Proteases, protease activated receptors and atherosclerosis

Coagulation activation by the tissue factor (TF) pathway plays pivotal roles in triggering platelets and precipitating acute coronary syndromes. While dual anti-platelet therapy is effective in secondary cardiovascular prevention, combining platelet antagonism with low dose aspirin and the oral coagulation FXa antagonist rivaroxaban has a synergistic clinical benefit over monotherapy in preventing the composite outcome of cardiovascular death, stroke, or myocardial infarction¹. It is therefore of considerable interest to understand the roles of coagulation proteases and their cell signaling effects in the development of atherosclerosis and vascular inflammation. Acute thrombosis in animal models typically requires the combination of FXII contact activation and the extrinsic TF pathway^{2, 3}, but vascular inflammation and hypertension in the absence of overt intravascular thrombosis can utilize a unique TF-initiated and platelet-assembled thrombin-FXI amplification loop independent of contact pathway FXII⁴. While atherosclerotic lesions in the mouse are typically not thrombogenic, interference with coagulation nevertheless attenuates lesion development. Genetic manipulation of thrombin generation⁵ or deletion of FXI⁶, as well as pharmacological inhibition of FXa⁷ or thrombin⁵ reduce atherosclerosis in apolipoprotein E (ApoE) knock-out mice, raising questions about cellular targets and proatherogenic mechanisms of coagulation proteases beyond precipitating intravascular thrombosis.

Two studies in the current issue of ATVB demonstrate roles for protease activated receptor (PAR) signaling in atherosclerosis. Rana et al. ⁸ used pharmacological intervention with pepducins, cell-penetrating peptides disrupting intracellular coupling of G protein-coupled receptors, to implicate PAR1, but not PAR2 in lesion development in ApoE−/− mice. Considering the previously demonstrated reduction of plaque burden by coagulation inhibition in this mouse model, it is surprising that thrombin inhibition with the hirudin analogue bivalirudin had no therapeutic benefit. Instead, pharmacological inhibition of matrix metalloproteinase (MMP) 1 attenuated lesion progression with similar efficacy as the PAR1 antagonist. In patients with angiographically confirmed coronary atherosclerotic burden, plasma MMP1 levels were found to be correlated with severity of disease. Macrophage infiltration of atherosclerotic lesions in the mouse was inhibited by the PAR1 pepducin PZ-128 (P1pal-7) and the MMP1 inhibitor FN-439. Although this inhibitor has some activity against MMP13 detected in atherosclerotic lesions, MMP13 plays no apparent role in atherosclerosis development, despite being a relevant PAR1 activator in heart failure⁹. In vitro studies further showed that MMP1-PAR1 signaling is crucial for tumor necrosis factor (TNF) α induction of vascular cell adhesion molecules (VCAM, ICAM) in endothelial cells, providing a mechanism for PAR1-mediated support of leukocyte recruitment to the vessel wall (Figure).

Figure.

Proposed contributions of endothelial cell (EC) MMP1-PAR1 and smooth muscle cells (SMC) PAR2 signaling in monocyte (Mo) recruitment in atherosclerosis.

Rana et al.⁸ found no effect of thrombin inhibition despite documented antithrombotic dosing of bivalirudin, which contrasts with effective lesion reduction by the small molecule thrombin inhibitor dabigatran given to in hyper-thrombotic thrombomodulin mutant (TM^Pro) mice⁵. Pharmacodynamic differences in the inhibition of free versus cell bound thrombin and vessel wall penetrance of the low molecular thrombin inhibitor may contribute to the observed differences. In addition, TM^Pro mice are impaired in thrombin binding and consequently protein C (PC) activation and activated PC has been implicated in MMP activation¹⁰. It is therefore conceivable that endothelial dysfunction, as mimicked by the TM^Pro mouse model, bypasses MMP-PAR1 signaling and enables thrombin-PAR1 driven vascular pathologies. It will be of interest to study and develop new mouse models that are resistant to thrombin cleavage of PAR1 at the canonical Arg⁴¹-Ser⁴² bond¹¹ or to alternative MMP1 cleavage of PAR1 at Asp³⁹-Pro⁴⁰ and MMP13 at Ser⁴²-Phe⁴³, which creates distinct tethered ligands potentially acting as biased agonists⁹.

In a seemingly contradictory second study in this issue, Jones et al.¹² demonstrate that PAR2, but not PAR1 deletion attenuates atherosclerotic lesion development in LDLR−/− mice. A similar atherosclerosis protection by PAR2-deficiency in ApoE−/− mice¹³ essentially excludes different roles for PAR2 in these two atherosclerosis models. Knock-out of PAR1 or PAR2 can result in partial compensation for the loss of the reciprocal receptor, as indicated by developmental studies¹⁴. Such compensation in knock-out mice may obscure receptor participation that are clearly demonstrated by pharmacological approaches. PAR receptor crosstalk is indeed likely, because PAR1 and PAR2 form functional heterodimers in endothelial cells¹⁵ and the anti-atherogenic PAR1 pepducin PZ-128 (P1pal-7) is known to inhibit PAR1/PAR2 heterodimer signaling¹⁶. In addition, an agonistic PAR1 pepducin produced injury-induced smooth muscle cell hyperplasia dependent on PAR1 as well PAR2 expression¹⁷, demonstrating that pepducins can modulate vessel wall PAR1/PAR2 heterodimer signaling.

The study by Jones et al.¹² adds to the increasing evidence that PAR2 is central to inflammatory processes, participating in innate immune signaling by toll like receptor 4^{18, 19} and promoting metabolic inflammation in obesity²⁰. While stromal and hematopoietic cell TF-PAR2 signaling contributes to weight gain and insulin resistance in obesity, atherosclerosis development is primarily driven by the radiation-insensitive vascular compartment. PAR2-deficient vascular smooth muscle cells stimulated with PAR2 agonist, TNFα, or interleukin (IL) 1β express lower levels of chemokines CCL2 and CXCL1 and consequently monocyte migration in vitro and vessel wall infiltration is reduced¹². Future studies with cell-type specific deletion of PAR2 are necessary to confirm the role for smooth muscle cell PAR2 in regulating leukocyte recruitment and identify potential roles of endothelial cell PAR2 in the upregulation of endothelial cell adhesion molecules by MMP1-PAR1 signaling proposed by Rana et al.⁸.

These studies provide important new insights into the contributions of protease signaling pathway to the development of atherosclerosis, but also demonstrate the need to evaluate these complex processes by combinations of pharmacological and genetic approaches. Taken together, these studies show that the key vascular signaling transducers for coagulation proteases, PAR1 and PAR2, both participate in the development of atherosclerosis and can be targeted for the attenuation of lesion progression. It is also clear that activation of PARs in the vascular compartment may involve interconnected proteolytic cascades and is not restricted to coagulation proteases. Understanding the limitations of target selective clinical anticoagulants, i.e. thrombin and FXa inhibitors, and the contributions of these proteases to PAR activation in vascular and immune cells will be essential for predicting long term benefit of antithrombotic therapy in secondary cardiovascular prevention.