DOI 10.1007/s00296-007-0346-1
O R I G I N A L A R T I C L E
Intra-articular injection of hyaluronate and indomethacin
in rabbits with antigen-induced arthritis
Yow-Jen Lo · Ming-Thau Sheu · Wen-Chi Tsai ·
Yun-Ho Lin · Jau-Le Li · Yu-Chih Liang · Chi-Ching Chang · Ming-Shium Hsieh · Chien-Ho Chen
Received: 20 July 2006 / Accepted: 3 March 2007 / Published online: 14 April 2007
© Springer-Verlag 2007
Abstract Combined eVects of hyaluronate and indometh-acin in the treatment of rabbits with antigen-induced arthritis (AIA) were evaluated by assessing joint swelling, C-reactive protein (CRP) and prostaglandin E2 (PGE2) levels with periodic intra-articular (ia) injections of hyalur-onate alone (HA group) and with either a low or high concentration of indomethacin (LI-HA or HI-HA group). End-point analyses included matrix metalloproteinases-3 (MMP-3) activity and macroscopic and histological joint examinations. Results demonstrated that treatment in LI-HA and HI-HA groups resulted in statistically signiWcant
suppression of CRP, PGE2, and MMP-3 in comparison with those of HA group. Inhibition of serum CRP was only observed in LI-HA group. The order of serum MMP-3 inhibition was LI-HA>HI-HA>HA. Based on macroscopic and histological analyses of pannus formation, hyperplasia, inXammation, joint leakage and erosion, and loss of proteo-glycan, the only statistically signiWcant improvement was shown in LI-HA group compared to HA group and HI-HA group compared to control group.
Keywords Antigen-induced arthritis · Hyaluronate · Indomethacin · CRP · PGE2 · MMP
Introduction
Arthritis is a chronic multifactorial disease induced when the immune system attacks and begins degrading the body’s joints. Common underlying symptoms of the above clinical manifestations include inXammation, destruction of cartilage and soft tissue, and dysfunction of the joints [1–3]. Therapeutic approaches using an elastoviscous HA solution and HA derivatives (hylans) for treatment of arthritis pain are based on the Wnding that long-lasting analgesia can be achieved in joints of arthritic horses by replacing patho-logic synovial Xuid with a highly puriWed HA solution of normal elastoviscosity but signiWcantly greater concentra-tion than that of healthy joint Xuid. Hyaluronic acid (HA) is an abundant non-sulfated glycosaminoglycan component of synovial Xuid and extracellular matrices. In normal human synovial Xuid, the molecular weight (MW) of HA is (6–7) £106 Da, and the concentration is 2–4 mg/ml. HA prepara-tions appear to be quite safe, with local reacprepara-tions at the injection site (e.g., pain and swelling) generally being mild and transient [4]. As a result of this discovery, highly Y.-J. Lo · M.-T. Sheu
College of Pharmacy, Taipei Medical University, Taipei, Taiwan, ROC
W.-C. Tsai
Department of Orthopedic Surgery, Min-Sheng Healthcare, Taoyuan, Taiwan, ROC
Y.-H. Lin
Department of Pathology, Taipei Medical University, Taipei, Taiwan, ROC
J.-L. Li · M.-S. Hsieh
Department of Orthopedics and Traumatology, Taipei Medical University Hospital,
Taipei Medical University, Taipei, Taiwan, ROC
Y.-C. Liang · C.-H. Chen (&)
School of Medical Technology, Taipei Medical University, 250 Wu-Hsing Street, Taipei, 110, Taiwan, ROC
e-mail: chenchho@tmu.edu.tw
C.-C. Chang
Department of Internal Medicine, Taipei Medical University Hospital, Taipei Medical University, Taipei, Taiwan, ROC
puriWed HA and hylan solutions became available world-wide for the treatment of arthritis pain in animals and humans [5,6].
Intra-articular (ia) injections of HA as treatment for knee OA were reviewed by Brabdt et al. [4]. Despite sev-eral investigators having reported that ia injections relieve joint pain and improve function in humans with knee OA [7,8], they concluded that there was insuYcient
informa-tion to draw conclusions concerning the eVect of this treat-ment, if any, on the progression of OA in humans. Nevertheless, HA has demonstrated a variety of eVects on cells in vitro that may be related to its reported eVects on joint disease. These include the inhibition of prostaglandin E2 (PGE2) synthesis induced by interleukin-1 (IL-1) [9,10] and protection against proteoglycan depletion and cytotox-icity induced by oxygen-derived free radicals, IL-1, and mononuclear cell-conditioned medium [11,12]. A recent study by Gonmis et al. further reported that elastoviscous properties of HA solutions are determining factors in reducing pain-eliciting nerve activity in both normal and inXamed rat joints [13]. Furthermore, local therapeutic eVects of intra-articular injections of high-MW HA in rats with AIA were reported by Roth et al. to be biphasic, with inhibition of inXammation and cartilage damage in the early chronic phase but with promotion of joint swelling, inXammation, and cartilage damage in the late chronic phase [14].
Since their introduction, NSAIDs have commonly been prescribed by physicians for the treatment of OA by oral administration. But in the 1990s, a series of research arti-cles cast doubt on the superior eYcacy of NSAIDs com-pared with acetaminophen [15,16]. However, a survey of 1,799 patients with OA, RA, and Wbromyalgia (FS) con-ducted by Wolfe et al. concluded that there was a consider-able and statistically signiWcant preference for NSAIDs compared to acetaminophen among these three groups of rheumatic disease patients. The report also stated that if safety and costs are not issues, there would hardly ever be a reason to recommend acetaminophen over NSAIDS, since patients generally preferred NSAIDS, and fewer than 14% preferred acetaminophen [17]. Furthermore, although NSAIDs are eVective in reducing symptoms, they do not reduce joint damage. Because of this and their potential for serious toxicity, they should be used at the minimum eVec-tive dosage, even in patients with inXammatory joint disease [18]. Recently, Yoon et al. reported that NSAIDs, such as indomethacin, have protective eVects against carti-lage damage, not only by alleviating inXammation but also by inhibiting NO-induced apoptosis and dediVerentiation of articular chondrocytes. The latter eVects require higher concentrations of NSAIDs that are 100- to 1,000-fold higher than those needed to inhibit prostaglandin synthesis [19]. This would imply that dosages of NSAIDs should be
higher than the oral dose used for anti-inXammatory pur-poses via a systemic route in order to achieve optimal local concentrations at arthritic joint sites to prevent cartilage damage.
The ideal treatment for arthritis would reduce pain and inXammation, maintain function, and at the same time, be safe. The purpose of this study was to evaluate the combined eVects on improving the inXammatory status, inXammatory pain, and cartilage degradation in rabbits with antigen-induced arthritis after treatment with hyaluro-nate alone and combined with indomethacin, which can be used as a local injection in order to avoid the serious side eVects of indomethacin but increase the local concentra-tion to a level that is suYcient to activate the mechanism which prevents cartilage damage while reducing pain and swelling at the local injection site. Also, the viscous property of the 1% HA solution in which indomethacin was dissolved was recognized to be capable of sustaining indomethacin’s pharmacological activity by providing a longer period of local retention of indomethacin within the knee joints.
Materials and methods Materials
Sodium pentobarbital was obtained from TCI (TOKYO, Japan). Indomethacin, ovalbumin (OVA), and Freund’s complete adjuvant were obtained from Sigma (St. Louis, MO, USA). A 0.9% isotonic NaCl solution was obtained from Sington (Taipei, Taiwan). ARTZDispo®, a commer-cial product containing 1% sodium hyaluronate, was sup-plied by Seikagaku (Tokyo, Japan) and was used in combination with indomethacin.
Preparation of the hyaluronate and indomethacin intra-articular injections
Indomethacin was dissolved in isotonic buVer and mixed well with 1% sodium hyaluronate (ARTZDispo®, Seikagaku) having a weight-averaged molecular weight of (6–12) £105. Indomethacin was dispensed as an intra-articular (ia) injection at Wnal concentrations of 5.6 (0.002 mg/ml) and 560(0.2 mg/ml) in a 1% sodium hyaluronate solution.
Induction of AIA
The experimental disease was induced as previously pub-lished [5]. The experimental design was for arthritis to be induced by OVA in New Zealand white (NZW) rabbits with an initial weight of 2.6–3.3 kg (Fig.1). Animal
experiments were carried out with the approval of the local and national ethics committees at the animal facility of Taipei Medical University. Sodium pentobarbital (0.2–0.3 mg/kg) was used for animal sedation. An emulsiWed mixture of 5 mg of OVA and Freund’s complete adjuvant containing 1 mg of Mycobacterium tuberculosis was prepared for immunization. NZW rabbits were immunized by subcuta-neously injecting this mixture into multiple sites of the shaved interscapular region on three occasions at 2-week intervals (days 1, 15, and 29). Arthritis was induced by a fourth injection of the emulsion containing 5 mg/ml of OVA in complete Freund’s adjuvant into both knee joints on day 19. All procedures were performed strictly in accor-dance with current local regulations.
Intra-articular treatment
One group (n = 3) served as the normal control. Another group (n = 5) served as the disease control and was sub-jected to four treatments at intervals of 5, 4, and 3 days (i.e. receiving treatments on days 1, 6, 10, and 13), respectively, consisting of ia injections of 0.5 ml of 0.9% NaCl into the right knee joint. Three treatment groups (n = 5 for each group) were also subjected to the same dosing regimen using ia injections of 0.5 ml of 1% sodium hyaluronate (the HA group), 1% sodium hyaluronate plus 5.6M indometh-acin (the LI-HA group), or 1% sodium hyaluronate plus 560.0M indomethacin (the HI-HA group) into the right knee joint. The left knee joint was not treated and served as the internal control. The rationale for the design of such a dosing regimen was the hypothesis that rabbits with AIA
would not be completely cured by treatment with hyaluro-nate alone or by hyalurohyaluro-nate and indomethacin which might still decrease inXammation and inXammatory pain and delay arthritic progression. All animals were sedated and sacriWced on day 40 by intravenous administration of sodium pentobarbital (50–80 mg/kg).
Clinical assessments of AIA
The eVects of treatment on the rabbits were monitored by analyzing weight loss, joint circumference, biochemical parameters (CRP, PGE2, and MMP-3), and macroscopic and histological evaluations of the blood and articular carti-lage [5]. The fur on the knees of all rabbits was shaved oV,
and the circumference of the knees was measured using a measuring tape before each injection. Measurements were taken in the anterior-to-posterior position, with the ankle extended, through the circumference of the knee.
Assay of CRP and PGE2
The CRP reagent, in conjunction with Beckman Array Systems and Calibrator or Calibrator 5 (USA), was used for the quantitative determination of human CRP by a rate nephelometer. The method employed in the Beckman CRP test measures the rate of increase in light scattered from particles suspended in solution as a result of complexes formed during an antigen–antibody reaction. After the antibody to CRP is brought into contact with CRP in a sam-ple, the increase in light scattering resulting from the antigen–antibody reaction is converted to a peak rate signal that is a function of the CRP concentration in the sample. Following calibration, the peak rate signal for a particular assay is automatically converted to concentration units by the analyzer.
PGE2 production was determined by measuring the serum level of rabbits using a PGE2 assay kit (Assay Designs, Ann Arbor, MI, USA), which was used to quan-tify the amount of PGE2 according to the manufacturer’s protocol. PGE2 levels were calculated against a standard curve of PGE2.
QuantiWcation and Western blot analysis of MMP-3 Enzyme activity assay kits were used to quantify the amounts of the pro-form and active form of MMP-3 present in the samples (Biotrak MMP-3 activity assay system, Amersham Pharmacia Biotech, Buckinghamshire, UK) fol-lowing the manufacturer’s protocol. Samples were compared to a serial dilution of standards (MMP-3, 0.5–16 ng/ml). The addition of 1 mM aminophenyl mercuric acetate (APMA) was used to activate proMMPs and thus measure total MMP-3 activity (i.e., both the pro-form and active form) of Fig. 1 Experimental design for arthritis induced by an emulsiWed
mix-ture of 5 mg of ovalbumin and Freund’s complete adjuvant containing 1 mg of Mycobacterium tuberculosis in NZW rabbits. All rabbits (n = 5 in each group) were immunized and treated with intra-articular injections of 0.5 ml of either 0.9% NaCl, 1% hyaluronate (the HA group), 5.6M indomethacin + 1% hyaluronate (the LI-HA group), or 560M indomethacin + 1% hyaluronate (the HI-HA group) into the right knee joint
8
1 15 19
2’
4’ Both joint cavities.
3’
22 28 33 37 40 days
1” Right joint treatment 2” 3” 4”
Sacrifice 1’Immunization
each sample. The absorbance of the colorimetric reaction product was read at a wavelength of 405 nm after incuba-tion at t = 0, 1, 2, 3, 4, and 5 h. The amounts of MMP-3 activity were calculated from the serial dilution data of the standards. Data are presented as the mean § SEM. The rate of change of MMP activity is expressed as (Abst=y ¡ Abst=0) x 1000/ty.
Serum diluted 1:10 in PBS was used for Western blot analysis of both pro-(59/57 kDa) and active-form (48/ 25 kDa) MMP-3. The protein was quantiWed by a Bio-Rad protein assay, size-fractionated by SDS-polyacrylamide gel electrophoresis, and transferred to a nitrocellulose membrane. Proteins were detected using the rabbit anti-MMP3 polyclonal antibody (Calbiochem, San Diego, USA) that recognizes the »57 kDa pro- and the »48 kDa active forms of MMP-3. Blots were developed using a per-oxidase-conjugated secondary antibody and an enhanced chemiluminescence system.
Macroscopic analysis of the joints
Opened joints were macroscopically evaluated for the extent of pannus formation as follows: 0, no involvement; 1, mild; 2, moderate; and 3, severe. The left and right nor-mal control cartilage was referred to as the baseline in this study, and its macroscopic pannus formation and erosion were graded 0 (A and B). The left knee joint which received no treatment in treated animal in each of the four treatment groups (0.9% NaCl, HA, LI-HA, and HI-HA) was used as its own internal control, and macroscopic pan-nus formation of both knees was graded as well. The mean§SEM of the grading score for both knees was calcu-lated and reported for each of the 4 treatment groups (n = 5 for each group).
Synovial and cartilage sample collection and histological analysis
Patellar and synovial samples from the infrapatellar, superolateral, and posterior locations were collected. Syno-vium was snap-frozen, or paraYn-embedded when the patellae were decalciWed and embedded in paraYn. The joints of rabbit knees were also removed and Wxed in 10% buVered formalin. The joints were decalciWed in 5% formic acid, embedded in paraYn, sectioned at 5-m thicknesses, and subsequently stained with hematoxylin-eosin (H&E) for general morphology examination and with toluidine blue for proteoglycan examination. Sections were exam-ined for 5 diVerent events: hyperplasia of cells of the syno-vial lining, inWltration of the sublining by mononuclear cells frequently organized in aggregates, leakage and ero-sion of the joint, and loss of proteoglycan [20]. Histological analysis was conducted, and the following grading scale for
hyperplasia and inWltration of synovial cells was scored as 0, no changes (·2 cell layers thick); 1, minimal hyperplasia and inWltration (2–3 cell layers thick); 2, mild hyperplasia and inWltration (5–10 cell layers thick); 3, moderate lining hyperplasia (>10 cell layers thick); and 4, marked hyperpla-sia and inWltration (organized pannus and lining layers). For leakage and erosion of joints, the grading scale was scored as 0, no changes; 1, minimal (size and number) erosion of hard tissues at the margins and in the central region of the joint; 2, small erosions in the notched region of the femur and the center of the tibia and marginal cartilage-pannus junctions; 3, larger deeper erosions in the same areas as in 2; 4, erosions beginning to break into the subchondral tra-becular epiphyses; and 5, erosions breaking into the full depth of the epiphyses and deeply into the marginal sub-chondral bone. The mean§SEM of the grading scale for each event (hyperplasia, inXammation, joint leakage, ero-sion, and loss of proteoglycan) was calculated and reported for each of the 4 treatment groups (n = 5 for each group). Inhibition of therapeutic responses and statistical analysis The percent inhibition of clinical parameters was calculated in each rabbit using the formula: [1 ¡ (C ¡ A)/(B ¡ A)] x 100, where A is the mean of clinical assessments in a nor-mal state or before joint injection with emulsion containing OVA in complete Freund’s adjuvant (day 1); B is the mean of clinical assessments in the immunized state (day 22) after injection with emulsion containing OVA in complete Freund’s adjuvant into both knee joints; and C is the mean of clinical assessments after various drug treatments (day 40) [21,22]. Student’s t-test was used to determine the sig-niWcance of the diVerence between values. Comparisons among multiple groups were assessed using one-way ANOVA and Bonferroni’s t-test (SigmaStat). A p value of <0.05 was considered statistically signiWcant.
Results
Development of arthritis
All rabbits had developed AIA by the end of the intra-articu-lar injections of the antigen on day 22 as indicated by the increased serum CRP in all treatment groups as shown in Table1. A statistically signiWcant diVerence was observed
between serum CRP levels before and after immunization for each of four treatment groups in the experiment. Serum levels of CRP detected on day 1 in AIA rabbits were in the range from 0.52 § 0.14 (in the HA group) to 0.66 § 0.03 mg/dl (in the LI-HA group). On day 22 after the Wnal intra-articular antigen injection, the serum CRP levels were almost fourfold higher (2.08 § 0.08 to
2.50 § 0.14 mg/dl) than those on day 1, with a signiWcance level of p < 0.05.
EVects on joint swelling
The fur of the knee of all rabbits was shaved oV, and the cir-cumference of the knee in the anterior-to-posterior position was measured with the ankle extended; results are listed in Table1. On day 20, AIA developed as a signiWcant swelling
of the right knee point in all treated rabbits. There were sta-tistically signiWcant diVerences in the amounts of joint swelling between days 1 and 22 (before and after immuniza-tion, at 11.56 § 0.10 and 12.64 § 0.11 cm, respectively, n = 20). However, there was a slight but statistically insig-niWcant decrease in joint swelling in each treatment group at the end of treatment on day 40. After initiation of AIA, the swelling remained statistically signiWcantly higher com-pared to the baseline levels on day 1 and had not recovered to the baseline levels at the end of any treatment.
EVects on serum CRP, PGE2, and MMP-3 levels
Clinical assessment of serum levels of CRP, PGE2, and MMP-3 was conducted at the initiation of AIA induction (on day 1), and before and after treatment (on days 22 and 40), and results are illustrated in Table1. As shown in Table1, there was a statistically signiWcant increase in
serum CRP levels after the Wnal AIA induction on day 22 (0.59 § 0.04 and 2.25 § 0.10 mg/dl, respectively) for all treatment groups. At the end of treatment on day 40, there were statistically signiWcant decreases in serum CRP levels compared with those before treatment on day 22 for the LI-HA and HI-LI-HA groups (2.50 § 0.14 vs. 1.48 § 0.06 mg/dl and 2.11 § 0.07 vs. 1.64 § 0.02 mg/dl, respectively), whereas slight but statistically insigniWcant decreases were recorded for the control group and HA group (2.32 § 0.33 vs. 2.18 § 0.17 mg/dl and 2.08 § 0.08 vs. 1.83 § 0.06 mg/dl, respectively). There were statistically signiWcant increases in serum PGE2 levels at the end of AIA induction on day 22 in all treatment groups (7.46 § 0.78 vs. 22.92 § 1.69 ng/ ml, n = 20). At the end of treatment on day 40, serum PGE2 levels had continually increased in the control and HA groups, whereas they had dropped in the LI-HA and HI-HA groups. There were statistically signiWcant increases in serum MMP-3 levels on day 22 in all treatment groups compared to respective values on day 1 (12.69 § 0.66 vs. 3.95 § 0.48 ng/ml). At the end of treatment on day 40, serum MMP-3 levels had continually increased in compari-son to values on day 22 in all treatment groups.
Inhibition of therapeutic responses
Therapeutic responses of each treatment were compared by the percent inhibition (%) of serum levels of CRP, Table 1 Clinical assessments
of the drug treatments (four intra-articular injections in the right knee joint) in rabbits with antigen-induced arthritis (AIA)
Assessment Treatment
0.9% NaCl HA LI-HA HI-HA
Weight (kg)
Day 1 3.12 § 0.11 3.09 § 0.07 3.11 § 0.17 3.04 § 0.04
Day 22 3.01 § 0.10 3.01 § 0.06 3.01 § 0.03 3.07 § 0.03
Day 40 2.86 § 0.05 2.93 § 0.09 2.95 § 0.06 2.98 § 0.05
Circumference (cm) of the left joint
Day 1 11.38 § 0.31 11.38 § 0.13 11.50 § 0.20 11.88 § 0.13 Day 22 12.50 § 0.20§ 12.50 § 0.29§ 12.63 § 0.31§ 12.88 § 0.13§ Day 40 12.38 § 0.13 12.25 § 0.14 12.18 § 0.28 12.50 § 0.20 CRP (mg/dl) Day 1 0.63 § 0.07 0.52 § 0.14 0.66 § 0.03 0.54 § 0.08 Day 22 2.32 § 0.33§ 2.08 § 0.08§ 2.50 § 0.14§ 2.11 § 0.07§ Day 40 2.18 § 0.17 1.83 § 0.06 1.48 § 0.06* 1.64 § 0.02* PGE2 (ng/ml) Day 1 9.15 § 0.71 2.63 § 0.15 9.47 § 0.74 7.90 § 0.77 Day 22 24.63 § 1.80§ 13.87 § 0.96§ 29.33 § 1.20§ 23.20 § 2.69§ Day 40 32.75 § 2.75* 16.87 § 1.74 26.67 § 0.88 17.90 § 2.09 MMP-3 (ng/ml) Day 1 4.71 § 1.19 2.27 § 0.30 4.53 § 0.97 4.28 § 0.86 Day 22 13.61 § 1.69§ 13.04 § 1.98§ 12.12 § 0.36§ 11.99 § 1.36§ Day 40 28.02 § 3.57* 21.38 § 3.24* 17.24 § 1.81 17.94 § 2.46 Changes in serum CRP, PGE2,
and MMP-3 on days 1 and 22 (before and after the intra-articu-lar injection, respectively) were statistically analyzed using Student’s t-test (§ p < 0.05). Changes in serum CRP, PGE2, and MMP-3 on days 22 and 40 (before and after various drug treatments, respectively) were statistically analyzed using Bonferroni’s t-test. *p < 0.05
PGE2, and MMP-3 calculated as deWned in “Methods”, and the results are displayed in Table2. The order of serum CRP inhibition after treatment was the LI-HA group (54.6 § 0.17%) > HI-HA group (30.3 § 9.2%) > HA group (2.2 § 0.8%) t control group (5.0 § 9.5%). Statis-tically signiWcant inhibition of serum CRP levels was only observed in the LI-HA and HI-HA groups, while treatment with HA alone demonstrated an insigniWcant eVect on serum CRP similar to that treated with 0.9% NaCl. The treatment with 0.9% NaCl (¡53.2 § 10.6%) or HA alone (¡26.0 § 6.1%) was not able to inhibit the formation of serum PGE2, whereas the inhibition of serum PGE2 levels was parallel to indomethacin concentrations with the per-cent inhibition in the LI-HA group (13.3 § 2.9%) lower than that in the LI-HA group (35.9 § 3.3%). No signiWcant inhibition but an increasing serum level of MMP-3 was observed in all treatment groups as indicated by the nega-tive values. However, statistically signiWcant retardation of the increase in serum MMP-3 levels in the HA, LI-HA, and HI-HA groups was found when compared to the con-trol group treated with an injection of 0.9% NaCl. The order of serum MMP-3 retardation of the increase was the LI-HA group (¡64.6 § 12.0%) > HI-HA group (¡75.8 § 5.0%) > HA group (¡81.7 § 26.7%). MMP-3 protein expression levels after the 4 treatments were also assessed by Western blot analysis, and results are shown in Fig.2. There was no signiWcant change in the serum
pro-form MMP-3 protein (calculated as the total intensity of two bands) in the normal group (no immunization) during the treatment course, but there were continual increases at a lower extent than the normal group in serum levels of the pro-form of the MMP-3 protein (calculated as the total inten-sity of two bands) in all treatment groups (immunization) from days 1 to 40.
Macroscopic analysis of the joints
After three intervals (5, 4, and 3 days) of ia treatment, rabbits were sacriWced, and a macroscopic examination revealed that all treated animals had developed typical arthritic lesions (Fig.3). The left and right normal control cartilage was used as references in this study, and macroscopic pannus
formation and erosion shown in Fig.3 were graded 0 (A and B). In the group treated with an injection of 0.5 ml 0.9% NaCl, macroscopic pannus formation of the left knee joint as the internal control which received no treatment was graded 2–3 (Fig.3C), whereas that for the right knee joint was graded 3 (Fig.3D). In the HA group, macroscopic pannus formation of the left knee joint as the internal control which received no treatment was graded 3 (Fig.3E), whereas that for the right knee joint was graded 2¡3 (Fig.3F). In the LI-HA group, macroscopic pannus formation of the left knee joint as the internal control which received no treatment was graded 3 (Fig.3G), whereas that for the right knee joint was graded 2¡3 (Fig.3H). In the HI-HA group, macroscopic pannus formation of the left knee joint as the internal control which received no treatment was graded 3 (Fig.3I), whereas that for the right knee joint was graded 3 (Fig.3J). Table3
summarizes the overall grading scores for all treated animals (n = 5) in each treatment group. There were no obvious diVerences in macroscopic pannus formation between the left and right knee joints for the treatment group injected Table 2 Inhibition (%) of drug treatments (four intra-articular injections in the right knee joint) on antigen-induced arthritis (AIA) in rabbits in comparison to the therapeutic response on day 40
a
The right joints of rabbits with AIA were injected with 0.5 ml 0.9% NaCl, hyaluronate, LI-HA, or HI-HA. Values are the mean § SEM * p < 0.05 versus the baseline for the variables that decreased by one-way ANOVA
Inhibition (%) 0.9% NaCla HAa LI-HAa HI-HAa
CRP 5.0 § 9.5 2.2 § 0.8 54.6 § 0.17* 30.3 § 9.2*
PGE2 ¡53.2 § 10.6 ¡26.0 § 6.1 13.3 § 2.9* 35.9 § 3.3*
MMP-3 ¡160.0 § 11.8 ¡81.7 § 26.7* ¡64.6 § 12.0* ¡75.8 § 5.0*
Fig. 2 MMP-3 protein-expression after the various drug treatments assessed by Western blot analysis. A For pro-form MMP-3 detection, rabbits were neither immunized nor treated in the normal group (n = 3), and were treated with 0.9% NaCl (n = 5) or 1% hyaluronate (n = 5) in the immunized groups. B For pro-form MMP-3 detection, rabbits were treated with 1% hyaluronate alone (HA, n = 5), 5.6M indomethacin + 1% hyaluronate (LI-HA, n = 5), or 560M indomethacin + 1% hyaluronate (HI-HA, n = 5) in the immunized groups. Day 1 was before immunization; day 22 was the immunized state after injection with 0.5 ml of an emulsiWed mixture into both knee joints; and day 40 was the immunized state after the various drug treat-ments. The results shown are representative of four independent exper-iments A 53 kD Pro Normal 0.9% NaCl HA 1 22 40 1 22 40 1 22 40 B 53 kD Pro HA LI-HA HI-HA 1 22 40 1 22 40 1 22 40
with 0.9% NaCl or for the HA group. However, there was obviously improved eYciency when the overall grading of macroscopic pannus formation of the right knee joint was compared to that of the left knee joint in the LI-HA and HI-HA groups with statistically insigniWcant diVerences between these two groups (89.6 § 17.8% for the LI-HA vs. 98.3 § 14.5% for the HI-HA group).
Histological analysis of the joints
Figures4 and 5 show the microscopic features of syno-vial tissues and joint cartilage, respectively, after staining with hematoxylin and eosin for general morphology examination and staining with. BrieXy, these alterations can be grouped into 5 diVerent events: hyperplasia of cells of the synovial lining, inWltration of the sublining by mononuclear cells frequently organized into aggre-gates, leakage and erosion of the joint, and loss of proteo-glycan. The mean § SEM of grading scale scores on both knees and the ratio of the right to the left knee for each animal calculated for 5 diVerent events in the 4 treatment groups are summarized in Table3. There were greater extents of lining hyperplasia and inXammatory cell inWl-tration in all treatment groups than in the normal syno-vium (p < 0.05). In comparisons among the 4 treatment groups, a greater extent of hyperplasia in the treatment group injected with 0.9% NaCl was noted (2.00 § 0.41 for the right knee vs. 1.75 § 0.25 for the left knee). Nev-ertheless, improved eYciency in retardation of the pro-gression of hyperplasia was seen in the LI-HA group (1.67 § 0.58 for the right knee vs. 1.83 § 0.29 for the left knee) or at least no further deterioration in the extent of hyperplasia for the HA and HI-HA groups (2.0 § 0.0 for both knees in these 2 groups). Regarding inXamma-tion, no diVerence was shown between the 2 knees in the group treated with 0.9% NaCl, whereas a slight decrease in the extent of inXammation in the treated knee was observed for the other 3 groups in the order of LI-HA (75.0 § 8.3%) t HI-HA (80.0 § 8.2%) > HA (91.7 § 8.3%). There were greater extents of leakage and erosion of the joints of both knees in all treatment groups than in the normal joints (p < 0.05) after treatment. In compari-sons between the internal control left knee and the treated right knee, however, all 4 treatments demonstrated improvements in joint leakage and erosion but to di Ver-ent extVer-ents in the order of the LI-HA group (72.2 § 14.7% and 61.1 § 5.6%) t HI-HA group (76.7 § 10.0% and 66.7 § 11.8%) > HA group (87.5 § 12.5% and 66.7 § 33.3%) > NaCl group (91.7 § 8.3% and 91.7 § 8.3%). Similarly, the loss of proteoglycan examined after staining with toluidine blue (data not shown) was greater for both treated and untreated knees in all treatment groups than for the corresponding site of normal joints (p < 0.05). However, comparisons between the untreated left knee as the internal control and the treated right knee demonstrated no improvement in the loss of proteoglycan in the HA group, whereas statistically signiWcant decreases in the loss of proteoglycan were observed for the other 3 treatment groups, with that for the LI-HA group (62.5 § 12.5%) being greatest.
Fig. 3 Macroscopic appearance of an opened joint with graded scales as follows for pannus formation and erosions: 0, none; 1, mild; 2, mod-erate; and 3, severe involvement. The left and right normal control car-tilage was used in this study, and macroscopic pannus formation and erosion were graded 0 (A and B). The left knee joint as the internal con-trol received no treatment and macroscopic pannus formation was graded 2¡3 (C), while the right knee joint was injected with 0.5 ml of 0.9% NaCl and macroscopic pannus formation was graded 3 (D). The left knee joint as the internal control received no treatment and macro-scopic pannus formation was graded 3 (E), while the right knee joint was injected with 0.5 ml of 1% hyaluronate (HA group) and macro-scopic pannus formation was graded 2¡3 (F). The left knee joint as the internal control received no treatment and macroscopic pannus forma-tion was graded 3 (G), while the right knee joint was injected with 0.5 ml of 5.6M indomethacin + 1% hyaluronate (LI-HA group) and macroscopic pannus formation was graded 3¡4 (H), while the left knee joint as the internal control received no treatment and macro-scopic pannus formation was graded 3 (I), while the right knee joint was injected with 0.5 ml of 560M indomethacin + 1% hyaluronate (HI-HA group), and macroscopic pannus formation was graded 3 (J). Only one typical image of the left and the right knees for each treat-ment group is shown
Left Normal 0.9% NaCl HA LI-HA HI-HA Right A C D F G I J H E B
Discussion
The eVects after treatments of ia injections with HA alone and combined with 2 levels of indomethacin on alleviating inXam-mation and improving cartilage degradation in rabbits with antigen-induced arthritis were evaluated by examining the CRP level for the resolution of inXammation, the PGE2 level for the reduction of inXammatory pain, and the serum MMP-3 level and macroscopic and histological analyses of the joints for protection against cartilage damage. Results indicated that treatment by an ia injection with HA alone (the HA group) was able to ineYcaciously decrease the CRP serum level, to only moderately suppress the serum level of PGE2, to slightly but signiWcantly inhibit the sustained elevated serum level of MMP-3, and to protect against cartilage damage only to some extent as evidenced by no diVerence in macroscopic pannus formation, no further deterioration in the extent of hyperpla-sia, a slight decrease in the extent of inXammation, some improvement in joint leakage and erosion, but no improve-ment in the loss of proteoglycan. As expected, treatimprove-ments with an ia injection of HA in combination with 2 levels of indo-methacin (in the LI-HA and HI-HA groups) were both more beneWcial to arthritic therapy than HA alone as evidenced by statistically signiWcant improvements in the responses of all parameters examined, with the low concentration of
indomethacin (the LI-HA group) being more eVective than the high concentration of indomethacin (the HI-HA group).
CRP is produced by the liver in response to circulating IL-6, tumor necrosis factor (TNF), or IL-1, and is a tra-ditional marker of systemic inXammation, including that originating in joints [23]. CRP has been employed since the 1950s as a laboratory marker of the inXammatory response [24]. The serum CRP level was also found to be a good parameter for prognostic purposes and for monitoring the treatment eVect if the inXuences of other stimuli of the acute phase response are excluded [25]. Results in Table2
show that there were statistically signiWcant decreases in serum CRP levels in the LI-HA (54.6 § 0.2%) and HI-HA groups (30.3 § 9.2%) compared to the HA group (2.2 § 0.8%), whose eVect was limited and little diVerent from the control group injected with 0.9% NaCl (5.0 § 9.5%). This indicates that an ia injection of HA alone was insuYcient for alleviating inXammation in rab-bits with AIA, and the combination of HA with indometha-cin, regardless of the level, was able to enhance alleviation of inXammation. However, the alleviating eVect on inXam-mation by the low concentration of indomethacin was more profound than that by the high concentration of indometha-cin as indicated by the greater extent of suppression of serum CRP levels. Since serum CRP is released by hepatoma Table 3 Macroscopic
appearance and histological score analysis of the knee joint after various drug treatments (four intra-articular injections in the right knee joint) on antigen-induced arthritis (AIA) in rabbits
Variables Normal 0.9% NaCl HA LI-HA HI-HA
Pannus formation Left joint 0.00 § 0.00 2.25 § 0.48 2.50 § 0.29 3.50 § 0.29 3.00 § 0.32 Right joint 0.00 § 0.00 2.25 § 0.48 2.50 § 0.29 3.00 § 0.41 2.80 § 0.20 Right/left ratio (%) 0.00 § 0.00 100.0 § 0.00 100.0 § 0.00 89.6 § 17.8 98.3 § 14.5 Synovial hyperplasia Left joint 0.00 § 0.00 1.75 § 0.25 2.00 § 0.00 1.83 § 0.29 2.00 § 0.00 Right joint 0.00 § 0.00 2.00 § 0.41 2.00 § 0.00 1.67 § 0.58 2.00 § 0.00 Right/left ratio (%) 0.00 § 0.00 112.5 § 12.5 100.0 § 0.00 87.5 § 12.5 100.00 § 0.00 InXammation Left joint 0.00 § 0.00 3.00 § 0.41 2.25 § 0.25 3.00 § 0.00 2.80 § 0.20 Right joint 0.00 § 0.00 3.00 § 0.41 2.00 § 0.00 2.50 § 0.29 2.20 § 0.20 Right/left ratio (%) 0.00 § 0.00 100.0 § 0.00 91.7 § 8.33 75.0 § 8.33 80.0 § 8.16 Joint leakage Left joint 0.33 § 0.33 2.25 § 0.63 1.75 § 0.63 2.33 § 0.33 2.60 § 0.51 Right joint 0.00 § 0.00 2.00 § 0.41 1.50 § 0.65 1.67 § 0.33 1.80 § 0.20 Right/left ratio (%) 0.00 § 0.00 91.7 § 8.33 87.5 § 12.5 72.2 § 14.7 76.7 § 10.0 Erosion Left joint 0.00 § 0.00 2.00 § 0.41 1.00 § 0.41 2.67 § 0.33 2.50 § 0.85 Right joint 0.00 § 0.00 1.75 § 0.25 0.75 § 0.48 1.67 § 0.33 1.50 § 0.29 Right/left ratio (%) 0.00 § 0.00 91.7 § 8.33 66.7 § 33.3 61.1 § 5.56 66.7 § 11.8 Loss of proteoglycan Left joint 0.00 § 0.00 7.50 § 1.44 7.50 § 1.44 8.75 § 1.25 9.00 § 1.00 Right joint 0.00 § 0.00 5.00 § 0.00 7.50 § 1.44 5.00 § 0.00 6.00 § 1.00 Right/left ratio (%) 0.00 § 0.00 75.0 § 14.4 100.0 § 0.00 62.5 § 12.5 80.0 § 12.5 Values are the mean § SEM
score. The macroscopic appearance and histological analyses of the knee joint (Right/ left ratio, %) after the various drug treatments were statistically analyzed using Bonferroni’s t-test. *p < 0.05
cells in response to the induction of IL-6 secreted by mac-rophages during the acute-phase response, the serum level or receptor expression of the IL-6 receptor should corre-spondingly have been more-strongly suppressed in the LI-HA group than in the HI-LI-HA group [26].
The increasing serum level of PGE2 was moderately suppressed in the HA group (¡26.0 § 6.1%) in comparison to that in the control group injected with 0.9% NaCl (¡53.2 § 10.6%). This reveals that an ia injection of HA alone is able to inhibit the formation of PGE2 but not to a suYcient extent. Indomethacin, a NSAID, is therapeutically used for its ability to block COX activity resulting in a smaller amount of PGE2 being formed. As expected, there was more-eYcient reduction in the serum PGE2 level in the HI-HA group (35.9 § 3.3%) than in the LI-HA group (13.3 § 2.9%) since a higher concentration of indometha-cin was combined with HA in the former group. Therefore,
the higher concentration of indomethacin in the HI-HA group did produce greater suppression of the serum PGE2 level, but less suppression of the serum CRP level, whereas the opposite was true for the low concentration of indo-methacin in the LI-HA group. It has been reported that PGE2 is a potent downregulator of IL-6 receptor expression in the NFS-60 cell line and mediates its eVects through an EP2-receptor-mediated cAMP signaling pathway [27]. Thus, this may provide a mechanistic explanation for the increased serum level of PGE2 possibly downregulating IL-6 receptor expression in hepatoma cells leading to a greater reduction in the extent of CRP secretion in the LI-HA group than in the HI-HA group. Therefore, it was con-cluded that an ia injection of HA combined with either a low or high concentration of indomethacin is beneWcial for the treatment of AIA, with the low concentration of indo-methacin capable of suppressing the formation of CRP to Fig. 4 Histological analysis of the synovial membrane after various
drug treatments by intra-articular injections in the right knee joint of rabbits with antigen-induced arthritis (AIA). A hematoxylin-and-eosin (H&E)-stained 5-m longitudinal section (see Fig. 3, green section) through the knee joint of a normal rabbit and a similar section in a rabbit exhibiting arthritis are shown for general morphology examination. The arthritic joint has increased proteinaceous Xuid and abundant neutroph-ils in the joint space. The synovial lining is thickened with an increase in type II synovial epithelium and adjacent inWltrates of lymphocytes and plasma cells. The left and right normal control cartilage was used in this study (A and B). The left knee joint as the internal control received
no treatment (C), while the right knee joint was injected with 0.5 ml of 0.9% NaCl (D). The left knee joint as the internal control received no treatment, while the right knee joint was injected with 0.5 ml of 1% hy-aluronate (HA group, F). The left knee joint as the internal control re-ceived no treatment (G), while the right knee joint was injected with 0.5 ml of 5.6M indomethacin + 1% hyaluronate (LI-HA group, H). The left knee joint as the internal control received no treatment (I), while the right knee joint was injected with 0.5 ml of 560M indomethacin + 1% hyaluronate (HI-HA group, J). Only one typical image of the left and the right knees at two magniWcations (*-1: £25 and *-2: £30¡80) for each treatment group is shown
Normal 0.9% NaCl C-1 C-2 D-1 D-2 HA E-1 E-2 F-1 F-2 LI-HA G-1 G-2 H-1 H-2
HI-HA I-1 I-2 J-1 J-2
Left Right
alleviate inXammation and decrease complement activation [28] and the high concentration capable of inhibiting the formation of PGE2 to reduce the inXammatory pain.
In contrast to CRP, a marker of systemic inXammation, MMP-3 is produced in the joint in response to local IL-6, TNF, and IL-1 and is a more-speciWc marker of synovial inXammation [23]. Elevation of MMP-3 serum levels might be restricted to inXammatory diseases associated with joints, as it is intensively expressed in the rheumatoid syno-vium. Since MMP-3 has potent activity of degrading the proteoglycan of cartilage [29], serum levels of MMP-3 have previously been shown to be correlated with radiological
damage in RA patients [29–31]. The results of this study demonstrate that the serum level of MMP-3 was elevated and sustained (Table1) in the group treated by injecting 0.9% NaCl, whereas elevation of MMP-3 serum levels was signiWcantly suppressed in the other 3 treatment groups (Table2), especially in the 2 treatment groups that com-bined HA with indomethacin at a low and high level. Therefore, it is expected that indomethacin at both levels in combination with HA would be beneWcial in the prevention of cartilage damage, as reXected by further suppression of serum levels of MMP-3, with the low concentration of indo-methacin exhibiting a greater extent of suppression than the high concentration.
Treatment with an ia injection of HA alone was expected to be eYcacious in the prevention of cartilage damage as reXected by a reduction in serum MMP-3 levels. MMP-3 is also recognized as an enzyme which plays a part in the destruction of cartilage and bone in rheumatoid arthritis (RA) [32–33]. Patients with RA have increased serum levels of MMP-3, which is thought to originate from the synovium [34,35] and strongly suggests that it reXects
syno-vial inXammation [25]. In addition, serum levels of MMP-3 are correlated with the number of joints aVected and are decreased after an ia injection of steroids [36]. Therefore, serum MMP-3 levels were expected to be elevated after induction of AIA and be continuously sustained during the treatment period for the control group (which received an ia injection of a 0.9% NaCl solution in the right knee), whereas increases in the serum levels of MMP-3 were sup-pressed in the 3 other treatment groups (HA, LI-HA, and HI-HA) since their arthritic right knee joints were eYca-ciously treated. It has been documented that ia administra-tion of HA in the rabbit knee inhibits MMP-3 and TIMP-1 production at the mRNA level in cartilage and synovium [37]. It was also reported that one mechanism of the thera-peutic eVect of HA by ia injection into the ACLT (anterior cruciate ligament transection) of rabbit knees is downregu-lation of MMP-3 and IL-1 in the synovium [38,39]. An in vitro study by Sasaki et al. revealed that HA inhibits the expression and production of MMP-1 and MMP-3 in IL-1-stimulated human synovial cells [40]. A similar eVect of inhibiting MMP-3 synthesis induced by IL- in human OA chondrocytes was disclosed for both chondroitin sulfate and HA (500¡730 kDa) [41]. Therefore, it was concluded that the observed suppression of increasing MMP-3 serum levels by an ia injection of HA was due to the inhibition of MMP-3 expression and production in the synovium and possibly in chondrocytes.
Indomethacin in combination with HA is beneWcial for pre-venting cartilage damage by further suppressing serum levels of MMP-3 with the low concentration of indomethacin pro-ducing a greater extent of suppression than the high concen-tration. It was reported by Sadowski and Steinmeyer that Fig. 5 Histological analysis of knee joint damage after the various
drug treatments by intra-articular injections in the right knee joint of rabbits with antigen-induced arthritis (AIA). A hematoxylin-and-eosin (H&E)-stained 5-m longitudinal section (see Fig. 3, green section) through the knee joint (articular surface of the lateral coudyle of femur) of a normal rabbit and a similar section in a rabbit exhibiting arthritis are shown for general morphology examination. Left and right normal control cartilage was used in this study (A and B). The left knee joint as the internal control received no treatment (C), while the right knee joint was injected with 0.5 ml of 0.9% NaCl (D). The left knee joint as the internal control received no treatment (E), while the right knee joint was injected with 0.5 ml of 1% hyaluronate (HA group, F). The left knee joint as the internal control received no treatment (G), while the right knee joint was injected with 0.5 ml of 5.6M indomethacin + 1% hyaluronate (LI-HA group, H). The left knee joint as the internal con-trol received no treatment (I), while the right knee joint was injected with 0.5 ml of 5.6M indomethacin + 1% hyaluronate (HI-HA group,
J). Only one typical image (£20¡40 magniWcation) of the left and the
right knees for each treatment group is shown.
Left Normal 0.9% NaCl HA LI-HA HI-HA Right A B C D E F G H I J
when tested at a concentration of 10M in bovine articular chondrocytes, indomethacin inhibited MMP-3 expression [42]. This Wnding is in agreement with a previous study by
Yamada et al. in which indomethacin reduced MMP-3 pro-duction by human chondrocytes [43]. Further, both studies demonstrated that the addition of exogenous PGE2 did not reverse the eVect of indomethacin, and it was noted that MMP-3 inhibition by this drug is independent of PGE2 syn-thesis. Further, oral administration of nimesulide, a COX-2 selective inhibitor, signiWcantly reduced serum levels of MMP-3, whereas that of ibuprofen, a COX-1/COX-2 inhibi-tor, moderately but signiWcantly increased the serum concen-trations of MMP-3 in patients with OA [44]. It was also found that indomethacin (a COX-2/COX-2 inhibitor) and NS-398 (a COX-2 selective inhibitor) enhanced IL-1-induced MMP-3 production in human PDL (periodontal ligament) cells, but both agents completely inhibited IL-1-induced PGE2 pro-duction. Since exogenous PGE2 reduced IL-1-induced MMP-3 production in a dose-dependent manner and both EP2 and EP4 agonists signiWcantly inhibited IL-1-induced MMP-3 production, it was concluded by this study that COX-2-dependent PGE2 downregulates IL-1-elicited MMP-3 production by c-AMP-dependent pathways via EP2/EP4 receptors in human PDL cells [45]. However, although this fact is in conXict with the results disclosed in the above study, it provides a possible mechanistic explanation for how the low or high concentration of indomethacin inXu-ences serum levels of MMP-3. A higher serum level of PGE2 as a result of giving a low dose of indomethacin in the LI-HA group should have reduced MMP-3 production following the mechanism discussed above to a greater extent than that by the low serum level of PGE2 resulting from administering a high dose of indomethacin in the HI-HA group.
Furthermore, MMP-3 serum levels are reported to be cor-related with parameters of inXammation including the eryth-rocyte sedimentation rate (ESR), CRP, and interleukin (IL)-6 levels [46–48]. This is consistent with the results demon-strated by the low serum level of MMP-3 correlating with the low level of CRP in the LI-HA group and the high serum level of MMP-3 with the high level of CRP in the HI-HA group in this study. As discussed above, the change in CRP serum levels might also have been mediated by PGE2 through the EP2-receptor-mediated cAMP signaling path-way to downregulate the receptor expression of IL-6 that is responsible for secretion of CRP in hepatoma cells. Hypo-thetically, PGE2 might play a determining role in the corre-lation between serum levels of MMP-3 and CRP. However, this could create a dilemma in the choice of which level of indomethacin to combine with HA to produce eVective improvement in antiarthritic therapy, since a low dose of indomethacin could lead to a higher serum level of PGE2 but a low level of MMP-3 and vice versa for a high dose of indomethacin. If destruction of cartilage and bone by
MMP-3 plays the determining role in the disease progression of arthritic joints, a low dose of indomethacin might be the best choice in combination with HA for ia administration.
For the macroscopic and histological analyses, a statisti-cally insigniWcant diVerence in most of those parameters examined was found when comparing the treated groups with the control group. This could be attributed to either a high variability of macroscopic and histological analyses or the treatment period not being long enough to allow signiWcant improvements to develop. Since diVerences in the serum levels of CRP, PGE2, and MMP-3 were statisti-cally determined to have suYcient power with this sample size, it would be reasonable to accept the former. Although there were no statistically signiWcant diVerences in most of the parameters examined by macroscopic and histological analyses in the 3 treatment groups (HA, LI-HA, and HI-HA) in comparison with the control group (0.9% NaCl), there was a little more improvement in the therapeutic eYciency in the LI-HA group in comparison to the HA group in terms of pannus formation, synovial hyperplasia and inXammation, leakage and erosion of the joint, and loss of proteoglycan (Table3). As discussed above, correspond-ingly lower serum levels of CRP and MMP-3 in the LI-HA treatment group in comparison with those in the HA and HI-HA groups might have been responsible for this.
Conclusion
In conclusion, it was found in this study that hyaluronate combined with a low dose of indomethacin (5.6M) might provide substantially more clinical beneWts to arthritic rab-bits based on suppression of the serum levels of CRP and MMP-3 with correspondingly improved therapeutic eYciency observed in the macroscopic and histological analyses. The results of this study demonstrated that MMP-3 was elevated and that this elevation was sustained (Table1) in all drug treatment groups, which suggests that MMP inhi-bition may be a potential therapeutic strategy. MMP inhibi-tors have frequently exhibited toxicity in clinical trials with systemic administration. A local injection combined with NSAIDs and an MMP inhibitor in a hyaluronate solution might be another appropriate strategy for arthritis therapy. Acknowledgments This study was sponsored by Min-Sheng Health-care (93MSH-TMU-17). The authors would like to express sincere thank Mr. Wen-Chung Lee for her assistance with the histological staining.
References
1. Walker JM, Helewa A (1996) Physical therapy in arthritis. Saun-ders, Philadelphia
2. Shahbaz H, James WS (2001) Septic arthritis. Curr Treat Opt In-fect Dis 3:279–286
3. Goldring SR, Gravallese EM (2000) Mechanisms of bone loss in inXammatory arthritis: diagnosis and therapeutic implications. Arthritis Res 2:33–37
4. Brandt KD, Smith GN Jr, Simon LS (2000) Intraarticular injection of hyaluronan as treatment for knee osteoarthritis: what is the evi-dence? Arthritis Rheum 43:1192–1203
5. Ceponis A, Waris E, Monkkonen J, Laasonen L, Hyttinen M, Solovieva SA, Hanemaaijer R, Bitsch A, Konttinen YT (2001) EVects of low-dose, noncytotoxic, intraarticular liposomal clodro-nate on development of erosions and proteoglycan loss in established antigen-induced arthritis in rabbits. Arthritis & Rheum 44:1908–1916
6. Campo GM, Avenoso A, Campo S, Ferlazzo AM, Altavilla D, Calatroni A (2003) EYcacy of treatment with glycosaminoglycans on experimental collagen-induced arthritis in rats. Arthritis Res Ther 5:R122–R131
7. Moreland LW (2003) Intra-articular hyaluronan (hyaluronic acid) and hylans for the treatment of osteoarthritis: mechanisms of ac-tion. Arthritis Res Ther 5:54–67
8. Aggarwal A, Sempowski IP (2004) Hyaluronic acid injections for knee osteoarthritis. Systematic review of the literature Can Fam Physician 50:249–256
9. Yasui T, Akatsuka M, Tobetto K, Hayaishi M, Ando T (1992) The eVect of hyaluronan on interleukin-1 alpha-induced prostaglandin E2 production in human osteoarthritic synovial cells. Agents Actions 37:155–156
10. Tobetto K, Yasui T, Ando T, Hayaishi M, Motohashi N, Shinogi M, Mori I (1992) Inhibitory eVects of hyaluronan on [14C] arachi-donic acid release from labeled human synovial Wbroblasts. Jpn J Pharmacol 60:79–84
11. Larsen NE, Lombard KM, Parent EG, Balazs EA (1992) EVect of hylan on cartilage and chondrocyte cultures. J Orthop Res 10:23–32 12. Presti D, Scott JE (1994) Hyaluronan-mediated protective eVect against cell damage caused by enzymatically produced hydroxyl (OH.) radicals is dependent on hyaluronan molecular mass. Cell Biochem Funct 12:281–288
13. Gomis A, Pawlak M, Balazs EA, Schmidt RF, Belmonte C (2004) EVects of diVerent molecular weight elastoviscous hyaluronan solu-tions on articular nociceptive aVerents. Arthritis Rheum 50:314–326 14. Roth A, Mollenhauer J, Wagner A, Fuhrmann R, Straub A, Venbrocks RA, Petrow P, Bräuer R, Schubert H, Ozegowski J, Peschel G, Müller PJ, Kinne RW (2005) Intra-articular injections of high-molecular-weight hyaluronic acid have biphasic eVects on joint inXammation and destruction in rat antigen-induced arthritis. Arthritis Res Ther 7:R677–R686
15. Williams HJ, Ward JR, Egger MJ, Neuner R, Brooks RH, Clegg DO, Field EH, Skosey JL, Alarcon GS, Paulus HE, Russell IJ, Sharp JT (1993) Comparison of naproxen and acetaminophen in a two-year study of treatment of osteoarthritis of the knee. Arthritis & Rheum 36:1196–1206
16. Zhang W, Jones A, Doherty M (2004) Does paracetamol (acetami-nophen) reduce the pain of osteoarthritis? A meta-analysis of randomised controlled trials. Ann Rheum Dis 63:901–907 17. Wolfe F, Zhao S, Lane N (2000) Preference for nonsteroidal
anti-inXammatory drugs over acetaminophen by rheumatic disease pa-tients: a survey of 1,799 patients with osteoarthritis, rheumatoid arthritis, and Wbromyalgia. Arthritis Rheum 43:378–385 18. Huang SH (2000) Rheumatology: 7. Basics of therapy. CMAJ
163:417–423
19. Yoon JB, Kim SJ, Hwang SG, Chang S, Kang SS, Chun JS (2003) Non-steroidal anti-inXammatory drugs inhibit nitric oxide-induced apoptosis and dediVerentiation of articular chondrocytes independent of cyclooxygenase activity. J Biol Chem 278:15319– 15325
20. Dawson J, Engelhardt P, Kastelic T, Cheneval D, MacKenzie A, Ramage P (1999) EVects of soluble interleukin-1 type II receptor on rabbit antigen-induced arthritis: clinical, biochemical and his-tological assessment. Rheumatology (Oxford) 38:401–406 21. Shafer-Weaver KA, Sayers T, Kuhnd DB, Strobl SL, Burkett MW,
Baseler M, Malyguine A (2004) Evaluating the cytotoxicity of in-nate immune eVector cells using the GrB ELISPOT assay. J Transl Med 2:31
22. Hanke JH, Gardner JP, Dow RL, Changelian PS, Brissette WH, Weringer EJ, Pollok BA, Connelly PA (1996) Discovery of a novel, potent, and Src family-selective tyrosine kinase inhibitor. Study of Lck- and FynT-dependent T cell activation. J Biol Chem 271:695–701
23. Ribbens C, Porras M, Martin Y, Franchimont N, Kaiser MJ, Jaspar JM, Damas P, Houssiau FA, Malaise MG (2002) Increased matrix metalloproteinase-3 serum levels in rheumatic diseases: relationship with synovitis and steroid treatment. Ann Rheum Dis 61:161–166 24. Posthumus MD, Limburg PC, Westra J, Cats HA, Stewart RE, van
Leeuwen MA, van Rijswijk MH (1999) Serum levels of matrix metalloproteinase-3 in relation to the development of radiological damage in patients with early rheumatoid arthritis. Rheumatology 38(11):1081–1087
25. Atkinson JP (2001) C-Reactive protein: a Rheumatologist’s friend revisited. Arthritis Rheum 44:995–996
26. Weinhold B, Bader A, Poli V, Ruther U (1997) Interleukin-6 is necessary, but not suYcient, for induction of the human C-reactive protein gene in vivo. Biochem J 325:617–621
27. de Silva KI, Daud AN, Deng JP, Jones SB, Gamelli RL, Shankar R (2003) Prostaglandin E2 mediates growth arrest in NFS-60 cells by down-regulating interleukin-6 receptor expression. Biochem J 370:315–321
28. Familian A, Voskuyl AE, van Mierlo GJ, Heijst HA, Twisk JW, Dijkmans BA, Hack CE (2005) InXiximab treatment reduces com-plement activation in patients with rheumatoid arthritis. Ann Rheum Dis 64:1003–1008
29. Yamanaka H, Matsuda Y, Tanaka M, Sendo W, Nakajima H, Taniguchi A, Kamatani N (2000) Serum matrix metalloproteinase 3 as a predictor of the degree of joint destruction during the six months after measurement, in patients with early rheumatoid arthritis. Arthritis Rheum 43:852–858
30. Posthumus MD, Limburg PC, Westra J, Cats HA, Stewart RE, van Leeuwen MA, van Rijswijk MH (1999) Serum levels of matrix metalloproteinase-3 in relation to the development of radiological damage in patients with early rheumatoid arthritis. Rheumatology (Oxford) 38:1081–1087
31. Catrina AI, Lampa J, Ernestam S, af Klint E, Bratt J, Klareskog L, Ulfgren AK (2002) Anti-tumour necrosis factor (TNF)-alpha ther-apy (etanercept) down-regulates serum matrix metalloproteinase (MMP)-3 and MMP-1 in rheumatoid arthritis. Rheumatology (Oxford) 41:484–489
32. Vincenti MP, Clark IM, BrinckerhoV CE (1994) Using inhibitors of metalloproteinases to treat arthritis. Easier said than done? Arthritis Rheum 37:1115–1126
33. Cawston T (1998) Matrix metalloproteinases and TIMPs: proper-ties and implications for the rheumatic diseases. Mol Med Today 4:130–137
34. Sasaki S, Iwata H, Ishiguro N, Obata K, Miura T (1994) Detection of stromelysin in synovial Xuid and serum from patients with rheu-matoid arthritis and osteoarthritis. Clin Rheumatol 13:228–233 35. Ribbens C, Andre B, Jaspar JM, Kaye O, Kaiser MJ, De Groote D,
Malaise MG (2000) Matrix metalloproteinase-3 serum levels are correlated with disease activity and predict clinical response in rheumatoid arthritis. J Rheumatology 27:888–893
36. Taylor DJ, Cheung NT, Dawes PT (1994) Increased serum proM-MP-3 in inXammatory arthritis: a potential indicator of synovial inXammatory monokine activity. Ann Rheum Dis 53:768–772
37. Han F, Ishiguro N, Ito T, Sakai T, Iwata H (1999) EVects of sodi-um hyaluronate on experimental osteoarthritis in rabbit knee joints. Nagoya J Med Sci 62:115–126
38. Takahashi K, Goomer RS, Harwood F, Kubo T, Hirasawa Y, Amiel D (1999) The eVects of hyaluronan on matrix metallopro-teinase-3 (MMP-3), interleukin-1beta(IL-1beta), and tissue inhibitor of metalloproteinase-1 (TIMP-1) gene expression dur-ing the development of osteoarthritis. Osteoarthritis Cartilage 7:182–190
39. Qiu B, Liu SQ, Peng H, Wang HB (2005) The eVects of sodium hyaluronate on mRNA expressions of matrix metalloproteinase-1, -3 and tissue inhibitor of metalloproteinase-1 in cartilage and synovium of traumatic osteoarthritis model. Chin J Traumatol 8:8–12
40. Sasaki A, Sasaki K, Konttinen YT, Santavirta S, Takahara M, Takei H, Ogino T, Takagi M (2004) Hyaluronate inhibits the inter-leukin-1beta-induced expression of matrix metalloproteinase (MMP)-1 and MMP-3 in human synovial cells. Tohoku J Exp Med 204:99–107
41. Monfort J, Nacher M, Montell E, Vila J, Verges J, Benito P (2005) Chondroitin sulfate and hyaluronic acid (500–730 kDa) inhibit stromelysin-1 synthesis in human osteoarthritic chondrocytes. Drugs Exp Clin Res 31:71–76
42. Sadowski T, Steinmeyer J (2001) EVects of non-steroidal antiinXammatory drugs and dexamethasone on the activity and expression of matrix metalloproteinase-1, matrix metalloprotein-ase-3 and tissue inhibitor of metalloproteinases-1 by bovine articular chondrocytes. Osteoarthritis & Cartilage 9:407–415
43. Yamada H, Kikuchi T, Nemoto O, Obata K, Sato H, Seiki M, Shinmei M (1996) EVects of indomethacin on the production of matrix metalloproteinase-3 and tissue inhibitor of metallo-proteinases-1 by human articular chondrocytes. J Rheumatol 23:1739–1743
44. Bevilacqua M, Devogelaer JP, Righini V, Famaey JP, Manicourt DH (2004) EVect of nimesulide on the serum levels of hyaluronan and stromelysin-1 in patients with osteoarthritis: a pilot study. Int J Clin Pract Suppl 144:13–19
45. Yan M, Noguchi K, Ruwanpura SM, Ishikawa I (2005) Cycloox-ygenase-2-dependent prostaglandin (PG) E2 downregulates ma-trix metalloproteinase-3 production via EP2/EP4 subtypes of PGE2 receptors in human periodontal ligament cells stimulated with interleukin-1alpha. J Periodontology 76:929–935
46. Calabrese LH, Michel BA, Bloch DA, Arend WP, Edworthy SM, Fauci AS, Fries JF, Hunder GG, Leavitt RY, Lie JT, Lightfoot RW, Masi AT, McShave DJ, Mills JA, Stevens MB, Wallace SL, ZvaiXer NJ (1990) The American College of Rheumatology 1990 criteria for the classiWcation of hypersensitivity vasculitis. Arthri-tis & Rheum 33:1108–1113
47. Brennan FM, Browne KA, Green PA, Jaspar JM, Maini RN, Feldmann M (1997) Reduction of serum matrix metalloproteinase 3 in rheumatoid arthritis patients following anti-tumor necrosis factor- (cA2) therapy. Br J Rheumatol 36:643–650
48. Partsch G, Wagner E, Leeb BF, Dunky A, Steiner G, Smolen JS (1998) Upregulation of cytokine receptors sTNF-R55, sTNF-R75, and sIL-2R in psoriatic arthritis synovial Xuid. J Rheumatol 25:105–110
