Background
Genetically modified mouse models play a crucial role in the study and understanding of the mechanisms underlying atherogenesis. Among these models, in biomedical research, the apolipoprotein E-deficient (
apoE−/−) mouse is widely recognized as one of the most commonly used preclinical models for studying plaque formation and progression [
1]. ApoE acts as a ligand for receptors involved in the clearance of chylomicrons and remnants of very low-density lipoproteins (VLDL) [
1,
2]. In
apoE−/− mice, the absence of apoE results in impaired lipoprotein clearance, leading to elevated plasma cholesterol levels. This elevation stimulates atherogenesis, resulting in the development of complex lesions similar to those found in humans. These lesions consist of a fibrous cap containing smooth muscle cells, foam cells (lipid-loaden macrophages), and a necrotic core [
3]. To accelerate the formation of lesions, mice are often fed a Western or high-fat diet, which initiates progressive plaque formation starting at the age of 10–12 weeks [
4].
Positron emission tomography (PET) is an in vivo imaging technique that provides valuable insights into atherosclerotic processes in small animals. It involves labeling biomolecules such as proteins, metabolites, hormones, and drugs with a positron emitter, enabling their use as tracers to target specific biological processes. The metabolic PET tracer [
18F]-fluoro-2-deoxy-D-glucose ([
18F]FDG) is a radiolabeled glucose analog that is taken up by cells metabolizing glucose, including inflammatory cells involved in the atherosclerotic process, such as activated macrophages. Several clinical and preclinical studies on atherosclerosis have shown that the accumulation of [
18F]FDG correlates with regions of high macrophage density in atherosclerotic vessels and plaques [
5‐
8].
The combination of the
apoE−/− mouse model and [
18F]FDG PET imaging presents new research opportunities but also poses challenges. Atherosclerotic lesions in
apoE−/− mice are often smaller than the resolution of a dedicated small animal PET system (1–1.5 mm), making it difficult to distinguish lesions from surrounding tissues due to partial volume effects [
9].
Small animal PET imaging typically requires the use of anesthesia to prevent animal movement and motion artifacts. The combination of anesthesia and the high skin-surface-to-body-weight ratio of mice can lead to hypothermia [
10]. Hypothermia can activate brown adipose tissue (BAT), resulting in increased [
18F]FDG uptake, which can blur the [
18F]FDG signal from atherosclerotic lesions [
11‐
13].
Recently, an
apoE−/− rat model has become available, showing comparable characteristics to
apoE−/− mice but with the advantage of a larger size [
14‐
16], which allows for easier detection of small structures such as blood vessels and atherosclerotic tissue. Although hypothermia and activation of BAT can still be a concern in rats, the impact is, due to a lower skin-surface-to-body-weight ratio, less significant compared to mice.
In this study, we aim to further characterize the apoE−/− rat model and investigate its feasibility as an in vivo model for atherosclerosis (PET/CT) imaging.
Methods
Animals
Ten male apolipoprotein E-deficient rats (SD-ApoEtm1sage) (apoE−/−) from different litters, with a Sprague Dawley background and ten male Sprague Dawley wild-type control rats (Ctrl) at the age of 10 ± 1 weeks were obtained from Sage Labs Inc. located in Boyertown, Pennsylvania, U.S.A. Upon arrival, the rats were allowed a minimum acclimation period of 14 days to recover from transportation and adjust to their new housing conditions.
The rats were housed in Makrolon cages at a constant temperature of 21 ± 2 °C and maintained on a 12-h light/12-h dark cycle. Prior to the baseline scan, all animals were provided with a standard chow diet. Following the baseline scan, the rats were given ad libitum access to a Western diet high in fat (21%) and cholesterol (0.21%) (D12079B, Research Diets Inc., New Brunswick, New Jersey, U.S.A.). Additional details regarding the diet can be found in the Additional file
1: Table S1.
The experimental protocol was conducted in accordance with the ARRIVE guidelines and approved by the Central Committee on Animal Experiments of The Netherlands (license number AVD 105002016707) and the animal welfare body of the University Medical Center Groningen (protocol 16707-01-001).
Experimental design and procedure
At the baseline assessment, non-fasted rats were anesthetized using isoflurane mixed with oxygen (5% for induction and 2% for maintenance). A blood sample for biochemistry measurements and plasma glucose levels was collected from the tail vein. Subsequently, a bolus injection of approximately 64 ± 5 MBq of [18F]FDG was administered intravenously. After the injection, the rats were placed back in a pre-heated home cage to recover.
To optimize the contrast between plaque and background [
17‐
19], 3 h were allowed to elapse post-injection before the rats were re-anesthetized and positioned in a dedicated small animal PET/CT camera (D-PET, Inveon®, Siemens Preclinical Solutions, Knoxville, Tennessee, U.S.A.). The rats were positioned with their thorax at the center of the field of view. A CT scan was performed using the following parameters: 40 kV, 250 µA, total rotation of 360° in 360 steps, and an exposure time of 350 ms. Subsequently, a 20-min static PET scan was conducted.
During the PET scan, the animals were kept warm using heating pads and a thermostat set to a temperature of 38 °C (M2M Imaging, Cleveland, Ohio, U.S.A.). After the baseline scan, the rats were returned to a pre-heated home cage to recover from the procedure.
The entire procedure was repeated at weeks 4, 12, 26, and 51 ± 1 after the baseline assessment. Body weight was measured weekly (EL2000 scale, Shimadzu, Kyoto, Japan). Following the final scan, the rats were euthanized, and the heart and vessels were perfused using saline. The aortic arch, abdominal aorta, liver, cecum content, and faeces were collected and stored for further analysis.
PET and CT analysis
The PET data obtained from the scans were reconstructed using an ordered set expectation maximization-3D/maximum a posteriori (OSEM3D/MAP) algorithm. The reconstruction process involved 2 OSEM iterations followed by 18 MAP iterations. The final reconstructed PET datasets had an in-plane image matrix size of 256 × 256 pixels, with a voxel size of 0.388 × 0.388 × 0.796 mm and a resolution of 1.5 mm at the center of the field-of-view. The scans were corrected for decay, random coincidences, scatter, and attenuation.
The CT data were reconstructed using the Filtered Back Projection (FBP) algorithm. A low noise reduction technique, a Shepp-Logan filter, and a beam hardening correction were applied during the reconstruction process. The resulting CT images had a maximum voxel size of 0.099 × 0.099 × 0.099 mm and a pixel size of 99.24 µm.
To fuse the PET and CT images, the PMOD version 3.9 software (PMOD Technologies Ltd., Zürich, Switzerland) was used. A 20 × 20 × 20 mm box was created and positioned rostrally, immediately above the heart, to include the aortic arch within the box. Voxels outside of the box were masked, and background noise was eliminated by excluding voxels with values below 60 kBq/cc. For the abdominal aorta, a 15 × 15 × 15 mm box was created and positioned in line with the diaphragm. Similar to the aortic arch, voxels outside of the box were masked, and background noise within the box was removed by excluding voxels with values below 60 kBq/cc. Subsequently, the average and the maximum standardized uptake value (SUVmean, SUVmax) of [18F]FDG was calculated in the tissue in the selected boxes Additionally, SUVmean was corrected for plasma glucose levels using the following formula:SUVcorr = ((SUVmean * plasma glucose level) / plasma glucose levelmean per group per time point).
Extra regions of interest (ROI) were drawn, using the CT as anatomical reference, around the liver, lungs, kidneys, intestines, myocardium, vena cava and blood pool. Target-to-background ratio (TBR), defined as SUVmax aorta/ SUVmean vena cava, was calculated for the aortic arch and abdominal aorta.
The percentage of body fat was determined using procedures described in the previous work [
20].
Lipid, lipoprotein, and bile acid analysis
During the study, blood samples were collected from the tail vein of the rats. A drop of blood was used to measure the plasma glucose levels (Accu-Chek Roche Mannheim, Germany). The rest of the plama was stored in EDTA-coated tubes. Total plasma cholesterol and triglycerides were measured using commercially available kits from Roche Diagnostics (Basel, Switzerland).
To analyze the lipoprotein subspecies, the plasma pool of each group was injected onto a Superose 6HR10/300GL column (GE Health, Uppsala, Sweden). The column utilized gel filtration to separate the different lipoprotein subspecies, following a previously described method [
21]. At the time of animal euthanasia, the livers were excised and frozen in liquid nitrogen. For lipid analysis, lipids were extracted from liver homogenates using a modified version of the Bligh and Dyer procedure, as previously published [
22]. The extracted lipids were then dissolved in water containing 2% Triton X-100. Hepatic total cholesterol and triglyceride levels were measured using commercially available reagents from Roche Diagnostics (Basel, Switzerland).
After collection at the end of the study fecal samples from the cecum were dried, weighed, and grinded thoroughly. For neutral sterol and bile acid extraction, 50 mg of grinded faeces was heated at 80 °C in alkaline methanol for 2 h, followed by extraction with petroleum ether according to previously published methods [
23]. Bile acids were methylated using a mixture of acetyl chloride, trimethylsilytate with pyridine, N,O-bis(trimethyllysilyl)trifluoroacetamide, and trimethylchlorosilane. Gas–liquid chromatography was then used to measure the levels of fecal neutral sterols and bile acids, as detailed in a previous publication [
23].
Immunohistochemistry
To confirm the presence of atherosclerotic lesions, Oil Red O staining (ORO) for lipid accumulation was performed on the complete aortic arches using the en face procedure. Aortic arches were thawed and washed with water and then dehydrated using a 60% 1,2-propanediol solution. The dehydration process was repeated twice. Each aorta was incubated with ORO staining solution from Lifeline Cell Technology (Walkersville, Maryland, USA) for 10 min. After incubation, the aortas were washed with water. The unfolded arches were imaged using a Canon 200D digital camera with a Canon Macro EF-S 60 mm lens (Canon Inc., Tokyo, Japan). The obtained images were analyzed using ImageJ image analysis software (version 1.53i, U.S. National Institutes of Health, Bethesda, Maryland, U.S.A.). The area of Oil Red O-positive staining was divided by the total area to quantify the extent of staining.
To assess monocytes and macrophages, ED1 staining was performed on frozen abdominal aortas. ED1 is expressed by macrophages and monocytes. The 3 µm thick aorta sections were mounted on glass slides. The sections were fixed with acetone and washed with PBS. Endogenous peroxidase activity was blocked using 0.009% H
2O
2. The sections were then incubated for 60 min at room temperature with anti-ED1 antibody (MCA341R, Bio-Rad Laboratories, Veenendaal, The Netherlands) diluted in PBS with 1%. After several washes with PBS, the sections were incubated with a secondary antibody (RAMPO, P0260, Dako, Amstelveen, The Netherlands) diluted in PBS with 1% BSA and 1% normal rat serum. Subsequently, the sections were incubated with a tertiary antibody (GARPO, P0448, Dako) diluted in PBS with 1% BSA and 1% normal rat serum. Peroxidase activity was visualized using the chromogen AEC (3-amino-9-ethylcarbazole), and the sections were counterstained with hematoxylin. The stained slides were imaged using a slide scanner (Hamamatsu NanoZoomer 2.0 HT, Iwata, Japan). The quantification of ED1-positive macrophages was performed using Aperio ImageScope software (version v12.4.3.5008, Leica Biosystems Imaging Inc., Vista, CA, U.S.A.). The number of strongly positive pixels was divided by the total surface area to quantify the presence of ED1-positive macrophages (N
sp/A
tot) [
20] (algorithm is shown in Additional file
2: Table S2.)
Statistical analysis
For repeated measurements, the Generalized Estimating Equations (GEE) model was used to account for missing data at the different time points in the design. The independent correlation matrix was selected for the analysis, and the Wald test was used to report p-values which were considered statistically significant at p < 0.05 without correction for multiple comparisons. For data with a single time point, a T-test was used to compare groups or conditions.
Discussion
This study establishes the apoE−/− rat as suitable model for atherosclerosis imaging.
The phenotype observed in the
apoE−/− rats used in our study aligns with data from previous studies conducted by our institute and others. Using the same model, nearly all studies have reported signs of hypercholesterolemia, characterized by elevated levels of total plasma cholesterol and total plasma triglycerides, after 8 weeks of feeding an atherogenic diet [
14‐
16,
20,
25‐
27]. Some studies have noted the absence of lesions in different regions of the aorta [
14,
15,
25], which can be attributed to the relatively short duration of treatment with an atherogenic diet and the young to medium age of the rats. Other studies have observed lesions, but only after introducing a secondary trigger such as occlusal disharmony [
26] or challenging endothelial injury [
14]. Gao et al
. utilized a Paigen (high cholesterol/bile salt) diet for 10–12 weeks and observed severe coronary atherosclerosis characterized by significant lipid accumulation, macrophage accumulation, and collagen fibers deposits. However, the authors noted only mild atherosclerosis in the aortic root and over the full length of the aorta [
16].
In our previous work, which focused on the modulation of aortic relaxation by perivascular adipose tissue (PVAT), we concluded that the
apoE−/− rat serves as a model for early-stage atherosclerosis. The data revealed an increased intima/media thickness ratio and an influx of ED1-positive macrophages in the aortic intima of
apoE−/− rats. Consistent with the findings of this study, we observed deposits of plaque formation in the thoracic aortas of
apoE−/− rats [
20].
When comparing the phenotype of
apoE−/− rats with
apoE−/− mice, similar results were observed. Zhou et al
. reported hypercholesterolemia in mice aged 40–52 weeks treated with an atherogenic diet for 8 weeks [
28]. Similarly, Gogulamudi et al
. reported similar levels of cholesterol and triglycerides in 19-month-old
apoE−/− mice fed an atherogenic diet [
29]. Despite the comparable results in cholesterol and triglyceride levels, the severity and composition of the plaques in mice were significantly different. Several studies reported high plaque burden (around 20%) [
29,
30], and severe atheromas with necrotic cores and collagen deposition in the media in old
apoE−/− mice [
29]. In young mice (15 weeks of age), intermediate lesions containing foam cells and smooth muscle cells were found, and after 20 weeks, fibrous plaques appeared [
4]. The difference in plaque severity and composition between
apoE−/− mice and rats is not yet fully understood but could be explained by species differences [
25] or by the background strain of the rats, which may be more resistant to atherosclerosis [
27].
The results of this study were based on male rats only. Data from the supplier and previous studies in
apoE−/− mice and rats showed significantly elevated plasma cholesterol levels in males compared to females [
31], or a tendency for higher levels [
32]. Kong et al
. suggested that the differences between the two sexes disappeared after females reached menopause [
31], which suggests a potential effect of estrogen on plasma cholesterol. For future studies using this model, we recommend including both male and female rats to facilitate scientific and therapeutic discoveries for both sexes [
33].
The PET data in our study are consistent with the findings previously described by Zhuang et al
. [
34]. Although the SUV
mean and SUV
max were higher in our study, the differences in [
18F]FDG uptake between control and knockout rats, as well as the time points when the differences occurred, were comparable. However, despite the similarity in [
18F]FDG data between the two studies, the conclusions drawn are quite different. Zhuang et al
. concluded that [
18F]FDG uptake in the aortic arch is associated with HIF-1α gene expression (hypoxia) rather than inflammatory lesions. The authors observed high expression of HIF-1α in the aortic arch, which exhibited high [
18F]FDG uptake. Conversely, they noticed the opposite effect in the pulmonary arteries, with high CD68 expression but lower uptake of [
18F]FDG [
34]. Although no clear correlation data were presented, their conclusion was supported by Laurberg et al., who did not find [
18F]FDG accumulation in advanced atherosclerotic lesions of
apoE−/− mice [
13], and Myers et al., who did not find a correlation between [
18F]FDG and CD68 expression [
35]. An explanation for the disparities between Zhuang et al. and our study could stem from variances in the genetic models employed (Biocytogen vs Sage Labs Inc.) or the varying compositions of the diets utilized (42% fat vs 21% fat). An alternative explanation could be the chosen imaging procedure. It has been suggested in different studies to perform [
18F]FDG PET imaging 3 h after injection to maximize the contrast between plaque and background [
17‐
19,
36,
37]. In our study, we observed a weak-to-moderate correlation between [
18F]FDG uptake and ED1-positive cells. The correlation was lower than reported by Tawakol et al
. [
19] which could be explained by the rather poor co-registration of the PET imaging with the immunohistochemistry images. In this study, we quantified the number of macrophages in a random section of the abdominal aorta between the diaphragm and hepatic and splenic arteries. For [
18F]FDG uptake, we used a volume of interest (VOI) measuring 15 × 15 × 15 mm positioned in line with the diaphragm. Despite the limited overlap, we found a correlation between [
18F]FDG uptake and the ED1 signal, which is consistent with findings from various studies in patients and animals [
8,
18,
37,
38]. An alternative explanation for the observed lower correlation may stem from the timing of the conducted immunohistochemistry. Silvola et al. have recommended utilizing LDLR
−/−ApoB and IGF-II/LDLR
−/−ApoB mice that are at least 6 months old prior to imaging. This recommendation is based on their observation of highly inflamed, large, and extensive atherosclerotic plaques with the highest FDG uptake occurring in plaques characterized by a high macrophage density [
39]. It is conceivable that the time point of 51 ± 1 weeks might not be optimal for identifying plaques with a high macrophage density in
apoE−/− rats. Considering the relatively slow plaque development in our study, it is plausible that the opportune time point for such plaques may fall later than the 52-week mark.
Another noteworthy detail in our data is the increasing [
18F]FDG uptake over time observed in the control rats. This effect has been reported previously [
40] and is most likely associated with the advanced age of the control rats combined with the Western diet [
29].
For the PET analysis, we selected two volumes of interest (VOIs) located on different sections of the aorta. The aortic arch was chosen due to its curvature and the presence of various bifurcations, which cause disturbed shear stress. Disturbed shear stress is known to play a role in the pathophysiology of atherosclerosis and is associated with atheroma formation near bifurcations and curvatures [
41]. The second VOI was positioned on the abdominal aorta directly below the diaphragm. Although this section lacks a curve, it is rich in bifurcations. Additionally, the distance between the heart and this part of the aorta is sufficient to avoid spillover effects.
In this study, we exclusively used [
18F]FDG as the PET tracer. The choice of [
18F]FDG was based on its association with vascular inflammation [
42,
43], the extensive use of [
18F]FDG PET imaging in atherosclerosis studies, and its status as the most commonly used and clinically available PET tracer. Imaging with other tracers such as
18F-Sodium Fluoride (micro-calcification) [
44] and
11C/
18F-choline (for macrophages in atherosclerotic plaques) [
45] could yield different conclusions due to variations in uptake pathophysiology.
Bile acids and neutral sterols play a significant role in the development of dyslipidemia. Breuninger et al
. reported positive associations between fecal bile acids and markers of dyslipidemia, as well as between fecal cholesterol and hypertriglyceridemia [
46]. In our study, we did not observe a significant difference in neutral sterols between the two strains, but higher levels of bile acids were found in
apoE−/− rats, suggesting that the disparity in lipid metabolism in
apoE−/− rats extends to bile acids.
Limitations
Due to technical issues with sample preservation and preparations, we were limited to ORO staining in the aortic arch and ED1 staining in the abdominal aorta only. Additional histological analyses (smooth muscle cell activation, endothelial cells, fibroblasts, microcalcification) in the aortic tissue or immunohistochemistry in other tissues would have provided a more comprehensive characterization of the model.
For this study, we based our scanning procedure partly on the suggestion by Rudd et al. [
36]. We are aware that the choices in the procedure (non-fasting, isoflurane anesthesia, interval between injection and scan) could have an impact on the physiological uptake of [
18F]FDG. It is therefore possible that the chosen protocol may not be optimal for atherosclerotic imaging in this model. We suggest an additional study to explore the best scan procedure for [
18F]FDG PET/CT imaging of atherosclerosis in this model.
Secondly, it can be debated that the time points in this study are not the best moments to image inflammation in atherosclerotic plaques. To address this, we suggest an additional study with histological analysis at each imaging time point, alongside an assessment of the correlation between [18F]FDG signal and inflammation.
Although the apoE−/− rat is a promising model for atherosclerotic PET/CT imaging, the model still has some disadvantages which were also observed in apoE−/− mice. Therefore, it would be of great interest to explore other genetic models for atherosclerosis, like the LDLR−/− rat model.
Our studies show that the apoE−/− rat is a viable model for longitudinal atherosclerotic PET imaging. This model exhibits hypercholesterolemia, increased bile acids and triglycerides, leading to the formation of early atherosclerotic lesions accompanied by macrophage accumulation. The gradual progression of the disease in this model resembles the rather slow development of atherosclerosis in the human condition, making it particularly suitable for investigating early stages of atherosclerosis in a non-invasive manner.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.