NF-κB is one of the most important proinflammatory transcription factor families involved in the pathogenesis of rheumatoid arthritis (RA) and contributes significantly to synovial inflammation (1–3). NF-κB activation can be induced by many different stimuli and is well controlled by various endogenous mechanisms, including the IκB proteins that are under tight control of the IKK complex. Under inflammatory conditions, NF-κB is mainly activated via this canonical signal transduction pathway (for review, see ref.4). We previously established that IKKβ is a key regulator of synovial inflammation and demonstrated that intraarticular administration of a dominant-negative form of IKKβ using adenoviral gene therapy (AdIKKβ-DN) significantly reduced clinical signs of arthritis activity (5). The consequences of local IKKβ inhibition on synovial cytokine expression at the site of inflammation have as yet not been studied in in vivo models of RA.
The present exploratory study was conducted to provide more insight into the mechanism of action of AdIKKβ-DN gene therapy. We investigated the consequences of local inhibition of IKKβ on synovial cellularity, cytokine expression, and the expression of matrix-degrading enzymes.
Serial sections from paraffin-embedded ankle joints of rats with adjuvant-induced arthritis (AIA) treated with intraarticular AdIKKβ-DN gene therapy (n = 10) were analyzed and compared with AdGFP-injected control animals (n = 10) (5). After blocking of endogenous peroxidase activity (H2O2) and antigen retrieval (citrate buffer), sections were incubated overnight at 4°C with antibodies specific for rat interleukin-1β (IL-1β), tumor necrosis factor α (TNFα), IL-6, and IL-10 in phosphate buffered saline (PBS)/1% bovine serum albumin (BSA) (10 μg/ml; all from R&D Systems, Minneapolis, MN). In addition, immunohistochemical stainings were carried out to detect matrix metalloproteinase 3 (MMP-3) (10 μg/ml; Chemicon, Temecula, CA) and tissue inhibitor of metalloproteinases 1 (TIMP-1) (5 μg/ml; R&D Systems). After incubation with secondary horseradish peroxidase (HRP)–conjugated swine anti-goat or goat anti-mouse antibodies (Dako, Glostrup, Denmark) in PBS/1% BSA, signal amplification was performed using biotinylated tyramine (Perkin Elmer Life Sciences, Emeryville, CA) followed by streptavidin–HRP (Dako) in PBS/1% BSA as described previously (6). Finally, peroxidase activity was detected using 0.02% 3-amino-9-ethyl-carbazole (Vector Laboratories, Peterborough, UK) as dye and H2O2 as substrate. Sections were counterstained with Mayer's hemalum solution (Merck, Darmstadt, Germany) and mounted with Kaiser's glycerol gelatin mounting medium (Merck). Computer-assisted image analysis was used to evaluate stained sections in a random order (7). Results were expressed as integrated optical density/mm2, which is proportional to the cellular concentration of protein multiplied by the area of positive staining.
Data were analyzed for statistical significance using SPSS version 11.5.1, by multivariate analysis combining the results obtained for proinflammatory markers, and by multivariate analysis combining the results obtained for the antiinflammatory markers (Hotelling's Trace test), acknowledging the a priori hypothesis that results obtained with proinflammatory markers would be different from results obtained with antiinflammatory markers. Descriptive data on the results for individual markers were analyzed for statistical significance using the Mann-Whitney U test. P values less than 0.05 were considered significant.
Intraarticular AdIKKβ-DN gene therapy resulted in a 50% reduction of synovial cellularity (mean ± SEM 104 ± 23 versus 212 ± 35 cells/mm2 with IKKβDN and green fluorescent protein [GFP] treatment, respectively; P < 0.05). Multivariate analysis of the data on the proinflammatory markers, with IL-1β, TNFα, IL-6, and MMP-3 as dependent variables and AdGFP and AdIKKβ-DN as fixed factors, resulted in a Hotelling's Trace test value of 0.041. This signifies that statistical differences in the expression of individual proinflammatory markers are valid. We observed significantly reduced synovial IL-1β and TNFα expression in AdIKKβ-DN–treated rats (P < 0.05), (Figure 1). Expression of IL-6 was also reduced in the AdIKKβ-DN–treated group, although this did not reach statistical significance. Interestingly, expression of the antiinflammatory cytokine IL-10 was not altered by AdIKKβ-DN treatment (Figure 1). These data are consistent with the recognized clinical efficacy of AdIKKβ-DN gene therapy in animal models of RA (5). Next, we investigated the effects of AdIKKβ-DN gene therapy on the expression of MMP-3 and TIMP-1 in synovial tissue. Local IKKβ inhibition resulted in significantly decreased MMP-3 expression (P < 0.05), whereas expression of TIMP-1 was unaffected (Figure 1). Additionally, multivariate analysis of the data on the antiinflammatory markers (IL-10 and TIMP-1) resulted in a Hotelling's Trace test value of 0.554, which is consistent with the lack of effect of AdIKKβ-DN gene therapy on these parameters (factors without an NF-κB–responsive promoter).
All known effective antirheumatic therapies result in decreased cellularity of rheumatoid synovial tissue, and for that reason synovial cellularity is used as a sensitive biomarker for the evaluation of novel therapies (8, 9). Of importance, the sensitivity to change of key synovial biomarkers is high in RA patients receiving active treatment (9). The results presented here, obtained using computer-assisted image analysis, show that intraarticular AdIKKβ-DN gene therapy results in a dramatic reduction of synovial cellularity, supporting the notion that this may be an effective approach to treat arthritis.
TNFα and IL-1β are cytokines that play an important role in the pathogenesis of RA, and targeting these cytokines has been proven effective in reducing arthritis severity (10). Conversely, the expression of TNFα and IL-1β is greatly reduced in rheumatoid synovial tissue after treatment with effective antirheumatic therapies such as prednisolone (8). The significantly reduced synovial IL-1β and TNFα expression in AdIKKβ-DN–treated rats is consistent with the clinical efficacy of AdIKKβ-DN gene therapy in AIA (5). Furthermore, in vitro experiments have demonstrated that IKKβ is essential for IL-1β production by synovial membrane cells and TNFα production by macrophages following stimulation with CD40 ligand (11). Although the decrease was not significant, IL-6 expression was reduced after AdIKKβ-DN gene therapy, which is consistent with the fact that the IL-6 promoter has DNA binding sites for several transcription factors including NF-κB (12). Thus, local IKKβ inhibition in vivo results in a reduction of known proinflammatory cytokines involved in RA.
Importantly, expression of the antiinflammatory cytokine IL-10 was not altered by AdIKKβ-DN treatment, which is consistent with the lack of NF-κB/Rel–responsive promoters in the IL-10 gene (13). This has important implications, since IKKβ inhibition selectively blocked the expression of proinflammatory cytokines, whereas the levels of the antiinflammatory cytokine IL-10 were not affected. Taken together, these data indicate that local canonical NF-κB inhibition, in addition to reducing synovial cellularity, results in an antiinflammatory shift in synovial cytokine expression, which may largely explain the beneficial effects of AdIKKβ-DN gene therapy on arthritis severity.
Matrix degradation and erosion of the connective tissue start at sites of attachment of synoviocytes to cartilage (14). At the invasive front, the synovial fibroblasts are found to express high levels of MMPs, such as MMP-1 (collagenase 1) and MMP-3 (stromelysin 1) (15). The activity of these enzymes is tightly regulated on a transcriptional level by activator protein 1 (AP-1) and NF-κB–responsive promoters (3, 16), as well as by interactions with specific inhibitors of the enzymatic activity at the posttranslational level (17). TIMP-1 strongly binds to both MMP-1 and MMP-3 and, like MMP-1 and MMP-3, is produced by synovial fibroblasts at sites of synovial attachment to cartilage (17).
Although intraarticular injection of AdIKKβ-DN resulted in significantly decreased MMP-3 expression, this did not result in reduced bone erosion in our previous study (5). This may be due to the fact that MMP expression is also dependent to a large extent on activation of the transcription factor AP-1 (3), which is thought to peak earlier than NF-κB in arthritis. AdIKKβ-DN gene therapy may therefore have been applied too late to prevent early MMP-induced bone destruction. Alternatively, intraarticular treatment with AdIKKβ-DN could result in a reduction of total MMP levels that is not biologically sufficient to prevent bone destruction in this animal model of severe RA. TIMP-1 expression was not significantly increased in synovial tissue of AdIKKβ-DN–treated rats, which is in accordance with the fact that no NF-κB–responsive elements in the promoter region of the TIMP-1 gene have been described (18–20).
The present data validate earlier in vitro experiments that used an adenoviral technique of blocking NF-κB in RA synovial cells through overexpression of IκBα or IKKβ-DN (11, 21). Our findings demonstrate that selective inhibition of the canonical NF-κB pathway in vivo using AdIKKβ-DN gene therapy results in decreased expression of proinflammatory cytokines and destructive MMP-3, while expression of antiinflammatory IL-10 and TIMP-1 is not affected. Detailed immunohistochemical analysis of the synovial tissue results in a better understanding of the underlying mechanisms of clinical efficacy. The data presented here support the notion that intraarticular IKKβ-DN gene therapy may be effective as a novel approach to treat RA (5).