Association study of ABCA1 polymorphisms in singapore populations 2

CAD susceptibility is associated with a multitude of risk factors including cigarette smoking, diabetes mellitus, obesity, hypertension, age as well as plasma lipoproteins like elevated

2 Literature Review I: Atherosclersosis and HDL

2.1 Coronary Artery Disease (CAD)

CAD is a leading cause of death in many industrialized nations and by 2020, based on current trends, it is projected to be the number one killer in developing countries (Lopez and Murray, 1998) In Singapore, between 2000 and 2003, CAD alone has accounted for

a quarter of all case mortalities (Health Facts Singapore 2003, Ministry of Health, http://www.moh.gov.sg)

CAD susceptibility is associated with a multitude of risk factors including cigarette smoking, diabetes mellitus, obesity, hypertension, age as well as plasma lipoproteins like elevated low density cholesterol (LDL-C) and reduced high density lipoprotein-cholesterol (HDL-C) It can be modified by changes in behavior, such as diet, exercise, alcohol consumption and estrogen replacement therapy Many patients with CAD do not have elevated LDL-C but rather low HDL-C, either alone or accompanied by hypertriglyceraemia Although CAD and HDL-C can be modifiable with lifestyle and pharmacological interventions, genetic components still contribute substantially to variation in HDL-C

2.2 HDL Structure

HDLs are the smallest (diameter 7-12 nm, molecular weight 200-400 kDa) and densest (1.063-1.25 g/ml) of the plasma lipoproteins (Barter et al., 2003) Each HDL particle consists of a hydrophobic core of mostly cholesterol esters (CE) and a small amount of triglycerides (TG) surrounded by a hydrophilic surface monolayer of phospholipids, unesterifed (free) cholesterol and apolipoproteins Table 2.1 gives the breakdown of the composition of a HDL particle Principal apolipoproteins include ApoAI (70% of total protein content) and ApoAII (20% of total protein content), with small amounts of the C

apolipoproteins (ApoCI-III), ApoAIV, ApoD, ApoE, ApoJ and ApoL Also associated with the HDL particle are plasma factors involved in remodelling of lipoproteins: lecithin-cholesterol acyltransferase (LCAT), lecithin-cholesterol ester transfer protein (CETP) and phospholipid transfer protein (PLTP) Remodelling activities account for much of the heterogeneity of HDL

Table 2.1 Composition of HDL

Component Content Phospholipid 50%

Cholesterol ester 30%

Free cholesterol 10%

Triglycerides 10%

2.2.1 HDL Subpopulations

HDL is structurally heterogeneous and can be classified into various species according to size, lipid or protein content or a combination of these factors Each HDL species behaves differently with respect to important physiological functions such as cholesterol efflux (von Eckardstein et al., 1994)

Using modern ultracentrifugation techniques, HDLs are separated into two major subtractions, HDL2 (1.063-1.125 g/ml) and HDL3 (1.125-1.210 g/ml), as well as a minor subfraction, VHDL (1.12-1.25 g/ml) (Barter et al., 2003) HDL2 has a higher cholesterol ester but lower protein content than HDL3 Polyacrylamide gradient gel electrophoresis further resolves particles within HDL2 and HDL3 fractions: HDL2b (10.57 nm), HDL2a (9.16 nm), HDL3a (8.44 nm), HDL3b (7.97 nm), HDL3c (7.62 nm) (Blanche et al., 1981) After agarose gel electrophoresis followed by anti-ApoAI immunoblotting, α and pre-β fractions

of HDL can be identified The large, lipid-rich and spherical α-HDL, corresponding to HDL2 and HDL3, forms the bulk of the circulating form of HDL The minor (5-15% of

plasma HDL), slower migrating pre-β HDL particles are small discs composed of mainly ApoAI and very little lipids Relative to α-HDL, pre-β HDL particles are more abundant in the extravascular compartment where they are likely to be the major cholesterol accepting species of HDL (Assman and Nofer, 2003) Using two-dimensional electrophoresis with agarose gel in the first dimension and nondenaturing polyacrylamide gradient gel in the second dimension, β HDL has been further resolved into pre-β1, pre-β2, pre-β3 HDL species in increasing order of size Pre-β HDL has a density of

>1.21 g/mL (the VHDL fraction) and therefore it was not originally recovered in the HDL fraction isolated by traditional ultracentrifugation

On the basis of apolipoprotein content, specific species of HDLs are distinguishable A1-HDL (floating in the HDL2 range; α mobility) contains only ApoAI and

is without ApoAII, whereas A1/A2 HDL (HDL3 fraction; α mobility) contains both ApoAI and ApoAII Other minor species contain both ApoAI and ApoAIV, or just ApoAIV

Heterogeneity in HDL particles also extends to the phospholipid composition (Rye

et al., 1999) The major phospholipid found on the surface of HDL is phosphatidylcholine, with many species involved depending on the type of fatty acid residues Most have a saturated fatty acid in the γ position, either C16 palmitic acid or C18 stearic acid, and an unsaturated fatty acid in the β position, such as oleic acid, linoleic acid or arachidonic acid (Subbaiah and Pritchard, 1989) Other HDL phospholipids include sphinogomyelin, phosphatidylserine, phosphatidylinositol and phosphatidylethanolamine (Rye et al., 1999)

2.2.2 Laboratory Determination of HDL

In routine biochemical analysis, HDL in plasma is assayed as the total cholesterol content (free cholesterol and CE) following conversion of all CE to cholesterol, hence the term HDL-C This is equivalent to the HDL band obtained by density gradient

ultracentrifugation and the α-HDL resolved by agarose gel electrophoresis Nuclear magnetic spectroscopy provides a rapid and convenient method of measuring the levels

of the different lipoprotein subclasses, thereby overcoming the limitations of existing laborious analytical methods that utilize centrifugation and gel electrophoresis (Otvos, 2000)

Clinical Trials

A low level of HDL-C is a strong, independent risk factor for CAD This observation was noted as early as in the 1950s but ignored until its rediscovery two decades later, possibly because HDL-C is normally a minor constituent (20-25%) of total plasma cholesterol (Breslow, 1995) Since then, many case-control, prospective and intervention studies have reaffirmed the inverse relationship between HDL and CAD In the Framingham Heart Study, 2815 men and women between the ages of 49 and 82 were monitored for four years for development of CAD and lipid levels The HDL-C level was found to be a potential risk factor, especially among older participants, and was significant even after other risk factors were considered (Gordon et al., 1977) The participants were followed up for another eight years, and again, low HDL was confirmed

as an independent risk factor of CAD, even after accounting for multiple risk variables like total cholesterol, alcohol consumption, blood pressure, cigarette smoking and body mass index (BMI), and also after stratifying individuals according to total cholesterol levels (Castelli et al., 1986) The Prospective Cardiovascular Munster (PROCAM; Assman et al., 1996) and the Quebec Cardiovascular Studies (Despres et al., 2000) also supported the notion of an inverse relationship between HDL level and the risk of CAD A meta-analysis of four prospective observational studies estimates that every 1 mg/dL (0.025

mM) reduction in HDL-C translates into a 2-3% increase in CAD risk whereas a 1 mg/dL increment in LDL-C raises CAD risk by 1% (Gordon et al., 1989)

While reducing LDL with statins has successfully reduced coronary events and cardiovascular risks, HDL remains an independent risk factor for CAD, as evidenced in several statin trials in which increases in HDL levels with statin therapy were associated with modest reductions in coronary events (Sacks, 2001) In the West of Scotland Coronary Prevention Study (WOSCOPS), pravastatin-treated patients were monitored for five years for the incidence of major coronary events Despite the overall result that pravastatin treatment lowered the risk of coronary events, patients with a lower baseline

of HDL still had higher risks than those with a higher baseline, suggesting the statin therapy did not alter the risk of low HDL In the Helsinki Heart Study, the randomized, double-blind primary intervention trial of gemfibrozil administration to middle-aged men with primary dyslipidemia over five years showed a 11% increase in HDL and reduction in LDL-C was accompanied by a 34% reduction in CAD incidence, a rate much lower than that predicted based on the extent of LDL-C lowering (Frick et al., 1987) A similar effect

of gemfibrozil has been demonstrated in the Veterans Affairs HDL Intervention Trial (VA-HIT) in 2531 men with CAD selected for low HDL-C (average 40 mg/dL) and normal

LDL-C (140 mg/dL) (Rubins et al., 1999) After a five-year followup, the treatment group showed a 24% reduction in combined end points (defined as death from CAD, stroke, nonfatal myocardial infarction and stroke) Fibrates like gemfibrozil are potent agonists of peroxisome proliferator-activated receptor (PPAR)-α, a nuclear receptor with actions that include increased lipoprotein lipase (LPL)-mediated lipolysis leading to lower TG, and transcriptional induction of ApoAI and ApoAII culminating in higher HDL (Staels et al., 1998)

These clinical trials independently demonstrate that HDL level is an independent risk factor of CAD and that the elevation of HDL has beneficial effects The clinical

significance of HDL-C levels has been recognized by the National Cholesterol Education Program Adult Treatment Panel III with the latest guidelines raising the CAD risk threshold for low HDL-C from 35 to 40 mg/dL (Expert Panel on Detection, Evaluation, And Treatment of High Blood Cholesterol in Adults, 2001)

2.4 Atheroprotective Mechanisms of HDL

Against the background of numerous epidemiological studies providing overwhelming evidence that HDL is atheroprotective, direct mechanistic explanations have been sought The atheroprotective attributes of HDL are diverse and are reviewed below

2.4.1 Protection against LDL Oxidation

HDL, particularly the HDL3 fraction, protects against LDL oxidation The appearance of oxidized LDL (oxLDL) species in the subendothelial matrix is a key initiating event in early atherogenesis OxLDL serves as a chemoattractant for monocytes, transforms macrophages to foam cells, exerts cytotoxic effects on endothelium, activates platelets, stimulates movement and proliferation of smooth muscle cells, and antagonizes the vasorelaxation effect of nitric oxide (NO) (Nofer et al., 2002) The anti-oxidative properties

of HDL are in part attributed to its associated enzymes such as paraoxonase (PON), platelet activating factor acetylhydrolase (PAFH) and glutathione peroxidase, as well as its structural components, ApoAI and certain lipid constituents (Assman and Nofer, 2003)

PON acts by preventing formation of minimally oxidized LDL as well as destroying the biologically active oxidized phospholipids on LDL once they are formed (Watson et al., 1995; Navab et al., 2000) A direct evidence for the anti-atherogenic effect of PON comes from the finding that PON-deficient mice have larger aorta lesions than wild type

and the HDL from these mice fails to prevent LDL oxidation in vitro (Shih et al., 1998) In

animal models prone to atherosclerosis such as ApoE- or LDLR-knockout mice, raised levels of oxidation markers were accompanied by reduced paraoxonase activity (Shih et

al., 1996) PON colocalizes with the fraction of HDL that contains both ApoAI and ApoJ (Kelso et al., 1994)

Apart from hydrolyzing platelet activating factor which is a potent lipid mediator with proinflammatory properties, PAFH, like PON and glutathione peroxidase, can destroy oxidized phospholipids on modified LDL (Assman and Nofer, 2003)

In addition, natural antioxidants found on HDL such as ApoAI which possesses several methionine residues, α-tocopherol and other lipophilic molecules can scavenge reactive oxygen species, guarding against the formation and accumulation of lipid hydroperoxides and oxidized CE, seeding molecules that induce the oxidation of phospholipids non-enzymatically (Assman and Nofer, 2003)

2.4.2 Prevention of Platelet Aggregation

HDL inhibits the propensity of platelets to aggregate, possibly mediated through phospholipase A2 formation of lysophosphatidyl from phosphatidylcholine (Yuan et al., 1995), attenuation of 12-lipooxygenase activity in platelets (Fujimoto et al., 1994) and increased NO production (see below)

2.4.3 Stimulation of Prostacyclin Synthesis

HDL induces biosynthesis of prostacyclin, specifically via a cyclooxygenase-2-dependent mechanism involving the mitogen-activated protein kinase (MAPK) pathway (Vinals et al., 1999) Prostacyclin is a vasodilator prostaglandin synthesized by vascular smooth muscle cells Its other atheroprotective actions include inhibition of platelet aggregation and adhesion, smooth muscle cell proliferation, leukocyte activation and adhesion, as well as reduction of CE accumulation in the vessel wall

2.4.4 Modulation of Endothelial Adhesion

An early event in atherosclerosis is the attachment of monocytes and lymphocytes but not neutrophils to the artery wall, triggered by minimally oxLDL Cockerill et al (1995) found that HDL modulates the mRNA and protein expression levels of cell adhesion molecules intercellular adhesion molecule-1 (ICAM-1), vascular cell adhesion molecule-1 (VCAM-1) and E-selectin in stimulated vascular endothelial cells E-selectin enables tethering and rolling of monocytes and lymphocytes along the endothelial surface while subsequent adhesion is mediated by ICAM-1 and VCAM-1 The marked inhibition is observed within the physiological range of HDL levels and is apparent with both native HDL and reconstituted discoidal HDL containing only ApoAI, phosphatidylcholine and free cholesterol

2.4.5 Stimulation of Endothelial Nitric Oxide Synthase (eNOS)

In endothelial cells, HDL binding to scavenger receptor BI (SR-BI) activates the tyrosine kinase receptor Src and phosphatidylinositol-3-kinase, triggering both the MAPK and Akt pathways and leading to phosphorylation of eNOS (Mineo and Shaul, 2003) eNOS activation generates NO which mediates a plethora of anti-atherogenic activities including vasorelaxation, attenuation of platelet adhesion and aggregation, inhibition of smooth muscle cell proliferation, reduced adhesion and migration of leukocytes to the vessel wall (Nofer et al., 2002)

2.4.6 Participation in Reverse Cholesterol Transport (RCT)

Much of the anti-atherogenic property of HDL has been ascribed to its participation in RCT (Glomset, 1968) RCT is the pathway in which excess cholesterol from the peripheral cells is delivered to the liver for excretion in bile or for steroid hormone synthesis In the non-growing adult animal, an amount of cholesterol equal to those

synthesized and absorbed must be excreted daily Humans synthesize about 10 mg cholesterol/day/kg (Dietschy et al., 1993) The appearance and accumulation of foam cells in the arterial wall, which are primarily macrophages engorged with CE, is a hallmark of atherosclerosis (Lusis, 2000) Scavenger receptors of class A and CD36 on the macrophage permit excessive uptake of modified lipoproteins (Linton and Fazio, 2001) Several steps can be identified in RCT: (i) cholesterol efflux; (ii) LCAT-mediated esterification of cholesterol; (iii) remodelling and HDL maturation in the plasma; (iv) delivery to liver and steroidogenic tissues

RCT is depicted schematically in Figure 2.1 The first step of RCT begins with cholesterol efflux, the movement of excess cellular cholesterol to external acceptors Cholesterol efflux can proceed in both passive and active manners (Yancey et al., 2003) Aqueous diffusion of cholesterol, which occurs in all cells, is generally slow and inefficient, on the time scale of hours SR-BI-facilitated efflux is more rapid but like passive diffusion, it proceeds via a concentration gradient and therefore cholesterol flux can proceed both ways The principal cholesterol acceptors in aqueous diffusion and

SR-BI mediated efflux are phospholipid-containing molecules; in the former, the preferred acceptors are small, discoidal whereas SR-BI interacts with a wide range of acceptors such as HDL and LDL (Yancey et al., 2003) On the other hand, ABCA1 mediates cholesterol and phospholipid efflux to lipid-poor apolipoproteins (such as pre-β HDL regenerated from the remodelling activities in the plasma, free ApoAI or nascent HDL which are newly secreted by the liver and small intestine) in an energy-dependent and unidirectional manner (Yancey et al., 2003) Because ABCA1 has been identified as the molecular defect in Tangier disease, a disorder characterized by severe HDL-deficiency and impaired cholesterol efflux, it has been postulated that that ABCA1 accounts for the

bulk of the cholesterol efflux in vivo

Following the efflux of cholesterol onto the surface of the nascent HDL particle, it

is esterified by LCAT, an enzyme bound to HDL surface The hydrophobic CE then partitions into the core, leaving the surface free to acquire more cholesterol, thus maintaining a cholesterol gradient from cells to HDL As CE accumulates, the initially discoidal HDL is converted into the large, spherical HDL which is the predominant form of HDL in the vascular system

In the plasma, HDL undergoes further remodelling by several enzymes CETP catalyzes the reciprocal exchange of CE and TG between HDL and TG-rich apoB-containing lipoproteins (LDL, VLDL, chylomicron remnants) Another HDL-associated enzyme, PLTP, swaps CE of HDL for the phospholipids in TG-rich lipoproteins, regenerating pre-β HDL (van Tol, 2002) Hepatic lipase hydrolyses phospholipids and TG

on HDL as well as VLDL remnants and LDL, generating smaller HDL particles (Santamarina-Fojo et al., 1998) Endothelial lipase, primarily a phospholipase AI, also remodels HDL to smaller particles (Choi et al., 2002)

Much of the cholesterol initially received by HDL becomes re-partitioned to apoB-containing lipoproteins which are subsequently taken up as whole particles by LDLR-mediated endocytosis in the liver and steroidogenic tissues The remaining CE in remnant HDL particles is catabolized by being selectively taken up by SR-BI in liver and steriodogenic tissues (Acton et al., 1996)

Based on evidence from animal studies, it has been recently proposed that ABCA1 activity in the liver, and not peripheral macrophages, is a major source of plasma HDL-C (Haghpassand et al., 2001; Tall et al., 2001; Aiello et al., 2002; van Eck et al., 2002; Basso et al., 2003) This represents a major paradigm shift in the classical concept

of RCT as a one-way pathway from the periphery to the liver