Increased perivascular synovial membrane expression of myeloid-related proteins in psoriatic arthritis

Authors


Abstract

Objective

To analyze S-100 protein expression, in the form of myeloid-related protein 8 (MRP8), MRP14, and the heterodimer MRP8/MRP14, in psoriatic arthritis (PsA) patients compared with rheumatoid arthritis (RA) and spondylarthropathy (SpA) patients, and to determine the effect of methotrexate (MTX) on the MRP antigen expression in PsA patients.

Methods

Serum, synovial fluid (SF), and synovium (taken at arthroscopy) samples were obtained from PsA (before and after MTX treatment), RA, and SpA patients. Concentrations of MRP8/MRP14 in serum and SF were measured by enzyme-linked immunosorbent assay. Expression of MRP8, MRP14, and MRP8/MRP14 in synovium was determined by immunohistochemistry.

Results

MRP8, MRP14, and MRP8/MRP14 levels were increased in serum, SF, and synovium from PsA, RA, and SpA patients. In all 3 groups, paired samples of serum and SF showed significantly higher MRP8/MRP14 levels in SF (mean ± SD 15,310 ± 16,999 ng/ml [median 11,400]) than in serum (908 ± 679 ng/ml [median 695]) (P = 0.0001). MRP8/MRP14 levels in serum correlated with systemic parameters of disease activity (erythrocyte sedimentation rate [ESR] r = 0.55, P = 0.005; C-reactive protein [CRP] level r = 0.55, P = 0.005), whereas levels in SF correlated with local parameters of disease activity (white blood cell count r = 0.45, P = 0.01; acute-phase serum amyloid A level r = 0.32, P = 0.03). MRP expression was significantly higher in the synovial sublining layer (SLL) of PsA patients compared with RA and SpA patients. MRP antigens were predominantly expressed in perivascular areas of the SLL in PsA patients. Following MTX treatment, MRP expression in serum and synovium from PsA patients was significantly reduced. Serum levels of MRP were more sensitive to the effects of MTX than were the ESR, CRP, or clinical joint scores.

Conclusion

MRP levels in serum and SF correlate with local and systemic inflammation and are equally increased in PsA, RA, and SpA patients. In contrast, MRP8, MRP14, and MRP8/MRP14 expression in the SLL of PsA patients is increased, particularly in perivascular regions, compared with that in RA and SpA patients, suggesting a central role of MRP proteins in transendothelial migration of leukocytes in PsA. Moreover, MRP expression is reduced following MTX treatment. MRP proteins may represent a novel therapeutic target in inflammatory arthritis.

Myelomonocytic cells play a key role in the production and perpetuation of synovial inflammation in seropositive and seronegative inflammatory arthritis. The synovitis of rheumatoid arthritis (RA) is characterized by infiltration of macrophages and, to a lesser extent, neutrophils, which contribute directly to joint inflammation and destruction through the production of proinflammatory cytokines and proteolytic enzymes, such as metalloproteinases (1). Macrophage infiltration and metalloproteinase production in the synovium correlates with the development of joint erosions (2, 3), and specific therapies targeted at macrophage products, such as tumor necrosis factor (TNFα), retard joint inflammation and destruction (4). Seronegative spondylarthropathy (SpA), including psoriatic arthritis (PsA), is also characterized by macrophage and neutrophil infiltration of the synovium (5, 6) and production of proinflammatory cytokines (7) and proteolytic enzymes (8), with reversal of inflammation by treatment with anti-TNFα (9, 10). However, PsA is characterized by less joint destruction than RA (11), and PsA synovium has been reported to have fewer infiltrating macrophages (5) and reduced TNFα production (7). This may be due to a reduction in myelomonocytic cell trafficking and activation in the synovium.

Myeloid-related protein 8 (MRP8) and MRP14 are 2 calcium-binding proteins that are members of the S-100 family of proteins (S-100A8 and S-100A9, respectively) (12). The S-100 proteins play a role in both intracellular functions, such as cell differentiation and cell cycle progression, regulation of kinase activities and cytoskeleton–membrane interactions, and extracellular functions, such as inducing neutrophil extension, chemoattraction, and the induction of adhesion molecule expression (13). MRP8 and MRP14 are expressed in high concentrations in infiltrating granulocytes and monocytes and during the stages of early differentiation of monocytes, but are absent in lymphocytes and mature tissue macrophages (12, 14).

MRP8 and MRP14 form a noncovalently associated heterodimer, MRP8/MRP14, in a calcium-dependent manner. The heterodimer then translocates from the cytosol to membrane structures (15), which is associated with inflammatory activation of monocytes, as characterized by increased oxidative burst and TNFα secretion (16). Monocytes expressing MRP8/MRP14 represent a fast-migrating subpopulation that uses an intercellular adhesion molecule–dependent system to infiltrate tissues at sites of inflammation (17). MRP8 and MRP14 are also secreted by activated and transmigrating monocytes (18, 19), and secreted MRP8 and MRP14 may be involved in inducing cell adhesion during diapedesis (12, 19). Thus, MRP8 and MRP14 play a key role in determining the extent of neutrophil and macrophage infiltration and activation at sites of inflammation. Furthermore, very high levels of MRP8/MRP14 may cause dysregulation of zinc metabolism, which directly results in a rare disorder characterized by inflammation and recurrent infections (20).

The expression of MRP8 and MRP14 in normal human tissues is minimal. Increased expression has been described in myelomonocytic cells in a number of inflammatory diseases including RA, (14, 21) reactive arthritis (ReA) (22), juvenile arthritis (19), psoriasis (23), systemic lupus erythematosus (SLE) (24), inflammatory bowel disease (25), and renal allograft rejection (26). In RA, expression of MRP8, MRP14, and MRP8/MRP14 in synovial tissue was observed in patients with active disease but not in patients in clinical remission. Synovial lining layer (LL) expression of MRP8, MRP14, and MRP8/MRP14 was noted predominantly in tissues adjacent to the cartilage–pannus junction (CPJ), whereas synovial sublining layer (SLL) expression was greater in tissues distal to the CPJ (21). Levels of MRP antigens have been shown to be increased in RA synovial fluid (SF) and to be higher in SF from RA patients than in serum from patients with RA and other inflammatory arthritides (27). Serum MRP8/MRP14 levels have a stronger correlation than C-reactive protein (CRP) levels with measures of disease activity in RA (27), juvenile rheumatoid arthritis (JRA) (19), ReA (22), and SLE (24). MRP8/MRP14 expression has been shown to be increased in psoriatic plaques but not in normal skin (23). There are no data on MRP expression in serum, SF, or synovium in PsA or on the effect of treatment with disease-modifying antirheumatic drugs (DMARDs) on MRP expression in serum and synovium in inflammatory arthritis.

In this study, we hypothesized that there are quantitative differences in MRP expression in PsA as compared with RA and that these differences may explain the lesser degree of macrophage infiltration and lining layer hyperplasia reported in PsA. Such differences in MRP expression may provide an explanation for the different pattern and lesser degree of joint destruction in PsA as compared with RA (11). The relationship between MRP expression and intraarticular and systemic inflammation was also examined, and the utility of MRP expression in PsA as a potential marker of disease activity was analyzed. Since MRP antigens regulate myelomonocytic infiltration in the synovium, they may represent a potential therapeutic target; thus, the effect of the therapeutic agent methotrexate (MTX) on MRP expression in PsA was assessed from in vivo studies.

PATIENTS AND METHODS

Patients.

All SF and synovium samples were obtained from patients with inflammatory arthritis and active knee synovitis who were undergoing knee arthroscopy. PsA was defined according to established criteria (28), RA was diagnosed according to the American College of Rheumatology (formerly, the American Rheumatism Association) criteria (29), and SpA was defined according to European Spondylarthropathy Study Group criteria (30). A second arthroscopy with synovial biopsy was performed on patients with PsA who were treated with MTX as part of a separate study. Serum samples were also obtained from a subgroup of the patients who were undergoing arthroscopy and from a further group of patients with PsA who were started on MTX treatment as part of an early arthritis study but who did not undergo synovial biopsy. These studies were approved by the St. Vincent's University Hospital Ethics Committee.

Patients were assessed on the day of arthroscopy or serum collection by the same physician (DK). Assessments included the Ritchie Articular Index, the European League Against Rheumatism swollen joint count (maximum 44 joints), the erythrocyte sedimentation rate (ESR; measured by standard Westergren technique), the CRP level (measured by standard nephelometry), and the 3-variable Disease Activity Score (DAS) (31).

Synovial biopsy.

Arthroscopic synovial biopsy of the knee was performed on patients under local anesthesia using a 2.7-mm Storz arthroscope and 1.5-mm grasping forceps. Biopsy samples were obtained from all compartments of the knee joint, embedded in TissueTek OCT compound (Sakura, Zoeterwoude, Netherlands), snap frozen, and stored in liquid nitrogen until used. Synovial biopsy samples from a minimum of 3 separate intraarticular sites for each patient at each time point (except in 1 patient from whom adequate synovium was obtained from only 2 intraarticular sites after MTX treatment) were analyzed. Cryostat sections (7 μm) were mounted on 3-aminopropyltriethoxysilane–coated glass slides, air dried overnight, wrapped in foil, and stored at –80°C until immunohistochemical analysis was performed.

SF and serum samples were centrifuged within 4 hours of collection at 10,000 revolutions per minute for 10 minutes, and the supernatant was separated and stored at –70°C until analyzed. When required for analysis, SF was thawed, pretreated for 1 hour with hyaluronidase (Sigma, St. Louis, MO), and centrifuged at 13,700 rpm for 10 minutes, and the supernatant was separated.

Enzyme-linked immunosorbent assay (ELISA) for MRP8/MRP14.

Sandwich ELISA was performed as previously described (18, 32). Rabbit antisera against recombinant MRP8 and MRP14 were produced as described previously. The monospecificity of the antibodies was analyzed by immunoreactivity against recombinant MRP8 and MRP14, by Western blot analysis of lysates of monocytes and granulocytes, as well as by immunoreactivity against MRP8- and/or MRP14-transfected fibroblastic cell lines as described previously (15). For calibration, different amounts of MRP8/MRP14 (range 0.25–250 ng/ml) were used; MRP8/MRP14 was isolated from human granulocytes as described previously (33). The assay has a sensitivity of <0.5 ng/ml and a linear range between 1 ng/ml and 30 ng/ml. MRP8 and MRP14 form noncovalently associated complexes in the presence of extracellular calcium concentrations that are detectable by the ELISA system (32). The ELISA was therefore calibrated with the native MRP8/MRP14 complex, and the data are expressed as nanograms per milliliter of MRP8/MRP14 (values for the single monomers are not shown).

Synovial immunohistochemistry.

Sections of synovium were stained with monospecific affinity-purified rabbit antisera to MRP8 and MRP14, as well as with mouse anti-human monoclonal antibody 27E10, which exclusively recognizes the MRP8/MRP14 heterodimer, but not the single monomer. Immunostaining was performed using a standard 3-stage immunoperoxidase method as previously described (19). Negative controls were performed by replacing the primary antibodies with isotype-matched control antibodies.

Microscopic analysis.

Only synovial sections in which the lining layer was identifiable were included in the analysis. All tissue sections were evaluated randomly by one of us (DK), who was blinded to the identities of the sections at the time of scoring. A semiquantitative analysis was performed, and the results were scored on a 0–5 scale (0 = <1% positive cells, 1 = 1–10%, 2 = 11–25%, 3 = 26–50%, 4 = 51–75%, and 5 = 76–100%). This scoring system was adapted from a validated scale to allow quantification of smaller degrees of cellular infiltration (34). Scoring was performed in randomly selected high-power fields (hpf) at 400× magnification; a minimum of 17 hpf from 3 separate sections were scored, and the mean score was calculated. Vascularity was expressed as the number of vessels per high-power field. These techniques have all been previously validated and reported (35, 36).

Measurement of acute-phase serum amyloid A (A-SAA).

A-SAA was measured using an ELISA technique specific for A-SAA and with no cross-reactivity for its constitutive counterpart C-SAA (Biotrin, Dublin, Ireland), as previously described (37). For serum samples giving a reading that did not fall within the range of the standard curve of the assay, the ELISA was repeated at a lower or higher serum dilution. All samples were assayed in duplicate, and assays were repeated if there was a discrepancy in the paired results. The detection limit of the assay is 2.25 μg/liter, which is equivalent to 0.9 mg/liter for samples diluted 1:400.

Statistical analysis.

All values are given as the mean ± SD (median). Data were analyzed by nonparametric analysis using StatView software (SAS Institute, Cary, NC). The Mann-Whitney U test was performed to compare the medians of the groups. Simple regression and Spearman's correlation coefficient were used to test for correlation of the variables.

RESULTS

Clinical details of the patients and findings of SF analysis.

SF was obtained from 48 patients (22 with PsA, 11 with RA, and 15 with SpA). Rates of prednisolone (in 1 PsA, 2 RA, and no SpA patients) and DMARD (in 4 PsA, 1 RA, and 2 SpA patients) administration were low, since most samples were obtained following the patients' initial presentation to a rheumatology clinic. Patients with PsA had significantly fewer involved joints and a lower ESR than did patients with RA and a longer duration of disease than did patients with RA and SpA (P < 0.05) (Table 1). Total white blood cell counts in the SF were higher in PsA and SpA patients, but there were no significant differences between the 3 patient groups. SF from PsA patients had a significantly lower lymphocyte count and higher neutrophil count than did SF from RA patients. Levels of MRP8/MRP14 were increased to a similar extent in SF from PsA, RA, and SpA patients as compared with reported serum levels in healthy controls (19).

Table 1. Clinical features and findings of synovial fluid analysis of MRP8/MRP14 heterodimer expression in patients with PsA, RA, and SpA*
 PsA patients (n = 22)RA patients (n = 11)SpA patients (n = 15)
  • *

    Values are the mean ± SD (median). P values were determined by Mann-Whitney U test. MRP = myeloid-related protein; PsA = psoriatic arthritis; RA = rheumatoid arthritis; SpA = spondylarthropathy; ESR = erythrocyte sedimentation rate; CRP = C-reactive protein; SAA = serum amyloid A.

  • P < 0.05 versus RA patients.

  • P < 0.05 versus SpA patients.

  • §

    P < 0.001 versus RA patients.

  • P < 0.001 versus SpA patients.

Clinical features   
 Disease duration, months63.3 ± 65 (30.5)11.8 ± 15.3 (5)18.8 ± 30.6 (6)
 Ritchie Articular Index5.8 ± 2.4 (5)§14.3 ± 6.5 (13.5)3.3 ± 5.4 (2)
 Swollen joint count5.9 ± 6.9 (3)§15.4 ± 9.7 (13)3.5 ± 7.0 (1)
 ESR, mm/hour26.9 ± 28.8 (14)42.5 ± 25.1 (44)22.2 ± 15.7 (21.5)
 CRP, mg/liter37.0 ± 53.8 (15.1)46.2 ± 52.4 (20)25.2 ± 21.6 (25)
Synovial fluid analysis   
 White blood cell count, mm24,916 ± 3,917 (3,630)1,638 ± 1,164 (1,780)4,871 ± 4,838 (3,620)
 Lymphocytes, %51 ± 31 (45)81 ± 28 (92)53 ± 32 (49)
 MRP8/MRP14, ng/ml12,558 ± 13,540 (7,000)19,925 ± 23,669 (12,700)16,502 ± 17,497 (12,900)
 Acute SAA, μg/liter53.6 ± 70.8 (18.9)66.1 ± 79.8 (18.1)37.1 ± 57.8 (4.6)

Correlation of SF MRP8/MRP14 levels with serum MRP8/MRP14 levels and with local and systemic parameters of disease activity.

Paired serum and SF samples were obtained from 28 patients. Levels of MRP8/MRP14 in SF were significantly higher than those in serum (mean ± SD 15,310 ± 16,999 ng/ml [median 11,400] versus 908 ± 679 ng/ml [median 695]; P = 0.0001). MRP8/MRP14 levels in serum correlated with the ESR (r = 0.55, P = 0.005), CRP level (r = 0.55, P = 0.005), and the Ritchie Articular Index (r = 0.4, P = 0.04), but not with the swollen joint count or with the SF level of MRP8/MRP14. Levels of MRP8/MRP14 in SF were significantly correlated with intraarticular markers of inflammation (SF white blood cell count r = 0.45, P = 0.01; SF A-SAA level r = 0.32, P = 0.03) but not with systemic parameters of inflammatory arthritis.

Clinical details of the patients and findings of synovial membrane analysis.

Synovium was obtained from 30 patients (14 with PsA, 11 with RA, and 5 with SpA [2 with ReA, 3 with ankylosing spondylitis]). Rates of prednisolone (in no PsA, 2 RA, and no SpA patients) and DMARD (in no PsA, 3 RA, and 1 SpA patients) administration (data not shown) and systemic parameters of inflammation were similar in the 3 disease groups (Table 2). MRP8, MRP14, and MRP8/MRP14 expression in the SLL was significantly higher in PsA patients than in RA and SpA patients. In the SLL of PsA patients, the dense staining for all 3 MRP antigens was predominantly associated with blood vessels and the surrounding perivascular inflammatory infiltrate. In the SLL of RA patients, there was less MRP antigen expression in blood vessels and perivascular areas (Figure 1). MRP8, MRP14, and MRP8/MRP14 expression in the LL was similar in all 3 disease groups.

Table 2. Clinical features and immunohistochemical analysis of MRP8, MRP14, and MRP8/MRP14 expression in the synovial membrane of patients with PsA, RA, and SpA*
 PsA patients (n = 14)RA patients (n = 11)SpA patients (n = 5)
  • *

    Values are the mean ± SD (median). Synovial membrane findings were scored semiquantitatively on a scale of 0–5, where 0 = <1% positive cells, 1 = 1–10%, 2 = 11–25%, 3 = 26–50%, 4 = 51–75%, 5 = 76–100%. P values were determined by Mann-Whitney U test. MRP = myeloid-related protein; PsA = psoriatic arthritis; RA = rheumatoid arthritis; SpA = spondylarthropathy; ESR = erythrocyte sedimentation rate; CRP = C-reactive protein.

  • P < 0.05 versus RA patients.

  • P < 0.05 versus SpA patients.

Clinical features   
 Disease duration, months20.1 ± 18.8 (18)47.5 ± 52.3 (30)47.2 ± 42.2 (48)
 Ritchie Articular Index5.6 ± 5.3 (4.5)10.7 ± 7.0 (9.5)2.4 ± 1.3 (2)
 Swollen joint count5.3 ± 4.6 (3)10.5 ± 6.5 (9)2 ± 1 (2)
 ESR, mm/hour30.4 ± 38.3 (15)41.6 ± 26.4 (32)36.8 ± 35.7 (21)
 CRP, mg/liter44.3 ± 18.8 (18.9)54.6 ± 63.2 (22.5)51.7 ± 38.9 (31)
Synovial membrane analysis   
 MRP8   
  Lining layer0.3 ± 0.4 (0)0.3 ± 0.6 (0)0.1 ± 0.2 (0)
  Sublining layer1.9 ± 1.0 (1.9)0.8 ± 1.0 (0.5)0.7 ± 0.8 (0.3)
 MRP14   
  Lining layer0.3 ± 0.4 (0.3)0.2 ± 0.4 (0)0.1 ± 0.2 (0)
  Sublining layer2.3 ± 1.1 (2.3)1.1 ± 0.8 (1.0)0.9 ± 0.5 (0.8)
 MRP8/MRP14   
  Lining layer0.5 ± 0.4 (0.5)0.4 ± 0.5 (0.3)0.4 ± 0.4 (0.5)
  Sublining layer2.4 ± 0.9 (2.5)1.2 ± 0.8 (1.3)1.8 ± 1.4 (1.3)
Figure 1.

Myeloid-related protein 8 (MRP8)/MRP14 expression in synovium from the suprapatellar compartment of the knee of patients with psoriatic arthritis (PsA) and rheumatoid arthritis (RA). A, MRP8/MRP14 expression in synovium from a PsA patient is predominantly seen in perivascular regions of the sublining layer (SLL), with minimal expression in the lining layer (LL). B, CD68 expression in synovium from a PsA patient is predominantly distributed in the LL, with less expression in perivascular regions of the SLL. C, MRP8/MRP14 expression in synovium from the same PsA patient as in A, demonstrating the widespread nature of the vascular pattern. D, MRP8/MRP14 expression in synovium from an RA patient is predominantly distributed in the SLL but is not associated with vascular regions to the same degree as in PsA patients. (Original magnification × 20 in A and B; × 10 in C and D.)

There was no significant correlation between synovial MRP8, MRP14, and MRP8/MRP14 expression and local or systemic parameters of inflammatory arthritis in the PsA, RA, or SpA groups. In all patients analyzed together, LL and SLL expression of MRP8 (r = 0.47, P = 0.01), MRP14 (r = 0.54, P = 0.003), and MRP8/MRP14 (r = 0.54, P = 0.003) were significantly correlated. Synovial expression of MRP8, MRP14, and MRP8/MRP14 did not correlate with MRP8/MRP14 levels in serum (n = 9) and SF (n = 13) samples obtained at the time of arthroscopy.

MRP8, MRP14, and MRP8/MRP14 expression at the CPJ.

A total of 29 CPJ and 96 non-CPJ biopsy samples were compared for expression of MRP8, MRP14, and MRP8/MRP14. LL expression of MRP8 (mean ± SD 0.5 ± 1.0 [median 0] at the CPJ, 0.1 ± 0.4 [median 0] at the non-CPJ; P = 0.05) and MRP14 (0.4 ± 0.8 [median 0] at the CPJ, 0.1 ± 0.4 [median 0] at the non-CPJ; P = 0.08) was greater in tissues from the CPJ, but this did not reach statistical significance. Analysis of PsA and RA subgroups demonstrated a similar pattern of MRP expression in the LL in addition to greater SLL expression of MRP8 at the CPJ in both groups, but this also did not reach statistical significance.

Effect of MTX treatment on serum concentrations of MRP8/MRP14 in PsA patients.

Serum was obtained from 14 patients with PsA who were beginning treatment with MTX. Their mean ± SD disease duration was 14.6 ± 8.6 months (median 13.5 months), with a mean ± SD followup of 6.4 ± 1.3 months (median 6 months). The mean ± SD dosage of MTX was 12.9 ± 4.8 mg/week (median 12.5). Significant reductions were observed in the Ritchie Articular Index (P = 0.03), the swollen joint count (P = 0.03), and the CRP level (P = 0.04) following treatment with MTX. Serum levels of MRP8/MRP14 were reduced from a mean ± SD of 1,407 ± 1,623 ng/ml (median 670) before MTX to 551 ± 443 ng/ml (median 410) after MTX (P = 0.01).

Effect of MTX treatment on MRP expression in the synovial membrane of PsA patients.

Synovium was obtained before and after treatment from 8 patients with PsA who were beginning MTX therapy. Their mean ± SD disease duration was 13.7 ± 8.8 months (median 15 months), with a mean ± SD followup 11 ± 2.4 months (median 11.5 months). The mean ± SD dosage of MTX was 14.1 ± 3.8 mg/week (median 15). These patients were not taking prednisolone or other DMARDs.

Significant reductions in the Ritchie Articular Index, the swollen joint count, and the DAS were noted following treatment with MTX (Table 3). Before treatment, MRP8, MRP14, and MRP8/MRP14 expression was increased in PsA synovium, with predominant expression in perivascular cellular infiltrates in the SLL (Figure 2). These infiltrates consisted of mononuclear and polymorphonuclear cells, usually in association with positive endothelial cell staining. MRP8, MRP14, and MRP8/MRP14 expression in the synovial SLL decreased significantly following treatment with MTX. After treatment, minimal residual MRP expression was observed in cells within the lumen of blood vessels and, to a lesser degree, within the SLL. Perivascular expression of MRPs was markedly reduced following treatment, with only occasional vessels demonstrating MRP staining in the SLL. Expression of MRP antigens in the LL was less intense than that in the sublining, and MRP8/MRP14 expression in the LL was reduced by MTX treatment.

Table 3. Clinical details and immunohistochemical analysis of MRP8, MRP14, and MRP8/MRP14 expression in synovial membrane samples from 8 patients with psoriatic arthritis before and after methotrexate treatment*
 Before methotrexateAfter methotrexateP
  • *

    Values are the mean ± SD (median). Synovial membrane findings were scored semiquantitatively on a scale of 0–5, where 0 = <1% positive cells, 1 = 1–10%, 2 = 11–25%, 3 = 26–50%, 4 = 51–75%, 5 = 76–100%. P values were determined by Mann-Whitney U test. MRP = myeloid-related protein; ESR = erythrocyte sedimentation rate; CRP = C-reactive protein; DAS = Disease Activity Score.

Clinical features   
 Ritchie Articular Index7.3 ± 6.4 (5)0.9 ± 0.8 (1)0.003
 Swollen joint count7.4 ± 5.1 (7.5)1.6 ± 2 (1)0.007
 ESR, mm/hour37 ± 47 (19)8 ± 10 (4)0.05
 CRP, mg/liter49.5 ± 66.7 (22.6)5.4 ± 8.1 (1.5)0.09
 DAS (3-variable)3.0 ± 1.2 (2.9)1.2 ± 0.4 (1.1)0.002
Synovial membrane analysis   
 MRP8, ng/ml   
  Lining layer0.1 ± 0.1 (0)0.04 ± 0.1 (0)0.14
  Sublining layer1.4 ± 0.7 (1.8)0.6 ± 0.5 (0.3)0.02
 MRP14, ng/ml   
  Lining layer0.1 ± 0.2 (0)0.0 ± 0.0 (0)0.06
  Sublining layer2.3 ± 0.9 (2.5)0.7 ± 0.2 (0.5)0.01
 MRP8/MRP14, ng/ml   
  Lining layer0.5 ± 0.4 (0.5)0.1 ± 0.2 (0)0.01
  Sublining layer2.3 ± 0.9 (2.5)0.7 ± 0.8 (0.5)0.01
Figure 2.

Effect of methotrexate (MTX) on myeloid-related protein 8 (MRP8)/MRP14 expression in synovium from patients with psoriatic arthritis (PsA). A, MRP8/MRP14 expression in PsA synovium before MTX treatment. B, Higher-power view of boxed area in A, before MTX treatment. C, MRP8/MRP14 expression in PsA synovium after MTX treatment. D, Higher-power view of boxed area in C, after MTX treatment. (Original magnification × 10 in A and C; × 20 in B and D.)

DISCUSSION

In acute inflammation, MRP8, MRP14, and the noncovalently associated heterodimer MRP8/MRP14 are expressed by infiltrating granulocytes and monocytes (12). Expression of MRP antigens is also associated with increased inflammatory activation of monocytes (16) and has been reported to be present in both infiltrating (19) and resident (21) cells in the synovium. In this study, MRP8, MRP14, and MRP8/MRP14 were present in increased amounts in the synovial membrane of patients with PsA, RA, and SpA. Expression was most marked in infiltrating cells in the SLL, with a lesser degree of expression in cells in the LL, suggesting a role of MRP in facilitating the transmigration of circulating cells into the synovium.

Despite previous findings of greater macrophage infiltration in RA than in PsA synovium (5), we found that the expression of MRP antigens was significantly greater in PsA synovium. This difference may be due to increased neutrophil infiltration of the synovial SLL in PsA. MRP antigen expression in the PsA SLL occurred in a distinct pattern in perivascular cellular infiltrates, and in addition to leukocytes, endothelial cells demonstrated positive staining for MRP8 and MRP14. This may result from secretion of MRP antigens by leukocytes during endothelial transmigration (18, 19) or from transient binding of leukocyte-secreted MRP antigen to the surface of activated endothelium (38). Alternatively, the perivascular MRP expression may be directly related to synovial hypervascularity (5) or to the distinctive synovial macrovascular changes reported in PsA (39). The pattern of MRP expression in PsA synovium contrasted with that in RA synovium (21), and the perivascular pattern of MRP expression in PsA may be further evidence of an activated and phenotypically distinct synovial vasculature.

Lining layer expression of MRP antigens occurred to a lesser degree than SLL expression and was similar in PsA, RA, and SpA patients. It has been proposed that LL cells expressing MRP represent a distinct activated macrophage population that is involved in the development of bony erosions at the CPJ in RA (21). We found that there was a nonsignificant trend toward elevated MRP expression in the LL at the CPJ in both PsA and RA. Since a fully quantitative scoring system was not used, a small but significant difference may not have been detected. However, it is important to note that similar degrees of LL MRP expression were found at non-CPJ sites in another study of RA synovium (40).

Our hypothesis that alterations in MRP expression may explain the differences in the pattern and extent of bony damage in PsA and RA was not supported by a simple quantitative difference in MRP expression in the LL at the CPJ. Evidence that monocyte trafficking in the SLL leads to macrophage accumulation in the LL is well established (1). However, it is important to note that the significantly greater degree of MRP expression in the SLL of PsA tissues did not result in a significantly higher MRP expression in the LL. This suggests that despite higher myelomonocytic infiltration in perivascular areas in the SLL of PsA synovium, there is a lesser degree of subsequent trafficking of these cells to the LL than occurs in RA synovium. The differences in myelomonocytic cell trafficking to the LL in PsA and RA synovium would provide an explanation for the thinner lining layer observed in PsA (5) and may be an important mechanism in producing the subsequent distinct differences in bony damage.

Similar elevated levels of MRP8/MRP14 were found in SF from PsA, RA, and SpA patients, despite differences in disease duration and clinical and laboratory measures of arthritis activity. This suggests that similar levels of myelomonocytic infiltration and activation are occurring in these diseases regardless of whether the disease is early or established (41) or is oligoarticular or polyarticular. SF concentrations were significantly greater than serum concentrations, suggesting that MRP8/MRP14 release into the SF by infiltrating cells in the synovium and SF results in diffusion of MRP8/MRP14 into the serum. MRP8/MRP14 levels in SF correlated with local markers of joint inflammation (the SF white blood cell count and SF levels of A-SAA, an acute-phase protein) but not with systemic parameters of arthritis activity.

Serum MRP8/MRP14 levels did not correlate with SF MRP8/MRP14 levels, but they did correlate with systemic parameters of arthritis activity, including the ESR, CRP level, and the Ritchie Articular Index. This may be due to the lesser contribution of MRP production by a single knee joint to the serum concentration in polyarticular disease, such as in the disease groups studied. Serum and SF levels have been shown to be correlated in patients with oligoarticular disease (19). Studies of patients with RA, ReA, and JRA have also found that the MRP level correlates closely with clinical markers of arthritis activity, and in some cases, the correlation is superior to that with the ESR or CRP level (22, 27, 42, 43). Thus, accumulating evidence suggests that MRP8/MRP14 is released at sites of inflamed synovium and is a superior marker of disease activity in inflammatory arthritis compared with the CRP level or the ESR.

MRP expression is up-regulated in synovium from patients with inflammatory arthritis, but is minimal in normal synovium and synovium from patients with RA whose disease is in remission. We found that MRP antigens were increased in serum, SF, and synovial membrane from patients with PsA. MTX treatment in PsA patients resulted in a significant reduction in clinical and laboratory parameters of inflammation. Serum concentrations of MRP8/MRP14 were also reduced by MTX treatment and were more sensitive than the ESR, CRP level, or clinical joint scores in assessing the response to treatment. Following MTX treatment, a marked decrease in MRP expression in the synovium of PsA patients was observed, principally through a reduction in infiltrating cells expressing MRP in the sublining layer. While these studies were not designed or powered to prove the efficacy of MTX treatment in PsA, the marked changes observed in the synovium have not been reported in placebo-treated patients (34, 44).

MRP expression after MTX treatment in PsA patients was noted in polymorphonuclear cells that remained within the lumen of blood vessels within the SLL, suggesting a failure of transmigration despite adhesion. This may be mediated by a reduction in interleukin-8 production (45), a potent promoter of neutrophil chemotaxis. Alternatively, MRP expression may be modified by the effects of MTX on Th1/Th2 cytokine balance in the synovium (34, 46), either through a preferential reduction of proinflammatory Th1 cytokine expression in the synovium (34) or through increased Th2-mediated suppression of MRP (47).

MRP8 and MRP14 are secreted by monocytes during interaction with inflammatory activated endothelial cells (18, 19), and complexes of both proteins exhibit direct proinflammatory effects (13). MRP8/MRP14 heterodimers mediate the adherence of leukocytes to endothelial cells and promote subsequent transmigration (38). Recent reports present evidence that extracellular MRP14 induces the up-regulation of CD11b/CD18 integrin binding activity on neutrophils and monocytes (17, 48). In this context, our data on the expression of MRP8/MRP14 in inflammatory synovitis point to a positive feedback mechanism by which the contact of phagocytes with activated endothelium leads to the release of MRP8/MRP14, which induces firm adhesion and transmigration of infiltrating cells to the synovial tissue. Analysis of the molecular mechanisms of release and the extracellular functions of MRP8 and MRP14 may thus offer molecular targets for future therapeutic strategies aimed at modulating the important inflammatory response mechanisms of phagocytes.

Taken together, our findings showed that MRP concentrations in serum and synovial fluid correlate with clinical parameters of disease and may be superior to standard laboratory parameters, particularly when assessing treatment response. The significant differences in synovial SLL expression of MRP antigens in PsA suggest that myelomonocytic cell trafficking from the SLL to the LL may be reduced in PsA patients as compared with RA patients. Further studies of myelomonocytic cell trafficking to the synovial LL are required to determine how this mechanism influences the pathogenesis of PsA and RA, with particular reference to the development of bony erosions at the cartilage–pannus junction. This is the first study to demonstrate that the therapeutic effect of MTX resulted in reduced MRP expression in serum and synovium. Thus, MRP antigens are a useful marker of local and systemic inflammation and are a potential novel therapeutic target in PsA, RA, and SpA.

Acknowledgements

The authors gratefully acknowledge the assistance of Dr. Bruce Kirkham for providing additional synovial samples, and Dr. Peter Youssef for assistance in performing arthroscopy and quantitative microscopic analysis of synovium.

Ancillary