Crosstalk Between Reverse Cholesterol Transport and Innate Immunity

Effects on cholesterol transporters

LPS downregulates macrophage ABCA1 [1, 26, 27] and ABCG1 [27], thus impairing cholesterol efflux. This and TLR3-induced cholesterol transporter downregulation occur, at least in part, through the inhibitory action of interferon regulatory factor-3 (IRF-3), at LXR binding sites in the ABCA1 promoter [26] (Figure 2). Many pro-inflammatory cytokines also downregulate ABCA1 [1, 28], although TNFα and IL-6, which are both induced in FC-loaded macrophages by ER stress [29], were shown to upregulate macrophage ABCA1 [30, 31]. Adding further complexity to the picture, LPS and zymosan have also been reported to upregulate ABCA1 [23, 32], and apoA-I itself upregulates ABCA1 through the TLR adaptor MyD88 [33], the latter suggesting a possible role for innate immune signals in cholesterol homeostasis.

Effects on HDL composition and function

In addition to reducing serum HDL, the APR causes complex changes to HDL composition [34] that in general are thought to impair its cholesterol transport and anti-inflammatory functions. In humans, decreased LCAT and CETP activity reduce formation of CE in HDL and its transfer to apoB lipoproteins. In addition, reduced levels of paraoxonase 1 (PON1) and activity of platelet-activating acyl hydrolase (PAF-AH), two HDL-associated antioxidant proteins, are thought to impair HDL antioxidant function during the APR [28, 35].

It has been shown that following LPS injection, the activity of the group IIa secretory phospholipase A2 (sPLA₂-IIa) increases, while apoA-I protein levels decrease [28]. sPLA2-IIA also decreases HDL particle size, by enhancing the hydrolysis of PLs on the HDL surface [36]. Increases in ABCG1-dependent cholesterol efflux to HDL during the APR have nonetheless recently been attributed to enrichment in HDL PL [37]. Direct sPLA₂-IIA injection does not, however, reduce RCT in vivo [22].

Serum amyloid A (SAA), a major APR protein upregulated in the liver, also exerts complex effects on HDL. It is thought to displace lipid-free apoA-I from HDL, thus enhancing its clearance from the circulation [36], and reducing in vivo RCT [22]. SAA was also found to increase cholesterol uptake by macrophages and decrease cholesterol uptake by hepatocytes [28]. EL is also upregulated during inflammation and works in conjunction with SAA to reduce nascent HDL formation by impeding ABCA1-mediated lipidation of apoA-I [38]. Inflammation induced by zymosan, a yeast glucan, impairs RCT principally by decreasing the cholesterol acceptor ability of plasma in association with increased SAA incorporation into HDL [23]. On the other hand, SAA can itself act as an acceptor for cellular cholesterol via ABCA1- and SR-BI-dependent pathways, and has been reported to enhance HDL-induced cholesterol efflux during the APR [39].

Myeloperoxidase (MPO), a pro-oxidant protein present in neutrophils and macrophages and released during the APR, also has potent effects on HDL function. MPO binds to HDL [40], specifically targeting apoA-I for oxidative modifications that are associated with reduced cholesterol efflux and LCAT-activating function [41, 42], reduced SR-BI binding [40], and reduced RCT in vivo [22]. Elegant work by several groups has begun to map out the precise residues of apoA-I that are oxidized by MPO [34]. MPO has also interestingly been shown to convert HDL into a pro-inflammatory particle capable of activating NF-κB in endothelial cells [40]. Thus, HDL is a specific target of several APR mediators, with inflammatory mediators compromising and even deranging HDLs hallmark cholesterol-mobilizing and anti-inflammatory functions.

Effects upon cholesterol transport in the liver

Two groups have recently reported that, during endotoxemia, the liver acts as the primary obstacle to RCT. McGillicuddy and colleagues reported that LPS injection reduces in vivo RCT, by impairing the liver-to-biliary transit of cholesterol, in association with downregulation of ABCG5, ABCG8, ABCG11, and CYP7A1 [24]. Annema et al. came to similar conclusions and found reduced expression of hepatic ABCG5, ABCG8, and ABCG11 after LPS injection, but only reduction in hepatic CYP27A1 and not CYP7A1 expression [22]. Whether these expression changes are direct or indirect in response to LPS remains unclear. By contrast, a third group reported that systemic exposure to zymosan impairs RCT primarily through reducing HDL quantity and quality [23]. Taken together, these reports indicate that microbial molecules impair in vivo RCT at several steps.

Cholesterol loading activates TLRs

Perhaps one of the more exciting findings of recent years in the field of innate immunity is that FC loading of macrophage membranes is sufficient to activate TLRs, perhaps due to associated changes in lipid raft microdomains. Thus, biochemical loading of macrophage plasma membrane cholesterol activates TLR4, whereas TLR3 is responsive to cholesterol loading of late endosomes [43]. Niemann Pick C mutant cells, which have defective cholesterol trafficking to the plasma membrane associated with endosomal cholesterol overload, display constitutive overcrowding and activation of TLR4 in endosomes, suggesting that TLR4 trafficking and activity are also controlled by homeostatic membrane flow [44]. In another example of coupling between TLR traffic and membrane traffic, HDL was recently reported to induce internalization of the TLR4 adaptor TRAM to intracellular compartments (Figure 2), thus impairing subsequent TLR induction of the interferon response, although this appears to be independent of cell cholesterol stores [45].

Notably, cholesterol loading may also induce pro-inflammatory responses in cells by TLR-independent mechanisms, in part determined by the specific subcellular distribution of cholesterol. Thus, FC-loaded macrophages secrete TNFα and IL-6 through activation of the C/EBP homologous protein (CHOP) branch of the unfolded protein response, as well as other ER stress-related pathways [29]. Cholesterol crystals have also recently been shown to initiate proinflammatory signaling responses, by activating the Nlrp3 inflammasome in macrophages thus leading to activation of caspase-1 and processing of IL-1β and IL-18 [46].

Effect of ABCA1/ABCG1 deletion on TLRs

Macrophages with cholesterol overloading, due to deficient efflux mechanisms, also display enhanced TLR responses, likely due to enhanced TLR trafficking to cholesterol-overloaded rafts. Abca1^−/− macrophages exhibit enhanced pro-inflammatory response to LPS and to specific TLR2, TLR7, and TLR9 agonists, but not TLR3 agonists, when compared to wild type (wt) macrophages [47, 48] (Figure 2). Lipid rafts of ABCA1-deficient macrophages contain increased cholesterol and increased TLR4 and TLR9 after cell exposure to the cognate ligands for these receptors [47, 48]. This suggests that ABCA1 dampens inflammation by reducing TLR trafficking to rafts through reduction of raft cholesterol. ABCA1 may also dampen LPS responses through cholesterol-independent mechanisms that involve activation of Janus kinase 2 (JAK2) and signal transducer and activator of transcription 3(STAT3) [49]. Abcg1^−/− macrophages also have increased cell-surface display of TLR4 as well as augmented responsiveness to LPS, while macrophages with dual ABCA1/ABCG1 deficiency have even higher amplification of TLR4 responses [50], as well as enhanced TLR2/TLR4/MyD88-dependent apoptosis in response to lipid loading during efferocytosis [51]. It was also recently reported that Abcg1^−/− mice have enhanced inflammatory responses to K. pneumoniae infection in the lung, and improved clearance of this bacterium [52]. These observations suggest that cholesterol transporters are important regulators of the innate immune response.

Multifaceted roles for Liver X Receptors in Innate Immunity

LXRα and -β are oxysterol-activated nuclear receptors that promote RCT through induction of the transporters ABCA1, ABCG1, and other target genes [53]. LXRs have also been described to have potent anti-inflammatory function, suggesting that they may act as central integrators of sterol metabolism and immunity. For example, LXR agonists suppress proinflammatory gene induction in cells, in response to LPS and bacteria [54], and attenuate LPS-induced inflammation in vivo, in several organs including the lung [55]. While LXR agonists may reduce inflammatory signaling in part through the disruption of rafts in an ABC transporter-dependent fashion [56], they also suppress inflammation through transporter-independent mechanisms. For example, sumo-modified and liganded LXR is reported to transrepress LPS-induced activation of NF-κB, in a nuclear receptor corepressor (N-CoR and SMRT) dependent mechanism [57, 58]. Recently, Mer, a receptor tyrosine kinase that facilitates efferocytosis by interacting with surface proteins on apoptotic target cells, was also identified as an LXR target gene [59]. LXR-deficient macrophages exhibit defective uptake of, and anti-inflammatory responsiveness to, apoptotic cells [59].

Interestingly, a role for LXRs in antimicrobial host defense has also been established as Lxra^−/−Lxrb^−/− mice fail to mount effective early neutrophilic, Th1, and Th17 airway responses to Mycobacterium tuberculosis infection, whereas treatment of wt mice with LXR agonists increases Th1/Th17 function and bacterial clearance [60]. LXR also promotes macrophage survival and pathogen clearance in the setting of Listeria monocytogenes infection, through induction of its target gene Apoptosis Inhibitor of Macrophages (AIM) [61]. By contrast, treatment of mice with an LXR agonist reduced neutrophil influx to the lung in response to K. pneumoniae, thereby impairing pulmonary clearance of this extracellular bacterium [55]. Thus, LXR appears to regulate host defense in a pathogen-dependent manner.

LXRs also have pro-inflammatory roles. Oxidized LDL and synthetic LXR ligands activate LXR and increase expression of its target gene TLR4 in human macrophages, thereby promoting NADPH oxidase-dependent reactive oxygen species (ROS) generation [62]. Similarly, treatment of human dendritic cells with TLR4 (LPS) or TLR3 (polyI:C) ligands in the presence of LXR activation results in prolonged NF-κB activation, augmentation of pro-inflammatory cytokine production, and an increased capacity to activate CD4⁺ T cell proliferation [63].

Taken together, as central sensors of cellular oxysterols, the LXRs likely play a pivotal role in integrating signals from sterol balance and innate immunity during metabolic stress. It is important to note in this regard that LPS itself induces cellular synthesis of oxysterol LXR ligands [64].

Multiple regulatory roles for SR-BI in innate immunity

SR-BI has recently also been shown to impact the innate immune response at several levels. Macrophages from SR-BI-null mice produce significantly higher levels of inflammatory cytokines in response to LPS than wt controls, suggesting cell-intrinsic TLR-regulatory functions of SR-BI [65]. Whether this stems from changes to lipid rafts is unclear. However, as CLA-1 and its splicing variant CLA-2 (human orthologue of the rodent SR-BII) mediate the adhesion and uptake of Gram-negative and Gram-positive bacteria [66] as well as LPS [19], clearance of microbial stimuli at the cell level may also conceivably act to dampen TLR responses. SR-BI null mice also have deficient hepatic clearance of HDL-associated LPS from serum in vivo, as well as deficient SR-BI-dependent cholesterol delivery from HDL to the adrenals that leads to impaired glucocorticoid synthesis during sepsis [67], which together lead to heightened inflammation and lethality. Conversely, transgenic mice overexpressing SR-BI are more resistant to septic death [65]. Taken together, SR-BI may protect against lethality in sepsis through the downregulation of macrophage inflammatory responses, enhanced clearance of LPS from the circulation, and increased synthesis of anti-inflammatory glucocorticoid.

Role of CETP and PLTP in LPS clearance and sepsis

Mice expressing human CETP have increased RCT, likely due to enhanced LDL receptor-dependent clearance of apoB lipoprotein cholesterol by the liver. They also display increased liver uptake of LPS and enhanced survival during endotoxemia [68]. Conversely, the delayed association of LPS with lipoproteins in PLTP null mice results in decreased LPS clearance, higher LPS toxicity, and a marked increase in LPS-induced mortality compared to wt controls [21]. Thus, RCT regulatory proteins in the serum coordinately control cholesterol and LPS disposal, playing an important role in the sepsis phenotype.

HDL – vehicle for lipids or mediator on the front lines of innate immunity?

Neutralization of LPS by HDL has been the subject of numerous studies. HDL binds both LPS and lipoteichoic acid, and also promotes the release of cell-bound LPS, thus reducing cellular activation [17, 28]. Both PLs and apolipoproteins contribute to HDL-mediated neutralization of LPS [28]. Recent studies demonstrate that HDL from EL knockout mice displays enhanced LPS neutralization properties, providing further insight into its role in the metabolic regulation of the innate immune response [69]. Indeed, HDL binds and neutralizes viruses, protects against parasitic infections [28], and mediates the lysis of trypanosomes through apolipoprotein L1 (apoL1) and haptoglobin-related protein (Hpr) [70], suggesting a much broader role in host defense than just LPS neutralization. Trypanosome lytic factor, a minor subclass of HDL composed of apoL1, Hpr, and apoA-I is now known to play a critical role in protecting humans from most species of African trypanosomes, and also inhibits infection by Leishmania [71, 72].

In addition to its activities in the extracellular milieu, HDL also suppresses inflammatory responses through its interaction with cellular cholesterol transporters. HDL and apoA-I inhibit CD11b activation in human monocytes through cholesterol efflux-dependent raft disruption [73]. HDL also works in conjunction with ABCA1/ABCG1 to protect against oxidative stress-induced macrophage apoptosis that is mediated by TLR2/TLR4/MyD88 during efferocytosis [51]. It also inhibits hematopoietic stem cell proliferation and associated monocytosis through ABCA1/ABCG1-dependent cholesterol efflux from bone marrow stem cells [74].

It was recently reported that lipid-free apoA-I, but not HDL, activates NF-κB and induces cytokines in macrophages through a pathway involving CD14, TLR2, TLR4, and MyD88 [33]. Consistent with these findings, and perhaps suggesting a broader class effect of the exchangeable apolipoproteins, apoC-III was also recently reported by another group to activate NF-κB in THP-1 cells, through interactions with TLR2 [75]. HDL and apoA-I have both also been reported to induce ADAM metallopeptidase domain 17 (ADAM17)-dependent shedding of TNFα from cells, presumably in a TLR-independent manner [5]. On the other hand, apoA-I has been shown to suppress pro-inflammatory macrophage responses to LPS in part through an ABCA1- and STAT3-dependent mechanism [49, 76], and also to be anti-inflammatory in the context of autoimmunity [77, 78]. Thus, further investigation will clearly be required to fully define the pro- vs. anti-inflammatory effects of apoA-I in vivo under different pathophysiologic conditions. Conversely, the TLR adaptor MyD88 was found to play an important role in cholesterol efflux from macrophages in vitro and RCT in vivo [33].

We speculate that MyD88 transduces signals that support homeostatic macrophage RCT during health, and that, during the APR, altered partitioning of lipid-free apoA-I may stimulate MyD88-dependent signaling in a manner that also promotes cholesterol mobilization. Recent studies suggest that this effect does not, however, override multiple other inhibitory effects of the APR on RCT [22-24]. While it is established that MyD88-dependent inflammation contributes to atherosclerosis by recruiting monocytes to atheromas [79], future studies are warranted to discriminate whether macrophage and endothelial MyD88 serve distinct roles, in particular whether the former may counteract atherosclerosis through promoting RCT under permissive conditions.