The link between inflammation and carcinogenesis is well known; experiments have implicated many components of the inflammatory cascade such as prostaglandin E2 and IL-6 as key players in tumor development, growth, and metastasis [37, 38]. These inflammatory cytokines and growth factors, either generated by the tumor cells themselves in an autocrine manner or derived from inflammatory or stromal cells in the tumor microenvironment, have received much attention as potential targets for therapeutic intervention [37, 39, 40]. Indeed, these cytokines trigger the activation of many signaling pathways known to contribute to tumorigenesis and chemoresistance such as the JAK/STAT and Ras/Raf/MAPK pathways . We had previously shown that STAT3 activation was present in a substantial number of OSA cell lines and primary canine OSA tumor samples and that inhibition of STAT3 using either a small molecule inhibitor or siRNA resulted in death of OSA cells in vitro . The purpose of the following study was to identify possible drivers of the observed STAT3 activation.
Our data demonstrate that OSM, a member of the IL-6 subfamily of cytokines, and components of the OSM signaling pathway are expressed in OSA cell lines and tumor samples, and that activation of the JAK/STAT3 pathway with OSM stimulation leads to increased production of MMP2, VEGF, and enhanced tumor cell invasion. These results suggest that this pathway may be important in vivo for OSA cell metastasis by facilitating the process of invasion and angiogenesis. Interestingly, expression of IL-6 and IL-6R was either very low or absent in the OSA cells and the cells did not respond to stimulation with IL-6 indicating that this cytokine is likely not an important contributor to OSA pathobiology.
OSM is known to affect a variety of biological processes including cell growth and differentiation, hematopoiesis, and inflammation . It has also been implicated as having a role in bone remodeling [20, 41] in part through stimulating osteoblast differentiation and activation. OSM can be expressed in the bone marrow compartment  and is secreted from activated lymphocytes, monocytes, and neutrophils [11, 43]. Interestingly, breast cancer cells have been demonstrated to stimulate neutrophils to produce the cytokine  and experiments have shown that OSM is produced by multiple human osteoblast-like cell lines including the OSA cell line MG-63 and mouse osteoblasts and osteocytes [10, 20]. Co-expression of OSM and its receptor was noted in the fresh frozen tumor samples while only OSM receptor was identified in the cell lines. Based on these data, it is possible that the OSM found in the tumor specimens is derived from local inflammatory or stromal cells in the OSA tumor microenvironment independent of or, as demonstrated with the breast cancer cell lines, under the influence of the tumor cells.
OSM activates JAK2 and STAT3 upon binding to its receptor in many cells including murine, rat, and human osteoblastic cells and osteosarcoma cell lines [21, 22]. However, the role of this cytokine pathway in OSA tumor cell survival and metastasis has not been fully explored. Upon stimulation with OSM, we demonstrated marked increases in JAK2, STAT3, and Src phosphorylation in canine and human OSA cell lines. This signaling enhanced the production of VEGF which is consistent with activation of STAT3, as it could be blocked by the small molecule STAT3 inhibitor LLL3 . It has been shown that OSM stimulation enhances VEGF expression in adipocytes  and that OSM stimulates strong phospho-STAT3 (tyrosine 705) in normal and keloid fibroblasts . Given that OSM is present in all canine patient tumor samples, it is plausible to infer that OSM in the tumor microenvironment in vivo likely enhances OSA basal Src and STAT3 activation and JAK2 phosphorylation. Indeed, the enhanced phosphorylation of Src and STAT3 and co-localization of Src and STAT3 with gp130 in the OSA cell lines following OSM stimulation suggest that a similar functional and spacial relationship exists between STAT3 and Src as shown by Schaeffer et al in multiple myeloma cells in which binding of IL-6 to gp130 led to activation of the Src family kinase Hck .
OSM is known to confer multiple, often divergent functions to various cell types including inhibition of melanoma and astroglioma tumor cell growth [43, 46] and stimulation of proliferation of AIDS-related Kaposi's sarcoma cells and fibroblasts . In OSA cells, OSM has been shown to downregulate osteoblast markers and induce glial fibrillary acidic protein , promote an osteocyte-like differentiation , and sensitize rat OSA cells to the antitumor effect of midostaurin . However, our data indicate that treatment of canine and a human OSA cell lines does not impact their proliferation or viability. Other studies have shown that OSM has a role in regulating the MMPs as part of both wound healing and inflammation . Enhanced MMP9 expression has been observed in astroglioma cell lines following OSM exposure  and breast cancer cells treated with OSM demonstrated increased VEGF production associated with detachment and invasion . OSM stimulation has been linked to VEGF upregulation in normal adipocytes, liver, smooth muscle, and cardiac myocytes [44, 48–50]. Lastly, OSM stimulation of astroglioma cells led to increased STAT3-dependent VEGF expression .
We observed increased MMP2 activity and VEGF expression with OSM stimulation of OSA cell lines that was partially abrogated by the small molecule STAT3 inhibitor, LLL3 . Higher levels of VEGF expression in human OSA tumors have been shown to correlate with a significantly worse prognosis and the presence of lung metastasis [52, 53]. Higher VEGF expression also has predictive value for survival of OSA patients . With respect to canine OSA, one study found that pretreatment platelet-corrected serum VEGF levels correlated significantly with DFI in dogs with OSA following amputation and adjuvant chemotherapy . Lastly, higher levels of plasma VEGF were found in more aggressive neoplasms in a survey of spontaneous canine tumors including those of the bone . These data suggest that OSM stimulation of OSA cells may enhance VEGF production, thereby promoting angiogenesis, contributing to the metastatic cascade. Our data showed that OSM stimulation of OSA lines significantly enhanced the invasive behavior of OSA cells and that this was augmented in the presence of HGF. However, we have previously demonstrated that HGF stimulation of OSA cells does not promote STAT3 phosphorylation , and it is thus likely that HGF contributes to the observed invasion through mechanisms other than MMP2 production. As both OSM and HGF are likely to be relatively ubiquitous in the tumor microenvironment, it is possible they may work to promote early invasion and metastasis of OSA cells in vivo.