Lipoprotein(a) and Oxidized Phospholipids in Calcific Aortic Valve Stenosis

Abstract

Purpose of review

Calcific aortic valve stenosis (AS) affects over 1 million patients in the US. Effective medical therapies do not exist. With the aging of the population and increase in incidence of AS, improved understanding of this disease and novel therapies to reduce the progression of AS and need for aortic valve replacement (AVR) are urgently needed.

Recent findings

Lipoprotein(a) [Lp(a)] is the only monogenetic risk factor for AV calcification and AS. Elevated Lp(a) levels are a strong, causal, independent risk factor for AS, as demonstrated in epidemiological, genome-wide association studies and Mendelian randomization studies. Lp(a) is the major lipoprotein carrier of oxidized phospholipids (OxPL) which are pro-inflammatory and promote calcification of vascular cells, two key pathophysiological drivers of AS. Moreover, Lp(a)-associated enzymes and lipids derived from breakdown of OxPL have been implicated in the pathogenesis of AS. These mechanistic insights likely explain the recent findings that elevated plasma Lp(a) and OxPL on apoB containing lipoproteins (OxPL-apoB) predict progression of pre-existing AS and need for AVR. The failure of the statin trials in AS may be partially explained by the fact that statins, as monotherapy or in combination with ezetimibe, have neutral or Lpa(a) raising effects, as has been shown most recently in the ASTRONOMER trial. Antisense oligonucleotides targeted to apo(a) are in Phase 2 clinical development and shown to lower both Lp(a) and OxPL-apoB.

Summary

Lp(a) and OxPL are key therapeutic targets in AS. Strategies aimed at potent Lp(a) lowering to normalize levels and/or suppress the pro-inflammatory effects of OxPL may be beneficial for preventing progression of AS and need for AVR.

Introduction

Calcific aortic valve stenosis (AS) is a progressively debilitating and potentially fatal disease that is currently treated with surgical aortic valve replacement (AVR) in patients who can tolerate surgery or with transcatheter aortic valve replacement in those with high surgical risk. It is estimated that there are over 1 million patients with AS in the US and over 2.5 million worldwide [1]. Affected aortic valve (AV) leaflets are characterized by progressive fibrosis, thickening and most importantly, calcification [2, 3], resulting in decreased leaflet mobility and increased obstruction of blood outflow from the left ventricle. Progressive worsening of AS occurs in the majority of patients, eventually leading to symptoms, such as angina, syncope and heart failure requiring AVR. Approximately 50% of patients are not eligible for AVR due to co-morbidities, and generally develop progressive heart failure and death. Effective medical therapies do not exist, and statins have failed to reduce progression of AVR in four randomized trials [4–7]. Novel therapies to reduce the progression of AS are urgently needed, since as the population ages there will be a corresponding progressive increase in incidence and prevalence of AS. We review the literature examining the role of Lp(a) as a risk factor for AS in clinical studies, as well as evidence describing the mechanisms by which its oxidized phospholipids (OxPL) content may mediate progression of AS.

Lipoprotein(a) is a risk factor for AS

Epidemiology Studies

Lipoprotein(a) [Lp(a)] was initially identified as a risk factor for AV disease based on the results of several epidemiologic and observational studies (summarized in Table 1). Cross-sectional studies have found enrichment of individuals with elevated Lp(a) in those diagnosed with AV sclerosis or stenosis by echocardiography [8, 9]. Case-control studies revealed elevated Lp(a) was an independent risk factor for AS among patients who had AVR [10] and for AV calcification in consecutive patients who had outpatient echocardiograms [11]. Data from the Copenhagen City Heart Study (CCHS) and Copenhagen General Population Study (CGPS) showed an observational hazard ratio (HR) of aortic stenosis of 1.4 (95% confidence interval (CI) 1.2–1.7) for a 10-fold increase in Lp(a) plasma levels [12]. Another analysis of the same cohorts examined both causal and observational associations between AS and Lp(a) and reported an observational HR for AS of 1.23 (CI 1.06–1.41) for each SD increase in Lp(a) [13]. The European Prospective Investigation into Cancer and Nutrition (EPIC)-Norfolk study found that after adjusting for sex, age, smoking, and low density lipoprotein cholesterol levels, Lp(a) ≥ 50 mg/dL (~125 nmol/L) was associated with a HR of 1.98 (CI 1.25–3.09) for risk of AS, compared with Lp(a) < 50 mg/dL(~125 nmol/L) [14]. In patients with heterozygous familial hypercholesterolemia, Lp(a) was a predictor of AV calcification with an odds ratio per 10 mg/dL (~25 nmol/L) increase in Lp(a) of 1.11 (CI 1.01–1.20) after multivariate analysis including lifelong LDL-C burden [15]. A recent secondary analysis of the prospective Aortic Stenosis Progression Observation: Measuring Effects of Rosuvastatin (ASTRONOMER) trial involving patients with mild to moderate AS identified elevation of Lp(a) within the top tertile (> 58.5 mg/dL or ~146 nmol/L) as an independent risk factor for the rate of progression of AS and occurrence of AS related events (AVR, death), especially in patients ≤ 57 years [16]. In summary, multiple epidemiological studies provide strong support that elevated Lp(a) levels are associated with AS. Epidemiologically, the risk of AS seems to start at Lp(a) levels of ~30 mg/dL (~75 nmol/L) but substantial risk does not accrue until Lp(a) levels are ~>60 mg/dL (~>150 nmol/L).

Table 1.

Epidemiologic Studies of Lp(a) as a Risk Factor for Aortic Valve Disease

Author	Year	Study Design	Size	Key Findings
Gotoh et al. [8]	1995	Cross-Section	784	AV sclerosis present in 36.1% of those with Lp(a) > 30 mg/dL and 12.7% of those with Lp(a) < 30 mg/dL.
Stewart [9]	1997	Cross-Section	5,201	Lp(a) elevation is a risk factor for AV sclerosis or stenosis (OR 1.23, CI 1.14–1.32).
Glader [10]	2003	Case-Control	101 cases, 101 controls	Lp(a) ≥ 48 mg/dL present in 21/101 cases of AS resulting in AVR vs 5/101 controls (OR 5, CI 1.7–14.6).
Bozbas [11]	2007	Case-Control	285 (112 with AV calcification)	Lp(a) levels are higher among those with AV calcification (AVC) (27.4 mg/dL vs 19.9 mg/dL without AVC; p = 0.033).
Kamstrup [12]	2014	Cohort	77,680	Lp(a) levels > 95th percentile (> 90 mg/dL), associated with HR for AS of 2.9 (CI 1.8–4.9); overall observational HR of AS 1.4 (CI 1.2–1.7) for a 10-fold increase in Lp(a) plasma levels.
Arsenault [14]	2014	Cohort, with case-control replication	17,553	Lp(a) ≥ 50 mg/dL is associated with increased risk of AV stenosis, adjusted HR 1.98 (CI 1.25–3.09, p = 0.002), compared with Lp(a) < 50 mg/dL.
Capoulade [16]	2015	Cohort	220	Lp(a) elevation > 58.5 mg/dL is an independent risk factor for increased rate of AS progression, especially in those ≤ 57 years.
Langsted [13]	2015	Cohort	100,578	Observational HR for AV stenosis for an increase in 1 SD of Lp(a): 1.23 (CI 1.06–1.41).
Vongpromek [15]	2015	Cross-Section	129 (HeFH)	OR (per 10 mg/dL increase in Lp(a)) of AV calcification: 1.11 (CI 1.01–1.20).

Genetic Studies

In addition to epidemiologic studies describing a relationship between Lp(a) and AS, a genetically determined relationship exists between Lp(a) levels and calcific AV disease (summarized in Table 2). The single nucleotide polymorphism (SNP) rs10455872 in the gene LPA, encoding for apo(a), was found in a Genome Wide Association Study (GWAS) to be strongly linked to AV calcification as determined by computed tomography. This finding was reported by the Cohorts for Heart and Aging Research in Genome Epidemiology (CHARGE) consortium using data, in the discovery phase of their analysis, from three large cohorts: The Framingham Heart Study (FHS), the Age, Gene/Environment Susceptibility-Reykjavik Study (AGES-RS), and white European participants from the Multi-Ethnic Study of Atherosclerosis (MESA), a total of 6,942 participants. After multivariate adjustments, the SNP rs10455872 (risk allele (G)) was associated with an increased risk of AV calcium with an odds ratio of 2.05 (CI 1.63–2.57). Mendelian randomization analysis within the same study, demonstrated that the LPA SNP rs10455872, associated with elevated plasma Lp(a) and AV calcification in whites and blacks, was a genetically determined risk factor for AV calcification (OR 1.62 (CI 0.94–2.81) per log-nmol/L increase in Lp(a)) and that this risk was likely mediated by the biological activity of Lp(a). Using data from the Malmö Diet and Cancer Study (MDCS) (n = 28,193) and CHS (n = 10,400), an independent association between rs10455872 and incident AS was also demonstrated, with a HR of 1.68 (CI 1.32–2.15) per risk allele reported for the MDCS cohort and a HR of 1.6 (CI 1.12–2.28) for the presence of at least one risk allele reported for the CCHS cohort. Importantly, additional analysis demonstrated that the association of rs10455872 with AV calcium was independent of coronary artery calcium and clinical coronary artery disease [17].

Table 2.

Genetic Studies of Lp(a) as a Risk Factor for Aortic Valve Disease

Author	Year	Study Design	Size	Key Findings
Thanassoulis [17]	2013	GWAS, cohort	6,942 (AVC analysis), 28,193 (MDCS analysis of AS), 10,400 (CCHS analysis of AS)	SNP rs10455872 in LPA is associated with AVC (OR per G allele 2.05 (CI 1.63–2.57, p = 9.0×10−10)) and with incident AS: HR 1.68 per risk allele (CI 1.32–2.15; p = 3×10−5) in MDCS, HR 1.6 for presence of at least one risk allele (CI 1.12–2.28, p = 0.01) in CCHS.
Kamstrup [12]	2014	Cohort	77,680	Relative risk of AV stenosis of 1.6 (95% CI 1.2–2.1) for a 10-fold increase in genetically determined Lp(a) values (includes rs10455872, rs3798220, and KIV-2 repeats).
Arsenault [14]	2014	Cohort, with case-control replication	17,553	Risk of AV stenosis increases with number of rs10455872 G alleles: 1 allele, unadjusted HR 1.78 (CI 1.11–2.87); 2 alleles, unadjusted HR 4.83 (CI 1.77–13.20). Case-control: rs10455872 associated with AV stenosis, OR 1.57 (95% CI 1.10–2.26).
Langsted [13]	2015	Cohort	100,578	Causal risk ratio (CRR) for AS associated with LPA SNPs (rs10455872 and rs3798220) for 1 SD increase in Lp(a): 1.38 (CI 1.23–1.55). For LPA KIV-2, CRR for AS for 1 SD increase in Lp(a): 1.21 (CI 1.06–1.40).

A Mendelian randomization analysis with LPA SNPs was performed using clinical data containing individual diagnoses for AS from the CGPS and CCHS (n = 77,680). Median Lp(a) levels varied with SNP rs10455872 carrier status (11 mg/dL or ~27.5 nmol/L for non-carriers, 60 mg/dL or ~150 nmol/L for heterozygotes, and 108 mg/dL or ~270 nmol/L for homozygotes). Lp(a) levels were also increased in carriers of SNP rs3798220 and with decreasing numbers of LPA kringle IV type 2 (KIV-2) repeats. Carriers of SNP rs10455872 had a multivariable adjusted HR of 1.6 (CI 1.2–2.0) for heterozygotes and 1.5 (CI 0.5–4.8) for homozygotes (trend, p < 0.001) compared to non-carriers. The rs3798220 SNP was also associated with elevated Lp(a), but its prevalence was too low to rule out whether an association was also present with AS. When including all LPA genotypes (i.e. rs10455872, rs3798220, and KIV-2 percentile group), the authors reported a relative risk of AS of 1.6 (CI 1.2–2.1) for a 10-fold increase in genetically determined Lp(a) values [12]. Using a multidirectional Mendelian randomization design, a study with 100,578 subjects reported a causal risk ratio for AS of 1.38 (CI 1.23–1.55). for each SD increase in Lp(a) based on LPA SNPs (rs10455872 and rs3798220) and 1.21 (CI 1.06–1.40) for each SD increase in Lp(a) associated with LPA KIV-2 genotype [13]. Within the EPIC-Norfolk study (n= 17,553) described above, carriers with one rs10455872 G allele had an unadjusted HR for AS of 1.78 (CI 1.11–2.87), and those with two G alleles had an unadjusted HR for AS of 4.83 (CI 1.77–13.20), compared to non-carriers [14].

In summary, 9 unique clinical studies demonstrated the predictive value of Lp(a) towards risk of developing AS and 4 large genetic studies using Mendelian randomization demonstrated that Lp(a) is a genetically determined, and therefore likely causal mediator of AS. With this in mind, reduction of Lp(a) concentrations, unlike reduction of low-density lipoprotein–cholesterol (LDL-C) with statins, is a promising novel therapeutic approach for AS.

Statins and Lp(a)

Though LDL-C is also a risk factor for AS, LDL-C lowering by statins has not been successful in altering the natural history of the disease. A recent Mendelian randomization analysis of 6942 subjects from the CHARGE consortium and the MDCS demonstrated that a higher LDL-C genetic risk score (GRS), consisting of SNPs specifically associated with LDL-C and weighed by each SNP’s correlation to plasma LDL-C levels, was associated with increased risk of incident AS [18]. Although the LDL-C GRS did not correlate with Lp(a) mass, all conventional measurements “LDL-C” contains the content of Lp(a)-cholesterol (Lp(a)-C), which can be 30–45% of Lp(a) mass [19], and may be a confounding variable in this analysis. Despite the role of LDL-C as a risk factor for AS, and small (n=121) open-label trial demonstrating reduced progression of AS in patients treated with rosuvastatin [20], four randomized placebo-controlled LDL-C lowering trials have been conducted in well established AS with rosuvastatin (ASTRONOMER [6] and Tyrolean Aortic Stenosis Study (TASS) [20]), atorvastatin (Scottish Aortic Stenosis and Lipid Lowering Trial, Impact on Regression (SALTIRE) [4]), and simvastatin/ezetimibe (Simvastatin and Ezetimibe in Aortic Stenosis (SEAS) [5]). These collectively enrolled 2344 patients ages 58–68 years with mild-moderate AS (ASTRONOMER, TASS, and SEAS) or moderate-severe AS (SALTIRE), and all failed to show reduction in AS progression [21].

One possible explanation for the ineffectiveness of statins in these well-conducted trials is that LDL-C lowering must be initiated at an earlier age or stage of the disease (i.e. aortic sclerosis) in order to prevent AS [22]. However, it is also likely that Lp(a), instead of LDL-C, is the major modifiable risk factor in AS. Using the latter paradigm, it is not surprising that statins were not beneficial for AS. Lp(a), compared to LDL, is relatively refractory to statin therapy. The LPA SNP rs10455872, which is associated with higher Lp(a) levels, has been associated with poorer response to LDL-C lowering by statins [23–25]. As the Lp(a) mass rises, the contribution of Lp(a)-C to measured LDL-C also increases [19], therefore Lp(a) represents a pool of “LDL-C” which is resistant to statin therapy and contributing to AS risk and can confound interpretation of AS risk due to LDL-C. Studies on the precise effects of statins on Lp(a) are controversial, with differences in assay quality compromising the ability to perform robust meta-analyses [26, 27]. In certain statin trials, including those with rosuvastatin, atorvastatin, pravastatin, pitavastatin, and simvastatin/ezetimibe, Lp(a) increases 10–20% with statin therapy [28] (Figure 1). Specifically, in the ASTRONOMER trial Lp(a) increased by 20% (p<0.05) from baseline levels in the 112 out of 220 patients randomized to rosuvastatin and treated for 1 year [16]. Lp(a) increased by 23% (p<0.01) from baseline levels in hyperlipidemic patients treated with simvastatin/ezetimibe [28], which was also the therapeutic intervention studied in SEAS [5]. These findings suggest that the benefits of LDL-C lowering by statins towards aortic stenosis may be offset by a deleterious increase in Lp(a).

Figure 1. Effects of therapeutic interventions on OxPL-apoB and Lp(a) levels.

Systematic review of trials with OxPL-apoB and Lp(a) levels following intervention with statins (brown), beneficial diets (blue), and Lp(a) lowering therapies (black). Each filled symbol represents the mean percent change, or the delta mean percent change between the intervention and placebo group where available, from each respective trial. Diamond symbols represent the mean change within each respective interventional category and span the 95% confidence interval. Data from trials with larger of subjects are represented with larger symbol. Reprinted with permission from J Clinical Lipidology [28].

Mechanisms for Lp(a) and OxPL towards AV calcification and the pathogenesis of AS

Early lesions on AV leaflets in AS contain oxidized lipids, apoB and apo(a), macrophages and T-cells [29–32]. Apo(a), via its lysine binding domains can bind to fibrin on denuded or injured endothelium [33–35], such as that on AVs subjected to mechanical stress in vivo, as a likely mechanism for intravasation of Lp(a) into the valve. Thereafter, OxPL present on Lp(a) can mediate pro-inflammatory and pro-calcific pathways potentiating AS (Summarized in Figure 2).

Figure 2. Potential mechanisms for the causal role of Lp(a) and OxPL in AS.

(A) Molecular changes involved in progression of AS. Fibrin is exposed at sites of aorta endothelial injury (depicted by lightning bolt) and can bind to Lp(a), leading to its retention in the valve. Pro-inflammatory lipids on Lp(a), such as OxPL, can promote calcification and bone formation via VIC directly or via up-regulation of ROS and pro-inflammatory cytokines in macrophages. (B) Anatomic changes with progression of a normal valve (left) to a severely stenotic valve (right).

Abbreviations: ATX = autotaxin; Lp-PLA₂ = Lipoprotein-associated phospholipase A₂; LPC = lysophosphatidylcholine; LPA = lysophosphatidic acid; OxPL = oxidized phospholipids; Lp(a) = lipoprotein(a); ROS = reactive oxygen species; VIC = vascular interstitial cell; ALP = alkaline phosphatase

OxPL is covalently bound to the apo(a) moiety of Lp(a), but is also present in the lipid-phase of the particle [36]. In humans, more than 85% of plasma lipoprotein-associated OxPL are bound to Lp(a) while the remaining exist primarily on apoB containing lipoproteins [37–39]. A second large and independent pool of OxPL present on other proteins is covalently linked to plasminogen, which has no other lipid content [40, 41]. Consistent with the role of this apoB-100 containing lipoprotein as the preferential lipoprotein carrier of OxPL, OxPL on plasma apoB (OxPL-apoB) levels correlate with Lp(a) levels in the population as a whole. However, this is dependent on genetics and underlying apo(a) isoform composition, irrespective of race [42–45]. Further genetic evidence that Lp(a) levels determine plasma OxPL levels include the high genetic covariance of Lp(a) and OxPL-apoB in a study examining these parameters in monozygotic versus dizygotic twins [43], and that the LPA SNPs rs3798220 [45] and rs10455872 [43] are also associated with elevated levels. Lastly, in trials with pharmacologic [16, 38, 46–49] or dietary [50–52] interventions that raise Lp(a), OxPL-apoB is increased as well. Conversely, lipid apheresis [37], and antisense oligonucleotides to apo(a) [53] and niacin [28], all lower OxPL-apoB.

OxPL play a central role in the development of atherosclerosis, which shares a similar pathogenesis with AS, particularly in pro-inflammatory pathways [54, 55]. Using E06, a monoclonal IgM natural antibody that binds to the PC head group of oxidized but not native PC-containing phospholipids [56–58], we have developed and clinically validated an ELISA that detects OxPL on plasma apoB [57, 59], which primarily detects OxPL on Lp(a). Because most of the OxPL in human plasma is bound to Lp(a), the OxPL-apoB assay thus detects the most inflammatory and atherogenic Lp(a). In over 40 publications, we have shown that elevated OxPL-apoB levels predict death/MI/stroke in unselected populations followed prospectively [60, 61], correlate with endothelial dysfunction and progression of coronary calcification [51, 62], predict the progression of femoral/carotid disease [63], coronary artery disease (CAD) [64], and are elevated in patients with ACS [65] and following PCI [66].

Oxidized lipoproteins, which are highly enriched in OxPL [67–69], have been implicated in promoting inflammation, valvular ectopic calcification and bone formation, features that are pronounced in severe AS. OxLDL potentiates macrophage-derived reactive oxygen species (ROS) and cytokine production (including CXCL1, CXCL2, CCL9, CCL5, and IL-1β) via toll-like receptors TLR-2/4 and NFκB signaling, leading to increased oxidative stress and inflammation that can accelerate AS [70–73]. Specifically, monocytes exposed to OxPL on apo(a) up-regulate expression of the inflammatory cytokine IL-8 [74].

Bone formation within the diseased AV is driven by the differentiation of vascular cells into osteoblasts [29] via bone morphogenic protein (BMP) signaling and up-regulation of osteoblastic transcription factors including RUNX2 and MSX2. BMP2 [69, 75–77] as well as RUNX2 [78] and MSX2 [79] expression in vascular cells, including AV derived vascular interstitial cells (VIC), are up-regulated following exposure to OxLDL. OxLDL exposure also suppresses osteoprotegerin [80], an inhibitor of vascular calcification via the RANK-L pathway [81, 82]. Finally, exposure to OxLDL in vitro stimulates extracellular matrix calcium deposition by vascular cells [78, 79, 83, 84] as well as up-regulation of alkaline phosphatase (ALP) [67, 79, 83, 85, 86] promoting mineralization.

In addition to OxPL, Lp(a) associated enzymes and lipids have been implicated in the pathogenesis of calcific AS. Lipoprotein-associated phospholipase A₂ (Lp-PLA₂), enriched on Lp(a) [87] hydrolyzes OxPL to yield a free oxidized fatty acid and lysophosphatidylcholine (LPC). Lp-PLA₂ expression and activity levels are higher in mineralized AVs compared to controls [88] and Lp-PLA₂ co-localizes with Lp(a) in human calcific AVs [89]. Moreover, LPC promotes mineralization of VICs in culture [89]. Autotaxin (ATX), another enzyme present on Lp(a), hydrolyzes LPC to form lysophosphatidic acid (LPA), and has also been found in diseased valves in tight proximity to apo(a) [89]. Administration of exogenous LPA to a mouse model of AS resulted in higher peak velocities across the AV and more AV calcification compared with controls [89]. These findings consistently support the role of Lp(a) as a vehicle for harmful substrates that promote AS. Whether Lp(a), its associated OxPL or its metabolites will be viable therapeutic targets to mitigate AS remains to be determined.

Lp(a) and OxPL are Associated with Increased AS progression

Elevated Lp(a) and OxPL-apoB were predictive of a worse outcome in the ASTRONOMER trial, originally designed to evaluate the role of rosuvastatin in preventing the progression of AS in those with mild-moderate AS followed for 3.5 years [6]. In a prospective analysis of 220/269 subjects from the original cohort with baseline Lp(a) and OxPL-apoB measurements, those with the highest tertile of Lp(a) (>58.5 mg/dl or ~146 nmol/L) had AS which progressed ~1.5 times faster (average peak velocity ± SD) (0.26 m/s/yr ± 0.03 vs 0.17 m/s/yr ± 0.02) [Figure 3] and had ~2 fold increased risk of a composite outcome of AVR (n = 47 in the cohort) and cardiac death (n = 2 in the cohort) [16]. This relationship was the same between individuals with tricuspid or bicuspid valves (consisting of 48% of subjects in this trial), but the risk of high Lp(a) on AS progression and AVR was more pronounced in those younger than 57 years of age. The risk of AS progression conveyed by elevated Lp(a) levels in this cohort of subjects starting out with mild-moderate disease actually more closely reflects individuals with moderate-severe AS (Figure 3). Of great relevance, those with the highest tertile of OxPL-apoB (>5.50 nM) as well as OxPL on apo(a) [OxPL-apo(a)] (>33.5 nM) had increased risk of AS progression and AVR which was nearly identical to those with the highest tertile of Lp(a), consistent with the thesis that OxPL carried by Lp(a) is responsible with the biological activity of the particle.

Figure 3. Rates of AS progression in stain trials.

Annual rates of progression in those with mild-moderate (mod) AS (composite estimate using data from ASTRONOMER [6], SEAS [5], and TASS [20]), those with mild-mod AS and low Lp(a) and high Lp(a) [orange bars] (data from Capoulade et al. [16]), and those with severe AS (estimated using data from SALTIRE [4]).

Future directions

There is a growing body of literature describing the role of Lp(a) as a genetically determined, causal risk factor for AS. However, two major questions linger: 1) what are the precise mechanisms linking Lp(a) to progression of AS? and, 2) will Lp(a) lowering attenuate AV calcification and the progression of AS?

We are only beginning to understand the mechanisms by which Lp(a) promotes the progression of AS. The relative contributions of lipids associated with Lp(a) implicated in AS, namely OxPL, PLC, and LPA remain to be characterized. One approach towards this question would be to quantify and compare OxPL, PLC, and LPA levels in human stenotic AVs and from human AVs without significant disease (i.e. from explanted hearts from patients undergoing cardiac transplant) using liquid chromatography-mass spectrophotometry (LC-MS). This could be complemented by quantification of OxPL, PLC, and LPA on purified plasma Lp(a) in comparison to other lipoproteins such as LDL, also feasible by LC-MS, from subjects with AS enrolled in long term prospective studies tracking progression of the disease.

Whether Lp(a)-associated OxPL or its metabolites are directly responsible for calcific AS can be tested in vivo using animal models. To date, the only animal models that develop clinically significant calcific AS are aged transgenic mice on an apoB^{100 only}/LDLR^−/− background, fed a high cholesterol diet [89–91]. Based on these existing models, we have generated transgenic hyperlipidemic Lp(a) mice which express human apoB-100 and apo(a) [92] on an apoB^{100 only}/LDLR^−/− background. These models, along with transgenic models resulting in varying OxPL and/or Lp(a) levels in plasma may be useful in studying mechanisms of calcific AS development and progression.

Trials evaluating the effect of Lp(a) lowering on AS progression are currently underway. Niacin [93], CETP inhibitors such as anacetrapib [94], PCSK9 inhibitors including alirocumab and evolocumab [19] all lower plasma Lp(a), though only moderately (by 20–40%), and have diverse metabolic effects including alterations in LDL-C, HDL-C and triglycerides. Participants are currently being recruited for a randomized, double-blind placebo control trial designed to assess the effect of extended release niacin (1.5–2 g/day) on aortic valve disease progression over 2 years in those > 50 years and < 85 years with aortic sclerosis or mild AS and Lp(a) > 50 mg/dL (~125 nmol/L). The primary outcome is AV calcium score progression by cardiac CT, although the rate of hemodynamic progression of AS will be determined by echocardiography at 1 and 2 years [95]. Antisense oligonucleotides (ASO) targeted against mRNA encoding apo(a) [IONIS-APO(a)_Rx] (previously called ISIS-APO(a)_Rx), specifically lowered Lp(a) and its associated OxPL by up to 89% and 93%, respectively, in a recent Phase I trial [53]. Results from clinical trials in subjects with existing AS randomized to antisense to apo(a) or placebo will add valuable knowledge regarding the mechanism and role of Lp(a) and OxPL in AS and the efficacy of this novel therapeutic approach.

Conclusion

Lp(a) is a prevalent, genetically determined causal risk factor for calcific AS and its pro-inflammatory and pro-calcific lipid cargo, including OxPL, are likely mechanistically linked to the development of calcific AS. Elevated Lp(a) (>58.5 mg/dl or ~146 nmol/L) and OxPL-apoB (>5.50 nM) levels predict faster rate of AS progression as well as need for AVR. Whereas LDL-C lowering by statins has not been effective in altering the natural progression of AS, trials with potent Lp(a) lowering therapies including IONIS-APO(a)-L_Rx represent a promising novel therapeutic approach for AS.

Key Points.

• Lp(a) is a genetically determined, causal, independent risk factor for calcific AS.

• Lp(a) is the major lipoprotein carrier of OxPL, a set of bioactive lipids which likely contribute to AV inflammation and calcification in AS.

• Elevated Lp(a) [>58.5 mg/dl)] and OxPL-apoB (>5.50 nM) levels predict faster rate of progression of AS and need for AVR.

• Statins, which may raise plasma Lp(a) and OxPL-apoB, have not been shown to alter the natural progression of AS in clinical trials.

• Antisense oligonucleotides targeted to apo(a) potently reduce Lp(a) and OxPL-apoB levels and can be tested to reduce the progression of AS and need for AVR.