Tenosynovitis and osteoclast formation as the initial preclinical changes in a murine model of inflammatory arthritis.код для вставкиСкачать
ARTHRITIS & RHEUMATISM Vol. 56, No. 1, January 2007, pp 79–88 DOI 10.1002/art.22313 © 2007, American College of Rheumatology Tenosynovitis and Osteoclast Formation as the Initial Preclinical Changes in a Murine Model of Inflammatory Arthritis Silvia Hayer,1 Kurt Redlich,1 Adelheid Korb,1 Sonja Hermann,2 Josef Smolen,1 and Georg Schett3 Objective. To determine the nature of the initial changes of joint inflammation occurring before, at the time of, and shortly after onset of clinically apparent arthritis. Methods. Human tumor necrosis factor (TNF)– transgenic mice were assessed for clinical, histologic, immunophenotypic, serologic, and molecular changes at the preclinical phase of arthritis, at the onset of disease, and at the stage of early clinical disease. In addition, the effects of a genetic osteoclast deficiency and pharmacologic inhibition of TNF were studied in these initial phases of disease. Results. Initial articular changes were observed even before the start of clinical symptoms. Infiltration of the tendon sheaths by granulocytes and macrophages as well as formation of osteoclasts next to the inflamed tendon sheaths were the first pathologic events. Tenosynovitis rapidly led to remodeling of the sheaths into pannus-like tissue, which formed osteoclasts that invaded the adjacent mineralized cartilage. Early lesions were associated with up-regulation of interleukin-1 (IL-1) and IL-6 as well as activation of p38 MAPK and ERK. In contrast, absence of osteoclasts led to uncoupling of tenosynovitis from invasion into cartilage and bone. TNF blockade also attenuated the pathologic changes associated with tenosynovitis. Conclusion. Structural damage begins even before the onset of clinical symptoms of arthritis and involves the tendon sheaths as well as adjacent cartilage and bone. These results suggest that tenosynovitis is an initiating feature of arthritis and that joint destruction starts right from the onset of disease. Our findings thus underscore the importance of immediate initiation of an effective therapy in patients with rheumatoid arthritis. Rheumatoid arthritis (RA) is a systemic autoimmune disease characterized by synovial hyperplasia, cartilage damage, and bone erosions, leading to progressive disability. Information on the character of the initial destructive events in arthritis is limited, since the affected structures are not directly accessible in patients with very early disease, and imaging techniques such as radiography and magnetic resonance imaging (MRI) do not provide insight into the cellular processes involved in the initial destructive process. Thus, it is unclear whether joint destruction accompanies inflammation of the joints right from disease onset or whether there is a certain lag period between onset of inflammation and structural damage. The latter has been supported by evidence of synovial inflammation occurring even in the absence of clinically apparent symptoms of arthritis (1). Chronic arthritic diseases such as RA usually lead to joint damage after a short period of time. Even comparatively insensitive techniques such as plain radiography show destructive changes within 6 months from disease onset in more than 50% of patients with RA (2). The destructive processes in RA depend on a complex interplay of synovial fibroblasts, lymphocytes, macrophages, and monocyte-derived osteoclasts (3,4). The documented invasive properties of synovial fibroblasts Dr. Schett’s work was supported by a START Program award from the Austrian Science Fund. 1 Silvia Hayer, PhD, Kurt Redlich, MD, Adelheid Korb, MD, Josef Smolen, MD: Medical University of Vienna, Vienna, Austria; 2 Sonja Hermann, MD: University of Erlangen-Nuremberg, Erlangen, Germany; 3Georg Schett, MD: Medical University of Vienna, Vienna, Austria, and University of Erlangen-Nuremberg, Erlangen, Germany. Address correspondence and reprint requests to Georg Schett, MD, Department of Internal Medicine III and Institute for Clinical Immunology, University of Erlangen-Nuremberg, Krankenhausstrasse 12, D-91054 Erlangen, Germany. E-mail: georg.schett@ med3.imed.uni-erlangen.de. Submitted for publication June 28, 2006; accepted in revised form September 29, 2006. 79 80 HAYER ET AL include production of tissue-degrading enzymes such as cathepsins and matrix metalloproteinases (5). In addition, macrophages, activated synovial fibroblasts, and lymphocytes contribute to this process by synthesizing proinflammatory cytokines and chemokines. Tumor necrosis factor (TNF) is a key cytokine in RA (6), and selective overexpression of TNF in mice is sufficient for the development of destructive arthritis (7). Of note, synovial fibroblasts and lymphocytes drive the differentiation of osteoclasts from monocyte precursor cells via the expression of RANKL (8–11). Osteoclasts are cells that are specifically designed to resorb bone (12). The purpose of this study was to investigate the nature and timing of joint destruction in chronic arthritis. We hypothesized that microscopic signs of joint destruction are present as early as the onset of clinically apparent disease. To address this question, we used human TNF-transgenic (HuTNF-Tg) mice as an experimental arthritis model. This model is characterized by 1) a chronic progressive course of disease, 2) a symmetric polyarthritis with predominant involvement of the small joints, and 3) a destructive phenotype, all of which are features that closely resemble those of human RA. Structural changes in the joints were assessed at the stage of preclinical disease, at clinical onset of disease, and during early disease. Alterations in the periarticular tissue, such as the tendons, as well as in the articular cartilage and bone were assessed during these 3 initial phases of arthritis. Our results show that tenosynovitis is the first pathologic event manifest in arthritis. Moreover, tenosynovitis precipitates the rapid destruction of the adjacent juxtaarticular bone even in these very early stages of disease. MATERIALS AND METHODS Animals. Heterozygous HuTNF-Tg mice (Tg197 strain, C57BL/6 genetic background) spontaneously develop inflammatory, destructive arthritis upon constitutive HuTNF overexpression. Both the genotype and phenotype of these mice have been described previously in more detail (7). Briefly, the animals develop joint swelling in the small joints of the front and hind paws between weeks 5 and 6 after birth. This chronic polyarthritis continuously progresses over time and finally results in severe joint damage. In this experiment, young inbred littermates were assessed weekly for the onset of clinical symptoms of arthritis starting 3 weeks after birth. Based on the onset of paw swelling and loss of grip strength, early arthritis in 3–6-week-old HuTNF-Tg mice was categorized in 1 of the 3 stages, as follows: 1) the preclinical stage, with no signs of joint swelling or loss of grip strength, 2) the disease onset stage, characterized by the appearance of the first clinical signs of arthritis, and 3) the early arthritis stage, characterized by signs and symptoms of the disease 1 week after disease onset. Each of the groups comprised 6 animals. Moreover, 6 additional HuTNF-Tg mice were treated with anti-TNF antibodies (infliximab; Centocor, Leiden, The Netherlands) at 10 mg/kg 3 times per week from week 4 to week 6. Six c-fos⫺/⫺ ⫻ HuTNF-Tg mice, which are completely deficient in osteoclast formation, were also studied. (Generation of c-fos ⫺/⫺ ⫻ HuTNF-Tg mice has been described previously .) All animals were killed by cervical dislocation and the blood was withdrawn by heart puncture. Animal procedures were approved by the local ethics committee. Clinical assessment. Clinical signs of arthritis, including grip strength and paw swelling, were assessed weekly in the mice starting 3 weeks after birth. Paw swelling was assessed by using a well-established semiquantitative score: 0 ⫽ no swelling, 1 ⫽ mild swelling of the toes and ankle, 2 ⫽ moderate swelling of the toes and ankle, and 3 ⫽ severe swelling of the toes and ankle. Grip strength was also recorded using a semiquantitative score: 0 ⫽ normal grip strength, ⫺1 ⫽ mildly reduced grip strength, ⫺2 ⫽ severely reduced grip strength, and ⫺3 ⫽ no grip at all. Conventional histology and immunohistochemistry. Decalcified paraffin-embedded 2-m sections of the paws were stained with hematoxylin and eosin for evaluation of inflammation, with toluidine blue for cartilage proteoglycan loss, and with tartrate-resistant acid phosphatase (TRAP) (leukocyte staining kit; Sigma, St. Louis, MO) for the amount of osteoclasts and severity of bone erosion. To investigate cell populations, immune phenotyping was carried out on T cells (anti-CD3, 1:50 dilution; Novo Castra Laboratories, Newcastle, UK), B cells (anti-CD45 receptor, 1:200 dilution; BD Biosciences PharMingen, San Diego, CA), neutrophils (clone 7/4, 1:2,000 dilution; Serotec, Oxford, UK), and macrophages (clone F4/80, 1:300 dilution; Serotec). For further immunohistochemical analyses, sections were stained with interleukin-1 (IL-1) antibodies (1:25 dilution; R&D Systems, Minneapolis, MN) and IL-6 antibodies (1:25 dilution; Acris Antibodies, Hiddenhausen, Germany). Sections were pretreated with proteinase K (0.5 g/ml for 5 minutes at 37°C; Roche Diagnostics, Mannheim, Germany) for granulocytes, macrophages, T cells, and IL-6 staining, or were pretreated with heat (for 20 minutes at 96°C) for IL-1 staining, or were left untreated for B cell staining. Blocking of endogenous peroxidase was done with 0.3–3% hydrogen peroxide in phosphate buffered saline for 10 minutes. Sections were then incubated for 30 minutes with biotinylated species-specific secondary antibodies as follows: rabbit anti-rat antibody for the detection of granulocytes, macrophages, B lymphocytes, and T lymphocytes, goat antirabbit antibody for the detection of IL-6, and rabbit anti-goat antibody for the detection of IL-1 (all from Vector Laboratories, Burlingame, CA). Subsequently, sections were incubated for 10 minutes with StreptABComplex/horseradish peroxidase (HRP) (Dako, Glostrup, Denmark) for the detection of macrophages and B lymphocyte, T lymphocyte, IL-1, and IL-6 staining, or with avidin–biotin–HRP complex (Vectastain ABC kit; Vector Laboratories) for the detection of granulocytes. Antigen-expressing cells were visualized by incubation with 3,3-diaminobenzidine (Sigma), resulting in brown staining. Areas of synovial inflammation, bone erosion, cartilage damage, and bone marrow infiltrates were quantified by PRECLINICAL TENOSYNOVITIS AND OSTEOCLAST FORMATION 81 Figure 1. Definition of different phases of very early inflammatory arthritis and characterization of inflammation. A and B, Based on the severity of paw swelling (A) and reduction in grip strength (B), 3 phases of very early arthritis were defined: 1) the preclinical stage, before the onset of clinical symptoms of arthritis (weeks 3–4), 2) the clinical onset of arthritis, with start of swelling and deterioration of grip strength (weeks 4–5), and 3) early arthritis, up to 1 week after disease onset (weeks 5–6). Results are the mean and SEM in human tumor necrosis factor–transgenic (HuTNF-Tg) mice versus wild-type (WT) mice. C and D, Hematoxylin and eosin–stained paw sections were assessed for signs of inflammation in wild-type as well as HuTNF-Tg mice in various different phases of early disease, characterized by pannus formation in the hind paws at the onset of inflammation (C) and tenosynovitis of the long peroneal muscle (D). Boxes in C are shown at higher magnification in D. (Original magnification ⫻ 25 in C; ⫻ 100 in D.) histomorphometry using an Axioskop 2 microscope (Zeiss, Oberkochen, Germany) and OsteoMeasure Analysis software (OsteoMetrics, Decatur, GA). Analysis of kinase activity by Western blotting. Protein extracts were prepared from the hind paws of 9 HuTNF-Tg mice (3 from each time point) and 3 wild-type mice by mincing the tissue in lysis buffer containing 20 mM HEPES, pH 7.9, 0.4M NaCl, 1.5 mM MgCl2, 1 mM dithiothreitol, 1 mM EDTA, 0.1 mM EGTA, 20% glycerol, and proteinase and phosphatase inhibitors (both from Sigma) with an Ultra-Thurrax homogenizer. The extracts were centrifuged for 15 minutes at 14,000 revolutions per minute, separated by sodium dodecyl sulfate– polyacrylamide gel electrophoresis, and transferred onto nitrocellulose membranes. After blocking, membranes were stained with polyclonal antibodies against the phosphorylated forms of p38 MAPK, ERK (p44/42 MAPK), and SAPK/JNK (all from Cell Signaling, Beverly, MA). For control purposes, antibodies against total p38 MAPK, ERK, SAPK/JNK (Cell Signaling), and actin (Sigma) were used. Detection was performed using HRP-conjugated secondary antibodies (Dako, Copenhagen, Denmark) and an enhanced chemiluminescence detection kit (ECL Western Blotting Detection Reagents; Amersham Biosciences, Buckinghamshire, UK). Enzyme-linked immunosorbent assay (ELISA). Serum levels of IL-1 and IL-6 in the 3–6-week-old mice were analyzed by quantikine ELISA (R&D Systems) in accordance with the manufacturer’s protocol. Statistical analysis. Results are expressed as the mean ⫾ SEM. Group mean values were compared by MannWhitney U test. P values less than or equal to 0.05 were considered significant. RESULTS Differentiation into the preclinical, disease onset, and early arthritis stages. Three different phases of very early arthritis based on a weekly assessment of paw swelling and grip strength were defined in HuTNF-Tg mice (Figures 1A and B). The preclinical stage, defined 82 HAYER ET AL Figure 2. Immune phenotyping, cytokine expression, and intracellular signaling in the initial phases of tumor necrosis factor–mediated arthritis. Cellular distribution of the inflamed tendon sheaths was determined by immunohistochemical staining for granulocytes (A), macrophages (B), T lymphocytes (C), and B lymphocytes (D) in human tumor necrosis factor–transgenic mice in the various phases of early disease. Brown cells indicate positive labeling. (Original magnification ⫻ 100.) by the absence of joint swelling and by normal (score 0) grip strength, was generally found between weeks 3 and 4 after birth. The disease onset stage, defined by the appearance of the first signs of joint swelling and start of deterioration of grip strength, could usually be observed between weeks 4 and 5 after birth. The third stage, early arthritis, was characterized by continuing mild clinical signs and symptoms 1 week thereafter (at weeks 5–6 after birth). Manifestation of the first inflammatory event, tenosynovitis, at the preclinical stage. To address the first inflammatory musculoskeletal changes occurring upon TNF overexpression, we analyzed the paw sections of HuTNF-Tg mice during the 3 disease stages (Figures 1C and D). Interestingly, in preclinical disease, the initial pathologic change was a tendon effusion rather than synovitis (Figure 1C). Shortly thereafter, in the disease onset stage, a multitude of inflammatory cells PRECLINICAL TENOSYNOVITIS AND OSTEOCLAST FORMATION ingressed into the tendon sheath, culminating, by the early disease stage, in a massive hyperplasia of the synovialis of the tendon sheath (Figure 1D). The thickness of the tendon sheath, in particular of the long peroneal muscle of the hind paw, was markedly increased in HuTNF-Tg mice compared with wild-type mice from 3 weeks onward. Formation of inflammatory pannus at disease onset. In conjunction with increased cellular influx into the tendon sheaths at disease onset, progressive hyperplasia of the joint synovial lining was observed, whereas no increased thickness of the joint synovial lining was seen during preclinical disease. Inflammatory pannus formation occurred simultaneously with cell influx and synovial hyperplasia of the tendon sheaths and dominated at sites adjacent to the initially inflamed tendon sheaths. Inflammatory pannus formation further increased significantly in early disease compared with that at disease onset. (For immunohistochemical images and quantitative data, a printout is available from the corresponding author upon request.) Influx of granulocytes and macrophages prior to development of tenosynovitis. We next assessed the cell populations participating in the initial tendon sheath inflammation. Interestingly, the dominant cell population causing tenosynovitis was granulocytes (Figure 2A). Single granulocytes were found in the synovium of the tendon sheath, but most of them accumulated in the effusion, resulting in a dramatic increase in cell numbers. Apart from granulocytes, macrophages were also found in the effusion, accumulating in early clinical disease (Figure 2B). In contrast, only a few T lymphocytes were present in the effusions (Figure 2C). B cells were absent in initial tenosynovitis and did not accumulate (Figure 2D). (For quantitative data on the different leukocyte populations invading the tendon sheath, a printout is available from the corresponding author upon request.) Expression of IL-1 and IL-6 in early inflammatory tenosynovitis. To further characterize the initial events of inflammatory arthritis occurring in the HuTNF-Tg mice, we investigated the expression of 2 major TNF-dependent inflammatory cytokines, IL-1 and IL-6. Interestingly, even the first effusing cells in initial (preclinical) tenosynovitis expressed IL-1, suggesting that IL-1 is not only induced very early in disease but also linked to the initial process, since other areas of the joint did not express IL-1 at this stage (Figure 3A). IL-1–positive cells were mainly neutrophils in the synovial tendon sheaths, whereas the mesenchymal lining of these sheaths was IL-1 negative. Later in the early stage, 83 Figure 3. Cytokine expression in the initial phases of tumor necrosis factor (TNF)–mediated arthritis. Expression of cytokines in the inflamed tendon sheaths was determined by immunohistochemical staining for interleukin-1 (IL-1) (A) and IL-6 (B), and serum levels of IL-1 (C) and IL-6 (D) were determined by enzyme-linked immunosorbent assay in wild-type (WT) and human TNF–transgenic (HuTNF-Tg) mice in various phases of early disease. Results are the mean and SD numbers of positive cells. ⴱ ⫽ P ⬍ 0.05 versus HuTNF-Tg mice at disease onset and/or preclinical stage. IL-1–expressing cells were also located in the inflammatory pannus and at sites of bone erosion in early disease. IL-6 was also found to be expressed by neutrophils in the synovial effusions right from the preclinical stage of disease (Figure 3B). In contrast to the pattern of IL-1 expression, IL-6 was also expressed by mesenchymal cells lining the inflamed tendon sheaths. Similarly, expression of IL-6 further increased with disease progression. (For immunohistochemical images, a printout is available from the corresponding author upon request.) This elevation in local expression of IL-1 and IL-6 also resulted in an elevation of the serum levels of IL-1 and IL-6, which significantly increased at the stage of early disease and followed the local increase of these 2 cytokines (Figures 3C and D). Thus, the rise in levels of these cytokines in the serum paralleled the increasing cytokine expression in the tendon sheaths. Increased expression and activation of MAPKs in early inflammatory arthritis. To better characterize the molecular changes at the initiation of arthritis, articular tissue of HuTNF-Tg mice was investigated for the expression and activation of members of the MAPK family. Protein expression of activated p38 MAPK was unchanged at the start of disease, whereas activation of p38 MAPK became evident in preclinical disease and 84 HAYER ET AL Figure 4. Onset of bone erosion and cartilage damage in tumor necrosis factor (TNF)–mediated arthritis. A and B, Tartrate-resistant acid phosphatase–stained sections showing the formation of mononucleated osteoclast precursors (A), multinucleated osteoclasts (purple staining indicated by arrows in B), and bone erosions (B) in human TNF–transgenic mice in various phases of early disease as compared with wild-type mice. Boxes in A are shown at higher magnification in B (original magnification ⫻ 25 in A; ⫻ 400 in B). C, Formation of bone marrow inflammation demonstrated by CD45 receptor–positive B lymphocyte staining (brown color indicated by arrows) (original magnification ⫻ 400). increased continuously over the course of early disease. In contrast, total SAPK/JNK expression did not show any increase at the stage of early arthritis and was not accompanied by an increase in activation as disease progressed. Interestingly, ERK increased right from the preclinical stage of disease and also showed an increased activation over time. (For Western blot images of the kinase activities, a printout is available from the corresponding author upon request.) Formation of primary bone erosions preceding the onset of clinical signs of arthritis. To determine the onset of bone erosion, we next searched for osteoclasts emerging in the synovial tissue. Interestingly, the first mononuclear TRAP⫹ cells appeared at the preclinical stage of disease, at sites where the inflamed tendon sheaths passed by the cartilage–pannus junction (Figures 4A and B). These initial preosteoclasts were associated, even at this earliest stage, with what appeared to be shallow erosions of the mineralized cartilage. TRAP⫹ cells then rapidly increased in number, size, and nucleation during the stages of disease onset and early arthritis, and these changes were associated with an increase in the size of bone erosions. To investigate the participation of juxtaarticular bone marrow in the onset of disease, we investigated this compartment by immunopleiotypic labeling. Cellular accumulations in the bone marrow formed upon disease onset but not in the earlier, preclinical stage (Figure 4C). The number and size of the bone marrow infiltrates increased at the early arthritis stage. These abnormalities were almost exclusively composed of B cells. (For quantitative data on these destructive changes, a printout is available from the corresponding author upon request.) Onset of cartilage damage after bone destruction in TNF-mediated arthritis. To determine the pattern of onset of cartilage damage, we evaluated the content of proteoglycans at the articular surface as well as the structural integrity of cartilage. The cartilage surface decreased gradually during these early stages of arthritis, but damage in the cartilage occurred much slower than in synovial tissue and bone. Thus, even in early arthritis, PRECLINICAL TENOSYNOVITIS AND OSTEOCLAST FORMATION the cartilage surface, although somewhat diminished, was not significantly reduced as compared with that in wild-type mice. Loss of proteoglycans, however, started early, at the stage of disease onset, and continued to proceed to structural damage of cartilage. Nonetheless, initiation of the pathologic changes in the cartilage followed both the development of tenosynovitis and the formation of osteoclasts, the latter being the initial events in this inflammatory arthritis model. (For immunohistochemical images and quantitative data, a printout is available from the corresponding author upon request.) Effects of antiresorptive and antiinflammatory therapy on early arthritis. On the basis of the early appearance of osteoclasts, we wanted to assess whether complete lack of these cells could alter the course of early inflammatory arthritis. In c-fos–deficient HuTNF-Tg mice, which lack osteoclasts, we found that tenosynovitis became apparent at the onset of disease, as characterized by inflamed tendon sheaths next to the periosteal bone surface. However, inflamed tissue was unable to create osteoclasts and thus failed to invade the subchondral bone. In contrast, early initiation of TNF blockade affected the development of tenosynovitis by reducing cellular infiltration and edema. Moreover, formation of synovial osteoclasts was decreased. Thus, TNF blockade has a dual mode of action with respect to joint damage in inflammatory arthritis. It blocks inflammatory tenosynovitis as well as osteoclast formation at the regions prone to bone erosions. (For immunohistochemical images and quantitative data, a printout is available from the corresponding author upon request.) DISCUSSION The early phases of RA have not been studied extensively. Although it is known that the pattern of disease onset may differ between patients, in the majority of patients with RA the disease starts insidiously (14). Not infrequently, tenosynovitis accompanies the early steps of RA (15–17). The sequence of the earliest pathologic events is still unknown. It has been postulated that an influx of inflammatory cells into the synovium precedes the development of the signs and symptoms of the disease, since frank synovitis could be found in the joints of RA patients who were clinically unaffected (1). It is also evident that immune system abnormalities precede the development of RA, since autoantibodies can be detected months to years before onset of the disease (18–20). Moreover, ⬃10% of RA patients have evidence of bone erosions on plain radiographs as early 85 as 8 weeks after disease onset (Machold K: personal communication). These and other observations have led to the concept of prearthritis (21). However, the events governing the changes that occur before arthritis, i.e., at the time that inflammatory joint disease becomes manifest, cannot be easily studied in humans, since at the preclinical stage, tissue is not easily accessible and standard imaging techniques, including MRI, have an insufficient resolution to discern changes on the cellular level. To address this question, we therefore studied an animal model that closely resembles human RA in terms of overexpression of TNF and the characteristics of inflammation and bone and cartilage destruction. With the aim of studying the first pathologic events of TNF-mediated inflammatory arthritis in the joint, we used the HuTNF-Tg mouse model of arthritis and screened for the onset of inflammation and structural damage before, at the time of, and shortly after disease onset. In fact, even in animal models, the knowledge regarding these initial phases of arthritis is scarce. Both collagen-induced and adjuvant-induced arthritis have a very rapid disease onset and progress quickly, making the dissection of these initial phases of arthritis difficult (3,22). However, it is evident from these models that highly active arthritis needs only a few days to induce destruction of the affected joint (3). The disease in HuTNF-Tg mice has a mild and rather unspectacular onset, but it continuously progresses over time, leading to accumulation of an increasing amount of joint damage. Thus, with the use of this model, we were able to better study these initial phases of disease and relate them to the onset of clinical symptoms in more detail. Interestingly, our study revealed that tendons are the earliest structures affected by inflammation (Figure 5). These structures are subject to the most prominent shear forces and are under prolonged mechanical load. Tendons are surrounded by synovial tissue, and the tendon sheaths have a cell composition similar to that of the synovial lining. Accumulation of fluid as well as effusion of inflammatory cells, primarily granulocytes and macrophages, into these structures is the first pathologic event induced by overexpression of TNF. Many investigators have speculated as to why these tendon sheaths are the first structures affected by chronic inflammation, but mechanical factors are likely to play a key role in this process. Ultimately, a complete remodeling of the tendon sheaths to pannus-like synovial tissue occurs, thereby completely filling the synovial space with inflamed tissue. This obviously leads to massive distur- 86 HAYER ET AL Figure 5. Model of arthritis onset in mice. Arthritis starts before onset of clinical disease, with infiltration of tendon sheaths (purple) by granulocytes and monocytes, leading to tenosynovitis. Tenosynovitis precipitates local osteoclast formation (red), in which tendons pass by the joint edges, thus inducing resorption of mineralized cartilage and bone. As tenosynovitis becomes more severe, the synovial membrane of the joint also becomes inflamed and hyperplastic (green), which is reflected in the onset of clinical disease. Structural damage further increases (light yellow). In the next step, tendon sheaths become further infiltrated and more pronounced inflammation of the synovial membrane occurs, which leads to an increase in clinical symptoms and acceleration of joint damage. bance of the tendon and ultimately impedes joint function. Moreover, the tendons use the bony ends and their edges as anchors to increase function, thus bringing these structures into close contact with sites predisposed to the development of bone erosions. These areas in which tendon sheaths, mineralized cartilage, and subchondral bone are in close contact are very distinct. In these contact regions, the first osteoclasts appear and start to resorb subchondral bone and mineralized cartilage (10,23). Thus, even before the onset of synovial inflammation, osteoclasts start their destructive work and pave the way for early B cell accumulation in the adjacent bone marrow, which is closely linked to breaking of the cortical bone barrier (24,25). Since osteoclasts appear even before the first clinical signs of arthritis, destructive changes begin to occur even before arthritis is clinically apparent in the patients. This finding further strengthens the concept of a prearthritic state and confirms that there is a very narrow window of opportunity for achieving full remission of RA without any sequelae (26). The activity of TNF leads to activation of MAPK cascades, such as p38 MAPK and ERK, very early in the disease. This is consistent with the well-documented activation of MAPK families in human RA (27–30). Thus, even in the preclinical disease phase, the joints of the HuTNF-Tg mice showed an activation of these 2 MAPK members, suggesting that p38 MAPK and ERK mediate the cellular effects of TNF right from the start of molecular onset of arthritis. As a matter of fact, p38 MAPK plays a major role in facilitating the cytokine activation induced by TNF. Induction of proinflammatory mediators such as IL-1 and IL-6 by TNF is mediated by activation of p38 MAPK (31). Of note, these cytokines are inducibly expressed in these initial stages of arthritis and appear to be selectively linked to the areas of onset of inflammation, such as the cellular infiltration of the tendon sheaths. IL-1 and IL-6 are subsequently expressed extensively throughout the newly formed inflamed synovial tissue. We found that this increased expression of IL-1 and IL-6 was paralleled by a rise in the serum levels of these 2 cytokines. Early activation of ERK, in contrast, might influence the survival of cells in the synovial membrane. ERK activation by TNF plays an important role in the maintenance of cell survival and the inhibition of apoptosis (32). In fact, apoptosis is low in the synovial membrane, suggesting that cytokines like TNF act on intracellular survival factors rather than apoptosis-inducing signaling molecules (33,34). In the present study, we also observed that damage of unmineralized surface cartilage occurred subsequent to, and not before or concomitant with, tenosynovitis and osteoclast formation. In contrast to osteoclastmediated damage, which affects mineralized structures, breakdown of the surface cartilage depends on cytokine and enzyme production of cells in the joint cavity and the synovial lining. Cartilage damage was linked to a PRECLINICAL TENOSYNOVITIS AND OSTEOCLAST FORMATION certain level of inflammation and structural organization of the synovial surface (lining layer), suggesting that a threshold of synovial inflammation is critical for matrix resorption. Thus, whereas destruction of mineralized tissue was linked to tenosynovitis and occurred very early in the disease, cartilage damage started later and was linked to inflammation of the synovial membrane. In summary, we have shown that tenosynovitis is the first structural change to occur in TNF-mediated arthritis. Regions of close interaction between the tendons and mineralized tissue allow the spread of inflammatory changes to the joints, particularly leading to the resorption of mineralized cartilage and bone. Importantly, these changes occur before the first clinical signs of joint swelling are apparent. Our data suggest that newly occurring tenosynovitis should be taken seriously, since it might be followed by the development of RA. Moreover, structural changes in the joints are a very early feature of disease, starting right from its subclinical onset. Thus, the time window to protect joints from damage may, in fact, be very short, and this knowledge should give impetus to the efforts for early initiation of targeted therapy in patients with RA. 87 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. ACKNOWLEDGMENTS We thank Thomas Pangerl, Birgit Türk, and Margarete Tryniecki for excellent technical assistance, Prof. Erwin Wagner (Research Institute for Molecular Pathology, Vienna, Austria) for providing the c-fos–deficient mice, and Prof. George Kollias (Alexander Fleming Biomedical Research Center, Vari, Greece) for providing the HuTNF-Tg mice. AUTHOR CONTRIBUTIONS Dr. Schett had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Study design. Drs. Hayer, Redlich, Smolen, and Schett. Acquisition of data. Drs. Hayer and Korb. Analysis and interpretation of data. Drs. Hayer, Redlich, Hermann, and Schett. Manuscript preparation. Drs. Hayer, Smolen, and Schett. Statistical analysis. Drs. Hayer and Hermann. 16. 17. 18. 19. 20. 21. 22. 23. REFERENCES 1. Kraan MC, Versendaal H, Jonker M, Bresnihan B, Post WJ, ‘t Hart BA, et al. Asymptomatic synovitis precedes clinically manifest arthritis. Arthritis Rheum 1998;41:1481–8. 2. Van der Heijde DM. Joint erosions and patients with early rheumatoid arthritis. Br J Rheumatol 1995;34:74–8. 3. Schett G, Smolen JS. New insights in the mechanism of bone loss in arthritis. Curr Pharm Des 2005;11:3039–49. 4. Goldring SR. Bone and joint destruction in rheumatoid arthritis: what is really happening? J Rheumatol Suppl 2002;65:44–8. 5. Pap T, Aupperle KR, Gay S, Firestein GS, Gay RE. Invasiveness 24. 25. 26. of synovial fibroblasts is regulated by p53 in the SCID mouse in vivo model of cartilage invasion. Arthritis Rheum 2001;44:676–81. Feldmann M, Brennan FM, Foxwell BM, Maini RN. The role of TNF ␣ and IL-1 in rheumatoid arthritis. Curr Dir Autoimmun 2001;3:188–99. Keffer J, Probert L, Cazlaris H, Georgopoulos S, Kaslaris E, Kioussis D, et al. Transgenic mice expressing human tumour necrosis factor: a predictive genetic model of arthritis. EMBO J 1991;10:4025–31. Shigeyama Y, Pap T, Kunzler P, Simmen BR, Gay RE, Gay S. Expression of osteoclast differentiation factor in rheumatoid arthritis. Arthritis Rheum 2000;43:2523–30. Gravallese EM, Manning C, Tsay A, Naito A, Pan C, Amento E, et al. Synovial tissue in rheumatoid arthritis is a source of osteoclast differentiation factor. Arthritis Rheum 2000;43:250–8. Gravallese EM, Harada Y, Wang JT, Gorn AH, Thornhill TS, Goldring SR. Identification of cell types responsible for bone resorption in rheumatoid arthritis and juvenile rheumatoid arthritis. Am J Pathol 1998;152:943–51. Kong YY, Feige U, Sarosi I, Bolon B, Tafuri A, Morony S, et al. Activated T cells regulate bone loss and joint destruction in adjuvant arthritis through osteoprotegerin ligand. Nature 1999; 402:304–9. Teitelbaum SL. Bone resorption by osteoclasts. Science 2000;289: 1504–8. Redlich K, Hayer S, Ricci R, David JP, Tohidast-Akrad M, Kollias G, et al. Osteoclasts are essential for TNF-␣-mediated joint destruction. J Clin Invest 2002;110:1419–27. Masi AT. Articular patterns in the early course of rheumatoid arthritis. Am J Med 1983;75:16–26. Brown FE, Brown ML. Long-term results after tenosynovectomy to treat the rheumatoid hand. J Hand Surg 1988;13:704–8. McQueen F, Ostergaard M, Peterfy C, Lassere M, Ejbjerg B, Bird P, et al. Pitfalls in scoring MR images of rheumatoid arthritis wrist and metacarpophalangeal joints. Ann Rheum Dis 2005;64:48–55. Stewart NR, McQueen FM, Crabbe JP. Magnetic resonance imaging of the wrist in early rheumatoid arthritis: a pictorial essay. Australas Radiol 2001;45:268–73. Palosuo T, Tilvis R, Strandberg T, Aho K. Filaggrin related antibodies among the aged. Ann Rheum Dis 2003;62:261–3. Aho K, von Essen R, Kurki P, Palosuo T, Heliovaara M. Antikeratin antibody and antiperinuclear factor as markers for subclinical rheumatoid disease process. J Rheumatol 1993;20:1278–81. Aho K, Palosuo T, Raunio V, Puska P, Aromaa A, Salonen JT. When does rheumatoid disease start? Arthritis Rheum 1985;28: 485–9. Smolen JS. The science of rheumatoid arthritis: a prelude. Arthritis Res Ther 2005;7:S1–3. Lubberts E, Oppers-Walgreen B, Pettit AR, van den Bersselaar L, Joosten LA, Goldring SR, et al. Increase in expression of receptor activator of nuclear factor B at sites of bone erosion correlates with progression of inflammation in evolving collagen-induced arthritis. Arthritis Rheum 2002;46:3055–64. Bromley M, Woolley DE. Chondroclasts and osteoclasts at subchondral sites of erosion in the rheumatoid joint. Arthritis Rheum 1984;27:968–75. Jimenez-Boj E, Redlich K, Turk B, Hanslik-Schnabel B, Wanivenhaus A, Chott A, et al. Interaction between synovial inflammatory tissue and bone marrow in rheumatoid arthritis. J Immunol 2005;175:2579–88. Gortz B, Hayer S, Redlich K, Zwerina J, Tohidast-Akrad M, Tuerk B, et al. Arthritis induces lymphocytic bone marrow inflammation and endosteal bone formation. J Bone Miner Res 2004;19: 990–8. Boers M. Understanding the window of opportunity concept in early rheumatoid arthritis [editorial]. Arthritis Rheum 2003;48: 1771–4. 88 27. Chabaud-Riou M, Firestein GS. Expression and activation of mitogen-activated protein kinase kinases-3 and -6 in rheumatoid arthritis. Am J Pathol 2004;164:177–84. 28. Schett G, Tohidast-Akrad M, Smolen JS, Schmid BJ, Steiner CW, Bitzan P, et al. Activation, differential localization, and regulation of the stress-activated protein kinases, extracellular signal– regulated kinase, c-Jun N-terminal kinase, and p38 mitogenactivated protein kinase, in synovial tissue and cells in rheumatoid arthritis. Arthritis Rheum 2000;43:2501–12. 29. Neidhart M, Rethage J, Kuchen S, Kunzler P, Crowl RM, Billingham ME, et al. Retrotransposable L1 elements expressed in rheumatoid arthritis synovial tissue: association with genomic DNA hypomethylation and influence on gene expression. Arthritis Rheum 2000;43:2634–47. 30. Inoue T, Boyle DL, Corr M, Hammaker D, Davis RJ, Flavell RA, et al. Mitogen-activated protein kinase kinase 3 is a pivotal HAYER ET AL 31. 32. 33. 34. pathway regulating p38 activation in inflammatory arthritis. Proc Natl Acad Sci U S A 2006;103:5484–9. Suzuki M, Tetsuka T, Yoshida S, Watanabe N, Kobayashi M, Matsui N, et al. The role of p38 mitogen-activated protein kinase in IL-6 and IL-8 production from the TNF-␣- or IL-1␤-stimulated rheumatoid synovial fibroblasts. FEBS Lett 2000;465:23–7. Schett G, Steiner CW, Xu Q, Smolen JS, Steiner G. TNF␣ mediates susceptibility to heat-induced apoptosis by protein phosphatase-mediated inhibition of the HSF1/hsp70 stress response. Cell Death Differ 2003;10:1126–36. Baier A, Meineckel I, Gay S, Pap T. Apoptosis in rheumatoid arthritis. Curr Opin Rheumatol 2003;15:274–9. Firestein GS, Nguyen K, Aupperle KR, Yeo M, Boyle DL, Zvaifler NJ. Apoptosis in rheumatoid arthritis: p53 overexpression in rheumatoid arthritis synovium. Am J Pathol 1996;149:2143–51.