Plasma cholesteryl ester transfer protein (CETP) catalyzes the transfer of CEs from high-density lipoproteins (HDLs) to triglyceride-rich and low-density lipoproteins (LDLs). The hypothesis that CETP inhibition will prevent cardiovascular disease (CVD) was based on the fact that low activity increases HDL cholesterol and decreases LDL cholesterol. Early reports that CETP gene transfer increased atherosclerosis in mice, and that CETP inhibition reduced lesions in cholesterol-fed rabbits fuelled enthusiasm for the approach. Although some cautioned that the development of CETP inhibitors was premature owing to uncertainties about their impact on the remodeling of HDLs and reverse cholesterol transport (RCT) in humans (Fielding and Havel, 1996), drug discovery programmes proceeded. Two parallel research tracks then emerged. While industry progressed to clinical studies, academics sought to clarify the relations of CETP activity to RCT and atherosclerosis. Two drugs reached Phase 3 clinical trials. However, ILLUMINATE was terminated when torcetrapib was found to have had increased CVD. Five years later, Dal-OUTCOMES was aborted when it was evident dalcetrapib was not conferring any benefit. After reviewing the literature up to May 2014, Miller (2014) concluded that CETP inhibition is more likely to increase CVD than prevent it, and was of the opinion that ongoing trials should be stopped. Since then several further pertinent studies have been published. They include four meta-analyses of the Taq1B polymorphism of the CETP gene. Cao et al. (2014) concluded that B2B2 homozygotes (low CETP activity, high HDL cholesterol) have a lower risk of myocardial infarction (MI) than B1B1 subjects. Using Mendelian randomization, Wu et al. (2014) found the B2 allele to be associated with a lower risk of coronary heart disease (CHD). However, another Mendelian randomization (Niu and Qi, 2015) found the B2 allele to be associated with a higher risk of CVD than the B1 allele. In this context, an earlier meta-analysis by Dullaart and Sluiter (2008) is of interest. These authors found that B2B2 carriers were less frequent among cases drawn from samples at high CVD risk than among cases drawn from population-based samples. Furthermore, in the latter case the odds ratio for CVD was 1.45 (95% CI: 1.07–1.95) in B2B2 relative to B1B1, while in the former it was 0.84 (0.74–0.96), suggesting that in the general population the B2 allele is actually associated with higher CVD risk in spite of the higher HDL cholesterol. Regieli et al. (2008) had come to a similar conclusion in the REGRESS study. After following 812 men with CHD on statins for 10 years, the B2 allele was associated with a hazard ratio for CVD death of 1.59 (P = 0.01) despite the expected low CETP activity and high HDL cholesterol. Four additional meta-analyses looked at the impact of CETP inhibitors on CVD risk as part of larger studies of the effects of HDL cholesterol-raising agents in patients taking statins. All four concluded that the trials have not demonstrated a beneficial effect (Keene et al., 2014; Hourcade-Potelleret et al., 2015; Ip et al., 2015; Verdoia et al., 2015). In clinical studies, Gu et al. (2014) reported that although the A allele of the −629°C/A polymorphism of the CETP gene was associated with a lower plasma CETP concentration than the C allele, it had not reduced CVD events in patients taking atorvastatin. Kastelein et al. (2015) observed no significant effect of anacetrapib on CVD incidence (four events vs. zero in the treated and placebo groups, respectively) during 12 months of follow-up in patients with heterozygous familial hypercholesterolemia already on lipid-lowering treatment. Scharnagl et al. (2014) confirmed earlier reports that human plasma samples with low CETP concentrations were less effective in promoting cholesterol efflux from cultured macrophages than samples with high concentrations. In animal studies, Kuhnast et al. (2015) found that anacetrapib reduced atherosclerosis in APOE*3Leiden.CETP transgenic mice. Briand et al. (2014) compared anacetrapib with dalcetrapib in hamsters, a species with natural CETP. In normal animals, neither drug at doses equipotent for CETP inhibition (by 60%) had any effect on macrophage-to-feces RCT, although they did lower equally the fractional clearance rate of HDL-CE. In dyslipidaemic animals, anacetrapib increased RCT, whereas an equipotent dose of dalcetrapib reduced it. Liu et al. (2015) found that inhibition of DNA topoisomerase II (Topo II) by etoposide, tenipooside or Topo II siRNA increased CETP gene expression and CETP secretion in HepG2 cells. When given to CETP transgenic mice, teniposide induced CETP expression in the liver, and increased macrophage-to-feces RCT to a greater degree than in wild-type mice with no CETP. Using a computer model of lipoprotein metabolism to analyze the on/off kinetics of the short-acting potent CETP inhibitor RG7232, Lu et al. (2015) concluded that inhibition of CETP is likely to reduce prebeta HDL production in humans. During the past 25 years, lipidologists have become accustomed to the controversies in this field. While the decrease in LDL cholesterol has been assumed to be beneficial, the rise in HDL cholesterol has long prompted discussion on three fronts. First, the early concerns (Fielding and Havel, 1996) that HDL remodeling might be deranged, leading to lowered prebeta HDL production, have not been allayed. Second, the early assumption that the rise in HDL CE would increase the direct delivery of CEs to the liver via SR-B1 receptors was challenged by evidence that the receptors may be saturated at normal plasma HDL concentrations (Woollett and Spady, 1997; Nieland et al., 2011). Third, evidence was reported that the large CE-rich HDLs produced by CETP inhibition might be dysfunctional. Despite these concerns, momentum was maintained in the hope that the reduction of LDL would more than offset any adverse effects on HDL. However, as time has gone on the cumulative evidence has increasingly pointed to CETP having both a facilitative role in RCT and a net preventative effect on CVD. The recent work summarized above has not weakened this evidence. Although the debate has centered largely on whether CETP inhibition is likely to have a beneficial or detrimental impact on CVD, the possibility that the net effect of the combined changes in LDL and HDL metabolism might vary according to the prevailing physiologic conditions, and therefore from subject to subject, and from time to time in the same subject, also merits consideration. Being at a crossroad in the transport of lipids, and having multiple effects on lipoprotein metabolism, the overall effects of CETP activity and its inhibition might vary according to the ambient HDL and LDL particle concentrations, for example, or to interplay with other genes. Reports of interactions between CETP alleles and alleles of the lipoprotein lipase (Corsetti et al., 2011), hepatic lipase (Soyal et al., 2011), apolipoprotein E (Sun et al., 2014), and nitric oxide synthase (Rahimi et al., 2012) genes seem to add weight to this possibility. It remains that the tandem hypotheses that CETP inhibition will both enhance RCT and prevent CVD are without sound scientific bases. Indeed a growing body of evidence of three kinds now supports the contrary. First, in three out of three experimental studies of their type, macrophage-to-feces RCT in vivo was increased when CETP expression was enhanced (Briand et al., 2014; Miller, 2014). Second, each of five studies that examined the effect of CETP activity in human plasma on cholesterol efflux from cultured cells in vitro found that high activity plasma was more effective than low activity (Miller, 2014; Scharnagl et al., 2014). Third, each of six prospective cohort observational studies have found that CETP concentration or activity was related inversely to CVD incidence (Miller, 2014). In the author's opinion, this is sufficient to justify stopping all ongoing clinical trials of CETP inhibitors. Carrying on and hoping for the best, while participants are unaware of the evolving evidence base, is not acceptable.

Hyperlipidemia is a common feature of diabetes and is related to cardiovascular disease. The very low-density lipoprotein receptor (VLDL-R) is a member of the low-density lipoprotein receptor (LDL-R) family. It binds and internalizes triglyceride-rich lipoproteins with high specificity. We examined the etiology of hyperlipidemia in the insulin-deficient state. VLDL-R expression in heart and skeletal muscle were measured in rats with streptozotocin (STZ)-induced diabetes. STZ rats showed severe hyperlipidemia on d 21 and 28, with a dramatic decline in VLDL-R protein in skeletal muscle (>90%), heart (approximately 50%) and a loss of adipose tissues itself on d 28. The reduction of VLDL-R protein in skeletal muscle could not be explained simply by a decrease at the transcriptional level, because a dissociation between VLDL-R protein and mRNA expression was observed. The expression of LDL-R and LDL-R-related protein in liver showed no consistent changes. Furthermore, no effect on VLDL-triglyceride production in liver was observed in STZ rats. A decrease in postheparin plasma lipoprotein lipase activity started on d 7 and continued to d 28 at the 50% level even though severe hyperlipidemia was detected only on d 21 and 28. In rat myoblast cells, serum deprivation for 24 h induced a reduction in VLDL-R proteins. Insulin (10(-6) m), but not IGF-I (10 ng/ml), restored the decreased VLDL-R proteins by serum deprivation. These results suggest that the combination of VLDL-R deficiency and reduced plasma lipoprotein lipase activity may be responsible for severe hyperlipidemia in insulin-deficient diabetes.

High-density lipoproteins (HDLs) have several metabolic actions in vitro that are potentially anti-atherogenic. In addition to their role in reverse cholesterol transport, native HDLs have been shown to protect low-density lipoproteins (LDLs) against oxidative modification, to have anti-inflammatory properties, and to inhibit platelet aggregation. These actions have been shown to occur also in vivo in both experimental animals and humans, when plasma HDL concentration is raised by intravenous infusion of native HDLs or reconstituted discoidal particles composed of the major HDL protein, apolipoprotein (apo) A-I, in association with phosphatidylcholine (PC) and, in the case of protection of LDLs against oxidative change, by lipid-free apo A-I and apo A-I mimetic polypeptides. Intravenous infusion of native HDLs, lipid-free apo A-I and apo A-I/PC discs, and oral administration of apo A-I peptides, have been found to prevent or reverse experimentally induced atherosclerosis in animals. A mutant form of apo A-I discovered in Italy (apo A-IMilano), the biological properties of which differ somewhat from those of normal apo A-I appears to be even more potent in this regard. The hope that this approach will provide a new effective therapy for atherosclerosis has been supported by a multi-center clinical trial, in which five weekly infusions of apo A-IMilano/PC discs induced significant regression of coronary lesions, as quantified by intravascular ultrasound, in men with clinical coronary heart disease. This article reviews the data from animal and human studies in this rapidly developing area. Keywords: apolipoprotein a, atherosclerosis, cholesterol, coronary heart disease, high-density lipoproteins

We have previously shown that intravenous apolipoprotein (apo) A-I/phosphatidylcholine (apo A-I/PC) discs increase plasma high-density lipoprotein (HDL) concentration in humans. We have now studied the associated changes in two enzymes, paraoxonase (PON) and platelet-activating factor acetylhydrolase (PAF-AH) that are carried in whole or in part by HDLs, and are thought to influence atherogenesis by hydrolyzing oxidized phospholipids in lipoproteins. Apo A-I/PC discs (40mg/kg over 4h) were infused into eight healthy males. Although plasma apo A-I and HDL cholesterol increased on average by 178 and 158%, respectively, plasma total PON and total PAF-AH concentrations did not rise. By the end of the infusion, HDL-associated PAF-AH had increased by 0.56 ± 0.14μg/mL (mean ± S.D., P < 0.01), and nonHDL-associated PAF-AH had decreased by 0.84 ± 0.11μg/mL (P < 0.05). These changes were accompanied by an increase in the HDL-associated PAF-AH/apo A-I ratio from 0.19 to 0.35 (P < 0.05), and by a decrease in the nonHDL-associated PAF-AH/apo B ratio from 2.1 to 1.4 (P < 0.05). No changes in PON or PAF-AH concentrations were detected in prenodal lymph (tissue fluid), collected continuously from the leg. Our results show that the total concentrations of PON and PAF-AH in plasma are uninfluenced by plasma HDL concentration. PAF-AH transfers readily between HDLs and LDLs in vivo, and its distribution between them is determined partly by their relative concentrations and partly by HDL composition.

We have generated transgenic rabbits that express the entire human apoA-I/C-III/A-IV gene cluster. As in humans, h-apoA-I and h-apoC-III were expressed in liver and intestine, whereas h-apoA-IV mRNA was detected in intestine only. Transgenic rabbits had significantly higher plasma total cholesterol, HDL-cholesterol and total phospholipid concentrations than non-transgenic littermates. In contrast to similar transgenic mice previously generated, which have gross hypertriglyceridemia, triglyceride concentrations were only moderately raised in transgenic rabbits. Plasma and HDL from transgenic rabbits were more effective than those from controls in promoting cholesterol efflux from cultured hepatoma cells. They had lower LCAT, lower CETP and higher PLTP activities than non-transgenic littermates. Cholesterol-feeding produced major increases in plasma lipids. The qualitative response to the diet was not modified by cluster expression. Human apoA-I concentration was halved by cholesterol-feeding, whereas h-apoC-III and h-apoA-IV concentrations were not significantly altered. Cholesterol efflux from hepatoma cells to plasma and HDL was not altered by the diet. Since lipoprotein metabolism of rabbits closely resembles that of humans, human apoA-I/C-III/A-IV transgenic rabbits may provide a reliable model for studies of the transcriptional regulation of the cluster, and for evaluating the effects of different agents on the expression of the three genes.

BACKGROUND:In individuals heterozygous for ABCA1 transporter mutations, defective reverse cholesterol transport (RCT) causes low HDL-cholesterol and premature coronary heart disease (CHD). However, the extent to which impaired RCT underlies premature CHD in others with low HDL-cholesterol is not known. The primary acceptors of cell cholesterol are a minor subclass of lipid-poor pre-beta-HDLs. These are generated during remodeling of alpha-HDLs, which account for almost all HDL-cholesterol. We studied the strength of the association of CHD with pre-beta-HDL concentrations in Japanese men.METHODS:Blood was collected from 42 men with clinical CHD and 44 healthy controls 40-70 years of age. Pre-beta-HDL was assayed by crossed immunoelectrophoresis.RESULTS:Cases had lower HDL-cholesterol (-23%), total apolipoprotein A-I (-26%), and pre-beta-HDL (-55%; all P <0.001) concentrations; lower pre-beta-HDL:alpha-HDL ratios (-45%; P <0.001); and higher plasma triglycerides (20%; P <0.03) than the controls. On stepwise logistic regression, CHD was associated most strongly with pre-beta-HDL concentrations. On ROC analysis, pre-beta-HDL concentration discriminated between cases and controls better than any other lipoprotein measurement. When plasma was incubated for 16 h at 37 degrees C, mean (SD) pre-beta-HDL increased by 47 (36)% in controls, but was unchanged in cases (group difference, P <0.001).CONCLUSIONS:Our results suggest that inefficient RCT, secondary to a low pre-beta-HDL concentration and production rate in plasma, contributes to premature CHD in Japanese men with low HDL-cholesterol.

We studied the variations in the concentrations of cholesterol, triglycerides, phospholipids, apolipoproteins (apos) (A-I, A-II, B, C-III, E), free glycerol and albumin in human prenodal leg lymph during the 24 h cycle. Lymph was collected continuously for up to 96 h from nine healthy males on a low-fat isocaloric diet. In three free-living subjects, all lipid and apolipoprotein concentrations underwent synchronous variations, rising during the night and decreasing during the day. In three subjects who remained in supine rest for 48 h, the amplitude of circadian variation was much smaller. In three who alternated periods of supine rest with upright exercise, the highest concentrations occurred during rest. Lipid, apolipoprotein and albumin concentrations were inversely related to lymph flow rate. Free glycerol, much of which in tissue fluid is derived from local adipocytes, did not follow this pattern. On multiple regression, concentrations in lymph were related independently to the corresponding concentration in plasma (positive) and to lymph flow rate (negative) or lymph albumin concentration (positive). These results show that lipoprotein concentrations in human tissue fluid are determined only partly by their concentrations in plasma. They are also strongly affected by hemodynamic factors via their effects on fluid transport.

We studied a four-generation family (17 subjects) with familial lecithin:cholesterol acyltransferase (LCAT) deficiency. A 30-year-old Caucasian male with corneal clouding and HDL cholesterol <0.1 mmol/l was a compound heterozygote for a novel mutation (Phe382→Val), a previously reported mutation (Thr321→Met) and a common variant (Thr208→Ser) of the gene. Immunoreactive LCAT concentration (1.2 μg/ml), α-LCAT activity (13 nmol/ml per h) and cholesterol esterification rate (CER) (14 nmol/ml per h) in his plasma were, respectively, 14, 8 and 14% of the mean values in healthy subjects. The proband and 13 of his relatives also had familial defective apo B (FDB, Arg3500→Gln). Six subjects had LCAT Phe382→Val in combination with FDB. Plasma lipoprotein(a) (Lp(a)) was 24 nmol/l in the proband and 46–211 nmol/l in his father and siblings, consistent with expression of the 16 kringle 4 isoform. The proband had no signs of coronary heart disease (CHD), but his father, a paternal uncle and a female cousin had CHD before age 38 years.

OBJECTIVE:Androgens and other drugs that reduce plasma concentrations of high density lipoprotein (HDL) cholesterol are often considered to be pro-atherogenic. Tibolone lowers HDL-cholesterol by 20% but the clinical significance of this effect is unknown. METHODS:In a randomized, double-blind study, 34 women received 2.5 mg tibolone daily and 34 women received placebo. Serum concentrations of lipids, lipoprotein subclasses and apolipoproteins, together with plasma activities of lipid transfer proteins and lipolytic enzymes and the capacity of plasma to induce cholesterol efflux from cultured cells, were measured. RESULTS:Compared to placebo, tibolone reduced serum concentrations of HDL-cholesterol (-14%), HDL phosphatidylcholine (-14%), apolipoprotein (apo)A-I (-12%), HDL subclasses lipoprotein (Lp)A-I (-20%), HDL-apoE (-16%), pre beta-LpA-I (-10%) and alpha-LpA-I (-12%) and increased hepatic lipase activity (+25%) and HDL sphingomyelin : phosphatidylcholine ratio (10.5%), but did not alter serum concentrations of HDL sphingomyelin, apoA-IV and LpA-I/A-II, lipoprotein lipase, the plasma activities of lecithin : cholesterol acyl transferase, cholesteryl ester transfer protein, phospholipid transfer protein or the plasma capacity to release cholesterol from cultured fibroblasts or Fu5AH hepatocytes. CONCLUSIONS:Tibolone lowers HDL-cholesterol in part by increasing hepatic lipase activity. Conservation of sphingomyelin and apoA-II in HDL, as well as cholesteryl ester transfer protein activity, preserves the capacity of plasma to release cholesterol, despite the lower concentrations of HDL-cholesterol. This may have important implications for the use of steroid effects on HDL concentrations as surrogates for atherosclerosis.

Platelet-activating factor acetylhydrolase (PAF-AH) is a phospholipase A2 associated with lipoproteins that hydrolyzes platelet-activating factor (PAF) and oxidized phospholipids. We have developed an ELISA for PAF-AH that is more sensitive than previous methods, and have quantified HDL-associated and non-HDL-associated PAF-AH in healthy, hyperlipidemic, and diabetic subjects. In healthy subjects, plasma total PAF-AH concentration was positively correlated with PAF-AH activity and with plasma total cholesterol, triacylglycerol, LDL cholesterol and apolipoprotein B (apoB) concentrations (all P < 0.01). HDL-associated PAF-AH concentration was correlated positively with plasma apoA-I and HDL cholesterol. Subjects with hyperlipidemia (n = 73) and diabetes mellitus (n = 87) had higher HDL-associated PAF-AH concentrations than did controls (P < 0.01). Non-HDL-associated PAF-AH concentration was lower in diabetic subjects than in controls (P < 0.01). Both hyperlipidemic and diabetic subjects had lower ratios of PAF-AH to apoB (P < 0.01) and higher ratios of PAF-AH to apoA-I (P < 0.01) than did controls. Our results show that the distribution of PAF-AH mass between HDLs and LDLs is determined partly by the concentrations of the lipoproteins and partly by the mass of enzyme per lipoprotein particle, which is disturbed in hyperlipidemia and diabetes mellitus.

The central roles played by lipoproteins in atherosclerosis are well established. Increased plasma concentrations of low-density lipoproteins (LDLs) and triglyceride-rich remnant lipoproteins are highly atherogenic, whereas high-density lipoproteins (HDLs) are known to protect against lesion development. These effects are driven, in part, by the impact of these lipoproteins on inflammation—a process that is central to atherogenesis. In individuals with dyslipidaemia, LDLs and other atherogenic lipoproteins enter the arterial wall where they undergo chemical modification, including oxidation. These modified lipoproteins initiate the inflammatory process that culminates in atherosclerosis lesion development. The inflammation can be reversed by HDLs via several mechanisms. These include promotion of cholesterol efflux, inhibition of LDL oxidation and reduction of adhesion molecule expression. Recent work has shown that HDLs are also able to inhibit acute vascular inflammation. Given the central roles played by lipoproteins and inflammation in atherogenesis, effective anti-atherosclerotic treatments should both modify the lipid profile and target the ongoing inflammation. These criteria are fulfilled by statins, which reduce inflammation by both lipid-dependent and -independent mechanisms. Additional protection from atherosclerosis may be provided by novel therapies that aim to increase plasma levels and activity of HDLs.

The salutary effects of high-density lipoproteins (HDLs) in animal and human models of endotoxic shock have in the past been attributed to the ability of this lipoprotein to bind to lipopolysaccharide. However, the precise mechanisms for the protective effect of HDL are unclear. The first objective of this study was to determine the effects of HDLs on the organ injury and dysfunction associated with acute severe endotoxemia. Second, to gain insight into the mechanism of action of HDL, we also investigated the effect of HDLs on 1) the expression of P-selectin and intercellular adhesion molecule-1 in the kidneys of rats treated with endotoxin and 2) the rise in the plasma levels of tumor necrosis factor-alpha (TNF-alpha). Rats were given Escherichia coli lipopolysaccharide (6 mg/kg i.v.), pretreated with either vehicle (n = 9) or reconstituted HDL (rHDL; apolipoprotein A-I/phosphatidylcholine proteoliposomes, n = 10), and were monitored for 6 h. Here we report that rHDL attenuates the renal injury and dysfunction caused by endotoxin in the rat. In addition, rHDL reduced the degree of histological tissue injury in the lung, liver and intestine and attenuated the expression of P-selectin and intercellular adhesion molecule-1 in the renal glomerulus. Interestingly, pretreatment of rats with rHDL did not prevent the hypotension nor the rise in plasma levels of TNF-alpha (at 90 min) caused by endotoxin. Thus, rHDL reduces the organ injury/dysfunction, but does not affect the circulatory failure, nor the rise in plasma levels of TNF-alpha caused by endotoxin in the rat. We propose that the mechanisms of these beneficial effects of HDL may be related to direct inhibition of adhesion molecule expression.

Plasma phospholipid transfer protein (PLTP) plays an important role in lipoprotein metabolism and reverse cholesterol transport. We have recently reported that plasma PLTP concentration correlates positively with plasma HDL cholesterol (HDL-C) but not with PLTP activity in healthy subjects. We have also shown that PLTP exists as active and inactive forms in healthy human plasma. In the present study, we measured plasma PLTP concentration and PLTP activity, and analyzed the distribution of PLTP in normolipidemic subjects (controls), cholesteryl ester transfer protein (CETP) deficiency, and hypo-alphalipoproteinemia (hypo-ALP). Plasma PLTP concentration was significantly lower (0.7 ± 0.4 mg/l, mean ± SD, n = 9, P < 0.001) in the hypo-ALP subjects, and significantly higher (19.5 ± 4.3 mg/l, n = 17, P < 0.001) in CETP deficiency than in the controls (12.4 ± 2.3 mg/l, n = 63). In contrast, we observed no significant differences in plasma PLTP activity between controls, hypo-ALP subjects, and CETP deficiency (6.2 ± 1.3, 6.1 ± 1.8, and 6.8 ± 1.2 μmol/ml/h, respectively). There was a positive correlation between PLTP concentration and plasma HDL-C (r = 0.81, n = 89, P < 0.001). By size exclusion chromatography analysis, we found that the larger PLTP containing particles without PLTP activity (inactive form of PLTP) were almost absent in the plasma of hypo-ALP subjects, and accumulated in the plasma of CETP deficiency compared with those of controls.These results indicate that the differences in plasma PLTP concentrations between hypo-ALP subjects, CETP deficiency, and controls are mainly due to the differences in the amount of the inactive form of PLTP.

The effects of common variants of cholesteryl ester transfer protein (CETP) (TaqIB), hepatic lipase (HL) (−514C>T), lipoprotein lipase (LPL) (S447X) and lecithin cholesterol acyl transferase (LCAT) (S208T) on the determination of high density lipoprotein cholesterol (HDL‐C) and apolipoprotein AI (apoAI) levels were examined in 2773 healthy middle‐aged men participating in the second Northwick Park Heart Study. The extent of gene:gene, gene:smoking and gene:alcohol interactions were determined. For HDL‐C levels, only CETP genotype was associated with significant effects (p<0·0001), with the B2 allele being associated with higher levels in both smokers and non‐smokers. This interaction was significant at the lowest tertile of TG, suggesting that TG levels were rate limiting. As previously reported, CETP, LPL and HL genotypes were all associated with significant effects on apoAI levels (all p<0·01), with carriers of the rare alleles having higher levels and with no evidence of heterogeneity of effects in smokers and non‐smokers. LCAT genotype was not associated with significant effects on either trait. There was no significant interaction between any of the genotypes and alcohol consumption on either HDL‐C or apoAI levels. All genotypic effects were additive for HDL‐C and apoAI. Environmental and TG levels explained more than 20% and 5·5% of the variance in HDL‐C and apoAI, respectively. The novel aspect of this finding is that genetic variation at these loci explained in total only 2·5% of the variance in HDL‐C and 1·89% of the variance in apoAI levels. Thus despite the key roles played by these enzymes in HDL metabolism, variation at these loci, at least as detected by these common genotypes, contributes minimally to the variance in HDL‐C and apoAI levels in healthy men, highlighting the polygenic and multifactorial control of HDL‐C.

SummaryPostprandial lipaemia is associated with activation of factor VII (FVII) and efflux of cholesterol from tissues to nascent plasma high density lipoproteins (HDL) containing apolipoprotein A-I (apo A-I). To determine whether FVII activation and cholesterol efflux occur together in other situations, the responses to intravenous infusion of HDL-like apo A-I/phosphatidylcholine discs were measured in 10 healthy men. Disc infusion (40 mg apo A-I/kg body weight) over 4 h was followed by increases in HDL cholesteryl ester and plasma apo A-I (p <0.0001). Significant activation of FVII was apparent during infusion in fasting subjects (p = 0.03), activated FVII averaging 123% of baseline value by 12 h (p <0.0001). Plasma thrombin-antithrombin (TAT) complex increased to 156% of baseline level by 12 h (p >0.05) but individual responses differed considerably. Peak TAT post-infusion was associated inversely with peak HDL triglyceride concentration (p = 0.004). The coagulation responses to disc-infusion may be due to transfer of phosphatidylserine to cell surfaces during cholesterol efflux.