Wall shear stress analysis using 17.6 Tesla MRI: A longitudinal study in ApoE-/- mice with histological analysis

This longitudinal study was performed to evaluate the feasibility of detecting the interaction between wall shear stress (WSS) and plaque development. 20 ApoE-/- mice were separated in 12 mice with Western Diet and 8 mice with Chow Diet. Magnetic resonance (MR) scans at 17.6 Tesla and histological analysis were performed after one week, eight and twelve weeks. All in vivo MR measurements were acquired using a flow sensitive phase contrast method for determining vectorial flow. Histological sections were stained with Hematoxylin and Eosin, Elastica van Gieson and CD68 staining. Data analysis was performed using Ensight and a Matlab-based “Flow Tool”. The body weight of ApoE-/- mice increased significantly over 12 weeks. WSS values increased in the Western Diet group over the time period; in contrast, in the Chow Diet group the values decreased from the first to the second measurement point. Western Diet mice showed small plaque formations with elastin fragmentations after 8 weeks and big plaque formations after 12 weeks; Chow Diet mice showed a few elastin fragmentations after 8 weeks and small plaque formations after 12 weeks. Favored by high-fat diet, plaque formation results in higher values of WSS. With wall shear stress being a known predictor for atherosclerotic plaque development, ultra highfield MRI can serve as a tool for studying the causes and beginnings of atherosclerosis.


Introduction
Atherosclerosis is a pathological chronic inflammatory process in the arterial wall especially of the endothelial cells affecting cardiovascular disease [1]. Human atherosclerotic plaque formation in the coronary arteries is associated a 2.5fold higher risk to die due to cardiac events [2]. Hemodynamic parameters like wall shear stress (WSS) or pulse wave velocity (PWV) can influence the regional compliance of the vessel and is discussed as an essential predictor of the atherosclerotic plaque formation [3][4][5][6][7]. A change of the shear stress at the aortal wall is discussed as a predictor for the development of atherosclerotic plaque formation especially at the branch of the aorta and the inner curvature of the aortic arch [8][9][10][11]. The WSS is a force per area exerted on a surface by a liquid flowing parallel to this surface measured in N = m 2 [12,13]. The WSS is defined as with the deformation tensor with t ! = WSS, η = viscosity, n ! = inward unit normal of the surface, x i,j = spatial dimensions and v i,j = velocity components [12,14]. Previous studies described the possibility of evaluating the WSS via 1.5 and 3 Tesla MRI in human vessels with contrast agent and without a longitudinal MR follow up or histological analysis [12,15]. Furthermore, some studies performed WSS measurements also in murine vessels [16,17], and for example Zhao et al. measured the WSS via a phase contrast method in the murine abdominal aorta in even 9 month old mice [18]. For evaluating the atherosclerotic plaque development and the WSS over a longer period in time, Apolipoprotein E-deficient (ApoE -/-) mice are a feasible model due to their spontaneous atherosclerotic plaque development [19,20]. The longitudinal in vivo murine WSS evaluation using 17.6 Tesla magnetic resonance imaging (MRI) has previously not been explored.
Thus, the purpose of this study was performing a longitudinal analysis to evaluate the feasibility of detecting the interaction between the wall shear stress and the beginning atherosclerotic plaque development in the murine aortic arch using 17.6 Tesla MRI with histological analysis.

Animal protocol
This longitudinal animal study was performed with 20 eight-week-old female ApoE -/mice (Charles River Laboratories, Sulzfeld, Germany) at the Physical Institute Würzburg (Germany). These mice were randomly separated in two groups consisting of twelve ApoE -/mice placed on atherogenic Western Diet (E15721-347, ssniff special diets GmbH, Soest, Germany) with a fat content of 21.2% and eight ApoE -/mice placed on Chow Diet (V1534-000, ssniff special diets GmbH, Soest, Germany) with a fat content of 3.3%, starting at the age of eight weeks. The unequal group sizes were chosen to compensate for the higher expected mortality in the group placed on Western Diet. Western Diet accelerates the formation of atherosclerosis in ApoE -/mice [20]. Magnetic resonance (MR) scans and histological analysis were performed after one to two week, eight weeks, and twelve weeks according to the study plan (Fig 1, S1 Table). This study was designed as a feasibility study for sequential MR WSS measurements in established atherosclerosis. Therefore, only a limited number of mice were allowed for baseline measurements in accordance to the animal protection (S1 Table).

Animal handling
The mice were maintained on a 12h-light-and 12h-dark-cycle at 20-22˚C indoor temperature and a relative humidity of (55 ± 10)% and were provided with water access and food (one group with Western Diet and one group with Chow Diet) ad libitum. Mice were group housed (2-4 mice per cage) and prior to experiments step-by-step acclimatized to the laboratory environment.
All examinations were performed under isoflurane narcosis to prevent suffering. Isoflurane was used for anesthesia due to the smallest effect on the hemodynamic parameters in contrast to other anesthetics as demonstrated by Janssen et al. [21]. The mice were induced into anesthesia at dose of 3-4 Vol.% isoflurane, then maintained by continuous inhalation of 1.5-2 Vol. % isoflurane and 2 L = min O 2 during spontaneous breathing. Mouse physiology was continuously monitored using a breathing and electrocardiographic (ECG) monitoring unit during the measurement. By individually adjusting anesthesia for each mouse to maintain ECG periods of about 110 ms and respiratory periods of about 1200 ms a comparable depth of anesthesia in all experiments was ensured. Body temperature was maintained by using a heating bed during animal preparation and by adjusting the temperature of the gradient cooling unit to 36˚C during the MR measurements.
The mice had been painlessly euthanized under overdose isoflurane anesthesia with a following exsanguination according to Annex IV 'Methods of killing animals' of the Directives of the European Parliament and of the Council on the protection of animals used for scientific purposes [22].
All experimental procedures were in accordance with the institutional and internationally recognized guidelines [22] and were approved by the Regierung von Unterfranken (Government of Lower Franconia), Würzburg, Germany, to comply with German animal protection law under reference number 55.2.-2531.01-23/11.

Magnetic Resonance Imaging (MRI)
A 2D gradient echo imaging method with a 3D phase contrast flow encoding [23] was validated and optimized using an ultrahigh field 17.6 Tesla MRI (Avance 750WB) with an 89 mm vertical bore operated by ParaVision 4.0 (Bruker BioSpin, Rheinstetten, Germany). The spectrometer is equipped with a 1000 mT = m gradient unit offering a 40 mm bore for rf resonators and samples. A 27 mm inner diameter custom-built birdcage resonator was employed in all imaging experiments (Fig 2). This setup enables the analysis of blood flow velocity in very narrow vessels.
Prior to the longitudinal in vivo study measurement procedures were established using phantoms and WT C57Bl/6 mice. As phantoms served tubes with a diameter of 4 mm and of 1 mm, respectively, with an aqueous solution of copper sulfate (1.25 g = ml ) circulating at 1 cm = s driven by a gear pump. Volumetric flow rate was determined by measuring the water volume drained from the tube per time and compared with the results obtained from MRI measurements.
For in vivo measurements mice were placed head first in the resonator and then the entire setup was inserted into the scanner from the bottom. Measurements were performed breath triggered with a pneumatic sensing balloon at the mouse chest and a trigger unit (Rapid Biomedical, Rimpar, Germany) (S1 Fig frames. The maximum encodable velocity was 167 cm = s , the temporal resolution 5 ms and the total acquisition time about 10 minutes. The four MR planes were located in the thoracic part of the aorta before the Truncus brachiocephalicus, after the Arteria subclavia sinistra and in the thoracic aorta, orthogonal to blood flow (Fig 3).
The optimal position of the mouse in the resonator and the MR plane orthogonal to the blood flow and the wall of the aorta was important to achieve comparable and reproducible MR scans for the following analysis of the flow parameters. All MR scans were performed without using contrast agents or filters in postprocessing.

WSS analysis
Data was imported into Matlab (The MathWorks, Natick, MA, USA) and exported to Ensight (CEI, Apex, NC, United States), where manual segmentation of the relevant anatomic region

PLOS ONE
was performed [24]. In a following step these data were processed using the Matlab-based "Flow Tool" (Department of Radiology, Medical Physics, University of Freiburg, Freiburg, Germany). Information on the internal algorithms of "Flow Tool" are described in [12]. In this tool the individual vessel cross sections were manually segmented using a B-Spline interpolation on the morphological data. Subsequently the tool computed from the corresponding flow data velocity maps for the selected MR planes and based on these the WSS. For blood viscosity and density the preset values of "Flow Tool" (viscosity: 4.5�10 −3 Ns = m 2 and density: 1055 kg = m 3 ) were used [12,25]. All WSS data in this study are presented as mean WSS values averaged over the whole circumference.

Statistical analysis and WSS analysis
The continuous variables were described using mean ± standard deviation (SD) (range) or median (interquartile range, range). Shapiro-Wilk test was used for testing the normal distribution of the data and Levene's test for homogeneity of variances. Differences between Western Diet group and Chow Diet group were tested using the Student's t-test with unequal variables for normally distributed data or using the Mann-Whitney U test for not normally distributed data. Changes over time in the individual groups were not evaluated due to the low numbers of mice in each group. A p value �0.05 was considered statistically significant. All statistics were done with SPSS version 25 ( © IBM corporation and its licensors 1989, 2017).

Phantom measurements
Using the flow phantom with a diameter of 4 mm, the volume per time was (2.911 ± 0.054) ml = s and the flow via MRI was (2.979 ± 0.101) ml = s (S2 Table); in the flow phantom with a diameter of 1 mm, the volume per time was (0.220 ± 0.003) ml = s and the flow via MRI (0.215 ± 0.003) ml = s (S3 Table). This comparison showed a very good agreement of both methods.

Study population
The body weight of all 20 ApoE -/mice with Western Diet and Chow Diet increased significantly over the study period of 12 weeks (Table 1,

Wall shear stress (WSS)
WSS was evaluated at four MR planes illustrated in

Histological analysis
The histological analysis was performed in mice in the Western Diet group and the Chow Diet group at each measurement point after one week, eight weeks, and twelve weeks. Spatial correlation between histological analysis and MRI was established by observation and comparison

PLOS ONE
of morphological landmarks in both modalities. In the group of mice with Western Diet the histological analysis showed a small plaque formation with elastin fragmentations after eight weeks and big plaque formations and severe elastin fragmentations after twelve weeks. In contrast, in the group of Chow Diet mice after eight weeks there were only a few elastin fragmentations but no plaque formation. Only at the third measurement point the histological analysis could exhibit bigger plaque formations (Figs 5 and 6, Table 3). The plaque formations occurred especially at the inner curvature of the aortic arch and at the bifurcations in all mice (S5 Fig).
The CD68 staining of ApoE -/mice of the Western Diet group after 12 weeks demonstrated the rate of macrophages inside the plaque formation in the aortic arch (Fig 7).

Discussion
In this longitudinal study MR based WSS measurements were feasible in ApoE -/mice via 17.6 Tesla ultra highfield MRI using a 2D gradient echo imaging method with a 3D phase contrast flow encoding. ApoE -/mice tend to develop plaque formations in different manifestation depending on the respective diet types [26]. This study extends previous findings by demonstrating increasing plaque formations over a time period of 12 weeks with a higher severity in the group fed with Western Diet. The histological analysis found the majority of the plaques at the inner curvature of the aortic arch, which has also previously been described [6,10,17]. ApoE -/mice were used for both groups to achieve a good comparability of the aortic anatomy The change of the WSS is on the one hand the cause of the atherosclerotic plaque formation and on the other hand the consequence. A low WSS value is described as a predictor for plaque

PLOS ONE
development [15,28]. Low WSS affects an atherosclerotic plaque development [31-33], whereas the plaque itself affects a higher WSS [6]. MR plane 3, which exhibits on plaque formations, shows lower WSS values compared to the WSS values of the other MR planes. The WSS values of this study are in accordance with findings in literature [6,34]. Previous studies showed that mice had 7 to 12fold higher WSS values compared to human WSS measurements [17,34,35]. Moreover, Cheng et al. could show an inverse correlation with a 7fold higher WSS comparing a mouse with a body weight of 0.03 kg with a human person with 60 kg [34]. Feintuch et al. and Trachet et al. also showed that the diameter of the aorta can also affect the WSS value with a higher WSS value in case of a smaller aortic diameter [16,36]. These findings are in accordance with the WSS values in this study compared to previous findings from Stalder et al. [12]. The first step of longitudinal visualization of the WSS via a 2D imaging method is feasible in a high quality, thus, now the application of 3D acquisitions, as described in [37], in a longitudinal study is needed. However, the previous publication by Stalder et al. [12] showed that a 3D measuring method underestimates the WSS values in comparison to a 2D measuring method.

Study limitation
This longitudinal animal study was performed with a small number of mice, which was reduced at each measurement point due to the histological analysis of the murine aorta. Using a 2D MR imaging method the WSS measurements could only be performed at predefined MR planes. The measurement was breath triggered depending on a constant narcosis with stable vital parameters. The histological analysis was done parallel to the tissue tec block in which the aorta was imbedded, whereby the comparison with the WSS values was possible due to the whole histological analysis of the block and the aorta.

Conclusions
In conclusion wall shear stress measurement in the small vessels of the mouse aorta is feasible.
There is a tendency towards higher values of wall shear stress in vessels affected by plaque formation. This formation is favored by a high-fat diet and can as shown by histology predominantly be detected at predilection sites at the inner curvature of the aorta. Ultra highfield MRI can therefore serve as a tool for studying the causes and beginnings of atherosclerotic plaque formation.