Matrix metalloproteinases as predictive markers of coronary in-stent restenosis

Abstract

Background: Matrix metalloproteinases (MMPs) and the tissue inhibitors of matrix metalloproteinases (TIMPs) are involved in remodelling the extracellular matrix of the vascular system. Altered MMP expression has been linked to the development of atherosclerosis, acute coronary syndrome, myocardial dysfunction and complications after vascular procedures including in-stent restenosis (ISR) after percutaneous coronary intervention. Other studies have focussed on the inactive pro-form of matrix metalloproteinases. However, an enzyme-linked immunosorbent assay (ELISA) system that measures the activity of specific matrix metalloproteinases is available. This thesis aimed to investigate whether circulating active MMPs could act as risk markers for ISR, and progression of coronary atherosclerosis and myocardial dysfunction.

Methods: Two groups of patients were studied. Initially, a case-control study was recruited, in which patients with a history of bare-metal stent restenosis were compared to patients who were asymptomatic after coronary stenting. A panel of MMPs was measured in the plasma of patients with restenosis and compared to controls. MMPs and TIMPs were measured using both conventional and activity ELISA. The contribution of MMPs to the risk prediction, beyond that afforded by established clinical and demographic risk models was investigated. In the second study a prospective cohort of patients undergoing coronary stenting were followed for 12 months. This study provided an opportunity to confirm the results of the first study. Additionally, data on clinical presentation and echocardiographic function allowed correlation of active MMPs and TIMPs with acute coronary syndrome, myocardial infarction and diastolic myocardial dysfunction. Serial blood samples over six months allowed tracking of MMP activity with each condition. The occurrence of both ISR and other coronary end-points were collected and correlated with MMP activity. Finally, the seasonal variation and stability during storage of MMPs at -80ºC and the relationship between the ratio of pro-MMP-9:TIMP-1 and MMP-9 activity were assessed. Results: Active MMPs -3 and -9, as well as TIMP-1 were strongly associated with a history of ISR, and were independent of clinical and demographic risk factors. When these markers were included in conventional risk models, they appeared to provide additional predictive information, with improved area under the curve and net reclassification index. In the prospective study, there was no association between active MMP-9 or TIMP-1 prior to intervention, but the change in active MMP-9 at three months was significantly greater in the group that developed ISR. Conversely, there was no association between active MMP-9 and non-ISR cardiac events. Initial presentation with ST-segment elevation myocardial infarction was also associated with increasing active MMP-9 at three months. Both active MMP-9 and TIMP-1 were associated with the occurrence of diastolic, but not systolic, myocardial dysfunction. The above findings were independent of clinical and demographic variables. Both active MMP-9 and TIMP-1 appeared to be stable in storage for up to three years, and not affected by variation with the season of sample collection. The relationship between the pro-MMP-9:TIMP-1 ratio and active MMP-9 suggests that the pro-MMP-9:TIMP-1 ratio is not a surrogate measure for MMP-9 activity. Conclusions: Active MMPs-3, -9 and TIMP-1 appear to be associated with ISR, and measuring them may allow improved prediction of ISR. Additionally, increasing MMP-9, but not TIMP-1, after coronary stenting may indicate those patients who are developing ISR. Active MMP-9 also rises in the months after ST-elevation myocardial infarction, and adjustment for this may be necessary if active MMP-9 is to be used as a clinical predictor of ISR. Active MMP-9 and TIMP-1 do not appear to be associated with the development of non-ISR events. Both active MMP-9 and TIMP-1 appear to be markers of diastolic, but not systolic, dysfunction, and active MMP-9 may be a sensitive early marker for diastolic dysfunction. Active MMP-9 and TIMP-1 appear to be stable in storage for up to three years, and not subject to seasonal variation, therefore these factors do not appear to be important biases. The pro-MMP-9:TIMP-1 ratio does not indicate the activity of MMP-9, and therefore active MMP-9 should be measured directly.