Monocyte and Macrophage Dynamics in the Cardiovascular System: JACC Macrophage in CVD Series (Part III)

Abstract

It has long been recognized that the bone marrow is the primary site of origin for circulating monocytes that may later become macrophages in atherosclerotic lesions. However, only in recent times has the complex relationship between the bone marrow, monocytes/macrophages and atherosclerotic plaques begun to be understood. Moreover, the systemic nature of these interactions, which also involves additional compartments such as extra-medullary hematopoietic sites (i.e., spleen), are only just becoming apparent. In parallel, progressive advances in imaging and cell labeling techniques have opened new opportunities for in vivo imaging of monocyte/macrophage trafficking in atherosclerotic lesions and at the systemic level. In this part 3 of a 4-part review series covering the macrophage in cardiovascular disease we intersect systemic biology with advanced imaging techniques to explore monocyte and macrophage dynamics in the cardiovascular system, with an emphasis on how events at the systemic level might impact local atherosclerotic plaque biology.

Condensed Abstract

Macrophages are fundamental to local atherosclerotic plaque biology. However, there has been a growing appreciation of the importance of the systemic interactions between the bone marrow, monocytes/macrophages and atherosclerotic lesions. In parallel, advances in imaging techniques have opened new opportunities for in vivo imaging of monocyte/macrophage trafficking both at the local and systemic levels. In this part 3 of a 4-part review series covering the macrophage in cardiovascular disease we intersect systemic biology with advanced imaging techniques to explore macrophage/monocyte dynamics in the cardiovascular system, with an emphasis on how events at the systemic level impact local plaque biology.

Previously in this 4-part review series we covered the basic biology of macrophages in Part 1, and macrophage pathobiology in atherosclerosis in Part 2. Here in Part 3 of this review series we explore macrophage biology and imaging in the cardiovascular system at the systemic level.

Fundamental concepts in hematopoiesis

The bone marrow in cardiovascular disease

Although the mortality rate from cardiovascular disease has declined over the last 50 years (1), myocardial infarction and stroke continue to claim a large proportion of lives, rendering atherosclerosis, the chronic condition behind these events, the most lethal disease worldwide. It is estimated that by 2030 cardiovascular disease costs will increase to more than US $1 trillion globally (2). New strategies aimed at attenuating risk and preventing plaque rupture or erosion are therefore urgently needed. A recently completed clinical trial, CANTOS, has successfully provided a proof-of-concept for the role of inflammation in atherosclerosis and cardiovascular disease (3,4). The trial produced several important insights. First, it showed that blocking an inflammatory cytokine such as interleukin-1β protects against cardiovascular events without affecting lipid levels or other conventional cardiovascular risk factors, thus providing support for the inflammatory basis of cardiovascular disease. Second, the effects seen, though already impressive, indicated that additional studies are required to decipher how inflammation contributes to disease. Third, CANTOS showed that blocking interleukin-1β increases infection rates, thereby indicating that immunotherapy for cardiovascular disease must preserve host defense. Fourth, given that interleukin-1β is predominantly a leukocyte product, the trial indirectly implicated the bone marrow as a major site contributing to cardiovascular disease.

The bone marrow is a highly vascularized region responsible for different tasks that, collectively, maintain hematopoietic stem cells (HSCs) and produce nearly all leukocytes after birth. The bone marrow is the permanent dwelling of long term hematopoietic stem cells (LTHSC). As cells that can repopulate the entire leukocyte, red blood cell and platelet pool, LTHSCs are arguably the most important marrow residents, and therefore must be protected from destruction or exhaustion. For example, resident bone marrow macrophages, damage-associated molecular patterns, and pathogen-associated molecular patterns in the blood can be a threat, potentially eliminating or causing direct harm to the stem cells. This pressure to shield and protect LT-HSCs is at odds with the need to produce hematopoietic cells on a massive scale. Red blood cells, for example, circulate for approximately 120 days before becoming senescent, but because a human adult has 30 trillion of them at any given time, the bone marrow must produce 250 billion red blood cells every day just to keep up. As for platelets, which live for up to 10 days,150 billion are required daily. Among leukocytes, myeloid cells such as monocytes and neutrophils are short-lived, relying on constant replenishment, but T and B lymphocytes are long-lived and require less bone marrow output. Therefore, not only is the overall production massive, but it must also be precisely calibrated to cell type. Once cells are produced, the bone marrow actively releases them to the circulation, and it must do so in a timely and balanced manner, which adds another level of complexity because the exiting cells are heterogeneous. Whereas monocytes and neutrophils exit the bone marrow as “fully” mature (monocytes can further differentiate in destination tissues), T and B cells exit as progenitors before maturing in the thymus and spleen, respectively. Stem cells also exit the bone marrow, presumably for immunosurveillance (5). Mobilized stem cells can either return back to the bone marrow or, under specific conditions, settle in secondary lymphoid organs, where they can give rise to their progeny in a process called extramedullary hematopoiesis. Mobilization of many of these cells is rhythmic, obeying circadian fluctuations (6). Finally, we now understand that, in response to infection or injury, the bone marrow output increases so as to meet demand. There are, therefore, multiple requirements on the bone marrow to ensure an adequate cell supply while preserving a stem cell pool so that supply lasts throughout life.

In recent years, work focusing on the bone marrow medullary microenvironment, on hematopoietic ontogeny (i.e. the developmental history of a bone marrow-derived cell in an individual), and on dynamic in vivo interactions has brought us closer to deciphering how the bone marrow can balance the contrasting imperatives of preserving a cell population on the one hand and using it for massive cell deployment on the other. The idea that the bone marrow architecture determines hematopoiesis is deeply rooted in the concept of the bone marrow niche, which contends that the bone marrow is organized into enclaves, or niches, consisting of specialized cells producing a set of products that preserve “stemness” while meeting leukopoietic and erythropoietic demands (7–10). The niches can be defined by their composition of endothelial cells, mesenchymal stem cells, CXCL12-abundant reticular (CAR) cells, macrophages, and adipocytes, producing secreted and membrane-bound cytokines, chemokines, retention factors, growth factors, proteases, and other species jointly affecting cell quiescence, maintenance, proliferation, differentiation and mobilization. Beyond composition, the niches can also be defined by their relative proximity to the vasculature. The idea here is that specific quiescent HSCs reside in spatially defined regions of the bone marrow and, moreover, that specific niches are necessary for the maintenance of quiescence (11). Niche organization, it follows, is geared toward balancing between generating vast numbers of cells versus preserving long-term hematopoietic cells. Such a system might require either progenitor transfer between niches or niche reallocation relative to the vasculature. One wonders whether this concept can be extended to include the bone marrow as a whole, whereby specific organs recruit cells from specific strategically-located bone marrow regions.

The relative contribution of the medullary microenvironment notwithstanding, a compelling argument can be made that ontogeny, and specifically the constraints that accompany ontogenic relationships, preserve balance in the bone marrow. For this argument, the hematopoietic tree can be imagined to consist of discrete progenitor entities linked through ontogeny, where a progenitor upstream in the ontogenic hierarchy gives rise to a progenitor that has a more limited commitment potential, higher proliferative capacity, and, perhaps as a consequence, shorter lifespan (12). Any given stem or progenitor cell is linked to a cell that is directly above or below in the developmental hierarchy. For a HSC to give rise to monocytes, for example, it would first need to proliferate such that one of its daughter cells remains unchanged while the other differentiates to the next progenitor downstream (so called “asymmetric division”). The process would then continue, with each step generating a more committed progenitor. This intrinsic “clonal succession” chain-of-command ensures that upstream HSCs remain relatively quiescent, non-proliferative, and stem-like but can, through a series of discrete steps that include asymmetric division, eventually give rise to all leukocytes and erythrocytes. It is an elegant model that precludes a monocyte arising from upstream LT-HSCs, perhaps for good reason as such shortcuts might compromise a stem cell’s privileged status. Nevertheless, we now have evidence that there is no need for a sequential and discrete transition in hematopoiesis, as cells can indeed bypass intermediates, taking shortcuts to generate their downstream progeny. Although we understand little about these shortcuts’ inner workings, the findings do suggest that hematopoiesis’ hierarchy is not defined by a set of discrete and orderly transitions between clearly articulated progenitor entities. This clonal diversity model argues that cells can be committed for specific populations much earlier than previously thought. It is a compelling model with strong evidence in its support (13–18).

The bone marrow might be fixed inside the bone, but it is neither static nor isolated. On the contrary, it is part of a dynamic and integrated system that can be imaged in vivo (as discussed later in this article) (19–21). The bone marrow’s dynamism is exemplified by studies that have focused on the role of the circadian clock (22). We now have evidence for synchronized cell migration between the bone marrow, blood, and tissue over the course of a day, where the highest number of cells in the blood can be found during our rest period while lower numbers circulate when we are awake. One wonders why such rhythmicity evolved? One interesting idea is “anticipatory inflammation” (23). It proposes that a cell’s location is always strategic. We have evolved to anticipate injury or infection when we are awake and interacting with the world. It is therefore strategically advantageous for leukocytes to be dispersed in organs when we are awake, so that they can more quickly respond to danger. Conversely, when we are asleep, injury is less likely, so leukocytes return to the blood. The idea is appealing but not entirely satisfying. After all, the blood is a conduit that connects all organs: is circulating not akin to anticipating? Perhaps the mass transfer is less about anticipatory inflammation, but more a form of circadian cleansing, similar to what happens in the brain during sleep (24).

The value of understanding how the bone marrow functions on a daily basis, under homeostatic conditions, cannot be overstated. However, it is now clear that inflammatory stimuli have a profound influence not only on the quantity but also the quality of hematopoiesis. For one, infection and injury elicit considerable changes in hematopoietic cells (13,16,25–27). In cardiovascular disease such as myocardial infarction, the bone marrow produces and releases more leukocytes and more stem cells (28–31). During atherosclerosis, a considerable number of monocytes and neutrophils are produced that can be recruited to growing atheroma (32,33). Moreover, during chronic inflammatory conditions, such as atherosclerosis, the bone marrow releases HSCs which can settle in extramedullary sites, giving rise to leukocytes via extramedullary hematopoiesis.

Extramedullary hematopoiesis in cardiovascular disease

At its broadest, extramedullary hematopoeisis is the production of hematopoietic cells outside of bone marrow medullary spaces. Fate mapping studies show that unperturbed adult hematopoiesis is maintained by bone marrow-derived ‘short-term’ stem cells downstream of HSCs with little contribution from extramedullary sources (34). However, in several hematological disorders and pathological states that arise secondary to bone marrow failure, extramedullary blood production can occur in various tissues. During hematopoietic stress, the spleen is frequently a site of extramedullary hematopoiesis as stem and progenitor cells mobilized from the bone marrow enter splenic cords through an open blood system. Once in the red pulp, splenic hematopoietic stem and progenitor cell (HSPC) activity is supported by the elaboration of perisinusoidal endothelial and stromal cells and their products, extending the concept of HSPC niche requirements beyond the bone marrow to extramedullary sites (35). Blood production outside the bone marrow favors myeloerythroid lineages over other lymphocytes, presumably because lymphoid progenitors are supported by elements produced by osteoid tissue which is lacking outside of bone (36,37). And while recent studies suggest considerable overlap in HSPC requirements in and out of the bone marrow, extramedullary and medullary output are distinct physiological responses. Preferential inhibition of stem cell factor production in the spleen, for example, reduces circulating blood cell counts without affecting bone marrow hematopoiesis (35).

What purpose does extramedullary hematopoiesis serve? Facultative niches arise at extramedullary sites in response to numerous hematopoietic stresses, however the relative contribution of these sources to inflammation at non-production sites is difficult to assess. Does extramedullary output simply increase cell number to meet hematopoietic demand or does it support the development of cell types with distinct functions? During myocardial infarction, animal studies show that recruited monocytes remove debris and facilitate tissue repair, and their numbers are largely sustained by splenic monocytopoiesis (30). Furthermore, HSPCs derived from medullary and extramedullary hematopoiesis might display different attributes and differentiation patterns. In atherosclerosis, chronic inflammation is characterized by continuous mobilization of stem and progenitor cells to the spleen. Once seeded, these progenitors proliferate and mature into myeloid lineages. Under these conditions the hematopoietic hierarchy is refined; while limited in their capacity to self-renew in the steady state, during hypercholesterolemia splenic monocytes proliferate vigorously (38). Importantly, fate mapping studies show that extramedullary-derived monocytes exit into the circulation and gather in atheroma where they produce disease modifying mediators and become lipid-laden foam cells, thus contributing to the overall cell burden of developing plaques (Central Illustration) (38).

Central Illustration: Systemic pathways and feedback loops regulating monocyte and macrophage trafficking in atherosclerosis, and the role of imaging.

The bone marrow is the primary site for the production of monocytes, which are generated from HSPCs. Once produced, monocytes are mobilized to the blood. HSPCs likewise mobilize to the blood according to circadian fluctuations. In the context of inflammation, HSPCs may give rise to monocytes in extramedullary sites such as the spleen. Bone marrow and spleen-derived monocytes circulate and infiltrate atherosclerotic lesions. Ly6c^high monocytes preferentially enter lesions where they differentiate to macrophages that can produce interleukin-1β (IL-1β), among other inflammatory products. Lesional macrophages can then proliferate and thus aggravate atherosclerosis. The various imaging approaches described in this review have shown enormous promise for understanding these monocyte/macrophage pathways at the systemic level.

Resident macrophages

The first observations of phagocytosis by macrophages were made by Elie Metchnikoff in starfish larvae in the early 1880’s. Yet, the prevailing concept of the monocyte-phagocyte system, which asserts that monocytes and macrophages derive exclusively from circulating bone marrow-derived cells (39), has only recently been challenged. As reviewed in Part 1 of this review series, in several tissues, resident macrophages develop from two distinct developmental programs. During vascular development, macrophage colonization further associates with a brief period of circulating monocyte recruitment shortly after birth (40). Importantly, pre- and perinatal macrophages survive into adulthood where they are largely maintained by mechanisms of self-renewal rather than additional input from the bone marrow.

Resident macrophages from different tissues exhibit considerable functional heterogeneity (41). This diversity is due partly to extrinsic factors conferred by the tissue in which they reside (42–44). Although vessels contain macrophages that have seeded shortly after birth (40), as reviewed in Part 2 of this review series, atherosclerotic lesions grow through a combination of continuous monocyte recruitment (32,33) and local macrophage proliferation (45). The concept of tissue resident macrophages is further discussed in Part 1 of this review series.

Imaging of leukocyte dynamics in atherosclerosis

The immunological and mechanistic studies described above underscore the importance of assessing atherosclerotic plaque inflammation in the context of immune organ activation. Correspondingly, non-invasive atherosclerosis imaging is shifting its focus from the vessel wall to a more comprehensive systems approach, to investigate monocyte/macrophage dynamics between the immune organs and the vessel wall in vivo and non-invasively. In the following section, we describe the most relevant imaging techniques that probe key processes in atherosclerosis macrophage dynamics. Focus will be on plaque monocyte (i) recruitment, plaque macrophage (ii) accumulation and (iii) proliferation, and (iv) macrophage egress from the plaque. Furthermore, we describe contemporary approaches aimed at integrating these imaging parameters with the in vivo assessment of hematopoietic organ and immune system activation in atherosclerosis.

Plaque monocyte recruitment

During atherosclerosis progression and under the influence of pro-inflammatory stimuli, bone marrow HSPCs migrate to the spleen, where they differentiate into monocytes. These splenic monocytes are recruited to inflamed organs, such as the atherosclerotic vessel wall (46), via the systemic circulation. Monocytes enter the arterial vessel wall through the extensive network of highly permeable neovessels that develop during plaque progression (47). Plaque neovessels are lined with activated endothelial cells, expressing specific adhesion molecules (such as vascular cell adhesion molecule 1 [VCAM-1] and intercellular adhesion molecule, Pselectin and E-selectin), which facilitate the translocation of monocytes from the blood stream into the vessel wall (48). Below we describe the most commonly used in vivo imaging techniques to quantify the extent and permeability of plaque neovessels, and the presence of adhesion molecules on their endothelial surface (Figure 1, Figure 2).

Figure 1: Schematic depiction of contrast agents and tracers that can be used to map monocyte recruitment (R), and the accumulation (A), proliferation (P) and egress (E) of macrophages from the vessel wall, as well as metabolic activity in hematopoietic organs (bone marrow and spleen).

Radiolabeled nanobodies and peptides, iron oxides or gadolinium based agents targeted towards adhesion molecules (i.e. VCAM-1, P-selectin) are able to quantify monocyte vessel wall recruitment. Several radiotracers (18F-FDG, 68Ga-DOTATATE, 18F mannose, 18F-NaF) and MRI or CT contrast agents (iron oxides, gold nanoparticles) can be used to quantify macrophage plaque accumulation. Macrophage proliferation can be quantified with the thymidine analog 18F-FLT, while fluorescent microbeads have been reported as a marker of egress.

Figure 2: In vivo imaging of monocyte recruitment, and plaque macrophage accumulation, proliferation and egress.

Monocyte recruitment (upper left panel) can be probed using DCEMRI to quantify endothelial permeability (148,149), or adhesion molecule imaging with targeted iron oxides for MRI or antibodies and nanobodies for PET imaging (79,81). Macrophage plaque accumulation (top right panel) can be quantified using 18F-FDG PET (90,150), or iron oxide MRI (120). Macrophage proliferation (bottom left panel) in plaque or hematopoietic organs (bone marrow and spleen) can be instead measured 18F-FLT (136). While no in vivo methods are currently available to image macrophage plaque egress (lower right panel), fluorescent beads (arrows) have been previously used for this purpose in terminal experiments (139). All images were obtained with permission from the respective publishers associated with these citations.

Plaque neovascularization and endothelial permeability

Contrast enhanced magnetic resonance imaging (MRI) with gadolinium-based agents (49–51) and, to a lesser extent ultrasound imaging with microbubbles (52), have both been applied to quantify plaque neovascularization and permeability in atherosclerosis. The tissue accumulation of gadolinium-based MRI contrast agents is related to plaque neovascularization, permeability and, therefore, indirectly reflects the active recruitment of monocytes and accumulation of macrophages in plaques. Contrast agent uptake in the vessel wall can be assessed by measuring the extent of signal enhancement in T1-weighted MR images (53–57) and/or T1 relaxation time after injection (51,58–61). For better quantification, the temporal pattern of contrast agent uptake can be analyzed by looking at signal enhancement over time using dynamic contrast enhanced (DCE-) MRI (62). This technique has been widely used in oncology and, more recently, has been applied to measure neovasculature and permeability in atherosclerosis. DCE-MRI involves the rapid serial acquisition of images during contrast agent injection, which allows tracking contrast agent uptake kinetics in the vessel wall. From the characteristic pattern of plaque signal enhancement, quantitative parameters, such as the percentage of microvascular volume (v_p) and plaque permeability (K^trans), can be derived (62). Human studies have demonstrated that carotid v_p and K^trans by DCE-MRI are higher in macrophage-rich plaques with abundant neovessels (50,63–65). Carotid K^trans was also found to be higher in patients with low blood levels of high density lipoprotein cholesterol (63) and increased c-reactive protein (CRP) (66) (the latter being a marker of systemic inflammation). Treatment with standard-of-care lipid lowering statins, a class of drugs known for its pleiotropic anti-inflammatory effects, was found to reduce v_p proportional with treatment duration (67). These clinical findings have been mirrored by equally encouraging results in the preclinical arena. In a rabbit model of atherosclerosis, a positive correlation was observed between the area under the curve of contrast agent uptake (representing K^trans) and plaque neovessels, macrophages (68,69) and permeability (49). A decrease in this area under the curve by DCEMRI, indicative of reduced plaque neovascularization and permeability, was induced by atorvastatin treatment (70), as well as plaque inflammation-reducing glucocorticoidencapsulating nanoparticles (71).

Adhesion molecules

Quantifying adhesion molecule expression requires the use of targeted tracers and highly sensitive imaging modalities, because of the activated endothelium’s relatively small mass in the arterial vessel wall. While several approaches based on MRI (72,73), ultrasound (74), and, more recently, optical coherence tomography (75) have been proposed, highly sensitive positron emission tomography (PET) (76,77) or single photon emission computed tomography (SPECT) are particularly attractive techniques for this purpose. Previously, PET and SPECT relied on the use of radiolabeled antibodies to target specific adhesion molecules (e.g., P-selectin specific antibodies) (77). However, antibody-based radiotracers are not ideal for vessel wall imaging, because of their relatively slow clearance from the circulation. Their blood stream persistence at the time of imaging confounds the already weak signal from the plaque endothelium, rendering accurate quantification very challenging. Unfortunately, simply delaying imaging to allow for the radiolabeled antibodies to clear from the systemic circulation is not a viable option to circumvent this problem, because the labeling radioisotopes often decay faster than the antibody is removed from the blood stream. Nanobodies, antibody fragments with very short blood halflives, have been recently proposed to overcome this obstacle. For example, accumulation of a technetium (Tc) 99 labeled nanobody targeted against VCAM-1 (99mTc-cAbVCAM1), formulated for SPECT imaging, was found to correlate with this adhesion molecule’s expression in the aortic arch of ApoE^−/− atherosclerosis-prone mice. Moreover, it allowed probing reduced vessel wall VCAM-1-expression in statin-treated mice (78). Alternatively, the same probe was radiolabeled with Fluorine-18 (18F) to allow PET imaging, which, with respect to SPECT, offers the advantage of higher spatial resolution (79) and better definition of tracer uptake in the small arterial vessel wall. Specific peptides that are internalized by VCAM-1-expressing cells (72,73,76) have also been developed as an alternative to the use of radiolabeled antibodies. For example, an 18F labeled tetrameric VCAM-1-targeted peptide (18F-4V) showed significantly higher accumulation in the aortas of ApoE^−/− atherosclerotic mice compared to wild type or statintreated ApoE^−/− mice, by both in vivo PET imaging and ex vivo nuclear methods. In ApoE^−/− mice, the 18F-4V signal correlated with VCAM-1 and CD68 gene expression (a marker specific to plaque macrophages) (76). In addition to radiotracers, iron oxide nanoparticles can be decorated by adhesion molecule-targeted peptides (72,73) to serve as molecular MRI agents. Tissue accumulation of iron oxide causes shortening of T2 and T2* relaxation times, which manifest as a loss of signal in T2- or T2*-weighted MR images. For example, MR imaging of a VCAM-1targeted linear peptide (VINP-28) labeled with ultra-small superparamagnetic particles of iron oxide (USPIO) showed significant signal loss in the aortic root of ApoE^−/− mice, indicative of adhesion molecule presence and active monocyte recruitment (73). While generally iron oxide MRI is less sensitive compared to PET and SPECT, some solutions have been proposed to increase the detection sensitivity of this technique to quantify vascular adhesion molecules. The simultaneous targeting of more than one adhesion molecule, and labeling with bigger size iron oxide particles (such as superparamagnetic particles of iron oxide [SPIO] or microparticles of iron oxide [MPIO]) has been proposed to allow improved detection of activated plaque endothelium. Examples of such an approach are VCAM-1 and E-selectin dual-targeted SPIOs (80) and VCAM-1 and P-selectin dual-targeted MPIOs (81–83), whose accumulation was found to be higher in symptomatic human carotid plaques (80) and also in the aortic root and vulnerable carotid plaques of ApoE^−/− atherosclerotic mice (82,83), and to correlate with adhesion molecule expression and macrophage content (80,81). More recently, dual-labeled VCAM-1 and E-selectin targeted MPIOs have been proposed as imaging agents for optical coherence tomography (75), an intravascular technique which, compared to MRI, PET or SPECT allows imaging at the unparalleled spatial resolution of 10–15 µm (84), and holds promise to improve the characterization of atherosclerotic plaques.

Plaque macrophage accumulation

The presence of abundant, active resident macrophages is a hallmark of unstable atherosclerotic plaques that are at high risk for hemorrhage, rupture, thrombosis and, consequently, for causing acute clinical events such as myocardial infarction and stroke. While many tracers have been proposed and investigated (85), until now imaging of plaque macrophage accumulation has relied on the use of non-specific PET radiotracers such as 18F fluorodeoxyglucose (FDG), or MRI contrast agents, such as non-targeted iron oxides (Figure 2).

FDG is a glucose analog rapidly internalized via glucose transporters (such as GLUT1 and GLUT3) by metabolically active cells, such as plaque macrophages (86–88). Different from glucose, FDG is not readily metabolized, therefore it accumulates in cells and when radiolabeled with 18F it can be detected using PET imaging. The 18F-FDG PET signal is higher in vulnerable human carotid plaques (89) and consistent with this it is also associated with areas containing macrophages, neovessels (90) and the lipid-rich necrotic plaque core (91). 18F-FDG accumulation has been correlated with plaque macrophage content and CD68 gene expression in humans (92,93) and in pre-clinical models (94–96) of atherosclerosis. These findings, together with the excellent reliability of this technique (97,98), have facilitated the translation of 18FFDG PET into pre-clinical and clinical drug trials as a surrogate imaging endpoint to evaluate the efficacy of new anti-atherosclerotic drugs. For example, in atherosclerotic rabbits, while aortic plaque 18F-FDG uptake increases during disease progression, it instead was found to stabilize or decrease after interventions aimed at modulating lipid levels either systemically such as switching from high fat diet to regular chow (99), treatment with statins (70,100) and the antidiabetic drug pioglitazone (101)) or locally in plaques (e.g. liver X receptor agonists (70)), and also treatment with anti-inflammatory liposomal steroids (71,102). In line with these findings, a human study recently demonstrated a dose-dependent reduction of 18F-FDG signal as early as 4 weeks after treatment with high-dose atorvastatin, with a further reduction seen after 12 weeks of treatment (103). The same therapeutic regimen was also found to have anti-inflammatory effects on periodontal inflammation, an independent risk factor for cardiovascular disease, thereby confirming the known pleiotropic effects of statins (104). Other notable examples of human clinical trials using 18F-FDG PET to evaluate the local plaque anti-inflammatory action (105) of novel anti-atherosclerotic drugs are the dal-PLAQUE trial (106,107) that investigated the cholesterylester transfer protein inhibitor dalcetrapib, the VIA-2291 trial (108) that evaluated the efficacy of 5-lipoxygenase inhibition, and the BMS-582949 trial (109,110) that focused on P38 mitogen-activated protein kinase inhibition. In the dal-PLAQUE trial (106,107), 18F-FDG PET (together with vessel wall MRI) showed no evidence of harm of dalcetrapib on the arterial wall in 24 months of treatment. While 18F-FDG uptake was not different between treatment and placebo groups in the index vessel, a targeted analysis of the carotid arteries showed a 7% reduction of tracer uptake in subjects treated with dalcetrapib. Conversely, the VIA-2291 (108) and BMS-582949 (109,110) trials showed no difference in vessel wall inflammation by 18FFDG PET between treatment and placebo groups, which was concordant with no systemic reduction of circulating hsCRP.

As an alternative to 18F-FDG PET, iron oxide nanoparticles can also be used to measure plaque macrophage accumulation. As mentioned above, tissue accumulation of nanoparticles containing an iron oxide core causes shortening of T2 and T2* relaxation times, causing a characteristic loss of signal in T2- and T2*-weighted MR images. Long-circulating iron oxide nanoparticles enter the arterial vessel wall through the permeable microvasculature (111) and are avidly phagocytized by plaque macrophages. Studies in atherosclerotic rabbits reveal accumulation of SPIOs (112), USPIOs (ferumoxtran-10 (113,114) and ferumoxytol (113)), P904 (115) and monocrystalline iron-oxide nanoparticles-47 (116) in macrophage-rich lesions (112,114). In humans, ferumoxtran-10 was found to accumulate in diseased carotid (117) arteries and to correlate with biomechanical stress (118), but not necessarily with carotid stenosis (119).

In both rabbits and humans, iron oxide MRI has also been applied to study pharmacological anti-atherosclerotic interventions. For example, MRI following USPIO P904 administration in rabbits revealed a reduction in plaque macrophage content after treatment with statins that was correlated with a lower 18F-FDG signal (100). In humans, the ATHEROMA study found a reduction in carotid ferumoxtran-10 uptake in humans after 12-weeks of high dose atorvastatin treatment (120). While iron oxide MRI is less sensitive and intrinsically less quantitative than 18F-FDG PET, advantages of this technique for the quantification of plaque macrophage accumulation include the lack of ionizing radiation and its relatively high spatial resolution. In principle, these features allow easier monitoring of disease progression, or regression, with treatment by allowing longitudinal serial imaging evaluations that do not expose subjects to repeated doses of radiation. However, recent debate on the safety and current small clinical market of iron-based compounds has hindered their application for clinical MRI (121). Despite these issues, there is renewed interest in certain iron-based MRI agents due to safety concerns regarding gadolinium-based MR contrast agents. Furthermore, while animal studies have indicated significant overlap between iron oxide MRI and 18F-FDG PET as readouts of plaque macrophage accumulation (100), recent human studies have suggested a more complex relationship between the two techniques (122), which requires further investigation.

In addition to 18F-FDG, a variety of novel PET tracers are currently being explored to improve the specificity of macrophage quantification in plaque. While 18F-FDG has been thoroughly validated to quantify macrophage accumulation in plaques, its uptake reports nonspecifically on cellular metabolic activity, and is therefore not necessarily limited to plaque macrophages. Novel, more specific tracers may provide a more accurate quantification of plaque macrophage burden, especially in challenging vascular territories such as the coronary arteries, where vessel wall 18F-FDG (121,122) uptake is confounded by the prominent background signal from metabolically active cardiomyocytes (123). Examples of these exploratory PET tracers include sodium fluoride (18F-NaF), gallium 68 (68Ga) DOTATATE, or new tracers specifically designed to target macrophages such as 18F labeled mannose binding to the macrophage mannose receptor (124). Among these, 18F-NaF has shown particular promise and is a PET tracer commonly used to quantify bone growth and remodeling. It was recently repurposed as a quantitative marker of active plaque microcalcification (125,126); a feature related to the presence of intense, necrotic plaque inflammation, as opposed to “healed” or stable plaques where macrocalcification is instead present. While the 18F-FDG coronary signal is challenging to quantify, 18F-NaF uptake can be quantified more easily and is consistently higher in patients with atherosclerosis versus controls (126). As another novel tracer, 68Ga-DOTATATE was recently approved by the Food and Drug Administration to better visualize somatostatin receptor 2-positive neuroendocrine tumors (127). The finding that somatostatin receptor 2 is also expressed on activated plaque macrophages (128,129) has spurred interest for the application of this tracer as a marker of plaque inflammation. In fact, accumulation of 68Ga-DOTATATE has been detected in mouse macrophage-rich plaques (128,130), and retrospective studies in cancer patients have demonstrated 68Ga-DOTATATE uptake in the atherosclerotic vessel wall (131–134). As expected, comparison with 18F-FDG showed only partial co-localization of the signal from the two tracers (131–134). The first prospective 68Ga-DOTATATE PET human atherosclerosis study demonstrated preferential accumulation in coronary and carotid culprit lesions, and a positive correlation with Framingham risk score (129). While 68Ga-DOTATATE uptake correlated with 18F-FDG signal and the latter was also able to differentiate culprit and non-culprit coronary plaques, 64% of the coronary FDG scans could not be interpreted and were confounded by the high myocardial background signal while 68Ga-DOTATATE scans were readable in all patients (129). The use of 68Ga-DOTATATE in the detection of atherosclerotic disease warrants further investigation.

Plaque macrophage proliferation

The 18F-labeled thymidine analog fluorothymidine (18F-FLT) was developed to quantify tumor cell proliferation (135). More recently, this tracer has been explored to quantify macrophage proliferation in atherosclerotic plaques by PET (136). Experiments in mice have revealed a higher 18F-FLT signal in the aorta, bone marrow and spleen of atherosclerotic ApoE^−/− mice compared to controls, indicating active proliferation of macrophages in atherosclerotic plaques and of monocyte/macrophage progenitors in hematopoietic organs. Treatment with 5 fluorouracil resulted in decreased accumulation of 18F-FLT. In atherosclerotic rabbits, the 18FFLT signal was correlated with aortic inflammation by 18F-FDG PET, while in patients both 18F-FLT and 18F-FDG uptake were higher in the carotid arteries and aortas of atherosclerotic patients versus healthy controls (Figure 2).

Plaque macrophage egress

The contribution of macrophage emigration (egress) out of atherosclerotic plaques to the regression of atherosclerosis is the subject of active investigation. While some studies have reported CCR7-mediated egress of CD68⁺ cells (such as dendritic cells and macrophages), that may be enhanced by treatment with statins (137) or liver X receptor agonists (138), others have instead pointed at suppressed monocyte recruitment as the main mechanism driving macrophage removal from atherosclerotic plaques (139). While no clinically available tracers exist to quantify in vivo macrophage egress from plaques non-invasively, preclinical mouse studies have relied on the use of fluorescent beads to track the fate of labeled monocytes during plaque regression (Figure 2) (139). In the future, we foresee the translation of these methods to in vivo pre-clinical or clinical studies. For example, the use of labeling with radiotracers for quantitative PET or SPECT imaging may help to advance our understanding of the mechanisms behind plaque macrophage egress and the role it plays during plaque regression. Last but not least, several imaging tracers and contrast agents are currently being investigated to quantify plaque macrophage apoptosis (140), an important pathophysiological process at different disease stages, which may contribute to macrophage “removal”, formation of the lipid rich necrotic core, and plaque destabilization (141).

Hematopoietic organ activation and system-level imaging of atherosclerosis

As described above, most studies on monocyte/macrophage dynamics in atherosclerosis have focused on investigating specific aspects of this process (i.e. monocyte recruitment versus macrophage accumulation, proliferation or egress), and how they modulate atherosclerosis progression or regression in the arterial vessel wall. However, based on the wealth of data reviewed in the first part of this article regarding the complex cross-talk between the hematopoietic system, inflammation and the vessel wall, more recent 18F-FDG PET studies have strived for a more comprehensive analysis. In particular, there has been a shift toward imaging at the systemic level, in an effort to study the complex interplay between vessel wall inflammation and immune organ activation during disease progression or after cardiovascular events. Initial reports showed that patients with recent myocardial infarction present with higher aortic 18FFDG uptake and increased circulating CRP, compared to controls with stable angina, despite a similar or lower aortic and coronary plaque burden (142). Further studies have indicated that the higher aortic inflammation and 18F-FDG uptake after myocardial infarction is mediated by a simultaneous increase in metabolic activity in the spleen and bone marrow, also reflected by an up-regulation of pro-inflammatory genes in circulating leukocytes (143). Increased splenic activation was found to be an independent predictor of subsequent cardiovascular events, suggesting that a “cardiosplenic axis” which modulates atherosclerotic plaque phenotype and risk profile may exist not only in mice (46), but also in humans. Additional 18F-FDG PET studies have also implicated the activated bone marrow as the main mediator of increased cardiovascular risk seen as a consequence of chronic psychosocial stress (144). More recently, chronic hyperactivity of the spleen and bone marrow, defined as high 18F-FDG signal, was also found in patients with stable atherosclerosis, compared to non-atherosclerotic controls (145). This indicates that immune-mediated mechanisms may not only drive disease aggravation after physical or other stressors (i.e. myocardial infarction or chronic emotional stress, respectively (144)), but also its gradual progression. Collectively, these studies have provided a platform for an organ systems approach to the combined, non-invasive assessment of plaque inflammation and immune organ activation in atherosclerosis.

Concluding remarks

In the future, integrating this system-level framework together with advanced in vivo imaging techniques that report more specifically on monocyte recruitment, and macrophage accumulation, proliferation and egress may allow for a more comprehensive evaluation of atherosclerosis macrophage dynamics and cardiovascular risk in patients. While initial attempts at the combined evaluation of some of these parameters gave variable results (64,146), the recent advent and increasing availability of multi-modality scanners, such as combined PET/MRI (147), may dramatically change this landscape. If current trends continue, we expect that the coming decade could well bring truly integrated, simultaneous evaluations of all the aspects of macrophage dynamics in atherosclerosis for the improved assessment of disease stage, response to treatment and patients’ risk for cardiovascular events.

Table 1:

Non-invasive imaging of macrophage dynamics: imaging modalities and biological targets

Organ	Process	Imaging Technique
		PET		SPECT		MRI		Others
		Tracer	Physiological process/Targ et	Tracer	Physiologi cal process/T arget	Technique/Co ntrast agent	Physiological process/Targ et	Technique/Contrast agent	Physiological process/Target
Plaque	Monocyte recruitment	64Cu antibodies (77)	P-selectin	99Tc nanobo dies (78)	VCAM-1	DCE-MRI with Gd-based agents (49– 51,53–61,63,66– 71,101,102)	Plaque neovasculariz ation and permeability	Ultrasound with microbubbles (52)	Plaque neovascularizat ion and permeability
		18F nanobodies (79)	VCAM-1			USPIO labeled peptides (72,73)	VCAM-1
		18F tetrameric peptide (76)	VCAM-1			Dual-targeted SPIOs (80)	VCAM-1 and E-selectin	Optical coherence tomography with dual- targeted MPIOs (75)	VCAM-1 and E- selectin
						Dual-targeted MPIOs (81–83)	VCAM-1 and P-selectin
	Macrophage accumulatio n	18F-FDG (70,71,89– 103,106– 110,124,151, 152)	Metabolically active cells (plaque macrophages)			Iron oxide particles (111– 122,153,154)	Plaque macrophages		CT with gold nanoparticles
		68Ga- DOTATATE (128–131)	SSTR2 positive plaque macrophages
		18F-mannose (124)	Macrophages through the macrophage mannose receptor (MMR)
		18F-NaF (125,126)	Active Microcalcifica tion
	Macrophage proliferation	18F-FLT (136)	Proliferating cells
	Macrophage egress								Fluorescent beads (139)
Hemato- Poietic organs	Activation	18F-FDG (142– 145)	Metabolically active cells (hematopoietic c progenitors)

Abbreviations

18F-4V

18F labeled tetrameric VCAM-1-targeted peptide

18F-FLT

18F-labeled thymidine analog fluorothymidine

CAR

CXCL12-abundant reticular

CRP

c-reactive protein

DCE

dynamic contrast enhanced

HSC

hematopoietic stem cell

HSPC

hematopoietic stem and progenitor cell

K^trans

plaque permeability (assessed by DCE-MRI)

LT-HSC

long-term hematopoietic stem cells

MPIO

microparticles of iron oxide

SPIO

superparamagnetic particles of iron oxide

USPIO

ultra-small superparamagnetic particles of iron oxide

VCAM

vascular cell adhesion molecule

v_p

microvascular volume (assessed by DCE-MRI)