Introduction

Atherosclerosis is the leading cause of disease and mortality in most developed countries1. Although age-related deaths attributable to atherosclerosis are declining, in 2019, approximately 18 million people worldwide died from cardiovascular diseases, primarily coronary and cerebrovascular atherosclerosis. Atherosclerosis has several types: an intracranial lesion called intracranial atherosclerotic disease (ICAD), and an extracranial lesion called internal carotid artery stenosis (ICS). When the vascular intima-media is damaged, atherosclerosis develops due to plaque formation, resulting in changes in blood flow and cross-sectional area2. The evaluation of vascular characteristics is important for understanding the pathogenesis of atherosclerosis. Despite advances in our understanding of the pathophysiology, pathogenesis, and new treatment modalities, the absence of an adequate non-invasive imaging tool for early detection limits the prevention and treatment of patients with various degrees and anatomical localizations of atherothrombotic disease. An ideal clinical imaging modality for atherosclerotic vascular disease should be safe, inexpensive, non-invasive, minimally invasive, accurate, and reproducible. Moreover, the imaging results should correlate with the extent of atherosclerotic disease and have a high predictive value for future clinical events.

Ultrasonography (US) is a versatile diagnostic tool for arteriosclerosis3,4. Although US is simple and enables real-time observation, it is limited to the evaluation of temporal changes in a single cross-section. Magnetic resonance imaging (MRI) is the most promising technique for studying atherothrombotic diseases in humans5,6. MRI has various tissue-contrast imaging techniques and allows for the characterization of plaque composition, including the discrimination of the lipid core, fibrosis, calcification, and intraplaque hemorrhage deposits7. Therefore, MRI is considered the gold standard for non-invasive evaluation of carotid artery plaque morphology and composition. In addition to measuring plaque formation, MRI can be used to assess changes in vascular blood flow and wall shear stress (WSS), providing useful diagnostic information on the pathogenesis of atherosclerosis. 4D-flow MRI can quantitatively evaluate temporal changes in blood vessels in three dimensions and multiple cross-sections and can also evaluate basic factors related to arterial stiffness, such as WSS8,9. Several animal studies have demonstrated the usefulness of 4D-flow MRI in evaluating rodent models10,11,12.

The purpose of this study is to perform 4D-flow MRI on rat carotid arteries to evaluate changes in flow velocity, WSS, and vessel cross-sectional area due to endothelial injury in rats used as a model of atherosclerosis.

Materials and methods

Animal preparation

All experimental protocols were approved by the Research Ethics Committee of our University. All experimental procedures involving animals and their care were performed in accordance with the Osaka University Guidelines for Animal Experimentation and the National Institutes of Health Guide for the Care and Use of Laboratory Animals. Sprague-Dawley rats (10–12 weeks old) were purchased from Japan SLC (Hamamatsu, Japan). Seven control rats and 16 rats with right endothelial injury (VED) were used in this study. All rats were housed in a controlled vivarium environment (24 °C; 12:12 h light/dark cycle) and were fed a standard pellet diet and water ad libitum. Rat carotid artery balloon injury models (right carotid artery endothelial injury) were generated as follows: (1) The rat was given isoflurane 1.5 to 2.0% inhalation anesthesia using an inhalation anesthetic device. (2) The rat was placed in a supine position with the treated side up. (3) The skin near the right margin of the laryngeal cartilage was incised, and the external carotid artery (ECA) and common carotid artery (CCA) were dissected. (4) A balloon catheter was inserted into the CCA through an incision in the dissected ECA. (5) The balloon catheter was inserted into the CCA, the incision was incised, and a balloon catheter was inserted into the CCA. 10 µL of the balloon was inflated and passed through the artery one time to disrupt the endothelium, the balloon catheter was removed, the ECA was ligated and the incision was sutured. (6) An antibiotic (enrofloxacin) was administered subcutaneously at a dose of 5 mg/kg. (7) An analgesic (buprenorphine) of 0.05 mg/kg was administered subcutaneously once postoperatively.The time interval between the balloon injury and MR examination was 1 week.

MRI

MR images of the animal necks were acquired using a horizontal 7-T scanner (PharmaScan 70/16 US; Bruker Biospin, Ettlingen, Germany) equipped with a 60 mm inner diameter volume coil. The rats were positioned in stereotaxic frames with their mouths secured to prevent movements during acquisition13,14. The body temperature was maintained at 36.5 °C with regulated water flow and was continuously monitored using a physiological monitoring system (SA Instruments Inc., Stony Brook, NY, USA). All MR experiments were performed under general anesthesia using isoflurane (3.0% for induction and 2.0% for maintenance).

Morphological information was acquired using 3D time-of-flight magnetic resonance angiography (MRA) with the following parameters: pulse sequence, flow-compensated 3D gradient-echo sequence; TR, 12 ms; TE, 2.6 ms; flip angle (FA), 20°; field-of-view, 32 × 32 × 32 mm3; matrix size, 200 × 200 × 200; number of averages, 1; slice thickness, 32 mm; slab orientation, trans-axial; spatial resolution, 160 μm × 160 μm × 160 μm; scan time, 6 m 37 s. 4D-flow images were acquired using time-resolved 3D phase contrast angiography under ECG gating with the following parameters: pulse sequence, flow-compensated gradient-echo sequence with bipolar gradients; TR,12 ms; TE, 2.9 ms; flip angle (FA), 30°; field-of-view, 32 × 32 × 32 mm3; matrix size, 150 × 150 × 25 mm3; number of averages, 2; slice thickness, 5 mm; number of frames, 10; slab orientation, trans-axial; spatial resolution, 200 μm × 200 μm × 200 μm; scan time, around 2 h; VENC, 125 cm/s. Flow velocities along the three orthogonal spatial directions were calculated using four velocity-encoding gradient steps based on a Hadamard encoding scheme.

MRI data analysis

From the MRI image, the CCA was defined from the aortic arch bifurcation to the internal and external carotid bifurcations. The CCA was divided into three parts: proximal, middle, and distal, starting from the portion closest to the heart (Fig. 1). The dotted line indicates the area of each part and the area’s vertical width is about 1 cm. The values of the flow rate and cross-sectional area, we draw a line through the center of the area and measured the average value at the centerline of each part. The WSS was measured the average value from the entire area of each part. All values were calculated using a fluid analysis software iTFlow@ (Cardio Flow Design Inc. Tokyo, Japan).

Fig. 1
Fig. 1
Full size image

Definition of parts in MRA MIP image of control rat. The common carotid artery is divided into three parts: proximal, middle, and distal, starting from the section closest to the heart. The dotted lines indicate the areas of each part. The average values of the flow rate and cross-sectional area were measured at the centerline of each part, and the wall shear stress was measured from the entire area of each part.

Statistical analysis

All data are presented as mean ± standard deviation. Differences were compared using one-way ANOVA followed by Tukey’s post-hoc test to evaluate all parameters (velocity, WSS, and vessel area). All analyses were performed using Prism 8 software (GraphPad Software, San Diego, CA, USA). P < 0.05 was considered statistically significant.

Results

The average body weight of the control rats was 353.1 ± 23.4 g, while the average body weight of the VED model rats was 339.2 ± 5.0 g. There was no significant difference between the average body weights of the control and VED model rats.

Figure 2 shows the maximau intensity projection (MIP) images of the control (Fig. 2a) and VED model (Fig. 2b) rats. As indicated by the white arrow, the right CCA of the VED model was clearly thinner and narrower than that of the control. Figure 3 shows the average cross-sectional areas. In the control group, the cross-sectional area significantly decreased from proximal to distal on both sides ([left] proximal:1.3 ± 0.3 mm2, distal: 0.0 ± 0.2 mm2, p < 0.05 [right] proximal: 1.4 ± 0.3 mm2, distal: 0.9 ± 0.3 mm2, p < 0.01, Fig. 3a). A similar trend was observed in the VED model ([left] proximal: 2.1 ± 0.6 mm2, distal: 1.1 ± 0.4 mm2, p < 0.01 [right] proximal: 1.5 ± 0.5 mm2, distal: 0.6 ± 0.2 mm2, p < 0.05, Fig. 3b). The value of left proximal CCA in VED model significantly improved 60% (control: 1.3 ± 0.3 mm2, VED model :2.1 ± 0.6 mm2, p < 0.05).

Fig. 2
Fig. 2
Full size image

MRA MIP images of (a) control and (b) VED model rats. The white arrow indicates that the right common carotid artery of the VED model was thinner and narrow compared to that of the control group.

Fig. 3
Fig. 3
Full size image

Cross-sectional area. The cross-sectional area of the control group was significantly reduced from proximal to distal ([left] proximal:1.3 ± 0.3 mm2, distal: 0.0 ± 0.2 mm2, p < 0.05, [right] proximal: 1.4 ± 0.3 mm2, distal: 0.9 ± 0.3 mm2, p < 0.01). A similar trend was observed in the VED model ([left] proximal:2.1 ± 0.6 mm2, distal: 1.1 ± 0.4 mm2, p < 0.01, [right] proximal: 1.5 ± 0.5 mm2, distal: 0.6 ± 0.2 mm2, p < 0.05). The value of proximal of light common carotid artery was significantly different between control and VED model (control: 1.3 ± 0.3 mm2, VED model: 2.1 ± 0.6 mm2, p < 0.05). *p < 0.05, **p < 0.01.

Figure 4 shows the 4D-flow streamline images of the control (Fig. 4a) and VED model (Fig. 4b) rats. As seen in the velocity maps derived from 4D-flow imaging, the velocity of the right CCA in the VED model was lower than in the control group (Fig. 4b). Figure 5 shows the average values of the mean flow rate. In the control group, there was no significant left-right difference in flow velocity (Fig. 5a). While the flow rate of the right proximal CCA was significantly decreased compared to the flow rate of the left proximal CCA (L: 46.0 ± 17.8 µL/beat, R: 20.0 ± 7.8 µL/beat, p < 0.05, Fig. 5b), the value of the right CCA was reduced by 38% compared to that of control rats in the VED model (control: 32.1 ± 7.9 µL/beat, VED model: 20.0 ± 7.8 µL/beat, p < 0.05).

Fig. 4
Fig. 4
Full size image

4D-flow streamline images of (a) control and (b) VED model. The endothelial injury was made over the entire right common carotid artery in VED model. The velocity of the right common carotid artery in the VED model (b) was lower than that in the control (a).

Fig. 5
Fig. 5
Full size image

Flow rate. There was a significant difference between the flow rates of the right and left proximal common carotid artery in the VED model (L:46.0 ± 17.8 µL/beat, R:20.0 ± 7.8 µL/beat, p < 0.05). The flow rate of the proximal right CCA was significantly different between the control and the VED model (control: 32.1 ± 7.9 µL/beat, VED model: 20.0 ± 7.8 µL/beat, p < 0.05). * p < 0.05, unpaired t-test.

Figure 6 shows the 4D-flow WSS images of the control group (Fig. 6a) and VED rats (Fig. 6b). The WSS of the right CCA in the VED model was lower than that in the control group. Figure 7 presents the average values of mean WSS. There was no significant left-right difference in velocity in either group of rats (Fig. 7a, b). However, the WSS in right proximal CCA was significantly decreased 35% compared to that of the control group (control: 7.2 ± 1.2 Pa, VED model :4.7 ± 1.4 Pa, p < 0.05).

Fig. 6
Fig. 6
Full size image

4D-flow WSS images of (a) control and (b) VED model. The endothelial injury was made over the entire right common carotid artery in VED model. The WSS of the right common carotid artery in the VED model (b) was lower than that in the control (a).

Fig. 7
Fig. 7
Full size image

WSS (a) Control (b) VED model. Comparisons of the same model showed no significant differences between the parts. WSS of proximal of right common carotid artery was significantly different between the control and the VED model (control: 7.2 ± 1.2 Pa, VED model: 4.7 ± 1.4 Pa, p < 0.05).

Discussion

Our study is the first to demonstrate changes in flow velocity, flow rate, WSS, and cross-sectional area due to endothelial injury using 4D-flow MRI of the bilateral carotid arteries in a rat model of unilateral carotid artery injury. In the VED models, the blood flow rate and WSS were predominantly decreased on the proximal injured side compared to the ipsilateral control side. Furthermore, the cross-sectional area of the vessel distal to the injured side in the VED model was reduced compared to that of the non-injured side. A comparison of the proximal area of the uninjured side between the control group and the VED model showed that the vascular cross-sectional area was predominantly increased in the VED models. The area in which these results were seen was uniform throughout the VED models.

In this study, significant reductions in velocity and flow were observed in the proximal part of the CCA in the VED model using 4D-flow MRI. This may be attributed to the fact that the distal part of the CCA is smaller in diameter than the proximal part (Fig. 3) and that a strong stenosis occurs at the catheter insertion stage of modeling. In carotid vascular echocardiography, in humans, the maximum systolic blood flow velocity increases when carotid artery stenosis exceeds 50%, and the blood flow velocity increases as stenosis progresses15. In patients with stenosis, 4D-flow MRI was used to measure end-diastolic and systolic flow velocities in the common and internal carotid arteries; flow velocity was decreased in the CCA and increased in the internal carotid artery (ICA)16. In other words, the flow velocity increases at the site of stenosis and decreases proximal to the stenosis, a pattern observed during both end-diastole and systole. The trend of increased flow velocity at the stenosis and decreased flow velocity proximal to the stenosis is consistent with findings from previous studies. Even within the newly stenosed vessel in this study, there was a difference in flow velocity between the stenotic area and the proximal part of the stenosis, with the flow velocity being lower at the proximal part of the stenosis.

In the VED model used in this study, the WSS was also significantly lower in the proximal part of the stenotic vessel, likely due to the decreased flow velocity at the proximal end of the CCA, as mentioned above. This is because WSS is calculated by dividing the flow velocity by the vessel radius17. Furthermore, we believe that the lack of a significant change in vessel diameter, despite the significant decrease in flow rate, is another factor contributing to the lower WSS. In a previous study, after 6–12 weeks of clipping the unilateral CCA in rats, flow and vessel diameter were measured, and the clipped vessels showed an average 35% decrease in flow and a 10.2% decrease in diameter18. Alterations in WSS affect the function and shape of endothelial cells, potentially leading to the development of atherosclerosis and other vascular diseases. It is known that an increase in flow velocity results in an increase in vessel diameter, while a decrease in flow velocity reduces vessel diameter to maintain a constant WSS19. In this study, a significant increase in the cross-sectional area was observed in the proximal part of the non-stenotic vessel in the VED model. The WSS in this area was not significantly different from that of the controls. Although there was no statistically significant difference in the flow rate due to the size of the SD, there was a trend toward increased flow in the proximal part of the non-stenotic vessel in the VED model compared with the control group (Fig. 5). Because there was no significant change in WSS with increased flow, we considered that the cross-sectional area of the vessel increased with changes in flow to maintain constant WSS, consistent with previous studies.

Limitation

This study had some limitations. First, we used a VED model in which endothelial damage was induced by inserting a balloon catheter to promote endothelial thickening. However, it was difficult to control the degree of damage (severe or mild) due to the nature of the model, and there were individuals in whom blood flow could not be measured using MRI because of severe damage. Even in individuals in whom blood flow in the CCA could be confirmed, evaluating the distal portion of the carotid artery was challenging data variability increased as the evaluation moved farther distally. Although the VED model used in this study is widely accepted as a model of carotid artery stenosis, it has been noted that creating a stable model remains difficult. Various alternative models of carotid artery stenosis and endothelial thickening have been proposed, such as those involving hypertension or high-fat diet feeding. However, it is necessary to validate findings using these other models in future studies.

Second, the flow velocity was calculated based on the mean flow velocity per cardiac cycle. Many previous studies on human subjects measured flow velocities during the end-diastole and systole phases of the cardiac cycle. While the same evaluation was performed in this study, maintaining a constant heart rate was challenging due to the extended imaging time and the effects of anethesia. Consequently, evaluation was conducted in units of one cardiac cycle.

Finally, carotid screening in clinical practice typically involves US to measure the rate of stenosis caused by endothelial thickening of the carotid artery for diagnostic purposes. In this study, we attempted to confirm the degree of endothelial thickening by excising the CCA and staining it with HE. However, identifying the stenotic area and calculating the stenosis rate proved difficult.

In the future, we plan to review imaging time, anesthesia management, and other influencing factors to calculate and evaluate values for each parameter during end-diastole and systole. Additionally, we intend to conduct further research using alternative models to enhance clinical applicability.

Conclusion

4D-flow MRI with 7T-MRI enabled the analysis of changes in flow velocity, flow rate, WSS, and cross-sectional area caused by vascular endothelial injury in a rat model of unilateral carotid artery injury.