Signal Transduction by HDL: Agonists, Receptors, and Signaling Cascades
Numerous epidemiologic studies revealed that high-density lipoprotein (HDL) is an important risk factor for coronary heart disease. There are several well-documented HDL functions such as reversed cholesterol transport, inhibition of inflammation, or inhibition of platelet activation that may account for the atheroprotective effects of this lipoprotein. Mechanistically, these functions are carried out by a direct interaction of HDL particle or its components with receptors localized on the cell surface followed by generation of intracellular signals. Several HDL-associated receptor ligands such as apolipoprotein A-I (apoA-I) or sphingosine-1-phosphate (S1P) have been identified in addition to HDL holoparticles, which interact with surface receptors such as ATP-binding cassette transporter A1 (ABCA1); S1P receptor types 1, 2, and 3 (S1P1, S1P2, and S1P3); or scavenger receptor type I (SR-BI) and activate intracellular signaling cascades encompassing kinases, phospholipases, trimeric and small G-proteins, and cytoskeletal proteins such as actin or junctional protein such as connexin43. In addition, depletion of plasma cell cholesterol mediated by ABCA1, ATP-binding cassette transporter G1 (ABCG1), or SR-BI was demonstrated to indirectly inhibit signaling over proinflammatory or proliferation-stimulating receptors such as Toll-like or growth factor receptors. The present review summarizes the current knowledge regarding the HDL-induced signal transduction and its relevance to athero- and cardioprotective effects as well as other physiological effects exerted by HDL.
KeywordsHigh-density lipoprotein (HDL) Apolipoprotein A-I (apoA-I) ATP-binding cassette transporter A1 (ABCA1) Scavenger receptor type I (SR-BI) Sphingosine-1-phosphate (S1P) Signal transduction
ATP-binding cassette transporter A1
ATP-binding cassette transporter G1
AMP-activated protein kinase
Calcium/calmodulin-dependent protein kinase
Cyclic adenosine monophosphate
Cell division control protein 42
Coronary heart disease
Chronic kidney disease
Endothelial nitric oxide synthase
Glucose transporter type 4
Granulocyte macrophage colony-stimulating factor
Hematopoietic stem progenitor cells
Janus kinase 2
c-Jun N-terminal kinase
Mitogen-activated protein kinase
Monocyte chemotactic protein 1
Madin-Darby canine kidney
PC-specific phospholipase C
PC-specific phospholipase D
PDZ domain-containing adaptor protein
Phosphatidylinositol-specific phospholipase C
Protein kinase A
Protein kinase C
Serum amyloid A
Scavenger receptor type I
Signal transducer and activator of transcription 3
Transforming growth factor-β
Tumor necrosis factor α
F1F0 ATP synthase β-subunit
Innumerable epidemiological studies document the inverse relationship between plasma levels of high-density lipoprotein (HDL) cholesterol and cardiovascular risk. The first hypothesis aiming to mechanistically explain this relationship has been formulated by John Glomset (1968), who proposed that HDL particles take up excess cholesterol in peripheral tissues including atherosclerotically changed arterial walls and transport it to the liver for excretion with bile. The first step in this process termed reverse cholesterol transport encompassed direct interaction between HDL and cell membrane with the subsequent efflux of cholesterol to lipoprotein particles. Early concepts envisioned cholesterol efflux as a purely physicochemical phenomenon with the unidirectional flux of free cholesterol perpetuated by a steady concentration gradient maintained by the activity of lecithin:cholesterol acyltransferase (LCAT) in HDL particles. John F. Oram was the first to note that HDL binding to specific binding sites precedes cholesterol efflux and is required for the effective mobilization of free cholesterol from internal pools and its transport to the plasma membrane, where it is available for further transfer to lipoprotein particles (Graham and Oram 1987; Slotte et al. 1987). The seminal discovery of Oram put HDL function in a fully new perspective: ever since then HDL has been perceived not only as a cholesterol transporter but also as hormone-like particle exerting its activities through binding to specific receptors.
Signal transduction occurs when an extracellular signaling molecule activates a surface receptor, which in turn alters intracellular molecules creating a response, which ultimately modifies cellular functions. The recognition of specific HDL binding sites raised immediately a question, whether HDL-cell interaction is linked to signaling events relevant to intracellular cholesterol transport and/or cholesterol efflux. It was Oram again who demonstrated that the exposure of fibroblasts to HDL causes activation of protein kinase C (PKC), which in turn contributes to translocation of free cholesterol between intracellular membranes and plasmalemma (Mendez et al. 1991). This finding opened a new field of research focusing on dissection of intracellular signaling pathways that are induced by HDL or its constituents and mediate cellular responses to lipoprotein particles. In the following years, several ligands present in HDL and their receptors expressed on the cell surface could be identified that mediate one or more physiological effects attributed to this lipoprotein. In addition, the interaction of HDL with the cell membrane was found to potently modify signal transduction induced by other ligands such as pathogen-associated molecular patterns (PAMPs) or growth factors. This review attempts to summarize the current state of knowledge on HDL-induced signal transduction and its physiological relevance.
2 ApoA-I-Induced Cell Signaling Directly Mediated by ABCA1
ATP-binding cassette transporter A1 (ABCA1) has been originally identified as a protein defective in familial HDL deficiency (Tangier disease), which is characterized on the molecular level by the failure of apoA-I to mobilize cholesterol from intracellular stores, to induce cholesterol and phospholipid efflux from cells, and ultimately to initiate the maturation of HDL particles and the HDL-mediated reversed cholesterol transport (Bodzioch et al. 1999; Brooks-Wilson et al. 1999; Rust et al. 1999). Early models explaining ABCA1-mediated lipid efflux from cells assumed that this transporter facilitates translocation of cholesterol and phospholipids to the exofacial leaflet of the cell membrane, where they became accessible for the passive transfer to apoA-I. Further studies, however, provided several pieces of evidence suggesting that apoA-I may directly interact with ABCA1 for the effective unloading of cell cholesterol. For instance, cells obtained from patients with Tangier disease displayed reduced number of apoA-I binding sites, whereas increased apoA-I binding to the cell surface could be observed in cells overexpressing full-length cDNA of ABCA1 (Francis et al. 1995; Wang et al. 2000). Moreover, studies utilizing chemical cross-linking indicated that apoA-I and ABCA1 are in a very close proximity (<7 Å) characteristic of ligand-receptor interaction, while naturally occurring mutations in extracellular ABCA1 loops profoundly impaired the cross-linking efficiency (Wang et al. 2000; Fitzgerald et al. 2002, 2004; Chroni et al. 2004). The binding of apoA-I to ABCA1 was subsequently found to increase the cell content of ABCA1 protein but not the ABCA1 mRNA abundance suggesting that apoA-I-ABCA1 interaction might be critical to ABCA1 stabilization (Wang et al. 2003). Actually, apoA-I was found to promote both the dephosphorylation of the proline-, glutamic acid-, serine-, and threonine-rich (PEST) sequence in the cytoplasmatic domain of ABCA1 and the recruitment of calmodulin to the specific 1-5-8-14 binding motif localized in a close vicinity to the PEST sequence (Wang et al. 2003; Martinez et al. 2003a; Iwamoto et al. 2010). In this manner, apoA-I binding appears to protect ABCA1 against calpain-mediated proteolysis. In addition, lipid-free apoA-I was found to retard ABCA1 degradation mediated by thiol proteases and thereby to facilitate ABCA1 accumulation on the cell surface (Arakawa and Yokoyama 2002).
ABCA1 is well known to be regulated by cyclic adenosine monophosphate (cAMP) at the transcriptional level, and cAMP responsive elements have been localized in the ABCA1 promoter. The interaction of apoA-I with ABCA1, in turn, was demonstrated to activate adenylate cyclase and to increase cAMP content, while severely reduced apoA-I-induced cAMP production was found in cells obtained from carriers of ABCA1 mutations (Haidar et al. 2004). These observations strongly suggested that cAMP and its downstream target protein kinase A (PKA) may represent an important component of apoA-I/ABCA1-induced signaling cascade regulating intracellular lipid translocation and/or their removal from cells. Actually, PKA phosphorylation sites were identified in the nucleotide binding domains of ABCA1, and the targeted mutation of one of these sites, S2054, significantly impaired apoA-I-mediated lipid efflux (See et al. 2002). Moreover, pharmacological modulation of cAMP/PKA signaling pathway exerted profound effects on apoA-I-induced cholesterol efflux. For instance, forskolin, an adenyl cyclase activator, and 8-Br-cAMP, a stable cAMP analog, substantially increased both ABCA1 phosphorylation and cholesterol efflux to apoA-I in wild-type cells and partly corrected the impaired efflux in cells obtained from patients with Tangier disease (Haidar et al. 2002). Similar effects were more recently observed in macrophages treated with Ht31, a PKA-anchoring inhibitor, which robustly elevated PKA activity in the cytoplasm, elevated ABCA1-dependent cholesterol export from cells, and completely reversed foam cell formation (Ma et al. 2011). Conversely, interference with cAMP/PKA signaling accomplished by treating cells with adenyl cyclase inhibitors, PKA inhibitors, or eicosapentaenoic acid, which decreased both cAMP levels and PKA activity, led to concomitant reduction of ABCA1 phosphorylation and ABCA1-mediated cholesterol efflux (Haidar et al. 2002, 2004; Hu et al. 2009). The molecular mechanisms underlying the stimulatory effect of PKA-mediated ABCA1 phosphorylation on the function of this transporter remain unclear. In contrast to PKC, PKA neither improves ABCA1 stability nor increases apoA-I binding. It seems more likely that phosphorylation of ABCA1 by the cAMP/PKA pathway allows the transporter to assume the conformation favoring more effective lipid translocation across the cell membrane (See et al. 2002).
Janus kinase 2 (JAK2) represents the third kinase, the activity of which is induced by ApoA-I/ABCA1 interaction. Even short exposure of cells expressing ABCA1 to apoA-I produces substantial autophosphorylation of JAK2, which is a prerequisite for further signal transmission to downstream located targets (Tang et al. 2004). Active JAK2 does not seem to directly phosphorylate ABCA1, and it is more likely that it targets an as yet unknown ABCA1 partner protein. Nevertheless, several experimental approaches clearly documented that JAK2 signaling is required for ABCA1 to effectively mediate lipid export from cells. For instance, JAK2-specific inhibitors substantially reduced apoA-I-induced cholesterol efflux, and mutant cells lacking JAK2 exhibited a severely impaired lipid transport (Tang et al. 2004, 2006). In addition, studies exploiting various ABCA1 variants revealed close interdependency between lipid export and JAK2 activation (Vaughan et al. 2009). In contrast to PKC and PKA, JAK2 activation by apoA-I neither promotes ABCA1 stability nor enhances intrinsic lipid translocase activity of the transporter. However, JAK2 activation seems to be critical for the optimum formation of apoA-I/ABCA1 complex, as JAK2-deficient cells exhibit severely impaired apoA-I binding despite normal ABCA1 protein levels (Tang et al. 2006).
JAK2 autophosphorylation is commonly followed by phosphorylation of several downstream targets, out of which signal transducer and activator of transcription 3 (STAT3) represents its canonical target. As ABCA1 possesses STAT3 docking sites, this molecule was an obvious candidate for linking JAK2 with ABCA1 activation. As a matter of fact, apoA-I induces STAT3 phosphorylation (which is tantamount to its activation) in an ABCA1-dependent fashion, and this effect is attenuated in cells expressing ABCA1, in which STAT3 docking sites have been mutated (Tang et al. 2009). However, neither complete STAT3 deficiency nor interruption of ABCA1-STAT3 interaction altered ABCA1 capacity to bind apoA-I and to efflux cholesterol from cells (Tang et al. 2009). Hence, apoA-I-induced STAT3 activation seems dispensable for the ABCA1-mediated lipid export. STAT3 represents an intracellular target of IL-10—a cytokine with well-established anti-inflammatory properties—and activation of STAT3 was demonstrated to suppress several aspects of inflammatory macrophage activation. These observations raised the interesting possibility that apoA-I might produce anti-inflammatory effects by activating STAT3 in ABCA1-dependent fashion. Actually, exposure of ABCA1-expressing macrophages to apoA-I suppressed lipopolysaccharide (LPS)-induced production of proinflammatory cytokines such as interleukin-6 (IL-6), and this effects were abolished after pharmacological STAT3 blockade or knocking down STAT3 expression with siRNA (Tang et al. 2009). Further investigations demonstrated that the suppressing effect of apoA-I on LPS-induced cytokine production was mediated by mRNA destabilization brought about by mRNA-destabilizing protein tristetraprolin, the expression of which was induced in cells exposed to apoA-I in an ABCA1- and JAK2/STAT3-dependent manner (Yin et al. 2011). Recently, apoA-I was demonstrated to promote cyclooxygenase-2 (COX-2) activation and prostacyclin (PGI2) production in endothelial cells, and these anti-inflammatory effects were also found to depend on ABCA1 and JAK2 activation (Liu et al. 2011).
Cell division control protein 42 (Cdc42) is a member of a small G-protein family involved in the regulation of cytoskeleton organization and intracellular vesicular trafficking. Early studies revealed decreased Cdc42 expression and abnormal cytoskeleton architecture in cells obtained from Tangier patients (Hirano et al. 2000). Moreover, impaired cholesterol efflux observed in aged human fibroblasts as well as fibroblasts from Werner syndrome characterized by the early onset of senescent phenotypes including premature atherosclerotic cardiovascular disease could be corrected by forced expression of wild-type Cdc42 (Tsukamoto et al. 2002; Zhang et al. 2005). Expression of constitutively active Cdc42 variant in normal Madin-Darby canine kidney (MDCK) cells potentiated apoA-I-induced cholesterol efflux, whereas expression of dominant negative Cdc42 variants in normal MDCK cells or fibroblasts exerted opposite effects (Hirano et al. 2000; Nofer et al. 2003). ApoA-I-induced cholesterol efflux was also reduced in cells pretreated with Clostridium difficile toxin, which irreversibly modifies and disables Cdc42 (Nofer et al. 2003). Both Cdc42 and ABCA1 were found to co-segregate into lubrol-insoluble, triton-soluble lipid rafts (Drobnik et al. 2002). In addition, Cdc42 could be immunoprecipitated with wild-type ABCA1, but not with C-terminally truncated variant of this transporter indicating that both proteins interact with each other and suggesting that C-terminus of ABCA1 harbors a potential docking site for Cdc42 (Tsukamoto et al. 2001; Nofer et al. 2006). Exposure of cells expressing wild-type ABCA, but not cells obtained from Tangier patients or expressing C-terminally truncated variant of ABCA1 to apoA-I, induced Cdc42 activation and subsequent actin polymerization (Nofer et al. 2006). Taken together, these findings suggest that apoA-I signals through ABCA1 to activate Cdc42 and thereby to promote cholesterol efflux. How exactly Cdc42 activation promotes lipid export from cells remains to be elucidated. One possibility is that Cdc42 modulates intracellular vesicular traffic and thereby helps to supply cholesterol and/or phospholipids to an intracellular compartment, where they are available for the ABCA1-mediated release. The retardation of intracellular lipid transport in cells with reduced Cdc42 levels and its intensification after enforced expression of Cdc42 appears to support this concept (Tsukamoto et al. 2002). Alternatively, components of the signaling machinery localized downstream to Cdc42 might accelerate intracellular lipid transfer or the ABCA1-dependent lipid translocation across the cell membrane. Actually, apoA-I was demonstrated to produce consecutive activation of two serine/threonine kinases, namely, p21-activated kinase 1 (PAK1) and c-Jun N-terminal kinase (JNK) in ABCA1- and Cdc42-dependent manner (Nofer et al. 2003). The potential involvement of these two kinases in the processes of lipid export from cells has been not clarified to date.
Mobilization of intracellular Ca(2+) from intracellular stores or Ca(2+) influx from extracellular space represents an integral component of diverse intracellular signaling pathways. In one single report, evidence has been provided that apoA-I provokes extracellular Ca(2+) influx in an ABCA1-dependent manner and that inhibition of calcineurin, the downstream target of Ca(2+) influx, with cyclosporine A or FK506 completely abolished apoA-I lipidation and also interfered with JAK2 phosphorylation (Karwatsky et al. 2010). However, other authors failed to observe Ca(2+) mobilization in fibroblasts or smooth muscle cells after exposure to apoA-I or other apolipoproteins (apoA-II, apoC-III) (Nofer et al. 2000 and unpublished observations). Further studies are necessary, therefore, to fully understand the contribution of Ca(2+) mobilization to ABCA1-mediated signaling and lipid export.
3 ApoA-I-Induced Cell Signaling Indirectly Mediated by β-ATPase and P2Y12/13 ADP Receptor
F1F0 ATP synthase is the terminal enzyme of the oxidative phosphorylation pathway responsible for the lion part of ATP synthesis in humans and other living being except in Archaea and present in mitochondria, chloroplasts, and prokaryotic membranes. Unexpectedly, the F1F0 ATP synthase (with F1 domain facing outside) has been found on the surface of several cells including hepatocytes, adipocytes, and endothelial cells, and its β-chain (β-ATPase) has been characterized as a receptor for apoA-I (Martinez et al. 2003b; Fabre et al. 2006; Radojkovic et al. 2009; Howard et al. 2011). The binding of apoA-I to β-ATPase induces ATP hydrolysis and promotes extracellular ADP generation, which in turn stimulates intracellular signaling by activating purinergic receptors, which belong to the family of G-protein-coupled receptors stimulated by extracellular nucleotides (Martinez et al. 2003b; Fabre et al. 2006). In hepatocytes, the β-ATPase/P2Y13-induced signaling was reported to activate the small G-protein RhoA and the downstream-located Rho-associated protein kinase (ROCK-I) and thereby to promote HDL endocytosis by as yet poorly defined mechanism (Fabre et al. 2006; Malaval et al. 2009). In endothelial cells, binding of apoA-I to β-ATPase and the subsequent generation of extracellular ADP stimulated endothelial proliferation and survival, but components of intracellular signaling cascade responsible for these effects remained uncharacterized (Radojkovic et al. 2009). In a recent study, the activation of β-ATPase/P2Y12 was found to precede HDL binding and cell association as well as ABCG1- and SR-BI-dependent internalization, and transcytosis in endothelial cells (Cavelier et al. 2012). The identity of intracellular signaling pathway linking P2Y12 activation to HDL internalization has not been addressed in this study.
4 ApoA-I- and HDL-Induced Cell Signaling Indirectly Mediated by ABCA1 and/or ABCG1
ABCA1 is unique among ABC transporters in that it requires the direct binding of acceptors for the transported substances. ATP-binding cassette transporter G1 (ABCG1) was also found to be a crucial mediator of lipid export. Studies by Wang et al. demonstrated that ABCG1 promotes efflux to a variety of acceptors, including HDL, low-density lipoprotein (LDL), phospholipid vesicles, and cyclodextrin (Wang et al. 2004, 2006). In major contrast to ABCA1, ABCG1 does not engage into direct interaction with lipid acceptors and for this reason cannot serve as a triggering element of intracellular second messenger cascade. However, recent studies document that ABCG1 (and to lesser extent ABCA1) profoundly impacts on signaling processes taking place in contiguity. Several studies showed that ABCA1 regulates cholesterol distribution between raft and non-raft membrane fractions and pointed to increased lipid raft formation in macrophages and other cells deficient in ABCG1 and ABCA1 (Landry et al. 2006; Sano et al. 2007; Koseki et al. 2007; Zhu et al. 2008a, Yvan-Charvet et al. 2008). Lipid rafts provide a molecular platform securing efficient signal transduction through several receptor families including proinflammatory Toll-like receptors (TLRs). Not surprisingly, therefore, both ABCA1-deficient cells and ABCG1-deficient cells, and in particular cells lacking both ABC transporters, exhibited increased cell surface expression of TLR4 and sensitivity to LPS stimulation (Koseki et al. 2007; Zhu et al. 2008a, Yvan-Charvet et al. 2008; Mogilenko et al. 2012). Also the NF-κB activation-dependent cytokine and chemokine release in response to stimulation with agonists of other TLR receptors (TLR2, TLR3) or other proinflammatory receptors (CD40) revealed to be increased in macrophages deficient in ABCA1 and/or ABCG1 (Sun et al. 2009; Yin et al. 2012). Such cells were also found to be prone to apoptosis during efferocytosis (phagocytosis of apoptotic cells) as compared to wild-type cells (Yvan-Charvet et al. 2010a). The enhanced apoptotic response observed in ABCA1- and ABCG1-deficient macrophages was dependent on an excessive assembly of NADPH oxidase/NOX2 complexes within lipid rafts followed by oxidative burst and ensuing sustained JNK activation, which turned on the apoptotic cell death program. It is very likely that similar mechanisms consisting in amplification of transmembrane signaling under conditions of decreased ABCG1 and/or ABCA1 activity and the resulting accumulation of lipid rafts may substantially contribute to the enhancement of several proinflammatory processes. For instance, cholesterol accumulation within lipid rafts was suggested to enhance monocyte adhesion and neutrophil activation, and both processes could be reversed by cholesterol removal from cells carried out either with help of synthetic cholesterol acceptors (e.g., cyclodextrin) or through induction of ABCA1-dependent cholesterol efflux (Murphy et al. 2008, 2011). In addition, ABCG1- and ABCA1-dependent modulation of cell membrane cholesterol content was demonstrated to critically regulate cellular signaling related to proliferatory responses. For instance, mice deficient in ABCG1- and ABCA1 developed severe monocytosis and neutrophilia, which was dependent on uncontrolled proliferation and expansion of hematopoietic stem progenitor cells (HSPC) (Yvan-Charvet et al. 2010b). The mechanistic analysis revealed increased plasma membrane lipid raft formation in these cells accompanied by elevated cell surface presence of interleukin IL-3/granulocyte macrophage colony-stimulating factor (GM-CSF) receptor. Addition of IL-3 or GM-CSF to ABCA1/G1-deficient bone marrow led to increased proliferation of HSPCs, reflecting activation of signaling pathways downstream of the IL-3/GM-CSF receptor. Similarly, defects in cholesterol efflux pathways in macrophages and dendritic cells in spleens of ABCA1/G1-deficient mice led to increased production of IL-23 and activation of a proinflammatory and proatherogenic signaling axis involving IL-23/IL-17 (Westerterp et al. 2012). Other studies showed increased lipid raft content in plasma membrane and enhanced proliferation in response to anti-CD3 antibody stimulation in T cells obtained from ABCG1-deficient animals (Bensinger et al. 2008). Similar increase in lymph node T-cell proliferation has been noted after disruption of cholesterol efflux pathway by deletion of apoA-I on a hypercholesterolemic background (Wilhelm et al. 2009).
5 HDL-Induced Cell Signaling Mediated by SR-BI
Scavenger receptor type I (SR-BI) is a cell surface glycoprotein originally identified by its homology to the scavenger receptor CD36 and capacity to bind acetylated LDL. Later studies in vitro revealed that SR-BI serves as the binding partner for a variety of ligands including native and oxidized LDL, very-low-density lipoproteins (VLDL), Lp(a), modified serum proteins such as maleylated albumin, lipid vesicles containing negatively charged phospholipids, apoptotic cells, and HDL. Despite this promiscuous binding behavior SR-BI is believed to primarily function as a physiologically relevant HDL receptor because of its tissue expression pattern reflecting sites of HDL cholesterol uptake and because of the profound impact of complete or tissue-specific SR-BI deficiency on HDL metabolism. The classic cellular function of SR-BI is to mediate the selective uptake of lipids from HDL and to transfer them into cells. Cholesteryl esters represent the primary substrate for SR-BI, but this receptor also facilitates transfer of various other lipid species including unesterified cholesterol, phospholipids, triglycerides, and lipid-soluble vitamins. SR-BI also mediates the efflux of cholesterol from cells to HDL, which in contrast to ABCA1- and ABCG1-mediated efflux is a passive, concentration gradient-dependent, and energy-independent process. Importantly, SR-BI-mediated lipid transfer from and to HDL occurs in the absence of consecutive internalization and degradation of HDL particle, which is a precedent condition for effective signal transduction (for recent reviews, see Hoekstra et al. 2010; Mineo and Shaul 2012).
The concerted activation of both Akt and MAPK appears to constitute the centerpiece of SR-BI-triggered signaling cascades. Both kinases are required for phosphorylation and full activation of endothelial nitric oxide synthase (eNOS), nitric oxide generation, and vasorelaxation, which are observed in endothelial cells or isolated aortas in the presence of HDL (Yuhanna et al. 2001; Mineo et al. 2003; Assanasen et al. 2005). In addition, PI3K/Akt/eNOS signaling was found to be involved in the HDL-induced COX-2 expression and PGI2 release in endothelial cells (Zhang et al. 2012). The proliferation of bone-derived mesenchymal stem cells in response to HDL was likewise mediated via its binding to scavenger receptor-B type I and activation of PI3K/Akt/MAPK pathways (Xu et al. 2012). Activation of MAPK but not Akt with subsequent induction of Rac1-GTPase was demonstrated to promote HDL-induced endothelial cell migration, which is critical for carotid artery reendothelialization after perivascular electric injury (Seetharam et al. 2006). Conversely, activation of Akt but not MAPK was necessary to sustain the migration and the proliferation of breast cancer cells (Danilo et al. 2013). In adipocytes, SR-BI-mediated activations of Akt and AMPK were found to promote glucose uptake and translocation of the glucose transporter GLUT4 to the plasma membrane suggesting that SR-BI signaling might contribute to insulin-sensitizing effects of HDL (Zhang et al. 2011). Finally, augmentation of hypoxia-induced angiogenesis, which is observed in the presence of HDL, was demonstrated to depend on SR-BI-mediated activation of Akt linked to increased expression of E3 ubiquitin ligases Siah1 and Siah2 and accelerated degradation of prolyl hydroxylases (PHD1-PHD3) combined with stabilization of hypoxia-inducible factor 1α (HIF-1α) (Tan et al. 2013). Recent studies provide evidence that similar to ABCG1 also SR-BI may indirectly affect proinflammatory signaling by reducing membrane cholesterol content. Actually, macrophages obtained from SR-BI-deficient animals were characterized by increased inflammatory responses to Toll-like receptor agonists (Guo et al. 2009; Cai et al. 2012). Feng et al. found that SR-BI deficiency enhanced lymphocyte proliferation, caused imbalanced interferon-γ (IFNγ) and IL-4 production in lymphocytes, and led to elevated inflammatory cytokine production in macrophages (Feng et al. 2011). Umemoto et al. demonstrated that HDL-induced disruption of lipid raft in adipocytes, which is accompanied by reduced activation of NADPH oxidase and production of monocyte chemotactic protein 1 (MCP1), was reversed by silencing both ABCG1 and SR-BI expression (Umemoto et al. 2013).
It is well established that HDL inhibits the activation of platelets by strong agonists such as thrombin and collagen (for review, see Nofer and van Eck 2011). The presence of SR-BI has been identified on platelet surface (Imachi et al. 2003). Recent investigations demonstrated that HDL binding to platelets was reduced in a concentration-dependent fashion by SR-BI ligands such as negatively charged liposomes, which in addition potently inhibited thrombin-induced platelet aggregation, granule secretion, fibrinogen binding, and mobilization of intracellular Ca2+ (Nofer et al. 2011). Furthermore, both native HDL and other SR-BI ligands failed to inhibit thrombin-induced activation of platelets obtained from SR-BI-deficient mice. These findings together with the results of Imachi et al. (2003), who observed an inverse relationship between the levels of SR-BI expression and the platelet aggregation in response to ADP, corroborate the hypothesis that SR-BI is a true functional HDL receptor on platelets. The likely consequence of the HDL binding to SR-BI on platelets is the induction of intracellular second messenger cascade. It has been previously demonstrated that incubation of platelets with HDL leads to the release of DAG from the plasma membrane phosphatidylcholine and to subsequent activation of PKC, which in turn stimulates the Na+/H+ antiport, promotes the alkalization of the cytoplasmic compartment, and inhibits the release of calcium ions from intracellular storage sites (Nofer et al. 1996). In addition, activation of PKC inhibits the activity of phosphatidylinositol-specific phospholipase C (PI-PLC), which is one of the most important signal transduction mediators of agonists such as thrombin and collagen (Nofer et al. 1998). It is conceivable that this signaling cascade is launched by interaction of HDL with SR-BI on platelet surface.
6 HDL-Induced Cell Signaling Mediated by S1P
Early studies on HDL-induced signal transduction focused on apoA-I or other apolipoproteins as potential binding partners for HDL receptor responsible for initiating the cascade of intracellular signaling events. However, it has soon become clear that neither apoA-I nor other proteins encountered in HDL particles can exclusively account for HDL signaling activities. Studies utilizing fibroblasts demonstrated that PI-PLC activation, which is observed in cells exposed to HDL and accompanied by increased phosphatidylinositol bisphosphate (PIP2) turnover and the mobilization of intracellular calcium, is not emulated by purified apoA-I or apoA-I-containing proteoliposomes, but occurs in the presence of lipids extracted from HDL holoparticles (Nofer et al. 2000). Likewise, certain cellular responses triggered by HDL in fibroblasts of smooth muscle cells such as cell proliferation or proliferation-linked activation of MAPK could be observed in the presence of HDL lipids or HDL particles, in which proteins were covalently modified or enzymatically degraded, but not in the presence of proteins isolated from HDL (Deeg et al. 1997; Sachinidis et al. 1999; Nofer et al. 2001a). The intense search for lipid species accounting for the signal-inducing capacity of HDL revealed that these lipoproteins serve as a carrier for bioactive lysosphingolipids such as sphingosine-1-phosphate (S1P), sphingosylphosphorylcholine (SPC), and lysosulfatide (LSF) and that signaling events observed in cells exposed to native HDL or isolated lysosphingolipids largely overlap (Sachinidis et al. 1999; Nofer et al. 2000, 2004; Kimura et al. 2001).
S1P is the most prominent and best characterized member of the lysosphingolipid family. It is produced by phosphorylation of sphingosine by sphingosine kinases 1 and 2 (SK1 and SK2) in response to a variety of stimuli and degraded to phosphoethanolamine and hexadecanal by S1P-lyase or to sphingosine and organic phosphate by S1P phosphatase. S1P interacts with five related G-protein-coupled receptors termed S1P1–5, which regulate a wide spectrum of cellular functions, including proliferation and survival, cytoskeletal rearrangements, and cell motility, and exert potent cytoprotective effects. In the vasculature, S1P receptors were identified on endothelial and smooth muscle cells, and S1P provides essential contribution to the new vessel formation and the maintenance of vascular barrier integrity. In addition, S1P was demonstrated to interfere with proliferation, migration, and activation of monocytes/macrophages and to regulate their recruitment to sites of inflammation (for recent reviews, see Maceyka et al. 2012; Levkau 2013). Erythrocytes and platelets are major sources of S1P in plasma, where it is preferentially (up to 60 %) associated with HDL subfraction rich in apolipoprotein M (apoM) (Dahm et al. 2006; Hänel et al. 2007; Christoffersen et al. 2011). Recent studies demonstrate that apoM specifically binds S1P and stimulates its biosynthesis for secretion (Liu et al. 2013b). There is substantial evidence suggesting that S1P may directly contribute to atheroprotective effects attributed to HDL. For instance, S1P closely correlates with HDL in a concentration range, in which HDL most effectively protects against atherosclerosis, and decreased HDL-bound S1P levels were noted in patients with coronary artery disease and myocardial infarction (Karuna et al. 2011; Sattler et al. 2010; Argraves et al. 2011). In addition, S1P was found to emulate in vitro several atheroprotective effects of HDL, which themselves could be attenuated by interrupting signaling through S1P receptors (Levkau et al. 2013).
As pointed out above, HDL is strongly mitogenic for smooth muscle cells, and HDL-associated lysosphingolipids appear to account for this effect. On the other hand, HDL was identified as a potent inhibitor of smooth muscle cell migration, which is a hallmark of the development of advanced atherosclerotic plaque and restenosis (Tamama et al. 2005; Damirin et al. 2007). The inhibitory effect of HDL was likely mediated by its S1P content, as it was abolished in the presence of S1P2-specific siRNA and following desensitization of S1P receptors and much enhanced in cells overexpressing S1P2 receptor. Further effects exerted by HDL on smooth muscle cell through its S1P content include modulation of the activity of several vasoactive factors. Chrisman et al. found that HDL-associated S1P desensitizes guanylyl cyclase B, a receptor for C-type natriuretic peptide (CNP), and thereby inhibits CNP-induced cyclic guanosine monophosphate (cGMP) accumulation in vascular smooth muscle cells (Chrisman et al. 2003). Gonzalez-Diez and colleagues demonstrated induction of COX-2 expression and PGI2 production in vascular smooth muscle cells exposed to HDL, which was mediated through S1P receptors 2 and 3 and protein kinase p38MAPK (González-Díez et al. 2008). They have also shown that simvastatin potentiates the HDL- and S1P-induced COX-2 expression by upregulating cellular levels of S1P3. Finally, HDL-associated lysosphingolipids S1P and SPC were found to suppress the proinflammatory activation of smooth muscle cells brought about by stimulation with thrombin and characterized by the increased release of inflammatory cyto- and chemokines such as MCP1 (Tölle et al. 2008). The detailed examination of signaling pathways accounting for inhibitory effects exerted by HDL-associated lysosphingolipids revealed that they were related to the inhibition of Rac1-mediated activation of NADPH oxidase and generation of reactive oxygen species. S1P3 was identified to mediate anti-inflammatory effects of HDL, S1P, and SPC in smooth muscle cells.
Cardiomyocytes express several S1P receptors, and HDL was found to exert protective effects on these cells through its S1P content. HDL and S1P showed potent capacity to protect cardiomyocytes against hypoxia-reoxygenation- or doxorubicin-induced apoptosis, and these effects were attributed to the induction of protein kinases Src, MAPK, and Akt and the transcription factor STAT3 (Frias et al. 2009, 2010; Tao et al. 2010). However, investigations concerning the identity of receptor-mediating protective effects of HDL-associated S1P were inconclusive, and equally supportive evidence for the involvement of either S1P2 or S1P1 and S1P3 has been provided. Recently, short-term treatment of cardiomyocytes with HDL or S1P was found to induce a PKC-dependent phosphorylation of connexin43, which is a gap junction protein present in ventricular cardiomyocytes and involved in cardioprotection by modulating ischemic preconditioning (Morel et al. 2012). In addition, HDL and S1P improved gap junctional communication, but only incrementally affected conduction velocities in cardiomyocytes. The physiological relevance of these effects has been clearly documented using an ex vivo approach, in which both HDL and S1P protected perfused hearts against ischemia/reperfusion-induced cell death.
Several other cell types were found to be responsive to stimulation with HDL-associated lysosphingolipids, but the physiological relevance of these effects remains obscure. For instance, HDL and S1P were demonstrated to stimulate prostate cancer cell migration and invasion through S1P receptors 2 and 3 and activation of MAPK and STAT3, to stimulate plasminogen activator inhibitor-1 (PAI-1) release from adipocytes through S1P2 and activation of PKC and Rho kinase, and to promote release of fibroblast growth factor-2 (FGF2), a potent neurotrophic factor, from astroglial cells through activation of PI-PLC and MAPK (Malchinkhuu et al. 2003; Lee et al. 2010; Sekine et al. 2011). HDL particles isolated from body fluids other than plasma were also found to contain S1P and to stimulate intracellular signaling through their S1P content. For instance, HDL-like particles obtained from cerebrospinal fluid stimulated astrocyte migration and oligodendrocyte retraction in a manner sensitive to inhibitors of S1P receptors 1 and 3 (Sato et al. 2007). Follicular fluid-derived HDL was found to promote human granulosa lutein cell migration through S1P3 and activation of small G-protein Rac1 (Becker et al. 2011). Despite the prominent role played by monocytes and macrophages in the pathogenesis of atherosclerosis, very little is known about the modulatory effects exerted by S1P cargo of HDL particles on these cells. In one study, both HDL and S1P were found to blunt monocyte activation induced by agonists of TLR2, but the attribution of this effect to the HDL-associated S1P was not unequivocally demonstrated (Dueñas et al. 2008).
7 HDL-Induced Cell Signaling: Future Challenges and Opportunities
HDL constitutes a large and heterogenous macromolecule carrying more than 50 proteins, several amphipathic peptides, miRNAs, and hundreds of lipid species representing almost all known lipid classes. Several proteins and lipids identified in HDL—among others, apolipoprotein H, haptoglobin, and steroid hormones—are recognized receptor ligands and inductors of intracellular signaling events. However, the rigid evidence demonstrating that these proteins or lipids contribute to the HDL-induced signaling governing diverse functional responses to this lipoprotein has been not provided to date or is at best equivocal. For instance, several studies documented that HDL carries estradiol in the form of fatty acyl esters produced in a reaction catalyzed by LCAT. HDL-derived estradiol esters are internalized via SR-BI-mediated mechanisms and subsequently hydrolyzed intracellularly to unfold their effects. However, the specific signaling events and/or physiological effects clearly attributable to the estradiol content in HDL particles have not been unequivocally defined. In one study, HDL-associated estradiol was found to stimulate eNOS and vasodilation in an SR-BI- and Akt-dependent manner, but these results were questioned by others (Gong et al. 2003; Nofer et al. 2004). In other study, dehydroepiandrosterone (DHEA) fatty acyl esters once incorporated in HDL were found to induce stronger vasodilatory response, and this effect appeared also to be mediated by SR-BI (Paatela et al. 2011). However, no clear signaling link between the HDL-associated DHEA and vasodilation has been established in this study. These two examples clearly show that the assignment of ligands present in HDL to the appending signaling machinery and to various functions exerted by these lipoproteins in physiology will be a major challenge in the future.
The composition of HDL particles is variable and influenced both by normal metabolism and in various states of pathology. For instance, HDL particles isolated from subjects suffering from acute or chronic inflammatory diseases were found to lose proteins and enzymes with established or presumed antiatherogenic function such as apoA-I or LCAT and to concomitantly acquire proinflammatory or prooxidative factors such as serum amyloid A (SAA), ceruloplasmin, lipoprotein-specific phospholipase A2, or myeloperoxidase. Not surprisingly, such inflammatory HDL particles are severely impeded in their capacity to trigger certain signaling pathways, but at the same time may acquire new ligands and thereby alter their signaling properties. For instance, the relatively high content of apolipoprotein J (apoJ) in HDL particles obtained from healthy subjects endowed them with important antiatherogenic property to promote endothelial survival by stimulation of the expression of antiapoptotic protein Bcl-XL (Riwanto et al. 2013). By contrast, HDL particles isolated from patients with stable coronary heart disease (CHD) or acute coronary syndromes were found to be poor in apoJ, but instead enriched in apoC-I or apoC-III (McNeal et al. 2013; Riwanto et al. 2013). Furthermore, the increased content of these apolipoproteins turned HDL particles into potent apoptosis inducers in vascular smooth muscle and endothelial cells by virtue of stimulating the expression of proapoptotic protein tBid. In addition to CHD, dysfunctional HDL is also encountered in chronic conditions such as chronic kidney disease (CKD). Abnormal HDL composition characterized primarily by increased serum amyloid A and apoC-III content as well as compromised HDL functionality has been repeatedly reported in uremic patients (Holzer et al. 2011; Weichhart et al. 2012). In one recent study, SAA-enriched HDL encountered in CKD was found to induce production of a proinflammatory chemokine MCP1 in smooth muscle cells by interaction with formyl peptide receptor 2 (FPR2) (Tölle et al. 2012). In other study conducted in CKD patients, accumulation of symmetric dimethylarginine (SDMA) in HDL was found to turn this lipoprotein into a noxious particle reducing NO bioavailability, evoking endothelial dysfunction, and subsequently increasing arterial blood pressure. The deleterious effects of SDMA-modified HDL were attributed to the noncanonical activation of TLR2 (Speer et al. 2013). This exemplifies another challenge facing HDL research in the future, namely, identification of ligands, receptors, and signaling cascades attributable to HDL in various states of pathology.
Notwithstanding the physiological or pathological character of HDL-induced signal transduction, its elucidation may open new opportunities for the development of antiatherogenic, anti-inflammatory, or antithrombotic drugs. Several interventional studies conducted recently and aiming at cardiovascular risk reduction through simple elevation of plasma HDL cholesterol fall through. This epic failure emphasizes the necessity to develop more sophisticated strategies based on targeted exploitation of selected antiatherogenic or anti-inflammatory activities exerted by HDL particles. The proper understanding of HDL-induced signaling may help to specifically enhance beneficial physiological or to disable harmful pathological effects of HDL and thereby to design more efficacious therapeutic approaches for cardiovascular and other inflammatory diseases.
- Nofer JR, Walter M, Kehrel B et al (1998) HDL3-mediated inhibition of thrombin-induced platelet aggregation and fibrinogen binding occurs via decreased production of phosphoinositide-derived second messengers 1,2-diacylglycerol and inositol 1,4,5-tris-phosphate. Arterioscler Thromb Vasc Biol 18:861–869PubMedGoogle Scholar
- Tsukamoto K, Hirano K, Yamashita S et al (2002) Retarded intracellular lipid transport associated with reduced expression of Cdc42, a member of Rho-GTPases, in human aged skin fibroblasts: a possible function of Cdc42 in mediating intracellular lipid transport. Arterioscler Thromb Vasc Biol 22:1899–1904PubMedGoogle Scholar
- von Otte S, Paletta JR, Becker S et al (2006) Follicular fluid high density lipoprotein-associated sphingosine 1-phosphate is a novel mediator of ovarian angiogenesis. J Biol Chem 281:5398–5405Google Scholar
- Walter M, Reinecke H, Gerdes U et al (1996) Defective regulation of phosphatidylcholine-specific phospholipases C and D in a kindred with Tangier disease. Evidence for the involvement of phosphatidylcholine breakdown in HDL-mediated cholesterol efflux mechanisms. J Clin Invest 98:2315–2323PubMedCentralPubMedGoogle Scholar
- Wilkerson BA, Grass GD, Wing SB et al (2012) Sphingosine 1-phosphate (S1P) carrier-dependent regulation of endothelial barrier: high density lipoprotein (HDL)-S1P prolongs endothelial barrier enhancement as compared with albumin-S1P via effects on levels, trafficking, and signaling of S1P1. J Biol Chem 287:44645–44653PubMedCentralPubMedGoogle Scholar
- Yin K, Deng X, Mo ZC et al (2011) Tristetraprolin-dependent post-transcriptional regulation of inflammatory cytokine mRNA expression by apolipoprotein A-I: role of ATP-binding membrane cassette transporter A1 and signal transducer and activator of transcription 3. J Biol Chem 286:13834–13845PubMedCentralPubMedGoogle Scholar
- Yvan-Charvet L, Welch C, Pagler TA et al (2008) Increased inflammatory gene expression in ABC transporter-deficient macrophages: free cholesterol accumulation, increased signaling via toll-like receptors, and neutrophil infiltration of atherosclerotic lesions. Circulation 118:1837–1847PubMedCentralPubMedGoogle Scholar
Open Access This chapter is distributed under the terms of the Creative Commons Attribution Noncommercial License, which permits any noncommercial use, distribution, and reproduction in any medium, provided the original author(s) and source are credited.