Author contributions: J.J.: conception and design, collection and/or assembly of data, data analysis and interpretation, and manuscript writing; Y.D., Y.Y., R.G., Y.Z., Z.K., and X.S.: provision of study material and data analysis and interpretation; S.G.: conception and design, financial support, data analysis and interpretation, manuscript writing, and final approval of manuscript.
Disclosure of potential conflicts of interest is found at the end of this article.
First published online in STEM CELLSEXPRESS January 10, 2013.
Leukemia inhibitory factor and bone morphogenetic protein signaling pathways play important roles in maintaining the self-renewal of mouse embryonic stem cells (ESCs). In contrast, the supplementation of fibroblast growth factor 2 (FGF2) in culture promotes mouse ESC differentiation. It has been proposed that factors that are adverse for maintaining the self-renewal of ESCs might play detrimental roles in the transcription factor-mediated reprogramming of somatic cells to pluripotency. However, recent evidence has revealed that reprogramming efficiency could be improved by FGF2, while the underlying molecular mechanism remains unknown. In this study, we dissected the roles of FGF2 in promoting mouse fibroblast reprogramming and disclosed the molecular mechanism behind this process. We used both primary induction and secondary inducible reprogramming systems and demonstrated that supplementation with FGF2 in the early phase of induced pluripotent stem cell induction could significantly increase reprogramming efficiency. Moreover, we discovered that many extracellular matrix candidate genes were significantly downregulated in fibroblasts treated with FGF2, and in particular, the synthesis of collagen could be greatly reduced by FGF2 treatment. Subsequently, we demonstrated that collagen is a barrier for reprogramming fibroblast cells to pluripotency, and the decreasing of collagen either by collagenase treatment or downregulation of collagen gene expression could significantly improve the reprogramming efficiency. Our results reveal a novel role of the extracellular matrix in mediating fibroblasts reprogramming. STEM CELLS2013;31:729–740
Induced pluripotent stem (iPS) cells can be derived from differentiated somatic cells directly by the ectopic expression of four transcription factors: Oct4, Sox2, Klf4, and c-Myc (OSKM) [1–6]. Although the evidence from recent studies has shown that mouse iPS cells functionally resemble embryonic stem cells (ESCs), the molecular mechanism underlying reprogramming remains largely unknown [7–9]. One major obstacle to dissecting the molecular insights of reprogramming is due to the extremely low efficiency of iPS cell induction. Therefore, great efforts have been made to search for factors that can accelerate the reprogramming process and improve the reprogramming efficiency. Subsequently, culturing cells under hypoxic conditions, supplementing them with vitamin C, and downregulating genes involved in the p53 pathway have been found to be beneficial for promoting iPS cell generation [10–16].
Members of the fibroblast growth factor (FGF) family are among the most common growth factors used to expand stem cells, including neural stem cells and human ESCs. However, the requirement of growth factors between mouse and human ESCs for maintaining self-renewal activity in culture appears distinct. It has been demonstrated that leukemia inhibitory factor (LIF) and bone morphogenetic protein (BMP) signaling pathways are essential for maintaining the self-renewal of mouse ESCs, whereas the supplementation of basic FGF (FGF2) in the culture leads to differentiation [17–20]. In contrast, combining FGF2 and activin/nodal is required for human ESC self-renewal [21, 22]. Interestingly, the epiblast stem cells (EpiSCs) derived from postimplantation mouse embryos require FGF2 and activin/nodal to maintain their self-renewal ability [23, 24], but the pluripotency of EpiSCs is greatly reduced compared with ESCs. It has been believed that the supplementation of growth factors, which play positive roles in supporting ESC self-renewal, in the culture medium is essential for successful reprogramming. According to the opposite roles of FGF2 in self-renewal maintenance between mouse ESCs and human ESCs or mouse EpiSCs, the supplementation of FGF2 in culture media might be detrimental for reprogramming mouse somatic cells to iPS cells. However, evidence from recent studies has shown that FGF2 plays a positive role in mouse somatic cell reprogramming, and the molecular mechanism underlying this process remains elusive [25, 26].
To understand the molecular mechanism of FGF2 in promoting iPS cells generation, we dissected the role of FGF2 in iPS induction using the secondary iPS system, which has been established previously [7, 27, 28]. Our results indicate that FGF2 can promote iPS induction in the early phase but functions adversely in the late phase of iPS induction. Most interestingly, we found that the treatment of fibroblasts with FGF2 in the early stage of iPS induction could dramatically downregulate many extracellular matrix genes and greatly reduce the collagen synthesis. Subsequently, we demonstrate that the decreasing of collagen production either by knockdown of Col1a1 expression or collagenase treatment could significantly improve the somatic cell reprogramming efficiency.
MATERIALS AND METHODS
Mouse and Cell Culture
The specific pathogen-free mice were housed in the animal facility of the National Institute of Biological Sciences. All our study procedures were consistent with the National Institute of Biological Sciences Guide for the care and use of laboratory animals. Mouse embryonic fibroblasts (MEFs) were derived from 13.5-dpc embryos that were collected from female Oct4-green fluorescent protein (GFP) (OG2) transgenic mice  that had been previously mated with male Rosa26-M2rtTA mice. Tail tip fibroblasts (TTFs) were derived from either the new born all-iPS mice or the adult all-iPS mice, which were generated through tetraploid complementation . Control R1 ESCs and OSKM iPS cells were cultured on mitomycin C-treated MEFs in ES medium containing Dulbecco's modified Eagle's medium (DMEM) (Gibco Invitrogen, Carlsbad, CA, http://www.invitrogen.com/site/us/en/home/brands/Gibco.html) supplemented with 15% (vol/vol) fetal bovine serum (FBS), 1 mM L-glutamine, 0.1 mM mercaptoethanol, 1% nonessential amino acid stock, and 1,000 U/ml LIF (all from Chemicon, Temecula, CA, http://www.millipore.com).
iPS Cells Generation
For primary iPS cells generation, the plasmid preparation and the procedure of iPS cell derivation were performed according to the methods described previously [7, 27, 28]. In brief, 293T cells were transfected with TetO-FUW-Oct4, Sox2, Klf4, and c-Myc plasmids separately along with lentivirus packaging plasmids ps-PAX-2 and pMD2G. Media containing virus were collected 24 hours after transfection, and 1 × 105 MEFs were plated on a gelatin-coated dish and infected for 6–8 hours with the collected virus-containing media supplemented with 4 μg/ml polybrene (Sigma, St. Louis, MO, http://www.sigmaaldrich.com). MEFs were recovered for 12 hours in feeder culture medium containing DMEM (Invitrogen, Grand Island, NY, http://www.invitrogen.com) supplemented with 10% (vol/vol) FBS (Hyclone, Waltham, MA, http://www.thermoscientific.com) and 1 mM L-glutamine (Invitrogen, Grand Island, NY, http://www.invitrogen.com), and then, medium was replaced with feeder culture medium supplemented with 1 μg/ml doxycycline (Sigma, St. Louis, MO, http://www.sigmaaldrich.com) to induce expression of the four factors. After 2 days of induction, the medium was changed to ES medium supplemented with 1 μg/ml doxycycline and changed every day. After the ESC-like colonies appeared at day 12 postinfection, the medium was replaced with ES medium for 2–3 days. Then, the colonies were mechanically picked up and digested for culturing on mitomycin C (MMC)-treated MEF cells.
For secondary iPS cell induction, 1 × 105 TTFs of the all-iPS mice were plated on a 35-mm dish. Twelve hours after plating, the medium was changed to feeder culture medium supplemented with doxycycline. Then, the medium was further changed into ES medium containing doxycycline after 2 days of induction and was then exchanged every day. After the ESC-like colonies appeared at day 10–12, the medium was replaced with ES medium without doxycycline for 2–3 days. Then, the colonies were mechanically picked up and digested for culturing on mitomycin C-treated MEF cells. Once the iPS cell colonies were picked, neither FGF2 nor PD173074 was applied to the medium during long-term culture.
To investigate the effects of FGF2 on iPS cell induction, somatic cells either were treated with FGF2 using 20 ng/ml of FGF2 (Peprotech) at different time points or were treated with the FGF receptor one inhibitor by adding 20 nM PD173074 (Calbiochem, Darmstadt, Germany, http://www.merckmillipore.com) to the induction medium. The initial number of plating cells was 1 × 105 per 35-mm dish. The efficiency of iPS cell generation was calculated by the number of GFP-positive colonies per well in the OG2/rtTA MEFs. For the TTFs from the all-iPS mice, the efficiency was calculated by counting the number of colonies that were both alkaline phosphatase (AP) positive and morphologically resembled ES colonies. Each group of treatment contains three replicates.
Growth Curve and BrdU Assay
For growth curve analysis, 1 × 105 cells were plated onto 35-mm dishes, and cells were harvested at certain time points and counted with a cell counting chamber. Each group of treatment contained three replicates. For the BrdU incorporation assay, TTFs of the all-iPS mice were seeded onto six-well plate at a density of 50,000 cells per well. Twelve hours later, the feeder culture medium was changed into feeder medium supplemented with doxycycline (1 μg/ml). Two days later, fibroblasts were treated with BrdU for 45 minutes as described in the manufacturer's instructions (FITC BrdU Flow Kit, BD, Franklin Lakes, NJ, www.bd.com). The cells were stained and analyzed by fluorescence-activated cell sorting. Flow cytometry was performed using standard procedures (Moflo, New Haven, CT, www.beckmancoulter.com).
Immunofluorescence Staining and AP Staining
Immunofluorescence staining was performed according to a method described previously . The following primary antibodies were used: mouse anti-Oct4 (Santa Cruz, CA, www.scbt.com), goat anti-Sox2 (Santa Cruz, CA, www.scbt.com), mouse anti-stage-specific embryonic antigen 1 (anti-SSEA1) (Millipore, Darmstadt, Germany, www.millipore.com), rabbit anti-Nanog (Cosmo, Carlsbad, California, www.cosmobiousa.com, BioCo, Ltd.), goat anti-Col1a1 (Santa Cruz, www.scbt.com), and rabbit anti-MMP8 (Santa Cruz, CA, www.scbt.com). The following fluorochrome-conjugated secondary antibodies were applied: Alexa Fluor 594 goat anti-rabbit (Molecular Probes, Grand Island, NY, http://www.invitrogen.com), Alexa Fluor 633 goat anti-mouse IgM (Molecular Probes, Grand Island, NY, http://www.invitrogen.com), and Alexa Fluor 633 goat anti-mouse IgG (Molecular Probes, Grand Island, NY, http://www.invitrogen.com), and Alexa Fluor 633 rabbit anti-goat IgG (Molecular Probes, Grand Island, NY, http://www.invitrogen.com). The samples were incubated with the appropriate first and secondary antibodies and the DNA was labeled by DAPI. Stained cells mounted on slides were observed on a LSM 510 META microscope (Zeiss, Berlin, Germany www.zeiss.com) using Plan Neofluar ×63/1.4 Oil DIC objective. AP staining was performed with the Leukocyte Alkaline Phosphatase Kit (Sigma-Aldrich, St. Louis, MO, http://www.sigmaaldrich.com) according to the manufacturer's instructions. AP staining was applied to cells at reprogramming day 10–12 when iPS colonies numbers reached a plateau.
Total RNA was purified with TRIzol Reagent (Invitrogen, Grand Island, NY, http://www.invitrogen.com). One microgram of RNA was reverse-transcribed using M-MLV Reverse Transcriptase (Promega, Madison, WI, www.promega.com) and RNasein RNase Inhibitor (Promega, Madison, WI, www.promega.com). Primers for ESC marker genes are described elsewhere .
Quantitative Reverse-Transcription PCR
Quantitative reverse-transcription PCR was carried out with SYBR Premix Ex Taq (Takara, Tokyo, Japan, www.takara-bio.com). Reactions were performed in triplicate using 1/10 concentration of the cDNA obtained as described above. The expression level of the viral transcript in each sample was normalized to glyceraldehyde-3-phosphate dehydrogenase (GAPDH), and the relative quantification of expression was estimated using the comparative CT method.
Collagen Synthesis Analysis of Fibroblasts During Reprogramming
Fibroblasts from the all-iPS mice were seeded on 24-well plates at 1 × 105 per well and then reprogramming was initiated by adding doxycycline. After 24 hours of induction, 1 μCi of L-[2,3-3H]-proline (Perkin-Elmer, Waltham Massachusetts, www.perkinelmer.com) was added to each well of a 24-well plate, in which 1 μg/ml L-ascorbic acid was added to the medium 2 hours before. After 24 hours of incubation, the cell supernatants were harvested and separated into two groups with equal amount. The first group was treated with collagenase IV (Sigma, St. Louis, MO, http://www.sigmaaldrich.com final concentration 1 mg/ml) at 37°C for 90 minutes and then was precipitated by adding 50% trichloroacetic acid to reach a final concentration of 10% and incubated on ice for 1 hour. The second group was used as control and precipitated by adding 50% trichloroacetic acid to reach a final concentration of 10% and incubated on ice for 1 hour. The precipitations of both groups were harvested by centrifugation at 12,000g, 15 minutes, 4°C and washed twice with 10% trichloroacetic acid. The precipitations were then dissociated by 0.3 ml buffer consisting 0.3 M NaOH and 0.3% SDS and warmed to 37°C to reach a fully dissociation. The incorporation of radioactivity was determined using a scintillation counter. The collagen synthesis was calculated by the radioactivity difference between the collagenase-digested group and the control group. The results were normalized to cells number. Both the incorporation wells and the cell number counting wells were seeded in triplicates. The experiment was repeated three times.
Western Blot Analysis
Whole cell extracts were prepared using RIPA buffer, resolved on SDS-PAGE gels, and transferred to polyvinylidene fluoride (PVDF) membranes. Primary antibodies used were goat anti-Col1a1 (Santa Cruz, CA, www.scbt.com), rabbit anti-MMP8 (Santa Cruz, CA, www.scbt.com), mouse anti-p16Ink4a (Santa Cruz, CA, www.scbt.com), and mouse anti-Tubulin (Sigma, St. Louis, MO, http://www.sigmaaldrich.com). Enhanced chemiluminescence (ECL) peroxidase-labeled anti-mouse, rabbit, or goat antibodies (Amersham, Pittsburgh, PA, http://www.gelifesciences.com) were used as the secondary antibodies.
Knockdown of Gene Expression by shRNAs
Generation of p15Ink4b, p16Ink4a, and p19Arf shRNA lentiviral particles was described previously . Briefly, short hairpin was cloned into PsicoR-RFP lentiviral vector under the U6 promoter. Co1a1 shRNA lentiviral vectors were purchased from Sigma. The lentiviral vector was transfected into 293T cells by calcium phosphate transfection. Virus-containing supernatant was collected 24–72 hours post-transfection and was used to infect MEFs. Successful knockdown of p15Ink4b, p16Ink4a, and p19Arf was verified by qPCR.
One million iPS cells were injected subcutaneously into each flank of recipient the nonobese diabetic/severe combined immunodeficient mice (NOD/SCID) mice (Jackson Laboratory, Bar Harbor, Maine, www.jax.org). Paraffin sections of formalin-fixed teratoma specimens were prepared 3–5 weeks after injection, and analysis of H&E-stained tissue sections was performed for each specimen.
To produce chimeric mice, approximately 10 iPS cells were microinjected into the institute of cancer research (ICR) eight-cell embryos using piezo-actuated microinjection pipette. After culturing for 1 day, the embryos were transplanted into uterus of pseudopregnant mice. When the chimeric mice reached adulthood, they were mated with ICR mice to test the germ-line transmission.
Total RNA was extracted from TTFs treated with or without FGF2 for 6 days after induction using Trizol reagent (Invitrogen, Grand Island, NY, http://www.invitrogen.com). Samples of both treatments were prepared in three biological repeats. Affymetrix Mouse Gene 1.0 ST Array (Affymetrix, Santa Clara, CA, www.affymetrix.com, Inc.) was used, and all experiments were performed at Beijing Capitalbio Corporation. Data were analyzed using GCOS 1.4 software provided by Affymetrix. Signal values of probes presented in two samples were plotted in a scatter graph. Pearson's correlation coefficient (R) between samples was calculated using Excel.
FGF2 Can Accelerate Reprogramming Process and Increase Reprogramming Efficiency
The MEFs derived from Oct4-GFP transgenic mice were infected by the lentiviruses containing Oct4, Sox2, Klf4, and c-Myc transcription factors. One group of cells was treated with FGF2 at a concentration of 20 ng/ml, and the other group of cells was treated with 1% bovine serum albumin (BSA) as the control group. After the induction of the four factors, the cells treated with FGF2 exhibited a more rapid change in morphology. At day 6 following induction, there were immature colonies of the FGF2-treated cells, whereas no colonies could be observed at this time point in the control cells (Fig. 1A). After 12 days of induction, the number of Oct4-GFP-positive colonies in the FGF2 treatment group was approximately 5–10 times more than that of the control group (Fig. 1B).
We then used the secondary inducible reprogramming approach to further confirm the effects of FGF2 on reprogramming, and similar results were obtained. As reported, all-iPS mice could be generated with iPS cells through tetraploid complementation in our laboratory [7, 30]. After generation of all-iPS mice with iPS cells derived from lentiviruses with dox-inducible OSKM, the skin fibroblasts or TTFs were cultured from the newborn or adult all-iPS mice, respectively. Because lentiviral integration in the genome is homogenous in the somatic cells retrieved from the all-iPS mice, variation due to viral infection during conventional iPS induction could be excluded. Upon doxycycline supplementation, expression of the four transcription factors could be initiated in the fibroblasts, and iPS cell colonies could subsequently form after 10 days of induction. As expected, significantly more AP-positive colonies could be observed in the cells treated with FGF2 than the control group (Fig. 1C). Moreover, no iPS colonies appeared once the FGF2 pathway was blocked by adding a chemical inhibitor of FGF receptor 1 (PD173074) in the culture medium (Fig. 1C). The number of iPS colonies was counted from day 9 to day 15 after induction. The number of colonies in the FGF2 group was 5–10 times higher than the control group. Meanwhile, we observed that iPS colonies reached a peak when FGF2 was applied for 12 days, while iPS colonies decreased when cells were treated with FGF2 for more than 15 days (Fig. 1D). Interestingly, FGF2 appeared to be more effective at promoting the reprogramming of the TTFs collected from a very old all-iPS mouse (i.e., a 1.5-year-old mouse), with a greater than 20-fold increase (Fig. 1E, 1F).
Taken together, these results demonstrated that the supplementation of FGF2 in culture medium during iPS induction, using either the conventional reprogramming approach or the secondary inducible reprogramming approach, can significantly accelerate the reprogramming kinetics and increase the reprogramming efficiency. More strikingly, reprogramming efficiency could be increased by greater than 20-fold when FGF2 was used to reprogram TTFs that were derived from aged mice.
iPS Cell Lines Generated with FGF2 Are Pluripotent
We next sought to investigate whether the iPS cell lines established by FGF2 treatment were pluripotent. As shown in Figure 2A, the FGF2 iPS cell lines derived from the OG2 fibroblasts were GFP positive, and the secondary iPS cell lines established from the TTFs of the all-iPS mice were AP-positive (Fig. 2B). The expression of the key transcription factors, including Oct4, Sox2, Nanog, and the ES-specific surface marker, SSEA-1, in the iPS cell lines could be detected by immunofluorescence staining (Fig. 2C). Moreover, the expression level of Nanog was indistinguishable between the FGF2 iPS cell lines and the control iPS cell lines (supporting information Fig. S1A). In addition, the FGF2 iPS cell lines expressed the other ES marker genes as analyzed by RT-PCR (supporting information Fig. S1B).
We then sought to evaluate the capability of the iPS cell lines derived with FGF2 to differentiate into three germ layers by a teratoma assay. When injected into NOD/SCID mice, iPS cells derived with FGF2 were capable of inducing teratoma formation with the generation of derivatives from all three germ layers (Fig. 2D). Next, we performed blastocyst injection to further confirm the pluripotency of the FGF2 iPS cell lines. Chimeric mice with germline transmission ability could be efficiently produced from the FGF2 iPS cell lines (Fig. 2E). These results demonstrated that the iPS cells generated using FGF2 for the reprogramming process are pluripotent.
Because FGF2 could accelerate cell proliferation under many circumstances , we then asked whether FGF2 functions in reprogramming through accelerating cellular proliferation. As expected, the number of fibroblasts in the FGF2 group was twice that of cells in the control group at day 3 (supporting information Fig. S2A, S2B). This was further proved using the BrdU-incorporating assay, in which the BrdU-incorporating cells at day 4 in the FGF2 group were 10-fold that of the control group (supporting information Fig. S2C, S2D). We further compared the gene expression characteristics of several cell cycle-related genes in the cells treated with FGF2 or PD173074. Of the genes tested, p16Ink4a and p15Ink4b were found to be highly expressed in the cells treated with PD173074 (supporting information Fig. S3A, S3B). In contrast, expression of p16Ink4a and p15Ink4b was downregulated in the FGF2-treated group, which was further confirmed by Western blot of p16Ink4a (supporting information Fig. S3A). To test the possibility that early addition of PD173074 inhibits reprogramming by elevating the expression of p16Ink4a and p15Ink4b, we constructed siRNA against p16Ink4a, p19ARF, and p15Ink4b; and the knockdown efficiency was verified (supporting information Fig. S3C). A dsRed reporter gene was inserted in the plasmids, and the resulting red fluorescence indicated the cells infected by the virus with the siRNA cassette. We found that p16Ink4a and p19Arf knockdown in the PD173074-treated fibroblasts could rescue the inhibitory effect of PD173074 in reprogramming (supporting information Fig. S3D, S3E). These results indicate that FGF2 can increase reprogramming efficiency at least partially through accelerating cell proliferation by downregulating the expression of cell cycle inhibitor genes.
FGF2 Is Essential in the Early Phase of Reprogramming and Detrimental to Reprogramming in the Late Phase
To investigate the dynamic roles of FGF2 in increasing reprogramming efficiency, we harvested the mRNA of the somatic cells during reprogramming induction and analyzed the expression level of FGF2. We found that the expression level of FGF2 gradually increased and reached a peak on day 6 of induction, and then the expression of FGF2 was downregulated after day 7 (Fig. 3A). Based on the gene expression characteristics of FGF2 during reprogramming, we next sought to investigate whether FGF2 plays important roles in promoting somatic cell reprogramming in the early phase of reprogramming. To address this hypothesis, FGF2 was withdrawn from the medium at day 6, day 7, day 10, and day 13, and then the AP-positive colonies were counted. As shown in Figure 3B, the reprogramming efficiency was dramatically reduced when FGF2 was withdrawn from the induction medium on day 6, whereas the reprogramming efficiency was comparable among the groups for which FGF2 was withdrawn on day 7, day 10, or day 13. These results indicate that the treatment of somatic cells with FGF2 from the beginning of induction to day 6 is critical for improving reprogramming efficiency. Moreover, inhibition of FGF2 signaling by PD173074 from day 1 to day 6 led to no AP-positive colony formation, which further demonstrated the critical role of FGF2 in the early phase of reprogramming (Fig. 3B). Subsequently, the reprogramming efficiency could be increased approximately twofold when cells were treated with PD173074 at the late phase of reprogramming from day 7, or day 9, or day 11 (Fig. 3B). The results demonstrated that the combination of FGF2 treatment from day 1 to day 6 with FGFR1 inhibitor PD173074 treatment from day 11 to 12 is the most efficient combination. Although the reprogramming efficiency is comparable between this combination and FGF2 treatment alone, the morphology of the iPS colonies produced by this combination was superior (Fig. 3B, 3C). These results indicate that FGF2 is beneficial and essential in the first 6 days of reprogramming, and the inhibition of FGF2 in the late phase of reprogramming could lead to a better result by improving the quality of iPS colonies.
Pluripotency of iPS Cell Lines Generated with Activation of the FGF2 Pathway in the Early Phase of Reprogramming and Inhibition of the FGF2 Pathway in the Late Phase
Because treatment of cells with FGF2 in the early period of reprogramming and inhibition of the FGF2 signal in the late period could elevate the efficiency and improve the colony morphology, it is necessary to confirm that the iPS cells produced are pluripotent. The secondary iPS cell lines generated with the FGF2 and PD173074 combination are AP-positive (Fig. 4A) and express Oct4, Sox2, Nanog, and SSEA-1 as determined by an immunostaining assay (Fig. 4C). When injected into NOD/SCID mice, iPS cells were capable of inducing teratoma formation and generated derivatives from all three germ layers (Fig. 4D). Moreover, chimeric mice with germline transmission ability could be efficiently produced from the iPS cell lines that were derived using FGF2 and PD173074 combination (Fig. 4B).
Gene Expression Profile of TTFs Treated with FGF2 During Reprogramming
To further investigate the molecular mechanism of FGF2 in elevating reprogramming efficiency, we performed microarray analysis and compared global gene expression profiles of all-iPS mice-derived TTFs treated with or without FGF2 for 6 days. Pearson correlation analysis was used to cluster the cells. The results demonstrate that the gene expression profile of the FGF2 group appeared distinct from the control group (Fig. 5A). Of the differentially expressed genes, 95 genes were upregulated in the FGF2 group with a greater than twofold difference, and 389 genes were downregulated in the FGF2 group with a greater than twofold difference.
Gene ontology (GO) analysis of the FGF2-affected genes based on their localization in the cell showed that 115 genes that belonged to the extracellular category were the largest proportion of the affected genes. The genes that belong to the membrane and integral to membrane categories were the second and third most affected genes, respectively (Fig. 5B). Next, we analyzed the genes based on their function and found that genes that function in focal adhesion were the most commonly affected. Genes that function in the ECM (extra cellular matrix)-receptor interaction and cell communication were the second and third most affected genes (Fig. 5C). A collective list of the upregulated and downregulated genes affected by FGF2 was shown in supporting information Table S1. Taken together, the microarray data showed that the genes affected by FGF2 are largely related to cell adhesion function.
In addition to the extracellular matrix genes, many genes involved in the mesenchymal to epithelial transition (MET) were found to be differentially expressed between the FGF2 group and the control group. Further analysis showed that the expression level of several mesenchymal genes, including Slug, Snail, Zeb1, Zeb2, and Cdh2, were significantly downregulated in the FGF2 group when compared with the control group (Fig. 5E). Similarly, the downregulation of the TGFbeta family genes was also observed in the FGF2 group (supporting information Fig. S4A). In contrast, the expression of BMP2 and BMP7 was upregulated in the FGF2 group (supporting information Fig. S4B). All these results indicate that the MET is facilitated in the cells treated with FGF2 during reprogramming. In contrast, epithelial genes were not found to exhibit significant upregulation in the FGF2 group (supporting information Fig. S4C). Taken together, these results indicate that supplementation of FGF2 during reprogramming facilitates the repression of mesenchymal genes, which is a prerequisite for the successful reprogramming of fibroblasts to a pluripotent state, and moreover, the downregulation of the ECM genes might play important roles in improving the reprogramming efficiency.
FGF2 Could Downregulate Collagen Gene Expression During Reprogramming
As shown above, the comparison of the gene expression profiles between the FGF2 group and the control group revealed that a large number of genes that function in cell adhesion and the extracellular matrix were differentially expressed. We next sought to investigate whether the extracellular matrix (ECM) functions in the reprogramming process. During the analysis of the gene expression profile, we noticed that the family of collagen genes was significantly downregulated in the FGF2-treated cells (Fig. 6A). Real-time PCR was applied and confirmed the microarray data (Fig. 6B). Collagen is the most abundant protein in the ECM and forms the matrix to which cells attach and enables cell migration .
To investigate whether FGF2 treatment could decrease collagen production during reprogramming, we examined the collagen type-1 alpha 1 (COL1A1) expression using Western blot and immunofluorescent staining. Type-I collagen is the most abundant collagen in fibroblasts, and COL1A1 encodes the major component of type-I collagen . The Western blot results clearly showed that FGF2 treatment could greatly reduce the COL1A1 production (Fig. 6C). Similarly, immunofluorescent staining of COL1A1 also showed reduced protein expression in fibroblasts treated with FGF2 at reprogramming day 2 (Fig. 6D). To further confirm that FGF2 treatment could lead to reduced collagen synthesis, the [3H]-proline incorporation experiment was performed. Since the major biosynthetic destination of proline is collagen, the incorporation of the radioactive [3H]-proline into newly synthesized proteins provides a reliable index of collagen synthesis. Significantly reduced collagen synthesis was detected in the FGF2-treated fibroblasts at reprogramming day 2 (Fig. 6E). Taken together, these results confirmed that the FGF2 treatment could significantly reduce the collagen production during reprogramming, which might play positive roles in promoting somatic cell reprogramming.
The Extracellular Matrix Is a Barrier for Reprogramming
Downregulation of collagen family genes due to FGF2 treatment might indicate that an impaired ECM might be beneficial for reprogramming. We therefore performed shRNA knockdown of Col1a1, and found that reduced expression of Col1a1 could indeed significantly elevate reprogramming efficiency as shown by more AP-positive colonies appeared at reprogramming day 10 (p < .001) (Fig. 7A, 7B, 7C).
To further test the possibility that the impaired ECM could facilitate reprogramming, we treated the TTFs of the all-iPS mice with collagenase IV. When 20 μg/ml of collagenase IV was applied to the culture medium, the number of colonies that were both AP-positive and morphologically similar to ES colonies in the collagenase IV-treated group was approximately two times more numerous than in the control group (Fig. 7D, 7E). In contrast, pretreatment of the cell culture dish with collagen is believed to strengthen the ECM. We reprogrammed cells on dishes that were precoated with collagen. As expected, additional collagen coated on the dish impaired the reprogramming process, as revealed by reduced iPS colony numbers (Fig. 7D, 7E). We also found that collagenase treatment itself could not accelerate cell proliferation (supporting information Fig. S5D).
We further investigated the effects of different combinations of FGF2, collagenase IV, and collagen on reprogramming outcome. We found that the treatment of fibroblasts with FGF2 and collagenase IV could not further elevate the reprogramming efficiency as compared to FGF2 treatment alone (Fig. 7D, 7E). But, FGF2 can rescue the adverse effects of collagen on reprogramming (Fig. 7D, 7E). These results indicate that the reduction of extracellular collagen compartments is part of positive effects of FGF2 in reprogramming.
The outcome of collagen content is influenced by both collagen synthesis and degradation mediated by matrix metalloproteinases (MMPs). We next examined whether the expression of MMPs could be affected by FGF2 treatment during reprogramming. Real-time PCR results showed that the expression of MMPs exhibited no difference in fibroblasts treated with or without FGF2 at reprogramming day 2 (supporting information Fig. S5A). Furthermore, Western blot and immunocytochemistry staining results showed that the expression of MMP8 appeared similar in both FGF2-treated and control fibroblasts at reprogramming day 2 (supporting information Fig. S5B, S5C). Taken together, this study demonstrated that FGF2 plays important and beneficial roles in the early phase of reprogramming. FGF2 promotes somatic cell reprogramming by accelerating cell proliferation, downregulating mesenchymal gene expression, and eliminating the extracellular matrix, particularly the extracellular collagen.
In this study, we demonstrated that FGF2 can improve the reprogramming efficiency in the early phase of reprogramming, whereas it functions adversely in the late phase of reprogramming. FGF2 promotes somatic reprogramming through accelerating cell proliferation, facilitating the MET, and eliminating extracellular collagens.
Reprogramming of differentiated somatic cells into pluripotent stem cells by transcription factors has been considered to be a stochastic event, and the number of cell divisions is a key parameter for driving somatic cell reprogramming . Previously, the inhibition of the p53 signaling pathways has been proven to be very efficient for improving the reprogramming efficiency, and it has been further demonstrated that reprogramming is a cell cycle-dependent event [13–16]. Either inhibition of p53 activity or inhibition of the downstream p16 activity could significantly accelerate cell division, which can further lead to a high efficiency of reprogramming. In this study, we observed that cell division could be significantly accelerated when FGF2 was applied in the induction medium. Subsequently, we demonstrated that the inhibition of the p16 cell cycle inhibitor by FGF2 greatly contributed to this acceleration of cell division and promotion of reprogramming efficiency.
The MET has been recently found to be crucial for reprogramming fibroblast cells into iPS cells, which has been considered as the most well-defined phenomenon in fibroblast reprogramming [35, 36]. Knocking down the expression of mesenchymal genes during reprogramming could significantly increase the reprogramming efficiency. By analyzing the characteristics of the gene expression of FGF2-treated fibroblasts during reprogramming, we found that many mesenchymal genes were significantly downregulated. Therefore, in addition to the beneficial effects of accelerating cell division, FGF2 can also play an important role in suppressing mesenchymal gene expression during reprogramming, which is a prerequisite for MET initiation.
Whether the extracellular matrix plays a role in somatic cell reprogramming has not been investigated before this study. Here, we clearly demonstrate that one major role FGF2 plays in improving the reprogramming efficiency is attributed to the elimination of the extracellular matrix, particularly extracellular collagens. We observed a significant downregulation of many extracellular collagen genes in the cells treated with FGF2, and we further showed that these extracellular collagens play an adverse role in somatic cell reprogramming and that the elimination of the collagens either by knocking down Col1a1 expression or collagenase treatment can significantly improve the reprogramming efficiency. Most recently, two reports showed that somatic cells in a suspension culture can be successfully reprogrammed, which indicates that cell attachment to the culture dish is not necessary for reprogramming [37, 38]. Our results further indicate that the attachment might even play adverse roles in reprogramming.
In conclusion, FGF2 plays important and beneficial roles in promoting somatic cell reprogramming in the early phase of reprogramming. FGF2 can promote somatic cell reprogramming partially through accelerating cell proliferation and facilitating the MET by downregulating mesenchymal-specific gene expression. Most importantly, we demonstrate for the first time that the extracellular matrix is a barrier for somatic cell reprogramming that can be significantly downregulated by FGF2 during reprogramming.
We are grateful to our colleagues in our laboratory for their assistance with experiments and in the preparation of this manuscript. This project was supported by the Ministry of Science and Technology (Grants 2010CB944900, 2011CB812700 and 2011CB964800).
DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
The authors indicate no potential conflicts of interest.