We assessed the effects of the Src inhibitor dasatinib on the osteogenic differentiation of bone marrow-derived mesenchymal stromal cells (MSC) treated or not with a combination of dexamethasone, ascorbic acid and β-glycerophosphate (DAG).
Firstly, we have searched for the most informative markers of osteogenic differentiation of MSC induced by DAG and found a set of 5 markers (mineralization, ALP activity and expression of RANKL, BSP and OPN) that best characterize the 4 stages of osteoblast differentiation: native stage (undifferentiated MSC), early stage (osteoprogenitors), intermediate stage (mature osteoblasts), and late stage (late osteoblasts). Of note, many genes related to the osteoblastic phenotype or activity have been previously used to define the osteoblastic status of different cellular types [30
]. However, in our model, we found that the expression of several of them, notably Runx2, COL-I, OSN and OPG, was already present in DAG-untreated MSC and remained unchanged during osteoblastic differentiation, while others have reported that the expression of Runx2 and COL-I increased along the osteoblast differentiation [32
]. This divergence can be explained by the fact that MSC preparations may contain multidifferentiated cells, which express both immature and mature proteins of different tissues [33
]. In addition, recent reports indicate that the phosphorylation level but not the expression level of Runx2 is related to the differentiation status of osteoblasts [35
]. Therefore, the expression of these genes should be considered rather as a weak indicator of MSC differentiation into osteoblasts. Interestingly and in accordance with previous observations [36
], ALP activity and gene expression increased during MSC differentiation into osteoblasts. However, we found that ALP activity and gene expression did not fully correlate; ALP activity is probably more reliable to the mineralization process. Indeed, in undifferentiated MSC and late osteoblasts, ALP activity was, respectively, very low and significantly downregulated, while ALP mRNA was expressed at high levels. These data indicate that ALP activity is more informative than ALP gene expression as a marker of MSC differentiation.
Based on our selection of markers, we have evaluated the effects of the Src inhibitor dasatinib on the differentiation of MSC into osteoblasts. Our interest in Src inhibitors stems from a recent study showing that the differentiation of the MC3T3-E1 pre-osteoblasts was accompanied by a decrease in phosphorylation of the activator site (416
Tyr) of Src and an increase in phosphorylation of its inhibitor site (527
]. This indicates that the post-translational inhibition of Src is associated with the activation of osteoblast differentiation. Moreover, targeted disruption of Src gene in mice leads to osteopetrosis [39
], and osteoblasts isolated from Src-/-
mice exhibit accelerated differentiation and elevated levels of osteoblast-related markers [40
]. Taken together, these data suggest that Src inhibition may stimulate osteoblast differentiation. To validate that Src was targeted by dasatinib in MSC, we confirmed that dasatinib inhibits the phosphorylation of the activator site of Src. We also showed that the specific Src inhibitor E804 stimulated osteoblast differentiation as it increased ALP activity in MSC exposed to DAG. By contrast, mice deficient in Abl, another potent target of dasatinib, are osteoporotic and have defects in osteoblast maturation [41
], excluding a positive role of Abl inhibition by dasatinib in the differentiation of MSC into osteoblasts. We also considered other targets of dasatinib, namely platelet-derived growth factor receptor (PDGFR) [42
] and cKIT/CD117 [43
]. The role of PDGFR in the osteoblastic differentiation of MSC remains, however, controversial [44
], and new data clearly demonstrated that PDGFR signaling is not involved in osteogenic differentiation of human MSC, while it affects MSC proliferation [46
]. On the other hand, we did not detect cKIT in MSC (data not shown, Western blotting using mouse mAb cKIT (Ab81) from Cell Signaling Technology, LND1 melanoma cell line as positive control), confirming previous observations [47
]. Altogether, these data indicate that Src is the most relevant dasatinib target involved in the process of osteoblastic differentiation of MSC.
In this study, we found that dasatinib promoted time-dependent changes in osteoblast-related markers in MSC. These changes matched the ones induced by DAG and corresponded to the progressive differentiation of MSC into osteoblasts (see Fig. ). Indeed, we documented that dasatinib, in combination with DAG, increased ALP activity after 7 days, and consequently, speeded up calcium deposition after 14 days without exceeding the maximum levels reached with DAG alone, and that it increased the expression of BSP and OPN, mainly in the late stage of MSC differentiation. Thus, dasatinib was able to stimulate each stage of the differentiation of MSC into osteoblasts. Importantly, our experiments were performed with a non-toxic but effective concentration of dasatinib (10-8
M) able to inhibit Src kinase activity (inhibition of Src phosphorylation) in accordance with a previous report showing inhibition of the kinase activity of purified Src protein with an IC50 of 3 × 10-9
]. Altogether, our data, added to previous studies from other groups [20
], strongly support that Src kinase activity is the main target for dasatinib in MSC differentiation process. Interestingly, a very recent study revealed that dasatinib strongly enhanced differentiation of primary mouse osteoblasts isolated from mouse calvaria (stimulation of ALP activity, osteocalcin secretion, matrix mineralization) [29
]. The authors showed that among the Src family kinases, only Src was activated at a high level in this model, and that 419
Tyr-Src phosphorylation was inhibited by dasatinib. Moreover, knockdown of Src by lenti-shRNA in osteoblasts enhanced their differentiation, suggesting that dasatinib stimulated osteoblast differentiation through the inhibition of Src. Of note, the authors also reported that osteoblast differentiation by dasatinib could be mediated through the inhibition of Abl, however, they found that Abl was expressed at a low level in osteoblasts, suggesting a limited impact.
In the other hand, we observed that dasatinib alone was able to decrease RANKL mRNA expression as well as RANKL protein released in culture medium (decrease in RANKL/OPG ratio) in undifferentiated MSC (no DAG induction) cultured up to 7 days, suggesting a possible indirect inhibitory effect on osteoclastogenesis and, consequently, on bone resorption. In this context, recent studies reported that tyrosine kinase inhibitors are effective in inhibiting the differentiation and activity of human osteoclasts [48
], and in preventing bone loss in animal models [17
]. Thus, our data indicate that, in addition to its previously reported direct inhibitory effects on the osteoclast formation and activity [49
], dasatinib may also act indirectly on osteoclasts through the modulation of RANKL expression in osteoblasts.
Furthermore, as far as the bone loss associated with metastases is concerned, recent preclinical and clinical studies indicated that dasatinib could potentially have inhibitory effects both on osteoclasts and tumor cells. Indeed, dasatinib decreased the osteolysis induced by prostate cancer cells injected into tibiae of SCID mice and inhibited the growth of cancer cells [51
], while it reduced the expression of markers of bone resorption in patients with advanced castration-resistant prostate cancer [52
]. Thus, our findings complete these observations by demonstrating a direct effect of dasatinib on osteoblasts, which could further contribute to the disruption of the vicious circle established between bone cells and tumor cells.
Our data also suggest that patients with bone loss could benefit from dasatinib therapy in accordance with critical findings using imatinib, another tyrosine kinase inhibitor currently used in chronic myeloid leukemia (CML). Indeed, short-term imatinib treatment increased OPG/RANKL ratio and osteocalcin levels in serum of CML patients [53
], while long-term (> 2 years) imatinib therapy promoted bone formation [54
] and increased bone mineral density (cortical bone mineralization) [55