The prevalence of obesity, metabolic syndrome, and diabetes is rapidly increasing in Western populations along with their secondary complications, particularly cardiovascular disease. Unfortunately, the acceleration of atherosclerosis is generally silent and difficult to monitor until the manifestations become symptomatic. Serum biomarkers, such as low-density lipoprotein or high-sensitivity C-reactive protein, offer nonspecific but possibly predictive indications of cardiovascular risk, but neither specifically quantifies the temporospatial microscopic and biochemical changes occurring in the atherosclerotic vessel wall (1- 3). As an alternative to monitoring systemic indexes, molecular imaging may allow noninvasive assessment of disease biomarkers directly within the involved tissues. For example, plaque angiogenesis is correlated with plaque rupture and often occurs alongside other morphological features of vulnerable lesions including macrophage infiltration, lipid-rich cores, and thin-cap shoulders (4- 8). Therefore, plaque neovasculature could be indicative of intramural atherosclerotic severity and overall cardiovascular disease risk.

Our laboratory developed an α_νβ₃-targeted paramagnetic nanoparticle agent (∼250 nm diameter) for noninvasive molecular imaging of angiogenesis in animal models of cancer and cardiovascular disease as well as for monitoring response to targeted antiangiogenic therapy (9- 12). The overall aim of this study was to use α_νβ₃-integrin paramagnetic nanoparticles for monitoring the temporal progression of aortic angiogenesis in a rat model of metabolic syndrome and for evaluating the effects of a known cardioprotective therapy, benfluorex, on adventitial neovascularization.

The James C. Russell:LA corpulent rat strain (JCR:LA-cp) is a unique model of metabolic syndrome characterized by spontaneous atherosclerosis and ischemic myocardial lesions due to obesity, hyperlipidemia, and insulin resistance (13- 16). This disease model arises from a genetic mutation in the leptin receptor, commonly referred to as the corpulent gene (cp) (15). Animals that are homozygous for the cp gene are denoted as JCR:LA-cp cp/cp obese, whereas animals that are heterozygous (cp/+) or homozygous for the functional leptin receptor gene (+/+) are indistinguishable and denoted as JCR:LA-cp +/? lean. The obese animals display a phenotype consistent with metabolic syndrome and the lean animals serve as normal controls. The prevalence of plaque angiogenesis as a vascular biomarker of atherosclerosis in metabolic syndrome is poorly understood and undocumented in the JCR:LA-cp rat model. Previous studies demonstrated that benfluorex, an anorectic and hypolipidemic drug, decreases insulin resistance, normalizes the lipid profile, and diminishes aortic plaque in JCR:LA-cp rats (17- 19).

Six-week old male JCR:LA-cp (+/?, lean rats, n = 5) and JCR:LA-cp (cp/cp, obese rats, n = 11) were studied. Six obese rats received dietary benfluorex (Sigma-Aldrich, St. Louis, Missouri) ad libitum (0.069% wt/wt; Purina Mills, St. Louis, Missouri) for 16 weeks, and the remaining obese (n = 5) and lean (n = 5) rats were fed a standard rodent diet (Purina Mills). Body weight and food consumption were recorded at baseline and weeks 4, 8, 12, and 16 of the study. Magnetic resonance (MR) molecular imaging with α_νβ₃-targeted paramagnetic nanoparticles (1.0 ml/kg, ∼0.005 mmol Gd³⁺/kg, ∼200 pmol nanoparticles/kg) was performed at baseline and at weeks 8 and 16. The study protocol was approved by the Animal Studies Committee of Washington University Medical School.

The nanoparticle contrast agent was prepared for MR experiments as described in previous reports (9- 12). Nanoparticle emulsions were composed of perfluorocarbon (20%, vol/vol) and excipient (80%, vol/vol). Each nanoparticle contained 98.0% perfluorooctyl bromide (Minnesota Manufacturing and Mining, St. Paul, Minnesota) with 2.0% (wt/vol) of a lipid surfactant comixture in a 1.7% (wt/vol) glycerin in water excipient. The surfactant comixture included 68 mole% lecithin (Avanti Polar Lipids, Inc., Alabaster, Alabama), 0.1 mole% peptidomimetic α_νβ₃-integrin antagonist conjugated to polyethylene glycol (molecular weight of 2,000) phosphatidylethanolamine (Avanti Polar Lipids, Inc.), 1.9 mole% phosphatidylethanolamine (Avanti Polar Lipids, Inc.), and 30 mole% gadolinium diethylenetriamine-pentaacetic acid-bis-oleate (IQsynthesis, St. Louis, Missouri).

Homing to angiogenic vessels was achieved with an α_νβ₃-integrin antagonist, originally developed for nuclear and optical imaging (20- 21). The specificity of the α_νβ₃-ligand mirrored that of LM609 as assessed by staining and flow cytometry and had a 15-fold preference for the Mn²⁺ activated receptor (21 nmol/l) (21). The half maximal inhibitory concentration for α_νβ₅, α₅β₁ and glycoprotein IIb/IIIa was determined to be >10 μM. Integrin-targeted nanoparticles presented approximately 300 ligands per particle with a half maximal inhibitory concentration of 50 pM for the Mn²⁺ activated α_νβ₃-integrin (12).

MR molecular imaging of rat abdominal aorta was conducted under anesthesia (1% to 2% isoflurane in 2 l/min oxygen) using a 3.0-T clinical magnet (Achieva; Philips Healthcare, Andover, Massachusetts); an 11-cm, 2-element phased array coil; and a T1-weighted, fat-suppressed, black-blood, multislice turbo spin echo imaging sequence (TR/TE = 296/10 ms, resolution = 213 μm × 213 μm × 2 mm, 4 slices, ∼11.6 min). Four contiguous slices were positioned immediately superior to the renal arteries, which provided an anatomical landmark for reproducible MR slice alignment over the course of the study. Saturation bands were placed proximal and distal to the region of image acquisition to null the blood signal, and fat suppression was achieved with spectral selection attenuated inversion recovery. MR images were acquired before and 120 min after tail vein injection of α_νβ₃-targeted paramagnetic nanoparticles at weeks 0 and 8. At week 16, aortic images were acquired before nanoparticle injection and dynamically at 20, 60, and 120 min post-treatment. All aortic images were normalized to a reference gadolinium standard placed within the field of view. In each transaxial image, the aortic wall was manually traced and the average signal intensity across all slices in each animal was calculated for each time point. MR signal enhancement on each day of the study was calculated for each animal relative to the aortic signal intensity before α_νβ₃-targeted paramagnetic nanoparticle injection.

At weeks 0, 8, and 16 of the study, blood samples were obtained and analyzed for fasted plasma triglyceride, cholesterol, and glucose using routine clinical chemistry procedures by the Washington University Department of Comparative Medicine. Plasma insulin and leptin levels were assayed by Millipore Bioscience Division (St. Charles, Missouri) at weeks 8 and 16.

After the last imaging time point, abdominal aortas were resected for histology. Formalin-fixed and paraffin-embedded (5 μm) sections were stained for von Willebrand factor, an endothelial cell marker (Chemicon International, Inc., Temecula, California). Four equally spaced histological sections within the imaged segment were examined for each animal. For each digitized section, adventitial microvessels, consisting of no more than 3 endothelial cells in circumference, were counted and divided by the adventitial area, which was traced and calculated with MATLAB 7.0 (The MathWorks, Inc., Natick, Massachusetts) to estimate microvessel density. Microvessel density (vessels/100 μm²) was averaged across the 4 microscopic sections to provide a single estimate for each rat.

All quantitative data are reported as mean ± SE. Two-way analysis of variance was performed to compare the experimental groups and the repeated measures (SAS Institute, Cary, North Carolina). For significant F tests (p < 0.05), group means were separated using the least significant difference technique.

At baseline, body weight and food consumption in the 6-week-old obese rat groups were 46.5% and 49.2% higher, respectively, than the lean animals (Figure 1). During the 16 weeks of the study, all treatment groups increased body weight and food consumption. The appetite suppression effect of benfluorex resulted in decreased (p < 0.05) weekly food consumption by obese rats relative to the untreated corpulent animals that was similar (p > 0.05) to the lean control intake. Although obese rats receiving benfluorex weighed less (517 ± 9 g, p < 0.05) than the untreated obese control animals (615 ± 18 g) at 16 weeks, these animals were heavier (p < 0.05) than the lean control rats (367 ± 18 g) despite equivalent dietary intakes. From the average daily food consumption, the amount of benfluorex incorporated into the feed and the average weights of the animals, the average drug dose over the course of this study was calculated as 31.8 ± 0.5 mg/kg/day.

The fat suppressed MR images of the heavier obese rats, both benfluorex-treated and untreated, revealed large, hypointense, periaortic adipose regions within the abdominal cavity that were not observed in lean animals (Figure 2). At 8 and 16 weeks of the study, marked neovascular signal enhancement was measured throughout the aortic slices of the untreated obese rats in contradistinction to the sparse background contrast level detected in the lean animals at 2 h post-injection. Benfluorex treatment reduced the extent and intensity of neovascular signal enhancement observed in the aortic wall of obese rats to a magnitude similar to that found in the lean animals, which paralleled the primary drug effect on feed consumption (Figure 3). Dynamic images obtained over 120 min after injection of α_νβ₃-targeted paramagnetic nanoparticles at week 16 clearly revealed a marked increase in neovascular contrast at 20 min for the untreated obese rats, whereas signal enhancement in both the lean and benfluorex-treated obese rats was much lower. These differences in contrast enhancement remained unchanged over the remaining time of scanning (Figure 4).

Fasting plasma glucose did not differ among the 3 treatment groups over the course of the study (Figure 5). As expected, cholesterol and triglyceride concentrations were lower (p < 0.05) in lean rats compared with the obese animals, but benfluorex treatment did not decrease (p > 0.05) plasma lipids significantly in the obese rats (Figure 5). Fasting hormonal concentrations of insulin and leptin were markedly higher (p < 0.05) in the obese versus lean animals at 8 and 16 weeks (Figure 5), similar to plasma lipid levels. However, benfluorex lowered (p < 0.05) insulin and leptin levels in the obese rats, but not to the level in lean animals.

Routine light microscopy of the aortas from obese rats showed increased adventitial adipose deposits compared with the lean animals, which paralleled the MR imaging results (Figure 6). No protruding aortic plaque was appreciated in the obese animals at 22 weeks of age, but subintimal plaque thickening was prevalent in the untreated obese rats and rarely appreciated in the lean control or obese rats fed benfluorex.

Anti–von Willebrand factor staining of endothelial cells revealed a greatly expanded vasa vasorum within the fatty adventitia of the abdominal aorta in untreated obese rats relative to the lean animals. The benfluorex-treated animals displayed only patchy areas of neovascular expansion within the adventitia, involving much less of the vessel wall compared with the untreated obese animals. Adventitial microvessel density in the untreated obese rats (5 ± 0.7 vessel/100 μm²) was greatly increased (p < 0.05) relative to benfluorex-treated obese rats (2.6 ± 0.3 vessel/100 μm²) and the lean control groups (2.7 ± 0.4 vessel/100 μm²), which were equivalent (Figure 7).

Metabolic syndrome is characterized by a clustering of metabolic abnormalities in conjunction with underlying insulin resistance and is closely associated with the onset of diabetes and accelerated cardiovascular disease. The corpulent JCR:LA-cp rat model of metabolic syndrome exhibited obesity, severe insulin resistance, hyperinsulinemia, and hypertriglyceridemia, as reported by others (13- 16). Vasa vasorum expansion and neovasculature proliferation, previously unreported in this model, was demonstrated and longitudinally assessed with MR molecular imaging at 3.0-T with α_νβ₃-targeted paramagnetic nanoparticles, whereas benfluorex-treated and lean rats had significantly lower MR signal enhancement throughout the study. The MR enhancement observed in the aortic wall of the lean animals was perhaps a result of circulating nanoparticles. In an atherosclerotic rabbit model, the MR enhancement in the aortic wall persisted up to 8 and 12 h for nontargeted and α_νβ₃-targeted nanoparticle formulations, respectively (22). Noninvasive MR molecular imaging results were closely corroborated by microscopic estimates of microvessel density within the aortic wall. Both MR and histology objectively measured overall neovascularity by encompassing all regions of the abdominal aorta, including the entire vessel circumference and all imaged slices and not only subjectively selected areas identified as “hot spots” of plaque development.

Neovascularization is a prominent feature of human atherosclerotic lesions characterized as unstable lipid-rich plaques, but only a minor component of stable fibrocalcific lesions (6). Angiogenesis not only provides a portal of entry for blood-borne cells and nutrients into the inflamed tissue (23), but these fragile neovessels are also prone to shear-induced disruption, resulting in intraplaque hemorrhage. Intraplaque hemorrhage induces amplification of intramural inflammatory and immune processes, which may ultimately progress to plaque rupture (8,24- 25).

The α_νβ₃-integrin is a heterodimeric transmembrane glycoprotein and biomarker that is differentially up-regulated in proliferating versus quiescent endothelial cells, but it is also expressed by numerous cell types prominently represented in atherosclerotic plaques, including endothelial cells (26- 27), macrophages (28), platelets (29), lymphocytes (29), and smooth muscle cells (30). We have shown that α_νβ₃-targeted paramagnetic nanoparticles specifically detect and quantify angiogenesis in atherosclerotic plaques of hypercholesterolemic rabbits (31). Microvessel counts have been closely correlated with the MR imaging signal enhancement, which increased monotonically at higher adventitial microvessel counts and decreased rapidly as neovessel counts decreased (32). Moreover, MR signal enhancement has been clearly shown to reflect a neovascular response to targeted antiangiogenic treatment given alone and in synergistic combination with atorvastatin therapy (32- 33). Histological staining of the α_νβ₃-integrin itself was not performed in this study because numerous cell types, including macrophages, platelets, lymphocytes, and smooth muscle cells (32), express this integrin. Earlier publications have demonstrated that α_νβ₃-targeted nanoparticles are confined to the vasculature (12,32) where they can only interact with endothelial cells. Therefore, histological staining of an endothelial marker, such as von Willebrand factor, is more likely to reflect possible sites of nanoparticle binding than staining for the integrin itself, which would greatly overestimate particle binding.

Dietary benfluorex, which increases neurotransmitter serotonin levels to produce satiety and loss of appetite, decreased diet intake to the lean control consumption. Body weight and circulating levels of insulin and leptin were reduced relative to the untreated obese animals, but were still elevated compared with the lean controls, in agreement with previous reports (18- 19). In earlier studies, the pairing of obese rat food intake to control animals and the feeding of benfluorex or d-fenfluramine have been found to diminish cardiovascular disease and myocardial lesions in this corpulent rat model (14,18- 19,34- 35). Ad libitum feed consumption in benfluorex-treated obese rats was comparable to the intake of their lean control counterparts, which was mirrored by a low aortic neovascular MR signal in both groups of animals. Collectively, these findings support the contention that the progression of cardiovascular disease is correlated with increasing intramural angiogenesis, which can be noninvasively and serially interrogated and quantified with MR molecular imaging. Other changes in body weight and hormonal profiles of the obese rats were directionally similar to the appetite suppression effects of benfluorex, but only noninvasive MR assessments of angiogenesis closely paralleled the diminished food consumption effect of the lean control group.

One limitation of the present study and other animal model studies is the focus on very early atherosclerotic disease because many of the key features of human atherosclerotic plaque reported to amplify neovascular response, such as intraplaque hemorrhage and associated T cell–mediated immune response, are not present. Although the ameliorative effect of benfluorex on dietary intake, body weight increase, and aortic neovascular expansion in early plaque was evident, the regression or stabilization of advanced plaque remains to be demonstrated.

This study measured the MR imaging enhancement arising from only the α_νβ₃-integrin targeted nanoparticle formulation. Although we did not use nontargeted nanoparticles to determine the nonspecific accumulation of the contrast agent due to the hyperpermeability of angiogenic vessels, this issue has been explored in previous studies of animal models of cardiovascular disease and cancer (10,31,33). Typically, the nontargeted formulation produces approximately 50% less image enhancement compared with the targeted particles. Although it is reasonable to assume that the nontargeted agent would also yield 50% less enhancement in the JCR rat, further studies will be required to assess the nonspecific contribution and the dependence of particle dose and post-injection time point on the nontargeted component.

The clinical use of benfluorex has been discontinued in the U.S. due to safety concerns. However, the molecular imaging techniques described in this study are sensitive to neovascular proliferation in the aortic wall and are not limited to only detecting the response to benfluorex treatment. This method could be applied to monitoring the end-organ effects of a vast array of therapeutic approaches, such as lifestyle modifications, surgical interventions, lipid-lowering medications, and glucose normalization.

These data represent the first report of molecular imaging of angiogenesis at a clinical field strength to detect early metabolic vasculopathy and to quantify therapeutic response to benfluorex in the corpulent JCR:LA-cp rat model of metabolic syndrome. Benfluorex, an appetite suppressant, reduced ad libitum food intake equivalent to that in lean control rats, which reduced body weight, insulin, and leptin levels of the obese rats relative to those of untreated animals. However, none of these measures were reduced to the control level in the lean rats. MR neovascular signal enhancement, on the other hand, was markedly reduced by benfluorex to the contrast levels measured in the lean rats, which closely paralleled the appetite suppression effects. These results support the contention that MR molecular imaging of angiogenesis with integrin-targeted nanoparticles could provide a sensitive, high-resolution signal for monitoring progression and treatment of cardiovascular disease.

The authors thank Cordelia Caradine, John Allen, and Ralph Fuhrhop for their technical assistance.