To determine whether treatment with the chimeric anti–tumor necrosis factor α antibody infliximab could reduce cellularity by the induction of apoptosis in synovial tissue.
To determine whether treatment with the chimeric anti–tumor necrosis factor α antibody infliximab could reduce cellularity by the induction of apoptosis in synovial tissue.
Twenty-four rheumatoid arthritis patients with active disease were randomized to receive either infliximab (3 mg/kg) (n = 12) or placebo (n = 12) intravenously. All patients were subjected to arthroscopic synovial biopsy directly before initiation of treatment. A second arthroscopic synovial biopsy of the same index joint was performed 48 hours after the first arthroscopy. After the second arthroscopy, the patients who had initially received placebo were also treated with infliximab in an extension study. A third arthroscopy was performed in all patients on day 28. Immunohistologic analysis was performed to characterize the cell infiltrate. In situ detection of apoptotic cells was performed by TUNEL assay and electron microscopy.
At 48 hours after initiation of infliximab treatment, there was a significant reduction in the number of intimal macrophages; this was not observed in the placebo group. The number of sublining macrophages, T cells, and plasma cells also tended to be decreased in infliximab-treated patients, but not in the placebo group. Of interest, we did not detect any increase in the number of apoptotic cells after infliximab treatment.
Infliximab therapy may reduce the number of inflammatory cells in rheumatoid synovial tissue as soon as 48 hours after initiation of treatment, but apparently not by induction of apoptosis. Conceivably, decreased cell infiltration primarily results from early inhibition of cell migration.
Tumor necrosis factor α (TNFα) blockade by treatment with TNF inhibitors, such as infliximab, etanercept, and adalimumab, has emerged as an important therapy for patients with rheumatoid arthritis (RA) (1–3). The clinical experience with this approach clearly underscores the key role of TNFα in the pathogenesis of RA, but the exact mechanism of action of TNF inhibition remains to be elucidated. Of special interest is the rapidity of the effects of TNFα blockade compared with conventional treatment with disease-modifying antirheumatic drugs (DMARDs). Clinical improvement has, for instance, been described as early as 1 week after initiation of treatment with the chimeric anti-TNFα monoclonal antibody infliximab (4).
Studies of the effects of TNFα blockade on the primary site of inflammation, the synovium, have revealed that synovial cell infiltration is reduced 4 weeks after a single infusion of 10 mg/kg of infliximab (5). Similarly, a reduction in synovial cellularity was observed 2 weeks after a single infusion of 10 mg/kg of infliximab (6). Decreased synovial cell numbers after infliximab therapy have also been reported in patients with spondylarthropathy (7). It has been suggested that the effects of TNFα blockade on the synovium could be explained in part by decreased cell migration (5, 6, 8). Anti-TNFα antibody treatment results in decreased expression of adhesion molecules in synovial tissue (5), reduced levels of circulating soluble adhesion molecules (8), and diminished chemokine levels in the synovium (6). Both adhesion molecules and chemokines are involved in the migration of inflammatory cells toward the synovial compartment. An inhibitory effect of infliximab therapy on synovial angiogenesis (9) could further reduce cell trafficking to the joints. The view that anti-TNF therapy may indeed affect cell migration is supported by the recent observation that infliximab treatment results in a significant decrease in signal on gamma camera images of the joints after intravenous injection of 111In-labeled granulocytes (6). Although there is strong circumstantial evidence, direct proof that anti-TNF treatment inhibits monocyte migration is not yet available.
In addition to recruitment of inflammatory cells, local retention and cell proliferation contribute to the increased cellularity of rheumatoid synovium (10). Moreover, recent studies suggest that impaired apoptosis could play an important role in the development of the hyperplastic synovium observed in RA patients (11). In fact, only very few apoptotic cells are found in the rheumatoid synovium in any stage of the disease (12, 13), despite the presence of fragmented DNA (12, 14, 15). Conceivably, anti-TNFα treatment could also decrease synovial cell infiltration by the induction of apoptosis.
In vitro studies using murine myeloma cells expressing mutant human TNFα have shown that infliximab may kill these cells by both antibody-dependent cellular toxicity and complement-dependent cytotoxicity effector mechanisms (16). Of interest, in vitro treatment with infliximab of peripheral blood monocytes from patients with Crohn's disease resulted in the induction of apoptosis through activation of caspase 8 and the mitochondrial pathway (17), a mechanism that is clearly different from complement-dependent cytotoxicity. A recent uncontrolled in vivo study in 10 patients with Crohn's disease revealed an increase in TUNEL-positive cells (cells exhibiting DNA fragmentation) in colonic biopsies at 24 hours after infusion of 5 mg/kg of infliximab (18). It has been suggested that the induction of apoptosis by anti-TNFα antibodies, but not by soluble TNF receptors, could explain the difference in therapeutic efficacy between infliximab and etanercept in Crohn's disease (19, 20).
It should be noted, however, that the experimental conditions in the in vitro studies (16, 17) were very different from the conditions in the inflamed joints of RA patients. The rheumatoid synovium is a particularly anti-apoptotic environment (11), and infliximab doses used for treatment of RA patients are usually lower than those used in Crohn's disease. Moreover, in contrast to Crohn's disease, there is, on average, no clear difference in the clinical response to treatment with either anti-TNFα antibodies or soluble TNF receptors in RA patients, suggesting that neutralization of the effects of TNFα may be more important than apoptosis induction in this disease (1–3). Therefore, we investigated whether treatment with 3 mg/kg of infliximab could induce apoptosis in the rheumatoid synovium. In addition, we studied for the first time the early effects of TNFα blockade on synovial cellularity. To this end, we conducted a placebo-controlled trial in 24 RA patients for 48 hours, followed by an open extension study for 28 days. Serial synovial biopsy samples were obtained before treatment, after 48 hours, and after 28 days.
Twenty-four patients with active RA, defined as a Disease Activity Score in 28 joints (DAS28) (21) of at least 4.8 (at the screening visit), despite maximal methotrexate treatment (either 25 mg/week or the maximum tolerable dose), were evaluated. All patients fulfilled the American College of Rheumatology (formerly, the American Rheumatism Association) criteria for RA (22) and treatment with at least 2 DMARDs had failed. Stable prednisone therapy at a dosage of ≤10 mg/day and nonsteroidal antiinflammatory drug (NSAID) treatment were allowed. In addition, all patients were receiving methotrexate at a dosage that had been kept stable since at least 4 weeks prior to inclusion in the study. All patients gave informed consent, and the study protocol was approved by the Medical Ethics Committee of the Academic Medical Center, University of Amsterdam.
Twenty-four patients were randomized to receive either 3 mg/kg of infliximab (n = 12) or placebo (n = 12) intravenously. All 24 patients were subjected to an arthroscopic synovial biopsy of a swollen knee, wrist, or ankle joint (23) immediately before initiation of treatment. In all 24 patients, a second arthroscopic synovial biopsy of the same index joint was performed 48 hours after the first arthroscopy. After the second arthroscopy, the patients who had initially received placebo were also treated with infliximab (3 mg/kg) in an extension study. A second administration of 3 mg/kg of infliximab was administered to all patients on day 15. A third arthroscopy was performed in 20 patients on day 28 (2 placebo- and 2 infliximab-treated patients refused the third arthroscopy).
Small-bore arthroscopy was performed under local anesthesia, and synovial tissue samples were obtained from the entire joint using a 2-mm grasping forceps (Storz, Tuttlingen, Germany), as previously described (23). On each occasion, an average of at least 6 biopsy samples were obtained. The synovial biopsy samples were collected and snap-frozen en bloc in TissueTek OCT (Miles, Elkhart, IN). Frozen blocks were stored in liquid nitrogen until sectioned for staining. Sections (5 μm) were cut in a cryostat and mounted on Star Frost adhesive glass slides (Knittelgläser, Braunschweig, Germany); slides were stored at −70°C until immunohistochemical analysis could be performed. For electron microscopic analysis, tissues were immediately transferred into fixative and stored at room temperature until analyzed.
Serial sections were stained with the following mouse monoclonal antibodies: anti-CD3 (Becton Dickinson, San Jose, CA) to detect T cells, anti-CD22 (CLB-B-ly; Central Laboratory of The Netherlands Red Cross Blood Transfusion Service, Amsterdam, The Netherlands) to detect B cells, anti-CD38 (Leu-17; Becton Dickinson) to detect plasma cells, anti-CD68 (EBM11; Dako, Glostrup, Denmark) to detect macrophages, and anti-CD55 (clone 67; Serotec, Oxford, UK) to detect fibroblast-like synoviocytes (FLS). For control sections, the primary antibodies were omitted or irrelevant antibodies were applied.
Staining for cell markers was performed as described previously (24). Following a primary step of incubation with monoclonal antibodies, bound antibody was detected according to a 3-step immunoperoxidase method. Horseradish peroxidase activity was detected using hydrogen peroxide as substrate and aminoethylcarbazole as dye.
Apoptotic cells in frozen synovial tissue sections were detected by TUNEL analysis of apoptosis-induced DNA strand breaks, using an in situ cell death detection assay, according to the manufacturer's instructions (Roche Diagnostics, Mannheim, Germany).
All sections were coded and randomly analyzed by computer-assisted image analysis. For all markers, 18 high-power fields were analyzed. CD68 expression was analyzed separately in the intimal lining layer and the synovial sublining. The images of the high-power fields were analyzed using the Qwin analysis system (Leica, Cambridge, UK), as described previously (25).
Small pieces of fresh synovial tissue were immersed in Karnovsky's fixative (Department of Pathology, Academic Medical Center, Amsterdam, The Netherlands) and embedded in Epon. Thin sections were briefly pretreated with diluted tannic acid for better visualization of extracellular matrix components before treating them with uranyl and lead according to standard procedures (26).
Wilcoxon's nonparametric signed rank test for matched pairs was used to compare data within each group. The Mann-Whitney U test was used to compare differences between the infliximab group and the placebo group. Values are expressed as the mean ± SEM.
The clinical study group consisted of 18 women and 6 men. Their mean age was 52 years (range 24–81 years), and their mean duration of disease was 256 months (range 16–765 months). Erosions were present in 15 of the 24 patients, and 18 patients were seropositive for IgM rheumatoid factor. Three patients were positive for antinuclear factor, and none of the patients had anti–double-stranded DNA antibodies. All patients had active disease, with mean DAS28 scores at study entry of 6.5 (range 5.6–8.0) in the infliximab-treated group and 6.1 (range 4.9–8.0) in the group initially treated with placebo. All patients were treated with stable doses of methotrexate; the mean dosage was 16.2 mg/week (range 5.0–30.0). On average, the patients had failed treatment with 4.3 DMARDs (range 2–9) before entering the study.
First, we investigated the early effects of TNFα blockade on the synovium by immunohistologic analysis of serial synovial biopsies of patients who were treated with infliximab for the first time compared with patients who received placebo. The mean scores for cell markers at baseline and after 48 hours are shown in Figure 1, and the changes in the mean number of cells after 48 hours compared with baseline are shown in Figure 2. Nonparametric analysis revealed a significant reduction in the scores for CD68+ intimal macrophages (mean ± SEM −142 cells/mm2 ± 67; P = 0.009) at 48 hours after initiation of infliximab therapy, whereas there were no significant changes in the placebo group (Figures 2 and 3). The numbers of sublining macrophages, T cells, and plasma cells also tended to decrease after infliximab therapy, but not in the placebo group; these differences did not reach statistical significance, possibly due to the small number of patients. We did not observe a clear-cut decrease in the numbers of B cells or FLS in infliximab-treated patients compared with the placebo group. The negative control sections were all negative. Thus, these data show that infliximab treatment results in decreased synovial cellularity early after the first infusion.
To determine whether increased apoptosis contributed to the early decrease in synovial cellularity, TUNEL assays were performed. This is a sensitive technique for the detection of apoptotic cells. The mean scores at baseline and after 48 hours are shown in Figure 1, and the changes after 48 hours compared with baseline are shown in Figure 2. Of interest, we did not observe any increase in the number of TUNEL-positive cells after treatment (mean ± SEM 0 ± 14 in the infliximab group, 5 ± 11 in the placebo group; both not statistically significant) (Figures 2 and 4). To exclude the possibility of a false-negative result as a consequence of decreased cellularity, we repeated the analysis after correction for total cell counts. There was also no increase in the number of TUNEL-positive cells/total nuclei after infliximab treatment.
Synovial tissue selected at the third arthroscopy was available from 20 patients who received 2 infusions (on days 0–3 and on day 15) of infliximab during the extension study. Again, there was no increase in the number of apoptotic cells in the synovium on day 28 compared with the baseline biopsy sample (mean ± SEM change in TUNEL-positive cells −10 ± 11; not statistically significant).
Electron microscopic analysis confirmed previous observations that apoptotic cells are virtually absent in the RA synovium, even when TUNEL-positive cells can be detected. There was no increase in the number of apoptotic cells after either infliximab or placebo treatment (Figures 4 and 5). Taken together, these data unambiguously show that the reduction in synovial cell infiltration after therapy with infliximab at a dose of 3 mg/kg cannot be explained by the induction of apoptosis in the rheumatoid synovium.
To determine the effects of infliximab therapy after more prolonged treatment, a third arthroscopy was performed on day 28. The patients who had initially received placebo were also treated with infliximab (3 mg/kg) on day 3. A second dose of 3 mg/kg infliximab was administered to all patients on day 15. Synovial tissue samples obtained at the first arthroscopy were compared with those selected at the third arthroscopy. The DAS28 was measured on the same day as the first and third arthroscopies. Clinical improvement was defined as a decrease in the DAS28 of >1.2 (27).
Synovial tissue selected at the third arthroscopy was available from 20 patients. The results were generally comparable to those obtained after 48 hours. There were clear trends toward decreased infiltration by macrophages, T cells, and plasma cells, although the differences did not reach statistical significance. B cells and FLS were not affected after 28 days, similar to the data obtained after 48 hours. Decreased numbers of T cells (mean ± SEM −28 ± 38), plasma cells (−30 ± 72), intimal macrophages (−38 ± 104), and sublining macrophages (−371 ± 187) were observed in clinical responders (n = 12), but not in patients who did not exhibit a decrease in the DAS28 of >1.2 after 28 days (n = 8) (Figure 6).
The results presented here are the first to show decreased synovial cellularity early after initiation of infliximab treatment. The effects could be seen as early as 48 hours after an infliximab infusion and were generally comparable to those observed on day 28. Interestingly, these changes could not be explained by induction of apoptosis in the synovium.
The infliximab dose used in the present study (3 mg/kg) is markedly lower than that used in previous synovial biopsy studies (10 mg/kg) evaluating the effects of infliximab in RA patients (5, 6, 28). We chose to use this dose since this is currently the usual starting regimen for RA. A second difference is that all patients in this study were also receiving stable doses of methotrexate, consistent with current clinical practice. Still, the effects on the synovial cell infiltrate were generally similar to those reported in the earlier studies, although perhaps less pronounced, which could be explained by the lower infliximab dose as well as the combination therapy with methotrexate. Previous studies have shown that methotrexate itself may inhibit synovial cell infiltration (29, 30). Therefore, it might be more difficult to detect an additional reduction in cellularity after adding infliximab to the therapeutic regimen.
A novel finding in this study is that the effects on synovial inflammation occur as soon as 48 hours after the first infusion of infliximab. Similarly, an open study of 10 RA patients revealed decreased expression of adhesion molecules and chemokines in the synovium, resulting in decreased cell trafficking to the joints, 24 hours after high-dose intravenous pulse corticosteroid treatment (31, 32). Consistent with the findings in the present study is the clinical experience that patients may start to feel better very early after initiation of infliximab treatment (4). The statistically significant decrease in macrophage numbers is of interest in light of the earlier studies showing an association between the number of synovial tissue macrophages and clinical signs of disease activity (33–35). After 1 month, the most pronounced reduction of macrophage numbers was found in the patients with clinical improvement. Taken together, these studies suggest that macrophage numbers could be used as a surrogate marker for arthritis activity in studies evaluating the effects of novel antirheumatic therapies.
TUNEL assays were used to detect apoptotic cells in the synovium. It should be noted that this method, using DNA nick end-labeling, is a very sensitive technique. We confirmed previous observations showing the presence of TUNEL-positive cells in rheumatoid synovial tissue (12, 12–15). The discrepancy between DNA fragmentation, which is probably a result of locally produced nitric oxide and oxygen radicals, and the low number of apoptotic cells in rheumatoid synovial tissue detected by electron microscopy, as shown in previous studies (12, 13) and confirmed here, could theoretically be explained by the rapid removal of apoptotic cells by macrophages (36). The proportion of cells showing DNA strand breaks is, however, so great that this disparity suggests impaired apoptosis (11). This could be explained by a variety of factors, including a relative deficiency of functional Fas ligand in the RA joint (37), nuclear factor κB activation (38), genotoxic changes in response to the toxic environment (39), and the expression of anti-apoptotic proteins such as interleukin-15 (40), FLIP-inhibitory protein (41), and sentrin (42).
To our surprise, the data shown here indicate that infliximab treatment at the dose used in this study did not alter the anti-apoptotic state of the synovium. In keeping with this idea is the fact that patients treated with infliximab do not develop symptoms of a cell lysis syndrome. In addition, infliximab treatment of RA patients may cause a transient leukocytosis (8), suggesting blockade of cell migration rather than apoptosis induction. On the other hand, an increase in TUNEL-positive cells was noted in intestinal biopsy tissues after infliximab treatment of patients with Crohn's disease (18). The discrepancy might be explained by differences in disease pathogenesis and perhaps by the difference in infliximab dose. Apparently, neutralization of the effects of TNFα is sufficient to induce clinical improvement in RA, even without induction of apoptosis at the site of inflammation. Consistent with this view is the observation that treatment with anti-TNFα antibodies and soluble TNF receptors appears, on average, to be equally effective in RA (although no direct comparative study between the two drugs has been conducted), which is different from the effects in Crohn's disease, where induction of apoptosis may be pivotal in achieving amelioration of the disease (19).
There tended to be more inflammatory cells in the synovial tissue of patients who received infliximab than in those who received placebo, although the differences did not reach statistical significance. This can be explained by interindividual variability and chance, since the study was completely blinded. Previous work has shown that there is marked interindividual variability in synovial inflammation among RA patients (34, 43). A possible difference between the 2 groups did not affect the main statistical analysis in this study, i.e., the analysis for matched pairs within each group. This approach has also been used in previous studies evaluating the effects of TNFα blockade on the synovium (5, 6, 28). Of importance, a possible difference in the severity of synovial inflammation between the 2 groups does not affect our major conclusion, that an increase in the number of apoptotic cells (in any group at any time point) was not detected.
This study shows significantly reduced macrophage infiltration as soon as 48 hours after initiation of infliximab treatment, which cannot be explained by induction of apoptosis in the synovium. Based on the literature discussed above, we hypothesize that neutralization of the effects of TNFα and decreased cell migration might be sufficient to reduce synovial cell infiltration and achieve clinical improvement.
We would like to thank Dr. K. P. Dingemans and Mr. M. A. van den Bergh Weerman, Department of Pathology, Academic Medical Center, Amsterdam, The Netherlands, for performing the electron microscopy studies.