Inflammation reprogramming and immunomodulation: Next-generation treatments for atherosclerosis

Summary

The current generation of highly successful atherosclerosis treatments, such as low-density lipoprotein (LDL)-cholesterol reduction, blood pressure management, and smoking cessation, has largely focused on ameliorating factors perceived to drive incident disease and its complications. The adverse contributions of these factors have typically been identified through epidemiological studies. The therapeutic strategies that arose in response focused on risk factors for disease development and tended to overlook the fact that patients already have established disease, by the time of presentation. However, by capitalizing on contemporary biological knowledge and technologies, it is becoming increasingly possible to shift from a model based on population-derived risk factor management to next-generation treatments (including monoclonal antibodies, small interfering RNA [siRNA], mRNA, epigenetic reprogramming, and gene editing) for atherosclerosis that are tailored to patient-level disease processes, informed by mechanistic characterization, offer potential to reverse or regress disease, and incorporate systems-level interventions that extend beyond the atherosclerotic plaque.

Graphical abstract

Choudhury et al. explore next-generation atherosclerosis therapies. They identify a shifting focus from risk factor management alone to patient-specific, mechanistically informed treatments. Leveraging advanced technologies (cell therapies, monoclonal antibodies, siRNA, mRNA, and gene editing), emerging approaches intend to reverse or regress established disease and incorporate systems-level interventions beyond the atherosclerotic plaque.

Introduction

Atherosclerosis is a chronic disease that is initiated by the deposition and retention of apolipoprotein B (ApoB)-containing lipoproteins in the walls of medium-sized and large arteries, notably the coronary arteries. Processes of atherogenesis are rehearsed in detail elsewhere.^1,2 In a persuasive “arterio-centric” hypothesis, atherosclerosis reflects a “response to retention.”³ In other words, extravasated lipoproteins become associated with the extracellular matrix, modified by oxidation, and drive an immune/inflammatory response. As the intimal lesion expands, smooth muscle cells and macrophages proliferate. The lesion is further elaborated by continued monocyte recruitment (and their differentiation to macrophages), smooth muscle cell migration, transdifferentiation, replication, and extracellular matrix production. Some ApoB on low-density lipoprotein (LDL) is modified in the extracellular space, consequently becoming DAMPs (damage-associated molecular patterns) with the potential to provoke inflammatory responses by macrophages. In another process, having internalized oxidized LDL particles, macrophages process them and present peptides derived from ApoB with major histocompatibility complex class II molecules. A T cell recognizes this complex through its specific T cell receptor. This interaction is further facilitated by the binding of co-stimulatory ligands to their corresponding receptors on the T cell surface. In response to co-stimulatory signals and cytokines (including interleukin [IL]-1β, IL-6, and tumor necrosis factor alpha) secreted by the antigen-presenting cells (primarily macrophages, but also dendritic cells), T cells activate transcription factors that promote differentiation into distinct T helper (Th) subsets. These Th cells then secrete specific cytokines that can exert either an athero-protective or pro-atherogenic effect.

Activated T helper 1 (Th1) lymphocytes produce interferon (IFN)-γ, which activates macrophages and exacerbates atherosclerosis. Other cell types generate protective mediators: T helper 2 (Th2) lymphocytes release the anti-inflammatory cytokine IL-10, and regulatory T (Treg) cells produce transforming growth factor β.⁴ Mononuclear phagocytes eliminate dying or dead cells undergoing apoptosis through a process called efferocytosis. When efferocytosis is inefficient, cellular debris accumulates, contributing to the formation of the central lipid core in atherosclerotic plaques.⁵ The overall disease process reflects a complex interleaving of pathological drivers and remedial homeostatic processes, some of which are inadequate or maladaptive.⁵ While the underlying processes are chronic, spanning decades, the complications of atherosclerosis, due to plaque erosion or rupture and consequent thrombotic occlusion, can be acute and catastrophic. Accordingly, therapies directed at atherosclerosis have focused on retarding the progression of plaque, by targeting clinically relatable factors relevant to its pathogenesis, for instance, through the aggressive lowering of LDL cholesterol, but it has become increasingly clear that additional systemic metabolic and immune factors also play important roles⁶ (Figure 1).

Figure 1.

Processes driving atherosclerosis and opportunities for therapeutic intervention

In the top left, the arterial wall is shown, highlighting the recruitment of immune cells such as monocytes and their differentiation into macrophage foam cells, along with the activation of T helper (Th) cells and B cells. Vascular cell adhesion molecule (VCAM) expression and the production of pro-inflammatory cytokines, including interleukin-1 (IL-1) and interleukin-6 (IL-6), contribute to the propagation of vascular inflammation. This section highlights where processes of inflammation can be targeted by interventions to ameliorate disease progression. The top right of the figure depicts mechanisms that promote the resolution of inflammation and the healing of atherosclerotic plaques. Regulatory T cells (Tregs) and regulatory B cells (Bregs) are involved in suppressing excessive immune responses, supporting plaque stabilization and repair. The liver is shown (bottom left) as a key organ in both the production of atherogenic apolipoprotein B (ApoB)-containing lipoproteins and the clearance of low-density lipoprotein cholesterol (LDL-C) via the low-density lipoprotein receptor (LDLR). Therapeutic approaches such as those targeting proprotein convertase subtilisin/kexin type 9 (PCSK9) enhance LDL clearance. Systemic inflammation is often measured by levels of C-reactive protein (CRP), which is made by the liver. The bone marrow (bottom right) is increasingly recognized as a contributor to atherosclerosis and its complications. Clonal hematopoiesis, particularly involving mutations in genes ten-eleven translocation methylcytosine dioxygenase 2 (TET2) and DNA methyltransferase 3A (DNMT3A), is potentially causally related to atherogenesis. These processes, together with trained immunity and enhanced myelopoiesis, amplify systemic and vascular inflammation.

Over the past 30 years, there has been a growing appreciation of “inflammation” as an important factor in atherogenesis and its complications.⁷ We have previously argued that the term inflammation is insufficiently precise to capture, or describe, the complexity of concomitantly active processes that span multiple organs and tissues and involve contributions from a huge variety of inter-communicating and competing cell types.⁸

A deeper understanding of plaque biology has been powerfully advanced by contemporary techniques, not least RNA sequencing. Unsupervised analyses of bulk sequencing of human carotid plaques at scale have suggested a new transcriptome-based functional categorization,⁹ while single-cell RNA sequencing has revealed ever-deepening complexity, including the identification of novel cell subtypes.^9,10,11,12 The addition of topographically resolved gene expression in cells obtained by laser capture microdissection¹³ and latterly spatial transcriptomic analyses has brought further mechanistic insight,¹⁴ though there are still limitations around the degree of resolution and sequencing depth.¹⁵

Analysis and interpretation of large, complex “-omics” datasets will be deepened and accelerated by expanding shared data repositories and tools to uncover cellular function and characteristics,¹⁶ including artificial intelligence that can interpret and infer functions that are not limited by the relatively crude “manual” annotations of static pathways.^17,18 In theory, AI can be used to build predictive models for treatment response based on single-cell sequencing data. These models can anticipate how specific cell types or patients will respond to drugs, helping to personalize treatment options and optimize therapeutic strategies.¹⁹ While powerful in principle, the accrual of relevant datasets is partly constrained by limitations due to the inaccessibility of disease-relevant tissues extracted from plaques that reflect disease-relevant processes. The major medium-term challenge might, therefore, not be in the rational identification of therapeutic targets or even in the design of agents to target them but in identifying the patients in whom such interventions might be beneficial. Unlike cancer, where the diseased tissue can be excised and its cellular and molecular characteristics determined, atherosclerotic plaques are inaccessible, and the molecular and cellular workings in individual patients are largely obscure. The search for windows into plaque biology through circulating factors (protein biomarkers, metabolites, peripheral blood mononuclear cells, etc.) has been keen but, so far, has not been influential in introducing or directing new treatments.

Here, we will review novel approaches to well-established clinical risk factors, e.g., through LDL-cholesterol reduction and will attempt to provide a contemporary context for a next generation of approaches, some of which extend far beyond the atherosclerotic plaque and draw in systems-level drivers of disease and opportunities for new classes of therapeutic intervention.

Novel approaches to LDL reduction

The central role for intensive LDL-cholesterol reduction in the primary and secondary prevention of cardiovascular disease is firmly established and accepted. Knowledge of the pathways that govern hepatic cholesterol biosynthesis and LDL receptor activity has allowed them to be effectively manipulated, by interventions targeting multiple points. In addition to HMG Co-A reductase inhibitors (statins) and bempedoic acid,²⁰ contemporary approaches target the LDL-receptor activity directly, e.g., through monoclonal antibodies to PCSK9 (proprotein convertase subtilisin/kexin type 9).^21,22 Small interfering RNA (siRNA) has emerged as an effective therapeutic tool in lipid management due to its ability to target and degrade messenger RNA (mRNA) transcripts, effectively “turning off” gene expression. Incliseran is an siRNA, given by a 6-monthly injection, that inhibits the translation of PCSK9 and lowers LDL cholesterol by around 50%.²³

Especially for statins, a large body of clinical evidence demonstrates efficacy and safety and suitability for long-term use.²⁴ Given that atherosclerotic cardiovascular disease remains prominent despite treatment, a next major question in respect of LDL-cholesterol reduction might not be how to reduce it but at which level (or other indication) to start treatment and how to best select patients who might benefit in the longer term.²⁵

A radical proof-of-principal approach has demonstrated that CRISPR base editing, delivered in vivo using lipid nanoparticles, led to near-complete knockdown of PCSK9 in the liver after a single infusion, with reductions in blood levels of PCSK9 and LDL cholesterol of approximately 90% and 60%, respectively, and which remained stably suppressed for at least 8 months.²⁶ In an alternative approach, Cappelluti et al. identified zinc-finger proteins as the most effective DNA-binding platform for silencing mouse Pcsk9 and have reported sustained epigenetic silencing following transient in vivo delivery. A single administration of lipid nanoparticles carrying editor mRNAs nearly halved circulating PCSK9 levels for almost a year in mice. Remarkably, Pcsk9 silencing and associated epigenetic repressive marks persisted, even after forced liver regeneration, demonstrating the cellular heritability of the induced epigenetic state. This “epigenetic” approach reduced PCSK9 levels in mice with an efficiency comparable to conventional gene editing but without inducing potentially harmful DNA breaks.²⁷ Thus, it will likely soon be technically possible to lower LDL cholesterol for life. It remains to be seen whether societies or individuals have appetite for this type of intervention.²⁸

While LDL cholesterol has been effectively targeted, lipoprotein (a) (Lp(a)) is genetically determined and confers cardiovascular risk but without currently available approved specific therapies.^29,30,31 The results of phase 3 clinical trials evaluating the effects of Lp(a) reduction with siRNA are ongoing,³² testing pelacarsen in the HORIZON trial³³ and olpasiran in the OCEAN(a) trial.^34,35 An additional exciting development is the development of an oral inhibitor of Lp(a), which is in an early-phase clinical trial.³⁶

Although treatment of LDL cholesterol has made substantial clinical impact, it is important to recognize that populations at risk do not remain static. The risk factor landscape and therapeutic backdrop has changed substantially in recent decades. In many parts of the world, the population is aging, the prevalence of smoking has diminished, while obesity, sedentary lifestyle and type 2 diabetes are increasing. Approaches to risk management should incorporate an additional reality that a high proportion of individuals are (or could be) already receiving effective disease-modifying therapies. The opportunities to further reduce cardiovascular risk by conventional interventions (e.g., by reducing LDL cholesterol to ever lower levels) are receding. In meta-analysis of 3 clinical trials, Ridker and colleagues have recently shown that, among patients receiving contemporary statins, with effective LDL-cholesterol reduction, inflammation assessed by high-sensitivity C-reactive protein (hsCRP) was a stronger predictor for risk of future cardiovascular events and death than LDL cholesterol. These data fall into a broader landscape that identifies a role of inflammation as a target for reducing risk in atherosclerotic cardiovascular disease. We have argued before that a shift in emphasis toward the consideration of processes of inflammation, inflammation resolution, and tissue homeostasis is helpful, but the term inflammation is too imprecise and superficial to be useful.⁸ In the following sections, we will consider (1) the expanded understanding of the inflammation biology of atherosclerotic plaques and (2) processes beyond the plaque that drive atherosclerosis and its complications and which might prove amenable to intervention, with particular emphasis on the bone marrow and the immune/inflammatory environment that it creates.

Processes of inflammation

It is over 30 years since markers of systemic inflammation (measured by the acute-phase proteins: C-reactive protein [CRP] and serum amyloid A) were associated with adverse outcomes, initially in patients with acute coronary syndromes³⁷ and subsequently in apparently healthy men.³⁸ Significantly, the precise “drivers” of this sub-clinical inflammation are still not clear. It is plausible that the levels of the non-specific markers of inflammation reflect a cumulative “inflammatory load,” components of which also drive pathogenic processes relevant to atherosclerosis. In support of this notion, adverse outcomes pertain across a range of rheumatological and autoimmune diseases,³⁹ and the precise nature of chronic inflammation seems to matter little.⁴⁰ Processes of inflammation within the walls of arteries contribute to the manifestations of ASCVD, but these (sometimes small, focal) lesions seem more likely to be responding to systemic pro-inflammatory cues rather than to be responsible for driving the systemic inflammatory response that is reflected by elevated hsCRP. In other words, systemic drivers of processes of inflammation probably serve to activate cells in the atherosclerotic plaque, as they do elsewhere. This interpretation is supported by interventions, such as experimental myocardial infarction (MI), that drive inflammation and accelerate atherosclerosis.⁴¹

As discussed in greater detail, in the following sections, the functional status of cells within and beyond the plaque is sensitive to a complex network of regulatory factors, including genetic, epigenetic, and microenvironmental cues from cytokines, DAMPs, pathogen-associated molecular patterns, interactions with reactive oxygen species, matrix components, and metabolites. Understanding their roles and interactions is opening new possibilities for therapeutic interventions.

Treating inflammation

Treatments of inflammation have reported benefit in the context of established atherosclerotic cardiovascular disease. The Colchicine Cardiovascular Outcomes Trial (COLCOT) was a placebo-controlled trial of colchicine 0.5 mg daily in patients who had recently had an MI. The primary major adverse cardiac event (MACE) endpoint occurred in 5.5% of the patients in the colchicine group, as compared with 7.1% in the placebo group (hazard ratio [HR] = 0.77; p = 0.02).⁴² This was followed by the LoDoCo2 clinical trial involving patients with chronic coronary disease in which the risk of cardiovascular events was significantly lower among those on 0.5 mg of colchicine daily versus placebo. A primary endpoint event occurred in 187 patients (6.8%) in the colchicine group and in 264 patients (9.6%) in the placebo group (HR = 0.69; p < 0.001).⁴³ However, the OASIS-9/CLEAR SYNERGY clinical trial in patients with acute MI, randomized to colchicine 0.5 mg daily (n = 3,528) vs. placebo (n = 3,534) following PCI (percutaneous coronary intervention),⁴⁴ showed no difference for colchicine vs. placebo. Although commentators have highlighted methodological constraints in trial delivery (during a pandemic) that brings the robustness of the absence of a treatment effect into question,⁴⁵ a recent meta-analysis has also raised questions about the efficacy of colchicine, applied in broadly unselected patient groups.⁴⁶

In part, clinical judgments around colchicine are hindered because the relevant mechanisms of its action remain unresolved. Colchicine functions as a microtubule disrupter with anti-inflammatory effects^47,48,49 that include inhibiting neutrophil function^50,51,52 and inflammasome activity.^53,54 In addition to modulating immune cells directly, colchicine also affects endothelial cells by suppressing the expression of vascular cell adhesion molecule-1 after an inflammatory challenge^50,55 and potentially has inflammation-independent effects on plaque stability.⁵⁶ However, neither the relevant mechanism of action nor the site of action is known with precision, and this knowledge gap impairs optimal patient selection.

The Canakinumab Anti-inflammatory Thrombosis Outcome Study (CANTOS) trial recruited patients with prior MI and elevated hsCRP.⁵⁷ As such, inclusion incorporated a recognized indication of systemic inflammation. The trial evaluated 3 doses (50, 150, and 300 mg) of the anti-IL-1β antibody, canakinumab. The 150-mg dose (but not the others) met the prespecified (multiplicity-adjusted) threshold for statistical significance for both the primary endpoint and the secondary endpoint, which included hospitalization for unstable angina requiring urgent revascularization (HR vs. placebo, 0.83; p = 0.005). There was no significant difference in all-cause mortality across all doses. Significantly, subsequent analysis, based on the outcome according to responsiveness of either hsCRP or IL-6, both showed that individuals in whom benefit accrued, on average, declared their responsiveness through changes in these biomarkers. Specifically, canakinumab recipients with hsCRP <2 mg/L saw a 25% reduction in major adverse cardiovascular events (HR = 0.75, p < 0.0001), while those with hsCRP ≥2 mg/L had no significant benefit. In the <2 mg/L group, cardiovascular and all-cause mortality were each reduced by 31% (HR = 0.69 for both; p = 0.0004 and p < 0.0001, respectively), whereas no mortality reduction was observed in the ≥2 mg/L group.⁵⁸ The equivalent analysis using IL-6 was even more striking. Canakinumab recipients with IL-6 <1.65 ng/L had a 32% reduction in major adverse cardiovascular events (HR = 0.68, p < 0.0001), a 30% reduction in MACE plus hospitalization for unstable angina (HR = 0.70, p < 0.0001), a 52% reduction in cardiovascular mortality (HR = 0.48, p < 0.0001), and a 48% reduction in all-cause mortality (HR = 0.52, p < 0.0001). However, those with IL-6 ≥ 1.65 ng/L had no significant benefit in any endpoint.⁵⁹

While the relative risk reductions in CANTOS were substantial, a major limitation was the small absolute benefit in the patients included in the trial. Patient selection based on perceived clinical event risk (secondary prevention after acute MI), augmented by hsCRP level, did not provide a population in which a high enough proportion of patients had events or accrued benefit. These secondary analyses (aforementioned) demonstrate that knowledge of responder vs. non-responder status would further enhance the extent of benefit. To advance effectively, this field needs better individual-level mechanistic characterization and should move beyond population-level risk. Furthermore, systemic treatment of inflammation in CANTOS had the drawback of immune suppression and infection, pointing to the need for more targeted immunomodulation as part of the next-generation therapies.⁶⁰

Mendelian randomization studies examining the effects of genetically determined CRP show that it does not have a casual role in coronary heart disease.⁶¹ In other words, while CRP shows a strong population-level correlation with coronary disease events, it is “reporting” on relevant processes but is not driving them. The cytokine IL-6 plays a key role in immune responses and inflammation, as a ligand that binds to specific receptors. It interacts with two forms of the IL-6 receptor (IL-6R): the membrane-bound IL-6R, found on certain immune cells, and the soluble IL-6R, which enables signaling in other cell types. Upon binding, IL-6 recruits gp130, a signal transducing protein, activating the JAK/signal transducer and activator of transcription (STAT) pathway, which regulates inflammation and immune responses.⁶² Data accrued from animal models and epidemiological and genetic studies suggest a (likely causative) role in disease-relevant processes of atherogenesis. RESCUE, a phase 2 trial, showed reductions in CRP in response to the anti-IL-6 ligand antibody, ziltivekimab, that were more marked than those in response to canakinumab in CANTOS.⁶³ The ongoing phase 3 trial ZEUS (recruitment completed) is testing a strategy of anti-IL-6 treatment in patients with pre-existing cardiovascular disease, impaired renal function, and elevated CRP.⁶⁴ Notably, interim analyses have not found any evidence of immunosuppression (such as more severe infections), as was found in CANTOS.

Cellular reprogramming-immunomodulatory therapies

Cellular processes within the plaque are complex and dynamic.⁵ Arguably, there has been a bias toward interventions that impede or mitigate pro-atherogenic processes. However, this mindset overlooks two important factors. Firstly, many patients with atherosclerosis present clinically when the disease has already progressed to an advanced stage. Secondly the processes of atherogenesis are not necessarily relentlessly progressive. There is a potential to harness endogenous healing and repair mechanisms to regress even advanced plaques.⁶⁵

Macrophages are a key cell type in atherosclerosis, occupy a range of functional states and, can alter their properties and behavior in response to the local chemokine/cytokine environment,⁶⁶ which is, in turn, modulated by T cells that are also a prominent component of the atherosclerotic plaque. Broadly speaking (and recognizing the inadequacy of this classification of convenience),¹² classically activated (M1-type) macrophages are pro-atherogenic, and alternatively activated macrophages (M2-type) are found in early lesions and have reparative characteristics.⁶⁶ Substantial evidence supports a role for Th (Th1) cell (characterized by IFN-γ production) in driving classical activation of macrophages and exacerbating atherosclerosis,⁶⁷ and IFN-γ deletion leads to 60% smaller plaques in ApoE^−/− mice.⁶⁸ Conversely, pro-atherogenic tendencies can be ameliorated by alternative cytokines, such as IL-10, which protects against atherosclerosis in mice.^69,70

The immune system also has endogenous capacity to regulate its activation. Central to this capability are subsets of cells that inhibit immune responses and mediate their effects by producing these regulatory cytokines such as IL-10. Of these, the most deeply characterized immunomodulatory cells are Treg populations.⁷¹ ApoE^−/− mice injected with Tr1 (a subset of Tregs) cells showed a significant decrease in Th1 responses with a reduced production of IFN-γ and an increase in IL-10 production (p < 0.001) by cultured T cells. This was associated with a significant reduction in atherosclerotic lesion size with lesser accumulation of macrophages and T cells.⁷²

Accordingly, there is a rationale to enhance the activity of the Treg population. Aldesleukin (recombinant IL-2) caused an enrichment of Tregs, without causing T effector activation, and was not associated with adverse effects in a small phase 1b/2a clinical trial.⁷³ Similarly, the administration of low-dose IL-2 (as aldesleukin) increased the number of circulating IL-10-producing B cells in patients after acute non-ST segment elevation MI.⁷⁴

An alternative approach is to administer IL-4 and/or IL-13 (which are physiologically produced by Th2 cells to induce alternative macrophage activation), which results in anti-inflammatory and tissue repair capacities.⁷⁵ Exogenous administration of IL-13 to cholesterol-fed LDLR^−/− mice promoted collagen formation and reduced macrophage content of existing lesions and induced a switch of lesional macrophages in vivo toward alternative activation.⁷⁶ One potential drawback is that IL-4 and IL-13 induce an immunosuppressive environment that favors the differentiation of tumor-facilitating M2-type tumor-associated macrophages.⁷⁷

There is also an expanding literature on the contributions of B lymphocytes in atherosclerosis based on clinical observations, studies in mouse models, and in vitro experiments.⁷⁸ Broadly, the B cell subtype “B2” is thought to be pro-atherogenic and the subtypes B1 and regulatory B cells (Bregs) atheroprotective. In keeping with these roles, B2 cells activate T cells to produce the pro-atherogenic cytokine IFN-γ. The atheroprotection of Bregs is mediated by the secretion of the inflammation-resolving cytokine IL-10, while B1 cells produce IgM, directed against oxidized LDL. This type of antibody has been shown to be a potent atheroprotective factor in mouse models,⁷⁹ and its potential as a therapeutic in human atherosclerosis is being investigated.

Sequencing over 1,700 B cells from mice with atherosclerosis and control mice revealed 56 antibodies from expanded B cell clones in atherosclerosis. About one-third targeted atherosclerotic plaques, indicating diverse lesion antigens, trigger responses. Proteomics identified ALDH4A1, a mitochondrial enzyme, as a target of antibody A12. Injection of A12 antibody in mice reduced plaque growth and LDL levels, highlighting its possible therapeutic role in cardiovascular disease.⁸⁰

In a small phase 2 clinical trial (ClinicalTrials.gov: NCT04776629), orticumab, a monoclonal antibody directed against oxidized LDL, reduced fat attenuation index (FAI, a measure of coronary inflammation) in patients with high baseline FAI on a background of the inflammatory condition, psoriasis.⁸¹ Another trial, scheduled to start in June 2025 (ClinicalTrials.gov: NCT06927739), will test whether the same antibody will be effective in secondary prevention in patients with a prior MI.

Overall, there is increasing interest in B cell-based therapeutics for atherosclerosis, including the antibody administration already noted, but also a vaccine approach and the adaptation of a multiplicity of strategies used in treating autoimmune diseases, and lymphoid cancers.

Cellular reprogramming-atherosclerosis regression

Clinical trials have evaluated the effects of drug interventions on atherosclerosis regression. In ASTEROID, an early observational study, intravascular ultrasound (IVUS) demonstrated a 6.1% reduction in total atheroma volume at 24 months in 507 patients taking 40 mg rosuvastatin daily.⁸² Indeed, multiple studies using IVUS have shown plaque regression through LDL-cholesterol reduction, and meta-analysis has even estimated an LDL cholesterol of around 1.8–2 mmol/L (70–77 mg/dL) as the “tipping point” between plaque progression and regression.⁸³ Plaque analysis after PCSK-9 inhibition has also shown apparent reductions in plaque lipid content. The Global Assessment of Plaque Regression with PCSK9 Antibody as Measured by Intravascular Ultrasound trial randomized almost 1,000 patients to evolocumab or optimal medical therapy. After 76 weeks, greater reduction in total atheroma volume was seen with evolocumab (2.9% vs. 0.4%; p < 0.05).⁸⁴ Similarly, non-invasive imaging studies of carotid arteries using MRI have shown rapid regression within 3 months of statin initiation.⁸⁵ Nevertheless, the changes measured in the volume of plaque or its constitutive elements have been small and insufficient to materially regress luminal stenosis. These methods of assessing plaque size or macroscopic composition are, however, insufficiently sensitive or refined to determine changes in cellular status.

Progression or persistence of atherosclerosis is a manifestation not only of pro-atherogenic processes but also of the failure to remove inflammatory and necrotic material, including through dysfunctional efferocytosis. Identification of the factors that determine cellular function and failed inflammation resolution offers possibilities for reprogramming and fundamental lesion remodeling.^86,87 As detailed earlier, in atherosclerosis progression, plaque macrophages are predominantly pro-inflammatory. Conversely, during resolution, there is an accumulation of inflammation-resolving, tissue remodeling macrophages, which express arginase (ARG) 1 and mannose receptor 1, and the concomitant loss of macrophages of the inflammatory subtype.⁸⁷ Since ARG1 expression is regulated through the transcription factor STAT6 and is necessary for resolution of murine atherosclerosis,⁸⁸ we tested whether the classical activators of STAT6, IL-4, and IL-13 were necessary for atherosclerosis regression. While IL-4 is necessary, there are also additional factors operating alongside, one of which is Wnt.⁸⁹ Regression and plaque cellular remodeling are also associated with induction (at mRNA and protein levels) of the chemokine receptor CCR7, an essential factor for dendritic cell migration. If CCR7 function is blocked by antibodies to CCR7 ligands CCL19 and CCL21, lesion regression is abolished, and foam cells persist.⁹⁰ Netrin-1 is a protein involved in immune responses, wound healing, and angiogenesis. In the context of atherosclerosis, it is secreted by macrophages in plaques, where it sustains inflammation by enhancing macrophage survival and blocking macrophage emigration. Macrophage-specific deletion of netrin-1 in mice with established atherosclerosis promotes plaque regression with reduced retention of plaque macrophages and smooth muscle cells within the lesion core, allowing a pro-resolving reorganization of the plaque.⁷⁵ This finding is reinforced by myeloid-specific deletion of Unc5b (the netrin-1 receptor), which reduced plaque size and led to enhanced efferocytosis, and a pro-resolving environment marked by an increase in local athero-protective T cells.⁷⁶

To date, clinical interventions have not targeted specifically the pathways and processes of atherosclerosis regression, but awareness of the resolution of atherosclerosis, distinct from retarded progression, and knowledge of its regulation will bring forward the possibility to harness endogenous repair processes and to treat established lesions. See Table 1 for summary of current approaches.

Table 1.

Emerging treatments for atherosclerosis; Emphasis on inflammation and immunomodulation

Drug	Clinical trial	Phase	Year	Population	Target pathway	Mechanism	Clinical outcome	Reference
Bempedoic acid	CLEAR outcomes	3	2023	high cardiovascular risk with intolerance to statin	LDL reduction	ATP citrate lyase inhibitor	significant lower MACE, hazard ratio of 0.87 (95% confidence interval [CI] 0.79–0.96)	Nissen et al.20
Evolocumab	FOURIER	3	2017	established atherosclerotic cardiovascular disease with elevated LDL despite statin	LDL reduction	mAb inhibitor of PCSK9 protein	significant lower MACE, hazard ratio of 0.85 (95% CI 0.79–0.92)	Sabatine et al.21
Alirocumab	ODYSSEY OUTCOMES	3	2018	history of acute coronary syndrome within a year with elevated LDL despite statin	LDL reduction	mAb inhibitor ofPCSK9 protein	significant lower MACE, hazard ratio of 0.85 (95% CI 0.78–0.93)	Schwartz et al.22
Inclisiran	ORION-10 & ORION-11	3	2020	established atherosclerotic cardiovascular disease with elevated LDL despite statin	LDL reduction	siRNA inhibitor of PCKS9 mRNA	significant reduction of LDL by 52.3% (95% CI 48.8–55.7) and 49.9% (46.6–53.1)	Ray et al.23
Pelacarsen	Lp(a)HORIZON	3	pending completion	evidence of myocardial infarction, ischemic stroke, or peripheral vascular disease within 10 years with raised Lp(a)	Lp(a) lowering	antisense oligonucleotide inhibitor of Apo (a) mRNA	pending	Malick32 and ClinicalTrials.gov33
Olpasiran	OCEAN (a)	3	pending completion	history of myocardial infarction and one additional risk factor with raised Lp(a)	Lp(a) lowering	siRNA inhibitor of Apo(a) mRNA	pending	ClinicalTrials.gov34 and O'Donoghue et al.35
Colchicine	COLCOT	3	2019	history of myocardial infarction within 30 days	immunomodulatory	unclear	significant lower MACE, hazard ratio of 0.77 (95% CI 0.61–0.96); higher frequency of pneumonia (0.9% vs. 0.4%, p = 0.03)	Tardif et al.42
Colchicine	LoDoCo2	3	2020	chronic coronary syndrome	immunomodulatory	unclear	significant lower MACE, hazard ratio of 0.69 (95% CI 0.57–0.83); higher incidence of death from noncardiovascular cause, hazard ratio of 1.51 (95% CI 0.99–2.31)	Nidorf et al.43
Colchicine	CLEAR	3	2024	presentation with ST-elevation and NST-elevation myocardial infarction	immunomodulatory	unclear	nil significant difference in MACE, hazard ratio of 0.99 (95% CI 0.85–1.16); higher frequency of diarrhea (10.2% vs. 6.6%, p < 0.001) but no difference in incidence of serious infections	Jolly et al.44
Canakinumab	CANTOS	3	2017	previous myocardial infarction and hsCRP of >2 mg/L	immunomodulatory	mAb inhibitor of IL-18	for 150 mg dose, significant lower MACE, hazard ratio of 0.85 (95% CI 0.74–0.98), higher incidence of fatal infection 0.28 vs. 0.18 per 100 person-years (p = 0.02)	Ridker et al.57
Ziltivekimab	RESCUE	2	2021	moderate to severe renal impairment and hsCRP of >2 mg/L	immunomodulatory	mAb inhibitor of IL-6	significant reduction in hsCRP across doses, 77% for 7.5 mg, 88% for 15 mg, and 92% for 30 mg	Ridker et al.63
Ziltivekimab	ZEUS	3	pending completion	chronic kidney disease, established atherosclerotic cardiovascular disease, and hsCRP of ≥2 mg/L	immunomodulatory	mAb inhibitor of IL-6	pending	ClinicalTrials.gov, U.N.L.o.M.64
Aldesleukin	ULACS	1/2b	2022	stable ischemic heart disease	immunomodulatory	recombinant IL-2	no significant adverse effects to drug of major concern	Zhao et al.73
Orticumab	NCT04776629	2	204	psoriasis, BMI ≥30 and LDL 100 mg/dL	oxLDL inhibition	mAb against oxLDL	reduction in FAI score in patients with high FAI score at baseline; no significant adverse effects to drug	Farina et al.81

Identifying processes of inflammation in vivo

Much of the early description of atherosclerosis was built on plaque topology and on the identification of cells and cell types based on a small number of surface markers, using immunohistochemistry and/or immunofluorescence. By combining small numbers of markers, relatively superficial inferences were made about potential function. The advent of single-cell sequencing, particularly combined with spatial data, provides hugely greater functional information and, potentially, insight.^13,91,92 However, even armed with detailed knowledge of molecular and cellular processes that are activated in atherosclerosis, the field is a long way from being able to map those processes in individual patients in vivo. We and others have attempted to relate disease-relevant characteristics from molecular and cellular analyses to features identified by non-invasive imaging. Plaque macrophage content and activity can be tracked through increased 18-FDG (18-fluorodeoxyglucose) glucose uptake and by the binding of cell-specific radioligands, such as 68Ga-DOTATATE that binds to somatostatin receptor type-2, which is expressed on activated macrophages,^93,94 while 18F-NaF PET (positron emission tomography)-computed tomography (CT) identifies ruptured coronary plaque after acute coronary syndromes.⁹⁵ Chai et al. have shown a relationship between the quantitative measurement of lipid-rich carotid atherosclerotic plaque core, using T2 mapping on MRI, with the upregulation of genes of the IFN/STAT1 pathways.¹³ Others have used AI-based approaches to relate the expression of genes representing inflammation, fibrosis, and vascularity to “radiomic features” extracted from tissue CT images.⁹⁶ However, there is a substantial disconnect between techniques to characterize plaque, which remain broad-brush strokes, and the fine granularity of spatial multi-omic data. None of these imaging approaches has been paired with a specific therapeutic intervention. The limitations are amplified by the appreciation that plaque biology is dynamic and that it has not yet been possible to reflect this by the application of occasional (and expensive) imaging.

Considerations in diabetes

Globally, approximately half a billion people have diabetes.⁹⁷ Type 2 diabetes accelerates the first presentation of cardiovascular disease by more than a decade and doubles the risk of death from a cardiovascular cause.⁹⁸ As discussed (earlier), a recent meta-analysis of three trials showed that in patients with pre-existing cardiovascular disease treated to contemporary lipid targets, the relationship between LDL-cholesterol level and future cardiovascular events was lost almost completely.⁹⁹ By contrast, analysis of quintiles of hsCRP showed a strong and enduring relationship even in patients in whom the LDL cholesterol had been effectively lowered. Notably, a high proportion (circa 70%) of patients aggregated in that meta-analysis had a diagnosis of type 2 diabetes. Much evidence suggests that type 2 diabetes is associated with disordered processes relevant to inflammation,^{100,101,102,103} and a recent review comprehensively considers the special case of atherosclerosis in diabetes.¹⁰⁴ Of particular interest are observations that hyperglycemia alters bone marrow function and enhances myelopoiesis^100,105,106 and that transient bursts of hyperglycemia do this more potently than sustained elevation.¹⁰⁷

Elevated blood glucose drives vascular complications,¹⁰⁸ and clinical outcomes for patients who have sustained MI are persistently worse in patients with diabetes.¹⁰⁹ While lowering blood glucose is a central plank of diabetes management,¹¹⁰ remarkably, reducing blood glucose has consistently little or no effect, or markedly delayed effects, in reducing rates of MI.^{110,111,112,113,114} Intriguingly, the UK Prospective Diabetes Study has shown that a period of sub-optimal glucose control (shortly after diagnosis but with convergence in glucose levels with a control group thereafter) manifests in adverse cardiovascular outcome >24 years later.¹¹⁵

We have reasoned that these “legacy effects” might be “recorded” in a cell type with (1) unusually long persistence, (2) agency in the pathogenesis of cardiovascular disease, and (3) evidence of relevant modifiable programming. Elevated glucose reprograms hematopoietic stem cells (HSCs) culminating in distinct chromatin modifications (e.g., histone 3 lysine 4 trimethylation and histone 3 lysine 27 acetylation) in HSCs and their macrophage progeny.^116,117,118 This “hyperglycemia-induced trained immunity” (HITI) promotes atherosclerosis and persists in vivo despite complete correction of hyperglycemia.¹¹⁶ The changes identified in response to hyperglycemia form part of a broader spectrum of trained immunity in response to metabolic and immunological stimuli, including oxidized-LDL and Western-type high-fat content diet.^{118,119,120,121,122} The epigenetic changes due to posttranslational modifications of histones (or DNA methylation) tend to be persistent.^123,124 Histone modifications play a crucial role in epigenetic reprogramming, as they influence the chromatin structure and gene expression patterns without altering the underlying DNA sequence. These modifications can be heritable and reversible, which makes them essential for cellular processes like development, differentiation, and response to environmental signals. Their occurrence in immune cell precursors in the bone marrow explains how terminally differentiated cells with short-half lives, such as macrophages, can also demonstrate metabolic “memory.” Acetylation groups are added to the lysine residues on histone tails, leading to a more relaxed chromatin structure (euchromatin), which is typically associated with gene activation. To effect fundamental “reprogramming,” histone modifications need to be “erased” or “written” anew, and a wide variety of target and potential drugs exist for this.^125,126,127 However, this might not be the most immediate or accessible way to deliver therapy. Knowledge of genes and pathways activated through epigenetic modifications means that the downstream effectors can be targeted, potentially using existing or repurposed drugs. For instance, in their description of HITI, Edgar et al. identified the activation of IL-6 and IL-1β pathways, both of which can be targeted by existing drugs (described previously).¹¹⁶

Clonal hematopoiesis in atherosclerosis

Clonal hematopoiesis of indeterminate potential (CHIP) has emerged as a novel and substantial age-associated risk factor for atherosclerosis. Work to elucidate the etiology is ongoing (see the following text) but promises to offer an attractive avenue for targeted therapeutic intervention at a large scale. CHIP is defined as the acquisition of a somatic mutation of a known hematological driver gene, identified in peripheral blood cells.^128,129 Originating within a hematopoietic stem/progenitor cell (HSPC), such mutations will confer a proliferation advantage, leading to clonal expansion with time. The mutation is detectable in the daughter cells of the mutated HSPC, and the percentage of mutant compared to wild-type is referred to as the variant allele frequency (VAF). Conventionally, a VAF threshold of 2% is used to diagnose CHIP, but with improved sequencing depth, there is increasing recognition of clonal hematopoiesis at much lower VAFs. Indeed, studies designed to detect a VAF of 2% identified CHIP in 10%–20% of those aged over 70 years.¹³⁰ However, when using a VAF threshold of 0.03%, clonal hematopoiesis is almost universal.¹³¹ Importantly, blood counts must be normal for CHIP to be diagnosed, and there must be no evidence of a hematological malignancy.

Initial work by Jaiswal et al. demonstrated the association of CHIP with a 1.4-fold increased risk of death, unaccounted for by the increased risk of hematological malignancy.¹³² Subsequently, the same group demonstrated that CHIP associated with a 1.9-fold increased risk of coronary heart disease, defined as a history of MI or coronary revascularization.¹³³

The majority of CHIP involves a single mutation, and 80% of cases involve mutations in the epigenetic regulator genes ASXL1, ten-eleven translocation methylcytosine dioxygenase 2 (TET2), and DNA methyltransferase 3A (DNMT3A).^130,134 However, the underlying mutation confers a variable risk of MI, with JAK2 posing the highest risk (HR 12.0) followed by ASXL1 (HR 2.0) and TET2 (HR 1.9).¹³³ The pattern of onset also varies by mutation, with DNMT3A CHIP, for example, developing earlier in life but associating with a lower risk of cardiovascular disease.

Translational work has uncovered the role of dysregulated inflammation in CHIP-associated atherosclerosis. Transplantation of TET2-deficient bone marrow into atherosclerosis-prone LDLR^−/− mice increased plaque size, driven by NLRP3 inflammasome-mediated IL-1-β signaling in plaque macrophages.¹³⁵ Other mouse models show that mutant JAK2 is capable of driving AIM2 (another type of inflammasome)-associated atherosclerosis, potentially amenable to IL-1-β inhibition.¹³⁶ Meanwhile, ASXL1 mutations have identified the loss of suppression of nuclear factor κB via IRAK1-TAK1 as being important for atherogenesis.¹³⁷

Understanding the sequence of events in the development of atherosclerosis among individuals with CHIP is pivotal in determining whether CHIP is causative, or merely a biomarker, in atherosclerosis. HPSCs remain largely dormant but exit quiescence in response to physiological stress—a process that provokes DNA damage.¹³⁸ Increased stem cell proliferation in response to atherosclerosis could account for the observed association with CHIP.¹³⁹ Additionally, the description of hyperglycemia-induced trained immunity points toward a persistent epigenetic reprogramming of HSPCs following hyperglycemic challenge, supporting the notion that these cells alter their behavior in response to metabolic challenges.¹¹⁶ When taking these findings and the aforementioned translational experiments into consideration, it is plausible that CHIP could play both a contributory and a reactive role as part of a vicious cycle in atherosclerosis development.

Recent work from Díez-Díez et al. sought to assess the temporal development of CHIP and femoral atherosclerosis measured with ultrasound.¹⁴⁰ By tracking a longitudinal cohort of healthy individuals over 6 years, they observed that participants with CHIP at enrollment, but no vascular disease, were more likely to develop overt atherosclerosis during follow-up. The converse was not observed, supporting the hypothesis that CHIP occurs prior to overt atherosclerosis.

Notwithstanding the complexities of large-scale genetic screening, clinicians must consider what future preventative intervention for CHIP-associated atherosclerosis might look like. To date, much of the discussion has centered around targeting inflammatory pathways. In the CANTOS trial, the use of canakinumab to induce IL-1β blockade showed a signal toward benefit among patients with TET2 CHIP, which needs further exploration.¹⁴¹ Similarly, an inactivating IL-6R coding mutation associates with a reduced cardiovascular risk among patients with CHIP in the UK Biobank cohort.¹⁴² The anti-inflammatory drug colchicine targets IL-1β signaling in mice, impeding atherosclerosis development in TET2 mutant mice; a reduction in MI among patients with TET2 CHIP taking colchicine was also seen in humans.¹⁴³

It remains to be seen what effect such therapies might have on the development of atherosclerosis in humans with CHIP. The ultimate therapy might instead be one aimed at eradicating the mutant HSPC clone, rather than one aimed at mitigating its effect. JAK2 can be targeted using selective inhibitors, although the effect on clone size in CHIP is uncertain. The effect of anti-inflammatory medications on CHIP clone size remains unknown but could plausibly reduce clonal burden in addition to inflammatory signaling.

More work is required to fully understand the relationship between the mutant CHIP clone and atherogenesis. Collaboration between cardiology and hematology clinicians and scientists has been essential to the progress so far and will likewise be instrumental in implementing future clinical interventions.

Conclusion and future prospects

Emerging from many of the interlinking strands surveyed earlier is a complex systems biology in which atherosclerosis reflects disordered local and systemic processes of homeostasis, which are intimately entwined in the fabric of “metabolism” and inflammation. Data are emerging on further specific neural interconnections in relation to immune function,¹⁴⁴ which are of particular relevance to sleep,¹⁴⁵ physical exercise,¹⁴⁶ circadian rhythm,¹⁴⁷ and psychological stress.^148,149

The future of atherosclerosis treatment is poised to benefit from a range of emerging technologies, offering innovative and personalized approaches. RNA-based therapeutics, such as siRNA therapies like inclisiran, are a rapidly advancing treatment option by targeting specific genes, such as PCSK9, with infrequent dosing. Similarly, antisense oligonucleotides targeting apolipoprotein C3 or Lp(a) are being explored to manage lipid profiles, while mRNA therapies hold promise for enhancing the expression of beneficial proteins to promote vascular health.

Immunomodulatory approaches are also advancing. Beyond existing treatments such as colchicine and canakinumab, targeted anti-inflammatory agents are being developed to modulate the immune response and reprogram processes of inflammation. Vaccines targeting oxidized LDL and other atherosclerosis-specific antigens, as well as Treg modulation, are being investigated to maintain a balanced immune response, reducing pathological inflammation.

Advanced drug delivery systems are enabling more precise therapies. Nanomedicine platforms are being designed to deliver drugs directly to atherosclerotic plaques, with plaque-specific targeting using antibodies against macrophage markers or neovascular components.

Looking further ahead, opportunities are already emerging in proof of concept for CRISPR-based gene editing and epigenetic reprogramming and for interventions that capitalize on expanded appreciation of the roles of circadian rhythm, sleep, stress, diet, gut microbiota, and in utero programming. Together, these technologies show concrete possibilities for a raft of transformative solutions for next-generation treatments of atherosclerosis.

Acknowledgments

R.P.C. and J.C. are funded by the BHF-MRC Centre of Research Excellence in Advanced Therapeutics (REACT: MRC Award MR/Z504865/1; BHF Award: SI/F/24/21170016), Oxford, and by the MeRIAD (Metabolite-Related Inflammation and Disease) Consortium (Novo Nordisk Foundation). E.A.F. acknowledges NIH grants HL084312, HL131481, and HL151328. R.H. is funded by Bristol Myers Squibb as part of the Oxford-BMS Fellowship Program.

Declaration of interests

R.P.C. is the UK Chief Investigator for the ZEUS trial (sponsor Novo Nordisk) and serves on the Global Expert Panel for that trial. He is engaged by Oxford University Consulting to undertake paid consultancy work for Novo Nordisk, NodThera, Velakor, and Tourmaline Bio. He serves on the Scientific Advisory Board of Tourmaline Bio. His laboratory receives, or has received, research funding from the Novo Nordisk Foundation, Novo Nordisk Research Centre Oxford, the Chan Zuckerberg Initiative, the Kusuma Trust, and Novartis. He served on the data safety and monitoring board for trials of low-dose interleukin-2 therapy (IVORY and LILACS): sponsor University of Cambridge. There are no stock/equity/patent declarations for R.P.C. or immediate family.