Protease Activity in Vascular Disease

The many roles of extracellular proteolysis in vascular pathology

Protease activity has been implicated in the pathogenesis of vascular diseases including atherosclerosis, thrombosis, and aneurysm. A wide variety of proteases representing different proteolytic families and their corresponding inhibitors are involved. These proteases contribute to vascular disease through a series of overlapping pathways that affect overall inflammatory status and structural integrity of the vessel wall. By activating protease-activated receptors (PARs), these enzymes propagate inflammatory signaling, cytokine production, and inflammatory cell recruitment. Additionally, proteases can degrade components of the extracellular matrix, elastic lamina, and fibrous cap in atheroma. The prevailing paradigm is that excessive proteolytic activity is a significant contributor to the initiation and progression of vascular disease. Recent approaches to treatment of vascular pathologies have attempted to modulate protease activity in an effort to reduce inflammation and preserve structural integrity of the vessel wall.

This Highlight article will review foundational evidence for the role of extracellular protease activity in vascular pathology (Figure 1) with a focus on the most recent discoveries and approaches to treatment. Advances related to atherosclerosis and aneurysm will be primarily emphasized as the role of protease activity in thrombosis has been recently reviewed ¹.

Figure 1. Expression of proteases and protease activated receptors in the vasculature.

Inflammation in the vasculature results in the generation of proteases by both stromal and inflammatory cells. Protease activity can contribute to vascular pathogenesis by propagating inflammation and causing structural damage, increasing risk of an atherothrombotic event. Protease-activated receptors (PARs) sense protease activity and drive cellular responses. These responses are typically proinflammatory and are dependent on the specific PARs that are expressed and the cell type. This figure emphasizes the distribution of proteases and PARs examined in recent literature. NE – neutrophil elastase; PR3 – proteinase 3; Cat-G – cathepsin G; uPA – urokinase-type plasminogen activator; MMP – matrix metalloproteinase.

Protease families commonly implicated in vascular pathogenesis

Serine proteases and their inhibitors

Serine proteases are one of the largest subclasses of proteolytic enzymes. Due to their diversity and vital functional roles, they are a crucial component of the human degradome. These enzymes are involved in many of the body’s essential processes including inflammation, blood coagulation, and apoptosis, all of which contribute to vascular disease ^{2, 3}. Several of these enzymes and their inhibitors have been highlighted in recent literature for their roles in vascular pathogenesis. A few key examples are summarized below.

Neutrophil elastase is a serine protease that contributes to the breakdown of elastin, an extracellular matrix protein that allows elasticity of various tissues throughout the body. In atherosclerotic lesions, elastase is produced by neutrophils and macrophages as part of the inflammatory response to cholesterol accumulation ⁴. Elastase degrades elastin in the artery wall, contributing to vascular stiffness and plaque instability, potentially increasing the risk of an atherothrombotic event. Recently, in vivo fluorescence imaging has been used to monitor elastase activity in atherosclerotic lesions ⁵. In this study, an activatable imaging agent, which is cleaved by elastase to produce a fluorescent signal, was used to image plaque at various time points in Ldlr^−/− mice on high-fat diet. After 4 weeks on diet, elastase activity was significantly increased within lesions ⁵. This activity progressively declined at 8 and 12-week time points. This suggests that during atherogenesis, an initial burst of elastase production occurs that declines with time and neutrophil number or that its activity is overcome by endogenous inhibitors such as alpha-1-antitrypsin.

Thrombin is synthesized by the liver as the zymogen prothrombin and secreted into the circulation, where it is activated in the coagulation cascade ^{6, 7}. It’s role in blood coagulation and platelet activation makes it essential for proper circulatory function, as well as injury response ⁷. In atherosclerosis, thrombin can exacerbate the inflammatory cascade by promoting platelet production and by activation of protease activated receptors (PARs) ⁸. Recent research on endogenous thrombin inhibitors, such as biglycan, suggest a protective role for thrombin inhibition in atherosclerosis ⁸.

Protein C is a vitamin K-dependent glycoprotein synthesized as a zymogen that is activated by the thrombin-thrombomodulin complex on endothelial cells to generate activated protein C (APC) ⁹. When activated, APC plays a crucial role in anticoagulation and the prevention of thrombosis ¹⁰. Genetic deficiency of protein C correlates with a higher risk of thrombotic events ¹⁰. In addition to its antithrombotic capabilities, APC also has important anti-inflammatory roles ¹¹. Recent evidence suggests that APC can prevent apoptosis and stabilize the endothelial barrier through a mechanism involving apolipoprotein E receptor 2 (APOER2) ¹². APOER2 is widely recognized for its role in the nervous system, however this study demonstrates an important function in the endothelium alongside APC ¹². Another recent ATVB article reports that knockdown of APC using siRNA in Apoe^−/− mice results in spontaneous atherothrombosis, further underscoring its relevance to the pathology of atherosclerosis and providing a potential new murine model for atherothrombosis ¹³.

Serine protease inhibitors (SERPINs) are a large family of protease inhibitors with 36 members, 29 active inhibitors and 7 inactive. They are distinct in their unusual mechanism, which constitutes irreversible inhibition of the protease due to a dramatic conformational change that blocks the active site. SERPINs have been implicated in many processes, including coagulation and inflammation. The possible use of SERPINs as therapeutic agents in cardiovascular disease has been suggested due to their anti-apoptotic effects ¹⁴.

Alpha-1 antitrypsin (A1AT) is an abundant SERPIN in humans and is the primary physiological inhibitor of neutrophil elastase. Genetic deficiency of A1AT can cause chronic obstructive pulmonary disease due to excessive elastase-mediated elastin degradation in the lungs. This condition has also been associated with increased carotid intima-media thickness ¹⁵, a marker of cardiovascular disease. Furthermore, A1AT gene variants correlate with atherosclerosis progression ¹⁶. This evidence suggests that unregulated elastolytic activity in the vessel wall may contribute to atherosclerotic plaque formation ¹⁶. One common A1AT variant, p.V213A, is strongly associated with risk of large artery stroke likely by affecting A1AT structure and its interaction with lipoproteins ¹⁷.

Plasminogen activator inhibitor-1 (PAI-1) is a potent inhibitor of the tissue-type and urokinase-type plasminogen activators. This results in the inhibition of fibrinolysis and the breakdown of blood clots, and preserves a normal clotting response to bleeding. Decreased or inactive PAI-1 causes potentially life-threatening bleeding disorder, and can lead to scar tissue buildup causing heart failure ¹⁸. Dysfunctional PAI-1 is linked to several diseases associated with vascular remodeling including pulmonary hypertension, vascular injury, and atherosclerosis ¹⁹. Elevated levels of PAI-1 are associated with the progression of coronary artery disease ²⁰, and transgenic mice overexpressing PAI-1 develop spontaneous coronary artery thrombosis ²¹. The mechanistic contribution of PAI-1 unclear, however it has recently been described as a driver of cellular senescence in vascular disease and ageing ²². PAI-1 has also recently been demonstrated to play an important role in inflammatory cell recruitment during ischemia-reperfusion injury ^{23, 24}.

Metalloproteinases and their inhibitors

Matrix metalloproteinases are a family of 23 calcium-dependent and zinc-binding enzymes that can degrade the extracellular matrix and are involved in the processing of several bioactive molecules ²⁵. Similar to other protease families, they are produced as zymogens that require activation. Many MMPs have been reported to contribute to vascular disease. For example, MMP-1 significantly correlates with coronary artery disease (CAD) in humans ²⁶. MMP2 and MMP-9 are elevated in acute coronary syndrome (ACS) ²⁷, and also play a role in pathology of abdominal aortic aneurysm ²⁸.

The ADAMs (a disintegrin and metalloproteinase) are a related protease family, which have also been implicated in vascular pathologies ^{29, 30}. ADAMTS-1, −4, and −5 activities have been implicated in thoracic aortic aneurysm (TAA) formation ^31–33. Recent studies published in ATVB have focused on ADAMTS-13. ADAMTS-13 cleaves von Willebrand factor (VWF) preventing the accumulation of excessive VWF aggregates and platelet-rich thrombi. Deficiency results in thrombotic thrombocytopenic purpura and is caused by rare genetic mutations or the generation of autoantibodies against ADAMTS-13 ³⁴. Current therapy for autoimmune deficiency involves removal of autoantibodies by plasma exchange. However, transfusion of ADAMTS-13 loaded platelets has recently been proposed as a novel treatment approach ³⁵.

Tissue Inhibitors of Matrix Metalloproteases (TIMPS) are a family of proteins responsible for inhibition of MMPs and ADAMs. TIMPs 1–4 have varying affinities for different MMPs and multiple enzymatic binding sites, allowing them to inhibit a variety of metalloproteinases ³⁶. TIMPs can prevent over activity of MMPs in degrading the extracellular matrix and may provide protection against atherosclerosis. For example, overexpression of TIMP3 decreases atherosclerotic inflammation and severity in low density lipoprotein receptor knockout mice ³⁷.

Extracellular Traps (ETs) are vast webs of protease activity

Extracellular traps (ETs) were originally described as a novel form of neutrophil cell death with a role in innate immunity ³⁸ and have more recently been discovered to also be produced by macrophages and mast cells ³⁹. This process, referred to as ETosis, is distinct from other forms of programmed cell death (e.g. necrosis and apoptosis), and releases a web-like structure of decondensed chromatin decorated with granular proteins (Figure 1). The proposed function of these structures is to maintain high local concentrations of effector proteins. In the case of neutrophil extracellular traps (NETs), these proteins include reactive oxygen species generating enzymes (e.g. myeloperoxidase) and proteases (e.g. neutrophil elastase, proteinase 3, cathepsin G, and MMP-9). NETs have been detected in both human and mouse atherosclerotic lesions ^{40, 41}, and elevated levels of circulating NET material are associated with severe coronary atherosclerosis ^{42, 43}. The mechanisms by which NETs contribute to atherosclerosis, thrombosis, and aneurysm have been the focus of recent investigation.

Neutrophil recruitment to the vessel wall can be promoted by activated macrophage cells. Accumulation of cholesterol in atherosclerotic lesions results in the formation of foam cells, which are dysfunctional cholesterol-loaded macrophages. The formation of cholesterol crystals within these cells can single-handedly activate the inflammasome by acting as both “signal 1” and “signal 2” for inflammasome assembly ⁴⁴. Inflammasome activation produces mature IL-1β which can then activate endothelial cells, triggering recruitment of neutrophils to plaque and their subsequent activation to promote NETosis ⁴⁴. IL-1β induced NETosis may also play a role in abdominal aortic aneurysm (AAA). NETs are detected in AAA aneurysms in humans and mice, and are colocalized with IL-1β ⁴⁵. In mouse AAA, knockout of IL-1β results in decreased neutrophil infiltration and is protective against AAA ⁴⁵. Additionally, treatment with the NETosis inhibitor Cl-amidine reduced AAA formation ⁴⁵. Another recent development suggests that NET serine protease activity contributes to the pathogenesis of AAA. Using a mouse model deficient in three neutrophil serine proteases (elastase, proteinase 3, and cathepsin G), Yan et. al. demonstrated that these enzymes contribute significantly to AAA formation ⁴⁶.

NETs can have direct effects on the vascular endothelium. Human endothelial cells exposed to NETs become activated, increasing expression of cell adhesion molecules ICAM-1 and VCAM-1, resulting in increased monocyte adhesion ⁴⁷. Endothelial cell tissue factor expression is also increased following NET exposure, resulting in accelerated clotting ^47–49. These effects on the endothelium are mediated by NET associated IL-1α and the serine protease cathepsin G ⁴⁷. The contributions of other NET-bound proteases (i.e. neutrophil elastase and proteinase 3) were not investigated due to rapid inhibition by abundant inhibitors in the serum component required for these assays. In addition to promoting recruitment of inflammatory cells, NET associated protease activity can also degrade endothelial cell barrier function. NET elastase activity can disrupt intracellular junctions by proteolysis of the tight junction protein VE-cadherin, resulting in vascular leakage ⁵⁰. NET elastolytic activity can also drive endothelial-to-mesenchymal transition through activation of β-catenin signaling, an observation that was blocked by the elastase inhibitor sivelestat ⁵⁰. This cellular trans-differentiation, stimulated by NET elastase activity, may contribute to the local accumulation of extracellular matrix proteins and vascular fibrosis in atheroma.

While it is clear that NETosis is contributing to the pathogenesis of atherosclerosis and aneurysm, the influence of biological factors that regulate NET formation is not well understood. A recent study published in ATVB identified age as a factor contributing to NETosis in vascular disease. NETs were significantly more prevalent in the atherosclerotic lesions of aged mice compared to lesions of younger mice ⁵¹. Using mitochondrial catalase transgenic mice, this observation was attributed to mitochondrial oxidative stress and suppression of mitochondrial oxidative stress in cultured neutrophils from aged mice prevented 7-ketocholesterol induced NET formation.

Sensing extracellular proteases by Protease-Activated Receptors (PARs)

The complex milieu of proteolytic enzymes present in the inflamed vessel can mediate a variety of physiological responses through stimulation of protease-activated receptors (PARs). The PARs are a family of four (PAR1–4) transmembrane g-protein coupled receptors. Nearly all cells in the body express at least one PAR, with most cell types expressing multiple PARs (Figure 1). These receptors are distinct from other g-protein coupled receptors in that their activation typically occurs when extracellular proteolytic cleavage of the receptor’s N-terminal domain results in exposure of a tethered ligand. This ligand then self-activates the receptor, triggering intracellular signal transduction. The responses mediated by PAR activation can vary significantly depending on the specific PAR and the activating protease. Cellular responses can even differ when the same PAR is activated by different proteases. A complex story is unfolding regarding which proteases activate which PARs and their subsequent cellular responses in vascular pathology.

Recent studies have focused on the contributions of PAR1 and PAR2 in atherosclerosis ⁵². Expression of these PARs is increased in atherosclerotic lesions from mice and humans ^53–55. Furthermore, knockout mouse models for PAR1 and PAR2 have each demonstrated reduced atherosclerotic burden ^55–57, suggesting that both contribute to atherogenesis. In a recent study by Rana et. al., circulating levels of MMP-1, a non-canonical activator of PAR1 ⁵⁸ that is colocalized with PAR1 in human atherosclerotic lesions ⁵⁹, was correlated with angiographically documented coronary atherosclerosis burden ²⁶. Treatment of Apoe^−/− mice with the MMP-1 inhibitor, FN-439, or the PAR1 antagonizing pepducin, PZ-128, resulted in reduced atherosclerosis burden and reduced vascular inflammation. The influence of canonical (thrombin) vs non-canonical (MMP-1, NE, PR3, APC) activation of PAR1 on intracellular signal transmission has recently been examined using a comprehensive phosphoproteomics analysis ⁶⁰. Despite the generation of varying tethered ligands and previous literature suggesting “biased activation” ^{58, 61, 62}, these non-canonical ligands triggered phosphoregulation patterns that largely resembled those of thrombin ^{60, 63}. PAR2 differs from PAR1 in its complement of activating proteases, and thus may play a distinct role in the vascular response to proteolytic activity. Using a bone marrow transplant model, Jones et. al. demonstrated that PAR2 on nonhematopoietic cells contributes significantly to development of atherosclerosis in Ldlr^−/− mice ⁵⁵. Lack of PAR2 resulted in improved plaque morphology in a chronic progression model and reduced plasma cytokines Ccl2 and Cxcl1. Taken together, these findings support significant proatherogenic roles for both PAR1 and PAR2 activation in mouse models.

Understanding the impact of individual PARs on intracellular signaling and their respective contributions to vascular pathology is further complicated by cross talk between PARs. For example, activation of PAR1 can result in transactivation of PAR2 ^{64, 65} and PAR4 ⁶⁶. Upon activation, each pair can form heterodimers that are internalized, triggering a distinct cellular response ^{67, 68}.

Structural implications of vascular protease activity

In addition to the extensive network of proinflammatory signaling triggered by PAR activation, proteases also cause damage to the underlying structure of the vasculature. Degradation of the extracellular matrix (ECM) components elastin, collagen, and fibronectin ⁶⁹ can impact endothelial barrier function and vessel elasticity. This increases propensity for an atherothrombotic event or aortic dissection. Several articles recently published in ATVB have furthered our understanding of the roles of protease activity in maintaining vascular structure.

Endothelial cells play an important role in maintenance of the sub-endothelial environment ⁷⁰ and may be able to regulate vascular protease activity. Chen et. al. recently reported that signaling through the nucleotide receptor P2Y₂ on endothelial cells can influence subendothelial MMP-2 activity and features of atherosclerotic plaque stability using an endothelial cell-specific P2Y₂R deficient Apoe^−/− mouse model ⁷¹.

Macrophages contribute to the vascular proteolytic environment through secreted and membrane-associated protease activity. Classically activated (i.e. proinflammatory) macrophages produce elastase, MMPs, and urokinase-type plasminogen activator (uPA), and are capable of ECM degradation whereas alternatively activated macrophages do not degrade ECM ⁷². Hohensinner et. al. recently described that this suppression of proteolytic activity upon alternative activation is due to expression of the uPA inhibitor, PAI-1 ⁷². However, the role of PAI-1 in vascular remodeling is complicated. While it appears to prevent macrophage driven ECM degradation, it may actually promote other aspects of vascular remodeling. In a murine carotid artery ligation model, inhibition of PAI-1 prevented smooth muscle cell migration and neointima formation ⁷³. Macrophage apoptosis can contribute to the development of the necrotic core in atherosclerotic lesions. Apoptotic signaling results in the release of markers such as FADD and caspases. In a Swedish cohort of over 4,000 subjects, plasma levels of these markers were associated with increased incidence of coronary events ⁷⁴. These findings suggest that the degree of atherosclerotic apoptosis may significantly contribute to event risk.

Vascular smooth muscles cells also contribute to inflammatory signaling and vessel remodeling ⁷⁵. Apoptosis of smooth muscle cells is associated with accelerated atherosclerosis and degradation of the ECM. Yu et. al. demonstrate that activation of the transcriptional regulator forkhead transcription factor O subfamily member 3a (FOXO3a) induces apoptosis in smooth muscle cells and promotes atherosclerosis via ECM degradation in a process involving MMP-13 ⁷⁶.

A recent article in ATVB reports a significant role for ADAMTS-5 as a key structural regulator of the large ECM proteoglycan versican ²⁹, and this activity was implicated in aortic aneurysm using an AngII infusion model. Mice lacking the catalytic domain of Adamts5 displayed increased aortic dilatation compared to control mice.

Protease-mediated structural damage is not limited to ECM proteins but can also impair the protective functions of soluble proteins. One key example is the protein apolipoprotein A-I (apoA-I). This core structural component of high-density lipoproteins (HDL) has numerous anti-inflammatory and atheroprotective properties ⁷⁷. Many of these functions have been shown to be impaired when apoA-I is cleaved at different sites by various proteases ^78–80.

Therapeutic approaches to modulate protease activity or PAR responses

The vast implications of extracellular proteolysis in vascular inflammation and structural integrity have driven the development of several approaches for therapeutic suppression of protease activity. Strategies include direct inhibition of protease activity, targeting of NETs ⁸¹, blockade of extracellular PAR activation, and inhibition of intracellular PAR signaling ⁸². Each of these has its merits and potential caveats. Many are faced with the challenge of specificity. Due to the tremendous diversity and overlapping functions of both proteases and their targets, it has proven difficult to develop inhibitors by any of the above strategies that will not produce off-target effects.

One novel strategy, recently published in ATVB, used nanoparticles to deliver the peroxisome proliferator-activated receptor-γ (PPAR-γ) agonist pioglitazone to circulating monocytes and aortic macrophages. Activation of PPAR-γ in these cells results in a shift toward a less inflammatory phenotype. Weekly administration of pioglitazone-nanoparticles to Apoe^−/− mice receiving western diet and angiotensin II infusion resulted in reduced MMP and cathepsin activity in the brachiocephalic artery and a more stable plaque phenotype ⁸³.

PAR4 is expressed on platelets, along with PAR1, where activation by thrombin initiates aggregation and thrombus formation. Current anti-platelet therapies, such as aspirin and P2Y₁₂ antagonists, do not affect PAR activation by thrombin and are associated with increased risk of bleeding. The results of a recent phase I trial of the PAR4 inhibitor BMS-986120 were recently published in ATVB ⁸⁴. BMS-986120 is the result of a high-throughput screen of a 1.1 million compound library for targets that inhibit PAR4-activation peptide induced calcium signaling and γ-thrombin induced platelet aggregation ⁸⁵. Wilson et. al. demonstrated the high selectivity of the compound for PAR4 activation and favorable clotting characteristics in a prospective randomized open-label blinded endpoint trial using plasma from participants receiving oral BMS-986120 ⁸⁴. Another selective PAR4 inhibitor, SCH-28, has also shown inhibition of platelet aggregation without preventing coagulation ⁸⁶. Specific antagonism of PAR4 on platelets, while leaving PAR1 signaling intact, appears to produce potent inhibition of thrombin induced activation while maintaining hemostasis ⁸⁷. Whether or not these strategies will be influenced by common PAR4 gene variants that affect platelet activation ⁸⁸ has not yet been reported. Additional anticoagulant approaches directly targeting coagulation factor proteins and platelets have been reported ^{1, 89–94}.

Pepducins are palmitoylated peptides designed to target the intracellular domain of G-protein coupled receptors ⁹⁵. Such peptides have been developed for protection against atherothrombotic disease or complications associated with coronary interventions by targeting signaling through PAR1 and PAR4 ^{82, 96}. The high specificity, rapid action, and reversible nature of these agents is expected to result in reduced off-target effects and reduced incidence of major bleeding as seen with other anticoagulant strategies.

One under-examined potential mechanism of protease regulation in vascular disease is the role of lipoproteins. Circulating lipoproteins, particularly HDL, carry a diverse protein cargo that includes a large proportion of proteases and protease inhibitors. In fact, over 20% of the almost 100 proteins commonly identified on HDL have known roles in protease regulation ³ and HDL-bound levels of some correlate with atherosclerosis burden in humans ⁹⁷. Furtado et. al. recently characterized the existence of novel HDL subspecies defined by the presence of the SERPINs A1AT and α−2-macroglobulin ⁹⁸. So far, it is unclear exactly what impact HDL-association has on the function of protease inhibitors or their involvement in atherogenesis.

Summary

These recent reports, and the data accumulated over decades, overwhelmingly support a critical role for protease activity in the pathogenesis of various vascular dysfunctions. These activities clearly evolved with critical roles in innate immune defense and wound healing ^{48, 99}. However, the typical vascular pathologies requiring medical attention today result from sterile inflammation and/or ageing. Therefore, an important consideration as we move forward with proteolysis suppressing therapies is whether it can be accomplished with sufficient specificity in terms of the molecular targets and tissue localization. This is imperative to avoid detrimental off-target effects that may impair host defense or wound healing. While investigating these mechanisms and potential targets in animal studies, it is important to carefully consider the model used, experimental design, and method of reporting ¹⁰⁰. While developing these options as treatments for vascular disease and considering their use in patients, it is also important to keep in mind the potential contributions of race and gender in response to therapy ^{88, 101}. Significant progress has been made in the fundamental understanding of proteostasis and the potential benefits of manipulating proteolytic balance to preserve vascular integrity and function. It appears that exciting new developments are on the horizon.