unlimited quantities of different cell types belonging to all three embryonic germ layers (
Thomson et al., 1998;
Reubinoff et al., 2000). Thus, hESCs are very valuable in developmental biology studies and will find many applications in regenerative medicine. Considerable efforts have been made to efficiently differentiate hESCs toward individual cell lineages, including ectodermal neurons and astrocytes (
Reubinoff et al., 2001;
Kato et al., 2006;
López-González et al., 2009), mesodermal cardiomyocytes and blood cells (
Chadwick et al., 2003;
Wang et al., 2005), and endodermal insulin-producing cells and liver cells (
Naujok et al., 2008;
Sharma et al., 2008). However, generation of specialized cell lineages from a mixture of differentiated hESCs remains technically challenging. Traditional methods for identifying target cells, such as immunohistochemistry and real-time polymerase chain reaction (RT-PCR), are time-consuming and result in a significant loss of viable cells. Fluorescent reporter genes under the control of tissue-specific promoters offer a method to identify and enrich specific hESC derivative cells, without the need for fixing or fractionation (
Giudice and Trounson, 2008). MultiSite gateway technology is an efficient, simple, and effective method for lentivector construction, which enables vectors containing different promoters, reporter genes, and selection markers to be prepared using a single recombination reaction (
Suter et al., 2006). Here, we construct modular lentivectors containing the constitutive promoter of
EF1α and three different tissue-specific promoters (
Tα1 of α-tubulin,
aP2 of adipocyte Protein 2, or
AFP of alpha fetoprotein) driving transcription of the
hrGFP (humanized
Renilla GFP) gene. The vectors are based on self-inactivating HIV-1 lentivirus (
Dull et al., 1998) and are successfully used to monitor hESC differentiation from H1 cells (
Thomson et al., 1998).
Materials and methods
Cell culture
H1 cells were maintained in an undifferentiated state on a layer of irradiated mouse embryonic fibroblasts (MEFs). The culture medium contained 80% (v/v) knock-out DMEM (Invitrogen, Carlsbad, CA, USA) supplemented with 20% (v/v) knockout serum replacement (Invitrogen), 0.1 mmol/L β mercaptoethanol (Sigma, St Louis, MO, USA), 1% (w/v) nonessential amino acids (Hyclone, Logan, UT, USA), 100 IU/mL penicillin (Hyclone), 100 μg/mL streptomycin (Hyclone), and 8 ng/mL human basic fibroblast growth factor (bFGF) (Chemicon, Temecula, CA, USA).
Vector construction
To generate entry vectors, we obtained the human
EF1α promoter, the rat
Tα1 promoter (both kindly provided by Dr. David M. SUTER, University of Geneva Medical School, Geneva, Switzerland) (
Suter et al., 2006), the rat
aP2 promoter (a kind gift of the Children’s Hospital Oakland Research Institute, Oakland, CA, USA), and the mouse
AFP promoter (courteously provided by Dr. Sai-Kiang KIM, National University of Singapore, Singapore) (
Yin et al., 2002), all flanked with
attB sites (Table 1). Promoter PCR products were cloned into pDONR
TMP4–P1R (Invitrogen) (Fig. S1a) utilizing the Gateway BP recombination method following the manufacturer’s instructions. The
hrGFP gene was cloned into pDONR
TM221 (Invitrogen) (Fig. S1b), using the same method. The resulting vectors, termed pUp-
EF1α, pUp-
Tα1, pUp-
aP2, pUp-
AFP, and pDown-
hrGFP, were next recombined into the pDest
puro vector (
Li et al., 2010) employing a recognized LR recombination reaction protocol using the Gateway LR kit and a clonase enzyme mix (Invitrogen) (Fig. S1c). The final lentiviral expression vectors were termed pLV/Final-puro-EF1α-hrGFP (
EF1α-
hrGFP), pLV/Final-puro-Tα1-hrGFP (
Tα1-
hrGFP), pLV/Final-puro-aP2-hrGFP (
aP2-
hrGFP), and pLV/Final-puro-AFP-hrGFP (
AFP-
hrGFP) (Fig. 1).
Lentivirus production and transduction of H1 cells
Lentiviral particles were prepared by transient cotransfection of 293FT cells with EF1α-hrGFP, Tα1-hrGFP, aP2-hrGFP, or AFP-hrGFP constructs together with pRSV-REV, pCMV-VSVG, and pMDL-G/P-RRE, using Lipofectamine 2000. Three days after transfection, supernatants containing viral particles were harvested and filtered through 0.45 μm-pore-sized polyethersulfone membranes and concentrated by ultracentrifugation (at 75 000 g for 90 min at 4°C).
Prior to transduction, undifferentiated H1 cells were washed twice with phosphate buffered saline (PBS) and dissociated into single cells by incubation with 0.25% (w/v) TrypLE™ Select (Invitrogen) for 2-3 min at 37°C. A selective Rho-associated kinase (ROCK) inhibitor, Y-27632, was added to the culture medium to enhance survival of dissociated ES cells before and after trypsinization, as described elsewhere (
Watanabe et al., 2007). Next, dissociated cells were plated onto a fresh MEF feeder layer, together with lentiviral particles. Twelve hours after infection, the medium was changed and cells were incubated at 37° for a further 12 h. Seven days after transduction, puromycin (1-2 μg/mL) was added to the culture medium and maintained at this level for a further 7 days. Single colonies were next picked for analysis.
Fluorescence-activated cell sorting
The level of green fluorescent protein (GFP) expression at day 7 in untransduced and transduced EF1α-hrGFP H1 cells was determined by flow cytometry. Cell sheets were mechanically separated from MEFs and dissociated into single cells using TrypLE™ Select, as described above. Samples were analyzed by collection of 10 000 events using Cell-Quest® software (BD Biosciences, San Diego, CA). Untransduced cells served as negative controls.
Teratoma formation by EF1α-hrGFP H1 cells
About (1-2)×107 undifferentiated fluorescent H1 cells were collected and injected subcutaneously into the inguinal groove of 6 week-old nude mice. After 8 weeks, tumors were removed, fixed in 4% (v/v) paraformaldehyde, dissected, and subjected to histological examination after staining with hematoxylin and eosin. All animal experimental procedures were approved by the Animal Ethics Committee of Sun Yat-sen University, Guangzhou, China.
Committed differentiation of transduced H1 cells
To assess neural differentiation,
Tα1-
hrGFP H1 cells at 80% confluence were dissociated (using 1 mg/mL collagenase type IV [Invitrogen]) into small clumps. The clumps were next cultured in Petri dishes to permit formation of embryoid bodies (EBs). Culture was performed over 7 days in a medium containing 20% (v/v) knockout serum replacement, without bFGF, and EBs were next plated on polyornithine- and fibronectin-coated six-well plates (Sigma) for 20 days under neural progenitor medium (NPM) (
Zhang et al., 2001) supplemented with 20 ng/mL bFGF and 20 ng/mL epidermal growth factor (EGF) (Chemicon).
To measure adipogenic differentiation, dissociated
aP2-hrGFP H1 cell clumps were cultured in Petri dishes to permit formation of EBs over 12 days using a culture medium containing 20% (v/v) fetal bovine serum (FBS) (Hyclone), without bFGF. EBs were next plated on 2% (w/v) gelatin-coated six-well plates for 18 days, under an adipogenesis-inducing medium (AIM) containing 1 mol/L dexamethasone, 0.2 mmol/L indomethacin, 0.5 mmol/L 3-isobutyl-1-methylxanthine (IBMX), 0.01 mg/mL insulin, and 10% (v/v) FBS (all supplements were from Sigma) in DMEM (
Yu et al., 2008).
To assess hepatic differentiation of
AFP-hrGFP H1 cells, a differentiation protocol similar to that previously reported (
Agarwal et al., 2008) was used; minor modifications were made. In brief, when H1 cells reached 50%-70% confluence, the enriched natural seawater medium was changed to RPMI 1640 medium (Invitrogen) containing 100 ng/mL human activin A (R&D Systems, Minneapolis, MN), and culture continued for a further 5 days to allow endoderm progenitors to develop. Next, endoderm cells were allowed to differentiate into hepatic-like cells upon sequential addition of induction factors, including FGF4, hepatocyte growth factor (HGF), Oncostatin M (all from Peprotech, Rocky Hill, NJ) and dexamethasone. Differentiation continued for a further 15 days.
Alkaline phosphatase assay and immunocytochemistry
Alkaline phosphatase activity was assessed by histochemical staining. Transduced EF1α-hrGFP H1 cells were fixed in 4% (v/v) paraformaldehyde at room temperature for 20 min, washed with PBS, and stained using the alkaline phosphatase substrate BCIP/NBT (Sigma) for 20-30 min. Color reactions within cells were examined by light microscopy.
For immunocytochemical assessment, differentiated cells were fixed in 4% (v/v) paraformaldehyde for 20 min and next washed twice with PBS. Non-specific antibody binding was blocked using a mixture of goat serum and 1% (w/v) bovine serum albumin (Sigma), and cells were next incubated with primary antibodies directed against Oct-4 (1∶100; Santa Cruz Biochemicals, Santa Cruz, CA), SSEA-4 (1∶200; DSHB, Iowa City, IO), TRA-1-60 (1∶200; DSHB), Nestin (1∶100; Abcam, Cambridge, UK), βIII-tubulin (Tuj-1, 1∶200; R&D Systems), Desmin (1∶200; Neomarker, Fremont, CA), Sox 17 (1∶50; R&D Systems), or AFP (1∶100; R&D Systems), overnight at 4°C. Incubation with secondary antibody proceeded at room temperature for 50 min in the dark: goat Cy3-conjugated anti-mouse antibody (1∶200 dilution; Jackson, West Grove, PA); goat R-phycoerythrin-conjugated anti-mouse antibody (1∶200; Southern Biotech, Birmingham, AL); or donkey Cy3-conjugated anti-goat antibody (1∶200; Jackson), was employed. Nuclei were counterstained with Hoechst 33342 (Sigma). MEFs were used as negative controls.
Oil Red O staining
Cells were washed twice with PBS and fixed with 4% (v/v) formaldehyde in PBS for 1 h at room temperature. Cells were next stained for 1 h at room temperature with filtered Oil Red O solution (0.3% [v/v] Oil Red O in 60% [v/v] isopropanol), washed twice with distilled water, examined using light microscopy and photographed (
Yu et al., 2008).
Periodic Acid-Schiff staining
Periodic acid-Schiff (PAS) is primarily used to identify glycogen in cells and tissues. Hepatically differentiated H1 cells were fixed in 4% (v/v) paraformaldehyde for 15 min and stained with PAS (Sigma) at room temperature. Briefly, fixed cells were oxidized with 1% (v/v) periodic acid for 5 min and next rinsed three times in PBS. Cells were exposed to Schiff’s reagent for 15 min. After rinsed with PBS, cells were stained with Mayer’s hematoxylin for 1 min (
Baharvand et al., 2006).
Results
Generation and characterization of EF1α-hrGFPH1 cells
We first transduced H1 cells with the EF1α-hrGFP lentivector. Dissociated cells on an MEF feeder layer were exposed to culture medium supplemented with concentrated lentivirus preparations for 12 h. Seven days after transduction, the percentage of hrGFP-positive (hrGFP+) H1 cells was assessed both microscopically and by FACS (Fig. 2a). Transduction efficiency was approximately 85%. After antibiotic selection, purified hrGFP-positive H1 cells were attained and single colonies formed by such cells were picked (Fig. 2b). Expression of hrGFP was maintained during prolonged in vitro culture.
To examine the pluripotent potential of EF1α-hrGFP H1 cells, several assays were employed to assess the properties of stem cells transduced in this manner. Cell morphology was typical of H1 cells; the transformed cells had high levels of alkaline phosphatase activity (Fig. 2c) and continued to express pluripotency markers, including Oct4 (Fig. 2d), SSEA-4 (Fig. 2e), and TRA-1-60 (Fig. 2f). Fluorescence microscopy showed high-level hrGFP expression in EBs (Fig. 3a). Immunocytochemistry revealed that infected H1 cells gave rise to cells expressing Nestin (an ectodermal marker), Desmin (a mesodermal marker), and Sox17 (an endodermal marker) (Fig. 3b-3d) in vitro, and that transgene expression was sustained. An in vivo differentiation assay showed that transduced H1 cells formed teratomas 8 weeks after implantation. Frozen section analysis revealed that cells inside tumors were genotypically hrGFP+ (Fig. 3e). Histological analysis of teratomas showed that the tumors contained derivatives of all three germ layers (Fig. 3f-h), thus including ectoderm-derived neural-lineage cells (Fig. 3f), mesoderm-derived smooth muscle cells (Fig. 3g), and endoderm-derived glandular epithelium (Fig. 3h).
Neural differentiation of Tα1-hrGFP H1 cells
To initiate neural differentiation, ES cell colonies were picked and grown in suspension to form EBs over 7 days. EBs were then plated on polyornithine- and fibronectin-coated six-well plates under the N2B27 medium containing both bFGF and EGF. Plated EBs generated outgrowths of flattened cells, although small elongated cells were evident in the centers of differentiated EBs. Expression of hrGFP was first detected in a few cells of the EBs, commencing at day 5 (Fig. 4a). Fluorescence microscopy allowed estimation of the percentage of cells expressing hrGFP; protein expression levels increased in a time-dependent manner. On day 25, hrGFP-expressing cells with neuroepithelial characteristics were found in the centers of most attached EBs (Fig. 4a). We next used immunocytochemistry to assess the expression levels of neural-specific markers in fluorescent cells. Only a low proportion of hrGFP-positive cells coexpressed Nestin (Fig. 4b), and about half such cells were βIII-tubulin-positive (Fig. 4c).
Adipogenic differentiation of aP2-hrGFP H1 cells
To evaluate the adipogenic potential of aP2-hrGFP H1 cells, ES cell colonies were subjected to digestion by collagenase type IV, and cells released by this method were next placed into bacterial culture dishes that did not allow for cell adherence and permitted to aggregate in medium with 20% (v/v) FBS, but without bFGF for 12 days. Next, EBs were induced to differentiate into adipocytes (as assessed by attachment to 2% [w/v] gelatin-coated wells of six-well plates over 18 days) using an adipogenic medium containing dexamethasone, indomethacin, IBMX, and insulin. Differentiated cells spread rapidly from the EBs, and hrGFP-positive cells without obvious cytoplasmic droplets first appeared 8 days after EB attachment (thus on day 20) (Fig. 4d). On day 30 of differentiation, about 10%-20% of hrGFP-positive cells, containing small cytoplasmic lipid droplets, were observed to be spreading outward from attached EBs (Fig. 4d). Oil Red O staining showed high-level lipid accumulation was inside such fluorescently differentiated cells (Fig. 4e).
Hepatic differentiation of AFP-hrGFP H1 cells
A monoculture protocol was employed to induce hepatic differentiation of AFP-hrGFP H1 cells. The induction factors activin A (stage 1; day 2-day 5), FGF4, HGF (stage 2, day 6-day 10), oncostatin M, and dexamethasone (stage 3, day 11-day 20) were sequentially added to the induction medium. Cell death was obvious during the first 3 days of differentiation. In stages 2 and 3, differentiated cells proliferated quickly and hrGFP gene expression was examined every day, as assessed by fluorescence microscopy. Some differentiated AFP-hrGFP H1 cells expressed the hrGFP protein at day 12, and the percentage of cells expressing hrGFP increased with extended culture time (Fig. 4f). Cells that were hrGFP+ in genotype tended to clump and were similar in size and morphology at day 20. Immunochemical analysis showed that all hrGFP+ cells expressed AFP (Fig. 4g), thus demonstrating excellent agreement between hrGFP+ status and immunostaining data. Moreover, differentiated cells showed high-level glycogen storage activity, reminiscent of the behavior of functional hepatocytes (Fig. 4h).
Discussion
In the present study, we successfully transduced H1 cells with HIV-1-based lentiviral vectors. After antibiotic selection, we obtained EF1α-hrGFP+ H1 cells constitutively expressing hrGFP. After long term in vitro culture, transduced cells maintained characteristics distinctive of hESCs, including continuous expression of the stem cell markers, AKP, Oct4, SSEA-4, and TRA-1-60; the ability to form embryoid bodies in vitro; and the capacity for teratoma formation in vivo with expression of derivatives of the three forms of embryonic germ cells. Transgene expression was well maintained both during prolonged culture and throughout differentiation. These results indicate that the transduced H1 cells retained both self-renewal capacity and pluripotency.
Tissue-specific promoters driving reporter gene expression have been used to monitor the differentiation status of hESCs and to purify cells of a particular lineage (
Wang et al., 2000;
Aubert et al., 2003;
Suter et al., 2006). The
Tα1 α-tubulin gene is a member of a multigene family, abundantly expressed in developing neurons and likely gives rise to most of the tubulin protein required for neuronal growth (
Miller et al., 1987;
Gloster et al., 1999).
Tα1 expression is high in all neurons during development, the expression is downregulated after target contact, and is re-induced after axonal injury (
Gloster et al., 1999). In the present study, we found that
Tα1-positive neuronal precursors occasionally expressed nestin. This suggests that
Tα1 promoter activity partially overlapped with that of the nestin promoter, as indicated in previous reports (
Wang et al., 2000;
Suter et al., 2006). An earlier study (
Suter et al., 2006) used a
Tα1-promoter-driven eGFP-expressing lentivector to monitor neural differentiation in mouse embryonic stem cells.
Tα1 activity was evident in most βIII-tubulin-positive cells. However, in our present study, only 40%-50% of
Tα1-
hrGFP+ cells expressed βIII-tubulin. This between-study difference may be due to human/mouse genetic variation.
The
aP2 gene is not only a specific marker of mature adipocytes, it is also a target of peroxisome proliferator-activated receptor γ (PPAR-γ)(
Tontonoz et al., 1995). The interplay between PPARγ, a central regulator of adipogenesis, and
aP2 gene expression has been well studied, and it has been suggested that an
aP2 promoter-driven reporter gene assay would be optimally rapid and sensitive when adipogenesis agents are to be detected (
Rival et al., 2004).
AFP is expressed at high levels in the embryonic yolk sac and fetal liver; however, the protein levels gradually decrease after birth (
Kwon et al., 2006).
AFP expression is reactivated during liver regeneration or tumorigenesis. Thus, assessment of AFP level should reliably aid the study of tissue- and developmental stage-specific regulation of gene expression (
Spear, 1999;
Kwon et al., 2006). In the present study, we found that
aP2-driven and
AFP-controlled hrGFP concentrations correlated precisely with endogenous gene expression levels, as revealed by immunocytochemistry and functional assays.
Here, we utilized a novel fluorescent reporter gene, thus differing from the commonly employed reporters encoding β-galactosidase, luciferase, or alkaline phosphatase. Fluorescent reporters permit real-time noninvasive observation of cells and should be preferred over other markers because signals can be detected without the need for cell fixation or exposure of cells to chromogenic substrates. Moreover, cells expressing fluorescent markers can be purified by FACS. Such markers can also be used to track the behavior of transplanted cells
in vivo (
Giudice and Trounson, 2008). To date, we have constructed entry and destination vectors containing over 30 promoters (including those of the genes encoding
PDX1,
albumin,
αMHC, and
NKX2.5), using eight reporters (e.g.,
hrGFP,
dsRed E2,
CyPet, and
yPet) and four antibiotic-selection markers (mediating resistance to puromycin, neomycin, blasticidin, and hygromycin).
In conclusion, lentivectors containing tissue-specific promoter-driven fluorescent reporter genes, constructed using MultiSite gateway technology, are versatile tools enabling cell tracing and high-throughput screening that seeks to identify cell fate-determining molecules. This novel approach will be valuable in the study of gene function in human embryogenesis and in the generation of unlimited amounts of pure cell types for transplantation studies.
Higher Education Press and Springer-Verlag Berlin Heidelberg
PDF
(694KB)
