gms | German Medical Science

Objectives: In the osteoarthritic (OA) knee joint, CXCL12 is primarily produced by synoviocytes in elevated levels. Human articular chondrocytes (hAC) represent potential recipient cells expressing the corresponding receptors CXCR4 and CXCR7. In vivo studies indicate that CXCL12 favors OA progression and systemic application of the CXCR4 inhibitor AMD3100 might be a feasible treatment. However, multiple side-effects are described for AMD3100, presumably, because it is an allosteric agonist of CXCR7. In this study, we investigate a novel CXCR4-inhibitory peptide, EPI-X4 JM#21, as a potential treatment option of OA.

Methods: Macroscopically intact (OARSI≤1) and highly degenerated (OARSI≥3) human cartilage was obtained from donors undergoing knee joint replacement. Cartilage was used for IHC, qPCR or ex vivo experiments. hAC were isolated from intact cartilage. Explants or hAC were stimulated with 100 or 200 ng/mL CXCL12, 1 μM AMD3100, and/or 1 μM EPI-X4 JM#21. Cytotoxicity was analysed in explants (7 d, Live-Dead staining) and hAC (24 h, alamarBlue). For osteogenic differentiation (21 d, monolayer, assessed by Alizarin Red S), hAC were cultured in differentiation medium supplemented with CXCL12, AMD3100, and/or EPI-X4 JM#21. Effects on hAC migration and proliferation were determined by a wound healing assay, Ki-67 staining and live-cell imaging. Statistical analysis: Student's t-test and one-way ANOVA.

Results and conclusion: CXCR4 was upregulated on mRNA (P=0.006) and protein level (P=0.02) in degenerated cartilage, compared to intact tissue. For CXCR7, no difference between intact and degenerated tissue was observed. However, CXCR7 was generally higher expressed than CXCR4 (325-fold, P=0.0006). Next, we demonstrated that CXCL12, AMD3100, and EPI-X4 JM#21 have no cytotoxic effects on hAC. Osteogenic differentiation was not altered by CXCL12 or inhibitors, but CXCR4 expression was upregulated (P=0.04) after 14 days, while CXCR7 was reduced (P=0.04). Finally, we found increased hAC migration in a wound healing assay, when stimulated with CXCL12 (P=0.04). Simultaneous treatment with EPI-X4 JM#21 (P=0.01) counteracted this effect. For AMD3100, an inhibitory trend was detected (P=0.2). Accordingly, CXCL12 increased migration distance (P=0.02) during live-cell imaging and both, AMD3100 (P=0.03) and EPI-X4 JM#21 (P=0.01), reduced migration distance. Proliferation (KI67 staining) was not affected by CXCL12 or the inhibitors.

Overall, CXCR4 antagonists AMD3100 and EPI-X4 JM#21 exhibited similar inhibitory effects in our experiments. Thus, EPI-X4 JM#21 might represent an alternative to therapeutically address CXCL12/CXCR4 signaling, without potential side effects due to CXCR7 activation as described for AMD3100. However, as AMD3100 treatment prevented OA progression in previous animal studies, the activation of CXCR7 might have contributed to the therapeutic effect. Future studies are needed to clarify the role CXCR7 in cartilage homeostasis and if selective targeting of CXCR4 is preferable in OA treatment.