S100A9 induces nociceptive pain but is not involved in allodynia in acute experimental synovitis

Abstract

Purpose: Synovitis-associated pain is an important aspect of arthritis pathology. Several inflammatory mediators released by the synovium have been implicated in the regulation of pain, including S100A8 and S100A9 which may regulate pain either via direct stimulation of TLR4 on the nerve endings in the synovium or via stimulation at the site of the dorsal root ganglia (DRG), thereby enabling an increased phagocyte infiltration, which may cause sensitization. We aimed to elucidate the role of S100A9 in the pain response after induction of an acute synovitis using streptococcal cell walls (SCW) as a trigger, comparing S100A9-/- mice and their WT controls. Methods: Synovial biopsies of 25 patients with early symptomatic OA were studied by micro-array analysis and subsequent functional annotation clustering (FAC). Patients with pain at t=5 years after biopsy isolation according to WOMAC score, were compared to those without pain. Acute synovitis was induced by a single i.a. injection of SCW in the knee joint of C57Bl6 (WT) mice and S100A9-/- mice, control mice received an i.a. saline injection. Serum S100A8/A9 levels were investigated by ELISA and expression of S100A8 and S100A9 in synovium and DRG by immunohistochemistry. Joint swelling and cell influx was assessed by 99mTc accumulaiton and histology, respectively. Pain response were investigated Incapacitance Tester (weight bearing), Catwalk (gait analysis) and von Frey’s filaments (mechanical allodynia). Gene expression of inflammatory mediators and neuron activation markers in DRG were determined by q-PCR. Monocyte influx and protein expression of NAV1.7 and GAP43 was monitored by immunohistochemistry. Results: In synovial biopsies of early symptomatic OA patients, approximately 750 genes were upregulated in patients that had pain 5 years after inclusion compared to patients without pain. Functional annotation clustering showed that inflammatory genes were highly significantly enriched (p=6,4*10-12). Among the upregulated genes were the alarmins S100A8 and S100A9. In mice, a single i.a. injection of SCW resulted in increased synovial expression of S100A8 and S100A9 and subsequent increased serum S100A8/A9 levels (2.6-fold, P < 0.001) 1 day p.i., which returned to basal levels at 7 days p.i. The increased expression of S100A9 did not contribute to the development of inflammation since joint swelling and cell influx were similar in WT and S100A9-/- mice 1 day p.i. Using the Incapacitance Tester, WT mice showed a marked and significant decrease in percentage of weight bearing on the SCW injected hindpaw (28%) compared to saline injection (47%, P < 0.001) 1 day p.i., whereas S100A9-/- mice did not. In addition, gait analysis showed that the stand-phase of the unaffected paws were significantly increased in WT mice 1 day p.i., which will reduce the load on the inflamed paw, while in S100A9-/- mice these parameters were not altered. No difference in mechanical allodynia was observed, both mouse strains showed a similar reduction of paw withdrawal threshold (4.2 and 4.5 fold decrease respectively). Analysis of DRG showed no increased phagocyte infiltration after SCW injection as determined by S100A8 and S100A9 immunohistochemistry and no change in gene expression of MCP-1, KC, IL-1β or TNF was measured. In addition, nu change in F4/80 staining was seen in both WT and S100A9-/- mice. However, expression of neuron activation markers NAV1.7, ATF3 and GAP43 were significantly increased at 1 day after SCW injection in WT mice as compared to saline injected mice (P = 0.022, 0.004 and 0.030, respectively) while SCW injection in S100A9-/- mice did not show increased expression, which is in line in with the reduced pain response observed earlier in S100A9-/- mice. Immunohistochemically the difference in NAV1.7 expression in the DRG was further confirmed at protein level. Conclusions: These findings show that S100A9, which is released from the synovium upon inflammation, is an important mediator of inflammatory nociceptive pain response in the knee, rather than by being involved in peripheral sensitization. During the acute phase of inflammation S100A8/A9 is likely regulated via direct activation of TLR4 on nerve endings in the synovium and not via increased infiltration of phagocytes in the DRG.