Disorders of high‐density lipoprotein biogenesis

Abstract

The characterization of the atheroprotective role of high‐density lipoproteins (HDL) made in the past two decades has rekindled interest in modulating HDL for therapeutic purposes. Rare deficiencies of HDL have allowed the identification of specific proteins acting as structural moieties, enzymes, lipid transfer proteins, cellular lipid transporters, and ligands for cellular receptors; these, in turn, represent potential drug targets. The study of several of these HDL deficiency states has shown the importance of cellular cholesterol transport in HDL metabolism. Based on a better understanding of the physiology of HDL formation, many cases of severe HDL deficiency in man can now be explained at the cellular level. Disorders of HDL biogenesis in man—apolipoprotein (apo) AI defects, mutations at the adenosine triphosphate (ATP) binding cassette AI (ABCA1), defects of specific lipases that modulate HDL, such as sphingomyelinase, the HDL deficiency seen in Niemann‐Pick disease type C (NPC)—can be linked to abnormal formation of nascent HDL particles via the ABCA1 transporter. Deficiency in lecithin:cholesterol acyl transferase impedes the formation of mature HDL particles, once nascent HDL particles are formed. As a consequence, modulating cellular cholesterol efflux and apo AI secretion—and thereby nascent HDL particles—may be an appealing strategy to raise HDL for therapeutic purposes.

• Genetics
• high‐density lipoproteins
• lipoproteins
• prevention

Introduction

Genetic lipoprotein disorders are frequently seen in patients with premature coronary artery disease (CAD) with a low high‐density lipoprotein‐cholesterol (HDL‐C) being one of the most common lipoprotein disorders 1. Isolated familial HDL deficiency is seen in approximately 4% of patients with premature CAD 2, 3. It should be pointed out that a low HDL‐C level is frequently encountered in the clustering of abdominal obesity, dyslipidemia with elevated triglyceride‐rich lipoproteins, high blood pressure, and insulin resistance that constitute the cardiometabolic syndrome. Most cases of decreased HDL‐C levels are a consequence of multiple metabolic disorders 4. Unraveling the causes of rare HDL deficiency has allowed a greater understanding of processing enzymes that modulate HDL lipids and has led to fundamental discoveries in the molecular physiology of cholesterol transport. These findings, in turn, provide potential therapeutic targets to prevent CAD and its recurrence.

The formation of HDL particles, referred to as HDL biogenesis, comprises many steps taking place in a carefully orchestrated series of events taking in both intracellular and interstitial fluid or plasma compartments.

Background

The epidemiology of HDL and risk of coronary artery disease (CAD) has been reviewed elsewhere 5, 6. It is estimated that approximately 40% of patients with premature CAD have a low HDL‐C 2, 3 and this represents the most common lipoprotein disorder in patients with CAD.

HDL metabolism has recently been reviewed in detail 7, 8. The metabolism of HDL is complex and incompletely understood. This complexity arises because HDL particles acquire their components from several sources while these components also are metabolized at different sites. High‐density lipoproteins have many roles. The best studied, and therefore considered the most important, is the reverse cholesterol transport. There is remarkably a paucity of evidence that reverse cholesterol transport constitutes a major role of HDL. Other roles of HDL include anti‐inflammatory, antioxidant, antithrombotic, vasomotor modulation, and possibly antiapoptotic 9, 10.

Apolipoprotein AI, the main protein of HDL, is synthesized in the intestine and the liver. Based on tissue‐specific knockout experiments of the adenosine triphosphate (ATP) binding cassette AI (ABCA1) transporter, it has been determined that approximately 80% of HDL‐derived cholesterol mass originates from the liver 11 and 20% from the intestine 12. Lipid‐free apo AI acquires phospholipids from cell membranes and from redundant phospholipids shed during hydrolysis of triglyceride‐rich lipoproteins. Lipid‐free apo AI binds to ABCA1 and promotes its phosphorylation via the cyclic adenosine monophosphate (cAMP)/protein kinase A pathway 13, 14, which increases the net efflux of phospholipids and cholesterol onto apo AI to form a nascent HDL particle 15. This particle, containing apo AI and phospholipids (and little cholesterol) resembles a flattened disk in which the phospholipids form a bilayer surrounded by two molecules of apo AI arranged in a circular fashion at the periphery of the disk 16, 17. We have found that the majority of ABCA1 exists as a tetramer in human living cells, supporting the concept that the homotetrameric ABCA1 complex constitutes the minimum functional unit for the formation of nascent HDL particles 18. This observation, that ABCA1 exists physiologically in an oligomeric form, has been confirmed 19, 20. These nascent HDL particles will mediate further cellular cholesterol efflux (Figure 1). Currently, standard laboratory tests do not measure these HDL precursors because they contain little or no cholesterol. Upon reaching a cell membrane, the nascent HDL particles will capture membrane‐associated cholesterol and promote the efflux of free cholesterol onto other HDL particles. Conceptually, the formation of HDL particles appears to involve two steps, the first step ABCA1‐dependent and the second probably does not require ABCA1 21. The identification of a novel two‐binding site model for ABCA1‐mediated nascent HDL genesis was supported by the identification of an ABCA1‐dependent phospholipid‐rich plasma membrane high‐capacity binding site for the lipidation of apo AI 22, 23. We have also provided evidence that there is speciation of nascent HDL into pre‐β and α‐HDL linked to specific cell lines, and this occurs by both ABCA1‐dependent and independent pathways 24. The efflux of cellular cholesterol from peripheral cells, such as macrophages, does not contribute importantly to overall HDL‐C mass but may have an important effect on export of cholesterol from atheromas. Macrophages can efflux cholesterol onto apo AI and apo E, onto nascent HDL particles via the ABCA1 transporter, or onto spherical HDL particles via the ABCG1 transporter. The ABCG1 transporter does not promote cellular cholesterol efflux to lipid‐free or lipid‐poor apo AI but to mature HDL particles 25. The plasma enzyme lecithin cholesterol acyl transferase (LCAT), an enzyme activated by apo AI, then esterifies the free cholesterol. LCAT transfers an acyl chain (a fatty acid) from the sn2 position of a phospholipid to the 3'‐OH residue of cholesterol, resulting in the formation of a cholesteryl ester.

Figure 1 Schematic diagram of high‐density lipoprotein (HDL) biogenesis in man. Please refer to Table I for abbreviations. LDL = low‐density lipoproteins; LDL‐R = LDL receptor; ACAT = acyl‐coenzyme A:cholesterol acyl transferase; Smase = sphingomyelinase; sPLA2 = secretory phospholipase A2; Lp‐PLA2 = lipoprotein associated PLA2; PON = paraoxonases 1, 2, and 3. HMG CoA Red: hydroxymethylglutaryl coenzyme A reductase and sER: smooth endoplasmic reticulum.

Table I. Lipoprotein‐processing enzymes, receptors, modulating proteins in HDL metabolism.

Abbreviation	Name	Role	Chromosome	Human disease
ABCA1	ATP binding cassette AI	Cellular phospholipid efflux	9q31	Tangier disease
Apo AI	Apolipoprotein AI	Major structural protein of HDL	11q23	Apo AI deficiency
CETP	Cholesteryl ester transfer protein	Lipid exchange in plasma	16q21	Elevated HDL cholesterol
EL	Endothelial lipase	PL hydrolysis	18q21.1
HL	Hepatic lipase	Tg hydrolysis	15q21	Remnant accumulation
LCAT	Lecithin:cholesterol acyl transferase	Cholesterol esterification	16q22.1	LCAT deficiency, low HDL
LPL	Lipoprotein lipase	Tg hydrolysis	8p22	Hyperchylomicronemia, low HDL
NPC1	Niemann‐Pick C gene product	Cellular cholesterol transport	18q11–12	Niemann‐Pick type C
PLTP	Phospholipid transfer protein	Lipid exchange in plasma	20q12
SMPD1	Sphingomyelinase phosphodiesterase	Sphingomyelin hydrolysis	11p15.4	Niemann‐Pick types A and B
SR‐B1	Scavenger receptor B1	HDL CE uptake	12

PL = phospholipid; Tg = triglycerides; CE = cholesteryl esters.

The triglyceride‐enriched HDL_2b is hydrolyzed by endothelial lipase and hepatic lipase, while phospholipids are altered by the secretion of phospholipase A2 (sPLA2) and sphingomyelinase (SMase). The roles of lipoprotein‐associated phospholipase A2 (Lp‐PLA2), present in low‐density lipoproteins (LDL) and in a smaller portion in HDL, and of paraoxonases 1, 2, and 3 remain presently unclear in the normal metabolism of HDL 8. In a process called selective uptake of cholesterol, HDL also provides cholesterol to steroid hormone‐producing tissues and liver through the scavenger receptor SR‐B1 26.

Because of their hydrophobicity, cholesteryl esters move to the core of the lipoprotein, and the HDL particle now assumes a spherical configuration (a particle denoted HDL₃). With further cholesterol esterification, the HDL particle increases in size to become the more buoyant HDL₂. Cholesterol within HDL particles can exchange with triglyceride‐rich lipoproteins via cholesteryl ester transfer protein (CETP), which mediates an equimolar exchange of cholesterol from HDL to triglyceride‐rich lipoprotein and triglyceride movement from triglyceride‐rich lipoprotein onto HDL. Inhibition of CETP increases HDL‐C in the blood and represents a potential therapeutic target for cardiovascular disease prevention. In a disappointing series of clinical trials, a selective CETP inhibitor, Torcetrapib, showed increased mortality over placebo, despite a marked increase in HDL‐C and improvement in the cholesterol/HDL‐C ratio 27. Phospholipid transfer protein (PLTP) mediates the transfer of phospholipids between triglyceride‐rich lipoprotein and HDL particles. Phospholipid transfer protein plays an essential role in the remodeling of model nascent HDL particles with the participation of apo B‐containing lipoproteins. Triglyceride‐enriched HDL is denoted HDL_2b. Hepatic lipase can hydrolyze triglycerides, and endothelial lipase can hydrolyze phospholipids within these particles, converting them back to HDL₃ particles.

One mechanism of reverse cholesterol transport includes the uptake of cellular cholesterol from extrahepatic tissues, such as lipid‐laden macrophages, and its esterification by LCAT, transport by large HDL particles, and exchange for one triglyceride molecule by CETP. Originally on an HDL particle, the cholesterol molecule can now be taken up by hepatic receptors on a triglyceride‐rich lipoprotein or LDL particle. HDL particles, therefore, act as shuttles between tissue cholesterol, triglyceride‐rich lipoprotein, and the liver.

Most HDL originates in the liver 11 and intestines 12. Reverse cholesterol transport by HDL constitutes a small but potentially important portion of the plasma HDL mass. Indeed, selective inactivation of macrophage ABCA1 does not change HDL‐C levels in mice but there is an increase in atherosclerosis 28. The catabolism of HDL particles has engendered debate among lipoprotein researchers. The protein component of HDL particles is exchangeable with lipoproteins of other classes. The kidneys appear to be a route of elimination of apo AI and other HDL apolipoproteins. The lipid components of HDL particles also follow a different metabolic route.

Disorders of HDL biogenesis represent the most frequent cause of severe genetic HDL deficiency in man (Figure 2). Most of these cases are due to mutations at the ABCA1 transporter gene. We have previously reported that approximately 20% of French‐Canadian patients with HDL deficiency have a mutation within the ABCA1 gene, a prevalence higher than reported in German or American cohorts 29, 30. Using a candidate gene sequencing strategy, we have identified mutations in the sphingomyelinase phosphodiesterase‐1 (SMPD1) gene causing Niemann‐Pick disease types A and B and mutations within apo AI, the major structural apolipoprotein of HDL. We were thus able to identify a molecular cause in one quarter (13/54) of severe HDL deficient patients 29–31. In addition, family studies have shown several chromosomal localizations for the low HDL trait, using both a quantitative trait loci (QTL) and a traditional affected subject approach. There is evidence for a major gene effect on chromosomes 4 and 16 and on several loci as well (reviewed in 32).

Figure 2 Monogenic causes of low HDL‐C (high‐density lipoprotein cholesterol) in humans. Genetic defects causing HDL deficiency in man are underlined. Proteins, processing enzymes, or transporters causing disorders of HDL biogenesis are framed. Lipases contributing to HDL metabolism are in italics. Please refer to Table I for abbreviations. sPLA2 = secretory phospholipase A2; Smase = sphingomyelinase. Note that LCAT has phospholipase A2 activity.

ABCA1

The ABCA1 gene codes for a phospholipid (and possibly cholesterol) cellular transporter. The role of ABCA1 is to transport cellular lipids (phospholipids and possibly cholesterol) onto amphipathic helical proteins, characteristic of HDL‐associated apolipoproteins. While the preferred substrate for ABCA1 appears to be apo AI, cellular cholesterol efflux onto apo AII, apo E and the apo C's has been shown 33. A defect in ABCA1 results in Tangier disease 34–36 and familial HDL deficiency in the homozygous and heterozygous form, respectively 37. Multiple mutations of ABCA1 have been identified in Tangier disease and in severe HDL deficiency. At least three studies have examined the prevalence of ABCA1 mutations in patients with low HDL‐C. Cohen et al.29 examined patients from the US and Canada and identified an ABCA1 mutation or gene variant in 10%–16% of patients with low HDL‐C; a similar finding was reported in Germany with approximately 10% of low HDL‐C being associated with ABCA1 mutations 30. Alrasadi et al. reported a 20% prevalence of ABCA1 mutation in subjects of French‐Canadian descent 31, possibly reflecting a founder effect. In a study of patients with heterozygous ABCA1 mutations, the risk of coronary artery disease was increased 3.5‐fold 38. The risk of vascular disease in Tangier disease remains a matter of debate. Despite a very low plasma HDL‐C level, not all patients with Tangier disease develop severe, premature atherosclerosis. This may be due, at least in part, to the very low plasma levels of LDL‐C seen in Tangier disease. Some patients develop a severe demyelinating neuropathy that contributes to increased morbidity and mortality in Tangier disease. Apo AI—and other apolipoproteins that contain amphipathic α‐helical motifs—bind to ABCA1 and mediate the transport of cellular lipids onto acceptor proteins via signaling molecules (protein kinase A) 13. In competitive binding studies, we have shown that lipid‐free apo AI binds with greater affinity to ABCA1 than reconstituted HDL particles; HDL binds with the least affinity 15. This suggests that once lipidated, the topology of apo AI changes in such a way as to decrease ligand‐receptor interaction. We postulate that lipidated apo AI leaves the ABCA1 binding site, making way for another lipid‐free apo AI moiety to initiate another cycle of lipidation. In a series of experiments, we have shown that apo AI activates ABCA1 phosphorylation through the cAMP/protein kinase A‐dependent pathway 13. We 23, and others 22 have shown that the apo AI–ABCA1 interaction creates a novel, phospholipid‐rich binding site on the plasma membrane. This would explain, at least in part, the two‐step model of HDL lipidation observed in previous studies. Data on the tertiary structure of ABCA1 suggest that the functional unit of ABCA1 is oligomeric in nature and that a tetrameric structure is the minimal functional unit of the ABCA1 efflux complex 18. This structural arrangement of ABCA1 has recently been confirmed by Trompier et al. 19. Other ABC transporters such as ABCG2 20 have been shown to form physiologically functioning oligomeric structures. This would explain the clinical observation that some heterozygous mutations of ABCA1 may behave in a dominant negative fashion. If indeed one member of the tetrameric ABCA1 efflux complex is abnormal, it is likely that the entire complex might be dysfunctional. Interestingly, some patients with low HDL‐C and a cellular cholesterol efflux defect do not have mutations within ABCA1. In rare cases, a defect in NPC1, the defective gene in Niemann‐Pick disease type C, causes a cellular cholesterol efflux defect not due to structural modifications in ABCA1 (39; see below). We have also identified non‐ABCA1 efflux defects in two kindred. The molecular basis for this disorder is yet unknown. Thus, mutations at the ABCA1 transporter lead to the decreased formation of nascent HDL particles by impaired cellular phospholipid and cholesterol efflux. The prototypical disorder of HDL deficiency, Tangier disease, is associated with rapidly catabolized apo AI (presumably in the kidneys) as these poorly lipidated apo AI particles are rapidly cleared from plasma.

Apo AI

Apo AI constitutes the major apolipoprotein in HDL, accounting for approximately 70% of the protein mass within HDL. The role of apo AI is structural—the recent characterization of the apo AI crystal structure 17 and molecular modeling of the lipidated apo AI peptide 16 have shed considerable light on the structure‐function of apo AI domains critical for nascent HDL particle assembly. Apo AI also activates LCAT and binds ABCA1 15. Mutations within the apo AI gene can lead to altered apo AI and HDL‐C levels, but not all apo AI mutations lead to decreased apo AI. Yet other mutations lead to amyloidosis 40. At least 47 mutations within the apo AI gene have been identified 40, 41. Most of the 18 reported mutations that affect apo AI and HDL‐C levels appear to cluster between residues ∼100 and 200, i.e. within amphipathic helices 5, 6, and 7, critical for HDL assembly. Two of these mutations, apo AI_Milano (Apo AI_R173C) and apo AI_Paris (Apo AI_R151C) are associated with a low HDL‐C, but, paradoxically, with no increase in the incidence of heart disease 42. In a proof‐of‐concept clinical study, AI_Milano proteoliposomes were injected intravenously weekly in 47 patients with acute coronary syndromes for a period of 5 weeks. At the end of the study, a small, but significant (−4.2%) regression of atheroma volume was documented by intravascular ultrasound 43. Interestingly, the infusion of wild‐type apo AI in proteoliposomes (with a 2:198 apo AI:phosphatidyl choline molar ratio) did not show changes in atheroma volume, suggesting that apo AI_Milano might have antiatherogenic properties 44. Mutations within the apo AI gene clustering in the amino terminus are associated with amyloidosis, a disease that has multisystemic effects but is not associated with atherosclerosis. Other mutations, scattered throughout the apo AI gene, are not associated with low HDL‐C or amyloidosis; their physiological significance is unknown. Most cases of apo AI mutations have been described in single families and account for a minority of HDL deficiency cases seen in humans. The structure‐function relationship between apo AI mutations and low HDL‐C is likely to involve multiple mechanisms, including decreased binding to ABCA1, decreased ability to form nascent HDL particles and decreased LCAT activation.

SMPD1

Sphingomyelinase deficiency as the results of mutations in the gene coding for SMPD1 causes Niemann‐Pick disease type 1 (subtypes A and B—NPD A and B) 45. While the clinical manifestations of the Niemann‐Pick group of diseases encompass pulmonary infiltrates, hepatosplenomegaly, the formation of lipid‐laden histiocytes in the bone marrow, and mental retardation, a low HDL‐C is a frequent biochemical feature 46. Studies on the composition of HDL phospholipids in HDL from patients with NPD B revealed an enrichment of sphingomyelin in HDL particles, despite a normal choline phospholipid mass 47. Sphingomyelinase exists in a lysosomal form (acid sphingomyelinase) and a secretory form, both from the same gene, but the posttranscriptional mechanisms leading to processing to the lysosomal compartment of the secretory pathway remain poorly understood. The enrichment in sphingomyelinase prevents activation of the LCAT reaction and impaired formation of cholesteryl esters. This mechanism is postulated to cause the low HDL‐C seen in NPD A and B subjects 47.

NPC‐1

Niemann‐Pick disease type C is caused by mutations at the Niemann‐Pick type C gene (NPC‐1). NPC disease is characterized by developmental abnormalities, mental retardation, and lymphoid infiltrates of lipid‐laden cells. The product of the NCP1 gene, NPC1 protein, appears to play a pivotal role in the transport of cholesterol from the late endosomal pathway to the endoplasmic reticulum and other subcellular organelles. Choi et al.39 have reported that in NPC disease, impaired cholesterol transport from the late endoplasmic pathway impairs the regulation of ABCA1 and cellular cholesterol efflux. In effect, the cholesterol accumulation in the late endosomal pathway observed in cells from NPC patients prevents its conversion to oxysterols (presumably predominantly in the mitochondrion), the endogenous ligand for the liver‐specific receptor LxR. Since the physiological regulation of ABCA1 is mediated via the LxR pathway, cells from patients with NPC fail to upregulate ABCA1 in response to cholesterol loading. This cellular phenotype can be rescued by LxR agonists, raising the possibility that such agents can be used for therapeutic purposes in NPC patients.

LCAT

Mutations in this gene cause Norum disease and fish‐eye disease, so called because of the pathognomonic corneal infiltrations seen with LCAT deficiency. The pathogenesis of LCAT deficiency stems from an inability to form cholesteryl esters in plasma and thus an inability to form mature HDL particles. Clinically, patients with LCAT deficiency may present with normochromic anemia, proteinuria, corneal opacifications, and increased foam cell formation in bone marrow and the kidneys 48, 49. While patients with LCAT deficiency have a marked decrease in HDL‐C, there appears to be no increase in premature coronary artery disease. LCAT deficiency is a very rare cause of HDL deficiency.

HL

In reports of subjects with hepatic lipase (HL) deficiency, the biochemical phenotype is characterized by triglyceride enrichment of LDL and HDL. The LDL‐C is normal or mildly elevated, and HDL‐C is elevated. Heterozygote carriers do not display significant lipoprotein abnormalities; postprandial clearance of triglycerides is delayed. Premature atherosclerosis has been reported in affected individuals 50.

LPL and Apo CII

Lipoprotein lipase (LPL) and its activator Apo CII deficiency cause severe hypertriglyceridemia and hyperchylomicronemia with a secondary HDL deficiency. The inability to hydrolyze triglycerides within chylomicrons and Very Low‐Density Lipoproteins (VLDL) leads to accumulation in plasma of these particles. Heterozygous mutations in the LPL gene are associated with elevations in plasma triglycerides, a reduction in HDL‐C, and an increased risk of CAD 51–53.

CETP

Mutations in the CETP gene are associated with increases in HDL‐C 54. CETP mediated the equimolar transfer of cholesteryl esters from HDL particles to triglyceride‐rich VLDL and intermediate-density lipoproteins (IDL) particles in exchange for triglycerides 55. A deficiency of CETP increases the cholesteryl ester content of HDL and the formation of large, buoyant HDL₂ particles. It has recently been found that these particles maintain their functional abilities, in terms of promoting cellular cholesterol efflux 56. It remains unclear whether CETP inhibition represents a valid therapeutic target, in light of the recently reported clinical trials of Torcetrapib, a potent CETP antagonist that showed potential toxicity in spite of beneficial improvements in lipoprotein fractions 27, 57.

In the past few years, major discoveries, especially the identification of ABC transporters in lipoprotein metabolism, has shed light on the complex series of events that lead to the biogenesis of HDL. Not surprisingly, much of this discovery process is the result of the careful study of severe lipoprotein disorders in man. As our understanding of basic cellular and interstitial processes increases, so will the identification of novel proteins involved in the cellular phospholipid and cholesterol efflux machinery and the finely tuned enzymes that modulate HDL in plasma. Indeed, comparative genomics 58 and HDL proteomics 59, 60 have allowed the identification of many novel proteins involved in HDL physiology. In turn, a better understanding of HDL components and metabolism might lead to therapeutic modulation of HDL for cardiovascular disease prevention.

Key messages

• High‐density lipoproteins (HDL) are formed principally in the liver and intestine.

• The ABCA1 transporter is necessary and essential for HDL biogenesis.

• Novel HDL metabolic pathways may represent potential therapeutic targets.