Lipoprotein(a) gets worse

Lipoprotein(a) [Lp(a)] is an independent genetic risk factor for cardiovascular disease and calcific aortic valve stenosis¹. The Emerging Risk Collaborators’ meta-analysis of epidemiologic cohorts including 126,634 participants has shown a continuous, independent, association of plasma Lp(a) concentrations (above 30-50mg/dL) and risk of coronary heart disease (CHD) or stroke¹. Further, Mendelian randomization studies have shown that Lp(a) is a genetically determined causal risk factor for myocardial infarction and aortic valve stenosis². Despite Lp(a)’s strong and causal association with CHD, its underlying molecular mechanisms are poorly understood, and this limits our ability to identify biologic targets for development of therapeutics.

Lp(a) is the product of the covalent binding by disulfide bond of apoB to the plasminogen-like glycoprotein apo(a) in a 1:1 molar ratio³. Just like an low density lipoprotein (LDL) particle, these proteins surround a lipid core of cholesteryl esters, free cholesterol, triacylglycerols, and many different phospholipid species. Both circulating lipoproteins can infiltrate to subendothelial space, undergo oxidation, and trigger or enhance atherogenesis. Both lipoproteins carry oxidized phospholipids (ox-PL), a cause of sustained local inflammation and contributor to cardiovascular disease⁴. Inappropriate LDL levels are found in >40% of people 40 years and older, and LDL management is a well-established target of CHD prevention and therapy. Inappropriately high Lp(a) levels are inherited and affect ~30% of the population. Evidence from large statin and proprotein convertase subtilisin/kexin type 9 (PCSK9) inhibitor trials suggests that elevated Lp(a) levels predict higher event rates at any achieved LDL level suggesting that despite its LDL-like structure, the presence of apo(a) adds unique functions and effects to Lp(a) ⁵. Indeed, synthesis, processing, clearance, and association with oxidized phospholipids, all make Lp(a) different from LDL ⁴(see Figure).

Figure:

Apo(a) and plasminogen share extensive structural homology. The variation in kringle IV type 2 repeats encoded by the LPA gene gives rise to a over 40 apo(a) size isoforms. Plasma Lp(a) levels are genetically determined and can vary 1000 fold. Plasma levels above 30-50mg/dL are linked to enhanced CVD risk. High plasma levels of Lp(a) and associated oxidized phospholipids promote pro-atherogenic and pro-thrombotic activities and inhibit the anti-thrombotic effects of plasminogen.

As we said, Lp(a) is the major carrier of oxidized phospholipids (ox-PL) in human plasma. In this issue of Circulation Research, Schnitzler et al. ⁶ through a series of elaborate in vitro, in vivo and human studies, show that ox-PL on Lp(a) stimulate vascular inflammation and leukocyte extravasation. The authors present convincing evidence that Lp(a) increases adhesion molecule expression, monocyte binding to the endothelium, and transmigration into the subendothelial space. Mutations in the ox-PL binding sites of apo(a) abolished the effects attributed to Lp(a), linking directly ox-PL to the vascular processes under study. They also show that Lp(a)-associated ox-PL exacerbate the inflammatory status and promote a pro-adhesive state by targeting the endothelial glycolysis pathway through activation of 6-phosphofructo-2-kinase/fructose-2,6-biphosphatase 3 (PFKFB3, a positive regulator of glycolysis). Further, the enhanced inflammatory effects of serum from patients with extremely high Lp(a) levels (mean 203 mg/dL) were significantly diminished after treatment with an antisense against apo(a) that lowers Lp(a) levels by as much as 80%. These findings are novel, important, and have a translational take, as they paint a wider and more complete picture of the ways in which Lp(a) acts as insult to the artery wall.

The apo(a) gene (LPA), evolved from a duplication of the plasminogen gene (PLG) and contains multiple kringle domains and a mutated protease domain that lacks the proteolytic activity of plasminogen⁴. Ox-PL are carried by both Lp(a) and plasminogen. Plasminogen is abundantly present in atherosclerotic plaques⁷. Because of their structural similarity, it is not surprising that Lp(a) and plasminogen can compete, or perhaps synergize, for different functions. For example, Lp(a) inhibits the effect of plasminogen on fibrinolysis, thus interfering with a protective vascular effect⁸. More recently, plasminogen has been shown to drive cellular cholesterol efflux⁹, a beneficial role previously attributed almost exclusively to HDL, and this effect appears to be inhibited in the serum of subjects with very high levels of Lp(a)¹⁰. So, there are many ways in which Lp(a) can be causing trouble for our arteries.

The studies of Schnitzler et al., though carefully designed, rigorously performed, and cautiously interpreted, warrant additional work to address how the ox-PL cargo on Lp(a) impacts plasminogen-mediated sterol efflux by the macrophages, and whether this in turn affects macrophage migration.

Lp(a) exhibits wide heterogeneity as a result of the multiple apo(a) length isoforms in the population³. Lp(a) varies in size (300-800 kDa) among individuals due to polymorphism in the LPA locus resulting in a variable number of kringle IV type 2 (KIV-2) repeats³, with over 40 different length isoforms of the protein. Kringle repeat length is inversely correlated with plasma Lp(a) levels, which are believed to be the drivers of risk¹¹. However, a recent Mendelian randomization study from the Pakistan Risk of Myocardial Infarction Study (PROMIS) cohort determined that smaller apo(a) isoform size and increased plasma Lp(a) levels both independently and causally associated with coronary heart disease². In the study by Schnitzler et al., levels of plasma or Lp(a)-associated ox-PL were not reported. Thus, it is not possible to deduce if the effect on endothelial inflammation was dependent on kringle repeat number, if the burden of ox-PL on Lp(a) particles changed with apo(a) length, and if the biologic effect of Lp(a) on CHD is dependent on this parameter.

Besides apoB and apo(a), which constitute ~80% of its proteome, Lp(a) associates with over 35 additional proteins with different functions, from lipid metabolism to wound healing to immune response¹². The wound healing properties of Lp(a) likely provided an evolutionary advantage in humans, and the protein cargo associated with this biological function is represented by coagulation factors, complement activation cascade proteins, and immune response proteins. The background proteome may be critical in driving Lp(a) to the site of endothelial injury. In the few proteomic studies published, no evaluations were made of proteome variations based on isoform size ¹². Conformational changes driven by proteome and kringle repeats may alter ox-PL content, thus affecting the atherogenicity of Lp(a). In future studies, we need to identify whether apo(a) isoform-specific protein signatures attract ox-PL cargo in terms of type or amount. Identification of the interplay between apo(a) isoforms, protein cargo, and ox-PL content on endothelial inflammation and leukocyte extravasation may lead to novel therapeutics going beyond the intuitive and still unattained target of reducing Lp(a) levels.

Lp(a) contributes to CHD risk mainly when its plasma concentrations are above 30 mg/dL, which is the case for ~30% of the population. This means that about 70% of people at risk have low Lp(a) levels, and therefore their CHD risk is driven by other factors and co-morbidities in the absence of a contribution by Lp(a). This is a different scenario than that of LDL, which always contribute to CHD risk and is always a target of therapy. This raises questions on the larger meaning of the results from the paper by Schnitzler et al., as, for example, statins lower LDL levels, reduce inflammation, and reduce CVD risk but do not lower Lp(a) levels. On the other hand, PCSK9 inhibitors lower both LDL and Lp(a), reduce CHD risk, but do not affect inflammation¹³. It must be noted that none of the PCSK9 inhibitor clinical trials have reported on baseline or treatment levels of plasma ox-PL.

To summarize, the ox-PL pool in plasma is primarily associated with Lp(a) and plasminogen. While 70% of the population has low plasma Lp(a) levels, plasma plasminogen levels are likely to be the primary modulators of the plasma ox-PL pool in most subjects. Plasma levels of ox-PL plasminogen are lower in survivors of an MI than in those with stable CAD¹⁵. It remains to be determined if the expected vascular benefits of Lp(a) reduction are mediated by its ox-PL cargo.