Introduction

Endothelial dysfunction and increased endothelial permeability play an important role in all stages of atherogenesis, from disease initiation, to the final stages of plaque hemorrhage and thrombosis, leading to major cardiovascular events (1,2). The development and validation of non-invasive MR imaging techniques to quantify these parameters in vivo may greatly improve our current ability to stratify the risk of cardiovascular event associated with specific plaque phenotypes and is currently an active area of research (3). Here, we investigate the quantification of endothelial permeability in the aortic root of atherosclerotic mice using 3D, black blood self-gated T1 mapping (4,5) before and after injection of a gadolinium (Gd) based contrast agent.

Methods

Animal model and study design. Seven-week-old male ApoE -/- mice (n= 14) were fed a high-fat diet (HFD) for 8 weeks, after which animals were randomized into two groups: 1) controls (n=8), continuing on HFD for 16 weeks; 3) statin treated (n=6), fed medicated HFD with atorvastatin (10mg/kg/day) for 16 weeks. All animals were imaged before HFD initiation (0W time point), and at end of study (24W time point). Image acquisition. Imaging was performed on a 7T preclinical MRI scanner. Endothelial permeability was quantified using a 3D, self-gated gradient recalled echo (GRE) sequence with variable flip angles (2, 5, 8, 11 and 14°) for T1 mapping, before and ~30 minutes after injection of 0.3 mmol/Kg of gadopentetate dimeglumine. Imaging parameters were: slice thickness 0.3mm with a spatial resolution of 0.117x0.117mm2, echo time 1.7ms, repetition time 7.6ms defining a 6 cardiac cycles sequence with 20 slices per cardiac cycle. Image analysis. For each flip angle series acquired before and after contrast agent injection, an experienced observer traced the inner and outer boundary of the aortic root (Figure 1), and a region of interest (ROI) in the background noise. Aortic root ROI signal intensity at each flip angle acquisition was used to calculate R1 (s-1) values before and after contrast agent injection using the DESPOT method (6). Aortic root signal to noise ratio (SNR) was calculated both pre and post contrast agent injection, by dividing ROI signal in the vessel wall by noise standard deviation. Pre and post contrast R1 values (R1pre and R1post, Figure 3), and the ratios R1post/R1pre (R1_Ratio, Figure 3) and SNRpost/SNRpre (SNR_Ratio, Figure 3) were compared within each group of mice between the 0 and 24 weeks time points, and at 24 weeks between statin-treated and untreated mice. Statistical analyses were performed, using paired Student test to compare R1pre, R1post , R1_Ratio, and SNR_Ratio within each group between the 0 and 24 weeks time points. Unpaired Student t tests were used to compare the same variables at 24 weeks between statin-treated and untreated mice. P values <0.05 were used to define statistical significance. Significant differences are indicated in the Figures with a star (*).

Results

Post-contrast R1 values were consistently higher than pre-contrast values for all groups, and at all time points (Figure 2), indicating extravasation of contrast agent in the aortic root. Post to pre-contrast SNR ratio, and pre-contrast R1 were not significantly different between 0 and 24 weeks for both groups. Post-contrast R1 (p=0.0102 for HFD and p=0.0255 for statin) and post to pre-contrast R1 ratio (p=0.0053 for HFD and p=0.0059 for statin) were significantly different between 0 and 24 weeks for both HFD and statin groups (Figure 3), indicating higher contrast agent extravasation in older mice. At 24 weeks SNR ratio and pre-contrast R1 were similar between HFD and statin treated group, while post-contrast R1 and post to pre contrast R1 ratio were lower in the statin treated group (Figure 4), indicative of lower contrast agent extravasation and endothelial permeability, although the difference was not statistically significant.

Discussion

This study suggests that 3D, self-gated, black blood T1 mapping after contrast agent injection is a promising technique to measure endothelial permeability in the aortic root of ApoE-/- atherosclerotic mice. Other studies in the literature investigating permeability in the mouse brachiocephalic artery, have used prospectively ECG triggered, bright blood acquisitions (7,8). Compared to these approaches, the technique that we propose allows for better vessel lumen/wall delineation of the mouse aortic root throughout the cardiac cycle, thereby allowing to quantify permeability in this challenging and important vascular territory. Compared to studies in humans, or larger animal models, we foresee that developing and validating the quantification of endothelial permeability by MRI in atherosclerotic mice will allow us to investigate in greater detail its relationship with the genetic, molecular and cellular make up of atherosclerotic plaques during the various stages of disease development, and after therapeutic intervention.

Acknowledgements

The authors would like to acknoledge the Société Française de Chirurgie Vasculaire et Endovasculaire for the support of a part of the study.