Exercise increases interleukin-10 levels both intraarticularly and peri-synovially in patients with knee osteoarthritis: a randomized controlled trial

Osteoarthritis (OA) is associated with cartilage erosion and bony changes as well as with intermittent periods of synovial membrane inflammation and subsequent release of biomarkers for inflammation [1]. Excessive loading of the joint in these patients can lead to a worsening of the pathology with an enhanced inflammatory response, joint pain and swelling [2]. On the other hand regular moderate exercise like strength training, cycling or walking is known to be advantageous in OA patients with no sign of deterioration of the inflammatory periods [3‐5]. It is unknown how exercise exerts its beneficial role in OA, and whether it mainly is accomplished via a stabilisation of the joints through muscle strength and control, or whether exercise has a direct effect upon the joint cartilage and the synovium. Exercise has been proposed to positively modulate low-grade inflammation in elderly patients [6] and has been shown to have a positive effect on the glycosaminoglycan (GAG) content in cartilage of subjects at increased risk for OA [7]. It is therefore possible that acute exercise may induce changes in the intraarticular and peri-synovial milieu that encourages anti-inflammatory activity as well as releases potential chondroprotective substances, for example, Interleukin (IL)-10.

We have previously shown that microdialysis can be used as a method to investigate, continuously over time, the peri-synovial interstitial tissue and the joints space in parallel [8]. Several biomarkers for both cartilage (Aggrecan, COMP and CTX-II) and inflammation (Interleukins) have been studied in the circulating blood, the urine and intraarticularly [9‐11]. In the present study the microdialysis method enabled us to investigate biochemical changes within and around the joint simultaneously in both the resting state and in a joint that had been subjected to exercise. The aim of the present study was, by applying the microdialysis technique, to monitor markers of cartilage breakdown and inflammation intraarticularly and in the synovium in a group of human females with knee OA over a period of three hours.

The microdialysis method was applied to a group of human females with OA of the knee in order to obtain information about the effect of a single bout of mechanical loading on cartilage biomarkers and cytokines both inside the joint as well as in the synovium over a period of three hours. The present study demonstrates that the anti-inflammatory cytokine, IL-10, increased significantly over time post exercise in the Ex group in both compartments, but not in the NEx group, indicating that the increase observed in the Ex group could be ascribed to exercise (Figure 3). Most previous studies on IL-10 and OA have been conducted on animals or ex vivo, and only a few investigators have also measured the concentration of IL-10 in synovial fluid in humans with OA, revealing concentrations in accordance with the present findings [14, 15] or in some cases showing non detectable concentrations [16, 17]. To our knowledge, the intraarticular and the peri-synovial concentration responses to exercise have never been investigated in OA previously. It is of interest that IL-10 responds positively to mechanical joint loading in this patient group, as IL-10 has been found to display chondroprotective properties by antagonizing important steps in the suggested pathogenesis of OA, such as suppressing the release of inflammatory mediators by macrophages and the activation of synoviocytes and chondrocytes [18, 19]. The previously mentioned study by Fraser et al [14] found that patients with early psoriatic arthritis and early rheumatoid arthritis have higher levels of IL-10 compared to patients in later stages of the disease. It appears that development of the disease leads to an impaired capacity to suppress the release of inflammatory mediators and thus a decreased immunoresponse beyond a certain point in the disease course. Taking into account the present data, this would argue in favour of stimulating the joints of patients with early OA, or patients at high risk of developing OA, with exercise, as this would result in a release of IL-10, a response that potentially could have beneficial effects upon the control of the immune response. The general influence of exercise on IL-10 has not been investigated formerly in human joint aspirates of synovial fluid and previous studies on serum levels show some inconsistency [20‐23]. One study has, in support of our results, demonstrated that higher levels of regular physical activity are associated with increased levels of IL-10 in the blood of healthy older males, whereas studies on young, moderately and well-trained males in different training sessions have shown both increases as well as a decrease in serum IL-10 concentrations. Increased production of anti-inflammatory cytokines during exercise can possibly restrict the production of pro-inflammatory cytokines such as IL-6, IL-8 and TNF-α. In vitro studies suggest that IL-6 act as a negative regulator of chondrocyte proliferation and articular cartilage metabolism [24] and that IL-8 possibly act as a modulator of both IL-6 and TNF-α as well as a chemotactic agent for neutrophils [25]. Hence, these cytokines play an important role in the low-inflammatory response of OA. In our study we determined highly significant increases in IL-6 and IL-8 concentrations from the first hour and a half of sampling to the second hour and a half of sampling; regardless of exercise and position of the catheter (Figure 3). Our levels reached medians of approximately 600 pg/ml and 5,000 pg/ml for each of the sampling periods irrespective of the position of the catheter, which contrasts with the findings of other investigators, who discovered levels of IL-6 and IL-8 in synovial fluid obtained by direct joint puncture of approximately 50 to 200 pg/ml [11, 26, 27]. A major trauma to the knee such as a tear of the anterior cruciate ligament has been found to result in IL-6 levels above 20,000 pg/ml within the first couple of days [28], and as the concentrations in the present study increased over time our findings could indicate that the insertion of the catheters per se induced a production of these cytokines due to the tissue injury. TNF-α concentrations show a pattern similar to IL-6 and IL-8, but for the peri-synovial level an increase was found only in the Ex group. The overall intraarticular concentrations of TNF-α at t1 in our study are somewhat below what has previously been reported [14, 27, 29]. We have no obvious explanation for this finding, but it could be speculated that the known suppressive effect of IL-10 on TNF-α [18, 30] could have contributed to the present findings.

Many biomarkers of cartilage measured in blood are known to present a circadian variation with concentrations being lower during the night and higher during daytime [9, 31], most likely due to an effect of the tissue resting. In the present study we have confirmed this resting effect as serum concentrations of COMP and Aggrecan decrease over the course of the test day (Figure 4). Other studies have detected a temporary increase in serum COMP following exercise [32, 33], which was not the case in our study, maybe due to the exercise protocol not being strenuous enough to induce systemic changes. We have, however, demonstrated for the first time that unloading of the joints for three hours results in an immediately measurable decrease in Aggrecan inside a single joint as well as in the synovium (Figure 5) despite previous physical activity. This is in agreement with the suggestion that Aggrecan is one of the first fragments to be released during cartilage breakdown. A high metabolism of Aggrecan would lead to an increased excretion through the urine, and it could then be expected that a higher urinary concentration would be seen in the Ex group, reflecting a faster turnover. The difference between the two groups at that point (T1, where the Ex group had already performed the exercise protocol as compared to the NEx group) did, however, not reach statistical significance. Instead a significant difference was found within the Ex group, indicating a faster reversion to baseline level (Figure 6). The concentration of CTX-II, a degradation product from collagen II, was not affected by a single bout of exercise in our subjects, and it is likely that the systemic measurement was too crude and insensitive to be able to detect any contribution from a single joint. It probably requires long term adaptation to loading to change the CTX-II level which has been demonstrated by O'Kane et al. in elite athletes performing different kinds of sports [34] and in OA patients followed for years [35]. Interstitial concentration of COMP did not reveal any changes in the NEx group in either of the compartments, but in the Ex group we found a significant decrease intraarticularly but not peri-synovially (Figure 5). A plausible reason for the intraarticular decrease could be the increased blood-flow and hydrostatic pressure following exercise, which could lead to a faster elimination of the COMP molecules in the Ex group. The turnover of cartilage is very slow with an estimated half life of collagen II of more than 100 years and for Aggrecan of 3 to 24 years [36, 37]. It is therefore not realistic that a single bout of exercise should have caused a molecular rearrangement of the cartilage; hence, the measured molecules must have been present in the joint space or very close to the cartilage surface during the exercise.

The microdialysis method used in the present study has been applied to a variety of human tissues including brain, adipose tissue and peritendinous tissue [38‐40] and is generally considered a minimally invasive procedure compared to other tissue sampling techniques. It must be acknowledged, though, that the microdialysis method still causes tissue injury during the insertion of the catheters which by itself can generate an inflammatory response as shown by Langberg et al in peritendinous tissue and by Clough et al. [41] in relation to skin wounding and allergen-induced inflammation. A sharp rise in IL-6 and IL-8 concentrations was found in the area with inflammation, although not to the same extent for IL-6 as in the present study. Importantly, Clough et al [41] also found an increase at a 1 cm distance from the insertion site, which indicates that the increased production of cytokines found in our study may be a result of a larger involvement of the already inflamed tissue next to the insertion site. Another important issue regarding the microdialysis method is the choice of tracer for determination of relative recovery. We used radioactively labelled glucose, which is a small molecule of only 0.18 kDa compared to Aggrecan and COMP (exceeding 500 kDa), and the chosen cytokines with a molecular weight of 11.1 to 25.6 kDa. This probably leads to an underestimation of the true concentrations (due to an overestimated RR) with regards to the cartilage markers. In addition, larger molecules such as markers of cartilage turn-over, do not move readily, which creates a possible risk of drainage from the area around the membrane. This must be considered since the concentrations of Aggrecan and COMP in our study are far from what joint puncture has shown in other studies on similar patients [42‐44]. However, even taking these limitations into consideration, it is important to note that procedures were identical for Ex and NEx groups, and this should not influence the detection of potential differences between groups.

Because of the specific use of anaesthetic method in this study and due to the risk of infection it was not considered pertinent to have catheters inserted before and again after exercise. Even with a block of only sensory nerves the risk of damage to the catheters during exercise would be too high, so we were constrained to look at two different groups and compare these. We believe the two groups are very homogenous as the KOOS score and other demographics showed no differences between groups. The presented results are valid for at selected group of female subjects, all with a relatively low Body Mass Index compared to other OA patients (because the insertion of the peri-synovial catheter would otherwise be complicated by penetrating too much adipose tissue) and with no history of other inflammatory diseases or regular intake of anti-inflammatory drugs.

With the limited conservative treatment options for OA patients it is of great importance to uncover the benefits of different kinds of exercise in order to provide a possibility of self-management. It could be speculated if specific groups of knee OA patients with no contraindications should be encouraged to perform more loadbearing activities to further improve the effect of exercise.

The present study demonstrates that it is possible to evaluate the effect of a single bout of exercise on cartilage and inflammatory markers in a group of female subjects with OA of the knee, over time and simultaneously in two different compartments of the knee, using the microdialysis method. We discovered a significant increase of the chondroprotective IL-10 only in the group that performed exercise compared to the non-exercising group intraarticularly as well as peri-synovially. This might contribute to explain the beneficial effect that exercise has on patients with knee OA.