CXCL16 expression during post-traumatic osteoarthritis development: implications for mesenchymal stem cell recruitment

Abstract

Purpose: Post-traumatic osteoarthritis (PTOA) develops following major joint trauma. Recently, we demonstrated that mesenchymal stem/multipotent stromal cells (MSCs) are recruited to the injured joint after ACL rupture (ACLR) in rats. While it is known that MSCs are recruited by chemokines, the temporal, injury-induced, intraarticular expression of chemokines and their specific roles in MSC recruitment remain largely undescribed. The purpose of this study was to characterize the in vivo expression of the understudied chemokine CXCL16, which can recruit MSCs via CXCR6, following ACLR. We also sought to assess whether primary fibroblast-like synoviocytes (FLS) from injured joint synovium are “reprogrammed” to exhibit differential CXCL16 expression compared to cells from a healthy joint. Finally, we assessed CXCL16- and FLS-mediated recruitment of MSCs in an in vitro chemotaxis model. Methods: With IACUC approval, ACLR was induced in adult C57BL/6 mice via tibial compression. 3- and 14-days post-injury, the synovium, menisci, and femoral and tibial epiphyses were analyzed with rtPCR for CXCL16 and ADAM10, the protease that cleaves CXCL16 from the cell membrane, (n=8 mice per time point). Further mice were injured and FLS were isolated to establish cultures (n=3 each) from injured and contralateral limbs 7-days post-injury. Cultured FLS were induced with cytokines and analyzed with rtPCR for CXCL16 expression. CXCL16 (10 and 50 ng/mL) was assessed for its role in recruitment of C3H10T1/2 cells, an MSC-like cell line, in an in vitro 3D chemotaxis system (Incucyte). Chemotaxis was quantified by automated calculation of cell-occupied area on the bottom membrane. Primary FLS were isolated from uninjured mouse synovium, induced with 10 ng/mL TNFα, co-cultured with C3H10T1/2 in the chemotaxis system, and analyzed for the ability to recruit C3H10T1/2 MSCs. Results: Histology grading shows mild PTOA and moderate synovitis in mice following ACLR, as evidenced by mild cartilage degradation, early osteophyte formation, and pronounced synoviocyte hyperplasia at 3 and 14d (Fig 1). Compared to contralateral knees, the synovium of injured limbs exhibited no upregulation of IL-17, but significant increases in TNFα (P=0.026) and IL-1β (P=0.017) at 3d and in TNFα at 14d (P=0.006). Both CXCL16 and ADAM10 are significantly upregulated by the synovium in vivo 3d after injury and remain elevated at 14d (Fig 2). The menisci of injured limbs upregulated CXCL16 at both 3d (P<0.001) and 14d (P<0.001), but there was no change in ADAM10 expression in meniscus. Results from both epiphyseal tissues showed no upregulation in chemokine expression between injured and contralateral limbs (Fig 2). Primary FLS from injured joints upregulated CXCL16 when treated with TNFα (P=0.025) and IL-1β (P=0.006), and interestingly, FLS from contralateral joints showed similar upregulation in response to TNFα (P=0.002) and IL-1β (P=0.006) (Fig 3). There was no difference in CXCL16 expression between injured and injured FLS. 3D chemotaxis assays demonstrate that 50 ng/mL CXCL16, but not 10 ng/ml, induces C3H10T1/2 chemotaxis. Further, TNFα-stimulated FLS induce C3H10T1/2 chemotaxis to a greater extent than unstimulated FLS or TNFα-containing media alone, demonstrating the ability of cytokine activated FLS to recruit MSC-like cells (Fig 4).View Large Image Figure ViewerDownload Hi-res image Download (PPT)View Large Image Figure ViewerDownload Hi-res image Download (PPT)View Large Image Figure ViewerDownload Hi-res image Download (PPT) Conclusions: Our data demonstrate that noninvasive ACLR causes mild PTOA and moderate synovitis up to 14d post-ACLR, with concurrently increased expression of cytokines. In vivo injury-induced upregulation of CXCL16 and ADAM10 by the synovium provide evidence for CXCL16’s possible role in recruiting MSCs, which highly express CXCR6. CXCL16 was upregulated by the meniscus of the injured joint at 3d with persistent upregulation at 14d. Our in vitro data demonstrate that the FLS is a contributor to CXCL16 expression, and this is mediated by both IL-1β and TNFα. Similar CXCL16 expression between FLS isolated from contralateral and injured limbs suggests that although profound inflammation is observed in vivo following ACLR, this inflammatory environment did not “reprogram” the FLS to exhibit greater CXCL16 expression in response to cytokines. Further studies will explore later time points to determine if such “stromal memory” is demonstrated by FLS in later stages of PTOA. Finally, we have established an in vitro model to study MSC recruitment and FLS-MSC interactions, as we demonstrate both CXCL16 and cytokine-induced FLS can recruit MSC. Ongoing work is employing inhibitors to assess the role of CXCL16 in FLS-mediated MSC recruitment.