European Journal of Nuclear Medicine and Molecular Imaging

, Volume 33, Issue 2, pp 111–116

Identification of interleukin-2 for imaging atherosclerotic inflammation


    • Imaging Science Laboratories—Department of Radiology and The Zena and Michael A. Wiener Cardiovascular Institute, Marie-Josée and Henry R. Kravis Center for Cardiovascular HealthMount Sinai School of Medicine
  • Vardan Amirbekian
    • Johns Hopkins University School of Medicine
    • Sarnoff FellowThe Sarnoff Endowment for Cardiovascular Science
  • Jean-François Toussaint
    • Département de Physiologie et Radioisotopes Hopital Européen Georges Pompidou (2ème C)
  • Valentin Fuster
    • The Zena and Michael A. Wiener Cardiovascular Institute—Marie-Josée and Henry R. Kravis Center for Cardiovascular HealthMount Sinai School of Medicine
Editorial Commentary

DOI: 10.1007/s00259-005-1981-y

Cite this article as:
Fayad, Z.A., Amirbekian, V., Toussaint, J. et al. Eur J Nucl Med Mol Imaging (2006) 33: 111. doi:10.1007/s00259-005-1981-y

Current perspective

It is predicted that cardiovascular diseases are going to become the main cause of death globally within the next 15 years [1]. Atherosclerosis is a systemic disease that, in Western societies, accounts for roughly half of all deaths [2]. For example, in the United States close to 5% of the population is afflicted with symptomatic or diagnosed coronary artery disease [3]. These patients represent a mere fraction of all those harboring atherosclerotic pathology. In fact, more than half of patients affected by coronary atherosclerosis experience sudden death or myocardial infarction as their first clinical manifestation of disease [3, 4]. Plaque rupture is thought to be the culprit in the majority of detrimental events [5]. There is significant evidence that the risk of plaque rupture is associated with plaque burden and plaque composition [68].

A major goal of atherosclerosis imaging is to attain the ability to accurately identify high-risk atherosclerotic plaques, so called vulnerable plaques [9]. Molecular imaging is an emerging methodology that will help solve this problem. Molecular imaging has the potential to give clinicians unique and functional pathophysiological information about disease processes such as atherosclerosis. It is likely that this information will allow clinicians to prognosticate which patients are vulnerable to vascular events. Furthermore, functional information gained about individual patients will provide guidance regarding therapeutic choices. The sum of information attained will give clinicians “intelligence” about pathology that can then be translated into aggressive management and prevention strategies that will save lives and decrease healthcare costs while ushering in the era of individualized medicine.

Molecular imaging of atherosclerosis is still in its early stages; however, important landmarks have already been reached. In a short period of time, imaging of oxidized low-density lipoprotein (ox-LDL), high-density lipoprotein (HDL), smooth muscle cell proliferation, neovascularization, biodynamics, proto-oncogenes, macrophages, inflammation, and apoptotic cells has been obtained using single-photon emission computed tomography (SPECT), positron emission tomography (PET), intravascular ultrasound (IVUS), or magnetic resonance imaging (MRI). Molecular imaging of atherosclerosis has two broad goals: the first is effective and accurate clinical application coupled with aggressive therapeutic and preventative strategies and the second is elucidation of pathophysiological pathways of atherosclerosis primarily in animal models of disease. A full discussion of molecular imaging of atherosclerosis is beyond the scope of this commentary; however, below we present some landmarks or promising developments in the field of atherosclerosis molecular imaging.

Molecular imaging of atherosclerosis

Nuclear imaging

Tsimikas et al. have used radiotracer-labeled monoclonal antibodies (murine monoclonal antibody, MDA2) to image ox-LDL. These studies have been performed in mouse and rabbit models of atherosclerosis. It was demonstrated that radiolabeled antibodies that indirectly image ox-LDL could be used to detect atherosclerosis [10]. They also showed that it is potentially possible to estimate plaque volume as well as to follow the progression and regression of atheromatous plaques [11, 12]. Most recently, Torzewski et al. performed a study to validate the ability of 125I-radiolabeled oxidation-specific antibodies to follow progression and regression of plaques [13]. However, due to the low target-to-background ratios and the difficult production of these radiotracers, this work has not yet been translated into clinical application. Another barrier is the well-known limited resolution of current nuclear techniques. Resolution will be particularly important for imaging atherosclerosis in arteries such as the coronaries.

Macrophages and monocytes represent good cellular targets for molecular imaging of atherosclerosis because they have been shown to play a critical role in the inflammation thought to be central to atherosclerosis [1417]. Macrophages have been targeted by Ohtsuki et al. using 125I-radiolabeled monocyte chemoattractant peptide-1 (MCP-1) in rabbits [18]. A strong relation was shown between radiotracer uptake and atherosclerotic plaque macrophage concentration [18].

There are data indicating that apoptosis of macrophages contributes to the size of a necrotic core, which is believed to contribute to instability [19]. There is also evidence that apoptosis of smooth muscle cells is associated with a thin fibrous cap [20]. Kolodgie et al. used 99mTc-radiolabeled annexin V to target apoptotic cells in a rabbit model of atherosclerosis. There was tenfold higher uptake in the aorta of atherosclerotic rabbits compared with control rabbits [21]. Quite interestingly, these methods are now being used in humans to study usefulness in assessment of carotid atherosclerosis. Although preliminary, the results are suggesting that imaging carotid atherosclerosis with radiolabeled annexin V may discern carotid plaque features indicative of instability [22]. Though it is noteworthy that this is now being studied clinically in humans, a lot more work is necessary before a conclusion can be drawn about the potential clinical impact. In particular, long-term randomized controlled studies are necessary to evaluate clinical utility.

Smooth muscle cells found in atherosclerosis were targeted by Carrió et al. using 111In-labeled negative charge-modified Z2D3 F(ab’)2 specific for an antigen expressed by proliferating smooth muscle cells [23]. After testing this approach in rabbits, the investigators went on to perform a study using SPECT in patients scheduled to undergo carotid endarterectomy. The radiotracer-labeled antibody uptake corresponded well with the angiographic location of the carotid plaques. The study demonstrated the feasibility of use in humans, but further clinical investigation has not been forthcoming.

An important pathologic process in the development of atherosclerosis is angiogenesis and neovascularization of plaques [2426]. Matter et al. imaged atherosclerosis by indirectly targeting angiogenesis with an antibody (L19) directed against extra-domain B of fibronectin, which is expressed during neovascularization but is not present in normal tissues [27].

PET has shown potential for imaging atherosclerosis. With newer PET/CT technology, especially Multi-Detector CT (MDCT), information on vascular stenoses and calcifications may be obtained simultaneously with PET [28]. Recently, Ogawa et al. performed a PET/CT study correlating atherosclerosis pathology to 18F-fluorodeoxyglucose (FDG) uptake in the aorta of atherosclerotic rabbits. They found a strong correlation between FDG uptake and the number of macrophages found in plaque [29]. Tawakol et al. confirmed these findings, showing a strong relationship between macrophage staining on immunohistopathologic slides and the amount of FDG uptake in plaques of atherosclerotic rabbits [30].

Several studies in humans have examined FDG uptake in the region of the aorta [3133]. However, these studies yielded little information regarding frequency, location, and intensity of uptake. Rudd et al. used FDG to perform PET on patients with symptomatic carotid atherosclerotic disease [34]. Their study found that FDG accumulated to a significantly greater degree in unstable plaques compared with the stable contralateral-sided plaques. Pathological analysis implied FDG accumulation in macrophages. In a different study, using PET/CT to image human aortic atherosclerosis, it was found that FDG uptake was associated with age and it was mostly distinct and separate from areas of calcification in the aorta [35]. Current vascular-related clinical applications of FDG PET are to assess inflammatory vascular diseases such Takayasu’s arteritis and giant cell arteritis [36, 37]. It has been shown that steroid therapy reduced the uptake of FDG in one of these studies [37], suggesting that FDG PET may then be used to assess anti-inflammatory therapeutic response.

The sum of evidence implies that PET using FDG may be suitable for assessing plaque vulnerability [16]. There are now a growing number of investigations in humans, although, for now, with small numbers of patients. Further study is necessary to examine whether FDG uptake correlates with future risk of atherosclerosis-related clinical events or plaque rupture. True sensitivity and specificity also need to be studied. However, there remain significant challenges to the use of FDG in assessing atherosclerosis. Unfortunately, FDG is nonspecific and is taken up into any metabolically active tissue. In fact, of all tissues, myocardium shows the highest uptake of FDG, which currently excludes FDG from use in imaging coronary atherosclerosis.

Molecular MRI and other techniques

Nuclear methods were first to be applied to the molecular/functional approach to imaging. Over time, however, it has become increasingly clear that multimodality techniques may work best. We think that no one imaging technology is going to dominate molecular and functional imaging. The diversity and variability of pathologic processes will require the strengths of different, and likely multimodal, imaging technologies to yield the best sensitivities and specificities.

MRI has emerged as an important modality for molecular imaging. MRI offers powerful anatomical resolution that may become vital for imaging small structures such as the coronary arteries. The idea of molecular MRI is based on delivering MRI contrast agents to locations of interest using molecular targeting techniques [38].

A unique and potentially potent molecular MRI contrast agent for atherosclerosis is gadolinium-loaded recombinant high-density lipoprotein (Gd-rHDL). This agent takes advantage of HDL’s physiologic role in migrating in and out of atherosclerosis to provide contrast enhancement on MRI. Frias et al. tested Gd-rHDL in vivo in apolipoprotein E (Apo-E) knockout mice and demonstrated a significant enhancement of atherosclerotic plaque at 24 h after injection of Gd-rHDL [39]. Using confocal microscopy it was shown that fluorescently labeled Gd-rHDL accumulated in the macrophages of atheromatous plaques [39]. Because of its efficacy and endogenous nature, rHDL is likely to be studied in humans in the very near future.

αvβ3 belongs to the family of integrins and is strongly associated with angiogenesis and neovascularization [4043]. Both neovascularization and αvβ3 have been targeted for molecular MRI of atherosclerosis [4446]. In a rabbit model, Winter et al. showed that regions of neovascularization had a good increase in signal intensity after administration of a αvβ3-targeted nanoparticle MRI contrast agent [44]. More investigation is necessary to move this contrast agent further toward possible clinical testing.

Gadofluorine-M is a possible molecular MRI contrast agent for atherosclerosis. It is a lipophobic gadolinium-based agent with unique properties that provide for either accumulation or retention in plaque. In a study by Sirol et al., there was a 1.5-fold increase in signal intensity at 1 h and a twofold increase at 24 h after injection of the contrast agent Gadofluorine-M in a rabbit model of atherosclerosis [47].

Thrombus is associated with complex atherothrombotic plaques and there is evidence that it may contribute to progression of lesions [48]. Sirol et al. performed a study in guinea pigs using a fibrin-specific MRI contrast agent. They found that after injection of the fibrin-specific contrast agent there was a 400% increase in the signal intensity of thrombus and they were able to detect 100% of all thrombi compared with only 42% prior to injection [49].

MRI investigators have also targeted macrophages found in atherosclerosis with iron particles such as ultra-small superparamagnetic particles of iron oxide (USPIOs) [5053]. Kooi et al. performed an investigation using USPIOs on symptomatic patients who then underwent endarterectomy. They showed that in areas of interest, where MRI changes were seen, there was a 24% decrease in the signal intensity on T2*. They also demonstrated, via pathology, that 75% of ruptured or rupture-prone plaques exhibited uptake of USPIOs whereas only one lesion representing 7% of stable lesions exhibited USPIO uptake [53]. Recently, Kelly et al. used a peptide conjugated to “magnetofluorescent” CLIO (an iron oxide-based agent) particles to image vascular cell adhesion molecule-1 (VCAM-1) on endothelial cells [54]. The study showed that this contrast agent could be used to image VCAM-1 overexpression on endothelial cells in Apo-E-deficient mice [54]. This approach is in line with attempts to image inflammation, which is believed to be a central process in atherosclerosis.

Intravascular ultrasound (IVUS) has also been used for atherosclerosis assessment. Hamilton et al. used echogenic microbubbles attached to antibodies to image a myriad of molecular targets in pigs. Specifically, they used IVUS to molecularly image tissue factor (TF), intercellular adhesion molecule-1 (ICAM-1), VCAM-1, fibrin, and fibrinogen [55]. It was shown that these targeted microbubbles provided significant contrast enhancement of plaque compared with untargeted microbubbles and normal saline [55]. However, IVUS is a fairly invasive technique and will probably play more of an adjunctive role in patients undergoing coronary interventions for primary indications. For broad clinical applications, such as identification of vulnerable patients, non-invasive techniques are more appealing.

Imaging with interleukin-2

Interleukin-2 (IL-2) was discovered in 1975 and was found to be a growth-promoting factor of bone marrow-derived T lymphocytes [56]. The discovery of IL-2 made it possible to generate, culture and study T lymphocytes. IL-2 has an array of pleiotropic effects on a wide variety of cells such as CD4+ T cells, CD8+ T cells, B cells, and natural killer cells [57]. Most prominently, it is thought to be important in T cell activation and inflammatory consequences thereof.

In this issue, Annovazzi et al. present the use of 99mTc-radiolabeled interleukin-2 (99mTc-IL-2) for carotid atherosclerosis imaging in humans. The concept behind the study is to use 99mTc-IL-2 in order to visualize T lymphocytes that may be found in inflammatory atherosclerotic plaques. They demonstrate a relationship between 99mTc-IL-2 uptake and immunohistopathology of carotid endarterectomy specimens obtained from the study patients. The study also examines the response of 99mTc-IL-2 uptake to atorvastatin [a 3-hydroxy-3-methylglutaryl coenzyme A (HMG-CoA) reductase inhibitor (statin)] and hypocholesterolemic diet. The results of the investigation show a significant correlation between the target-to-background (T/B) ratio of 99mTc-IL-2 uptake and the percent of IL-2 receptor (IL-2R)-positive cells on flow cytometry and standard histology as well as the absolute number of IL-2R-positive cells on flow cytometry. These results are exciting although current spatial resolution remains limited. Future nuclear functional imaging must either be improved significantly or it must be combined with high-resolution techniques such as MRI or MDCT. It is interesting what newly proposed MDCT-SPECT will bring to the field. Given the limitations on spatial resolution, 99mTc-IL-2 uptake due to perivascular or lymphatic inflammation may potentially be a problem if broader clinical studies are pursued.

B-mode ultrasonography was used during the study as a tool for clinical assessment of the carotid plaques. There were no significant correlations between observed 99mTc-IL-2 T/B ratios and plaque morphology and luminal anatomy as assessed by B-mode ultrasound. Furthermore, there was no relationship between ultrasound findings and percentage of IL-2R-positive cells. These findings are not at all surprising. There are considerable data indicating that ultrasound per se cannot precisely characterize carotid plaque especially in small numbers of patients. Ultrasound is useful in screening for significant stenoses due to carotid plaques. Significant stenosis of the carotid is associated with a high risk of stroke.

In the longitudinal portion of the study, Annovazzi et al. found that atorvastatin decreased 99mTc-IL-2 uptake by 19% as measured by the T/B ratios. This finding naturally leads to further questions. First, what does a reduction in the T/B ratio mean clinically? Does it mean there is less inflammation? Does it mean there is less plaque? Does it mean that these patients have reduced their risk of stroke or other adverse events? Of course, we do not yet know the answers, but these questions fundamentally ask: “How will the findings of Annovazzi et al. translate into the clinical world?” The short response is that much more work is necessary before we can even begin to know the answers. It should be noted that, in previous work by other groups, statins have been shown to inhibit IL-2 release and T lymphocyte activation [58]. In fact, it was shown that lowering LDL with fibrates also had a similar effect. However, the effect of statins on IL-2 was beyond comparable LDL reductions achieved with fibrates. This suggests that statins may actually have biological activity beyond LDL reduction.

As part of the longitudinal portion of their study, the authors also present data implying that hypocholesterolemic diet did not change carotid 99mTc-IL-2 uptake. However, the conclusions reached about diet are somewhat debatable. There were only five (n=5) patients who were in the diet group at the outset of the study. Two patients out of the five dropped out because of non-compliance with the diet (in reality leaving n=3). In the data presented, two of the five diet patients did have decreases in 99mTc-IL-2 uptake comparable with those in patients in the atorvastatin group. Furthermore, the period of diet was only 3 months, which is likely too short to produce effects. Investigations of diet in larger numbers of patients, with high compliance, over longer periods of time are needed before conclusions can be reached about the ability of diet to modify carotid 99mTc-IL-2 uptake.

Summary and conclusions

Although radiolabeled IL-2 has been used in humans to image inflammatory conditions of the gastrointestinal system [59, 60], this is the first time that it has been used to image inflammation associated with carotid atherosclerosis. Whether IL-2 is the best cytokine for targeted imaging of atherosclerosis remains to be seen. Nevertheless, this study points out to us the potential of using endogenous cytokines to assess inflammation as well as to monitor response to treatment.

It is exciting to see the various approaches researchers are taking to develop techniques for molecular imaging of atherosclerosis. As techniques with the best potential emerge, investigators must develop and move these techniques forward toward broader long-term clinical investigations in humans. Only then can we begin to evaluate true clinical utility as well as translate this utility into personalized medicine.


Funding for this research was provided in part through grants from the NIH/NHLBI R01 HL071021 and R01 HL078667 to ZAF and the Stanley J. Sarnoff Endowment for Cardiovascular Research, Inc to VA.

Copyright information

© Springer-Verlag 2005