Ras-MAPK Signaling in Osteogenic Differentiation: Friend or Foe?

The role of Ras-mitogen–activated protein kinase (MAPK) signaling in osteogenic differentiation is currently a source of great controversy. On one side is robust evidence from in vitro studies, pharmacological therapy, and genetic disease indicating that increased Ras-MAPK signaling can antagonize bone formation. In this model, committed osteoprogenitors have a choice of either proliferation or further differentiation, with Ras-MAPK being a key regulator of that balance. In contrast, several conflicting in vitro studies have implied that Ras-MAPK is vital for supporting Runx2/Cbfa1 activity and subsequent osteogenic gene expression. As evidence mounts on both sides of the argument, it is important to step back and critically analyze the approaches taken thus far and the relative strength of their conclusions. This perspective attempts to provide a concise and unbiased summary of both lines of evidence to better understand the conflicting models. A particular focus has been placed on comparing differences in methodology. Whereas no single factor emerges as expedient to account for the divergent views, a clear lack of uniformity in experimental methods is evident. Perhaps the greatest variation was seen in the use of numerous cell lines and primary cell types at various stages of osteogenic commitment and differentiation. We speculate that uncommitted multipotent progenitors, committed osteoprogenitors, mature osteoblasts, and cells derived from other lineages may respond differently to identical stimuli, thus explaining some of the apparent conflict in the literature. As well as experimental consistency, we propose that systematic approaches for the independent examination of osteogenic commitment and osteogenic differentiation are of paramount importance for future progress. The Ras-MAPK signaling pathway can profoundly modulate cell proliferation, differentiation, and survival.1-4 Ras is a small GTPase that serves as a central point of convergence and divergence for a number of critical cell signaling pathways. Numerous receptor tyrosine kinases can activate Ras by stimulating guanosine diphosphate (GDP) guanosine triphosphate/(GTP) exchange, and Ras proteins are also functionally regulated by post-translational farnesylation and palmitoylation.5 There are several Ras isoforms (H-ras, K-ras, N-ras), and they are differentially stimulated by the diverse upstream effectors of Ras. These isoforms also show distinct patterns of post-translational modification and subcellular trafficking.5, 6 Many targets have been identified as downstream of Ras. These include the small GTPases Rac and Ral as well as the Jun N-terminal kinase (JNK), phosphatidylinositol 3 phosphate kinase (PI3-K), p38 kinase (p38K), and NF-κB pathways, but the classical and oft-cited target is the Raf-1/mitogen-activated, ERK-activating kinase (MEK)/MAPK axis.1 Ras-MAPK signaling operates through a well-described kinase cascade that translates extracellular stimuli into changes in gene expression (Fig. 1).7, 8 This pathway has been implicated in regulating a number of interrelated processes: proliferation, cell cycle, oncogenesis, development, and differentiation in many cell types. Increased MAPK activity often correlates with increased proliferation, and activating Ras/MAPK mutations can result in unrestrained proliferation and oncogenesis.7, 8 However, the effects of Ras-MAPK signaling on cell differentiation are less well defined, particularly in the case of osteogenesis. Cross-talk between the BMP-Smad and Ras-MAPK signaling pathways. In the classical BMP signaling pathway, osteogenic BMPs (i.e., BMP-2, BMP-4, BMP-7) bind and signal through BMPR dimers. BMPRs phosphorylate Smads1/5/8, which bind with co-Smad4 and translocate to the nucleus to promote osteogenic gene expression (left). In contrast, growth factor signals (FGF, EGF, etc.) and binding of integrin receptors to extracellular matrix (ECM) proteins result in the activation of the small GTPase Ras. Ras can initiate a number of signaling cascades (right). The Ras/Raf-1/MEK/ERK pathway has been strongly implicated in the regulation of proliferation, survival, and cell cycle regulation, but may also impinge on osteogenic signaling. Potential mechanisms behind this impact include (1) the capacity of ERKs to phosphorylate Smads and repress their nuclear import, (2) the potential for ERKs to increase the activity of transcription factor Runx2 through phosphorylation, and (3) ERKs may directly affect the expression of osteogenic target genes by unknown means. Conversely, a number of studies have indicated that activation of the BMP pathway can lead to the increased activity and expression of Ras-MAPK components and target genes. Osteoblast commitment and differentiation are required for the maintenance of normal bone homeostasis. This progression is coordinated by extracellular signals that are transduced by cell signaling pathways to guide specialized changes in gene expression. One of the key master regulators of osteoblast specification is Runx2, which precedes a cascade of other bone markers including osterix, alkaline phosphatase (ALP), collagen type I (COL1), osteonectin, osteocalcin (OCN), bone sialoprotein (BSP), and osteopontin.9, 10 Bone morphogenic proteins (BMPs), members of the TGF-β superfamily, are intimately involved in osteoblast commitment and differentiation.11 TGF-β proteins bind and activate specific receptor serine/threonine kinases at the cell surface, which in turn leads to the phosphorylation of Smad proteins. Smad signaling involves a complex interplay of Smad dimerization, competitive interactions of R-Smads, I-Smads, and co-Smads, nuclear translocation, and finally, the transcriptional activation of target genes (Fig. 1).10-12 Whereas BMPs are critical and sufficient to induce bone formation, cross-talk from other signaling pathways may also impact on osteogenic commitment and differentiation. The Wnt-frizzled pathway,13, 14 PI-3 kinase signaling,15-17 p38 kinase signaling,18-20 focal adhesion kinase (FAK)-mediated signals from the extracellular matrix interactions,21, 22 and the Ras-MAPK axis (i.e., the focus of this review) have all been implicated in this cross-talk. Cell signaling pathways are likely to be an important nexus for balancing and managing pro- and anti-osteogenic signals. However, in many cases, the interaction of these pathways and their relative importance to the determination of cell fate remain ambiguous. As noted, cell culture studies have produced two opposing theories regarding the role of Ras-MAPK signaling in the regulation of osteogenesis. The first theory is that increased MAPK signaling is detrimental to cells progressing down an osteogenic pathway (Table 1). In a seminal paper by Kretzschmar et al.,23 epidermal growth factor (EGF)-stimulated MAPK signaling was found to antagonize the BMP-mediated activation of Smad1, albeit in an epithelial cell line. Based on mutagenesis experiments, MAPK and BMP pathways were proposed to phosphorylate different and discrete Smad1 sites. Growth factor treatment resulted in MAPK activation, increased Smad1 phosphorylation, and nuclear exclusion of Smad1. Ablation of MAPK-responsive sites in the Smad1 linker region showed constitutive nuclear localization, even after growth factor treatment. Antagonism between TGF-β and EGF signals has been put forward as a general phenomenon, and this is supported by a subsequent study showing analogous sites in Smad2 and Smad3.24 This model is also strengthened by observations from type 1 neurofibromatosis (NF1). The NF1 disease state results from a defect in one copy of the NF1 gene. Neurofibromin, encoded for by the NF1 gene, is a Ras-GTPase–activating protein that catalyzes hydrolysis of Ras-GTP to Ras-GDP. NF1 haploinsufficiency leads to a constitutive increase in activated Ras-GTP.25, 26 NF1 patients show an increased incidence of orthopedic defects,26 and the NF1 disease state has been more recently linked with generalized osteopenia and osteoporosis.27, 28 In a heterozygous null mouse model of NF1, an increase in MAPK activity was noted in the developing skeleton, although this did not lead to global defects in bone morphogenesis.29 Cultured bone marrow osteoprogenitors30 and calvarial/long bone osteoblasts31 from NF1-deficient mice possess reduced os teogenic potential compared with wildtype mouse cells. A similar deficiency was also noted in cultured bone-derived cells isolated from NF1 human fetuses.32 These data suggest that increased Ras-MAPK activity in NF1 may have a mild but negative affect on osteoblast differentiation. Whereas an increase in Ras-MAPK activity has been associated with decreased osteogenesis, several in vitro studies have examined the inverse scenario. Inhibition of Ras-MAPK signaling through the PD98059 MEK inhibitor, dominant negative Ras, or dominant negative MEK-1 expression have been found to enhance osteogenic differentiation in cultured osteoblasts.33-35 A similar response was also seen when BMP-treated C2C12 myoblasts were treated with PD98059.34 However, in both instances, the magnitude of osteogenic improvement associated with MAPK inhibition was minor compared with the magnitude of BMP treatment alone. We infer that Ras-MAPK signals subtly modulate osteogenesis rather than acting as a crucial anti-osteogenic pathway. These observations are further corroborated by studies using pharmacological compounds that inhibit Ras action. Nitrogen-containing bisphosphonates and statins impede the mevalonate pathway, leading to blockade of Ras prenylation.36-38 Several in vitro studies have examined the effect of these drugs on osteoblast differentiation. In the case of bisphosphonates, this can lead to slight enhancement of osteoblast differentiation at specific concentrations39, 40 but results in apoptosis at higher doses.41 Statins such as simvastatin and lovastatin also display a capacity to promote osteogenic differentiation and mineralization.42, 43 Whereas these drug studies further support a subtle antagonistic relationship between Ras signals and osteogenesis, bisphosphonates and statins can disrupt other prenylated proteins in addition to Ras (e.g., Rac and Rho), which may also influence osteogenic signaling. A series of reports support a conflicting theory that advocates Ras-MAPK signaling is both essential and advantageous for osteogenic differentiation (Table 2). Inhibitor studies form the foundation for this theory and refute the data from several other research teams.23, 33, 34 MEK inhibitors U0126 and PD98059 were shown to significantly decrease OCN and BSP mRNA expression44, 45 and OCN promoter activity45, 46 in MC3T3-E1 cells. However, OCN and BSP are mature osteoblast markers,47 and cells in the aforementioned studies were harvested at relatively early time-points, possibly before the peak of OCN/BSP mRNA expression. ALP activity, an earlier marker of osteogenic differentiation, was examined in a more recent study. Multipotent KS483 cells were grown under osteogenic conditions and treated with either U0126 or PD98059.48 Under these conditions, PD98059 more strongly inhibited osteogenic commitment than U0126. This difference was attributed to pro-estrogenic activity in PD98059 not present in U0126.48 ALP activity and proliferation were also reduced in BMP-2–treated C3H10T/12 cell lines expressing dominant negative ERK2.49 Whereas technical concerns involving culture times and bone marker selection overshadow several studies, there is a growing body of evidence indicating that Ras-MAPK blockade can be detrimental to a cell's osteogenic potential. Along the same lines, increased Ras activity has been shown to augment osteogenesis in several systems. Physical shock wave therapy was found to increase membrane hyperpolarization and Ras activity in CRL-11882 bone marrow stromal cells.50 This was associated with increases in numerous bone markers (ALP activity, OCN, and COL1 mRNA) and mineralized nodule formation. These effects could be blocked by the expression of a dominant negative Ras mutant.50 Dominant negative Ras and MAPK inhibitors were also found to have anti-osteogenic effects in another culture study using MC3T3-E1, HOS-TE85, and COS1 cells.21 This study predominantly used a Gal4-Smad1/Gal4UAS-lux reporter system to study Ras-MAPK–mediated changes in Smad1 transcriptional activity. However, because Smad1 trafficking has been shown to be highly responsive to MAPK signals,22-24 this makes interpretation of the data challenging. Perhaps the most detailed in vitro report supporting an agonistic effect of Ras-MAPK on osteogenesis also attempts to describe a mechanism for this effect. Xiao et al.46 proposed that MAPK signaling directly stimulates osteogenesis through the phosphorylation and subsequent activation of the transcription factor Runx2. Purified full-length Runx2 protein could be phosphorylated by recombinant ERK2 in vitro, and this could be enhanced by co-incubation with constitutively active MEK or inhibited with dominant negative MEK.46 The authors showed in a variety of Runx2-transformed cell lines that increased MEK-MAPK activity increased OCN promoter activity and/or expression and that MEK inhibition decreased OCN promoter activity.46 It has been frequently and logically hypothesized that the presence of bone matrix proteins may have a positive influence on osteogenic differentiation. The adhesion of human osteoblasts to bone matrix proteins leads to increased spreading, migration, and expression of osteogenic markers, and this can be impeded by dominant negative ERK1 expression.51 Matrix–integrin binding is purported to activate a pro-osteogenic FAK-Ras-MAPK signaling pathway.22, 51 Concordant with this, expression of antisense FAK (asFAK) was found to prevent MC3T3-E1 differentiation.22 Curiously, the nuclear translocation of Smad1 was decreased by FAK inhibition,22 when this translocation is conventionally associated with Ras-MAPK activation.23, 24 Finally, in the context of bone healing, the magnitude of an osteogenic response is dependent on progenitor proliferation/migration and osteogenic specification/differentiation. Thus, even if Ras-MAPK signaling is mildly antagonistic to osteogenic differentiation, an increase in Ras-MAPK–driven osteoblast proliferation and migration52 may enhance increase bone healing overall. This concept is supported by in vitro experiments showing sequential treatment with fibroblast growth factor (FGF)-2 or FGF-9 (to stimulate proliferation) followed by BMP-2 (to stimulate osteogenesis) generated greater overall mineralization in chick calvarial osteoblast cultures than either treatment alone.53 In an additional layer of complexity, several studies have shown that BMP treatment can also increase Ras-MAPK activity (Table 3). In a straightforward in vitro study using C3H10T1/2 fibroblasts, BMP-2 treatment was found to increase cell proliferation, and this could be inhibited by dominant negative ERK2 expression.49 A sustained increase in activated MAPK activity was also noted after 10 h or 1 day of BMP-2 exposure. The increase in MAPK activity was suggested to result from an increase in both total and phosphorylated ERK2 rather than just an increased proportion of phosphorylated ERK2.49 Similar results were found when MC3T3-E1 cells were treated with BMP-4.19 These data imply that BMP signaling may affect the expression of some classically “proliferative” genes in addition to its “osteogenic” targets. Increases in ERK, p38K, and JNK signaling have all been noted during the osteogenic differentiation of mesenchymal stem cells (MSCs).54 Blockade of ERK signals during differentiation resulted in reduced osteoblast commitment (determined by ALP expression) and also shifted MSC differentiation toward an adipogenic fate.54, 55 Whereas these studies suggest an important role for Ras-MAPK signaling in the osteogenic commitment of MSCs, it is unclear how applicable these findings are to cells that already express Runx2. Assuming that all of the aforementioned peer-reviewed studies were executed and reported accurately, it must be concluded that differences must result from underlying variations in experimental methodology. Differences in cell lines, culture conditions, osteogenic stimuli, methods for blocking or stimulating Ras-MAPK activity, treatment durations, and osteogenic outcome measures varied considerably. Although there is no obvious common technical difference that distinguishes the studies supporting each particular viewpoint, a number of important distinctions need to be made. First, a more critical appraisal needs to be made regarding the osteogenic outcomes used in cell culture experiments. In particular, the premature measurement of mature osteoblastic markers such as OCN may give confounding results. Many of the seminal papers supporting Ras-MAPK as agonistic for bone formation based their conclusions simply on OCN mRNA expression or the activation of artificial OCN promoter constructs in poorly differentiated cells.22, 44-46, 51 Clearly, the standardization of osteogenic markers would be advantageous in comparing the responses of different cell lines to different stimuli. Second is that a fundamental difference exists between measuring osteogenic commitment and osteogenic differentiation. Osteogenic commitment is epitomized by the expression of the early bone fate marker Runx2 and indicates a primary fate decision made by a previously multipotent cell. In contrast, osteogenic differentiation involves the additional progressive, sequential expression of numerous osteogenic genes, including those involved with bone matrix production and calcification.47 Several key studies supporting an agonistic role for Ras-MAPK signaling have focused on or required the osteogenic commitment of multipotent progenitor cells.50, 54, 55 One logical synthesis would be that Ras-MAPK phosphorylation of Runx2 acts to promote the osteogenic commitment of progenitor cells, whereas Ras-MAPK phosphorylation of Smad proteins hinders their subsequent differentiation. In conclusion, the duality seen in the signaling literature is a conflict that is not easily resolved. The possibility that Ras-MAPK signals may be essential for the commitment of multipotent progenitors to an osteoblast cell fate, yet antagonize further osteogenic differentiation, certainly warrants further study. Clearly more elegant and directed experiments need to be considered to resolve this dilemma along the lines suggested.