Readapting the adaptive immune response – therapeutic strategies for atherosclerosis

Abstract

Cardiovascular diseases remain a major global health issue, with the development of atherosclerosis as a major underlying cause. Our treatment of cardiovascular disease has improved greatly over the past three decades, but much remains to be done reduce disease burden. Current priorities include reducing atherosclerosis advancement to clinically significant stages and preventing plaque rupture or erosion. Inflammation and involvement of the adaptive immune system influences all these aspects and therefore is one focus for future therapeutic development. The atherosclerotic vascular wall is now recognized to be invaded from both sides (arterial lumen and adventitia), for better or worse, by the adaptive immune system. Atherosclerosis is also affected at several stages by adaptive immune responses, overall providing many opportunities to target these responses and to reduce disease progression. Protective influences that may be defective in diseased individuals include humoral responses to modified LDL and regulatory T cell responses. There are many strategies in development to boost these pathways in humans, including vaccine‐based therapies. The effects of various existing adaptive immune targeting therapies, such as blocking critical co‐stimulatory pathways or B cell depletion, on cardiovascular disease are beginning to emerge with important consequences for both autoimmune disease patients and the potential for wider use of such therapies. Entering the translation phase for adaptive immune targeting therapies is an exciting and promising prospect.

Linked Articles

This article is part of a themed section on Targeting Inflammation to Reduce Cardiovascular Disease Risk. To view the other articles in this section visit http://onlinelibrary.wiley.com/doi/10.1111/bph.v174.22/issuetoc and http://onlinelibrary.wiley.com/doi/10.1111/bcp.v82.4/issuetoc

Abbreviations

ACS

acute coronary syndrome

APCs

antigen presenting cells

ATLO

adventitial tertiary lymphoid organs

CAD

coronary artery disease

CVD

cardiovascular disease

HLA

human leukocyte antigens

hsp

heat shock protein

MDA

malondialdehyde

MHC

major histocompatibility complexes

MI

myocardial infarction

PC

phosphorylcholine

Tregs

regulatory T cells

Tables of Links

TARGETS
Other protein targets a	Enzymes b
CD20	HMG CoA reductase
CD80	Catalytic receptors c
CD86	TNFRSF13B (TACI)
ICOS, CTLA‐4, CD152	Transporters d
PD‐1, CD279	SLC12A3, Na‐Cl symporter

LIGANDS
Abatacept	IL‐6
BAFF	IL‐12
Belimumab	IL‐15
CXCL13	IL‐18
hsp60/65	Rituximab
IFN‐γ	Tabalumab
IL‐2	TNFSF13 (APRIL)
IL‐5

These Tables list key protein targets and ligands in this article that are hyperlinked to corresponding entries in http://www.guidetopharmacology.org, the common portal for data from the IUPHAR/BPS Guide to PHARMACOLOGY (Southan et al., 2016), and are permanently archived in the Concise Guide to PHARMACOLOGY 2015/16 (^a,b,c,dAlexander et al., 2015a,b,c,d).

Introduction

Cardiovascular diseases (CVD) remain a major global health issue. A major underlying cause of myocardial infarction (MI) and stroke is atherosclerosis, the chronic inflammatory response to injury in the vascular wall of major arteries at specific sites of susceptibility (Lusis, 2000). Atherosclerotic plaque is initiated by accumulation and modification of LDL cholesterol below the endothelial cell layer. Subsequent progression, which may occur discontinuously over decades, includes inflammatory activation, infiltration and proliferation of both immune cells and local vascular smooth muscle cells, and matrix deposition (Lichtman et al., 2013; Bentzon et al., 2014). The balance between accumulation of lipid‐filled foam cells, lack of clearance of apoptotic or necrotic cells, deposition of extracellular matrix and immune cell polarization intrinsically determines the growth and stability of the plaque. Extrinsically, many factors affect atherosclerosis, but perhaps none more prominently than circulating LDL levels. This is emphasized by the effectiveness of statins in reducing the risk of cardiovascular events (Collins et al., 2016). Plaque rupture, more specifically of the fibrous cap, is the most common cause of acute events. For example, rupture underlies 73% of fatal coronary thrombi (Falk et al., 2013), with plaque erosion as the most common alternative event. In the era of cholesterol‐lowering drugs (e.g. statins), there is a trend towards diminishing occurrence of plaque rupture (Libby and Pasterkamp, 2015). Our treatment of CVD has improved greatly over the past three decades. For example, the UK has seen a 70% drop in age‐adjusted cardiovascular mortality since 1979. However, over the same period, prevalence has remained stable (Bhatnagar et al., 2016). Current priorities include identifying at‐risk patients in more precise and timely ways, reducing the progression of atherosclerosis to clinically significant stages and preventing plaque rupture or erosion. Inflammation and recruitment of the adaptive immune system may be important in all of these aspects and is one focus for future therapeutic development.

Adaptive immune cells, T cells and B cells, have a unique ability to create clonally distinct antigen receptors through somatic rearrangement of genomic DNA and (in B cells) through further mutation. This creates an antigen receptor repertoire of near‐global specificity; B cells recognize diverse epitopes within whole molecules; T cells recognize peptides presented within major histocompatibility complexes (MHC) on antigen presenting cells (APCs). The MHC Class I (for CD8⁺ T cells) and MHC Class II (for CD4⁺) are named human leukocyte antigens (HLA) in humans. Normally, several layers of tolerance mechanisms limit a significant involvement of T cells in non‐infectious inflammation. These include deletion of autoreactive clones during thymic development, anergy through overstimulation or dominance of regulatory T cells (Tregs) that curtail initial T cell responses. T cell clones may overcome these mechanisms via defective negative selection, continued endurance of otherwise anergic T cells or failed Treg control. Key parameters dictating these changes include T cell specificity, APC stimulatory capacity and cytokine milieu. For example, CD4⁺ T cells are polarized to functional subtypes including Th1, Th2, Th17, follicular helper (Tfh) or Treg. B cells are also tamed by tolerance checkpoints. In addition, the most potent pro‐inflammatory responses are T cell‐dependent, providing another level of control.

The atherosclerotic vascular wall is invaded from both sides (arterial lumen and adventitia), for better or worse, by the adaptive immune system. We now understand a great deal about the effects of adaptive immune responses at different stages of CVD, with both pathogenic and protective influences delineated. We are now in a phase of translating these findings to the clinic and trialling therapeutic strategies in patients. Inflammation may represent a significant proportion of the residual risk for CVD remaining in patients on current therapies. If adaptive immune responses occur early on during disease development or before therapies are initiated, then despite reduction in risk factors (LDL cholesterol, innate cytokines), the memory compartment may endure and continue to produce pathogenic effectors. Strategies targeting adaptive immune components could therefore provide key future therapies to mitigate residual risk in patients. We will first discuss the analysis of human atherosclerotic patient samples and functional studies in experimental models. We then summarize some of the therapeutic strategies targeting adaptive immunity currently being considered.

Atherosclerosis and the adaptive immune system in patients

Local evidence – T cells

The most clinically relevant sites of atherosclerotic plaque development are the coronary and carotid arteries, with the aorta and its major branches also providing important sites in both humans and experimental models. Recent evidence significantly strengthened the association of adaptive immune infiltration with adverse outcomes or unstable status of plaques. T cells are found within coronary atherosclerotic plaques at many stages of disease progression (Boyle, 1997; Liuzzo et al., 2000; Watanabe et al., 2007; Kortelainen and Porvari, 2014; Otsuka et al., 2015). Otsuka et al. (2015) recently found CD3⁺ T cells in early lesions classified as pathological intimal thickening, but only once CD68⁺ cells (macrophages) were also present. T cells were then continuously present in later stages (fibroathermomas) of coronary arteries. Many studies have also detected T cells within carotid plaques (Kleindienst et al., 1993; Palma et al., 2006; de Boer et al., 2007; Chen et al., 2009; Erbel et al., 2011; Dietel et al., 2013; Rohm et al., 2015). Plaques of otherwise similar status in older patients contain higher levels of infiltrating T cells (Najib et al., 2012). Both macrophages and T cells are present in higher proportions in the fibrous cap of ruptured plaques (Falk, 1983; Lendon et al., 1991; van der Wal et al., 1994; Boyle, 1997; Otsuka et al., 2015). Erbel et al. (2011) recently showed that T cell infiltration was significantly associated with ischaemia. In the aorta, an increased T cell response in both plaque and adventitia peaks in correlation with plaque instability, with reduced T cell infiltration in healed plaques (van Dijk et al., 2015). Increased infiltration of effector T helper cells rather than Treg cells is a hallmark of unstable, compared with stable, carotid artery plaques (Dietel et al., 2013; Rohm et al., 2015). Altogether, T cell responses are involved in both early plaque progression and late stage degradation of fibrous caps, but the links and similarities between these responses are still unclear. In addition to T cells themselves, the molecules able to locally activate T cells are abundantly expressed by APCs in atherosclerotic plaque, for example, HLA (Jonasson et al., 1985; Xu et al., 1990) and co‐stimulatory molecules (de Boer et al., 1997). In the adventitia of atherosclerotic vessels, adaptive immune cell infiltration is commonly detected and can advance to formation of organized tertiary lymphoid organs (Boyle, 1997; Houtkamp et al., 2001; Clement et al., 2014). CD8⁺ T cells are quite common in advanced human atheroma and are found both in plaque and adventitia (Zhou et al., 1996; van Dijk et al., 2015). Early studies suggested less involvement of CD8⁺ compared with CD4⁺ T cells (Zhou et al., 1996), but the potential importance of CD8⁺ T cells has now been re‐assessed (see below).

Studies more than 20 years ago demonstrated that human plaque T cells respond to oxidized LDL (oxLDL)‐derived peptides presented on HLA molecules (Stemme et al., 1995) and subsequently to heat shock protein (hsp)60 (Benagiano et al., 2005; Almanzar et al., 2012). In the study by Stemme et al. (1995), isolation and stimulation of T cell clones from atherosclerotic plaques showed that a significant proportion proliferated in response to oxLDL and were Th1 polarized, producing IFN‐γ. This seminal study established the paradigm that oxLDL represents a neo‐self antigen responsible for activating and recruiting a T cell‐driven autoimmune response against the artery wall. Benagiano et al. (2003) similarly found a predominance of CD4⁺ Th1 polarized clones. An alternative T cell antigen is hsp60, which is up‐regulated on inflamed endothelial cells. Hsp60 bears close similarity with bacterial hsp (GroEL), so T cells initially activated by bacterial antigens could also regulate atherosclerosis (Mosorin et al., 2000; Benagiano et al., 2005).

Local evidence – B cells and antibodies

B cells are very rarely found within atherosclerotic plaques themselves, but are more often present in adventitial inflammatory cell infiltrates (Boyle, 1997; van Dijk et al., 2015). PCR approaches have been successful in detecting the presence of B cells in the plaque and/or adventitia of coronary and carotid plaques (Burioni et al., 2009; Hamze et al., 2013). These studies showed the presence of class‐switched, somatically mutated B cell receptor genes and the presence of limited but expanded clones. In addition, IgG‐ and IgA‐expressing cells were most common in one study (Hamze et al., 2013). These studies confirm the presence of an active ongoing B cell humoral response in advanced stages of atherosclerosis. Interestingly, Hamze et al. (2013) concluded that adventitial and plaque B cell clones were distinct and did not represent a directly linked response. How systemic responses that originate in spleen or draining lymph nodes relate in terms of antigen specificity and functional effects to local plaque and adventitial responses is currently unknown.

Antibodies, however, abundantly bind within atherosclerotic plaques. The assumed major specificity of these antibodies is modified lipids; however, many other autoantigens may be targeted (Merched et al., 2016). Modified lipid epitopes are certainly present, as exemplified by anti‐phosphorylcholine (PC; clone E06) staining in plaques (van Dijk et al., 2012). As in other autoimmune diseases, a major target could also be extracellular matrix or nuclear antigens. High levels of anti‐nuclear antibodies are not a prominent feature of coronary artery disease (CAD) patients unless autoimmunity is present. Nevertheless, one study associated enhanced circulating anti‐nuclear antibodies in symptomatic compared to asymptomatic carotid artery stenosis patients (Döring et al., 2012). IgE is also found in human carotid atherosclerotic plaques in areas of high CD68 expression and TUNEL⁺ apoptotic cells (Wang et al., 2011). In terms of antibody effector pathways, the classical complement pathway (primarily initiated by immune complexes) appears to be active in human atherosclerotic plaques (Kimoto et al., 1996; Oksjoki et al., 2007). Antibodies also influence cellular functions via ligation of cell surface Fc receptors. There is a family of Fcγ receptors (I–IV) for IgG, FcεR (CD23) for IgE, FcμR (Toso) for IgM and the neonatal Fc receptor (FcRn), which is important in IgG recycling (Smith and Clatworthy, 2010). Macrophages within atherosclerotic plaques are influenced by IgG immune complexes via Fcγ receptors and vulnerable plaques exhibit enhanced levels of FcγR and downstream signalling molecules (Lennartz et al., 2011). FcγRIII (CD16) is an important marker of non‐classical patrolling monocytes, compared with classical inflammatory monocytes that express high levels of the toll‐like receptor 4 co‐receptor CD14. Endothelial cells and vascular smooth muscle cells also express FcγR (Tanigaki et al., 2015). In summary, there is evidence for an adaptive immune cell response ongoing within atherosclerosis, as well as the presence of upstream activating and downstream effector pathways. The levels of these responses correlate closely with clinically relevant plaque status.

Systemic, genetic and risk factor‐related evidence

The analysis of circulating T cell subsets has also informed us about potential roles in CVD (see Ammirati et al., 2015). As in atherosclerotic plaques, T cells reactive to oxLDL are present in the blood of CVD patients (Frostegård et al., 1992). Various functional T cell subsets have been associated with CAD patients. Ammirati et al. found a positive correlation between effector memory T cells and CVD (Ammirati et al., 2012). In contrast, Treg levels are reduced in CAD patients (Mor et al., 2006; Ji et al., 2009; Wigren et al., 2012). A prominent feature of the T cell system of cardiovascular patients is the expansion of an otherwise rare population of CD4⁺ CD28⁻ T cells. This population increases with chronic inflammation, produces IFN‐γ and acquires cytotoxic functions not normally associated with CD4⁺ T cells (Dumitriu et al., 2012; Téo et al., 2013). One study found that CD4⁺ CD28⁻ cells responded to hsp60 (Zal et al., 2004). Chronic low‐grade inflammation may promote re‐stimulation of these cells, leading to loss of CD28 (Bobryshev, 2010; Chistiakov et al., 2016), as well as resistance to apoptosis (Kovalcsik et al., 2014). Acute coronary syndrome (ACS) patients have a reduced T cell repertoire compared with healthy controls, suggesting accelerated ageing or systemic expansion of certain, perhaps autoreactive, clones (Klingenberg et al., 2014).

A large number of studies have correlated modified lipid‐binding antibody levels to various aspects of CVD, as reviewed in more detail elsewhere (Carbone et al., 2013). A consistent finding is the inverse correlation of CVD with anti‐oxidized lipid IgM antibodies (Karvonen et al., 2003). These ubiquitous and common antibodies may be important in limiting the accumulation of oxidized lipid debris and preventing foam cell formation. An alternative possibility could involve acting on PC and malondialdehyde (MDA)‐positive bacteria. Low levels of anti‐PC IgM can predict mortality (Carrero et al., 2009) whereas high levels are associated with slower atherosclerosis progression (Su et al., 2005; Ravandi et al., 2011). Single nucleotide polymorphisms close to genes encoding TNFSF13 (APRIL), a known plasma cell survival factor, and its receptor TNFRSF13B (TACI) were associated with IgM levels in humans (Osman et al., 2012). IgG antibodies to the same epitopes have more inconsistent correlations. For example, Lehtimäki et al. (1999) found a significant positive association with atherosclerosis, whereas in a prospective observational study, Wilson et al. (2012) found no effect on risk. Another recent study demonstrated that high levels of antibodies binding ApoB100 in LDL are associated with reduced CVD (Björkbacka et al., 2016). Khamis et al. (2016) recently found total IgG levels correlated with CVD and adjusting for total levels negated the correlation of oxLDL‐specific IgG. Increased serum IgE levels have been associated with enhanced risk of CAD, and in particular plaque stability (Kounis and Hahalis, 2016). In addition, IgE, eosinophil and mast cell responses occur as part of immune responses in ACSs (Kritikou et al., 2016).

A recent study used genomics data to determine that two B cell regulatory genes, TNFSF13B (B cell activating factor; BAFF) and SPIB, were the core regulatory factors in a module of genes co‐regulated in non‐CVD subjects but deregulated in CVD patients (Huan et al., 2013). Although neither BAFF nor SPIB is entirely B cell specific, this study supports the idea that significant changes in B cell responses are associated with CVD. A major function for IL‐5, and one of many functions of IL‐6, is to promote B cell humoral responses, and both IL‐5 and IL‐6 receptors are candidate target genes of two genomic loci associated with CVD through GWAS analysis (Nikpay et al., 2015), as is the chromosome 6p21 region, which contains the Class I HLA locus (Davies et al., 2012). Soluble IL‐2Rα, a decoy receptor for the lymphocyte growth factor IL‐2, associates with CVD even after adjustment for other risk factors; genetic polymorphisms at chromosome 10p15–14 influence the level of sIL‐2Rα, but do not directly link to increased incidence of CVD (Durda et al., 2015). The chemokine CXCL13 was recently shown to correlate closely with germinal centre B cell responses in humans (Havenar‐Daughton et al., 2016), although it does have innate immune functions as well. Higher CXCL13 levels were found in CVD patients, particularly those with symptomatic status (Smedbakken et al., 2012). This systemic data supports the histology data showing adaptive immune responses correlating closely with plaque status (van Dijk et al., 2015; see above).

Many central risk factors for CVD and atherosclerosis may also elicit adaptive immune responses. The immune system declines with age (see Linterman, 2014). However, a decrease in the strength of responses to infections or vaccinations is matched with a paradoxical increase in autoimmune propensity. The roles of adaptive immune responses in hypertension, obesity and type II diabetes have mainly been addressed in animal models and are considered briefly below.

Atherosclerosis and the adaptive immune system in experimental models

The functional importance of adaptive immunity in atherosclerosis is demonstrated by a plethora of in vivo studies in animal models, most extensively in the complementary mouse models of hypercholesterolemia, Apoe ^−/− and Ldlr ^−/− mice, but also in hypercholesteraemic rabbits. A limit to translating experimental findings into therapies is the differences between the murine and human immune systems. Examples of significant differences include distinct functional properties of antibody isotypes and Fc receptor‐dependent responses between mice and humans (Smith and Clatworthy, 2010) and the ability of mouse but not human thymus to maintain the naïve T cell pool (den Braber et al., 2012). A recent study suggested entirely different transcriptional profiles of mouse and human responses to a range of diseases (Seok et al., 2013), although this study was challenged soon after (Takao and Miyakawa, 2014). Accepting these caveats, major findings on the involvement of B cell and T cell responses in humans (as far as they are understood) are faithfully reproduced in mice. Several consistent findings allow us to assign important anti‐atherogenic status to Tregs and to natural IgM‐producing B1 cells. Depletion of Tregs leads to increased atherosclerosis (Ait‐Oufella et al., 2006; Gotsman et al., 2006). Interference with conventional dendritic cells also leads to reduced Treg levels and increased atherosclerosis (Choi et al., 2011; Subramanian et al., 2013). As atherosclerosis progresses, local Treg levels are reduced in Ldlr ^−/− mice (Maganto‐García et al., 2011). This could potentially relate to their conversion to Il‐17⁺ or IFN‐γ producing Th1‐like cells (Butcher et al., 2016). Tregs dampen immune responses in many ways and can both prevent the intitiation of responses and the acceleration of the resolution phase. Tregs are more sensitive than effector or naïve T cells to IL‐2. Once Tregs recognize antigen, they can repress APC by removing co‐stimulatory molecules [e.g. through cytotoxic T lymphocyte antigen‐4 (CTLA4)] and by secreting IL‐10 and TGF‐β.

The protective role of B1 cells is supported by many studies in mouse models. The enhanced atherosclerosis in splenectomized mice was reversed by transfer of B1 but not B2 cells (Kyaw et al., 2011). Both B1a (CD5⁺) and B1b (CD5⁻) cells are capable of producing atheroprotective IgM (Binder et al., 2003; Rosenfeld et al., 2015). Unlike IgG antibodies, IgM does not bind to pro‐inflammatory Fc receptors on macrophages and neutrophils, although it can be associated with promoting antigen presentation to T Cells. The oxidized epitopes targeted by atheroprotective IgM are found on various circulating particles including oxLDL, Lipoprotein(a), apoptotic cells and microparticles, all of which may also accumulate within atherosclerotic plaques (Tsiantoulas et al., 2015). It remains to be determined whether IgM removal of all or only particular carriers of these epitopes is important for atheroprotection. An atheroprotective clonotype of anti‐PC IgM, T15/E06, is encoded by the V1 heavy chain gene. Although IgM encoded by this gene is important for protection from Staphylococcus pneumoniae infection, atherosclerosis was not affected, and overall IgM reactive with PC was not significantly affected (Centa et al., 2015). As the most prominent antagonist to these protective pathways, Th1 type CD4⁺ T cells promote atherosclerosis. Production of the cytokines IFN‐γ and IL‐12 are important mediators of this role (Davenport and Tipping, 2003; Buono et al., 2005). The IL‐1 family cytokine IL‐18 is also a strong signal for Th1 differentiation when in combination with IL‐12 or IL‐15 (Novick et al., 2013). In mice, IL‐18 promotes atherosclerosis (Mallat et al., 2001), and a recent paper showed it can act via both the IL‐18 receptor and an alternative receptor, SLC12A3 (Wang et al., 2015). IFN‐γ can activate macrophages towards a pro‐inflammatory phenotype and also directly targets vascular cells. B cell production of GM‐CSF (Hilgendorf et al., 2014) and antigen presentation by plasmacytoid dendritic cells (Sage et al., 2014) are upstream regulators of Th1‐driven atherosclerosis.

Other lymphocyte subsets have less well‐defined roles, with studies in mice showing opposing effects on atherosclerosis (Legein et al., 2013; Witztum and Lichtman, 2014; Taleb et al., 2015). This can be due to differences in experimental design and conditions between labs as well as to the multifunctional roles of these subsets. For example, B2 cells have been ascribed both pathogenic and protective roles based on different experimental approaches. B2 cell depletion, leaving B1 cells intact, leads to reduced atherosclerosis (Ait‐Oufella et al., 2010; Kyaw et al., 2012; Sage et al., 2012). However, adoptive transfer of purified B2 cells into B cell deficient (lacking both B1 and B2 cell) mice was either pro‐ or anti‐atherogenic in two different studies (Kyaw et al., 2010; Doran et al., 2012). Autoimmune disease models associated with B2 cell production of autoantibodies are generally associated with increased atherosclerosis (Gautier et al., 2007; Clement et al., 2014; Merched et al., 2016). Indirectly, knockout of pro‐inflammatory IgG (Fcγ) receptors results in reduced atherosclerosis (Hernández‐Vargas et al., 2006; Kelly et al., 2010), supporting a pathogenic role for autoantibodies. Although there is evidence of pathogenic influence, anti‐oxLDL IgG is most to be likely protective in the same way as IgM of the same specificity. Treatment of atherosclerotic mice in a regression model (switched to chow from high‐fat diet) with human monoclonal IgG1 to MDA‐modified ApoB100 resulted in accelerated regression and 50% less atherosclerosis than the control group (Schiopu et al., 2007). Initial studies with mice lacking inhibitory FcγRIIb showed more atherosclerosis. However, more recently, the opposite effect was shown, with a mixed genetic background in the early studies a critical difference (Mendez‐Fernandez et al., 2011; Harmon et al., 2014; Ng et al., 2015). Given the diverse subsets and response types of adaptive B2 cells, more models that can target specific functions or prevent antigen‐specific responses are required to more accurately probe the potential roles for B2 cells. Several studies have suggested that CD8⁺ T cells, not commonly found in mouse atherosclerotic plaques, do not play a significant role in atherosclerosis (Zhou et al., 1996; Kolbus et al., 2012); however, CD8⁺ T cells can promote atherosclerosis in some studies (Kyaw et al., 2013a). A recent study demonstrated that in addition to their classical cytotoxic function, the lack of CD8⁺ T cells significantly modulated monocyte production in the bone marrow (Cochain et al., 2015). Given the recent genetic association of CVD with the CD8⁺ T cell activating MHC Class I locus (Davies et al., 2012), further study of the role of CD8⁺ T cells is certainly warranted.

An adventitial adaptive immune response in the descending aorta of Apoe ^−/− mice can occur quite early during experimental models (Galkina et al., 2006; Koltsova et al., 2012), and in old Apoe ^−/− mice, these responses become organized into adventitial tertiary lymphoid organs (ATLO) (Srikakulapu et al., 2016). A recent study used smooth muscle cell‐specific deletion of lymphotoxin β receptor to prevent the formation of ATLO and found that this exacerbated atherosclerosis (Hu et al., 2015). Microvessels are abundantly present in atherosclerotic plaque (Collett and Canfield, 2005), responding to plaque hypoxia (Marsch et al., 2014). Such microvascular networks provide the possibility for direct interaction of adventitial and plaque immunity. Adventitial responses can alternatively have systemic consequences for CVD. T cell responses in perivascular adipose tissue are essential for the development of experimental hypertension (Itani et al., 2016), although human evidence of these links is sparse. In early atherosclerosis, dendritic cells scanning sites of atherosclerosis susceptibility were recently shown to exit back into the vessel lumen (intravasation) (Roufaiel et al., 2016), providing another potential route and suggesting systemic T cell recruitment.

Beyond atherosclerosis models, T cell and B cell responses also play prominent roles in co‐morbidities for CVD. Autoimmune‐driven arthritis and systemic lupus erythematosus (SLE) have a clear adaptive immune component. Increasingly, models of type 2 diabetes reveal important potential roles for the adaptive immune system. As shown for atherosclerosis, in these models, the T cells that display an effector/memory phenotype, and are predominantly Th1‐polarized, infiltrate obese adipose tissue and promote insulin resistance (see Majdoubi et al., 2016). B cell responses originating in adipose tissue may also be important in type 2 diabetes. For example, transfer of wildtype but not MHCII^−/− B cells to B cell‐deficient obese mice induced insulin resistance (Winer et al., 2011).

Therapeutic strategies targeting the adaptive immune system

Current therapies for CVD designed to primarily target other disease processes may also act in part through actions on the adaptive immune system (see (van Vré et al., 2011). The anti‐inflammatory properties of cholesterol‐lowering statins are well known (Jain and Ridker, 2005) including direct potential effects on adaptive immune responses. For example, simvastatin reduces MHCII expression (Kwak et al., 2000). The complex relationship between lymphocytes and systemic lipid metabolism is receiving more attention (Chyu et al., 2014; Sorci‐Thomas and Thomas, 2016). Angiotensin‐converting enzyme inhibitors used to lower blood pressure may also directly target T cells, which express angiotensin AT₁ receptors and play an important role in angiotensin II‐induced vascular pathology (Guzik et al., 2007).

Given the presence of both protective and pathogenic responses, therapeutic strategies designed against the adaptive immune system can be broadly divided into those attempting to boost protective pathways or those attempting to correct or inhibit pathogenic pathways. Strategies can also be global, by targeting co‐stimulatory or cytokine signalling pathways, or antigen‐specific. Co‐stimulatory (or co‐inhibitory) proteins form ligand–receptor pairs between interacting cell types and can dictate the outcomes for both cells and thus play a prominent role in guiding adaptive immune responses (den Haan et al., 2014). A prominent and well‐studied co‐stimulatory pairing is CD40‐CD40 ligand, members of the TNF family. However, as this pathway has many different effects on atherosclerosis via important roles in innate immune cell or platelets as well as the adaptive immune response, we do not discuss this pathway further here (Gerdes et al., 2016; Jansen et al., 2016). Methods to isolate locally induced (i.e. ATLO) responses from systemic responses are currently lacking in the clinical setting but could form the basis for future strategies. Below, we discuss antigen‐independent strategies to boost Treg‐mediated protection or attenuate pathogenic responses. Antigen‐specific immunization and tolerization strategies, targeting both T cells and B cells, are discussed finally. The targeted pathways of these strategies are illustrated in Figure 1. Most strategies outlined below could potentially attenuate both plaque progression and the risk of plaque rupture. Given the lack of widely used models of plaque rupture in mice, it is hard to assess the effects separately in experimental models. Immunization and tolerization strategies are the most likely to be effective at earlier stages of disease.

Figure 1.

Potential therapeutic strategies targeting adaptive immunity in atherosclerosis. Adaptive immune responses are initiated in lymphoid organs (spleen, lymph nodes and, in atherosclerosis, in inflamed adventitia). Cell–cell interactions between dendritic cells, naïve T cells and B2 cells, dependent on co‐stimulatory molecules and cytokines, result in atherogenic responses such as effector T cells and pathogenic autoantibodies from B2 cells but also protective responses such as Tregs. B1 cells (and B2 cells) are stimulated by antigen and cytokines such as IL‐5 to produce atheroprotective anti‐oxidation specific epitope (OSE) antibodies. Atherosclerosis is driven by macrophage foam cell formation leading to necrotic core formation. Prevention of the rupture of the fibrous cap of smooth muscle cells and collagen matrix is a key therapeutic target. Pathogenic T cells (e.g. Th1), antagonized by Tregs, promote inflammation and cell death of endothelial cells, smooth muscle cells and macrophages. Pathogenic autoantibodies that bind plaque cells or extracellular matrix may also promote inflammatory activation of vascular and immune cells. Protective antibodies neutralize and promote uptake of apoptotic/necrotic cells, modified lipids and microparticles. The points of intervention for potential therapeutic strategies are indicated in the red boxes.

Regulatory T cell enhancing therapies

Interference with CD80/86–CD28 co‐stimulation is an effective targeted immunosuppressive strategy, mimicking the effects of Tregs. Co‐stimulation of CD28⁺ T cells via CD80/86 on APCs drives atherosclerosis in mice (Buono et al., 2004; Matsumoto et al., 2016). Biological therapies based on CTLA4 are approved for rheumatoid arthritis and kidney transplantation. In mice, translational studies using CTLA‐Ig (abatacept) have shown promise. Ewing et al. showed decreased intimal hyperplasia and inflammation with abatacept (Ewing et al., 2013), whereas Ma et al. showed effective inhibition of homocysteine‐accelerated atherosclerosis in Apoe ^−/− mice (Ma et al., 2013). IL‐2 triggers expansion and differentiation of both effector T cells and Tregs. Tregs have a significantly enhanced sensitivity to IL‐2, and this underlies the approach of treating patients with low dose IL‐2 to preferentially expand Tregs with minimal impacts on effector T cells or other IL‐2R expressing cells (e.g. eosinophils) (Klatzmann and Abbas, 2015). Experimental studies characterized that low dose IL‐2, or IL‐2 in complex with an anti‐IL2 antibody to enhance stability, preferentially expands Tregs (see Pham et al., 2016). This therapeutic strategy holds promise in reducing many T cell‐mediated pathologies or where an increased Treg response can limit inflammation‐driven damage. In human clinical studies, low dose IL‐2 has been tested against type 1 diabetes, systemic lupus erythromatosus and hepatitis C virus‐induced vasculitis (Klatzmann and Abbas, 2015; Pham et al., 2016). One ‘side effect’ of this strategy is that IL‐2 may also expand CD25⁺ innate lymphoid cells. The role of these cells in atherosclerosis is just beginning to be investigated (Engelbertsen et al., 2015). Furthermore, there is incomplete knowledge on how IL‐2 therapy affects the emerging plasticity of Tregs (Butcher et al., 2016).

Co‐stimulatory pathway targeting therapies

As an alternative to boosting Tregs, therapies directly blocking effector T cell development or function could be developed to neutralize pathogenic T cell responses in atherosclerosis. Two prominent co‐stimulatory pathways that inhibit or promote effector T cell responses are programmed cell death‐1 (PD‐1) and inducible T cell co‐stimulator (ICOS) respectively. PD‐1 and its ligands PD‐L1 and PD‐L2 are important in restraining pathogenic T cell responses. High PD‐1 expression on T cells can also be associated with anergic or exhausted T cells in chronic inflammation in humans (McKinney et al., 2015) and may mark a pathogenic subset of T cells refractory to normal inhibitory signals. In murine atherosclerosis, PD‐1 is associated with a protective role. Ldlr ^−/− mice lacking both PD‐L1 and PD‐L2 develop enhanced systemic T cell activation, serum TNF and increased atherosclerosis (Gotsman et al., 2007). Pd1 ^−/− mice also develop enhanced atherosclerosis (Bu et al., 2011; Cochain et al., 2014). In addition to systemic T cell activation, this was associated with increased apoptotic cell (both smooth muscle cells and immune cells) accumulation within atherosclerotic plaques. Surprisingly, this result was not replicated in bone marrow chimeras with PD‐L1/2 deficiency (Bu et al., 2011). Anti‐PD1 antibody did not increase atherosclerotic size over a 3 week treatment but greatly increased T cell infiltration (Bu et al., 2011). Promoting PD‐1‐mediated immune restraint could therefore be beneficial in CVD, in contrast to cancers where removing this inhibitory pathway enables anti‐tumour T cell responses.

ICOS and its ligand ICOSL are members of the B7 family and are very important in promoting T helper cell responses, including Tregs. Deletion of ICOS in bone marrow cells reduced Tregs and increased atherosclerosis in Ldlr ^−/− mice (Gotsman et al., 2006). ICOS‐Fc treatment also led to increased atherosclerosis, although a significant level of anti‐mouse ICOS antibodies was induced by the treatment (Afek et al., 2005). More recently, targeting ICOS‐ICOSL was effective in reducing atherosclerosis in a mouse model of combined autoimmunity and atherosclerosis (Clement et al., 2014). Deficiency in CD8⁺ Tregs led to a prominent expansion of CD4⁺ T cell activation, particularly B helper Tfh. ICOS‐ICOSL interactions are crucial in B2 cell‐Tfh interactions, and anti‐ICOSL treatment reduced Tfh cell levels and atherosclerosis. The role of Tfh‐mediated responses and effects of anti‐ICOSL therapies in other models of atherosclerosis remain to be determined. A number of other co‐stimulatory pathways have been associated with atherosclerosis in experimental models, but progress towards the clinic is less advanced. For example, OX40–OX40 ligand interactions promote T cell responses through stimulating both survival and expansion. In a regression model, where switching from Western to chow diet (10 weeks of each) induces little change in plaque size but a very significant reduction in macrophages and expansion of collagen, anti‐OX40 ligand induced plaque regression (Foks et al., 2013). Information on the effects of immunotherapies on cardiovascular parameters is constantly emerging, and this information is important in initiating novel strategies to directly counteract atherosclerosis progression.

B cell depletion

B cell depletion therapy uses monoclonal antibodies against B cell‐specific cell surface molecules such as CD20 or survival factors such as BAFF. B cell depletion was developed against B cell‐derived cancers and is also now approved for use against autoimmune diseases. Although autoantibody production may also be modulated, the primary mode of action of B cell depletion is likely to be via its effects on effector/memory T cells (Lykken et al., 2014). Some human data on the effects of B cell depletion on CVD in autoimmune disease patients are beginning to emerge. Novikova et al. (2015) detected a reduced cIMT in rituximab‐treated individuals while Provan et al. (2015) found that after 12 months rituximab, pulse wave velocity, a measure of arterial stiffness, was reduced. A previous study (Raterman et al., 2013) found no effects on 6 months follow up. In atherosclerosis mouse models, primary studies showed reduced atherosclerosis when anti‐CD20 treatment was administered prior to atherosclerotic development (Ait‐Oufella et al., 2010; Kyaw et al., 2010). It was also effective in treating progression of pre‐established atherosclerosis (Kyaw et al., 2010), with further mechanistic insight into how B cell‐T cell interactions regulate atherosclerosis still required. In MI models, B cell depletion was also protective but through a distinct mechanism. B cells producing the chemokine CCL7 are critical in recruiting destructive monocytes to the infarction site, and thus, B cell depletion may represent an acute strategy post‐MI in reducing infarct inflammation (Zouggari et al., 2013). We believe that current broad targeting of B cells using depletion therapies may be more practical for the acute treatment of ischaemic heart disease post‐MI rather than for the long‐term treatment of atherosclerosis. However, more targeted strategies with similar mechanisms may well be more suitable. Several anti‐BAFF monoclonal antibodies, belimumab, tabalumab and blisibimod, are in use in the clinic against autoimmune diseases and chronic lymphocytic leukaemia. In contrast to B cell depleting antibodies targeting B cell surface molecules (CD19, CD20, CD22), anti‐BAFF therapies will also block other BAFF functions and so may have distinct modes of action. For example, BAFF may act directly on activated T cells (Scapini et al., 2010), and the BAFF receptor TACI is reported to be expressed on macrophages in mice (Allman et al., 2015). BAFF may also affect cholesterol metabolism and adipose tissue function (Jackson et al., 2016; Kim and Hyun, 2016). Antibodies against the BAFF receptor target more specifically the pathogenic B2 cell subset and are effective in mouse models (Kyaw et al., 2013b) but are not yet translated to humans. To date, there have been no studies investigating cardiovascular parameters in patients treated with BAFF‐targeting therapies.

Immunization and tolerization

In most cases, immunization and tolerization against a particular antigen should result in opposite responses in the host. In the case of modified LDL, studies in rabbits and mice have shown that both immunization (injection of antigen with adjuvants) and tolerization (non‐adjuvant administration of peptides via oral or subcutaneous routes) can significantly reduce atherosclerosis (Palinski et al., 1995; Ameli et al., 1996; George et al., 1998). The target of tolerization is intentionally the Treg response, whereas immunization increases both innate and adaptive B cell responses, as well as Treg responses. The increased Treg response is proposed to be most active in reducing atherosclerosis and may be the reason why both immunization and tolerization can protect from atherosclerotic development (Freigang et al., 1998), with increased IL‐5 that boosts innate B1 cell responses also a functional protective pathway (Binder et al., 2004). Complexing the PC antigen into liposomes was also effective in reducing atherosclerosis (Hosseini et al., 2015). Immunization with human ApoB100 induced a Th2 response and an IgG1 antibody response to both human and mouse LDL, but did not affect atherosclerosis (Engelbertsen et al., 2014). An alternative target of such adaptive immune‐targeted strategies is hsp60/65, a known target of pathogenic T cell responses in mice (Zhong et al., 2016). Others include β2‐GPI (George et al., 2004) and modified fibronectin (Dunér et al., 2011). Interestingly, different peptides from hsp60 had opposite effects on T cell‐driven immune responses and on atherosclerosis (Grundtman et al., 2015), suggesting that the peptide itself as well as the stimulatory profile of the APC and the environment can modulate outcomes. The development of peptides promoting a tolerogenic (anti‐atherogenic) response could form the basis for future targeted therapies for CVD. Similarly, a study in mice demonstrated the protective effect of immunization with Streptococcus pneumonia polysaccharides, which contain PC epitopes (Binder et al., 2003). In humans, combined analysis of eight observational studies suggests an association with pneumococcal polysaccharide vaccination and decreased risk of acute cardiovascular events in the older population (Ren et al., 2016). Recently, the AUSPICE trial was initiated to directly investigate the potential of this strategy (Ren et al., 2016).

The same IgM‐targeted epitopes of oxLDL (MDA, PC) are also present on apoptotic cells. Apoptotic cell treatment was successful in inducing a strong IgG anti‐PC response as well as boosting anti‐PC IgM levels, leading to reduced atherosclerosis (Grasset et al., 2015). This led to a decrease in circulating cholesterol through an as yet ill‐defined mechanism. Thus, antibody responses could directly modulate atherosclerosis at the level of circulating cholesterol upstream of vessel accumulation and oxidation. However, previous studies of lymphocyte deficient mice (Sage et al., 2014) and mice in which Tregs are depleted also show significant changes in circulating cholesterol levels (Klingenberg et al., 2013). The effects of the modulation of dendritic cell levels, the archetypal T cell stimulating APC, on cholesterol have also complicated interpretation of those studies in understanding the importance of adaptive immune reactions in atherosclerosis (Gautier et al., 2009). Conversely, direct modulation of cholesterol in Apoe ^−/− mice can modulate CD4⁺ T cell responses (Chyu et al., 2014). Vaccination for atherosclerosis is an appealing concept, and the integration of new knowledge both from ongoing human trials and from advanced interventions in mouse models will give enhanced insight into the most effective mechanism to boost protective adaptive immunity.

Conclusion

Atherosclerosis is affected at several stages by adaptive immune responses, providing many opportunities to target these responses to reduce disease progression. Protective influences defective in diseased individuals include humoral responses to modified LDL and Treg responses. There are diverse strategies in development to boost these pathways in humans. Alternative strategies to neutralize pro‐inflammatory adaptive responses include blocking critical co‐stimulatory pathways or other indirect anti‐inflammatory treatments such as B cell depletion. Entering the translation phase is an exciting and promising prospect, but critical unresolved issues include showing efficacy above existing therapies, such as statins, which may also have direct anti‐inflammatory effects on top of lipid lowering. In addition, our knowledge of the network interactions within the immune system and crosstalk with metabolic homeostasis must improve to develop accurately targeted treatments.

Author contributions

A.S. performed literature searches, designed the structure and prepared draft versions; A.S. and Z.M. wrote and prepared the final manuscript.

Conflict of interest

The authors declare no conflicts of interest.

Sage, A. P. , and Mallat, Z. (2017) Readapting the adaptive immune response – therapeutic strategies for atherosclerosis. British Journal of Pharmacology, 174: 3926–3939. doi: 10.1111/bph.13700.