OUP user menu

★ Editor's choice ★

Spotlight on HDL biology: new insights in metabolism, function, and translation

Daniel J. Rader
DOI: http://dx.doi.org/10.1093/cvr/cvu164 337-340 First published online: 15 July 2014

High-density lipoproteins (HDLs) have captured the imagination of scientists and physicians from virtually the time they were originally described, but especially after the cholesterol carried in HDL (HDL-C) was shown to be strongly and inversely associated with risk of coronary heart disease (CHD).1 This observation, replicated in many populations around the world, has led to the popular view of HDL-C as the ‘good cholesterol’ and to the belief that intervention to raise plasma levels of HDL-C would lead to reduction in risk of CHD (the ‘HDL hypothesis’). Because of its obvious clinical relevance, the structure, function, metabolism, and regulation of HDL has been studied in great detail by hundreds of laboratories around the world. The science and biology of HDL is a mature field that has been informed by structural biology, lipid and protein biochemistry, model systems, human genetics and physiology, and pharmacology. Although much has been learned, there remain major questions surrounding HDL, not the least of which includes recent scepticism about the fundamental HDL hypothesis.

Two major developments over the last few years, in human genetics and clinical trials with pharmacological agents, have led to doubts about the HDL hypothesis. Human genetics have taught us a great deal about HDL metabolism. Mendelian disorders of extreme low and high HDL-C levels have been extremely informative regarding the key roles of specific proteins.2 More recently, unbiased genome-wide studies have led to scores of additional loci associated with HDL-C levels.3 A focus of intense interest has been the relationship between genetic factors that influence HDL-C levels and risk of CHD. It is interesting that despite decades of study, the three Mendelian causes of extremely low HDL-C — apoA-I structural mutations, ABCA1 deficiency (Tangier disease), and familial lecithin-cholesterol acyltransferase (LCAT) deficiency — have yet to be convincingly linked to an increased risk of CHD.2 Conversely, the sole Mendelian cause of extremely high HDL-C — cholesteryl ester transfer protein (CETP) deficiency — has not been shown to be protective against CHD. However, because these conditions are rare, it is challenging to perform adequately powered observational studies, and thus anecdotal reports and subclinical atherosclerosis studies predominate and uncertainty remains. Only within the last few years has there been a serious attempt to ask whether genetic variants associated with HDL-C that exist at some frequency in the general population are associated with CHD, an approach known as ‘Mendelian randomization’. For example, heterozygotes for loss-of-function mutations in ABCA1 were reported to be at no significantly increased risk of CHD.4 Conversely, heterozygotes for a loss-of-function mutation in endothelial lipase (gene LIPG) that raises HDL-C levels5 do not appear to be protected from CHD.6 While common non-coding variants at CETP that influence HDL-C levels are weakly associated with CHD in the expected (inverse) direction, these variants are generally also associated with LDL-C levels.7 On the other hand, common variants at most other loci associated with HDL-C have no significant association with CHD.6,8 This is particularly notable in light of the observation that variants associated with LDL-C and even triglycerides (TGs) are often significantly associated with CHD, even after adjusting for their effects on HDL-C.8 Thus, the human genetics suggest that HDL-C may not be causal with regard to its association with CHD risk.

The other strand leading to uncertainty about the HDL hypothesis relates to clinical trials with niacin and CETP inhibitors. Niacin has been used as a pharmacological approach to raising HDL-C (and lowering LDL-C and TGs) for decades. One trial in the pre-statin era, the Coronary Drug Project, showed that in hypercholesterolaemic men with CHD treatment with immediate-release niacin reduced cardiovascular events. However, two trials in the last several years involved the addition of niacin to a statin in patients with CHD who had well-controlled LDL-C levels and failed to show evidence of a significant reduction in cardiovascular events. Both trials have caveats with regard to their ultimate interpretations. The AIM-HIGH trial9 was a relatively underpowered given the modest increase in HDL-C compared with placebo. It was terminated early due to futility with no evidence of even a trend towards benefit from extended-release niacin. The HPS2-THRIVE trial was much larger but included a second drug, laropiprant, intended to reduce the flushing associated with niacin.10 It failed to demonstrate a significant benefit of extended-release niacin therapy (although it did show an overall trend towards benefit and a significant benefit in the subgroup with higher baseline LDL-C levels). The experience with CETP inhibitors to date is an even stronger argument against the HDL hypothesis. Torcetrapib was the first CETP inhibitor to enter Phase 3 and the trial, ILLUMINATE, was stopped early due to increased cardiovascular events and mortality in the torcetrapib group despite a robust increase in HDL-C.11 The interpretation of this trial is complicated by the discovery that torcetrapib has substantial off-target effects on adrenal hormones and blood pressure that could increase cardiovascular risk.12 However, a ‘cleaner’ CETP inhibitor, dalcetrapib, failed in a large phase 3 trial dal-Outcomes to reduce cardiovascular events despite an HDL-C increase of ∼25% and the trial was stopped early for futility without even a trend towards benefit.13 This is perhaps the single greatest challenge to the HDL hypothesis. While it has been argued that the HDL-C increase was modest, the observational studies do not suggest a threshold effect but rather demonstrate a continuous inverse relationship between HDL-C and CHD risk. Nevertheless, two potent CETP inhibitors, anacetrapib and evacetrapib, are in phase 3 cardiovascular outcome trials, the results of which are awaited with great interest by the biomedical community.

The human genetics and clinical trials have led to a re-evaluation of the HDL cholesterol hypothesis and to increased focus on other aspects of HDL biology, structure, and function. A classic, putatively atheroprotective, function of HDL and its major protein apoA-I has been to promote cellular cholesterol efflux, especially from macrophages, and return of that cholesterol to the liver in a process known as ‘reverse cholesterol transport’ (RCT). Detailed experimentation has elucidated a number of specific transporters, such as ABCA1 and ABCG1, responsible for mediating cellular cholesterol efflux to HDL and their regulation and biology.14 Other studies have detailed the many physiological processes involved in the metabolism of cholesterol once it has effluxed to HDL, including esterification by LCAT, selective uptake by the liver via scavenger receptor class B type I (SR-BI), transfer to apoB-containing lipoproteins by CETP, excretion into bile, direct transport from plasma to intestinal lumen, and excretion into faeces. In animal models, an integrated dynamic measure of macrophage to faeces RCT in response to genetic or pharmacological manipulation has been shown to predict the effect on atherosclerosis of that intervention substantially better than the effect on a static plasma HDL-C concentration.15 Studies on the capacity of human HDL to promote cellular cholesterol efflux from macrophages have suggested that this is a better predictor of prevalent atherosclerotic disease than HDL-C itself,16,17 though uncertainty remains regarding its ability to predict incident cardiovascular events.17 The idea of moving from the ‘HDL cholesterol hypothesis’ to the ‘HDL flux hypothesis’ has been raised18 and has important implications for the development of new therapeutic approaches targeted to HDL and RCT.

Macrophage cholesterol efflux/RCT is only one of many ‘functions’ of HDL that have been described over the last two decades. HDL has effects on the biology of other cell types by promoting cholesterol efflux. For example, cholesterol efflux from hematopoietic stem cells influences their maturation and release into the circulation of mature leucocytes.19 Cholesterol efflux from megakaryocytes influences the rate of platelet production.20 Cholesterol efflux from pancreatic beta cells appears to influence the rate of insulin secretion.21,22 Cholesterol efflux to apoA-I and apoE in the central nervous system appears to influence neurodegenerative disease and probably other physiological processes.

Moreover, HDL has properties that go well beyond its ability to promote cholesterol efflux.23 For example, HDL has long been recognized to have anti-inflammatory properties on a variety of cell types, including endothelial cells and macrophages.24 The molecular basis of these effects are still being elucidated, but effects on signal transduction play a role.25 HDL is a major transporter of sphingosine-1-phosphate (S1P) through binding to the HDL-associated protein apoM,26 and HDL-associated S1P is bioactive in several ways through activating its receptors on cells such as endothelial cells.27 HDL was recently demonstrated to exert anti-inflammatory effects on macrophages by up-regulating the protein ATF3.28 While HDL and apoA-I have been shown to exert anti-inflammatory effects in vivo in animal models,29 the relevance to humans remains to be firmly established. Nevertheless, this property of HDL is one of the most consistently observed and there is much more to be learned.

Another consistent property of HDL is its ability to promote nitric oxide (NO) synthesis by endothelial cells. Through a process that involves endothelial SR-BI, HDL results in a signal transduction cascade that increases endothelial nitric oxide synthase activity and NO production.26 The benefits of this property of HDL have been shown clearly in mouse model systems. This property of human HDL has elegantly been shown to be strongly associated with the content of HDL paraoxonase,30 an enzyme that hydrolyzes lipid substrates, and importantly with risk of CHD. Thus, a potentially important atheroprotective property of HDL may be its ability to promote endothelial NO production.

A theme of the preceding paragraphs is that other proteins and lipids on HDL have bioactive properties lending HDL additional functions. Indeed, there has been an intense effort to describe in detail the full HDL proteome31 and lipidome32 and their relationship to cardiovascular disease. These studies have led to the recognition that HDL is a highly complex and heterogeneous macromolecule, with dozens of discrete subspecies of defined protein and lipid composition.33,34 A developing concept is that HDL evolved as a ‘scaffold’ for plasma proteins and lipids to assemble in specific ways for specific purposes, often in support of innate immune function. Indeed, an elegant example of this concept derives from the observation that HDL harbours a lytic factor for a specific species of Trypanosomes called Trypanosoma brucei. This lytic factor turns out to be a complex of the proteins apoL-I and haptoglobin-related protein (HPR), which individually have no activity but when assembled on an HDL particle form a potent lytic complex.35 Humans are resistant to T. brucei, but most other species, which lack apoL-I and HPR, are vulnerable to it. Other paradigms in which HDL-associated complexes provide protection from pathogens are likely to emerge. Another amazing discovery has been that HDL serves to transport microRNAs (miRNAs) within the blood, and that after picking up miRNAs from one tissue it can transport them to another tissue where they are biologically active.36 The possibilities of how this ‘miRNA endocrine function’ of HDL may influence physiology and disease are only now beginning to emerge and indicate a whole other layer of HDL functionality.37

Remarkably, the decline of the classic (simplistic) HDL cholesterol hypothesis has occurred in parallel with an explosion in our broader understanding of HDL structure, composition, metabolism, and biology. In this Spotlight issue, several aspects of our evolving insights in HDL biology and potential for translation are discussed in depth.3846 While we certainly need to reassess the simple construct that raising HDL-C will reduce cardiovascular risk, we should not conclude that HDL is biologically unimportant or that therapeutic intervention targeted towards specific HDL functions has no future. Indeed, a variety of therapeutic approaches targeted to various aspects of HDL metabolism and function continue to be actively explored.47 The next decade of research in HDL is bound to be exciting, with additional unexpected discoveries and the potential that the substantial investment in understanding HDL biology may result in effective new therapies for not only atherosclerotic cardiovascular disease, but even other conditions as well.


  • The opinions expressed in this article are not necessarily those of the Editors of Cardiovascular Research or of the European Society of Cardiology.