Immune complex induced arthritis in rats: role of lipid mediators on cell infiltration

We investigated the participation of lipid mediators in an experimental immune complex (IC) arthritis model in rats. The animals were subjected to intraarticular injection of anti-bovine sertLm albumin (BSA) IgG antibodies followed by i.v. injection of BSA. Histopathological analysis of the synovial membranes disclosed infiltration of polymorphonuclear (PMN) cells and vascular congestion. Slight increase in vascular permeability, measured by Evans blue dye extravasation into the joints, was detected after 3 h of arthritis. Cellular influx into the articular cavities was most evident at the sixth hour of arthritis with predominance of PMN. Pretreatment with either indomethacin, a cyclooxygenase inhibitor, or L-660,711, a peptido-leukotriene antagonist, did not inhibit cell infiltration, whereas pretreatment with either L-663,536, a 5-lipoxygenase inhibitor, or L-655,240, a thromboxane antagonist, significantly inhibited the phenomenon. Pretreatment with WEB 2170, a platelet activating factor (PAF) antagonist, also significantly inhibited cell influx. These results suggest that thromboxane, LTB4 and PAF mediate cell infiltration in this IC arthritis model.


Introduction
Immune complex (IC) deposition is an important event in the pathogenesis of rheumatoid arthritis (RA). This assumption is supported by the detection of IC deposits in the synovium and articular cartilage, as well as by the possibility of rheumatoid factor self-aggregation in large insoluble immune complexes and the association of circulating immune complexes with disease severity in some RA patients. The most commonly used experimental models of RA, the antigen-induced arthritis and the adjuvantinduced arthritis, exhibit close resemblance to the pathophysiology of RA. However, they do not allow a precise investigation of the isolated contribution of IC to the development of arthritis.
Induction of a reverse passive Arthus reaction in rabbit knee joints promotes a modest increase in vascular permeability, pronounced cellular infiltration, mostly polymorphonuclear (PMN) neutrophils, and cartilage destruction, being considered as a suitable experimental model to study the mechanisms involved in articular lesions dependent on IC formation. 2 Although the recruitment and activation of these cells in IC-induced lesions is considered essential to the full-blown development of the reaction, the factors governing these phenomena are still obscure.
Previous data have shown that eicosanoids and platelet activating factor (PAF) are involved in IC disease models. These lipid mediators can alter vascular permeability and microcirculatory blood flow. In addition, they promote cell migration and activation, causing release of other inflammatory mediators as well as reactive oxygen intermediates and lysosomal enzymes, which are effectors of local tissue damage. Detection of prostaglandin E2, leukotriene B4 and thromboxane B 2 in the bronchoalveolar lavage fluid 4 and peritoneal exudate 5 in Arthus reactions in rats argues for the participation of these substances in IC-induced lesions. Moreover, lung haemorrhagic lesions induced by IC were shown to be mediated by PAF and LTB4 .6 However, the role of these mediators in IC-induced arthritis has not been established yet.
The aim of the present study was to characterize the early events of an arthritis caused exclusively by immune complexes. To this purpose, a reverse passive Arthus reaction was induced in the knee joint of rats. Synovial histopathological alterations, vascular permeability, and cell influx into the joints were evaluated. The contribution of PAF and eicosanoids to cell infiltration into were determined using a Neubauer chamber. Difthe joints was also investigated, ferential counts were performed on fixed and stained cell suspensions (0.05% ceystal violet dis-Materials and Methods solved in 30% acetic acid). (10mg/kg, per os), was injected 20h prior to injection of chloral hydrate (400mg/kg). Rabbit arthritis. IgG antibodies to BSA (2001.tg in 0.1 ml saline) were injected into the knee joints, followed by Drugs and reagents.. The compounds L-663,536, i.v. injection of I mg BSA in 0.2 ml saline. Control L-660,711, and L-655,240 were kindly supplied by animals received the same amount of the anti-Merck Frosst Lab., Canada, and WEB 2170 by body intraarticularly followed by saline injection Boehringer-Ingelheim, Germany. Indomethacin, BSA, rabbit IgG anti-BSA antibodies, and FITClabelled goat anti-rabbit IgG were purchased Collection of synovial exudate: The animals were from Sigma Chem Co., USA, and FITC-labelled anaesthetized and killed by ceeeical dislocation goat anti-human IgG from Hemagen, USA. and exsanguinated. The synovial cavity of the knee joints was then washed with 0.4ml saline Statistical analysis.. ANOVA or Student's t-test containing 5 U/ml heparin. The exudates were were used to compare means, p < 0.05 was concollected by aspiration, centrifuged (500 x g for sidered significant. 630nm and compared to a standard curve, a synovium specimen obtained 6h after arthritis Results were expressed as l.tg of Evans blue per induction, showing interstitial haemorrhage and ml of exudate, infiltration of PMN cells. The control groups exhibited none of these alterations (not Histopathological analysis and immunofluores-shown). The direct immunofluorescence, using cence technique: Synovial tissue sections were FITC anti-rabbit IgG, aspect of a synovium speprocessed for routine haematoxylin-eosin stain-cimen from a rat subjected to the Arthus reacing and also for direct immunofluorescence, tion is depicted in Fig. I(B), showing granular Briefly, 3 l.tm thick cryosections of synovial IC deposits which were not seen in the control tissues were mounted on glass slides, fixed in group (not shown). A direct immunofluoresacetone at 20C for 8min, and air-dried. Tissue cence using anti-human IgG conjugate did not sections were incubated with FITC-labelled goat show fluorescence staining, thus ensuring the anti-rabbit IgG or FITC-labelled goat anti-human specificity of the anti-rabbit IgG conjugate used IgG diluted to 1:100 in PBS (0.1 M NaCl, pH7.2) (data not shown). for 30 min at 37C in wet chamber. After washing twice in PBS for 10 min, slides were mounted in Increase in vascular permeability: The alteration glycerine at pH 9.0 and observed by using fluorin vascular permeability, measured as the concenescence microscopy, tration of Evans blue dye extravasated into the joint cavity was evaluated 10 min, 30 min, I h, and Cell counting: The synovial exudates were centri-3 h after arthritis induction. A significant (p < fuged (500 x g for 10min) and the cell pellet 0.05) increase in extravasation of the dye was resuspended in 0.1ml saline. Total cell counts observed only in the 3h exudates, being 0.59tg/ F A C Rocba et al. Role of lipid mediators on the cellular influx.. In order to investigate the role of eicosanoids and PAF in the cell influx into the knee joints, we studied the effect of treating the animals with indomethacin (a cyclooxygenase inhibitor), L-663,536 (a lipoxygenase inhibitor), L-655,240 (a thromboxane antagonist), L-660,711 (a peptidoleukotrienes antagonist) or WEB 2170 (a PAF antagonist) prior to arthritis induction. Figure 3 shows that pretreatment with either indomethacin or L-660,711 did not alter the number of cells present in the exudate collected 6 h after induc- In this IC arthritis model, the groups subjected (213.32 _ _ _ 37.56 x 10 cells/ml), with a preto the reverse passive Arthus reaction were dominance of PMN cells. After 24h of arthritis, always compared to the group that received the there was still a statistically significant difference antibody intraarticularly, but not the antigen.
between the two groups although, at this time, Since the mere administration of antibody into the difference was due to a five times greater the joints may result in a mild synovitis, the comnumber of mononuclear cells in the IC group, parison of both groups was essential to allow the study of the inflammatory process induced solely by the deposition of immune complexes.
The histopathological analysis revealed vascular congestion and cellular infiltrate, composed mainly of neutrophils at 6 h and mononuclear cells at 24h. Similar findings were reported in rabbits subjected to IC arthritis. 2 The demonstration of granular immunofluorescent deposits in the synovia of the rats subjected to the Arthus reaction, but not in the control group, provides a strong indication that the inflammatory changes were mostly caused by IC deposition in the synovial tissue.
The increase in vascular permeability as measured by the analysis of the concentration of Evans blue dye extravasated to the synovial cavity was significant only when measured over a period of 3 h following induction of the reaction.
These results differ from those obtained in a model of active Arthus reaction in rabbits, where a significant increase in vascular permeability was observed] Since active immunization induces production of precipitating antibodies as well as antibodies cytophylic for mast cells, the latter would promote release of mast cell derived mediators which would account for the increased vasopermeability. In the present study, we used rabbit IgG antibodies which are not cytophylic for rat mast cells, in order to ensure a reaction mediated predominantly by immune complexes. In this case, increase in vascular permeability is not a prominent event, as we have previously reported in IC-induced alveolitis. One possible speculation is that, in this type of reaction, there is a release of vasoconstricting substances which may limit plasma extravasation. In support of this assumption, it was shown that release of slow reacting substance of anaphylaxis (SRS-A) in passive Arthus reaction markedly inhibited carrageenin oedema, s The infiltration of inflammatory cells into the synovial cavity peaked 6 h after arthritis induction, with a predominance of PMN cells, and declined after 24 h, when mononuclear cells became predominant. These results are in accordance with previous data showing leukocyte infiltration in the reverse passive Arthus reaction in rabbit joints 2 as well as in the antigen-induced arthritis model. 7,9 Pretreatment with specific inhibitors or antagonists of eicosanoids and PAF showed that That the concentrations of the inhibitors used in this study were adequate was confirmed by the reduction of LTB4 and PGE2 levels in the synovial exudates by L-663,536 and indomethacin, respectively. The concentration of the PAF antagonist, WEB 2170, was assayed against PAFinduced vasopermeability in rats, being 5mg/ kg--the dose that provoked the best inhibition. A recent publication from our group demonstrates the efficacy of this dose on PAF-induced vascular permeability in rat lungs. 14 The concentration of L-655,240 was taken from a relevant published report where this compound selectively antagonized the effects of a thromboxane analogue both in vivo and in vitro. 1 The peptido-leukotriene antagonist L-660,711 was used at doses ten times higher than those reported to inhibit antigen-induced bronchoconstriction in sensitized rats. 13 The fact that an inhibitor of 5-1ipoxygenase inhibited cell influx whereas an antagonist of F. A C Rocha et al.
peptido-leukotrienes had no effect suggests that illustrated by its detection in the synovial fluid of leukotriene B4 is responsible for this phenomexperimental models of arthritis. 25 Furthermore, enon. Taken together, these results suggest that it was reported that PAF contributes to cell LTB4, thromboxane, and PAF contribute to cell migration in the antigen-induced arthritis model infiltration in IC arthritis, in rabbits, an effect that could result from either The presence of LTB 4 in the synovial fluid of a direct action of PAF or an indirect one via the arthritis patients has already been demon-release of other inflammatory mediator(s). 2 We strated. 15 Moreover, neutrophils and macro-have recently observed that the PAF antagonist phages from RA patients stimulated with calcium WEB 2170, but not indomethacin, L-663,536 or Lionophore A23187 release more LTB 4 than do 655,240 inhibited TNF release in this IC arthritis those of healthy volunteers. 16 Besides promoting model (Rocha et al., submitted). These results PMN adhesion to vascular endothelial cells, 1 suggest that the effect of PAF on cell migration LTB 4 is also a potent chemoattractant to both might be through TNF release. It also points to neutrophils and macrophages. 18 However, few interesting interactions between lipid mediators studies have focused on the importance of leukoand cytokines. Another mechanism that might trienes in arthritis. To our knowledge, this is the contribute to cell migration is related to complefirst demonstration that LTB 4 contributes to leument activation. However, we believe that C5 kocyte infiltration in arthritis, fragments are not essential for cell migration in Non-steroidal anti-inflammatory drugs IC reactions, based on previous observations that (NSAIDs) are the most commonly prescribed the temporal profile of neutrophil accumulation substances to treat RA patients. It is well known in response to IC deposition is similar in coisothat these drugs provide only symptomatic relief genic strains of mice, where one of them is to RA patients, without altering the disease genetically deficient in the C5 component. 2 outcome. This could be due to the cessation of The exact role of neutrophils in the pathogenan inhibitory effect of the cyclooxygenase prodesis of arthritis is a subject of debate. Although it ucts on leukotrienes synthesis 19 or by a preferenseems indisputable that these cells are responsitial shift of the substrate pool into the lipoxygen-ble for the release of inflammatory mediators, ase pathway, thus enhancing leukotriene lysosomal enzymes and reactive oxygen species, 20 production. This putative increase in leukoa direct role for them in mediating cartilage triene release could partially explain the lack of breakdown is not clearly established. In an IC efficacy of NSAIDs in modifying the natural arthritis model in mice it was suggested that course of RA.
these cells are involved in cartilage destruction. 2s Previous data have shown that thromboxane However, in the antigen-induced arthritis model A2 is chemotactic to mouse neutrophils in in rabbits, it was suggested that neutrophils vitro. 21 Additionally, econazole, an imidazole might not be essential for cartilage degradation, compound which inhibits thromboxane synthsince the animals still presented cartilage breakesis, inhibited cell infiltration in carrageenin down after being rendered neutropenic by prepleurisy. 22 This effect of thromboxane could be treatment with nitrogen mustard. 9 More recently, explained by its ability to enhance neutrophil in an in vitro study, it was demonstrated that adhesion to endothelial cells, via activation of the neutrophils were able to degrade both proteogly-CD18 receptor on neutrophils. 23 In our experi-cans and collagen in human cartilage, an effect ments, the thromboxane antagonist significantly that was facilitated by previous stimulation of the inhibited cell influx, suggesting that thromboxneutrophils with immunoglobulins. 29 Such data ane is involved in cell migration in arthritis. We are corroborated by previous studies suggesting have previously demonstrated that thromboxane that immunuglobulins might facilitate neutrophil is involved in PAF-induced vasoconstriction in adhesion to cartilage surface, where these cells the synovial microcirculation of normal rabbit could then be activated. Another explanation joints. 24 Our present data indicating the participa-could be that activation of the Fc receptor of tion of LTB 4 and TXB2 in the cell infiltration in previously primed neutrophils might potentiate this IC arthritis model suggest that these lipid the release of inflammatory mediators, as well as mediators may have a more important regulatory release of reactive oxygen species and lysosomal role on inflammatory arthropathies than has pre-enzymes by these cells, thus resulting in viously been ascribed to them. increased tissue damage. Our results also showed that pretreatment of Since IC deposition is involved in some rheuthe animals with WEB 2170, a specific PAF matic conditions, including rheumatoid arthritis, antagonist, significantly reduced cell infiltration, systemic lupus e/thematosus and vasculitis, the thus implicating PAF in this phenomenon. The possibility of studying the mechanisms governing potential importance of PAF in arthritis can be IC-induced lesions in joints makes this model relevant to the understanding of the pathophysiology of synovitis. The IC arthritis model described here proved to be a suitable and reproducible model of experimental arthritis. The participation of eicosanoids, mainly leukotrienes and thromboxane, as well as PAF in the regulation of cell migration described here suggests that inhibitors or antagonists of these substances might be valuable alternatives to the therapy for inflammatory arthropathies.