Liver X Receptors in Atherosclerosis and Inflammation

Abstract

Liver-X-receptors (LXRs) are cholesterol sensing nuclear receptors that are not only key regulators of lipid metabolism and transport, but they also suppress inflammatory signaling in macrophages through a unique mechanism of transrepression. In this brief review, we focus on the regulatory actions of LXR primarily in macrophages responding to a proatherogenic environment. LXR potentially interferes with atherosclerosis by two different agonist dependent signaling pathways. The first is through promoting reverse cholesterol transport (RCT) by directly activating genes of cellular cholesterol export. The second is through a general inhibitory action on pro-inflammatory genes where sumo-modified and agonist bound LXR recruits negative co-regulatory proteins to NF-κB at immune response gene promoters through protein-protein interactions. The anti-inflammatory actions of LXR may be a direct response to the pro-inflammatory actions recently proposed for cholesterol on inflammasome activity in the vessel wall.

Introduction

The peroxisome proliferatior-activated receptor gamma (PPARγ) and liver X receptor (LXR) nuclear receptors play important roles in both lipid metabolism and inflammation ¹. The endogenous agonists for PPARγ and LXR are fatty acids ² and oxysterols, ³ respectively and there is a large body of literature that has established both of these receptors as fundamental regulators of lipid metabolism. Two other reviews in this series focus on PPARs ⁴ (and Plutzky review). The focus here is on LXR.

When sub-classified by degree of similarity within the larger nuclear receptor superfamily, the proteins encoded from the two LXR genes, LXRα (NR1H3) and LXRβ (NR1H2), are placed into the nuclear receptor-like subfamily 1 ⁵. The human LXRα gene is located on chromosome 11p11.2, while the LXRβ gene is located on chromosome 19q13.3. The expression patterns for LXRα and LXRβ mRNA vary significantly from tissue to tissue ⁶. LXRα is more restricted and is mainly expressed in intestine, fat tissue, macrophage, kidney, lung, and in liver where it is most abundant. On the other hand, LXRβ is more ubiquitous and expression has been observed in most cell types and tissue systems examined. At the protein level, the DNA and ligand binding domains of LXRα and LXRβ are highly similar (>75% identity), they both require heterodimerization with an retinoid X receptor (RXR) partner for DNA binding activity and the ectopically expressed proteins behave similarly in many assay systems ⁷. Thus, there is likely a significant level of redundancy at the molecular level suggesting that distinct functional roles may mostly be due to differences in patterns and levels of expression of the receptors than any inherent differences between the receptor proteins themselves. However, a recent report suggests that the two LXRs may play slightly different roles in regulating gene expression and atherogenesis ⁸.

LXRα in cholesterol metabolism

An important physiological role for LXRα in cholesterol metabolism was revealed when normal appearing chow-fed LXRα^−/− mice were challenged with a high cholesterol diet ⁹. These mice developed severe hepatic pathology due to toxic accumulation of cholesterol from a failure in the feed-forward mechanism that normally increases bile acid production from cholesterol ⁹. Oxysterols were identified as endogenous LXR agonists ³ and over the last several years, models for how LXR functions in a signaling pathway involving hepatic/intestinal communication and several nuclear receptors that together modulate flux through the cholesterol/bile acid synthetic pathway have been developed ^{10, 11}.

The mechanism assumes that elevated hepatic cholesterol levels are sensed by LXR, which then activates expression of the gene encoding cholesterol 7α-hydroxylase (CYP7A1), the rate-limiting enzyme in the neutral bile acid synthetic pathway in the mouse liver. This provides a metabolite based feed-forward mechanism that delivers cholesterol for conversion into bile acids to prevent cholesterol over-accumulation ^{12, 13}. CYP7A1 is a regulatory hub and is also subject to feedback regulation by bile acids and it also integrates physiologic input from several other signaling pathways that influence whole body metabolism ¹¹. Interestingly, the feed-forward mechanism for CYP7A1 regulation by LXR is not conserved in humans ¹⁴, which emphasizes the limitation of rodent studies in this system as predictors of human metabolism.

Comparing the effects of treating wild-type (Wt) and LXR^−/− mice with synthetic LXR agonists has revealed major insights into LXR action in vivo. Lipid profiling showed that LXR agonist treatment results in severe hypertriglyceridemia and hepatic steatosis in Wt but not LXR^−/− mice ^{6, 15}. When combined with additional animal studies and complementary cell culture and in vitro mechanistic analyses, a fundamental role for LXRα in activating hepatic lipogenesis was documented. In this mechanism, LXR acts indirectly by binding to the promoter and activating expression of the sterol regulatory element binding protein-1c (SREBP-1c) gene ^{6, 15}. LXR also directly activates expression of the gene encoding carbohydrate response-element binding protein (ChREBP), which is the other major transcriptional activator of lipogenesis in the liver ¹⁶. Thus, LXR increases levels of SREBP-1c and ChREBP proteins and together they directly activate most, if not all, of the genes required for hepatic lipogenesis and triglyceride secretion.

Additional studies demonstrated that LXR agonist still activated several lipogenic genes in the liver of SREBP-1c knockout mice ¹⁷ and LXR was also shown to directly activate key lipogenic genes directly ⁷. Why there would be a need for LXR to activate lipogenic genes directly as well as indirectly through SREBP-1c is not entirely clear. A subset of oxysterols bind the ER resident Insig protein and could inhibit the proteolytic maturation of SREBP-1 ¹⁸. This would limit the effectiveness of the SREBP-1c pathway and direct activation by LXR would be required to stimulate lipogenesis.

In contrast to CYP7A1 regulation, this pathway appears to be conserved in humans. The functional significance of why a cholesterol sensor should increase hepatic fatty acid biosynthesis is not totally clear and does not appear to occur in other normal cells where LXRα is expressed. However, a reasonable scenario suggests that activation of hepatic lipogenesis would ensure adequate fatty acids for conversion of the otherwise toxic (and possibly pro-inflammatory) free sterol to sterol-ester.

LXRs and Atherosclerosis

Even though activation of LXR signaling can have adverse effects on serum and hepatic lipids in mice⁷, studies have demonstrated that chronic administration of LXR agonists dramatically decrease lesion formation in both low density lipoprotein receptor (LDLR) and apolipoprotein E (apoE) knockout-mediated atherosclerosis mouse models ¹⁹. Because LXRs are expressed in several tissues and LXR agonist administration results in elevated serum and hepatic lipids as well as decreased cholesterol absorption, simple comparative studies in Wt vs. LXR knockout mice could not directly address the importance of individual tissues in the anti-atherogenic response. To address this issue, LXR knockout and agonist feeding studies have been evaluated and when taken in total, they suggest that LXR expressed in macrophages is key to the reduction in atherosclerosis ²⁰. The most direct experiments demonstrating this come from combinatorial transplantation studies using bone marrow from Wt or LXRα/β double knockout mice to repopulate bone marrow from irradiated LDLR and apoE knockout mice ^{21, 22}. These studies indicate that disease protection is provided mainly by LXR activity in bone marrow-derived cell lineages and also provide strong evidence that the effects are at least somewhat independent of LXR action in other tissues of the body; most notably liver and intestine. However, some studies have demonstrated that whereas bone marrow-derived LXR accounts for a majority of the lesion reducing effects through LXR agonist treatment, there was a greater reduction when LXRα was also expressed in other tissues ⁸.

Macrophages develop from the bone marrow and are known to play key roles in lipid metabolism and atherosclerosis ²³. LXR agonists increase reverse cholesterol transport (RCT) from macrophages by increasing expression of macrophage apoE and cholesterol efflux transporters ABCA1 and ABCG1. This is likely an important part of the mechanism for LXR dependent protection from atherosclerosis as these effects are not observed in agonist treated LXR knockout mice or their isolated macrophages ²². Excess accumulation of cholesterol within macrophages at sites of atherosclerotic lesions converts them into foam cells and accounts for the major fraction of lesion deposited cholesterol ²⁴. Thus by stimulating reverse cholesterol transport, LXR reduces foam cell formation and lesion cholesterol content directly.

Cholesterol crystals within macrophages have been recently proposed as an initiator of a pro-inflammatory signaling response at developing atherosclerotic lesions ^{25, 26} and it is well documented that inflammation and inflammatory signaling plays a significant role in the pathology of atherosclerosis²⁷. Similar to the mechanism whereby uric acid crystals increase inflammation associated with gout ²⁸, crystalline cholesterol can also increase NACHT, LRR and PYD domains-containing protein 3 (Nlrp3) inflammasome dependent activation of Caspase-1 (Fig. 1). Caspase-1 then converts pro IL-1β and pro IL-18 to their active cytokine forms and stimulates an NF-κB dependent pro-inflammatory cascade. These observations suggest that excess intracellular cholesterol may also contribute to severity of atherosclerosis by directly promoting inflammation. Interestingly, LXR not only increases expression of genes for RCT in macrophages but also independently attenuates the inflammatory response ²⁹. This suggests LXR agonists probably amplify a natural inhibitory response to intracellular cholesterol. In this unusual role, LXR is not recruited to LXR response elements within pro-inflammatory gene promoters but is recruited through a “transrepression” mechanism where sumo-modified and agonist bound LXR forms a complex with transcriptional co-repressors nuclear receptor co-repressor (N-CoR) and/or silencing mediator for retinoid and thyroid hormone receptors (SMRT) ^{30, 31}. The protein complex interacts with NF-κB, which targets the entire complex to promoters through NF-κB binding sites. As a result, the recruitment of the co-repressor activities dampens the NF-κB activation response at inflammatory gene promoters (Fig. 1). It is intriguing that intracellular sterols might initiate an inflammatory response as well as activate LXR. Thus, the anti-inflammatory effect of LXR action may have evolved as a response to the pro-inflammatory actions of intracellular sterols.

Figure 1.

Macrophages also initiate an inflammatory response upon bacterial infection and it is well-documented that infection by Chlamydia pneumoniae is associated with enhanced foam cell formation and atherosclerosis development which are reduced in animal models that lack toll-like receptor (TLR) signaling pathway components ^{32, 33}. This suggested that an inflammatory response to infection by this pathogen contributes to its effects on atherosclerosis. Naiki et al. ³³ also showed that LXR agonist treatment reduced foam cell formation enhanced by C. pneumoniae infection as well. Thus, it is likely that LXR has a general role in limiting inflammation linked to atherosclerosis and probably other diseases and LXR agonists could be therapeutic for combating inflammation regardless of the initiating insult.

Thus, LXR interferes with the pathology of atherosclerosis by two complementary agonist dependent signaling pathways in macrophages. One is through directly activating genes to promote RCT and limit cholesterol deposition and the other is through inhibiting pro-inflammatory gene expression, which reduces lesion-associated inflammation.

In contrast to the well-studied role of LXR signaling in macrophages, how LXR action in other cells that may directly influence atherosclerosis such as endothelial and smooth muscle cells (SMCs) is less well studied. Smooth muscle cells play an essential role in blood vessel integrity and rhythmic contracture. SMCs also influence atherosclerosis progression by protecting mature plaques from rupture ³⁴. LXRβ is expressed in vascular SMCs and LXR agonist treatment prevents neointima formation following vascular injury, which is consistent with a role in decreased SMC proliferation as well ³⁵. There is also evidence that genes involved in RCT are also upregulated in smooth muscle so they may help limit cholesterol deposition during plaque development as well ³⁶.

LXR target gene expression in endothelial cells is increased when they are exposed to high laminar flow in culture; conditions that mimic normal high volume arterial blood flow. Whereas LXR expression is significantly lower in vessel regions subjected to turbulent blood flow such as in the aortic arch ³⁷. These results suggest that in endothelial cells, LXR may respond to changes in blood flow to possibly influence RCT and reduce inflammation similar to LXR activity in macrophages. This would also have beneficial effects in limiting atherogenesis as well.

LXRs as regulators of the innate immune response

The role of LXR signaling in inflammation suggested it might also be a more general modulator of the innate immune system responses to pathogen engagement. Transcription profiling of LXR null and Wt mice subjected to bacterial challenge by Listeria monocytogenes identified AIM, apoptotic inhibitor of macrophages (also known as SPα, scavenger receptor cysteine-rich repeat protein, or API6, apoptosis inhibitor 6) as a target of LXR signaling ³⁸. Mechanistic studies documented there is a functional LXR binding site located in the AIM gene distal 5’ flanking region. AIM functions to prevent cells from undergoing apoptosis ³⁹ and increases survival of infected cells after bacterial challenge ³⁸. It was also demonstrated that SPα increased macrophage cell survival and decreased bacterial load after infection by L. monocytogenes and other pathogenic bacteria in an LXR-dependent manner ³⁸. Interestingly, in a follow-up study using AIM null/LDLR null double knockout mice, expression of AIM was shown to contribute to macrophage survival in the atherosclerotic plaque, which may also contribute to the protective effects of LXR signaling that limit atherosclerosis ⁴⁰. Paradoxically, AIM knockout mice were protected from developing atherosclerosis presumably due to increased apoptosis of macrophages during the early stages of plaque development. Thus, the mechanisms for LXR signaling in development of atherosclerosis may be complex with some aspects being beneficial while others may complicate the disease process.

LXR null macrophages have also been used to demonstrate that LXR plays a key role in phagocytosis of apoptotic thymocytes, a role that couples innate with adaptive immune responses ⁴¹ and would be predicted to be beneficial in limiting plaque depostion in advanced atherosclerotic lesions ⁴². The Mer receptor tyrosine kinase was identified as a candidate LXR gene target involved in this response. Mer is one of a group of molecules that facilitate phagocytosis by interacting with cell surface proteins on apoptotic cells. A-Gonzalez et al. ⁴¹ showed that LXR binds and induces Mer expression in a feed-forward mechanism to enhance apoptotic T cell clearance and a deficiency of either LXR or Mer reduced this effect. Interestingly, LXR and Mer are specifically required for the phagocytosis of apoptotic cells and not other phagocytic responses that occur in response to other initiating events. The phagocytosis of apoptotic thymocytes is also known to be associated with the inhibition of inflammatory signaling. This response was also reduced in LXR null macrophages suggesting that the suppression of inflammatory signaling during phagocytosis of apoptotic cells could be an additional physiologic context for the transrepressive effects of LXR on inflammation.

Translational Perspectives

Studies in mice using LXR knockout animals and synthetic LXR agonist treatment have provided a wealth of information on the roles that LXRs play in coordinating lipid metabolism with other physiological processes. This has established LXR as a key factor linking mammalian lipid metabolism with the innate immune response.

How well the results from mouse studies will be predictive of potentially targeting LXR pathways in humans for therapy is not clear. The mouse models of atherosclerosis are limited and are not useful models for more advanced human atherosclerotic lesions. There is evidence that LXRs not only prevent but also may reverse plaque development so that targeting early lesions in humans with LXR intervention may be more likely to be effective. Many of the LXR targets in the mouse are also LXR targets in humans and play roles in RCT, which would prove beneficial for LXR based therapy. However, other LXR target genes such as AIM appear to play a more complicated role in atherosclerosis by protecting macrophages from apoptosis at lesion sites. This may be advantageous later in plaque development but paradoxically AIM knockout mice have reduced lesion development. This is likely because when AIM knockout macrophages are recruited to early lesion sites, they undergo apoptosis and fail to nucleate the growth of a developing plaque ⁴⁰.

Key molecular targets of LXR are also different in humans versus mice. As mentioned earlier, CYP7A1 is regulated by LXR in mice but not humans. The small heterodimer partner (SHP) nuclear receptor contains a dimerization domain but lacks the canonical Zn-finger DNA binding domain. SHP dimerizes with other nuclear receptors but inhibits their activity to help mediate bile acid feedback regulation. SHP is an LXR target in humans but not in mice ¹⁴ and the human LXRα gene is autoregulated in humans but not in mice ⁴³. There are also probably additional examples of species differences in LXR action as well. Thus, because of significant molecular genetic differences between humans and mice and the limitations of current mouse models available for comparative atherosclerosis studies, it is challenging to predict how best to therapeutically target the LXR pathway for treatment in humans.

Finally, because of its role in stimulating hepatic lipogenesis, LXR agonist treatment results in significant hepatic steatosis in mice ⁷. This could significantly contribute to Metabolic Syndrome abnormalities as well as progress to steatohepatitis with severe specific liver damage as well ⁴⁴. Thus, despite the positive effects on atherosclerosis in mouse models, targeting LXR for therapeutic purposes in humans must overcome this significant obstacle. With this in mind, recent studies suggest that at least some synthetic LXR agonists may retain the beneficial effects on atherosclerosis without the effects on liver triglyceride accumulation ⁴⁵. There is also more recent evidence that selective LXR agonists may also have similar effects in non-human primates ⁴⁶. Thus, there is promising evidence that it will be possible to produce selective LXR modulators that retain the beneficial effects on atherosclerosis without the complicating effect on liver triglyceride metabolism.

Non-standard Abbreviations and Acronyms

LXR

liver-X-receptor

PPAR

peroxisome proliferator activated receptor

CYP7A1

cholesterol 7α-hydroxylase

LDL

low density lipoprotein

ApoE

apolipoprotein E

RCT

reverse cholesterol transport

Wt

wild type

SREBP

sterol regulatory element binding protein

ChREBP

carbohydrate response element binding protein

NF-kB

nuclear factor kappa B

SHP

small heterodimer partner

NACHT

NAIP, CIITA, HET-E and TP1 containing

LRR

leucine rich repeat region

PYD

pyrin domain

Nlr

NACHT, LRR, PYD containing receptor

ABCA1

ATP-binding cassette, sub- family A, member 1, ATP-binding cassette, sub- family G, member 1 NACHT, LRR and PYD containing

AIM

apoptosis inhibitor in macrophages

N-CoR

nuclear receptor co-repressor

SMC

smooth muscle cells

SMRT

silencing mediator for retinoid and thyroid hormone receptors