Rheumatoid arthritis (RA) is a chronic inflammatory joint disease in which both humoral (1, 2) and cellular (3) immune mechanisms are considered to be responsible for the induction of the disease. There is now a growing awareness that RA is a heterogeneous disease composed of several distinct disease pathways (4). This heterogeneity is also reflected in the large number of different animal models of RA that have been developed.
Antigen-induced arthritis (AIA) (5–7) and collagen-induced arthritis (CIA) (8–10) are 2 commonly used murine models of RA. AIA is induced by intradermal immunization and subsequent intraarticular injection of methylated bovine serum albumin (mBSA) (5–7). The disease is chronic and antigen-specific, and has been suggested to be T cell dependent (5). CIA is induced by intradermal immunizations with the major cartilage protein component type II collagen (CII) (8, 11). CIA is dependent both on B and T cells as well as the complement system (12, 13).
The plasminogen activator (PA) system is a general proteolytic enzyme system that has been suggested to play an important role in the degradation of the extracellular matrix (ECM) during the development of RA (14, 15). Plasmin, the key component of the PA system, is generated by conversion of the precursor plasminogen by 2 distinct physiologic PAs, tissue PA (tPA) and urokinase PA (uPA) (16). The main function of the PA system is fibrinolysis, which is important for the maintenance of the hemostatic balance (16). However, by acting in concert with other proteinases, the PA system has also been proposed to play a role in the remodeling and degradation of the ECM during many physiologic and pathologic processes, including ovulation (17), skin wound healing (18), tumor invasion (19, 20), and inflammatory cell infiltration, fibrin deposition, and joint destruction associated with RA (21). However, despite the rapid progress in our understanding of the involvement of the PA system in RA, the molecular mechanisms by which the PA system participates in the pathogenesis of RA are still poorly understood.
In a previous study using the AIA model, it was shown that plasminogen-deficient mice develop much more severe arthritis than do wild-type control mice (22). In contrast, our recent studies using the CIA model have shown that plasminogen-deficient mice are completely resistant to CIA, and we propose that plasmin may play its essential roles by activation of the complement system (23). The aim of the present study was to gain a deeper understanding of the molecular mechanism by which plasmin affects the pathogenesis of RA in the 2 different RA models. To do this, we developed a new experimental animal model of RA, which we called local injection–induced arthritis (LIA). In this model, the induction was administered the same way as with AIA, but instead of mBSA (used in AIA as the antigen), we substituted CII. The differences between LIA and AIA are the antigen (CII is used in LIA and mBSA is used in AIA) and the administration (local trauma is induced in the knee joint in LIA, whereas there is no local trauma in CIA). The local trauma also initiates a wound healing–like process in which the innate immune response against trauma is involved. By using these different models, the different functional roles of plasminogen in AIA and CIA can be interpreted and the mechanistic differences between the models can be elucidated more precisely.
- Top of page
- MATERIALS AND METHODS
Exacerbation of the severity of AIA by plasminogen deficiency. Previous studies on the role of plasminogen in arthritis have shown that in an AIA model, plasminogen-deficient mice develop a much more severe arthritis than do wild-type control mice (22). In contrast, our recent study showed that in a CIA model, plasminogen-deficient mice were unable to develop arthritis (23). To determine whether the disease profile of AIA varies due to strain differences, we first induced AIA in our mouse strain. Wild-type and plasminogen-deficient mice with a DBA/1 background were immunized with mBSA on days 0 and 7, and AIA was induced by intraarticular injection of mBSA on day 21. As shown in Figure 1, comparable levels of arthritis were observed in wild-type and plasminogen-deficient mice on both day 3 and day 10. Thereafter, in wild-type mice, the disease subsided, with eventual recovery on day 60. Plasminogen-deficient mice, however, had significantly more severe arthritis as compared with wild-type controls on day 30 and onward. In both wild-type and plasminogen-deficient mice, cartilage degradation, pannus formation, and bone erosion were observed on days 3 and 10. However, on days 30 and 60, severe cartilage degradation, pannus formation, and bone erosion were found only in plasminogen-deficient mice (data not shown). The anti-mBSA IgG levels in serum were comparable at all time points after AIA induction in wild-type and plasminogen-deficient mice (data not shown). In both wild-type and plasminogen-deficient mice, no development of arthritis was observed in the control knee joints that were injected with PBS. These results were consistent with those of previous studies (22) and indicate that plasmin plays a protective role in the pathogenesis of AIA.
Figure 1. Development of arthritis. The severity of arthritis in the knee joints was scored histologically at different time points after the induction of antigen-induced arthritis (AIA). Plasminogen-deficient (Plg−/−) mice were compared with wild-type (Plg+/+) control mice. Results are expressed as the mean and SEM. ∗ = P < 0.05 versus control mice, by Mann-Whitney U test.
Download figure to PowerPoint
Induction of arthritis by local knee joint trauma in CII-immunized plasminogen-deficient mice. To study the different functional roles of plasmin during AIA and CIA and to further investigate the different pathogenic mechanisms in these models, we created a new experimental animal model of RA, which we called LIA. In this model, immunization and local injection were performed as for AIA, but the mBSA that was used in AIA as the antigen was replaced by CII. Using this model, arthritis development 3, 10, 30, and 60 days after local treatment in knee joints was investigated in wild-type and plasminogen-deficient mice that had previously been immunized with CII. When mice were only immunized with CII without local treatment in the knee joints, arthritis developed only in wild-type controls and not in plasminogen-deficient mice (Figure 2A). This result confirmed our previous findings that plasminogen-deficient mice are resistant to CIA (23). However, both wild-type and plasminogen-deficient mice that were immunized with CII and locally injected with 0.9% NaCl in the knee joints developed arthritis (Figure 2B). At all observation time points though, the severity of arthritis was significantly greater in wild-type mice than in plasminogen-deficient mice. In both genotypes, the severity of arthritis seemed to decline with time.
Figure 2. Time-course effects of local treatments on arthritis development in mice. The severity of arthritis was scored histologically, as described in Materials and Methods. Wild-type (Plg+/+) and plasminogen-deficient (Plg−/−) mice that received A, no local treatment, B, local treatment with NaCl, and C, injections with type II collagen (CII) were compared during the experimental period. Results are expressed as the mean and SEM. ∗ = P < 0.05; ∗∗ = P < 0.01 versus control mice, by Mann-Whitney U test.
Download figure to PowerPoint
When mice were immunized with CII and subsequently injected with CII in the knee joints, wild-type mice developed significantly more severe arthritis than did plasminogen-deficient mice on days 3 and 10 (Figure 2C). On days 30 and 60, the severity of arthritis became comparable between the 2 genotypes. Over the entire experimental period, the severity of arthritis appeared to decline in wild-type mice, whereas it appeared to be more constant in the plasminogen-deficient mice. Taken together, these results indicate that plasminogen-deficient mice are susceptible to arthritis after local injections of 0.9% NaCl or CII, which demonstrates that local trauma in the knee joints is critical for the induction of arthritis in plasminogen-deficient mice in this model.
The results obtained in the experiment shown in Figure 2 were further analyzed with regard to the influence of different treatments at each time point (Figure 3). On days 3 and 10 in wild-type mice, local injection of CII significantly increased the severity of arthritis as compared with the wild-type group that did not receive local treatments (P < 0.01) (Figures 3A and B). However, on days 30 and 60, this difference became insignificant (Figures 3C and D). In wild-type mice, there was no significant difference at any time point in the severity of arthritis between mice that received local injections of CII and those that received 0.9% NaCl (Figure 3). Compared with the group that received no local treatment, local injection of 0.9% NaCl did not seem to significantly increase the severity of the arthritis in wild-type mice during the entire experimental period either, except on day 10 (Figure 3). However, in plasminogen-deficient mice, the severity of the arthritis in the group that was locally injected with CII was significantly greater than that in the group that was locally injected with 0.9% NaCl during the entire experimental period, except at the beginning (day 3) (Figure 3). Taken together, these results indicate that the difference in the components of the local injection played a more important role in plasminogen-deficient mice than in wild-type mice, especially at the later stages of arthritis.
Figure 3. Effects of treatments on arthritis development in mice at each time point. The severity of arthritis was scored histologically, as described in Materials and Methods. Wild-type (Plg+/+) and plasminogen-deficient (Plg−/−) mice that received different local treatments were compared on A, day 3, B, day 10, C, day 30, and D, day 60 after treatment. Results are expressed as the mean and SEM. ∗ = P < 0.05; ∗∗ = P < 0.01, by Mann-Whitney U test. CII = type II collagen.
Download figure to PowerPoint
Lack of arthritis development in nontraumatized joints of mice during LIA. During LIA and AIA, local injections were performed only at the knee joints. Thus, it was interesting to further investigate the development of arthritis in paw joints, which were not traumatized during the induction of arthritis in these models. On day 60 after the induction of CIA, LIA (locally injected with 0.9% NaCl, CII, or with no local treatment), and AIA, paw joints were removed from wild-type and plasminogen-deficient mice and histologic analysis was performed. Consistent with our previous findings, arthritis developed only in the paw joints of wild-type mice, and not in plasminogen-deficient mice, after CIA induction (Figures 4A and B). During the development of LIA, severe arthritis was observed in >80% of paws of wild-type mice that were injected with 0.9% NaCl or CII in the knee joints, or even in the paws of wild-type mice that received no local treatment (Figure 4C, and data not shown). In the corresponding groups in plasminogen-deficient mice, however, there was no arthritis development in the paws examined (Figure 4D, and data not shown). After the induction of AIA, no arthritis development was observed in nontraumatized paw joints of both wild-type and plasminogen-deficient mouse groups (Figures 4E and F), indicating that AIA is a monarthritis model restricted to locally injected knee joints.
Figure 4. Histologic analysis of paw joints. The nontraumatized paw joints of wild-type (Plg+/+) and plasminogen-deficient (Plg−/−) mice were assessed for arthritis development 60 days after the induction of collagen-induced arthritis (CIA) (A and B), local injection–induced arthritis (LIA) (with local type II collagen treatment) (C and D), and antigen-induced arthritis (AIA) (E and F). Note that arthritis developed in the paw joints of Plg+/+ mice, but not Plg−/− mice, after the induction of CIA and LIA. (Original magnification × 100.)
Download figure to PowerPoint
Accumulation of necrotic tissue in plasminogen-deficient mice during AIA and LIA. Trauma-induced intravascular thrombosis may lead to ischemia and cause necrosis. Previous studies have suggested that the PA system plays a role during the tissue necrosis process (28–30). We also consistently observed synovial tissue necrosis and apoptosis in plasminogen-deficient mice during AIA and LIA. Tissue sections were analyzed both by histology and TUNEL staining, and the data are summarized in Table 1. During AIA, necrosis was observed in 100% of the plasminogen-deficient mice but in none of the wild-type mice. During LIA, 54% of the knee joints of plasminogen-deficient mice locally injected with CII developed necrosis, but only 25% of the knee joints of plasminogen-deficient mice treated with 0.9% NaCl developed necrosis. No necrosis was observed in plasminogen-deficient mice that did not receive a local injection or in any groups of the wild-type mice. These data revealed that plasmin seems to be an important factor in cleaning necrotic tissue in AIA and LIA, and further suggest that the disease profiles involved in wild-type and plasminogen-deficient mice are different in these models.
Table 1. Incidence of synovial tissue necrosis in wild-type and plasminogen-deficient mice during AIA and LIA*
|Arthritis model||No. of samples||Incidence of necrosis|
|AIA|| || |
|LIA|| || |
| Plg+/+ (CII)||41||0|
| Plg+/+ (NaCl)||40||0|
| Plg+/+ (no treatment)||40||0|
| Plg−/− (CII)||43||54|
| Plg−/− (NaCl)||40||25|
| Plg−/− (no treatment)||40||0|
- Top of page
- MATERIALS AND METHODS
Plasmin seems to play contrasting roles in the pathogenesis of arthritis in different murine arthritis models. In AIA, plasmin plays a protective role, possibly due to its fibrinolytic capacity (22). However, plasmin seems to be critically required for induction of the disease in the CIA model (23). It is possible that the apparent discrepancy of the plasminogen-deficient phenotype during CIA and AIA could be explained by differences in the experimental setup or the mouse strains used. However, it is more likely that the apparent discrepancy may reflect the different pathogenic mechanisms of these 2 arthritis models.
To study the pathogenic mechanisms of arthritis and the cause of this discrepancy, we developed the LIA model, where CII or 0.9% NaCl is injected intraarticularly into the knee joints of previously CII-immunized mice. Our data show that plasminogen-deficient mice that are normally resistant to CIA can develop arthritis following an intraarticular injection of CII or even 0.9% NaCl. The trauma that is induced by this treatment causes a wound healing–like process that involves an innate immune response, whereby the requirement for plasmin can be passed. Taken together, these data suggest that plasmin may be essential for disease induction in arthritis models that are critically dependent on complement activation. In addition to previously proposed protective and destructive roles that plasmin may play in fibrin degradation and tissue destruction, our data clearly reveal that plasmin may also play a pivotal role in the induction of the disease.
Susceptibility to CIA is associated with class II MHC (25). In this model, mice are immunized with rat CII, which is taken up and presented by antigen-presenting cells and recognized by lymphocytes to generate CII-specific CD4+ T cells and B cells. These lymphocytes further induce an autoimmune response against CII that leads to the development of arthritis. It has been shown that CIA is critically dependent on B cells as well as the complement system, whereas T cells do not seem to play an essential role (12, 13, 31). CIA is a permanent and relapsing disease involving inflammation, tissue destruction, and tissue repair of the affected joints.
In AIA, the mice are first immunized with the antigen mBSA, followed by a local injection of mBSA into the knee joint that starts an adaptive immune response against mBSA (5). The positively charged mBSA is bound to negatively charged cartilage, which initiates an immune attack against the surface-bound mBSA as well as the cartilage. It is likely that local immune complexes are formed, but it has also been shown that mBSA-specific T cells recognize the antigen in the joints and induce the arthritis (32, 33). The inducing agent mBSA is not a self protein, and the arthritis induced is a local reaction that is dependent on the intraarticular injection of mBSA. The susceptibility of AIA is not associated with class II MHC, which is evidence against a role of T cells that cross-react specifically with a cartilage autoantigen. AIA is not a self-perpetuating process. When the mBSA immunogen has been used up, the inflammation declines and the arthritis recedes.
Both models have several features in common with human RA, including intraarticular deposition of fibrin, infiltration of neutrophils and macrophages, formation of an inflammatory pannus, and destruction of articular cartilage and bone. Although both AIA and CIA are experimental models that resemble RA, the pathogenic mechanisms of AIA and CIA differ in several respects, including antigen challenge, disease induction, and clinical features (34, 35). Whereas AIA is a monarthritis model restricted to the locally injected knee joints, CIA is a polyarthritis model in which the arthritis can develop in all of the peripheral joints. Disease severity in CIA can be evaluated using a macroscopic scoring system, whereas the severity of AIA is evaluated histologically.
The LIA model described here is very similar to the CIA model, except that in LIA, a local injection that induces intraarticular trauma is included. Both CIA and LIA are self-perpetuating arthritis models triggered by autoimmune responses in which the arthritis inflammation is dependent not only on an autoimmune response consisting of arthritogenic antibodies, but also on other inflammatory components, such as T cells, macrophages, and fibroblasts. In both AIA and LIA, a trauma is generated at the knee joint after the local injection of antigen. This trauma, however, will also initiate an innate immune response against the trauma caused by the injection. The pathogenesis of AIA and LIA thus appears to involve the combined effects of both an adaptive immune response against the antigens as well as a trauma-induced innate immune response. Our finding that a local injection of 0.9% NaCl is sufficient to induce arthritis in plasminogen-deficient mice in the LIA model indicates that a trauma-induced innate immune response totally changes the disease profile of arthritis from being resistant, as in CIA, to being sensitive, as in LIA.
Our previous study suggests that plasmin plays an essential role in CIA induction in a step between the binding of arthritogenic antibodies on joint cartilage and the infiltration of inflammatory cells, possibly by activating the complement system (23). Recent studies in mice deficient in members of the complement system have shown that both alternative and classical pathways of complement activation are important in the development of arthritis in CIA as well as in CII antibody–induced arthritis (36–38). Provided that plasmin plays an essential role in complement activation, these findings are consistent with our results in plasminogen-deficient mice in CIA. LIA is similar to CIA except that a trauma is induced at the knee joints. It has also been shown that the complement system is not a primary mediator of inflammation following a wound or trauma, whereby neutrophils can be activated in the absence of active complement (39). This may be the underlying reason why, in LIA, a local injection of 0.9% NaCl is sufficient to induce mild arthritis in the injected knee joints of plasminogen-deficient mice. However, the exact mechanism by which plasmin is functionally correlated with the complement system remains unclear in vivo. Such investigations are being performed in our laboratory.
Interestingly, local injections of 0.9% NaCl or CII induced less severe arthritis in plasminogen-deficient mice than in the wild-type control group, suggesting that plasmin may be involved in both the induction and the development of the disease. AIA is critically dependent on T cells, and the T cell–mediated inflammation cascade can directly recruit and activate macrophages. Thus, the B cell and complement activation cascade may not be essentially involved in the pathogenesis of AIA. The severe arthritis and the presence of necrosis observed in plasminogen-deficient mice with AIA strongly indicate that plasmin is not important in the induction of AIA but that plasmin may play an important role in fibrinolysis and tissue remodeling in this disease.
Several mechanisms, including a beneficial role in intraarticular fibrin degradation and deleterious effects via degradation of joint matrix proteins, have been proposed by which plasmin may play an important role during arthritis. Studies of mice lacking uPA, plasminogen, or plasminogen activator inhibitor 1 (PAI-1) in the development of the AIA model indicated significantly increased levels of fibrin in the joints of uPA- or plasminogen-deficient mice, whereas lower levels of fibrin were detected in PAI-1–deficient mice (22, 40). The levels of fibrin deposition seemed to correspond to the severity of AIA. It has also been shown that fibrin deposition occurs in the first few days after CIA induction (41).
However, our data showing that plasminogen-deficient mice are resistant to CIA (23), develop a less severe disease in LIA, and a more severe disease in AIA as compared with wild-type controls suggest that plasminogen deficiency plays a profound role in the pathogenesis of arthritis. In CIA, fibrin deposition is predominant during the first few days after inflammation starts (41). Thereafter, fibrin is eliminated from the joints, and tissue degradation starts and predominates during the remaining stages of disease (41). However, injection of mBSA into knee joints creates a trauma. Persistent deposition of fibrin has been observed in traumas in plasminogen-deficient mice (18). In this respect, fibrin deposition in plasminogen-deficient mice in AIA may be a consequence of the inflammation associated with trauma rather than being the reason for the inflammation of arthritis, or both. Thus, the accumulated fibrin deposition in plasminogen-deficient mice in the AIA model may not be a main pathogenic factor for the arthritis phenotype of these mice.
The results of this and other studies reflect the complex nature of RA and the contrasting roles that plasmin seems to play in the pathogenesis of this disease. In addition to the previously proposed beneficial roles that plasmin may play by degrading intraarticular fibrin and the deleterious effects it has in joint matrix degradation, based on our data, this protease also seems to play a prominent role in the induction of certain autoimmune inflammatory cascades, most likely in the complement activation step. If this is also the case in humans, plasmin may be an alternative therapeutic target for interfering with an early stage of the disease.