OCT4 is a master transcriptional regulator, which mediates pluripotency in ESCs through inhibition of tissue-specific and promotion of stem cell-specific genes. Suppression of OCT4, along with other regulators of pluripotency, such as SOX2 and NANOG, has been correlated with cell-fate specification and lineage-specific differentiation. Recent reports have shown the expression of OCT4 in adult MSCs but have not ascribed functional homology with ESCs. MSCs are mesoderm-derived cells, primarily resident in adult bone marrow, that undergo lineage-specific differentiation to generate specialized cells such as stroma, fat, bone, and cartilage. We have previously demonstrated the plasticity of MSCs through their ability to generate neuronal cells. Here, we show that OCT4 provides similar regulatory circuitries in human MSCs and ESCs, using chromatin immunoprecipitation-DNA selection and ligation technology and loss-of-function studies. MSCs were found to express the embryonic transcription factors OCT4, NANOG, and SOX2. In addition, OCT4 was found to (a) target similar genes in MSCs and ESCs, (b) promote the expression of MSC-specific genes, and (c) regulate MSC cell cycle progression. The results suggest similar regulatory mechanisms for OCT4 in MSCs and ESCs and have implications regarding MSC plasticity.
Disclosure of potential conflicts of interest is found at the end of this article.
The therapeutic utilization of stem cells in regenerative medicine holds vast potential for treating diseases and disorders with high mortality [1, , , –5]. ESCs generate tissues of all three germ layers when placed in an uncontrolled environment . This pluripotent nature of ESCs makes them attractive for stem cell therapy. However, ESCs come with inherent problems, mostly tumorigenicity .
Adult stem cells (ASCs) show reduced risk for tumorigenesis; however, their therapeutic potential remains questionable . Although ease in harvesting, particularly those in bone marrow (BM), makes ASCs appealing, the ability to generate cells of each germ layer is a subject of investigation. ASCs undergo lineage-specific differentiation to give rise to progeny within the same germ layer as the stem cell. Although there have been many reports investigating the transdifferentiation of ASCs, limited efforts have been made to ascribe the underlying molecular mechanisms responsible for such plasticity.
Mesenchymal stem cells (MSCs) are mesoderm-derived cells found in the fetus and the adult . In adults, the primary organ of residence for MSCs is the BM . MSCs show lineage-specific differentiation along adipogenic, chondrogenic, and osteogenic paths . MSCs have been reported to transdifferentiate into cells of ectodermal and endodermal tissue [11, , , , –16]. Given their plasticity and reduced risk of tumor formation, MSCs may show potential for controlled differentiation along multiple lineages in vivo. These properties make MSCs attractive candidates for regenerative therapies.
The plastic behavior of MSCs suggests some level of functional similarities with ESCs on their maintenance and differentiation potential. Another commonality between the two stem cells comes from recent reports showing expression of the embryonic transcription factor OCT4 in MSCs . OCT4 is expressed in ESCs, where it inhibits tissue-specific genes and enhances self-renewal and pluripotency . Specifically, OCT4 interacts with other embryonic regulators, such as SOX2 and NANOG, to oversee a vast regulatory network to maintain pluripotency and inhibit differentiation . Whether OCT4 demonstrates similar functions and regulatory circuitries in MSCs is a question that is investigated in this report. Here we report on the role of OCT4 in MSCs and compare this with ESCs using a novel approach that combines global gene target profiling and loss-of-function studies.
Materials and Methods
Reagents and Antibodies
Dulbecco's modified Eagle's medium (DMEM) with high glucose, DMEM/Ham's F-12 medium (F12), knockout serum replacement, nonessential amino acids, l-glutamine, l-alanyl-l-glutamine, and B-27 supplement were purchased from Gibco (Carlsbad, CA, http://www.invitrogen.com); fetal calf serum (FCS), β-mercaptoethanol, Ficoll-Hypaque and all-trans retinoic acid from Sigma-Aldrich (St. Louis, http://www.sigmaaldrich.com); defined FCS from Atlanta Biologicals (Lawrenceville, GA, http://www.atlantabio.com); propidium iodide from Calbiochem (San Diego, http://www.emdbiosciences.com); and 4′,6-diamidino-2-phenylindole, dilactate (DAPI) and Texas Red phalloidin from Molecular Probes Inc. (Carlsbad, CA, http://www.probes.invitrogen.com). Rabbit anti-OCT4, -SOX2, -NANOG, and fluorescein isothiocyanate (FITC)-goat anti-rabbit were purchased from Abcam (Cambridge, M.A., http://www.abcam.com); β-actin monoclonal antibody, horseradish peroxidase (HRP)-conjugated goat anti-rabbit IgG, and HRP goat anti-mouse IgG antibodies from Sigma-Aldrich; and goat anti-ribosomal protein L28 from Santa Cruz Biotechnology Inc. (Santa Cruz, CA, http://www.scbt.com).
Culture of Human MSCs
Human MSCs were cultured from BM aspirates as described . The use of human BM aspirates followed a protocol approved by the Institutional Review Board of The University of Medicine and Dentistry of New Jersey-Newark campus. Unfractionated BM aspirates (2 ml) were diluted in 12 ml of DMEM containing 10% defined FCS (Atlanta Biologicals), hereby referred to as D10 medium. The defined FCS is permissive to the expansion of MSCs with retention of differentiation potential . Cells were transferred to vacuum-gas plasma treated Falcon 3003 tissue culture plates (BD Biosciences, San Diego, http://www.bdbiosciences.com). Plates were incubated at 37°C, and at day 3, mononuclear cells were isolated by Ficoll-Hypaque density gradient and replaced in the culture plates. Fifty percent of the medium was replaced with fresh D10 medium at weekly intervals until the adherent cells were approximately 80% confluent. Cells were then trypsinized and subcultured at a ratio of 1:4. After four cell passages, the adherent cells were symmetric, CD14−, CD29+, CD44+, CD34−, CD45−, SH2+, prolyl-4-hydroxylase−, and capable of differentiating along osteogenic, adipogenic, and chondrogenic paths.
For telomerase assay, Western and Southern analyses, polymerase chain reaction (PCR), propidium iodide staining, and chromatin immunoprecipitation (ChIP), 105 MSCs were seeded in 100-mm Falcon 3003 tissue culture plates. For immunofluorescence and morphology/growth curve studies, 103 MSCs were seeded in 35-mm Falcon 3001 plates. All experiments used MSCs that had undergone 5–20 population doublings (passages 3–10) except for immunofluorescence and telomere studies, which examined cells at passages 3 and 25. The experiments were performed at <80% confluence to control for changes in gene expression due to contact inhibition. MSCs isolated from different BM donors were used to repeat all experiments.
Culture of Human ESCs
Propagation of human ESCs (hESCs) followed the protocol outlined by WiCell Research Institute . Briefly, murine embryonic fibroblasts (MEFs) were purchased from American Type Culture Collection (Manassas, VA, http://www.atcc.org) and grown in DMEM containing 15% FCS (Sigma-Aldrich) on six-well tissue culture plates (Falcon 353046; BD Biosciences). Confluent feeder cells (MEFs) were irradiated with 70 Gy, delivered by a Mark II Cesium Unit to halt proliferation while maintaining metabolic activity. The human ESC line BG0IV is among those listed on the NIH registry and was purchased from American Type Culture Collection . After 4 hours of irradiation, human ESCs were added to the feeder layer and propagated in complete hESC medium, which comprised DMEM/F12, 5% knockout serum replacement, 15% FCS, 4 ng/ml basic fibroblast growth factor, 0.1 mM β-mercaptoethanol, 0.1 mM nonessential amino acids, and 2.0 mM l-alanyl-l-glutamine. hESC medium was changed every 48 hours, and the cells were passaged in new feeder cultures at 2-week intervals. Overgrowth was avoided, since this would result in differentiation. At various time intervals, the cultures were tested for markers of ESCs. In addition, the cells were tested for pluripotency, based on the ability to form teratomas in SCID mice. The profile of hESCs was SSEA-1−, SSEA-4+, TRA-1–60+, TRA-1–81+.
Neuronal Induction of MSCs
At 20% confluence, MSCs subcultured in D10 medium were induced with neuronal induction medium (which was composed of DMEM/F12, 2% FCS (Sigma-Aldrich), B27 supplement, 20 μM retinoic acid, and 12.5 ng/ml basic fibroblast growth factor), and cells were cultured as previously described . In parallel studies, MSCs were cultured in medium alone with vehicle used for reconstitution of retinoic acid and growth factors, hereafter referred to as uninduced MSCs (day 0 [D0]). The data in all experiments were similar for D0 MSCs and those cultured in expansion medium (D10 medium), with the latter referred to as undifferentiated MSCs. Experimental endpoints for induced MSCs were 6 and 12 days (D6 and D12) of induction, which correspond to partially and fully differentiated neuronal phenotype, respectively [15, 16].
Telomerase activities from uninduced (D0) and induced (D6 and D12) MSCs were determined using the Telo TAGGG telomerase PCR ELISAPlus kit purchased from Roche Applied Sciences (Indianapolis, IN, http://www.roche-applied-science.com). The procedure followed the manufacturer's specified instructions. Briefly, cell-free supernatants were added to biotin-labeled synthetic primers for amplification by PCR. Amplified products, proportional to the cell's telomerase activity, were hybridized to digoxigenin-labeled detection probes and immobilized on streptavidin-coated plates. Telomerase activities were quantitated by chemiluminescence.
Semiquantitative Reverse Transcription-PCR and Quantitative PCR
Total RNA (2 μg) was reverse transcribed, and 200 ng of cDNA was used in PCR. Semiquantitative PCRs were normalized by amplifying the same sample of cDNA with primers specific for glyceraldehyde-3-phosphate dehydrogenase (GAPDH). The cycling profile for GAPDH (25 cycles) and all other primer sets (35 cycles) was as follows: 94°C for 30 seconds, 55°C for 30 seconds, and 72°C for 30 seconds, with a final extension at 72°C for 10 minutes. For quantitative PCR, the Platinum SYBR Green qPCR SuperMix-UDG Kit was used (Invitrogen, Carlsbad, CA, http://www.invitrogen.com). Quantitative PCRs (qPCRs) were normalized by amplifying the same sample of cDNA with primers specific for β-actin. qPCRs were performed with a 7500 Real Time PCR System (Applied Biosystems, Foster City, CA, http://www.appliedbiosystems.com). The cycling profile for real-time PCR (40 cycles) was as follows: 94°C for 15 seconds and 60°C for 45 seconds. Gene expression analysis was performed using the 7500 System SDS software (Applied Biosystems). Normalizations were performed with β-actin, and values were arbitrarily assigned a value of 1. Primer sequences and additional information are presented in supplemental online Table 1.
Suppression of OCT4 in MSCs
OCT4 was suppressed in MSCs using short interfering RNA (siRNA) duplexes with the following wild-type RNA sequence: (sense) 5′-cga gag gau uuu gag gcu g-3′ and (antisense) 5′-cag ccu caa aau ccu cuc g-3′. The OCT4 target (NM_002701) spans +937/+955, with the following sequence: 5′-cga gag gat ttt gag gct g-3′. Negative control siRNA sequence containing five nucleotide mutations, shown as underline/italics, was as follows: (sense) 5′-cua gau gcu cuu gau gcu g-3′ and (antisense) 5′-cag cau caa gag cau cua-g 3′. At 50% confluence, undifferentiated MSCs were transfected with siRNA duplexes using the siPORT NeoFX transfection reagent (Ambion, Austin, TX, http://www.ambion.com). After 48 hours, transfectants were assayed for morphology, cell proliferation, and gene expression.
Nuclear proteins were extracted with the Nxtract kit according to the manufacturer's specified guidelines (Sigma-Aldrich). Total protein was determined with a Bio-Rad (Hercules, CA, http://www.bio-rad.com) DC protein assay kit. Extracts (15 μg) were treated with protease inhibitor and analyzed using 4%–20% SDS-polyacrylamide gel electrophoresis precast gels (Bio-Rad). Proteins were transferred onto polyvinylidene difluoride membranes (PerkinElmer Life and Analytical Sciences, Boston, http://www.perkinelmer.com) and incubated overnight with primary antibodies. Detection was performed with HRP-conjugated IgG. Primary and secondary antibodies were used at dilutions of 1/1,000 and 1/2,000, respectively. Membranes were stripped with Restore Stripping Buffer (Pierce, Rockford, IL, http://www.piercenet.com) for reprobing with other antibodies. Cytoplasmic contamination of nuclear extracts was determined by reprobing the membranes with anti-ribosomal protein L28.
ChIP was performed using the ChIP-IT Enzymatic Kit from Active Motif, according to the manufacturer's specific guidelines. Briefly, chromatin from uninduced (D0) and induced (D6 and D12) MSCs was incubated overnight with 2 μg of anti-OCT4, anti-IgG, or anti-RNA polymerase II antibodies and precipitated with protein G-agarose. Enriched chromatin was analyzed by PCR using primers flanking the OCT4 binding sites within NANOG and SOX2 that were predicted using the MatInspector transcription factor analysis tool from Genomatix (Munich, Germany, http://www.genomatix.de) . The primers span +1,807/+1,950 (NC_000003) and +297/+445 (NC_000012) for SOX2 and NANOG, respectively, with the following sequences: SOX2 (forward), 5′-agt ctc caa gcg acg aaa-3′ and (reverse) 5′-tca gca aga agc ctc tcc-3′; NANOG, (forward) 5′-gat ttg tgg gcc tga aga-3′ and (reverse) 5′-ttg ttt gcc ttt ggg act-3′. Control chromatin enriched with RNA polymerase II and negative control IgG antibodies were analyzed by PCR using primers within the kit. Total chromatin served as the input loading control.
ChIP-DNA Selection and Ligation Assay
Chromatin was immunoprecipitated with ant-OCT4 from undifferentiated MSCs and ESCs by ChIP assay, as described above. The enriched chromatin fraction was processed by Aviva Systems Biology (San Diego, http://www.avivasysbio.com) for genome-wide microarray profiling as described previously [24, 25]. Briefly, input and anti-OCT4-enriched genomic DNA (gDNA) were randomly biotinylated and annealed to the DNA selection and ligation (DSL) oligo pool for Taq ligation and T3 and T7 PCR amplification. Input gDNA was labeled with Alexa Fluor 647 and anti-OCT4-enriched gDNA with Cy3. PCR products were hybridized to the 40-mer Hu20K array (Aviva Systems Biology), and slides were scanned on a GenPix4000B scanner (Axon Instruments/Molecular Devices Corp., Sunnyvale, CA, http://www.moleculardevices.com).
The single-error model was applied for data analysis, as previously described [26, 27]. The top 2% of gene promoters enriched with anti-OCT4 were classified based on physiological function and compared between MSCs and ESCs; n = 3.
Propidium Iodide Staining
MSCs transfected with wild-type or mutant OCT4 siRNA or untransfected were treated with RNase A (1 mg/ml) and fixed with cold 70% ethanol. Cells were stained with 20 μg/ml propidium iodide (PI) solution and transferred to round bottom tubes for DNA analysis by FACScan (BD Biosciences). All analyses were performed using CellQuest software (BD Biosciences), and percentage statistics are given.
Undifferentiated MSCs, subcultured for 3 or 25 passages (P3 and P25, respectively), were established on glass coverslips placed in 35-mm culture dishes. On the day of immunofluorescence labeling, cells were washed with phosphate-buffered saline (PBS) and then fixed with 3.7% formaldehyde and permeabilized in 1% Triton-X. Cells were incubated overnight at 4°C with rabbit anti-OCT4 at a concentration of 1/250 in 0.1% bovine serum albumin (BSA)/PBS. Primary antibodies were developed with secondary FITC-goat anti-rabbit at a concentration of 1/500 for 2 hours at room temperature. Cells labeled with FITC-IgG served as isotype controls. Following labeling, cell nuclei and cytoskeletons were counterstained with 300 nM DAPI and 6.6 μM Texas Red phalloidin diluted in 0.1% BSA/PBS, respectively. Coverslips were transferred to glass coverslides and examined on a three-color fluorescent microscope (Nikon Instruments Inc., Melville, NY, http://www.nikoninstruments.com).
Telomere Length Assay
Assessment of telomere length in early- and late-passage MSCs was performed using the TeloTAGGG Telomere Length Assay (Roche Applied Sciences), according to manufacturer's guidelines. Briefly, 2 μg of isolated gDNA was digested using restriction enzymes provided within the kit. Digested gDNA was then separated by gel electrophoresis and transferred by Southern blotting as previously described . Membranes were hybridized with telomere-specific digoxigenin (DIG)-labeled probes and detected by chemiluminescence using anti-DIG-alkaline phosphatase.
Statistical analyses were performed with analysis of variance and Tukey-Kramer multiple comparisons test. p < .05 was considered significant.
Telomerase Activity in MSCs
Elevated telomerase activity has been linked to stem cells, whereas decreased levels are consistent with maturation and terminal differentiation . Since MSCs express high levels of telomerase in vivo and have been reported to possess some telomerase activity in vitro, we compared activities with developing and mature neuronal cells derived from MSCs [30, 31]. We have previously shown that MSCs generate neuronal cells, which exhibit properties consistent with true neurons [15, 16]. The purpose of this question is to show that loss of stemness by MSCs correlates with lowered telomerase activity.
We examined timeline changes in telomerase activity beginning with uninduced cells (D0) and cells neuronally induced for 6 (D6) and 12 (D12) days [15, 16]. Our previous studies have shown high telomerase activity in undifferentiated MSCs . Uninduced MSCs showed significantly (p < .05) increased telomerase activities as compared with D6 and D12 induced cells (Fig. 1A). Parallel studies with ESCs yielded similar results (data not shown). In summary, the results show elevated telomerase activity in the undifferentiated MSCs as compared with cells undergoing differentiation.
Expression of Embryonic Transcription Factors in Uninduced MSCs
Since MSCs have been reported to generate cells of ectodermal and endodermal tissues, we next asked whether MSCs share other similarities with ESCs besides telomerase activity [11, , , , –16]. To address this question, we examined expression of the embryonic transcription factors OCT4, SOX2, and NANOG in uninduced MSCs. Induced MSCs served as a model to study the effects of differentiation on transcription factor expression.
By semiquantitative reverse transcription (RT)-PCR for OCT4, we observed strong band intensities in D0 cells, which became undetectable in D6 and D12 cells (Fig. 1B). The changes in OCT4 mRNA were next studied at the protein level by Western blots with nuclear extracts from D0, D6, and D12 MSCs. Dense bands were observed in D0 cells but were undetectable at D6 and D12 induction (Fig. 1C). We next assessed whether uninduced MSCs express SOX2 and NANOG. Dense bands were observed in D0 cells for SOX2 and NANOG mRNA and protein, with undetectable bands at D6 and D12 (Fig. 1D, 1E). Total RNA and nuclear extracts from ESCs served as positive control for RT-PCR and Western blot, respectively (Fig. 1B–1E). In summary, uninduced MSCs express the embryonic transcription factors OCT4, NANOG and SOX2.
OCT4 Interaction with NANOG and SOX2 in Uninduced MSCs
The expression of OCT4 in uninduced MSCs does not determine the status of OCT4 interaction with known target genes. To address this question, we performed ChIP assay with anti-OCT4 immunoprecipitated chromatin from uninduced (D0) and induced (D6 and D12) MSCs. The precipitated gDNA was subjected to RT-PCR with primers flanking the OCT4 binding regions within NANOG and SOX2. Dense bands were observed for the uninduced cells (Fig. 1F, top row) but were undetectable in the induced cells (Fig. 1F, middle and bottom rows). In summary, OCT4 was found to interact with the NANOG and SOX2 genes in uninduced MSCs. These findings are consistent with the regulatory actions of OCT4 in ESCs .
OCT4 Gene Targets in MSCs and ESCs
We next asked whether OCT4 regulates a similar profile of target genes in MSCs as observed in ESCs. To address this question we explored the genome-wide profile of OCT4 targets in MSCs and ESCs using conventional ChIP coupled with a DSL strategy. ChIP-DSL is more sensitive than traditional ChIP on chip, thus allowing for analysis of fewer cells . ChIP assay was performed with anti-OCT4 immunoprecipitated chromatin from undifferentiated MSCs and ESCs. The precipitated and input gDNA was then subjected to DSL technology and hybridized to the 40-mer Hu20K promoter array (Aviva Systems Biology). In parallel control studies, ChIP assays were performed with D12 induced cells and stromal fibroblasts. However, since these cells did not express OCT4, no chromatin was enriched for hybridization to the promoter array (data not shown). Tables 1 and 2 show the 50 top-ranked genes targeted by OCT4 in ESCs and MSCs, respectively.
Table Table 1.. Top-ranked genes targeted by OCT4 in human ESCs
Table Table 2.. Top-ranked genes targeted by OCT4 in human MSCs
We next compared similarities among the top 2% of genes (400 total genes for each stem cell type) enriched with anti-OCT4 in MSC and ESC samples, and have determined 104 common to both cell types (Fig. 2A). Table 3 shows the 50 top-ranked genes common to ESCs and MSCs. The 800 selected genes were next classified based on physiological function using the Entrez Gene algorithm and are presented as pie charts (Fig. 2B, 2C) . The results indicate similar profiles among classes of genes targeted by OCT4 in both stem cell types.
Table Table 3.. Top-ranked genes targeted by OCT4 in human MSCs and human ESCs
OCT4-Regulated Genes in MSCs
OCT4 inhibits tissue-specific gene expression and promotes self-renewal and pluripotency in ESCs . Our data indicate similar regulatory circuitries for OCT4 in MSCs and ESCs (Fig. 2). To this end, we examined whether suppression of OCT4 in undifferentiated MSCs would lead to downregulation of stem cell-specific and upregulation of differentiation-specific genes. Undifferentiated MSCs were knocked down for OCT4 and compared with untransfected cells for OCT4, SOX2 and NANOG mRNA and protein expression. In control studies, cells were transfected with OCT4 mutant siRNA. Cells transfected with wild-type OCT4 siRNA showed efficient knockdown at the levels of mRNA and protein, as compared with controls (Fig. 3A, 3B). OCT4 knockdowns also showed a marked decrease in SOX2 and NANOG mRNA (Fig. 3A) and protein (Fig. 3B).
Next, we asked whether OCT4 knockdown affected genes linked to (a) differentiation of MSCs or (b) stem cell maintenance. By RT-PCR, we showed decreased expression of genes associated with mesodermal stem cells and found undetectable bands in each compared with controls: MYF5 for myogenesis, peroxisome proliferator-activated receptor-γ for adipogenesis, SOX9 for chondrogenesis, and PH-4 for stroma (Fig. 3C). A dense band was observed solely in the knockdown cells for p63, which has been linked to cellular differentiation (Fig. 3C) . We last studied genes linked to stem cell maintenance, specifically Notch1, Wnt1, FoxD3, β-Catenin, HDAC1, and Myst1. Except for a light band for β-Catenin, we observed undetectable bands for all, suggesting some loss of stemness. All knockdown MSCs and control cells were negative for ectoderm- and endoderm-specific genes (data not shown). In summary, the results demonstrate that OCT4 directly or indirectly regulates stem cell- and mesoderm-specific genes in undifferentiated MSCs.
Cell Cycle Regulation by OCT4 in MSCs
The next set of studies determined whether loss of OCT4 correlates with cells exiting the cell cycle. Among the cell cycle-specific genes studied, those linked to cycle progression (Cdks and cyclins) showed undetectable bands in the knockdowns, whereas a strong band was observed for p21, which is associated with cell cycle arrest and differentiation (Fig. 3D). Upregulation of p63 (Fig. 3C) and p21 in the knockdowns suggests that the cells may be undergoing a program of differentiation.
We subsequently investigated whether OCT4 knockdown lead to changes in MSC morphology and/or growth rate. Knockdowns exhibited a distinctly different morphology compared with wild-type and siRNA control cells (Fig. 3E). Cells displayed decreased cell volume with reduced cellular projections and significantly (p < .05) lowered growth rates (Fig. 3F) up to 3 days following transfection. This difference in growth rates between wild-type and knockdown MSCs is consistent with the results from PI cell cycle analysis, which shows a shift in cells from G2/M to G1/G0 phase following OCT4 knockdown (Fig. 3G). In summary, knockdown of OCT4 in MSCs appears to shift cells from a cycling to a noncycling state.
Comparison of Early- and Late-Passage MSCs
Our final investigations determined whether there is a potential correlation between extended passaging of MSCs and loss of differentiation potential. As a putative readout for MSC differentiation potential, we compared early- and late-passage cells for (a) OCT4 and telomerase expression and (b) telomere length.
Loss of differentiation potential has been shown in MSCs extensively passaged for up to 100 population doublings . If OCT4 is a valid indicator of MSC differentiation potential, then it would be predicted that OCT4 would decrease in late-passage cells. To this end, MSCs cultured for 3 or 25 passages (P3 and P25) were examined for OCT4 expression by immunofluorescence (Fig. 4A, 4B) and real-time RT-PCR (Fig. 4C). Early-passage MSCs (P3) showed bright perinuclear fluorescence in 61% of cells (Fig. 4A, green). In contrast, late-passage MSCs (P25) did not show any observable staining for OCT4 (Fig. 4B). These results were consistent with real-time RT-PCR analyses, which showed a significant (p < .05) decrease in OCT4 expression in P25 MSCs (Fig. 4C).
To examine whether the decrease in telomerase activity observed in differentiated MSCs (Fig. 1A) is consistent with increased differentiation due to extended passaging, we performed real-time RT-PCR analyses for telomerase in P3 and P25 cells (Fig. 4D). P25 MSCs showed a significant (p < .05) decrease in telomerase levels compared with P3 cells. These results were consistent with a decrease in telomere length from 14.9 to 6.1 kilobase pairs (Fig. 4E, 4F). In summary, the extended passaging of MSCs results in decreased OCT4 and telomerase expression, which may directly correlate with decreased differentiation potential.
The in vitro plasticity that MSCs have shown has wrought both excitement and doubt within the stem cell field. Proponents believe that MSCs have the potential to form any tissue in the body and provide an autologous source of cells for regenerative therapies. Critics argue that the observed phenotypes are artifacts of culture and are not clinically relevant. Although neither side has definitively been proven right or wrong, a basic understanding of the biology of MSCs is necessary to make a confident assessment. To this end, our present study has focused on the reported expression of the embryonic transcription factor OCT4 in MSCs. These studies compared the known actions of OCT4 in ESCs to those in MSCs as a means of illustrating conserved regulatory mechanisms between the two stem cells. A similar role for OCT4 in MSCs may underscore the observed plasticity seen with these cells.
We have previously reported the generation of neuronal cells from human MSCs [15, 16]. We thus used neuronal induction as a model to explore the expression of embryonic transcription factors (OCT4, NANOG, and SOX2) in stem cells versus differentiated cells. If expression of these transcription factors is assumed to be consistent with MSC function, then their levels would be predicted to decrease in the induced cells (Fig. 1). OCT4 was found to interact with NANOG and SOX2, as shown by ChIP assay in uninduced MSCs (Fig. 1F). These results suggest a potential autoregulatory loop among the three transcription factors, as supported by SOX2 and NANOG downregulation following OCT4 knockdown (Fig. 3A, 3B). The findings are consistent with known interactions between the transcription factors, as observed in ESCs . Current studies are examining whether a specific subpopulation of MSCs coexpresses OCT4, NANOG, and SOX2, since only 61% of early-passage MSCs were positive for OCT4 (Fig. 4A). This observation illustrates the heterogeneity often found within populations of MSCs and points to the need for optimizing culture conditions within the field. Further work must be done to determine whether this differential behavior among MSCs is due to specific subpopulations of stem cells or whether some cells have begun to undergo lineage-specific differentiation.
An interesting observation from the above results was the detection of SOX2 in the uninduced MSCs but not the D6 induced cells (Fig. 1D, 1E). SOX2 is highly expressed in neural stem and progenitor cells, where it functions in stem cell maintenance . We have previously demonstrated that MSCs induced for 4 and 6 days exhibit a phenotype consistent with neural stem and progenitor cells . Thus, it would be predicted that D6 cells express SOX2. Additional studies are necessary to determine whether SOX2 is expressed at 4 days induction to support the hypothesis of a putative neural stem/progenitor cell.
We demonstrated that uninduced MSCs have elevated telomerase activity compared with induced cells (Fig. 1A). These results supported the findings of a previous report . Although several studies have observed similar results, others have shown low expression of telomerase in MSCs [36, 37]. Recent studies have used MSCs immortalized through ectopic expression of telomerase to enhance differentiation potential . A potential explanation for the conflicting findings may be the expression of telomerase in early-passage MSCs with concomitant loss during consecutive passages. Indeed, a comparison of telomerase levels and telomere length (Fig. 4D–4F) in early- and late-passage MSCs revealed a significant decrease with extended passaging.
We used a genome-wide approach in examining whether OCT4 regulates similar gene targets in MSCs and ESCs (Fig. 2). ChIP-DSL technology provides a novel method that combines conventional ChIP methods with sensitive microarray analyses [24, 25]. Using the Hu20K promoter array, we compared the genome-wide promoter targets of OCT4 in MSCs and ESCs to better understand whether similar regulatory circuitries exist in the two stem cells. OCT4 has been predicted computationally to bind approximately 1,000 target genes in mouse ESCs . We used this finding as a rationale to study the top 2% of 20,000 human promoters analyzed with anti-OCT4-enriched gDNA. OCT4 was found to regulate many different genes in MSCs and ESCs (Fig. 2A). However, the physiological function ascribed to those genes was remarkably similar (Fig. 2B, 2C).
Several of the genes targeted by OCT4 in ESCs (Table 1; Fig. 2) have been previously described as known targets . However, there are many genes targeted by OCT4 in MSCs (Table 2), as well as in ESCs (Table 1), that have not been identified previously. Since OCT4 mainly acts to repress specific target genes in ESCs, it would be interesting to see whether OCT4 acts as a negative regulator in MSCs as well. The observation that OCT4 targets analogous genes in both stem cells (Fig. 2B, 2C) demonstrates similar regulatory circuitries but does not necessarily ascribe a similar role for stem cell maintenance in MSCs. One consideration in the comparison between MSCs and ESCs is their varied culture conditions and heterogeneity among MSCs. If passaging of MSCs produces differentiation in some cells and self-renewal in others, then comparing the population of MSCs as a whole to ESCs may be somewhat inaccurate. Individual analysis of expression patterns in distinct populations of MSCs would further an appropriate comparison between the two types of stem cells.
To investigate whether OCT4 is involved in the maintenance of MSCs, we performed loss-of-function studies using siRNA (Fig. 3). Mesodermal stem cell-specific genes were decreased following OCT4 knockdown (Fig. 3C). The results suggest that OCT4 positively regulates expression of mesodermal genes. These findings are in contrast to the role for OCT4 in ESCs, where it inhibits tissue-specific gene expression. However, OCT4 was also found to positively regulate NANOG and SOX2, as observed in ESCs (Fig. 3A, 3B) [19, 39]. Further investigation is needed to accurately determine whether OCT4 is central to the retention of MSC plasticity. For example, OCT4 knockdown MSCs should not be able to generate neuronal cells if OCT4 is involved in MSC plasticity.
OCT4 has been shown to regulate genes linked to cell cycle progression and cytoskeletal organization in ESCs . We thus examined whether OCT4 exhibits similar functions in MSCs. Profiling of cell cycle-specific genes in OCT4 knockdowns demonstrated a loss of genes involved in cycling, an increase in the cell cycle inhibitor p21, and expression of p63, which has been linked to differentiation (Fig. 3C, 3D) . These results suggest that the knockdown cells may be undergoing growth arrest or a program of differentiation. To confirm the cell cycle profiling studies, we performed DNA content flow cytometry analysis and found a shift from G2/M to G1/G0 phase in the knockdowns (Fig. 3G). These results are consistent with the delay in cell growth and change in morphology observed following OCT4 knockdown (Fig. 3E, 3F) and suggest a link among OCT4, cell cycle regulation, and cytoskeletal organization in MSCs.
Our final studies investigated the expression of OCT4 in early- and late-passage MSCs. A problem often encountered with in vitro expansion of MSCs is a decline in their growth rate and differentiation potential with extended passaging. ESCs, on the other hand, can be maintained in culture for long periods of time without differentiating. The observed differences in the two stem cells may be a result of constitutive telomerase expression by ESCs.
Elevated expression of OCT4 was observed in P3 MSCs as compared with low to undetectable expression at P25 (Fig. 4A–4C). Given the effect of OCT4 knockdown on MSC-specific genes (Fig. 3), OCT4 expression could be an indicator of MSC differentiation potential as an experimental tool and also in clinical diagnostics. Studies are necessary to determine whether the OCT4-negative cells can differentiate along multiple mesodermal lineages to demonstrate that they are not differentiating progeny of MSCs. The observed results cannot be explained by changes in patterns of gene expression due to overgrowth, since all experiments were performed at <80% confluence.
In summary, the present investigation provides pilot studies to compare the molecular mechanisms regulating ESC and MSC phenotype. OCT4 was shown to target analogous genes in both stem cell types. This finding formed the impetus for loss-of-function studies implicating OCT4 in the regulation of MSC-specific genes, and it alludes to a potential role in MSC plasticity.
Disclosure of Potential Conflicts of Interest
The authors indicate no potential conflicts of interest.
This work was supported by a grant from the F.M. Kirby Foundation. This work constitutes partial fulfillment for a Ph.D. thesis (S.J.G.).