Although, extensive animal models for OA study have been developed over the past half century, there is no gold standard animal model used in OA research. In the ideal animal model, the disease must be induced reliably, with 100% penetrance, and within a suitable time frame, and yet still present with disease characteristics that are comparable to the human condition. [17, 18] As an important chemokine produced in the pathological process of human OA, we found that intra-articular injection of SDF-1 as a novel biochemical method can successfully induce knee OA in rabbits, and the effect with SDF-1 40ug/kg injection after 12 weeks is similar to the classic surgical modeling. This modeling method may be another choice closer to the natural OA pathology.
Macroscopic observation and HE staining of SDF-1 treated group in our experiment showed typical characteristics of early OA described in the literature , and Mankin score indicating a histological improvement of cartilage deterioration by SDF-1 and the degeneration of SDF-1 40ug/kg group on articular cartilage went to middle stage of OA at 12 weeks that was similar to the surgical group. This result is also consistent with our previous research.
As pivotal pro-inflammatory cytokines in AO pathology, IL-1 and TNF-α play roles in promoting cartilage matrix degradation. [21, 22] In the immunohistochemical study, the positive expression of IL-1 and TNF-α are significantly increased in all SDF-1 groups, and uneven extracellular matrix staining, cracks forming and cartilage defects are obvious. After 8 weeks, the expression level of IL-1 and TNF-α in SDF-1 40ug/kg group was higher than the surgical group. It indicated high dose SDF-1 intra-articular injection promotes inflammatory factors secretion earlier than Hulth surgery modeling, and inflammation was more severe. Wang et al. also demonstrated that activation of SDF-1/CXCR4 pathway significantly promoted the secretion of IL-1 β, IL-6, IL-15, IL-17, IL-18, and TNF-α. These inflammatory factors activate MAPKs pathway and TLRs pathway to promote the apoptosis of chondrocytes, and finally lead to the further imbalance of homeostasis in the knee joint. 
MMPs play a key role in the degradation of cartilage matrix and damages the stability of articular cartilage, especially MMP-3,9,13. [24–26] MMPs are rarely expressed under normal physiological conditions, however significantly overexpressed in synovial fluid and articular cartilage of human OA.  We examined the levels of MMP-3,9,13 in synovial fluid and the expressions of MMPs mRNA in chondrocytes. Either in joint fluid or chondrocytes, the mRNA expression and secretion levels of the MMP-3,9,13 increased with SDF-1 injection concentration and time, indicating that the modeling was successful. At 12 weeks, the content of MMP-3,9,13 in SDF-1 40ug/kg group was beyond surgical group. After 8 weeks, MMP-13 mRNA expression in the SDF-1 40ug/kg group increase had not much change demonstrated that the model only stood on the middle stage of OA. Similarly, Wang, G et research showed TN14003 can target blocking SDF-1/CXCR4 signaling pathway in vivo, reduce the expression and secretion of MMP-3, MMP-9, and MMP-13 in cartilage tissue, and reduce the degradation of collagen II and aggregating proteoglycan, thus delaying the degeneration of articular cartilage. 
Col II and ACAN are the principal component of cartilage ECM, and their expression and apoptosis are the best indicators of cartilage degeneration.  In our study, the alteration of Col II and ACAN mRNA and protein was downregulated as a result of SDF-1 injected increase of concentration and time. In SDF-1 40ug/kg injected group the degradation of Col II and ACAN after eight weeks were more than the surgical group. Compared with classical surgical method, the loss of COL Ⅱ ACAN is more, and the time of building model are shorter. Other researchers also obtained the similar result as our experiment. By changing the content of SDF-1 in the joint, SDF-1/CXCR4 signaling pathway can be regulated, which affects the apoptosis of chondrocyte, and inflammation, and matrix Catabolism [20, 28–30]
The results of current research were all highly consistent with the predictions that exogenous SDF-1 intra-articular injection could establish the OA rabbit model. High dose of SDF-1 40ug/kg can cause the degeneration of rabbit knee joints faster, reaching middle stage of OA, and the efficacy similar to traditional surgical modeling. All rabbits were alive, no adverse reactions and complications of joint such as infection and instability occurred in surgical modeling. [31, 32] This provides a new method for modeling animals knee joint. It is more beneficial to the research of pathophysiology of human knee OA.
Some limitations remain in the present study. Firstly, only 12 weeks were observed in the current study, and the pathological features at the early and middle stages of OA were mainly presented in our model. Whether the model is applicable to the later stage of OA still needs a longer period of observation. Secondly, OA is a complex disease and the pathological changes of OA are the result of many cytokines and chemokines synergistic effect. Only SDF-1 was used to induce modeling, which may ignore the influence of other factors. Thirdly, this experiment has not been compared with other chemical modeling methods, and cannot fully explain the advantages of the same type of modeling methods. Further studies of SDF-1 induced OA model will be carried out to overcome above shortages in the future.