Background A swine model of carotid atherosclerosis may greatly facilitate the identification of imaging characteristics of vulnerable plaques and the preclinical evaluation of endovascular intervention. In this study we assess the association of matrix metalloproteinase (MMP)-9 expression and neovascularity in carotid atherosclerotic plaques with MRI patterns in a swine model.
Methods Carotid atherosclerosis models were created in miniswine using a combination of partial ligation and a high cholesterol diet. The animals were imaged in a 1.5 T MR scanner at 3 months and carotid arteries were obtained for histopathological and immunohistochemical examination. Contrast-enhanced T1-weighted imaging (T1WI) was used to match the histology findings. The contrast-to-noise ratio (CNR) of the plaques on T1WI and contrast-enhanced T1WI were measured and the association of MMP-9 expression and neovascularity in the carotid plaque with CNR on MRI was analyzed.
Result Forty carotid artery segments were matched between MRI and histology. All segments were advanced carotid atherosclerotic plaques. The matched contrast-enhanced T1WI and histology slices showed good correlation for ratio of plaque size to lumen diameter (r=0.94, p<0.001). Plaque CNR on contrast-enhanced T1WI was higher in plaques with strong MMP-9 expression than in those with weak MMP-9 expression (p=0.05). Plaque CNR on contrast-enhanced T1WI was also higher in plaques with marked neovascularization than in those without (p=0.02).
Conclusions Increased plaque CNR on contrast-enhanced T1WI is associated with MMP-9 expression and neovascularization in carotid atherosclerotic plaques and may be used to identify vulnerable plaques.
Statistics from Altmetric.com
Atherosclerosis is the underlying cause of the majority of cardiovascular and cerebrovascular events.1 Ischemic stroke frequently occurs in patients with carotid stenosis and vulnerable atherosclerotic plaques. Treating all patients with severe asymptomatic carotid stenosis requires a high number of patients needed to treat to prevent one major ischemic stroke. Identifying patients at greater risk for stroke in asymptomatic patients remains a clinical challenge. MRI evaluation of carotid atherosclerotic plaques may help identify those patients with unstable plaques that are prone to rupture leading to acute thrombosis of the artery.
In previous studies, vulnerable atherosclerotic plaques with distal atheroembolism were created in swine with combined dietary hyperlipidemia and partial surgical ligation.2 ,3 Increased expression of matrix metalloproteinase (MMP)-9 associated with intraplaque hemorrhage was also shown in this swine model.4 This swine model may thus be useful to improve the diagnostic and therapeutic tools for carotid disease. However, the feasibility of detection of carotid atherosclerotic plaque by MRI in this swine model is unclear.
Neovascularity and intraplaque hemorrhage contribute to the progression and rupture of atherosclerotic plaques.5 ,6 Marked neovasculization has been shown in atherosclerotic plaques from symptomatic patients compared with those from asymptomatic patients.7 ,8 MMP-9 may contribute to neoangiogenesis, induce the proliferation and migration of macrophages and smooth muscle cells as well as the degradation of extracellular matrix, and subsequently enhance the risk of plaque rupture and intraplaque hemorrhage.9 In this study we assess the association of MMP-9 expression and neovascularity in carotid atherosclerotic plaques with MRI patterns in this swine model.
Materials and methods
The experimental protocol was approved by our institutional Animal Research Committee in accordance with policies set by the National Institutes of Health guidelines. Five miniswine aged 10–12 weeks and 15–20 kg body weight were used to develop carotid atherosclerosis using the combination of partial surgical ligation and hyperlipidemia. All swine were fed with a high fat and high cholesterol diet containing 2% cholesterol, 10% yolk powder, 8% lard and 1% supplemental choline to induce hyperlipidemia. Two weeks later the animals were selected for surgery if the atherogenic diet resulted in a more than fourfold increase in their plasma cholesterol levels. The swine were maintained on this atherogenic diet for 12 weeks after surgery.
The animals were anesthetized with an intramuscular injection of midazolam 5–10 mg and an intravenous infusion of 3% pentobarbital 10 mg/kg/h. The bilateral common carotid arteries were exposed and dissected approximately 5 cm in length through a midline skin incision at the neck. The carotid arteries were ligated with a 4/0 silk suture along with a 2F microcatheter segment, which was subsequently pulled out from inside the suture to leave a small lumen in the artery. Bilateral carotid arteries in each animal were partially ligated to create approximately 70% stenosis. Doppler ultrasonography of the carotid arteries was obtained to confirm the degree of stenosis and patency of the carotid arteries immediately after ligation and at the time of sacrifice.
The swine were anesthetized and placed in a supine position 12 weeks after surgery. MRI examinations were performed in a 1.5 T MR scanner (Achieva 1.5 T, Philips Medical Systems) with a four-channel head-neck array coil. Cross-sectional images of the bilateral common carotid artery were obtained through the aortic root with a slice thickness of 4.0 mm. The image parameters included: TR/TE 950/15 ms, field of view (FOV) 205 mm, flip angle 90°, matrix 348×222, average number of 3 for T1-weighted imaging (T1WI) sequence; TR/TE 3950/100 ms, FOV 250 mm, flip angle 70°, matrix 436×324, average number of 2 for T2-weighted imaging sequence; and TR/TE 1800/30 ms, FOV 230 mm, flip angle 90°, matrix 256×207, average number of 3 for proton density-weighted imaging sequence. Contrast-enhanced T1WI sequence was performed 5 and 30 min after intravenous injection of gadopentetate dimeglumine at a dose of 0.2 mmol/kg body weight and a rate of 3 ml/s.
The swine were killed with an overdose injection of 3% pentobarbital and perfused with 10% formalin 12 weeks after MR examinations. Bilateral common carotid arteries were collected and fixed in 10% formalin. The carotid arteries were cut into 4 mm segments beginning from the aortic root, embedded in paraffin, sectioned at 5 μm thickness and stained with H&E to determine the presence of atherosclerotic lesions. The atherosclerotic changes of these carotid arteries were classified using Stary stage types I–VI set out by the American Heart Association classification of coronary atherosclerosis.10 In brief, type I lesions are characterized by the presence of smooth muscle proliferation within the intima; type II lesions have intimal proliferation with foamy macrophages; type III lesions contain small pools of extracellular lipid; type IV lesions have extracellular lipid cores; type V lesions have type IV features plus fibrous thickening; and type VI lesions are characterized by surface defects, intraplaque hemorrhage or surface thrombus. Types IV, V and VI lesions were defined as advanced plaques. The following features of plaque vulnerability were recorded: large lipid core, marked foam cells, marked inflammation, plaque rupture, thin fibrous cap, intraplaque hemorrhage, surface thrombus and marked neovascularization. Neovascularization was defined as the presence of at least 10 new vessels in the plaque per section on histological examination.11
Serial paraffin-embedded cross-sections of carotid arteries were used for immunostaining according to the streptavidin-biotin complex technique. They were incubated with rabbit polyclonal MMP-9 antibody at 1:100 dilution. Controls in the absence of primary antibodies were also performed.
The immunostains were examined and images were acquired by high-resolution digital color camera. Digital images were analyzed using Image-Pro Plus V.5.0 software (Media Cybernetics, Bethesda, Maryland, USA). The percent of MMP-9 immunopositive area (immunopositive area/total intimal area×100) was determined by averaging several images per section that covered all the plaque regions. A strong expression of MMP-9 in the carotid plaque was recorded if the value of the MMP-9 positive area was more than the mean value of the MMP-9 positive area in this group.
MRI data analysis
The segments of carotid artery were started at the aortic root to facilitate co-localization between the MR findings and histological specimen. T1WI and contrast-enhanced T1WI were used to match the histology findings based on morphological features such as the shape of the atherosclerotic plaque and vessel lumen. Because of carotid artery shrinkage occurring after formalin fixation, only plaque size and lumen diameter measurements were performed on contrast-enhanced T1WI and H&E-stained sections. The ratio of plaque thickness to lumen diameter measured by MR and histology were compared.
The plaque region with a significant enhancement on the contrast-enhanced T1WI and the corresponding region on the T1WI were selected as the region of interest for the following measurement. The signal intensity (SI) of atherosclerotic plaque and adjacent muscle within the matched vessel segments on the T1WI and contrast-enhanced T1WI were measured. The noise was obtained from the standard deviation of air signal located outside the swine body. The signal-to-noise ratio (SNR) was calculated by dividing the plaque SI by the noise and the contrast-to-noise ratio (CNR) was calculated by dividing the difference of plaque segment SI and adjacent muscle SI by the noise.12 ,13 CNR and SNR of the carotid atherosclerotic plaques were measured on both the T1WI and contrast-enhanced T1WI. All signal measurements were performed on a workstation equipped in the MR scanner (Achieva 1.5 T). Two investigators who were blinded to the histological results performed the MRI assessment.
The correlation for ratio of plaque thickness to lumen diameter between MR and histological findings was evaluated with the Pearson coefficient correlation. The association of MMP-9 expression in the plaque with the value of CNR and SNR on MRI in the matched atherosclerotic plaques was compared. Continuous data were assessed for normality by the Kolmogorov–Smirnov test, normally distributed continuous data were analyzed by the Student t test and unevenly distributed continuous data were analyzed by the Mann–Whitney U test. SPSS software was used to perform the analysis; p values <0.05 were considered statistically significant.
Histology of carotid atherosclerotic plaques
Ten carotid atherosclerotic artery models in five swine were developed and scanned by MR before sacrifice. Forty carotid artery segments with advanced atherosclerotic plaque (Stary IV–VI) were matched between MRI and the histological study. There was one segment with Stage IV plaque, 13 segments with Stage V plaques and 26 segments with Stage VI plaques. Among the 40 carotid plaques we observed a large lipid core in 25 segments (62.5%), an abundance of foam cells in 34 (85.0%), intraplaque hemorrhage in 20 (50.0%), marked plaque inflammation in 40 (100%), cap rupture in 13 (32.5%), marked neovascularization in 26 (65.0%) and calcification in 5 segments (12.5%). Twenty-three segments were considered as thin fibrous cap plaques with the minimal thickness of fibrous cap <200 μm. The typical features of vulnerable atherosclerotic plaque are shown in figure 1.
Positive staining of MMP-9 was mostly distributed around the plaque core, especially in the plaque shoulder and fibrous cap. There were concentrated foam macrophages and marked neovascularization in the area of positive staining. Two carotid segments were not used for the analysis because of poor quality MMP-9 staining. Strong expression of MMP-9 was found in 15 carotid plaques while weak MMP-9 expression was found in the remaining 23 carotid plaques. Areas positive for MMP-9 expression were greater in carotid plaques with marked neovascularization (9.88±7.75% vs 5.15±5.21%; p=0.05) and in those with intraplaque hemorrhage (10.4±8.0% vs 5.89±5.69%; p=0.04).
MRI of atherosclerotic plaques
A crescent- or irregularly-shaped abnormal signal within the vessel lumen was found on the axial section of MRI, which represented as an atherosclerotic plaque. The typical presentation of a carotid atherosclerotic plaque was shown as an isointense or slight hypointense signal on the T1WI and proton density-weighted imaging. They had a varying degree of enhancement signal on the contrast-enhanced T1WI.
Table 1 shows the average thickness of the atherosclerotic plaques and the diameter of the vessel lumen on H&E-stained sections and contrast-enhanced T1WI, respectively. The matched contrast-enhanced MRI and histology slices showed good correlation for the ratio of plaque thickness to lumen diameter (r=0.94, p<0.001).
Plaque CNR, SNR and MMP-9 expression
Table 2 shows the average SNR and CNR on T1WI and contrast-enhanced T1WI, respectively. The average difference of plaque SNR between contrast-enhanced T1WI and T1WI was 20.9±16.9 and the average difference of plaque CNR between contrast-enhanced T1WI and T1WI was 13.7±13.0.
CNR of plaque on the contrast-enhanced T1WI was higher in 15 carotid plaques with strong MMP-9 expression than in 23 plaques with weak MMP-9 expression (25.5±16.5 vs 16.3±11.7, p=0.05; figure 2). CNR of plaque on contrast-enhanced T1WI was also higher in 26 carotid plaques with neovascularity than in 14 plaques without neovascularity (23.8±16.0 vs 12.7±5.0; p=0.02).
There were no associations between strong MMP-9 expression or neovascularity and CNR on T1WI, SNR on T1WI and SNR on contrast-enhanced T1WI. There were also no associations between intraplaque hemorrhage and CNR or SNR on MRI.
This study shows that this swine carotid model has many morphological features of vulnerable atherosclerotic plaques which are similar to the pathological findings of carotid plaques in symptomatic patients with carotid stenosis. The vulnerable carotid atherosclerotic plaques in this swine model can be identified by MRI. MMP-9 expression and neovascularity in the carotid plaque can be detected by contrast-enhanced MRI. The value of plaque CNR on contrast-enhanced T1WI increased in carotid plaques with strong MMP-9 expression and marked neovascularization. These results support the notion that CNR is a marker of active vessel remodeling, which occurs commonly in vulnerable plaques.
MRI can be used to differentiate atherosclerotic plaque compositions and to track atherosclerotic lesion progression and response to treatment.14 ,15 The feasibility of detecting carotid atherosclerosis by MRI has been shown in previous animal model studies. Lin et al16 performed contrast-enhanced MR angiography in a balloon-injured swine model to detect carotid arteries containing atherosclerotic-like lesions. They suggested an accumulation of contrast agent in the enhanced carotid artery wall because of an increased vascular supply associated with thrombosis and neointimal thickening. Ma et al17 found that carotid atherosclerotic plaques in a rabbit model induced with a high-lipid diet following balloon-injured endothelium can be identified by a 1.5 T clinical MR scanner. Increased SI on T1WI was detected in a higher degree of atherosclerotic plaques. However, vulnerable features of plaque rupture and intraplaque hemorrhage were not found in this rabbit model. We found that advanced carotid atherosclerotic plaques in this swine model can be detected on the contrast-enhanced T1WI sequence. Advanced atherosclerotic plaques have been induced in our swine model whereas early atherosclerosis lesions were mostly induced in previous animal models. In a recent clincal study, Millon et al18 found that gadolinium enhancement of carotid plaque on MRI was significantly more frequent in symptomatic than in asymptomatic patients. Gadolinium enhancement was associated with vulnerable plaque, neovascularization and macrophages on histological analysis.
The present study shows that the value of CNR on the contrast-enhanced T1WI was associated with MMP-9 expression and marked neovascularization. The value of CNR may be used as a marker of plaque vulnerability on MRI. Zheng et al12 found that the neointima areas in a rabbit femoral artery atherosclerosis model appeared to be significantly more enhanced than the fibrotic connective tissue and non-injured arteries after administration of contrast agent. Increased CNR of femoral plaque was also shown on the contrast-enhanced MRI. Amirbekian et al19 found that MRI enhancement in atherosclerotic lesions was produced after MMP-targeted contrast agent injection in atherosclerotic apolipoprotein E-deficient mice. MMP-targeted contrast agents facilitated a 93% increase in the value of CNR of atherosclerotic aortas. Co-localization between fluorescent contrast agents and MMP staining in atherosclerotic plaques was found. Fluorescent contrast agents were particularly present in the fibrous cap of plaques. Hyafil et al20 found that this MMP-targeted contrast agent can be used to detect arterial wall remodeling of atherosclerotic aortas in a rabbit model.
Our result in the swine model is consistent with previous studies of the association between plaque neovascularity and increased CNR and the extent of enhancement on MRI in human specimens and animal models. Kerwin et al21 showed that the extent of neovasculature within the carotid atherosclerotic plaque can be detected by contrast-enhanced MRI. Sirol et al22 found that advanced plaques and early atheroma were both enhanced on MRI after injection of contrast agent in a rabbit model with aortic atherosclerotic plaque. However, the average CNR value was significantly higher in advanced plaques than in early atherosclerotic lesions. Histological analysis demonstrated an association between neovessel density and plaque enhancement on MRI.
This study has several limitations. The findings of increased plaque SNR on the contrast-enhanced T1WI in our swine model is consistent with the study by Koktzoglou et al.13 However, no association was found between MMP-9 expression or neovascularity and SNR on contrast-enhanced T1WI owing to the small sample size of animal models and carotid plaques. We performed the study on a 1.5 T clinical MR scanner equipped with a head-neck array coil but not a specific volumetric array coil for animal study. High field MRI or dedicated surface array coils for carotid imaging may generate better resolution and SNR, allowing better characterization of atherosclerotic plaques.
Increased plaque CNR on the contrast-enhanced T1WI is associated with MMP-9 expression and neovascularization in the carotid atherosclerotic plaques and may be used to identify vulnerable plaques. High resolution MRI technique for assessment of the composition of the plaque in this swine model is warranted.
We thank Fernando Viñuela and Lei Feng, Division of Interventional Neuroradiology, University of California, Los Angeles, for advice to perform this study and critical reading of this manuscript.
XBJ and WSY contributed equally.
Contributors Guarantor of integrity of entire study: ZSS. Study concepts and study design: ZSS. Data acquisition, data analysis and interpretation: all authors. Manuscript drafting and manuscript revision for important intellectual content: ZSS, XBJ, WSY. Manuscript editing: ZSS, XBJ. Manuscript final version approval: all authors.
Funding This work is supported by the National Natural Science Foundation of China (81070949), Program for New Century Excellent Talents in University of China (NCET-11-0539) and the Fundamental Research Funds for the Central Universities, Sun Yat-sen University (09ykpy38) (to ZSS).
Competing interests None.
Provenance and peer review Not commissioned; externally peer reviewed.