The Role of Endoplasmic Reticulum Stress in the Progression of Atherosclerosis

Abstract

Prolonged activation of the endoplasmic reticulum (ER) stress pathway known as the Unfolded Protein Response (UPR) can lead to cell pathology and subsequent tissue dysfunction. There is now ample evidence that the UPR is chronically activated in atherosclerotic lesional cells, particularly advanced lesional macrophages and endothelial cells. The stressors in advanced lesions that can lead to prolonged activation of the UPR include oxidative stress, oxysterols, and high levels of intracellular cholesterol and saturated fatty acids. Importantly, these arterial-wall stressors may be especially prominent in the settings of obesity, insulin resistance, and diabetes, all of which promote the clinical progression of atherosclerosis. In the case of macrophages, prolonged ER stress triggers apoptosis, which in turn leads to plaque necrosis if the apoptotic cells are not rapidly cleared. ER stress-induced endothelial cell apoptosis may also contribute to plaque progression. Another potentially important pro-atherogenic effect of prolonged ER stress is activation of inflammatory pathways in macrophages and, perhaps in response to athero-prone shear stress, endothelial cells. While exciting work over the last decade has begun to shed light on the mechanisms and in vivo relevance of ER stress-driven atherosclerosis, much more work is needed to fully understand this area and to enable an informed approach to therapeutic translation.

Introduction: Atherogenesis and Endoplasmic Reticulum Stress

Brief Overview Atherogenesis and Plaque Progression

Atherosclerotic vascular disease is the leading cause of death in the industrialized world, soon to be world-wide.^{1, 2} Atherogenesis begins with the subendothelial retention of apolipoprotein B-containing lipoproteins in focal areas of the arterial subendothelium.^{3, 4} These lipoproteins, perhaps after oxidative or other types of modification, then trigger a series of maladaptive inflammatory responses.^3-7 Key among these responses is the attraction of blood-borne monocytes to activated endothelial cells overlying areas of lipoprotein retention, followed by monocyte differentiation into macrophages. The macrophages internalize the retained lipoproteins, leading to foam cell formation, and the macrophages become activated and inflammatory.^6-9 Dendritic cells, T cells, and mast cells, B cells, and possibly neutrophils also enter the growing lesion and contribute to the inflammatory response to retained lipoproteins. As lesions develop, the inflammatory response amplifies and does not resolve, which is likely due to the continuing and even accelerated process of lipoprotein retention.^{3, 10, 11} As such, the subendothelial space expands with cells and extracellular matrix.

In the vast majority of lesions, several processes prevent the key clinical consequence of atherosclerosis, namely, acute thrombotic vascular occlusion.¹² Among these are outward remodeling of the vessel wall, which maintains the patency of the arterial lumen; phagocytic clearance (“efferocytosis”) of dead cells, mostly apoptotic macrophages, which prevents plaque necrosis; and scar formation by collagen-producing intimal smooth muscle cells (myofibroblasts), which helps defend against matrix protease-mediated erosion or rupture of the intima into the lumen.^13-16 However, in a small percentage of lesions, one or more of these processes fail, leading sequentially to plaque erosion or rupture, exposure of the blood to coagulation and thrombotic factors in necrotic lesions, and acute lumenal thrombosis.^{13, 17, 18} Examples of cellular processes that lead to this failure are increased lesional macrophage apoptosis, defective efferocytosis, and death of collagen-producing intimal smooth muscle cells.^13-16

As is obvious from the above summary, a complex interplay among cell biological and physiologic factors promote the initiation and progression of atherosclerosis. At one level, non-arterial wall systemic factors, especially those leading to elevated levels of circulating apoB-lipoproteins, play essential roles. At the level of the arterial wall, exogenous and endogenous factors contribute to atherogenesis and plaque progression by promoting (a) apoB lipoprotein retention and modification; (b) activation of endothelial cells; (c) entry and activation of inflammatory cells, notably macrophages; (d) entry and proliferation of intimal smooth muscle cells; (e) regulation of collagen biosynthesis and collagen turnover; (f) alterations in intimal cell death and in clearance of the dead cells; and (g) activation of coagulation factors and platelets. As will become evident in this review, a number of these processes can be influenced by prolonged ER stress.

Physiologic and Pathophysiologic Endoplasmic Reticulum Stress

Over the last decade, ER stress has emerged as a factor that is relevant to a number of systemic and arterial-wall factors that promote atherosclerosis. As reviewed in this Series, ER stress represents a response by cells to transient or prolonged perturbations in ER function, especially function related to protein synthesis, calcium regulation, and intracellular redox potential.^19-21 ER stress signaling, often referred to as the Unfolded Protein Response, is triggered by three upstream proteins, IRE1, ATF6, and PERK. IRE1, by promoting the expression of XBP-1, and ATF6 play key roles in chaperone production, which helps relieve physiologic and pathophysiologic imbalances between nascent proteins and the chaperones required to ensure proper protein folding and assembly. The XBP-1 pathway also promotes the degradation of misfolded proteins. Other IRE1 pathways can lead to apoptosis (below) and mRNA degradation.²² PERK, by phosphorylating eIF2α, temporally slows protein translation, which allows perturbations in protein translation to be corrected in an optimal manner. PERK, through phospho-eIF2α-mediated translational up-regulation of ATF4, also leads to the induction of CHOP (GADD153; gene name Ddit3), which participates in various corrective functions during transient ER stress. However, the fact that CHOP-deficient mice develop normally and exhibit good heath,²³ suggests that many of CHOP's salutary functions may be redundant with other ER stress effectors in the laboratory animal setting. As described below, prolonged expression of CHOP is a potent inducer of apoptosis.

Despite the essential beneficial functions of the UPR during transient ER stress, pathologically chronic ER stress often leads to tissue dysfunction and disease.¹⁹ ER stress can be prolonged by chronic disturbances in protein folding, oxidative stress, and other processes that lead to sustained ER dysfunction. In one scenario relevant to atherosclerosis, chronic ER stress affects systemic risk factors at the level of hepatic lipid metabolism and pancreatic beta cell function, particularly in the settings of obesity, insulin resistance, and diabetes. This topic will not be covered here, and the reader is referred to a number of excellent reviews in this area.^24-28 Rather, this review will focus on the effects of ER stress on plaque cells, particularly macrophages and endothelial cells. Relevant to this topic is the general principle that two segments of the UPR can trigger pathologic cell death. Chronic activation of IRE1 promotes apoptosis through the signal transducers ASK1 and JNK and by altering the balance and activity of Bcl family members.²⁹ Prolonged elevation of CHOP triggers apoptosis through effects on intracellular calcium metabolism and by alterations in Bcl family members.^{23, 29-32} There is increasing in vitro and in vivo evidence that ER stress-induced apoptosis of intimal cells, notably macrophages, plays an important role in atherosclerotic plaque progression. In addition, chronic ER stress can adversely affect endothelial cell biology. This review will summarize the latest findings related to these topics and discuss therapeutic implications and future directions.

Pro-Atherogenic Effects of ER Stress in Endothelial and Smooth Muscle Cells Endothelial Cells

As reviewed above, alterations in endothelial function and protein expression play important roles in attracting inflammatory cells during atherogenesis. Lusis and colleagues³³ treated cultured human aortic ECs with the UPR activator tunicamycin and found induction of interleukin 8, interleukin 6, monocyte chemoattractant protein 1 (MCP1), and the chemokine CXC motif ligand 3 (CXCL3). The expression of these molecules could be blocked by gene silencing of ATF4 and/or XBP1. UPR activation in ECs was also induced by oxidized 1-palmitoyl-2-arachidonyl-sn-3-glycero-phosphorylcholine (oxPAPC), an oxidized phospholipid found in advanced atherosclerotic lesions. A possible in vivo link was suggested by the finding that oxidized phospholipid-rich areas of human lesional endothelium showed evidence of UPR activation.³³ Interestingly, the ability of oxPAPC to activate the UPR in human aortic ECs differed significantly among individual donors, and this phenotypic variation was associated with genetic variation in a locus affecting action of USP16, a histone H2A deubiquitinase.³⁴

Atherosclerosis occurs at sites of disturbed blood flow, and biomechanical transduction of disturbed flow through the endothelial cells (ECs) at these sites may be a key process in promoting pro-atherogenic EC alterations in response to retained lipoproteins.³⁵ In this context, Davies group found that the ER stress transducers IRE1α, ATF6α, and XBP1 were increased in ECs from athero-susceptible regions in normal swine aorta.³⁶ In cultured ECs, athero-prone flow (shear stress) increased the expression of the UPR effector GRP78 via a pathway involving the mitogen-activated protein kinase p38 and the integrin α2β1.³⁷ There is also some evidence that endothelial ER stress in pre-lesional areas is exacerbated in the setting of diabetes, perhaps due to hyperglycemia-driven accumulation of glucosamine in the cells.³⁸ Xu and colleagues³⁹ also found that endothelial XBP1 expression was increased by disturbed flow in cultured ECs. This finding was correlated with endothelial proliferation, which may reflect a protective response. However, when XBP1 was overexpressed in ECs, a key endothelial junctional adhesion protein, VE-cadherin, was decreased and endothelial apoptosis ensued. In Apoe^-/- mice, endothelial XBP1 expression correlated with areas of lesion severity, and when XBP1 was overexpressed in lesions, atherosclerosis was accelerated. One interpretation of these findings is that transient, limited XBP1 induction protects ECs against disturbed flow, while chronically elevated levels of XBP1 damages the cells and thereby promotes atherosclerosis. However, it is also possible that other UPR effectors that are induced in athero-prone endothelium may play a pro-atherogenic role. For example, there is evidence that pre-lesional endothelial cells in these sites are primed through changes in gene expression for subsequent LPS- or high fat diet-induced activation of the NF-kB pathway^{40, 41}. Given links between the UPR and NF-κB activation,⁴² it is possible that the early ER stress response may be involved in these changes. Further mechanistic investigation and, most importantly, endothelial-specific XBP1 targeting studies in mouse models of atherosclerosis will be needed to support this idea.

Other possible athero-relevant inducers of pathologic ER stress in ECs include homocysteine and modified forms of LDL. Homocysteine induces ER stress-induced apoptosis in cultured human umbilical vein ECs. While one group showed that disabling mutants of IRE1 can block apoptosis in this model,⁴³ another group showed that over-expression of a PERK pathway effector called T cell death–associated gene 51 (TDAG51), which is up-regulated in homocysteine-treated ECs, can trigger a form of detachment-mediated cell death.^{44, 45} Hyperhomocysteinemia is associated with atherothrombotic vascular disease in humans and accelerated atherosclerosis and heightened ER stress in mouse models of atherosclerosis.^{46, 47} When hyperhomocysteinemic, fat-fed rabbits were treated with taurine to reduce homocysteine levels in the plasma, coronary artery atherosclerosis, endothelial apoptosis, and endothelial CHOP were decreased.⁴⁸ However, the pro-atherogenic effect of homocysteine in animal models likely involves other lesional cell types in addition to ECs,⁴⁷ and how much of the proatherogenic effect of hyperhomocysteinemia is attributable to prolonged ER stress awaits genetic causation studies in vivo.

Various forms of LDL modification, such as through oxidation, glycation, or lipid hydrolysis, may occur during the course of atherogenesis.⁷ Oxidized and glycated LDL is an inducer of the UPR in cultured ECs through a mechanism involving disturbed ER calcium metabolism.⁴⁹ In particular, the oxidized-glycated LDL induces oxidative stress in ECs, which in turn inhibits the ER calcium pump, sarcoplasmic/endoplasmic reticulum calcium-dependent ATPase (SERCA). Another link between SERCA oxidation and ER stress was revealed by a mouse model lacking an endogenous inhibitor of SERCA oxidation called AMP kinase α2 (AMPKα2). The authors showed that Ldlr^-/- mice lacking this protein had increased lesional ER stress and atherosclerosis.⁵⁰ Finally, LDL hydrolyzed by secretory phospholipase A₂ induces the UPR in cultured ECs, which in turn stimulates inflammatory cytokine production through activation of p38.⁵¹

Smooth muscle cells

Less is known about the potential atherogenic role of ER stress in vascular smooth muscle cells (VSMCs) than in ECs or macrophages. In theory, ER stress-induced SMC apoptosis could lead to decreased collagen production and thus thinning of the protective collagen cap in advanced lesions.⁵² A proof-of-concept study using an ER stress-inducing drug called bortezomib was used to test the hypothesis that UPR-induced apoptosis in VSMCs could trigger death of the cells and thereby adversely affect collagen production in lesions. In this study, bortezomib, which is a proteasome inhibitor, was shown to cause ER stress and apoptosis in cultured VSMCs but not in macrophages.⁵³ In aortic explants from Apoe^-/- mice, bortezomib induced the UPR, including CHOP, and cell death was observed in regions that were populated with SMCs. In fat-fed Apoe^-/-mice, bortezomib was associated with a marked decreased in VSMCs, a modest decrease in collagen, and a marked increase in necrotic cores.⁵³ Whether or not these effects were actually caused by UPR induction remains to be shown.

Possible endogenous inducers of ER stress-induced apoptosis in VSMCs include 7-ketocholesterol, unesterified cholesterol, homocysteine, and glucosamine. Exposure of human coronary artery-derived VSMCs to 7-ketocholesterol, which is found in rupture-prone shoulder regions of thin-capped human coronary atheromata, led to the induction of CHOP and UPR chaperones.⁵⁴ Induction of these effectors was associated with the accumulation of reactive oxygen species and could be blocked by the anti-oxidant N-acetylcysteine. Most importantly, 7-ketocholesterol induced apoptosis in the VSMCs, and silencing of CHOP through siRNA blocked apoptosis.⁵⁴ Unesterified cholesterol induces ER stress-induced apoptosis in cultured SMCs,⁵⁵, as it does in macrophages,⁵⁶ but there is little evidence that SMCs in atherosclerotic lesions accumulate large amounts of unesterified cholesterol. Evidence that the UPR inducer homocysteine could affect SMCs comes from the study described in the previous section examining hyperhomocysteinemic Apoe^-/- mice.⁴⁷ The authors found evidence of UPR induction in intimal SMCs in these mice, which may be mediated through changes in ER calcium release.⁵⁷ Moreover, homocysteine activates sterol response element binding protein-2 (SREBP-2) in cultured VSMCs, leading to an increase in intracellular lipid accumulation.^{58, 59} Finally, intracellular glucosamine, which is elevated in the setting of diabetes, has been shown to induce the UPR effector glucose-regulated protein 78 (Grp78) in cultured human aortic SMCs, but whether this is a protective or maladaptive response was not addressed.⁶⁰ Future studies are needed to determine more precisely the extent to which the UPR is activated in lesional SMCs during lesion development and how ER-stressed SMCs might affect atherosclerosis progression.

Pro-Atherogenic Effects of ER Stress in Macrophages

Evidence and Roles for ER Stress in Advanced Lesional Macrophages

Initial work by Feng et al.³⁰ and then Austin and colleagues⁶¹ demonstrated that macrophages are prominent among atherosclerotic lesional cells undergoing ER stress in general, and expressing CHOP in particular, during the progression of atherosclerosis in chow-fed or Western diet-fed Apoe^-/- mice. These data provided evidence that CHOP expression increases as lesions progress, a concept that was later supported and expanded to human atherosclerosis in an important study by Myoishi et al.⁶² These investigators examined human coronary artery lesions in autopsy samples and fresh human carotid endarterectomy specimens for lesion stage, UPR markers, and apoptosis, as measured by the terminal deoxynucleotidyl transferase-mediated dUTP nick end-labeling (TUNEL) assay. In both sets of human lesions, the authors found a striking relationship among advanced lesion stage, CHOP expression, and lesional apoptosis such that only advanced, “vulnerable” plaques showed evidence of robust CHOP expression and apoptosis.

The human data support a working model developed by the author's laboratory and others over the last decade to explain how ER stress-induced macrophage apoptosis is a key event in the generation of necrotic cores and thus clinically dangerous plaques.^{11, 15, 63} In early lesions, macrophage apoptosis is difficult to detect, because the apoptotic cells are rapidly cleared by neighboring macrophages through a phagocytic process called “efferocytosis.”^{15, 63} Efferocytosis engulfs and safely destroys apoptotic cells before they become necrotic, and the process also triggers a distinct anti-inflammatory response involving cytokines such as interleukin-10 and transforming growth factor-β.⁶⁴ Indeed, genetic manipulations that promote macrophage apoptosis in early lesions actually suppress lesion cellularity and inhibit plaque progression through this efferocytic mechanism, and vice versa.^{65, 66} In advanced lesions, however, the situation is very different. First, apoptosis is almost certainly increased and is triggered by mechanisms unique to the milieu of the advanced plaque, notably chronic ER stress.⁶⁷ Second, efferocytosis is not as efficient as in normal physiology, and so the apoptotic cells become secondarily necrotic, and the aforementioned anti-inflammatory response does not occur.⁶⁸ The cumulative result is the formation of inflammatory necrotic cores and, potentially, plaque disruption and acute lumenal thrombosis.^{11, 15, 63}

Experimental Models of ER Stress in Cultured Macrophages

Mechanistic data with cultured macrophages and molecular genetic causation data in vivo using mouse models of atherosclerosis have revealed a critical role for the UPR in advanced lesional macrophage death and plaque necrosis.^{30, 69-71} The cell culture studies have used experimental proof-of-concept inducers of ER stress as well as stressors that are likely to be relevant to advanced lesions. Examples of the former type include the protein glycosylation inhibitor tunicamycin⁴² and the SERCA inhibitor thapsigargin.⁷² The athero-relevant inducers are chosen from among the molecules and processes in advanced lesions that can lead to prolonged UPR activation, such as high levels of intracellular unesterified cholesterol, oxysterols, oxidant stress, hypoxia, and peroxynitrite.⁷³ One type of cultured macrophage model uses robust ER stress, which can be achieved with thapsigargin or 7-ketocholesterol, an ER stressor that is the most abundant oxysterol in human atherosclerotic lesions.³¹

Another type of model takes into account the possibility of more subtle ER stress in vivo. In this model, macrophage apoptosis is induced by the combination of a low-dose ER stressor and an athero-relevant “second hit,” each of which are unable to induce apoptosis by themselves. Examples include low-dose 7-ketocholesterol, thapsigargin, or the peroxynitrite donor SIN-1 as the ER stressor and combinatorial pattern recognition receptor (PRR) activation as the second hit.^74-76 Combinatorial PRR activation triggers apoptosis in the setting of subtle ER stress by both enhancing apoptotic pathways, including a pro-apoptotic NADPH oxidase-reactive oxygen species (ROS) pathway, and by suppressing cell-survival pathways that are compensatorily triggered by ER stress (see below).^{72, 75, 76} This pathway of macrophage apoptosis may have evolved as a host defense mechanism against disease-causing organisms that require living macrophages to survive.⁷⁷ One example of combinatorial PRR activation that can trigger apoptosis in macrophages undergoing subtle ER stress are activators of the type A scavenger receptor (SRA) and toll-like receptor 4 (TLR4), such as oxidized LDL or acetylated LDL.⁷⁵ A second example are activators of CD36 and TLR2/6 heterodimer, such as oxidized phospholipids (oxPLs), oxidized LDL, or saturated fatty acids (SFAs).^{70, 76} Another variation of this model is a “combined two-hit” scenario in which macrophages unable to re-esterify lipoprotein-derived cholesterol are incubated with atherogenic lipoproteins. In this setting, the lipoproteins activate TLRs and scavenger receptors at the cell surface, and lipoprotein-derived unesterified cholesterol that accumulates in the ER membrane activates the UPR.^{74, 75, 78} For each of these models, there is evidence in vivo. For example, advanced lesional macrophage apoptosis and plaque necrosis are blocked in advanced murine atheromata that have genetically induced deletion or inhibition of scavenger receptors,⁷⁹ TLRs,⁷⁶ or cholesterol trafficking to the ER membrane.⁸⁰

Mechanisms of CHOP-Induced Apoptosis in ER-Stressed Macrophages

Using one or more of these cultured macrophage models, mechanistic studies have shown that a full apoptosis response requires the CHOP pathway upstream and activation of both the Fas and mitochondrial pathways of apoptosis downstream.^{30, 81, 82} Recent studies have provided evidence for specific molecular signaling pathways linking CHOP with the downstream apoptosis processes (Figure 1). In particular, CHOP promotes apoptosis primarily by stimulating a persistent release of calcium from ER stores.^{31, 32} The mechanism involves activation of IP3-receptor-mediated calcium release by the action of the CHOP transcriptional target ER oxidase 1α (ERO1α).³¹ Evidence suggests that ERO1α activates IP3R1 by hyper-oxidizing the ER lumen, leading to disulfide bond formation in a lumenal loop of IP3R1, which in turns increases IP3R1 calcium channel activity.^{83, 84} The persistent increase in cytoplasmic calcium activates calcium/calmodulin-dependent protein kinase II (CaMKII), which then activates a number of pro-apoptotic pathways, including those involving the death receptor Fas, the pro-apoptotic pathway mediated by release of apoptogenic factors from the mitochondria, a pro-apoptotic pathway involving signal transducer and activator of transcription-1 (STAT1), and NADPH oxidase-mediated ROS^{32, 76} (G. Li, C. Scull, L. Ozcan, and I. Tabas, submitted manuscript in revision, [2010]).” Most importantly, interruption of these events, most notably through gene targeting of CHOP, lessens advanced lesional macrophages apoptosis and plaque necrosis in both the Ldlr^-/- and Apoe^-/- models of atherosclerosis.^{69, 71} In vitro studies have suggested another role of prolonged CHOP induction in macrophages, namely, activation of inflammatory signaling pathways.⁸⁵ In particular, prolonged CHOP expression leads to induction of interleukin-6 through an ERK1/2 pathway.⁸⁵ Interestingly, CHOP-induced inflammation in macrophages can be stimulated by anti-HIV protease inhibitors, which are associated with accelerated atherothrombotic vascular disease.⁸⁶

Figure 1. Model of ER stress-induced apoptosis in advanced lesional macrophages.

Advanced lesions contain a number of molecules and processes that are known ER stressors. Activation of the UPR in macrophages leads to elevated levels of CHOP, which through a pathway involving the CHOP target ERO1 and IP3R-mediated calcium release, leads to elevation in cytosolic calcium (Ca²⁺_i) and activation of CaMKII. CaMKII triggers apoptosis execution pathways through a number of mechanisms, including NADPH oxidase activation and subsequent ROS generation. In vivo evidence suggests that this scenario is necessary but not sufficient by itself for apoptosis. Rather, apoptosis in ER-stressed macrophages may require a “second hit” consisting of combinatorial pattern recognition receptor (PRR) activation, and there are many factors in advanced lesions that can activate PRRs. The mechanism by which PRR activation tips the balance toward apoptosis in ER-stressed macrophages involves amplification of apoptosis pathways, such as NADPH oxidase-mediated ROS generation, and suppression of ER stress-induced compensatory cell survival pathways. See text for details. (Illustration Credit: Cosmocyte/Cameron Slayden)

CHOP deletion does not fully suppress apoptosis in ER-stressed macrophages, and so other mechanisms of ER stress-induced apoptosis involving the IRE1-JNK pathway and/or alterations in Bcl family members may also be involved (see Introduction). In this regard, Li et al.⁸⁷ showed that that IRE1α was activated (phosphorylated) in lesional extracts from Western-diet fed Apoe^-/- mice and that cholesterol-induced apoptosis was partially suppressed in cultured macrophages treated with Ire1a siRNA or with a JNK inhibitor. However, even here there may be a link with the CHOP pathway, because silencing of IRE1α in ER-stressed macrophages was associated with a decrease in CHOP expression.⁸⁷ In terms of Bcl family members, Bax levels are increased in the cholesterol model of ER stress-induced macrophage apoptosis.⁸² Moreover, deletion of Bcl2 in macrophages renders them more susceptible to ER stress-induced apoptosis in vitro and is associated with accelerated advanced lesional apoptosis and plaque necrosis in Western diet-fed Apoe^-/- mice.⁸⁸ Defining CHOP-independent pathways of ER stress-induced macrophage apoptosis in advanced atherosclerosis is an important goal for the future.

Another important consideration in the study of ER stress-induced macrophage apoptosis is the fact that the initial ER stress response is often accompanied by activation of cell-survival signaling pathways, which then become de-activated soon before apoptosis occurs. Examples of cell-survival modules that are activated and then eventually fail in macrophages exposed to prolonged ER stress include pathways mediated by interferon-β, Akt, NF-κB, p38α, ERK,^{72, 75, 89-92} and autophagy⁹³ (X. Liao, J. Sluimer, B. Levine, and I. Tabas, unpublished data, [2010]). Thus, understanding how these pathways are activated and, more importantly, why they eventually fail represents a critical area in this field.

Insulin Resistance, Saturated Fatty Acids, and Lipoprotein(a) as Inducers of ER Stress-Induced Macrophage Apoptosis

The concept that ER stress/CHOP-induced macrophage apoptosis is a critical step in necrotic plaque formation may have particular relevance to the rapidly growing epidemic of atherothrombotic vascular disease driven by insulin resistance and diabetes.^{92, 94} Indeed, atherosclerotic lesions in diabetic subjects are characterized by especially large necrotic cores, even when corrected for overall lesion size.⁹² In this context, recent studies have shown that a potent inducer of prolonged ER stress in macrophages is insulin resistance.⁹⁵ Macrophages have a functional insulin receptor signaling pathway, and down-regulation of insulin receptor signaling due to hyperinsulinemia promotes ER stress and apoptosis in these cells.⁹⁶ The mechanism involves processes that elevate cytosolic calcium through SERCA inhibition (see above).⁹⁵ Moreover, apoptosis is further exacerbated in the setting of macrophage insulin resistance through a mechanism involving suppression of the NF-κB cell survival pathway.⁹⁰ Most importantly, both advanced lesional macrophage apoptosis and plaque necrosis are increased in a model of macrophage insulin resistance in Western diet-fed Ldlr^-/- mice.⁹⁶

Obesity and insulin resistance may further promote ER stress in lesional macrophages through elevations of saturated fatty acids (SFAs).^{70, 76} Although the mechanism of SFA-induced ER stress is not fully known, SFAs lower the fluidity of the ER membrane bilayer, which is a known inducer of the UPR.^{78, 97} Recent work has suggested that an intracellular “lipid chaperone” called macrophage fatty acid-binding protein-4, also known as adipocyte fatty acid-binding protein aP2, mediates SFA-induced ER stress and apoptosis in macrophages.⁷⁰ Previous work had shown that Apoe^-/- mice deficient for aP2 are protected against atherosclerosis.⁹⁸ In the new study, macrophages lacking aP2 were shown to be protected from palmitate-induced ER stress and apoptosis, and this protection was associated with activation of the transcription factor LXRα. LXRα is an inducer of an enzyme that converts SFAs into mono-unsaturated fatty acids, which are much less potent inducers of ER stress. Thus, aP2, by preventing SFA desaturation, is a mediator of SFA-induced ER stress in macrophages. In the Western diet-fed Apoe^-/- model, aP2 deficiency resulted in decreased P-PERK, XBP-1, and apoptosis in macrophage-rich regions of atherosclerotic lesions. Whether aP2 deficiency protects against plaque necrosis, an important consequence of ER stress-induced macrophage apoptosis, has not yet been demonstrated. Furthermore, molecular genetic proof that the protective effects of aP2 deficiency in Western diet-fed Apoe^-/- mice are mediated by suppression of SFA-induced UPR activation remains to be demonstrated, although the role of the UPR was supported by data from a drug experiment in this study, as will be discussed below.

The ability of oxPLs to trigger apoptosis in ER-stressed macrophages (above) may provide a clue to a 60-year-old mystery in heart disease. Approximately 25-30% of Caucasians have elevated levels of a lipoprotein called lipoprotein(a) [Lp(a)], which is an LDL-like lipoprotein with a covalently bound kringle-containing protein called apolipoprotein(a).⁹⁹ Based on many clinical studies, including large epidemiological and human genetic studies, Lp(a) has been shown to be a potent, independent risk factor for advanced atherothrombotic vascular disease.^100-103 Nonetheless, the mechanisms linking Lp(a) with specific cellular events associated with advanced plaque progression and necrosis, the sine qua non of clinically relevant atherosclerotic disease, are not known. Important observations over the last decade have revealed that Lp(a) is a major carrier of oxPLs in human plasma,¹⁰⁴ which, based on the two-hit ER stress-PRR model described above, begged the question as to whether Lp(a) could trigger apoptosis in ER-stressed macrophages. Exciting new data have shown that Lp(a), but not native LDL, is a potent inducer of apoptosis in ER-stressed macrophages, but not unstressed macrophages.⁷⁶ As with oxPLs, Lp(a)-induced apoptosis depended on TLR2/6, CD36, and oxidative stress. In vivo support for the role of Lp(a) in advanced lesional macrophage death is provided by a transgenic rabbit model in which modest plasma levels of Lp(a) markedly promoted plaque necrosis without affecting en face lesion area.¹⁰⁵ Additional mechanistic and in vivo studies, including studies with human lesional material, will be needed to provide further support for this new concept.

Therapeutic Implications

Given the role of prolonged ER stress in a number of important diseases, notably those associated with advanced age and obesity, there has been increasing interest in therapeutic strategies focused on relieving ER stress, including for the purpose of preventing atherosclerosis.^{25, 106, 107} In this regard, compounds that seem to act as so-called “chemical chaperones” may be useful in decreasing the adverse effects of prolonged ER stress.^{106, 107} Among these are two compounds—4-phenyl butyric acid (PBA) and tauroursodeoxycholic acid (TUDCA)—that have been tested in animal studies for the aforementioned purpose and used in humans for a variety of disorders.¹⁰⁸

As mentioned in the Introduction, prolonged ER stress in the setting of obesity can lead to systemic atherosclerotic risk factors related to insulin resistance, and both compounds have been shown to have benefit in relieving insulin resistance in experimental mouse models of obesity.¹⁰⁸ In terms of direct effects on the arterial wall, the mouse study described above on SFA/aP2-induced ER stress in lesional macrophages showed that PBA relieved palmitic acid-induced ER stress and apoptosis in cultured macrophages and decreased lesion area and both lesional ER stress and apoptosis in Western diet-fed Apoe^-/- mice.⁷⁰ In a different setting, namely, diabetes-induced atherosclerosis in male hamsters, PBA failed to relieve ER stress or atherosclerosis.¹⁰⁹ TUDCA treatment was used in yet another model of ER stress in atherosclerosis—Western diet-fed Ampka2^-/-Ldlr^-/- mice (above)—and was found to have beneficial effects on both lesional ER stress and atherosclerosis lesion area.⁵¹ However, definitive proof that the anti-atherogenic mechanisms of PBA and TUDCA are linked to relief of ER stress per se is lacking.

Statins are the mainstay of preventative therapy for atherothrombotic vascular disease in humans.¹¹⁰ While their predominant and perhaps sole beneficial effect is through their ability to lower apoB lipoproteins in the circulation, thus decreasing arterial wall lipoprotein retention, the possibility of so-called “pleiotropic” protective effects by mechanisms other than through lowering plasma cholesterol is a topic of great interest in the field.¹¹⁰ One study showed a possible beneficial effect of statins in preventing SFA-induced ER stress in a human monocyte-macrophage cell line and in decreasing in ER stress in lesions of statin-treated Ldlr-/- mice.¹¹¹ However, much more mechanistic and in vivo work is needed to further substantiate this point. Furthermore, any effect of these drugs on lesional ER stress may be through the decrease in exposure of lesional cells, predominantly macrophages, to excess apoB lipoprotein-derived cholesterol.⁵⁶

A proposed strategy of preventative treatment that has not yet shown to be beneficial in humans—anti-oxidant therapy—is also relevant to a discussion of therapeutically targeting ER stress in lesions, because oxidative stress can be both a cause and a consequence of prolonged ER stress.¹¹² For example, the aforementioned example of ER stress being triggered by oxidation of the ER calcium pump SERCA can be relieved by the anti-oxidants apocynin and tempol (4-hydroxy-2,2,6,6-tetramethylpiperidine-N-oxyl or 4-hydroxy-tempo),⁵⁰ and tempol has been shown to decrease lesional ER stress and atherosclerosis in Western diet-fed Apoe^-/- and Apmka2^-/-Apoe^-/- mice (above).⁴⁹ Moreover, oxidized PLs, Lp(a), SFAs, and 7-ketocholesterol activate NADPH oxidase and induce oxidative stress in cultured macrophages⁷⁶ (G. Li, C. Scull, L. Ozcan, and I. Tabas, submitted manuscript in revision, [2010]). These processes are crucial for ER stress-induced apoptosis in macrophages and can be prevented by inhibitors of NADPH oxidase and by the anti-oxidant N-acetylcysteine⁷⁶ (G. Li, C. Scull, L. Ozcan, and I. Tabas, submitted manuscript in revision, [2010]). The major anti-oxidant that has been tested in humans is vitamin E, with most studies showing no beneficial effect.¹¹³ However, vitamin E may not target the specific mechanisms of oxidative stress associated with ER stress-induced perturbations in lesional cells.¹¹⁴ For example, the author's laboratory found that while N-acetylcysteine was effective in preventing ER stress-induced macrophage apoptosis, vitamin E was not (G. Li and I. Tabas, unpublished data). Another strategy may be to target specific oxidative enzymes that may play critical roles in oxidative stress-induced cell death. In addition to NADPH oxidase, the ER oxidase ERO1α plays a crucial role in CHOP-mediated death in ER-stressed macrophages,³¹ and Ron and colleagues¹¹⁵ have recently described small molecular inhibitors of this enzyme that protected mouse embryonic fibroblasts from ER stress-induced cell death.

A discussion of “anti-ER stress therapy” to prevent or treat atherosclerosis must consider several important points. First, as mentioned above, the beneficial effects of PBA and TUDCA, and particularly anti-oxidants, in mouse models of atherosclerosis have not yet been proven to be related to their ER stress-relieving properties. For example, PBA has a number of other biological actions, including histone de-acetylase inhibitor activity.¹¹⁶ Second, the UPR is an essential protective pathway in cells, and attempts to relieve ER stress by blocking this pathway will likely cause adverse effects unless the approach is targeted and limited. Moreover, different cell types in lesions may have different mechanisms of ER stress-induced damage. For example, prolonged CHOP expression appears to be an important cause of macrophage apoptosis,^{30, 62, 69, 71} while excessive induction of XBP1 has been shown to induce apoptosis in ECs.³⁹ However, physiologic levels of XBP1 may play a protective role in ECs subjected to atherogenic perturbations in blood flow.³⁶ Thus, the success of an ER stress-relieving strategy will likely depend on which branch of the UPR is targeted and whether it can prevent excessive UPR activation without perturbing physiologic ER stress signaling. In this light, recent developments in understanding how IRE1α functions to promote cell death vs. cell survival at a protein structural level raise the possibility of much more specific drugs that can block IRE1α-dependent cell death while retaining the protective effects of the IRE1α-XBP1-chaperone pathway.^{22, 115} Third, the process in atherosclerosis that actually leads to clinical disease is the development of necrotic, thin-capped plaques.^{13, 17, 18} Clinical studies in humans have shown that this type of “vulnerable” plaque morphology is much more predictive of atherothrombotic vascular disease than overall plaque size.¹¹⁷ Yet most animal studies use decrease in lesion area rather than improvements in plaque morphology as the measure of drug effectiveness. An ER stress-induced process specifically associated with plaque necrosis is advanced lesional macrophage apoptosis.¹¹ However, even here the situation is complex in terms of therapy, because while preventing death of advanced lesional macrophages may decrease plaque necrosis, it would also increases in the number of living macrophages, which may promote plaque disruption through other mechanisms, such as secretion of matrix-degrading proteases.¹¹⁸ In summary, careful consideration to mechanism, site of action, potency, and adverse of effects of ER stress-relieving therapy will be essential if this strategy is ever to be useful in preventing atherothrombotic vascular disease.

Summary and Future Directions

Prolonged ER stress has been identified as a pathogenic mechanism in a large number of disease processes, particularly chronic diseases associated with aging, obesity, and diabetes. Atherothrombotic vascular disease is particularly important to consider in this arena, because ER stress can adversely affect both systemic atherosclerotic risk factors and cell biological processes occurring at the level of the arterial wall (Figure 2). With regard to the arterial wall, the focus of this review, there is now increasing in vitro and in vivo evidence that prolonged ER stress is an important cause of macrophage and possibly endothelial cell apoptosis in advanced lesions (Figure 2B). Additional ER stress-mediated pro-inflammatory effects in these cells may also affect early atherogenesis (Figure 2A). Nonetheless, much more work is needed to truly understand the role of ER stress in atherosclerosis and to enable a well-informed effort at translating this knowledge into useful therapeutic strategies. Given the complexity of the multiple branches and functional consequences of ER stress signaling, a major goal over the next several years will be to define which branches of the UPR are activated and to elucidate the consequences of individual or integrated branch activation. Moreover, each cell type in atherosclerotic lesion development plays distinct roles, and these roles vary throughout the course of atherogenesis. Thus, determining the effects of interrupting prolonged ER stress in individual cell types at various stages during lesion development is a critical goal for the future. The development and use of conditionally gene-targeted mice will be essential to meet this goal. Temporal control over UPR gene silencing will also help avoid the potential confounding effects of compensatory alterations due to germline deletion. In terms of linking ER stress with clinically relevant atherosclerosis progression, mouse models of atherosclerosis are useful to study atherogenesis from lesion initiation up to the stage of necrotic plaques, but they are not a good model for plaque disruption or acute lumenal thrombosis.¹¹⁹ Although future mouse models may help address this problem, there is no substitute for human studies. Continuing work on describing what happens to ER stress pathways in different cell types in human lesions, particularly the types of lesions that are likely to cause clinical disease, will help inform animal studies. In addition, more precise causation information may be gleaned from genome-wide association studies¹²⁰ and UPR gene re-sequencing studies in which subjects with accelerated atherothrombotic vascular disease are compared with age-matched control subjects. Only through the knowledge gained from this combined approach will we be able to fully understand the roles of ER stress in atherosclerosis, to determine whether specific elements of prolonged ER stress are rationale drug targets, and to intelligently approach drug discovery and testing in this area.

Figure 2. Possible roles of prolonged ER stress in early atherogenesis and advanced plaque progression.

A, In early atherogenesis, extracellular matrix-retained and modified apoB-containing lipoproteins (LPs) trigger the expression of adhesion molecules and chemokines in endothelial cells (“EC activation”), leading to attraction of monocytes and other inflammatory cells to the nascent lesion. Early lesional endothelial cells, perhaps in response to disturbances in laminar blood flow at athero-susceptible sites, show evidence of ER stress, which may further promote EC activation. B, In advanced lesions, ER stress is prominent in macrophages (Mϕs) and can lead to inflammation and apoptosis in these cells. When apoptotic macrophages are not rapidly cleared by neighboring phagocytes, they become secondarily necrotic and lead to the generation of the necrotic core, a key feature of clinically dangerous plaques. ER stress in advanced lesions may also cause the death of endothelial cells, which may further amplify plaque progression and disruption, and of smooth muscle cells, which may contribute to the thinning of the protective fibrous cap. See text for details.

Nonstandard Abbreviations and Acronyms

aP2

adipocyte fatty acid-binding protein-2 (macrophage fatty acid-binding protein-4)

apoB

apolipoprotein B

Apoe

apolipoprotein E

AMPK

AMP kinase

ASK1

apoptosis signal-regulating kinase-1

ATF6

activating transcription factor-6

CaMKII

calcium/calmodulin-dependent protein kinase II

CHOP

CEBP-homologous protein

EC

endothelial cell

eIF2α

eukaryotic initiation factor 2α

ER

endoplasmic reticulum

ERK

extracellular signal-regulated kinases

ERO1α

ER oxidase-1α

GRP78

78-kDa glucose-regulated protein

GADD

growth arrest and DNA damage

IP3R

inositol 1,4,5-triphosphate-aftivated receptor

IRE1

inositol requiring-1

JNK

c-Jun amino-terminal kinase

Ldlr

LDL receptor

Lp(a)

lipoprotein(a)

oxPAPC

oxidized 1-palmitoyl-2-arachidonyl-sn-3-glycero-phosphorylcholine

PBA

4-phenyl butyric acid

PERK

RNA-dependent protein kinase (PKR)-like endoplasmic reticulum (ER) kinase

PRR

pattern recognition receptor

ROS

reactive oxygen species

SERCA

sarco/endoplasmic reticulum calcium-dependent ATPase

SFA

saturated fatty acid

SMC

smooth muscle cell

SRA

type A scavenger receptor

Tempol

4-hydroxy-2,2,6,6-tetramethylpiperidine-N-oxyl or 4-hydroxy-tempo

TLR

toll-like receptor

TUDCA

tauroursodeoxycholic acid

UPR

unfolded protein response

VSMC

vascular smooth muscle cell

XBP1

XhoI site-binding protein 1