Initiation of degenerative joint damage by experimental bleeding combined with loading of the jointA possible mechanism of hemophilic arthropathy.код для вставкиСкачать
ARTHRITIS & RHEUMATISM Vol. 50, No. 6, June 2004, pp 2024–2031 DOI 10.1002/art.20284 © 2004, American College of Rheumatology Initiation of Degenerative Joint Damage by Experimental Bleeding Combined With Loading of the Joint A Possible Mechanism of Hemophilic Arthropathy Michel J. J. Hooiveld, Goris Roosendaal, Kim M. G. Jacobs, Marieke E. Vianen, H. Marijke van den Berg, Johannes W. J. Bijlsma, and Floris P. J. G. Lafeber altered, as shown by histologic damage. Histologic analysis also revealed signs of synovial inflammation. These effects were not observed 10 weeks after the experimental bleedings in joints that did not undergo forced loading. Conclusion. Experimental joint bleedings when combined with loading of the affected joint resulted in features of progressive degenerative joint damage, whereas similar joint hemorrhages without joint loading did not. This might reflect a possible mechanism of joint damage in hemophilia. Objective. To investigate the effect of a limited number of experimental joint bleedings, combined with loading of the affected joint, on the development of progressive degenerative joint damage. Methods. The right knee of 8 mature beagle dogs was injected with freshly collected autologous blood 3 times per week for 4 weeks, to mimic a limited number of joint hemorrhages occurring over a short period. To ensure loading of the experimental joint, the contralateral control knee of the animals was fixed to the trunk 4 hours per day, 3 days per week. Ten weeks after the last injection, cartilage tissue and synovium were collected from both knees to analyze features of joint degeneration. Cartilage was prepared for analysis of proteoglycan turnover (synthesis, retention, release, and content) and histologic features. Synovium was prepared for histologic analysis. Results. The rate of proteoglycan synthesis was significantly increased, characteristic of degenerative cartilage damage as seen in osteoarthritis. Release of newly formed proteoglycans (as a measure of retention) and total loss of proteoglycans from the cartilage matrix were increased. Cartilage matrix integrity was adversely Blood-induced arthropathy is the most common clinical phenomenon in patients with severe hemophilia (⬍1% clotting factor VIII or IX) (1). Current opinion regarding the pathogenesis of hemophilic arthropathy is that recurrent intraarticular bleeding leads to production of catabolic factors by inflamed synovium. Although synovectomy improves clinical status in these patients, it does not end the progressive cartilage damage (2,3). Moreover, hemophilic arthropathy has been described to be more degenerative in nature, as seen in osteoarthritis, as opposed to inflammatory, as in rheumatoid arthritis (4). Previous studies have shown that synovitis is indeed not primarily responsible for the cartilage damage after intraarticular bleeding, but rather that exposure of cartilage to blood induces direct adverse effects (5,6). These cartilage alterations are primarily induced by the combination of mononuclear cells and red blood cells. This combination of blood cells has been found to act on chondrocyte activity, leading to a long-lasting inhibition of cartilage matrix synthesis and an increased loss of matrix components in vitro (6). These in vitro effects have been sustained for at least 10 weeks (7). Supported by grants from Aventis Behring (Zaventem, Belgium) and the van Creveld Foundation (Utrecht, The Netherlands). Michel J. J. Hooiveld, PhD, Goris Roosendaal, MD, PhD, Kim M. G. Jacobs, BSc, Marieke E. Vianen, BSc, H. Marijke van den Berg, MD, PhD, Johannes W. J. Bijlsma, MD, PhD, Floris P. J. G. Lafeber, PhD: University Medical Center Utrecht, Utrecht, The Netherlands. Address correspondence and reprint requests to Floris P. J. G. Lafeber, PhD, Department of Rheumatology & Clinical Immunology, F02.127, University Medical Center Utrecht, PO Box 85500, 3508 GA, Utrecht, The Netherlands. E-mail: F.Lafeber@azu.nl. Submitted for publication May 20, 2003; accepted in revised form February 16, 2004. 2024 JOINT BLEEDINGS AND LOADING IN HEMOPHILIC ARTHROPATHY Moreover, in vitro experiments showed that chondrocyte apoptosis is involved in these unfavorable events (8). In vivo animal studies have shown that adverse changes in cartilage matrix can be induced by a single 4-day experimental intraarticular bleeding. These changes were still present after a 12-day recovery period and appeared to be independent of primary synovial inflammation (9). However, the in vivo effects were less pronounced than those observed in vitro. Apparently, factors in vivo that are not present in vitro are able to partially counteract the unfavorable effects of blood on cartilage. In the clinical setting it has been shown that early treatment of joint bleedings in patients with hemophilia may protect against arthropathy in later years; starting treatment with clotting factors after the third joint hemorrhage during early childhood results in better clinical outcome after 22 years of followup (10). However, this early prophylaxis cannot completely prevent arthropathy. This corroborates our findings, from in vitro and animal studies, that only a few bleeding episodes may induce permanent adverse alterations in cartilage matrix turnover and, consequently, cartilage degeneration after a period of years. In contrast with these data, however, are the results of a previous in vivo study in which we found that a limited number of experimental joint bleedings in dogs did not lead to adverse changes in cartilage matrix turnover, evaluated 10 weeks after the last injection (7). This suggests that in vivo, factors in addition to blood exposure are needed for induction of progressive degenerative features of joint damage. Therefore, in the present study we combined a series of experimental joint bleedings with loading of the affected joint to investigate whether the latter is of importance in the development of characteristics of degenerative joint damage after joint bleeding. MATERIALS AND METHODS Animals. Eight skeletally mature beagle dogs (mean ⫾ SD age 1.8 ⫾ 0.1 years, mean ⫾ SD weight 9.9 ⫾ 1.2 kg) were obtained from the animal facilities of Utrecht University. They were housed in groups of 2 and were let outside daily for at least 2 hours on a large patio. The animals were fed a standard commercial diet with water ad libitum. Approval for this study was obtained from the University Animal Experiments Ethical Committee. Experimental hemorrhages. Freshly collected autologous blood was injected into the right knee with the animal under short-term anesthesia (Domitor/AntiSedan; Orion Farmos, Turku, Finland). The intraarticular injection of blood was confirmed by placing a second needle on the contralateral 2025 side of the joint, through which blood flowed after injection. When this occurred, the second needle was removed and 3 ⫾ 1 ml of blood (mean ⫾ SD) was injected. Thus, blood was added to synovial fluid, closely mimicking conditions that occur in hemophilia. Distension of the knee was moderate and was visible in the first 24 hours and palpable until 48 hours after the injection. No cumulative effect in this respect was observed. Anticoagulants were not used. The contralateral control knee was not injected at all, to prevent possible traumatic bleeding. Blood was injected 3 times per week for 4 consecutive weeks. We have found that up to 14 repeated injections with vehicle in 7 weeks (as a control for the experimental injections), using the same protocol of 2 needles entering the joint, did not result in changes to cartilage or synovial tissue (ref. 11, and Hooiveld MJJ, et al: personal observation). In addition to the blood injections, the animals underwent a forced loading protocol for the injected joints. The contralateral control knee was fixed to the trunk for 4 hours per day, 3 days per week. Thus, during 7% of the duration of the experiment, the blood-injected knee was forced to be loaded. Importantly, this not only results in loading during the fixation procedure, but also facilitates normal usage of the blood-injected joints during the time between the periods of forced loading. The frequency of loading was based on the findings of previous studies in which mechanically induced cartilage damage combined with this loading profile led to features of osteoarthritis (12). At the end of the experiment, 10 weeks after the last injection (week 14), the dogs were killed by intravenous injection of 10 ml of Euthesate (200 mg/ml; Cefa Sante Animal BV, Maassluis, The Netherlands). Results from this study were compared with the results obtained in a study using an experimental procedure that was identical with respect to animals, housing, feeding, blood injections, and followup, except that in the earlier study (of 6 dogs) no forced loading was applied (7). Importantly, in that study, in which the dogs did not undergo forced loading, they were reluctant to load the experimental blood-injected joints during and shortly after the period of injection, as indicated by the fact that they walked on 3 limbs and did not load the bloodinjected joint. After the dogs were killed at the end of the experiments, both hind limbs were amputated and synovium and cartilage were collected and processed within 2 hours. Procedures were carried out under laminar flow conditions. Synovium from the suprapatellar pouch and cartilage from the condyles and plateau of both knee joints were collected under aseptic conditions and analyzed according to standard procedures as described below. Cartilage preparation. Cartilage samples were obtained from weight-bearing areas and were scored according to a slight modification of the Mankin criteria (13). The cartilage was cut as thick as possible, taking care to exclude the underlying bone, and was then cut into square pieces (mean ⫾ SD 6.2 ⫾ 3.1 mg). Within 3 hours after death of the animals, the cartilage tissue samples were individually placed in culture medium in 96-well round-bottomed microtiter plates, for a 1-hour equilibration period. For both femoral and tibial cartilage, the proteoglycan synthesis rate, DNA content, and levels of proteoglycan retention, release, and content were determined. In addition, samples were prepared for histologic analysis. 2026 Measurement of chondrocyte activity. Proteoglycan synthesis rate. As a measure of chondrocyte matrix synthesis, the synthesis of proteoglycan, one of the main matrix components, was determined by assessment of the rate of sulfate incorporation (14). Briefly, after 1 hour of preincubation, 148 kBq of Na235SO4 (NEX-041-H, carrier free) (DuPont, Kortrijk, Belgium), in 10 l Dulbecco’s modified Eagle’s medium was added to 200 l of culture medium. After a 4-hour labeling period, samples were washed twice with ice-cold phosphate buffered saline and digested with papain (Sigma, Zwijndrecht, The Netherlands). Sulfated proteoglycans (glycosaminoglycans [GAGs]) were precipitated by addition of cetylpyridinium chloride (Sigma). The 35SO42⫺-labeled proteoglycans in the precipitate were measured by liquid scintillation analysis. The rate of sulfate incorporation was normalized to the specific activity of the medium and expressed as nmoles of sulfate incorporated per hour per gram wet weight of cartilage tissue. Because the proteoglycan synthesis rate is dependent on cell numbers, the DNA content of the samples was determined as a measure of cellularity. Part of the papain digest was used for measurement of DNA, using the fluorescent dye Hoechst 33258. Calf thymus DNA (D-4764; Sigma) was used as a reference. The amount of sulfate incorporated was also normalized to the DNA content of the samples and expressed as nmoles per hour per milligram DNA. Proteoglycan retention and release. As a measure of retention of newly synthesized proteoglycans in the cartilage, the release of 35SO42⫺-labeled proteoglycans in a 3-day culture medium was determined, using experimental procedures that have been described previously (14). Labeling with 35SO42⫺ was performed as described above. After this 4-hour labeling, samples were rinsed thoroughly and subsequently incubated in fresh medium for 3 days. GAGs were precipitated with Alcian blue. The 35SO42⫺-labeled GAGs were measured by liquid scintillation analysis, and release was normalized to specific activity of the medium and the wet weight of the explants, expressed as nmoles per milligram released in 3 days. Because production of newly formed proteoglycans depends on the proteoglycan synthesis rate, release of these newly formed proteoglycans was also normalized to proteoglycan synthesis and expressed as the percentage release of newly formed proteoglycans. In addition, the Alcian blue–precipitated GAGs were quantified spectrophotometrically with chondroitin sulfate as a reference. This revealed the total release of proteoglycans (resident and newly formed), expressed as GAG released per wet weight of the cartilage sample (milligrams per gram). Because release depends on the content of proteoglycans, total release was also normalized to proteoglycan content and expressed as the percentage release of the total amount of proteoglycans. Determination of cartilage matrix characteristics. Proteoglycan content. The amount of GAG in cartilage was determined as a measure of the proteoglycan content of the cartilage samples (14). In the papain digests, GAGs were precipitated and stained with Alcian blue, as described above. The GAG content was expressed as GAG weight normalized to the wet weight of the cartilage explants (milligrams per gram). Cartilage histology. Four cartilage samples from the femoral condyle and 4 samples from the tibial plateau in the weight-bearing area of each knee were fixed in 4% phosphate HOOIVELD ET AL buffered formalin for histochemistry analysis. After fixation, samples were dehydrated, placed in xylene, and subsequently embedded in paraffin. Sections (4 m) were deparaffinized and stained with Safranin O–fast-green–iron hematoxylin and scored according to a slight modification of the Mankin criteria (13). With this modification, the maximum possible score is 11 instead of 14, with an average score of 5 for cartilage with established osteoarthritis (4,13). At least 3 sections from each of the 8 specimens from each knee were stained and scored independently, in random order, by 2 observers. Assessment of synovial inflammation. Three pieces of synovial tissue were obtained from the suprapatellar pouch of each knee. They were fixed in 4% phosphate buffered formalin for histochemistry analysis. Deparaffinized sections were stained with hematoxylin and eosin. To determine the degree of inflammation, at least 3 sections from each of the 3 specimens from each knee were stained and scored. Synovial inflammation was defined based on lymphocyte infiltration, hyperplasia of the synovial lining cells, and hypertrophy of the synovial membrane, using a slight modification of the scoring system of Goldenberg and Cohen (15) as described previously (14). Two independent observers, unaware of the source of the tissue, scored the tissue samples separately and in random order. Statistical analysis. Power analysis was performed to calculate the number of animals needed in each group to achieve statistical power of ␣ ⫽ 0.05 and ␤ ⫽ 0.10, based on the differences observed in our previous experiments (7,9). We calculated that 6 animals per group would be sufficient to achieve reliable statistical outcomes. Cartilage parameters were determined in 6 cartilage explants from fixed locations from both femoral condyles and tibial plateaus. Because, for all parameters, no differences were found for values obtained from the 6 samples from the condyles and the 6 samples from the plateau, the means of the 12 samples for all parameters were taken as representative of that particular knee. A retrospective comparison with previously published data (7) was made. Mean ⫾ SEM values (n ⫽ 8 animals for the current data, n ⫽ 6 for the retrospective data [the group with joint loading and group without joint loading, respectively]) are presented. To determine the statistical significance of differences between control and experimental knees within a group, we used a nonparametric test (Wilcoxon’s 2-sample test). Differences in the contralateral control joints between the 2 groups were tested with a nonparametric test for independent samples (Mann-Whitney U test). For comparison of the effects of blood injections with forced loading compared with blood injections alone, the difference between blood-injected joints and contralateral control joints was calculated and averaged for each group. Percentage differences were calculated for biochemical parameters and absolute differences for histologic parameters. For comparison of the 2 groups in terms of these parameters, the Mann-Whitney U test was again used. Statistical analyses were performed using SPSS, version 10.0. P values less than or equal to 0.05 were considered significant. RESULTS Effects of blood injection when combined with forced loading. Chondrocyte activity. The rate of proteoglycan synthesis, expressed per milligram wet weight of JOINT BLEEDINGS AND LOADING IN HEMOPHILIC ARTHROPATHY 2027 Figure 1. Proteoglycan turnover, measured 10 weeks after the last intraarticular injection of blood into the right knee of the animals. Values are the mean and SEM from 8 animals, for the blood-injected joints (solid bars) and the contralateral control joints (open bars). Asterisks indicate a statistically significant difference versus control joints. A, Proteoglycan synthesis rate as measured by the rate of sulfate incorporation. Proteoglycan synthesis per mg wet weight of tissue was significantly increased in the experimental joints (39 ⫾ 8%; P ⬍ 0.01). The proteoglycan synthesis rate per cell was equally increased (30 ⫾ 10%; P ⬍ 0.01) (data not shown). B, Release of newly formed proteoglycans, as a measure of the retention of these newly formed proteoglycans, during 3 days. An increased release of newly formed proteoglycans was observed in the experimental joints (30 ⫾ 14%; P ⬍ 0.03). C, Total release of glycosaminoglycan (GAG) during 3 days of ex vivo culture. A 16 ⫾ 6% increase in the release of GAG was found in the experimental joints (P ⬍ 0.03). When normalized to the GAG content, a similar increase was observed (14 ⫾ 4%; P ⬍ 0.02) (data not shown). These results show that after blood exposure, GAGs (resident and newly formed) are lost from the cartilage matrix. tissue, was significantly enhanced in the experimental joints (blood-injected with forced loading) compared with the contralateral control joints. The mean ⫾ SEM increase was 39 ⫾ 8% (P ⬍ 0.01) (Figure 1A). This increase in the synthesis rate was not due to a significant change in chondrocyte number, since the rate of proteoglycan synthesis per DNA content (as a measure of chondrocyte number) was also significantly higher in the experimental joints (30 ⫾ 10%; P ⬍ 0.01). The difference in DNA content between control and experimental joints was only 7 ⫾ 4% (P not significant). The release of newly formed proteoglycans, as a measure of retention of these newly formed proteoglycans, was significantly increased in the blood-injected joints with forced loading compared with the contralateral control joints (30 ⫾ 14%; P ⬍ 0.03) (Figure 1B). Even when normalized to the proteoglycan synthesis rate, a 5% increase in release of newly formed proteoglycans was observed. Release of the total amount of proteoglycans (resident and newly formed [measured as Alcian blue–precipitated GAG]) 2028 HOOIVELD ET AL Figure 2. Changes in cartilage matrix integrity after exposure to blood. Values are the mean and SEM from 8 animals, for the blood-injected joints (solid bars) and the contralateral control joints (open bars). Asterisks indicate a statistically significant difference versus control joints. A, Total content of glycosaminoglycan (GAG) in cartilage, as measured by Alcian blue precipitation of digested cartilage samples. There was a small, statistically nonsignificant, decrease in proteoglycan content in the experimental joints. B, Cartilage integrity based on histologic findings. Cartilage samples were stained with Safranin O, and cartilage degeneration was scored according to a modification of the Mankin score. There was an increase in the histologic score in cartilage from the experimental joints (0.9 ⫾ 0.2 units; P ⬍ 0.01). was significantly increased in the experimental joints (16 ⫾ 6%; P ⬍ 0.03) (Figure 1C). When total proteoglycan release was normalized to proteoglycan content (percentage release), a similar increased loss of these matrix molecules was found (14 ⫾ 4%; P ⬍ 0.02). Cartilage integrity. As a measure of cartilage integrity, proteoglycan content was determined. There appeared to be a small decrease (mean ⫾ SEM 2 ⫾ 3%) in the amount of proteoglycans in the blood-injected joints with forced loading (P not significant) (Figure 2A). However, when cartilage integrity was evaluated histologically, a statistically significant increase in the modified Mankin score in the blood-injected joints with forced loading, compared with the contralateral control joints, was found (change of 0.9 ⫾ 0.2 units; P ⬍ 0.01) (Figure 2B). Changes consisted mainly of slight (1 point) to moderate (2 points) losses of Safranin O staining, and surface irregularities (1 point). Although the mean difference of 0.9 units might seem small, it should be kept in mind that the average score of cartilage with established osteoarthritis has been reported to be ⬃5 (4,13). Synovial inflammation. Synovial inflammation was assessed 10 weeks after the last experimental hemorrhage, a period with intermittent forced loading of the blood-exposed joint. The score was slightly but statistically significantly increased in experimental versus contralateral joints (mean ⫾ SEM 1.5 ⫾ 0.3 versus 0.6 ⫾ 0.1; P ⬍ 0.01). Validity of use of the contralateral joint as a control during forced loading. It could be questioned whether the contralateral control joints of the animals with intermittent forced loading were proper controls, although forced loading was applied for only 7% of the total experiment time. Fixing the contralateral joint to the trunk could induce changes in cartilage that would make these joints inappropriate to use as controls. For this reason we retrospectively compared the findings in contralateral control joints of the animals used in the present study with the values from animals that underwent an identical protocol but without the intensified loading (7). These data are presented in Table 1. There were no significant differences observed for any of the chondrocyte activity, cartilage integrity, or synovial in- Table 1. Parameters measured in the contralateral control joints of the animals with experimental hemorrhages and forced loading and of the animals with experimental hemorrhages but no forced loading* Chondrocyte activity Proteoglycan synthesis, nmoles sulfate/hour/gm wet weight Proteoglycan synthesis, nmoles sulfate/hour/mg DNA Release of newly formed proteoglycans, nmoles/gm in 3 days Release of newly formed proteoglycans, % Total release of proteoglycans, mg GAG/gm wet weight Total release of proteoglycans, % Cartilage integrity Proteoglycan content, mg GAG/gm wet weight Histologic damage, modified Mankin score Synovial inflammation, modified Goldenberg/Cohen score Group with joint loading (n ⫽ 8) Group without joint loading (n ⫽ 6) 6.9 ⫾ 0.3 45.7 ⫾ 2.0 0.8 ⫾ 0.1 26.3 ⫾ 1.0 3.0 ⫾ 0.2 10.7 ⫾ 0.5 6.6 ⫾ 1.1 36.7 ⫾ 6.5 NA NA 3.0 ⫾ 0.3 9.9 ⫾ 0.9 26.1 ⫾ 1.0 0.9 ⫾ 0.1 0.6 ⫾ 0.1 26.6 ⫾ 0.7 1.0 ⫾ 0.2 0.6 ⫾ 0.2 * The group without joint loading consisted of animals from a previous study (7). Values are the mean ⫾ SEM. There were no statistically significant differences between the 2 groups for any of the parameters investigated in the contralateral control joints. NA ⫽ not available; GAG ⫽ glycosaminoglycan. JOINT BLEEDINGS AND LOADING IN HEMOPHILIC ARTHROPATHY Figure 3. Effect of loading of the experimental joint on cartilage matrix turnover. Values are the mean and SEM percentage change in experimental versus contralateral control joints in the animals without intensified joint loading (n ⫽ 6) (open bars) and the animals with intensified joint loading (n ⫽ 8) (hatched bars). P values for differences between the groups with and without loading are indicated; asterisks indicate statistically significant change within each group, compared with the contralateral joints. A, Changes in proteoglycan synthesis. Both groups showed a significant increase in proteoglycan synthesis in the experimental knee compared with that in the contralateral knee, suggesting repair activity. The difference between the group with and the group without loading was not significant. B, Changes in total release of glycosaminoglycan (GAG). In experimental joints that had not been loaded, release of GAG was decreased compared with the contralateral joints, but the difference was not significant. However, if loading was applied, an increase in release of GAG, and consequently a loss of cartilage matrix, was observed (P ⬍ 0.05). flammation parameters measured. Thus, fixation of the contralateral control joint does not alter cartilage matrix turnover in this joint. This indicates that the use of the contralateral joint as a control as was done in the present study, even in combination with the forced loading protocol, is appropriate. Effect of forced loading versus no forced loading after blood exposure. In a group of animals on which we have reported previously (7), beagle dogs were studied using a protocol identical to the one in the present study, but without the forced loading. To evaluate the effects of experimental bleeding with versus without forced loading, the results of that study were compared with the present findings. Chondrocyte activity. Figure 3A shows the percentage difference in proteoglycan synthesis compared with the contralateral joints, in animals that underwent repeated hemorrhages with joint loading and animals that underwent these hemorrhages without loading. No significant difference between the 2 groups was observed. In both groups, repair activity was observed. When normalized to DNA content, similar results were found; DNA content in both groups did not change significantly as a result of the intraarticular blood injec- 2029 tions (mean ⫾ SEM 7 ⫾ 11%; P not significant). Unfortunately, no data on retention were available for the group that did not undergo joint loading. The change in release of proteoglycans, however, was significantly different in the animals that did not undergo joint loading compared with the forced loading group. Release was decreased by a mean ⫾ SEM of 12 ⫾ 6% in the group without loading and increased by 16 ⫾ 6% in the group that underwent loading (P ⬍ 0.02) (Figure 3B). The same differences between groups were observed when the release was normalized to proteoglycan content (⫺11 ⫾ 6% and ⫹14 ⫾ 4% in the group without and the group with loading, respectively; P ⬍ 0.01). Cartilage integrity. The proteoglycan content of the cartilage was increased in the previous study in which the animals did not undergo forced loading of the experimental joints (7). In the present study with forced loading, a decrease in proteoglycan content was observed, although the difference from the value in the group that did not undergo loading was not statistically significant (Figure 4A). The findings in the different groups were paralleled by the histologic results. In the study without forced loading of the experimental joints, no significant differences between the control and the Figure 4. Effect of loading of the experimental joint on cartilage matrix integrity. Values are the mean and SEM percentage changes in experimental versus control joints in the animals without intensified joint loading (n ⫽ 6) (open bars) and the animals with intensified joint loading (n ⫽ 8) (solid bars). P values for differences between the groups with and without loading are indicated; asterisks indicate statistically significant change within each group, compared with the contralateral joints. A, Changes in total proteoglycan content. There was a nonsignificant increase in proteoglycan content in the group without loading, and a nonsignificant decrease in the group with loading. The difference between the group with and the group without loading was not significant. B, Changes in cartilage histologic findings (graded according to a modification of the Mankin score). In the group without loading the change was not significant, whereas in the group with loading there was a significant increase in the histologic score (P ⬍ 0.01). 2030 HOOIVELD ET AL experimental joints were found (mean ⫾ SEM score 1.0 ⫾ 0.2 versus 0.8 ⫾ 0.1, respectively), whereas in the present study a slight increase in histologically determined cartilage damage was evident. The change in score was ⫺0.2 ⫾ 0.3 in the group without loading versus ⫹0.9 ⫾ 0.2 in the group with loading (P ⬍ 0.05) (Figure 4B). Synovial inflammation. Inflammation of synovial tissue as scored histologically did not change significantly in the group that did not undergo loading. In contrast, as described above, a statistically significant increase was observed in the forced loading group. DISCUSSION The present data demonstrate that a limited series of joint bleedings with loading of the affected joint adversely influences cartilage matrix turnover and integrity. In addition, increased (although still very mild) synovial inflammation was found in the experimental joints. The changes in chondrocyte activity are characteristic of early osteoarthritic cartilage (12,16). The enhanced synthesis of proteoglycans is ineffective in protecting against osteoarthritis, because the additionally formed proteoglycans are lost from the matrix, i.e., retention of these newly formed proteoglycans is decreased. Also, the total release of proteoglycans, mainly resident ones, is increased. As a consequence, not unexpectedly, decreased cartilage integrity is observed. Although the effects are small and may not appear dramatic, such changes observed 10 weeks after the last blood injection might be the first step in the development of osteoarthritis. Comparable changes in chondrocyte activity, matrix integrity, and synovial inflammation have been reported to occur 10 weeks after induction of experimental osteoarthritis by anterior cruciate ligament transection in dogs (12,17) or by surgically applied chondral damage accompanied by transient intensified loading of the affected joint (12,16). The observed damage corroborates our previous in vitro and short-term in vivo data showing that blood has direct harmful effects on cartilage (5,6,9). In addition, long-term studies in which animals were injected with blood have shown that cartilage matrix is adversely affected (18,19). Moreover, these data corroborate clinical findings that a limited number of joint hemorrhages in childhood, before the initiation of prophylactic treatment, may result in joint damage more than a decade later (10). Whether the changes observed in the present study will indeed lead over a period of years to fullblown joint damage remains to be investigated. Interestingly, in a recent study with identical conditions except for the loading protocol, no significant changes that would lead to progressive degenerative joint damage were found 10 weeks after the last injection (7). This in contrast to the results of previous studies which have shown that even a less intensive injection protocol (only 2 injections) was associated with adverse changes in cartilage matrix turnover that were observed shortly after the last injection and were sustained for 2 weeks (9). Apparently, such changes can be reversed completely over a successive period of 8 weeks. Because, after several experimental bleedings, dogs can easily walk on 3 limbs (moving the affected fourth limb without loading it), it is possible that the absence of loading might have been protective against the development of joint damage after a series of experimental joint hemorrhages in our recent study. This may have provided the necessary “rest” to the joint to allow normalization of the adversely altered parameters and prevention of progressive degenerative changes. In the present study the animals were deprived of this rest. Although still a matter of debate, it has been suggested that articular cartilage can easily recover from repeated injuries (20). The direct effect of intensified loading of the affected joint in this study was not investigated. A delicate balance determines the equilibrium between growth and breakdown of cartilage. It is known that mechanical signals can lead to cartilage degeneration (21,22). However, in vitro and in vivo experiments have shown that mildly to moderately increased cartilage loading leads to stimulation of cartilage matrix synthesis rather than to damage (21–23). In addition, the amount of time that the animals were exposed to intensified loading in the present study, being less than 7% of the total time, is not expected to lead on its own to adverse changes in cartilage matrix turnover (12). Priming of the cartilage, by diminishing repair capacity as a result of the experimental joint bleedings, appears to be necessary for the development of the progressive joint damage as observed (16). This adverse priming of cartilage by blood has also been demonstrated in the case of anterior cruciate ligament transection, where joint bleeding during surgery has been reported to facilitate the development of experimental osteoarthritis (24). The slight increase in synovial inflammation might be important in relation to the slow but steady progression of joint destruction 10 weeks after joint bleeding. In contrast to the present study with forced loading, in the study without forced loading no inflammation was observed 10 weeks after injection, and no joint damage was observed (7). Although the initial JOINT BLEEDINGS AND LOADING IN HEMOPHILIC ARTHROPATHY cartilage degeneration is caused directly by blood, synovitis is initiated and might be essential in establishing the progressive character of joint destruction. In contrast, it could be possible that inflammation proceeds because of progressive cartilage damage. The data from the present study could have clinical implications. From general practice in treatment of hemophilia, it is known that a limited number of joint bleedings has much more impact in childhood than during adulthood. It is possible that due to differences in pain perception between children and adults, loading of the affected joint after a joint bleed may occur earlier and with more intensity in children than in adults. The fact that loading has an important part in blood-induced joint damage, as described herein, might be explanatory in this respect. Moreover, the notion that joint loading plays a role in the process of hemophilic arthropathy is corroborated by the centrifugal, progressive character of cartilage damage in these patients (4,25). The peripheral, generally much less loaded, cartilage usually appears normal in patients with hemophilia, whereas the central cartilage in weight-bearing areas becomes diseased. These findings suggest that avoiding or minimizing joint loading (use, but no weight bearing) might be helpful in protecting against hemophilic arthropathy. REFERENCES 1. Madhok R, York J, Sturrock RD. Haemophilic arthritis. Ann Rheum Dis 1991;50:588–91. 2. Rodriguez-Merchan EC. Methods to treat chronic haemophilic synovitis. Haemophilia 2001;7:1–5. 3. Roosendaal G, Mauser-Bunschoten EP, de Kleijn P, Heijnen L, van den Berg HM, van Rinsum AC, et al. Synovium in haemophilic arthropathy. Haemophilia 1998;4:502–5. 4. Roosendaal G, van Rinsum AC, Vianen ME, van den Berg HM, Lafeber FP, Bijlsma JW. Haemophilic arthropathy resembles degenerative rather than inflammatory joint disease. Histopathology 1999;34:144–53. 5. Roosendaal G, Vianen ME, van den Berg HM, Lafeber FP, Bijlsma JW. Cartilage damage as a result of hemarthrosis in a human in vitro model. J Rheumatol 1997;24:1350–4. 6. Roosendaal G, Vianen ME, Marx JJ, van den Berg HM, Lafeber FP, Bijlsma JW. Blood-induced joint damage: a human in vitro study. Arthritis Rheum 1999;42:1025–32. 7. Hooiveld M, Roosendaal G, Vianen M, van den Berg M, Bijlsma J, Lafeber F. Blood-induced joint damage: longterm effects in vitro and in vivo. J Rheumatol 2003;30:339–44. 8. Hooiveld M, Roosendaal G, Wenting M, van den Berg M, Bijlsma 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 2031 J, Lafeber F. Short-term exposure of cartilage to blood induces chondrocyte apoptosis. Am J Pathol 2003;162:943–51. Roosendaal G, TeKoppele JM, Vianen ME, van den Berg HM, Lafeber FP, Bijlsma JW. Blood-induced joint damage: a canine in vivo study. Arthritis Rheum 1999;42:1033–9. Fischer K, Der Bom JG, Mauser-Bunschoten EP, Roosendaal G, Prejs R, de Kleijn P, et al. The effects of postponing prophylactic treatment on long-term outcome in patients with severe hemophilia. Blood 2002;99:2337–41. DeGroot J, Verzijl N, Wenting-van Wijk MJ, Jacobs KM, van El B, van Roermund PM, et al. Accumulation of advanced glycation end products as a molecular mechanism for aging as a risk factor in osteoarthritis. Arthritis Rheum 2004;50:1207–15. Marijnissen AC, van Roermund PM, TeKoppele JM, Bijlsma JW, Lafeber FP. The canine groove model, compared with the ACLT model of osteoarthritis. Osteoarthritis Cartilage 2002;10:145–55. Mankin HJ, Dorfman H, Lippiello L, Zarins A. Biochemical and metabolic abnormalities in articular cartilage from osteo-arthritic human hips. II. Correlation of morphology with biochemical and metabolic data. J Bone Joint Surg Am 1971;53:523–73. Lafeber FP, van der Kraan PM, van Roy JL, Huber-Bruning O, Bijlsma JW. Articular cartilage explant culture: an appropriate in vitro system to compare osteoarthritic and normal human cartilage. Connect Tissue Res 1993;29:287–99. Goldenberg DL, Cohen AS. Synovial membrane histopathology in the differential diagnosis of rheumatoid arthritis, gout, pseudogout, systemic lupus erythematosus, infectious arthritis and degenerative joint disease. Medicine (Baltimore) 1978;57:239–52. Marijnissen AC, van Roermund PM, Verzijl N, TeKoppele JM, Bijlsma JW, Lafeber FP. Steady progression of osteoarthritic features in the canine groove model. Osteoarthritis Cartilage 2002;10:282–9. Brandt KD, Braunstein EM, Visco DM, O’Connor B, Heck D, Albrecht M. Anterior (cranial) cruciate ligament transection in the dog: a bona fide model of osteoarthritis, not merely of cartilage injury and repair. J Rheumatol 1991;18:436–46. Roy S, Ghadially FN. Pathology of experimental haemarthrosis. Ann Rheum Dis 1966;25:402–15. Hoaglund FT. Experimental hemarthrosis: the response of canine knees to injections of autologous blood. J Bone Joint Surg Am 1967;49:285–98. Martin JA, Buckwalter JA. Human chondrocyte senescence and osteoarthritis. Biorheology 2002;39:145–52. Buckwalter JA, Mankin HJ. Articular cartilage: degeneration and osteoarthritis, repair, regeneration, and transplantation. Instr Course Lect 1998;47:487–504. Sah RL, Kim YJ, Doong JY, Grodzinsky AJ, Plaas AH, Sandy JD. Biosynthetic response of cartilage explants to dynamic compression. J Orthop Res 1989;7:619–36. Van Valburg AA, van Roy HL, Lafeber FP, Bijlsma JW. Beneficial effects of intermittent fluid pressure of low physiological magnitude on cartilage and inflammation in osteoarthritis: an in vitro study. J Rheumatol 1998;25:515–20. Myers SL, Brandt KD, O’Connor BL, Visco DM, Albrecht ME. Synovitis and osteoarthritic changes in canine articular cartilage after anterior cruciate ligament transection: effect of surgical hemostasis. Arthritis Rheum 1990;33:1406–15. Rippey JJ, Hill RR, Lurie A, Sweet MB, Thonar EJ, Handelsman JE. Articular cartilage degradation and the pathology of haemophilic arthropathy. S Afr Med J 1978;54:345–51.