Matrix metalloproteinases involvement in rheumatoid arthritis
Tchetverikov, I.Citation
Tchetverikov, I. (2005, February 17). Matrix metalloproteinases involvement in rheumatoid arthritis. Retrieved from https://hdl.handle.net/1887/625
Version: Corrected Publisher’s Version
CHAPTER
In this chapter, findings of this thesis are summarized and discussed in light of the current understanding of the pathogenesis of joint diseases; future prospective of the research is outlined. Chapter 1 is the general introduction to this thesis. Chapters 2 and 3 show evidence of Į2Macroglobulin/Matrix Metalloproteases (Į2M/MMP) complex formation in
the synovial fluid (SF) and in the systemic circulation and describe the characteristics of the assay used to determine MMP activity in Į2M/MMP complexes. Chapter 4 provides
MMP profiles as found in SF and in the circulation of rheumatoid arthritis (RA) and osteoarthritis (OA) patients in comparison to the levels found in control population. Based on these findings candidate markers for disease activity and joint tissue degradation were determined. In Chapter 5 the results of the study investigating the value of the systemic MMP levels as predictors of joint damage progression are presented. Chapter 6 shows the results of the study designed to investigate the imbalance between MMPs and tissue inhibitors of matrix metalloproteases (TIMPs) in joint pathology. The effects of Disease Modifying Drugs (DMARD’s) frequently used for RA treatment on levels of Į2M/MMP
complexes in systemic circulation are described in Chapter 7. Summary and discussion
Chapter 1 is a general introduction reviewing the current knowledge with regard to the pathogenesis of rheumatoid arthritis in general and involvement of proteinases in degradation of articular cartilage in particular. Especially, the involvement of MMPs in the process is discussed.
Rheumatoid arthritis is a chronic systemic disease characterized by systemic inflammation and destruction of joint tissue. Joint deformities are important features of the disease which impairs patients’ ability to work and to fully participate in events of every day life. Care of RA patients includes treatment aimed at reduction of the inflammation which is often a prominent feature of the disease. The ultimate long term goal in the RA treatment is however to prevent loss of the articular cartilage and as such to prevent permanent damage to the joint. Although more and more is becoming clear about the strategies to achieve this goal, it is still a challenge for the rheumatologist. The unraveling of the disease pathophysiology contributes to the improvement of the used treatments regimens and the development of new therapeutic approaches.
SUMMARY AND DISCUSSION
109 systems are indicated to be involved in the process and, among others, matrix metalloproteinase are likely to play a major role. MMPs are a family of related enzymes which all have a Zn2+ion in the catalytic domain. The family consists of more than 20 enzymes and five main groups can be distinguished, e.g. collagenases, gelatinases, stromelysins, matrilysins and membrane type (MT)-MMPs. It has been shown in vitro that MMPs have a combined ability to degrade almost all components of extra cellular matrix (ECM), which is crucial for the proper function of the joint. Ample body of evidence has been accumulated indicating the involvement of MMPs in joint pathology. The presence of increased amounts of mRNA in RA has been extensively documented. Virtually all known MMPs are shown to be present in the inflamed synovium. Also high protein levels of pro-matrix metalloproteases (proMMP) are found in the joint tissue as well as in the synovial fluid. Moreover, increased levels of proMMPs are found also in the circulation, which could indicate leakage of the locally produced MMPs or production of MMPs in the circulation in response to proinflammatory cytokines. Either way, these findings provide useful information about the current disease activity or may even be indicative of the probable disease course with regard to the joint destruction.
Thus, the goal of the present work was to elaborate on the role of MMP in the process of joint inflammation and joint tissue degradation.
Chapter 2 provides information on complex formation between activated MMPs and a general protease scavenger Į2M in the systemic circulation of RA patients. As such, the
presence of D2M/MMP complexes in serum or synovial fluid indicates activation of the
produced proMMPs since D2M/MMP complex formation requires presence of activated
MMPs. Size exclusion analysis showed similarity between D2M/MMP complex formation
in buffer and in normal plasma spiked with activated MMPs, indicating D2M/MMP
complex formation in the systemic circulation of RA patients. The results also showed that active MMPs were fully inhibited by Low Molecular W eight (LMW ) inhibitor BB94 in the presence of Į2M, whereas no inhibition was achieved by High Molecular W eight (HMW )
inhibitor TIMP-1. Further, MMP activity in D2M/MMP complexes in serum of RA patients
was significantly higher than in serum of healthy controls (P < 0.001). Thus, using HMW and LMW substrates and inhibitors and size separation analysis it is shown that activated, but not TIMP inhibited MMPs form complexes with Į2M. The results indicate that MMP
activity in Į2M/MMP complexes in systemic circulation of RA patients is increased as
compared to OA patients and healthy controls. Moreover, since levels of MMP activity in D2M/MMP complexes in serum of RA patients were correlated with levels of a systemic inflammatory marker, erythrocytes sedimentation rate (ESR, r = 0.72, P < 0.001), MMP activity measurements using LMW fluorogenic substrates became a candidate marker of disease activity.
To describe the characteristics of the assay used to measure MMP activity in the systemic circulation, several experiments were performed and the results are presented in Chapter 3. To measure MMP activity in Į2M/MMP complexes, LMW fluorogenic substrate
with Į2M is proportional to the amount of MMPs present in the sample. To increase the
specificity (also other proteases such as serine and cysteine proteases are likely to convert TNO211-F) a LMW MMP specific inhibitor BB94 was added to the samples of interest. The difference in fluorescence increase between samples with or without BB94 is considered MMP related. The increase in fluorescence was proportional to MMP-9 concentration used in calibration curves. The variation of the method used to measure MMP activity was tested. The intra and inter assay variations were 9 and 12 % respectively and were comparable to the limits for other, commercially available assay kits. Moreover, the results showed that neither gender nor age of the patients influenced the measured MMP activity in serum. Also no circadian variation in MMP activity in Į2M/MMP
complexes in serum of RA patients was found. Taken together, these findings indicate reliability of the newly developed method. In combination with the results described in chapter 2 these results show that a reliable method to measure activated MMPs in complex with D2M was developed. This method provides additional information about the MMP
system since unlike the protein based ELISA’s it measures not the proactive enzymes but the proportion of MMPs which was actually activated before it was inhibited by D2M.
Experiments described in Chapter 4 were designed to provide inventory of MMPs present in serum and synovial fluid of RA and OA patients to allow comparison to MMP levels found in control population. Also levels of collagen degradation hydroxyproline (Hyp) products were determined in SF as well as levels of a marker of inflammation, C-reactive protein (CRP) in the systemic circulation. The results showed that levels of all studied MMPs in SF of both RA and OA groups were higher than in the control population, except for proMMP-13 levels which were below the detection limit. Comparison of RA and OA groups revealed that except for proMMP-2 and TIMP-1 levels, higher levels were found in the RA population. In serum, levels of proMMP-3, -8 and -9 were increased in RA patients as compared to OA or controls, whereas only proMMP-8 and -9 were increased in serum of OA patients as compared to controls. Thus, the results provide a comprehensive analysis of MMPs found in SF and in the circulation of RA patients and allow comparison between normal situation and joint diseases.
The presented results also provide evidence to support the imbalance hypothesis of MMPs and TIMP in arthritic diseases (TIMPs levels are likely to be insufficient to fully inhibit present MMPs). In our study not only the molar MMP/TIMP ratio was high in SF of RA and OA patients, but also MMP activity in D2M/MMP complexes was higher than in
controls indicating that higher levels of activated but not TIMP inhibited MMPs were present in samples.
SUMMARY AND DISCUSSION
111 MMPs are the principal enzymes involved in the process. Although it should be noticed that there was only a trend towards correlation between proMMP-8 and Hyp in SF, levels of other measured MMP which is involved in joint pathology, proMMP-3, were not correlated with Hyp levels whatsoever. When comparison between systemic levels of inflammation (CRP) and MMP levels was performed the opposite situation was observed. Only proMMP-3 levels were correlated with CRP and not proMMP-8 or -9 levels. Combined, these findings indicate that MMP-8 and -9 may reflect the degradation of joint tissue whereas MMP-3 may be indicative of the inflammatory process.
Based on these observations the above mentioned enzymes (in addition to MMP activity in Į2M/MMP complexes measurements) became candidate markers for the
pathophysiological processes of joint inflammation and degradation. The experimental results provided sufficient basis to advance to the following phase of the project and to investigate the predictive role of MMP levels in the circulation for the disease course with regard to joint inflammation and/or joint tissue degradation.
Previously, conflicting reports have been published with regard to the predictive values of the systemic levels of the MMPs for joint degradation. Whereas Posthumus et al. and Yamanaka et al. suggested that serum proMMP-3 levels may be predictive of joint cartilage degradation, Cunnane et al. indicated that proMMP-3 levels merely represent the ongoing inflammation in the joints. Of other MMPs of interest, little is known. Goldbach-Mansky et al. suggested that MMP-2 may be associated with development of the joint erosions. Since in our previous studies no sufficient evidence of MMP-2 as a possible marker of disease activity was found, measurement of proMMP-2 was not included in the study protocol. The value of the serum levels of another gelatinase, MMP-9, as a marker of rheumatic disease was not clear, whereas the results of the study described in chapter 4 were promising. The data on MMP-8 in arthritic diseases was fairly limited. Kullich et al. showed high MMP-8 levels in serum of RA patients, whereas the value of this MMP as prognostic marker was, to our knowledge, not yet investigated.
Thus, the following candidate markers were included in the protocol of the study described in Chapter 5: proMMP-3, -8, -9 and MMP activity in D2M/MMP complexes. One hundred
and eight patients from Early Arthritis Clinic cohort of the Leiden University Medical Centre who were diagnosed with RA of recent-onset were entered in this longitudinal study. The patients were followed for 2 years; clinical data, blood samples and radiographs were obtained at baseline, at 1 year and at 2 years. The results showed that joint damage progressed during the 2-year follow up from 0 to 10 (median, Sharp-van der Heijde joint damage score progression, P < 0.001). Serum levels of proMMP-3 remained stable during the study period. Systemic proMMP-8, -9 and Į2M/MMP levels as well as disease activity
(as shown by modified Disease Activity Score, DAS) significantly decreased during the 2-year study. A stepwise log linear regression analysis was performed to analyze which parameters at baseline were associated with progression of joint damage as determined by progression of the total Sharpvan der Heijde joint damage score. Next to proMMP3, 8, -9 and MMP activity in Į2M/MMP complexes other potential predictors such as rheumatoid
model. The strongest association between joint damage progression and the input factors was found for proMMP-3 levels (B: 0.37, 95% CI: [0.13-0.61], P = 0.003); this association was independent from all other parameters. The analysis using the erosion and joint space narrowing sub-components of the total Sharp-van der Heijde joint damage score as outcome measure yielded similar results: proMMP-3 levels at onset of the disease was the strongest independent predictor. Other studied MMPs (MMP-8 and -9) showed an association with DAS: the levels steadily decreased during the study period in the whole group and in mild and severe joint damage subgroups (the whole group was divided into mild and severe subgroups based on the total Sharp-van der Heijde joint damage score). MMP activity in Į2M/MMP complexes was however decreased only in the subgroup with
mild joint damage progression and not in the severe joint damage progression subgroup. Thus, the results of our study showed that proMMP-3 levels at onset of the disease were predictive of cartilage loss at the end of the second year of the follow up period. Moreover, proMMP-3 predicted joint damage progression independently from other known parameters such as SE or RF. Interestingly, the latest publication by Green and colleagues confirms our findings. In their study a large cohort of RA patients of recent onset was used and the results yielded significant correlation between serum proMMP-3 levels and the progression of cartilage loss as documented by the Larsen score. Thus, it is now shown by two independent studies in large cohorts and in longitudinal setting that proMMP-3 levels at onset of the disease are indeed predictive of severe disease with regard to the loss of the articular cartilage. The study of Green et al. also indicated that proMMP-1 as well as proMMP-3 predicted development of joint erosion during the first 12 month of presentation in patients without increased CRP levels at presentation. In our study no such analysis was performed and no conformation of these findings can be given. Our results, however, showed that another collagenase (MMP-8) was not predictive of joint damage and neither were high levels of proMMP-9 at presentation. It is also shown that levels of MMP-8 and -9 steadily decreased during the follow up period, which was accompanied by a decrease in DAS.
SUMMARY AND DISCUSSION
113 Further studies will be needed to provide evidence for this hypothesis. Moreover, recent studies, e.g. by Constantijn et al., indicate that the polymorphism in MMP-3 gene may be associated with more erosive disease, indicating that certain MMP-3 genotype may lead to increase in joint cartilage degradation. It is certainly of interest to investigate whether there is a genetic predisposition leading to high production levels of proMMP-3.
Involvement of the MMP in the process of joint disease was further investigated in experiments described in Chapter 6. This chapter presents the results of the cross-sectional study aimed at determination of protein and activity levels of MMP-1 and -3 in SF of patients with knee joint injury (injury group, N = 283), primary osteoarthritis (OA group, N = 108) and acute pyrophosphate arthritis (inflammatory group; N = 65). The levels of studied MMPs and TIMP-1 were compared to those in knee-healthy controls (CTRL, N = 35). The results showed that mean levels of proMMP-1 and -3 and TIMP-1 were increased in injury group, OA and inflammatory study groups when compared to the CTRL group. MMP-1 activity was increased in inflammatory and injury groups over CTRL levels, whereas MMP-3 activity levels were increased only in the inflammatory group. Moreover, the significant increase in MMP-1 activity in Į2M/MMP complexes
coincided with a decrease in TIMP-1 levels in the injury group, indicating that MMP activity in Į2M/MMP complexes may in fact provide indication of MMP/TIMP imbalance
in joint pathology.
To summarize, the findings presented in this thesis provide a significant body of evidence to support the presence of MMP/TIMP imbalance in joint pathology. It has been extensively documented that protein levels of TIMP-1 in synovial fluid and in the circulation of RA patients are likely to be insufficient to counteract the increased production of MMPs. However, little is known about the activation of proMMPs in vivo. It was previously indicated that activated, but not TIMP inhibited MMPs can be neutralized by Į2M. We hypothesized that in pathological situations involving joint inflammation and
destruction, the high production of proMMPs leads to increased levels of activated MMPs and because of the MMP/TIMP imbalance high levels of Į2M/MMP complexes may be
found in these conditions. Thus, if high levels of Į2M/MMP complexes were found in the
rheumatoid joint, it would support the idea of an MMP/TIMP imbalance. To test this hypothesis a method to measure MMP activity in MMP/Į2M complexes using small
fluorogenic substrates, such as TNO211-F was developed. Based on the results described in chapters 2, 3, 4, 5 and 6 it was concluded that increased production of proMMPs in fact results in increased amount of activated MMPs in complexes with Į2M in patient with joint
pathology. Although not all TIMP subclasses are accounted for, the presence of increased levels of Į2M/MMP complexes as such shows that TIMP levels are insufficient to inhibit
all activated MMPs. Moreover, this theory is also supported by our findings that an increase in Į2M/MMP complexes in SF of injury patients coincides with a decrease in
TIMP-1 levels.
hundred and two RA patients, who were randomly assigned to be treated with either leflunomide 20 mg/day after a loading dose of 100 mg/day for the first 3 days or methotrexate 15 mg/week (initial dose 7.5 mg/week, increase to 10 mg/week 4 weeks after baseline, and increase to 15 mg/week 8 weeks after baseline), were followed for 1 year. Clinical variables included DAS and CRP, measured at identical time points. Blood samples were available at baseline, 4 month and one year and were tested for MMP activity using fluorogenic substrate TNO211-F. The results show that levels of Į2M/MMP
complexes in systemic circulation were significantly reduced upon treatment in both study groups. The mechanism of action of methotrexate in RA is currently not completely understood but appears to be more than an effect on the purine biosynthesis, and appears to be not cell type specific. At least two modes of action of leflunomide have been documented: inhibition of dihydroorotate dehydrogenase (DHODH), by which leflunomide influences the de-novo pyrimidine biosynthesis, and interaction with primary and secondary signaling events. Moreover, leflunomide has been previously shown in vitro to reduce proMMP production in synovial tissue of RA patients as well as increase of the measured amount of TIMPs, leading to restoration of MMP/TIMP imbalance. Our results further extend these findings and provide in vivo indication that the net inhibitory effect on the proteolytic system of these drugs can be monitored in the systemic circulation. In the future, more sensitive or specific assays may provide detailed information with regard to MMPs involved in the disease and help to understand the mode of action of different drugs used to control it.
Future prospects
The research line described in this thesis could be expanded in several directions. Firstly, it is shown in chapter 6 that it is possible to modify the MMP activity assay to be able to perform measurements of specific MMP, such as MMP-1 or MMP-3 in synovial fluid. The specificity improvement could be achieved by re-designing LMW fluorogenic peptide substrates in order to have an MMP specific built-in cleavage site. The size of the substrate is important for specificity because bigger substrates allow a more complex structure which allows a better fit between the active site of the specific MMP subclass and the cleavage site of the substrate. Also new emerging techniques such as phage-display can be used to develop new MMP specific substrates. Given the almost unlimited number of possible combinations not only specificity but also sensitivity of the assays can be improved. This approach is however not likely to result in a major increase in specificity since the size of the substrate still has to be small enough to ensure the ability of the substrate to penetrate Į2M.
SUMMARY AND DISCUSSION
115 MMP specific inhibitors are however still the limiting factor for this approach since high specificity is rarely achieved in low molecular weight inhibitors.
The combination of the specific substrates and inhibitors seems to be the most logical solution for development of MMP specific activity assays. The work presented in chapter 6 gives an example of such assays. MMP-1 activity was measured using TNO113-F which is preferentially cleaved by activated MMP-1, whereas MMP-3 measurements were performed using TNO003-F which is in turn preferentially cleaved by activated MMP-3. To minimize the non-specific substrate cleavage by other enzymes (MMPs, serine or cysteine proteases) non-MMP-1 and a non-MMP-3, respectively, were used. The difference in fluorescence increase between samples with addition of general MMP inhibitor and non-MMP-1 or non-MMP-3 specific inhibitor was regarded as MMP-1 or MMP-3 specific.
In theory it should be also possible to further improve the sensitivity of the methods by combining existing ELISA techniques and fluorogenic substrates. Complex formation between activated MMPs and Į2M causes conformational changes of the Į2M molecules. If
specific antibodies can be developed against MMP containing Į2M molecules using
conventional antibodies development techniques or newly developed techniques such as aptamer libraries, than much more sensitive assay formats are possible, such as the measurements of MMP activity in captured Į2M/MMP complexes.
Also another area of research, not covered in this thesis, may provide additional information with regard to the MMP system. It is conceivable that using existing techniques assays can be developed to determine the amount of MMP/TIMP complexes in body fluids. This is the part of MMP activation/inhibition equation which up to date remains unaccounted for.
To expand the field of application of MMP activity assay one could think of cancer research. It has become clear that MMPs play a crucial role in metastasis of different types of cancer by the breakdown of connective tissue in the proximity of the malignancy and also in the process of neovascularisation. High systemic proMMP levels have been shown to be associated with shortened survival in several malignancies and it is therefore of interest to investigate whether MMP activity measurements in Į2M/MMP complexes can
provide a new tool to describe or to predict the course of the disease.
In conclusion, the results described in the present thesis show that:
- high proMMPs production leads to increased levels of activated MMP in SF and systemic circulation of RA patients;
- upon activation MMPs form complexes with Į2Macroglobulin in synovial fluid and in
the systemic circulation of RA patients;
- MMP-1, -3, -8 and -9 are involved the pathogenesis of RA;
- proMMP-3 levels at the onset of RA are predictive of the joint damage progression; - MMP/TIMP imbalance is present in various joint pathologies;
