Abnormal host defense against pathogens has been implicated in the pathogenesis of spondylarthropathy (SpA), a disease characterized by abundant synovial infiltration with innate immune cells. Given the role of Toll-like receptors (TLRs) in activation of innate inflammation and the occurrence of TLR-dependent infections after tumor necrosis factor α (TNFα) blockade treatment, the present study was undertaken to analyze TLRs and their modulation by TNFα blockade in SpA.
Peripheral blood mononuclear cells (PBMCs) were obtained from SpA and rheumatoid arthritis (RA) patients during infliximab therapy, and from healthy controls. TLR-2 and TLR-4 expression and TNFα production upon lipopolysaccharide (LPS) stimulation were analyzed by flow cytometry on different monocyte subsets. Synovial biopsy specimens from 23 SpA patients before and after infliximab or etanercept treatment, from 15 RA patients, and from 18 osteoarthritis (OA) patients were analyzed by immunohistochemistry.
Expression of TLR-4, but not TLR-2, was increased on PBMCs from patients with SpA, whereas both TLRs were increased in RA patients. TLR expression was particularly increased on the CD163+ macrophage subset. Infliximab reduced TLR-2 and TLR-4 expression on monocytes of SpA and RA patients, leading to lower levels than in controls and to impaired TNFα production upon LPS stimulation. In inflamed synovium, the expression of both TLRs and of CD163 was significantly higher in patients with SpA than in those with RA or OA. Paralleling the systemic effect, TLRs in synovium were down-regulated following treatment with infliximab as well as etanercept, indicating a class effect of TNFα blockers.
Inflammation in SpA is characterized by increased TLR-2 and TLR-4 expression, which is sharply reduced by TNFα blockade. These findings suggest a potential role of innate immunity–mediated inflammation in SpA and provide an additional clue regarding the mechanism of action as well as the potential side effects of TNFα blockade.
The spondylarthropathies (SpA) are a group of chronic inflammatory joint diseases characterized by axial involvement and peripheral arthritis. Although the pathogenesis of SpA is still unclear, a major clue is provided by findings in the reactive arthritis subtype, in which chronic joint inflammation is triggered by gastrointestinal or urogenital infection with bacteria (1). The absence of evidence of viable microbes in the joint (2–4), the frequent occurrence of gut inflammation in other SpA subtypes independent of gastrointestinal infections (5–8), and the influence of the germ-free state on the development of SpA-like gut and joint disease in HLA–B27–transgenic rats (9) suggest that bacterial triggering of the immune system, rather than infection itself, is important in SpA. Considering the strong genetic link with HLA–B27, it was originally proposed that microbial products might be presented, in the context of this specific class I major histocompatibility complex molecule, to cytotoxic CD8+ lymphocytes, which in turn may cross-react with self peptides in the joint. Despite extensive investigation, however, the role of antibacterial CD8+ lymphocytes in the pathogenesis of the disease remains to be formally demonstrated (10).
Recently, we presented the hypothesis that cells of the innate immune system, such as polymorphonuclear cells and macrophages, may be more important than lymphocytes in SpA inflammation. First, both peripheral blood and gut lymphocytes had impaired cytokine production in SpA (11, 12). Second, macrophages expressing the scavenger receptor CD163 were increased in both the gut and the joints of SpA patients, and local production of soluble CD163 inhibited T cell activation in the joint (13, 14). Third, in addition to the increased levels of CD163+ macrophages, we observed a significant increase of polymorphonuclear cells in the synovium of SpA versus rheumatoid arthritis (RA) patients (15, 16). Finally, levels of both CD163+ macrophages and polymorphonuclear cells, but not of CD3+ or CD20+ lymphocytes, correlated with global disease activity in SpA (17). Taken together, these data suggest that activation of innate immune cells in the gut as well as in the joints, by microbial products and/or cross-reactive self molecules, may be relevant to inflammation in SpA.
Recently, the activation of innate immune cells by pathogen-associated molecular pathways has been linked to the Toll-like receptors (TLRs) (18–21). The TLR family comprises 11 different members, of which membrane TLR-2 and TLR-4 are ligated by lipoproteins and peptidoglycans from gram-positive bacteria and by lipopolysaccharides (LPS) from gram-negative bacteria, respectively. Upon binding of their ligands, TLRs activate a complex signaling cascade which results ultimately in the production of mediators of inflammation. This pathway is crucial for host defense against a wide variety of pathogens including the invasive and intracellular bacteria involved in reactive arthritis (22–25), but can also lead to sterile inflammation in the absence of microbes (20). This may relate to the recognition by TLRs of self motifs such as heat-shock protein 70, fibronectin, hyaluronic acid, heparan sulfate, and fibrinogen (26–30). Given that these self ligands are present in abundance in the joint and that TLRs have been found in the synovial membrane (31–34), this pathway may be involved in the innate immune inflammation of the joint. Indeed, there is evidence for a role of TLRs in murine arthritis models (35, 36).
Based on these observations on microbial triggering and innate immune cells in SpA and on the involvement of TLRs in murine arthritis models, the present study was undertaken to investigate the expression and potential role of TLR-2 and TLR-4 in systemic and synovial inflammation in SpA in humans. Moreover, the impressive down-modulation of inflammation and the occurrence of serious infectious side effects related to TLR-dependent pathogens during tumor necrosis factor α (TNFα) blockade treatment (37–41) led us to investigate the effect of TNFα blockers on systemic and local expression of TLRs.
PATIENTS AND METHODS
Patients and samples.
The study protocol was approved by the Ethics Committee of Ghent University Hospital. Eighty-two subjects were included. All subjects provided written informed consent before enrollment. All SpA and RA patients had active disease that fulfilled the European Spondylarthropathy Study Group classification criteria for SpA (42) or the American College of Rheumatology (formerly, the American Rheumatism Association) criteria for RA (43). All osteoarthritis (OA) patients had active synovitis affecting at least 1 knee joint. Baseline demographic and clinical characteristics of the patient groups are shown in Table 1, and response to treatment among the patients treated with TNFα blockade is shown in Table 2.
Table 1. Demographic and clinical data on the SpA, RA, and OA patients included in the analysis of PBMCs or synovial tissue*
SpA (n = 8)
RA (n = 9)
SpA (n = 23)
RA (n = 15)
OA (n = 18)
Values for sex and treatment are the number of patients; other values are the median (range). SpA = spondylarthropathy; RA = rheumatoid arthritis; OA = osteoarthritis; PBMCs = peripheral blood mononuclear cells; CRP = C-reactive protein; ESR = erythrocyte sedimentation rate; NSAIDs = nonsteroidal antiinflammatory drugs; DMARDs = disease-modifying antirheumatic drugs; MTX = methotrexate; LEF = leflunomide; SSZ = sulfasalazine.
Disease duration, years
Swollen joint count
Serum CRP, gm/liter
7 MTX, 1 LEF,
6 MTX, 2 SSZ,
1 MTX + LEF
Table 2. Clinical response in the SpA and RA patients included in the analysis of PBMCs or synovial tissue*
Peripheral blood samples from 8 SpA patients (all with ankylosing spondylitis [AS]), 9 RA patients, and 9 healthy controls (median age 28 years [range 23–35]) were obtained at baseline. The 8 SpA and 9 RA patients were treated with infliximab (5 mg/kg and 3 mg/kg, respectively), at weeks 0, 2, and 6. Additional blood samples were obtained from the SpA and RA patients at weeks 2 and 6 (prior to infusion). Peripheral blood mononuclear cells (PBMCs) were isolated by Ficoll-Paque gradient centrifugation (Pharmacia, Uppsala, Sweden) and used for phenotypic and functional analysis.
Synovial tissue samples (16 biopsy specimens from each patient) were obtained from clinically involved knee joints of 23 SpA patients (8 with AS, 7 with psoriatic arthritis, and 8 with undifferentiated SpA), 15 RA patients, and 18 OA patients, by needle arthroscopy as described previously (44). Eight of the SpA patients were treated with infliximab (5 mg/kg at weeks 0, 2, and 6) and 15 were treated with etanercept (25 mg twice weekly); in these SpA patients, paired biopsy samples were also obtained at week 12. Five-micrometer sections of these samples were used for histologic and immunohistochemistry analysis.
PBMCs were washed in phosphate buffered saline (PBS) and incubated for 30 minutes with the appropriate amount of the following fluorochrome-labeled monoclonal antibodies (mAb) for phenotypic characterization: fluorescein isothiocyanate (FITC)–conjugated anti–TLR-2 (Immunosource, Halle-Zoersel, Belgium), FITC-conjugated anti–TLR-4 (Serotec, Oxford, UK), phycoerythrin (PE)–conjugated anti-CD163 (BD Biosciences PharMingen, San Diego, CA), PE/Cy5-conjugated anti-CD33 (BD Biosciences PharMingen), and allophycocyanin (APC)–conjugated anti–HLA–DR (BD Biosciences PharMingen). The labeled PBMCs were washed, fixed in PBS/1% formaldehyde, and analyzed by 4-color flow cytometry (FACSCalibur; Becton Dickinson, San Diego, CA) using CellQuest software (Becton Dickinson). Monocytes were identified using a forward and side scatter gate in combination with a gate on CD33high cells (previous experiments showed that all of these cells were CD14+). Within the global monocytic population identified by CD33, specific monocyte subpopulations were analyzed based on the expression of CD163 (13, 14). Nonspecific staining and autofluorescence were determined using isotype-matched control mAb. Samples from different disease groups and samples obtained pre- and posttreatment were analyzed in a single run under blinded conditions, to avoid interassay variability. Results are expressed as mean fluorescence intensity (MFI). Differences in MFI for TLR-2 and TLR-4 were consistent in multiple experiments, and expression levels were comparable with reported data (33, 34).
For analysis of TNFα production, PBMCs were resuspended in RPMI 1640 medium (Invitrogen, Merelbeke, Belgium) and stimulated for 6 hours with either 10 ng/ml LPS (Escherichia coli O26:B6; Sigma, St. Louis, MO) or 25 ng/ml phorbol 12-myristate 13-acetate (PMA), both in the presence of 10 μg/ml brefeldin A (Sigma) for the last 5 hours in order to inhibit the secretion of produced cytokines. Unstimulated cells incubated for 6 hours in RPMI 1640 were used as controls. The cells were subsequently incubated for 30 minutes with PE/Cy5-conjugated anti-CD33 (BD Biosciences PharMingen). Next, 2 ml of lysing buffer (BD Biosciences PharMingen) was added for 10 minutes, cells were centrifuged, and 500 μl of permeabilization buffer (BD Biosciences PharMingen) was added for another 10 minutes. The cells were then incubated for 30 minutes with APC-conjugated anti-TNFα (BD Biosciences PharMingen). Isotype- and concentration-matched control mAb were again used to assess nonspecific binding. The labeled cells were analyzed by flow cytometry as described above.
Synovial biopsy specimens were fixed, stained, and scored as described in detail previously (13–17, 44–47). Briefly, 8 paraffin-embedded specimens from each patient were stained with hematoxylin and eosin for histologic evaluation of inflammatory infiltration. The remaining 8 samples were embedded in tissue freezing medium. Frozen sections were fixed in acetone and incubated with anti-CD68 and anti-CD163 mouse mAb (both from Dako, Glostrup, Denmark) for detection of macrophages (13, 14) and with mAb against TLR-2 and TLR-4 (both from Abcam, Cambridge, UK). After rinsing of the specimens, endogeneous peroxidase was blocked with 1% hydrogen peroxide. The sections were subsequently incubated for 15 minutes with a biotinylated anti-mouse secondary antibody and then for 15 minutes with streptavidin–peroxidase complex (LSAB+ Kit; Dako). The color reaction was developed using 3-amino-9-ethylcarbazole substrate (Dako) as chromogen. Finally, the sections were counterstained with hematoxylin. The stained sections were masked with regard to diagnosis and time of sampling and scored by 2 independent observers (DB and LdR). The synovial lining layer and sublining layer were scored separately for each parameter, using a 4-point semiquantitative scale (0 = the lowest and 3 = the highest level of expression) that had been extensively validated previously (13–17, 44–47). Interobserver correlation was high for all parameters (r > 0.90, P < 0.01), with interobserver agreement reached in >85% of cases. In instances of discordant scores between the 2 observers (which never differed by >1 point), the mean of the 2 scores was used. Semiquantitative scoring was previously shown to generate results similar to those obtained with manual counting and digital image analysis (48, 49). Samples from different disease groups and pre- and posttreatment samples in the SpA group were stained in a single run to exclude interassay variability.
The flow cytometric data were normally distributed and are expressed as the mean ± SD. The significance of the differences between groups was determined using Student's t-test. Correlation coefficients were calculated with Pearson's correlation test. The nonparametric immunohistochemical data are expressed as the median (range), and the significance of the differences between groups was determined using the Mann-Whitney U test for unpaired data and the Wilcoxon signed rank test for paired data. P values less than 0.05 were considered significant.
Increased expression of TLR-4, but not TLR-2, on monocytes from patients with SpA.
The expression of TLR-2 and TLR-4 on PBMCs from SpA patients, as identified by their forward and side scatter and their bright expression of CD33, was measured by flow cytometry. As shown in Figure 1, the MFI of TLR-4 (mean ± SD 106 ± 33 versus 73 ± 7; P = 0.005), but not of TLR-2 (82 ± 24 versus 87 ± 9), was increased in SpA versus healthy controls. In order to assess whether these alterations were specific to SpA or were more generally related to chronic arthritis, patients with RA were analyzed as well. The RA cohort exhibited increased expression of both TLR-4 (MFI 111 ± 45) and TLR-2 (119 ± 44) (P = 0.011 and P = 0.025, respectively, compared with healthy controls), and there was a trend toward higher expression of TLR-2 in RA cells versus SpA cells (P = 0.055). Since this difference between the findings in SpA patients and those in RA patients suggested that the alterations of TLR-4 expression in SpA were not merely a reflection of global monocyte activation, we additionally investigated the expression of HLA–DR as an activation marker on PBMCs and found no significant difference between levels in SpA patients (mean ± SD MFI 303 ± 172) and healthy controls (278 ± 167). Accordingly, TLR-4 expression correlated with neither HLA–DR nor TLR-2 expression in SpA, whereas there was a strong correlation between these latter 2 markers (r = 0.85, P = 0.007). Taken together, these data indicate a specific increase of TLR-4 expression on PBMCs from patients with SpA.
Increased expression of TLR-2 and TLR-4 on the CD163+ monocyte subset.
Since findings in our previous studies suggested a role for CD163+ macrophages in SpA inflammation (13–17), we investigated in more detail the expression of TLRs on CD163+ versus CD163− cells within the overall peripheral blood CD33bright monocyte population. Compared with CD163− monocytes, CD163+ monocytes in healthy controls showed clearly increased expression of TLR-2 (mean ± SD MFI 118 ± 38 versus 85 ± 16; P = 0.011) and TLR-4 (95 ± 55 versus 72 ± 19; P = 0.033). Similar findings were obtained in monocytes from patients with SpA (TLR-2 102 ± 19 versus 79 ± 19; P = 0.012 and TLR-4 114 ± 17 versus 102 ± 22; P = 0.046) and patients with RA (TLR-2 147 ± 28 versus 107 ± 22; P < 0.001 and TLR-4 123 ± 20 versus 103 ± 20; P = 0.016) (Figure 1). In contrast to findings in studies of target tissues such as synovium and gut (13, 14), the percentage of CD163+ monocytes in peripheral blood tended to be lower in patients with SpA (mean ± SD 1.2 ± 0.2%) than in healthy controls (3.6 ± 3.6%) or patients with RA (7.6 ± 6.9%). Therefore, the increased expression of TLR-4 in SpA patients compared with healthy controls was not due to an increased percentage of CD163+ monocytes in the peripheral circulation, but rather reflected an increase in TLR-4 levels on both CD163+ and CD163− monocytes in SpA.
Increased expression of TLR-2 and TLR-4 in SpA synovium.
The intrinsic increase of TLR-4 on monocytes from patients with SpA, the increased TLR expression on CD163+ versus CD163− cells in all 3 study cohorts, and the previously reported increase in the CD163+ macrophage subset in SpA synovitis (13, 14) raised the question of the local expression of TLR-2 and TLR-4 in SpA synovium. As seen in Figure 2, immunohistochemical analysis revealed abundant cellular staining for TLRs in SpA synovial tissue. Positively staining cells were observed in both the synovial lining and sublining layers, where they were predominantly located in the perivascular regions or in close proximity to lymphocytic infiltrates. The staining pattern was similar for TLR-2 and TLR-4, with a strong correlation between levels of the 2 TLRs in the lining (r = 0.63, P = 0.001) and sublining (r = 0.60, P < 0.001). Of interest, TLR-2 and TLR-4 expression in the lining also correlated with the number of CD163+ macrophages (r = 0.40, P = 0.07 for TLR-2; r = 0.50, P = 0.02 for TLR-4), but not with CD68+ cells.
Whereas the overall staining pattern was similar in RA and OA synovium, the expression level of both TLRs was significantly higher in SpA than in RA or OA synovium (Table 3). Compared with RA and OA synovium, the expression of TLR-4 in SpA synovium was higher in both the lining layer and the sublining layer (both P ≤ 0.001). In contrast to findings in peripheral blood, the degree of expression of TLR-2 was also increased in SpA synovium compared with RA and OA synovium, in both the lining and sublining layers (both P ≤ 0.001). Whereas overall inflammatory infiltration was lower in OA than in SpA samples (P ≤ 0.001), the absence of a significant difference between SpA and RA indicates that the increased TLR expression in SpA synovitis was not merely inflammation-related, but was also disease-specific. Moreover, comparing SpA with RA, there was no difference in the number of CD68+ macrophages (median [range] score 2.5 [0.5–3] versus 1.5 [1–3] in the lining and 2 [0–3] versus 2 [1–3] in the sublining) but an increase in the number of CD163+ macrophages (2.5 [1–3] versus 0 [0–2.5]; P < 0.001 in the lining and 2 [0–3] versus 0 [0–3]; P = 0.007 in the sublining), thereby providing further evidence that TLR expression was related to the CD163+ macrophage subset rather than to the degree of synovial inflammation per se.
Table 3. Histologic and immunohistochemical analysis of overall inflammatory infiltration and TLR-2 and TLR-4 expression in the lining and sublining layers of synovial tissue from patients with SpA, RA, and OA*
SpA (n = 23)
RA (n = 15)
OA (n = 18)
Values are the median (range) semiquantitative scores on a 4-point scale. TLR-2 = Toll-like receptor 2 (see Table 1 for other definitions).
Down-regulation of TLR-2 and TLR-4 expression on SpA monocytes by infliximab treatment.
Given the major effect of TNFα blockade on inflammation in SpA, we next investigated whether treatment with infliximab in vivo would modulate the expression of TLRs on PBMCs in SpA. As shown in Figure 3, the expression of TLR-4 decreased gradually over a 6-week period of infliximab treatment (mean ± SD MFI 106 ± 33 at baseline, 93 ± 14 at week 2, and 78 ± 12 at week 6; P = 0.011 at week 6). Although not increased at baseline, the expression of TLR-2 also gradually declined during TNFα blockade treatment (82 ± 24 at baseline, 73 ± 26 at week 2, and 53 ± 16 at week 6; P = 0.003 at week 6), reaching an expression level 40% lower than that in healthy controls (P < 0.001). In RA patients, analyzed as a control group in these experiments, infliximab treatment had a similar effect on monocyte expression of TLR-4 (MFI 111 ± 45 at baseline, 64 ± 23 at week 2, and 69 ± 25 at week 6; P = 0.019 at week 6) and TLR-2 (119 ± 44 at baseline, 85 ± 36 at week 2, and 82 ± 26 at week 6; P = 0.065 at week 6). Underscoring the specificity of the decrease of TLR-2 and TLR-4 during TNFα blockade in vivo, the expression of HLA–DR on monocytes did not decrease in SpA patients (MFI 303 ± 172 at baseline versus 454 ± 326 at week 6).
Functional impairment of the TLR-4 pathway after infliximab treatment.
Since we had previously demonstrated the ability of CD163+ monocytes to produce high amounts of TNFα upon LPS stimulation (13), which could potentially be related to their high TLR-4 expression, we next investigated whether down-
regulation of TLR expression on monocytes by TNFα blockade in vivo would result in a functional impairment of their capacity to produce proinflammatory cytokines such as TNFα. As shown in Figure 4, activation of SpA PBMCs through the TLR-4 pathway by LPS resulted in significantly reduced TNFα production after infliximab treatment (mean ± SD MFI 179 ± 84) compared with baseline (259 ± 70) (P = 0.002). In contrast, TNFα production upon PMA stimulation was not affected by infliximab treatment in vivo (MFI 272 ± 174 versus 269 ± 67).
Down-regulation of synovial TLR-2 and TLR-4 expression in SpA by infliximab treatment.
Given the increased expression of TLR-2 and TLR-4 in SpA synovium and the down-regulation of the expression and function of these TLRs on peripheral blood monocytes with infliximab treatment, we next assessed whether this effect of infliximab also extended to the inflamed SpA synovium, by analyzing synovial biopsy samples obtained at baseline and week 12. As shown in Figure 5 and Table 4, TLR-2 expression in the synovial sublining layer (P = 0.023) and TLR-4 expression in the lining layer (P = 0.034) and sublining layer (P = 0.018) decreased significantly after infliximab treatment. A similar trend was observed for TLR-2 expression in the synovial lining layer (P = 0.074).
Table 4. Effect of 12-week infliximab or etanercept treatment on TLR-2 and TLR-4 expression in the lining and sublining layers of synovial tissue from patients with SpA*
Infliximab (n = 8)
Etanercept (n = 15)
Values are the median (range) semiquantitative scores on a 4-point scale. As indicated by the medians, the lower P values in the infliximab group than in the etanercept group are due not to a differential effect between the 2 tumor necrosis factor α blockers, but to the lower number of samples in the infliximab group. TLR-2 = Toll-like receptor 2; SpA = spondylarthropathy.
Down-regulation of synovial TLR-2 and TLR-4 expression in SpA by etanercept treatment.
In order to assess whether this down-regulation of synovial TLR-2 and TLR-4 expression in SpA occurs only with infliximab or is a more general class effect of TNFα blockers, we additionally investigated synovial biopsy specimens obtained at baseline and week 12 of etanercept treatment. As was observed with infliximab, there was a pronounced down-regulation of both TLR-2 and TLR-4 expression in the synovial lining layer (P < 0.001 for both TLRs) and in the sublining layer (P < 0.001 and P = 0.005, respectively) (Figure 5 and Table 4). The median decrease in TLR-2 and TLR-4 expression was similar in the infliximab-treated and etanercept-treated groups, although the P values were smaller in the latter group due to the larger number of samples (15 etanercept-treated patients versus 8 infliximab-treataed patients).
In view of the role of bacterial triggering in the pathogenesis of SpA and our recent observations indicating a relative increase of monocyte/macrophages and polymorphonuclear cells in SpA synovitis (13–17), the present study was undertaken to explore the hypothesis that alterations in the TLR pathway could be involved in an abnormal activation of innate immunity–mediated inflammation in SpA. Since TLR-2 (which associates with TLR-1 or TLR-6) and TLR-4 recognize products of gram-positive and gram-negative bacteria, respectively, and are expressed on the cell surface of monocyte/ macrophages and neutrophils, the present study focused on the expression of these 2 members of the TLR family.
A new finding of our study was that PBMCs from SpA patients have a nearly 50% increase in TLR-4 expression compared with those of healthy controls. Of interest, neither TLR-2 nor HLA–DR expression was increased on these cells, indicating that this phenomenon is not just a reflection of systemic inflammation or nonspecific activation of phagocytes. This was further emphasized by the absence of a correlation of TLR-4 levels with expression of the 2 other markers on phagocytes, or with parameters of systemic inflammation such as C-reactive protein level and erythrocyte sedimentation rate (data not shown). Moreover, RA patients with comparable systemic inflammation exhibited a different profile, with increases of both TLR-4 and TLR-2. The TLR-4 and TLR-2 expression levels were roughly comparable with those previously reported for RA PBMCs (33, 34), and similar findings were obtained when the flow cytometric data were analyzed as the percentage positive cells rather than by MFI (data not shown).
Although it is unclear which stimuli and mechanisms are responsible for the up-regulation of TLRs in vivo (50–52), differential regulation of TLR-2 and TLR-4 was recently demonstrated in septic shock as well as in RA (33, 53, 54). In a recent study, it appeared that TLR-4 was increased on all monocytes whereas TLR-2 was specifically up-regulated on the CD16+ subset (33). These CD16+ monocytes produce high amounts of proinflammatory cytokines such as TNF and have been implicated in the pathogenesis of RA (55–57).
Transposing this observation to SpA, we investigated whether the increase in TLR-4 expression could be observed on all phagocytes or was due to a specific subset. We focused particularly on phagocytes expressing the scavenger receptor CD163, which are increased in SpA inflammation and are also able to produce high amounts of TNFα (13, 14). Consistent with the notions of a proinflammatory role of these cells and a link between TLR and scavenger receptor expression (58), the present study showed that CD163+ phagocytes expressed higher levels of TLR-2, TLR-4, and HLA–DR than did their CD163− counterparts. However, this difference was observed not only in patients with SpA, but also in RA patients and healthy controls, and the fraction of CD163+ monocytes in the peripheral circulation tended to be lower in SpA. Therefore, the increased systemic TLR-4 expression in SpA could not be explained solely by the CD163+ subset, as was confirmed by the fact that, as in RA (33), TLR-4 expression was increased on all monocyte subsets in SpA patients compared with healthy controls (data not shown).
In contrast to the situation in peripheral blood, the demonstration of higher TLR-2 and TLR-4 expression on CD163+ phagocytes may be of importance with regard to tissue inflammation in SpA, since we previously demonstrated a specific increase of this macrophage subset in the synovium, as well as the gut, in patients with SpA (13–17). In RA, recent studies demonstrated the expression of TLR-2 and TLR-4 on macrophages as well as fibroblasts in the synovial lining and sublining layers (31–34). Results of the present study extend these findings by indicating not only that TLR-2 and TLR-4 are expressed with a similar pattern in SpA synovitis, but also that the expression of both TLRs is significantly higher in SpA compared with RA synovium despite a similar degree of local inflammation. Given the similar TLR-4 expression and the higher TLR-2 expression on RA versus SpA PBMCs, it is likely that the specific increase of CD163+ macrophages, rather than systemic alteration, contributes to the increased synovial TLR expression in SpA. Alternatively, it should be considered that TLR expression might also be abnormal on other cell types, such as synovial fibroblasts.
Independent of these considerations, the elevated expression in SpA and the fact that both synovial phagocytes and fibroblasts produce inflammation mediators such as cytokines and chemokines upon TLR-2 or TLR-4 stimulation (13, 31, 33, 34, 59) indicate a need for further investigation of a potential role of the TLR pathway in inducing and/or perpetuating synovial tissue inflammation in SpA. Whereas it is difficult to provide direct functional evidence in human SpA and thereby to demonstrate the biologic significance of our findings, animal studies have indicated the importance of TLR-4 in the early inflammatory cytokine response of phagocytes upon infection with SpA-associated gram-negative bacteria such as Salmonella, Yersinia, and Chlamydia (22–25, 60). Accordingly, the role of bacterial products such as streptococcal cell wall and LPS in the induction and/or perpetuation of experimental arthritis is critically dependent on TLR-2 and TLR-4, respectively, and the associated adaptor molecule, myeloid differentiation factor 88 (35, 36). Moreover, self molecules such as proteoglycans, which are abundant in the joint and are involved in SpA synovitis, also signal through the TLR pathway and can induce inflammatory responses in vivo (26–30, 61).
In an attempt to address the functional importance of the increased TLR expression in human SpA, we investigated whether anti-TNFα therapy, which is highly effective in SpA, was associated with modulation of the expression and/or function of TLRs. Analysis of PBMCs indicated that in both RA and SpA, TLR-4 expression was significantly down-modulated by infliximab treatment over a period of 6 weeks. Consistent with the decrease in inflammatory cytokine production after down-regulation of surface TLR-4 expression in endotoxin tolerance (62), the infliximab-induced down-regulation of TLR-4 was associated with impairment of TNFα production by monocytes upon LPS stimulation. The fact that TNFα production upon PMA stimulation was unaltered is compatible with the notion of specific interference with the TLR-4 pathway, rather than a generalized hyporesponsiveness of the phagocytes. Differential regulation of the 2 pathways has previously been demonstrated with other drugs, such as indomethacin and rosiglitazone (63, 64). Although TLR-2 expression on SpA phagocytes was not increased at baseline, it was also significantly down-regulated by infliximab treatment, resulting in lower levels than in healthy controls.
Parallelling the findings in peripheral blood, infliximab also decreased the expression of TLR-2 and TLR-4 in inflamed synovial membrane, which is probably the combined result of the systemic effect of infliximab on TLR expression and the treatment-induced reduction of inflammatory phagocyte infiltration (46, 47). Moreover, the similar down-regulation of synovial TLR-2 and TLR-4 expression with etanercept treatment indicates that this is a class effect of TNFα blockers rather than an infliximab-specific mechanism.
Taken together, our results suggest that TNFα blockade in vivo interferes with the innate immune system in SpA, although this remains to be proven by functional studies. Such interference could then lead to a reduction in the systemic and local inflammation. However, the down-regulation of TLRs, and especially TLR-2, which declines to levels below those found in healthy controls, might also interfere with normal host defense. In RA, the increased susceptibility to infections with essentially intracellular pathogens during TNFα blockade treatment has previously been related to decreased TLR-4 expression and interferon-γ production by myeloid cells (40, 41). The present in vivo data and our recent finding of tuberculosis reactivation and abscesses with streptococci in infliximab-treated SpA patients (37–39) are compatible with the hypothesis that TNFα blockade in vivo interferes with the normal innate immune response to pathogens and suggest that this might even be more pronounced in SpA than in RA.
In conclusion, the present findings indicate that inflammation in SpA is characterized by strongly increased levels of TLR-2 and TLR-4 expression, which are sharply reduced by TNFα blockade. These data suggest a potential role of innate immunity–mediated inflammation in SpA and could provide additional clues regarding the efficacy as well as the potential side effects of TNFα blockade in this disease. The demonstration of altered TLR-2 and TLR-4 expression on phagocytes and in inflamed synovium in SpA suggests the need for further evaluation of TLR expression and function on other synovial cell types such as fibroblasts (31, 32, 59) and polymorphonuclear cells (65, 66), assessment of other target tissues of SpA inflammation such as the gut mucosa (67–69), analysis of the relationship to genetic risk factors such as HLA–B27 and the innate immune system–related NOD2 polymorphisms (70, 71), and investigation of other TLRs such as TLR-9 (72).
The authors wish to thank Ms Virgie Baert and Mrs. Annemie Herssens for excellent technical assistance.