OUP user menu

Atherosclerosis and proteinase activation

Clare M. Dollery, Peter Libby
DOI: http://dx.doi.org/10.1016/j.cardiores.2005.11.003 625-635 First published online: 15 February 2006


Rapidly accumulating evidence points to the matrix metalloproteinases (MMPs) as major molecular mediators of arterial diseases. Findings from human pathological specimens, animals, and cell and molecular biology implicate matrix metalloproteinases in all stages of atherosclerosis including lesion initiation and progression and ultimately in plaque complication and triggering of thrombosis. The complex interactions within the proteolytic cascade allow multiple levels of control over these functions. This review weighs the evidence for the role of MMPs in arterial biology with particular reference to their activation in the atherosclerotic plaque.

  • Atherosclerosis
  • Metalloproteinase
  • Plaque
  • Matrix

1. Introduction

Cardiovascular disease already constitutes the leading cause of death in Europe (4.35 million per year) and the United States (2600 per day) and likely will become the leading cause of death worldwide [1,2]. Atherosclerosis causes about half of these deaths from coronary heart disease and one third from stroke [1]. In face of this health burden, researchers have sought a greater understanding of the mechanisms of atheromatous vascular disease.

It is now 14 years since pivotal work by Henney et al. furnished evidence for expression of the mRNA encoding stromelysin (MMP-3), a proteinase formerly associated with cancer metastasis, in the human atherosclerotic plaque [3]. Since then the perception of atheroma as an inert obstruction in the vessel has given way to the concept of inflammation as a pivotal mechanism underlying atherosclerosis. Atheromatous plaques depend on recruitment of inflammatory cells, such as monocyte-derived macrophages, into the vessel wall and production of proinflammatory mediators such as interleukin-1 (IL-1), tumour necrosis factor-α (TNF-α), and CD40 ligand [4]. These signals encourage smooth muscle cell (SMC) accumulation, lipid deposition, and the production and activation of a complex cascade of proteinases that alter the structure of plaques and affected arteries. In the early stages of atheroma, MMPs may facilitate migration of both SMC and monocyte-macrophages. When focused at the fibrous cap or shoulders of advanced plaques proteinases can promote rupture exposing the thrombogenic core to circulating blood and provoke acute coronary syndromes or silent rupture and lesion progression. Here we review the evidence supporting the role of MMPs in atheroma. For detailed summary tables listing the various MMPs, their localisation, structures, and substrates, the reader is referred to several other reviews in this issue (articles by Nagase et al., Vanhoutte et al., and Newby).

2 Proteinases in human atheroma

2.1 Proteinases in the atherosclerotic plaque

Much of our knowledge of proteinases in human atherosclerotic plaques comes from studies of surgical carotid endarterectomy specimens and transcatheter coronary atherectomy samples. Fewer studies involve post mortem tissue or explanted hearts from patients undergoing transplantation for ischaemic cardiomyopathy. Proteinases also participate in vascular response to mechanical injury but here we weigh their role in atheroma. This review will focus on MMPs and consider their interaction with other proteinases present in atheroma.

2.1.1 The metalloproteinases

Henney et al. used in situ hybridization to localise MMP-3 to SMC and particularly to lipid-laden macrophages in human plaques. They first suggested that the concentration of proteinase-rich foam cells at the shoulder of the plaque might destabilize the plaque. Vine and Powell then identified MMP-9 as the major source of gelatinolytic activity in dissected layers of aortic lesions, seeing only relatively sparse MMP-3 [5]. Our group characterized the expression and function of the interstitial collagenase MMP-1, as well as MMP-2, MMP-3, MMP-9 and tissue inhibitors of MMPs (TIMP)-1 and-2 (see Section 2.1.5), revealing shoulder regions of carotid plaques with abundant foam cells, in situ proteolysis and increased expression of MMPs-9, -3, and -1 [6,7].

Nikkari et al. subsequently localized MMP-1 with intraplaque haemorrhage, a histological marker of instability, while our group identified MMP-1 in areas of increased wall stress in human carotid atheroma [8,9]. MMP-9 localises in human coronary atherectomy tissue. Further, Formato et al. correlated levels of MMP-2 and MMP-9 with inflammatory mediators IL-6 and IL-8 in carotid artery extracts. Higher levels of MMP-9 occurred in plaques histologically classified as unstable while local proteolysis of apolipoproteins suggested that these proteinases are active [10]. Other groups have colocalised MMP-9 to inflammatory infiltrates in the carotid and found higher MMP-9 levels in patients within 1 month of symptomatic embolic stroke compared to those without recent symptoms [11]. The same group showed that prior statin therapy lowered intraplaque MMP-1, MMP-9, and IL-6 confirming, in human plaque, concepts concordant with the results of our rabbit studies [12–14].

Collagen, a major structural component of the atherosclerotic lesion, provides biomechanical strength for fibrous plaques. Destruction of interstitial collagen destabilizes atheromatous plaques. The collagenases (MMP-1, -8 and -13) share the unusual ability to initiate collagenolysis by cleaving intact triple helical collagen. Our group showed MMP-1 and MMP-13 colocalised with cleaved collagen fragments in situ in human lesions (Fig. 1) [15]. MMP-8 was overlooked in atheroma because of its traditional association with neutrophils; however, transcriptional profiling reveals CD40 ligand-dependent production of MMP-8 by macrophages and its association with collagen cleavage in situ in carotid atheromatous plaques [16].

Fig. 1

Increased collagenolysis in human atheromatous plaques colocalises with MMP interstitial collagenases. Immunofluorescence using antibodies that detect all type I collagen (top middle panel–Texas red) or only recognize a neo-epitope created by MMP-mediated cleavage (top left hand panel–Fluorescein green) show distinct areas of collagen breakdown which co-localizes with immunostaining for MMP-1 (Collagenase I), MMP13 (Collagenase III), and macrophage CD68 (lower panels). (from Circulation 1999; 99:2503–9).

Although a plethora of MMPs concentrate in inflammatory cells at the plaque shoulders, patterns of MMP expression vary, e.g. MMP-7 and -12 occur along the perimeter of the cellular lipid core adjacent to but not within the cells of the fibrous cap. This localization suggests a demarcation zone along the fibrous cap that might peel away from the lipid core during rupture, promoting thrombosis and infarction [17].

Membrane-type MMPs (MT-MMP) form a sub-group of MMPs characterized by an anchoring transmembrane domain that focuses proteolysis at the cell surface and resist inhibition by some TIMPs [18,19]. MT1-MMP (MMP-14) localises in both normal and atheromatous human coronary arteries at autopsy although medial SMC expression of MMP-14 is more pronounced in normal vessels [20]. In plaque macrophages MMP-14 colocalises with MMP-2, and activates it by proteolytic cleavage, suggesting functional significance despite low abundance [20]. SMC in normal coronary arteries express MT3-MMP (MMP-16) as do plaque macrophages [21].

Unusually for the MMP family, MMP-11 (stromelysin-3) is secreted as an active proteinase after intracellular propeptide cleavage by furins in the Golgi apparatus. MMP-11 has little direct matrix-degrading ability but can degrade serpins, relieving serine proteinases and some cathepsins from inhibition and promoting proteolysis within the plaque. MMP-11 abounds in SMC, endothelial cells (EC) and macrophages in carotid plaques and is regulated by CD40 ligation [22]. The case of MMP-11 illustrates the interrelation between seemingly distinct groups of proteinases and overlapping specificities in a complex cascade.

2.1.2 The ADAMs (A Disintegrin and Metalloproteinase)

Only two disintegrin metalloproteinases (ADAMs) are known in human atheroma, ADAM-9 and -15, and they have unclear function. ADAMs have an integral transmembrane domain and may thus act at the cell surface to assist migration.

2.1.3 The cysteine proteinases

Breakdown of elastin, a structural component of the arterial wall, may have importance in expansive and occlusive remodeling. Our group found SMC-derived cathepsins S and K, both potent elastases, in human atheroma and showed localization to breaks in the elastic laminae [23]. Interestingly, normal vessels contained little or no cathepsins while fatty streaks, thought to be an early stage in atheroma, had widely distributed cathepsin S and K. More advanced atheroma macrophages also express these proteinases which degrade elastin and collagen [24]. Additionally, plaque macrophages contain cathepsins B and D [25]. Oorni et al. have now identified cathepsin F in shoulder macrophages of advanced coronary atheroma and have shown that cathepsins S, K, and F can degrade apolipoprotein B [26]. In vitro, this yields release of lipid droplets from low density lipoprotein (LDL) particles, implicating cathepsins not only in elastolysis but also in the genesis of the plaque's lipid core [26].

2.1.4 The serine proteinases

The serine proteinases most common in atheroma belong to the fibrinolytic cascade, where urokinase-type plasminogen activator (u-PA) and tissue plasminogen activator (t-PA) activate plasminogen to plasmin, which in turn breaks down fibrin to its degradation products. U-PA is itself subject to proteolytic processing. Its single chain form binds to its surface receptor u-PAR, thus focusing more potent proteolysis in the adjacent microenvironment. In atherosclerotic aortae and carotid arteries, tPA and uPA occur in intimal SMC and in macrophage-derived foam cells [27]. Macrophages localised on the necrotic core margin express particularly high levels of uPA, while both tPA and uPA abound in the neomicrovessels of plaques, suggesting a role in plaque angiogenesis. Raghunath et al. found u-PAR, tPA and uPA in the coronary plaque shoulders of explanted hearts and described widespread plasminogen activator inhibitor (PAI-1) throughout the plaque [28]. Unlike some MMPs, the plasmin/plasminogen cascade does not directly degrade collagen or its breakdown products but can degrade laminin, fibronectin, and proteoglycans. Their predominant role in the plaque may be activation of MMPs [29–32].

We recently described neutrophil elastase in macrophages in the shoulders of human carotid plaques and showed its regulation by CD40 ligand [33]. The potent serine proteinase, neutrophil elastase activates MMPs and inactivates TIMP-1 [34–37]. A putative endogenous vascular elastase may participate in pulmonary hypertensive vasculopathy but is not described in human atheroma [38].

Mast cell serine proteinases, chymase and tryptase also localise in the fibrous cap and shoulder regions of atherosclerotic plaques, with a preponderance of activated mast cells in this key region [39,40]. Chymase and tryptase can degrade type IV collagen and fibronectin and also activate MMPs e.g. MMP-1, -2, and -9 [41,42]. Johnson et al. demonstrated chymase- and tryptase-dependent MMP-1 activation in situ in human atherosclerotic plaques [40].

2.1.5 The proteinase inhibitors

The surfeit of proteinase activity described above would predict rapid arterial expansion, plaque disruption, and adherent thrombotic events. The chronic indolent nature of atherosclerosis in face of this plethora of proteinases probably relates in part to the local opposing action of proteinase inhibitors. Of the four known TIMPs, at least three occur in human atheroma. Neointimal and medial SMC in normal vessels express TIMP-1 and -2. However, in atheroma TIMP-1, -2, and -3 cluster in shoulder regions and macrophages overlying the necrotic core, mirroring the expression of the MMPs they inhibit [6,43,44]. PAI-1, the major inhibitor of the plasmin system, distributes throughout the plaque, particularly in diabetic subjects but is more sparse in key areas of inflammatory macrophage infiltration at the plaque shoulders [28]. Macrophages synthesize α-1 proteinase inhibitor, the cognate inhibitor of neutrophil elastase, in atheroma [45]. However, plaque MMPs may inactivate this inhibitor as well as the mast cell proteinase inhibitor α-1 antichymotrypsin [46]. This example illustrates another feedback loop within the proteolytic cascade. The cysteine proteinase inhibitor, cystatin C, localises in the normal vasculature but is sparse in atherosclerotic plaques, likely boosting net cysteinyl proteinase action in plaques [47].

2.2 Regulation of MMPs in the vasculature

MMPs are tightly regulated at three levels–transcription, proenzyme activation, and inhibition [48]. These aspects of proteinase regulation doubtless influence the biology of atheromata.

2.2.1 Transcriptional regulation

Cells within the plaque elaborate proinflammatory cytokines, among these interferon-γ, TNF-α, and IL1β. IFNγ decreases collagen synthesis by vascular SMC [49]. SMC and EC basally express MMP-2 and show little regulation of TIMPs 1 and 2 [44,50,51]. IL1β and TNF-α induce MMPs-1, -3, -8 and -9 in both SMC and EC [51–53]. Mechanical stretch stimulates ECs and SMC to produce MMP-2 and may reflect events in hypertension [54–56]. Platelet-derived growth factor (PDGF) and fibroblast growth factor (FGF) induce MMP-9 and MMP-1, respectively, in SMC [52,57]. Oxidized LDL, a possible promoter of atherogenesis, induces MMP-1 in EC, MT1-MMP in SMC, and MT3-MMP in macrophages [20,21,58].

Binding of CD154 on T cells to CD40 on plaque EC, SMC and macrophages potently stimulates inflammation and the induction of MMPs-1, -2, -3, -8, -9 and -11 [16,22,59]. Macrophages probably provide the bulk of MMPs within the plaque [60]. IL-1 and TNF-α cause selective induction of MMP-9 in monocyte-derived macrophages and do not effect the major collagenases, MMP-1 and -3 [61]. However, lipid-laden macrophages elaborate more MMP-1 and -3 than alveolar macrophages [62]. Exposure to oxidized LDL may also enhance activation of constitutive MMP-2 via MT3-MMP induction [21].

Numerous stimuli boost production of MMPs within the cells of the plaque–few of the same stimuli induce TIMPs. Both SMC and ECs constitutively express TIMP-1 and -2, macrophages in atheromata make most TIMP 1–3 [51,52]. PDGF and transforming growth factor β induce TIMP-1 and -3 in SMC [63]. The cytokines and inflammatory cells in the plaque can augment production of MMPs out of proportion to levels of their inhibitors, swinging the balance in favor of proteolysis [51].

2.2.2 Proenzyme activation

Almost all MMPs arise as zymogens requiring activation; thus, MMP protein does not equate with enzymatic activity. Multiple pathways of activation exist within the MMPs and by other proteinases or factors. The activation of MMP-2 typifies the complexity of this system (Fig. 2) [64].

Fig. 2

Activation of MMP-2. The most potent means of activation of MMP-2 may depend on interaction with complexes of MT1-MMP (MMP-14) and TIMP2. MT1-MMP is activated by furins in the Golgi and translocates to the cell membrane where it is inhibited by TIMP-2 forming a complex. ProMMP-2 then binds to TIMP-2 via its hemopexin domain (see insert bottom right for scheme illustrating aspects of MMP structure). This leaves the proMMP-2 molecule vulnerable to cleavage by adjacent active molecules of MT1-MMP, which partially activates MMP-2. Other active MMP-2 molecules accomplish the final cleavage of the propeptide. Other mechanisms that involve serine proteinases or reactive oxygen species can also activate MMP-2 (as shown in the upper figure). Red arrows show steps which produce active MMP-2 while blue arrows indicate steps in the pathway. The expert review of Mott and Werb forms the inspiration for the depiction of MT1-MMP–TIMP-2 complex activation mechanism [64].

The main mechanism of proMMP-2 activation at the cell surface involves binding of its inhibitor TIMP-2 to the membrane-associated MT1-MMP. This docking inhibits MT1-MMP but allows TIMP-2 to anchor proMMP-2 by binding its non-catalytic haemopexin domain. Thus, proMMP-2 localizes on the cell membrane and can be cleaved by an adjacent MT1-MMP molecule. Finally, another MMP-2 molecule removes the remaining pro domain [65]. Use of the MT1-MMP–TIMP-2 complex as an MMP-2 surface receptor recruits TIMP-2 as a critical part of MMP-2 activation, establishing the paradox that TIMP-2, an MMP inhibitor, can enhance MMP activity [66]. Other proteinases present in the plaque also activate MMP-2, e.g. plasmin (in the presence of MT1-MMP), neutrophil elastase, cathepsin G, thrombin, and mast cell chymases and tryptases [42,67,68].

Not only proteinases mediate MMP activation; foam cell-derived reactive oxygen species can activate pro MMP-2, a pathway of potential pathological relevance (Fig. 3) [69]. Serine proteinases, such as neutrophil elastase, may also favour matrix breakdown by inactivating TIMPs [35]. In turn, MMPs can inactivate serine proteinase inhibitors α-1 proteinase inhibitor (MMP-9 and -8) and α-1 antichymotrypsin (MMP-8 activated by MT1-MMP) [70,71]. These mechanisms may not all occur in atheroma but they illustrate the redundancy available within this proteolytic cascade and hint at the difficulty of switching off proteolysis.

Fig. 3

MMPs activate MMPs. The cascade of MMPs that activate themselves or each other is shown. Using the scheme from Fig. 2 processing from inactive zymogens (blue catalytic site) to activate proteinases (pink catalytic site) is illustrated. Red arrows indicate activation steps and black arrows show the MMPs contributing to this process by cleavage of the pro domain.

Other MMPs, such as MMP-9, depend predominantly on plasmin for activation, while upstream activation of plasmin may depend more on uPA than tPA. MMPs also activate each other, in particular MMP-3 activates a number of MMPs including MMP-9 and itself (Fig. 3)[72]. MMP-7 is activated both by MMP-3 and by hypochlorous acid, a product of myeloperoxidase found in plaque macrophages, and MMP-7 in turn can activate MMP-1 [73]. Figs. 2 and 3 illustrate some of the complex interactions which occur within the MMP family and with other proteinases.

2.2.3 Specific inhibitors

TIMPs, naturally occurring specific inhibitors, limit MMP activity (see above), as do the less specific inhibitors α2-macroglobulin and exogenous substances such as heparin [74]. Alpha2-macroglobulin's large size limits its contribution. TIMPs are secreted multifunctional proteins that regulate connective tissue metabolism. The family includes at least four members [63,75]. TIMP-1 and TIMP-2 share only 42% amino acid identity, but overlap considerably in their ability to inhibit MMPs; however, they interact differently with the progelatinases [65]. Unlike other TIMPs, TIMP-3 binds to the extracellular matrix and has emerging roles in apoptosis and inflammation (the latter via inhibition of TNF-α activation) [76,77]. Vascular injury regulates TIMP-4 in rodents but has not been described in human atheroma [78].

3 What do MMPs do in the vasculature?

3.1 Evidence from isolated cells

The family of MMPs can degrade most components of the vasculature including collagen and elastin, functions that contribute to both constrictive and expansive remodeling [65,79]. Although MMPs may permit cells to pass through basement membrane barriers, these enzymes might have a more active role than simply degrading the structural components of the vessel wall.

Migration of cells through the extracellular matrix likely represents a major function of the MMPs. TIMPs 1–4 delivered directly or by gene transfer significantly inhibit migration of SMC in vitro [78,80–82]. More recent studies of cells isolated from MMP-deficient mice implicate MMPs-2, -9, and -8 in migration of SMC and macrophages through a variety of substrates [83–85]. MMPs not only facilitate migration by focusing proteolysis but also directly enhance migration. In epithelial cells, cleavage of laminin 5 by MMP-2 reveals a cryptic site that stimulates migration [86]. Recent findings have shown MMP-1 promotes growth and invasion of cells by binding to the proteinase activated receptor-1. Cleavage of this receptor reveals a tethered ligand that initiates signaling via this G protein-coupled receptor and activates migration [87]. In essence, this mechanism allows the cell to sense a proteolytic environment and actively move towards an area of degraded matrix. Although described in the context of tumourigenesis, human macrophages and SMC in atheroma express the thrombin receptor PAR-1 [88]. The action of many MMPs focuses at the cell surface, allowing tight control of cell behaviour rather than more haphazard destruction of matrix by diffusion of secreted proteinases [89].

MMPs also liberate growth factors that are stored in the extracellular matrix such as transforming growth factor β (TGFβ) and vascular endothelial growth factor (VEGF). MMP-9 releases VEGF bound to proteoglycans in the ECM, enhancing its bioavailablity, which may influence plaque neovascularisation [90]. MMPs-2, -9, -13 and -14 can activate TGFβ stored in the ECM while MMP-2 releases TGFβ from the matrix [64]. Plasmin generated by uPA also activates TGFβ in human vein SMC [91].

3.2 Evidence for the function of MMPs from atherosclerotic and injured animals

Hyperlipidaemic mice and rabbits provide useful preparations for studying atheroma, and with vascular injury, assessing the processes of acute remodeling. Atherosclerotic rabbits have helped define our understanding of pathological findings in human plaques. Our studies of rabbits revealed expression of MMPs-1, -2, -3, and -9 in lesions produced by lipid-rich diet with balloon injury, while others have shown a partial compensation by augmented TIMPs [14,92]. Importantly, shifting to a low-fat diet reduces plaque proteolysis, while MMP-1 levels and macrophage content progressively decline [14]. Further, administration of HMG-CoA reductase inhibitors can reduce macrophage accumulation and MMPs in rabbits [13], linking MMPs and the proteolytic state within the plaque to established treatments which influence outcomes after acute coronary syndromes [93].

Mice provide a convenient model of atherosclerosis because of the relative ease of genetic manipulation. Their small size, lack of spontaneous plaque formation, and absence or low abundance of MMP-1 do not reflect the human condition. Most studies employ mice deficient in apolipoprotein E or the LDL receptor that combined with dietary manipulation causes hypercholesterolemia and formation of atheroma in the aorta and to a lesser extent the coronary arteries[94]. Carmeliet and coworkers crossed ApoE-deficient mice with uPA- or tPA-deficient mice and found that uPA deficiency impaired development of aortic ectasia probably via reduced activation of MMP-12 [30]. They also described a protective role of PAI-1 in atheroma development. ApoE-deficient mice crossed with PAI-1-deficient mice show more disorganization and degradation of collagen and complex plaques–possibly via enhanced activation of MMPs by plasmin released from the control of its inhibitor [95]. Bone marrow transplantation showed PAI-1 was from SMC in the recipient vessel wall rather than circulating monocyte-derived macrophages.

D'Armiento and colleagues exploited the lack of endogenous MMP-1 in mice to study overexpression of human MMP-1 in macrophages in ApoE-deficient mice. Unexpectedly the mice with MMP-1-producing macrophages showed smaller plaques than controls and no evidence of plaque rupture [96]. These mice have altered MMP-1 from birth, which could reduce collagen accumulation; thus, MMPs may be more critical in destabilization of established plaques than in atherogenesis [97]. MMP-3 also appears to have a dual role. While mice lacking both MMP-3 and ApoE show reduced aneurysm formation, they have more extensive atheroma [98]. Absence of MMP-3 causes increased collagen and fewer plaque macrophages–a characteristic associated with greater stability in human plaques. These observations suggest that MMPs may attenuate atherogenesis but do not exclude an adverse role in plaque rupture.

Two groups have studied TIMP-1 deficiency in ApoE-deficient mice, and both found that TIMP-1 deficiency produced macrophage-rich lesions with active proteinases and medial destruction [99,100]. Lesions in TIMP-1-deficient mice were 30% smaller than controls in Silence's study [100] and the same as controls in LeMaitre's [99]. Studies using pharmacological MMP inhibitors did not show any effect on lesion size in atheroma-prone mice [101,102]. Adenoviral gene transfer of TIMP-1 into ApoE mice 6 weeks after commencing a high-fat diet reduced both lesion size and macrophage content, supporting the prevailing concept that MMPs adversely influence established plaques [103].

ApoE−/− mice deficient in either MMP-9 or MMP-12 show reduced medial destruction, while MMP-9 deficiency but not lack of MMP-12 reduced lesion size [104]. After carotid ligation, hypercholesterolaemic MMP-9-deficient mice show a reduced plaque burden and less cellular lesions than MMP-9+/+ animals [105]. Thus, MMP-9 may promote atherogenesis.

Collagenase-resistant mice provide a novel approach to the redundancy that may influence mouse models of MMP deficiency in atheroma. Rather than a specific proteinase defect, these mice have genetically modified collagen that resists digestion by all MMP collagenases. In an atherogenic background the lesion size of collagenase-resistant mice was similar in size to controls, but SMC number in the intimal lesions decreased and collagen was more abundant [106]. These findings implicate MMP-mediated collagenolysis in SMC accumulation in the atheromatous plaque and in collagen homeostasis in the vessel wall.

Plaque rupture is rare in hypercholesterolaemic mice, and many have sought a useful model of this process [94]. A model that garnered recent attention involves analysis of breaks in the elastic laminae of plaques in the brachiocephalic artery [107,108]. These experiments differ from those discussed above in the artery studied, diet that contains 21% lard, and the mixed Sv129 and C57Bl6 strain. This strain absorbs more cholesterol and has a lower platelet TGFβ concentration than the commonly used C57Bl6 strain [109]. Importantly, plaques defined as ruptured in this model were larger and more likely to be in an occluded vessel while buried elastic laminae within the plaques suggest that multiple ruptures might underlie atheroma progression-consistent with our current concept of human atheroma. In a recent study, pravastatin treatment of these mice did not reduce the incidence of plaque rupture after 9 weeks. However, 40 weeks of statin therapy produced a marked reduction from 37% to 5% (pravastatin did not change cholesterol levels at the early time point) [110]. As in rabbits, pravastatin reduced atheroma area at both 9 and 40 weeks. Early or unpublished observations suggest MMP-12 promotes the development of brachiocephalic lesions, while, in contrast to previous results, MMP-9 protects against plaque rupture, and TIMP-1 may have no effect [97]. Predictably, no single animal preparation reflects all aspects of human disease.

Another emerging avenue of research involves the use of molecular imaging of MMP function in vivo. This field is in its infancy but in murine tumours a near infrared fluorescent MMP substrate detects the activity of MMPs in vivo, and activity decreases after intervention with an MMP inhibitor [111]. Similar approaches have identified active cathepsin B in atherosclerotic mice and the evolution of MMP-2 and -9 activities after murine myocardial infarction [112,113]. Although this technology presents challenges, such as limited tissue penetration and autofluorescence of elastin in arteries, initial results have provided proof of principle of proteinase imaging in vivo.

3.3 Genetic studies and clinical trials of inhibitors

As MMPs act locally in regions of inflammation or at the cell surface, and given the abundance of proteinase inhibitors in serum, measurement of circulating MMP levels may poorly reflect MMP function in compartments critical to atherosclerosis. Genetic studies seeking a systemic predisposition to proteolysis may help to reveal the roles of MMPs. Henney and colleagues explored this concept when they studied a functional 5A/6A polymorphism in the MMP-3 promoter in patients from a 3-year atherosclerosis regression study [114,115]. The 6A active variant correlates with progression of luminal narrowing [115] and associates with acute myocardial infarction in a case control study [116]. The MMP-12 gene shows a functional polymorphism (A to G at position 82). In a relatively small study, the A allele, which enhances promoter activity, correlated with smaller lumen diameter in diabetic patients but not in the wider group. The more active T allele of an MMP-9 functional promoter polymorphism (C to T at −1562) is commoner in patients with three vessel disease in a subgroup of the ECTIM study of acute myocardial infarction but does not predict myocardial infarction [117]. A more powerful study of 1127 patients from Blankenburg et al. showed higher MMP-9 serum levels associated with the T allele but it did not predict cardiovascular death and serum MMP-9 did not independently predict cardiovascular risk [118]. The extensive redundancy in the proteolytic cascades may permit compensation to prevent the effects of excess of one MMP. Interestingly less active polymorphisms of the serine proteinase inhibitor α-1 antiproteinase promoter correlate with atheroma progression in two studies [119].

Enthusiasm for cardiovascular trials of MMP inhibitors waned after cancer trials showed side effects such as tendonitis (possibly related to inhibition of ADAMs) combined with a lack of efficacy and in one trial the possibility of harm [120]. These MMP inhibitors have not been used clinically in atherosclerosis, and most animal studies of post-angioplasty or instent stenosis have shown no effect or a catch-up phenomena after short-term promise [101]. The antibiotic doxycyclin can reduce MMP levels in serum and in carotid endarterectomy or aneurysm specimens explanted at surgery, but this agent has shown no effect on atheroma progression. [121,122]

4 Conclusion

A large body of evidence asserts the importance of MMPs in atherosclerosis. Studies showing activity of these enzymes in situ in the plaque and correlating cleaved collagen with MMP protein have increased our understanding of MMP functions in atheroma. Genetically modified mice expanded our knowledge of the functions of MMPs and promise greater insight into proteinases and their inhibitors in arterial diseases. The redundancy of MMPs and their elaborate interactions with other proteinases may allow compensation for the defects in some mice. Novel approaches such as studies of mice resistant to collagenolysis help to address these issues. Initial attempts to take laboratory findings to the bedside have been hampered by toxicity and efficacy issues with MMP inhibitors. The last 14 years have witnessed extensive progress in understanding basic proteinase biology, and it is clear that the vascular role of MMPs goes beyond ‘good’ inhibitors and ‘bad’ proteinases. Molecular imaging of MMPs may afford an opportunity to study the biology of plaques in vivo and allow better understanding of pathophysiologic mechanisms and evaluation of new therapies.


This work was supported by grants from The Jules Thorn Charitable Trust to Dr Dollery, and from the NIH (HL-080472-01) and The Leducq Transatlantic Network of Excellence for Cardiovascular Research to Dr Libby.The authors acknowledge the editorial assistance of Karen Williams.


  • Time for primary review 26 days


  1. [1]
  2. [2]
  3. [3]
  4. [4]
  5. [5]
  6. [6]
  7. [7]
  8. [8]
  9. [9]
  10. [10]
  11. [11]
  12. [12]
  13. [13]
  14. [14]
  15. [15]
  16. [16]
  17. [17]
  18. [18]
  19. [19]
  20. [20]
  21. [21]
  22. [22]
  23. [23]
  24. [24]
  25. [25]
  26. [26]
  27. [27]
  28. [28]
  29. [29]
  30. [30]
  31. [31]
  32. [32]
  33. [33]
  34. [34]
  35. [35]
  36. [36]
  37. [37]
  38. [38]
  39. [39]
  40. [40]
  41. [41]
  42. [42]
  43. [43]
  44. [44]
  45. [45]
  46. [46]
  47. [47]
  48. [48]
  49. [49]
  50. [50]
  51. [51]
  52. [52]
  53. [53]
  54. [54]
  55. [55]
  56. [56]
  57. [57]
  58. [58]
  59. [59]
  60. [60]
  61. [61]
  62. [62]
  63. [63]
  64. [64]
  65. [65]
  66. [66]
  67. [67]
  68. [68]
  69. [69]
  70. [70]
  71. [71]
  72. [72]
  73. [73]
  74. [74]
  75. [75]
  76. [76]
  77. [77]
  78. [78]
  79. [79]
  80. [80]
  81. [81]
  82. [82]
  83. [83]
  84. [84]
  85. [85]
  86. [86]
  87. [87]
  88. [88]
  89. [89]
  90. [90]
  91. [91]
  92. [92]
  93. [93]
  94. [94]
  95. [95]
  96. [96]
  97. [97]
  98. [98]
  99. [99]
  100. [100]
  101. [101]
  102. [102]
  103. [103]
  104. [104]
  105. [105]
  106. [106]
  107. [107]
  108. [108]
  109. [109]
  110. [110]
  111. [111]
  112. [112]
  113. [113]
  114. [114]
  115. [115]
  116. [116]
  117. [117]
  118. [118]
  119. [119]
  120. [120]
  121. [121]
  122. [122]
View Abstract