Abstract
Objectives
To investigate the long-term prognostic significance of baseline plasma MMP-1 levels in a group of well-characterized male patients with known or suspected coronary artery disease, including those presenting with the acute coronary syndrome.
Background
MMP-1 is an interstitial collagenase that is considered the primary enzyme responsible for collagen degradation. In addition, MMP-1 can lead to platelet activation through the PAR1 pathway that is independent of thrombin.
Methods
Baseline plasma MMP-1 levels were measured in 364 male patients who were referred for coronary angiography and followed prospectively for the development of all-cause mortality for 5 years.
Results
After adjustment for a variety of baseline clinical, angiographic and laboratory parameters, baseline plasma MMP-1 levels (analyzed as a continuous variable) was an independent predictor of all-cause mortality at 5 years (HR, 1.49; 95% CI, 1.23-1.80; P<0.0001). Furthermore, in 3 additional multivariate models that included a wide variety of contemporary biomarkers with established prognostic efficacy (i.e., ST2, GDF-15, cystatin C, hs-CRP, myeloperoxidase, TIMP-1, adiponectin, RDW, hemoglobin, erythropoietin), MMP-1 remained an independent predictor of all-cause mortality at 5 years. Similar results were obtained when the analyses were restricted to the subpopulation of patients presenting with acute coronary syndrome.
Conclusions
Elevated levels of MMP-1 are associated with an increased risk of long-term all-cause mortality in patients with known or suspected coronary disease that is independent of a variety of clinical, angiographic, laboratory variables, including a whole host of contemporary biomarkers with established prognostic efficacy representing multiple different pathophysiologic processes.
Keywords: matrix metalloproteinase 1, biomarkers, acute coronary syndrome, prognosis, mortality
Introduction
Matrix metalloproteinases (MMPs) are a family of extracellular matrix degrading enzymes implicated in plaque destabilization (1,2). MMPs are expressed by macrophages, VSMC, and endothelial cells in response to inflammatory stimuli (2,3). MMP-1 is an interstitial collagenase that is considered the primary enzyme responsible for collagen degradation (3). MMP-1 is primarily detected in the vulnerable shoulder region and areas of foam cell formation in the atherosclerotic plaques where it colocalizes with degraded collagen fragments (1,4-6). Accordingly, MMP-1 has been associated with total plaque burden (2). Furthermore, platelets are known to harbor MMP-1, which is released after exposure of platelets to thrombin (3,7). MMP-1 is then subsequently capable of directly cleaving PAR1 on the platelet surface (3,8). PAR1 is a G-protein-coupled receptor that is classically activated through cleavage of the N-terminal exodomain by the serine protease thrombin (9). Thus, MMP-1 can lead to platelet activation through the PAR1 pathway that is independent of thrombin. Collectively, these basic observations suggest that MMP-1 may play a pathogenic role in atherosclerosis and its complications through two different, yet complementary, mechanisms – namely, matrix-degradation in the vessel wall and thrombin-independent effects on PAR1 and platelet activation (3).
Despite these basic science observations potentially linking MMP-1 to the development and propagation of atherosclerosis, no studies to date have specifically examined the prognostic significance of baseline plasma levels of MMP-1 in patients with ischemic heart disease. Accordingly, the objective of the current study was to investigate the long-term prognostic significance of baseline MMP-1 levels in a group of well-characterized male patients with known or suspected coronary artery disease who were referred for coronary angiography, including those presenting with the acute coronary syndrome. An additional and specific objective of the study was to compare the utility of MMP-1 in this regard with other well-established and contemporary biomarkers representing multiple different pathophysiologic processes, such as hemodynamic stress, inflammation, and myocardial necrosis, in an attempt to better delineate the clinical significance of the basic science observations pertaining to MMP-1. The aim was to test whether this novel protein could improve risk prediction above and beyond established and existing biomarkers.
Methods
The study population and design have been described in detail elsewhere (10). The study database was generated at an urban Veterans Administration (VA) Medical Center and was approved by the local IRB. Written informed consent was obtained from all patients. All patients who were referred to the Cardiac Catheterization Laboratory for coronary angiography at the Bronx Veterans Affairs Medical Center between January 13, 1999 and October 17, 2002 were eligible for inclusion in the database. Patients with active gastrointestinal bleeding or a hemoglobin concentration less than 8 gm/dL were excluded. During the period of study enrollment, a total of 523 unique and consecutive male patients underwent diagnostic coronary angiography. Of these 523 patients, 50 could not be enrolled because of an unexpected loss of key study personnel between January 29, 2001 and July 2, 2001. Of the remaining 473 patients, 84 were either unwilling or unable to provide informed consent. Thus, a total of 389 patients provided informed consent and constituted the final study population. Of these 389 patients, MMP-1 values were not available for 25 patients. The present study represents an analysis of both the entire cohort of 364 patients as well as the subpopulation of 180 patients who underwent diagnostic coronary angiography for the evaluation of ACS (12% with ST-segment elevation MI, 43% with non-ST-segment elevation MI, and 45% with unstable angina pectoris).
Fasting whole blood was obtained from all patients at the time of angiography prior to injection of contrast. The whole blood was centrifuged (4g) for 10 minutes at 4°C, the plasma aliquoted into 1.5 mL vials and then stored at -70°C for subsequent analysis. The samples were thawed only once when ready for analysis. All patients were followed prospectively for the development of all-cause mortality.
Patients were divided into tertiles according to baseline MMP-1 values (<=0.503 ng/mL, >0.503 to <1.082 ng/mL, and >=1.082 ng/mL). Summary statistics for continuous variables were presented both as means (with standard deviations) as well as medians (with interquartile ranges), and comparisons between the three groups were performed with the non-parametric Kruskall-Wallis test. All biomarkers were log transformed and normalized to reduce skewness and kurtosis of data. Categorical data were summarized as frequencies and percentages, and comparisons between the 3 groups were performed with Pearson chi-square test or Fisher's exact test.
Predictors of all-cause mortality at 60 months were identified by univariate Cox proportional hazards regression with results were presented as hazards ratios (HR) and corresponding 95% confidence intervals. For the entire cohort as well as the ACS subpopulation, each of the clinical, angiographic and laboratory variables (including biomarkers) listed in Table 1 were studied for their association with all-cause mortality; predictors with p<0.05 on unadjusted analyses were subsequently entered into multivariate Cox proportional hazards regression models. Independent predictors were then identified using the backward elimination procedure. All biomarkers were analyzed as continuous variables and the hazards ratios represented an increase in one standard deviation in the respective log transformed biomarker. Four different multivariate models were created for the entire cohort and for the ACS subpopulation. The first model adjusted only for baseline clinical, angiographic and laboratory variables. The other three “biomarker only” models, adjusted for a variety of biomarkers with established prognostic efficacy in the literature (hs-CRP, MPO, Cystatin-C, ST2, NT-proBNP, Adiponectin, TIMP-1, hemoglobin, RDW, and Erythropoietin). All biomarkers were checked for collinearity and three different “biomarker only” models were created so as not to include any 2 markers whose Spearman correlation was > 0.5. In this way, we sought to determine the relative prognostic significance of MMP-1 with respect to contemporary biomarkers with established prognostic significance and to better understand the pathophysiologic basis for the predictive power of MMP-1. Because this study was exploratory in nature and the objective was to determine the relationship between various covariates (including MMP-1) and the outcome of all-cause mortality in this specific patient population, there were no a priori intentions to create predictive models that could subsequently be used to predict outcomes in validation cohorts. Accordingly, prediction metrics such as the c-index were not reported for the current analyses.
Table 1. Baseline Characteristics of the Entire Population Stratified by Tertiles of Plasma MMP-1 Values.
| Characteristic | Statistics | 1st tertile (<=0.503 ng/mL) N=121 |
2nd tertile (>0.503 to <1.082 ng/mL) N=122 |
3rd tertile (>=1.082 ng/mL) N=121 |
p Value |
|---|---|---|---|---|---|
| Age | Mean (SD) | 62.6(10.2) | 65.9(10.2) | 67.1(9.2) | |
| Median (25th, 75th) | 62.5(53.9, 69.9) | 67.3(57.2, 74.3) | 67.3(61.8, 74.2) | 0.0018 | |
| Race | Black, N (%) | 37(30.6) | 33(32.0) | 41(33.9) | 0.4412 |
| Hispanic, N (%) | 41(33.9) | 29(23.8) | 32(26.5) | ||
| White, N (%) | 43(35.5) | 54(44.3) | 48(39.7) | ||
| Family History of Premature CAD | N (%) | 37(30.6) | 31(25.4) | 22(18.2) | 0.0804 |
| Diabetes mellitus | N (%) | 51(42.2) | 55(45.1) | 53(43.8) | 0.8987 |
| Hypertension | N (%) | 95(78.5) | 103(84.4) | 105(86.8) | 0.2073 |
| Any history of tobacco use | N (%) | 98(81.0) | 97(79.5) | 105(86.8) | 0.2910 |
| Active tobacco use | N (%) | 39(32.2) | 36(29.5) | 42(34.7) | 0.6857 |
| Hyperlipidemia* | N (%) | 69(57.0) | 61(50.0) | 64(52.9) | 0.5444 |
| Atrial fibrillation | N (%) | 8(6.6) | 8(6.6) | 13(10.7) | 0.3855 |
| CHF on presentation | N (%) | 29(24.0) | 27(22.1) | 41(33.9) | 0.0838 |
| MI on presentation | N (%) | 41(33.9) | 28(23.0) | 35(28.9) | 0.1678 |
| Aspirin | N (%) | 106(87.6) | 105(86.1) | 97(80.2) | 0.2384 |
| Beta blocker | N (%) | 83(8.6) | 82(67.2) | 83(68.6) | 0.9650 |
| ARB | N (%) | 6(5.0) | 9(7.4) | 7(5.8) | 0.7236 |
| ACE inhibitor | N (%) | 70(57.9) | 73(59.8) | 76(62.8) | 0.7302 |
| Statin | N (%) | 63(52.1) | 59(48.4) | 64(52.9) | 0.7531 |
| Fibrate | N (%) | 7(5.8) | 5(4.1) | 4(3.3) | 0.6372 |
| LV function | EF >= 55%, N(%) | 43(38.7) | 45(38.5) | 35(31.0) | 0.8777 |
| EF 45-54%, N(%) | 25(22.5) | 29(24.8) | 28(24.8) | ||
| EF 31-44%, N(%) | 26(23.4) | 26(22.2) | 32(28.3) | ||
| EF <= 30%, N(%) | 17(15.3) | 17(14.5) | 18(15.9) | ||
| No. of diseased coronary arteries** | 0, N(%) | 31(25.6) | 18(14.8) | 21(17.4) | 0.1068 |
| 1, N(%) | 24(19.8) | 22(18.0) | 13(10.7) | ||
| 2, N(%) | 24(19.8) | 35(28.7) | 35(28.9) | ||
| 3, N(%) | 34(28.1) | 42(34.4) | 47(38.8) | ||
| 4, N(%) | 8(6.6) | 5(4.1) | 5(4.1) | ||
| Prior CABG | N(%) | 8(6.6) | 14(11.5) | 11(9.1) | 0.4182 |
| ACS | N(%) | 61(50.4) | 57(46.7) | 62(51.2) | 0.7545 |
| STEMI | Yes | 11(9.1) | 6(4.9) | 6(5.0) | 0.3083 |
| NSTEMI | Yes | 30(24.8) | 22(18.0) | 29(24.0) | 0.3843 |
| Unstable angina | Yes | 20(16.5) | 29(23.8) | 27(22.5) | 0.3347 |
| BMI | Mean(sd) | 29.3(6.2) | 29.3(5.9) | 27.3(4.8) | |
| Med(25th, 75th) | 28.2(25.6, 31.8) | 28.6(24.6, 32.6) | 27.1(23.8, 30.2) | 0.0461 | |
| hs-CRP | Mean (SD) | 20.7(33.3) | 26.3(47.2) | 25.9(41.9) | |
| Median (25th, 75th) | 8.9(4.7,18) | 8.4(3.2, 17.3) | 11.5(4.8, 21.9) | 0.1962 | |
| Troponin | Mean (SD) | 13.8(68.1) | 5.9(20.3) | 12.2(43.3) | |
| Median (25th, 75th) | 0.3(0.2, 3.1) | 0.3(0.2, 1.7) | 0.3(0.2, 1.5) | 0.8277 | |
| Serum creatinine | Mean (SD) | 1.08(0.29) | 1.13(0.42) | 1.55(2.10) | |
| Med(25th, 75th) | 1.0(0.9, 1.2) | 1.0(0.9, 1.2) | 1.1(0.9, 1.3) | 0.0515 | |
| NT-pro-BNP | Mean(SD) | 1252(982) | 1196(940) | 1591(1230) | |
| Med(25th, 75th) | 974(488, 1723) | 848(529, 1580) | 1215(646, 2198) | 0.0247 | |
| Erythropoietin | Mean(SD) | 16.1(13.3) | 13.5(10.5) | 19.4(20.4) | |
| Med(25th, 75th) | 11.8(7.8, 18.1) | 10.5(7.7, 15.5) | 12.6(8.7,19.4) | 0.0405 | |
| GFR-MDRD | Mean(SD) | 83.7(22.6) | 82.4(25.3) | 76.0(28.2) | |
| Med(25th, 75th) | 81(70, 94) | 79(69,94) | 72(62,90) | 0.0249 | |
| TIMP-1 | Mean(SD) | 81.5(38.1) | 87.8(41.4) | 110.7(52.3) | |
| Med(25th, 75th) | 70.2(63.1, 90.7) | 78.4(64.6, 93.8) | 96.1(78.3, 125.7) | <0.0001 | |
| Myeloperoxidase | Mean(SD) | 25.3(23.0) | 22.1(13.1) | 28.9(19.8) | |
| Med(25th, 75th) | 18.4(13.3, 28.1) | 17.9(12.8, 27.4) | 22.5(16.0, 35.5) | 0.0041 | |
| Adiponectin | Mean(SD) | 7.5(5.7) | 7.0(4.7) | 9.7(6.9) | |
| Med(25th, 75th) | 5.8(3.5, 9.4) | 5.7(3.7, 8.7) | 7.7(4.9, 14.0) | 0.0028 | |
| ST2 | Mean(SD) | 22.9(13.6) | 23.5(14.3) | 28(15.5) | |
| Med(25th, 75th) | 19.7(15.8, 26.7) | 19.2(15.6, 26.2) | 21.8(17.7, 34.7) | 0.0041 | |
| Cystatin-C | Mean(SD) | 957(240) | 1055(348) | 1289(671) | |
| Med(25th, 75th) | 941(770,1055) | 997(814,1160) | 1104(921, 1356) | <0.0001 | |
| GDF-15 | Mean(SD) | 1212(787) | 1342(886) | 1859(1391) | |
| Med(25th, 75th) | 992(757, 1367) | 1100(856, 1532) | 1492(1034,2056) | <0.0001 | |
| RDW | Mean(SD) | 14.1(1.2) | 13.9(1.2) | 14.9(2.6) | |
| Med(25th, 75th) | 13.9(31.1, 14.7) | 13.7(13.2, 14.6) | 14.2(13.3, 15.3) | 0.0094 | |
| Hemoglobin | Mean(SD) | 13.7(1.6) | 13.6(1.6) | 13.1(1.9) | |
| Med(25th, 75th) | 13.8(12.8, 14.7) | 13.7(12.6,14.7) | 13.2(12.0, 14.3) | 0.0364 |
Hyperlipidemia was diagnosed in patients who had been given lipid-lowering medication or had a history of total cholesterol levels > 240 mg/dl [19].
Takes into account the left main, left anterior descending, left circumflex, and right coronary arteries (minimum = 0, maximum = 4
Time-to-event at 60 months was presented with Kaplan-Meier curves for the individual endpoint of all-cause mortality. Comparisons between the 3 groups identified by tertiles of MMP-1 as described above were performed using the log-rank test.
All analyses used two-sided tests with an overall significance level of α = 0.05. All statistical analyses were performed using SAS version 8 (SAS Institute Inc, Cary, NC).
Fasting blood was obtained from all patients at the time of angiography for subsequent analysis. Commercially available kits were used to measure the plasma levels of high-sensitivity C-Reactive Protein (hs-CRP; Life Diagnostics, West Chester, PA, USA), N-Terminal-Pro-B-Type Natriuretic Peptide (NT-proBNP; Diagnostic Automation, Calabasas, CA, USA), Myeloperoxidase (MPO; Assay Designs, Ann Arbor, Michigan, USA), Adiponectin (R & D Systems, Minneapolis, MN, USA), total TIMP-1 (EMD Biosciences, San Diego, CA, USA), GDF-15 (R & D Systems, Minneapolis, MN, USA), ST2 (R & D Systems, Minneapolis, MN, USA), Cystatin C (BioVendor, Asheville, NC, USA), erythropoietin (R & D Systems, Minneapolis, MN, USA), and MMP-1 (R & D Systems, Minneapolis, MN, USA).
Patients were followed for the occurrence of all-cause mortality. The information regarding the date of death was obtained using the following modalities: death certificate, social security death index, conversation with next of kin and/or primary physician, and review of medical records.
Results
Baseline characteristics
A total of 364 male patients were enrolled in the study. Five-year clinical data in the form of all-cause mortality were available for all of the patients. The baseline clinical, laboratory, and angiographic characteristics of the study population stratified by the lower, middle, and upper tertiles of MMP-1 values are shown in Table 1.
Association of MMP-1 with baseline clinical variables and other biomarkers
Elevated MMP-1 levels were associated with older age, lower BMI, lower GFR, and lower hemoglobin values. In addition, the levels of MMP-1 were also positively correlated with those of NT-proBNP, erythropoietin, TIMP-1, myeloperoxidase, adiponectin, ST2, Cystatin C, GDF-15 and RDW.
Clinical outcomes for the entire population
There were a total of 109 deaths (28.02%) at 5 years. The following baseline variables were significant for their association all-cause mortality at 5 years with p<0.05 on univariate analysis: age/10 years, a family history of premature coronary artery disease, diabetes mellitus, myocardial infarction on presentation, congestive heart failure on presentation, atrial fibrillation, ACE-inhibitor use, left ventricular systolic function, serum creatinine, GFR-MDRD, the number of diseased coronary arteries, a history of CABG surgery, as well as the following biomarkers all analyzed as continuous variables: TIMP-1, hs-CRP, Fibrinogen, Adiponectin, NT-proBNP, hemoglobin, Erythropoietin and RDW. Together with MMP-1 (analyzed as a continuous variable), these significant univariate predictors of outcome were entered into multivariate models to identify the independent predictors of all-cause mortality at 5 years (Table 2). Congestive heart failure on presentation was not included as a covariate in multivariate models because of collinearity with NT-proBNP and left ventricular systolic function. Model 1 adjusted for clinical and angiographic variables only, Models 2-4 adjusted only for other established biomarkers. TIMP-1 and Cystatin C, GDF-15 and Cystatin C, and TIMP-1 and GDF-15 were not included in the same multivariate models due to a high degree of collinearity (>0.5) between these biomarkers. In each and every one of these models, MMP-1 emerged as a strong and independent predictor of all-cause mortality at 5 years.
Table 2. Multivariate Cox Regression Models for All-Cause Mortality at 5 Years for the Entire Cohort.
|
All variables included in the selection were significant predictors of all-cause mortality at 5 years on univariate analysis, p<0.05
Model 1 included only clinical and angiographic variables in the selection, along with MMP-1
Model 2 included MMP-1 with other biomarkers, except GDF-15 and TIMP-1 due to a high degree of collinearity between them
Model 3 included MMP-1 with other biomarkers, except Cystatin C and GDF-15, due to a high degree of collinearity between them
Model 4 included MMP-1 with other biomarkers, except Cystatin C and TIMP-1, due to a high degree of collinearity between them
Kaplan-Meier survival analysis showed a significantly reduced survival in patients in the highest MMP-1 tertile (Figure 1). At 5 years, the all-cause mortality rate was 18.2% in the lowest tertile, 22.1% in the middle tertile, and 43.8% in the highest tertile (P < 0.0001 by log-rank test).
Figure 1.
Kaplan-Meier curves for all-cause mortality according to baseline plasma MMP-1 levels stratified by tertiles in the entire cohort of patients. At 5 years, the number of events (deaths) in the lowest tertile was 22 (18.2%), 27 (22.1%) in the middle tertile, and 53 (43.8%) in the upper tertile (p < 0.0166 by log-rank test). The upper numbers on the x-axis represent the number of patients at risk.
Clinical outcomes for the ACS subpopulation
To determine whether the predictive power of MMP-1 was either related to or affected by baseline risk, similar analyses were performed in the subgroup of patients who had presented with an acute coronary syndrome at baseline. The ACS group, which consisted of 180 patients, was comprised of those with unstable angina, NSTEMI and STEMI. For this group, in whom follow-up data for all-cause mortality were similarly available for 100% of the patients, there were a total of 59 deaths.
In the ACS subpopulation, the following baseline variables were significant for their association all-cause mortality at 5 years with p<0.05 on univariate analysis: age/10 years, diabetes mellitus, CHF on presentation, myocardial infarction on presentation, left ventricular systolic function, serum creatinine, GFR-MDRD, the number of diseased coronary arteries, as well as the following biomarkers all analyzed as continuous variables: TIMP-1, hs-CRP, adiponectin, NT-proBNP, hemoglobin, and RDW. As was done for the entire cohort of patients, MMP-1 (analyzed as a continuous variable) and these significant univariate predictors of outcome were entered into multivariate models to identify the independent predictors of all-cause mortality at 5 years (Table 3). Congestive heart failure on presentation was not included as a covariate in multivariate models because of a very high and significant correlation between this variable and both NT-proBNP and left ventricular systolic function). Model 1 adjusted for clinical and angiographic variables only, Models 2-4 adjusted only for other established biomarkers. TIMP-1 and Cystatin C, GDF-15 and Cystatin C, and TIMP-1 and GDF-15 were not included in the same multivariate models due to a high degree of collinearity (>0.5) between these biomarkers. In each and every one of these models, MMP-1 emerged as a strong and independent predictor of all-cause mortality at 5 years. As was the case for the entire cohort of patients, MMP-1 emerged as a strong and independent predictor of all-cause mortality at 5 years in each and every one of these models.
Table 3. Multivariate Cox Regression Models for All-Cause Mortality at 5 Years for the ACS Subpopulation.
|
All variables included in the selection were significant predictors of all-cause mortality at 5 years on univariate analysis, p<0.05
Model 1 included only clinical and angiographic variables in the selection, along with MMP-1
Model 2 included MMP-1 with other biomarkers, except GDF-15 and TIMP-1, due to a high degree of collinearity between them
Model 3 included MMP-1 with other biomarkers, except GDF-15 and Cystatin C, due to a high degree of collinearity between them
Model 4 included MMP-1 with other biomarkers, except Cystatin C and TIMP-1, due to a high degree of collinearity between them
Kaplan-Meier survival analysis showed a significantly reduced survival in patients in the highest MMP-1 tertile (Figure 2). At 5 years, the all-cause mortality rate was 16.7% in the lowest tertile, 30.0% in the middle tertile, and 51.7% in the highest tertile (P = 0.0002 by log-rank test).
Figure 2.
Kaplan-Meier curves for all-cause mortality according to baseline plasma MMP-1 levels stratified by tertiles in the ACS subpopulation of patients. At 5 years, the number of events (deaths) in the lowest tertile was 10 (16.7%), 18 (30.0%) in the middle tertile, and 31 (51.7%) in the upper tertile (p = 0.0002 by log-rank test). The upper numbers on the x-axis represent the number of patients at risk.
Discussion
The most significant finding of our study was that elevated baseline plasma levels of MMP-1 were strong and independent predictors of long-term all-cause mortality in a cohort of male patients with known or suspected coronary artery disease referred for coronary angiography. This association remained significant even after adjustment for a wide variety of clinically relevant covariates, including a whole host of contemporary biomarkers with established prognostic significance in the literature. Furthermore, the independent relation between MMP-1 levels and all-cause mortality was also evident in the subgroup of patients presenting with an acute coronary syndrome at baseline. It is particularly noteworthy that MMP-1 emerged as an independent predictor of all-cause mortality even after adjustment for numerous well-established biomarkers representing different pathophysiologic processes and shown to have prognostic significance in a variety of cardiac populations. To this end, we methodically performed several different multivariate models which included in the selection traditional biomarkers of inflammation (such as hs-CRP and MPO) (11,12), biomarkers of prognostic value in heart failure (such as NT-proBNP, ST2, cystatin C) (13-15), biomarkers of hematologic significance with known prognostic utility (such as RDW, Hb and erythropoietin) (16-18), and recently-described novel biomarkers with incompletely understood and/or multiple different mechanisms (such as adiponectin, TIMP-1 and GDF-15) (10,19-20). In all of these multivariate models, MMP-1 emerged as the most significant and only independent biomarker which predicted all-cause mortality at 5 years. This was true despite the fact that the biomarkers included in these analyses were significant univariate predictors of all-cause mortality. Not only do these data suggest that MMP-1 is an especially powerful biomarker but we believe that the consistency of the association between MMP-1 and mortality in all of these multivariate models is a testimony to the validity of the findings.
Matrix metalloproteinases are a family of zinc-dependent endopeptidases that have recently emerged as important mediators of platelet and endothelial function, and atherothrombotic disease (3). MMP-1 is an interstititial collagenase that is considered the primary enzyme responsible for collagen degradation (3,5,7). It is expressed in most human tissues, including cells of the blood vessel wall, inflammatory cells, and platelets (3). The matrix-degrading functions of MMP-1 in the vessel wall have been considered to play a critical role in the pathogenesis of atherosclerosis and its complications. Indeed, consistent with a role for MMP-1 in atherothrombosis, MMP-1 expression has been shown to be increased in atherosclerotic plaques, and in particular in the cells in the vulnerable region of the plaque (4-6). Furthermore, platelets express several metalloproteinases on their surface, including MMP-1 (3,7). MMP-1 has been demonstrated to have agonist activity for PAR1 (8), which is a G-protein-coupled receptor that is classically activated through cleavage of the N-terminal exodomain by the serine protease thrombin (3,9). Exposure of platelets to collagen results in activation of MMP-1, which in turn directly cleaves PAR1 on the surface of platelets (8). This in turn leads to collagen-dependent thrombogenesis, arterial thrombosis, and clot retraction. Thus, there may be a dual role for MMP-1 in the pathogenesis of atherosclerosis: destabilization of the collagenous structure of plaques coupled with the induction of a procoagulant state through its effects on platelets (3). This unique dual role could potentially explain our data on how and why MMP-1 emerged as a particularly powerful and independent biomarker in a cardiac population. Furthermore, with respect to therapeutic implications of these findings, blockade of the MMP-1/PAR-1 pathway with an MMP-1 inhibitor greatly curtailed arterial thrombosis in an animal model of arterial denudation, suggesting that such inhibition may represent a point of early intervention in preventing arterial thrombosis (3,8). Thus, inhibition of the MMP-1/PAR-1 axis may represent a novel approach to the treatment of vascular thrombosis.
The mechanism(s) by which high MMP-1 levels are associated with adverse outcomes in patients with known or suspected coronary artery disease is unknown. While it is certainly tempting to attribute the prognostic utility of MMP-1 to its known effects on platelet activation and/or its destabilizing actions on the collagenous structures of the atherosclerotic plaque, demonstration of such a cause-and-effect association is far beyond the scope of the present study. Furthermore, it is also conceivable that the prognostic power of MMP-1 may in fact be completely unrelated to any of its cardiovascular effects. For example, MMP-1 levels have been shown to have prognostic significance in patients with malignancy (21-23). Finally, as is the case for the vast majority of biomarkers in the published literature, it is unknown whether MMP-1 represents an actual mediator of increased risk or simply a marker of that risk. Regardless, however, what is clearly evident from the current study is that MMP-1 likely represents a novel and powerful prognostic biomarker for the prediction of long-term all-cause mortality that is independent not only of traditional cardiovascular risk factors but also independent of many of the well-established biomarkers in the literature.
There are several limitations to this study. First, the study was exploratory and observational in nature and as such is subject to the limitations of this type of analysis. However, we had formulated the hypothesis that elevated MMP-1 levels might be associated with adverse outcomes (based on review of the available literature in this regard) before performing analyses on our database. We therefore feel that this approach reduced the risk of spurious conclusions. Second, the study did not have a ‘control’ group consisting of patients who were not referred for cardiac catheterization. Third, the size of our population was small and the study was not designed with a priori calculations with respect to sample size or statistical power. As such, the findings need to be confirmed in larger and prospectively designed studies. Fourth, the population was a high-risk cohort as evidenced by both clinical and laboratory parameters as well as by the high event rate for death. Therefore, it is unknown whether MMP-1 levels would be similarly predictive of events in a low-risk population. Finally, this study was conducted in an exclusively male population and the results cannot be extrapolated to women.
Conclusions
In conclusion, we found that elevated MMP-1 levels were a strong and independent predictor of all-cause mortality in a broad population of male patients referred for coronary angiography, including those with the acute coronary syndrome. Furthermore, in multivariate models adjusting for numerous other contemporary biomarkers with established prognostic utility in a variety of cardiac populations, MMP-1 consistently emerged as the most important and only independent predictor of all-cause mortality at 5 years. While these data are certainly consistent with the known biologic actions of MMP-1 on a basic science level, it remains unknown whether MMP-1 represents a marker or a mediator of risk. The answer to this question could have potential therapeutic implications, as there are a variety of MMP inhibitors currently under development.
MMP-1 is an interstitial collagenase that is considered the primary enzyme responsible for collagen degradation
MMP-1 can lead to platelet activation through the PAR1 pathway that is independent of thrombin
Elevated levels of MMP-1 are associated with an increased risk of long-term all-cause mortality in patients with known or suspected coronary disease that is independent of a variety of clinical, angiographic, laboratory variables
Acknowledgments
Financial Support: This study was sponsored, in part, by grants related to biomarkers from the American Heart Association (Grant-in-Aid 10GRNT3700003) and the NIH-NHLBI (5R21HL098661-02).
Abbreviations
- MMP-1
matrix metalloproteinase 1
- eGFR
estimated glomerular filtration rate
- PAR1
protease-activated receptor-1
- MI
myocardial infarction
- NSTEMI
non-ST-segment elevation MI
- STEMI
ST-segment elevation MI
- ACS
acute coronary syndrome
- CAD
coronary artery disease
- CABG
coronary artery bypass graft surgery
- CHF
congestive heart failure
- LV
left ventricular
- EF
ejection fraction
- BMI
body mass index
- hs-CRP
high-sensitivity C-Reactive Protein
- TIMP-1
tissue inhibitor of matrix metalloproteinase 1
- NT-proBNP
N-Terminal-Pro-B-Type Natriuretic Peptide
- MPO
myeloperoxidase
- GDF-15
Growth Differentiation Factor 15
- RDW
red blood cell distribution width
- ACE-I
angiotensin converting enzyme inhibitor
- ARB
angiotensin receptor blocker
Footnotes
Relationship with Industry: All authors have reported they have no relationships relevant to the contents of this paper to disclose.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
References
- 1.Newby AC. Metalloproteinase expression in monocytes and macrophages and its relationship to atherosclerotic plaque instability. Arterioscler Thromb Vasc Biol. 2008;28:2108–14. doi: 10.1161/ATVBAHA.108.173898. [DOI] [PubMed] [Google Scholar]
- 2.Lehrke M, Greif M, Broedl UC, et al. MMP-1 serum levels predict coronary atherosclerosis in humans. Cardiovasc Diabetol. 2009;8:50. doi: 10.1186/1475-2840-8-50. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Austin KM, Covic L, Kuliopulos A. Matrix metalloproteinases and PAR1 activation. Blood. 2013;121:431–9. doi: 10.1182/blood-2012-09-355958. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Sukhova GK, Schonbeck U, Rabkin E, et al. Evidence for increased collagenolysis by interstititial collagenases-1 and -3 in vulnerable human atheromatous plaques. Circulation. 1999;99:2503–9. doi: 10.1161/01.cir.99.19.2503. [DOI] [PubMed] [Google Scholar]
- 5.Galis ZS, Sukhova GK, Lark MW, et al. Increased expression of matrix metalloproteinases and matrix degrading activity in vulnerable regions of human atherosclerotic plaques. J Clin Invest. 1004;94:2493–503. doi: 10.1172/JCI117619. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Galis ZS, Muszynski M, Sukhova GK, et al. Enhanced expression of vascular matrix metalloproteinases induced in vitro by cytokines and in regions of human atherosclerotic lesions. Ann N Y Acad Sci. 1995;748:501–7. doi: 10.1111/j.1749-6632.1994.tb17348.x. [DOI] [PubMed] [Google Scholar]
- 7.Galt SW, Lindemann S, Allen L, et al. Outside-in signals delivered by matrix metalloproteinase-1 regulate platelet function. Circ Res. 2002;90:1093–9. doi: 10.1161/01.res.0000019241.12929.eb. [DOI] [PubMed] [Google Scholar]
- 8.Trivedi V, Boire A, Tchemychev B, et al. Platelet matrix metalloproteinase-1 mediates thrombogenesis by activating PAR1 at a cryptic ligand site. Cell. 2009;137:332–43. doi: 10.1016/j.cell.2009.02.018. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Leger AJ, Covic L, Kuliopulos A. Protease-activated receptors in cardiovascular diseases. Circulation. 2006;114:1070–7. doi: 10.1161/CIRCULATIONAHA.105.574830. [DOI] [PubMed] [Google Scholar]
- 10.Cavusoglu E, Ruwende C, Chopra V, et al. Adiponectin is an independent predictor of all-cause mortality, cardiac mortality and myocardial infarction in patients presenting with chest pain. Eur Heart J. 2006;27:2300–9. doi: 10.1093/eurheartj/ehl153. [DOI] [PubMed] [Google Scholar]
- 11.Lindahl B, Toss H, Siegbahn A, et al. Markers of myocardial damage and inflammation in relation to long-term mortality in unstable coronary artery disease. FRISC Study Group. Fragmin during Instability in Coronary Artery Disease. N Engl J Med. 2000;343:1139–47. doi: 10.1056/NEJM200010193431602. [DOI] [PubMed] [Google Scholar]
- 12.Cavusoglu E, Ruwende C, Eng C, et al. Usefulness of baseline plasma myeloperoxidase levels as an independent predictor of myocardial infarction at two years in patients presenting with acute coronary syndrome. Am J Cardiol. 2007;99:1364–8. doi: 10.1016/j.amjcard.2006.12.060. [DOI] [PubMed] [Google Scholar]
- 13.Richards AM, Doughty R, Nicholls MG, et al. Plasma N-terminal pro-brain natriuretic peptide and adrenomedullin: prognostic utility and prediction of benefit from carvedilol in chronic ischemic left ventricular dysfunction. Australia-New Zealand Heart Failure Group. J Am Coll Cardiol. 2001;37:1781–7. doi: 10.1016/s0735-1097(01)01269-4. [DOI] [PubMed] [Google Scholar]
- 14.Dupont M, Wu Y, Hazen SL, et al. Cystatin C identifies patients with stable chronic heart failure at increased risk for adverse cardiovascular events. Circ Heart Fail. 2012;5:602–9. doi: 10.1161/CIRCHEARTFAILURE.112.966960. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Rehman SU, Mueller T, Januzzi JL., Jr Characteristics of the novel interleukin family biomarker ST2 in patients with acute heart failure. J Am Coll Cardiol. 2008;52:1458–65. doi: 10.1016/j.jacc.2008.07.042. [DOI] [PubMed] [Google Scholar]
- 16.Cavusoglu E, Chopra V, Gupta A, et al. Relation between red blood cell distribution width (RDW) and all-cause mortality at two years in an unselected population referred for coronary angiography. Int J Cardiol. 2010;141:141–6. doi: 10.1016/j.ijcard.2008.11.187. [DOI] [PubMed] [Google Scholar]
- 17.Cavusoglu E, Chopra V, Gupta A, et al. Usefulness of anemia in men as an independent predictor of two-year cardiovascular outcome in patients presenting with acute coronary syndrome. Am J Cardiol. 2006;98:580–4. doi: 10.1016/j.amjcard.2006.03.031. [DOI] [PubMed] [Google Scholar]
- 18.Belonje AM, Voors AA, van der Meer P, et al. Endogenous erythropoietin and outcome in heart failure. Circulation. 2010;121:245–51. doi: 10.1161/CIRCULATIONAHA.108.844662. [DOI] [PubMed] [Google Scholar]
- 19.Cavusoglu E, Ruwende C, Chopra V, et al. Tissue inhibitor of metalloproteinase-1 (TIMP-1) is an independent predictor of all-cause mortality, cardiac mortality, and myocardial infarction. Am Heart J. 2006;151:1101. doi: 10.1016/j.ahj.2006.02.029. [DOI] [PubMed] [Google Scholar]
- 20.Wollert KC, Kempf T, Peter T, et al. Prognostic value of growth-differentiation factor-15 in patients with non-ST-elevation acute coronary syndrome. Circulation. 2007;115:962–71. doi: 10.1161/CIRCULATIONAHA.106.650846. [DOI] [PubMed] [Google Scholar]
- 21.McGowan PM, Duffy MJ. Matrix metalloproteinase expression and outcome in patients with breast cancer: analysis of a published database. Ann Oncol. 2008;19:1566–72. doi: 10.1093/annonc/mdn180. [DOI] [PubMed] [Google Scholar]
- 22.Cheng S, Tada M, Hida Y, et al. High MMP-1 mRNA expression is a risk factor for disease-free and overall survivals in patients with invasive breast carcinoma. J Surg Res. 2008;146:104–9. doi: 10.1016/j.jss.2007.05.032. [DOI] [PubMed] [Google Scholar]
- 23.Li M, Xiao T, Zhang Y, et al. Prognostic significance of matrix metalloproteinase-1 levels in peripheral plasma and tumor tissues of lung cancer patients. Lung Cancer. 2010;69:341–7. doi: 10.1016/j.lungcan.2009.12.007. [DOI] [PubMed] [Google Scholar]