Activation of PKA/CREB Signaling is Involved in BMP9-Induced Osteogenic Differentiation of Mesenchymal Stem Cells

Mesenchymal stem cells (MSCs) are non-hematopoietic multipotent cells and can differentiate into osteoblastic, chondrogenic and other lineages [1,2,3,4,5,6,7,8]. Bone morphogenetic proteins (BMPs) are a group of secreted signaling proteins belonging to the Transforming Growth Factor-β (TGF-β) superfamily. Originally discovered by their ability to induce bone and cartilage formation, BMPs are now proved to play pivotal roles in almost all organ systems [9,10,11,12]. Genetic disruptions of the major players in BMPs signaling cause embryonic lethality, or result in various skeletal and/or extra-skeletal abnormalities during development [5,9,13]. To date, about 20 BMPs family members have been identified and characterized. Several forms of recombinant BMPs, most notably BMP2 and BMP7, have been validated to promote osteogenesis and are now used as adjunctive therapy in the clinical setting [14,15,16,17,18]. However, it is uncertain if BMP2 and BMP7 are in fact the most potent BMPs in inducing bone formation.

BMP9 (also known as growth differentiation factor 2, or GDF2) was originally isolated from fetal mouse liver cDNA libraries [19]. BMP9 was shown to induce the cholinergic phenotype of embryonic basal forebrain cholinergic neurons [20], regulate glucose and lipid metabolism [21], maintain the homeostasis of iron metabolism [22], and control lymphatic vessel maturation [23]. In addition, BMP9 participated in various pathogenic processes including hereditary hemorrhagic telangiectasia, breast cancer, ovarian cancer, osteosarcoma and hepatocellular carcinoma [24,25,26,27,28]. In recent studies, BMP9 has been proved to be most highly capable of inducing osteogenic differentiation of MSCs by regulating a distinct set of downstream targets [11,29,30,31]. Our previous study demonstrated that ALK1 and ALK2 are the type I TGF-β receptors responsible for BMP9 osteogenic signaling in MSCs [32]. We also revealed that BMPR II and ActR II are the functional type II TGF-β receptors facilitating BMP9 osteogenic signaling [33]. Many signaling molecules such as IGF, MAPKs, Wnt, growth hormone (GH), FGF2, Notch and TGF-β have been found to modulate BMP9-induced osteogenic differentiation of MSCs [34,35,36,37,38,39,40,41]. Moreover, a recent study demonstrated that TSA (a potent inhibitor of histone deacetylase) can potentiate BMP9-induced early osteogenic differentiation of MSCs, implying that epigenetic modification may be involved in BMP9-induced osteogenesis [42]. Despite these valuable discoveries, BMP9 remains as one of the least studied BMPs. The detail molecular mechanism underlying BMP9-induced osteogenic differentiation of MSCs is still largely unknown and warrants extensive studies.

Protein kinase A (PKA, also known as cAMP-dependent protein kinase) is an enzyme whose activity is highly regulated by fluctuating levels of cyclic adenosine monophosphate (cyclic AMP, cAMP) within cells [43,44]. PKA, when allosterically activated by cAMP, can activate cAMP response element binding protein (CREB) by phosphorylating CREB at serine and threonine residues [43,44,45]. Phosphorylated CREB then become competent to bind promoter regions of responsive genes and regulate transcription of CREB targets. PKA/CREB signaling is shown to participate in various physiological and pathological processes such as metabolism, proliferation, differentiation, memory, inflammation and carcinogenesis [46,47,48,49,50,51,52,53]. The role of PKA/CREB signaling in osteogenesis and bone formation has been studied previously in different cell types with cytokines/compounds to activate or inhibit PKA/CREB signaling [54,55,56,57,58,59,60,61,62,63], although the obtained results are contentious. Nevertheless, these studies strongly support the notion that PKA/CREB signaling is important in osteogenesis and bone development.

In the present study, we try to probe the possible involvement and exact role of PKA/CREB signaling in BMP9-induced osteogenic differentiation of MSCs. We find that BMP9 can effectively promote activation of PKA/CREB signaling in MSCs, possibly by down-regulating the expression of protein kinase A inhibitor γ (PKIγ). Using PKA inhibitors H89 and KT5720, we demonstrate that inhibition of PKA activity reduces BMP9-induced early osteogenic marker ALP activity. Accordingly, BMP9-induced expression of late osteogenic markers OPN and OCN, as well as matrix mineralization is significantly reduced by PKA inhibitors. PKA inhibitor H89 is shown to disrupt BMP9-activated p38 signaling, and inhibit Runx2 activation. Moreover, we find that interference of CREB function by a dominant negative form of CREB protein (A-CREB) suppresses BMP9-induced osteogenic differentiation of MSCs as well. Together, our results strongly suggest that PKA/CREB signaling may play a regulatory role in BMP9-induced osteogenic differentiation of MSCs.

C3H10T1/2, C2C12 were obtained from ATCC (Manassas, VA). The immortalized mouse embryo fibroblasts (iMEFs) were established and characterized as previously described [64]. Cell lines were maintained in complete DMEM medium supplemented with 10% fetal calf serum (FCS, Gibco) and antibiotics. H89 was obtained from Sigma (St. Louis, MO), and KT5729 was obtained from Santa Cruz Biotechnology (Dallas, TX). A-CREB plasmid (pCMV500-A-CREB) was obtained from Addgene (Cambridge, MA). Unless otherwise indicated, all chemicals were purchased from Sigma (St. Louis, MO).

Recombinant adenoviruses expressing BMP9 (Ad-BMP9) were generated previously using the AdEasy system, as demonstrated [29,30,31,32,33,34,35,36,37,38,39,40,41,42]. Analogous adenoviruses expressing only GFP (Ad-GFP) were used as a control.

The PKA activity in MSCs was measured by a nonradioactive assay kit (Enzo Life Sciences, New York, NY). Briefly, MSCs (30% confluence) were infected by Ad-BMP9. At 12h, 24h, 36h and 48h post infection, cells were lysed in PKA extraction buffer (25 mM Tris-HCl pH 7.4, 0.5 mM EDTA, 0.5 mM EGTA, 10 mM β-mercaptoethanol, 1 µg/ml leupeptin, 1 µg/ml aprotinin) and subject to the PKA assay according to manufacturer's instructions.

Western blot was performed as previously described [29,30,31,32,33,34,35,36,37,38,39,40,41]. At the indicated time point, cells were lysed. Then, proteins were transferred to a polyvinylidene difluoride (PVDF) membrane. Membrane was blocked with Super-Block Blocking Buffer, and probed with the primary antibody, followed by incubation with a secondary antibody conjugated with horseradish peroxidase. The proteins of interest were detected with SuperSignal West Pico Chemiluminescent Substrate (Thermo Scientific, IL) and semi-quantitated with Quantity One software. Primary antibodies were obtained from Santa Cruz, as follows: anti-CREB, anti-phosphor-CREB, anti-PKIγ, anti-OPN, anti-OCN, anti-phosphor-p38, anti-p38, anti-phosphor-ERK1/2, anti-ERK1/2, anti-phosphor-Smad1/5/8, anti-Smad1/5/8, anti-Runx2 and anti-β-actin.

The intracellular cAMP level was detected by a direct cAMP enzyme immunoassay kit (Enzo Life Sciences, New York, NY). Briefly, MSCs (30% confluence) were infected with Ad-BMP9. At the indicated time point, cells were lysed in 0.1M HCl. The cell lysates were centrifuged at room temperate and the supernatant was directly subject to the cAMP assay according to manufacturer's recommendations.

Total RNA was isolated with Trizol Reagents (Invitrogen, Carlsbad, CA) and used to generate cDNA templates by RT reaction with hexamer and Superscript II RT (New England Biolabs., Ipswich, MA). Semi-quantitative RT-PCR was carried out as described previously [29,30,31,32,33,34,35,36,37,38,39,40,41,42]. The specificity of PCR products was confirmed by 1.5% agarose gel electrophoresis and semi-quantitated with Quantity One software. All samples were normalized by the expression level of GAPDH. For qPCR analysis, SYBR Green-based qPCR assay was carried out by employing 7000 Real-Time PCR system with a standard pUC19 plasmid as described elsewhere [32,33].

ALP activity was assessed with a modified Great Escape SEAP chemiluminescence quantitative assay (BD Clontech, Mountain View, CA) and/or a histochemical staining assay (using a mixture of 0.1 mg/ml napthol AS-MX phosphate and 0.6 mg/ml Fast Blue BB salt) as described [29,30,31,32,33,34,35,36,37,38,39,40,41,42].

Matrix mineralization was detected by Alizarin Red S stain, as described [29,30,31,32,33,34,35,36,37,38,39,40,41,42]. The staining of matrix mineralization was recorded under bright field microscopy.

Cells were seeded in 25 cm² cell culture flasks at 30% confluence and transfected with 2mg of Smad1/5/8 responsive element luciferase reporter, p12xSBE-Luc [32,33], or Runx2-binding sites (OSE2) luciferase reporter, p6xOSE-Luc [39] using Lipofectamine (Invitrogen, Carlsbad, CA). At 16h after transfection, cells were seeded to 24 well plates and infected with Ad-BMP9 in the presence of H89. At the indicated time point, cells were lysed and subjected to luciferase assay using Promega's Luciferase Assay Kit (Promega, Madison, WI), as described [32,33,39].

All quantitative experiments were performed in triplicate and/or repeated three times. Data were expressed as mean ± S.D. Statistical significances between treatment vs control were determined by one-way analysis of variance and Student's t test. A value of p<0.05 was taken as the level of statistically significance.

First of all, we sought to determine if BMP9 can activate PKA/CREB signalingin common-used MSC lines C3H10T1/2, C2C12 and our recently established MSC line iMEFs. Using a nonradioactive assay, we found that PKA activity in MSCs was significantly increased at 12h, 24h, 36h and 48h post BMP9 stimulation (Fig. 1A). Accordingly, we found that BMP9 was able to promote CREB phosphorylation, without altering the total amounts of CREB protein in MSCs (Fig. 1B, Fig. 1C and Fig. 1D). However, H89, a potent PKA inhibitor, was found to reduce BMP9-induced phosphorylation of CREB effectively (Fig. 1B, Fig. 1C and Fig. 1D). These results strongly suggest that BMP9 may effectively induce activation of PKA/CREB signaling in MSCs.

Next, we sought to analyze the possible mechanism underlying BMP9-induced activation of PKA/CREB signaling in MSCs. As cAMP is a well-characterized activator of PKA/CREB signaling, we therefore postulated that BMP9 may increase cAMP level and subsequently activate PKA/CREB signaling in MSCs. Interestingly, we found that the intracellular cAMP level was not altered by BMP9 treatment (Fig. 2A), implying that cAMP level was probably not relevant to BMP9-induced PKA/CREB signaling. Notably, we found that the expression of protein kinase A inhibitor γ (PKIγ) was down-regulated at both mRNA and protein level upon BMP9 treatment in MSCs (Fig. 2B and Fig. 2C). Taken together, these above results suggest that BMP9 may activate PKA/CREB signaling in MSCs, possibly by down-regulating the expression of PKIγ.

As PKA/CREB signaling was activated by BMP9 treatment, we sought to explore if PKA/CREB signaling has any effects on BMP9-induced early osteogenic differentiation of MSCs. When C3H10T1/2 cells were treated with BMP9 in the presence of varying concentrations of H89 (0, 1, 2.5, 5 and 10 µM), we found that H89 can inhibit BMP9-induced ALP activity mostly in a concentration dependent manner (Fig. 3A and Fig. 3B). In addition, we obtained similar results of H89 on BMP9-induced ALP activity in C2C12 and iMEFs cells (Fig. 3C and Fig. 3D). To avoid nonspecific and/or overlap effects of H89, we employed another PKA inhibitor KT5720 and found that this PKA inhibitor can reduce BMP9-induced ALP activity in MSCs as well (Fig. 3E and Fig. 3F). These results strongly suggest that PKA may regulate BMP9-induced early osteogenic differentiation of MSCs.

Although ALP is a well-established early osteogenic marker, it is hardly an accurate predictor of the late stage of osteogenic differentiation [11,32,33]. Thus, we sought to determine if PKA inhibitors have any effects on BMP9-induced late osteogenic markers, such as matrix mineralization, OPN and OCN expression. When C3H10T1/2 and C2C12 cells were exposed to BMP9 and treated with H89 and KT5720, we found that these two PKA inhibitors can both impair BMP9-induced in vitro matrix mineralization (Fig. 4A and Fig. 4B). Moreover, we found that H89 can effectively decrease BMP9-induced OPN and OCN expression at mRNA (Fig. 4C and Fig. 4D) and protein levels (Fig. 4E) in C3H10T1/2 MSCs. Taken together, these above results strongly suggest that PKA may regulate BMP9-induced osteogenic differentiation of MSCs.

Runt-related transcription factor 2 (Runx2) is a key transcription factor associated with osteogenesis [65,66]. Runx2 can directly stimulate expression of most of the well-established osteogenic markers, including osteocalcin (OCN) [66]. Our previous studies have demonstrated that BMP9 can promote Runx2 activation in MSCs [11,39]. Herein we asked whether BMP9-induced activation of Runx2 was also affected by PKA activity in MSCs. We found that BMP9-induced Runx2 expression was reduced by PKA inhibitor H89 at mRNA (Fig. 5A) and protein levels (Fig. 5B). Moreover, using a commonly used Runx2-regulated OCN promoter reporter (p6xOSE-luc), we found that H89 treatment was able to reduce BMP9-induced luciferase activity of p6xOSE-Luc reporter (Fig. 5C), which contains Runx2-responsive elements and reflects Runx2 transcriptional activity [32,33,39]. These results suggest that PKA may modulate BMP9-induced Runx2 activation in MSCs.

Next, we sought to explore the possible mechanism behind the effect of PKA on BMP9-induced osteogenic differentiation of MSCs. Smad1/5/8 signaling is the classical pathway for BMP9 to induce osteogenic differentiation of MSCs [32,33], so we tried to confirm whether PKA can regulate this signaling or not. We found that PKA inhibitor H89 had no apparent effect on phosphorylation of Smad1/5/8 induced by BMP9 in C3H10T1/2 MSCs (Fig. 5D). Moreover, using luciferase reporter assay, we demonstrated that H89 did not alter BMP9-induced Smad1/5/8-responsive luciferase activity (Fig. 5E). These date suggest that PKA may not affect BMP9-induced activation of Smad1/5/8 signaling. Our previous studies showed that p38 and ERK1/2 MAPKs may play important role in BMP9-induced osteogenic differentiation of MSCs as well [35,36]. We next sought to analyze if PKA would exert an influence on these two signals. We found that PKA inhibitor H89 effectively suppressed phosphorylation of p38 in BMP9-stimulated C3H10T1/2 MSCs, while H89 did not affect ERK1/2 phosphorylation (Fig. 5F). These above results imply us that PKA may crosstalk with p38 signaling to regulate BMP9-induced osteogenic differentiation of MSCs.

CREB is a well-characterized mediator of PKA signaling [43,44,45]. As BMP9 can promote phosphorylation/activation of CREB, we next sought to investigate the role of CREB in BMP9-induced osteogenic differentiation of MSCs. Using a dominant negative form of CREB (A-CREB), we found that disruption of CREB function by A-CREB diminished BMP9-induced ALP activity (Fig. 6A, Fig. 6B and Fig. 6C) and matrix mineralization of MSCs (Fig. 6D). Furthermore, we demonstrated that suppression of CREB function by A-CREB effectively inhibited BMP9-induced protein expression of OPN, OCN and Runx2 in C3H10T1/2 MSCs (Fig. 6E and Fig. 6F). Collectively, these above results suggest that activation of PKA/CREB signaling is probably pivotal in mediating BMP9-induced osteogenic differentiation of MSCs.

Previous studies have evidenced that BMP9 is probably the most potent BMPs in inducing osteogenic differentiation of MSCs by regulating several important downstream molecules [11,29,30,31]. However, the precise molecular mechanism involved remains to be fully elucidated. Therefore, we are particularly interested in exploring downstream signaling pathway(s) involved in BMP9 osteoinductive activity.

The functions of PKA/CREB signaling in osteogenesis have been previously investigated both in vitro and in vivo. However, the obtained results are contentious. Some studies proposed a stimulatory role of PKA/CREB signaling in osteogenesis [53,54,55,56,57,58], others substantiated an inhibitory role of PKA/CREB [59,60], and still others supported that the role of PKA/CREB in osteogenesis is diverse depending on the stimulus, dosage and timing [61,62,63]. Although these studies did not reach complete unanimity, it is well accepted that PKA/CREB signaling may play a functional role in osteogenesis and bone development [53,54,55,56,57,58,59,60,61,62,63]. In this report, we investigate the exact role of PKA/CREB signaling in BMP9-induced osteogenic differentiation of MSCs, and the possible mechanism involved. We find that BMP9 is able to activate PKA/CREB signaling probably possibly by down-regulating PKIγ expression in MSCs. PKA inhibitors are shown to suppress BMP9-induced osteogenic differentiation of MSCs. Mechanistically, we find that inhibition of PKA activity by PKA inhibitor reduces BMP9-induced Runx2 activation and p38 phosphorylation in MSCs. Furthermore, interference of CREB function by A-CREB decreases BMP9-induced osteogenic differentiation of MSCs as well. These results strongly suggest that PKA/CREB signaling may play a critical role in regulating BMP9-induced osteogenic differentiation of MSCs.

Generally, PKA activity is controlled by intracellular cAMP level. cAMP is ubiquitously found in all mammalian cells and acts as second messenger in many biological processes such as proliferation, differentiation and metabolism. cAMP-induced PKA activation has been found to control several physiological and pathological processes including chondrogenesis, retina regeneration, tumor angiogenesis and cancer cells metastasis [67,68,69,70,71]. In this current report, however, we did not detect any obvious increase of intracellular cAMP level in BMP9-treated MSCs. It seems that other event(s) rather than cAMP level may contribute to BMP9-induced PKA activity in MSCs. PKA activity can also be modulated by endogenous protein kinase inhibitors (PKIs), which act as pseudosubstrates for PKA by binding and inactivating C-subunits through their consensus sequence [72,73,74]. To date, three specific PKI isoforms have been characterized in both human (PKIA, PKIB and PKIG) and mouse (PKIα, PKIβ and PKIγ) [72,73]. The tissue expression pattern of each mouse PKIs isoform is unique. The PKIα is most highly expressed in heart, skeletal muscle, cerebral cortex, and cerebellum, the PKIβ is predominately expressed in testis, whereas the PKIγ is detected in most tissues throughout the body [72]. A previous study demonstrated that BMP2-mediated PKIγ down-regulation may be prerequisite for the PKA activation during the osteoblastic differentiation of C2C12 cells [74]. In this current study, we demonstrated that the PKIγ expression in MSCs was decreased upon BMP9 treatment. These results suggest that down-regulation of PKIγ expression is probably an important event responsive for BMP9-induced activation of PKA/CREB signaling in MSCs.

CREB is a transcription factor belonging to ATF/CREB transcription factor family which consists of different ATFs including ATF, CREB, CREM (cAMP response element modulator) and related proteins [44,75,76]. It has been described by various studies that ATF/CREB members including CREB may play a role in osteogenesis. However, the exact effects of ATF/CREB family on osteogenesis are diverse and disputable, depending on the extracellular stimuli, the type of ATF/CREB members and context of specific cells. For example, ATF4 promotes osteocalcin gene expression and osteoblastic differentiation [77]. ATF4 mediates parathyroid hormone (PTH)-induced osteoblastic differentiation and bone formation [78]. Ablation of ATF4 in mice leads to severe skeletal defects, including delayed ossification and low bone mass, short stature and short limbs [79]. On the other hand, another report described that ATF3 suppresses BMP2-induced ALP expression and activation in MC3T3-E1 cells [80]. Notably, Long F, et al disrupted the function of CREB protein in transgenic mice, and found that the expression pattern of certain key signaling molecules in developing bones is markedly attenuated [81]. This in vivo study implied that CREB is required for bone development. In this current report, we found that inhibition of CREB function can reduce BMP9-induced osteogenic differentiation of MSCs. Thus, CREB may be a key regulatory molecule required for BMP9 osteoinductive activity.

In conclusion, our results demonstrated that BMP9 can activate PKA/CREB signaling in MSCs possibly through down-regulation of PKIγ expression. Notably, using specific inhibitors for PKA and CREB, we found that activation of PKA/CREB signaling is essential for BMP9-induced osteogenic differentiation of MSCs. This knowledge will provide insights into the molecular mechanisms by which BMP9 induces osteogenic differentiation of MSCs. Future studies should be devoted to the elucidation of detail cross-talk between PKA/CREB signaling and other signal molecules in the context of BMP9-induced osteogenic differentiation of MSCs and bone formation.

This work was supported by grant from the National Natural Science Foundation of China (81272006), and by the Natural Science Foundation Project of Chongqing Science and Technology Commission (cstc2013jcyjA10061).

The authors declare no conflict of interest.