The authors have read the journal’s policy and have the following conflicts: CTC owns stock in Becton, Dickinson and Company. This does not alter the authors’ adherence to all the PLoS ONE policies on sharing data and materials.
Conceived and designed the experiments: SH LS BR MJT. Performed the experiments: SH EYT MJT BL. Analyzed the data: SH EYT SMC LS BR MJT BL. Contributed reagents/materials/analysis tools: SH SW ME SMC CTC BR. Wrote the paper: SH LS BR MJT.
Embryonic stem cells (ESCs) and induced pluripotent stem cells (iPSCs) offer great promise in regenerative medicine and disease modeling due to their unlimited self-renewal and broad differentiation capacity. There is evidence that the growth properties and critical signaling pathways differ between murine and human ESCs; therefore, it is essential to perform functional studies to test the putatively conserved mechanisms of pluripotent stem cell self-renewal between species. Previously, we identified the transcription factor Zfx as a key regulator of self-renewal in murine ESCs. Here we extend those findings to human ESCs. ZFX knockdown in hESCs hindered clonal growth and decreased colony size after serial replating. ZFX overexpression enhanced clone formation in the presence of Y-27632, increased colony size at low density and decreased expression of differentiation-related genes in human ESCs. ZFX-overexpressing hESCs resisted spontaneous differentiation but could be directed to differentiate into endodermal and neural cell fates when provided with the appropriate cues. Thus, ZFX acts as a molecular rheostat regulating the balance between self-renewal and differentiation in hESCs, revealing the close evolutionary conservation of the self-renewal mechanisms in murine and human ESCs.
Embryonic stem cells (ESCs) and the related induced pluripotent stem cells (iPSCs) are unique cells capable of giving rise to all tissues of the adult organism. These pluripotent stem cells (PSCs) can be exponentially expanded in culture while retaining their differentiation potential. The traits of pluripotency and continuous self-renewal underlie the value of PSCs as a potential source for cell replacement therapies and disease modeling, as well as a tool to study normal human development
Although ESCs from different species share the same key properties of pluripotency and self-renewal, major differences were found between murine (mESCs) and human ESCs (hESCs) including expression of different sets surface markers and distinct growth factor requirements
Previously, we demonstrated a role for the transcription factor Zfx in the self-renewal of mESC and adult hematopoietic stem cells
Here, we used a genetic approach to analyze the role of ZFX in hESC self-renewal and differentiation. Lentiviral shRNA knockdown of ZFX impaired self-renewal of hESCs. A novel bacterial artifical chromosome (BAC) transgenic strategy
H9 (WA-09) hESCs were maintained on MEFs (GlobalStem, Inc., Rockville, MD) before feeder-free expansion on Matrigel with conditioned media and 10 ng/ml FGF2 (3–5 days). To prepare cells for transduction, hESCs were exposed to Accutase (Innovative Cell Technologies, Inc., San Diego, CA) to create a single-celled suspension before plating at 18,000 cells/cm2 in the presence of 10 µM Y-27632. The following day, cells were transduced at a multiplicity of infection of 1 and 0.1. For each knockdown, phenotypes were observed at both multiplicities relative to the scrambled control. Puromycin selection began two days after transduction and no mock transduced cells remained two days after selection began. At this time, 5,000 puromycin resistant cells were replated per condition and were expanded in conditioned media +10 ng/ml FGF2 for seven days before crystal violet staining.
Cells were transduced at a multiplicity of ∼0.1 as described above and puromycin selection was applied as described. Cells were fixed in 4% paraformaldehyde for 15 minutes before rinsing four times with PBS. Fixed cells were blocked/permeabilized with 1× Perm/wash buffer (BD Biosciences, San Jose, CA) for 20 minutes before exposure to ZFX primary antibody overnight at 4°C. The next day, cells were washed three times in Perm/wash buffer before exposure to secondary antibodies conjugated to Alexa 488 and Hoechst 33258. Cells were washed three times after the hour, and PBS was added to the cells for imaging.
Imaging was performed on an Operetta High Content Screening System. The Hoechst was exposed for 20 mSec and the Alexa 488 signal was captured by a 300 mSec exposure using 50% intensity from the light source. Harmony software version 3.0 was used to identify cell nuclei (Hoechst signal) and quantitated the Alexa 488 pixel intensity (ZFX protein) for each cell nucleus. The mean pixel intensity from 50 cell fields per sample were imaged from 3 independent experiments for quantitation. All knockdowns were significantly different from the scrambled control (p<0.0001 for pairwise t-tests). 11,803 (Scr), 10,296 (Z2), 4,765 (Z3) and 6,905 (Z4) nuclei were measured in these experiments.
Human ESCs were maintained as previously described
Real-time PCR analysis was performed using the following QuantiTect primers: POU5F1, CXCR4, PAX6, GAPDH and ACTN2. Other gene expression levels were determined by qPCR using the following primers:
ZFX
F:
R:
EOMES
F:
R:
ID2
F:
R:
VIM
F:
R:
Septin 5
F:
R:
RNA was isolated using Trizol (Invitrogen, #15596-018) and reversed transcribed with the QuantiTect Reverse Transcription kit (Qiagen, #205313). SYBR green PCR reactions were performed using the PerfeCta SYBR Green SuperMix (Quanta Bioscience, #95054-500) on an Eppendorf Mastercycler Realplex2.
Direct immunofluorescence was performed using SSEA-4 Alexa Fluor® 488 (BD Pharmingen, #560308, La Jolla, CA), TRA1-81 Alexa Fluor® 555 (BD Pharmingen, #560123, La Jolla, CA), and Oct3/4 Alexa Fluor® 647 (BD Pharmingen, #560307, La Jolla, CA) according to the manufacturers recommendations.
Protein samples of H9 hESC and ZFXOver hESC clones were prepared in RIPA buffer plus protease inhibitors (Roche Diagnostics Ltd.). Protein concentration was determined using Bradford assay and 30 ug or 15 ug of each sample were separated by 10% polyacrylamide gel. Gels were transferred to PVDF membranes using the Bio-Rad mini-gel transfer apparatus. Membranes were blocked with 3% milk for 1 hour at room temperature and probed with affinity-purified polyclonal anti-Zfx antibody (in 3% Milk, 1∶1000 overnight at 4°C). After washing 3X with TBST, membranes were next incubated with peroxidase-labeled goat anti-rabbit IgG (1∶10,000 in 3% milk) for 1 hour at room temperature before visualization with ECL.
Human ESCs were expanded on Matrigel-coated dishes in standard human ESC media containing 6 ng/ml FGF2 without conditioned media. After seven days, single-celled suspensions were made by dissociating cells with Accutase (Innovative Cell Technologies, Inc., San Diego, CA) for 45 minutes before being washed twice with human ESC media. Aliquots of the cells were stained with anti-SSEA-3 (SSEA-3 Alexa Fluor® 647; unpublished, BD Pharmingen, La Jolla, CA) and anti-SSEA-1 antibodies (SSEA-1 Alexa Fluor® 647; BD Pharmingen #560120, La Jolla, CA) before quantitation on a FACSAria. For this assay, a number of unrelated BAC transgenic cell lines were used as normal clone controls (see
Figure | Control | Source |
2 | H9 and ZFXNormal | This manuscript |
3 | H9 and ZFXNormal | This manuscript |
4 | H9, DLL1::GFPc277 and c281, HES5::GFP c10 and ID1::YFPc2 | Placantonakis et al., 2009 and James et al., 2010 |
5 | H9, ID1::YFPc2 and DLL1::GFPc277 | Placantonakis et al., 2009 and James et al., 2010 |
6 | H9 and ZFXNormal | This manuscript |
S1 | H9 and ZFXNormal | This manuscript |
S2 | H9 and ZFXNormal | This manuscript |
S3 | H9 and ZFXNormal | This manuscript |
S4 | H9 and ZFXNormal | This manuscript |
S5 | H9, DLL1::GFPc277 and c281, HES5::GFP c10 and ID1::YFPc2 | Placantonakis et al., 2009 and James et al., 2010 |
For colony-forming cell assays, human ESC were dissociated with Accutase (Innovative Cell Technologies, Inc., San Diego, CA) for 45 minutes. Dissociated cells were washed twice with human ESC media. Serial dilutions of hESCs were plated onto Matrigel-coated 6 well dishes in conditioned media with 10 ng/ml FGF2. Cells were exposed to Y-27632 for 24 hours after plating, and colonies were visualized after 7 days by staining with crystal violet. The data presented are the number of colonies from the 1∶100 dilution and represent the number of colonies derived from 2,666 cells in 9.6 cm2.
Crystal violet staining was used because it is easier and cheaper and works as well for determining the number of colony forming units in a given culture.
To direct human ESCs to endoderm, we used a previously published protocol
RNA was isolated from the Tra1-81HI/SSEA-3HI fraction of ZFXOver1,2, ZFXNormal and H9 hESCs using Trizol (Invitrogen). Samples were labeled and hybridized to Illumina human 6 oligonucleotide arrays. Normalization and model-based expression measurements were performed using the Illumina analysis package (LUMI) available through open-source Bioconductor project (
Statistical analysis was performed using Prism for Mac version 5.0a. Unpaired, two-tailed T tests were performed comparing control cell lines versus ZFXOver clones.
Sub-confluent cultures were treated with 0.1 ug/ml Colcemid (Karyomax, Invitrogen) for 60–90 minutes before harvesting according to standard cytogenetics procedures. Briefly, cells were trypsinized to a single cell suspension, pelleted at 250 g for 5 minutes and resuspended in warm 0.075M KCl. After 8 minutes incubation at 37°C, the hypotonic was diluted with approximately 1/4 volume of 3∶1 methanol/glacial acetic acid fixative, gently mixed, and the cells pelleted as before. The supernatant was removed and the cell pellet loosened by gently flicking the base of the tube. The cells were then fixed in 3 changes of fixative. Fixed cell suspensions were stored at −20°C. Fixed metaphase preparations were dropped onto dry slides and the quality of spreading assessed by phase microscopy. Slides were then air-dried and aged at 37°C overnight.
DNA for human BAC clone RP11-1107D4 (BACPAC Resources, Children’s Hospital and Research Center at Oakland), spanning the ZFX locus, was labeled by nick translation with Red dUTP (Enzo Life Sciences Inc., supplied by Abbott Molecular Inc.), and hybridized to metaphase slides. Briefly, approximately 100 ng of labeled probe and 1 ug human Cot-1 DNA (Invitrogen) was mixed with hybridization buffer (50% formamide, 2×SSC,10% dextran sulfate, 0.1% SDS, 1×Denhardt’s solution, 40 mM sodium phosphate buffer, pH7) and applied to each slide before sealing under a coverslip with rubber cement. The slides were then placed in a HYBrite™ (Abbott Molecular), denatured at 72°C for 3°minutes, and then hybridized at 37°C overnight. After coverslip removal in 2×SCC, 0.1% Igepal CA 630 at room temperature, the slides were washed in 0.4×SSC, 0.3% Igepal at 73°C for 2°minutes, then in 2×SCC, 0.1% Igepal at RT, and rinsed briefly in 2×SSC. The slides were then stained in 0.08 µg/ml DAPI in 2×SSC for 3 minutes, rinsed, air-dried, then mounted in antifade solution (Vectashield, Vector Labs), and stored at 4°C. Slides were scanned using a Zeiss Axioplan 2i epifluorescence microscope equipped with a megapixel CCD camera (CV-M4+CL, JAI) controlled by Isis 5.2 imaging software (Metasystems Group Inc, Waltham, MA). At least 10 metaphases and 5 interphase nuclei were examined for each preparation.
Zfx is essential for the self-renewal of mouse ESCs and hematopoietic stem cells. To test whether ZFX is required in human ESCs, we used a lentiviral vector system to efficiently and stably knockdown (KD) ZFX in hESCs. To characterize the extent of ZFX knockdown, initial experiments were performed in leukemia cell lines where three of the five shRNAs tested caused a strong knockdown of ZFX expression levels and a concomitant growth impairment (data not shown). H9 hESCs cells were transduced with these 3 ZFX knockdown viruses before puromycin selection. All 3 of the vetted ZFX knockdowns caused a marked decrease in the number and size of colonies relative to the scrambled shRNA control (
Human ESCs were transduced with ZFX knockdown lentiviral constructs and a scrambled control before clonal replating. (A) Live cell images and (B) and the entire well stained with crystal violet seven days after replating. (C) ZFX immunofluorescence after knockdown in hESCs, and (D) ZFX quantitative immunofluorescence analysis. Each dot is the average pixel intensity of nuclear ZFX protein averaged from all cells in one microscopic field. The average pixel intensity from 50 microscopic fields derived from 3 independent experiments is shown. The crosshairs and whiskers represent the mean and SEM. All knockdowns were significantly different from the scrambled control (p<0.0001 for t-tests of each knockdown versus scrambled). 11,803 (Scr), 10,296 (Z2), 4,765 (Z3) and 6,905 (Z4) nuclei were measured in these experiments. (E) ZFX RNA levels as measured by quantitative PCR after knockdown.
Similar to many transcription factors, the overexpression of ZFX cDNA from a heterologous promoter is toxic to cells (B.R., unpublished data). To overexpress murine Zfx under its native regulation, the entire genomic Zfx locus has been introduced into mESC as a bacterial artificial chromosome (BAC) transgene
As shown in
A. Normalized
Under standard growth conditions, all hESC clones had similar morphologies, growth and apoptosis rates as the parental H9 hESC line (
A. ZFXOver clones, ZFXNormal and H9 hESCs (together grouped as controls) were dissociated into single cells before clonal replating. Cells were expanded for 10 days before fixation and staining with crystal violet. B. Colony counts between ZFXOver and control hESCs in 3 independent experiments with error bars representing the S.E.M.
Next we tested the growth of ZFXOver clones under conditions suboptimal for hESC self-renewal. ZFXOver and control hESC lines were cultured in the absence of conditioned media or feeder cells to promote spontaneous differentiation, and the percentage of cells expressing SSEA-3 and SSEA-1 was measured to quantify the ratio of undifferentiated versus differentiated cells, respectively. In control experiments, there was no statistical difference in marker expression when conditioned media was used to expand cells without feeders (data not shown). However, in suboptimal conditions, H9 and the control clones showed strong signs of differentiation while the ZFXOver cells resisted spontaneous differentiation (
A. ZFXOver clones and controls were expanded in conditions promoting self-renewal before SSEA-3 and SSEA-1 FACS analysis. B. ZFXOver clones and controls were expanded in suboptimal conditions before SSEA-3 and SSEA-1 FACS analysis. The quantitation compares ZFXOver clones to controls in three independent experiments and error bars represent the S.E.M. See
Reciprocally, H9 and control clones expressed high levels of SSEA-1, a marker of differentiated cells (combined mean 80.05±1.14% SEM, n = 13; p<0.0001) compared to ZFXOver1 (32.50±8.98 AU SEM, n = 3; p<0.0001) and ZFXOver2 (39.23±11.33 AU SEM, n = 3; p<0.0001). The mean fluorescent intensity of SSEA-1 expression was also significantly higher in controls (11,326±1008 AU SEM, n = 13) compared to ZFXOver1 (760.7±188 AU SEM, n = 3; p = 0.002) and ZFXOver2 (772.7±211.2 AU SEM, n = 3; p = 0.0002). Taken together, these data show that ZFX overexpression promotes self-renewal of hESCs and inhibits differentiation in suboptimal culture conditions.
To test whether higher ZFX levels globally prohibit ESC differentiation, ZFXOver clones, two BAC transgenic control clones (ID1::YFPc2 and Dll1::GFPc277) and H9 were directed into endoderm or neural tissue and we examined lineage marker expression during differentiation using quantitative RT-PCR. We detected no significant difference in the loss of Nanog expression or gain in CXCR4 expression in ZFXOver clones during endodermal differentiation (
ZFXOver clones, two control clones that express normal ZFX levels (ID1::YFPc2, Dll1::GFPc277) and H9 were directed to endoderm or neural cells, and the level of Nanog, Pax6 (neural) and CXCR4 (endoderm) mRNA at each time point was measured by quantitative PCR.
To gain insight into the genome-wide expression changes due to ZFX overexpression, we performed microarray analysis in ZFXOver hESCs. In order to avoid any systematic bias in the gene expression analysis due to varying levels of spontaneous differentiation in ZFXOver versus controls, we isolated the most undifferentiated (Tra1-81HI/SSEA-3HI) hESCs from each line. Hierarchical clustering analysis of the microarray data (
Two independent samples of Tra1-81HI/SSEA-3HI hESCs were isolated from ZFXOver clones, ZFXNormal and H9 hESC array analysis. A. Dendrograms of each cell line after clustering analysis. B. Up- and down-regulated genes in ZFXOver compared to H9 using an adjusted p-value of 0.05 as a cutoff. C. Quantitative PCR validation of selected genes on the array.
In the current study we demonstrate a role for the transcription factor ZFX in modulating the self-renewal of hESC using gain- and loss-of-function approaches. ZFX reduction caused a loss of self-renewal while BAC-mediated ZFX overexpression increased the clonogenicity and decreased spontaneous differentiation of hESCs. Importantly, ZFX-overexpressing clones retained their ability to undergo differentiation in response to appropriate stimuli. The use of BAC transgenesis was critical to circumvent general toxic effects of ZFX overexpression observed using heterologous promoters. Gene expression driven by the endogenous gene locus in a BAC provides advantages over heterologous promoters, such as native gene regulation, reduced position effect
Because the extrinsic self-renewal signals, morphology and clonogenicity differ between human and mouse ESCs, it is critical to identify the self-renewal regulators that are conserved between the two species. Here we provide evidence for the functional conservation of ZFX, a critical member of the self-renewal transcriptional network. Our gain-of-function studies in hESCs are compatible with our previous work in mESCs demonstrating enhanced self-renewal and reduced spontaneous differentiation in both murine and human ESC. However, ZFX-overexpressing hESC underwent normal lineage-specific differentiation
The enhanced self-renewal observed in ZFX-overexpressing clones could reflect a reduction in the baseline heterogeneity of cultured ESC. This hypothesis is supported by the increased plating efficiency of ZFXOver clones and by a small but reproducible increase in the percentage of cells expressing undifferentiated hESC markers (data not shown). It is possible that high ZFX levels stabilize a chromatin conformation that favors self-renewal over differentiation. This ‘locked’ state could also explain the kinetic delay in neural differentiation observed in ZFX-overexpressing hESC. Alternatively, ZFX overexpression may convert hESC from their primed state into a less differentiated, naïve state characteristic of murine ESC. Indeed, many of the genes downregulated in ZFXOver clones are expressed in mouse epiblast stem cells isolated from the postimplantation embryo
(TIF)
(TIF)
(TIF)
(TIF)
(TIF)
We thank Margaret Leversha and Lei Zhang (Molecular Cytogenetics Core Lab at MSKCC) for the BAC FISH and karyotyping. We thank Lorraine Clark for the use of the Illumina Analyzer.