Novel HDL-directed pharmacotherapeutic strategies

Recombinant apo A-I Milano/phospholipid

The recombinant apo A-I Milano protein differs from wild-type apo A-I by a cysteine to arginine substitution at amino acid 173. The effects of apo A-I Milano complexed with phospholipid (1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine), also known as ETC-216, have been studied in animals and humans. The apo A-I Milano mutation was first identified in a cohort of Italian patients who exhibited a decreased prevalence of atherosclerosis despite very low levels of HDL cholesterol (10–20 mg/dl).¹⁸ In support of an antiatherogenic activity for this drug, infusions of ETC-216 in rabbit and mouse models of atherosclerosis were associated with considerable reductions in the lipid and macrophage content of plaque.^10,19,20 The superiority of apo A-I Milano over native apo A-I, however, remains uncertain. Using genetic mouse models of atherosclerosis, cholesterol efflux assays in vitro, and examination of macrophage-specific reverse cholesterol transport in vivo, some studies suggested that apo A-I Milano had greater antiatherogenic effects than wild-type apo A-I,^21,22 whereas others indicated equivalence to the wild-type protein.^23–26 Head-to-head intervention studies comparing administration of the two types of apo A-I are lacking.

In a small clinical study of 47 patients with acute coronary syndrome, five weekly doses of 15–45 mg/kg of ETC-216 significantly reduced total atheroma volume by 14.1 mm³ compared with baseline (4.2%; P <0.001), as measured by coronary intravascular ultrasonography (IVUS) (Table 1).¹⁴ Total atheroma volume did not change significantly with the administration of placebo (–0.2 mm³; P = 0.97), and HDL-cholesterol and apo A-I levels were not reported. Two of 15 patients (9%) in the high-dose ETC-216 group were withdrawn from the study owing to adverse events, one because of symptomatic cholelithiasis and the other because of a hypersensitivity reaction.¹⁴ A subsequent dose-ranging study in a rabbit model suggested that five infusions of ETC-216 at doses equivalent to 8–15 mg/kg in humans might be adequate to maximize plaque stabilization or regression.⁸ Substantial interest remains in investigating the therapeutic potential of rHDL infusion containing apo A-I Milano. Clinical development of this drug, which was halted owing to manufacturing difficulties, is scheduled to resume in 2011 following completion of technology transfer to a different pharmaceutical company.²⁷

Purified native apo A-I/phospholipid

CSL-111 consists of complexes of native apo A-I and phosphatidylcholine, the former isolated from precipitates obtained by cold ethanol fractionation of human plasma and the latter derived from soybean.^16,28 A randomized, placebo-controlled study in which serial intravascular ultrasonography was used to assess 145 patients with acute coronary syndrome revealed that administration of four weekly 40 mg/kg infusions of CSL-111 reduced atheroma volume by 5.3 mm³ compared with baseline (3.4%; P <0.001) (Table 1).¹⁶ No statistically significant difference in atheroma volume was observed when comparing patients receiving placebo and those treated with CSL-111 (P = 0.48). Quantitative coronary angiography, on the other hand, demonstrated that CSL-111 use was associated with less progression of coronary atherosclerosis than placebo (–0.039 mm and –0.071 mm, respectively; P = 0.03), as measured by the coronary score (per-patient mean of the minimal lumen diameter for all lesions measured) among patients with less-severe stenoses at baseline.¹⁶ Lower coronary scores have been associated with an increased risk of incidence of coronary events.²⁹ By comparison, 40 mg daily doses of pravastatin or 40–80 mg daily doses of lovastatin administered for 2 years attenuated coronary atherosclerosis progression as assessed by coronary score to a similar degree as only 4 weeks of CSL-111 treatment.^30,31 One-third of patients randomly assigned to receive 80 mg of CSL-111 had reversible alanine aminotransferase elevations exceeding 10 times the normal upper limit, but the group administered CSL-111 at the lower dose of 40 mg/kg did not differ from the placebo group with regard to transaminase levels. More participants randomly assigned to 40 mg/kg of CSL-111 exhibited hypotension compared with those receiving placebo (13.8% and 7.1%, respectively). Further details regarding the observed changes in blood pressure were not provided in the original publication.¹⁶

A subsequent study in patients with lower extremity peripheral artery disease demonstrated that treatment with a single CSL-111 infusion was associated with significant reductions in lipid content (P <0.05; assessed by oil red O staining) and endothelial adhesion molecule expression (P <0.05; assessed by vascular cell adhesion molecule 1 [VCAM-1] expression) in plaque excised by atherectomy.³² These changes were observed at the time of percutaneous superficial femoral artery revascularization, which occurred 5–7 days following CSL-111 infusion, indicating a rapid onset of action of the drug. In addition, CSL-111 considerably increased the levels of HDL cholesterol and enhanced the capacity of plasma depleted in apolipoprotein B (apo B) to promote cholesterol efflux from macrophages. Among patients with diabetes mellitus, infusion of CSL-111 increased the levels of HDL cholesterol up to 40%, inhibited platelet aggregation ex vivo, and reduced monocyte activation and neutrophil adhesion.^33,34 A reformulated version of CSL-111, called CSL-112, has been reported in preclinical studies to provide greater cholesterol efflux capacity and have less hepatotoxicity than CSL-111.³⁵ A phase I study of intravenous CSL-112 has been initiated.³⁵

Autologous delipidated HDL

A novel HDL-directed pharmacotherapeutic approach utilizes the infusion of autologous delipidated HDL.¹⁷ The Plasma Delipidation System-2 (PDS-2) from Lipid Sciences involves the collection of 1 l of plasma by apheresis over 1.5–2.0 h followed by the selective removal of lipids from HDL particles using organic solvents; the lipid-poor preβ-HDL, an efficient acceptor of cholesterol from ABCA1, is subsequently reinfused over 1 h.

Preclinical evaluation of selective delipidated HDL in a nonhuman primate model of dyslipidemia achieved a significant 6.9% reduction in aortic atheroma volume, assessed by IVUS.³⁶ In a small clinical study, among 28 patients with acute coronary syndrome, seven weekly infusions of autologous delipidated HDL increased preβ-HDL levels approximately 30-fold and decreased total atheroma volume by 12 mm³ (5.2%) from baseline (Table 1).¹⁷ As in the studies of intravenous apo A-I therapy discussed above, no statistically significant differences were observed between actively-treated patients and control groups. Autologous delipidated HDL infusions did not induce transaminitis or hypersensitivity reactions; however, apheresis resulted in hypotension in a third of the participants undergoing this treatment.

Oral upregulator of endogenous apo A-I

A small synthetic molecule (termed RVX-208) belonging to the quinazoline family, which are known primarily for their antimalarial properties, was reported in cellular screening assays to raise apo A-I production.³⁷ In a monkey model of atherosclerosis, administration of a compound containing RVX-208 over approximately 2 months increased plasma apo A-I and HDL-cholesterol levels up to 60% and 97%, respectively, in a dose-dependent manner. Monkeys treated with RVX-208 had enhanced cholesterol efflux from macrophage foam cells ex vivo. A doubling in the levels of triglycerides was observed with the highest dose of the drug; however, the mid-range dose led to a 43% increase in apo A-I concentration without affecting triglyceride levels.³⁷ No effect on apo B or LDL-cholesterol levels was observed with any dose of RVX-208. A short exploratory study in humans demonstrated increases in total plasma apo A-I and preβ-HDL levels (10% and 42%, respectively), as well as augmentation of cholesterol efflux with the use of RVX-208.³⁷ Findings from the ASSERT study were presented at the 2010 AHA conference, and the full results are anticipated to be released in 2011. Modest changes in HDL-C and apo A-I were reported to have been observed in patients with stable coronary artery disease on chronic statin therapy.³⁸ According to Dr Stephen J. Nicholls, a study of RVX-208 is being planned in patients with acute coronary syndrome using coronary intravascular ultrasonography and is currently scheduled to begin in 2011.³⁸

CETP inhibitors

CETP shuttles cholesteryl esters from HDL to LDL and VLDL in exchange for triglycerides.³⁹ Genetic CETP deficiency causes marked elevation in HDL-cholesterol levels and modest elevation in apo A-I levels. Some,^40–44 but not all,^45–47 epidemiological studies indicate that CETP has an atherogenic role, with reduced plasma CETP mass or activity associated with reduced incidence of atherothrombotic cardiovascular disease; however, this issue remains far from settled. Preclinical studies involving reduction in CETP activity through various means, including the use of small-molecule inhibitors, antisense oligonucleotides, and vaccine-induced antibodies, have generally demonstrated decreased plaque formation with these agents, lending further support to CETP inhibition as a viable therapeutic strategy for the prevention and management of atherothrombotic disease.^48–52

Despite these supporting data, the demise of the CETP inhibitor torcetrapib raised doubts about the concept of CETP inhibition. In the Investigation of Lipid Level Management to Understand its Impact in Atherosclerotic Events (ILLUMINATE),⁵³ which involved patients receiving atorvastatin, torcetrapib worsened the primary combined cardiovascular end point (death from coronary heart disease, nonfatal myocardial infarction, stroke, and hospitalization for unstable angina) (hazard ratio [HR] 1.25, 95% CI 1.09–1.44, P = 0.001) and all-cause mortality (HR 1.58, 95% CI 1.14–2.19, P = 0.006) compared with placebo after 12 months despite increasing HDL-cholesterol levels by 72% and decreasing LDL-cholesterol levels by 25%.⁵³ The worrisome results of ILLUMINATE, paralleled by negative imaging findings in trials in which coronary and carotid ultrasonography were used,^54,55 have been at least partly attributed to the off-target effects of torcetrapib, such as the raising of systolic blood pressure by an average of 5.4 mmHg.^56,57 As suggested by aldosterone measurements from participants in ILLUMINATE,⁵³ findings from animal models⁵⁸ and human adrenal cell assays⁵⁹ indicate that these off-target effects are related to stimulation of aldosterone synthesis by torcetrapib via pathways independent of CETP inhibition.

Could inhibition of CETP activity explain some of the adverse effects observed in the clinical development program of torcetrapib? Initial concerns about CETP inhibition highlighted the possibility that the formation of large cholesterol-rich HDL particles resulting from CETP inhibition might be associated with impaired cholesterol efflux from peripheral macrophages to these particles.⁶⁰ However, studies of patients either deficient in CETP or treated with a CETP inhibitor demonstrated not reduced, but rather enhanced, efflux of cholesterol via ATP-binding cassette sub-family G member 1 (ABCG1) to HDL.^59,61,62 An important arm in the proximal step of reverse cholesterol transport, therefore, remains intact, if not improved, in the setting of CETP inhibition. On the other hand, the net effect of CETP inhibition on reverse cholesterol transport might depend, in part, on the effects of CETP on the hepatic uptake of HDL-derived cholesterol. The hepatic uptake of HDL cholesterol can be direct, or occur after transfer of HDL cholesterol via CETP to apo B-containing lipoproteins, which are then taken into the liver via the LDL receptor. As illustrated by macrophage-specific reverse cholesterol transport assays, in the setting of highly effective LDL clearance, CETP can actually enhance reverse cholesterol transport and have an atheroprotective role; however, when LDL clearance is impaired, CETP can slow down reverse cholesterol transport and thus be proatherogenic.⁶³ The interdependence of terminal reverse cholesterol transport pathways suggests that the hypothesis that the individuals who are most likely to benefit from CETP inhibition are those with suboptimal LDL-receptor-mediated hepatic uptake of cholesterol. Fortunately, at least two novel compounds apparently lacking the off-target effects of torcetrapib enable further testing of the CETP-inhibition strategy.

Niacin receptor agonists

Niacin, the first antidyslipidemic agent identified, remains the most potent drug for increasing levels of HDL-cholesterol and apo A-I. Evidence suggests that HDL mediates the potential atheroprotective effects of niacin,⁷⁹ although the role of beneficial changes in other lipid parameters, including a reduction in the concentration of LDL particles and lipoprotein(a), cannot be excluded. Randomized trials have shown a clinical benefit of niacin among patients with coronary heart disease, with low cardiovascular event rates^80,81 and improvements in surrogate measures of vascular function, as assessed by imaging,^79,82 associated with niacin administration. One of these studies, for example, revealed regression of carotid intima–media thickness (IMT), the primary imaging end point, associated with niacin treatment, whereas no change in IMT was observed among patients treated with the LDL-lowering drug ezetimibe.⁸² The niacin group achieved better outcomes in IMT compared with the ezetimibe group in the setting of 23% higher HDL-cholesterol levels, despite 7% higher LDL-cholesterol levels, lending further credence to an HDL-targeted approach.⁸² Definitive data on the incremental value of niacin when added to optimum medical therapy, which includes aggressive statin use, might emerge from two large trials, the Atherothrombosis Intervention in Metabolic Syndrome with Low HDL/High Triglycerides and Impact on Global Health Outcomes (AIM-HIGH) study⁸³ and the Treatment of HDL to Reduce the Incidence of Vascular Events (HPS2-THRIVE) study,⁸⁴ which are anticipated to be communicated in 2012 or 2013.

After 50 years of clinical use,⁸⁵ the mechanisms by which niacin raises the levels of HDL (and other lipoproteins) remain elusive. Studies of cultured hepatocytes suggest that these effects might be mediated by decreased catabolism of apo A-I owing to downregulation of the HDL holoparticle (protein plus lipids) receptor, which is a terminal step in the reverse cholesterol transport pathway.⁸⁶ Kinetic studies in humans, on the other hand, demonstrate either decreased clearance^87,88 or increased synthesis of apo A-I in patients receiving niacin,⁸⁹ the latter finding being notably absent from analyses in vitro. The discovery of the niacin receptor GPR109A⁹⁰ promised to usher in a new era in which the molecular mechanisms underlying the effects of niacin on lipids, and its adverse effects on the skin, such as flushing, could be clearly defined. This discovery might also open the way for the development of new synthetic agonists of GPR109A, which would ideally have the same therapeutic effects as niacin, but not the skin-related adverse effects of this drug, which substantially limit its therapeutic use. Activation of GPR109A on skin immune cells, such as Langerhans cells and keratinocytes, stimulates phospholipase A₂, enhancing synthesis of prostaglandin D₂ (PGD₂) and prostaglandin E₂ (PGE₂), which subsequently induce cutaneous capillary vasodilation by binding to subtypes of the PGD₂ receptor (DP1) and the PGE₂ receptor (EP2 and EP4).⁴ Coadministration of the DP1 antagonist laropiprant reduces the incidence of the skin-related adverse effects of niacin.⁹¹ However, flushing still occurs in over half of the patients, and discontinuation of niacin treatment owing to severe skin-related symptoms remains a problem.⁹¹ A study in mice implicates keratinocyte-produced PGE₂ as a key mediator of niacin-induced skin flushing, suggesting that this molecule might be another potential target to minimize flushing.^92,93

Although activation of GPR109A on adipocytes clearly leads to an acute reduction in lipolysis and release of free fatty acids,⁹⁰ the role of this receptor in mediating the effects of niacin on plasma lipoproteins remains uncertain. Partial agonism of GPR109A might differentially activate lipolysis and flushing pathways. A study in mice supports this dissociation and implicates ERK1/2 mitogen-activated protein kinase phosphorylation in vasodilation.⁹⁴ Administration of the partial agonist MK-0354 in humans resulted in a lack of flushing; however, despite effective suppression of lipolysis and free-fatty-acid release, this partial agonist had no effect on HDL-cholesterol or other plasma lipoproteins.⁹⁵ These results have brought into question the role of GPR109A agonism in mediating the effects of niacin therapy on plasma lipoproteins. This question is critical for the field, and will determine the future of GPR109A agonists as a therapeutic strategy.

Endothelial lipase inhibitors

Synthesized by and bound to vascular endothelial cells, endothelial lipase exhibits predominant phospholipase activity and affinity for HDL, unique characteristics among the lipoprotein–lipase family.⁹⁶ An association between the expression of endothelial lipase and HDL-cholesterol levels was identified in overexpression and loss-of-function mouse models.^97–99 Genetic studies indicate a similar association in humans, with rare and low-frequency loss-of-function variants of endothelial lipase, identified through deep sequencing, resulting in elevated HDL-cholesterol levels.¹⁰⁰ How changes in HDL-cholesterol levels attributed to endothelial lipase ultimately affect atherosclerosis remains uncertain. Some human studies point to an atherogenic role of endothelial lipase, with a positive association between plasma levels of this enzyme and coronary artery calcification, features of the metabolic syndrome (such as impaired fasting glucose, hypertriglyceridemia, low HDL-cholesterol, hypertension, and waist circumference), and inflammation.^101,102 Carriers of endothelial-lipase variants associated with raised HDL-cholesterol levels have been reported to have a decreased risk of atherothrombotic disease,¹⁰³ although this link has not been observed in other studies.^104,105 Of note, animal studies have shown that HDL-cholesterol levels do not always correlate with physiological changes associated with decreased atherosclerotic burden, with an endothelial-lipase knockout mouse model failing to demonstrate improved macrophage reverse cholesterol transport with raised HDL-cholesterol levels.¹⁰⁶ More worrisome, in the setting of hepatic-lipase deficiency, is that the absence of endothelial lipase resulted in an accumulation of small dense LDL,¹⁰⁶ a particularly atherogenic subpopulation of LDL. These findings suggest that endothelial-lipase inhibition could potentially exert a detrimental effect on atherosclerosis, even though HDL-cholesterol levels are raised. Despite this uncertainty, endothelial lipase has been the object of substantial interest as a therapeutic target. High-throughput screening identified several compounds sharing a sulfonyl furan urea core as potent and selective endothelial-lipase inhibitors.¹⁰⁷ Further evaluation of these and other small-molecule inhibitors of endothelial lipase has not been performed yet.

Agonists of the liver X receptor

Liver X receptors (LXRs), which are members of the nuclear receptor superfamily, have a central role in lipid metabolism. Endogeneously activated by oxysterols, LXRs regulate the transcription of a myriad of target genes by binding to their promoters together with the retinoic acid receptor.¹²⁵ With regard to HDL and reverse cholesterol transport, LXR activation has been demonstrated to promote mobilization of intracellular cholesterol,¹²⁶ increase macrophage cholesterol efflux via macrophage ABCA1 and ABCG1,¹²⁷ and augment intestinal HDL generation.¹²⁸ Two LXR isoforms have been identified—LXRα and LXRβ. These isotypes have structural homology, but differ in tissue distribution, with the former being primarily expressed in the liver, macrophages, intestine, kidney, and adipose tissue, whereas the latter is found ubiquitously.¹²⁹

Therapeutic development of LXR agonists has been hindered by hepatic steatosis and increased plasma tri-glyceride concentrations reported in preclinical studies of these drugs.^130,131 Fortunately, dissociating LXR efficacy and toxicity might be possible owing to the differential effects of LXR agonism by receptor isoform and by tissue-specific effects. For example, gene knockout models in mice implicate LXRα,¹³² and not LXRβ,¹³³ as the primary stimulus for hepatic lipogenesis and hypertriglyceridemia. Administration of a nonselective LXR agonist to LXRα-deficient mice increased HDL-cholesterol levels and stimulated macrophage ABCA1 expression and cholesterol efflux without inducing fatty liver and with minimal transcription of hepatic sterol regulatory element-binding protein isoform 1c (SREBP1c) mRNA, which upregulates the synthesis of triglycerides in the liver.^74,134,135 Moreover, in a mouse model susceptible to atherosclerosis that lacked LXRα and apo E, treatment with the LXRα/β agonist GW3965 not only improved cholesterol in vitro (decreased total and unesterified cholesterol levels and increased HDL-cholesterol levels) but reduced aortic plaque development.¹³⁶ These results indicate that a selective LXRβ agonist might augment reverse cholesterol transport without engendering hypertriglyceridemia or fatty liver.

A second approach to safer LXR development might be to selectively activate intestinal LXR. Fatty liver arises from activation of hepatic LXR, which, through upregulation of SREBP1c, stimulates lipogenesis.^137,138 Elevation of triglyceride levels occurs via SREBP1c activation¹³⁹ and the subsequent suppression of apolipoprotein A-V (apo A-V),¹⁴⁰ which is an inhibitor of VLDL synthesis (apo A-V can also stimulate VLDL hydrolysis).¹⁴¹ On the other hand, LXR expression on both macrophages and the small intestine contributes to the regulation of reverse cholesterol transport. The absence of macrophage LXRs in a mouse model attenuated, but did not abolish, the increase in macrophage-specific reverse cholesterol transport observed following the administration of the systemic LXRα/β agonist GW3965, which indicates that macrophage LXR expression, but also nonmacrophage LXR expression, contribute to this transport.¹⁴² An intestine-specific LXRα/β agonist, GW6340, promoted macrophage-specific reverse cholesterol transport, augmenting the fecal excretion of radiolabeled sterol by 52% via increased intestinal HDL production and intestinal excretion of HDL-derived cholesterol.¹⁴² Consistent with these findings, a transgenic mouse model of intestine-specific LXRα overexpression demonstrated increased preβ-HDL levels, enhanced macrophage-specific reverse cholesterol transport, and decreased aortic atheroma volume.¹⁴³ Intestine-specific expression of LXRα also protected mice taking a high-cholesterol diet from cholesterol and triglyceride accumulation in the liver.¹⁴³ Intestine-specific LXR agonism, therefore, has potential antiatherogenic effects through the enhancement of reverse cholesterol transport and might avoid the toxicity associated with hepatic LXR activation.

LCAT activators

LCAT esterifies cholesterol, enabling its incorporation into the hydrophobic core of the HDL particle and transforming nascent HDL into spherical mature HDL in the process. Although the importance of LCAT in HDL metabolism is clear, with decreased activity of this enzyme leading to lower HDL-cholesterol levels, its role in reverse cholesterol transport and atherosclerosis in both animals^144–146 and humans^147,148 has been increasingly controversial. The evidence further calls into question the relevance of LCAT in the development of atherosclerosis and throws doubt on the therapeutic potential of stimulating LCAT activity. A cross-sectional examination showed that homozygous and heterozygous carriers of mutations associated with LCAT deficiency have no increase in carotid IMT despite low HDL-cholesterol levels (homozygous 9 mg/dl, heterozygous 39 mg/dl).¹⁴⁹ In a case–control analysis of the European Prospective Investigation of Cancer (EPIC)-Norfolk study¹⁵⁰ population, low plasma levels of LCAT were not associated with the development of coronary artery disease, in contrast to results from a prior study by the same group examining families with LCAT gene mutations.¹⁵¹ Perhaps the most striking data on this issue are those from the Prevention of Renal and Vascular Endstage Disease (PREVEND) study,¹⁵² in which high LCAT activity, despite elevated HDL-cholesterol levels, independently predicted an increased risk of cardiovascular events. These findings are consistent with mouse studies of LCAT overexpression, which demonstrated attenuated macrophage-specific reverse cholesterol transport in vivo and ABCA1-mediated macrophage cholesterol efflux ex vivo.¹⁵³ Only one LCAT activator has reached early clinical development, ETC-642, and little data are available on the outcomes of treatment with this agent.¹²¹ Despite these uncertainties, LCAT activation remains an interesting potential therapeutic target owing to its ability to raise HDL-cholesterol levels and its possible influence on reverse cholesterol transport.