To examine the role of sirtuin-1 (SIRT-1)/FoxO3a in the expression of cysteine-rich protein 61 (CYR-61) in rheumatoid arthritis synovial fibroblasts (RASFs) and the influence of simvastatin on this pathway, and to determine the relationship between disease progression and FoxO3a/CYR-61 signaling in synovial fibroblasts in vivo using a rat model of collagen-induced arthritis (CIA).
In RASFs, the expression of CYR-61 and SIRT-1, the localization of FoxO3a in the nucleus/cytoplasm, and the phosphorylation/acetylation of FoxO3a were examined by Western blotting. Secretion of CCL20 was assessed by enzyme-linked immunosorbent assay. Promoter activity of the Cyr61 gene was evaluated by luciferase assay, with or without forced expression of FoxO3a and SIRT-1 by lentiviral transduction. FoxO3a–Cyr61 promoter interaction was examined by chromatin immunoprecipitation. In rats with CIA, the expression of CYR-61 and phosphorylated FoxO3a in synovial fibroblasts was examined by immunohistochemistry.
In RASFs, simvastatin suppressed the tumor necrosis factor α (TNFα)–induced production of CYR-61 and CCL20. Nuclear levels of FoxO3a were decreased after TNFα stimulation of RASFs, and forced expression of FoxO3a reversed the inductive effects of TNFα on CYR-61. Simvastatin inhibited the nuclear export, phosphorylation, and acetylation of FoxO3a and maintained its binding to the Cyr61 promoter. Forced expression of SIRT-1 in RASFs led to decreased levels of CYR-61 and deacetylation of FoxO3a. Following treatment with simvastatin, the expression of SIRT-1 was up-regulated and SIRT-1/FoxO3a binding was enhanced in RASFs. In rats with CIA, intraarticular injection of simvastatin alleviated arthritis and suppressed CYR-61 expression and FoxO3a phosphorylation in synovial fibroblasts.
CYR-61 is important in the pathogenesis of RA, and SIRT-1/FoxO3a signaling is crucial to induction of CYR-61 in RASFs. Simvastatin plays a beneficial role in inflammatory arthritis through its up-regulation of SIRT-1/FoxO3a signaling in synovial fibroblasts. Continued study of the pathways linking sirtuins, FoxO proteins, and the inflammatory responses of RASFs may provide new insights into the pathophysiology of RA.
Rheumatoid arthritis (RA) is a chronic autoimmune inflammatory disease that affects the synovial joints. Development of RA is characterized by hyperplasia of the synovial lining, in which synovial fibroblasts and other immune cells interact with one another to create a pathologic tissue that provokes the destruction of cartilage and bone. Activated synovial fibroblasts play a key role in RA pathogenesis by producing proteases, cytokines, and chemokines that contribute to joint destruction and perpetuate the inflammatory process (1). Although previous studies showed that a network of cytokines, including tumor necrosis factor α (TNFα) and interleukin-1β (IL-1β), contribute to the activation of RA synovial fibroblasts (RASFs), the exact signaling mechanisms that participate in the inflammatory responses of RASFs are still not fully understood (1).
Cysteine-rich protein 61 (CYR-61; also known as CCN1) belongs to the CCN family of matricellular signaling molecules (2). Functionally, CYR-61 has been shown to regulate angiogenesis and the proliferation, adhesion, migration, and differentiation of cells (2) and is important for wound healing (3) and embryo development (4). Previous studies have implicated CYR-61 in the pathogenesis of inflammatory diseases, including RA. A complementary DNA (cDNA) microarray analysis of B cells from monozygotic twins revealed significantly higher expression of CYR-61 in the twin with RA compared with the healthy twin (5).
In a recent study, we found that CYR-61 stimulated CCL2 expression in human osteoblastic cells, and the levels of CYR-61 in osteoblasts correlated with disease progression in an animal model of inflammatory arthritis (6). Zhang et al demonstrated that CYR-61 was overexpressed in RA synovial tissue and that it played a critical role in IL-17–mediated proliferation of RASFs (7). Those authors further reported that CYR-61 activated the production of IL-6 in RASFs via the αvβ5/Akt/NF-κB signaling pathway, and stimulation of the differentiation of IL-17–producing Th17 cells by RASFs was dependent on CYR-61 (8).
Since RASFs also produce a large amount of CCL20, a chemokine required for the recruitment of arthritogenic Th17 cells (9), it would be of interest to investigate whether CYR-61 is also involved in the regulation of CCL20 production by RASFs. More importantly, the regulatory mechanism of CYR-61 expression in RASFs needs to be clarified.
Recently, Lee et al (10) found that the promoter of the Cyr61 gene contains the forkhead factor binding motif, and FoxO3a is a novel negative regulator of the Cyr61 gene in vascular smooth muscle cells. The FoxO subfamily of forkhead transcription factors plays several key roles in cellular survival and metabolism, and these factors are crucial in immune regulation (11). A growing body of evidence indicates that FoxO3a and other FoxO proteins are important for maintaining immune cell homeostasis in RA (12–15). However, the role of FoxO proteins in the effector mechanisms of RASFs has not been studied.
The activity of FoxO proteins is tightly regulated by posttranslational modifications, including phosphorylation, acetylation, and ubiquitination (16, 17). In general, the phosphorylation status governs the nucleus–cytoplasm shuttling, and acetylation/deacetylation can alter the transcriptional program of FoxO proteins (16, 17).
The sirtuins (SIRT-1–SIRT-7) are a family of NAD+-dependent protein deacetylases that modulate the function of many cellular proteins, including FoxO transcription factors (18). Brunet et al (19) reported that SIRT-1, the best characterized member of the family, has a dual effect on FoxO3a function. SIRT-1 enhances the ability of FoxO3a to confer cell cycle arrest and resistance to oxidative stress, but also inhibits the ability of FoxO3a to induce cell death. It is therefore interesting to examine the regulatory role of SIRT-1/FoxO3a in CYR-61 expression in RASFs.
Although hydroxymethylglutaryl-coenzyme A reductase inhibitors (statins) are not routine treatment agents for RA patients, mounting evidence indicates that they have beneficial effects against RA progression (20, 21). However, previous studies using collagen-induced arthritis (CIA) in rodent models to examine the therapeutic effects of systemically delivered statins have yielded conflicting results (22–25). The disparity may be due to the liver-specific pharmacokinetics and poor distribution of statins to the bone and joints when given systemically. In an earlier study using intraarticular drug administration in a rat CIA model, we confirmed the beneficial effect of simvastatin on inflammatory arthritis (6), which suggests that intraarticular injection of statins may be a useful therapy for RA. Although previous studies showed that the protective effects of statins on RA may derive from multiple mechanisms (21, 25, 26), their action on SIRT-1 signaling in RASFs has not been investigated.
The aims of this study were to examine the role of SIRT-1/FoxO3a in the expression of CYR-61 in RASFs and determine the influence of simvastatin on this pathway. In addition, with the use of a rat model of CIA, the relationship between disease progression and FoxO3a/CYR-61 signaling in synovial fibroblasts was assessed in vivo.
MATERIALS AND METHODS
Antibodies against human/rat CYR-61, β-actin, and rat CD55 were from Santa Cruz Biotechnology. Anti–lamin B1 and anti–acetylated lysine antibodies were from Abcam. Anti–α-tubulin antibodies were obtained from Sigma-Aldrich. Anti-FoxO3a and human/rat phosphorylated (phospho)–FoxO3a (Ser253) antibodies were from Millipore. Anti–SIRT-1 antibodies and recombinant human CYR-61 were purchased from Abnova. Recombinant human TNFα was obtained from PeproTech. The human CCL20 enzyme-linked immunosorbent assay (ELISA) kit was purchased from R&D Systems. Bovine type II collagen (CII) was obtained from Chondrex. Inactivated Mycobacterium tuberculosis and Freund's incomplete adjuvant (IFA) were from Difco.
Synovial tissue specimens were obtained during synovectomy or joint replacement surgery from RA patients. Informed consent was obtained from all patients and the study was approved by the Research Ethics Committee of National Taiwan University Hospital. All RA patients fulfilled the American College of Rheumatology 1987 criteria for the classification of RA (27).
To isolate synovial fibroblasts, synovial tissue specimens were minced into small pieces and incubated for 2 hours with 1 mg/ml type I collagenase (Sigma-Aldrich) in Dulbecco's modified Eagle's medium (DMEM) at 37°C. Cells were collected by filtering the suspension through nylon mesh (70 μm). The cells were extensively washed and cultured in complete high-glucose DMEM (Hyclone; Thermo Fisher Scientific) supplemented with 10% fetal calf serum (Gibco, Life Technologies), 2 mML-glutamine, 100 units/ml penicillin, and 100 μg/ml streptomycin in a humidified atmosphere of 5% CO2. Synovial fibroblasts at passages 4–6 were used in our study.
Preparation of simvastatin.
The active form of simvastatin was prepared according to the protocol described previously (6). Briefly, 25 mg simvastatin (Merck) was dissolved in 0.2 ml ethanol (95–100%) followed by the addition of 0.3 ml NaOH. After heating at 50°C for 2 hours, the solution was neutralized with 1N HCl to pH 7.2 and brought to a 1 ml volume with normal saline. The final concentration of the stock solution was 4 mg/ml.
Nuclear/cytoplasmic fractionation and Western blot analysis.
Fractionation of nuclear and cytoplasmic proteins was carried out with the ProteoExtract Subcellular Proteotome Extraction kit (Calbiochem) according to the manufacturer's instructions. Western blot analysis was performed as described previously (28). Data were quantified by densitometric analysis, with results expressed as the mean ± SD fold change relative to the values in untreated controls, from 3 independent experiments.
ELISA and immunoprecipitation.
The ELISA and immunoprecipitation experiments were performed in the RASFs as described previously (28).
Lentiviral-based gene transfection.
Forced expression of FoxO3a, wild-type SIRT-1, and a catalytically inactive SIRT-1 H363Y mutant in cultured RASFs was performed using a lentiviral-based technique. The Lenti-X system (Clontech) was used. Plasmids with cDNA constructs for FoxO3a (pBluescriptR-FoxO3a) and SIRT-1 (pCMV-SPORT6-SIRT-1) were obtained from Open Biosystems. The plasmid with a SIRT-1 mutant construct (pECE-flag-SIRT-1 H363Y) was purchased from Addgene. The inserts were amplified with polymerase chain reaction (PCR), and the PCR fragments were then ligated to the multiple cloning site of pLVX-AcGFR1-N1 (Clontech). Recombinant lentiviruses were produced by cotransfecting 293T cells with lentiviral vector DNA (pLVX-AcGFP-N1-FoxO3a, pLVX-AcGFP-N1-SIRT-1, or pLVX-AcGFP-N1-SIRT-1 H363Y) and Lenti-X HTX packaging mixes (expressing the Pol, Tat, Rev, Gag, and VSV-G envelope) using the Xfect transfection reagent (Clontech).
Cell culture supernatants were harvested at 48 hours after transfection and filtered through a 0.45-μm filter. Culture supernatants containing lentivirus were added to RASFs in the presence of polybrene. Twenty-four hours later, the efficiency of transduction was assessed by detecting green fluorescent protein expression under flow cytometry. Stably transduced cells were selected using puromycin.
The pGL2-Cyr61 P1-Luc construct containing a 972-bp fragment of the 5′-flanking region of the Cyr61 gene, starting from −169-bp upstream of the ATG start codon (A = +1) (29), was kindly provided by Dr. N. Schütze (Orthopedic Center for Musculoskeletal Research, University of Würzburg, Würzburg, Germany). A luciferase reporter assay for the analysis of Cyr61 promoter activity was performed as described previously (6).
Chromatin immunoprecipitation (ChIP) assay.
A ChIP assay was performed as described previously (6), using an assay kit from Upstate Biotechnology. The DNA–protein complexes from treated RASFs were precipitated using anti-FoxO3a antibodies, and immunocomplexes were collected. The bound DNA was recovered by phenol-chloroform-ethanol precipitation and used for quantitative PCR on a StepOne Plus apparatus with SYBR Green reagents (Applied Biosystems) to identify the region of human Cyr61 promoter flanking the FoxO3a binding site. The primer sequences were as follows: forward 5′-CCAACCAGCATTCCTGAGAT-3′ and reverse 5′-CGTATAAAAGGCGGGCTCC-3′.
Animal model of CIA.
CIA was induced in 20 male Sprague-Dawley rats weighing 220–250 grams, as described previously (6). The experimental protocol was approved by the Laboratory Animal Center at the College of Medicine of National Taiwan University, and the animals were maintained in accordance with the National Science Council of Taiwan's recommendations on the management and use of experimental animals. Bovine CII in acetic acid (4 mg/ml) was emulsified in an equal volume of Freund's complete adjuvant (CFA; containing 2 mg/ml of inactivated Mycobacterium tuberculosis in IFA).
On day 1, CIA was elicited by intradermal injections into 5 sites (2 on the tail base, 3 on the back) with a total of 500 μg bovine CII in 500 μl CFA. Booster injections were administered on day 7 into 3 sites (1 on the tail base, 2 on the back) with a total of 300 μg bovine CII in 300 μl IFA. Starting on day 0, rats were given intraarticular injections (0.5 ml/kg) of simvastatin (0.5 mg/ml) to the right ankle joint and normal saline to the left ankle joint (as control) every 5 days, until the animals were killed on day 21.
The rats were examined every day by an investigator (KLH) who was blinded with regard to the treatment protocols. The ankle joints were evaluated on a 0–4 scale, in which grade 0 = no swelling or erythema, grade 1 = slight swelling and/or erythema, grade 2 = low-to-moderate edema, grade 3 = pronounced edema with limited joint usage, and grade 4 = excess edema with joint rigidity (6).
The ankle joints were placed in position on Eastman Kodak X-OMAT TL high-resolution specimen-imaging film and radiographed with a Faxitron x-ray system (model 43855A). Images were shot at 26 kV for 10 seconds. Erosive changes were analyzed using a semiquantitative scale according to the degree of bony destruction/erosions (on a scale of 0–4), assigning 1 point for an erosion in the tibia, calcaneus, talus, and small tarsal bones (all considered together). For example, the maximum score that an ankle could receive was 4 if erosions were present in the tibia, calcaneus, talus, and any one or more of the small tarsal bones (6).
Histologic examination and immunohistochemistry.
The ankle joints were fixed, decalcified, embedded in paraffin, sectioned, and subjected to histopathologic examination. The histopathologic score was evaluated using a scale of severity ranging from 1 to 4, in which grade 1 = hyperplasia of the synovial membrane and presence of polymorphonuclear infiltrates, grade 2 = pannus and fibrous tissue formation and focal subchondral bone erosion, grade 3 = articular cartilage destruction and bone erosion, and grade 4 = extensive articular cartilage destruction and bone erosion (6).
Immunohistochemical staining of the joint tissue was performed with antibodies against rat CYR-61 and phospho-FoxO3a (Ser253). For identification of cell types in the serial sections, an anti-CD55 antibody (synovial fibroblast marker) was used. For each animal, quantitative analysis was performed on the 3 sections with the strongest inflammatory reactions. The field in each section exhibiting the highest inflammatory activity was selected and examined microscopically at 400× magnification. The total number of CD55+ RASFs and the number of CYR-61+ and phospho-FoxO3a+ RASFs were counted. The data were converted to the percentage of marker-positive RASFs.
Data were subjected to analysis of variance for multiple comparisons, and then to Fisher's protected least significant difference test. P values less than 0.05 were considered significant.
Suppression of TNFα-induced CYR-61 and CCL20 production in RASFs by simvastatin.
Western blot analysis showed that TNFα stimulated CYR-61 expression in RASFs (Figure 1A). The stimulating effect was time-dependent and peaked at 24 hours. Since CCL20 is an important inflammatory effector produced by RASFs (9), we also examined the effects of TNFα and CYR-61 on CCL20 synthesis. Assessment of the RASF culture medium by ELISA revealed that CCL20 secretion was enhanced after TNFα stimulation of RASFs (Figure 1B). Treatment of the cells with simvastatin attenuated the TNFα-induced production of CYR-61 and CCL20 in RASFs (Figures 1C and D).
Exogenous CYR-61 also stimulated CCL20 production, and the combination of TNFα and CYR-61 had an additive effect (Figure 1D). Simvastatin treatment inhibited the CCL20-inductive effects of CYR-61 and TNFα (Figure 1D).
Decreased nuclear expression of FoxO3a after TNFα stimulation, and reversal of the CYR-61–inductive effect of TNFα following forced expression of FoxO3a in RASFs.
To examine the role of FoxO3a in CYR-61 production in RASFs, we assessed the localization of FoxO3a after TNFα treatment and the effect of its expression on CYR-61 synthesis. In untreated RASFs, prominent expression of FoxO3a in the nuclei was noted. Upon TNFα stimulation, FoxO3a translocated from the nucleus to the cytoplasm within 60 minutes, and nuclear expression of FoxO3a remained low after 120 minutes of TNFα treatment (Figures 2A and B).
Forced expression of FoxO3a in RASFs was performed by lentiviral-mediated gene transfer, and Western blots confirmed the induced overexpression of FoxO3a in the nuclei (Figure 2C). More importantly, although a nucleus-to-cytoplasm redistribution of the overexpressed FoxO3a was noted after TNFα stimulation, nuclear levels of the transcription factor remained relatively high in the transduced cells (Figures 2C and D). Correspondingly, RASFs with forced expression of FoxO3a were resistant to the inductive effect of TNFα on CYR-61 production (Figure 2E). Results of luciferase reporter assay confirmed that overexpression of FoxO3a in the cells inhibited the TNFα-stimulated activation of the Cyr61 promoter (Figure 2F).
Inhibition of nuclear export, phosphorylation, and acetylation of FoxO3a and maintenance of FoxO3a binding to the Cyr61 promoter by simvastatin.
To investigate the effects of simvastatin on the activities of FoxO3a, localization of the protein after drug treatment was assessed first. Western blot analysis demonstrated that, in a dose-dependent manner, simvastatin maintained the nuclear localization of FoxO3a after TNFα treatment (Figure 3A). At the same time, simvastatin inhibited the Ser253-phosphorylation (Figure 3B) and lysine-acetylation (Figure 3C) of FoxO3a that was induced by TNFα. Furthermore, results of ChIP assay showed that TNFα treatment of RASFs resulted in decreased FoxO3a binding to the Cyr61 promoter, and treatment with simvastatin restored the occupancy of FoxO3a on the promoter (Figure 3D).
Involvement of SIRT-1 in decreasing CYR-61 expression and deacetylating FoxO3a.
Since FoxO3a is a substrate of the deacetylase SIRT-1, and the acetylation status of the protein plays a key role in its function (18), we next examined the involvement of SIRT-1 in the modulating effect of FoxO3a on CYR-61 expression. Forced expression of SIRT-1 in RASFs was performed by lentiviral-mediated gene transfer (Figure 4A). Western blot analysis revealed that overexpression of SIRT-1 suppressed the TNFα-induced CYR-61 synthesis in RASFs. Moreover, the inhibitory effects of SIRT-1 and simvastatin on CYR-61 production were additive (Figure 4B).
In addition, Western blot analysis of anti-FoxO3a immunoprecipitates using anti–acetylated lysine antibodies demonstrated that overexpression of SIRT-1 decreased the TNFα-induced acetylation of FoxO3a (Figure 4C). Results of luciferase reporter assay confirmed that forced expression of SIRT-1 in the cells suppressed the TNFα-enhanced Cyr61 promoter activity (Figure 4D). The inhibitory effect of SIRT-1 seemed stronger than that of FoxO3a. Compared to SIRT-1 alone, the forced expression of FoxO3a and SIRT-1 together had little additive effect on promoter suppression (Figure 4D). To assess the significance of deacetylase activity in the suppressive action of SIRT-1 on Cyr61 transcription, forced expression of a catalytically inactive SIRT-1 mutant was performed. These experiments showed that the SIRT-1 mutant had little effect on the promoter activity of Cyr61 (Figure 4D).
Up-regulation of SIRT-1 and SIRT-1/FoxO3a binding in RASFs by simvastatin.
We next examined the effect of simvastatin on SIRT-1 expression. Western blot analysis showed that simvastatin enhanced SIRT-1 expression in RASFs, whereas TNFα treatment for 24 hours had little effect on the expression of SIRT-1 (Figure 4E). Results of immunoprecipitation assay revealed that simvastatin also increased the binding between SIRT-1 and FoxO3a (Figure 4F).
In vivo suppression of CYR-61 expression and FoxO3a phosphorylation in rat synovial fibroblasts after treatment with simvastatin.
To examine the relationship between FoxO3a phosphorylation and CYR-61 expression in synovial fibroblasts in vivo, a rat CIA model was used. The experiment confirmed the protective effect of intraarticular delivery of simvastatin on inflammatory arthritis (6). At the end of the experiment, treatment with simvastatin had significantly reduced the arthritis inflammation index (mean ± SD 2.2 ± 0.2 with treatment versus 3.6 ± 0.4 without treatment; P < 0.05). Semiquantitative analysis of arthritis progression by radiography revealed that the radiographic scores of joint damage were elevated in the left ankles (untreated control joints), whereas administration of simvastatin in the right ankles reduced radiographic bone destruction (mean ± SD radiographic score 3.4 ± 0.3 in the left ankles versus 1.6 ± 0.2 in the right ankles; P < 0.05).
Histopathologic examination of the ankle joints revealed irregular bone resorption associated with extensive pannus formation in the control joints (Figure 5A). Immunohistochemical staining of the joint tissue demonstrated marked CYR-61 expression in synovial fibroblasts (Figure 5B). Strong staining for phospho-FoxO3a in synovial fibroblasts was also noted (Figure 5C).
In contrast, intraarticular injection of simvastatin markedly diminished the severity of joint inflammation and bone destruction in the joints of rats with CIA (Figure 5D). More importantly, simvastatin reduced the number of CYR-61+ synovial fibroblasts (Figure 5E) and phospho-FoxO3a+ synovial fibroblasts (Figure 5F). The histopathologic score (mean ± SD) was 3.8 ± 0.4 in the control joints and 1.9 ± 0.4 in the simvastatin-treated joints (P < 0.05) (Figure 5G). Significant differences in the percentage of CYR-61+ synovial fibroblasts (mean ± SD 60.1 ± 9.6% versus 22.7 ± 7.2%; P < 0.05) (Figure 5H) and phospho-FoxO3a+ synovial fibroblasts (55.2 ± 8.5% versus 20.7 ± 5.5%; P < 0.05) (Figure 5I) were found between the control and simvastatin-treated joints in this experimental CIA model.
The present study demonstrated that CYR-61 is an important effector molecule in RA pathogenesis. In cultured RASFs, TNFα enhanced CYR-61 synthesis, and the two molecules additively induced CCL20. Our experiments further showed that simvastatin has a therapeutic effect by reducing the expression of CYR-61 and CCL20 in RASFs. The transcription factor FoxO3a plays a major role in the inhibitory action of simvastatin on CYR-61 induction by TNFα. Based on our findings in a rat CIA model, the critical role of FoxO3a/CYR-61 signaling in the therapeutic effect of simvastatin was confirmed.
Previous studies have shown that CYR-61 contributes to the pathogenesis of RA by mediating IL-17–induced proliferation of RASFs (7) and stimulating the differentiation of IL-17–producing Th17 cells (8). Th17 cells are now believed to play a crucial role in the pathogenesis of RA (30, 31), and CCL20 is required for the migration of Th17 cells to initiate self-destructive immune reactions in RA (9). Once synovial inflammation occurs, RASFs may further recruit Th17 cells through CCL20 production (9). In our study, we demonstrated, for the first time, that CYR-61 up-regulated CCL20 secretion by RASFs, and an additive effect was observed when cells were exposed to both TNFα and CYR-61. Since CYR-61 plays pivotal roles in both cell proliferation and chemokine secretion by RASFs, it may serve as a therapeutic target in the treatment of RA.
Jonsson et al showed that FoxO3a is required for the survival of neutrophils in inflammatory arthritis (12), but its function in RASFs has not been studied. In our experiments, we found that FoxO3a is a negative regulator of CYR-61 expression in RASFs. TNFα treatment abrogated the function of FoxO3a by inducing its nuclear export. Forced expression of FoxO3a inhibited TNFα-enhanced synthesis of CYR-61 and suppressed the promoter activity of the Cyr61 gene.
The activity of FoxO proteins is tightly regulated by phosphorylation and acetylation (16, 17). Akt-dependent phosphorylation of the transcription factors inhibits their DNA-binding activity and results in nuclear export (16, 17). TNFα may inhibit the function of FoxO3a in RASFs through the activation of Akt, since previous studies (32, 33) and our experiments (results not shown) revealed that TNFα is an activator of Akt in RASFs. In support of this notion, we found that TNFα induced phosphorylation of FoxO3a at the Akt-specific site Ser253 (34).
Besides Akt-dependent phosphorylation, other mechanisms may also be involved in TNFα-induced suppression of FoxO3a in RASFs, since we observed that TNFα enhanced acetylation of the transcription factor. Acetylation by CREB binding protein/p300 has been identified as an important regulatory pathway for the function of FoxO proteins (35). Although the exact consequences of acetylation remain unclear, one study demonstrated that acetylation of FoxO proteins attenuates their transcriptional activity (35). In our study, we found that FoxO3a acetylation was associated with a decrease in its repressive action on Cyr61 transcription. TNFα has been shown to enhance the binding between p300 and other transcription factors and promote their acetylation (36, 37). Whether this pathway is involved in TNFα-induced acetylation of FoxO3a remains to be determined.
Recently, the therapeutic effects of statins against RA progression have drawn the attention of many researchers. In our study, we confirmed that simvastatin is beneficial to RA by its ability to maintain the repressive action of FoxO3a on Cyr61 transcription in RASFs. In TNFα-treated RASFs, simvastatin kept FoxO3a in the nucleus and retained its binding to the Cyr61 gene promoter. The effect of simvastatin on FoxO3a may derive from its inhibitory action of this treatment on Akt and the enhancement of SIRT-1 activity. In RASFs, it was reported that statins suppressed the TNFα-induced activation of Akt (38), and we observed a similar effect on RASFs after treatment with simvastatin (results not shown). We also showed that simvastatin treatment resulted in a decrease in Akt-specific Ser253-phosphorylation of FoxO3a. On the other hand, the other pathway through which simvastatin influences FoxO3a activity is via deacetylation of the protein. For the first time, we demonstrated that simvastatin decreased the acetylation of FoxO3a in RASFs and maintained its repressive action on gene expression. It is noteworthy that acetylation/deacetylation of FoxO proteins may also influence the sensitivity for phosphorylation. It was shown that, in addition to an inhibitory effect on DNA binding, acetylation of FoxO1 caused an increase in Akt-mediated phosphorylation, leading to the subsequent nuclear export of the transcription factor (39).
The mechanism by which simvastatin promotes deacetylation of FoxO3a appears to involve SIRT-1. SIRT-1 is a key regulator of the transcriptional activity of FoxO3a (19), and therefore its function in the modulating effect of FoxO3a on CYR-61 synthesis and its relationship to the action of simvastatin were examined. In our study, we documented that SIRT-1 suppressed TNFα-induced CYR-61 expression in RASFs, possibly through deacetylation of FoxO3a. It is obvious that the deacetylase activity of SIRT-1 is pivotal to its action, since forced expression of a catalytically inactive SIRT-1 mutant instead of the wild-type protein had little effect on the promoter activity of the Cyr61 gene. More importantly, we further observed that simvastatin enhanced the level of SIRT-1 in RASFs and promoted SIRT-1/FoxO3a binding. Our results indicate that SIRT-1/FoxO3a signaling plays a critical role in the inhibitory effect of simvastatin on CYR-61 expression in RASFs.
Sirtuins have attracted much attention recently because of their role in lifespan regulation in lower organisms and their capacity to regulate mammalian cell growth and survival in response to stress. The identification of several transcription factors known to play a role in the immune system as substrates of sirtuins has suggested that they may also participate in the regulation of inflammation, but a conclusive role for sirtuins as pro- or antiinflammatory regulators is still a matter of debate (40). In the case of arthritis, Niederer et al (41) found that persistent overexpression of SIRT-1 inhibited apoptosis and promoted proinflammatory cytokine production in RASFs. In contrast, Nakayama et al (42) demonstrated that resveratrol-induced apoptosis of synovial cells was sirtuin dependent. Moreover, SIRT-1 was shown to have the ability to inhibit osteoclastogenesis (43) and enhance the survival of osteoarthritic chondrocytes (44). In the present study, we also found that SIRT-1 is antiinflammatory, through its suppression of CYR-61 production in RASFs. Obviously, careful investigation of the cell- and disease-specific effects of sirtuins is needed before clinically relevant applications of agents targeting these molecules could be envisioned.
Our experiments with the rat CIA model confirmed the importance of CYR-61 expression and FoxO3a signaling in the pathogenic mechanism of RASFs. Overexpression of CYR-61 has been demonstrated in synovial fibroblasts from RA patients (7), but a relationship between expression levels of CYR-61 and the extent of disease activity in arthritis was not reported. In our study, we showed that alleviation of arthritis in a rat CIA model by intraarticular delivery of simvastatin was accompanied by a reduction in CYR-61 expression in synovial fibroblasts.
With regard to the relationship between FoxO3a expression and disease activity in inflammatory arthritis, previous studies demonstrated that FoxO3a may favor the survival of neutrophils in the blood (12, 15) and T cells in the synovium from RA patients (13, 15), and high phosphorylation levels of FoxO3a were found in synovial T cells from patients displaying low disease activity (13). In contrast, we observed increased accumulation of inactive phospho-FoxO3a in the synovial fibroblasts from rat ankle joints with a higher arthritis inflammation index. It is obvious that the exact mechanisms by which FoxO3a is involved in RA are complex and the transcription factor may exert disparate functions in different cell types important for the pathogenesis of RA.
In conclusion, we demonstrated that CYR-61 is important in the effector mechanisms of RASFs and confirmed the beneficial effect of simvastatin on inflammatory arthritis. Our study is the first to reveal that SIRT-1/FoxO3a signaling is crucial to the induction of CYR-61 in RASFs and that simvastatin has a modulating effect on this pathway. Continued study of the pathways linking sirtuins, FoxO proteins, and inflammatory responses of RASFs may provide new insights into the pathophysiology of RA, with possible therapeutic applications.
All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. Dr. S.-K. Lin had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study conception and design. Kok, L.-D. Lin, Hou, Hong, Chang, Hsiao, S.-K. Lin.