Inhibition of interleukin-1 but not tumor necrosis factor suppresses neovascularization in rat models of corneal angiogenesis and adjuvant arthritis.код для вставкиСкачать
ARTHRITIS & RHEUMATISM Vol. 46, No. 10, October 2002, pp. 2604–2612 DOI 10.1002/art.10546 © 2002, American College of Rheumatology Inhibition of Interleukin-1 but Not Tumor Necrosis Factor Suppresses Neovascularization in Rat Models of Corneal Angiogenesis and Adjuvant Arthritis Angela Coxon, Brad Bolon, Juan Estrada, Stephen Kaufman, Sheila Scully, Alana Rattan, Diane Duryea, Yi-Ling Hu, Karen Rex, Efrain Pacheco, Gwyneth Van, Debra Zack, and Ulrich Feige Objective. To assess the capacities of the cytokine inhibitors interleukin-1 receptor antagonist (IL-1Ra; anakinra) and PEGylated soluble tumor necrosis factor receptor I (PEG sTNFRI; pegsunercept) to suppress neovascularization. Methods. A corneal angiogenesis assay was performed by implanting nylon discs impregnated with an angiogenic stimulator (basic fibroblast growth factor or vascular endothelial growth factor) into one cornea of female Sprague-Dawley rats. Animals were treated with IL-1Ra or PEG sTNFRI for 7 days, after which new vessels were quantified. In a parallel study, male Lewis rats with mycobacteria-induced adjuvant-induced arthritis were treated with IL-1Ra or PEG sTNFRI for 7 days beginning at disease onset, after which scores for inflammation and bone erosion as well as capillary counts were acquired from sections of arthritic hind paws. Results. Treatment with IL-1Ra yielded a dosedependent reduction in growth factor–induced corneal angiogenesis, while PEG sTNFRI did not. IL-1Ra, but not PEG sTNFRI, significantly reduced the number of capillaries in arthritic paws, even though both anticyto- kines reduced inflammation and bone erosion to a similar degree. Conclusion. These data support a major role for IL-1, but not TNF␣, in angiogenesis and suggest that an additional antiarthritic mechanism afforded by IL-1 inhibitors, but not anti-TNF agents, is the suppression of the angiogenic component of pannus. Inflammatory arthritis is initiated and sustained by the release of myriad proinflammatory cytokines (for review, see refs. 1 and 2). In patients with rheumatoid arthritis (RA), 2 critical proinflammatory cytokines are interleukin-1 (IL-1) and tumor necrosis factor ␣ (TNF␣) (3). Inhibition of IL-1 and/or TNF␣ reduces the extent of inflammation in RA (4,5) and lessens inflammation and bone destruction in various experimental models of arthritis (6–12). Therefore, therapeutic agents that inhibit the action of these 2 cytokines are gaining rapid acceptance as early, aggressive treatments for RA. IL-1 (13–15) and TNF␣ (16–19) both also have been implicated as angiogenic factors. This finding is provocative because angiogenesis is a critical component in the initiation and maintenance of pannus, the aggressive fibrovascular extension of synovial tissue that is responsible for the extensive bone and cartilage damage in RA (20) and experimental models of arthritis (21). Human vascular endothelial cells activated by exposure in vitro to either IL-1 or TNF␣ increase their expression of receptor activator of nuclear factor B ligand (RANKL, an osteoclast differentiation factor) and its soluble receptor osteoprotegerin (OPG, an osteoclastogenesis inhibitory factor) (22). However, OPG expression (and thus, RANKL inhibition) peaks early and then falls, while RANKL increases slowly and is sustained (22). These in vitro results implicate cytokine-activated Presented in part at the 65th Annual Scientific Meeting of the American College of Rheumatology, San Francisco, CA, November 2001. Angela Coxon, PhD, Brad Bolon, DVM, PhD, Juan Estrada, MD, Stephen Kaufman, MS, Sheila Scully, BS, Alana Rattan, BS, Diane Duryea, Yi-Ling Hu, BS, Karen Rex, BS, Efrain Pacheco, BS, Gwyneth Van, BS, Debra Zack, MD, PhD, Ulrich Feige, PhD: Amgen Inc., Thousand Oaks, California. Drs. Coxon and Bolon contributed equally to this work. Address correspondence and reprint requests to Ulrich Feige, PhD, Amgen, One Amgen Center Drive, M/S 29-M-B, Thousand Oaks, CA 91320-1799. E-mail: email@example.com. Submitted for publication October 30, 2001; accepted in revised form June 18, 2002. 2604 INHIBITION OF IL-1, NOT TNF, SUPPRESSES NEOVASCULARIZATION endothelial cells in newly formed blood vessels as direct participants in the bone erosion by pannus. Numerous animal studies indicate that antiangiogenic treatments effectively reduce both the incidence and the severity of collagen-induced arthritis (CIA) (23–27) and adjuvant-induced arthritis (AIA) (28,29). The present study compared the antiangiogenic activities of anti–IL-1 and anti-TNF␣ biologic agents in conventional models of arthritis and vascular growth factor–induced neovascularization. Our data indicate that therapy targeting IL-1, but not TNF␣, will reduce angiogenesis in arthritic joints. MATERIALS AND METHODS Animals. Rats (Charles River, Wilmington, MA) were acclimated for 1 week, after which they were randomly assigned to treatment groups. Animals were given tap water and fed pelleted rodent chow (8640; Harlan Teklad, Madison, WI) ad libitum; calcium and phosphorus contents were 1.2% and 1.0%, respectively. Surgical procedures were performed using isoflurane anesthesia and standard operating practices and sterile technique. All animals were killed by CO2 inhalation. These studies were conducted in accordance with federal animal care guidelines and were preapproved by the Amgen Institutional Animal Care and Use Committee. Treatments. The treatments assessed in this study were the recombinant human anticytokine biologics IL-1 receptor antagonist (IL-1Ra) (anakinra; Amgen, Thousand Oaks, CA) and PEGylated soluble TNF receptor I (PEG sTNFRI) (pegsunercept; Amgen). IL-1Ra was administered at 0, 0.5, 1.5, or 5 mg/kg/hour in cortical somatosensory evoked potential vehicle (140 mM saline containing 10 mM sodium citrate, 0.5 mM EDTA, and 0.1% [weight/volume] Tween 80) by subcutaneous (SC) infusion using implanted osmotic minipumps (model 2ML1, delivery rate 10 l/hour; Alza, Palo Alto, CA), while PEG sTNFRI was given at 0 or 4 mg/kg/day in phosphate buffered saline (PBS) by SC bolus. Minipumps were implanted in the dorsal subcutis, and wounds were sealed using steel clips. Doses of IL-1Ra and PEG sTNFRI were selected based on previous studies (12) demonstrating that these doses, routes, and schedule cover the region of the dose-response curve ranging from (almost) inactive to fully active against the severe polyarthritis characteristic of the AIA model in Lewis rats. In particular, we have shown that IL-1Ra (5 mg/kg/hour) and PEG sTNFRI (4 mg/kg/day) yield comparable antiinflammatory and bone-sparing effects when given for 7 days. Corneal implant angiogenesis assay. The vascular growth factor–induced corneal neovascularization bioassay was performed in triplicate using a standard experimental design (30), with the modification that a porous solid was used to dispense protein (rather than cells or a slow-release polymer) (31). Briefly, circular discs (0.6 mm diameter) were punched from nylon filter paper (Nylaflo; Gelman, Ann Arbor, MI) using a 20-gauge needle with a squared-off end, after which they were placed in PBS containing either 0.1% bovine serum albumin (BSA) alone (vehicle) or 0.1% BSA in combination with 1 of 2 recombinant human vascular growth factors 2605 (R&D Systems, Minneapolis, MN): 3.75 M basic fibroblast growth factor (bFGF) or 10 M vascular endothelial growth factor (VEGF). Discs were incubated for 1 hour at 4°C before use; each disc absorbed ⬃0.1 l of solution. Adult female Sprague-Dawley rats weighing 250–300 gm (n ⫽ 8 per group) were anesthetized with isoflurane, after which a vertical incision 0.8 mm in length was made on the cornea. A pocket was formed in the corneal stoma, and a single disc was inserted. The margin of each disc was located ⬃1.8–2.0 mm from the blood vessels of the lateral limbus. Prior immunohistochemical work in our laboratory has shown that the vascular growth factor persists in and around the corneal disc for more than 7 days (Coxon A: unpublished observations). Treatment with IL-1Ra (0, 0.5, 1.5, or 5 mg/kg/hour) or PEG sTNFRI (0 or 4 mg/kg/day) was initiated on the day of surgery and continued for 7 days. An additional cohort to control for nonspecific anti-angiogenic effects of stress associated with implantation of the osmotic minipump was infusion of BSA at 5 mg/kg/hour. All groups were run in parallel to limit variability associated with interstudy comparisons. At necropsy, the eyes were enucleated and immersed overnight in zinc formalin. The cornea and associated implant were removed, placed in distilled water, transilluminated, and photographed at 5⫻ using a Sony CatsEye DKC 5000 digital camera (A.G. Heinze, Lake Forest, CA) mounted on a Nikon SMZ-U stereomicroscope (A.G. Heinze); a reference stage micrometer was photographed for image calibration. Numerical data were generated from the digital images using Metamorph image analysis software (v4.5; Universal Imaging, Downingtown, PA) installed on a Windows NT workstation (Microsoft, Redmond, WA). Three end points were analyzed on each corneal image: 1) disc placement distance from the limbal vessels, 2) the number of vessels intersecting a 2.0 mm–long perpendicular line drawn across the midpoint of the shortest line between the disc and the limbus, and 3) blood vessel area (as determined by digital thresholding and automated pixel counting). After 7 days of treatment with PEG sTNFRI or vehicle, selected corneas were stained with a proprietary mouse monoclonal antibody (mAb) specific for human sTNFRI (Amgen). Briefly, acetone-fixed 6 M–thick frozen sections were blocked with CAS Block (Zymed, South San Francisco, CA) and incubated with either the anti-sTNFRI antibody or an isotype-matched mAb control. Binding of the primary antibody was detected using a biotinylated horse anti-mouse secondary antibody (Vector, Burlingame, CA). Slides were quenched with 3% H2O2 followed with avidin– biotin–peroxidase complex (Vector). Reaction sites were visualized with diaminobenzidine (Dako, Carpinteria, CA) and counterstained with hematoxylin. Selected corneas were processed for in situ hybridization. The eyes were harvested, after which the corneal region containing the disc was removed and fixed by immersion in zinc formalin (Z-Fix; Anatech, Battle Creek, MI). Tissue was processed in paraffin by routine methods, after which antisense RNA probes for rat CD31 (platelet endothelial cell adhesion molecule [PECAM], corresponding to nucleotides 220–474 of the mouse sequence; GenBank accession no. L06039), rat IL-1RI (nucleotides 777–1126; GenBamk accession no. M95578), and rat IL-1RII (nucleotides 683–933; GenBank 2606 COXON ET AL Table 1. Semiquantitative criteria for histopathologic lesion scores Score Inflammation 0 1 2 3 4 Bone erosion 0 1 2 3 4 5 Grading scale Normal Few inflammatory cells Mild inflammation Moderate inflammation (often but not always diffuse) Marked inflammation (diffuse and dense, with large periarticular abscesses) Normal Minimal loss of cortical or trabecular bone at a few sites Mild loss of cortical or trabecular bone at modest numbers of sites (generally tarsals) Moderate loss of bone at many sites (usually the trabeculae of the tarsals, but sometimes the cortex of the distal tibia) Marked loss of bone at many sites (usually as extensive destruction of trabeculae in the tarsals, but sometimes with partial loss of cortical bone in the distal tibia) Marked loss of bone at many sites (with fragmenting of tarsal trabeculae and fullthickness penetration of cortical bone in the distal tibia) accession no. Z22812) labeled with 33P-rUTP (Amersham, Arlington Heights, IL) were applied to 4 M–thick sections according to standard protocols (32). Following emulsion autoradiography, sections were counterstained with hematoxylin and eosin (H&E) and examined under both darkfield and standard illumination. Induction of AIA. Adjuvant arthritis was induced in adult male Lewis rats weighing 180–200 gm (n ⫽ 6 per group) on day 0 as described (12), by a single intradermal injection, at the tail base, of heat-killed Mycobacterium tuberculosis H37Ra (0.5 mg; Difco, Detroit, MI) suspended in 0.05 ml paraffin oil (Crescent Chemical, Islandia, NY). A refined volume displacement method (12) was used beginning on day 8 to measure hind paw volume to determine the clinical onset of arthritis. At onset (typically, day 9), a 7-day course of therapy was initiated using IL-1Ra (5 mg/kg/hour) or PEG sTNFRI (4 mg/kg/day). A concurrent control group consisted of untreated arthritic rats; additional control cohorts treated with vehicle(s) were not included in this study because our past experience with this model has shown that clinical and histopathologic responses of untreated and vehicle-treated animals are equivalent (Feige U: unpublished observations). All group studies were run in parallel. The small group size was used because interindividual variability between untreated arthritic rats is minimal (12). At necropsy (day 16 postimmunization), hind paws were removed at the fur line (just proximal to hock), fixed in 70% ethanol, decalcified, divided longitudinally along the median axis, and processed in paraffin. One 4 M–thick section was stained with H&E. As previously described (12), inflammation and bone erosion scores were acquired in a “blinded” analysis using semiquantitative grading scales (Table 1). In addition, a serial section of each hind paw was stained using a commercial indirect immunoperoxidase kit (Vectastain Elite ABC Kit; Vector) with a rabbit anti-human mAb (Dako) directed against the endothelial marker von Willebrand factor (also known as factor VIII–related antigen). Labeled capillaries were enumerated in periarticular soft tissues at 4 sites (Figure 1) because these zones exhibited a marked leukocyte infiltrate in untreated arthritic animals. Counts were made at 200⫻ using an ocular reticule with 100 square divisions. The Figure 1. Photomicrograph of a hind paw from a rat with adjuvant-induced arthritis, compared with a control rat. Four sites at which capillary counts were acquired are denoted (hematoxylin and eosin stain; original magnification ⫻ 6). INHIBITION OF IL-1, NOT TNF, SUPPRESSES NEOVASCULARIZATION 2607 total area counted per site was 0.5 mm2. An attempt to assess messenger RNA expression levels for selected vascular markers (PECAM), vascular growth factors (bFGF, VEGF), and IL-1RI and IL-1RII in normal and arthritic joints was thwarted by degradation of nucleic acid integrity during processing. Statistical analysis. Statistical significance for the corneal angiogenesis assay was assessed by analysis of variance followed by Fisher’s exact test. Histopathologic data were compared using conservative nonparametric tests, the chisquare test for inflammation and erosion scores, and Wilcoxon’s rank sum test for the capillary counts. P values less than 0.05 were considered significant. RESULTS Figure 2. Effect of interleukin-1 receptor antagonist (IL-1Ra) on vascular endothelial growth factor (VEGF)–induced corneal neovascularization. Infusion of IL-1Ra (5 mg/kg/hour for 7 days) yielded almost complete inhibition of VEGF-induced corneal neovascularization between the implanted disc (top) and the limbus (bottom). In contrast, PEGylated soluble tumor necrosis factor receptor I (PEG sTNFRI; 4 mg/kg/day for 7 days) had essentially no effect. A–D show representative corneas from each of the 4 treatment groups. BSA ⫽ bovine serum albumin. Bar ⫽ 500 m. Corneal angiogenesis assay. Seven days after corneal implantation of discs, neovascularization extending from the limbus to the implant was grossly prominent in all rats in which discs contained either VEGF (Figures 2B and 3) or bFGF (Figure 3). As expected, angiogenesis was absent if vascular growth factors were not present in the disc (Figure 2A). Systemic administration of IL-1Ra for the duration of the study produced a significant, dose-dependent decrease in corneal neovascularization induced by either bFGF (Figure 3) or VEGF (Figures 2C and 3). Counts of both the new capillary profiles and the total vascularized area were lowered; in animals treated with 5 Figure 3. Effect of IL-1Ra on basic fibroblast growth factor (bFGF)– and VEGF-induced corneal angiogenesis. IL-1Ra infusion produced a significant (ⴱ ⫽ P ⱕ 0.05) dose-dependent decrease in both bFGF-induced (a) and VEGF-induced (b) corneal angiogenesis relative to rats treated with vehicle or with an inactive protein (BSA, a control for stress-induced anti-angiogenic effects associated with protein infusion). Representative results from 1 of 3 experiments are shown; values are the group mean ⫾ SEM (n ⫽ 8). CON ⫽ control (see Figure 2 for other definitions). 2608 COXON ET AL Figure 4. Effect of PEG sTNFRI on VEGF-induced corneal neovascularization. Administration of PEG sTNFRI (4 mg/kg/day for 7 days) had no impact on VEGF-induced corneal neovascularization. This result was obtained despite the fact that the cornea was permeated with PEG sTNFRI (upper photomicrograph), as indicated by uniform distribution of immunoreactivity following application of a mouse monoclonal antibody specific for human sTNFRI. In contrast, corneal control sections from vehicle-treated rats labeled with anti-sTNFRI (middle panel) and from PEG sTNFRI–treated animals labeled with an isotype control (CON) antibody (lower panel) exhibited no reactivity. Arrows indicate VEGF-induced blood vessels. Representative results from 1 of 3 experiments are shown; values are the group mean ⫾ SEM (n ⫽ 8). Bar ⫽ 50 M. See Figure 2 for other definitions. mg/kg/hour of IL-1Ra, group mean values were comparable with those observed in rats given implants that lacked vascular growth factors. In contrast, neither PEG sTNFRI nor BSA affected the angiogenic response elicited by bFGF (data not shown) or VEGF (Figures 2 and 4). The presence of PEG sTNFRI throughout the corneal stroma surrounding implants was confirmed by immunohistochemistry analysis (Figure 4). The expression patterns of CD31 (PECAM), IL-1RI, and IL-1RII were determined by in situ hybridization on corneas harvested 7 days after the induction of angiogenesis (Figure 5). CD31 expression was higher near the limbus and occurred only in endothelial cells, including those located at the leading edge of new capillary branches. Both IL-1RI and IL-1RII were substantially up-regulated in corneas near VEGFcontaining implants. IL-1RI exhibited a more diffuse pattern, though most expressing cells were located near the implant. The major cell types labeled with IL-1RI were mononuclear inflammatory cells (with small to large nuclei) and corneal stromal cells (with fusiform or serpentine nuclei), although large endothelial cells occasionally expressed this receptor. The IL-1RII signal was higher near the implant and was found only in mononuclear inflammatory cells and a few corneal stromal cells. AIA. Adjuvant-induced arthritis is characterized by a clearly defined pattern of structural lesions combining both destructive and reparative processes. Dense aggregates of mixed inflammatory cells and numerous osteoclasts were present in association with erosions of cortical and trabecular regions of the tarsal bones and tibiae. Production of new bony trabeculae along periosteal and endosteal surfaces was extensive. Joint cartilages often were isolated by erosion of the subjacent epiphyseal bone, but remained intact. Neutrophils, newly formed capillaries (i.e., angiogenesis), hyperplastic stromal cells, and sometimes edema were the most INHIBITION OF IL-1, NOT TNF, SUPPRESSES NEOVASCULARIZATION 2609 Figure 5. In situ hybridization for CD31 (platelet endothelial cell adhesion molecule), IL-1RI, and IL-1RII in the rat corneal angiogenesis model. Corneal implants impregnated with VEGF induced both angiogenesis and intrastromal leukocyte infiltration (A), while implants with an inactive protein (BSA) did not (B). Introduction of growth factor substantially up-regulated expression of CD31, as well as IL-1RI and IL-1RII, though their patterns of expression were different. CD31, an endothelial cell marker, was expressed in blood vessels extending from the limbus toward the disc. Expression of IL-1RI was more diffuse, was concentrated near the disc, and was found in leukocytes, corneal stromal cells, and some large endothelial cells. IL-1RII expression was also highest near the implanted disc but was restricted to leukocytes and some stromal cells. Sections are oriented with the limbus (L) on the left and the implanted disc (D) on the right. Dashed lines in insets A1 and B1 denote the plane of section. A and B are hematoxylin and eosin–stained brightfield images. The 3 panels below A and B show isotopic in situ hybridization under darkfield illumination, with higher-magnification insets below A1 and B1. Black arrowheads indicate capillaries; red arrows indicate nonvascular (leukocytic or stromal) cells. See Figure 2 for definitions. prominent inflammatory changes in the periarticular soft tissues. Semiquantitative grades of both inflammation and bone erosion generally were of marked severity in rats with untreated arthritis, while changes were absent in nonarthritic controls (Figure 6). Therapy with either IL-1Ra (5 mg/kg/hour) or PEG sTNFRI (4 mg/kg/day) significantly reduced inflammation and bone erosion to a similar degree (Figure 6). However, despite their similar antiinflammatory and bone-protective effects, IL-1Ra reduced the angiogenic response in arthritic joints, while PEG sTNFRI did not. The mean number of capillary profiles per mm2 in rats given IL-Ra was modestly but not significantly increased by 13% relative to counts in nonarthritic controls (Figure 7). In contrast, both untreated arthritic rats and arthritic animals treated with PEG sTNFRI had 35% more capillaries in periarticular soft tissues than did nonarthritic controls (Figure 7). DISCUSSION The development of new therapies that combat the angiogenic component of pannus could have a significant impact on the health and quality of life for RA patients. This treatment paradigm is supported by numerous animal studies demonstrating that antiangiogenic agents significantly suppress both the incidence and the severity of disease (23–29). The most efficient means of adding angiogenesis inhibitors to the clinical armamentarium for RA would be to determine whether current therapies exhibit anti-angiogenic properties. For example, 2 current RA therapies, methotrexate (33) and D-penicillamine (34), have been shown to inhibit neovascularization in vitro and in vivo, although the mechanisms by which this effect is mediated are not known. A new paradigm in this therapy is the early and aggressive deployment of biologic response-modifying agents to specifically inhibit the actions of proinflamma- 2610 tory cytokines in RA. Both IL-1 (13–15) and TNF␣ (16,17), 2 master cytokines that play a significant role in RA (3), are known to promote angiogenesis. Given this intense interest in anticytokine biologics, our experiments provide important new information regarding the ability of such novel therapeutic molecules to regulate neovascularization in the arthritic joint. One key finding afforded by our present data is that administration of IL-1Ra to block IL-1 activity significantly inhibits angiogenesis in vivo. This effect was observed in both the corneal implant system (a noninflammatory setting) and the AIA model (an inflammatory condition). IL-1Ra previously has been found to inhibit angiogenesis in a model of inflammatory corneal neovascularization in mice (15). The exact mechanism by which IL-1Ra exerts this effect is unknown. IL-1 has many pro-angiogenic activities, including increased expression of vascular growth factors (35,36), endothelial Figure 6. Effect of IL-1Ra and PEG sTNFRI on inflammation and bone erosion. Infusion of IL-1Ra or injection of PEG sTNFRI reduced inflammation (A) and bone erosion (B) to a comparable degree in Lewis rats with adjuvant-induced arthritis. ⴱⴱ ⫽ P ⱕ 0.05 versus untreated arthritic animals; the percent reduction is indicated. Values are the mean and SEM (n ⫽ 6). See Figure 2 for definitions. COXON ET AL Figure 7. Effect of IL-1Ra on the number of capillaries in periarticular soft tissues of Lewis rats. Infusion of IL-1Ra reduced the number of capillaries in periarticular soft tissues of Lewis rats with adjuvantinduced arthritis relative to untreated arthritic rats. In contrast, injection of PEG sTNFRI had no effect on blood vessel numbers. ⴱⴱ ⫽ P ⱕ 0.05 versus normal animals; the percent difference is indicated. Values are the mean and SEM (n ⫽ 6). See Figure 2 for definitions. mitogenesis (37), and induction of matrix metalloproteinases (36); IL-1Ra presumably can negate all of them. However, our data indicating that IL-1Ra blocks bFGF- and VEGF-induced angiogenesis in the cornea are particularly intriguing in that increasing evidence suggests that these vascular growth factors play a prominent role in the pathogenesis of RA. For example, VEGF levels in serum (38,39) and synovial fluid (40–42) are substantially elevated in RA and are correlated with disease activity. Furthermore, VEGF is expressed widely in arthritic joints, especially macrophages and fibroblasts in RA patients (40,43), as well as mice with CIA (25) and rats with AIA (42). Significantly, IL-1 induces VEGF expression in cultured fibroblast-like synoviocytes isolated from RA synovial tissue (35,44). Similarly, IL-1, acting via cyclooxygenase 2 (45,46) and nitric oxide (36), also mediates the expression of bFGF. Notwithstanding a conflicting report that IL-1 can inhibit bFGF-induced angiogenesis in a rabbit corneal angiogenesis model (47), the preponderance of evidence, including the data acquired in our corneal angiogenesis model, suggests that IL-1 plays some part in vivo in the induction of angiogenesis by bFGF and VEGF. More important, our findings clearly show that inhibition of IL-1 suppresses angiogenesis in arthritic joints. Surprisingly, another significant finding afforded by our present data is that administration of a TNF inhibitor, PEG sTNFRI, has no impact on angiogenesis in vivo. Again, this outcome was apparent in both the corneal implant system and the Lewis rat model of AIA. This finding was unanticipated because TNF␣ has been reported to initiate angiogenesis in the rabbit (16) and rat (17) models of corneal neovascularization, and anti- INHIBITION OF IL-1, NOT TNF, SUPPRESSES NEOVASCULARIZATION TNF␣ therapy significantly decreases serum VEGF in RA (38). The explanation for this phenomenon is unclear at this point. It is unlikely that the process of growth factor–induced neovascularization that occurs in the cornea is mediated by different biochemical events than pathologic angiogenesis in the inflamed joint. This inference is supported by a study documenting sustained elevation in circulating TNF␣ concentrations in conjunction with falling tissue levels of VEGF in arthritic rats treated with an angiogenic inhibitor (24). However, our data do not support this interpretation because PEG sTNFRI did not exhibit an anti-angiogenic effect whether applied in either the arthritic joint or the cornea. Alternatively, this discrepancy might reflect a difference in the arthritogenic functions governed by IL-1 and TNF␣. While many proinflammatory activities of these 2 cytokines are shared, their functions are not identical (2). Thus, the conclusion that best fits our present data is that the proinflammatory and jointdamaging effects of IL-1 and TNF␣ in arthritis represent functions held in common, while IL-1 alone controls the angiogenic response. This hypothesis is supported by numerous experimental studies of arthritis that suggest a central role for TNF␣ in mediating inflammation, while IL-1 controls bone and cartilage destruction as well as inflammation (48). The importance of IL-1 but not TNF␣ to angiogenesis in the joints of rats with AIA suggested by our present data could help to explain the better efficacy afforded by IL-1 inhibition with respect to preserving joint integrity. In conclusion, our experiments indicate that IL-1, but not TNF␣, plays a critical role in neovascularization, including pathologic angiogenesis that occurs in arthritic joints of rats with AIA. This important new finding suggests that IL-1 inhibitors, but not anti-TNF agents, will affect the process of arthritis not just by downregulating the inflammatory cascade, but also by thwarting the invasive, highly vascular pannus reaction that is responsible for much of the bone and cartilage destruction characteristic of RA. These data are significant because agents that block IL-1 should preserve joint integrity by 2 distinct mechanisms: averting pathologic angiogenesis (as demonstrated by our results) and moderating osteoclast expansion (49,50). Given the high treatment costs and lost productivity associated with the crippling bone sequelae of chronic RA, the capacity of IL-1Ra to have an effect on bone and cartilage destruction via 2 pathways should afford a central place for this agent as well as other IL-1 inhibitors in the anti-RA armamentarium. 2611 ACKNOWLEDGMENT The authors thank Dr. Robert Radinsky for critical review of the manuscript. REFERENCES 1. Arend WP, Dayer JM. Cytokines and cytokine inhibitors or antagonists in rheumatoid arthritis. Arthritis Rheum 1990;33: 305–15. 2. Brennan FM, Maini RN, Feldmann M. Role of pro-inflammatory cytokines in rheumatoid arthritis. Springer Semin Immunopathol 1998;20:133–47. 3. Brennan FM, Field M, Chu CQ, Feldmann M, Maini RN. Cytokine expression in rheumatoid arthritis. Br J Rheumatol 1991;30 Suppl 1:76–80. 4. Bresnihan B, Alvaro-Gracia JM, Cobby M, Doherty M, Domljan Z, Emery P, et al. Treatment of rheumatoid arthritis with recombinant human interleukin-1 receptor antagonist. Arthritis Rheum 1998;41:2196–204. 5. Richard-Miceli C, Dougados M. Tumour necrosis factor-␣ blockers in rheumatoid arthritis: review of the clinical experience. BioDrugs 2001;15:251–9. 6. Williams RO, Feldmann M, Maini RN. Anti-tumor necrosis factor ameliorates joint disease in murine collagen-induced arthritis. Proc Natl Acad Sci U S A 1992;89:9784–8. 7. Geiger T, Towbin H, Cosenti-Vargas A, Zingel O, Arnold J, Rordorf C, et al. Neutralization of interleukin-1␤ activity in vivo with a monoclonal antibody alleviates collagen-induced arthritis in DBA/1 mice and prevents the associated acute-phase response. Clin Exp Rheumatol 1993;11:515–22. 8. Wooley PH, Dutcher J, Widmer MB, Gillis S. Influence of a recombinant human soluble tumor necrosis factor receptor FC fusion protein on type II collagen-induced arthritis in mice. J Immunol 1993;151:6602–7. 9. Wooley PH, Whalen JD, Chapman DL, Berger AE, Richard KA, Aspar DG, et al. The effect of an interleukin-1 receptor antagonist protein on type II collagen–induced arthritis and antigen-induced arthritis in mice. Arthritis Rheum 1993;36:1305–14. 10. Joosten LA, Helsen MM, van de Loo FA, van den Berg WB. Anticytokine treatment of established type II collagen–induced arthritis in DBA/1 mice: a comparative study using anti-TNF␣, anti–IL-1␣/␤, and IL-1Ra. Arthritis Rheum 1996;39:797–809. 11. Kuiper S, Joosten LA, Bendele AM, Edwards CK III, Arntz OJ, Helsen MM, et al. Different roles of tumour necrosis factor ␣ and interleukin 1 in murine streptococcal cell wall arthritis. Cytokine 1998;10:690–702. 12. Feige U, Hu YL, Gasser J, Campagnuolo G, Munyakazi L, Bolon B. Anti-interleukin-1 and anti-tumor necrosis factor-␣ synergistically inhibit adjuvant arthritis in Lewis rats. Cell Mol Life Sci 2000;57:1457–70. 13. Mantovani A, Bussolino F, Dejana E. Cytokine regulation of endothelial cell function. FASEB J 1992;6:2591–9. 14. Fan TP, Hu DE, Guard S, Gresham GA, Watling KJ. Stimulation of angiogenesis by substance P and interleukin-1 in the rat and its inhibition by NK1 or interleukin-1 receptor antagonists. Br J Pharmacol 1993;110:43–9. 15. Dana MR, Zhu SN, Yamada J. Topical modulation of interleukin-1 activity in corneal neovascularization. Cornea 1998;17:403–9. 16. Fràter-Schroder M, Risau W, Hallmann R, Gautschi P, Bohlen P. Tumor necrosis factor type ␣, a potent inhibitor of endothelial cell growth in vitro, is angiogenic in vivo. Proc Natl Acad Sci U S A 1987;84:5277–81. 17. Leibovich SJ, Polverini PJ, Shepard HM, Wiseman DM, Shively V, Nuseir N. Macrophage-induced angiogenesis is mediated by tumour necrosis factor-␣. Nature 1987;329:630–2. 2612 18. Paleolog E. Target effector role of vascular endothelium in the inflammatory response: insights from the clinical trial of antiTNF␣ antibody in rheumatoid arthritis. Mol Pathol 1997;50: 225–33. 19. Qian Y, Dekaris I, Yamagami S, Dana MR. Topical soluble tumor necrosis factor receptor type I suppresses ocular chemokine gene expression and rejection of allogeneic corneal transplants. Arch Ophthalmol 2000;118:1666–71. 20. Walsh DA, Pearson CI. Angiogenesis in the pathogenesis of inflammatory joint and lung diseases. Arthritis Res 2001;3:147–53. 21. Buma P, Groenenberg M, Rijken PF, van den Berg WB, Joosten L, Peters H. Quantitation of the changes in vascularity during arthritis in the knee joint of a mouse with a digital image analysis system. Anat Rec 2001;262:420–8. 22. Collin-Osdoby P, Rothe L, Anderson F, Nelson M, Maloney W, Osdoby P. Receptor activator of NF-B and osteoprotegerin expression by human microvascular endothelial cells, regulation by inflammatory cytokines, and role in human osteoclastogenesis. J Biol Chem 2001;276:20659–72. 23. Peacock DJ, Banquerigo ML, Brahn E. Angiogenesis inhibition suppresses collagen arthritis. J Exp Med 1992;175:1135–8. 24. Oliver SJ, Cheng TP, Banquerigo ML, Brahn E. Suppression of collagen-induced arthritis by an angiogenesis inhibitor, AGM1470, in combination with cyclosporin: reduction of vascular endothelial growth factor (VEGF). Cell Immunol 1995;166: 196–206. 25. Lu J, Kasama T, Kobayashi K, Yoda Y, Shiozawa F, Hanyuda M, et al. Vascular endothelial growth factor expression and regulation of murine collagen-induced arthritis. J Immunol 2000;164:5922–7. 26. Miotla J, Maciewicz R, Kendrew J, Feldmann M, Paleolog E. Treatment with soluble VEGF receptor reduces disease severity in murine collagen-induced arthritis. Lab Invest 2000;80:1195–205. 27. Sone H, Kawakami Y, Sakauchi M, Nakamura Y, Takahashi A, Shimano H, et al. Neutralization of vascular endothelial growth factor prevents collagen-induced arthritis and ameliorates established disease in mice. Biochem Biophys Res Commun 2001;281: 562–8. 28. Peacock DJ, Banquerigo ML, Brahn E. A novel angiogenesis inhibitor suppresses rat adjuvant arthritis. Cell Immunol 1995;160: 178–84. 29. Badger AM, Blake S, Kapadia R, Sarkar S, Levin J, Swift BA, et al. Disease-modifying activity of SB 273005, an orally active, nonpeptide ␣V␤3 (vitronectin receptor) antagonist, in rat adjuvant-induced arthritis. Arthritis Rheum 2001;44:128–37. 30. Polverini PJ, Cotran PS, Gimbrone MA Jr, Unanue ER. Activated macrophages induce vascular proliferation. Nature 1977;269: 804–6. 31. Kenyon BM, Voest EE, Chen CC, Flynn E, Folkman J, D’Amato RJ. A model of angiogenesis in the mouse cornea. Invest Ophthalmol Vis Sci 1996;37:1625–32. 32. Wilkinson DG. In situ hybridization: a practical approach. Oxford (UK): IRL Press; 1993. 33. Hirata S, Matsubara T, Saura R, Tateishi H, Hirohata K. Inhibition of in vitro vascular endothelial cell proliferation and in vivo neovascularization by low-dose methotrexate. Arthritis Rheum 1989;32:1065–73. 34. Matsubara T, Saura R, Hirohata K, Ziff M. Inhibition of human endothelial cell proliferation in vitro and neovascularization in vivo by D-penicillamine. J Clin Invest 1989;83:158–67. 35. Ben-Av P, Crofford LJ, Wilder RL, Hla T. Induction of vascular endothelial growth factor expression in synovial fibroblasts by COXON ET AL 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. prostaglandin E and interleukin-1: a potential mechanism for inflammatory angiogenesis. FEBS Lett 1995;372:83–7. Sasaki K, Hattori T, Fujisawa T, Takahashi K, Inoue H, Takigawa M. Nitric oxide mediates interleukin-1-induced gene expression of matrix metalloproteinases and basic fibroblast growth factor in cultured rabbit articular chondrocytes. J Biochem (Tokyo) 1998; 123:431–9. Detmar M, Tenorio S, Hettmannsperger U, Ruszczak Z, Orfanos CE. Cytokine regulation of proliferation and ICAM-1 expression of human dermal microvascular endothelial cells in vitro. J Invest Dermatol 1992;98:147–53. Paleolog EM, Young S, Stark AC, McCloskey RV, Feldmann M, Maini RN. Modulation of angiogenic vascular endothelial growth factor by tumor necrosis factor ␣ and interleukin-1 in rheumatoid arthritis. Arthritis Rheum 1998;41:1258–65. Maeno N, Takei S, Imanaka H, Takasaki I, Kitajima I, Maruyama I, et al. Increased circulating vascular endothelial growth factor is correlated with disease activity in polyarticular juvenile rheumatoid arthritis. J Rheumatol 1999;26:2244–8. Fava RA, Olsen NJ, Spencer-Green G, Yeo KT, Yeo TK, Berse B, et al. Vascular permeability factor/endothelial growth factor (VPF/VEGF): accumulation and expression in human synovial fluids and rheumatoid synovial tissue. J Exp Med 1994;180:341–6. Koch AE, Harlow LA, Haines GK, Amento EP, Unemori EN, Wong WL, et al. Vascular endothelial growth factor: a cytokine modulating endothelial function in rheumatoid arthritis. J Immunol 1994;152:4149–56. Yamashita A, Yonemitsu Y, Okano S, Nakagawa K, Nakashima Y, Irisa T, et al. Fibroblast growth factor-2 determines severity of joint disease in adjuvant-induced arthritis in rats. J Immunol 2002;168:450–7. Nagashima M, Yoshino S, Ishiwata T, Asano G. Role of vascular endothelial growth factor in angiogenesis of rheumatoid arthritis. J Rheumatol 1995;22:1624–30. Jackson JR, Minton JA, Ho ML, Wei N, Winkler JD. Expression of vascular endothelial growth factor in synovial fibroblasts is induced by hypoxia and interleukin 1␤. J Rheumatol 1997;24: 1253–9. Crofford LJ, Wilder RL, Ristimaki AP, Sano H, Remmers EF, Epps HR, et al. Cyclooxygenase-1 and -2 expression in rheumatoid synovial tissues: effects of interleukin-1␤, phorbol ester, and corticosteroids. J Clin Invest 1994;93:1095–101. Majima M, Hayashi I, Muramatsu M, Katada J, Yamashina S, Katori M. Cyclo-oxygenase-2 enhances basic fibroblast growth factor-induced angiogenesis through induction of vascular endothelial growth factor in rat sponge implants. Br J Pharmacol 2000;130:641–9. Cozzolino F, Torcia M, Aldinucci D, Ziche M, Almerigogna F, Bani D, et al. Interleukin 1 is an autocrine regulator of human endothelial cell growth. Proc Natl Acad Sci U S A 1990;87: 6487–91. Van den Berg WB. Uncoupling of inflammatory and destructive mechanisms in arthritis. Semin Arthritis Rheum 2001;30:7–16. Kimble RB, Matayoshi AB, Vannice JL, Kung VT, Williams C, Pacifici R. Simultaneous block of interleukin-1 and tumor necrosis factor is required to completely prevent bone loss in the early postovariectomy period. Endocrinology 1995;136:3054–61. Xu LX, Kukita T, Nakano Y, Yu H, Hotokebuchi T, Kuratani T, et al. Osteoclasts in normal and adjuvant arthritis bone tissues express the mRNA for both type I and II interleukin-1 receptors. Lab Invest 1996;75:677–87.