Characterization and molecular detection of atherothrombosis by magnetic resonance—potential tools for individual risk assessment and diagnostics

Abstract

This review focuses on recent non‐invasive or minimally invasive magnetic resonance (MR) approaches to study atherothrombosis. The potential benefits of combining diverse metabolic information obtained by the variety of MR techniques from tissues in vivo and ex vivo and from body fluids in vitro are also briefly discussed. A well established methodology is available for lipoprotein subclass quantification from plasma by ¹H MR spectroscopy providing information for assessing the long‐term risk of atherosclerosis. Multi‐contrast MR imaging in vivo relying on endogenous contrast allows partial characterization of components in atherothrombotic plaques. The use of exogenous contrast agents in MR angiography enhances blood‐tissue contrast and provides functional information on plaque metabolism, improving plaque characterization and assessment of plaque vulnerability by MR imaging. Recent applications of molecular targeted MR imaging have revealed novel opportunities for specific early detection of atherothrombotic processes, such as angiogenesis and accumulation of macrophages. Currently, MR imaging and spectroscopy can produce such metabolic in vivo and in vitro information that in combination could facilitate the screening, identification and follow‐up of cardiovascularly vulnerable patients in research settings. The recent developments imply that in the near future MR techniques will be part of clinical protocols for individual diagnostics in atherothrombosis.

• Atherosclerosis
• magnetic resonance imaging
• magnetic resonance spectroscopy
• molecular diagnostic techniques
• molecular probe techniques
• thrombosis

View correction statement:

ERRATUM

Abbreviations
ACS	=	acute coronary syndrome
CAD	=	coronary artery disease
CHD	=	coronary heart disease
HDL	=	high‐density lipoprotein
LDL	=	low‐density lipoprotein
MR	=	magnetic resonance
MRA	=	magnetic resonance angiography
MRI	=	magnetic resonance imaging
MRS	=	magnetic resonance spectroscopy
rHDL	=	reconstituted high‐density lipoprotein
VLDL	=	very low‐density lipoprotein

Atherothrombosis—life is full of risks

Atherosclerosis is a diffuse systemic disease that is characterized by the local build‐up of lipid‐rich plaques within the walls of large arteries. The processes are multigenetic, being influenced also by dietary and environmental factors, and are apparent as early as the second decade in life with an increased incidence in the elderly 1. Inflammatory processes are in a key role in the development of atherothrombosis 2. In advanced disease plaque rupture as well as erosion can lead to arterial thrombosis and thereby to an acute coronary syndrome (ACS) 3.

A common practice for assessing the risk of an individual for atherosclerosis is to measure the lipoprotein lipids in blood plasma or serum: high levels of low‐density lipoprotein (LDL) and low levels of high‐density lipoprotein (HDL) cholesterol are pro‐atherogenic 4. The lipoprotein subclasses are also characteristically related to the risk of atherosclerosis 5. Thus, in principle, with the lipoprotein profile of a person one can approximate the risk value for a clinical manifestation of atherosclerosis, i.e., the probability of a cardiovascular event, in a specified period of time. However, the predictive accuracy of the lipoprotein and other blood compartment‐associated risk factors is too poor to be clinically useful in predicting the individual incidence of ACS in the near future 6,7. Therefore, the biochemistry of plasma seems not to be able to consistently reflect the atherothrombotic processes in the vessel wall. This is not surprising since atherothrombotic plaques take many forms depending on the relative proportion of, e.g., a lipid‐rich core, a fibrous cap, vascular smooth‐muscle cells, extracellular matrix components, and inflammatory cells 3.

Currently the standard method for the evaluation of coronary artery disease (CAD) is invasive angiographic visualization and quantification of significant coronary stenosis 8. Angiography (with either X‐ray or magnetic resonance (MR)) images only the vessel lumen and the silhouette of lesions that impinge on the lumen 9. However, it is possible that lumen calibre preserves unaltered or minimally altered despite the presence of a large plaque 10. In view of that, for example, the coronary artery which subsequently occluded had less than a 50% stenosis on the first angiogram in two‐thirds of patients 11. In other words, ACSs often result from disruption of mildly stenotic plaques with large extracellular lipid cores, thin fibrous caps and inflammatory cell infiltrates 3. In the coronary arteries, the pathobiological behaviour of plaque is consequently determined by its molecular composition and not by its size. In addition, there can be striking heterogeneity in the composition of human atherothrombotic plaques even in the same individual 3. Accordingly, angiography alone is not adequate for diagnosing a CAD or the risk for an ACS.

Hence, there is a fundamental requirement for non‐invasive in vivo imaging tools, firstly, to allow for direct detection of atherothrombotic tissues, i.e., to identify and probe atherothrombosis 12, and, secondly, to characterize the composition of various plaques in order to understand and assess their risk for rupture and thrombotic complications, i.e., to predict an ACS on an individual basis 13. As a summary, Figure 1 presents a potential scheme for the utilization of MR methodologies in the assessment of long‐term and short‐term risk for atherothrombotic events. This scheme can be seen as one option to elucidate the potential of MR in detecting individual intermediate atherothrombotic end‐points and utilizing their prognostic value before the occurrence of a definite end‐point as discussed in Raggi et al. 14.

Figure 1. A potential scheme utilizing magnetic resonance(MR) methodologies in the risk assessment of long‐term risk for atherothrombotic events (non‐symptomatic individuals) and of short‐term risk for recurrent cardiovascular events after an experienced acute coronary syndrome (ACS) (symptomatic patients). At risk assessment point I the molecular constituents of serum, including lipoprotein subclasses, could be assessed by in vitro¹H magnetic resonance spectroscopy (MRS) metabonomics for non‐symptomatic individuals. If high long‐term risk for atherothrombotic events is indicated, non‐invasive in vivo magnetic resonance imaging (MRI) could follow for the potential detection of plaque (risk assessment point IIa) and subsequent compositional evaluation of the vulnerability of the detected plaque(s) for rupture or erosion (IIb). Depending on the outcome from the plaque detection and assessment by MRI the individual could accordingly be directed for further actions. If vulnerable plaque at point IIb would be detected, considerations for aggressive drug therapy or invasive therapies such as angiographic stenting or bypass surgery would be needed. In the case of an individual with an experienced ACS (III) in vitro¹H MRS metabonomics could be used to complement the clinical protocols when evaluating the risk for recurrent cardiovascular events and the proper individual treatment options. For some symptomatic patients in vivo MRI might also be feasible at point III for direct assessment of plaque composition and vulnerability. This scheme can be seen as one option to elucidate the potential of MR in detecting individual intermediate atherothrombotic end‐points and utilizing their prognostic value before the occurrence of a definite end‐point. The recent MR findings and developments summarized in this review awaken confidence that this kind of scheme might be operational in the near future saving both human suffering and societal health costs.

Key messages

• Magnetic resonance imaging (MRI) is a non‐invasive method allowing detection, characterization and follow‐up of atherothrombosis in vivo.

• The use of in vitro¹H magnetic resonance spectroscopy (MRS), to detect lipoprotein subclasses and plasma metabolites, together with in vivo multi‐contrast MRI, to characterize plaque components, could provide a genuinely new approach to uncover individual intermediate atherothrombosis before any clinical manifestations.

• In general, contrast agents enhance characterization of atherothrombotic processes, and, in particular, molecular targeted MRI offers a kind of in vivo equivalent to immunohistochemistry for detecting early atherosclerotic processes, such as accumulation of macrophages and angiogenesis.

A vulnerable patient—magnetic resonance imaging and spectroscopy

Vulnerable plaques are not the only culprit factors for the ACSs but vulnerable blood (prone to thrombosis) and vulnerable myocardium (prone to fatal arrhythmia) also play an important role in the outcome. Hence, a concept of a ‘cardiovascular vulnerable patient’ was recently proposed to define subjects susceptible to an ACS or sudden cardiac death based on plaque, blood, or myocardial vulnerability (for example, 1‐year risk ⩾5%) 6,7. For the detection of a vulnerable plaque, five major criteria were presented, namely active inflammation, a thin cap with a large lipid core, endothelial denudation with superficial platelet aggregation, fissured or injured plaque and severe stenosis. Interestingly, magnetic resonance imaging (MRI) seems to be the only non‐invasive methodology capable of detecting all these major criteria for plaque vulnerability 6. Several serum markers of vulnerable blood (reflecting metabolic and immune disorders), e.g., abnormal lipoprotein profiles, diabetes, hypertriglyceridaemia, and lipid peroxidation are also such that they influence characteristically the MR spectroscopy (MRS) information of plasma and serum 15.

In this review we will discuss the potential use of MRI and MRS in the characterization and risk assessment of atherothrombosis. The applications of endogenous contrast in MRI will be dealt with before discussing the use of exogenous contrast agents. Molecular targeted MRI is then discussed in relation to the molecular biology of atherothrombosis followed by a brief note on the role of MRS in lipoprotein quantification and plaque characterization.

Magnetic resonance imaging and angiography

It should be noted that because the MR method operates at the radio frequency region of the electromagnetic radiation it is intrinsically non‐ionizing, non‐destructive and non‐invasive. In routine clinical settings MRI offers three‐dimensional data with high spatial resolution (approximately 1 mm) and in advanced research settings for two decades less (∼10 µm). The MRI techniques for the evaluation of vascular anatomy are collectively called magnetic resonance angiography (MRA). Currently, MRI and MRA are often seen as a favourable technology for the detection, visualization and characterization of atherothrombosis: also two extensive text‐books of cardiovascular magnetic resonance (CMR) have recently appeared 16,17.

Anatomical aspects

X‐ray tomography, with the aid of ionizing radiation and nephrotoxic contrast media, is at present the most accurate and robust way to image coronary arteries 18. However, a new non‐invasive MRA technique, that does not even use a contrast agent, now allows imaging of the whole coronary artery tree within 30 minutes with 0.55×0.55×0.75 mm³ voxel size 19.

Characterization and quantification of plaque using endogenous contrast

By using multi‐contrast MRI, i.e., a combination of T1‐, T2‐, and proton density‐weighted images, with sub‐mm spatial resolution, plaque anatomy and components have been determined in various test animal and human artery specimens ex vivo and in vivo20. Unfortunately, in general there is an overlap of signal intensities between the different MRI pulse sequences (e.g., T1‐, T2‐, and proton density weighting) and the various plaque components (e.g., fibrous cap, lipid core, calcification, thrombus, media and adventitia), which makes unique identification of the plaque components difficult and requires particular attention on the image analysis. Recently, however, a spatially enhanced iterative cluster analysis has been developed for the automated analysis of multi‐contrast ex vivo MR images of human coronary artery specimens. It turned out that the resulting MR segmentation had an excellent match with histopathology images for all American Heart Association plaque classification types I–VI 21. A result using the new cluster analysis method is illustrated in Figure 2 together with an applied colour composite MR image as well as a histopathology for comparison. In addition, a stepwise logistic regression model has been developed to identify human atherosclerotic carotid plaque components based on multi‐sequence in vivo MRI 22. As a complement to the above plaque classifications, a detailed description of the MRI appearance of a plaque rupture‐associated thrombosis in histologically validated platelet‐rich thrombi has recently been presented 23. These studies jointly suggest that MRI seems to be intrinsically capable of imaging the whole cascade of atherothrombotic phenomena non‐invasively without contrast agents.

Figure 2. Illustration of the spatially enhanced iterative cluster analysis for the multi‐contrast ex vivo magnetic resonance (MR) images of human coronary artery specimens. On the left are shown the T₁‐, proton density, and T₂‐weighted (T₁W, PDW and T₂W, respectively) axial plane images that were assigned red (R), green (G), and blue (B) colours, respectively, and combined to yield a colour composite and pseudo‐colour cluster image of the coronary artery with colour distributions corresponding to different tissue types. On the right a comparable histopathology image via a combined Masson's elastic (CME) stain is shown. This type Vb‐Vc lesion exhibits eccentric fibrous deposits and a dark area of calcification. The identified areas include dense fibrous tissue (df), calcium deposits (cal), media (med), and fibrous cap (fc). The CME histopathology shows an excellent agreement with the MR images. (With permission from Itskovich et al. 2004 21.)

In apoE‐/‐ mice, MRI has been shown to accurately quantify aortic atherosclerosis in vivo24. Importantly for genetic and pharmacological studies of plaque composition in the mouse models of atherosclerosis, a new method, based on ex vivo three‐dimensional MRI at 11.7 T and at a resolution of 47×47×62.5 µm³ in apoE‐/‐ mice, was recently developed that enables accurate quantification of lipid‐rich/necrotic core and cell‐rich cap areas in atherosclerotic lesions 25.

While the non‐invasive MRI of sub‐mm aortic plaques tackles such problems as obtaining sufficient sensitivity and exclusion of artefacts due to respiratory motion and blood flow, the ultimate goal to image coronary artery plaques tends to amplify all these problems 20. However, human coronary artery lumen and wall imaging has been achieved in vivo by the so‐called black‐blood MRI method, the application of which maximizes the signal from static tissue and results in a signal void in blood 26. In addition, three‐dimensional black‐blood MRI of human coronary vessel wall has detected positive arterial remodelling in patients with non‐significant CAD 27. These results suggest that non‐invasive in vivo MRI investigations of human coronaries has potential to improve risk stratification in individual patients, thereby being able to guide and follow‐up treatment options and ultimately to help in preventing ACSs.

Due to its non‐invasiveness MRI is particularly well suited for in vivo follow‐up and pharmacological studies. An MRI study of the participants of the Framingham Heart Study offspring cohort, who were free of clinically apparent CAD, identified aortic atherosclerosis in 38% of women and in 41% of men. The prevalence of atherosclerosis increased with age and was more apparent in the abdominal than in the thoracic aorta 28. MRI can also detect lowered aortic distensibility and impaired flow‐mediated vasodilation indicative of endothelial dysfunction in young healthy smokers 29. These data demonstrate the ability of MRI in detecting subclinical atherosclerotic disease and to assist in the risk stratification and treatment of patients with an asymptomatic disease.

The clinical significance of carotid plaque characterization has been demonstrated in a case‐control study of patients undergoing carotid endarterectomy: a recent (within 90 days) history of transient ischaemic attack or stroke was strongly associated with the presence of thin or ruptured plaque identified preoperatively by MRI. The risk of recent ischaemic neurological symptoms was increased by a remarkable 23‐fold in cases in which ruptured plaque was identified compared with a thick fibrous cap 30. These observations are enhancing the use of MRI in prospective evaluation of plaque behaviour. However, in clinical practice detection of fibrotic caps or plaque components is not currently used, but the treatment decision for endarterectomy is based on the luminal stenosis of carotid arteries according to clinical multi‐centre trials such as the European Carotid Surgery Trial 31.

Statins are widely used to treat hypercholesterolaemia and atherosclerotic disease. A regression of aortic as well as carotid plaques after lipid‐lowering therapy with simvastatin has been demonstrated using MRI. After 12 months of treatment, MRI showed that statistically significant reductions occurred in vessel wall thickness and vessel wall area both in carotid arteries and aorta, but no increase in vessel lumen area was detected. The increase in lumen size measurement achieved statistical significance after 18 and 24 months of treatment for both carotid and aortic lesions 32. More recently, 1‐year lipid‐lowering therapy with 20‐mg atorvastatin was demonstrated to induce a significant plaque regression in the thoracic aorta (along a marked LDL cholesterol reduction). In the abdominal aorta, even the 20‐mg dose resulted in only a retardation of plaque progression, and a significant progression was observed with the 5‐mg dose treatment 33. These results demonstrate that MRI can be used to monitor the effects of pharmacological interventions to atherosclerosis and suggest plaque‐specific follow‐up to optimize the individual treatment options. However, monitoring the effects within an individual patient calls for high inter‐study reproducibility that may not be currently available in clinical routine at 1.5 T.

It has also been revealed, using MRI and multivariate analysis, that LDL cholesterol levels correlate with the plaque extent in the thoracic aorta but not in the abdominal aorta 34. Therefore, plaque formation in the thoracic aorta appears to be more closely associated with hypercholesterolaemia than in the abdominal aorta. Hence, lipid‐lowering therapy is more likely to be effective for plaque regression in the thoracic aorta. In general, these data point towards a specific role of plasma lipoproteins in the progression and regression of atherothrombotic plaques and thereby call for follow‐up studies that combine quantification of plasma lipoprotein levels and plaque components. A genuinely new MR approach would be the use of in vitro¹H MRS of plasma for the lipoprotein subclass quantification 5,15 and in vivo multi‐contrast MRI to characterize the plaque components 20; see Figure 1.

It has recently been shown in rabbits at 2 T that MRI can be used to monitor progression and regression of atherosclerosis also in a short‐term model (cholesterol feeding and normal diet both for 12 weeks). The results indicated that positive remodelling, a process in which the arteries compensate for the progressive growth of atherosclerotic plaques by increasing their external circumference and thereby maintaining the lumen diameter, occurs early during progression and persists through regression of atherosclerotic lesions 35. An interesting finding, also from an MRI study of rabbits treated with thromboxane A2 receptor inhibitor (to potentially improve endothelial function and to reduce the inflammatory component of atherosclerosis in addition to antiplatelet activity), is that regression of already established advanced atherosclerotic lesions occurs after 6 months of treatment while maintaining the atherogenic diet. Also, a change in the phenotype of the atherosclerotic plaques was observed in the treatment group using histological stainings 36. The reduction in the content of macrophages, apoptotic cells, metalloproteinases and endothelin‐1, and the increase in vascular smooth muscle cells, suggested that selective inhibition of thromboxane A2 pathway may transform lesions towards a more stable plaque phenotype. These findings also demonstrate the additional value of non‐invasive MR techniques, enabling molecular and cellular specification, in the clarification and follow‐up of pathophysiological processes of atherothrombosis in vivo.

Plaque imaging with the aid of exogenous contrast agents

Contrast agents serve as a fascinating option to enhance the capability of MRI and MRA in cardiovascular imaging and they have been extensively used in the detection of ischaemic myocardial injuries and vascular stenosis 37. A typical MR contrast agent, such as widely clinically used gadolinium chelate Gd‐DTPA (gadopenetate, Magnevist), is paramagnetic material containing unpaired electrons, which influence the relaxation properties and thereby the intensity (contrast) of the detectable nuclei. The MR contrast agents (unlike in X‐ray studies) are not directly observable. The MR contrast agents are a heterogeneous group of molecules, the effects of which in the actual MR experiment depend on, e.g., the structural properties, dosage, distribution, tissue, pulse sequence, and magnetic field strength 38.

The advantage of using contrast agents in the MRI studies of atherothrombosis is two‐fold. First, they may improve tissue contrast between the plaque components as illustrated in Figure 3 showing an example of Gd‐DTPA‐enhanced MRI differentiation of a fibrous cap 39. Second, the distribution of contrast agents in the body as well as the time course of observed contrast may provide valuable structural and functional information in tissue characterizations, such as the extent of vascularization and the size of an intercellular matrix. In MRA gadolinium is used to enhance the blood‐tissue contrast, i.e., to improve the detection of anatomical details such as a vessel lumen.

Figure 3. A T₁‐weighted image from a patient with abdominal aortic aneurysm demonstrating bright signal within thrombus and darker signal in lipid core towards vessel wall (on the left). The cap region is faintly seen on the T₁‐weighted image. A comparable T₁‐weighted image, but several minutes after infusion of Gd‐DTPA, is shown on the right. Note the improved delineation of fibrous cap by bright signal within (arrow). (With permission from Kramer et al. 2004 39.)

In atherothrombosis new microvessels form that may be associated with local inflammation 3. The presence of new vessels has also been associated with carotid plaque instability, and in addition they have been connected to abnormal permeation of plasma proteins, such as albumin and fibrinogen 37. Using a gadolinium‐based contrast agent in MRI of carotid arteries, plaque microvascularization was found to affect the contrast enhancement thereby suggesting potential improvement in the detection of vulnerable carotid plaque 40. Using MS‐325, a gadolinium‐based contrast agent that binds albumin, areas of high signal intensity have been observed in the aortic or iliac arterial wall. This finding may not only reflect increased plaque vascularity but also leaking of these microvessels due to active inflammatory processes 37.

The most extensively used MR contrast agent Gd‐DTPA is highly hydrophilic and thus does not penetrate cellular membranes. In contrast, gadofluorine, also a gadolinium‐based contrast agent, is lipophilic. As a matter of fact, gadofluorine has recently been shown to enhance MRI of atherosclerotic plaques and to improve plaque detection in non‐stenotic lesions, not visible when using endogenous contrast 41. It has also been suggested that gadofluorine has a high affinity towards lipid‐rich plaque areas 42. On the other hand, a recent study using a Gd‐DTPA‐type of a contrast agent reports good results in quantifying intact fibrous cap and lipid‐rich necrotic core size in human atherosclerotic carotid plaques 43.

Molecular targeted MRI

The detection of contrast in conventional MRI relies on anatomical, physiological, or metabolic heterogeneity in the tissues. On the other hand, the rising interest in applying ‘molecular imaging’ is based on an assumption that disease processes, i.e., atherothrombosis in this particular case, can be detected and characterized via specific detections along the molecular and cellular development and pathways. It has been proposed that molecular imaging offers an in vivo equivalent to immunohistochemistry, in situ hybridization or in situ zymography 1.

The general idea in molecular imaging is to target the specific molecule of interest using a specific ligand, which is conjugated to a carrier particle loaded with signal affecting elements. For example, an α_vβ₃‐integrin may be targeted using a monoclonal antibody conjugated to a suitable nanoparticle loaded with paramagnetic chelates. In an effective situation the carrier particles, and thereby also the MR signal‐affecting elements, accumulate in the vicinity of the target molecules in the tissue thus potentially modifying the detectable signal. Though the molecular imaging appears very appealing there are several pitfalls in both the biological and MRI side of potential applications 1,12,37. The contrast agent must identify the target tissue with high specificity and induce distinctive signal changes within the corresponding voxels in comparison to voxels containing untargeted tissue. It is notable that if only a small number of paramagnetic chelates are conjugated to the targeting molecule, sufficient contrast is difficult to achieve. Therefore, specific nanoparticle approaches have been developed in which extensive shipment of paramagnetic or superparamagnetic agents is feasible 1. However, a limitation of the nanoparticle approach is the steric hindrance imposed by such relatively large particles. Also the issues of toxicity and systemic clearance of the agent are important. The following chapters briefly introduce a few recent molecular targeted MRI studies of fundamental importance in atherothrombosis research.

Angiogenesis and α_vβ₃‐integrins—perfluorocarbon nanoparticles

Angiogenesis is involved in the early phases of atherosclerotic plaque development and it may also be one of the issues relating to vulnerable plaque. The α_vβ₃‐integrin is expressed on endothelial cells during angiogenesis, but not in the resting state, and atherosclerotic lesions are known to be highly vascular in comparison to normal vessel tissue 44. An MRI application at 1.5 T using α_vβ₃‐integrin‐targeted, paramagnetic perfluorocarbon nanoparticles in hyperlipidaemic rabbits was successful in detecting angiogenesis within the aortic wall (but it could not be spatially resolved to a specific vascular layer, e.g., adventitia) 45. The nanoparticles used were ∼273 nm in diameter and contained ∼94 200 Gd³⁺‐atoms per particle. These data indicated that α_vβ₃‐integrin‐targeted, paramagnetic nanoparticles might be used for molecular imaging of neovascularization throughout the vascular wall and that α_vβ₃‐integrin expression shows considerable heterogeneity within individual aortic segments substantiating that early atherosclerosis evolves as a diffuse process. These findings are demonstrated in Figure 4. Variation in the gross severity of atherosclerosis from the diaphragm to the renal arteries generally corresponded to the calculated MRI signal enhancement and the magnitude of neovascular proliferation observed histologically. These findings are in agreement with an analogous heterogeneous distribution of aortic foam cells in cholesterol‐fed rabbits 46,47. In contradiction to the MRI results with α_vβ₃‐integrin‐targeted nanoparticles, delayed contrast enhancement of the vessel wall with Gd‐DTPA revealed no significant difference between cholesterol‐fed and control diet rabbits. Thus, non‐specific contrast agents seem insensitive to the subtle vascular changes associated with the early development of atherosclerosis.

Figure 4. A magnetic resonance imaging(MRI) application at 1.5 T using α_vβ₃‐integrin‐targeted, paramagnetic perfluorocarbon nanoparticles in a hyperlipidaemic rabbits to detect angiogenesis within the aortic wall. Per cent enhancement maps (false‐coloured from blue to red) are illustrated from individual aortic segments at renal artery (A), mid‐aorta (B), and diaphragm (C) 2 hours after infusion of the α_vβ₃‐integrin‐targeted nanoparticles. (With permission from Winter et al. 2003 45.)

In addition to the detection of neovascularization in early atherosclerotic lesions, it has been demonstrated that α_vβ₃‐integrin‐targeted nanoparticles can also be used to specifically and locally deliver potent pharmaceutical agents, such as fumagillin, to elicit a marked antiangiogenic effect with even a single dose 44. Atherosclerotic rabbits were treated with α_vβ₃‐integrin‐targeted nanoparticles including 0 or 0.2 mole % fumagillin. The enhancement of the MRI signal, averaged over all imaged slices from the renal artery to the diaphragm at 2 hours after injection, was used as an integrated, quantitative assessment of the atherosclerotic burden. This was noted identical at baseline for animals treated with or without fumagillin. One week later, reassessment of residual aortic angiogenic activity with α_vβ₃‐integrin‐targeted paramagnetic nanoparticles (with no drug) revealed that MR signal enhancement in rabbits given α_vβ₃‐integrin‐targeted nanoparticles including fumagillin was markedly reduced whereas the MR signal from the neovasculature in control animals was unchanged. Thus, it has been postulated that α_vβ₃‐integrin‐targeted nanoparticles are an entity that might be used for specific detection, treatment and follow‐up of early atherosclerotic lesions 44.

Macrophages—ultrasmall particles of iron oxide

Iron oxides are superparamagnetic materials with potentially huge effects on the relaxation contrast of protons. A wide variety of iron oxide‐based nanoparticles have been developed that differ in hydrodynamic particle size and surface coating material 44. In general, these particles are categorized into superparamagnetic iron oxides (SPIOs) (diameter from 50 to 500 nm) and ultrasmall superparamagnetic iron oxides (USPIOs) (diameter <50 nm), the size of which has an important effect on their physicochemical and pharmacokinetic properties. USPIOs provide adequate circulation time for tissue penetration (the half‐life of USPIOs in human circulation is ∼30 hours) but they also require significant time, usually 24 hours, for sufficient background signal clearance. Because these agents often create an extended image void, the proximate anatomy surrounding key targeted points of interest may be obscured, particularly in high‐resolution MRI. This opposes the behaviour of paramagnetic contrast agents that can create rapid contrast enhancement signal which also allows for subtle tissue contrast features to be clearly discerned at high spatial resolutions.

Contrast agents which specifically identify components of vulnerable plaque are of considerable interest. Morphological characteristics of vulnerability are a predominance of macrophages, ulceration, intraplaque haemorrhage, and thrombosis 3. Angiogenic vessels within atherosclerotic plaques or inflammatory allografts permit extravasation of USPIOs, which are phagocytosed by resident tissue macrophages. The accumulated iron oxide is readily visualized in MRI as regions of signal void, which grossly correlate spatially with active atherosclerosis 48.

In an MRI study of rabbits, conventional extracellular contrast agent Gd‐DOTA (gadoterate, Dotarem) failed to reveal any abnormality in either the hyperlipidaemic or control rabbits in the aortic lumen before administration of USPIOs. In case of hyperlipidaemic rabbits, 5 days after the administration of USPIOs, the aortic wall exhibited marked irregularities in MRI that was not observed in controls 46. Histopathological analysis showed uptake of Fe‐particles in the plaque macrophages of the hyperlipidaemic rabbits but not in the controls. Fe‐particles are known to cause susceptibility effects and this can explain signal intensity irregularities in the aortic wall in the MRI of hyperlipidaemic rabbits. Interestingly, visual examination of the aortic walls of hyperlipidaemic as well as control rabbits did not reveal any appreciable irregularities. Thus, accumulation of USPIOs in macrophages in the aortic wall seems to provide delicate structural information on atherosclerotic processes beyond visual inspection. These observations may have some clinical relevance since increased macrophage activity is associated with active inflammation, a marker of vulnerable plaque 3,49.

In a retrospective analysis in bladder or prostate cancer patients who had originally received USPIOs for staging lymph node metastases, signal loss in the arterial wall was observed in 7 of 19 patients, indicating some accumulation of USPIOs in human atherosclerotic plaques 50. In another study, in which MRI was performed before and after administration of USPIOs on 11 symptomatic patients, scheduled for carotid endarterectomy, histology and electron microscopy analyses of the plaques showed USPIOs primarily in macrophages within the plaques in 10 of 11 patients 48. Histology showed USPIOs in macrophages of carotid plaques in 75% of the ruptured and rupture‐prone lesions and in 7% of the stable lesions. Of the patients with USPIO uptake, significant signal decrease in the post‐USPIO MRI were observed in the vessel wall in 54% of the quadrants after 24 hours. These results indicate that the tested USPIOs accumulate predominantly in macrophages in ruptured and rupture‐prone human atherosclerotic lesions and that this can induce significant signal changes in in vivo MRI 24 hours after intravenous administration of USPIOs. Thus, USPIO‐enhanced MRI seems to be a promising way to identify high‐risk plaques.

In human studies the dosages of USPIOs have been noticeably smaller than in animal studies. As a result, the observed signal intensity reductions are relatively minor calling for particular attention on image analysis and reproducibility 51. In addition, the predictive value of USPIO‐enhanced plaque imaging to detect abnormality associated with increased stroke risk has to be confirmed in a prospective study. Nevertheless, the current evidence suggests that the USPIO‐enhanced imaging has a considerable potential for becoming a part of clinical routine in the assessment of patients with carotid atherosclerosis. The use of USPIOs has clear pathophysiological background. The histopathological and experimental studies have also confirmed the capability of USPIOs to indicate regional inflammation using clinical scanners.

Macrophages—recombinant high‐density‐lipoprotein‐like nanoparticles

A quite common problem with many available paramagnetic MR contrast agent constructs is that they are not capable of delivering a sufficiently large amount of gadolinium ions to the tissue in order to induce detectable changes in the MRI contrast. Also, some of the MR contrast agents may be too large to have free access to biochemical epitopes within the vascular subendothelium of atherosclerotic plaques. Recently a high‐density lipoprotein (HDL)‐like nanoparticle contrast agent, that selectively targets atherosclerotic plaques, has been introduced 52. The HDL particles are the key players in reverse cholesterol transport and have several advantages: a small size (diameter of 7–12 nm); protein components that are endogenous, biodegradable, and do not trigger immunoreactions; and the particles are not recognized by the reticuloendothelial system. Furthermore, HDL‐like particles can easily be reconstituted and they can carry a considerable contrast agent pay‐load.

Thus, reconstituted HDL particles (rHDL) of approximately 9 nm diameter were prepared containing 15–20 molecules of Gd‐DTPA‐DMPE (1,2‐dimyristoyl‐sn‐glycero‐3‐phosphoethanolamine diethylenetriaminepentaacetic acid), a phospholipid‐based contrast agent, and also a fluorescent phospholipid with a green emission, NBD‐DPPE (1,2‐dipalmitoyl‐sn‐glycero‐3‐phosphoethanolamine‐N‐(7‐nitro‐2,1,3‐benzoxadiazol‐4‐yl)), for confocal fluorescence microscopy 52. The general structure of the rHDL particles is shown in Figure 5. In hyperlipidaemic apoE‐knockout mice, sequential MRI showed that Gd‐DTPA‐DMPE localized predominantly at the atherosclerotic plaque by 24 h after injection (see Figure 5). Furthermore, by 48 h after rHDL injection, the intensity of the plaque decreased to a value similar to that observed for the plaque immediately after injection. Importantly, the enhancement was also related to plaque composition: the more cellular the content was (as judged by histology), the more intense was the signal. At 24 h post‐injection, when MRI of the rHDL‐like contrast agent showed its maximum intensity in plaques, aortas were removed and imaged by confocal fluorescence microscopy. The fluorescence was located mainly in the intimal layer, where the lipids accumulate and a small number of cells had internalized and retained the fluorescence. Histology characterization identified these cells primarily as macrophages.

Figure 5. An application of a high‐density lipoprotein (HDL)‐like nanoparticle contrast agent for macrophage‐targeted in vivo magnetic resonance imaging (MRI). Phospholipid‐based contrast agent Gd‐DTPA‐DMPE and NBD‐DPPE‐labelled phospholipid for fluorescence confocal microscopy were used for the reconstruction of HDL. The MR images of an apolipoprotein E knockout mouse at 1.5 T illustrate the pre‐ and 24 h post‐injection of contrast agent at a rHDL dose of 4.36 μmol/kg. White arrows point to the abdominal aorta; the insets denote a magnification of the aorta region. (Modified from Frias et al. 2004 52 with permission.)

Fibrin—various approaches

Plaque rupture with subsequent thrombosis is a common underlying pathophysiology in ACS and stroke. Direct in vivo visualization of thrombi would thus be beneficial both for diagnosis and follow‐up of therapy. Therefore, there has recently been an active search for fibrin‐targeted MRI contrast agents as an aid for vulnerable plaque detection. The successful approaches include fibrin‐targeted paramagnetic perfluorocarbon nanoparticles 53 as well as fibrin‐binding gadolinium‐labelled peptides EP‐1873 54 and EP‐1242 55. In recent studies, a remarkable progression in the detection of the components of ruptured or highly vulnerable plaque has been achieved. It has been demonstrated that imaging with a Gd‐based fibrin‐binding contrast agent, EP‐2104R, can specifically detect thrombotic material with platelets and fibrin 56,57. In the aforementioned studies tissue contrast on images has been promising, suggesting the method to be sensitive enough to detect thrombus also in clinical practice. Imaging with EP‐2104R may provide new information on the aetiology and pathogenesis of recent thromboembolic events but its predictive value in case of thrombotic events still remains to be clarified. EP‐2104R is currently being tested in human subjects in phase I and phase II clinical trials 57. In addition, quantitative ‘magnetic resonance immunohistochemistry’ with the fibrin‐targeted perfluorocarbon nanoparticles has recently been illustrated in the case of an ex vivo human carotid endarterectomy sample (Figure 6), showing the potential use of ¹⁹F imaging and spectroscopy for identifying unstable atherosclerotic lesions which exhibit microdeposits of fibrin indicative of the possibility for a future rupture 58.

Figure 6. Illustration of quantitative‘magnetic resonance immunohistochemistry’ with the fibrin‐targeted perfluorocarbon nanoparticles in the case of an ex vivo human carotid endarterectomy sample: (a) An optical image of a 5‐mm cross‐section of a human carotid endarterectomy sample. This section showed moderate luminal narrowing as well as several atherosclerotic lesions. (b) A ¹⁹F projection image acquired at 4.7 T through the entire carotid artery sample shows high signal along the lumen due to nanoparticles bound to fibrin. (c) A concentration map of bound nanoparticles in the carotid sample. (With permission from Morawski et al. 2004 58.)

Miscellaneous

The new field of molecular targeted MRI is in an active phase of continuous development, and the role of nanotechnology in detecting the various phases of atherothrombosis is rapidly expanding 59. In addition to the abovementioned applications there are several emerging techniques that are also likely to find their way to the studies of atherothrombosis. These include in vivo MRI visualization of gene expression and detection of stem cells 60. Contrast agents targeted to endothelial cell adhesion molecules as well as to enzymatic and proteolytic processes in the extracellular matrix are also likely to benefit atherosclerosis research 1. Moreover, ‘smart’ agents (that undergo a physicochemical change due to their intended molecular interaction with the target) and the holistic antibody‐mediated targeting of a certain disease stage‐related characteristic vascular proteins (applying the phage display method) are fascinating options likely to induce the development of molecular specific ways for the imaging of atherothrombotic processes 1,60. It is also currently envisioned that the site‐directed delivery of therapeutic substances, concomitantly with the detection of early atherosclerotic complications, may well have a substantial impact on clinical cardiology in the near future 59.

Magnetic resonance spectroscopy

Lipoprotein subclasses by ¹H MRS of plasma

In the mid‐1990s two research groups independently developed ¹H MRS‐based lipoprotein lipid quantification methodology 61,62, and later on several groups have been assessing this option 63. One of the original approaches 61 was developed for a commercial assay, named NMR LipoProfile® by LipoScience Inc. According to the company it has performed over one million NMR LipoProfile® tests (http://www.liposcience.com). In contrast to other existing methods of lipoprotein subclass quantification, involving both physical separations of the subclasses (e.g., ultracentrifugation) and separate determinations of subclass specific lipid values (e.g., enzymatic measurements of cholesterol), the ¹H MRS method relies on a single spectrum of a plasma sample and subsequent mathematical deconvolution of the subclass specific information. Recent clinical applications of the MRS‐based methodology have evidently shown its usefulness in biomedical research 5.

Is ¹H MRS of serum able to diagnose coronary heart disease?

It has been suggested that pattern recognition techniques applied to ¹H MR spectra of human serum could diagnose not only the presence, but also the severity, of coronary heart disease (CHD) 64. Results were reported that more than 90% of subjects with stenosis of all three major coronary vessels were distinguished from subjects with angiographically normal coronary arteries, with specificity greater than 90%. According to the current understanding of the complexity of atherothrombosis and the array of clinical data on lipoprotein lipids and their relation to CHD, the above indication is quite unexpected. Particularly, since conventional ¹H MRS characteristics were used resulting in spectra that many independent research groups have shown to reflect mostly variations in the concentrations of lipoprotein subclasses, the variations of which are not statistically distinctive for diagnostic purposes 15. Also the study group of patients was very limited and not indicative of clinical patient material which is likely to explain the fortuitous conclusion concerning ¹H MRS‐based CHD diagnosis 64. (A note added in proof: A report to support these conclusions has recently been appeared stating that ¹H MRS analysis of plasma is only a weak predictor of CAD (Kirschenlohr HL, Griffin JL, Clarke SC, Rhydwen R, Grace AA, Schofield PM, et al. Proton NMR analysis of plasma is a weak predictor of coronary artery disease. Nat Med. 2006;12:705–10.).)

On the other hand, ¹H MRS of plasma has been shown to detect a proatherogenic state in healthy, non‐diabetic subjects who subsequently develop diabetes 5. Furthermore, it has recently been indicated that an overall measure combining information from seven lipids (total, HDL, LDL and VLDL cholesterol, triglycerides, apolipoprotein A‐I, and apolipoprotein B) is a superior predictor of CHD risk compared with conventional lipid measures 65—since the ¹H MR spectrum of plasma is currently known to contain not only these data but specific information on lipoprotein subclasses and on an array of other metabolites 15—it would be unexpected if the total ¹H MR spectrum of plasma would not lead to good estimation and classification of atherosclerosis risks. Thus, as also suggested in Figure 1, the potential usage of in vitro¹H MRS is in the risk evaluation of atherothrombotic events, the true clinical value of which also remains to be elucidated in extensive clinical studies.

Molecular characterization of plaque—¹H and ¹³C MRS

The molecular characteristics of various atherosclerotic plaques have been studied using ¹H and ¹³C MR spectroscopy ex vivo66,67. Interestingly, ¹³C MRS of human arterial plaques revealed the relation of decreased polyunsaturation for increased stenosis 68. In the light of recent evidence, that apoptosis would be an indicator of atherothrombotic processes 69 and that in vivo¹H MRS can detect apoptosis‐related accumulation of polyunsaturated fatty acids 70, MRS studies might be useful in figuring out the relation of apoptosis to the development of atherothrombosis and to plaque vulnerability in vivo. In relation to animal models of atherosclerosis it is good to note that ¹H MRS has shown clear molecular differences between rabbit and human plaques 71.

The lipid components of atherosclerotic plaques are abundant in liquid‐like cholesterol esters (CEs) derived largely from apolipoprotein B‐100 containing lipoproteins 72. Exploiting this, a study demonstrating the usefulness of MR image‐guided, single‐voxel ex vivo¹H MRS to characterize lipid‐rich and lipid‐poor regions of atherosclerotic plaque (see Figure 7) has just appeared 73. This procedure may show the way for non‐invasive detection and quantification of CEs in atherosclerotic plaques in vivo since preliminary work at 3.0 T on human carotid artery atherosclerosis has given promising results 73. Another recently established methodology, based on in vivo¹H MRS with cardiac and respiratory gating, for myocardial lipid quantification in humans, supports the future rise of cardiac applications of in vivo¹H MRS 74. It is also anticipated that in addition to traditional individual signal quantifications, a holistic approach of MR metabonomics will find its way to in vivo MRS.

Figure 7. Demonstration of the magnetic resonance(MR) image‐guided, single‐voxel ex vivo¹H magnetic resonance spectroscopy (MRS) to characterize lipid‐rich and lipid‐poor regions of atherosclerotic plaque. Spectra elicited from 1‐mm³ voxels are illustrated with insets showing the parametric map and voxel position for reference. Panels A and C illustrate the intense water signal without water suppression. Weak lipid signals are evident in panel B only after vertical expansion (inset). In contrast, strong lipid signals (consistent with cholesterol esters) are seen in panel D. (With permission from Ruberg et al. 2006 73.)

Combined MR schemes for individual diagnostics—a future strategy?

It would be advantageous to detect molecular and cellular processes related to the early stages of developing atherosclerosis. This would clinically facilitate early individual primary prevention and also give a personal rationale to comply with life‐style modifications and potential drug therapies. The recent applications of MR in atherothrombosis research, as indicated in this review, suggest the potential usage of in vitro MRS for risk assessment and of in vivo MRI for direct detection of plaque composition and vulnerability. The scheme presented in Figure 1 can be seen as one option to elucidate the potential of MR in detecting individual intermediate atherothrombotic end‐points and utilizing their prognostic value before the occurrence of a definite end‐point. Molecular targeted MRI also shows notable possibilities for the non‐invasive detection of early atherothrombotic processes at various pre‐clinical stages of atherosclerosis. The field of MR technologies, in combination with nanotechnology and bioinformatics, is progressing rapidly. Consequently, it is tempting to envision that combinatorial MRI and MRS schemes could be operational at the clinic in the near future.

Acknowledgements

Dr Juhana Hakumäki is acknowledged for fruitful discussions about molecular imaging and MRI technology. This study was supported by the Academy of Finland.