Lipoprotein oxidation biomarkers for cardiovascular risk: what does the future hold?

Keywords::

• autoantibodies
• cardiovascular disease
• LDL
• Lp(a)
• oxidation biomarkers
• oxidized phospholipids

Despite the availability of established cardiovascular risk factors, novel biomarkers for enhanced cardiovascular disease (CVD) risk prediction and determination of response to therapy are clinically needed. Patients continue to be unrecognized as being at moderate or high risk by standard measures, and currently utilized biomarkers often do not predict risk at the individual level. Novel biomarkers should be independently predictive of CVD, widely available, relatively inexpensive, and have high sensitivity and specificity for the disease process. For clinical impact, it is important that they reclassify a significant proportion of subjects either into lower or higher risk categories, in order to allow for adjustment in treatment recommendations. Biomarkers reflecting the pathophysiology of oxidized lipoproteins have developed a mature clinical database and are poised to have an important role in clinical decision making in the near future.

Indirect biomarkers of oxidized lipoproteins

The early phases in the development of biomarkers of oxidative stress consisted primarily of indirect, nonspecific and time-consuming measures, such as lag time of oxidation of LDLs exposed to copper, thiobarbituric reactive substances or lipid hydroperoxides. These gave way to the measurement of indirect biomarkers of oxidized lipoproteins, such as autoantibodies to the model oxidation-specific epitopes (OSEs) copper-oxidized LDL and malondialdehyde LDL [1]. Based on experimental and clinical data, IgG autoantibodies to OSEs appear to be atherogenic, but IgM autoantibodies confer atheroprotection [2]. Although this may seem paradoxical, recent evidence demonstrates that IgM autoantibodies to OSE represent a high proportion (up to 30%) of all circulating IgM antibodies, and often represent ‘natural’ antibodies present at birth in the germ line [3]. Such IgM antibodies represent an arc of the innate immune response and were probably selected evolutionarily to bind to and detoxify ‘danger-associated molecular patterns’ [3] present on a variety of nonself, proinflammatory OSE antigens, which are present on apoptotic cells and infectious pathogens. The OSE antigens on oxidized LDL (OxLDL) mimic those on apoptotic cells and are also recognized by autoantibodies. Since these autoantibody biomarkers are intimately involved with the atherosclerotic disease process, particularly from an immunological perspective, they have the potential to provide unique information in clinical risk prediction. However, this field has been limited by inadequately powered studies for clinical events and different methodologies for measuring levels. In particular, the lack of standardized antigens has been a major impediment in comparing studies among groups, as reported values are strongly influenced by the nature of the modified antigen used to determine the levels. In order for these measures to have a more prominent clinical role, both standardized antigens, such as chemically defined peptide mimotopes of malondialdehyde-LDL [4], and uniform methodologies for measurement, are needed. Definitive, well-powered clinical studies need to be performed to assess their independent value in predicting CVD events.

Measures of oxidized phospholipids as both biomarkers & risk factors for CVD

The oxidation of apolipoprotein B-containing lipoproteins is an important, if not obligatory, event in the formation of foam cells, the initiation of chronic inflammatory responses and the progression and destabilization of atherosclerotic lesions. Oxidized phospholipids (OxPLs) are proinflammatory mediators of atherogenesis, key components of OxLDLs and apoptotic cells, are immunogenic, accumulate in atherosclerotic lesions and contribute to both early and late CVD events [5]. Therefore, OxPLs maintain a more proximal role to the disease than LDLs and do not only reflect atherogenesis, but contribute to it.

Within the last 15 years, antibody-based ELISAs have been developed to directly measure OxPL on lipoproteins, and a large clinical database exists to assess their role in CVD risk prediction [6,7]. These assays include measuring:

• • OxPL on lipoproteins;

• • Carriers of OxPL, such as lipoprotein (a) (Lp[a]);

• • Enzymes that cleave the oxidized fatty acids from OxPLs, such as secretory phospholipase A₂ (sPLA₂) and lipoprotein-associated phospholipase A₂ (Lp-PLA₂).

These phospholipases are enzyme biomarkers of increased CVD risk and also targets of emerging therapeutic agents in Phase III trials [8].

Oxidized phospholipids on apolipoprotein B-100 assay (OxPL/apoB)

OxPL/apoB levels are measured using a double-antibody sandwich ELISA by first capturing apoB on a microtiter well plate with an apoB-specific antibody and then detecting OxPL with murine monoclonal antibody E06, binding to the phosphocholine head group of oxidized but not native phospholipids [6]. By design, OxPL/apoB levels are independent of plasma apoB and LDL-C levels, are highly sensitive to the number of OxPL epitopes on apoB and, importantly, reflect the content of OxPL present on the apoB-containing lipoprotein Lp(a). OxPL/apoB are associated with endothelial dysfunction, progression of coronary calcification, the presence and progression of coronary, femoral and carotid disease, and are elevated in ACS and percutaneous coronary intervention [6]. Importantly, elevated OxPL/apoB levels predict risk of cardiac death, myocardial infarction and stroke, as shown in the 15-year prospective Bruneck Study [9,10] and EPIC-Norfolk Study [11]. OxPL/apoB levels are either equivalent to or often independent of Lp(a), depending on the patient population. For example, they predict the presence of angiographically determined coronary artery disease [12] or CVD events, even with Lp(a) in the multivariable model [10]. OxPL/apoB levels also enhance the predictive value of the Framingham Risk Score and increase the receiver operating characteristic curve c-statistic in predicting CVD events [10]. Furthermore, OxPL/apoB, in combination with IgG and IgM autoantibodies to OxLDL, allows reclassification of approximately 33% of patients in the moderate Framingham Risk Score category, enhancing the ability to reclassify subjects into lower or higher risk categories and potentially enhancing clinical utility in a broad range of patients at risk for CVD.

This assay is currently being transitioned to be made available to the broad research and clinical communities in 2012. Areas of continued research will involve:

• • Further validation cohorts in CVD prediction, particularly in unique populations (i.e., diabetics, familial hypercholesterolemia and different racial/ethnic groups);

• • Assessing its role in predicting response to therapeutic interventions in established and emerging therapies, such as antisense oligonucleotides to apoB [13] or apo(a) [14], cholesterol ester transport protein and proprotein convertase subtilisin/kexin type 9 inhibitors;

• • Application and clinical validation in non-CVD areas of enhanced oxidative stress, such as nonalcoholic steatohepatitis, macular degeneration [15], multiple sclerosis and Alzheimer’s disease.

Lipoprotein (a)

Lp(a) is composed of liver-derived apo(a) covalently bound to apoB, and levels >25 mg/dl, representing the atherogenic threshold, are present in approximately 30% of Caucasians and 60–70% of blacks. This probably represents a worldwide prevalence of over 1 billion people. Mendelian randomization studies reflecting the random assortment of LPA gene variants have provided strong evidence that Lp(a) is an independent, causal, genetic risk factor for CVD [16,17]. In view of this, the European Atherosclerosis Society and National Lipid Association have advocated measuring Lp(a)in moderate- and high-risk patients [18,19]. Since Lp(a) levels are predominantly determined by the LPA gene locus with little environmental influence, they do not change over a lifetime, except in extreme acute-phase responses, and therefore theoretically only need to be measured once. Identifying index cases will allow screening of affected relatives, akin to cascade screening in familial hypercholesterolemia, as Lp(a) is transmitted in an autosomal dominant fashion. In the near future, the clinical evolution of Lp(a) from an emerged to an established risk factor should include the implementation of broader measurement of Lp(a) levels among populations at risk, as suggested by the European Atherosclerosis Society and National Lipid Association. This may bring about controversy, particularly among purists who would first want to see a clinical efficacy trial prior to widespread testing. However, until the development of specific agents is achieved, such a trial will not occur in the very near future. In the meantime, a very large number of patients that carry the residual risk of Lp(a) will continue to accrue CVD events. Such patients identified with high Lp(a) should have all other risk factors treated appropriately, along with low-fat dietary measures, caloric restriction, statins and niacin, which is the only practical drug that can currently lower Lp(a) (by ∼30%). It should also be appreciated that statins often raise Lp(a) levels, despite their overall clinical benefit, for unknown reasons. In the future, potent agents to specifically lower Lp(a) should be developed to allow the design of trials recruiting patients with elevated Lp(a) a priori and assessing whether lowering Lp(a) results in decreased CVD risk.

Phospholipases sPLA₂ & Lp-PLA₂

sPLA₂ and Lp-PLA₂ are enzymes that cleave the oxidized fatty acid side chain at the sn-2 position of OxPL. This reaction generates lysophosphatidylcholine and an oxidized free fatty acid, which are lipid mediators of multiple proinflammatory and proatherogenic pathways [20]. Phospholipases are expressed by macrophages and are upregulated by proinflammatory compounds and OxLDL. Lp-PLA₂is carried by all lipoproteins, whereas sPLA₂ is not carried by lipoproteins to any significant extent. Prospective epidemiological studies in clinically stable populations have shown that elevated levels of Lp-PLA₂ and sPLA₂ mass and/or activity are associated with an increased risk of cardiovascular events [20,21]. The database for Lp-PLA₂ as a predictor of CVD, and also of stroke, is quite large, with 32 major studies of primary and secondary prevention. However, it does not appear to predict subsequent risk in patients presenting with acute coronary syndrome (ACS). In addition, the odds or hazard ratios are modest and are often significantly attenuated when adjusting for LDL-C, which is the main carrier of Lp-PLA₂. Similarly, changes with lipid-lowering therapies reflect the change in LDL-C, and therefore do not seem to have a significant clinical role in following therapeutic interventions. The scientific evidence for sPLA₂ is not as well developed as that for Lp-PLA₂, but the evidence base is growing that sPLA₂ is a robust predictor of CVD events. In fact, the hazard ratio is often significantly higher than it is for Lp-PLA₂ when measured in the same studies [11,22]. We have recently shown in the MIRACL trial that sPLA₂ mass independently predicts death during a 16-week period following ACS and that high-dose atorvastatin significantly reduces sPLA₂ and Lp-PLA₂ mass and activity after ACS, and mitigates the risk of death associated with sPLA₂ mass [22]. Finally, the hazard ratio for new CVD events is approximately doubled when either OxPL/apoB or Lp(a) is combined with sPLA₂[11] or Lp-PLA₂[9], suggesting pathophysiological synergy between these biomarkers in risk prediction. Additional studies are also needed to assess whether the mass or the activity of these phospholipases provide complementary or similar information, as their values do not necessarily have a high correlation.

In conclusion, these data suggest that oxidative biomarkers are robust, independent biomarkers of CVD risk emanating from oxidative/inflammatory pathways. Going forward, combinations of these biomarkers may add significant prognostic value to CVD risk prediction and are poised to play a significant clinical role. Additional studies of reclassification are needed for all of these biomarkers. If they continue to show that they can affect management in the setting of currently available clinical risk models, as suggested for OxPL/apoB [10], then they should ultimately take their place in clinical guidelines for assessing the risk of CVD.

Financial & competing interests disclosure

S Tsimikas is named as an inventor of and receives royalties from patents and patent applications awarded to the University of California (CA, USA) for the use of oxidation-specific antibodies, is a cofounder and has equity interest in Atherotope, is a consultant to ISIS, Genzyme, Quest and Aterovax, and has received investigator-initiated grants from Merck and Pfizer. The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.

No writing assistance was utilized in the production of this manuscript.