Sphingosine 1-phosphate/sphingosine 1-phosphate receptor 1 signaling in rheumatoid synovium: Regulation of synovial proliferation and inflammatory gene expression

Authors


Abstract

Objective

Sphingosine 1-phosphate (S1P) is involved in various pathologic conditions and has been implicated as an important mediator of angiogenesis, inflammation, cancer, and autoimmunity. This study was undertaken to examine the role of S1P/S1P1 signaling in the pathogenesis of rheumatoid arthritis (RA).

Methods

We examined S1P1 messenger RNA (mRNA) and protein levels in RA synoviocytes and MH7A cells by reverse transcriptase–polymerase chain reaction and Western blotting. We also performed S1P1 immunohistochemistry analysis in synovial tissue from 28 RA patients and 18 osteoarthritis (OA) patients. We investigated the effects of S1P on proliferation by WST-1 assay, and its effects on tumor necrosis factor α (TNFα)– or interleukin-1β (IL-1β)–induced cyclooxygenase 2 (COX-2) expression and prostaglandin E2 (PGE2) production in RA synoviocytes and MH7A cells by Western blotting and enzyme-linked immunosorbent assay, respectively. Finally, we examined whether these effects of S1P were sensitive to pertussis toxin (PTX), an inhibitor of the Gi/Go proteins.

Results

S1P1 mRNA and protein were detected in RA synoviocytes and MH7A cells. S1P1 was more strongly expressed in synovial lining cells, vascular endothelial cells, and inflammatory mononuclear cells of RA synovium compared with OA synovium. S1P increased the proliferation of RA synoviocytes and MH7A cells. S1P alone significantly enhanced COX-2 expression and PGE2 production. Moreover, S1P enhanced expression of COX-2 and production of PGE2 induced by stimulation with TNFα or IL-1β in RA synoviocytes and MH7A cells. These effects of S1P were inhibited by pretreatment with PTX.

Conclusion

These findings suggest that S1P signaling via S1P receptors plays an important role in cell proliferation and inflammatory cytokine–induced COX-2 expression and PGE2 production by RA synoviocytes. Thus, regulation of S1P/S1P1 signaling may represent a novel therapeutic target in RA.

Sphingosine 1-phosphate (S1P) is a bioactive lipid molecule that mediates a wide variety of cellular responses via interactions with members of the endothelial differentiation gene (Edg) family of G protein–coupled receptors expressed at the plasma membrane (1). S1P receptors, namely S1P1/EDG-1, S1P2/EDG-5, S1P3/EDG-3, S1P4/EDG-6, and S1P5/EDG-8, are now recognized as high-affinity receptors for S1P (2) and exhibit variable tissue distributions. S1P1, S1P2, and S1P3 are widely expressed in various tissues, whereas expression of S1P4 is confined to lymphoid and hematopoietic tissues and S1P5 expression is largely localized to the central nervous system (1).

In mammals, S1P receptors are thought to regulate important physiologic actions, including vascular development, vascular permeability, and immune cell trafficking (3). S1P is generated by the metabolism of sphingomyelin, with S1P levels being tightly regulated by a series of enzymes, including sphingosine kinase, S1P phosphatase, and S1P lyase. Although platelets are the main source of S1P in plasma and release S1P upon activation (4), other cell types, such as erythrocytes, neutrophils, and mononuclear cells, also secrete S1P (5, 6). S1P affects diverse biologic processes, including cell growth, differentiation, migration, and apoptosis. Recently, S1P-mediated modulation of cyclooxygenase 2 (COX-2) expression has been reported (7–9). Importantly, S1P is involved in various pathologic conditions and has been implicated as an important mediator in angiogenesis, cancer, and autoimmunity (1).

Rheumatoid arthritis (RA) is a disease characterized by polyarticular inflammation, in which hyperplasia of the synovial lining cells is observed along with angiogenesis and inflammatory mononuclear cell infiltration. This contrasts with the degenerative disease osteoarthritis (OA). Various cytokines produced by both infiltrating cells and synoviocytes are the underlying cause of many of the pathologic processes of RA. In particular, tumor necrosis factor α (TNFα) and interleukin-1β (IL-1β) are key proinflammatory molecules in the cytokine cascade in RA. High levels of TNFα and IL-1β have been demonstrated in the synovial fluid and plasma of patients with RA (10, 11). These cytokines are central to the pathogenesis of RA and ultimately trigger the production of matrix metalloproteinases and activation of osteoclasts, thereby resulting in irreversible damage to soft tissues and bones (12).

It is widely known that TNFα and IL-1β induce expression of COX-2 and production of prostaglandin E2 (PGE2) by RA synoviocytes (13, 14). Blockade of these cytokines inhibits the release of other proinflammatory cytokines by synoviocytes. In fact, anti-TNFα antibodies (infliximab or adalimumab), a soluble TNF receptor–fusion protein (etanercept), and IL-1 receptor antagonist (anakinra) are widely used in the treatment of RA (15–18). These biologic agents significantly reduce joint inflammation and improve physical function in early and advanced RA. Although the downstream mechanisms of their action are still poorly understood, an important pathway may involve the COX-2 enzyme.

COX is a rate-limiting enzyme in the synthesis of PGs. Two isoforms of the COX enzyme have been described: COX-1, which is constitutively expressed in a variety of cells and tissues (19), and COX-2, which is induced by mitogens, cytokines, and growth factors and is primarily responsible for the PGs generated at sites of inflammation (20). The inflammation of RA is closely related to the production of PGE2 by synoviocytes (21), and PGE2 plays a major role in angiogenesis through the expression of vascular endothelial growth factor (VEGF) in rheumatoid synovium (22). Moreover, PGE2 has been shown to trigger bone resorption by osteoclasts (23). Various cytokines and biologic response modifiers regulate COX-2 expression.

The goal of this study was to clarify the role of S1P in the pathogenesis of RA. We demonstrated that S1P1 is strongly expressed in RA synovium compared with OA synovium. In addition, S1P signaling via S1P1 enhances synoviocyte proliferation and inflammatory cytokine–induced COX-2 expression and PGE2 production. We conclude that S1P/S1P1 signaling may play an important role in the pathogenesis of RA.

MATERIALS AND METHODS

Human synovial tissue specimens.

Synovial tissues were isolated from patients with RA and OA at the time of arthroscopic biopsy or total joint replacement. All 28 RA patients met the 1987 American College of Rheumatology (ACR; formerly, the American Rheumatism Association) criteria for the classification of RA (24), and all 18 OA patients met the ACR criteria for OA classification (25). Synovial tissue specimens were preserved in 10% formalin, embedded in paraffin, and serially sectioned onto microscope slides at a thickness of 4 μm. All patients gave informed consent, and the institutional medical ethics committee approved the study protocol.

Cell line.

Human MH7A synovial cells isolated from intraarticular soft tissue of the knee joints of RA patients were obtained from Riken (Saitama, Japan). MH7A is a cell line established by transfection with the SV40 T antigen (26). MH7A cells were cultured in RPMI 1640 (Sigma, St. Louis, MO) containing 10% heat-inactivated fetal bovine serum (FBS; Whittaker, Walkersville, MD), 100 units/ml of penicillin, and 100 μg/ml of streptomycin (Invitrogen, San Diego, CA) at 37°C in an atmosphere of 5% CO2 in air.

Establishment of human fibroblast-like synoviocytes (RA synoviocytes).

Synovial tissue specimens from RA patients were minced into small pieces and treated with 1 mg/ml of collagenase (Sigma) for 2 hours at 37°C in serum-free RPMI 1640, filtered through a nylon mesh, and washed extensively (27). Next, the cells were suspended in RPMI 1640 containing 10% heat-inactivated FBS, 100 units/ml of penicillin, and 100 μg/ml of streptomycin. All cultures were incubated at 37°C in an atmosphere of 5% CO2 in air. When the cell cultures reached confluence, synoviocytes were trypsynized and passaged to other flasks. RA synoviocytes were used for experiments between the fourth and ninth passages.

RNA preparation and analysis of S1P1 and S1P3 messenger RNA (mRNA).

Total RNA from RA synoviocytes and MH7A cells was prepared using Isogen according to the instructions of the manufacturer (Nippon Gene, Toyama, Japan). Reverse transcription (RT) and complementary DNA amplification were performed with the TaKaRa RNA polymerase chain reaction (PCR) kit (Takara Shuzo, Otsu, Japan) (28). The primers used for the RT-PCR analysis have been reported previously (8). Forward and reverse primers were as follows: human S1P1 (429-bp product) sense 5′-TATCAGCGCGGACAAGGAGAACAG-3′, antisense 5′-ATAGGCAGGCCACCCAGGATGAG-3′; human S1P3 (394-bp product) sense 5′-CTGCCTGCACAATCTCCCTGACTG-3′, antisense 5′-GGCCCGCCGCATCTCCT-3′; GAPDH (246-bp product) sense 5′-GATGACATCAAGAAGGTGGTGAA-3′, antisense 5′-GTCTTACTCCTTGGAGGCCAT-GT-3′. The conditions for thermal cycling were as follows: 30 cycles of 94°C for 1 minute, 62°C for 1 minute, and 72°C for 1 minute. PCR products were electrophoresed on a 2% agarose gel and visualized by ethidium bromide staining.

Immunohistochemistry analysis for S1P1 and COX-2 expression in RA and OA synovium.

Immunostaining was performed with peroxidase labeling techniques (21). All procedures were performed at room temperature. Tissue sections were deparaffinized, and endogenous peroxidase activity was blocked by incubation in 0.3% peroxide in methanol for 45 minutes. The sections were preincubated with 1% FBS in phosphate buffered saline (PBS) for 60 minutes, followed by incubation overnight with either the primary antibody against human S1P1 (purified IgY from a chicken immunized with human S1P1) or preimmune chicken serum (1:200 dilution in PBS) (29). Sections were then washed in PBS and incubated for 60 minutes with horseradish peroxidase (HRP)–conjugated anti-chicken IgY (1:200 dilution in PBS) (29). The sections were further washed with PBS. Finally, color was developed by immersing the sections in a solution of 0.05% (weight/volume) 3,3′-diaminobenzidine (DAB; Sigma) and 0.01% hydrogen peroxide in 0.05M Tris (pH 7.4) for 3 minutes. The sections were counterstained with Mayer's hematoxylin (Wako, Osaka, Japan). Positive staining was indicated by brownish deposits, and background staining was purple.

COX-2 staining was performed with the Vectastain ABC kit according to the protocol suggested by the manufacturer (Vector, Burlingame, CA) (21), using primary antibody against human COX-2 (1:400 dilution in PBS; Santa Cruz Biotechnology, Santa Cruz, CA) (30, 31). Positive staining was indicated by brownish deposits. Double antibody staining with anti-S1P1 antibody and anti–COX-2 antibody was performed as previously described (32). For S1P1 staining, the color was developed with a DAB solution including 0.04% (w/v) nickel chloride. Positive staining was indicated by black deposits. COX-2 staining was performed with alkaline phosphatase labeling techniques. Positive staining was indicated by red deposits. Counterstaining was performed with methyl green solution (Vector). Double-positive staining was indicated by black-red deposits, and background staining was green. Control staining with preimmune chicken serum and normal goat IgG was negative in all cases.

Immunohistochemistry analysis for S1P1 expression in RA and OA synovial tissue.

For each tissue specimen from the RA and OA patients, the extent and intensity of anti– SP-1P1 staining in synovial lining cells, inflammatory mononuclear cells, and vascular endothelial cells was graded on a scale of 0–4+ by 2 blinded observers on 2 separate occasions, using coded slides as previously described (21, 30–32). The observer assigned the scores after assessment of the entire tissue section, such that a grade of 4+ indicated that staining was maximally intense throughout the specimen, while a grade of 0 indicated that staining was absent throughout the specimen.

Cell proliferation studies.

RA synoviocytes or MH7A cells were cultured with or without S1P (0–0.1 μM) in triplicate in flat-bottomed 96-well microplates at a concentration of 1 × 105 cells/ml in RPMI 1640 containing 10% (v/v) FBS. After 72 hours, cell viability was assessed by measuring mitochondrial NADA–dependent dehydrogenase activity with a cell counting kit (Dojindo, Kumamoto, Japan), using the sulfonated tetrazolium salt 2-(4-iodophenyl)-3-(4-nitrophenyl)-5-(2,4-disulfophenyl)-2H-tetrazolium monosodium (WST-1) (27).

Measurement of PGE2 levels.

RA synoviocytes or MH7A cells (2 × 104) were plated in flat-bottomed 24-well microplates and were cultured with S1P (0–0.1 μM) in the presence or absence of 100 ng/ml TNFα (Sigma) or 10 pg/ml IL-1β (Sigma). After 72 hours of incubation, PGE2 synthesis was determined by assaying supernatants with a PGE2 enzyme-linked immunosorbent assay (ELISA) kit according to the instructions of the manufacturer (Assay Designs, Ann Arbor, MI) (33).

Western blot analysis for COX-2.

RA synoviocytes or MH7A cells (1 × 107) were plated in flat-bottomed 24-well microplates and cultured with S1P (0–0.1 μM) in the presence or absence of 100 ng/ml TNFα. After 12 hours of incubation, cells were lysed in RIPA lysis buffer (Santa Cruz Biotechnology), and protein content was determined using Bio-Rad protein assay reagent (Bio-Rad, Hercules, CA) with bovine serum albumin as standard. Each sample (20 μg) was resolved on 10% polyacrylamide gels under denaturing conditions and then transferred to 0.45-μm nitrocellulose membranes. After blocking overnight at 4°C with 5% nonfat milk in Tris buffered saline–0.01% Tween 20 (Santa Cruz Biotechnology), membranes were incubated with primary antibody against COX-2 (1:400 dilution in PBS; Santa Cruz Biotechnology) overnight at 4°C. After washing the membranes with Tris buffered saline–0.05% Tween 20 (washing buffer), HRP-conjugated goat anti-rabbit secondary antibody (1:1,000 dilution in PBS; Santa Cruz Biotechnology) was added, followed by incubation for 45 minutes. After further washing, the color was developed with luminol reagent (Santa Cruz Biotechnology), and HRP activity of the blots was analyzed using an LAS1000 imager (Fuji, Tokyo, Japan).

Western blot analysis for S1P1.

Samples from RA synoviocytes or MH7A cells were prepared using RIPA lysis buffer and Bio-Rad protein assay reagent as described above. Detection of S1P1 protein was performed by Western blot analysis using chicken anti-S1P1 IgY (1:200 dilution in PBS) (primary antibody) and HRP-conjugated anti-chicken IgY (1:1,000 dilution in PBS) (secondary antibody). Color was developed with luminol reagent, and HRP activity of the blots was analyzed using an LAS1000 imager.

Measurement of the effect of pertussis toxin (PTX) on S1P-enhanced MH7A cell proliferation and TNFα- or IL-1β–induced PGE2 production by MH7A cells.

MH7A cells were preincubated for 24 hours in the presence or absence of 100 ng/ml PTX (Sigma). After rigorous washing, cells were stimulated with S1P (0–0.05 μM). After incubation for 72 hours, cell proliferation was assessed by WST-1 assay. PTX-pretreated cells were stimulated with S1P in the presence or absence of 100 ng/ml TNFα or 10 pg/ml IL-1β. After 72-hour incubation, PGE2 levels were measured by ELISA.

Statistical analysis.

Results are expressed as the mean ± SEM. The significance of the difference between the experimental results and control values was determined by Student's t-test. P values less than 0.05 were considered significant.

RESULTS

S1P1 and S1P3 mRNA expression in RA synoviocytes and MH7A cells.

To examine whether RA synoviocytes and MH7A cells express receptors for S1P, we first determined the expression of S1P1 and S1P3 receptor transcripts in synoviocytes from RA patients (n = 10) and in MH7A cells, using RT-PCR analysis (Figure 1A). RA synoviocytes (10 of 10) and MH7A cells expressed S1P1 and S1P3. Human umbilical vein endothelial cells (HUVECs) were included in these studies as positive controls for S1P1 and S1P3 expression (34). Negative control studies, performed with no RNA or no RT in the RT reaction, yielded no detectable bands (results not shown). Primers specific for human GAPDH generated the expected 246-bp band in RA synoviocytes and MH7A cells.

Figure 1.

Expression of sphingosine 1-phosphate 1 (S1P1)/S1P3 mRNA and S1P1 protein in rheumatoid arthritis (RA) synoviocytes and MH7A cells. A, Detection of S1P1 and S1P3 mRNA expression by reverse transcriptase–polymerase chain reaction (PCR). Primers for S1P1 generated the expected 429-nucleotide band, and S1P3 the expected 394-nucleotide band, after 30 cycles of PCR. Primers for GAPDH generated the expected 246-bp band in all subjects. Human umbilical vein endothelial cells (HUVECs) were used as a positive control. B, Detection of S1P1 protein by Western blot analysis. Shown are representative results from 1 of 5 RA patients. HUVECs were used as a positive control. Lane 1, MH7A cells. Lane 2, RA synoviocytes. Lane 3, HUVECs.

S1P1 protein expression in RA synoviocytes and MH7A cells.

To investigate whether RA synoviocytes and MH7A cells express S1P1 protein, we performed Western blot analysis of S1P1 (Figure 1B). By Western blot analysis of chicken anti-human S1P1 antibody, S1P1 protein (40 kd) was detected in RA synoviocytes (5 of 5) and MH7A cells, as well as in HUVECs, used as a positive control for S1P1 expression.

S1P1 immunostaining of synovial tissue from patients with RA and OA.

To examine the expression and localization of S1P1 in RA and OA synovial tissue, we performed immunohistochemistry analysis (Figure 2). In RA synovial tissue (Figures 2A and E), we found markedly enhanced expression of S1P1 in synovial lining cells, vascular endothelial cells, and inflammatory mononuclear cells, compared with findings in OA synovial tissue (Figure 2B). In normal synovium obtained from patients with trauma (n = 3), S1P1 expression was very weak in synovial lining cells and vascular endothelial cells, as shown by immunohistochemistry analysis (results not shown). Control staining with preimmune chicken serum (Figure 2C) and HRP-conjugated anti-chicken IgY (Figure 2D) was uniformly negative. The S1P1 expression was localized to the cytoplasmic regions, with an absence of nuclear staining, consistent with the fact that G protein–coupled receptors are synthesized in the membranous compartments of the cytoplasm and transported to the plasma membrane. We also detected expression of S1P1 in both RA synoviocytes (Figure 2F) and MH7A cells (results not shown). The intensity and cellular localization of S1P1 in RA synoviocytes and MH7A cells were comparable with those in synovial lining cells.

Figure 2.

Sphingosine 1-phosphate 1 (S1P1) immunostaining of synovial tissue from patients with rheumatoid arthritis (RA) or osteoarthritis (OA). Positive staining is indicated by brown deposits; background appears purple. A and E, RA synovium stained with anti-S1P1 IgY. B, OA synovium stained with anti-S1P1 IgY. C, RA synovium stained with preimmune chicken serum. D, RA synovium stained with horseradish peroxidase–conjugated anti-chicken IgY. E, RA synovium stained with anti-S1P1 IgY (SL = synovial lining cell layer; M = mononuclear cell; BV = blood vessel). F, RA synoviocytes stained with anti-S1P1 IgY (FLS = fibroblast-like synoviocytes). (Original magnification × 200 in A–D; × 400 in E; × 800 in F.)

Detection of colocalization of S1P1 and COX-2 in RA synovium.

To investigate the expression of S1P1 and COX-2 in RA synovium, we immunostained the same sections from RA synovial tissue specimens using anti-S1P1 antibody and anti–COX-2 antibody, as shown in Figure 3. Immunoreactive S1P1 (Figure 3A) and COX-2 (Figure 3B) were detected in synovial lining cells, inflammatory mononuclear cells, and vascular endothelial cells. Control staining with normal goat IgG was uniformly negative (Figure 3C).

Figure 3.

Immunohistochemistry analysis for S1P1 and cyclooxygenase 2 (COX-2) in RA synovium. The same synovial sections were used to analyze both S1P1 and COX-2. Positive staining is indicated by brown deposits; background appears purple. A, RA synovium stained with anti-S1P1 IgY. B, RA synovium stained with anti–COX-2 IgG. C, RA synovium stained with normal goat IgG. D–H, Double staining with anti-S1P1 IgY and anti–COX-2 IgG. The same synovial sections were stained with the first antibody (S1P1 antibody) using the peroxidase method and with the second antibody (COX-2 antibody) using the alkaline phosphatase method. Positive staining for peroxidase or alkaline phosphatase resulted in black or red deposits, respectively. Double-positive staining yielded black-red deposits. Counterstaining was performed with methyl green solution. D and H, RA synovium stained with anti-S1P1 IgY/anti–COX-2 IgG. E, RA synovium stained with anti-S1P1 IgY/normal goat IgG. F, RA synovium stained with preimmune chicken serum/anti–COX-2 IgG. G, RA synovium stained with preimmune chicken serum/normal goat IgG. See Figure 2 for other definitions. (Original magnification × 100 in A–C; × 200 in D–G; × 400 in H.)

We also immunostained the same sections from RA synovium by a double antibody staining method. The first antigen (S1P1) was stained with peroxidase and the second (COX-2) was stained with alkaline phosphatase. Black or red deposits, respectively, represented positive staining. Intense black-red deposits were demonstrated in synovial lining cells, inflammatory mononuclear cells, and vascular endothelial cells upon double staining with both anti-S1P1 and anti–COX-2 (Figures 3D and H). Control staining with preimmune serum/normal goat IgG (Figure 3G) was uniformly negative. Staining with anti-S1P1/normal goat IgG (Figure 3E) showed only black deposits. Staining with preimmune serum/anti–COX-2 (Figure 3F) showed only red deposits. These observations suggest that S1P1 and COX-2 have very similar patterns of expression and localization in rheumatoid synovium.

S1P1 immunostaining in RA and OA synovial tissue.

The extent and intensity of S1P1 immunostaining of synovial lining cells, inflammatory mononuclear cells, and vascular endothelial cells in 28 RA and 18 OA synovial specimens was graded on a scale of 0–4+ by 2 blinded observers (21). S1P1 staining was significantly more extensive and intense in rheumatoid synovial specimens compared with specimens from OA patients (Figures 4A–C).

Figure 4.

Quantitative analysis of S1P1 immunostaining in RA and OA synovial tissue. The extent and intensity of staining with anti-S1P1 antibody in A, synovial lining cells, B, inflammatory mononuclear cells, and C, vascular endothelial cells from 28 patients with RA and 18 patients with OA were assessed by 2 observers under blinded conditions and graded on a scale of 0–4+. A grade of 4+ indicates that staining was maximally intense throughout the specimen; a grade of 0 indicates that staining was absent throughout the specimen. Values are the mean ± SEM. ∗ = P < 0.01. See Figure 2 for definitions.

Effect of S1P on proliferation of and COX-2–induced PGE2 production by rheumatoid synoviocytes.

To investigate the effect of S1P on the proliferation of RA synoviocytes, we performed a modified MTT assay (WST-1 assay). Treatment with S1P (0.01–0.05 μM) enhanced the proliferation of RA synoviocytes (n = 13) (Figure 5A). The maximal effect observed was a 1.25-fold increase in proliferation, induced by 0.01 μM S1P. Additionally, we investigated the effect of S1P on the proliferation of MH7A cells. Treatment with S1P at concentrations ranging from 0.01 μM to 0.05 μM enhanced the proliferation of MH7A cells in a dose-dependent manner (Figure 5B). The maximal effect was a 1.6-fold increase in proliferation induced by 0.05 μM S1P. At higher concentrations of S1P (0.1 μM), proliferation returned to baseline levels.

Figure 5.

Effect of S1P on proliferation of A, RA synoviocytes (n = 13) and B, MH7A cells. Values are the mean and SEM from 3 experiments. ∗∗ = P < 0.05; # = P < 0.01. See Figure 2 for definitions.

Next, to investigate whether S1P has a role in the production of PGE2 by RA synoviocytes, PGE2 levels were measured by ELISA. Interestingly, S1P, at the same concentrations used in the proliferation assay, enhanced PGE2 production by RA synoviocytes (Figure 6A). Addition of S1P resulted in a 1.5-fold increase in PGE2 production compared with control cells, with maximal activity evident at an S1P concentration of 0.05 μM. Moreover, S1P (0.01–0.05 μM) enhanced TNFα- or IL-1β–induced PGE2 production by RA synoviocytes. Addition of S1P resulted in an ∼1.5-fold increase in PGE2 production compared with TNFα or IL-1β alone, with maximal activity evident at S1P concentrations of 0.01–0.05 μM. In MH7A cells, stimulation with S1P alone (0.01–0.1 μM) did not influence PGE2 production. However, S1P enhanced TNFα- or IL-1β–induced PGE2 production by MH7A cells (Figure 6B), as was observed with RA synoviocytes. Addition of S1P resulted in a 1.8-fold increase in PGE2 production compared with cells stimulated with TNFα alone, with maximal activity evident at an S1P concentration of 0.05 μM. Addition of S1P induced a 1.5-fold increase in PGE2 production compared with IL-1β stimulation alone, with maximal activity evident at an S1P concentration of 0.01 μM.

Figure 6.

Effect of sphingosine 1-phosphate (S1P) on tumor necrosis factor α (TNFα)– or interleukin-1β (IL-1β)–induced prostaglandin E2 (PGE2) production by A, rheumatoid arthritis (RA) synoviocytes (n = 13) and B, MH7A cells. Open bars show stimulation with S1P alone; solid bars show stimulation with both S1P and TNFα; hatched bars show stimulation with both S1P and IL-1β. 0 S1P indicates stimulation with TNFα or IL-1β alone. Values are the mean and SEM from 3 experiments. ∗ = P < 0.02; # = P < 0.01; ∗∗ = P < 0.05.

To further investigate whether S1P-enhanced PGE2 production is mediated through COX-2 induction, the expression level of COX-2 protein was examined by Western blotting using a specific anti–COX-2 antibody. S1P also enhanced COX-2 protein expression by RA synoviocytes (n = 5). The 72-kd COX-2 protein was increased in a concentration-dependent manner (0–0.05 μM) compared with control cells (Figure 7A). The maximal effect (1.7-fold increase) was observed with an S1P concentration of 0.05 μM.

Figure 7.

Effect of S1P on A, cyclooxygenase-2 (COX-2) protein expression in RA synoviocytes (representative data from 1 of 5 patients) and B, TNFα-induced COX-2 protein expression in MH7A cells. The levels of COX-2 protein were examined by Western blot analysis. Western blotting was also performed with an antiactin antibody to confirm equal protein loading. Results are depicted quantitatively as the percent of control, with baseline COX-2 protein expression being regarded as 100%. See Figure 6 for other definitions.

In MH7A cells, stimulation with S1P alone did not affect PGE2 production, and we therefore investigated the enhancement of TNFα-induced COX-2 protein expression by S1P in MH7A cells. When TNFα (100 ng/ml)–treated MH7A cells were stimulated with S1P at various concentrations (0–0.1 μM), the 72-kd COX-2 protein was increased in a concentration-dependent manner compared with its production in control cells treated with TNFα alone (Figure 7B). The maximal effect (2.2-fold increase) occurred with an S1P concentration of 0.1 μM. The concentration dependence of S1P-induced COX-2 expression was similar to that of S1P-induced PGE2 production. These results indicate that S1P is an important factor in the induction of COX-2 expression and the enhancement of PGE2 production in rheumatoid synoviocytes.

Mediation of S1P-enhanced proliferation and TNFα- or IL-1β–induced PGE2 production in MH7A cells via a PTX-sensitive pathway.

We demonstrated in this study that rheumatoid synoviocytes express S1P1 and S1P3, and it has been reported that S1P1 and S1P3 couple to G proteins of the Gi/Go family (35, 36). Therefore, we investigated the role of PTX-sensitive G proteins in S1P enhancement of rheumatoid synoviocyte proliferation. Unfortunately, the viability of RA synoviocytes was diminished by pretreatment with PTX (100 ng/ml). Therefore, we used MH7A cells, but not RA synoviocytes, in this assay. MH7A cells were preincubated with 100 ng/ml of PTX for 24 hours. After rigorous washing, the cells were stimulated with S1P (0–0.05 μM). After 72 hours, we performed a modified MTT (WST-1) assay. Pretreatment with PTX inhibited S1P-enhanced proliferation in MH7A cells (Figure 8A).

Figure 8.

Effect of pertussis toxin (PTX) on S1P-enhanced MH7A cell proliferation and on S1P-enhanced TNFα or IL-1β induction of PGE2 by MH7A cells. A, MH7A cell proliferation. ∗ = P = 0.027. B, S1P-mediated enhancement of TNFα-induced PGE2 production by MH7A cells. ∗ = P < 0.01. C, S1P-mediated enhancement of IL-1β–induced PGE2 production by MH7A cells. ∗ = P < 0.01. Values are the mean and SEM (n = 3 experiments) with (solid bars) or without (open bars) PTX pretreatment. NS = not significant (see Figure 6 for other definitions).

We further investigated the effect of PTX-sensitive G proteins on S1P enhancement of PGE2 production induced by the inflammatory cytokines TNFα or IL-1β. MH7A cells were preincubated with 100 ng/ml of PTX for 24 hours prior to stimulation with each of the inflammatory cytokines and S1P. Pretreatment with PTX inhibited the S1P enhancement of TNFα- or IL-1β–induced PGE2 production (Figures 8B and C). These results show that S1P enhancement of rheumatoid synoviocyte proliferation and TNFα- or IL-1β–induced PGE2 production by rheumatoid synoviocytes occurs via a Gi/Go-dependent pathway.

DISCUSSION

Several novel findings were demonstrated in this study: 1) S1P1/S1P3 mRNA and S1P1 protein are present in RA synoviocytes and MH7A cells. 2) S1P1 is markedly expressed in synovial lining cells, vascular endothelial cells, and inflammatory mononuclear cells of RA synovium. 3) S1P1 expression is significantly more extensive and intense in RA synovium than in OA synovium. 4) The localization of S1P1 expression is similar to that of COX-2 expression in RA synovium. 5) S1P increases the proliferation of RA synoviocytes and MH7A cells. 6) S1P alone induces COX-2 expression and PGE2 production by RA synoviocytes, and, in addition, S1P enhances inflammatory cytokine (TNFα or IL-1β)–induced COX-2 expression and PGE2 production by RA synoviocytes and MH7A cells. 7) PTX partially blocks S1P modulation of these effects.

S1P has been shown to affect diverse biologic processes, including cell growth, differentiation, migration, and apoptosis. S1P affects many different cells, such as lymphocytes (37), macrophages (38), and endothelial cells (35). Previous evidence has implicated S1P in the pathophysiology of disease states, such as autoimmunity, inflammation (39), cancer (40, 41), and angiogenesis (34, 35).

In this study, we examined the participation of S1P in the pathogenesis of RA. First, we confirmed the abundant expression of S1P1 protein in RA and OA synovium. Surprisingly, S1P1 was more highly expressed in synovial lining cells, vascular endothelial cells, and inflammatory mononuclear cells of RA synovium compared with OA synovium. The histologic hallmarks of RA include hyperplasia of the synovial intimal lining, angiogenesis, and infiltration of mononuclear cells, especially synovial T cells and macrophages. When we examined the content of S1P in synovial fluid from patients with RA and OA by high-performance liquid chromatography, we found that S1P levels were higher in RA patients (mean 1,078.92 pM/ml; n = 17) compared with OA patients (765.01 pM/ml; n = 13). S1P levels in synovial fluid from RA and OA patients were much higher than levels in serum (∼400 pM/ml) or plasma (∼100 pM/ml). S1P/S1P1 signaling in RA synoviocytes may thus play an important role in the pathogenesis of RA.

It has been previously reported that vascular endothelial cells express S1P1 and that S1P signaling via the S1P1 receptor is important for angiogenesis (34, 35). Indeed, deletion of the S1P1 receptor in mice revealed an important function of S1P in the regulation of vascular stabilization (42). Thus, S1P1 signaling in the endothelial compartment is critical for the regulation of vascular maturation. S1P has a variety of effects on endothelial cells, including endothelial migration, differentiation, and survival. It is one of the most potent inducers of in vitro chemotaxis of endothelial cells (34) and also induces endothelial cell proliferation (43). Indeed, S1P regulates the morphogenetic differentiation of endothelial cells into a capillary-like, tubular phenotype (35).

S1P stimulates new blood vessel formation or angiogenesis in combination with VEGF or basic fibroblast growth factor in vivo (44), and it is therefore pertinent that VEGF and fibroblast growth factors are produced at high levels by RA synoviocytes. Moreover, S1P can activate VEGF receptor 2 in the absence of exogenous VEGF, by receptor crosstalk (45). VEGF is an endothelial cell mitogen and is present in high concentrations in RA synovial fluid and tissue (46). VEGF expression is especially high in the synovial intimal lining and is produced by cultured fibroblast-like synoviocytes that have been exposed to hypoxia and IL-1 (47). These data strongly suggest that S1P may be a potent inducer of angiogenesis in RA.

A role of S1P in inflammation was suggested by the observations that S1P can protect leukocytes and lymphocytes from undergoing apoptosis induced by Fas, TNFα, and ceramide (48). S1P can activate monocytes (38) and endothelial cells (35) and can also mediate lymphocyte infiltration (37). It enhances TNFα and IL-1β mRNA expression and secretion from mouse macrophages (38). Low concentrations of S1P are directly chemotactic for T cells and enhance T cell chemotactic responses to diverse chemokines (37). Therefore, the increased expression of S1P1 in infiltrating mononuclear cells and the resultant increase in S1P signaling may be intimately associated with the inflammation of the synovium in RA.

We also investigated the effect of S1P stimulation in RA synoviocytes and MH7A cells that expressed S1P1 and S1P3 receptors. S1P increased the proliferation of RA synoviocytes and MH7A cells. Moreover, S1P induced COX-2 expression and PGE2 production in RA synoviocytes, and enhanced TNFα- or IL-1β–induced COX-2 expression and subsequent PGE2 production in both RA synoviocytes and MH7A cells.

We then investigated the effect of PTX (which specifically inactivates Gi/Go-mediated signaling pathways) on enhancement of cell proliferation and inflammatory cytokine–induced PGE2 production by MH7A cells. Both of these actions of S1P were inhibited by PTX. Since S1P1 couples only to Gi, most of its effects are PTX sensitive (35). However, S1P3 binds to PTX-sensitive as well as to PTX-insensitive Gq and G13 (36). These results suggest that S1P stimulation of Gi protein–dependent signaling pathways plays an important role in modulating cell proliferation and inflammatory cytokine–induced COX-2 expression and PGE2 production by MH7A cells. A role of S1P in the modulation of COX-2 expression has been reported previously. Pettus et al have demonstrated that S1P mediates COX-2 induction and PGE2 production in response to TNFα in murine L929 fibroblast cells and human adenocarcinoma A549 cells (9). In addition, Kim and colleagues have demonstrated that S1P regulates COX-2 expression in amnion-derived WISH cells (8). In WISH cells, COX-2 induction promoted by S1P is partially blocked by PTX (8). Our results are consistent with these reports.

The observation that S1P is a critical mediator of COX-2 induction has potentially important implications regarding therapy in RA. COX-2–induced PGE2 production by rheumatoid synoviocytes plays an important role in inflammation, angiogenesis through VEGF expression, and osteoclastic bone destruction. Specifically, S1P may aggravate synovial hyperplasia, inflammation, and angiogenesis through the induction of COX-2 in RA synovium.

In conclusion, we have demonstrated increased S1P1 expression in inflamed RA synovium. S1P1 expression is closely connected with synovial hyperplasia, angiogenesis, and inflammation in RA. In particular, S1P signaling via S1P1 plays an important role in synoviocyte proliferation and COX-2–induced PGE2 production by RA synoviocytes. Therefore, regulation of S1P signaling via S1P1 may be a novel therapeutic target in RA.

Acknowledgements

We wish to thank Eriko Kagawa for immunohistochemical analysis, Sachie Kitano for excellent technical support in Western blot analysis, Yuko Taki for PCR analysis, Atsushi Omoto for supplying RA synoviocytes, Yuen Yee and Julie Saba for performing high-performance liquid chromatography, and Ronald L. Wilder for useful comments.

Ancillary