Differential mRNA Expression of COX-2 and Proinflammatory Mediators in Patients with Rotator Cuff Tears and Osteoarthritis of the Hip

Purpose: Bursal 
inflammation is thought to be a major cause of pain in degenerative rotator 
cuff tears (RCTs). While the expression of proinflammatory mediators, such as COX-2, 
TNF-α, IL-1β, and IL-6, is crucial for the pathophysiology of osteoarthritis 
(OA), their role in degenerative RCTs remains unknown. The aim of this study 
was to determine the expression of COX-2 
and proinflammatory mediators in the development of RCT-induced pain by 
comparing their levels 
in patients with hip OA or RCTs. Methods: We included samples obtained 
from 31 shoulders of 31 patients with RCTs and samples from 30 hips of 27 patients with hip OA. The mRNA levels of COX-2, TNF-α, IL-1β, 
and IL-6 were determined using 
RT-PCR, and were compared between the subacromial bursa and 
hip joints. We also analyzed IL-1β-induced COX-2 
expression in the subacromial bursa and synovial blast of the hip. Results: COX-2, IL-1β, and IL-6 expression levels were 
significantly lower in the subacromial bursa of RCTs than in hip OA samples, 
while no significant difference was observed for TNF-α. No 
significant difference in the fold increase was observed between subacromial 
bursa and hip OA samples, even though IL-1β-induced COX-2 expression increased in both 
samples. Conclusion: Our findings suggest that the main 
mechanism underlying pain development differs between patients with RCTs and 
those with hip OA.


Introduction
Most patients with rotator cuff tears (RCTs) first come to the hospital because of pain [1]. Inflammation is normally implicated as the source of shoulder pain in patients with rotator cuff impairments [2] [3]. Inflammatory processes occur in the acute phase of RCTs following tendon injury, and bursal inflammation is thought to be a major cause of shoulder pain in degenerative RCTs [2] [3] [4], particularly when compared with inflammation in the glenohumeral (GH) joint [5]. However, the expression of proinflammatory mediators in degenerative RCTs remains largely unknown.
In contrast to RCTs, the origin of pain inosteoarthritis (OA) has been extensively studied. Inflammatory stimuli initiate a cascade of events, including the release of cytokines and other mediators from the cartilage and synovium. These induce metabolic disturbances and enhance catabolism in joint tissues of patients with OA, which promotes joint pain [6]. Proinflammatory mediators, especially tumor necrosis factor-alpha (TNF-α), interleukin (IL)-1β, and IL-6, are crucial for the pathophysiology of OA [6]. In addition, anti-inflammatory effects are related to the downregulation of cyclooxygenase-2 (COX-2) and reduced proinflammatory mediator expression, regardless of joint [7] [8].
The purpose of this study was to investigate the role of COX-2 and proinflammatory mediators in the development of pain in patients with degenerative RCTs. To this end, we used real-time polymerase chain reaction (RT-PCR) to compare these mediators' mRNA expression in RCT patients and patients with hip OA-associated pain. We show that COX-2 and proinflammatory mediator levels were generally lower in patients with RCTs than in those with hip OAassociated pain.

Patient Selection
The experimental protocol was conducted in accordance with the guidelines of  [9], whereas hip OA grade was assessed using the anteroposterior view of plain radiographs and the Kellgren and Lawrence X-ray grading system [10]. Preoperative clinical assessments of RCTs and hip OA were performed using the Constant score and modified Harris Hip Score, respectively [11] [12]. Cartilage condition was assessed arthroscopically or macroscopically according to the International Cartilage Repair Society (ICRS) classification system [13].
All arthroscopic rotator cuff repair surgeries were performed by a single experienced shoulder surgeon (K.T.). We harvested a synovial membrane sample from the subacromial space of each RCT patient because subacromial bursitis is generally accepted as the main source of pain in degenerative RCTs compared with inflammation in the GH joint [2] [3] [4] [5]. For total hip arthroplasty, every patient was operated using the anterolateral approach in the supine position by two experienced hip surgeons (T.N. and F.K.). We harvested a sample of the synovial membrane from the site of strongest synovitis for further evaluation when capsulectomy was performed.

Isolation of RNA and Quantitative RT-PCR
Total RNA was extracted from the harvested synovial samples using TRIzol  Table 1. Quantitative PCR was performed using the CFX-96 RT-PCR detection system (Bio-Rad, Hercules, CA, USA). PCR cycle parameters were as follows: initial denaturation at 95˚C for 1 min, followed by 40 cycles of 95˚C for 5 s, and 60˚C for 30 s. mRNA expression of target genes was normalized to that of glyceraldehyde 3-phosphate dehydrogenase (GAPDH) mRNA using delta-delta method. The levels of four molecular biomarkers of inflammation in synovial tissue samples were measured: COX-2, TNF-α, IL-6, and IL-1β. We compared these levels between the RCT and OA patient groups, and between the subacromial bursa (SAB) of RCTs and hip joints of hip OA.

Effect of IL-1β on COX-2 Expression in the Synovial Fibroblasts of SAB and Hip Synovium
Synovial tissue was excised from the SAB and hip joints for the assessment of molecular biomarker levels, as previously described [2] [3] [14]. However, SAB is pathologically different from joint synovial blast [15], and no reports on the difference in COX-2 expression between SAB and joint synovial blast are available. Therefore, we analyzed IL-1β-induced COX-2 expression in the SAB and synovial blast of the hip. The excised tissue was minced, digested for 2 h at 37˚C with type 1 collagenase (0.1%; Sigma-Aldrich, Lakewood, NJ, USA), and passed through a 40-μm filter (Becton Dickinson, Franklin Lakes, NJ, USA), to yield single-cell suspensions. Cell numbers were determined using a hemocytometer. Nucleated cells isolated from the synovium were seeded at 1 × 10 4 cells/cm 2 in 6-well plates containing minimum essential medium (Gibco, Carlsbad, CA, USA) supplemented with 10% fetal bovine serum, 100 U/mL penicillin, and 100 μg/mL streptomycin, and the cells were incubated at 37˚C in an incubator with 5% CO 2 for seven days. After one week, synovial fibroblasts were left untreated or treated with 50 ng/mL of human recombinant IL-1β (Biolegend, San Diego, CA, USA) for 24 h. Cells were harvested, total RNA was isolated as described, and COX-2 mRNA in synovial fibroblasts was determined by RT-PCR. The experiment was performed four times.

Statistical Analysis
Statistical analyses and power analyses were performed using commercial software (JMP Pro version 14.3, SAS Institute Inc., Cary NC, USA). Results are expressed as the mean ± standard deviation (SD). The chi-square test was used for gender comparisons in RCT and hip OA groups. Mann-Whitney's U-test was used to compare age, and mRNA levels of COX-2 and proinflammatory mediators between two groups. For all statistical analyses, significance was defined as P < 0.05.

Patient Characteristics
Patients' characteristics are presented in Table 2. There was no significant difference in age between the groups, whereas significant differences in gender and ICRS classification were observed.

Proinflammatory Cytokine Levels in Patients with RCTs and Hip OA
Mann-Whitney's U-test revealed a significant difference in COX-2, IL-6, and  IL-1β mRNA levels between RCT and hip OA patients. However, no significant difference in TNF-α expression was observed between three synovial fibroblast samples ( Figure 1, Table 3).

Discussion
The present study revealed that COX-2, IL-6, and IL-1β mRNA levels were significantly lower in patients with RCTs than in those with hip OA, whereas no significant difference in TNF-α expression was detected. In addition, both SAB and hip joint synovium showed an increase, albeit non significant, in COX-2 mRNA expression following treatment with IL-1β. Hence, our findings suggest that the main mechanism underlying the development of pain in patients with RCTs after the acute phase differs from that underlying the development of pain in patients with hip OA because, with the exception of TNF-α, the levels of critical pain mediators involved in the pathophysiology of OA were lower in RCT patients. Even though they are both ball joints, the shoulder joint differs from the hip joint in that it is a non-weight bearing joint. The destructive response of the synovium, which is induced by cartilage breakdown, plays a major role in the development of OA pain [6]. In contrast to OA, SAB is affected by a torn rotator cuff. Our RCT subjects exhibited a much lower grade of ICRS classification compared with the hip OA group, suggesting that in their case, subacromial bursitis was not affected by cartilage debris as much as it was inhip OA. Accordingly,    [16], however, TNF-α can also act independently to initiate or propagate inflammation [6]. In a previous animal study, we confirmed continued TNF-α expression over a period of two weeks, even though COX-2 and IL-1β were significantly decreased [17]. That result may explain why only TNF-α mRNA levels failed to exhibit any significant difference between RCT and hip OA  [18]. The significant increasing of IL-6 expression has been confirmed, however, the relationship between IL-6 and RCTs pain is still controversial [2] [4] [18].
We examined IL-1β-induced expression of COX-2 in synovial fibroblasts, a treatment that induces considerable COX-2 expression [7] [16] [19] [20]. However, we did not observe any significant differences in COX-2 expression between SAB and hip synovial fibroblasts, suggesting that these cells expressed similar levels of both COX-2 and its downstream effector, prostaglandin E2 (PGE2). COX-2 and IL-1β mRNA levels decreases with time in rat RCT model [17]. Accordingly, the observed differences in COX-2, IL-1β, and IL-6mRNA levels may not reflect differential basal levels of expression between tissues. Instead, they may indicate the difference in inflammation states observed in these diseases.
In spite of their frequent prescription in this context, the effectiveness of nonsteroidal anti-inflammatory drugs (NSAIDs) in the management of pain in patients with degenerative RCTs is controversial [21] and not recommended for the conservative treatment and perioperative management of RCTs [22]. In contrast, NSAIDs are strongly recommended for the management of hip OA pain [23] [24], as COX-2 and other inflammation mediators are produced in the synovial tissues of OA joints [14] [25]. The binding of NSAIDs to COX isozymes inhibits the synthesis of prostanoids [26] [27] [28]. PGE2 is the dominant prostanoid produced during inflammation, and inhibition of its synthesis by NSAIDs is believed to be the main mechanism underlying potent analgesic and anti-inflammatory properties of these agents [26] [27] [28]. Moreover, the efficacy with which pain and PGE2 production are inhibited depends on the dose of NSAIDs [25]. Hence, if pain correlates with COX-2 expression, a normal dose of oral NSAIDs can be sufficient to control both OA and RCT pain. Our findings suggest that degenerative RCT-associated pain depends on factors other than COX-2.
Thus patients with RCTs, who do not feel any efficacy of taking NSAIDs, may not benefit from their use for pain management. Moreover, discontinuing NSAID treatment would allow such patients to avoid the unwanted side effects caused by these drugs. However, in cases where NSAIDs are efficacious to the patient with RCTs, their use for pain management can be kept in place.

Limitations
The major limitation of this study was the impossibility to compare mRNA expression of proinflammatory mediators between RCTs and shoulder OA, as rotator cuff degeneration occurs spontaneously in patients with shoulder OA. The target of pain control medication is to lower the level of pain-related substance rather than that of pain inducers. Our purpose was to compare the mRNA expression patterns of proinflammatory cytokines in synovial tissues between RCTs and hip OA. To this end, our study clearly shows the difference in mRNA Another limitation was our failure touse enzyme-linked immunosorbent assay (ELISA) or western blotting to support RT-PCR data. Assessment of COX-2 expression by RT-PCR is considered appropriate to determine the activity of this enzyme in synovial tissues, as it is consistent with immunohistochemical data [15]. In addition, the amount of PGE2 measured by ELISA is related to the expression of COX-2, TNF-α, and IL-1β measured by RT-PCR [7] [8]. Therefore, COX-2, TNF-α, IL-6, and IL-1β levels assessed using RT-PCR are consistent with the detection of these four molecular biomarkers of inflammationby ELISA or western blotting.
A minor limitation is constituted by the significant difference in gender between RCT and hip OA groups due to higher prevalence of hip dysplasia among females [29]. However, we could not detect any difference in mRNA expression levels of COX-2, TNF-α, IL-6, and IL-1β nor in the efficacy of NSAIDs or corticosteroids between genders. Therefore, we believe that the different gender composition between RCT and hip OA groups had no bearing on our results.

Conclusion
We compared the mRNA expression of COX-2 and proinflammatory mediators in patients with degenerative RCTs and hip OA, and demonstrated that COX-2, IL-1β, and IL-6 levels in SAB samples obtained from patients with RCTs were significantly lower than those from patients with hip OA. Our findings suggest that the main mechanism underlying the development of pain in patients with degenerative RCTs differs from that underlying the development of pain in patients with hip OA.

Funding
This study was supported by research grants from Japan Society for the Promotion of Science (JSPS) Grant-in-Aid for Scientific Research (KAKENHI) Grant No. 18K16633.

Conflicts of Interest
The author(s), their immediate families, and any research foundations with which they are affiliated have not received any financial payment or other benefits from any commercial entity related to the subject of this article.