Sage Journals: Discover world-class research

Abstract

The aim of this study was to investigate the feasibility of noninvasive monitoring of plaque burden in apolipoprotein E–deficient (ApoE^−/−) mice by Gadospin F (GDF)-enhanced magnetic resonance imaging (MRI). Gadolinium uptake in plaques was controlled using transmission electron microscopy (TEM) and x-ray fluorescence (XRF) microscopy. To monitor the progression of atherosclerosis, ApoE^−/− (n = 5) and wild-type (n = 2) mice were fed a Western diet and imaged at 5, 10, 15, and 20 weeks. Contrast-enhanced MRI was performed at 7 T Clinscan (Bruker, Ettlingen, Germany) before and 2 hours after intravenous injection of GDF (100 μmol/kg) to determine the blood clearance. Plaque size and contrast to noise ratio (CNR) were calculated for each time point using region of interest measurements to evaluate plaque progression. Following MRI, aortas were excised and GDF uptake was cross-validated by TEM and XRF microscopy. The best signal enhancement in aortic plaque was achieved 2 hours after application of GDF. No signal differences between pre- and postcontrast MRI were detectable in wild-type mice. We observed a gradual and considerable increase in plaque CNR and size for the different disease stages. TEM and XRF microscopy confirmed the localization of GDF within the plaque. GDF-enhanced MRI allows noninvasive and reliable estimation of plaque burden and monitoring of atherosclerotic progression in vivo.

ATHEROSCLEROSIS is characterized by the formation of inflammatory plaques that occupy the intimal layer of arterial walls. Despite advances in understanding of the pathogenesis of atherosclerosis, resulting cardiovascular diseases such as cardiac infarction, stroke, and renal failure remain the main cause of mortality in industrialized and developing nations.¹ Growing evidence suggests that the decisive factor determining plaque vulnerability is more dependent on the plaque configuration than on the degree of luminal narrowing.^2,3 However, luminographic techniques, such as conventional angiography and contrast-enhanced computed tomography (CT) or magnetic resonance angiography, are still used in the clinical diagnostic, although these techniques underestimate the true burden of atherosclerosis.⁴

Recent improvements in conventional and specifically molecular magnetic resonance imaging (MRI) proved to be a promising tool in allowing evaluation of plaque composition at the cellular and molecular levels, thus improving the discrimination between stable and vulnerable plaque.^5,6 In particular, contrast-enhanced MRI using extracellular gadolinium (Gd)-based contrast agents has shown promise for highlighting specific plaque components.^7,8

Gadofluorine M (GDM; Bayer Schering Pharma AG, Berlin-Wedding, Germany) has been described as a reliable, noninvasive tool to identify and quantify plaque burden in animal models of atherosclerosis⁹ as it binds to plaque and specifically targets the extracellular matrix of plaque,¹⁰ thus making it an important marker of plaque staging.¹¹ Furthermore, a Gd-labeled elastin-specific magnetic resonance contrast agent (ESMA) that allows visualization and quantification of atherosclerotic plaque burden in apolipoprotein E–deficient (ApoE^−/−) mice at 3 T MRI has been reported.¹²

Gadospin F (GDF; Miltenyi Biotec GmbH, Bergisch Gladbach, Germany) is an amphiphilic Gd-based contrast agent with a high protein-binding affinity specifically formulated for preclinical MRI. With a perfluorinated side chain, it is similar to GDM; however, it has not been established for preclinical 7 T MRI studies on ApoE^−/− mice, and the sensitivity and specificity of this contrast agent for monitoring plaque progression are not yet known.

Typically, well-established ApoE^−/− mice are used for the evaluation of atherosclerotic disease.^13,14 Examination by noninvasive imaging techniques is preferable as longitudinal preclinical studies on the development of plaque burden, the progression of disease, and the evaluation of therapy response become more reliable. The first sites examined for the development of plaques included the aortic root, the lesser curvature of the aortic arch, the branches of the brachiocephalic artery, and the right common carotid artery.¹⁴

To get an ideal spatial and chemical resolution of GDF within the aortic plaque, imaging analysis was performed using x-ray fluorescence (XRF) microscopy, which, to the best of our knowledge, has not been reported previously. We obtained precise trace element ion maps in situ of atherosclerotic tissues and the spatial distribution of Gd and other metal ion localizations, such as P, S, Cl, K, Ca, Zn, Cu, and Fe, in this specific compartment of aorta tissue. We used the original double ring store (DORIS III) as a dedicated synchrotron radiation source at DESY (Hamburg, Germany), and the circumference of 289 meters allowed us to store positrons and electrons at an energy of 4.45 GeV and in a bunched package. The DESY campus at Hamburg (there is another campus at Zeuthen near Berlin) is one of the most intense x-ray sources in the world. The radiation in the extreme ultraviolet and x-ray regimen is used in a broad range of applications, from biology and chemistry to materials science, for example, to characterize the complex anatomic structures of the human inner ear.¹⁵

In this study, we established a reliable and easy to use imaging strategy to define plaque burden by using GDF, an amphiphilic small-molecular-weight MRI contrast agent. The focus was put on a short imaging time without electrocardiography (ECG) or respiratory triggering while still keeping good image quality. Moreover, the efficacy of GDF to assess plaque progression by signal intensity measurements was evaluated. Additionally, the localization and distribution pattern of GDF within the atherosclerotic plaque were verified by transmission electron microscopy (TEM) and XRF microscopy.

Material and Methods

Contrast Agent

GDF has a low molecular weight (1.3 g/mol) and is a derivate of Gd-DO3A, containing a perfluorinated side chain. GDF has an r1 relaxivity of 18 mmol/L⁻¹s⁻¹ in plasma and 15 mmol/L⁻¹s⁻¹ in water (1.5 T and 37°C). Due to the hydrophobic character of the fluorinated side chain, GDF assembles like small aggregates or micelles in diluted solution. The driving force is the hydrophobic character of the fluorinated site chain. Therefore, GDF is lipophilic compared to agents such as Gd–diethylene triamine pentaacetic acid (DTPA) but is water soluble.

Animals

All B6.129P2-apoE^tm1Unc/J (C57BL/6J background) mice (ApoE^−/−) were purchased from Charles River Laboratories/Jackson Laboratories (Bar Harbor, ME) and were housed in the animal unit of the University Medical Center Hamburg-Eppendorf, Germany. C57BL/6J wild-type mice were used as controls. ApoE^−/−mice were fed a Western diet (ssniff EF R/M from TD88137 mod., ssniff Spezialdiäten GmbH, Soest, Germany) to induce atherosclerotic plaques.

Magnetic Resonance Imaging

MRI examinations were performed using a 7 T preclinical animal MRI scanner (Bruker Clinscan, Ettlingen, Germany). Animals were anesthetized with 1 to 2% isoflurane in O₂ gas mixture with a flow rate of 600 mL/min. The vital parameters were controlled by measuring the respiratory breathing cycle (SA Instruments, Stony Brook, NY), and the body temperature was maintained at 37°C during examination using a water-bed heating system. A four-channel mouse surface coil, with a coil diameter of 20 mm for signal receiving and a rat body coil for radiofrequency transmission, was used (Bruker, Ettlingen, Germany). The scan protocol included the survey scan followed by a parasagittal T₁-weighted two-dimensional (2D) turboflash sequence to cover the aortic arch (Figure 1, Table 1). Based on the parasagittal scan, the T₁-weighted, three-dimensional (3D) inversion recovery sequence was planned in transversal direction, covering the thoracic aorta. No ECG or breath triggering was used.

Figure 1.

MRI protocol and region of interest (ROI) settings. A, Transversal planning of the T₁-weighted 3D inversion recovery sequence (indicated by white lines) covering the aorta on the parasagittal image. B, For image analysis, contrast to noise ratio calculation and plaque size estimation ROI were placed around the atherosclerotic plaque and in the muscle. a = aortic root; b = brachiocephalic artery.

Table 1.

MRI Sequence Parameters

MRI Sequence	TR (ms)	TE (ms)	TI (ms)	NSA	Matrix	FA	Voxel (mm³)	Slices	Time (min)
Parasagittal 2D turoflash	2,450	2.6	1,450	10	320 × 320	20	0.09 × 0.09 × 0.7	1	00:25
Transversal 3D turboflash	650	2	250	6	192 × 192 × 64	20	0.18 × 0.18 × 0.18	64	09:49

FA = flip angle; MRI = magnetic resonance imaging; NSA = number of signal averages; TE = echo time; TI = inversion time; TR = repetition time.

Earlier and Late Contrast-Enhanced MRI

To determine the signal intensity (SI) in blood over time of GDF in vivo, ApoE^−/− mice (n = 3) were scanned using the 3D inversion recovery sequence (see Table 1), each for 10 minutes. First, a scan to determine the baseline was performed before GDF (100 μmol/kg) was injected intravenously followed by MRI for 2 hours.

Monitoring Disease Progression by MRI

ApoE^−/− mice (n = 5, 16 weeks old) and wild-type mice (n = 2, 16 weeks old) were fed a high-fat diet (HFD) for a period of 20 weeks. To follow disease progression, animals were imaged after 5, 10, 15, and 20 weeks of the HFD. Images were obtained before and 2 hours after intravenous injection of GDF (100 μmol/kg).

Image Analysis

MRIs were analyzed by Image Processing and Analysis in Java (ImageJ, version 1.45, National Institutes of Health, Bethesda, MD). To obtain blood clearance, regions of interest (ROI) were placed in the ascending aorta and muscle, and intensities of every single image in a time series S(t) were exported and normalized to the average intensity S₀ contrast before GDF injection. Moreover, ROI were placed in muscle and plaque, as seen in Figure 1, for each slice to estimate plaque size and CNR. Plaque area was calculated by adding the single plaque ROI of each slice. To determine the CNR over time in the plaque of the aortic root and the brachiocephalic artery, the signal difference relative to the muscle was calculated according to the following equation:

C N R = \frac{S I_{p l a q u e} - S I_{m u s c l e}}{0.5 \times (σ_{p l a q u e} + σ_{m u s c l e})} .

Statistical Analysis

Statistical analysis was performed using GraphPad Prism 5 (GraphPad Software, La Jolla, CA). Values are expressed as means ± standard error of the mean (SEM). For all statistical comparisons, a Student t-test (unpaired, two-tailed) was applied. The size and CNR of plaque were compared using Pearson correlation coefficient to assess the validity. Correlation coefficients > 0.7 were considered strong, 0.7 to 0.5 moderate, 0.5 to 0.3 weak, and 0.3 to 0 almost nonexistent. Probability values < .05 were considered statistically significant.

Transmisson Electron Microscopy

For TEM analysis, mice were perfused with 2.5% glutaraldehyde and 4% paraformaldehyde and aortas were removed from surrounding tissue.

After preselection of the aortic plaque areas by light microscopy, the samples were postfixed with 1% osmium in phosphate-buffered saline (PBS) and 0.5% gallic acid in PBS, dehydrated, and embedded in EPON resin. For the evaluation of the Gd contrast by TEM, samples were prepared according to the Tokuyasu cryosectioning technique protocol, excluding the uranyl acetate staining to prevent false-positive findings.¹⁶ To compare the TEM analysis of native aortic sections to GDF-injected sections, the settings of the camera and the acquisition software were equal. Images were acquired with a Philips CM120 electron microscope equipped with a BioTWIN lens and a Gatan MSC 794 Camera (Gatan GmbH, München, Germany).

XRF Imaging

XRF imaging was performed at beamline L at the synchrotron light source DORIS III, and sample preparation was performed as previously described.¹⁷ Briefly, immediately after tissue isolation, 20 μm thin cross sections were prepared using a cryomicrotome (Leica, Biosystems, Wetzlar, Germany). The samples were directly attached to a 200 nm thick silicon-nitride membrane, freeze-dried, and stored over silica gel in a desiccator prior to x-ray measurements. Hard x-ray microscopy at E = 10.2 keV was used to mainly target Gd L emission lines and K lines of P, S, Cl, K, Ca, Zn, Cu, and Fe. To identify Gd, two L lines were chosen: Lα1 at 6057.2 eV and Lβ1 at 6137.2 eV. The experimental setup included a multilayer monochromator for high-photon flux and a polycapillary half-lens to focus the incoming x-ray beam to a spot of 15 μm diameter. A silicon drift detector positioned at 90° to the incoming x-ray beam equipped with a standard collimator with a 2 mm pinhole was used to collect the fluorescence signal. A 100 μm thick National Institute of Standards and Technology standard (μM 612) measured with the same setup was used for signal calibration to allow quantification of element concentrations in the samples. Two-dimensional maps were created by scanning the sample through the x-ray beam. At each point, XRF spectrum was collected. The spectra were subsequently fitted using AXIL (IAEA Laboratories Seibersdorf, Seibersdorf, Austria) and PyMca (ESRF, Grenoble, France) softwares.^18,19

Ethics Statement

The experiment was supervised by the institutional animal welfare officer and approved by the local licensing authority (Behörde für Soziales, Familie, Gesundheit und Verbraucherschutz, Amt für Gesundheit und Verbraucherschutz, Hamburg, Germany, project no. 33/10).

Results

Earlier and Late Contrast-Enhanced MRI

Earlier and late contrast-enhanced MRIs revealed that 2 hours after intravenous application of GDF, the majority of the contrast agent was cleared from the blood, so the vessel lumen appeared dark on T₁-weighted MRIs, as seen prior to injection (Figure 2). Muscle tissue was used as a control because it showed low uptake.

Figure 2.

Plaque imaging and blood clearance by MRI. Parasagittal slice orientation through the aortic arch. Atherosclerotic plaques were not detectable on precontrast scans (A). A significant signal enhancement within the atherosclerotic plaque along the aortic arch was achieved 2 hours after intravenous injection of GDF in ApoE^−/− mice (B), which were fed on a high-cholesterol diet for 20 weeks. Signal enhancement was completely lost 2 days after injection (C). No signal difference was detectable in the wild-type mice (control) (D, E). The white arrows are pointed at aortic plaque lesions along the aortic arch. Earlier and late contrast-enhanced MRIs (F) of ApoE^−/− mice showed the highest signal intensity within the blood after 10 minutes and exhibited a continuous decrease over the next 2 hours. Contrast agent was cleared from blood 2 hours after application of GDF. Muscle as control tissue showed no GDF uptake. a = aortic root; b = brachiocephalic artery; c = carotid artery; p = pulmonary artery; s = subclavian artery; SI₀ = signal intensity contrast before injection of GDF; SI(t) = signal intensity at the different time points after injection of GDF.

Plaque Detection and Disease Monitoring by MRI

Significant signal enhancement along the aortic arch was achieved 2 hours after injection of GDF in ApoE^−/− mice, whereas no signal differences between pre- and postcontrast images were detected in wild-type mice (see Figure 2). At that time point, GDF was completely cleared from the blood. When imaging was repeated 2 days postinjection, the observed signal enhancement in the plaque of ApoE^−/− mice completely vanished.

To monitor disease progression by MRI, GDF was injected into five ApoE^−/− mice at four different time points on the HFD and images were taken before and 2 hours after GDF application. A gradual and considerable increase in plaque CNR on MRI was observed for all time points (Figure 3). These temporal changes in plaque CNR of the aortic root (Figure 4) were significantly different between mice on an HFD for 10 and 15 weeks (p < .001) and for 15 and 20 weeks (p < .001), whereas no significant differences were detected by CNR between the group on an HFD for 5 and 10 weeks. The CNR for brachiocephalic plaque over time (Figure 3B) was significantly different between 5 and 10 weeks (p < .01) and 15 and 20 weeks (p < .001), whereas no significant difference was measurable between 10 and 15 weeks.

Figure 3.

Monitoring of plaque progression by contrast to noise ratio (CNR) estimation. The temporal changes in plaque CNR of the aortic root were significantly different between 10 and 15 weeks (p < .001) and 15 and 20 weeks (p < .001), whereas no significant differences were detectable in the CNR between 5 and 10 weeks. The CNR over time for brachiocephalic plaque was significantly different between 5 and 10 weeks (p < .01) and 15 and 20 weeks (p < .001), whereas the difference between 10 and 15 weeks was not significant. The white arrows are pointed at the vessel wall of the aortic root or brachiocephalic artery. a = aortic root; b = brachiocephalic artery.

Figure 4.

Monitoring of plaque progression by plaque size estimation and in correlation with contrast to noise ratio (CNR) measurements. For the aortic root, the plaque size increased from 0.52 ± 0.01 mm³ after 5 weeks on a high-cholesterol diet to 0.83 ± 0.01 mm³ after 10 weeks (p < .01), to 1.57 ± 0.02 mm³ (p < .001) and to 2.25 ± 0.01 mm³ after 20 weeks (p < .001) (A). Plaque size in the brachiocephalic artery was smaller but also gradually increased over time. After 5 weeks on a high-cholesterol diet, the plaque size was 0.15 ± 0.01 mm³; after 10 weeks, 0.26 ± 0.02 mm³ (p < .001); after 15 weeks, 0.37 ± 0.02 mm³ (p < .05); and after 20 weeks, 0.54 ± 0.02 mm³ (p < .01) (A). A strong correlation was achieved for plaque size and CNR measurements in the aortic root (r > .78; p < .001) (B), whereas a moderate correlation was determined between plaque size and CNR measurements of the brachiocephalic artery (r > .68; p < .01) (C).

These temporal changes in plaque CNR were in agreement with the estimated progression in plaque size (see Figure 4A). For the aortic root, plaque size increased from 0.52 ± 0.01 mm³ after 5 weeks on an HFD to 0.83 ± 0.01 mm³ after 10 weeks (p < .01), to 1.57 ± 0.02 mm³ (p < .001), and to 2.25 ± 0.01 mm³ after 20 weeks (p < .001). Plaque size in the brachiocephalic artery was smaller but also increased gradually and substantially over time. After 5 weeks on an HFD, the plaque size was 0.15 ± 0.01 mm³; after 10 weeks, 0.26 ± 0.02 mm³ (p < .001); after 15 weeks, 0.37 ± 0.02 mm³ (p < .05); and after 20 weeks, 0.54 ± 0.02 mm³ (p < .01).

Plaque size and CNR measurements in the aortic root revealed a strong correlation (r = .787; p < .01; 95% confidence interval). Correlation between plaque size and the CNR of the brachiocephalic artery was somewhat lower but still moderate (r = .682; p < .01; 95% confidence interval).

Transmission Electron Microscopy

The contrast of uninjected (Figure 5A) and post–GDF-injected aortic plaque (Figure 5E) images obtained by TEM of conventionally stained with uranyl acetate/osmium tetroxide and resin-embedded samples is similar. The images show a cross section of a typical peripheral plaque area with endothelial cells and foam cell residues. Gd of the GDF could not be detected from contrast differences within the aortic plaque by using conventional TEM. However, after performing TEM analysis of the unstained Tokuyasu cryosections, the contrast of the aortic plaque injected with GDF is higher (Figure 5, F and H) compared to the control (Figure 5, B–D) as Gd acts as a known electron-dense material.

Figure 5.

Transmisson electron microscopy (TEM) of EPON and TEM of Tokuyasu cryosections of plaque. The contrast of uninjected (A) and post–GDF-injected aortic plaque (E) images obtained by TEM of EPON-embedded samples is similar. The images show a cross section of a typical peripheral plaque area with endothelial cells (ec) and foam cell residues (fcr). Gadolinium of the GDF could not be detected from contrast differences within the aortic plaques by using conventional TEM. However, after performing TEM analysis of the Tokuyasu cryosections, the contrast of the aortic plaque injected with GDF was higher (F–H) compared to the control (B–D) as gadolinium acts as a known electron-dense material. The white arrow indicates gadolinium. vl = vessel lumen.

XRF Imaging

To determine the Gd distribution within the plaque after GDF injection, we performed an XRF study of aortic 20 μm thick cross sections after MRI analysis.

Figure 6 shows a typical image taken with the XRF microscope. The image was collected of an area of 320 × 1,140 μm, with 32 × 114 pixels with a dwell time of 10 seconds per pixel. The absolute area concentration is expressed in arbiter units (color bar). Figure 6A shows the summed fluorescence spectrum collected from a freeze-dried cross section of atherosclerotic tissues, previously observed in MRI (Figure 6B) and visible by light microscopy (Figure 6C). The fluorescence maps of Gd L lines and K lines of Fe, Ca, Zn, P, S, Cl, K, and Cu are shown in Figure 6D. By using two L emission lines of Gd (Gd Lα1 Gd Lβ1) (see Figure 6D) from the fluorescence spectra (Figure 6A), we were clearly able to localize the GDF in the tissue. Strong Gd contributions, as well as P, S, Ca, K, and Zn, were observed in the plaque. However, Fe was differently localized than Gd within that particular tissue, whereas S and Cl were distributed quite homogeneously over the whole aorta cross section (see Figure 6, D and E). The overlaying images (see Figure 6, E and F) show the colocalization of the Gd and P, probably originating from the phospholipids.

Figure 6.

X-ray fluorescence of atherosclerotic plaque obtained at 10.2 keV photon energy. A, The summed fluorescence spectrum of the scanned area along with a fit (sum: black line; fit: red line; individual peaks: see legend). B, MRI of atherosclerotic plaque. C, Visible light microscopy images taken with a Nikon Eclipse Ti microscope of the sample before mounting for the x-ray microscopy experiment. D, Fluorescence images of gadolinium Lα1 and Lβ1 and K lines of Fe, Ca, Zn, P, S, Cl, K, and Cl, respectively. E, Overlying distribution of Gd (blue), P (red), and Fe (green). F, Overlying distribution of Gd (blue), P (red), and S (green). Image size 320 × 1,140 μm, pixel size 10 μm, with a dwell time of 10 seconds per pixel. The color scale with the corresponding minimum (blue) and maximum (red) values indicates the number of normalized counts.

Discussion

In the present study, we performed multimodal imaging of atherosclerotic tissue in mice. We established the detection of atherosclerotic plaques in ApoE^−/− mice at 7 T MRI by using GDF, an amphiphilic small-molecular-weight MRI contrast agent. Earlier and late contrast-enhanced MRIs determined the optimal time point for plaque imaging 2 hours after intravenous injection. Changes in plaque development at various stages of atherosclerotic disease were assessed through CNR calculation, which increased over time and was proportional to the increase in plaque size. A moderate to strong correlation could be achieved between plaque size and CNR measurements. TEM analysis from ultramicrotome aortic sections¹⁶ further proved that post-GDF sections seemed to be higher in contrast than plaques not injected with GDF. In addition, we were able to confirm the elementary structure of GDF-bound Gd within the plaque by XRF microscopy.

Gd-based contrast agents are known to distribute into the extracellular fluid space, and the detection of atherosclerotic plaques using Gd-DTPA, GDM, or Gd-labeled ESMA has been described before.^10,12,20,21 Recently, Gd-DTPA was applied to image atherosclerosis in human subjects by surface MRI.^22–24 When T₁-weighted surface MRI was applied, a 29 to 80% increase in signal intensity of fibrous tissue with Gd-DTPA in carotid arteries occurred.^22,24 This is typically limited to one vascular bed (or even one MRI slice) due to the rapid wash-in/out plaque kinetics of currently approved agents, including Gd-DTPA. This leads to a short imaging window of about 20 minutes and limits the use of these agents for imaging multiple plaques in different vascular beds.²¹ Due to the short imaging window, which is even smaller in mice, the use of Gd-DTPA in ApoE^−/− mouse models for the detection and estimation of CNR in atherosclerotic plaque is, from our experience, not feasible. The wash-in/out plaque kinetic of GDF is slower, making it a reliable tool for the diagnosis of atherosclerosis in preclinical animal studies. The diagnosis of atherosclerotic plaque even at early stages using GDF is feasible. A disadvantage of GDM is the late imaging time point, 2 days after intravenous injection, no matter if rabbits or mice were used as an animal model. Compared to GDM, the best imaging time point for GDF used in the same concentration is 2 hours after intravenous injection.

The high-resolution MRI protocol enables us to quantify plaque burden by CNR and plaque size measurements. The differentiation between early and advanced plaques according to the CNR was also shown in New Zealand White rabbits in different young (fed on an HFD for 2 months) and older (fed on an HFD for 8 months) animals.⁹ Our results proved that the monitoring of plaque progression within the same ApoE^−/− mouse at 7 T MRI with short time intervals of 5 weeks is realistic. We achieved significant differences for almost all time points for the aortic root and the brachiocephalic artery. This may help in the discovery of novel therapeutics that slow or reverse the disease and represents a good tool for preclinical longitudinal MRI studies.

To analyze the distribution pattern of GDF within atherosclerotic plaque and to establish a second-line tool for the detection of GDF within the aortic plaque, TEM and XRF analysis was performed. To the best of our knowledge, TEM and XRF analysis of atherosclerotic plaque to prove the uptake of GDM or GDF has not been reported previously. Makowski and colleagues showed the specific distribution of Gd-labeled ESMA by TEM.¹² However, to clarify the detection of Gd, a microscope was equipped with an EDAX energy dispersive spectroscopy detector (EDAX, Mahwah, NJ). The evidence of Gd-DTPA within neural and mesenchymal stem cells has also been performed by TEM.^25,26 In those cells, the Gd-DTPA forms clusters, which make visualization much easier. That Gd acts as an electron-dense material was also confirmed by the study of Cao and colleagues, in which Gd nanoparticles were synthesized and the spherical shapes were analyzed by TEM.²⁷ In agreement with those studies, our TEM analysis was evaluated without uranyl acetate staining to prevent false-positive findings. Thus, we did not detect clear differences between the GDF-treated plaque and the negative control, the native plaque, by TEM of resin-embedded samples. This may be due to the fact that the GDF is diffusely distributed within the plaque. This is in accordance with the electron microscopy analysis of dermal deposits in a patient with nephrogenic systemic fibrosis. Here the question of whether anorganic material deposits such as calcium and/or magnesium also contain Gd²⁸ remained unsolved. However, when the aortic sections were prepared according to the Tokuyasu cryosectioning protocol, we achieved a higher contrast for the injected plaque compared to the control due to the fact that Gd acts as an electron-dense material.^25–27

The XRF analysis provided elemental mappings, especially Gd, in situ within the atherosclerotic plaque. It is possible to distinguish clearly the part of the plaque where Gd was localized (see Figure 6). The measurements of the typical Gd mapping within the plaque were in agreement with previous studies.¹² We observed a pronounced enrichment of Gd in the atherosclerotic mouse tissues. Until now, XRF was not used for the detection of GDF in atherosclerotic plaques; however, this method gave an ideal spatial and chemical resolution to answer the question of where the Gd-lipid deposits are localized in the aorta. We were also able to detect the endogenous elements P, K, Ca, Cl, S, Fe, Cu, and Zn, which has not been shown before. In the long term, this analytical method represents a good opportunity to make eventual diagnosis in the early stages of disease.

Limitations

This study has several limitations. As the aim of the study was to evaluate plaque progression for the same individual over time, we did not perform a correlation of plaque size and CNR to histologic analysis. Sometimes it was difficult for the wild-type mice and the first measured time point at 5 weeks on an HFD to discriminate the vessel and the aortic plaque from surrounding tissue. Also, for TEM analysis, only a small section of the atherosclerotic plaque can be included. Although electron microscopy requires relatively invasive sample preparation, typically including fixation and thin slicing of the sample, x-ray microscopy can provide spectra without the risk of changing the cell organization by chemical fixation. Another important difference with respect to TEM-EDAX (Energy Disperse X-Ray Analysis) is related to the depth of the probe. Whereas EDAX is surface sensitive to the outer molecular layers, XRF excited by synchrotron radiation also collects signal from entire 20 μm thick sections.

Conclusion

We demonstrated the successful use of GDF-enhanced MRI at 7 T for atherosclerotic plaque detection in ApoE^−/− mice. The optimal time point for plaque imaging was determined 2 hours after GDF injection. Due to the assessment of changes in plaque sizes and CNR at various stages of atherosclerosis, the monitoring of plaque progression is feasible. In addition, we were able to show that the detection of GDF by TEM (Tokuyasu) and by XRF microscopy within the atherosclerotic plaque is reliable.

Footnotes

Acknowledgments

We would like to thank Dr. K. Appel for assistance in using the beamline L.

Financial disclosure of authors: This work was supported from the Hamburg State excellence initiative Nanotechnology in Medicine (NAME). T. Dučić carried out the XRF experiments at the DORIS III light source (beamline L), a member of the Helmholtz Association (HGF), supported through projects II-20080239 and I-2011094: “X-Ray Microscopy of Gadolinium in Atherosclerotic Lesions.” M.H. is supported by EU FP7 project RESOLVE (FP7-HEALTH-2012-305707).

Financial disclosure of reviewers: None reported.

References

Ross

. Atherosclerosis is an inflammatory disease. Am Heart J 1999;138(5 Pt 2):S419–20, doi:10.1016/S0002–8703(99)70266–8.

Falk

Shah

Fuster

. Coronary plaque disruption. Circulation 1995;92:657–71, doi:10.1161/01.CIR.92.3.657.

Takaya

Yuan

Chu

. Association between carotid plaque characteristics and subsequent ischemic cerebrovascular events: A prospective assessment with MRI—initial results. Stroke 2006;37:818–23, doi:10.1161/01.STR.0000204638.91099.91.

Barkhausen

Ebert

Heyer

. Detection of atherosclerotic plaque with Gadofluorine-enhanced magnetic resonance imaging. Circulation 2003;108:605–9, doi:10.1161/01.CIR.0000079099.36306.10.

Tardif

Lesage

Harel

. Imaging biomarkers in atherosclerosis trials. Circ Cardiovasc Imaging 2011;4:319–33, doi:10.1161/CIRCIMAGING.110.962001.

Quillard

Libby

. Molecular imaging of atherosclerosis for improving diagnostic and therapeutic development. Circ Res 2012; 111:231–44, doi:10.1161/CIRCRESAHA.112.268144.

Cai

Hatsukami

Ferguson

. In vivo quantitative measurement of intact fibrous cap and lipid-rich necrotic core size in atherosclerotic carotid plaque: Comparison of high-resolution, contrast-enhanced magnetic resonance imaging and histology. Circulation 2005;112:3437–44, doi:10.1161/CIRCULATIONAHA.104.528174.

Kerwin

Hooker

Spilker

. Quantitative magnetic resonance imaging analysis of neovasculature volume in carotid atherosclerotic plaque. Circulation 2003;107:851–6, doi:10.1161/01.CIR.0000048145.52309.31.

Sirol

Moreno

Purushothaman

. Increased neovascularization in advanced lipid-rich atherosclerotic lesions detected by gadofluorine-M-enhanced MRI: Implications for plaque vulnerability. Circ Cardiovasc Imaging 2009;2:391–6, doi:10.1161/CIRCIMAGING.108.801712.

10.

Meding

Urich

Licha

. Magnetic resonance imaging of atherosclerosis by targeting extracellular matrix deposition with Gadofluorine M. Contrast Media Mol Imaging 2007;2:120–9, doi:10.1002/cmmi.137.

11.

Briley-Saebo

Mulder

Mani

. Magnetic resonance imaging of vulnerable atherosclerotic plaques: Current imaging strategies and molecular imaging probes. J Magn Reson Imaging 2007;26:460–79, doi:10.1002/jmri.20989.

12.

Makowski

Wiethoff

Blume

. Assessment of atherosclerotic plaque burden with an elastin-specific magnetic resonance contrast agent. Nat Med 2011;17:383–8, doi:10.1038/nm.2310.

13.

Zhang

Reddick

Piedrahita

Maeda

. Spontaneous hypercholesterolemia and arterial lesions in mice lacking apolipoprotein E. Science 1992;258:468–71, doi:10.1126/science.1411543.

14.

Nakashima

Plump

Raines

. ApoE-deficient mice develop lesions of all phases of atherosclerosis throughout the arterial tree. Arterioscler Thromb 1994;14:133–40, doi:10.1161/01.ATV.14.1.133.

15.

Lareida

Beckmann

Schrott-Fischer

. High-resolution x-ray tomography of the human inner ear: Synchrotron radiation-based study of nerve fibre bundles, membranes and ganglion cells. J Microsc 2009;234:95–102, doi:10.1111/j.1365–2818.2009.03143.x.

16.

Tokuyasu

. A technique for ultracryotomy of cell suspensions and tissues. J Cell Biol 1973;57:551–65, doi:10.1083/jcb.57.2.551.

17.

Ducic

Quintes

Nave

. Structure and composition of myelinated axons: A multimodal synchrotron spectro-microscopy study. J Struct Biol 2011;173:202–12, doi:10.1016/j.jsb.2010.10.001.

18.

Sole

Papillon

Cotte

. A multiplatform code for the analysis of energy-dispersive X-ray fluorescence spectra. Spectrochim Acta B 2007;62:63–8, doi:10.1016/j.sab.2006.12.002.

19.

Vekemans

Janssens

Vincze

. Analysis of x-ray-spectra by iterative least-squares (AXIL) – new developments. X-Ray Spectrom 1994;23:278–85, doi:10.1002/xrs.1300230609.

20.

Sirol

Itskovich

Mani

. Lipid-rich atherosclerotic plaques detected by gadofluorine-enhanced in vivo magnetic resonance imaging. Circulation 2004;109:2890–6, doi:10.1161/01.CIR.0000129310.17277.E7.

21.

Ronald

Chen

Belisle

. Comparison of gadofluorine-M and Gd-DTPA for noninvasive staging of atherosclerotic plaque stability using MRI. Circ Cardiovasc Imaging 2009;2:226–34, doi:10.1161/CIRCIMAGING.108.826826.

22.

Yuan

Kerwin

Ferguson

. Contrast-enhanced high resolution MRI for atherosclerotic carotid artery tissue characterization. J Magn Reson Imaging 2002;15:62–7, doi:10.1002/jmri.10030.

23.

Kramer

Cerilli

Hagspiel

. Magnetic resonance imaging identifies the fibrous cap in atherosclerotic abdominal aortic aneurysm. Circulation 2004;109:1016–21, doi:10.1161/01.CIR.0000116767.95046.C2.

24.

Wasserman

Smith

Trout

3rd

Carotid artery atherosclerosis: In vivo morphologic characterization with gadolinium-enhanced double-oblique MR imaging initial results. Radiology 2002;223:566–73, doi:10.1148/radiol.2232010659.

25.

Shen

Cheng

Zhong

. Efficient in vitro labeling rabbit neural stem cell with paramagnetic Gd-DTPA and fluorescent substance. Eur J Radiol 2010;75:397–405, doi:10.1016/j.ejrad.2009.04.040.

26.

Shen

Zhong

Duan

. Magnetic resonance imaging of mesenchymal stem cells labeled with dual (MR and fluorescence) agents in rat spinal cord injury. Acad Radiol 2009;16:1142–54, doi:10.1016/j.acra.2009.03.016.

27.

Cao

Shen

Wang

. Controlled intracellular self-assembly of gadolinium nanoparticles as smart molecular MR contrast agents. Sci Rep 2013;3:1024.

28.

Schroeder

Weingart

Coras

. Ultrastructural evidence of dermal gadolinium deposits in a patient with nephrogenic systemic fibrosis and end-stage renal disease. Clin J Am Soc Nephrol 2008;3:968–75, doi:10.2215/CJN.00100108.

Gadospin F—Enhanced Magnetic Resonance Imaging for Diagnosis and Monitoring of Atherosclerosis: Validation with Transmission Electron Microscopy and X-Ray Fluorescence Imaging in the Apolipoprotein E—Deficient Mouse

Abstract

Material and Methods

Contrast Agent

Animals

Magnetic Resonance Imaging

Earlier and Late Contrast-Enhanced MRI

Monitoring Disease Progression by MRI

Image Analysis

Statistical Analysis

Transmisson Electron Microscopy

XRF Imaging

Ethics Statement

Results

Earlier and Late Contrast-Enhanced MRI

Plaque Detection and Disease Monitoring by MRI

Transmission Electron Microscopy

XRF Imaging

Discussion

Limitations

Conclusion

Footnotes

Acknowledgments

References