Immunity and Inflammation in atherosclerosis

Abstract

There is now overwhelming experimental and clinical evidence that atherosclerosis is a chronic inflammatory disease. Lessons from genome-wide association studies, advanced in vivo imaging techniques, transgenic lineage tracing mice, and clinical interventional studies have shown that both innate and adaptive immune mechanisms can accelerate or curb atherosclerosis. Here, we summarize and discuss the pathogenesis of atherosclerosis with a focus on adaptive immunity. We discuss some limitations of animal models and the need for models that are tailored to better translate to human atherosclerosis and ultimately progress in prevention and treatment.

Introduction

Atherosclerosis is the most common underlying pathology of coronary artery disease (CAD), peripheral artery disease (PAD), and cerebrovascular disease^{1, 2}. The chronic build-up of vessel-occluding plaques in the subendothelial intimal layer of large and medium sized arteries eventually results in significant stenosis that restricts blood flow and causes critical tissue hypoxia³. The most common complications, myocardial infarction (MI) and stroke, are caused by spontaneous thrombotic vessel occlusion and represent the most common cause of death worldwide^{4, 5}. Current clinical guidelines focus on the treatment of these complications⁶. Clinically used therapies that efficiently prevent or curb the progression of atherosclerosis are limited to drugs that lower low-density lipoprotein (LDL) cholesterol. Traditionally, atherosclerosis was regarded as a cholesterol storage disease caused by the retention of lipoproteins including LDL in the intimal space of arteries. Retained LDL is modified and taken up by scavenger receptor-mediated phagocytosis. This process results in the continuous growth of fatty infiltrates rich in inflammatory leukocytes that macroscopically appear as plaques. Levels of plasma cholesterol, LDL cholesterol, and apolipoproteins, including Apolipoprotein B (ApoB), are highly correlated with clinical atherosclerosis^{7, 8}. Mice along with other animal models suggest causality: Elevating plasma cholesterol levels, as achieved by genetic knock-outs of LDL-receptor (LDLR) or Apolipoprotein E (ApoE) in mice, causes atherosclerosis in C57BL/6 mice that otherwise do not develop spontaneous disease^{9, 10}. Genome-wide association studies (GWAS) have correlated many single nucleotide polymorphisms (SNPs) in or near the genes encoding lipid-associated proteins. Examples include LDLR, APOB and proprotein convertase subtilisin/kexin type 9 (PCSK9), which modulate LDL cholesterol levels, as risk factors in atherosclerosis and MI^{11, 12}. In addition, atherosclerosis is accompanied by a chronic, low-grade inflammatory response that attracts cells of the innate and adaptive immune systems into the atherosclerotic plaque³, some of them recognizing ApoB, the core protein of LDL particles. Thus, atherosclerosis is a chronic inflammatory disease with an autoimmune component¹³. This autoimmune response is clinically best documented by antibodies against LDL and other atherosclerosis antigens, which are found in all patients and animal models. In many studies, low-affinity ‘natural’ antibodies against oxidation epitopes in LDL were found to be negatively correlated with atherosclerosis, while high-affinity antibodies secreted by IgG-producing plasma cells were positively correlated¹⁴. Here, we will summarize and discuss the adaptive autoimmune mechanisms that accompany and modify atherosclerotic disease.

LDL accumulation initiates vascular inflammation

The atherogenic process starts with the accumulation of several plasma lipoproteins in the subendothelial space at sites of flow perturbation and endothelial dysfunction. This is best documented for LDL, whose accumulation correlates with classical risk factors, such as smoking, hypertension, and metabolic dysregulation in obesity and diabetes¹⁵. In the intima, LDL undergoes oxidative modifications by reactive-oxygen species (ROS), which promote the uptake of oxLDL into macrophages¹⁶. In addition, oxidized phospholipids per se trigger inflammation of the arterial wall¹⁷ by binding to Toll-like receptors (TLRs), a group of widely expressed pattern-recognition receptors (PRRs) that cause pro-inflammatory signaling¹⁸. Clinically, oxLDL is a marker of plaque inflammation¹⁹. Native LDL can also be taken up by macrophages by micropinocytosis, or in its aggregated form as cholesterol complexes or -crystals by phagocytosis. The sustained influx of cholesterol eventually exceeds the phagocytes’ metabolic capacity and intracellular lipid droplets form. Microscopically, cholesterol-laden macrophages are ‘foam cells’. Cholesterol loading is thought to cause a myeloid cell response with pro-inflammatory cytokine secretion, in situ macrophage proliferation, and further recruitment of myeloid cells (summarized in²⁰). A clinically important consequence of cholesterol loading is the formation of intracellular cholesterol microcrystals that activate the inflammasome, a molecular machinery comprising molecules of the cytosolic-nucleotide binding domain and leucine-rich repeat gene family (NLRP3) that cleaves pro-IL-1β into its biologically active form²¹. IL-1β serves as an inflammatory master cytokine that enhances the expression of many pro-inflammatory cytokines, as well as of CRP²². Notably, attenuating cholesterol storage and enhancing cholesterol efflux pathways may favor the resolution of plaque inflammation end even promote plaque regression (summarized in²³). The myeloid response is accompanied by the infiltration of cells of the adaptive immune system, B and T cells^{24, 25}. Notably, the plaque’s growing content of myeloid cells and lymphocytes correlates with clinical complications and may predispose for future thromboembolic events caused by large cellular infiltrates and a thin fibrous cap (‘unstable plaque’)^{26, 27}.

Evidence for an autoimmune response in atherosclerosis

The presence of T and B cells in the plaque²⁸ sparked the hypothesis that atherosclerosis includes an autoimmune response. Adaptive immunity in infection and autoimmunity proceeds by a humoral arm that comprises specific antibodies against the antigen secreted by plasma cells, and a cellular arm with T cells that either activate B cells during co-stimulation or differentiate into effector T cells with pro- or anti-inflammatory cytokine production²⁹. CD8⁺ and CD4⁺ T cells only initiate immune responses to peptides presented MHC-I on all nucleated cells or MHC-II on antigen-presenting cells (APCs), respectively. Such responses are MHC-restricted, i.e. they only occur in individuals expressing a specific MHC-allele with the ability to bind the relevant peptide epitope. Binding of a specific T cell receptor (TCR) concomitant with co-stimulatory events provided by APCs activates T cells and causes their clonal proliferation³⁰. In mouse atherosclerosis, 2-photon microscopy has revealed an increased rate of APC-CD4⁺ T-helper cell interactions in the plaque specifically in the setting of hypercholesterolemia that resulted in pro-inflammatory cytokine secretion³¹. In addition, T-helper cells show an increasing maturation into antigen-experienced effector/memory (T_EM) and central-memory (T_CM) T cells in the lymph nodes (Figure 1a) that is also observed in atherosclerotic plaques^{28, 31}. Sequencing of the TCR revealed an oligoclonal origin of lesional T cells^{32, 33} suggesting that some (antigen-specific) T cell clones actively expand in the plaque. The enhanced activation of T cells is accompanied by an expansion of lymph nodes draining the atherosclerotic aorta in aged atherosclerotic Apoe^−/− mice (Figure 1b) and a local and systemic pro-inflammatory response that is further enhanced by a hypercholesterolemia-inducing diet^34-36. These findings support the concept that specific antigens drive an immune-response in the aorta and lymph nodes during atherosclerosis.

Figure 1: Activation of T cells is a hallmark of atherosclerosis.

(a) During feeding with a Western Diet (WD), CD4⁺ T-helper cells from atherosclerosis-prone Apoe^−/− build-up a significant immune memory with more than one half of T cells express markers of CD62L⁻ CD44⁺ T-effector memory cells (T_EM) and CD62L⁺ CD44⁺ central-memory cells (T_CM) when compared to atherosclerosis-free wildtype (WT) mice. (b) Along with enhanced T cell activation, lymph nodes draining the aorta and supra-aortic arteries (cervical, axillary lymph nodes) massively increase in size. Courtesy of D. Wolf and K. Ley

LDL – an autoantigen within the plaque

Of all candidates that may serve as B- and T cell activating antigens, plasma levels of LDL and its core protein ApoB show the strongest clinical and causal link with atherosclerosis in humans³⁷. ApoB-containing triglyceride-rich remnant particles also show a strong association with CVD, inflammation, and immune pathways⁷. Indeed, LDL as (auto-) antigen was first suggested by Gero et al in 1959: immunization with LDL protected against atherosclerosis in rabbits³⁸, suggesting that autoimmune response against LDL can be atheroprotective³⁹. Many CD4⁺ T cells in human plaques recognize oxLDL⁴⁰ by binding to MHC-presented peptide epitopes from ApoB-100^{41, 42}. A tetramer of recombinant MHC-molecules loaded with an ApoB-derived peptide – a tool to detect antigen-specific T-helper cells in vivo⁴³ – found a naturally occurring population of CD4⁺ T cells in the blood that recognizes the human peptide ApoB_3036-3050⁴². Furthermore, atherosclerosis is accompanied by IgG-antibodies against LDL, oxLDL, and ApoB⁴⁴. Collectively, these findings strongly suggest LDL as a relevant self-antigen that drives an autoimmune response against self-proteins in the atherosclerotic plaque. Besides LDL/ApoB, heat shock proteins (HSPs)^45-47 and some foreign peptides from pathogens such as Cytomegalovirus (CMV), Hepatitis C Virus (HCV), Human Immunodeficiency Virus (HIV), Human Papilloma Virus (HPV), and others^48-50 have been proposed as atherosclerosis-relevant antigens.

T-helper cell dependent immunity in atherosclerosis

Early evidence from immunohistochemistry studies^{28, 51}, more recent scRNAseq^{24, 52}, and CyTOF approaches^{24, 53} have estimated that ~ 25-38 % of all leukocytes in mouse aortic and human atherosclerotic plaques are CD3⁺ T cells, with CD3⁺CD4⁺ T-helper cells accounting for ~ 10 %. T cells predominantly populate atherosclerotic lesions with an enrichment in the fibrous cap^{28, 51}, but are also found in the adventitia of older lesions^{24, 54}. Their recruitment to the plaque occurs via chemokine receptors C-C chemokine receptor type 5 (CCR5), -X-C Motif Chemokine Receptor 6 (CXCR6), and others^{55, 56}. CD4⁺ T cells are critical regulators of the adaptive immune response with the ability to differentiate into distinct T-helper subtypes that can either be immune-dampening or activating to other T cells, exert direct anti- or pro-inflammatory effects on tissue resident cells, provide B cell help to induce the production of high-affinity IgG antibodies, or exhibit cytolytic activity²⁹ (Figure 2). Thus, the function of T-helper cells in atherosclerosis is multi-faceted and depends on specific transcriptional programs and patterns of cytokine secretion that can either fuel or attenuate atherosclerosis. Early evidence from Rag-1 deficient mice, which cannot produce mature T- and B cells, suggested a pathogenic role for T and B lymphocytes only in early atherosclerosis with moderately enhanced lipid levels, but not in severely hypercholesteremic Apoe^−/− mice^{57, 58}. Genetic absence of T cells in athymic nu/nu mice or a depletion of CD4⁺ T cells by anti-CD4 antibodies protected from lesion development⁵⁹. After antigen presentation by APCs, lesional T cells differentiate into functionally distinct T-helper subtype (T_H) −1, −2, −17, T-regulatory cells (T_reg), T-follicular helper cells (T_FH) and Type 1 regulatory (T_R1) cells⁶⁰. Atherosclerosis is a known T_H1-disease. Many CD4⁺ T cells in the plaque express the pro-inflammatory, T_H1-associated cytokines IFN-γ, IL-2, IL-3, TNF, and lymphotoxin (LT), which can activate macrophages, T cells, and other components of the plaque, and thereby aggravate the inflammatory response⁶¹. T cells that express the plaque-homing chemokine receptor CCR5 in lymph nodes, and T cells from atherosclerotic lesions secrete IFN-γ and express T-bet, the T_H1-lineage defining transcription factor^{55, 62}. Knocking out IFN-γ, its receptor, or T-bet protects mice from atherosclerosis^63-65. IFN-γ may directly reduce plaque stability by inhibiting smooth muscle cell proliferation⁶⁶, affecting macrophage polarization, and modulating cardiovascular risk factors⁶⁷. On the other hand, regulatory CD4⁺ T cells (T_regs) that express the transcription factor FoxP3 and the high affinity IL-2 receptor CD25 protect mice from atherosclerosis^{68, 69}. T_regs exert their atheroprotective properties by secreting the anti-inflammatory cytokine IL-10⁷⁰, plaque-stabilizing TGF-β⁷¹, and by suppressing the proliferation of pro-inflammatory T-effector cells⁷². Atheroprotective effects of in-vivo treatment with IL-2 complexes⁷³ and anti-CD3 treatment^{74, 75} have been attributed to a relative increase of T_regs. In addition, T_R1 cells that lack FoxP3 expression but express CD49b and Lag-3 secrete IL-10 and are atheroprotective^{76, 77}. In the atherosclerotic plaque, a substantial proportion of T cells moreover express transcripts for the T_H2 cytokines IL-4, IL-5, and IL-13²⁴. In contrast to abdominal aortic aneurysm formation, which is a clear T_H2-dependent disease⁷⁸ and the negative correlation of IL-4 with clinical atherosclerosis⁷⁹, the relevance of T_H-2 immunity in atherosclerosis remains unclear. The T_H2 cytokine IL-4 antagonizes T_H1 responses and diminished lesion formation in one study⁸⁰, while depletion of IL-4 has also been reported to be atheroprotective⁸¹. Likewise, the role of T_H17 cells in atherosclerosis is controversial: Deletion or neutralization of the master cytokine IL-17 protected from atherosclerosis^82-84, while other studies reported that T_H17 immunity protected from atherosclerosis and induced a stable plaque phenotype^85-87. T-follicular helper cells (T_FH), which are required to co-stimulate B cells and to induce an Ig-class switch, have also been proposed to be pro-atherogenic⁸⁸ or to protect from atherosclerosis by secreting LDL-lowering/neutralizing anti-LDL/ApoB secreting antibodies⁸⁹. The different findings in these studies may reflect different but unknown antigen specificities of the T cells studied.

Figure 2: T cell polarization in atherosclerosis.

Naïve T helper cells (T_H) acquire the complete phenotype of an effector T cell in the plaque after presentation of antigenic peptides from ApoB by antigen-presenting cells (APCs). Therefore, an APC takes up (oxidized) LDL-cholesterol particles, processes, and presents peptides from ApoB on MHC-II. The T cell recognized this complex by a specific T cell receptor (TCR). This process is guided by the binding of co-stimulatory ligands to their corresponding receptors on T cells. As a result of co-stimulatory signals and cytokines secreted by the APC, T cells express transcription factors (denoted in the cells) that favor the differentiation into distinct T_H-types. These express specific cytokines that can either act in an atheroprotection or pro-atherogenic manner. The relevance for atherosclerosis remains controversial for some T_H-phenotypes.

It is noteworthy that antigen presentation initiates and modulates the CD4⁺ T-helper cell response atherosclerosis. T cell activation in an antigen-specific manner is an exclusive consequence of antigen-presenting cells (APCs) that present antigenic peptides displayed on MHC molecules²⁹. Blocking MHC-II during co-stimulation or on APCs abrogates the downstream CD4⁺ T cell response^{31, 90}. T cell immune responses are typically initiated by antigen-loaded dendritic cells (DCs) migrating to lymph nodes. Several cells in the atherosclerotic plaque act as APCs for recall responses of antigen-experienced effector and memory T cells, including macrophages in the plaque, B cells in the adventitia, along with conventional DCs and plasmacytoid dendritic cells (pDCs). Depending on co-stimulatory signals and cytokines provided by these APCs, the immune response can be skewed into a tolerogenic (immune-suppressive) or an immunogenic response (summarized in^{91, 92}).

The role of other T cell subsets remains less well defined. It has been suggested that MHC-I dependent cytotoxic CD8⁺ T cells contribute to plaque inflammation and the build-up of the necrotic core^{93, 94}, but antigen specificity has not been considered⁹⁵. Natural killer (NK) cells regulate antigen-specific T cell immunity besides the killing of infected and tumor cells. They are detected at low frequencies in the plaque and may therefore modulate atherosclerosis⁹⁶. In contrast to earlier studies, a recent report, however, suggested that NK cells do not affect atherogenesis^97-100. In addition, CD1d-restricted NK-cells can recognize glycolipid antigens. Some NKT cell subsets were reported to aggravate atherosclerosis, but the atherosclerosis-relevant glycolipids detected by these NKT-cells remain unknown^101-103.

The function of ApoB-specific, auto-reactive T-helper cells

It has been challenging to determine the phenotype of ApoB-specific CD4⁺ T cells, i.e. the fraction of T cells with a TCR recognizing ApoB-peptides presented on MHC-II, within the pool of all lesional T cells. In animal models, ApoB-specific T cells have been expanded in-vitro or by vaccination against LDL or ApoB-peptides in-vivo. A direct transfer of vaccination-induced T cells in one study aggravated atherosclerotic disease¹⁰⁴. T cells re-stimulated with oxLDL ex-vivo promoted atherosclerotic disease after adoptive transfer in a model of immunodeficient scid/Apoe^−/− mice¹⁰⁵. Neutralization of T cells that responded to oxLDL stimulation by a monoclonal antibody directed against the TCRBV31 chain protected from atherosclerosis¹⁰⁶, suggestive of a pro-atherogenic function of ApoB-reactive T cells. However, more recent technologies to specifically detect ApoB-specific T cells in-vivo suggest the opposite: Tracking of ApoB-reactive T cells in mice and humans suggests that a majority of antigen-specific T cells are immunosuppressive T_regs⁴²_. This is consistent with recent work from Gistera et al. who transferred ApoB-reactive T cells from a mouse with a transgenic TCR directed against oxLDL/ApoB, which protected from atherosclerosis⁸⁹. These mixed results obscure the exact function of antigen-specific T cells. Likely, their phenotype and function depend on presented peptides, the microenvironment, and cytokine milieu, which potentially affects T cell polarization. Some pro-atherogenic antigen-specific T cells^{104, 106} were isolated and cloned for in vitro assays by a procedure known to pre-dispose and select pathogenic T_H1 and to neglect T_reg clones. It is possible that the population of antigen-specific T cells may be multi-potent to give rise to several T_H-lineages in-vivo – an idea consistent with the recent observation that MHC-II multimer selected ApoB-reactive T cells can express several T_H-defining transcription factors simultaneously⁴².

The T_reg-switch hypothesis – how protective immunity turns into a pathogenic response

The notion that atherosclerosis has an autoimmune component raised the question whether atherosclerosis is prevented in an antigen-specific manner by ApoB-reactive T_regs³⁹ in healthy individuals. T_regs prevent the onset on autoimmune disease¹⁰⁷. Naturally occurring T_regs are generated in the thymus (nT_regs) and peripheral are induced T_regs from naïve T cells (iT_regs). Despite the proven atheroprotective role of bulk T_regs^{68, 69} it has been unclear whether T_regs reactive to ApoB exist and how these may contribute to disease. Interestingly, ample clinical data suggest a strong inverse relationship between T_regs and atherosclerosis: Numbers of T_regs and IL-10 expression are lower in patients with myocardial infarction^{108, 109}. Low T_reg numbers predict cardiovascular events¹¹⁰. Blood T_reg numbers in established murine atherosclerosis decline in later disease, while effector T cells increase³⁶ (Figure 3a). However, in subclinical human atherosclerosis, T_reg numbers correlate positively with LDL¹¹¹. Likewise, in mice hypercholesterolemia initially favors the differentiation of T_regs¹¹², an effect that may be a counter-regulatory response to enhanced inflammation³⁶, intracellular lipid accumulation¹¹³, or an antigen-specific response. The latter hypothesis was supported by enhanced T cell receptor (TCR) downstream signaling events in hypercholesterolemic mice¹¹⁴, suggesting that a sub-population of T cells responds to antigens associated with increased LDL levels or to components of LDL particles itself. Thus, these data indirectly suggest the existence of LDL/ApoB-reactive T_regs that bear a TCR specifically responding to ApoB auto-peptides. These cells respond when the corresponding peptides are presented by MHC-II molecules by various APCs. Indeed, we directly demonstrated the existence of such ApoB-reactive T cells by MHC-II tetramers loaded with the human and mouse auto-peptide ApoB_3036-3050. Using this tool, we showed that among all ApoB_3036-3050-reactive CD4⁺ T-cells in patients free of cardiovascular disease, two thirds exclusively expressed FoxP3, indicative of a large population of ApoB-reactive T_regs. In patients with subclinical atherosclerosis, the percentage of exclusively FoxP3-positive T cells declined to ~ 30%, while a substantial proportion of the remaining FoxP3⁺ T cells acquired simultaneous expression of ROR-γT or T-bet, the T_H17 and T_H1-defining transcription factors, respectively⁴². These observations, along with the diminished pool of T_regs in later mouse and human disease, support the idea that the immunosuppressive phenotype of T_regs disappears as atherosclerosis progresses. Consist with this hypothesis, T_regs in late atherosclerosis in mice simultaneously express T-bet, lose their ability to regulate and to protect from atherosclerosis, while retaining some phenotypic similarity with T_regs, such as some residual expression of FoxP3^{36, 55, 62}. Gaddis et al. recently proposed that FoxP3 expression may be lost in favor of the transcription factor Bcl-6, the defining transcription factor for follicular-helper T cells⁸⁸. Adoptively transferred ApoB⁺ T-helper cells turned into T_FH cells after adoptive transfer⁸⁹. In other autoimmune conditions, such as experimental autoimmune encephalitis (EAE) and arthritis, an instability of FoxP3 expression triggers the formation of antigen-specific, but dysfunctional, partially non-protective former T_regs (exT_regs)^115-117. The instability of FoxP3 may be caused by increased methylation of the FoxP3 locus, which is observed in patients with severe cardiovascular disease¹¹⁸ and that may be prevented by modifications of lipid metabolism or anti-cytokine interventions^{88, 115}. In addition, the function of FoxP3 may be regulated by alternative splicing favoring pathogenic transcriptional programs¹¹⁹. These data suggest that the initial protective immune response by T_regs switches into a pathogenic response as atherosclerosis progresses³⁹ (Figure 3b,c).

Figure 3: Decline of protective T-regulatory cells (T_reg) in the course of atherosclerosis.

(a) As disease progresses, the pool of T_reg-dominated antigen-specific T cells is overwhelmed by effector T cells (T_eff) with a presumably pro-atherogenic function. (b) Over time, T_regs expressing their defining transcription factor FoxP3 start to express alternative T_H-transcription factors, such as RORγ-T (T_H17), Bcl-6 (T_FH), or T-bet (T_H1). FoxP3 either remains co-expressed or disappears. Likely, this switch into FoxP3-low expressed or -negative exT_regs may be caused by antigen-specificity of the T cell, the cytokine milieu in the atherosclerotic plaque, or the loading of the T cell with intracellular cholesterol. (c) These observations have built the concept of an increasing replacement of (athero-) protective immunity with a pro-atherogenic response.

Pro- and anti-atherogenic B cell responses in atherosclerosis

Classically, two types of B cells can be distinguished: B1 cells that are part of the innate immune system and secrete germ-line encoded IgM antibodies in a T cell independent manner, and B2 cells that need to be activated by T follicular helper cells (T_FH) to differentiate into plasma cells that secrete IgG antibodies. In infection and vaccination against pathogens, B cell-derived plasma cells secrete IgG antibodies that neutralize or opsonize bacteria and viruses²⁹. In addition, B cells can secrete numerous cytokines that distinctly affect inflammation. Examples include IRA-B cells, which are pro-atherogenic and secrete GM-CSF to drive myeloid cell activation and to induce pro-atherogenic T_H1 immunity¹²⁰. B-regulatory cells (B_reg) secrete the anti-inflammatory cytokine IL-10 and induce protective T-regulatory cells or directly act anti-inflammatory¹²¹, although the relevance for atherosclerosis is controversial¹²². The role of other cytokine producing B-effector (Be) cells is unclear. Only a few B cells are found in the atherosclerotic plaque²⁴; the majority of B cells reside in the adventitia, in particular in aged atherosclerotic animals, where arterial tertiary lymphoid organs (ATLOs) form¹²³. B cells in the spleen respond to a high cholesterol diet¹²⁴, suggesting local and systemic B cell responses in atherosclerosis. Global gain and loss of function experiments have suggested an overall protective role of B cells^{125, 126}. In general, innate B1 response seem to be atheroprotective and adaptive B2 responses pro-atherogenic (Figure 4):

Figure 4: Distinct role of B cells in atherosclerosis.

B cells on a developmental stage (Pre-B) turn into innate-like B1 cells or adaptive, conventional B2 cells (right panel). B1 cells recognize epitopes on LDL and oxLDL particles, which leads to their activation and expression of low-affinity IgM antibodies by proliferation. Often, these IgM show cross-reactivity with epitopes on bacteria such as Streptococcus pneumoniae or on apoptotic cells. Interfering with B1 functionality aggravates atherosclerosis. B2 cells require co-stimulation by T_FH cells by MHC-II:peptide:TCR interactions and co-stimulatory signaling events to fully differentiate into plasma cells that express high-affinity IgG antibodies against atherogenic antigens, such as ApoB, oxLDL, or heat-shock proteins (HSP). Neutralizing B2 cells is atheroprotective, while the role of IgG-antibodies remains controversial with reported pro- and anti-atherogenic functions. Independent of the classification of B1/2 cells, distinct B cell subsets have been shown to express non-exclusive sets of cytokines, which allows the definition of cytokine-secreting B-effector (Be) −1 and −2 cells, regulatory B cells (B_regs), and innate-response activator (IRA) B cells (left panel).

B1 cells:

B 1 cells represent a first-line, innate defense against common pathogens. In mice, they are characterized as CD11b⁺CD43⁺CD23⁻B220^lowCD19⁺ cells and may be sub-divided into B1a and B1b cells depending on their location and surface markers¹²⁷. Typically, most B1 cells reside in the peritoneal cavity. In the atherosclerotic plaque, a few CD11b⁺ B220_negCD19⁺ B1-like cells are found that further decrease in more advanced disease²⁴. B1 cells secrete germ-line encoded IgM. Typically, B1-derived IgM recognize phosphocholine (PC) head groups of polysaccharides in the wall of bacteria, such as S. pneumoniae. The same IgMs also bind oxidation-specific neo-epitopes on LDL and epitopes on apoptotic cells^128-130. Oxidative neo-epitopes also seem to be generated in the spleen during sterile inflammation¹³¹. In cardiovascular disease, IgM recognizing epitopes on LDL or ApoB are inversely correlated with atherosclerosis, complications, and outcome^132-138. It has been shown that IgMs directed against oxLDL inhibit its uptake by macrophages and prevent myeloid-cell inflammation^{139, 140}. Consistently, several studies with gain- and loss-of-function experiments have established an atheroprotective role for B1 cells^141-145.

B2 cells:

IgG antibodies originate from plasma cells that have undergone B cell maturation with the help of T_FH cells in germinal centers, which causes a switch from low-affinity IgM to high-affinity IgG¹⁴⁶. IgG antibody titers to native and oxidized LDL or ApoB are positively correlated with atherosclerotic disease in mice and humans^{133, 147-149}. Inhibiting B2 cells is reportedly atheroprotective^150-152, while specifically interfering with plasma cell functioning seems to be proatherogenic¹⁵³. The role of IgG antibodies in atherosclerosis is controversial: It was suggested that IgGs against ApoB aggravate¹⁵⁴ or protect from atherosclerosis^{155, 156}. A clinical phase II study (Goal of Oxidized LDL and Activated Macrophage Inhibition by Exposure to a Recombinant Antibody, ‘GLACIER’) using a monoclonal IgG antibody against a human ApoB-peptide failed to show its expected atheroprotective effect¹⁵⁷. The design of the study with the use of 8F-fluorodeoxyglucose (FDG) PET-imaging as surrogate for plaque inflammation instead of cardiovascular endpoints, the short observation period of 85 days, and the small study population may have contributed to its lack of efficacy.

Vaccination against atherosclerosis – a translatable strategy?

The discovery of the autoimmune component of atherosclerosis has sparked the idea of immunizing with LDL or peptides from ApoB to prevent atherosclerosis by inducing or maintaining the traits of protective immunity against ApoB. Almost 60 years ago, it was shown that rabbits develop smaller atherosclerotic lesions after subcutaneous injection of LDL³⁸. That vaccination with LDL can be atheroprotective was confirmed in a variety of species, LDL preparations, routes, and adjuvants^158-160. At least seven MHC-II-restricted peptides from ApoB, which contains the immunodominant epitopes of LDL, are protecting from atherosclerosis when used in vaccines: p3, p6, p101, p102, p103, p18, p210^{42, 161-163}. An ongoing challenge is to decipher the mechanism of action, which is critically required to define vaccination protocols translatable to humans. It has been proposed that either T_regs^{42, 164-167}, IL-10^{42, 161, 167, 168}, or vaccination-induced IgG-antibodies may confer atheroprotection, depending on doses, routes, and adjuvants used⁴⁴. Recent studies, however, suggest that atheroprotection does not require IgG-antibodies¹⁶⁹ and primarily proceeds by IL-10⁺ ApoB-specific T_regs⁴².

Whether vaccination strategies can be translated to humans remains unclear. A first step towards a translatable approach was the identification of human ApoB-peptide epitopes accessible to immunomodulation in two recent studies^{42, 170}. In mice, ApoB-peptides have been delivered in the non-translatable classical adjuvants Complete Freund’s Adjuvant (CFA), an emulsion of mineral oil supplemented with inactivated mycobacteria, or Incomplete Freund’s Adjuvant (IFA), which lacks the mycobacteria component of CFA. Subcutaneous or intraperitoneal injections of both, CFA and IFA, were shown to elicit non-specific inflammation^{171, 172}. This limitation was recently overcome by the discovery that a squalene oil, a class of adjuvants already used in clinical practice, can be used as an adjuvant for ApoB-peptides¹⁶⁹. In addition, it remains unclear whether vaccination is effective in established atherosclerosis as most studies tested the prevention of de-novo atherosclerosis in rodents.

Limitations of animal models

The principles of the cellular and humoral adaptive immune response in experimental murine atherosclerosis have been established. The efficacy of anti-inflammatory therapy in human atherosclerosis has been validated in the CANTOS trial recently. However, significant challenges remain for the translation of animal studies to humans. First, mice, which represent the most widely employed atherosclerosis model, neither develop spontaneous atherosclerosis, nor do atherosclerotic knockout mice develop coronary artery disease. In addition, spontaneous atherothrombotic events resembling heart attacks and strokes do not occur in atherosclerotic mice. Also, blood lipoprotein profiles in mice are unlike those in humans, even in the genetic atherosclerosis models. Second, most atherosclerosis studies are conducted in a single mouse strain (C57BL/6) that cannot capture the genetic diversity seen in humans. Genetic diversity is known to modulate the response against antigens and atherosclerosis-relevant stimuli within a spectrum from pro- to anti-inflammatory^173,174. Third, some cytokines and immune receptors are not conserved between mice and humans, because the immune systems of both species are under intense evolutionary pressure. Fourth, mice represent a simplified model system for antigen presentation and recognition. Unlike humans, mice are housed in specific-pathogen free (SPF) facilities, which neglects the likely pathogen-driven activation, antigenic repertoire, and differentiation of immune cells¹⁷⁵. While C57BL/6 mice bear just one MHC-II allele/molecule (I-A^b), humans express several alleles of a large pool of different MHC-II variants that are termed human leukocyte antigens (HLA) with over 10,000 different HLA allelic forms. This extreme variability renders the direction and amplitude of autoimmunity in humans difficult to predict.

Clinical considerations

Decreasing LDL levels and attenuating the inflammatory response represent the two fundamental therapeutic strategies against atherosclerosis available today. The most successful causal medication as measured by event-free person years are inhibitors of endogenous cholesterol synthesis by the HMG-CoA reductase (statins)^{176, 177}, which lower LDL-cholesterol and have pleiotropic anti-inflammatory effects beyond what can be expected from the reduction of LDL¹⁷⁸. Statins can prevent, reduce, and even reverse atherosclerotic plaque burden¹⁷⁹. Monoclonal antibodies to Proprotein convertase subtilisin/kexin type 9 (PCSK9) lower LDL-cholesterol even more dramatically by blocking LDL degradation^{180, 181} without apparent impact on levels of CRP levels¹⁸², a biomarker of systemic inflammation. However, even after LDL-lowering with statins and PCSK9-inhibition, a substantial residual inflammatory risk remains¹⁸³. These observations have established the distinct, but overlapping, roles of inflammation- and lipid associated risk. Low-dose treatment with the anti-proliferative and anti-inflammatory drug colchicine prevented cardiovascular events in a small prospective clinical trial¹⁸⁴. In addition, the Canakinumab Anti-inflammatory Thrombosis Outcomes Study (CANTOS) showed that inhibition of inflammation by the Interleukin 1-β (IL1-β) antibody canakinumab reduced cardiovascular end-points in patients with established atherosclerosis by 15%¹⁸⁵. Strikingly, these observations have proven the inflammatory hypothesis on a conceptual basis, yet it is unclear, which patients may benefit from novel anti-inflammatory therapies: First, inhibition of IL1-β impaired host defense, which was reflected by an increased incidence of lethal infections¹⁸⁵. Second, the recent Cardiovascular Inflammation Reduction Trial (CIRT) that tested low-dose anti-inflammatory methotrexate in patients with coronary heart disease did not reach its endpoints¹⁸⁶. This lack of efficacy was partially explained by the inclusion of patients at low inflammatory risk and calls for a future personalized risk stratification (inflammatory versus lipid risk) and treatment once anti-inflammatory therapy is available in clinical practice. Whether the autoimmune component of atherosclerosis may already be addressable by unspecific anti-inflammatory therapy is currently unknown. However, vaccination and immunomodulation may provide a future antigen-specific therapy that is unlikely to impair host defense. The first validation of MHC-II tetramers to quantify the ApoB-reactive T cell responses⁴² and the measurement of auto-antibodies¹⁸⁷ in humans may provide feasible risk stratification tools in the challenge to define patients at a high immune risk for atherosclerosis in future.

Conclusion

Atherosclerosis is a chronic inflammatory disease of the vessel wall that is largely driven by an innate immune response through myeloid cells as monocytes and macrophages. Autoimmunity against ApoB and other antigens involve CD4⁺ T-helper cells that instruct myeloid cells and antigen-specific antibodies that may directly modify the pathogenicity of these antigens. This autoimmune response is detectable in humans and animal models with atherosclerosis. While the classical perception is that autoimmunity is pathogenic per se, recent evidence suggests that ApoB-specific CD4⁺ T-helper cells are already detectable in subjects without clinical atherosclerosis, where many of them show atheroprotective features. As atherosclerosis progresses, the protective auto-immune response converts into a pathogenic one. It is unknown whether this switch in functionality represents a cause or a consequence of atherosclerosis and inflammation. It is clear that the adaptive immune system in atherosclerosis can be pro- or anti-inflammatory and thus pro- or anti-atherogenic. Manipulating the adaptive immune system by immunomodulatory strategies or vaccination is an attractive concept. Limitations in the predictive power of animal models and a lack of a full understanding of the role of auto-antibodies, B- and T cells present formidable hurdles to clinical translation.

Non-Standard Abbreviations and Acronyms

scRNAseq

Single-cell RNA sequencing

CyTOF

Cytometry by Time of Flight (mass cytometry)

T_H

T-helper cell

T_reg

T-regulatory cell

ATLO

Artery tertiary lymphoid organ

FACS

Fluorescence-activated cell sorting

LDL

Low-density lipoprotein

LDLR

Low-density lipoprotein receptor

ApoE

Apolipoprotein E

PCSK9

Proprotein convertase subtilisin/kexin type 9

GWAS

Genome-wide association studies

CRP

C-reactive protein

MHC

Major Histocompatibility complex

HDL

High-density lipoprotein