Characteristics of the study patients
Demographic and clinical characteristics of all RA patients are shown in Table . Age, gender or disease duration did not differ among the patients with RA, OA and Orth.A. Ten of the RA patients were male and 34 were female. The mean age was 51.5 (range 23 to 75) years. Mean disease duration was 80.0 (range, 1 to 480) months. 80% (35/44) of the RA patients had positive RF, and positive ACPA was detected in 84% (37/44) of the RA patients. All RA patients had DAS28-CRP values > 3.2, suggesting active disease. Mean DAS28-CRP score was 5.85 (range 3.22 to 7.92). 75% (33/44) of the patients were in the high activity group, 25% (11/44) were in the moderate. A total of 25% (11/44) of the RA patients had never been treated with corticosteroids or DMARDs, except for Chinese herbs and/or pain-killers to relief arthralgia. A total of 27% (12/44) had taken corticosteroids alone before being referred to our hospital. A total of 41% (18/44) had taken one or more DMARDs, including methotrexate, leflunomide, sulfasalazine or hydroxychloroquine, while 32% (14/44) patients were actively on DMARDs and corticosteroids. Only 7% (3/44) had been treated with TNF-α blocker (etanercept). There was no washout period before synovial biopsy.
Synovial expression of TRAF6 in RA, OA and Orth.A
To explore whether synovial expression of TRAF6 was aberrant in RA patients, we determined synovial expression of TRAF6 in 44 RA patients with a blinded manner compared with 9 patients with OA and 7 with Orth.A as "less inflamed" disease control. TRAF6 expression was found in the intimal and subintimal layer of RA synovium, with intense staining found in the endochylema as well as nucleus of intimal synoviocytes and subintimal inflammatory cells (Figure ). Double immunofluorescence staining showed TRAF6 was expressed in most of the intimal cells and obviously expressed in CD68+ cells and some other CD68- cells in the subintimal area (Figure ). Significantly enhanced synovial expression of TRAF6 was detected from RA patients compared to those from OA or Orth.A patients (2.53 ± 0.94 vs 0.72 ± 0.44 and 0.71 ± 0.49, P < 0.0001) (Figure ). There was no significant difference in synovial TRAF6 expression between OA and Orth.A patients.
Figure 1 Representative immunohistochemical findings of synovial TRAF6 expression. A, Mild synovial TRAF6 expression in an OA patient; B, Intensive synovial TRAF6 expression in a RA patient; C, Expression of TRAF6 in intimal layer of the synovium in a RA patient. (more ...)
Figure 2 Representative confocal microscopic images showed TRAF6 and CD68 by indirect double immunofluorescence staining in RA synovium. A, original magnification, ×400; B, original magnification ×1,200. (a, TRAF6 (red); b, CD68 (green); c, DAPI (more ...)
Figure 3 Synovial expression of TRAF6 in RA, OA and Orth.A. Scatterplots showing synovial TRAF6 expression scores in RA, OA and Orth.A patients as measured by immunohistochemistry. Synovial expression of TRAF6 was scored semiquantitatively on a five-point scale (more ...)
In the RA group, 34.1% (15/44) patients showed low synovial TRAF6 expression, 65.9% (29/44) showed high synovial TRAF6 expression, and the mean (range) synovial TRAF6 expression score was 1.5 (0.0 to 2.0) and 3.1 (2.0 to 4.0), respectively.
Correlation of synovial TRAF6 expression with the synovitis score
Synovitis score was assessed in all 44 RA patients and the mean (range) score was 3.7 (1.0 to 6.5). 63.6% (28/44) patients had low-grade synovitis, 34.1% (15/44) had high-grade synovitis and 2.3% (1/44) patient had no synovitis. The mean synovial TRAF6 expression score in all RA patients was 2.5 (range 0.5 to 4.0). There was no significant difference in synovial TRAF6 expression between low-grade and high-grade synovitis group (2.4 ± 1.0 vs 2.7 ± 0.9, P = 0.320). Moreover, synovitis score was not significantly different between low and high synovial TRAF6 expression group (3.3 ± 1.4 vs 3.9 ± 1.2, P = 0.124). However, a significant correlation was found between synovial TRAF6 expression and synovitis scores from all 44 RA patients (r = 0.412, P = 0.006) (Table and Figure ). When the three components of synovitis score (hyperplasia of intima, cellular density of synovial stroma and inflammatory cell infiltration) were subanalyzed, the inflammatory cell infiltration was found to correlate best with synovial TRAF6 expression (r = 0.367, P = 0.014) (Table and Figure ).
Correlation between synovial TRAF6 expression and parameters of inflammation or joint destruction in RA patients▲
Figure 4 Spearman's rank correlation analysis for synovial TRAF6 expression and synovitis score in RA. A, Correlation between synovial TRAF6 expression and synovitis score; B, Correlation between synovial TRAF6 expression and hyperplasia of intima; C, Correlation (more ...)
To explore the possible different effect of synovial TRAF6 expression, we subanalyzed intimal and subintimal TRAF6 expression in different grades of synovitis groups and found no significant difference in intimal or subintimal TRAF6 expression between low-grade and high-grade synovitis RA patients. Our data suggested only subintimal TRAF6 expression correlated significantly with synovitis scores (r = 0.465, P = 0.001) and the inflammatory cell infiltration (r = 0.450, P = 0.002) (Table ).
Synovial TRAF6 expression correlated with the density of the common mononuclear inflammatory cells
Sequential slides from those used for TRAF6 staining were then stained for the aforementioned CD markers. In RA synovium, intimal CD68+ cells, cell density of subintimal CD68+ cells, CD3+ cells and CD34+ cells were significantly higher in the high TRAF6 expression group than that in the low TRAF6 expression group. Moreover, significant correlation of TRAF6 expression was detected with intimal CD68+ cell density (r = 0.636, P < 0.0001), and subintimal CD68+ cell density (r = 0.532, P = 0.005), CD3+ cell density (r = 0.478, P = 0.004), CD20+ cell density (r = 0.313, P = 0.047), CD38+ cell density (r = 0.390, P = 0.020), CD79a+ cell density (r = 0.398, P = 0.032) as well as MCV (r = 0.596, P = 0.006) (Table ).
When intimal and subintimal TRAF6 expression were subanalyzed, intimal CD68+ cells and subintimal CD3+ cell density were shown to be correlated significantly with intimal TRAF6 expression (r = 0.631 and 0.367, P < 0.0001 and P = 0.033, respectively). However, subintimal TRAF6 expression correlated significantly with subintimal CD3+ cell density (r = 0.526, P = 0.001), CD20+ cell density (r = 0.359, P = 0.034), CD38+ cell density (r = 0.391, P = 0.020), CD79a+ cell density (r = 0.417, P = 0.024), CD68+ cell density (r = 0.562, P = 0.003) and MCV (r = 0.522, P = 0.013), as well as intimal CD68+ cell density (r = 0.604, P < 0.0001) (Table ).
In addition, the density of subintimal common mononuclear inflammatory cell types, including CD68+ cells, CD3+ cells, CD38+ cells, CD20+ cells, and CD79a+ cells, were significantly higher in high-grade synovitis RA patients than that in low-grade synovitis RA patients. Moreover, the density of subintimal CD68+ cells, CD3+ cells, CD38+ cells, CD20+ cells and CD79a+ cells correlated positively with the synovitis score (r = 0.306, 0.539, 0.635, 0.752, and 0.771, respectively; P = 0.028, P = 0.001, P < 0.0001, P < 0.0001, and P < 0.0001, respectively). Therefore, subintimal common mononuclear inflammatory cells may be useful indicators of histologic disease activity in RA.
Correlation of synovial TRAF6 expression with clinical parameters
Spearman's rank order correlation test was performed to investigate the correlation between synovial TRAF6 expression and ESR, CRP, RF, ACPA, TJC in 28 joints, SJC in 28 joints, HAQ, morning stiffness, gripping power or DAS28-CRP, all of which are serological or clinical parameters that reflect disease activity or severity of RA. No significant correlation was found between synovial TRAF6 expression and DAS28-CRP (P = 0.308) in RA. None of the above parameters correlated significantly with synovial TRAF6 expression (Table ). No significant correlation was found between the presence/absence of synovial TRAF6 expression and age or gender or disease duration. Moreover, the above parameters were not different significantly between low TRAF6 expression group and high TRAF6 expression group.
In addition, about 80% of RA patients in our study were positive for RF and 84% were positive for ACPA, and synovial TRAF6 expression were similar in seropositive and seronegative patients (data not shown).
Correlation of synovial TRAF6 expression with joint destruction
Unfortunately, no significant correlation was found between synovial TRAF6 expression and total Sharp score (P = 0.583), the erosion subscore (P = 0.347) or the joint space narrowing subscore (P = 0.493) (Table ). Moreover, total Sharp score, the erosion subscore and the joint space narrowing subscore were not different significantly between low TRAF6 expression group and high TRAF6 expression group. However, disease duration correlated significantly with total Sharp score (r = 0.586, P < 0.0001), the erosion subscore (r = 0.533, P < 0.0001), and the joint space narrowing subscore (r = 0.562, P < 0.0001). Additionally, DAS28-CRP correlated significantly with the erosion subscore (r = 0.325, P = 0.032).
Subanalysis of treated and untreated RA patients
To exclude the influence of therapy on patient recruiting, we further subanalyzed RA patients between treated and untreated group. There was no significant difference in disease duration, age, gender or the age of disease onset between treated and untreated RA patients. There was no significant difference in clinical, serological or radiological parameters related to disease activity or severity, either. Between treated and untreated RA patients, there was no significant difference in total synovial TRAF6 expression, intimal or subintimal TRAF6 expression. But higher TRAF6 expression was found both in treated and untreated RA patients than in OA or Orth.A patients.
Four RA patients treated with methotrexate and etanercept were followed for 12 months and underwent repetitive synovium biopsy. These patients were between 30 and 51 years of age, 1 male and 3 female. Disease activity and pain decreased after 12 months compared to baseline, and synovial TRAF6 expression decreased obviously, paralleled with the decrease of synovitis score, DAS28-CRP, and Sharp score (Table ). Wilcoxon Signed Ranks Test showed the difference of DAS28-CRP and synovial TRAF6 expression between baseline and 12 months later was at borderline significance (P = 0.068 and 0.059, respectively). No significant difference was found in synovitis score and Sharp score between baseline and 12 months later (P = 0.257 and 0.141, respectively). As there were only four patients biopsied repetitively after follow up, we presumed there was actually a lack of power to detect this difference in a statistically significant manner.
Comparative analysis of synovial TRAF6 expression and disease activity in four RA patients during follow-up