PCSK9 Goes “DAMP”

The story of Proprotein convertase subtilisin/kexin 9 (PCSK9) is a triumph of 21^st century scientific discovery leading to rapid translation to clinical utility. The story began with genetic discoveries in 2003 showing that gain of function mutations in the PCSK9 gene associated tightly with a rare form of familial hypercholesterolemia¹, and then in 2005 that rare loss of function mutations were associated with very low levels of LDL cholesterol (LDL-C) – 30% lower in African American populations and 15% in Caucasian² – and with an even more impressive reduction in risk of coronary artery disease (CAD) – approximately 90% and 50% respectively³. Shortly thereafter the cell biology was worked out, revealing that PCSK9 is produced in the liver, secreted, and then binds to the LDL receptor (LDLR), interfering with normal intracellular trafficking and accelerating lysosomal degradation, so that LDLR surface expression is decreased and circulating LDL levels thereby increased.⁴ The lack of any obvious detrimental effect on health or lifespan in individuals with loss of function mutations and very low LCL-C levels suggested that PCSK9 inhibition could be a viable therapeutic target for CAD. Humanized inhibitory monoclonal antibodies were quickly developed and early phase clinical trials were published in 2012, less than 10 years after the gene discovery, and showed efficacy in lowering LDL-C with no serious adverse side effects. In 2015 the FDA approved the first two antibodies, evolocumab and alirocumab, for patients with certain forms of hypercholesterolemia, based on well-designed randomized clinical trials showing improved lipid profiles and CAD outcomes compared to statin therapy alone. These agents, while expensive and cumbersome to deliver, are now part of the standard armimentarium for CAD risk reduction, and clinical research is ongoing to develop effective long-term strategies to inhibit PCSK9, such as with CRISPR/Cas mediated gene therapy.

As with all good stories, more complexity was revealed as additional drafts were produced. We learned, for example that PCSK9 regulates expression of many other cellular receptors in addition to LDLR, including metabolic receptors such as the VLDL receptor, apoER2, CD36 and LRP-1.⁴ Some of these, including CD36 and LRP-1, are potent signaling receptors expressed on vascular and hematopoietic cells and thus PCSK9 might very well regulate important hemostatic systems, including inflammation, hemostasis, and tissue repair. Observations that PCSK9 inhibitory antibodies offer CAD risk reduction, even in the presence of normal or low LDL-C; and that plasma PCSK9 levels associate with CAD risk independently of LDL-C levels support the hypothesis that some of the beneficial effects of anti-PCSK9 therapy might relate to these other receptor systems⁵.

The elegant studies by Qi et al from Fudan University published in this issue of Circulation⁶ convincingly demonstrate a role for PCSK9 in regulating platelet function through the type 2 scavenger receptor CD36. They showed that recombinant PCSK9 when added to washed platelets at concentrations consistent with what are seen in normal human plasma, enhanced platelet activation ex vivo in response to low doses of “classical’ agonists, as assessed by multiple parameters, including aggregation, integrin α_2bβ₃ activation, granule secretion, and cell spreading. This was reflected in an increase in intravascular thrombosis in vivo in a model of FeCl₃-induced arteriolar injury in which the mice were administered recombinant PCSK9 or in which LDLR deficient mice which are known to have increased PCSK9 plasma levels were studied.

What makes this work truly compelling is the mechanistic studies linking the phenotype to CD36. The investigators used pharmacologic, immunologic and genetic approaches to show that the enhanced platelet responsiveness to classical agonists and the enhanced arterial thrombosis induced by PCSK9 were dependent upon platelet CD36 expression and CD36 signaling. PCSK9 addition to platelets activated the CD36 signaling cascade, which has been characterized by our lab and others and shown to involve src-family kinases, ERK5, JNK, and P38 MAP kinases; Vav-family guanine nucleotide exchange factors, phospholipase A2; and intracellular reactive oxygen species.^7,8 Immunologic blockade of CD36, genetic deletion of CD36, or pharmacologic inhibition of CD36 signaling pathways abrogated the effects of PCSK9 in both ex vivo and in vivo studies. Interestingly, the effect of PCSK9 on CD36 activity was not related to suppression of CD36 expression, but rather was due to direct PCSK9 interaction with CD36. In other words. PCSK9 was presumably functioning as a ligand for CD36-mediated platelet activation. The authors showed by co-immunoprecipitation and immuno-fluorescence microscopy that CD36 and PCSK9 likely interacted closely with each other, supporting this hypothesis. Importantly, the apparent binding affinity of PCSK9 for CD36 was within the range of reported circulating PCSK9 plasma concentrations, and was similar to the affinity of known ligands for CD36, such as oxidized LDL (oxLDL).

CD36 is very pleiotropic in function⁹, serving as a fatty acid transporter on adipocytes and myocytes, a mitochondrial metabolic regulator on macrophages, an anti-angiogenic thrombospondin receptor on microvascular endothelial cells, a pro-atherosclerotic oxLDL uptake receptor on macrophages, and a pro-inflammatory signaling receptor on macrophages. The work of Qi et al reveals a potentially important connection between PCSK9 and platelet CD36. A role for CD36 as a pro-thrombotic receptor on platelets is now well established. Podrez et al initially showed that oxLDL interactions with CD36 promoted a platelet-dependent prothrombotic phenotype manifest as enhanced platelet activation to classical agonists¹⁰ – similar to what Qi et al showed for PCSK9. Human genetic studies have shown strong associations between platelet CD36 surface expression levels with platelet reactivity to oxLDL and with genetic polymorphisms associated with CAD risk¹¹, and mouse studies have shown that deletion of CD36 or its downstream signaling partners protects animals from high fat diet¹⁰ or diabetogenic diet-induced pro-thrombotic states¹².

The prothrombotic role of CD36 highlights an important recent recognition that the hemostatic system behaves very differently in settings of chronic inflammatory diseases compared to “normal” conditions. While studies of platelet function and coagulation enzymes in normal human subjects have yielded enormous insights into human physiology and led to life saving pharmacologic approaches to prevention and treatment of athero-thrombosis and venous thrombosis, studies of platelet and coagulation systems in the setting of chronic inflammation, such as cancer, diabetes, hyperlipidemia, atherosclerosis, and obesity have revealed that receptors and enzymes previously not thought to be relevant to normal hemostasis, may in fact be quite relevant to thrombosis. Examples include coagulation factor XII and the contact activating enzyme system (high molecular weight kininogen and prekalikrein), which are now being studied as contributors to venous thrombosis, especially in the setting of cancer. CD36 also falls into this category as a contributor to arterial thrombosis in the settings of hyperlipidemia, chronic systemic inflammation, and diabetes.

The contribution of CD36 to platelet-mediated thrombosis is due to its ability to transmit signals from extracellular cues to the platelet cytoplasm. Relevant CD36 extracellular ligands include oxLDL, advanced glycated (AGE) proteins¹², S100A family inflammatory peptides¹³, and neuropathic amyloid peptides¹⁴. These molecules fall into the general category of “danger associated molecular patterns” or DAMPs, a broad family of modified endogenous structures generated during inflammation, tissue injury and oxidant stress that interact with specific scavenger receptors and toll-like receptors to trigger pro-inflammatory and prothrombotic responses by the innate immune system. Published studies with mouse models show that experimental conditions that mimic type 1 and type 2 diabetes mellitus, obesity, hyperlipidemia and chronic vascular inflammation generate CD36 ligands in vivo and promote platelet hyper-reactivity and arterial thrombosis. The studies by Qi et al strongly support the addition of PCSK9 to the CD36 ligand family, raising the question of whether PCSK9 should be considered a “danger associated molecular pattern” or DAMP, similar to oxLDL, AGE-proteins, S100A, and cell-derived microvesicles. Interestingly, PCSK9 expression is enhanced in macrophages by activation of the NLRP3 inflammasome and interleukin-1β¹⁵, consistent with this hypothesi.

Is there plausibility for high affinity interactions between PCSK9 and CD36? Qi et al’s studies clearly show physical and functional interactions between the two proteins, but direct binding has not yet been proven. Structural studies of PCSK9 binding to LDLR show a critical role for a transient amphipathic helix in the PCSK9 prodomain-4. Since neuropathic amphipathic amyloid peptides are known to bind and activate CD36 in microglial cells¹⁴, it is certainly reasonable to hypothesize that PCSK9 may bind to CD36 via its “DAMP-like” amphipathic domain.

The study by Qi et al adds to the growing body of literature supporting an important role for CD36 in promoting platelet hyperactivity and atherothrombosis in disease states and importantly identifies PCSK9 inhibition as a potentially safe therapeutic target to prevent arterial thrombosis in individuals who may be at increased risk because of elevated PCSK9 levels related to genetic or acquired causes, such as inflammation or even use of statins.