• We are sorry, but NCBI web applications do not support your browser and may not function properly. More information
Logo of nihpaAbout Author manuscriptsSubmit a manuscriptNIH Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
Stem Cells. Author manuscript; available in PMC Aug 2, 2012.
Published in final edited form as:
PMCID: PMC3410557

Fgf2 Inhibits Differentiation of Mesenchymal Stem Cells by Inducing Twist2 and Spry4, Blocking Extracellular Regulated Kinase Activation and Altering Fgfr Expression Levels


Mesenchymal stem cells (MSCs) are known to differentiate into connective tissue lineages but intracellular signaling pathways that maintain cells in an undifferentiated state remain largely unexplored. Previously we reported that fibroblast growth factor 2 (Fgf2) reversibly inhibited multi-lineage differentiation of primary mouse MSCs and now identify a unique compliment of signaling proteins that are dynamically regulated by this mitogen and whose expression levels are strongly correlated with inhibition of cell differentiation. Fgf2 selectively induced expression of Twist2 and Sprouty4 (Spry4) and repressed expression of soluble frizzled related receptor 2 (Sfrp2), runt-related transcription factor 2 (Runx2), and peroxisome proliferation activated receptor gamma (Pparg). In contrast, Wnt3a induced expression of Twist but not Twist2 or Spry4 and bone morphogenetic protein 2 (Bmp2) failed to alter expression of all three genes. Moreover, pre-treatment of MSCs with Fgf2 delayed extracellular regulated kinase 1 and 2 (Erk1/2) phosphorylation and repressed bone-specific gene expression during an osteo-induction time course. Alternatively, pre-treatment with Wnt3a had no effect whereas BMP2 pretreatment augmented ERk1/2 activation and bone-specific gene expression. Fgf2 also induced expression of Fgfr1 and 4 and repressed Fgfr2 and 3 expression in MSCs, whereas Wnt3a and Bmp2 had the opposite effect. Finally, immuno-staining revealed that Twist and Spry4 were co-expressed in MSCs and that Fgf2 treatment altered their sub cellular distribution in a manner consistent with their mode of action. Collectively, these studies demonstrate that inhibition of mouse MSC differentiation by Fgf2 is strongly correlated with up regulation of Twist2 and Spry4 and suppression of Erk1/2 activation.

Keywords: Mesenchymal stem cells, mesenchymal stromal cells, FGF2, Twist, Sprouty 4, extracellular regulated kinase, osteogenic


A number of specific signaling pathways have been identified that regulate differentiation of MSCs to connective tissue cell lineages [16]. Less well characterized are the signals and/or pathways that maintain MSCs in an undifferentiated state. In the latter case, various studies have demonstrated a role for canonical Wnt signaling and Fgf2 in regulating cell growth and differentiation but their effects on MSCs appear complex. For example, exogenous application of Wnt3a [7] or ectopic expression of the Lpr5 receptor [8] has been shown to stimulate MSC growth by inducing cyclin D (Ccnd1) and c-Myc expression. In contrast Dkk1, an inhibitor of the canonical Wnt pathway has been shown to stimulate quiescent human MSCs to enter the cell cycle [9]. Activation of canonical Wnt signaling in MSCs has also been reported to inhibit osteogenic differentiation [10, 11] or stimulate it by directly inducing Runx2 expression (12) and enhancing the activity of Bmp’s [13]. In mice, loss of function of the Wnt antagonist secreted frizzled related receptor 1 (Sfrp1) results in increased bone mass and enhanced healing of bone fractures [14]. These mice also exhibit accelerated chondrocyte maturation, consistent with reports that Wnt signaling promotes chondrogenic differentiation of human MSCs [15]. Based on these findings, the affects of Wnt signaling on MSC growth and differentiation are believed to be context dependent [16] and also depend upon the relative strength of the signal transmitted in cells [17].

Fgf2 also exhibits positive and negative effects on growth and differentiation of MSCs. For example, Fgf2 has been shown to stimulate growth and preserve the differentiation potential of MSCs during long-term culture expansion in vitro [1823]. Alternatively, it has also been shown to promote osteoblast differentiation by inducing osteocalcin gene expression in MSCs and enhancing calcium deposition [24, 25]. Fgf2 also reportedly stimulates chondrogenic and adipogenic differentiation of human [26, 27] and rat [28] MSCs, respectively. Recent studies by Debiais et al. [29] revealed that Fgf2 exhibits differentiation stage-specific effects on cellular differentiation by showing that it stimulates growth of immature osteoblast progenitors but induces osteogenic differentiation of more mature precursors. Therefore, the fact that MSC populations are functionally heterogeneous with respect to their differentiation potential [30] may explain the varied effects of both Wnt signaling and Fgf2 on their growth and differentiation.

Previously, we reported that Fgf2 inhibited multi-lineage differentiation of primary mouse MSCs enriched from bone marrow via immundepletion [19]. Moreover, this effect was reversible as multi-lineage differentiation potential was restored in MSCs following culture in the absence of Fgf2 for 7 days prior to induction of cell differentiation. In this report we demonstrate that MSCs express a unique compliment of mRNAs encoding a variety of Fgf2, Wnt, and Bmp signaling pathway components and downstream targets, such as Twist, Twist2, and Spry4. Moreover, we show that Fgf2, Wnt3a, and Bmp2 differentially regulate expression of these genes, which is strongly correlated with effects on cell differentiation. For example, pre-treatment of MSCs with Fgf2 strongly up regulated Twist2 and Spry4, altered the sub cellular distribution of these proteins, and suppressed activation of Erk1/2 following stimulation with osteogenic inducers, resulting in a block in bone-specific gene expression. Conversely, pre-treatment of cells with Wnt3a up regulated expression of Twist but not Twist2 or Spry4, had no effect on Erk1/2 activation, and only modestly affected cell differentiation. In contrast, Bmp2 pre-treatment had no effect on Twist, Twist2, and Spry4 expression or Erk1/2 activation but significantly augmented osteogenic differentiation. Fgf2 also up regulated expression of Fgfr1 and Fgfr4 and suppressed expression of Fgfr2 and Fgfr3 whereas Wnt3a and Bmp2 had the opposite effect on receptor expression. Collectively, these studies demonstrate that Fgf2 specifically alters expression of its own receptors and several downstream signaling proteins and these changes are strongly correlated with inhibition of cell differentiation. To our knowledge this is the first report to demonstrate that Spry4 is a target of Fgf2 signaling in MSCs and that its expression is strongly correlated with suppression of Erk1/2 activation and inhibition of cell differentiation.


Validation of SAGE Tags

Construction of a mouse MSC cDNA library was described previously [31]. Aliquots (1 μl) of the primary pool of the amplified phage library were used as input in PCR reactions (100 μl) containing 100 pmoles of forward and reverse gene-specific primers, 1x PCR buffer, 0.2 mM dNTPs and 0.5 U Taq polymerase (Sigma-Aldrich, St. Louis, MO, http://www.sigmaaldrich.com). After an initial denaturation step at 94°C for 3 min, reactions were amplified for 30 cycles at 94°C for 30 sec, 55–65°C for 45 sec, and 72°C for 90 sec, followed by a final incubation at 72°C for 5 min. PCR product were electrophoresed through a 1% agarose gel, excised from the gel and purified using GeneElute columns (Sigma-Aldrich) and then cloned using the AdvanTAge PCR cloning kit (Clontech, Palo Alto, CA, http://www.clontech.com). Plasmid DNA was isolated and sequenced as described above to confirm the identity of each product. Primers sequences are listed in Supplementary Table 1.

MSC Cultivation

MSCs were enriched from the long bones of FVB/n mice by immundepletion as previously described [19, 32]. Following immuno-depletion MSCs were cultured 48 hrs in complete culture media (CCM) and then transferred to osteogenic or chondrogenic induction media for up to 6 weeks as indicated. CCM was composed of α-MEM containing L-glutamine (Invitrogen, Carlsbad, CA, http://www.invitrogen.com) and supplemented with 10% fetal calf sera (Lot # F0091, Atlanta Biologicals, Atlanta, GA, http://www.atlantabio.com), penicillin (100U/ml) and streptomycin (100U/ml)..

Cell Differentiation

MSCs were plated in 24-well plates at 0.5 × 104 cells/well in CCM alone or CCM supplemented with 20 ng/ml Fgf2 (Sigma-Aldrich), 10 ng/ml Wnt3a, or 100 ng/ml Bmp2 (R&D Systems Inc., Minneapolis, MN, http://www.rndsystems.com) or combinations thereof for 7 days. Thereafter, the CCM was removed and cells were incubated in osteogenic induction media (OIM) for an additional 21 days with media changes every other day. OIM consisted of CCM supplemented with 10% FBS, 50ug/ml ascorbic acid 2-phosphate, 10mM β-glycerol-phosphate, and 10−8M dexamethasone. The cell monolayer was subsequently washed twice with PBS, fixed in 4% paraformaldehyde for 10 minutes, and then incubated an additional 10 minutes in 1% Alizarin Red S stain (pH 4.1). The extent of mineralization was assessed by extracting the bound stain with a solution of 20% methanol and 10% acetic acid in water for 15 minutes and analyzing the absorbance of the solution at 450 nm using a micro-plate reader (BioRad Laboratories, Hercules, CA, http://www.bio-rad.com). All experiments were performed in duplicate and data were reported as mean ± S.D. Alternatively, cells were collected by trypsinization, suspended in chondrogenic induction media (CIM), transferred to a 15 ml polypropylene tube that was centrifuged for 10 min at 300xg, and then cultured for 21 days. CIM consisted of high glucose Dulbecco’s Minimal Eagle Medium (Invitrogen) supplemented with 10 ng/ml transforming growth factor (TGF)-3 β (R&D Systems Inc), 0.5μm dexamethasone, 50 μg/ml ascorbate-2-phosphate, 40 μg/ml proline, and 1% insulin-transferin-selenium-A (Invitrogen). Micro mass pellets were then cultured an additional 21 days in hypertrophic media, which consisted of DMEM supplemented with 1nM dexamethasone, 50μg/ml ascorbic acid-2-phosphate, 20 nM β-glycerol phosphate, and 50ng/ml thyroxine with media changes every 2–3 days. The micro mass pellets were then washed in PBS, fixed in formalin, embedded in paraffin, cut into 5 μm sections and stained with 1% (w/v) toluidine blue sodium borate.

Real-Time PCR

Total RNA was isolated using the RNeasy Kit (Qiagen, Alameda, CA, http://www.qiagen.com), converted to cDNA using the Transcriptor First Strand cDNA Synthesis Kit (F. Hoffmann-La Roche, Basel, Switzerland, http://www.roche.com), and amplified by PCR using the TaqMan® EZ RT-PCR kit (Applied Biosystems, Carlsbad, CA, http://www.appliedbiosystems.com) according to the manufacturer’s instructions. Reactions were performed on a 7900 HT sequence detector (Applied Biosystems) and transcript levels quantified using the relative Ct method by employing GAPDH mRNA as an internal control. The following Assay-On-Demand Taqman® probes (Applied Biosystems) were used for analysis: Twist, Mm00442036; Twist2, Mm00492147; Spry4, Mm00442345; Ctnnb1, Mm00483033; Sfrp1, Mm00489161; Sfrp2, Mm00485986; Pparg,Mm00440939; Runx2, Mm00501578; Ccnd1, Mm00432360. Quantification of bone-specific gene and Fgfr1-4 was performed using the following primers and FastStart Universal SYBR Green Master Mix (F. Hoffman-La Roche). Bglap, 5’-agactccggcgctacctt-3’ and 5’-ctcgtcacaagcagggttaag-3’; Sparc, 5’-ggtggaggagacaggggta-3’ and 5’-tgtcagccaccacctcct-3’; Alp1, 5’-ctgactgacccttcgctctc-3’ and 5’-gtggtcaatcctgcctcct-3’; Bsp, 5’-aagaggagggggaggaaga-3’ and 5’-cgagagtgtggaaagtgtgg-3’; Fgfr1, 5’-atccgcagcctcacattc-3’ and 5’-ggtggtattaactccagcagtctt-3’; Fgfr2, 5’-cctgcggagacaggtaacag-3’ and 5’-cgcgttgttatcctcacca-3’; Fgfr3, 5’-ggcctgcgtgctagtgtt-3’ and 5’-tcaggccctggaacctct-3’; Fgfr4, 5’-cctctgaggaaatggagcag-3’ and 5’-ccacagcacagcctcaca-3’; Ccnd1, 5-attggtctttcattgggcaacggg-3’ and 5’-ggccaattgggttgggaaagtcaa-3’; Ctnnb1, 5’-tgcagcttctgggttccgatgata-3’ and 5’-agatggcaggctcagtgatgtctt-3’; Gapdh, 5’-tcaacagcaactcccactcttcca-3’ and 5’-accctgttgctgtagccgtattca-3’..


MSCs (2 × 104) were plated in 8-well chamber slides (BD Biosciences) and after 24–48 hrs fixed with 4% paraformaldehyde for 15 min at room temperature. Cells were then incubated for 1 hour in PBS containing 8% Tween-20 and 30% goat or rabbit serum. The cells were then washed with PBS and incubated an additional hour at room temperature with a polyclonal anti-Twist (1:100, H-81) or anti-Sprouty 4 (1:100, V-15) antibody (Santa Cruz Biotechnology, Inc., Santa Cruz, CA, http://www.scbt.com). After washing cells were incubated for 40 min with the appropriate secondary antibody (1:400) conjugated to either FITC or PE. Cells were then mounted with DAPI mounting media and imaged using a Leica RX-DMV upright fluorescent microscope (Meyer Instruments Inc., Houston, TX, htpp://www.meyerinst.com) attached to a Cooke Sensicam High Performance digital camera (Hamamatsu Corp., Middlesex, NJ, http://www.hamamatsu.com) and rendered using Slidebook® software (Intelligent Imaging Innovations, Denver, CO, http://www.intelligent-imaging.com). Color intensities for all images were set above the cut-off threshold determined for cells stained with the fluorochrome-conjugated secondary antibodies alone.

Western blot

Protein lysates were prepared using the Qproteome Mammalian Protein Prep Kit (Qiagen) according to the manufacturer’s instructions. Protein samples (20 μg) were prepared in Laemmli Sample Buffer (BioRad Laboratories) containing β-mercaptoethanol (1:20), denatured at 95°C for 5min, electrophoresed on NuPAGE 10% Bis-Tris gels using 1X NuPAGE MES SDS Running Buffer and then transferred to 0.45um nitrocellulose membranes in 1X NuPAGE Transfer Buffer containing 10% methanol (Invitrogen). Membranes were washed with Tris Buffered Saline (TBS) for 5 min, incubated in TBS with 0.1% Tween-20 (TBST) and 5% w/v nonfat dry milk overnight at 4°C, washed an additional 3x in TBST, and then incubated with the appropriate primary antibody in TBST containing 5% w/v BSA for 2 hours at room temperature with gentle agitation. Antibodies for Erk1/2 (9102, 1:5000), phosphoErk1/2 (9101, 1:5000), and Ctnnb1 (9582, 1:2000) were obtained from Cell Signaling Technology (Beverly, MA, http://www.cellsignal.com) and Gapdh (IMG-5143A, 1:1000) was obtained from Imgenex (San Diego, CA, http://www.imgenex.com). Membranes were then washed 3x in TBST and probed with a HRP conjugated goat anti-rabbit antibody (sc2004, Santa Cruz Biotechnology Inc.) at a 1:20,000 dilution in TBST with 5% w/v nonfat dry milk for one hour. Blots were developed using the Immuno-Star Western C Kit (BioRad Laboratories). X-ray films were developed using a Kodak X-OMAT Processor (Eastman Kodak Co., Rochester, NY, http://www.Kodak.com ) and quantified using ImageJ software (National Institutes of Health).

Statistical Analysis

All data were expressed as mean ± standard deviation. Data from treated groups were compared to untreated groups using the Student’s t test. Differences between treatment groups were considered significant if the p-value was <0.05.


MSCs express a unique compliment of Wnt, Fgf, and Bmp signaling pathway components

Interrogation of the mouse MSC transcriptome, which we catalogued previously by serial analysis of gene expression (SAGE) [32], identified numerous SAGE tags corresponding to transcripts encoding an array of Fgf, Bmp, and Wnt signaling pathway components (Supplementary Table 2). Screening of a mouse MSC cDNA library by PCR confirmed the SAGE tags corresponded to expressed mRNAs encoding the Frizzled (FZD) receptor family members Fzd1, Fzd2, Fzd7, and Fzd10, the Wnt pathway components Apc, Akt2, Axin1, and β-catenin (Ctnnd1) and the Wnt antagonists Sfrp1, Sfrp2, and Dickkopf 3 (Dkk3) (Figure 1A). The cells also expressed transcripts encoding Fgf2, Wnt5a, Bmpr1a, Gremlin and Spry4, an inhibitor of the receptor-transduced mitogen-activated protein kinase (MAPK) signaling pathway [33]. Several transcription factors important for mesoderm specification were also expressed including Runx2, sine-oculis homeobox 1 (Six1), Twist, and Twist2 (not shown). Both Runx2 and Six1 are known to play important roles in the specification of bone and muscle, respectively, during development [1, 34]. Similarly, Twist and Twist2 are evolutionary conserved basic helix-loop-helix transcription factors expressed during development in the mesenchyme of the brachial arches, paraxial and lateral plate mesoderm, and somites where they function to inhibit cellular differentiation [3537].

Figure 1
Validation of catalogued SAGE tags. A) Screening of a mouse MSC cDNA library by PCR using gene specific primers revealed expressed transcripts encoding various receptors, signaling intermediates, and transcription factors that modulate cellular responses ...

Fgf2 dynamically regulates expression of Wnt and Fgf signaling pathway components in MSCs

Culture of MSCs for seven days in media supplemented with 20 ng/ml Fgf2, which was shown previously to inhibit multi-lineage differentiation [19] also significantly altered expression of a subset of the aforementioned genes (Figure 1b). Specifically, levels of Twist2 and Spry4 were induced by 4.5 and 16.5-fold, respectively, while expression of Sfrp2 and Runx2 were down regulated by 3.9-fold and 2.6-fold, respectively. Each of these changes were significantly different (p<0.05) as compared to untreated cells. In addition, expression of Twist, Ccnd1, and β-catenin (Ctnnb1) were up regulated by 1.8, 1.5 and 2.1-fold, respectively, while Pparg was down regulated 1.5-fold but these changes were not significantly different as compared to untreated cells. Exposure over a 14 day time course further revealed that Fgf2 induced dynamic changes in gene expression in MSCs (Figure 2). For example, levels of Spry4 and Ccnd1 were rapidly up regulated within the first 24 hrs of treatment. Thereafter Spry4 remained highly elevated while Ccnd1 mRNA levels gradually declined to near baseline levels by 6–7 days post-treatment. Twist2 mRNA levels were also highly induced but with delayed kinetics as compared to Spry4 and Ccnd1. In contrast, both Twist and Ctnnb1 exhibited modest levels of induction and Sfrp2 and Runx2 were strongly down regulated throughout the time course. Additionally, Sfrp1 mRNA levels were also significantly down regulated over the first week of Fgf2 treatment but then steadily increased to several fold above baseline by the end of the time course. Consequently, expressed levels of Sfrp1 appeared to be inversely correlated to that of Sfrp2.

Figure 2
Fgf2 dynamically regulates gene expression in MSCs. MSCs were cultured for 14 days in CCM alone or CCM supplemented with 20 ng/ml Fgf2. At the indicated time points total RNA was isolated form cells and analyzed by real-time PCR using primers and probes ...

Fgf2 and Wnt3a regulate different signaling intermediates in MSCs

Because Fgf2 altered expression of several known targets of Wnt signaling, such as Cnnd1 and Ctnnb1, we sought to determine whether treatment of MSCs with Wnt3a resulted in similar effects on gene expression. Therefore, MSCs were cultured for 7 days in CCM supplemented with Wnt3a (10 ng/ml), Fgf2 (20 ng/ml) or both growth factors and changes in gene expression were quantified by real-time PCR. As demonstrated previously, Fgf2 up regulated expression of Twist2, Spry4, and Ccnd1 and depressed expression of Sfrp1 to a significant extent as compared to untreated cells (Figure 3A). In contrast, Wnt3a stimulated expression of Twist (4.1-fold) and Ctnnb1 (1.4-fold) and down regulated expression of Sfrp1 (3.2-fold) to a significant extent. Wnt3a also induced expression of Twist2 and Spry4 by 2.1 and 1.6-fold, respectively, and down regulated Sfrp2 expression 1.5-fold but these changes were not significant as compared to untreated cells. Changes in mRNA expression levels observed in cells treated with both Fgf2 and Wnt3a mirrored that seen for each individual growth factor (Figure 3A). To confirm these results, MSCs were exposed to doses of Wnt3a ranging from 2.5–100 ng/ml and expressed levels of Twist, Twist2, and Spry4 were quantified by real-time PCR. As shown in figure 3b, Wnt3a induced Twist expression in a dose dependent fashion, had only a modest effect on Twist2 expression but significantly up regulated Spry4 expression at all concentrations tested as compared to untreated cells. However, overall Spry4 induction levels were modest (~ 2-fold) as compared to that seen with Fgf2 (> 8-fold). The same dose of Wnt3a shown to modulate Twist expression in MSCs was also shown to induce Ctnnb1 mRNA and protein expression, thereby confirming that the mitogen was biologically active (Supplemental Figure 1).

Figure 3
Expression of Twist, Twist2, and Spry4 is differentially regulated by Fgf2 and Wnt3a in MSCs. A) Real-time PCR was used to quantify changes in expressed levels of the indicated genes in primary mouse MSCs cultured for seven days in CCM alone or CCM supplemented ...

Fgf2 and Wnt3a exhibit varying capacities to block cellular differentiation of MSCs

To determine if both Fgf2 and Wnt3a inhibit MSC differentiation, cells were cultured in CCM supplemented with each mitogen alone or in combination for seven days. Thereafter cells were washed repeatedly with CCM and cultured in OIM for 21 days. Staining cultures with Alizarin Red S revealed significant differences in the extent of calcium deposition between treatment groups. Specifically, pre-treatment with Fgf2 was shown to strongly inhibit osteogenic differentiation (Figure 4C) as compared to control cultures (Figure 4A, B). In contrast, pre-treatment with Wnt3a only modestly reduced the extent of Alizarin Red S staining (Figure 4D). As expected, pre-treatment with both mitogens significantly repressed osteogenic differentiation (Figure 4E). Quantifying the amount of Alizarin Red S dye extracted from each cell monolayer confirmed that Fgf2 abrogated mineralization while Wnt3a reduced mineral deposition by a modest but significant (p<0.05) extent (Figure 4F). A similar outcome was observed with respect to chondrogenic differentiation. Pre-treatment of cells with Fgf2 for seven days followed by exposure to CIM for six additional weeks resulted in a complete lack of differentiation as evidenced by the stunted size of the micro-mass pellet and complete absence of chondrocytes (Figure 4G). In contrast, micro mass pellets produced by MSCs pre-treated with Wnt3a appeared to be reduced in size as compared to controls but still retained identifiable chondrocytes and stained weakly with toluidine blue, which binds to sulfated proteoglycans in the extracellular matrix.

Figure 4
Fgf2 but not Wnt3A inhibits osteogenic and chondrogenic differentiation of MSCs. MSCs were cultured in either CCM alone (A, B) or CMM supplemented with 20 ng/ml Fgf2 (C), 10 ng/ml Wnt3a (D), or both growth factors (E) for seven days. Thereafter cells ...

Spry4 induction by Fgf2 is correlated with suppression of Erk1/2 phosphorylation and bone-specific gene expression during osteo-induction

Spry proteins are known to antagonize receptor tyrosine kinase signaling by blocking activation of the Ras/Raf/Erk pathway [38]. Therefore we quantified changes in Erk1/2 phosphorylation in MSCs pre-treated with Fgf2, Wnt3a, and Bmp2 during a time course of osteogenic differentiation. Pretreatment of cells with Fgf2 strongly suppressed Erk1/2 phosphorylation up to day 6 (D6) of the osteo-inductive time course whereas pre-treatment with Wnt3a failed to alter the kinetics of Erk1/2 activation as compared to untreated cells and Bmp2 accelerated the time course of Erk1/2 activation (Figure 5A). Differences in Erk1/2 activation at D6 between the different treatment groups were clearly evident when levels of phosphorylated Erk1/2 were normalized to the total amount of Erk1/2 protein (Figure 5B). Moreover, suppression of Erk1/2 activation by Fgf2 was strongly correlated with its capacity to specifically induced Spry4 mRNA in MSCs, which was not up regulated to a significant extent by Wnt3a or Bmp2 (Figure 5C). Consistent with these results, Fgf2 also strongly suppressed expression of several bone-specific genes in MSCs during the pre-treatment phase and blocked their up regulation during the osteo-induction time course (Figure 5D–G). Some of these proteins promote mineralization of the extra cellular matrix by binding to calcium ions and as such their lack of induction reflects the lack of Alizarin Red S staining in Fgf2-treated MSC cultures (Figure 4C). In contrast, pre-treatment of MSCs with Bmp2 significantly enhanced expressed levels of these genes as compared to untreated cells while the effect of Wnt3a was less pronounced but also positive (Figure 5D–G). Importantly, these effects of Fgf2 appeared to be transient. For example, Spry4 mRNA levels were rapidly down regulated in cells following withdrawal of Fgf2 and exposure to OIM. Also, Erk1/2 phosphorylation levels appeared to increase by day 11 of the osteo-inductive time course and there is a clear trend toward increased bone-specific gene expression at late phases of the osteo-induction time course. These results are consistent with our previous studies showing that osteogenic differentiation was restored in Fgf2-treated MSCs if the mitogen was withdrawn for 7 days prior to osteo-induction [19]. Overall these data demonstrate a strong correlation between Fgf2-induced expression of Spry4 and suppression of Erk1/2 activation with inhibition of bone-specific gene induction and lack of extracellular matrix mineralization in response to osteo-induction.

Figure 5
Effect of mitogens on Erk1/2 activation and bone-specific gene expression

Mitogens differentially regulate expression of fibroblast growth factor receptors in MSCs

We also evaluated expressed levels of Fgfr1-4 in MSCs during a time course of mitogen pre-treatment and subsequent osteogenic differentiation as described above. As shown in Figure 6A, pre-treatment with Fgf2 induced expression of Fgfr1 and Fgfr4 and suppressed expression of Fgfr2 and Fgfr3 when normalized to untreated cells (Figure 6A). Moreover, Fgfr1 expression levels were significantly (p<0.05) greater than Fgfr2 and Fgfr3 at late phases of pre-treatment (P7) and both Fgfr1 and Fgfr4 were expressed at significantly (p<0.05) higher levels than Fgfr3 at early phases of osteo-induction (D2). Expression of the four receptors was strongly down regulated at D11 and D17 of osteo-induction but by D21 expression of these receptors was re-established. Alternatively, Bmp2 pre-treatment had the opposite effect on Fgfr expression levels. For example, Fgfr2 expression was significantly (p<0.05) higher than that of Fgfr1 and Fgfr4 throughout the pre-treatment period and early phases of osteo-induction and Fgfr3 levels were also highly elevated during these early time points but the differences failed to reach statistical significance (Figure 6B). Nevertheless, Fgfr3 expression levels were negatively correlated (r2 = −0.773) between Fgf2 and Bmp2 pre-treated MSCs from P3-D6. At later time points Fgfr4 expression levels remained low. Wnt3a also appeared to suppress expression of Fgfr4 and up regulate expression of Fgfr2 and Fgfr3 but with slightly different kinetics as compared to Bmp2 (Figure 6C). However, Wnt3a had the unique effect of suppressing expression of all four receptors on D21 of osteo-induction. Collectively, these data demonstrate a strong correlation between osteogenic differentiation and up regulation of Fgfr2 and Fgfr3 in MSCs. They also suggest that preferential down regulation of these receptors by Fgf2 contributes to its inhibitory effect on cell differentiation.

Figure 6
Effect of mitogens on Fgfr expression in MSCs

Fgf2 alters the sub cellular distribution of Twist and Spry4 proteins in MSCs

Immuno-fluorescent staining of MSCs cultured in the absence of Fgf2 using an antibody that recognizes both Twist and Twist2 revealed a peri-nuclear staining pattern for this protein (Figure 7A). It also demonstrated that Twist proteins were not uniformly expressed throughout the population. At present it is unclear if this result is due to the fact that Twist proteins are expressed at variable levels, sequestered in complexes that prevent immuno-detection, or are actually restricted to a specific subset of cells with defined function. Staining with an anti-Spry4 antibody yielded a similar expression pattern (Figure 7C, E) and co-staining with both antibodies revealed that Twist and Spry4 were co-expressed approximately 70% of the time in cells wherein one protein was detected (Figure 7D–F). Moreover, when MSCs were cultured in Fgf2 for several days Twist proteins were detected in the nucleus of most cells (Figure 7G, H) and Spry4 became localized to the inner surface of the cell membrane (Figure 7I). These results indicate that Fgf2 not only modulates expression of Twist and Spry4 mRNA but also regulates the function of the encoded proteins by altering their sub cellular distribution in a manner consistent with their mode of action.

Figure 7
Fgf2 alters the sub cellular distribution of Twist and Spry4 proteins in MSCs


To our knowledge, this is the first study to demonstrate that Fgf2 specifically induces Twist2 and Spry4 expression in MSCs and implicate Spry4 as a negative regulator of MSC differentiation via its capacity to inhibit Erk1/2 activation. Similar to our findings, previous studies have linked both canonical Wnt signaling and Fgf2 with expression of Twist. For example, Wnt signaling was shown to be necessary to maintain twist expression in Xenopus neural crest cells [39] and Howe et al. [40] showed that Twist is up regulated in response to Wnt1 in mouse mammary epithelial cells. In the latter case, analysis of the Twist promoter revealed the presence of a consensus TCF binding site and several Ets binding core sequences and transfection studies confirmed the promoter was responsive to Ctnnb1. Moreover, ectopic expression of Twist in epithelial cells was shown to suppress β-casein induction in response to lactogenic hormones consistent with its role as an inhibitor of cellular differentiation. Wnt signaling is also known to play an important role in epithelial-mesenchymal transitions during embryonic development and cancer metastasis by inducing expression of Twist and Slug, which repress E-cadherin gene transcription [41, 42]. Additionally, Fgf signaling was initially implicated in regulating Twist expression due to the fact that mutations in the human genes encoding FGFR2, FGFR3, or TWIST produce craniosynostosis, indicating that these proteins lie along the same signaling pathway that controls osteoblast maturation [4345]. Fgf2 has subsequently been shown to up regulate Twist expression in calvarial explant cultures from mouse embryos and in the developing mouse tooth and palate [46, 47]. Consistent with our studies, Fgf2 has also been shown to up regulate expression of Twist2 in osteoblast cell lines where it was postulated to maintain cells in a progenitor-like state [48]. However, to our knowledge we are the first to report that Wnt3a, which activates the canonical Wnt pathway, and Fgf2 differentially regulate expression of Twist and Twist2 in MSCs. The latter suggests these genes may have distinct functions in MSCs similar to their role in mouse development. For example, Twist null mice die by day 11.5 of embryonic development but mice that lack Twist2 survive into post-natal life [49]. A recent report indicates that Twist2 also represses expression of adipocyte determination and differentiation dependent factor 1 (ADD1), a transcription factor that regulates fatty acid metabolism and insulin-dependent gene expression [50]. Similarly, Foxoa1 was recently shown to play an important role in regulating chondrogenic differentiation of human MSCs, and signaling through PIP2 via Fgfr activation is known to modulate expression of this gene. Collectively, these findings demonstrate how Twist proteins may block multi-lineage differentiation of MSCs following up regulation by Fgf2.

Our data are also the first to demonstrate that Spry4 is strongly up regulated in MSCs following exposure to Fgf2 but not Wnt3a or Bmp2. This result is consistent with studies showing that Spry4 plays an important role in limb development by modulating the activity of Fgfs [54]. Spry4 belongs to a family of proteins that are known to inhibit receptor tyrosine kinase signaling by interacting with GRB2 and/or other accessory proteins, thereby blocking activation of the MAPK pathway [52]. A number of studies have demonstrated that this pathway plays important roles in MSC differentiation. For example, osteogenic differentiation of human MSCs was shown to be accompanied by a sustained phase of Erk1/2 activation and could be blocked by pharmacological inhibitors of Erk [53]. In addition, studies have shown that phosphorylation of Erk1/2 during the initial phase of adipogenic differentiation plays a critical role in cellular differentiation by regulating expression of Pparg and CCAAT/enhancer binding protein (C/EBP), alpha [57]. These results are consistent with our findings demonstrating that Fgf2-induced up regulation of Spry4 is strongly correlated with a block in Erk1/2 phosphorylation and inhibition of osteogenic differentiation. These data are also consistent with our previous reports showing the effect of Fgf2 is reversible since after mitogen withdrawal Erk1/2 activation still occurs at later time points and there is a trend toward increase bone-specific gene expression at late phases of osteo-induction (19). Further studies are underway to dissect the specific role played by Spry4 in regulating adipogenic and chondrogenic differentiation and identifying binding partners for this protein in MSCs.

Our data also indicate that Fgf2 alters the sub cellular localization of Twist and Spry4 proteins. Specifically, exposure of MSCs to Fgf2 resulted in redistribution of Twist proteins into the cell nucleus and Spry4 to the cell membrane. This is consistent with the fact that Twist proteins function as transcriptional repressors and that Spry4 is known to bind to accessory proteins that interact with phosphorylated domains within various receptor tyrosine kinases [38]. Our immunostaining data also suggested that Twist and Srpy4 proteins are not uniformly expressed within MSC populations. At present it is unclear if lack of detection of these proteins in all cells is due to variations in expression levels or if restriction of these proteins to a specific sub population imparts the latter with a specialized function. Recently, we have found that TWIST and TWIST2 mRNA expression levels are significantly higher in tri-potent human MSC clones as compared to clones of lesser potency and parental populations (DGP, personal communication). Therefore, dosage rather than absolute expression levels of these proteins may be critical in determining their function. This is consistent with studies showing that hetero-dimer composition of Twist proteins affects overall protein stability [55] and forced over expression of TWIST or TWIST2 in human MSCs blocks osteogenic and chondrogenic differentiation [56].

Finally, our studies also indicate that extrinsic regulators of cell differentiation including Fgf2 and Bmp2 may regulate fate determination of MSCs, in part, by modulating expression levels of Fgfr proteins. Specifically, both Wnt3a and Bmp2 appeared to selectively up regulate expression of Fgfr2 and Fgfr3 in MSCs, while Fgf2 suppressed expression of these receptors and up regulated Fgfr1 and Fgfr4. Therefore, selective up or down regulation of receptor expression at the cell surface may enhance or reinforce mitogen-induced signals in MSCs. Further examination of these processes will provide new insight into the signaling pathways that regulate MSC self-renewal.

Supplementary Material

Supp Fig S1

Supplemental Figure 1. Wnt3a induces expression of Ctnnb1 mRNA and protein in MSCs:

MSCs were cultured for 3 days in CCM alone or CCM supplemented with Wnt3a (2.5–100 ng/ml). Total RNA prepared from cells was then analyzed by real-time PCR (A) and Western blot (B) to quantify changes in expressed levels of Ctnnb1. A) Relative Ctnnb1 mRNA levels were quantified by real-time PCR and normalized to Gapdh levels. All experiments were run in duplicate and plotted values represent fold differences (mean ± SD) in mRNA levels between control and Wnt3a treated cells. B) Western blot analysis of cytoplasmic extracts from MSCs cultured as described in A demonstrate an increase in Ctnnb1 protein levels following Wnt3a treatment. *,p<0.05.

Supp Table S1&S2


This research was supported in part by a grant from the National Institute of Health to (D.G.P., 1 R01 NS052301-01A2), the Louisiana Gene Therapy Research Consortium (New Orleans, LA), and HCA-the Health Care Company (Nashville, TN).



The authors indicate no potential conflicts of interest.

Author contribution

W-T. L.: Collection and/or assembly of data, manuscript writing.

V. K.: Collection and/or assembly of data.

D. P.: Conception and design, financial support, data analysis and interpretation, manuscript writing, final approval of manuscript.


1. Komori T, Yagi H, Nomura S, et al. Targeted disruption of Cbfa1 results in a complete lack of bone formation owing to maturational arrest of osteoblasts. Cell. 1997;89:755–764. [PubMed]
2. Shui C, Spelsberg TC, Riggs BL, et al. Changes in Runx2/Cbfa1 expression and activity during osteoblastic differentiation of human bone marrow stromal cells. J Bone Miner Res. 2003;18:213–221. [PubMed]
3. Rosen ED, Spiegelman BM. PPARgamma: a nuclear regulator of metabolism, differentiation, and cell growth. J Biol Chem. 2001;276:37731–37734. [PubMed]
4. Gimble JM, Robinson CE, Wu X, et al. Peroxisome proliferator-activated receptor-gamma activation by thiazolidinediones induces adipogenesis in bone marrow stromal cells. Mol Pharm. 1996b;50:1087–1094. [PubMed]
5. Bi W, Deng JM, Zhang Z, et al. Sox9 is required for cartilage formation. Nat Genet. 1999;22:85–89. [PubMed]
6. Tsuchiya H, Kitoh H, Sugiura F, et al. Chondrogenesis enhanced by overexpression of sox9 in mouse bone marrow-derived mesenchymal stem cells. Biochem Biophys Res Commun. 2003;301:338–343. [PubMed]
7. Karow M, Popp T, Egea V, et al. Wnt signaling in mouse mesenchymal stem cells: impact on proliferation, invasion and MMP expression. J Cell Mol Med. 2009;13:2506–2520. [PubMed]
8. Baksh D, Boland GM, Tuan RS. Cross-talk between Wnt signaling pathways in human mesenchymal stem cells lead to functional antagonism during osteogenic differentiation. J Cell Biochem. 2007;101:1109–1124. [PubMed]
9. Gregory CA, Singh H, Perry AS, et al. The Wnt signaling inhibitor dickkopf-1 is required for reentry into the cell cycle of human adult stem cells from bone marrow. J Biol Chem. 2003;278:28067–28078. [PubMed]
10. Boland GM, Perkins G, Hall DJ, et al. Wnt3a promotes proliferation and suppresses osteogenic differentiation of adult human mesenchymal stem cells. J Cell Biochem. 2004;93:1210–1230. [PubMed]
11. de Boer J, Siddappa R, Gaspar C, et al. Wnt signaling inhibits osteogenic differentiation of human mesenchymal stem cells. Bone. 2004;34:818–826. [PubMed]
12. Gaur T, Lengner CJ, Hovhannisyan H, et al. Canonical WNT signaling promotes osteogenesis by directly stimulating Runx2 gene expression. J Bio Chem. 2005;280:33132–33140. [PubMed]
13. Tang N, Song WX, Luo J, et al. BMP-9-induced osteogenic differentiation of mesenchymal progenitors requires functional canonical Wnt/beta-catenin signaling. J Cell Mol Med. 2009;13:2448–2464. [PubMed]
14. Bodine PV, Zhao W, Kharode YP, et al. The Wnt antagonist secreted frizzled-related protein-1 is a negative regulator of trabecular bone formation in adult mice. Mol Endocrinol. 2004;18:1222–1237. [PubMed]
15. Tuli R, Tuli S, Nandi S, et al. Transforming growth factor-beta-mediated chondrogenesis of human mesenchymal progenitor cells involves N-cadherin and mitogen-activated protein kinase and Wnt signaling cross-talk. J Biol Chem. 2003;278:41227–41236. [PubMed]
16. Eijken M, Meijer IM, Westbroek I, et al. Wnt signaling acts and is regulated in a human osteoblasts differentiation dependent manner. J Cell Biochem. 2008;104:568–579. [PubMed]
17. Ling L, Nurcombe V, Cool SM. Wnt signaling controls the fate of mesenchymal stem cells. Gene. 2009;43:1–7. [PubMed]
18. van den Bos C, Mosca JD, Winkles J, et al. Human mesenchymal stem cells respond to fibroblast growth factors. Hum Cell. 1997;10:45–50. [PubMed]
19. Baddoo M, Hill K, Wilkinson R, et al. Characterization of mesenchymal stem cells isolated from murine bone marrow by negative selection. J Cell Biochem. 2003;89:1235–1249. [PubMed]
20. Choi SC, Kim SJ, Choi JH, et al. Fibroblast growth factor-2 and -4 promote the proliferation of bone marrow mesenchymal stem cells by activation of the PI3K-Akt and ERK1/2 signaling pathways. Stem Cells Dev. 2008;17:725–736. [PubMed]
21. Ahn HJ, Lee WJ, Kwack K, et al. FGF2 stimulates the proliferation of human mesenchymal stem cells through transient activation of JNK signaling. FEBS Lett. 2009;583:2922–2926. [PubMed]
22. Tsutsumi S, Shimazu A, Miyazaki K, et al. Retention of multilineage differentiation potential of mesenchymal stem cells during proliferation in response to FGF. Biochem Biophys Res Commun. 2001;288:413–419. [PubMed]
23. Bianchi G, Banfi A, Mastrogiacomo N, et al. Ex vivo enrichment of mesenchymal cell progenitors by fibroblast growth factor 2. Exp Cell Res. 2003;287:98–105. [PubMed]
24. Hanada K, Dennis JE, Caplan AI. Stimulatory effects of basic fibroblast growth factor and bone morphogenetic protein-2 on osteogenic differentiation of rat bone marrow-derived mesenchymal stem cells. J Bone Miner Res. 1997;12:1606–1614. [PubMed]
25. Scutt A, Bertram P. Basic fibroblast growth factor in the presence of dexamethasone stimulates colony formation, expansion, and osteoblastic differentiation by rat bone marrow stromal cells. Calcif Tissue Int. 1999;64:69–77. [PubMed]
26. Solchaga LA, Penick K, Proter JD, et al. FGF-2 enhances the mitotic and chondrogenic potentials of human adult bone marrow-derived mesenchymal stem cells. J Cell Physiol. 2005;203:398–409. [PubMed]
27. Varas L, Ohlsson LB, Honeth G, et al. Alpha10 integrin expression is up-regulated on fibroblast growth factor-2-treated mesenchymal stem cells with improved chondrogenic differentiation potential. Stem Cell Dev. 2007;16:965–978. [PubMed]
28. Neubauer M, Fischbach C, Bauer-Kreisel P, et al. Basic fibroblast growth factor enhances PPARgamma ligand-induced adipogenesis of mesenchymal stem cells. FEBS Lett. 2004;577:277–283. [PubMed]
29. Debiais F, Hott M, Graulet AM, et al. The effects of fibroblast growth factor-2 on human neonatal calvaria osteoblastic cells are differentiation stage specific. J Bone Miner Res. 1998;13:645–654. [PubMed]
30. Russell KC, Phinney DG, Lacey MR, et al. In vitro high-capacity assay to quantify the clonal heterogeneity in tri-lineage potential of mesenchymal stem cells reveals a complex hierarchy of lineage commitment. Stem Cells. 2010;28:788–798. [PubMed]
31. Phinney DG, Hill K, Michelson C, et al. Biological activities encoded by the murine mesenchymal stem cell transcriptome provide a basis for their developmental plasticity and broad clinical efficacy. Stem Cells. 2006;24:186–198. [PubMed]
32. Phinney DG. Isolation of mesenchymal stem cells from murine bone marrow by immunodepletion. Methods Mol Biol. 2008;449:171–186. [PubMed]
33. Christofori G. Split personalities: the agonistic antagonist Sprouty. Nature Cell Biology. 2003;5:427–432. [PubMed]
34. Ridgeway AG, Skerjanc IS. Pax3 is essential for skeletal myogenesis and the expression of Six1 and Eya2. J Biol Chem. 2001;276:19033–19039. [PubMed]
35. Getelman I. Twist protein in mouse embryogenesis. Dev Biol. 1997;189:205–214. [PubMed]
36. Fuchtbauer EM. Expression of M-twist during postimplantation development of the mouse. Dev Dyn. 1995;204:316–322. [PubMed]
37. Wolf C, Thisse C, Stoetzel C, et al. The M-twist gene of Mus is expressed in subsets of mesodermal cells and is closely related to the Xenopus X-twi and the Drosophila twist genes. Dev Biol. 1991;143:363–373. [PubMed]
38. Guy GR, Jackson RA, Yusoff P, Chow SY. Sprouty proteins: modified modulators, matchmakers or missing link? J Endorcrinology. 2009;203:191–202. [PubMed]
39. Borchers A, David R, Wedlich D. Xenopus cadherin-11 restrains cranial neural crest migration and influences neural crest specification. Development. 2001;128:3049–3060. [PubMed]
40. Howe LR, Watanabe O, Leonard J, et al. Twist is up-regulated in response to Wnt1 and inhibits mouse mammary cell differentiation. Cancer Res. 2003;63:1906–1913. [PubMed]
41. Heuberger J, Birchmeier W. Interplay of cadherin-mediated cell adhesion and canonical wnt signaling. Cold Spring Harb Perspect Biol. 2010;2:a002915. [PMC free article] [PubMed]
42. Yang J, Mani SA, Donaher JL, et al. Twist, a master regulator of morphogenesis, plays an essential role in tumor metastasis. Cell. 2004;117:927–939. [PubMed]
43. Howard TD, Paznekas WA, Green ED, et al. Mutation in TWIST, a basic helix–loop−helix transcription factor in Saethre-Chotzen syndrome. Nat Genet. 1997;15:36–41. [PubMed]
44. El Ghouzzi V, Le Merrer M, Perrin-Schmitt F, et al. Mutations of the TWIST gene in the Saethre- Chotzen syndrome. Nat Genet. 1997;15:42–46. [PubMed]
45. Muenke M, Gripp KW, McDonald-McGinn, et al. A unique point mutation in the Fibroblast Growth Factor Receptor 3 gene (FGFR 3) defines a new craniosynostosis syndrome. Am J Hum Genet. 1997;60:555–564. [PMC free article] [PubMed]
46. Rice DP, Aberg T, Chan Y, et al. Integration of FGF and TWIST in calvarial bone and suture development. Development. 2000;127:1845–1855. [PubMed]
47. Rice R, Thesleff I, Rice DP. Regulation of Twist, Snail, and Id1 is conserved between the developing murine palate and tooth. Dev Dyn. 2005;234:28–35. [PubMed]
48. Lee MS, Lowe G, Flanagan S, et al. Human Dermo-1 has attributes similar to twist in early bone development. Bone. 2000;27:591–602. [PubMed]
49. Chen ZF, Behringer RR. Twist is required in head mesenchyme for cranial neural tube morphogenesis. Genes Dev. 1995;9:686–699. [PubMed]
50. Lee YS, Lee HH, Park J, et al. Twist2, a novel ADD1/SREBP1c interacting protein, represses the transcriptional activity of ADD1/SRECP1c. Nucleic Acids Res. 2003;31:7165–7174. [PMC free article] [PubMed]
51. Taniguchi K, Ayada T, Ichiyama K, et al. Sprouty2 and sprouty4 are essential for embryonic morphogenesis and regulation of FGF signaling. Biochem Biophys Res Commun. 2007;352:896–902. [PubMed]
52. Cabrita MA, Christofori G. Sprouty proteins, masterminds of receptor tyrosine kinase signaling. Angiogenesis. 2008;11:53–62. [PubMed]
53. Jaiswal RK, Jaiswal N, Bruder SP, et al. Adult human mesenchymal stem cell differentiation to the osteogenic and adipogenic lineage is regulated by mitogen-activated protein kinase. J Biol Chem. 2000;275:9645–9652. [PubMed]
54. Prusty D, Park BH, Davis KE, et al. Activation of MEK/ERK signaling promotes adipogenesis by enhancing peroxisome proliferator-activated receptor gamma (PPARgamma) and C/EBPalpha gene expression during the differentiation of 3T3-L1 preadipocytes. J Biol Chem. 2002;277:46226–46232. [PubMed]
55. Cakouros D, Raices RM, Gronthos S, Glackin CA. Twist-ing cell fate: mechanistic insights into the role of twist in lineage specification/differentiation and tumorigenesis. J Cell Biochem. 2010;110:1288–1298. [PubMed]
56. Isenmann S, Arthur A, Zannettino AC, et al. Twist family of basic helix-loop-helix transcription factors mediate human mesenchymal stem cell growth and commitment. Stem Cells. 2009;27:2457–2468. [PubMed]
PubReader format: click here to try


Cited by other articles in PMC

See all...


Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...