Inflammatory processes in cardiovascular disease: a route to targeted therapies

Abstract

Inflammatory processes are firmly established as central to the development and complications of cardiovascular diseases. Elevated levels of inflammatory markers have been shown to be predictive of future cardiovascular events. The specific targeting of these processes in experimental models has been shown to attenuate myocardial and arterial injury, reduce disease progression, and promote healing. However, the translation of these observations and the demonstration of clear efficacy in clinical practice have been disappointing. A major limitation might be that tools currently used to measure ‘inflammation’ are insufficiently precise and do not provide information about disease site, activity, or discriminate between functionally important activation pathways. The challenge, therefore, is to make measures of inflammation that are more meaningful, and which can guide specific targeted therapies. In this Review, we consider the roles of inflammatory processes in the related pathologies of atherosclerosis and acute myocardial infarction (AMI), by providing an evaluation of the known and emerging inflammatory pathways. We highlight contemporary techniques to characterize and quantify inflammation, and consider how they might be used to guide specific treatments. Finally, we discuss emerging opportunities in the field, including current limitations and challenges that are the focus of ongoing study.

Inflammation and its failure to resolve are firmly established as central to the development and complications of several cardiovascular diseases^1–3. Elevated levels of markers of inflammation, such as C-reactive protein (CRP) and serum amyloid A (SAA), have been shown to be predictive of future cardiovascular events across a range of clinical settings^4–6. The role of the innate immune system in the pathogenesis of cardiovascular diseases has been an area of particular focus, with the appreciation that targeting innate immune function in experimental models can variously attenuate disease progression and injury, and promote healing^7–10.

Processes of inflammation contribute to a broad range of cardiovascular diseases; however, in this Review, we consider the roles of inflammatory processes in the related pathologies of atherosclerosis and acute myocardial infarction (AMI), by providing an evaluation of known and emerging inflammatory pathways that are thought to be central to their pathogenesis, and the consequences of their activation. We highlight contemporary techniques to characterize and quantify inflammation, and consider how these might be used to guide specific treatments. Finally, we discuss emerging opportunities in the field, including current limitations and challenges that are the focus of ongoing study.

As interest in this field has expanded, one of the main challenges has been the use of the term ‘inflammation’ to cover a wide variety of processes and observations, some of which are very poorly defined. For instance, ‘inflammation’ can include observation of nonspecific changes in acute-phase proteins (such as CRP and SAA) that are produced in the liver, remotely from the site of pathology, but also encompasses a nuanced understanding of local molecular and cellular processes in the heart or vessel wall that are intimately involved in disease causation. Therefore, our approach is to try to identify and elaborate on mechanistically important inflammatory processes, to link these to clinically relevant biomarkers, and to focus on pathways that are therapeutically tractable.

A major limitation has been that the biomarkers that are currently used to gauge inflammation are crude and nonspecific (for example, CRP). They reflect the downstream consequences of inflammatory activity, but do not provide information relating to the site(s) of activation and cannot be used to discriminate functionally important activation pathways. The challenge, therefore, is to make measures of inflammation that are meaningful to the pathology of interest and which can guide specific, targeted therapies.

Inflammation and cardiovascular disease

Following the first observations of the role of inflammatory processes in cardiovascular disease³, our understanding of these complex processes has greatly expanded. Their complexity and central roles in the pathogenesis of cardiovascular disease are now at least partially appreciated. Indeed, the term ‘inflammation’ is now used to describe a broad range of processes including those at the site of disease, in the blood, at remote sites, and also as downstream sequelae of disease — with each representing very different biology. Importantly for clinical purposes, techniques have not evolved to enable the reliable, specific, and accurate characterization of these processes.

Acute myocardial infarction

Pathogenesis

AMI most commonly results from the acute rupture of a coronary atherosclerotic plaque resulting in the rapid formation of thrombus in the infarct-related epicardial artery and loss of blood flow distal to the site of occlusion¹¹. Other causes include plaque erosions and calcific nodules¹². Myocardial ischaemia results in the activation of a well-orchestrated immune response, both locally at the site of injury, but also in the circulating blood and at remote sites (FIG. 1). These processes are important as both mediators of injury and subsequently in repair and recovery². Elevated plasma levels of nonspecific indicators of inflammation (such as high-sensitivity CRP and IL-6) have indicated that their continued upregulation is associated with worse outcomes^13,14.

Figure 1.

Biological pathways central to the pathogenesis of acute myocardial infarction (AMI). Immediately following AMI, a number of local processes are activated with release of reactive oxygen species and cytokines with infiltration of circulating neutrophils and monocytes resulting in acute myocardial injury. Simultaneously a number of remote sites are also activated (e.g. spleen, bone marrow) via signalling pathways that result in further activation of the immune system and injury. Following this, a reparative phase ensues predominantly mediated by monocytes and T-lymphocytes resulting in tissue repair and recovery with upregulation of processes involved in angiogenesis and extracellular matrix deposition. Abbreviations: ROS: reactive oxygen species; TLR: toll-like receptors; DAMPS: damage associated molecular patterns; HSP: heat shock proteins; HMGB1: high mobility group box 1 protein; VCAM: vascular cell adhesion molecule; NLR: NOD-like receptor; NLRP3: NOD-like receptor family pyrin domain containing 3; IL: interleukin; TNF: tumour necrosis factor; IFN: interferon; CX3CR1: CX3 chemokine receptor 1; miRNA: micro ribonucleic acid; CRP: C reactive protein; SAA: serum amyloid A.

Inflammatory pathways locally in the myocardium

Myocardial ischaemia leads to cardiomyocyte injury and, if prolonged, death, with release of their intracellular contents and changes to the extracellular matrix. This process results in endogenous signals that trigger a cytokine ‘burst’, activation of platelets that mediate leukocyte activation and their tissue infiltration¹⁵, and the rapid recruitment of neutrophils to the myocardium, which following degranulation cause direct injury to endothelial cells and production of reactive oxidation species (ROS), proteases, and cytokines¹⁶. In response, endothelial cells increase expression of leukocyte and platelet adhesion molecules and compromise of intercellular tight junctions facilitates the binding and transmigration of circulating monocytes¹⁷ that coincides with a reduction of in-situ neutrophils secondary to a combination of local cell death and egress from the site^18,19.

These endogenous signals are often referred to as ‘danger associated molecular patterns’ (DAMPs) and include heat-shock proteins²⁰, high-mobility group protein B²¹, low-molecular hyaluronic acid²², and fibronectin fragments²³, which can all activate innate immune pathways, including Toll-like receptors (TLRs)²⁴, nucleotide-binding domain leucine-rich repeat containing receptors (NLRs), the complement cascade, and reactive oxygen species (ROS) production.

Of the TLRs, cardiomyocyte expression of TLR4, also a major feature of activated macrophages, is substantially increased in response to DAMPs²⁵. TLR4 deficiency is associated with decreased infarct size, a reduction in systemic inflammation²⁶, and improved left ventricular remodelling²⁷ in experimental models. In patients after AMI, TLR4 activation in monocytes (the precursors of macrophages) is associated with the development of heart failure²⁸. Conversely, TLR2 deficiency is associated with reduced myocardial fibrosis after AMI and improved left ventricular remodelling²⁹. TLR4, therefore, seems to be an important component in the initial inflammatory response, and TLR2 central to repair and left ventricular remodelling³⁰.

Among the various NLRs, the NLRP3 (NACT, LRR and PYD domains-containing protein 3 inflammasome complex has been identified as an important mediator of injury after myocardial ischaemia and includes NLRP3, apoptosis-associated speck-like protein containing a CARD, and the cysteine protease caspase 1³¹. The activated NLRP3 inflammasome binds and activates caspase 1 that converts interleukin-1β (IL-1β) to its active form, and is central to the activation of other cytokines³². The complement system consists of cell membrane (and plasma) proteins, many of which circulate as pro-enzymes. Complement activation has an important role in mediating neutrophil and monocyte recruitment acutely to the injured myocardium by the downstream complement effectors C3a and C5a, and the release of monocyte chemotactic protein 1 (also known as C-C motif chemokine 2; CCL2 by vascular smooth muscle cells, microvascular endothelial cells, and macrophages^33–36.

ROS are also released immediately by ischaemic cardiomyocytes after AMI, and induce leukocyte chemokine upregulation³⁷, resulting in an increased capacity of endothelial intercellular adhesion molecule 1 (ICAM1) ligands to bind to neutrophils³⁸, complement activation³⁹, and the recruitment of additional leukocyte subpopulations that subsequently become activated. Cytokines also contribute to the pathophysiology of AMI, with tumour necrosis factor (TNF), IL-1β, and IL-6 all having central roles. For example, TNF can promote the release of other proinflammatory chemokines and adhesion molecule synthesis in the infarcted myocardium, augmenting further leukocyte infiltration⁴⁰. TNF has interestingly also been shown to be protective of apoptosis, though its net actions are thought to be deleterious in AMI⁴¹. IL-1β similarly mediates injurious processes and leukocyte infiltration⁴². IL-6 is also upregulated after myocardial ischaemia and modulates its actions through activation of the JAK/STAT (Janus tyrosine kinase/signal transducer and activator of transcription) cascade^43,44.

Activation of remote inflammatory pathways and processes

Initially, ‘classical’ inflammatory monocyte subsets are mobilized and recruited from the blood, but in experimental models, there follows a rapid and massive mobilization of monocytes from the spleen and bone marrow⁴⁵, which is also likely to occur in humans⁴⁶. The precise stimuli to effect remote monocyte mobilization are incompletely understood, but include β₃-adrenoceptor activation in the bone marrow⁴⁷ and local action of CCL2 via C-C chemokine receptor type 2 (CCR2)¹⁷.

After recruitment to the injured myocardium, blood-derived macrophages in synergy with resident macrophages initially function to scavenge necrotic debris and are active in local proteolysis and phagocytosis⁴⁸. This stage is followed by a second phase consisting of the accumulation of nonclassical macrophages promoting reparative processes, such as angiogenesis and extracellular matrix deposition^17,49. The precise roles of the various cardiac macrophage subsets either in the steady state or after AMI remain largely unknown. Additionally, the macrophages also communicate with other cell types (such as B cells, neutrophils, and mast cells) to exert further local and remote actions in both tissue injury and repair^50,51. Leukocyte numbers then return to baseline within 2 weeks in both the heart and blood.

Role of adaptive immunity

In contrast to the innate immune system (consisting primarily of monocyte-derived cells), which provides immediate defence against tissue injury by responding to general factors without pre-exposure, the adaptive immune system (that is, lymphocytes) provides the capacity to react to specific antigens after exposure, and to maintain memories of them.

In AMI, lymphocytes constitute a very small proportion of the infiltrating cell population⁵². However, the higher systemic frequency of activated T lymphocytes in patients with AMI compared with those with stable angina^53,54 suggests that the adaptive immune system is also activated. In particular, CD4⁺CD28^null T-effector cells are increased in the peripheral blood of patients after AMI⁵⁵, and release proinflammatory cytokines (such as interferon-γ), activating monocytes and tissue macrophages, with the level of increase positively correlated with the risk of future acute coronary events⁵⁶. Conversely, CD4⁺CD25⁺ regulatory T cells (T_REG) release anti-inflammatory cytokines, such as IL-10. After AMI, there seems to be an imbalance in T-cell subsets, with a reduction in T_REG numbers⁵⁷. This imbalance might indicate that, in these individuals, the counter-regulatory response to the activation of effector T cells is impaired. T lymphocytes have also been shown to have a role in recovery after AMI in experimental models, with the activation of CD4⁺ T cells required for optimal wound healing and remodelling of the left ventricle⁵².

Evidence of inflammation at remote sites

The activation of the immune system after AMI also results in general systemic inflammation, as evidenced by increased plasma levels of inflammatory cytokines, which are positively correlated with adverse outcomes⁵. Observational human studies have also demonstrated that plasma levels of TNF are strongly associated with the occurrence of reperfusion injury after recanalization of the infarct-related artery⁵⁸, and with an increased risk of death⁵⁹. Levels of TNF, IL-1, and IL-6 are significantly elevated in patients with AMI complicated by cardiogenic shock when compared with patients with uncomplicated AMI⁶⁰. The extent to which systemic levels of cytokines give meaningful insight into the state of local inflammatory pathways is uncertain, but the complexities of the local processes of inflammation in injury and repair seem to argue against a predictive approach dependent on a single circulating reporter biomarker.

Targeting of biological processes

The ability to target therapeutically these inflammatory central biological processes could potentially attenuate myocardial injury, augment repair and recovery, and reduce the risk of future events. For example, in experimental myocardial infarction, the targeting of inflammatory ‘classical’ monocytes to reduce the total number of monocytes recruited to infarcted, inflamed myocardium by splenectomy or by the administration of angiotensin-converting-enzyme inhibitors has been shown to result in a reduction in infarct size and improved left ventricular function⁶¹. Similarly, targeting their recruitment with the administration of CCR2-siRNA delivered by nanoparticles also resulted in a reduction in the number of recruited monocytes, with a consequent reduction in infarction size and improved left ventricular function^8,62. On the basis of these (and other) promising experimental data, a number of agents have been or are currently being investigated in clinical trials (TABLE 1).

Table 1.

Anti-inflammatory agents in acute myocardial infarction

Target Drug	Preclinical studies			Clinical studies
	Animal model	Outcomes	Result	Population	Outcomes	Result
IL-6 Tocilizumab	Mice	Improved LV remodelling163	✓	117 patients with NSTEMI	Reduction in troponin T level164	✓
IL-1β Canakinumab	Mice	Improved LV remodelling and reduction in cardiomyocyte apoptosis after AMI165 Impaired LV remodelling and suppression of pro-collagen gene expression166	–	NA	NA	NA
IL-1R Anakinra	Mice and rats	Improved LV remodelling and reduction in cardiomyocyte apoptosis after AMI167	✓	182 patients with NSTEMI 20 patients with STEMI	Reduction in hsCRP at 14 days, same event rate at 30 days, significant excess in MACE in IL-1Ra group at 1 year168 No difference in clinical outcomes, reduction in incidence of heart failure with treatment169	-
Tumour necrosis factor Etanercept	Rats	Improved LV ejection fraction after AMI, significant reduction in infarct size170	✓	26 patients with AMI	Reduced neutrophil count and IL-6 level, increased monocyte–platelet aggregates171	–
MMP-2 PG-116800	Mice	Reduction in leukocyte infiltration and cardiac rupture, and improved survival rate172	✓	203 patients with STEMI and low LV ejection fraction	No improvement in LV ejection fraction at 6 months173	–
MMP-2, MMP-9 Doxycycline	Mice	Reduction in LV dilatation after AMI174	✓	110 patients with STEMI	Reduction in ventricular dilatation at 6 months175	✓
C-reactive protein	Pigs	Removal of C-reactive protein reduced infarct size176	✓	NA	NA	NA
P-selectin Inclacumab	Dogs	Significant reduction in infarct size and neutrophil infiltration177	✓	544 patients after NSTEMI	Reduction in biomarkers after PCI178	✓
α1-Antitrypsin	Mice	Reduction in caspase 1 activity, preservation of viable myocardium, and improved LV remodelling179	✓	10 patients with STEMI	Reduction of hsCRP level180	✓
C3/C5 (complement) Pexelizumab	Rats	Reduction in myocardial infarction size by 44%181	✓	5,745 patients with STEMI	No difference in outcomes with treatment182	–
Integrin / CD18	Dog Primate	Nearly 50% reduction in infarction size183 Reduction in infarct size184	✓	420 patients with STEMI 394 patients with STEMI	No difference in infarct size185 No difference in coronary blood flow, infarct size, or rate of ECG normalization186	–
Immunoglobulin	Rats	No improvement in survival or LV function187	–	62 patients with STEMI	No effect on LV remodelling at 6 months188	–
P38 MAPK Losmapimod	NA	NA	NA	3,503 patients with AMI	No difference in outcomes189	–

AMI, acute myocardial infarction; ECG, electrocardiogram; hsCRP, high-sensitivity C-reactive protein; LV, left ventricular; MAPK, mitogen-activated protein kinase; MMP, matrix metalloproteinase; NA, not available; NSTEMI, non-ST-segment elevation myocardial infarction; PCI, percutaneous coronary intervention; STEMI, ST-segment elevation myocardial infarction; ✓, positive study; –, neutral study; X, negative study.

To date, however, clinical efficacy has not been demonstrated with any agent. The reasons for this lack of success are likely to be multifactorial. One possible reason might be that the experimental model that has been used in initial therapeutic evaluation is not representative of human disease. Secondly, a common theme to each of these studies is that, in general, enrolment has been on the basis of the crude diagnosis of AMI, without attempts at stratification or ‘mechanistic staging’. As discussed above, the underlying pathogenesis of AMI is a complex interaction between multiple biological processes. In combination with the historical ineffectiveness of general anti-inflammatory drugs in this setting, a ‘one size fits all’ approach might not be appropriate to determine efficacy, and a more-targeted approach to identify patients who would benefit from specific interventions is required.

Molecular techniques

The quantification of a number of different blood components has commonly been used to measure ‘inflammatory’ processes and has included soluble adhesion molecules (such as ICAM1, P-selectin, and vascular cell adhesion molecule 1 [VCAM1]), cytokines (such as IL-1β, IL-6, and TNF)⁶³, SAA⁶⁴, CRP⁶⁵, and peripheral white cells^66,67. With greater understanding of the underlying biology, it is increasingly apparent that these markers actually represent very different biological processes, can be ‘upstream’ or ‘downstream’ of the pathological process of interest, might not be specific, and provide little indication of activity, site, or status. Levels of high-sensitivity CRP, for example, have been observed to be acutely elevated after ischaemic events⁶⁸ and, when chronically elevated, have been shown to be associated with an increased risk of future vascular events^69,70. Mendelian randomization studies, however, have indicated that no causative relationship exists between single nucleotide polymorphisms associated with elevated CRP levels and coronary heart disease⁷¹, and the specific targeting of the molecule was not associated with benefit in preclinical studies⁷².

One approach to improve characterization of circulating immune cells after AMI has been to use flow cytometry-based techniques. This powerful approach has enabled more-precise phenotyping of leukocyte populations (such as identification of monocyte subsets) and considerable heterogeneity in both phenotype and function has been observed. Using monocytes as an example, flow cytometry can be used to phenotype the total cell population into subsets on the basis of the cell-surface expression of CD14 and CD16⁷³, with CD14⁺⁺CD16^– monocytes that express CCR2 thought to be ‘inflammatory’. However, even with these more sophisticated phenotyping techniques, measuring circulating levels of individual soluble biomarkers or the identification of changes in crude counts of leukocyte subtypes is unlikely to provide sufficiently granular information to inform targeted drug selection.

To this end, transcriptomic analysis of circulating monocytes after AMI has identified patterns of gene expression that are conserved between mice and humans³⁰. Although not yet formally tested, defined patterns of gene expression in downstream effector cells might plausibly give a more-precise insight into the activity of relevant pathways than measuring soluble biomarkers of uncertain origin and or site of action.

In AMI, circulating monocytes display patterns of gene expression in common with ‘inflammation’ and ‘mitosis’, with CD14, IL-1, and TLR2 identified as highly relevant components of the response using gene-set-enrichment techniques^30,74. Analogous approaches have been used to validate preclinical models of disease, identify new therapeutic targets³⁰ and, clinically, gene sets in circulating leukocytes have been used to diagnose cardiac transplantation rejection⁷⁵. Such approaches also borrow from the field of oncology, where they have enabled the reclassification of many malignancies on the basis of molecular signatures that have different prognostic and therapeutic responses⁷⁶. In breast cancer, microarray analysis of frozen breast tissue has been used to develop a 70-gene panel called Mammaprint, which can provide a prognostic measure of the likelihood of distant metastasis within 5 years of completing treatment⁷⁷. Technological advances have enabled single-cell analysis to provide information about the heterogeneity of the human bone marrow cell population against which diseased states and response to drug treatment could be compared as a basis of mechanistic studies⁷⁸. Complementing transcriptome profiling, proteomic studies have also been used to investigate changes in serum protein levels and, using this approach, haptoglobin has been identified a possible prognostic biomarker after AMI⁷⁹. Similarly, metabolomics⁸⁰ and lipidomics⁸¹ methods, although in their infancy, can be also used to identify novel therapeutic targets, and possibly monitor disease activity. Future approaches might enable incorporation of complex, integrated datasets spanning multiple -omics domains and with much greater sampling frequency to allow higher resolution of temporal changes than is currently possible.

Extracellular vesicles have also gained interest as both a diagnostic tool and a potential therapeutic target. They have been shown to be important in cell-to-cell signalling, including in the context of inflammation. Extracellular vesicles carry a variety of cell-derived molecules including proteins⁸², mRNA, and microRNA⁸³, and include exosomes (30–100 nm), multivesicular bodies (also called ectosomes or microparticles: 100–1,000 nm)⁸⁴ and apoptotic bodies (1–5 μm)^85,86. Extracellular vesicles are continually produced by all cell types and, in the setting of AMI, those from endothelial cells and platelets have been shown to have roles in homeostasis of the cardiovascular system including coagulation⁸⁷ and tissue repair⁸⁸. After AMI, levels of multivesicular bodies increase⁸⁹ and are likely to exert their actions through mediating cell-to-cell signalling via transport of bioactive proteins between cells⁹⁰ resulting in direct activation of target cells^91,92. Several miRNAs found in extracellular vesicles might have roles in cardiovascular disease⁹³, for example miR-150 released from monocytes that enhances cell migration⁹⁴. Extracellular vesicles might, therefore, be a tool to predict future risk⁹⁵, determine disease activity, be a therapeutic target, or provide templates for the engineering of vesicles that consist of a payload with a specific, targeted function.

Imaging techniques

Measuring blood biomarkers, in however sophisticated a way, cannot provide precise information on the distribution of pathology or accurately assess its anatomical extent. In combination with molecular tools, imaging modalities can be employed to determine disease activity and site. MRI is commonly used in the clinical setting and provides information about location and extent of acute myocardial injury, with quantification of myocardial oedema and also of irreversible injury with late gadolinium enhancement indicating myocardial scar.

The administration of ultra-small particle of iron oxide seems to improve specificity with which the location of infiltrating macrophages can be identified^96,97. In experimental models of AMI, the application of ¹⁸fluorodeoxyglucose (¹⁸FDG) PET in combination with MRI further improved localization of infiltration of metabolically active leukocytes in experimental models⁹⁸, with cellular imaging studies supporting these conclusions⁹⁹. Clinical studies have demonstrated similar findings, which provide the capacity to quantify metabolic activity in the myocardium after infarction¹⁰⁰, monitor continued disease activity, and might also provide a prognostic marker of functional outcome¹⁰¹.

Further information that we have termed ‘mechanistic staging’ might be obtained from quantitative molecular imaging techniques that target important known processes, ideally on a mechanistic pathway that can be paired to a specific therapeutic agent or intervention. For example, quantitative molecular imaging has been described for VCAM1 and P-selectin using ‘leukocyte-mimetic’ particles to examine extent and distribution of endothelial activation^102,103, which has also been applied to measure the ‘imprint’ of previous ischaemia¹⁰⁴. An alternative approach has been to image proteases that are critical in the pathogenesis of disease and has been applied to preclinical models¹⁰⁵ and humans after AMI¹⁰⁰. Therefore, by utilizing a combination of molecular and imaging techniques, it might be possible not only to identify therapeutic targets, but also to identify patients with ‘active’ disease, localize sites of activity, and monitor response to therapy.

Atherosclerosis

Pathogenesis

Atherosclerosis is a chronic inflammatory process¹⁰⁶. Immune cells are important mediators from the earliest fatty streaks to late-stage complex plaques. As knowledge of the immune processes in atherogenesis has expanded, opportunities for therapy through their manipulation or modification have become apparent. However, tools to determine disease severity and activity remain fairly crude.

Endothelial cells, lymphocytes, smooth muscle cells, monocytes, and macrophages are all involved in the pathogenesis of atherosclerosis from earliest foam cell formation through to development of advanced plaques (FIG. 2).¹⁰⁷ Initial activation of the endothelium from disruptions to normal shear stress result in an increase in permeability to lipoproteins and upregulation of adhesion receptors¹⁰⁸, and facilitate deposition of lipid from apolipoprotein B (apoB)-containing lipoproteins in the subendothelial space. Endothelial activation also promotes recruitment of circulating monocytes that originate from either the bone marrow¹⁰⁹ or the spleen⁴⁵. Monocyte adhesion depends on the upregulation of a number of cell-adhesion molecules on the luminal surface of the endothelium,¹¹⁰ notably, ICAM1, P-selectin, and VCAM1; followed by the expression of three major chemokine families, CCR2, CCR5, and CX3C chemokine receptor 1 (CX3CR1), that facilitate the transmigration process^111,112. After recruited from the spleen and bone marrow¹¹³, monocytes can terminally differentiate into macrophages¹¹⁴, differentiate and locally proliferate into distinct functional phenotypes¹¹⁵, or directly influence the phenotype of in situ cells (for example, lesional macrophages)¹¹⁶. Lipids from retained apoB-containing lipoproteins are taken up by activated macrophages by a number of processes, including phagocytosis of aggregated LDL¹¹⁷, pinocytosis of LDL¹¹⁸, uptake of modified apoB-lipoproteins by scavenger receptors, and uptake by CD36. Ongoing lipid accumulation induces their transformation into macrophage-derived foam cells. As the foam cell population grows within lesions in the arterial wall, the rate of accumulation exceeds the rate of clearance¹¹⁹, and eventually the foam cells coalesce into a lipid-rich necrotic core.

Figure 2.

Biological processes central to the pathogenesis of atherosclerosis. Endothelial cells, lymphocytes, smooth muscle cells, monocytes, and macrophages are all involved in the pathogenesis of atherosclerosis from earliest foam cell formation through to development of advanced plaques. Initial activation of the endothelium from disruptions to normal shear stress result and facilitate deposition of lipid in the subendothelial space. Endothelial activation also promotes recruitment of circulating monocytes where they terminally differentiate into macrophage or differentiate and locally proliferate into distinct functional phenotypes. Activated macrophages take up lipid and results in their transformation into macrophage-derived foam cells. As the foam cell population grows within lesions in the arterial wall, the rate of accumulation exceeds the rate of clearance, and eventually the foam cells coalesce into a lipid-rich necrotic core. Abbreviations: VCAM: vascular cell adhesion molecule; ICAM: intercellular cell adhesion molecule; CCR: chemokine-chemokine receptor; CX3CR1: CX3 chemokine receptor; IL: interleukin; TNF: tumour necrosis factor; ApoB: Apoliporotein B; CRP: C reactive protein; SAA: serum amyloid A; MMP: matrix metalloproteinase; NLRP3: NOD-like receptor family pyrin domain containing 3; VSMC: vascular smooth muscle cells.

A number of secreted cytokines are also implicated in the pathogenesis of atherogenesis. Cells of both the innate (macrophages) and adaptive (T and B lymphocytes, dendritic cells) immune systems can modulate the local inflammatory milieu. Depending on their local environment, T lymphocytes can be stimulated to secrete proinflammatory T_h1 cytokines such IL-1, IL-6, and TNF; or T_h2 cytokines such as IL-4, IL-10, and IL-13, which promote resolution of inflammation. IL-1 and TNF signalling is primarily mediated by p38 mitogen-activated protein kinase (MAPK) / nuclear factor (NF)-κB pathways¹²⁰. Conversely, activation of IL-6 signals via signal transducing protein gp130, which activates JAK1 and STAT1 and 3¹²¹ resulting in the activation of endothelial cells and macrophages to produce adhesion molecules and chemokines¹²².

The physical properties within the plaque can also have an important role in the propagation of disease. Neovascularization in human plaques originating from the vasa vasorum is thought to contribute to intraplaque haemorrhage, which not only accelerates plaque expansion and inflammation, but also predicts future plaque rupture^123,124. Angiogenesis is driven by vascular endothelial growth factor, which is a major downstream target of hypoxia inducible factor (HIF) 1α that is induced by a hypoxic environment, especially within the large lipid-rich necrotic core present in advanced atheroma^125,126. Hypoxia and HIF1α expression has been shown in vitro to alter macrophage lipid handling and suppress cholesterol efflux via ABCA1 (the ATP-binding cassette subfamily A member 1) in both mouse¹²⁷ and human¹²⁸ macrophages. Moreover, in vitro and ex vivo experiments have implicated hypoxia and HIF1α in driving enhanced glucose uptake, metabolic activity, and polarization of macrophages in human atheroma¹²⁹.

In 2003, Rong and colleagues¹³⁰ reported that cholesterol-loaded vascular smooth muscle cells lost their differentiation markers and instead expressed macrophage markers. A number of experiments in mice, and notably humans^131–133, have produced in vivo data consistent with these phenomena: 30–40% of cells classified as plaque macrophages were of vascular smooth muscle cell origin. It is too early to discern the contribution of these cells to the pathological process during atherogenesis, but at least in cell culture, they do not share the molecular signature or the functional properties of activated macrophages¹³⁴.

Targeting of biological processes

The ability to characterize specifically and target the biological processes central to the pathogenesis of atherosclerosis might prevent progression of disease and development of vulnerable plaque. Even more exciting would be the regression of disease, a possibility suggested by a number of animal models and limited clinical data¹³⁵. Although the targeting of various pathways in experimental models of atherosclerosis has been associated with benefit, clear efficacy in clinical studies has not yet been demonstrated for any single agent (TABLE 2). This apparent lack of efficacy might be related to poor representation of human disease by preclinical models used for evaluation therapeutic agent, or a failure to target patients rationally based on mechanistic staging (disease stage, activity, and severity).

Table 2.

Anti-inflammatory agents in atherosclerosis

Target Drug	Preclinical studies			Clinical studies
	Animal model	Outcomes	Result	Population	Outcomes	Result
IL-1β Canakinumab	Rats	Reduction of carotid neointima formation after injury190	✓	17,200 patients with MI 189 patients with type 2 diabetes mellitus and pre-existing vascular disease	First occurrence of MACE, change in carotid plaque burden (ongoing)163 No effect on vascular structure and function, but trend towards retarded progression of atherosclerosis by MRI164	–
Recombinant IL-1ra	Apoe−/− mice	Reduction in fatty lesion formation191	✓	NA	NA	NA
Tumour necrosis factor Etanercept, infliximab	Apoe−/− Tnf−/− mice	Reduction in atherosclerosis plaque progression192	✓	2,126 patients with rheumatoid arthritis (meta-analysis)	Trend towards reduction in all cardiovascular events193	✓
IL-12 subunit p40 Ustekinumab	Apoe−/− Il12−/− mice	Reduction in atherosclerosis194	✓	3,117 patients with psoriasis (pooled analysis)	Overall cardiovascular events were similar between treatment and placebo groups195	–
IL-17A Secukinumab	Apoe−/− mice	Reduction in atherosclerosis lesion area196	✓	3,993 patients with psoriasis (pooled analysis)	Overall cardiovascular events were similar between treatment and placebo groups197	–
p38 MAPK (Losmapimod)	Ldlr−/− mice	Reduction in foam cell formation and expression of CCL2 and VCAM1198	✓	99 patients with atherosclerosis receiving stable statin therapy	Reduction in vascular inflammation in most inflamed areas and reduction in inflammatory biomarkers199	✓
CC-chemokines 35K protein to inactivate CC-chemokines	Apoe−/− mice	Reduction in atherosclerosis, macrophage content, and lipid content in plaque200	✓	NA	NA	NA
P-selectin Inclacumab	Rabbits	Reduced intimal hyperplasia after injury201	✓	380 patients with saphenous vein graft disease	No effect on saphenous vein graft disease progression202	–
VCAM1	Apoe−/− mice	Reduction in plaque formation and reduced infiltration of CD45+ cells203	✓	NA	NA	NA
sPLA2 Varespladib	Apoe−/− mice	Reduction in atherosclerosis204	✓	5,145 patients after ACS	Significant increase in the risk of AMI, no reduction in risk of recurrent events205	X
LpPLA2 Darapladib	Pigs	Reduction in atherosclerosis206	✓	15,828 patients with stable CAD	No reduction in the primary end point of cardiovascular death, myocardial infarction, or stroke207	–
Serine protease inhibition	Rabbits	Reduction in plaque growth after injury208	✓	NA	NA	NA
Colchicine	Rabbits	Reduction in lipid deposition in aortic plaque209	✓	532 patients with stable coronary disease	Significant reduction in major cardiac events210	✓
Methotrexate	Rabbits	Reduction in neoatheroma formation and intimal thickening211	✓	7,000 patients with stable CAD and diabetes mellitus or metabolic syndrome	Effects on nonfatal MI, nonfatal stroke, and death (ongoing)212	–

ACS, acute coronary syndrome; AMI, acute myocardial infarction; CAD, coronary artery disease; CCL2, C-C chemokine 2 (also known as monocyte chemoattractant protein 1); LpPLA₂, lipoprotein-associated phospholipase A₂; MACE, major adverse cardiovascular event; MAPK, mitogen-activated protein kinase; MI, myocardial infarction, sPLA₂, secretory phospholipase A₂; VCAM1, vascular cell adhesion molecule 1.

Molecular techniques

The relationship between high-sensitivity CRP (which has been used as a surrogate biomarker for ‘inflammation’) and atherosclerosis has been the subject of much interest. A number of studies and meta-analyses have demonstrated an association between elevated CRP levels and an increased risk of cardiovascular disease⁶⁵. Indeed, a reduction in CRP in the context of a statin trial was noted to be an independent predictor for outcome¹³⁶. In the JUPITER study¹³⁷ that used high-sensitivity CRP to identify patients at increased risk of vascular events (without elevated LDL-cholesterol levels), patients with high basal high-sensitivity CRP levels were found to be at a significantly increased risk of future vascular events. Although the association between high-sensitivity CRP and risk is well established, what exactly this biomarker represents is currently unclear.

Similarly, other circulating proteins including VCAM1¹³⁸, ICAM1¹³⁹ and P-selectin¹⁴⁰ have been measured in the plasma of patients and have been shown to be associated with the level of atherosclerosis. But again, in common with high-sensitivity CRP, what these ‘downstream’ measures represent in uncertain.

The application of cellular biomarkers has been of interest in atherosclerosis and, by virtue of their central role in pathogenesis, circulating monocytes have been the focus of research¹⁴¹. Using flow cytometry, the number of circulating CD14⁺CD16⁺⁺ monocytes has been shown to be inversely related to plasma HDL levels, and CD16⁺ monocytes are proportional to the severity of atherosclerosis¹⁴², whereas a reduction in the monocyte subpopulation is associated with a reduction in intima–media thickness¹⁴³. Again, these measures do not provide any information about to the underlying pathogenesis of disease. To elucidate the functional characteristics of these ‘inflammatory’ cells at the level of atherosclerotic plaque, techniques such as laser-capture microdissection (LCM) have been successfully employed to procure intralesional cells in a cell-type-specific and location-specific fashion. In mice, immuno-LCM coupled with downstream gene or protein expression analysis has been used to study lesional macrophages in various stages of atherogenesis, including atherosclerosis regression. The inflammatory state of plaque macrophages was shown to be dynamically regulated as the severity of disease varied¹⁴⁴. In humans, studies using LCM on surgical plaque explants has also identified genes implicated in lipid metabolism (such as FABP4 and LEP) and activation of the adipokine/peroxisome proliferator-activator receptor (PPAR) signalling pathways¹⁴⁵, as well as a repertoire of inflammatory gene ‘signatures’¹⁴⁶, that might be important in mechanistic staging. Early findings utilizing this approach seem to corroborate observations from genome-wide association studies as well as plaque compositional changes in plaque revealed by high-resolution MRI Chai, personal communication. Circulating multivesicular bodies derived from platelets, monocytes, and red blood cells have also been characterized specifically with regard to the miRNA that they contain, and have been shown to have roles in VCAM1 inhibition, activation of plaque macrophages, and in cell-to-cell signalling¹⁴⁷ that are central to atherosclerosis progression¹⁴⁸.

Imaging techniques

To overcome some of the limitations of molecular tools, imaging techniques can provide better characterization of pathogenesis while also providing information relating to the anatomical site of disease by specifically targeting components of the process of interest. Conventional imaging modalities (such as CT and MRI) focus on the presence and extent of disease with limited qualitative information (for example, plaque components), but in combination with molecular contrast agents with specific targets, provide the capacity to determine the presence and site of biological processes of interest. For example, MRI after the administration of recombinant HDL particles containing gadolinium¹⁴⁹ or ultra-small particles of iron oxide¹⁵⁰ can provide information about macrophage infiltration in plaque, or with microparticles of iron oxide targeting VCAM1 can provide information about endothelial activation — a critical step in atherosclerosis pathogenesis¹⁰³. The use of ¹⁸FDG-PET in combination with MRI can also improve characterization of plaque ‘activity’, with uptake indicating macrophage activity¹⁵¹; high carotid FDG uptake is an independent predictor of cardiovascular events in asymptomatic patients¹⁵². The use of other tracers, such as ¹⁸F-NaF, have also been used to identify microcalcification in the identification and localization of high-risk coronary plaques¹⁵³. Co-registration of fluorescence molecular tomography and CT with the aid of protease sensors has also been demonstrated to be useful in imaging atherosclerosis in a murine model, and has also been applied to determine response to therapy¹⁵⁴.

Intravascular fluorescence imaging techniques have increasingly been utilized in clinical practice with intravascular ultrasonography and optical coherence tomography (OCT) increasingly used to identify atherosclerotic plaque and diagnose rupture, erosions, calcific nodules, and the presence and distribution of calcification, lipid, and fibrous tissue¹². Targeted molecular agents have been used to identify lipid-rich, metabolically active plaque in rabbit arteries^155,156. Dual-modality OCT and near-infrared autofluorescence (NIRAF) using a specialized OCT–NIRAF catheter has been shown to be feasible and safe in humans and, indeed, NIRAF enhances the sensitivity to detect vulnerable thin-cap fibroatheroma¹⁵⁷.

In summary, the combination of molecular tools and imaging techniques provides the capacity to characterize active inflammatory processes, the sites of disease, and also enables longitudinal monitoring of disease progression and response to therapy.

Future implications

In combination with improved understanding, the ability to characterize and mechanistically stage underlying pathological biological processes will enable a more-tailored approach to the management of patients presenting with, or at high risk of, cardiovascular disease. New therapies will be focused on targeting specific pathological processes, especially those related to inflammation, at the appropriate stage of disease and on the basis of current activity. The same tools can also be used to monitor response to treatment¹⁵⁸ (including toxicity¹⁵⁹). Some preclinical examples include studies in which either IL-13 (which polarizes macrophages to the M2 state) or a ligand of resolving receptors on macrophages were effective in retarding the progression of atherosclerosis^160,161.

New biomarkers and imaging techniques could also be applied when evaluating the efficacy of existing and new therapies and be used as surrogate clinical end points (as opposed to the ‘hard’ clinical end points), resulting in increased statistical power of studies, a reduction in the number of patients that need to be enrolled, reduced cost, and a shorter timescale. Finally, the majority of reports to date using many of these novel tools have demonstrated only ‘proof of concept’. Future studies demonstrating clinical benefits associated with their application in conjunction with technological advances (for example, platform integration and miniaturization) enabling the feasibility of point-of-care testing are required to make these interventions more broadly applicable to patients beyond research centres¹⁶².

Conclusions

A number of biological processes contribute to the pathogenesis of cardiovascular disease. These have historically all been termed ‘inflammatory’, but this oversimplification does not accurately convey the heterogeneity of the processes involved. Experimental data suggest that the specific targeting of multiple biological processes can result in attenuation of injury and augmentation of reparative processes. However, despite these very promising results, clear efficacy of anti-inflammatory therapeutics has not been demonstrated in clinical practice. Currently available tools to measure and monitor inflammation are often nonselective, represent downstream sequelae of inflammation, and do not provide any information about presence, extent of activity, or location. By using improved informative techniques to characterize inflammation, a more selective, rational targeted therapeutic approach can be employed. Future efforts will focus on these areas, with a goal of personalized medicine to complement existing treatment strategies, and the ultimate aim of improving clinical outcomes.

Key points.

• Inflammation and its failure to resolve are firmly established as central to the development and complications of several cardiovascular diseases

• Targeting of inflammatory processes in experimental models has been demonstrated to be beneficial in attenuating myocardial and arterial injury, reducing disease progression, and promoting healing, but clinical translation has been disappointing

• Current tools to measure ‘inflammation’ are nonspecific and represent downstream sequelae of biological processes, but provide little insight into disease state, site, or activation pathways

• Contemporary molecular techniques (such as proteomics and gene-expression analysis) improve our ability to characterize underlying biological processes, and identify activation pathways as biomarkers and as a basis to develop new therapeutics

• Noninvasive imaging tools enable the identification of activation of specific pathways and their sites, and can be used to monitor response to therapy

Biographies

Neil Ruparelia received his medical degree from University College London, UK, and obtained his DPhil from the University of Oxford, UK, after being awarded a clinical research training fellowship from the British Heart Foundation. He is a clinical cardiologist and has completed fellowships in coronary and structural intervention in Oxford, London and Milan. His principal research interests are focused on the pathogenesis of acute myocardial infarction. He is the recipient of the President’s Medal in cardiology awarded by the Royal Society of Medicine, and awards from the American Heart Association and European Society of Cardiology.

Joshua T. Chai is an interventional cardiology fellow in London, UK. After graduating from Cambridge University and clinical training in London, he received his DPhil from University of Oxford on a Medical Research Council funded Clinical Research Training Fellowship. His main research interests are in atherosclerosis imaging and plaque biology. His doctoral research focused on integrated characterization of carotid atherosclerosis using novel MRI techniques, combined with plaque macrophage transcriptomic analysis utilizing laser capture microdissection. He is the recipient of a Young Investigator Award from the European Atherosclerosis Society; and a Finalist of the AHA Melvin Judkins Young Investigator Award.

Edward A. Fisher is the Leon H. Charney Professor of Cardiovascular Medicine at the NYU School of Medicine, USA, where he directs the vascular biology research programme and the Preventive Cardiology Center. He is also a visiting professor of cardiovascular medicine at the University of Oxford, following his tenure as the George Eastman Professor. Dr Fisher received his MD from NYU and a PhD from MIT for a thesis on lipoprotein metabolism. His clinical training was at the Duke and Harvard Medical Schools. His research interests include the cell biology of lipoprotein formation, the regression and imaging of atherosclerosis, and the development of nanomedicines to target plaque inflammation.

Robin P Choudhury is Professor of Cardiovascular Medicine at the University of Oxford, where he is director of the Acute Vascular Imaging Centre and an interventional cardiologist. His training was in Oxford, London, and New York. He is a Fellow of Balliol College and of the Royal College of Physicians. His interests are in the characterization of patients with atherosclerosis and vascular inflammation and in particular in the development, validation, and application of new approaches to vascular imaging. He is the principal investigator of the Tripartite Immunometabolism Consortium (TrIC).