Synovial inflammation in anterior cruciate ligament injury knees in mice: surgical vs non-surgical models

Abstract

Purpose: Anterior-cruciate-ligament (ACL) rupture is one of the most common knee injuries, resulting in pain, disability, and increased risk of developing OA. However, only ∼50% of individuals with ACL rupture develop OA, and the factors underlying this variable risk remain unclear. Changes in joint stability as a result of ACL deficiency drive mechanobiological dysfunction and the development of OA. However, specific factors are known to increase OA risk such as concurrent injury to the meniscus, cartilage, and subchondral bone. Also, the degree of initial joint inflammation and its resolution are implicated in long-term outcomes. Surgical and nonsurgical models of ACL rupture in mice result in similar joint instability but distinct risks or trajectories of OA. We used these two models to investigate if differences in synovial inflammatory response are associated with this risk of OA disease progression Methods: Ten-week-old, male C57BL/6 mice were randomly assigned to ACL transection (ACLT), ACL rupture (ACLR), sham-surgery or uninjured groups. Mice in the ACLT and sham-operated groups were subjected to anaesthesia, right unilateral medial arthrotomy, patella luxation, and fat-pad elevation to visualise the ACL, at which time the surgeon was informed to transect or spare the ACL before joint closure. Uninjured and ACLR animals were anaesthetised, the latter having a single compressive load (9-12N) to posteriorly displace the femur and traumatically rupture the ACL. Animals were housed in mixed injury groups (up to 5/cage) and allowed unrestricted activity. At 1, 2, 4 and 8 weeks post-injury, mice were euthanised and joints harvested for fixation, paraffin embedding, sectioning and OA histopathology scoring (n=7/group/timepoint); or synovial tissue (en bloc fat-pad, synovial lining and joint capsule anterior to the collateral ligaments) isolation for RNA extraction and quantitative RT-PCR analysis of cytokines (IL1, IL6, TNF), inflammatory cell markers (CD11b, CD4, CD8), and enzymes implicated in OA pathophysiology (ADAMTS4, ADAMTS5, MMP3, MMP9, MMP13) (n=6/group/time point). Results: ACL injury resulted in progressive OA pathology with cartilage pathology significantly greater in ACLR compared with ACLT at 2 and 4 weeks (p<0.001), but by 8 weeks was equivalent in the two models; posterior osteochondral erosion significantly worse in ACLR than ACLT at 4 and 8 weeks (p<0.01); and osteophyte size (p<0.001) and maturity (p<0.01) greater at 1 and 2 weels in ACLR compared with ACLT. Synovitis was increased acutely in both ACL-injuries and then decreased with time, with ACLT>ACLR (p=0.003) at week-1 but similar thereafter; both still greater than uninjured joints through to week 8 (p<0.001). Sham and ACLT had similar synovitis scores until week 8 when the latter was more severe. Despite histomorphological similarities in synovitis in ACLT and ACLR, there were differences in gene expression patterns. Synovial expression of TNF, CD11b and CD8 was unchanged at any time when comparing between groups. Surgery alone (ACLT and Sham) increased synovial IL6 (week-1: ACLT p=0.02, Sham p<0.001 vs uninjured) while ACLR did not. While CD4 was also increased by surgery (week-1 and -2: ACLT p<0.01, Sham p<0.05 vs uninjured), it was also increased in ACLR versus uninjured at these times (p<0.05; Figure 1). ADAMTS5 expression was not detected in synovial tissue, and MMP9 mRNA did not differ between groups. In contrast, ADAMTS4, MMP3, and MMP13 expression were increased by surgery (ACLT/Sham) and ACLR at week-1 and -2 compared with uninjured joint, although both surgeries>ACLR (Table 1). Conclusions: While both models caused joint instability, OA trajectory and ultimate severity were greater in ACLR. In contrast, surgery induced greater histo-pathological synovitis acutely but was similar in both injuries thereafter. There was a difference in the “phenotype” of the synovial inflammation in the two injury models. In ACLT there was upregulation of IL6 not evident in ACLR but similar to Sham, suggesting this was a response to surgery and wound healing. While helper T-cell (CD4) increase was greater in ACLT and Sham, this also occurred in ACLR suggesting a role of adaptive immunity in OA after nonsurgical injury. Expression of ADAMTS4, MMP3 and MMP13 were similarly increased acutely in both injury models, suggesting synovial tissues as a source of these enzymes implicated in the structural pathology of OA. While their expression was not correlated with ultimate pathological severity they may be broad-spectrum target across varying phenotypes of post-traumatic OA. Our data suggest that synovitis severity per se was not associated with differential OA risk with ACL injury; it likely still contributes to long-term OA structural pathology.View Large Image Figure ViewerDownload Hi-res image Download (PPT)