microRNAs and HDL life cycle

Abstract

miRNAs have emerged as important regulators of lipoprotein metabolism. Work over the past few years has demonstrated that miRNAs control the expression of most of the genes associated with high-density lipoprotein (HDL) metabolism, including the ATP transporters, ABCA1 and ABCG1, and the scavenger receptor SRB1. These findings strongly suggest that miRNAs regulate HDL biogenesis, cellular cholesterol efflux, and HDL cholesterol (HDL-C) uptake in the liver, thereby controlling all of the steps of reverse cholesterol transport. Recent work in animal models has demonstrated that manipulating miRNA levels including miR-33 can increase circulating HDL-C. Importantly, antagonizing miR-33 in vivo enhances the regression and reduces the progression of atherosclerosis. These findings support the idea of developing miRNA inhibitors for the treatment of dyslipidaemia and related cardiovascular disorders such as atherosclerosis. This review article focuses on how HDL metabolism is regulated by miRNAs and how antagonizing miRNA expression could be a potential therapy for treating cardiometabolic diseases.

1. Introduction

Cholesterol is a major component of the plasma membrane in mammalian cells. In addition to its structural requirement, cholesterol is important for other cell functions such as cell proliferation and bile acid and hormone biosynthesis.^1,2 Despite its pivotal role in controlling multiple physiological processes, abnormal levels of cholesterol can trigger a number of cardiometabolic diseases, including atherosclerosis and type-II diabetes.³ Because mammalian cells cannot degrade cholesterol, cholesterol removal is indispensable in order to prevent cholesterol accumulation in cells. Excess cholesterol must be removed and transported from the peripheral tissues to the liver for reutilization and excretion into feces in a physiological process traditionally known as reverse cholesterol transport (RCT).⁴ During RCT, plasma high-density lipoprotein (HDL) is thought to function as a sterol transporter that facilitates the movement of sterols from the peripheral cells to the liver. HDL integrates a heterogeneous class of lipoproteins with a density > 1.063 g/mL.⁵ HDL particles contain various apolipoproteins of which ApoA1 and ApoA2 are quantitatively the most abundant.^6,7

HDL formation occurs in the liver and intestine. The interaction between lipid poor ApoA1 with the ATP binding cassette (ABC) A1 mediates this first step in HDL formation (Figure 1).⁶ ABCA1 is a member of the ABC family of membrane transporters that promotes phospholipid and cholesterol transfer from cells to poorly lipidated ApoA1. Even though the mechanism by which ABCA1 regulates this process is not fully characterized, it is thought that poorly lipidated ApoA1 binds to ABCA1 in the membrane surface leading to increased stability and activity of the transporter in the plasma membrane. In response to ATP hydrolysis, ABCA1 promotes the trans-bilayer transport of phospholipids from the inner to outer leaflet of the plasma membrane. The uneven phospholipid packing in the plasma membrane bilayer leads to the formation of extravesiculated lipid domains. Lipid poor ApoA1 binds to this phospholipid and cholesterol-rich domain and promotes the spontaneous solubilization to form pre β-HDL particles.⁸

Figure 1.

miRNAs regulate reverse cholesterol transport (RCT). ABCA1 is regulated by a number of miRNAs that reduce the cholesterol efflux to lipid-poor ApoA1 that originates nascent HDL particles. miR-33 also inhibits the expression of bile acid transporters (ABCB11 and ATP8B1) in the liver, thereby regulating the last step of the RCT. ABCA1 is also regulated by miRNAs in the intestine and in the macrophages accumulated in atherosclerotic plaques. Free cholesterol in the nascent HDL is further esterified to cholesteryl esters by lecithin-cholesterol acyltransferase (LCAT) leading to the formation of mature HDL particles. HDL particles deliver cholesterol to the liver via SRB1 receptor, which is also regulated by several miRNAs including miR-185, miR-223, miR-96, and miR-185. This figure was performed using the Servier Medical Art illustration resources (http://www.servier.com).

The efflux of cholesterol and HDL formation through the ABCA1 pathway remains predominant. This finding was brought to light when the mutation in Tangier disease, a condition characterized by low plasma HDL levels, was attributed to mutations in the ABCA1 gene.^9–11 However, other proteins also play a role in HDL metabolism, including ABCG1 and the scavenger receptor class B type 1 (SRB1), which is involved in the maturation of HDL particles. Studies in vitro have shown that ABCA1 and ABCG1 synergistically mediate cholesterol efflux to HDL.¹² These models propose that the lipidated ApoA1 formed after the ApoA1/ABCA1 interaction serves as an acceptor for cholesterol that is effluxed from cells in an ABCG1-dependent process.¹² The importance of ABCG1 in regulating cholesterol efflux is well established; however, the mechanism of ABCG1 action is controversial. In contrast to ABCA1, it has been recently reported that ABCG1 is localized in endocytic vesicles where facilitate the distribution of specific intracellular sterols away from the endoplasmatic reticulum.¹³ Absence of ABCA1 and ABCG1 in mice results in massive accumulation of macrophage foam cells in various tissues such as in the spleen, heart, thymus, liver, and lung.^14,15 However, there are conflicting data regarding the role of both transporters during the progression of atherosclerosis. While transplantation of bone marrow from Abca1^−/− mice into Ldlr^−/− or ApoE^−/− recipients caused an increase in atherosclerosis,¹⁶ deficiency of ABCG1 in bone marrow cells resulted in either a modest increase, or decrease of atherosclerosis.^17,18 Despite the well-established role of ABCA1 in HDL formation, its effect on atherogenesis is less clear. Even though the increase in the incidence of atherosclerosis has been reported in people affected with Tangier disease, not all the subjects develop atherosclerosis. Moreover, mice lacking ABCA1 in the liver develop similar atherosclerotic lesions than Ldlr^−/− mice.¹⁹ Interestingly, absence of ABCA1 in the liver markedly diminish plasma HDL-C levels (less than 50%) but cause also a marked reduction in circulating VLDL and LDL.

In addition to ABCA1/ABCG1-mediated cholesterol efflux, excess intracellular cholesterol can also be eliminated by aqueous diffusion. This process consists of desorption of free cholesterol molecules from the plasma cell membrane into the surrounding aqueous phase. Collision of desorbed cholesterol molecules with HDL particles diffusing in the extracellular aqueous space leads to their rapid uptake into the lipoprotein acceptor.

SRB1 mediates bidirectional flux of cholesterol between cells and HDL, modulating changes in the composition and structure of HDL particles.^20,21 SRB1 facilitates the delivery of cholesterol to steroidogenic tissues and liver. Mouse studies have revealed the importance of this receptor in controlling lipoprotein metabolism and the progression of atherosclerosis.^22,23 Remarkably, absence of SRB1 results in a dramatic increase of atherosclerosis at a very young age. Srb1^−/−ApoE^−/− mice develop atherosclerotic plaques in the coronary arteries, a rare phenotype observed in mouse models of atherosclerosis.²²

HDL particles are constantly remodelled by a number of plasma enzymes and proteins. The lecithin-cholesterol acyltransferase (LCAT), a liver synthesized glycoprotein, mediates the transfer of fatty acids from phospholipids to free cholesterol present in the pre β-HDL to form cholesteryl ester.^24,25 Glomset proposed that esterification of cholesterol would drive the net efflux from cells because esterification would prevent the back-exchange of cholesterol from HDL to cells.^4,26 The cholesteryl ester generated by LCAT in HDL could be transferred to other lipoproteins, including very-low density lipoproteins (VLDL). This process is catalysed by the cholesteryl ester transfer protein (CETP), an enzyme that facilitates the transport of cholesteryl esters and triglycerides between the lipoproteins.^27,28 This neutral lipid transfer process results in a net gain of triglycerides and net loss of cholesteryl ester in HDL.²⁸ In addition to CETP, human plasma also contains a second lipid transfer protein, designated phospholipid transfer protein (PLTP), which mediates transfer of phospholipids from ApoB-containing lipoproteins to HDL. Moreover, two members of the triglyceride lipase family, the hepatic lipase (HL) and the endothelial lipase (EL), also participate in HDL metabolism. HL is a glycoprotein synthesized primarily by the liver and has both triglyceride lipase and phospholipase and hydrolyzes HDL, generating smaller subspecies, including pre β-HDL.²⁹ EL acts primarily as a phospholipase and hydrolyzes HDL phospholipids and is primarily located in vascular endothelial cells.³⁰

HDL particles deliver cholesterol and cholesterol esters to the liver and steroidogenic tissues through the scavenger receptor SRB1 (Figure 1). Within the liver, a portion of cholesterol is enzymatically converted in bile salt molecules. This process is unique to the liver because only hepatocytes express high levels of the enzyme cholesterol 7a-hydroxilase (CYP7A1), which initiates and rate-limits the multi-step conversion process.³¹ Cholesterol and bile acid molecules have different physical properties. While cholesterol is insoluble in water, bile acid salts are biological amphiphiles and highly soluble allowing the transport of cholesterol in the digestive system by forming micelles. Biliary lipids are secreted across the apical (canalicular) membrane of hepatocytes by three different transmembrane transporters: ABCB11 (aka BESP), ABCG5/ABCG8 (an obligate heterodimer that facilitates cholesterol efflux), and ABCB4 (aka MDR3; which pumps phospholipids).³² Another transporter, ATP8B1 maintains the asymmetry of phospholipids to promote the required lipid packing of the canalicular membrane for resistance to hydrophobic bile salts and canalicular membrane transport.^33,34 All these transporters play an essential role in the removal of cholesterol from the peripheral tissues towards the liver for excretion.^35,36 This process represents the last step of RCT, a protective mechanism against the development of atherosclerosis.

2. Transcriptional regulation of ABCA1 and ABCG1 expression

As mentioned earlier, ABCA1 regulates cellular cholesterol efflux and HDL biogenesis. Importantly, ABCA1 mRNA and protein half-life is very short (1–2 h), suggesting that de novo transcription and translation are critical for controlling its expression in response to environmental changes, such as cholesterol loading. ABCA1 and ABCG1 are oxysterol-regulated genes and both are transcriptionally regulated by the liver X receptors (LXRs).^12,37,38 These nuclear hormone receptors [(LXRα (NR1H3) and LXRβ (NR1H2)] form heterodimers with the retinoic-X-receptor (RXR)³⁹ leading to the transcriptional activation of both ABC transporters. Genetic deletion of LXRα and LXRβ results in a massive accumulation of cholesterol in peripheral tissues, suggesting the critical role of both transcription factors in controlling cholesterol removal from cells.^40,41 Moreover, absence of LXRα and LXRβ markedly increase the progression of atherosclerosis in mice.⁴⁰ In addition to the transcriptional regulation of ABCA1 and ABCG1 by LXRs, several groups have recently identified that the expression of both transporters are significantly post-transcriptionally regulated by microRNAs (miRNAs)^42–44 (Table 1).

Table 1.

miRNA regulation of HDL metabolism

MiRNA	Target genes	Cell type /tissue	Ref.
miR-33a/b	ABCA1, ABCG1, NPC1, CPT1A, SIRT6, AMPK, HADHB, CROT, CYP7A1, ABCB11, ATP8B1, NSF, SRC3, PCK1, G6PC, IRS2, RIP140, NFYC, SREBP1	Macrophages (primary mouse peritoneal macs, THP-1), endothelial cells (EAhy926), hepatic cells (HepG2, HEPA, Fu5aH, Hep3B), and mouse liver	45–52
miR-33*	NPC1, RIP140, SRC3, NFYC, IRS2, CROT	Hepatic cells (Huh-7), macrophages (THP-1)	53
miR-758	ABCA1	Macrophages (primary mouse peritoneal macs, J774, THP-1), hepatic cells (HepG2, Huh-7, HEPA), and neuroglyomal cells (H4)	54
miR-26	ABCA1, ARL7	Macrophages (Raw264.7 and THP-1)	55
miR-145	ABCA1	Hepatic cells (HepG2) and pancreatic β cells (MIN6 and primary mouse β cells)	56
miR-106b	ABCA1	Mouse primary hippocampal neurons (DIV14) and mouse neuroblastoma cells (Neuro2a)	57
miR-10b	ABCA1, ABCG1	Macrophages (primary mouse peritoneal macs, J774,THP-1)	58
miR-144	ABCA1	Macrophages (mouse peritoneal macrophages, J774, THP-1), hepatic cells (HepG2, Huh-7, Hepa), endothelial cells (EAhy926), mouse liver, human hepatic cells (Hep3B), and mouse primary hepatocytes	59–61
miR-27	ABCA1	Macrophages (THP-1)	62
miR-206	LXRα	Macrophages (THP-1)	63
miR-613	LXRα	Hepatic cells (HepG2)	64
miR-96	SRB1	Hepatic cells (HepG2)	65
miR-223	SRB1	Carried on HDL	65
miR-185	SRB1	Hepatic cells (HepG2)	65
miR-125a	SRB1	Hepatic cells (Hepa), steroidogenic cell lines (MLTC-1, granulosa cells)	66
miR-455	SRB1	Hepatic cells (Hepa), steroidogenic cell lines (MLTC-1, granulosa cells)	66

3. microRNAs and lipid metabolism

MicroRNAs (miRNAs) are small (18–25 nucleotides), evolutionarily conserved, non-coding RNAs that have an important function in gene regulation, acting predominantly at the post-transcriptional level.^67,68 Since miRNAs have been described in Caenorhabditis elegans, hundreds of miRNAs have been identified in animals, plants, and viruses.^67,68 They have been shown to participate in almost every cellular process investigated, including cholesterol homeostasis and lipoprotein metabolism.^{42–44,67,68} Mature miRNA products are generated from precursors called pri-miRNAs, which are composed of hundreds or thousands of nucleotides through sequential processing by the ribonuclease DROSHA.^69,70 This ultimately produces a nuclear hairpin precursor called the pre-miRNA, which is then exported to the cytoplasm where it is processed by DICER to produce the mature miRNA.^69,70 miRNAs typically control the expression of their target genes by imperfect base pairing to the 3′-untranslated region (3′UTR) of mRNAs. miRNAs are preferentially incorporated into the RISC complex where they associate with Argonaute proteins directing the binding of the RISC complex to the 3′UTR of their target mRNAs. This association produces mRNA repression either by transcript destabilization, translational inhibition, or both.^69,70 One miRNA often regulates multiple genes that are involved in a specific signalling cascade or cellular mechanism, thus making miRNAs potent biological regulators. In the past few years, several groups have demonstrated the important role of miRNAs including miR-33, miR-122, and miR-30c in controlling lipoprotein metabolism.^{45–47,71–73} In the following sections of this review article, we will discuss the most recent findings regarding the importance of miRNAs in regulating HDL metabolism.

4. miR-33 regulation of HDL metabolism and atherogenesis

The miR-33 family consists of two intronic miRNAs, miR-33a and miR-33b, which are encoded within the introns of the sterol regulatory element-binding proteins (SREBP) 2 and 1 genes, respectively.^45–48 The SREBPs are a family of membrane-bound transcription factors that regulate cellular lipid synthesis and clearance of pro-atherogenic lipoproteins.^74,75 SREBP2 regulates the expression of genes involved in the cholesterol biosynthetic pathway as well as the low-density lipoprotein receptor (LDLR). The function of SREBP1 is more complex because the same gene encodes two different SREBP isoforms (SREBP1c and SREBP1a). SREBP1c increases the expression of genes that regulate fatty acid synthesis and SREBP1a regulates genes that control cholesterol metabolism and fatty acid synthesis. SREBP1c is the predominant SREBP1 isoform in adult liver and it is activated in response to insulin. The fact that miR-33a and miR-33b are co-transcribed with their respective host genes suggests that miR-33a/b regulate related physiological processes controlled by SREBP2 and SREBP1.^45–48 Indeed, it has been demonstrated that miR-33a and miR-33b help boost cellular cholesterol and fatty acid levels during times of need. Under conditions that stimulate SREBP transcription, miR-33a/b are co-expressed with their host genes and reciprocally regulate genes involved in cellular cholesterol efflux/HDL biogenesis (ABCA1 and ABCG1) and fatty acid degradation (CPT1A, CROT, HADHB, AMPK1A).^45–48 These findings illustrate an elegant genetic regulatory mechanism by which miR-33a/b and their host genes cooperate to tightly regulate intracellular cholesterol and fatty acid levels. In addition to the genes mentioned above, Baldan and colleagues⁴⁹ have reported that miR-33 also regulates the expression of a number of bile acid transporters, including ABCB11 and ATP8B1 that control bile secretion. Moreover, a recent study also identified CYP7A1 as a miR-33 target gene.⁵⁰ Altogether, these findings support the hypothesis that miR-33 controls whole-body cholesterol homeostasis by affecting HDL biogenesis (via ABCA1), cellular cholesterol efflux from peripheral tissues (via ABCA1 and ABCG1) and bile acid synthesis (via CYP7A1) and secretion (via ATP8B1 and ABCB11).

Given that miR-33 levels markedly regulate ABCA1 expression, several groups assessed the efficacy of anti-miR-33 therapy for increasing circulating HDL-C. Three independent studies demonstrated that silencing of miR-33 in mice using modified anti-sense oligonucleotides, or viral delivery of hairpin inhibitors, increased hepatic ABCA1 expression and plasma HDL-C levels by 25–35%.^45–47 In addition to the elevated circulating HDL-C levels observed in mice treated with anti-miR-33 oligonucleotides, antagonism of miR-33 in vivo also enhanced RCT.^47,49 The effect of miR-33 inhibitors on plasma HDL-C levels was later confirmed in miR-33 deficient mice, which showed a significant increase in hepatic ABCA1 expression and a 25% increase in serum HDL-C compared with wild-type mice.⁵¹ Most importantly, anti-miR-33 therapy also resulted in increased plasma HDL-C levels in non-human primates.^76,77

A number of observational studies, including the Framingham Heart Study, have shown a strong inverse correlation of plasma HDL-C levels with coronary heart disease. To demonstrate whether anti-miR-33 therapy reduces the progression and enhances the regression of atherosclerosis in atherosclerosis-prone mouse models, several groups inhibited miR-33 using chemically modified antisense oligonucleotides.^78–80 The first study reported that a 4-week treatment with 2′-fluoro/methoxylethyl (2′F/MOE) anti-miR-33 oligonucleotides of Ldlr null mice fed previously a western diet (WD) for 14 weeks increased circulating HDL-C and enhanced the regression of atherosclerosis.⁷⁹ The results of the atherosclerosis progression studies, however, are somehow conflicting. While Baldan's group showed that prolonged anti-miR-33 therapy failed to raise plasma HDL-C and did not prevent the progression of atherosclerosis,⁷⁸ our group demonstrated that antagonism of miR-33 reduced atherogenesis despite the fact that HDL-C levels were not affected.⁸⁰ The different outcomes observed in the last two studies might be explained by different cholesterol content in the western diets (0.3 and 1.25%), oligonucleotide chemical modifications, and length of treatment. Indeed, while we demonstrated that the (2′F/MOE) anti-miR-33 oligonucleotides enhance ABCA1 expression in the artery wall, Baldan's study did not assess the efficacy of anti-miR-33 therapy in increasing ABCA1 expression in atherosclerotic plaques. Finally, the fact that miR-33 null mice have significant protection against the progression of atherosclerosis strongly suggests that inhibiting miR-33 in vivo might be useful for treating atherosclerotic vascular disease.⁸¹

5. miR-33 regulation of fatty acid and glucose metabolism

Besides the main role of miR-33 in regulating cholesterol metabolism, miR-33 also contributes to the regulation of additional metabolic pathways such as fatty acid metabolism and insulin signalling.^48,52 miR-33 regulates the expression of carnitine O-octanyl transferase (CROT), carnitine palmitoyltransferase 1A (CPT1A), and hydroxyacyl-coenzyme A dehydrogenase-3-ketoacyl-coenzyme A thiolase-enoyl-coenzyme A hydratase (trifunctional protein) β-subunit (HADHB), thereby controlling fatty acid β-oxidation.^48,52 CROT and CPT1A regulate the transport of fatty acids to the mitochondria for their degradation and HADHB is required for the last steps of the mitochondria β-oxidation pathway. Inhibition of miR-33 in human hepatic cells increases the degradation of fatty acids, suggesting that anti-miR-33 therapy may be useful for treating hepatic steatosis by increasing the degradation rate of fatty acids in the liver.^48,52

miR-33 also regulates the post-transcriptional expression of the AMPK-activated protein kinase (AMPK), sirtuin 6 (SIRT6), and insulin receptor substrate 2 (IRS2), thus controlling fatty acid and glucose metabolism.⁴⁸ This observation suggests that anti-miR-33 therapy could increase insulin sensitivity. In addition to the regulation of insulin signalling in human hepatic cells, it has recently been shown that miR-33 also modulates the expression of ABC transporters and insulin secretion in human and mouse pancreatic islets. Of note, inhibition of miR-33 in pancreatic islets increases ABCA1 expression and enhances insulin secretion while overexpression of miR-33 has the opposite effects.⁸² Altogether, these findings indicate that antagonism of miR-33 might increase plasma HDL-C and insulin sensitivity and reduce hepatic lipid accumulation and plasma triglyceride levels. However, it has been recently shown that miR-33-deficient mice develop obesity, hepatic steatosis, and insulin resistance.⁸³ Mechanistically, Horie et al.⁸³ found that miR-33 inhibits SREBP1 expression, thereby increasing fatty acid synthesis. Moreover, we have also reported that miR-33 regulates gluconeogenesis, suggesting that derepression of miR-33 might increase hepatic glucose production.⁸⁴ Together, these observations indicate that miR-33 regulates multiple metabolic processes and that further experiments are warranted to fully understand the molecular mechanism by which miR-33 controls lipid and glucose metabolism.

Although most of the studies have focused on the role of miR-33a-5p (guide strand), we have also recently discovered that the guide strand (miR-33-3p; aka miR-33*) accumulates in a number of tissues and targets similar genes as miR-33.⁵³ These findings suggest that both strands of the miR-33 locus may work together to control cellular lipid metabolism. Collectively, these studies have illuminated the key role of miR-33 in regulating lipid and glucose metabolism and how targeting miR-33 might be a useful therapy for treating cardiometabolic disorders.

6. Other miRNAs that regulate ABCA1 expression and HDL metabolism

miR-33 was the first miRNA described to regulate hepatic ABCA1 expression and plasma HDL-C levels in vivo. However, in the last years, it has become clear that ABCA1 expression is highly regulated at the post-transcriptional level by multiple miRNAs, including miR-758, miR-26, miR-106b, miR-27, miR-145, miR-10b, and miR-144.^{53–58,85,86} These miRNAs can regulate ABCA1 expression and function in a variety of cell types such as macrophages, neurons, pancreatic β-cells, enterocytes, and hepatocytes. The relative importance of these miRNAs in controlling ABCA1 expression will most likely be dictated by their relative abundance and the expression of other miRNA targets in specific cells or tissues. Additionally, the expression of these miRNAs and mRNA targets can also be regulated by physiological stimuli that alter miRNA expression levels. Importantly, several reports have demonstrated that cellular lipid metabolism influences miRNA expression, representing positive or negative feedback models that contribute to the complex regulation of ABCA1 expression. This is the case of miR-758, an intergenic miRNA that, similarly to miR-33, is downregulated after cholesterol loading in macrophages and in the liver of mice fed a high-fat diet.⁵⁴ Additionally, two reports have recently demonstrated the role of miR-144 in regulating cholesterol metabolism.^59,60 miR-144 is synthesized as a polycistronic transcript together with miR-451. In vertebrates, this conserved miRNA cluster plays an important role in eritropoiesis and cancer and was first described to bypass the classic Dicer processing step during miRNA biogenesis.^87–89 We identified miR-144 using an unbiased genome-wide screen of miRNAs modulated by LXR ligands in combination with bioinformatic tools for miRNA target predictions. We found that miR-144 directly targets ABCA1 and its overexpression markedly reduces ABCA1 protein levels in human and mouse macrophages and hepatic cell lines. Importantly, our in vivo results indicated that delivery of miR-144 mimics to mice inhibits hepatic ABCA1 expression levels and reduces circulating HDL-C.⁸⁶ More importantly, inhibition of endogenous miR-144 levels using anti-miR-144 conjugated particles in mice increases hepatic ABCA1 expression and raises plasma HDL levels. In a second report, de Aguiar Vallim et al.⁶⁰ identified miR-144 using a genome-wide screening aimed at identifying miRNAs regulated by farnesoid X receptor (FXR), a nuclear receptor that controls hepatic sterol and bile acid levels. Similar to our findings, gain- and loss-of-function experiments also showed that changes in hepatic miR-144 levels influence hepatic ABCA1 expression and circulating HDL-C. These results suggest a novel model by which miR-144 contributes to the FXR effect by inhibiting hepatic ABCA1 and promoting the redirection of hepatic cholesterol to biliary excretion.⁶⁰ The role of miR-144 in regulating plasma HDL levels has also recently been confirmed in a mouse model of atherosclerosis.⁶¹ Overexpression of miR-144 accelerates the progression of atherosclerosis by impairing RCT and promoting pro-inflammatory cytokine production in ApoE^−/− mice.

A recent study has shown that the intestinal microbiota can regulate RCT by modulating the expression of miR-10b expression. The authors found that protocatechuic acid (PCA), a metabolite produced by the gut microbiota from cyaniding-3 to O-β-glucoside (Cy-3-G) inhibits ABCA1 and ABCG1 expression. Importantly, PCA accelerates macrophage cholesterol efflux and Cy-3-G consumption promotes RCT and regresses atherosclerosis in ApoE^−/− mice. miR-145 also regulates ABCA1 expression in HepG2 cells and in murine pancreatic islets. Overexpression of miR-145 inhibits cholesterol efflux in HepG2 cells and causes cholesterol accumulation in pancreatic islets resulting in a marked decrease in glucose-stimulated insulin secretion. Finally, miR-26 and miR-27 inhibit ABCA1 expression and cholesterol efflux in mouse and human macrophage cell lines, respectively.^55,56,62

miRNAs also regulate the expression of LXR, thereby controlling the transcriptional activation of ABCA1. LXR is directly targeted by miR-1, miR-206, miR-613, and miR-155.^63,64,90 miR-1, miR-206, and miR-613 suppress lipogenesis by inhibiting LXRα and its target genes including SREBP1, acetyl-CoA carboxylase (ACC), and fatty acid synthase (FAS).^63,64,90 miR-155 also inhibits LXR expression and its absence confers protection against hepatic steatosis in mice.⁹⁰ LXR activation also regulates the expression of miRNAs that control the expression levels of some LXR-induced genes such as ABCA1 and the ADP-ribosylation factor-like-7 (ARL7).⁵⁵ Interestingly, miR-26 expression is inhibited in cells treated with LXR agonists, suggesting that the downregulation of miR-26 might cooperate with the LXR transcriptional activation to increase ABCA1 expression.⁵⁵

RXRα is also regulated by miRNAs including miR-128-2.⁹¹ This miRNA targets RXRα, thus inhibiting LXR-induced ABCA1 expression. Interestingly, miR-128-2 increases SREBP2 expression and decreases SREBP1, ABCA1, ABCG1, and RXRα expression. Overall, these data suggest the existence of a complex miRNA network operating under different physiological conditions in a number of different cell types and tissues to orchestrate post-transcriptional regulation of lipid metabolism (Figure 1).

7. Post-transcriptional regulation of ABCA1 by HuR

Although the importance of miRNAs in the regulation of ABCA1 and ABCG1 expression has been demonstrated in a number of cells and tissues, recent findings suggest that miRNA action might occur in conjunction with RNA-binding proteins (RBPs).^92–94 RBPs bind to AU-rich elements (AREs) in the 3′UTR of genes, thereby modulating their expression by increasing or decreasing translation and/or mRNA stability. Around 20 RBPs, including the well-known members of the ELAV family (HuR, HuB, HuC, and HuD), have been identified.⁹⁵ Interestingly, we have recently found that HuR binds to the 3′UTR of ABCA1 and increases its expression by enhancing protein translation.⁹⁶ A number of studies have shown that RBP compete or cooperate with miRNAs to control gene expression.^97,98 As such, the miRNA-mediated regulation of ABCA1 expression may be influenced by HuR binding to the ARE motifs in the 3′UTR (Figure 2). Further studies would be important to determine whether HuR might compete or cooperate with miRNAs in the regulation of ABCA1 expression.

Figure 2.

Interplay between miRNAs and HuR in the post-transcriptional regulation of ABCA1 expression. HuR regulates ABCA1 expression at post-transcriptional level. HuR might cooperate with miRNAs to enhance translational repression or activation and ARE-mediated mRNA decay (A) or compete by counteracting the miRNA binding to the 3′UTR (B).

8. miRNA regulation of SRB1 expression

SRB1 regulates the cholesterol transport from HDL to the liver for excretion.^35,36 Absence of SRB1 impairs RCT and causes massive atherosclerosis in mice.²² Indeed, SRB1 null mice develop coronary atherosclerosis, a rare phenotype observed in few mouse models of atherosclerosis.^22,99,100 SRB1 expression is regulated by different transcription factors including LXR, SREBP, liver receptor homologue 1 (LRH-1), and peroxisome proliferator-activated receptors (PPAR).¹⁰¹ Moreover, SRB1 expression is controlled at the post-transcriptional level by alternative splicing and its interaction with PDX domain containing 1 (PDZK1), a scaffolding protein that regulates SRB1 cellular localization and function.^102,103 In addition to this mechanism of regulation, SRB1 expression is also controlled by miRNAs. In this regard, a number of recent studies have uncovered several miRNAs, including miR-455, miR-125a, miR-185, miR-96, and miR-223 that bind directly to the 3′UTR of SRB1 and suppress its expression.^65,66 Overexpression of miR-455, miR-125a, miR-185, miR-96, and miR-223 reduces SRB1 protein levels and HDL-C uptake. Conversely, antagonism of these miRNAs enhances SRB1 expression and increases HDL-C uptake. Interestingly, the levels of miR-96 and miR-185 inversely correlate with the increased expression of SRB1 in the livers of ApoE knockout mice fed a high-fat diet.⁶⁵ Overall, these observations suggest that SRB1 expression is regulated at the post-transcriptional level by alternative splicing, protein localization, and miRNAs.

9. HDL transports endogenous miRNAs and delivers them to recipient cells

In addition to the classic view of HDL as sterol transporter that facilitates the movement if sterols from peripheral cells to the liver for its reutilization or excretion, it is now recognized that HDL function is much more complex in terms of differential lipids and proteins that it transports. Surprisingly, it has been recently reported that HDL can transport miRNAs and deliver them to the receiving cells influencing their gene expression.¹⁰⁴ Among of them, miR-223 is one of the most abundant miRNAs (10000 copies/μg of HDL). Notably, HDL-derived miR-223 is transferred to human hepatic cells (Huh7) via SRB1 receptor leading to a significant reduction of miR-223 target genes expression.

Similarly, HDL-derived miR-223 is also transferred to endothelial cells and inhibits the intracellular adhesion molecule 1 (ICAM-1), thereby reducing monocyte adhesion and inflammation¹⁰⁵ (Figure 3). This finding might explain in part the well-known anti-inflammatory effects of HDL. However, additional studies are necessary to address some important questions including the mechanism of miRNA loading in HDL particles and the receptors that mediate the transfer of HDL-derived miRNAs and the recipient cells within the atherosclerotic plaques (endothelial cells, macrophages, and smooth muscle cells).

Figure 3.

HDL-derived miR-223 is transferred to endothelial cells and reduces inflammation. HDL transports miRNAs including miR-92a, miR-126, and miR-223. miR-223 can be transferred to endothelial cells (ECs). The receptor that facilitates the transfer in ECs is unknown but previous studies demonstrate that SRB1 regulates the transport in human hepatic cells (Huh7). miR-223 inhibit ICAM-1 expression in ECs thereby reducing monocyte adhesion and inflammation.

10. Conclusions

miRNAs have emerged as critical regulators of almost all biological processes including lipoprotein metabolism. Work over the last years has demonstrated that miRNAs play an important role in regulating HDL metabolism (Figure 1). A number of genes associated with HDL biogenesis, cellular cholesterol efflux, and biliary secretion are post-trancriptionally regulated by miRNAs. Most of the studies have identified miRNAs that regulate ABCA1 and SRB1 expression (Table 1); however, it is not known whether or not the expression of other key players that control HDL metabolism, such as CETP and LCAT, are modulated by miRNAs. This possibility is unlikely as both genes have a very short 3′UTR (less than 200 nt) and no conserved miRNA-binding sites within their 3′UTR across species.

Antagonizing the expression of some miRNAs, including miR-33, has shown to markedly increase circulating HDL-C in mice and non-human primates. Moreover, anti-miR-33 therapy reduces the progression and enhances the regression of atherosclerosis in mice. These findings suggest that antagonizing a set of miRNAs in the liver to increase ABCA1 and SRB1 expression might enhance RCT. Most of the therapies aimed to increase plasma HDL-C levels, such as CETP inhibitors, fail to protect against coronary artery disease. Moreover, recent results from Mendelian randomization studies also fail to demonstrate an association between circulating HDL-C levels and cardiovascular risk.¹⁰⁶ Even thought, these studies argue about the benefit of HDL-C to protect against myocardial infarction, it is clear that HDL-C levels are not necessarily reflective of the broad antiatherogenic properties of HDL particles, including RCT. Indeed, Rader and colleagues¹⁰⁷ have shown that the cholesterol efflux capacity from macrophages, a metric of HDL function, has a strong inverse association with both carotid intima-media thickness and the likelihood of angiographic coronary artery disease, independently of HDL-C. It may be possible that anti-miR-33 therapies, which influence HDL metabolism by controlling HDL biogenesis, cellular cholesterol efflux, and bile excretion might be useful for treating dyslipidaemias and cardiovascular related disorders.

Conflict of interest: none declared.

Funding

C.F.-H. laboratory is supported by the National Institutes of Health (R01HL107953 and R01HL106063) and The Leducq Foundation.