Identification of biomarker panels as predictors of severity in coronary artery disease

Abstract Matrix metalloproteinases (MMPs) are implicated in atherosclerotic plaque rupture and recondition. Specific tissue inhibitors (TIMPs) control MMP functions. Both MMPs and TIMPs are potential biomarkers of plaque instability. Elevated Apo‐CII and CIII and Apo‐E levels are recognized as cardiovascular disease risk factors. We aimed to establish the best blood biomarker panel to evaluate the coronary artery disease (CAD) severity. Plasma levels of MMP‐3 and MMP‐9, TIMP‐1 and TIMP‐2, Apo‐CII, Apo‐CIII and Apo‐E were measured in 472 patients with CAD evaluated by coronary angiography and electrocardiography, and in 285 healthy controls. MMP‐3 and MMP‐9 plasma levels in CAD patients were significantly increased (P < 0.001) compared to controls (3.54‐ and 3.81‐fold, respectively). Furthermore, these increments are modulated by CAD severity as well as for Apo‐CII and Apo‐CIII levels (P < 0.001). TIMPs levels were decreased in CAD versus controls (P < 0.001) and in inverse correlation to MMPs. Standard ROC curve approach showed the importance of panels of biomarkers, including MMP‐3, MMP‐9, TIMP‐1, TIMP‐2, Apo‐CII and Apo‐CIII, for disease aggravation diagnosis. A high area under curve (AUC) value (0.995) was reached for the association of MMP‐9, TIMP‐2 and Apo‐CIII. The unbalance between MMPs and TIMPs in vascular wall and dyslipidaemia creates favourable conditions for plaque disruption. Our study suggests that the combination of MMP‐9, TIMP‐2 and Apo‐CIII values (‘CAD aggravation panel’) characterizes the severity of CAD, that is electrophysiological state, number of involved vessels, stent disposal and type of stent.


| INTRODUC TI ON
The main cause of coronary artery disease (CAD) is atherosclerosis, which is a multifactorial disorder identified by artery intima damage. The normal vascular intima is in balance with the synthesis and degeneration of extracellular matrix (ECM), thus any shortage in ECM synthesis causes degradation and disruption of atherosclerotic plaques, specifically the vulnerable plaque areas leading to stenosis and myocardial infarction (MI). 1,2 The major cellular and extracellular protein components of atherosclerotic plaques are apolipoproteins, fibrin(ogen)-related antigens and matrix metalloproteinases (MMPs). 3 MMPs, or matrixins, a class of estimated 28 endopeptidases, perform an important function in vascular inflammation and atherosclerotic plaque remodelling with potential plaque rupture via fibrous cap deterioration. 2,4 MMP-3 is implicated in the formation of atherosclerotic plaque, has an important role in the turnover of ECM in the vascular wall leading to plaque rupture. 5 MMP-9 is involved in collagen degradation, vasculature remodelling, angiogenesis, inflammation and atherosclerotic plaque rupture. 5 MMPs activities are controlled by endogenous inhibitors, that is tissue inhibitors of metalloproteinases or TIMPs. 6 An imbalance between MMPs and TIMPs could cause large increase in MMP activity and may lead to pathological changes in the vessel wall structure associated with vascular disease, enhancing atherosclerotic plaque development and instability. 7 Thus blood levels of MMPs and TIMPs appear as CAD risk biomarkers. 8 We hypothesized that MMP-3, MMP-9, TIMP-1 and TIMP-2 circulating levels were irregular in patients at various severity stages of CAD. We also measured the circulating levels of apolipoproteins, namely Apo-E, Apo-CII and Apo-CIII, all being biological risk factors for cardiovascular diseases and their different events, especially CAD. 9 Apo-CII and -CIII lead to hypertriglyceridemia which favours atherosclerosis development by promoting inflammation and plaque vulnerability. 10 Apo-E is involved in the stimulation of cholesterol efflux from macrophage foam cells in the atherosclerotic lesion, and thus in the pathogenesis of cardiovascular disease. 10 Previous studies found that the degradation of ApoC-II by MMPs may play an important role in the development of pathophysiological conditions with Apo-CII deficiency such as atherosclerosis, thus MMPs are directly involved in the accumulation of cholesterol in atherosclerotic lesions 11 ; also, a silenced Apo-CIII could increase the activity of MMP-9. 12 We aimed to study the association between MMPs, TIMPs and apolipoproteins in Tunisian patients with CAD and established a biomarker panel able to evaluate CAD aggravation and to assess a diagnostic panel.

| Studied cohort
In this case-control multicentric study, we enrolled 472 patients based on electrocardiography and coronary angiography results and 285 healthy controls who were used as a referent group for the different studied parameters. Controls and patients with CAD were matched in age and gender (no sex-based or race/ethnicity-based differences were present).
We eventually included 472 consecutive CAD patients undergoing coronary angiography (angiographically defined MI) either in emergency with characteristic symptoms (typical prolonged chest pain), electrocardiographic changes or worsening symptoms during hospitalization in a period of two years from December 2014 to December 2016, from 3 hospitals: in the Cardiology department of Fattouma Bourguiba hospital in Monastir, Farhat Hached hospital in Sousse, and the regional hospital Mohamed Ben Sassi in Gabes, all are in Tunisia.
We collected the clinical data in a database, including a detailed medical history, sex, age, gender, weight and height (BMI), smoking, drug abuse and alcohol intake, the presence or absence of high blood pressure (HBP), as systolic blood pressure >140 mm Hg and/or diastolic blood pressure >90 mm Hg or treated with antihypertensive treatment, the presence or absence of diabetes mellitus or hyperlipidemia, family history of CAD and significant increase in biochemical markers of myocardial injury.
The MI diagnosis was formulated independently by the cardiologists depending on the following criteria: clinical symptoms, laboratory workup, electrocardiogram and a confirmed coronary angiography (a minimum of 50% stenosis in one major coronary vessel). The clinical severity classification has to be distinguished from angiographic complexity scores (atherosclerotic burden).
The case group was divided into 3 different subgroups according to the degree of complication: (a) patients with an ST-segment elevation MI (STEMI; n = 162), the most serious cardiovascular complication, that requires urgent medical care, (b) patients without ST-segment elevation acute MI (NSTEMI; n = 242), patients who also have a severe presentation requiring prompt medical care and (c) patients with stable angina (n = 67) being the most benign clinical presentation, but all explored by a coronary angiography. Exclusion criteria for the patient group were the presence of non-atherosclerotic vascular disease, cardiomyopathy, valvular disease, congenital heart disease, hepatic, renal and thyroid disease, autoimmune and infectious disease, and cancer. Patients under treatment with anti-inflammatory drugs or with any surgical proceeding in the previous 6 months were not included in this study. The medical treatment used for patients is RAASI (renin-angiotensin-aldosterone system inhibitors) either an ACE inhibitor or an angiotensin receptor blocker. Diabetes was treated with insulin or oral antidiabetic drugs. Lipid-lowering agents (statin/fibrate) were introduced just after CAD diagnosis using coronarography (1-7 days later).
285 volunteers included in the control group with no coronary arteriography abnormality and no cardiovascular disease (CVD) history, no echocardiographic regional wall motion abnormalities, no MI electrocardiographic signs, no indicative symptoms of CAD and no significant valvular abnormalities in echocardiograms. Subjects presenting any history of heart disease, severe illness limiting life expectancy, cardiomyopathy, or neoplasm, or refusal to consent were excluded.
The clinical research ethics committee of CHU Farhat Hached, Sousse approved the study protocol and total of procedures followed was in conformity with conventional guidelines (ethics committee for human research and conform to the 1975 Helsinki declaration ethical guidelines), and to participate in the study, the patients provided a written informed consent.

| Coronary angiography examination
In order to document the coronary stenosis, all patients underwent coronary angiography. Angiocoronarographic exams were performed and interpreted using Siemens ARTIS ZEE catheter laboratory and its dedicated QCA analysis software; they were routinely executed in at least two perpendicular projections, and follow-up angiography was performed using the same projections. The minimal luminal, per cent of stenosis diameter and lesion length were obtained. The presence of ≥50% stenosis in at least one coronary artery defines the CAD diagnosis. The coronary atherosclerosis burden was assessed using the number of stenosed coronary vessels (one, two or three-vessel disease) as well as the severity of the proximal lesions and left main coronary artery stenosis. Moreover, the absence (stent-) or presence (stent+) of stent implantation was recorded as well as its type, that is bare-metal (BMS) or drug-eluting (DES) stent. But, in our study the stent choice is irrelevant because it was never based on clear practice guidelines.

| Biochemical analysis
Venous blood samples were collected just before coronarography, while respecting 12-hour fasting, from all participants and collected into 5-mL tubes containing EDTA, Na-heparinate and no anticoagulant serum in function of dosed parameters. Serum and plasma obtained after centrifugation (3400 g, 4°C) were stored until assay at −80°C.
MMPs and TIMPs were measured in plasma heparin samples with ELISAs using the Human MMP-9 kit from Boster Biological Technology or the Human MMP-3, TIMP-1 and TIMP-2 kits from Clinisciences following the manufacturer's instructions, all including calibrators and quality controls. Each kit contains specific MMPs or TIMPs antibodies covalently bound to sets of microspheres, each having a unique spectral signature. Apo-E, Apo-CII and Apo-CIII were quantified with immunoturbidimetric methods in plasma heparin samples on Konelab 20XT from Thermofisher using the corresponding kits from Kamiya Biomedical Company including calibrators and quality controls. The samples were measured in duplicates and they were diluted or not according to manufacturer's instructions. The inter-assay and intra-assay coefficients of variation were less than 6%. Receiver operating characteristic (ROC) curves were selected from a homemade analysis workflow in order to find the best logistical regression model. Details are described below.

| Baseline characteristics and classical risk factors for cardiovascular diseases
The totality of 472 CAD subjects (mean age, 57.1 ± 11.4 years; 67% male) and 285 controls (mean age, 55.01 ± 13.82 years; 65.6% male) matched in terms of age and gender were enrolled in this study. The patient group included 34.5% STEMI, 51.3% NSTEMI and 14.2% stable angina subjects. After coronary angiography examination, depending on the number of touched vessel, patients with CAD manifested either one (33.5%), two (27.8%) or three (38.8%) vessel disease. Regarding the therapeutic procedures, a total of 42.4% of patients were treated with a bare-metal stent (BMS), 23.7% with a drug-eluting stent (DES), and for 33.9% of patients no stenting was performed. Compared to the control group, patients with CAD showed a significant increase in the percentage of peripheral artery disease history (P < 0.001), a family history of CAD (P < 0.001), diabetes mellitus (P < 0.001), hypertension (P < 0.001) and had more elevated BMI (P < 0.001). In addition, statistically significant discrepancies were shown in CAD patients when compared to controls in terms of higher consumption of alcohol and tobacco (P < 0.001) ( Table 1). Table 2 shows higher statistically significant plasma levels for most of the biochemical (lipids) parameters in CAD patients compared to controls (P < 0.001) ( Table 2). Concerning the patient treatments, Table 3 shows no statistical variation between CAD patient subgroups when comparing the percentage of different subgroups CAD patients (STEMI, NSTEMI, stable angina) under different medical treatments; so treatment had no influence.

| Association of different biomarkers with CAD severity, number of touched vessel and stent treatment
MMP-3 and -9 have higher plasma levels in patient subgroups than in the control group, the highest MMP-3 and-9 levels were obtained in the most severe presentation of CAD (STEMI) compared to NSTEMI and stable angina patients (P < 0.0001, for both). Inversely, TIMP-1 and TIMP-2 levels were increased in controls compared to the three CAD subgroups of severity: controls vs stable angina (P < 0.0001), controls vs NSTEMI (P < 0.0001), controls vs STEMI (P < 0.0001), But the  Figure 1A-C).
Significant higher plasma MMP-3 and -9 levels were found in CAD patients with two-or three-vessel disease compared to controls or patients with one vessel disease (P < 0.0001). TIMP-1 and TIMP-2 levels were lower in all patient subgroups compared to controls (P < 0.0001), but not statistically different between them.
Apo-CII, -CIII and -E plasma levels were significantly increased in all subgroups compared to controls (P < 0.0001 for Apo-CII and -III, P < 0.001 for Apo-E) but not statistically different between them, but for Apo-CII slightly lower in the three-vessel disease subgroup compared to those with one vessel disease (P < 0.01; Figure 2A-C).
MMP-3 and -9 plasma levels were more elevated in patients treated with stents than without (P < 0.0001), and ever more in patients treated with drug-eluting stent than with bare-metal stent (BMS) (P < 0.001). TIMPs levels were lowest in all subgroups but without any statistical difference according to stent treatment choices. Apo-CII, -CIII and -E plasma levels were significantly higher in all subgroups than in controls (P < 0.0001), but for Apo-E slightly in both stent treatment subgroups (P < 0.01), and for Apo-CII and -III even higher in 'no stent treatment' subgroup (from P < 0.001 to P < 0.0001), and without any statistical difference between DES and BMS treatment subgroups ( Figure 3A-C). represents the patient group too with mostly higher levels of apolipoproteins and without elevation of MMP-3, -9 and TIMP-1, -2 levels, whereas control group (i.e. cluster 1) had higher TIMPs levels, lower MMPs and apolipoproteins levels. In cluster 1, we distinguish a small sub-cluster (red framed) with higher TIMP-1 levels, lower levels of metalloproteinases and higher Apo-E ( Figure 4).

| ROC Curve analysis: panel performance study
In order to find the minimum dosage variable that classes each patient with the minimum misclassifications, we decided to test all combinations of dosage variables to find the best models. We was found for the "Apo-CIII + TIMP-2 + MMP-9" model ( Fig 5A).

| D ISCUSS I ON
The In our study, MMP-9, MMP-3 blood levels are increased in CAD patients compared to healthy controls, these two biomarkers have a relevant positive correlation between them, also they are higher F I G U R E 2 A, MMPs levels variation between different number of stenosed vessels in patient subgroups and controls. Number of stenosed Vessels, C: Controls, 0: one stenosed vessel, 1: two stenosed vessels, 2: Three stenosed vessels, n.s: non-significant, *P < .01, **P < .001, ***P < .0001. B, TIMPs levels variation between different number of stenosed vessels in patient subgroups and controls. Number of stenosed Vessels, C: Controls, 0: one stenosed vessel, 1: two stenosed vessels, 2: three stenosed vessels, n.s: non-significant, *P < .01, **P < .001, ***P < .0001. C, Apolipoproteins levels variation between different number of stenosed vessels in patient subgroups and controls. Number of stenosed Vessels, C: Controls, 0: one stenosed vessel, 1: two stenosed vessels, 2:Three stenosed vessels, n.s: nonsignificant, *P < .01, **P < .001, ***P < .0001 MMPs levels, particularly MMP-9, are higher in subjects with acute coronary syndromes (ACS) and contributed to the molecular process of plaque destabilization. 16 Higher levels were shown for MMP-3 in patients with triple vessel disease. 17 In patients with and without ACS, an elevation of MMP-3 was related to an increased risk of CAD. 18 A correlation was found between MMP-3 and all mortality causes. 19 Similarly, a higher level of MMP-9 was found in patients with CAD 20,21 and unstable unruptured plaque, in ruptured plaque as well as in unstable plaque than in the stable plaque in another study. 22 Nishiguchi's 23 study showed a higher MMP-9 level in post stent STEMI patients, additionally in triple vessel stenosis patients. 24 Furthermore, patients with multiple sites of in-stent restenosis had higher MMP-9 levels. A previous study showed an elevated MMP-3 level in patients with arterial hypertension. 25 However, decreased MMP-3 levels were shown in stable and unstable CAD patients. 26 Most of these authors suggested that MMP-9 is a functional biomarker in CAD. A similar Tunisian study found only MMP-3 which could be used as a cardiovascular events predictor due to mostly for their endpoint study, the smaller population study (patients/ controls), patient group subdivision and the difference of cities sampling. 27 MMP-3 regulates vascular smooth muscle cell migration via MMP-9 activation and the MMP-3 expression and secretion may precede MMP-9 expression and secretion. 28 As shown in previous studies, MMP-9 would be related to the stability of atherosclerotic plaques, so a higher level may indicate that plaques are vulnerable or have ruptured. We suggested that, MMP-3, MMP-9 was related F I G U R E 3 A, MMPs levels variation between different stent treatments in patient subgroups and controls. Stent treatment, C: Controls, 0: No stent treatment, 1: Bare-metal stent, 2: Drug elution stent n.s: non-significant, *P < .01, **P < .001, ***P < .0001. B, TIMPs levels variation between different stent treatment in patient subgroups and controls. Stent treatment, C: Controls, 0: No stent treatment, 1: Bare-metal stent, 2: Drug elution stent n.s: non-significant, *P < .01, **P < .001, ***P < .0001. C, Apolipoproteins levels variation between different stent treatments in patient subgroups and controls. Stent treatment, C: Controls, 0: No stent treatment, 1: Bare-metal stent, 2: Drug elution stent, n.s: non-significant, *P < .01, **P < .001, ***P < .0001 MMP activity and increasing lesion stability. 36 It was suggested that a change in MMP/TIMP balance could have an importance in hypertensive heart disease. 37 Another author found that patients with CAD had higher MMP-9 and TIMP-1 plasma levels but a lower TIMP-2 level, 38 also both increased MMP-9 and TIMP-1 plasma levels in premature coronary atherosclerosis patients, but with lower MMP-3 and TIMP-2 levels. 39 Notably, increased plasma levels of both MMP-9 and TIMP-1 were identified in patients with CAD, 40 as well as in patients with angina or MI. 41 Higher plasma MMP-9 and TIMP-1 levels were also found in NSTEMI. 42 Giagtzidis's team has found that patients with peripheral arterial disease undergoing angioplasty/stenting had MMP-3 elevation, but not for MMP-9, TIMP-1 and -2. 43 Moreover, MMP-9 levels were higher in ACS than in stable angina, but TIMP-1 showed no difference. 44 MMP-9 level was increased in STEMI patients also in smokers and with no difference for TIMP-1, and with no correlation between MMPs and TIMPs. 45 Apo-CII and Apo-CIII were significantly increased in CAD patients compared to controls; they showed a statistically significant variation when comparing different patient subgroups and they were in positive correlation with MMPs and in negative correlation with TIMPs. It is the first study which investigates this correlation and we suggested a synergic action between MMPs and apolipoproteins in cardiovascular complications. In previous studies, Apo-CIII and Apo-E have been associated with a higher risk of diabetes 46 ; their F I G U R E 4 Heatmap showing study biomarkers and the associated discrimination between patient and control groups blood levels were increased in chronic heart disease (CHD) in relation to dyslipoproteinaemia 47 and they were elevated in CAD. 48 High Apo-CIII level was found in CAD patients, 49 as well as in patients with familial hyperlipidemia and cardiovascular disease. 50 Apo-CII and Apo-E were higher in function of stenosis progression. 51 Apo-CII and -CIII are considered as endogenous mediators of lipoprotein lipase activity and affect the concentration of triglycerides in blood which leads to aggravation of aortic atherosclerosis, thereby increasing the inflammatory response. 10 In discordance to our study, lower levels of Apo-CII, Apo-CIII and Apo-E were described in CAD patients. 52,53 Moreover, they showed no significant differences in peripheral vascular disease, 54 and a lower Apo-CII was found in STEMI patients 55 as well as depending on CAD severity with the number of stenosed vessels. 56   vulnerable plaque, and it was always better than TIMP-1. The choice of an Apo-C is more difficult as Apo-CII was a better marker than Apo-CIII for the number of stenosed vessels, but inversely Apo-CIII was better than Apo-CII for the type of stent treatment.
Apo-E never was discriminative in function of CAD aggravation and was discarded from our 'CAD aggravation panel', finally combining MMP-9, TIMP-2 and Apo-CIII with best AUC.
Our study is the first looking for the relation between MMPs, TIMPs and apolipoproteins; we support that this panel might be helpful to more detailed risk stratification in patients with CAD, helpful to predict cardiovascular disease and therapeutic procedure.

| CON CLUS IONS
Based on our findings and supportive evidence in the literature, we However, our study has some limitations such as the size of the studied population, their geographical localization and the follow-up period. The descriptive clinical study is used to draw hypothesis which must obviously be confirmed in animal or cellular models.

ACK N OWLED G EM ENTS
We thank Dr Rami Alzebdieh and Dr Modar Abdullatif for their helpful comments. We thank engineer Claudine Delom énie from UMS IPSIT.

CO N FLI C T O F I NTE R E S T
The authors whose names are enrolled in this study proclaim that there is no disagreement of interest, and declare that they have no association with or connection in any organization or entity with financial or non-financial relevance (such as knowledge or beliefs, affiliations, professional or personal relationships) in the subject issue or materials discussed in the article. There is no conflict of interest between authors.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from the corresponding author upon reasonable request.