Perspective The failure of cholesteryl ester transfer protein inhibitors: is it due to increased blood viscosity?

Overview

It is rare in the field of cardiovascular medicine that new data cause a rapid or decisive change in mainstream clinical practice, let alone one’s thought process. The failure of antioxidant supplementation to clearly impact cardiovascular disease [Vivekananthan et al. 2003] has not caused any decreased interest in their daily use for chronic oxidative stress. Besides this attention for a possible role in oxidative stress, other markers have continued to be explored for potential relationships in the development of cardiovascular disease. For example, continued interest in homocysteine as a risk factor for atherosclerosis persists, despite conflicting data. Likewise, it could be considered unusual that the failure of two inhibitors of cholesteryl ester transfer protein might have a positive impact in atherosclerotic cardiovascular disease. Despite significantly increasing serum high-density lipoprotein (HDL) levels by these inhibitors, there was generated doubt not only on the theory of ‘reverse cholesterol transport’, but also on 40 years of epidemiologic data suggesting that HDL protects against atherosclerosis. In 2013, the National Lipid Association concluded that HDL is not a therapeutic target at present [Toth et al. 2013].

In the ILLUMINATE trial, the cholesteryl ester transfer protein inhibitor, torcetrapib, was associated with statistically significant excesses in both cardiovascular events and total mortality [Barter et al. 2007]. More patients in the torcetrapib group died of cancer and infection, although there was no evidence that torcetrapib increased the total numbers of neoplasms or infections. Torcetrapib also resulted in a 5.4 mm Hg increase in blood pressure. The dal-OUTCOMES trial, investigating dalcetrapib, was halted early because of futility [Schwartz et al. 2012]. The rapid reversal in the mainstream opinion about HDL as a potential therapeutic marker speaks to the appeal of the simple but refuted notion that more HDL must be better than less, and the inability to consider any other mechanism of atheroprotection than the reversal in cholesterol transport.

We propose that the noted increase in HDL particle size caused by inhibitors of cholesteryl ester transfer protein can result in an increase in erythrocyte aggregation, an elevation in blood viscosity, resulting in negative clinical trial outcomes. Because blood viscosity is inversely proportional to flow, an increase in viscosity will decrease tissue perfusion, potentially reflected in noncardiovascular morbidity and mortality. Further, elevated blood viscosity is increasingly recognized as a cause of cardiovascular morbidity and mortality. Elevated blood viscosity is associated with major risk factors for atherosclerotic cardiovascular disease such as hypercholesterolemia, diabetes mellitus, cigarette smoking, hypertension, male gender, aging, metabolic syndrome and hyperfibrinogenemia. It is thought that increased blood viscosity in areas of low shear stress fosters the development of mural thrombi, which subsequently are organized into atherosclerotic plaques. The same mechanism can cause superimposed thrombosis, resulting in myocardial infarction or stroke [Sloop 1996, 1999; Sloop et al. 2014]. The epidemiologic evidence supporting a role in which an increase in blood viscosity plays a role in cardiovascular disease was recently reported [Holsworth et al. 2013].

In addition to explaining the increased cardiovascular and total mortality associated with inhibition of cholesteryl ester transport protein, the proposed paradigm explains the increased risk of cardiovascular events associated with very high HDL-cholesterol in the IDEAL study [van der Steeg et al. 2008] and the increased risk associated with increased HDL particle number in the EPIC-Norfolk cohort [El Harchaoui et al. 2009].

Very high HDL levels and increased particle size increase cardiovascular risk

Post hoc analysis of the IDEAL and EPIC-Norfolk studies, following the failure of the torcetrapib trial, showed that cardiovascular event risk increased with very high HDL levels or increased particle size. Data from the IDEAL study showed HDL cholesterol levels greater than 80 mg/dl imparted a 6-fold increase in relative risk for a major coronary event. Data from the EPIC-Norfolk study showed that HDL particle size greater than 10.07 nm imparted an 8-fold increased risk for a major coronary event compared with a particle size of less than 8.60 nm. This noted increase in cardiovascular disease risk was progressive with increasing particle size [van der Steeg et al. 2008].

In the ILLIUMINATE trial, torcetrapib, 120 mg daily given with atorvastatin, caused an increase in HDL particle size from 8.5 ± 0.2 nm to 9.1 ± 0.33 nm. Torcetrapib, 120 mg given alone, resulted in an increase in particle diameter from 8.4 ± 0.4 nm to 9.1 ± 0.65 nm. When the dose of torcetrapib was increased to 120 mg twice daily, HDL particle diameter increased from 8.4 ± 0.43 nm to 9.7 ± 0.65 nm. Upon analyzing dosage regimens in this trial, it appeared that the normal HDL particle size which did not reflect an increase in cardiovascular risk was approximately 8.60 nm. Other lipid molecules were also affected by torcetrapib. Torcetrapib monotherapy, 120 mg every day, increased low-density lipoprotein (LDL) particle size from 20.4 ± 0.89 nm to 21.4 ± 0.79 nm. without changing total serum cholesterol [Brousseau et al. 2004]. Other investigators have found similar results. Dalcetrapib therapy resulted in a dose-related increase of 46.5% to 247.7% in the number of large HDL particles and a 114.7% to 149.7% increase in the number of large LDL particles [Ballantyne et al. 2012]. Cannon and colleagues reported that inhibition of cholesteryl ester transfer protein with anacetrapib also increases HDL particle size [Cannon et al. 2010].

Erythrocyte aggregation and blood viscosity

Evidence suggests that the risk of atherosclerosis associated with abnormal LDL and HDL levels is due to their effect on blood viscosity via modulation of erythrocyte aggregation [Sloop 1996, 1999; Sloop and Garber, 1997]. Erythrocyte aggregation increases blood viscosity at low shear rates or depressed blood flow by increasing the inertia of the suspended particles in blood. These aggregates are weak and readily disrupted by increasing shear force or blood flow, resulting in decreased viscosity. Thus, erythrocyte aggregation is one cause of the non-Newtonian behavior of blood.

Models of erythrocyte aggregation

The electrostatic repulsive force between erythrocytes is due to negatively charged sialic acid in the glycocalyx. Erythrocytes can approach each other only to within 7.9 nm of their glycocalyx surface [van Oss and Absolom, 1983]. There are two models of erythrocyte aggregation: the bridging model and the depletion force model [Neu and Meiselman, 2002]. According to the bridging model, any molecule or particle which is sufficiently large to simultaneously bind two erythrocytes can foster erythrocyte aggregation. Examples of such molecules are LDL, fibrinogen, immunoglobulin M (IgM), and long chain polymers of dextran or polyethylene glycol.

The binding of LDL and HDL to the erythrocyte is by nonionic adsorption [Hui et al. 1981], which predominantly involves van der Waals interactions. The glycocalyx offers more opportunities for van der Waals interactions than the polar, hydrophilic surface of the plasma membrane. Therefore, the relevant intercorpuscular distance between glycocalyces is probably 7.9 nm, not the 18 nm between erythrocyte cell membranes. HDL particles should be able to span this distance, especially when particle size is increased following inhibition of cholesteryl ester transfer protein. Increased HDL particle size increases the stability of erythrocyte aggregates by at least two mechanisms. First, larger HDL particles result in a greater distance between the erythrocytes, as an aggregate complex. According to Coulomb’s law, electrostatic forces decrease with the square of the distance between the two charged particles. Because electrostatic repulsion is a major disaggregating force [Neu and Meiselman, 2002], the greater the distance between erythrocytes, the greater is the stability of the aggregate. Further, the adhesion energy between particles can increase with the contact radius or surface area, being greater with larger HDL particles.

The second model of erythrocyte aggregation is the ‘depletion force’ model [Neu and Meiselman, 2002]. The erythrocyte, like any large colloid, is surrounded by an ‘excluded volume,’ so-called because the center of smaller particles, ‘depletants,’ are excluded. The excluded volume for any particle is determined by its radius. When erythrocytes get sufficiently close to each other, their excluded volumes overlap, creating an ‘overlap volume’ and depletants are excluded from this intercorpuscular region. The higher depletant concentration in the surrounding solution produces an osmotic pressure that drives erythrocytes together, promoting aggregation. In this model, HDL particles are depletants. Larger HDL particles will have a larger ‘excluded volume’ around erythrocytes, increasing the chance for overlap and development of the depletion force. In addition, an increased number of HDL particles will increase the osmotic pressure, augmenting erythrocyte aggregation and increasing viscosity.

Both mechanisms probably operate in vivo. Competition for erythrocyte binding between LDL and HDL, as in the bridging model, explains the protective effect of HDL. The increased risk associated with an increase in HDL particle number is better explained by the depletion force model.

The increase in blood pressure and excess mortality noted with torcetrapib therapy support a role for increased blood viscosity. Increased blood viscosity will increase peripheral vascular resistance, which will increase blood pressure. Because viscosity is inversely proportional to perfusion, increased blood viscosity could easily influence mortality. For example, the life expectancy of untreated polycythemia vera is only 1.5–3.0 years.

Unlike torcetrapib, dalcetrapib was not associated with an increase in major adverse cardiovascular events [Mohammadpour and Akhlaghi, 2013]. This may be due to the preservation of a population of normal HDL₂ particles. Compared with the other cholesteryl ester transfer protein inhibitors, dalcetrapib has a different chemical structure and has been called a cholesteryl ester transfer protein ‘modulator’ [Niesor, 2011]. Unlike torcetrapib, dalcetrapib inhibits only the heterotypic cholesteryl ester transfer pathway (i.e., the transfer of cholesteryl esters and triglycerides between apolipoprotein B-containing lipoproteins), not the homotypic pathway (i.e., between the subclasses of HDL). Thus, dalcetrapib does not inhibit the conversion of HDL₃ to HDL₂ [Shinkai, 2012]. One study has shown that HDL₂ but not HDL₃ reduces blood viscosity [Sloop and Garber, 1997]. Only normal HDL₂ may possess the proper size, surface area and surface chemistry to compete with LDL for binding to erythrocytes, thereby antagonizing erythrocyte aggregation and decreasing blood viscosity.

Future directions

In terms of advancing science, abandoning the theory of reverse transport of cholesterol is welcome. In vivo evidence supporting this theory has been insufficient [Illingworth, 1993]. This scientific advance probably cost hundreds of millions of dollars of pharmaceutical capital. However, a smaller study may not have elicited such a decisive response. Hopefully, abandonment of the reverse transport theory will result in increased attention to alternative mechanisms for the protective effect of HDL, such as modulation of viscosity. In addition, the theory of reverse transport of cholesterol was used to explain the existence of atherosclerotic, non-lipid containing plaques, type Vc [Stary et al. 1995]. Now that reverse transport has been questioned, more attention may be given to alternative theories of plaque development, such as organization of mural thrombi as suggested by Duguid [Duguid, 1960]. The pursuit of these new avenues with sound data has the potential to generate novel possibilities for therapeutic, clinical intervention.