SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES

Objective

To investigate whether interleukin-6 (IL-6) is a regulator of vascular endothelial growth factor (VEGF) in rheumatoid arthritis (RA).

Methods

Serum VEGF levels in RA patients were assayed before and after 8 weeks or 24 weeks of maintenance therapy with humanized anti–IL-6 receptor monoclonal antibody (anti–IL-6R mAb). VEGF secreted by RA synovial fibroblasts cultured in the presence of IL-6, IL-1β, and/or tumor necrosis factor α (TNFα) was measured. The inhibitory effect of anti–IL-6R mAb, recombinant IL-1 receptor antagonist (IL-1Ra), and anti-TNFα mAb on VEGF production was also examined.

Results

Serum VEGF levels in RA patients before anti–IL-6R mAb therapy were significantly higher than those in healthy controls (P < 0.0005). Treatment of RA patients with anti–IL-6R mAb normalized serum VEGF levels. In the in vitro study, IL-6 and IL-1β each induced a slight amount of VEGF production in synovial cells, but TNFα did not. Although VEGF-inducing activity of these cytokines was not remarkable when they were added alone, IL-6 acted synergistically with IL-1β or TNFα to induce VEGF production. There was no synergistic effect between IL-1β and TNFα. In the presence of all of these cytokines, anti–IL-6R mAb eliminated the synergistic effect of IL-6, IL-1β, and TNFα, while IL-1Ra or anti-TNFα mAb did not.

Conclusion

Anti–IL-6R mAb therapy reduced VEGF production in RA. IL-6 is the pivotal cytokine that induces VEGF production in synergy with IL-1β or TNFα, and this may be the mechanism by which IL-6 blockade effectively suppresses VEGF production in synovial fibroblasts.

Rheumatoid arthritis (RA) is an autoimmune inflammatory disease characterized by the presence of rheumatoid factor (RF) and destruction of bone and cartilage in multiple joints. Pathologic features of the affected joints include hyperplasia of synovial tissue, with increased angiogenesis necessary to oxygenate the tissue (1, 2).

Angiogenesis consists of multiple processes such as degradation of vascular basement membranes and surrounding extracellular matrix, and migration and proliferation of endothelial cells (3). Several mediators have been reported as angiogenic factors that are important for these processes. Vascular endothelial growth factor (VEGF) is a potent angiogenic factor which promotes migration and proliferation of endothelial cells (4, 5). VEGF also induces vascular permeability (4, 5) and mediates inflammation (6).

There is a great deal of evidence suggesting that VEGF plays an important role in the pathogenesis of RA. VEGF is produced by macrophages, fibroblasts surrounding microvessels, vascular smooth muscle cells, synovial lining cells in synovium (7), neutrophils in synovial fluid (8), and peripheral blood mononuclear cells (PBMCs) (9) from patients with RA. VEGF levels are significantly higher in synovial fluids from patients with RA than in those from patients with osteoarthritis (OA) or other arthritides (6, 10, 11). In addition, serum VEGF levels correlate with disease activity scores and radiologic progression in patients with RA (12). However, the mechanism of VEGF production in RA is not fully understood. Hypoxia and interleukin-1 (IL-1) have been reported to induce VEGF production in synovial fibroblasts (13, 14). Tumor necrosis factor α (TNFα) up-regulates VEGF production in PBMCs in RA patients (9). Indeed, Paleolog et al reported that therapy with the anti-TNFα monoclonal antibody (mAb) cA2 (infliximab) not only improved clinical symptoms of RA, but also reduced serum VEGF levels in RA patients (15). They also showed that the combination of anti-TNFα mAb and IL-1 receptor antagonist (IL-1Ra) suppressed VEGF production from RA synovial membrane cells in vitro (15). Therefore, TNFα and IL-1 are believed to be capable of inducing VEGF production in RA.

IL-6 is a pleiotropic cytokine with various biologic activities, such as induction of acute-phase reaction, regulation of immune response, and promotion of hematopoiesis (16). Furthermore, IL-6 activates osteoclasts in the presence of soluble IL-6 receptor (sIL-6R) (17). Therefore, overproduction of IL-6 may be involved in the appearance of RF, elevation of acute-phase proteins, hypergammaglobulinemia, thrombocytosis, and destruction of the joints in patients with RA (18). Indeed, IL-6 levels both in synovial fluid and in serum from RA patients have been reported to be higher than those in patients with other arthritides (19, 20). We have reported that treatment of refractory RA with MRA, a humanized anti–IL-6R mAb, resulted in improvement of both clinical symptoms and laboratory findings (18). Thus, IL-6 plays an important role in the pathogenesis of RA. Moreover, it has been reported that IL-6 induces VEGF production in human myeloma cells (21), epidermoid carcinoma cells, rat skeletal muscle myoblasts, and rat glioma cells (22) in vitro. We have also demonstrated that IL-6 blockade reduced hyaline vascular vessels in the affected lymph node in a patient with Castleman's disease (23), in which VEGF is produced by infiltrating plasma cells (24). This evidence suggests that IL-6 may regulate VEGF production in RA.

To investigate whether IL-6 is indeed a regulator of VEGF production in RA, we analyzed serum VEGF levels before and after anti–IL-6R mAb therapy in RA patients. We also examined the ability of IL-6, acting synergistically with IL-1β and/or TNFα, to induce VEGF production in RA synovial fibroblasts in vitro.

PATIENTS AND METHODS

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES

Patients and humanized anti–IL-6R mAb.

In a previous study (18), 11 patients with RA that had been resistant to conventional treatment with various disease-modifying antirheumatic drugs, including methotrexate and corticosteroids, were treated with humanized anti–IL-6R mAb at Osaka University Hospital. All patients gave informed consent for anti–IL-6R mAb therapy under the auspices of and approval by the ethics committee of Osaka University. Humanized anti–IL-6R mAb (IgG1 class) (MRA) was produced in Chinese hamster ovary cells and purified on protein A (25). The appropriate amount of mAb was diluted to a volume of 50 ml in saline and administered intravenously over a period of 1 hour. Increasing doses (1 mg, 10 mg, 50 mg, and 100 mg/injection) of the mAb were administered intravenously twice weekly to establish the maximal efficacy in each patient.

All of the patients had active disease as defined by the presence of at least 10 swollen joints and at least 10 tender joints, an erythrocyte sedimentation rate (ESR) >40 mm/hour, a serum C-reactive protein (CRP) concentration of >3.0 mg/dl, a hemoglobin level of <10 gm/dl, and a ferritin level of >100 ng/ml. Patients could continue treatment with prednisolone at ≤15 mg/day; the dosage of prednisolone had to remain stable or decrease during the study period. Pregnant women, nursing women, and women of childbearing potential not using an effective method of contraception were excluded from the study.

Of the 11 patients in the previous study (18), 3 withdrew within the first few weeks (1 because of the development of antiidiotypic antibodies, 1 because of angina, and 1 at the patient's request). The remaining 8 patients received injections of 50–100 mg anti–IL-6R mAb once or twice weekly for 8 weeks and were analyzed for disease activity and serum VEGF levels. Four patients received the treatment for an additional 16 weeks (24 weeks total). Neither anti-DNA antibodies nor antinuclear antibodies developed in any patient, and both clinical and laboratory improvements were observed. After 8 weeks of treatment, clinical response was seen in 7 of 8 patients (88%) as assessed by the American College of Rheumatology 20% improvement criteria (ACR20) (26), and in 4 of 8 (50%) by the ACR50 (27). After 24 weeks of treatment, response was observed in 4 of 4 patients (100%) by the ACR20 and in 2 of 4 (50%) by the ACR50 (18). Serum samples were obtained before and after 8 and 24 weeks of maintenance therapy with anti–IL-6R mAb. As a control for baseline VEGF levels, serum samples were obtained from 32 healthy individuals who had provided informed consent. All serum samples were stored at −20°C until VEGF measurement. Sera from all patients and controls in the earlier study (18) were used to investigate the effects of anti–IL-6R mAb on VEGF in the present study.

Cytokines and antibodies.

Recombinant human IL-6 was provided by Ajinomoto (Kawasaki, Japan), recombinant sIL-6R by Tosoh (Kanagawa, Japan), and humanized anti–IL-6R mAb by Chugai Pharmaceutical (Tokyo, Japan). Recombinant human IL-1β was purchased from BioSource International (Camarillo, CA) and recombinant human TNFα from PeproTech EC (London, UK). Recombinant human IL-1Ra (catalog no. 280-RA) and anti-human TNFα mAb (catalog no. MAB210) were purchased from R&D Systems (Minneapolis, MN).

Cell preparation and cytokine stimulation.

Synovial fibroblasts were isolated from the synovial tissues of 3 RA patients and 1 OA patient, with their informed consent and in accordance with the guidelines of our institution's ethics committee. Synovial fibroblasts were prepared from synovial tissue as previously described (28). These synovial fibroblasts showed TNFα-dependent proliferation and IL-6 production (28). Normal human embryonic lung fibroblasts (HELs) were kindly provided by Dr. K. Kondo (Osaka University Medical School). Cells were cultured with Dulbecco's modified Eagle's medium containing 10% fetal calf serum (FCS), 100 μg/ml streptomycin, and 100 units/ml penicillin. We used synovial fibroblasts that had been passaged 5–9 times and HELs passaged 14 times for this experiment. All of the in vitro experiments were performed in triplicate. Cells (2 × 104/500 μl/well) were cultured in the presence or absence of IL-6 (1–100 ng/ml), sIL-6R (100 ng/ml), IL-1β (0.05–5 ng/ml), and TNFα (0.1–10 ng/ml) in 48-well flat-bottomed culture plates (Corning, Corning, NY) for 24–72 hours. The concentrations of IL-6, sIL-6R, IL-1β, and TNFα used in this experiment were comparable with those observed in synovial fluid from patients with active RA (29–31).

The IL-6 receptor system consists of a ligand binding receptor (IL-6R) and its signal transducer (gp130) on the cell surface (32). Soluble IL-6R has been identified in vivo, and the complex of IL-6 and sIL-6R can induce homodimerization of gp130 and mediate the signal in cells that do not express enough IL-6R. Since synovial fibroblasts express little IL-6R on their cell surface, cells were stimulated with IL-6 in the presence of sIL-6R in these in vitro experiments, as previously reported (28). The culture supernatants were collected and analyzed for VEGF production.

To study the effect of cytokine inhibitors, anti–IL-6R mAb (25 μg/ml), IL-1Ra (10 μg/ml), or anti-TNFα mAb (10 μg/ml) was added to the culture 30 minutes prior to cytokine stimulation. Anti–IL-6R mAb at 25 μg/ml is capable of inhibiting the action of IL-6 (100 ng/ml) with sIL-6R (100 ng/ml) on synovial fibroblasts in vitro (28). IL-1Ra at 10 μg/ml (2,000-fold in excess of the amount of IL-1β) is a high enough concentration to inhibit biologic response of 5 ng/ml of IL-1β (33). Anti-TNFα mAb at 10 μg/ml, administered according to the instructions of the manufacturer (R&D Systems) can inhibit the biologic response of 10 ng/ml of TNFα. To investigate the expression of VEGF messenger RNA (mRNA) in RA synovial fibroblasts, cells (1 × 106/10 ml/dish) were cultured with IL-6 (100 ng/ml)/sIL-6R (100 ng/ml), and a combination of IL-6/sIL-6R and IL-1β (5 ng/ml) for 48 hours in 100-mm culture dishes (Corning).

Enzyme-linked immunosorbent assay (ELISA) for VEGF.

VEGF levels in the culture supernatants of synovial fibroblasts and in the serum samples were determined by ELISA (R&D Systems). The VEGF immunoassay is designed to measure the VEGF165 and VEGF121 isoforms. All assays and calibrations were performed in duplicate.

Quantitative real-time reverse transcriptase–polymerase chain reaction (RT-PCR) for VEGF mRNA and for IL-6R mRNA.

Total RNA was isolated from cytokine-stimulated synovial fibroblasts with an RNeasy mini-kit (Qiagen, Hilden, Germany). Complementary DNA (cDNA) was synthesized for 1 hour at 37°C using 3 μg of total RNA, an oligo-dT primer, and Moloney murine leukemia virus RT (Promega, Madison, WI). The primers and probes used to quantify VEGF and GAPDH were as follows: for VEGF, forward primer 5′-GCACCCATGGCAGAAGG-3′ (in exon 2), reverse primer 5′-CTCGATTGGATGGCAGTAGCT-3′ (in exon 3), probe 5′-(FAM reporter dye)-ACGAAGTGGTGAAGTTCATGGATGTCTATCAC-(TAMRA quencher dye)-3′ (spanning exon 2–exon 3 junction); for GAPDH, forward primer 5′-GAAGGTGAAGGTCGGAGTC-3′ (6–24 nucleotides of cDNA), reverse primer 5′-GAAGATGGTGATGGGATTTC-3′ (212–231 nucleotides of cDNA), probe 5′-(JOE reporter dye)-CAAGCTTCCCGTTCTCAGCC-(TAMRA quencher dye)-3′ (183–202 nucleotides of cDNA).

Quantitative real-time RT-PCR was carried out in duplicate with the aid of the TaqMan Universal PCR Master Mix kit and the PE Biosystem 5700 sequence detector, according to the protocol recommended by the manufacturer (Applied Biosystems, Foster City, CA). Briefly, a reaction volume of 50 μl contained 25 μl of 2× master buffer, 15 pmoles of each primer, 10 pmoles of the dual-labeled probe, and 0.3 μg cDNA for each tested sample. Thermal cycling was initiated with 2 minutes of incubation at 50°C for activating uracil N-glycosylase, followed by 10 minutes of denaturation at 95°C for activating AmpliTaq Gold (Applied Biosystems), and 2-step thermocycling for 40 cycles (15 seconds at 95°C, and 1 minute at 60°C). During each PCR cycle, the amount of fluorescence that occurred when a fluorogenic oligonucleotide probe was activated by activity of Taq polymerase after binding to a specific PCR product was monitored (TaqMan PCR). Ct values (the number of PCR cycles needed to reach threshold fluorescence) were then determined for a series of standards, and a standard curve was generated by plotting Ct versus the log of the amount of VEGF cDNA added for the reaction. This curve was then used to compare the relative amount of VEGF mRNA in the samples from control and cytokine-treated cells. The values were normalized to GAPDH. RT-PCR for IL-6R expression in RA synovial fibroblasts was performed as described previously (28).

Statistical analysis.

The Mann-Whitney U test was used for statistical analysis of the comparison of serum VEGF levels between RA patients and healthy subjects. Wilcoxon's signed rank test was used for analysis of the comparison of serum VEGF levels before versus 8 weeks after the initiation of anti–IL-6R mAb therapy. Student's t-test was used for analysis of VEGF production from synovial fibroblasts.

RESULTS

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES

Serum VEGF levels in RA patients and the effect of anti–IL-6R mAb therapy.

Before the initiation of anti–IL-6R mAb therapy, serum VEGF levels were significantly higher in RA patients (n = 8) than in healthy controls (n = 32) (mean ± SD 568 ± 240 pg/ml versus 189 ± 130 pg/ml; P < 0.0005) (Figure 1A). Anti–IL-6R mAb therapy reduced the serum VEGF levels in these patients (Figure 1B). After 8 weeks of anti–IL-6R mAb therapy, the mean ± SD serum VEGF level in the RA patients was 228 ± 102 pg/ml, which was significantly lower than the level before treatment (P < 0.05). At 24 weeks, serum VEGF levels in the RA patients were normal (199 ± 24 pg/ml; n = 4) (Figure 1B). In these patients, CRP levels and ESR were decreased to near-normal by anti–IL-6R mAb administration (before treatment CRP 6.1 ± 3.3 mg/dl, ESR 85 ± 28 mm/hour; after 8 weeks of treatment CRP 0.5 ± 0.5 mg/dl, ESR 15 ± 8 mm/hour; n = 8). Painful/tender joint scores (maximum possible 49) and swollen joint scores (maximum possible 46) were also reduced (before treatment painful/tender joint score 21.0 ± 12.5, swollen joint score 13.9 ± 5.0; after 8 weeks of treatment painful/tender joint score 6.6 ± 6.9, swollen joint score 6.0 ± 2.4). As noted above, 7 of 8 patients achieved a response according to the ACR20 and 4 achieved a response according to the ACR50 after 8 weeks of treatment (18).

thumbnail image

Figure 1. Serum vascular endothelial growth factor (VEGF) levels measured by enzyme-linked immunosorbent assay. A, Box plots showing the levels of serum VEGF in 8 patients with rheumatoid arthritis (RA) and 32 healthy controls. Boxes show the 25th and 75th percentiles; horizontal lines within boxes show the median; vertical bars above and below boxes show the 10th and 90th percentiles; open circles show outlying values. P value was determined by Mann-Whitney U test. B, Serum VEGF levels in 8 RA patients before therapy with humanized anti–interleukin-6 receptor monoclonal antibody (anti–IL-6R mAb) and after 8 weeks and 24 weeks of therapy. VEGF levels normalized with anti–IL-6R mAb therapy.

Download figure to PowerPoint

Induction of VEGF production in RA synovial fibroblasts by proinflammatory cytokines.

To investigate whether IL-6 induces VEGF production in RA synovial fibroblasts, we analyzed VEGF in the supernatant of RA synovial fibroblasts cultured for 72 hours with IL-6 and sIL-6R (Figure 2). With no cytokine stimulation, RA synovial fibroblasts spontaneously secreted VEGF into the culture supernatant (mean ± SD 478 ± 16 pg/ml). IL-6 (1–100 ng/ml) in the presence of sIL-6R (100 ng/ml) induced VEGF production in RA synovial cells in a dose-dependent manner. With 100 ng/ml of IL-6, the increase in VEGF production (to 643 ± 23 pg/ml) was significant (P < 0.001 versus no cytokine), but the level of VEGF activity was not markedly high (Figure 2). Similarly, we examined the ability of TNFα and IL-1β to induce VEGF production (Figure 2). IL-1β induced VEGF production (with IL-1β at 5 ng/ml, 599 ± 43 pg/ml VEGF; P < 0.05 versus no cytokine), while TNFα did not (with TNFα at 10 ng/ml, 447 ± 36 pg/ml VEGF; P = 0.248 versus no cytokine). These data indicate that stimulation with IL-6 or IL-1β singly does not produce a major effect on VEGF production in synovial fibroblasts.

thumbnail image

Figure 2. Induction of vascular endothelial growth factor (VEGF) production in rheumatoid arthritis (RA) synovial fibroblasts by proinflammatory cytokines. RA synovial fibroblasts (from patient RA1 in Table 1) were cultured for 72 hours with various concentrations of interleukin-6 (IL-6; 1–100 ng/ml)/soluble IL-6 receptor (100 ng/ml), IL-1β (0.05–5 ng/ml), and tumor necrosis factor α (TNFα; 0.1–10 ng/ml). Release of VEGF into culture supernatants was measured by enzyme-linked immunosorbent assay. Values are the mean and SD of triplicate experiments under each condition. ∗ = P < 0.001; ∗∗ = P < 0.05 versus no cytokine.

Download figure to PowerPoint

Table 1. VEGF production in RA synovial fibroblasts and other fibroblasts treated with IL-6, IL-1β, and TNFα*
FibroblastsTreatment
NoneIL-6IL-1βTNFαIL-6 + IL-1βIL-6 + TNFαIL-1β + TNFαIL-6 + IL-1β + TNFαAnti–IL-6R mAbIL-1RaAnti-TNFα
  • *

    Cells were cultured for 72 hours, and vascular endothelial growth factor (VEGF) in the supernatant was assayed. Interleukin-6 (IL-6) was used at a concentration of 100 ng/ml, IL-1β at 5 ng/ml, tumor necrosis factor α (TNFα) at 10 ng/ml, anti–IL-6 receptor monoclonal antibody (anti–IL-6R mAb) at 25 μg/ml, IL-1 receptor antagonist (IL-1Ra) at 10 μg/ml, and anti-TNFα at 10 μg/ml. Values are pg/ml of VEGF (mean from triplicate wells). RA = rheumatoid arthritis; OA = osteoarthritis; HEL = normal human embryonic lung fibroblasts; ND = not done.

  • Each cytokine inhibitor was used in the presence of the triple combination of IL-6, IL-1β, and TNFα.

  • VEGF production in RA1 cells cultured in the presence of 10% fetal calf serum (RA1FCS+) and cells cultured in the absence of 10% FCS (RA1FCS−) was examined separately. The viability of RA1FCS− was the same as that of RA1FCS+.

RA14786436474981,5271,0457031,4897541,4981,440
RA21,1662,1121,5531,2873,5442,2451,2412,4601,3312,3112,878
RA35296691,1139421,8961,6291,0401,6291,0791,8311,684
OA3234335744041,041664485727513863825
HELs8742,1273,9921,3717,9971,9284,8508,9285,0623,3658,914
RA1FCS+9151,8721,1719522,7012,5459671,713NDNDND
RA1FCS−218450556386777719415490NDNDND

In order to test for synergistic effects of IL-6, IL-1β, and TNFα on VEGF production in RA synovial fibroblasts, we stimulated the cells with various combinations of IL-6/sIL-6R (100 ng/ml, respectively), IL-1β (5 ng/ml), and TNFα (10 ng/ml) (optimal doses as defined in the study of single cytokine stimulation). VEGF accumulated in the supernatant during cultures of 24–72 hours (Figures 3A–C). IL-6 in synergy with IL-1β induced VEGF production in synovial fibroblasts. At 48 hours, the mean ± SD level of VEGF induced by IL-6 + IL-1β was 1,235 ± 90 pg/ml, while the levels induced by IL-6 alone and IL-1β alone were 636 ± 40 pg/ml and 540 ± 32 pg/ml, respectively (P = 0.0005, IL-6 + IL-1β versus IL-6; P = 0.0002, IL-6 + IL-1β versus IL-1β). At 72 hours, the levels of VEGF induced by IL-6 + IL-1β, IL-6 alone, and IL-1β alone were 2,092 ± 134 pg/ml, 788 ± 9 pg/ml, and 846 ± 59 pg/ml, respectively (P < 0.0001, IL-6 + IL-1β versus IL-6; P = 0.0001, IL-6 + IL-1β versus IL-1β) (Figure 3A).

thumbnail image

Figure 3. Synergistic effects of IL-6, IL-1β, and TNFα on VEGF production in RA synovial fibroblasts. RA synovial fibroblasts (from patient RA1 in Table 1) were cultured with various combinations of IL-6 (100 ng/ml)/soluble IL-6 receptor (sIL-6R; 100 ng/ml), IL-1β (5 ng/ml), and TNFα (10 ng/ml) for 24, 48, or 72 hours. VEGF levels in the culture supernatants were measured by enzyme-linked immunosorbent assay. A, No cytokine, IL-6/sIL-6R, IL-1β, and a combination of IL-6/sIL-6R and IL-1β. B, No cytokine, IL-6/sIL-6R, TNFα, and a combination of IL-6/sIL-6R and TNFα. C, No cytokine, IL-1β, TNFα, and a combination of IL-1β and TNFα. Values are the mean ± SD of triplicate experiments. ∗ = P = 0.0005; ∗∗ = P < 0.0001; † = P < 0.05; †† = P < 0.001 versus IL-6/sIL-6R alone, by Student's t-test. See Figure 2 for other definitions.

Download figure to PowerPoint

Similarly, a synergistic effect between IL-6 and TNFα on VEGF production was observed at 48 hours (IL-6 + TNFα 747 ± 18 pg/ml; P < 0.05 versus IL-6) and 72 hours (IL-6 + TNFα 1,218 ± 81.0 pg/ml; P < 0.001 versus IL-6) (Figure 3B). However, no synergistic effect between IL-1β and TNFα on VEGF production was observed during the 72-hour culture (Figure 3C). VEGF levels induced by the triple combination of IL-6, IL-1β, and TNFα were not higher than those induced by the double combination of IL-6 and IL-1β (Table 1).

Depletion of FCS from culture media reduced basal VEGF production to one-fourth of that observed in the presence of FCS, indicating that there may be some factors inducing VEGF production in FCS. However, even in the absence of FCS, a similar synergistic effect of IL-6 with either IL-1β or TNFα was observed (Table 1). Similar results were obtained in experiments using synovial fibroblasts from 2 other RA patients (Table 1). In addition, synovial fibroblasts from an OA patient and lung fibroblasts also showed a similar profile of VEGF production in response to the cytokines tested (Table 1). Thus, the phenomena observed were not specific to RA fibroblasts.

Effect of cytokine stimulation on the expression of VEGF mRNA as determined by real-time RT-PCR.

We used real-time RT-PCR to quantitate the effect of IL-6 (100 ng/ml)/sIL-6R (100 ng/ml) in the presence or absence of IL-1β (5 ng/ml) on VEGF mRNA expression (Figure 4). The number of PCR cycles needed to reach threshold reporter fluorescence intensity showed a linear relationship with the amount of VEGF plasmids added in the PCR (correlation coefficient [R2] >0.99 over a range of 1 × 10−7–1 × 10−2 μg/ml) (Figure 4A). On the basis of this standard, VEGF expression in RA synovial fibroblasts was determined. IL-6/sIL-6R induced VEGF mRNA expression by a factor of 1.54 compared with no cytokine, and the combination of IL-6/sIL-6R and IL-1β augmented VEGF mRNA expression by a factor of 2.88 compared with no cytokine (Figure 4B).

thumbnail image

Figure 4. Real-time reverse transcriptase–polymerase chain reaction (RT-PCR) analysis of VEGF gene expression in RA synovial fibroblasts. A, Standard curve for quantitation of VEGF. Vertical axis shows the number of PCR cycles needed to reach threshold fluorescence (Ct) as determined by real-time quantitative PCR. Horizontal axis shows the concentrations for serial 10-fold dilution of the control VEGF cDNA. The value of Ct has a linear relationship with the amount of VEGF cDNA added in the PCR. B, Effect of cytokine stimulation on the expression of VEGF mRNA assayed by real-time RT-PCR. RA synovial fibroblasts (from patient RA1 in Table 1) were cultured for 48 hours with no cytokine, IL-6 (100 ng/ml)/soluble IL-6 receptor (sIL-6R; 100 ng/ml), or IL-6/sIL-6R with IL-1β (5 ng/ml). Complementary DNA (0.3 μg) synthesized from total RNA was loaded for the real-time RT-PCR assay. Values are the mean from a representative experiment performed in duplicate. See Figure 2 for other definitions.

Download figure to PowerPoint

IL-6R expression as determined by RT-PCR.

To test the possibility that the synergistic effect occurs via up-regulation of IL-6R expression by IL-1β, IL-6R expression in RA synovial fibroblasts was assessed by RT-PCR. IL-1β did not up-regulate IL-6R expression (data not shown).

Inhibition of VEGF production in synovial fibroblasts by anti–IL-6R mAb, IL-1Ra, and anti-TNFα mAb.

To examine the effect of cytokine-specific inhibitors on VEGF production in RA synovial fibroblasts, we treated the cells with anti–IL-6R mAb (25 μg/ml), IL-1Ra (10 μg/ml), or anti-TNFα mAb (10 μg/ml) in the presence of IL-6 (100 ng/ml), sIL-6R (100 ng/ml), IL-1β (5 ng/ml), and TNFα (10 ng/ml) (Figure 5). The VEGF level in the culture supernatant of RA synovial fibroblasts stimulated with the triple combination of IL-6, IL-1β, and TNFα was 3.1-fold higher than the level without any cytokine stimulation. Anti–IL-6R mAb treatment eliminated the synergistic effect of IL-6, IL-1β, and TNFα, leading to a mean ± SD VEGF level (754 ± 27 pg/ml) that was almost the same as that with IL-1β + TNFα stimulation (703 ± 30 pg/ml). However, no significant decrease in VEGF levels was observed in the RA synovial fibroblasts treated with IL-1Ra or anti-TNFα mAb, even though we used high enough concentrations of these inhibitors (2,000-fold and 1,000-fold in excess of the amount of IL-1β and TNFα, respectively) as indicated in a previous report (33) and according to the manufacturer's instructions. IL-1Ra at 20,000-fold in excess of the amount of IL-1β reduced VEGF production by 34% in the presence of IL-6 (100 ng/ml) + sIL-6R (100 ng/ml) + IL-1β (0.5 ng/ml), but not in the presence of IL-6 (100 ng/ml) + sIL-6R (100 ng/ml) + IL-1β (0.5 ng/ml) + TNFα (10 ng/ml) (data not shown).

thumbnail image

Figure 5. Inhibition of VEGF production in RA synovial fibroblasts by anti–IL-6 receptor monoclonal antibody (anti–IL-6R mAb), IL-1 receptor antagonist (IL-1Ra), and anti-TNFα mAb. RA synovial fibroblasts (from patient RA1 in Table 1) were cultured for 72 hours in the presence or absence of IL-6 (100 ng/ml)/soluble IL-6 receptor (sIL-6R; 100 ng/ml), IL-1β (5 ng/ml), TNFα (10 ng/ml), anti–IL-6R mAb (25 μg/ml), IL-1Ra (10 μg/ml), and/or anti-TNFα mAb (25 μg/ml). VEGF levels in the supernatants were determined by enzyme-linked immunosorbent assay. Values are the mean and SD from 3 independent experiments. See Figure 2 for other definitions.

Download figure to PowerPoint

DISCUSSION

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES

The fact that anti–IL-6R mAb therapy reduced serum VEGF levels in the RA patients indicates that IL-6 is involved in VEGF production in RA. Anti–IL-6R mAb therapy also improved disease activity in these patients, and the reduction in serum VEGF may be the result of disease amelioration. Another possible mechanism for the reduction in serum VEGF levels is that the blockade of IL-6 action directly inhibits VEGF production in RA. Because IL-6 reportedly induces VEGF production in some cells in vitro (21, 22), we analyzed VEGF production in RA synovial fibroblasts stimulated with IL-6. As we expected, IL-6 in the presence of sIL-6R induced VEGF production, but the activity was weak. Since IL-6 has been shown to suppress the growth of synovial fibroblasts (28), this increase in VEGF was not the result of an increase in the number of cells. Quantitative RT-PCR analysis confirmed that VEGF induction occurred through the augmentation of transcription.

IL-1β alone also significantly induced VEGF production in RA synovial fibroblasts, while TNFα did not. However, TNFα might have introduced the signal into the cells, since IL-6–induced VEGF production was augmented by TNFα. In addition, we have reported that synovial fibroblasts proliferate and produce IL-6 in response to TNFα (28). These results, taken together with the finding that the combination of IL-1β and TNFα did not have a synergistic or additive effect on VEGF production, indicate that the stimulatory signal of TNFα on VEGF expression may be weaker than that of IL-1β.

Although the VEGF-inducing activity of each cytokine was not very strong, IL-6, in synergy with IL-1β and TNFα, induced VEGF production in RA synovial fibroblasts. Of these cytokines, IL-6 is pivotal in VEGF induction in synovial fibroblasts, because a synergistic effect was observed between IL-6 and IL-1β and between IL-6 and TNFα, but not between IL-1β and TNFα. This in vitro finding has important implications. Since these cytokines all exist together in vivo, especially at high levels in affected joints, it is necessary to consider their interactions in order to understand their in vivo pathologic as well as physiologic roles in RA.

In this study, cells were stimulated with IL-6 in the presence of sIL-6R because they constitutively express gp130 signal transducer, but little IL-6R, on their cell surface. In addition, IL-1β did not induce IL-6R expression. Therefore, the synergistic effect of IL-6 and IL-1β may be mediated through second messengers. It has been reported that several signal-transducing molecules, such as signal transducer and activator of transcription 3, p42/p44 mitogen-activated protein kinase (MAP kinase), nuclear factor κB, and p38 MAP kinase, are involved in VEGF production (34–37). IL-6 reportedly activates the former two, while IL-1 and TNFα activate the latter two, and therefore the IL-6–mediated signal for VEGF production may be different from IL-1– and TNFα-mediated signals. The mutual action of these signaling molecules may exert a synergistic effect between IL-6 and IL-1 or between IL-6 and TNFα. IL-1 and TNFα may compensate for one another for VEGF-inducing activity because they share the same signaling pathways. Therefore, IL-6 blockade seems the most favorable for inhibition of VEGF production in the presence of the triple combination of IL-6, IL-1, and TNFα, as in the affected joints of patients with RA.

Anti–IL-6R mAb eliminated the synergistic effect of IL-6, IL-1β, and TNFα on VEGF production in RA synovial fibroblasts, while IL-1Ra and anti-TNFα mAb did not, even though we used high concentrations of IL-1Ra and anti-TNFα mAb to block the activity. Anti-TNFα mAb did not inhibit VEGF production. The combination of IL-6 and IL-1β induced the maximal response, similar to that obtained with the triple combination of IL-6, IL-1β, and TNFα. Since IL-1Ra up to 2,000-fold in excess of the amount of IL-1β present did not suppress the synergistic effect, IL-1Ra may not be strong enough to block the signal of IL-1β in its synergistic action with IL-6 on VEGF production. In fact, IL-1Ra at 20,000-fold in excess of the amount of IL-1β partially inhibited VEGF production. These findings, taken together with the short half-life of IL-1Ra in vivo, indicate that IL-1Ra does not seem favorable for suppression of VEGF production.

Our data suggest that IL-6 blockade is the most effective of the 3 inhibitors in suppressing VEGF production by synovial fibroblasts. This is supported by the evidence that serum VEGF levels in our patients treated with anti–IL-6R mAb were normalized. Treatment with anti-TNFα mAb cA2 has also been reported to significantly reduce serum VEGF levels in RA patients (15), although the VEGF-inducing activity of TNFα was weak and another anti-TNFα mAb did not inhibit VEGF production in the presence of IL-6 and IL-1β in our in vitro experiment. Since cA2 suppresses IL-6 and IL-1 production in vivo in RA patients (38), the decrease in serum VEGF may be mediated through the suppression of IL-6 and IL-1 production by cA2 therapy. However, the suppression of IL-6 by cA2 is transient (38), and this may be the reason that the reduction in serum VEGF in the patients treated with anti-TNFα mAb was only partial. Similar mechanisms may explain the partial reduction in CRP and serum amyloid A protein levels in such patients (38).

The findings of this study indicate that IL-6 blockade directly suppresses VEGF production in synovial fibroblasts and may consequently reduce serum VEGF levels in patients with RA. The inhibition of VEGF production could be one of the mechanisms by which anti–IL-6R mAb therapy is effective in this disease.

Induction of VEGF by proinflammatory cytokines has been observed not only in synovial fibroblasts from RA patients, but also in other cell types including malignant cells (21). Moreover, since VEGF is involved in the progression of malignant neoplasms, anti–IL-6R mAb therapy may be effective in the treatment of the neoplasms. Further studies will be needed to elucidate the molecular mechanism of VEGF production by proinflammatory cytokines.

Acknowledgements

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES

We wish to thank Ms C. Aoki and Ms K. Umetani for their technical assistance, and Ms A. Okajima for her secretarial work.

REFERENCES

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES