New Therapeutic Targets for Calcific Aortic Valve Stenosis

Abstract

Calcific aortic valve stenosis (CAVS) is the most common form of acquired valvular heart disease, present in 3% of the population more than 75 years of age (1). Although risk factors are similar for CAVS and atherosclerosis, ~50% of patients with CAVS do not have clinically significant cardiovascular disease (CVD), suggesting related, but unique, pathophysiology (1). Although surgical aortic valve replacement remains the gold standard treatment for most patients, at least one-third of symptomatic patients with CAVS may not undergo surgical aortic valve replacement.

To fill this clinical need, transcatheter aortic valve replacement is increasingly being used, but overall survival remains modest due to the advanced age and other comorbidities. With the aging of the population, the prevalence of CAVS will increase rapidly and portends medical, financial, and ethical burdens to healthcare systems worldwide. Hence, identification of causal pathways mediating CAVS may provide novel targets for earlier therapy before end-stage disease. One of these pathways may involve the lipoprotein(a) (Lp[a]), lipoprotein-associated phospholipase A₂ (Lp-PLA₂), and oxidized phospholipids (OxPL) axis (2).

Lp(a) is composed of apolipoprotein B-100 (apoB) covalently bound to apolipoprotein(a) [apo(a)]. On the basis of recent meta-analyses, genome-wide association study (GWAS), and Mendelian randomization studies, Lp(a) is now generally considered a causal risk factor for CVD (3). Recently, Thanassoulis et al. (4) showed in a GWAS study of multiple ethnic groups that genetic variation at the LPA gene locus, as manifested by single nucleotide polymorphism (SNP) rs10455872, was causally related to CAVS. Lp(a) is proatherogenic because of its low-density lipoprotein (LDL) moiety and through additional mechanisms mediated by apo(a) (3,5,6). A key proinflammatory property of Lp(a) may be its content of OxPL (7,8). Among lipoproteins, OxPL is predominantly found on Lp(a), with only small amounts on LDL and high-density lipoprotein (HDL) (8,9). A large clinical database supports the hypothesis that the risk of Lp(a) is driven by its content of OxPL, mainly present on small apo(a) isoforms associated with high Lp(a) levels (7). Elevated levels of OxPL on apoB (OxPL/apoB) predict death, myocardial infarction, and stroke, reclassify ~30% of patients to different risk categories, and reflect extent of CVD in multiple arterial beds (10).

Lp-PLA₂ hydrolyzes OxPL to yield a free oxidized fatty acid (OxFA) and lysophosphatidylcholine (11). Both the precursor (OxPL) and the 2 byproducts of Lp-PLA₂ manifest proinflammatory effects. It has not been established which is more clinically relevant, and clinical trials are being performed to assess whether inhibiting breakdown of OxPL with the Lp-PLA₂ inhibitor darapladib will improve clinical outcomes. Lp-PLA₂ is secreted by inflammatory cells and circulates primarily on LDL, but is also present on HDL and Lp(a), increasing proportionally on Lp(a) with higher Lp(a) levels (9). On average, it has been estimated that only 1 of 100 LDL particles carries Lp-PLA₂ (11). However, on an equimolar basis, Lp(a) has a 1.5-fold to 2-fold higher mass of Lp-PLA₂ and as much as a 7-fold higher specific-activity than LDL (12,13). Lp-PLA₂ mass and activity independently predict CVD events, and this risk is approximately doubled when Lp(a) and/or OxPL/apoB are also elevated (14), suggesting a common pathophysiological link. Interestingly, Lp-PLA₂ on HDL is associated with lower CVD risk (15), whereas on LDL, it is associated with higher risk. Overall, these observations suggest plausible mechanisms through which Lp(a)–Lp-PLA₂–OxPL may mediate CAVS, as has been suggested for atherosclerosis and cardiovascular disease (2). In this issue of the Journal, 2 reports provide such evidence (16,17).

In the first paper, Kamstrup et al. (16) confirm and extend the findings of Thanassoulis et al. (4) by showing that elevated Lp(a) levels (>74 mg/dl, representing ~10% to 15% of the general population) were independently associated with CAVS, predicting a 3-fold increased risk. Consistent with genetic causality, carriers of LPA SNPs rs10455872 and rs3798220 and of low number of KIV-2 repeats, all of which are associated with higher Lp(a) levels, were also associated with CAVS. The enhanced risk was driven by high Lp(a) levels and not an independent risk of the LPA SNPs. Lp(a) levels fluctuate minimally throughout life within pre-set genetically determined levels and are mediated almost exclusively by the LPA gene, including various LPA SNPs and variability in KIV-2 repeats. The relatively unchanging levels of Lp(a) among populations are akin to randomization in a trial with a therapy that results in significant on-treatment differences in Lp(a) levels.

In the second paper, Mahmut et al. (17) provide physical evidence that Lp-PLA₂ may play a role in CAVS. They evaluated 40 surgically excised stenotic aortic valves and 20 noncalcified aortic valves obtained after heart transplantation, and documented higher expression of Lp-PLA₂ at the transcription, protein, and activity level. Presumably, Lp-PLA₂ in the valves was derived from local secretion by inflammatory cells and also transported by LDL and Lp(a). Lp(a) has strong lysine binding sites and, after entering through disruptions in the endothelial layer of the valve leaflets, may accumulate and promote inflammation through several pathways (1,3). Mahmut et al. (17) further documented histological colocalization with OxLDL and lysophosphatidylcholine, and noted modest correlations of Lp-PLA₂ with indices of valve remodeling and hemodynamic severity of CAVS. Although this study is done well overall, several caveats should be borne in mind. One, it cannot be determined whether these observations are cause or consequence of CAVS, and trials to inhibit Lp-PLA₂ and assess effect on CAVS will be needed. Based on the fact that the gene that encodes Lp-PLA₂, PLA2G7, was not apparently associated with CAVS in the recent GWAS study by Thanassoulis et al. (4) suggests it may not be a causal mechanism. Nonetheless, it may still play a role in CAVS progression to clinical symptoms. Two, the correlations of CAVS Lp-PLA₂ expression with plasma Lp-PLA₂ and OxLDL were not adjusted for LDL cholesterol, Lp(a), HDL cholesterol, or apoB levels and may simply reflect higher levels of such particles, with their cargo of Lp-PLA₂, entering valve tissue from the circulation. Three, the antibody used in the OxLDL assay binds a conformational epitope of apoB rather than OxPL. Therefore, it does not necessarily reflect OxPL plasma levels, which would be important to link it to Lp-PLA₂ pathophysiologically. It is also underappreciated that this antibody cross reacts with apoB at levels widely present in humans (18) and may not measure, only OxLDL, but also apoB in plasma, as has been shown by experiments showing that samples of known content of OxLDL when spiked with unoxidized LDL result in a proportional increase in OxLDL levels (19). And four, most patients had bicuspid aortic stenosis, and it would be important to confirm such findings in a larger number of patients with tricuspid CAVS. However, it is possible that Lp(a)-Lp-PLA2-OxPL may be particularly important pathophysiologically in bicuspid aortic valve stenosis.

Valvular calcification precedes the development of CAVS and is actively regulated. Oxidative stress, calcific nodules, and inflammatory infiltrates play a significant role in CAVS (1), and the accumulation of OxLDL is associated with increased inflammation and may potentiate calcification and matrix remodeling (20,21). Treatment of calcifying aortic smooth muscle cells with OxPL has been shown to stimulate expression of alkaline phosphatase and formation of cellular aggregates containing calcium mineral, characteristic features of osteoblastic differentiation (22,23). Recently, OxLDL and Lp(a) have been shown to bind monocyte chemoattractant protein-1, which may play a role in attracting monocytes to subendothelial spaces (24).

There is currently no medical therapy to prevent or reduce the progression of CAVS in humans. Studies in hypercholesterolemic Ldlr−/− mice have shown that CAVS develops and that lowering cholesterol earlier in the course of disease progression will retard aortic stenosis, but if begun much later, will not influence the extent of stenosis (25,26). Retrospective human studies have suggested statins are effective in reducing progression. However, prospective randomized human trials, including the SALTIRE (Scottish Aortic Stenosis Lipid lowering Therapy Impact on Regression study, atorvastatin 80 mg/dl), SEAS (Simvastatin and Ezetimibe in Aortic Stenosis study, (simvastatin 40 mg/dl plus ezetimibe 10 mg/dl) and the ASTRONOMER (Aortic Stenosis Progression Observation: Measuring Effects of Rosuvastatin study, rosuvastatin 40 mg/dl), have shown no significant effect on CAVS progression. (1). There are several possibilities for the negative results, including that treatment of elderly patients studied in these trials, who already had moderate to heavily calcified valves at the end stage of the disease, was too late to influence disease outcomes. Therefore, therapeutic agents beyond standard lipid-lowering therapy are needed to have an impact on CAVS.

The preceding observations suggest that targeting the Lp(a)–Lp-PLA₂–OxPL pathways may be a viable approach in mitigating CAVS. That could be done by several currently available approaches targeting Lp(a), including antisense oligonucleotides specifically directed to apo(a) of Lp(a) currently in phase 1 trials, which lower Lp(a) levels >85% (27), inhibitors of Lp-PLA₂ such as darapladib currently being evaluated in phase 3 trials in patients with acute coronary syndromes and stable coronary artery disease, and therapies to prevent oxidation of lipoproteins or minimize their proinflammatory effects, such as sufficient antioxidants, oxidation-specific antibodies targeting OxPL, and other immunomodulatory agents (28,29). It also sets the stage for the discovery of new therapeutic agents.

In summary, these studies significantly enhance our understanding of the development of CAVS by highlighting the potential role of the Lp(a)-Lp-PLA₂–OxPL axis as mediators of CAVS and suggesting new therapeutic targets. These targets can initially be validated in animal models and ultimately in randomized clinical trials.