• We are sorry, but NCBI web applications do not support your browser and may not function properly. More information
Logo of pnasPNASInfo for AuthorsSubscriptionsAboutThis Article
Proc Natl Acad Sci U S A. Jul 10, 2007; 104(28): 11742–11747.
Published online Jul 3, 2007. doi:  10.1073/pnas.0610812104
PMCID: PMC1913858
Medical Sciences

TNF-induced structural joint damage is mediated by IL-1


Blocking TNF effectively inhibits inflammation and structural damage in human rheumatoid arthritis (RA). However, so far it is unclear whether the effect of TNF is a direct one or indirect on up-regulation of other mediators. IL-1 may be one of these candidates because it has a central role in animal models of arthritis, and inhibition of IL-1 is used as a therapy of human RA. We removed the effects of IL-1 from a TNF-mediated inflammatory joint disease by crossing IL-1α and β-deficient mice (IL-1−/−) with arthritic human TNF-transgenic (hTNFtg) mice. Development of synovial inflammation was almost unaffected on IL-1 deficiency, but bone erosion and osteoclast formation were significantly reduced in IL-1−/−hTNFtg mice, compared with hTNFtg mice based on an intrinsic differentiation defect of IL-1-deficient monocytes. Most dramatically, however, cartilage damage was absent in IL-1−/−hTNFtg mice. Chimera studies revealed that protection of cartilage is based on the loss of IL-1 on hematopoietic, but not mesenchymal, cells, leading to decreased expression of ADAMTS-5 and MMP-3. These data show that TNF-mediated cartilage damage is completely and TNF-mediated bone damage is partially dependent on IL-1, suggesting that IL-1 is a crucial mediator for inflammatory cartilage and bone degradation.

Keywords: cytokines, rheumatoid arthritis, cartilage

Rheumatoid arthritis (RA) is a chronic inflammatory joint disease presenting as symmetric polyarthritis with synovial inflammation, cartilage damage, and bone erosion. New treatment modalities in RA, especially the introduction of systemic cytokine blockade, have clearly shown that proinflammatory cytokines control arthritis. Despite the high complexity of this cytokine network of pro- and antiinflammatory molecules, the pharmacological neutralization of single cytokines achieves dramatic improvement of signs and symptoms of chronic arthritis. This result suggests a certain hierarchy of cytokines, with some of them being the “key players” and others important for the “fine tuning” of inflammation (1). Unraveling the interrelationship of those cytokines pivotally involved in chronic inflammatory disease is thus of key importance to understanding the molecular basis of this complex disorder.

A great number of clinical trials has shown that blockade of TNFα is a central tool to interfere with chronic inflammatory processes such as chronic inflammation in RA (2). These findings show that TNFα is a cytokine of key importance in arthritis. This idea is also supported by the fact that sole overexpression of TNFα in rodents is sufficient to trigger full-blown destructive arthritis with synovial inflammation, cartilage damage, and bone destruction (3). It is unclear, however, whether TNF acts directly or rather orchestrates a cascade of other cytokines that mediate synovial inflammation, cartilage damage, and bone erosion, which together lead to full-blown disease (4). Importantly, cytokines induced by TNF could preferentially mediate structural damage in RA, which leads to loss of joint function due to the destruction of the joint architecture after breakdown of cartilage and bone.

IL-1 is such a candidate cytokine because it is known to closely interact with TNF in various physiological and pathological situations. TNF can induce cellular IL-1 production and vice versa, and IL-1 is able to drive an erosive arthritis in rodents when its antagonist IL-1ra is genetically deleted (5). IL-1 has been introduced as the second cytokine target of antiinflammatory therapy in rheumatic diseases blockade in the clinical setting to treat chronic inflammatory arthritis, and it has been shown as clinically effective in a substantial proportion of RA patients (6). Interestingly, the interaction between TNF and IL-1 is poorly investigated, although blockers of both cytokines are in clinical use. Particularly, it is unclear whether TNF requires IL-1 to induce arthritis or can act independently for IL-1. We hypothesized that IL-1 might act as an important downstream mediator of the proinflammatory and destructive effects of TNF in vivo, and therefore we created mice overexpressing TNF, but completely lacking IL-1. These IL-1−/−hTNFtg mice still developed inflammation, but were completely protected from cartilage breakdown and had much less bone damage than mice with intact IL-1. This finding suggests that IL-1 is essential for inflammatory cartilage damage and a major contributor to inflammatory bone erosion.


No Weight Loss and Milder Arthritis in IL-1−/−hTNFtg Mice.

We interbred hTNFtg mice to develop a destructive form of arthritis, with IL-1a/b−/− mice generating mice that overexpress TNF but completely lack IL-1 (IL-1−/−hTNFtg). hTNFtg, but not IL-1−/−hTNFtg, mice were significantly smaller in size compared with wild-type mice and only slowly increased in body weight (Fig. 1 A and B), indicating that IL-1 deficiency completely blocks development of TNF-mediated wasting. Clinical arthritis started at week 6 in hTNFtg and IL-1−/−hTNFtg mice. Progression of arthritis was significantly more severe in hTNFtg mice compared with IL-1−/−hTNFtg mice (P < 0.05; Fig. 1 C and D). Hence, IL-1 regulates the activity of TNF-mediated arthritis, but does not mediate the entire effect of TNF to the joint.

Fig. 1.
IL-1 impacts clinical manifestations of TNF-induced inflammation. (A) Wild-type, IL-1−/−, hTNFtg, and IL-1−/−hTNFtg mice were photographed at 12 weeks of age. (B–D) Wild-type, IL-1−/−, hTNFtg, and ...

TNF-Induced Synovial Inflammation Almost Develops Normally in the Absence of IL-1.

We next assessed the development of synovitis in IL-1−/−hTNFtg mice by quantitative histological analysis (Fig. 2A). Synovitis was already present in 8-week-old TNFtg mice and steadily increased in size up to week 16. Unexpectedly, synovial inflammation was virtually identical in IL-1−/−hTNFtg mice at weeks 8 and 12. At the late stage of arthritis (week 16), there was reduced synovial inflammation (P < 0.05; hTNFtg vs. IL-1−/−hTNFtg mice). Immunohistochemical analysis revealed no signs of IL-1 expression in the inflammatory synovial tissue of IL-1−/−hTNFtg mice, whereas 15% of the infiltrating inflammatory cells of hTNFtg mice produced IL-1 (Fig. 2B). To determine whether IL-1 deficiency affects the trafficking of immune cells in inflammatory arthritis, cellular distribution of hematopoietic cells in the spleen, bone marrow, and joints was analyzed. Neither the overexpression of TNF nor the deficiency of IL-1 affected cellular distribution in these compartments [supporting information (SI) Fig. 6 A and B].

Fig. 2.
TNF-induced synovitis develops despite IL-1 deficiency. (A) Representative H&E-stained sections of hind paws from wild-type, IL-1−/−, hTNFtg, and IL-1−/−hTNFtg are shown. (Magnification: ×100.) (B) Mean ...

Partial Protection from TNF-Mediated Bone Erosion in the Absence of IL-1.

We next assessed the impact of IL-1 on TNF-mediated inflammatory bone destruction and analyzed the extent of arthritic bone damage in hTNFtg and IL-1−/−hTNFtg mice. MicroCT imaging of joints revealed bone erosions in hTNFtg mice, whereas the joints of IL-1−/−hTNFtg mice appeared structurally preserved (Fig. 3A). To quantify the extent of bone damage, we performed histological analyses of joints (Fig. 3B). hTNFtg mice developed severe bone erosions between weeks 6 and 8. Although IL-1−/−hTNFtg mice were not completely protected from bone erosions, structural damage to bone was <50% of that observed in hTNFtg mice and was associated with blunted osteoclastogenesis in the synovial tissue of IL-1−/−hTNFtg mice (Fig. 3 C and D; P < 0.05). Importantly, bone erosion was also significantly diminished at disease stages with identical inflammation (week 12).

Fig. 3.
IL-1 partially protects from TNF-induced local bone loss. (A) MicroCT-based reconstruction of knee (Right) and carpal (Left) joints from wild-type, IL-1−/−, hTNFtg, and IL-1−/−hTNFtg mice. (B–D) Histological sections ...

Decreased Receptor Activator of NF-κB Ligand (RANKL) Responsiveness in IL-1-Deficient Mononuclear Cells.

To determine the role of IL-1 in inflammatory osteoclastogenesis in vitro, osteoclast differentiation by RANKL was studied in the various genotypes (Fig. 3E). Osteoclastogenesis in wild-type mice was 3-fold higher than in IL-1−/− mice (P < 0.05). Cells from hTNFtg cells displayed a significantly enhanced osteoclastogenic response to macrophage colony-stimulating factor (M-CSF) and RANKL, whereas this enhanced response was completely reversed to wild-type levels in IL-1−/−hTNFtg mice (P < 0.05). This result indicates a general decrease in responsiveness of IL-1-deficient mononuclear cells toward RANKL, but maintained responsiveness toward TNF. In line, osteoclast formation in IL-1−/−hTNFtg mice was still significantly increased compared with IL-1−/− mice, and the exogenous addition of TNF could also significantly enhance differentiation of IL-1−/− cells (Fig. 3F). Analysis of the resorptive properties of osteoclasts from all four genotypes was similar to the results obtained for osteoclast differentiation, suggesting that these osteoclasts are functional, and the differences among genotypes are based on their differentiation profile. Markers of osteoclast-mediated bone resorption were significantly lower in mice lacking IL-1 both in cell culture supernatants of cultivated osteoclasts as well as systemically in the serum of mice (SI Fig. 7).

Molecular characterization of osteoclast precursors revealed a decreased expression of RANK as the underlying cause for a blunted response of IL-1-deficient cells to RANKL, whereas other surface and functional markers such as c-fms, cathepsin K, TRAP, and MMP9 were similarly expressed (Fig. 3G). Only expression of the calcitonin receptor as a marker of late osteoclast differentiation was impaired in IL-1−/− cells, which underlines the differentiation deficits of these cells. Moreover, analysis of RANK-dependent signaling cascades revealed an impaired activation of Akt and ERK pathway in IL-1-deficient cells, whereas p38 MAPK and JNK activation was normal.

IL-1 Is Essential for Cartilage Destruction in TNF-Mediated Arthritis.

We next turned to the role of IL-1 in structural damage of cartilage. Nonarthritic wild-type and IL-1−/− mice had minimal signs of proteoglycan loss throughout the observation period. hTNFtg mice progressively lost proteoglycan from articular cartilage between weeks 8 and 12 (Fig. 4 A and B). In contrast to hTNFtg mice, however, IL-1−/−hTNFtg mice exhibited a completely different phenotype, showing full protection of cartilage even at late stages of arthritis (P < 0.05). When assessing chondrocyte death by assessing empty cartilage lacunae, similar results were found (Fig. 4C). Nonarthritic genotypes did not experience significant chondrocyte apoptosis, whereas arthritis led to significant chondrocyte loss in hTNFtg mice. In IL-1−/−hTNFtg mice, no relevant chondrocyte loss was found even at the late stage of disease. Also the functional properties of remaining chondrocytes were maintained in IL-1−/−hTNFtg mice, as evident from the analysis of matrix macromolecule synthesis of cartilage explants (Fig. 4D). Cartilage from wild-type and IL-1−/− mice showed similar metabolic activity, whereas it was strongly diminished in hTNFtg mice. On the contrary, metabolic activity of IL-1−/−hTNFtg mice was comparable to wild-type and IL-1−/− mice, suggesting an intact function of chondrocytes in these arthritic IL-1-deficient animals.

Fig. 4.
TNF-mediated cartilage is completely IL-1-dependent. (A) Representative toluidine blue-stained sections from hind paws of wild-type, IL-1−/−, hTNFtg, and IL-1−/−hTNFtg mice are shown. (Magnification: ×100.) Reduced ...

We next asked whether a decreased activity of matrix-degrading enzymes underlies the structural protection of cartilage in IL-1-deficient mice. Analysis of matrix metalloproteinase (MMP)-induced VDIPEN neoepitope formation revealed complete absence of VDIPEN expression in IL-1−/−hTNFtg, similar to the results obtained for wild-type and IL-1−/− mice. In contrast, VDIPEN expression was highly positive in hTNFtg mice (P < 0.05 vs. other groups), suggesting that IL-1 is essentially required for inflammatory cartilage damage because of its regulatory function on matrix-degrading enzymes (Fig. 4E).

Cartilage Damage Depends on Hematopoietic, but Not Mesenchymal, IL-1 Production.

To dissect whether bone marrow or the mesenchyme is the source of IL-1 critical for the detrimental effects of TNF on cartilage, we generated radiation chimeras of TNFtg and IL-1−/−hTNFtg mice (Fig. 5A). hTNFtg mice challenged with hTNFtg bone marrow served as a positive control and developed arthritis with severe cartilage damage (Fig. 5 B and C). IL-1−/−hTNFtg mice challenged with IL-1−/−hTNFtg bone marrow served as a negative control and developed synovial inflammation, but only minimal signs of cartilage damage. Moreover, hTNFtg mice challenged with IL-1−/−hTNFtg bone marrow were also protected from cartilage damage, suggesting a critical function of IL-1 expression from hematopoietic cells (P < 0.05 vs. positive control). However, IL-1−/−hTNFtg mice challenged with hTNFtg bone marrow developed almost full-blown cartilage damage similar to the positive control. This result suggests that mesenchymal IL-1 production is not a driving force in cartilage damage. Consistent with this hypothesis, immunohistochemical stainings for IL-1 in chimeras showed exclusive expression of IL-1 on hematopoietic cells, which was completely absent in chimeras of hTNFtg hosts and IL-1-deficient donors (SI Fig. 8).

Fig. 5.
Cartilage damage depends on IL-1 signaling from hematopoietic cells. (A) Development of cartilage damage was assessed in radiation chimeras 7 weeks after bone marrow transplantation. Representative toluidine blue-stained sections. (Magnification: ×100.) ...

Defective Regulation and Expression of Aggrecanases and MMPs by Hematopoietic Cells from IL-1−/−hTNFtg Mice.

Next we addressed the molecular differences of mesenchymal and hematopoietic cells on IL-1-deficiency. Isolated synovial fibroblasts from hTNFtg mice showed a higher proliferation rate but also a higher rate of spontaneous and inducible apoptosis compared with cells from IL-1−/−hTNFtg mice (SI Fig. 9 A and B). This finding suggests that the net effect on overall turnover of synovial mesenchymal cells is balanced among the two genotypes. Moreover, secretion of MMP type 3 (MMP-3) was not significantly different among the two groups (SI Fig. 9C), indicating that cartilage protection in IL-1−/−hTNFtg is not based on functional defects of mesenchymal cells. In contrast to mesenchymal cells, profound differences in the function of mononuclear cells were evident. Monocytes from wild-type, but not IL-1−/−, mice provoked a strong up-regulation of MMP-3, MMP-13, and ADAMTS-5 in chondrocytes, whereas expression of ADAMTS-4 and TIMP-1 were not different (SI Fig. 10). In addition, MMP-3, MMP-13, and ADAMTS-5 expression was also impaired in monocytes from IL-1−/− mice and were associated with a blunted expression of c-jun and fra-1 proteins, which are key signal transduction elements for the expression of MMPs. Other compounds of the AP-1 transcription factor complex were identically expressed among the two genotypes (SI Fig. 11).


Based on several experimental models of arthritis, TNF and IL-1 are considered to be key players in inflammatory joint disease. This key role is also reflected by the introduction of a TNF and IL-1 blockade into the clinical therapy of RA and other inflammatory diseases (2). The actual role of TNF and IL-1 in the disease process may differ among individuals. Likewise, different models of arthritis favor TNF or IL-1 as the main trigger of disease (79). In key models of inflammatory arthritis, such as collagen-induced arthritis, serum transfer arthritis, and IL-1ra-deficient mice, IL-1 is the pivotal cytokine blockade that leads to complete suppression of disease (5, 10, 11). However, there are also strictly TNF-driven models of arthritis, which closely mimic human RA (3). It is unclear so far whether there are common pathophysiological mechanisms in TNF- and IL-1-mediated arthritis (12, 13), which affords a detailed analysis of the various pathophysiological processes in arthritis: synovial inflammation, cartilage damage, and bone erosion.

Surprisingly, TNF could induce joint inflammation in the absence of IL-1. IL-1−/−hTNFtg mice developed full-blown synovial inflammation with an identical distribution of disease as found in hTNFtg mice, suggesting that IL-1 is dispensable for the recruitment of inflammatory cells into the joints. Because a pharmacological blockade of IL-1 in TNF-driven models of arthritis, such as hTNFtg mice and adjuvant-induced arthritis, has given inconsistent results and has only poorly defined the role of IL-1 as a downstream mediator of TNF-induced inflammation in vivo, this genetic study gives more detailed insight into the interplay between TNF and IL-1 (1416). These pharmacological approaches also left open the possibility that an IL-1 blockade could be suboptimal and the remaining IL-1 could drive inflammation. This possibility has now been formally ruled out by these experiments because TNF can drive joint inflammation even in the complete absence of IL-1, suggesting that IL-1 blockers could easily fail to control inflammation in patients with TNF-mediated disease. In fact, this is well in line with the rather low efficacy of IL-1 blockade in diseases like RA and ankylosing spondylitis (17, 18).

TNF is a potent driver of osteoclastogenesis and bone resorption and acts indirectly by inducing RANKL on mesenchymal cells and activated T cells and directly through TNF-receptor type 1 on osteoclasts (19, 20). RANKL expression by synovial fibroblast-like as well as activated T cells appears crucial for the propensity of synovial tissue to generate osteoclasts, which is further enhanced by proinflammatory cytokine expression in the joint (21). IL-1 can also regulate osteoclast formation, and recently IL-1-dependent and IL-1-independent pathways have been demonstrated for inflammatory osteoclastogenesis. It has also been shown that TNF can induce IL-1 via p38 MAPK activation in bone marrow stromal cells and preosteoclasts, suggesting that IL-1 might truly modulate TNF-mediated osteoclastogenesis in vivo (22).

Bone erosions and osteoclast formation were significantly reduced in IL-1−/−hTNFtg mice by >50%, compared with hTNFtg mice with intact IL-1. We found a decreased differentiation and resorption potential of osteoclast precursors in the absence of IL-1. Molecularly, this intrinsic defect of mononuclear cells to differentiate in osteoclasts was based on reduced RANK expression, which makes them more resistant to RANKL stimulation. This result is well in line with previous reports showing an impaired osteoclast differentiation of mononuclear cells from IL-1R−/− mice on challenge with RANKL (23). Regulation of osteoclastogenesis through RANK is also known by MCSF, which increases RANK expression on osteoclast precursors (24). In contrast, IL-1-deficient cells were still responsive to TNF, emphasizing a selective modulatory role of IL-1 in RANKL–RANK interaction. Because TNF responsiveness was maintained in IL-1-deficient osteoclast precursors, it appears conceivable that osteoclasts could still form in the inflamed joints of IL-1−/−hTNFtg mice because TNF expression is most pronounced at these inflammatory sites (25). Still the osteoclast-forming potential of synovial tissue from IL-1−/−hTNFtg mice was dramatically reduced because precursor cells had a decreased susceptibility toward RANKL stimulation.

Arthritis is a major trigger for proteoglycan loss and cartilage damage. Current concepts of cartilage damage suggest that enzymes such as metalloproteinases and aggrecanases play a key role in matrix degradation, causing irreversible damage and structural disintegration of the cartilage (26, 27). In vitro and in vivo studies have suggested that IL-1 is a major driver of cartilage damage and even exceeds the effects of TNF (28). Animal models that show a particular high level of cartilage destruction, like collagen-induced arthritis, are highly IL-1-dependent (7). Cartilage was completely resistant to damage in IL-1−/−hTNFtg mice, which was intriguing given the highly destructive phenotype of TNFtg mice, the persistence of inflammation in this model despite complete absence of IL-1, and the full protection of cartilage, which did not even exhibit relevant proteoglycan loss. The absence of MMP-induced neoepitope formation did suggest that IL-1−/−hTNFtg mice show deficits in the synthesis of matrix-degrading enzymes. This lack of sufficient MMP and aggrecanase production is crucial for understanding the apparent uncoupling between inflammation and cartilage damage in IL-1−/−hTNFtg mice.

Because IL-1 can potentially induce the production of matrix-degrading enzymes in mesenchymal cells such as chondrocytes and fibroblasts, as well as in hematopoietic cells, it appeared crucial to dissect the relative contribution of the mesenchymal and inflammatory cells entering the joint in cartilage pathology (29). Surprisingly, creation of chimeric animals did suggest that IL-1 is virtually exclusively produced by hematopoietic, but not mesenchymal, cells, and that its production in hematopoietic cells was crucial for cartilage damage. This finding suggests that IL-1 may have a key regulatory role in the production of matrix enzymes by inflammatory cells. Interestingly, depletion of hematopoietic cells in osteoarthritis models by clodronate liposomes has also shown to protect from cartilage damage. Although the molecular mechanism is unclear, depletion of hematopoietically derived IL-1 might cause this finding. Isolated monocytes lacking IL-1 showed a defective production of MMP-3 and ADAMTS5 and were not capable of driving chondrocytic production of MMPs and ADAMTs. Molecularly, this defective aggrecanase and MMP production is associated with defective expression of c-jun and fra-1 by hematopoietic cells. Both genes are part of the AP-1 transcription factor complex. Fra proteins actually belong to the fos family of transcription factors, which built up the AP-1 complex on engaging jun proteins. Interestingly, among this family of transcription factors, c-jun proteins are particularly responsible for MMP production, and defective expression in IL-1−/− mice likely explains the impaired enzyme production (30).

In summary, TNF is sufficient to drive inflammation without any IL-1 in vivo, but completely fails to degrade cartilage and is severely impaired to drive bone resorption. This result suggests IL-1 is an essential mediator of the TNF-induced bone and cartilage damage.

Materials and Methods

For more details regarding the following procedures, see SI Materials and Methods.

Animals and Clinical Assessment.

All four genotypes [wild-type, IL-1−/− (31), hTNFtg (3), and IL-1−/−hTNFtg mice] of the F2 generation were comparatively analyzed in this study (n = 100). Clinical signs of arthritis and body weight were determined once weekly. The local ethics committee approved all animal procedures.


Quantitative real-time RT-PCR was performed by using LightCycler technology (Roche Diagnostics, Indianapolis, IN). Primer sequences are published in SI Materials and Methods.

FACS Analysis.

For immunophenotyping, cells were isolated from the spleen and the tibial and femoral bones (n = 4–5 per group); stained with antibodies against T cells (CD4, CD8), B cells (CD45R), and monocytes (CD11b) (BD Biosciences, San Jose, CA).

MicroCT Imaging.

MicroCT images of knee joints and hind paws were acquired on a vivaCT40 (Scanco, Bassersdorf, Switzerland).

Joint Histology.

Serial paraffin sections (2 μm) of hind paws were stained with H&E, toluidine blue, and tartrate-resistant acid phosphatase (TRAP). For immunohistochemistry (n = 4 per group), tissue sections were stained with monoclonal antibodies against surface markers of neutrophil granulocytes, macrophages, T- and B-lymphocytes, as well as polyclonal antibodies against IL-1 or VDIPEN.

Generation of Radiation Chimera.

Recipient 4-week-old hTNFtg and IL-1−/−hTNFtg mice received a lethal whole-body irradiation of 11 Gy. Bone marrow cells were isolated from long bones of age-matched donor hTNFtg or IL-1−/−hTNFtg mice, and 5 × 106 bone marrow cells were injected i.v. in recipient mice.

Isolation of Bone Marrow Macrophages.

Bone marrow cells were stimulated with M-CSF (25 ng/ml) for 4 days, floating cells were discarded, and adherent cells were used for assays. For immunoblotting analysis, protein extracts from RANKL-stimulated (50 ng/ml) cells were separated on SDS/12% PAGE and stained with antibodies against phosphorylated and total p38 MAPK, JNK, and ERK (Cell Signaling Technology, Beverly, MA).

Generation of Osteoclasts and Osteoclast Analyses.

Spleen cells were cultured overnight in αMEM containing 10% FBS and 30 ng/ml M-CSF. Nonadherent cells were harvested and cultured with 30 ng/ml M-CSF and 50 ng/ml RANKL (R&D Systems, Minneapolis, MN). TRAP staining was performed to identify osteoclasts. Bone resorption was carried out on 0.4-mm-thick bovine cortical slices.

Synovial Fibroblast Analyses.

Synovial tissue (n = 3 per group) was minced, digested, and cultivated in DMEM supplemented with 10% FBS. Cell numbers were counted after various periods of time to show the proliferative potential of cells. Apoptosis was determined by Apo-One Homogenous Caspase 3/7 Assay (Promega, Madison, WI).

Cartilage Explant Cultures and Macromolecule Biosynthesis.

Cartilage tissue (n = 3–5 per group) was obtained from knee joints and labeled with 20 μCi/ml of [35S]sulfate (Amersham Biosciences, Buckinghamshire, U.K.). Unincorporated isotope was removed by using Sephadex G-25 (Amersham Pharmacia Biotech, NJ) gel chromatography.

Primary Chondrocytes and Cellular Cocultures.

Primary mouse chondrocytes were prepared from ventral rib cages. For coculture experiments, primary monocytes were isolated from bone marrow of wild-type and IL-1−/− mice; 1 × 106 monocytes per ml were placed in transwell inserts, incubated with TNF (10 ng/ml). Transwells with adherent monocytes were then cocultured with confluent primary chondrocytes. Total RNA from chondrocytes was isolated from cell cultures by TRIzol reagent.

Statistical Analysis.

Data are shown as means ± SEM. Group mean values of histological data were compared by paired Student′s t test. For comparison of clinical assessments, a nonparametric Wilcoxon signed-rank test was used.

Supplementary Material

Supporting Information:


We thank Dr. Yoichiro Iwakura (Center for Experimental Medicine, Institute of Medical Science, University of Tokyo, Tokyo, Japan) for kindly providing the IL-1a/b−/− mice; Dr. George Kollias (Biomedical Sciences Research Center, Vari, Greece) for providing the hTNFtg mice; and Birgit Türk, Ivana Mikulic, Carl-Walter Steiner, and Margarete Tryniecki for their excellent technical assistance. This work was supported by Austrian Science Fund Grant P18223 (to K.R.), an Austrian Science Fund START prize (to G.S.), Interdisziplinäres Zentrum für Klinische Forschung Erlangen Project 22, the Doerenkamp Professorship in Innovations in Animal and Consumer Protection (to K. Brune, University of Erlangen–Nuremburg), and Deutsche Forschungsgemeinschaft Grant FOR 661 (to G.S.).


macrophage colony-stimulating factor
matrix metalloproteinase
rheumatoid arthritis
receptor activator of NF-κB ligand
tartrate-resistent acidic phosphatase.


The authors declare no conflict of interest.

This article is a PNAS Direct Submission.

This article contains supporting information online at www.pnas.org/cgi/content/full/0610812104/DC1.


1. McInnes IB, Schett G. Nat Rev Immunol. 2007;7:429–442. [PubMed]
2. Gartlehner G, Hansen RA, Jonas BL, Thieda P, Lohr KN. J Rheumatol. 2006;33:2398–2408. [PubMed]
3. Keffer J, Probert, Cazlaris H, Georgopoulos S, Kaslaris E, Kioussis D, Kollias G. EMBO J. 1991;10:4025–4031. [PMC free article] [PubMed]
4. Brennan FM, Chantry D, Jackson A, Maini R, Feldmann M. Lancet. 1989;2:244–247. [PubMed]
5. Horai R, Saijo S, Tanioka H, Nakae S, Sudo K, Okahara A, Ikuse T, Asano M, Iwakura Y. J Exp Med. 2000;191:313–320. [PMC free article] [PubMed]
6. Bresnihan B, Alvaro-Gracia JM, Cobby M, Doherty M, Domljan Z, Emery P, Nuki G, Pavelka K, Rau R, Rozman B, et al. Arthritis Rheum. 1998;41:2196–2204. [PubMed]
7. Joosten LA, Helsen MM, Saxne T, van De Loo FA, Heinegard D, van Den Berg WB. J Immunol. 1999;163:5049–5055. [PubMed]
8. Feige U, Hu YL, Gasser J, Campagnuolo G, Munyakazi L, Bolon B. Cell Mol Life Sci. 2000;57:1457–1470. [PubMed]
9. Redlich K, Hayer S, Maier A, Dunstan CR, Tohidast-Akrad M, Lang S, Turk B, Pietschmann P, Woloszczuk W, Haralambous S, et al. Arthritis Rheum. 2002;46:785–792. [PubMed]
10. Ji H, Pettit A, Ohmura K, Ortiz-Lopez A, Duchatelle V, Degott C, Gravallese E, Mathis D, Benoist C. J Exp Med. 2002;196:77–85. [PMC free article] [PubMed]
11. Saijo S, Asano M, Horai R, Yamamoto H, Iwakura Y. Arthritis Rheum. 2002;46:533–544. [PubMed]
12. van den Berg WB. Curr Rheumatol Rep. 2002;4:232–239. [PubMed]
13. Dinarello CA. Clin Exp Rheumatol. 2002;20:S1–S13. [PubMed]
14. Zwerina J, Hayer S, Tohidast-Akrad M, Bergmeister H, Redlich K, Feige U, Dunstan C, Kollias G, Steiner G, Smolen J, et al. Arthritis Rheum. 2004;50:277–290. [PubMed]
15. Schett G, Middleton S, Bolon B, Stolina M, Brown H, Zhu L, Pretorius J, Zack DJ, Kostenuik P, Feige U. Arthritis Rheum. 2005;52:1604–1611. [PubMed]
16. Probert L, Plows D, Kontogeorgos G, Kollias G. Eur J Immunol. 1995;25:1794–1797. [PubMed]
17. den Broeder AA, de Jong E, Franssen MJ, Jeurissen ME, Flendrie M, van den Hoogen FH. Ann Rheum Dis. 2006;65:760–762. [PMC free article] [PubMed]
18. Haibel H, Rudwaleit M, Listing J, Sieper J. Ann Rheum Dis. 2005;64:296–298. [PMC free article] [PubMed]
19. Lam J, Takeshita S, Barker JE, Kanagawa O, Ross FP, Teitelbaum SL. J Clin Invest. 2000;106:1481–1488. [PMC free article] [PubMed]
20. Abu-Amer Y, Erdmann J, Alexopoulou L, Kollias G, Ross FP, Teitelbaum SL. J Biol Chem. 2000;275:27307–27310. [PubMed]
21. Kong YY, Feige U, Sarosi I, Bolon B, Tafuri A, Morony S, Capparelli C, Li J, Elliott R, McCabe S, et al. Nature. 1999;402:304–309. [PubMed]
22. Wei S, Kitaura H, Zhou P, Ross FP, Teitelbaum SL. J Clin Invest. 2005;115:282–290. [PMC free article] [PubMed]
23. Lee SK, Gardner AE, Kalinowski JF, Jastrzebski SL, Lorenzo JA. Bone. 2006;38:678–685. [PubMed]
24. Kitaura H, Zhou P, Kim HJ, Novack DV, Ross FP, Teitelbaum SL. J Clin Invest. 2006;115:3418–3427. [PMC free article] [PubMed]
25. Butler DM, Malfait AM, Mason LJ, Warden PJ, Kollias G, Maini RN, Feldmann M. J Immunol. 1997;159:2867–2876. [PubMed]
26. Goldring SR. Rheumatology. 2003;42(Suppl 2):11–16. [PubMed]
27. van Meurs JB, van Lent PL, Holthuysen AE, Singer II, Bayne EK, van den Berg WB. Arthritis Rheum. 1999;42:1128–1139. [PubMed]
28. van de Loo FA, Joosten LA, van Lent PL, Arntz OJ, van den Berg WB. Arthritis Rheum. 1995;38:164–172. [PubMed]
29. Burrage PS, Mix KS, Brinckerhoff CE. Front Biosci. 2006;11:529–543. [PubMed]
30. Hosseini M, Rose AY, Song K, Bohan C, Alexander JP, Kelley MJ, Acott TS. Invest Ophthalmol Vis Sci. 2006;47:1469–1476. [PubMed]
31. Horai R, Asano M, Sudo K, Kanuka H, Suzuki M, Nishihara M, Takahashi M, Iwakura Y. J Exp Med. 1998;187:1463–1475. [PMC free article] [PubMed]

Articles from Proceedings of the National Academy of Sciences of the United States of America are provided here courtesy of National Academy of Sciences
PubReader format: click here to try


Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...


Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...