Identification of congenital aortic valve malformations in juvenile natriuretic peptide receptor 2-deficient mice using high-frequency ultrasound

Abstract

Mouse models of congenital aortic valve malformations are useful for studying disease pathobiology, but most models have incomplete penetrance [e.g., ∼2 to 77% prevalence of bicuspid aortic valves (BAVs) across multiple models]. For longitudinal studies of pathologies associated with BAVs and other congenital valve malformations, which manifest over months in mice, it is operationally inefficient, economically burdensome, and ethically challenging to enroll large numbers of mice in studies without first identifying those with valvular abnormalities. To address this need, we established and validated a novel in vivo high-frequency (30 MHz) ultrasound imaging protocol capable of detecting aortic valvular malformations in juvenile mice. Fifty natriuretic peptide receptor 2 heterozygous mice on a low-density lipoprotein receptor-deficient background (Npr2^+/−;Ldlr^−/−; 32 males and 18 females) were imaged at 4 and 8 wk of age. Fourteen percent of the Npr2^+/−;Ldlr^−/− mice exhibited features associated with aortic valve malformations, including 1) abnormal transaortic flow patterns on color Doppler (recirculation and regurgitation), 2) peak systolic flow velocities distal to the aortic valves reaching or surpassing ∼1,250 mm/s by pulsed-wave Doppler, and 3) putative fusion of cusps along commissures and abnormal movement elucidated by two-dimensional (2-D) imaging with ultrahigh temporal resolution. Valves with these features were confirmed by ex vivo gross anatomy and histological visualization to have thickened cusps, partial fusions, or Sievers type-0 bicuspid valves. This ultrasound imaging protocol will enable efficient, cost effective, and humane implementation of studies of congenital aortic valvular abnormalities and associated pathologies in a wide range of mouse models.

NEW & NOTEWORTHY We developed a high-frequency ultrasound imaging protocol for diagnosing congenital aortic valve structural abnormalities in 4-wk-old mice. Our protocol defines specific criteria to distinguish mice with abnormal aortic valves from those with normal tricuspid valves using color Doppler, pulsed-wave Doppler, and two-dimensional (2-D) imaging with ultrahigh temporal resolution. This approach enables early identification of valvular abnormalities for efficient and ethical experimental design of longitudinal studies of congenital valve diseases and associated pathologies in mice.

INTRODUCTION

Congenital aortic valve malformations include aortic atresia (1), aortic stenosis (2, 3) and bicuspid aortic valves (BAVs) (4). Among these, BAVs are the most common, affecting more than 2% of the global population (4, 5). In BAV disease, two of the three cusps in a normal tricuspid aortic valve (TAV) fuse, forming a central raphe between the cusp commissures. The morphological patterns of BAV fusion differ depending on which commissures fuse during valvulogenesis, with the fusion of the right coronary (RC) and left coronary (LC) cusps being the most common in humans (6, 7). Patients with BAVs are prone to a spectrum of valvular complications including mild-to-severe stenosis, accelerated calcification, and/or regurgitation (4, 5). Up to 84% of patients with a BAV also have ascending aortic dilatation (AAD) (8), with aortic diameters that are consistently reported to be larger than aged-matched TAV subjects (9). Dilatations can progress to aneurysms that are at risk of dissection and rupture, which represent two main causes of mortality in patients with BAV. Because of its complex sequelae, BAV disease is now classified as a group of diseases as opposed to a single entity (5, 8, 9), and this motivates improved understanding of BAV pathobiology and associated aortopathies.

To this end, mouse models of BAV disease are useful to investigate the causality of specific gene mutations on associated valvulopathies and aortopathies. However, a practical challenge is that mouse models with BAVs [including those with mutations in NOTCH1 (10–13), the GATA transcription factor family (14–16), endothelial nitric oxide synthase (eNOS) (12, 13, 17), and natriuretic peptide receptor 2 (Npr2) (18)] show incomplete penetrance of BAV (between 2 and 77%). For longitudinal studies of BAV-associated pathologies, which can take several months to develop (18), it would be operationally inefficient, economically burdensome, and unethical to blindly enroll large numbers of mice in a long-term study, when only a small fraction have the valve abnormality of interest. A novel method to diagnose BAVs in young mice is needed to overcome these challenges, and the lack thereof is partly responsible for the paucity of mouse studies in the field.

To address this need, we developed and validated a high-frequency ultrasound imaging protocol to diagnose congenitally abnormal aortic valves in 4-wk-old mice. High-frequency ultrasound imaging (≥30 MHz) has been widely used to evaluate cardiovascular structure, function, and hemodynamics in fetal and adult mouse models (19–21), including those of normal aortic valves (22). However, direct visualization of congenital valve defects in young mice is challenging because of the low echogenicity of their thin aortic valvular cusps and their high heart rates. Here, we used high-frequency color and pulsed-wave Doppler imaging to characterize transaortic blood flow, and electrocardiogram (ECG)-based kilohertz visualization (EKV), a two-dimensional (2-D) imaging modality with ultrahigh temporal resolution (23–25), to visualize the aortic valve cusps with enhanced contrast and their fine movement at a very high frame rate. To develop our protocol, we used Npr2 heterozygous mice on a low-density lipoprotein receptor-deficient background (Npr2^+/−;Ldlr^−/−), which have a reported BAV incidence rate of ∼11% (18). Notably, Npr2^+/−;Ldlr^−/− mice are the only mouse model reported to acquire AAD, with more disturbed hemodynamics and greater dilatation in mice with BAVs than those with TAVs (18). Here we defined imaging-based criteria that accurately identify congenital valve dysfunction in juvenile mice, which were validated by gross dissection and histology. Our ultrasound imaging approach to identify aortic valve malformations in young mice is accessible and accurate and will enable efficient and ethical experimental designs for longitudinal studies of congenital valve disease in mice.

MATERIALS AND METHODS

Animal Handling and Genotyping

All animal procedures were approved by the Animal Care Committees of the University of Toronto in accordance with the guidelines of the Canadian Council on Animal Care. Male and female heterozygous Npr2^+/− (JAX STOCK Npr2^tm1Gar/J; Cat. No. 007658) and homozygous Ldlr^−/− (JAX B6.129S7-Ldlr^tm1Her/J; Cat. No. 002207) mutant mice were obtained from The Jackson Laboratory (JAX; Bar Harbor, ME). Npr2^+/− mice were obtained from JAX on a mixed 129/B6 background at N2 to C57Bl/6 and were subsequently backcrossed with C57Bl/6J mice (JAX; Cat. No. 000664) to N3 before experimental use. Npr2^+/− mice (N3 to C57Bl/6) were then bred to Ldlr^−/− mice to generate Npr2^+/−;Ldlr^−/− mice. Npr2^+/− mice on the Ldlr^−/− background were maintained via heterozygous sibling breeding.

Mice were genotyped for Npr2 by the standard polymerase chain reaction (PCR) genotyping protocol for stock number 007658 (v.1.0, Jackson Laboratory). DNA was isolated from juvenile ear notches with 25 mM NaOH/0.2 mM EDTA at 98°C for 1 h followed by neutralization with 40 mM Tris-HCl. PCR was performed with JAX primers 9950 (common forward): 5′- CGGCTATCAGGCTCAGTTTT-3′, 9951 (wild-type reverse): 5′- CAGCATTCTGGAGGCTAAGG-3′, and oIMR7415 (mutant reverse): 5′- GCCAGAGGCCACTTGTGTAG-3′. Gel electrophoresis bands were produced by the wild-type allele (Npr2⁺) at 490 bp and by the mutant allele (Npr2^tm1Gar) at 234 bp. For this study, all Npr2^+/−;Ldlr^−/− mice (breeders and experimental mice) were fed a normal diet (Harlan Teklad Global 2918, 6.2% fat and 0% cholesterol by weight, 18% kcal/g from fat; Indianapolis, IN).

In Vivo Imaging Using High-Frequency Ultrasound

Fifty Npr2^+/−;Ldlr^−/− mice were imaged at 4 wk of age. The 4-wk time point was selected for screening, as mice are often put on a high-fat Western diet starting at 6 wk of age to expedite the progression of aortopathies and other cardiovascular diseases (18); therefore, a 4-wk screening time point would allow for categorization of healthy versus diseased mice before study enrollment and diet introduction. Mice were also imaged at 8 wk of age to assess changes, if any, in the congenital abnormalities identified in valves of at 4-wk-old mice. A high-frequency ultrasound imaging system (Vevo3100, FUJIFILM VisualSonics, Toronto, Canada) with 30-MHz transducers (MX400) was used. Mice were anesthetized for imaging using isoflurane (induced at 5% and then maintained at 1.5%). Mice were positioned in supine position with four paws taped to electrodes on a prewarmed platform for ECG recording and heart rate monitoring. A rectal thermometer was used to monitor body temperature, which was maintained at 37°C. Nair hair removal cream was used on the chest to remove fur.

Color and pulsed-wave Doppler for transaortic flow dynamics.

The first step of the imaging protocol (Fig. 1) assessed transaortic flow dynamics using both color and pulsed-wave Doppler through the right superior parasternal acoustic window. To achieve this, the transducer was positioned pointing downward and slightly to the left, at an angle of 30–45° in relation to the mouse’s coronal plane and the ultrasound stage (Fig. 1, A–C). Refinements were made to the positioning of the transducer relative to the mouse until a clear longitudinal cross section of the flow pathway encompassing the left ventricular outflow tract, aortic orifice, and ascending aorta was visible, allowing for observation of aortic valve movement (Fig. 1D). For the assessment of blood flow pattern and velocity distribution across the aortic orifice and ascending aorta, color Doppler-flow mapping was conducted using the highest pulsed repetition frequency (PRF) of 40 kHz (Fig. 1E). To capture the highest velocity distal to the valves, the largest gate size (1.08 mm) was used (Fig. 1F). Velocity time integral (VTI₂) at this same location, distal to the aortic valve, was measured. VTI₁ at the end of the left ventricular outflow tract (proximal to the aortic valve) was next quantified by repositioning the pulsed-wave Doppler gate and minimizing the gate size to 0.4 mm (Fig. 1G).

Figure 1.

Summary of the ultrasound imaging sections used to analyze the transaortic flow dynamics and morphology of the aortic valve in the Npr2^+/−;Ldlr^−/− mice. The midline of the mouse is indicated by the yellow line. A–D: transducer was oriented to obtain the right superior parasternal long-axis view of the flow channel from the left ventricular outflow tract (LVOT), aortic orifice, and aortic sinuses to the ascending aorta (AA). Fine adjustments were made to ensure that the smallest possible intercept angle (∼35°) between the mouse’s coronal plane and the ultrasound beam was obtained. E: Color Doppler mapped blood flow from the LVOT to the AA (flow moving toward the transducer is shown as red). F: pulsed-wave Doppler and a large gate (1.08 mm) that was placed distal to the aortic valve (within the aortic sinuses) were used to quantify the peak systolic velocity and the velocity time integral (VTI). G: pulsed-wave Doppler and a small gate (0.4 mm) that was placed proximal to the aortic valves (at the distal end of the LVOT) were used to quantify VTI. H–K: transducer was oriented to obtain the right inferior parasternal long-axis view of the flow channel proximal and distal of the aortic valve, with the mouse’s coronal plane at an angle of ∼45° to the transducer. L: diameter of the aortic annulus was measured. M: transducer was rotated 90° from the position in I for a right short-axis view of the aortic valvular cusps. N: commissures between the right-coronary (RC) cusp and noncoronary (NC) cusp were visualized. O–R: transducer was oriented to obtain left parasternal long-axis view of the flow channel proximal and distal of the aortic valve, with the mouse’s coronal plane ∼30° to the transducer. S: diameter of AA was measured at a location distal to the sinotubular junction (STJ), LVOT, and left ventricle (LV). T: transducer was rotated 90° from the position in P for a left short-axis view of the aortic valvular cusps. U: commissures LC and NC cusp were visualized.

EKV to visualize aortic valve morphology and movement.

Imaging sections were taken from the right parasternal acoustic window to assess aortic valve cusp morphology and dynamics using EKV. The transducer imaging plane was maintained at an angle of ∼45° from the mouse’s coronal plane, with the aorta’s longitudinal axis perpendicular to the ultrasound beam (Fig. 1, H–K). An EKV cineloop operating at 1,000 Hz was recorded to observe aortic cusp movements and measure the peak systolic diameter of the aortic annulus at the end of left ventricular outflow tract (Fig. 1L). With a 90° rotation of the transducer, a short-axis view of the aortic cusps was obtained to capture another EKV cineloop (Fig. 1M). The movement of the aortic valvular cusps throughout entire cardiac cycle and the echogenic commissure lines between the RC-noncoronary (NC) and RC-LC cusps were observed (Fig. 1N).

Subsequently, the transducer was repositioned to visualize the aortic valves through the left parasternal imaging window, maintaining an imaging section of ∼30° relative to the mouse’s coronal plane (Fig. 1, O–R). The longitudinal axis of the aorta was again aligned perpendicular to the ultrasound beam, and another EKV cineloop was recorded (Fig. 1S). Subsequently, a 90°-transducer rotation (Fig. 1T) produced an EKV cineloop for the short-axis view, highlighting commissures between the LC-NC and LC-RC cusps (Fig. 1U).

M-mode to visualize and quantify left ventricular morphology and function.

Finally, the left precordial acoustic window was used to visualize the left ventricle along its longitudinal axis (19, 20). An M-mode recording was made to visualize left ventricular morphology and quantify associated functionality.

Ultrasound data analysis.

The ultrasound images were analyzed using the VevoLab software (v.5.6.1, FUJIFILM VisualSonics). Color Doppler-flow mapping was qualitatively assessed for the overall flow pattern through the aortic orifice and ascending aorta (recirculation) during systole and signs of flow reversal into the left ventricle (regurgitation) during diastole. Pulsed-wave Doppler data were quantitatively analyzed with angle correction for the peak systolic velocities and VTI (mm), and the measurements were averaged over three cardiac cycles.

The EKV cineloops from both the right and left parasternal short-axis views were analyzed to identify morphological features of the aortic valves, including number of cusps, relative cusp size, and commissure morphology and fusion, if present. To identify potential aortic remodeling, the aortic diameter was measured distal to the sinotubular junction on the EKV cineloop of the left parasternal long-axis view.

To calculate cardiac output (CO), the diameter of the aortic annulus (DAA) measured from the EKV cineloop of the right parasternal long-axis view (Fig. 1L) was used to calculate the cross-sectional area of the aortic orifice (AAO; the end of the left ventricular outflow tract) as follows:

$$\text{AAO}=\mathrm{\pi }\left(\frac{\text{DAA}}{2}\right)$$ (1)

AAO, VTI₁ (proximal to the aortic valve), heart rate (HR), and the left ventricular stroke volume (SV in mL) were next used to calculate CO (in mL/min):

$$\text{SV}={\text{VTI}}_1\times \hspace{0.17em}\text{AAO}$$ (2)

$$\text{CO}=\text{SV}\times \text{HR}$$ (3)

The M-mode recording from the long-axis view of the left ventricle was used to quantify left ventricular fractional shortening (LVFS in %).

To calculate the effective aortic valve area (AVA in mm²), the continuity equation was used. AAO (1), the VTI₁ measurements from proximal to the aortic valve, and the VTI₂ measurements from distal to the aortic valve were used to calculate AVA as follows:

$$\text{AAO}\times {\text{VTI}}_1=\text{AVA}\times {\text{VTI}}_2$$ (4)

Gross Anatomy Analysis

To confirm the suspected aortic valvular abnormalities diagnosed using the in vivo ultrasound imaging data, the gross anatomies of the mouse aortic valves were analyzed by ex vivo dissection. After imaging at the 8-wk time point, mouse hearts were dissected and slowly flushed with 5 mL PBS^−/− with a 25-gauge needle through the apex of the left ventricle. The hearts were fixed in 4% formaldehyde for 24 h. Following that, three 70% ethanol washes were done, and the hearts were stored in 70% ethanol until needed. The hearts were subsequently examined using an Olympus SZ61 Olympus stereomicroscope. To check for presence of aortic valvular abnormalities, the ascending aorta was transected at the level of the sinotubular junction to expose the coronary sinuses and cusps of the aortic valve. Forceps were used to probe the aortic valves to identify cases of fusion or partial fusion. Images were taken under the stereoscope using an iPhone11 camera.

Histology

To analyze differences in leaflet thickness and extracellular matrix composition, the hearts were paraffin-fixed, longitudinally sectioned to 10-µm thickness, and stained using Movat’s pentachrome stain. Histological sections were examined on an Olympus BX31 light microscope and imaged using an Olympus SC30 camera.

Statistics

Quantitative data are presented as means ± SD. Normality was tested with the Shapiro–Wilk test. For normally distributed data, a two-way ANOVA and Tukey test for pairwise comparisons were performed on GraphPad Prism 9. For nonnormally distributed data, the Scheirer–Ray–Hare test with Dunn test for pairwise comparisons was performed with R (26). P ≤ 0.05 was considered statistically significant.

RESULTS

Npr2^+/−;Ldlr^−/− Mice Demonstrated Congenitally Abnormal Aortic Valves with an ∼14% Incidence Rate

A total of 50 Npr2^+/−;Ldlr^−/− mice (32 M and 18 F) were genotyped and screened using echocardiography at both 4 and 8 wk of age. Gross anatomical observation after dissection at 8 wk determined that 43 mice had normal TAVs (Fig. 2A) and 7 had abnormal aortic valves, comprising two abnormal TAVs (1 male and 1 female) with thickened cusps (Fig. 2B), two putative partially fused BAVs (both males) (Fig. 2C), and three Sievers type-0 BAVs (2 males and 1 female) (Fig. 2D). The abnormal valve incidence rate of ∼14% is similar to that reported previously (∼11%) for Npr2^+/− mice not on an Ldlr^−/− background (18). In contrast to normal TAVs (Fig. 3, A and B), the aortic valve cusps in the four mice with abnormal TAVs or partially fused BAVs were thickened and had increased proteoglycan, collagen, and cellularity evident by Movat’s pentachrome staining (Fig. 3, C and D). Although qualitatively thicker, there were no noticeable qualitative differences in cellularity or extracellular matrix composition in the cusps of Sievers type-0 BAVs (Fig. 3, E and F) compared with those of normal TAVs.

Figure 2.

Fifty Npr2^+/−;Ldlr^−/− mice (32 males and 18 females) were screened using echocardiography at 4 and 8 wk of age. Gross anatomical inspection of the aortic valves showed 43 mice had normal tricuspid aortic valves (TAVs; A); 2 mice had abnormal TAVs with thickened left-coronary (LC), right-coronary (RC), and noncoronary (NC) cusps (B); 2 mice had putative partial fusion between the RC-LC cusps or the RC-NC cusps (black arrows) [type IIC asymmetric bicuspid aortic valve (BAV)] with thickened RC, LC, and NC cusps (C); and 3 mice had lateral Sievers type-0 BAVs (D). Scale bar = 100 μm.

Figure 3.

Longitudinal histological sections of mouse aortic valves were stained with Movat’s pentachrome. A and B: representative images displaying thin right coronary (RC) and left coronary (LC), characteristic of a normal mouse tricuspid aortic valve (TAV). C and D: representative images showing thickened RC and LC cusps observed in a representative TAV with thickened cusps, with increased collagen content (yellow), cellularity (red), and proteoglycan content (blue) relative to normal TAVs. E and F: representative images from a Sievers type-0 lateral bicuspid aortic valve (BAV) suggested no apparent differences in cellularity, collagen, or proteoglycan content compared with a normal TAV. Scale bar = 200 μm (A, C, and E) and 50 μm (B, D, and F).

Echocardiography Identified Aortic Valve Congenital Abnormalities in 4-wk-Old Npr2^+/−;Ldlr^−/− Mice

Mice were imaged at 4 and 8 wk of age by color-Doppler, pulsed-wave Doppler, and EKV imaging to assess valve morphology and function and left ventricular function. Imaging findings were retrospectively correlated with valve morphotypes determined by gross anatomical and histological examination to establish imaging-based criteria to diagnose congenitally abnormal aortic valves.

Color-Doppler imaging of the ascending aorta in 4-wk-old mice revealed unidirectional flow through the aortic valve and no regurgitation in 43 mice with TAVs (Fig. 4, A and B). In contrast, recirculation distal to the aortic valve during systole and regurgitation during diastole were evident in six of the seven mice with anatomically abnormal valves (Fig. 4, C and D; Table 1). Regurgitation was qualitatively mild in three of the mice with abnormal valves and severe in the other three. Recirculation and regurgitation were not evident in one mouse that had an abnormal TAV with the least thickened cusps of the seven abnormal cases.

Figure 4.

Color-Doppler mapping blood flow through the ascending aorta at 4 wk of age. Color bar represents velocity associated with color of flow observed in frames. A: frame depicting laminar flow during peak systole through a representative normal tricuspid aortic valve (TAV). B: frame depicting lack of regurgitation during peak diastole in a normal TAV. C: frame depicting abnormal flow in a representative Sievers type-0 bicuspid aortic valve (BAV), with aliasing and recirculation, during peak systole, distal to the aortic valve. D: frame depicting severe regurgitation during peak diastole in a representative Sievers type-0 BAV.

Table 1.

Evaluation of color Doppler-associated parameters that ascertain the presence of abnormal flow through the ascending aorta at 4 wk of age

Valve Morphotype	Frequency of Recirculation	Frequency of Regurgitation
Normal TAV	0/43	0/43
TAV with mild cusp thickening	0/1	0/1
TAV with severe cusp thickening	1/1	1/1 (mild)
Putative partially fused BAV	2/2	2/2 (mild)
Sievers type-0 BAV	3/3	3/3 (severe)

All 43 mice with normal TAVs had normal flow through the ascending aorta during systole and diastole. Mice (6 out of 7) with an abnormal tricuspid aortic valve (TAV), partially fused bicuspid aortic valve (BAV), or Sievers type-0 BAV had flow disturbances, characterized by signs of recirculation and regurgitation. A mouse with an abnormal TAV with mild cusp thickening had no noticeable recirculation or regurgitation.

Average peak systolic velocities through the aortic valve were measured by pulsed wave-Doppler imaging. Transvalvular velocities in 4-wk-old mice with abnormal valves were significantly greater than those in mice with normal TAVs (1,416 ± 164 vs. 1,033 ± 108 mm/s, P < 0.0001; Fig. 5A). Similarly, average velocities measured at 8 wk were significantly different between normal TAV and abnormal valve mice (1,018 ± 123 vs. 1,482 ± 194 mm/s, P < 0.0001; Fig. 5A). Notably, differences between the average 4- and 8-wk velocities were not significant for both the normal TAV and abnormal valve mice (P > 0.99 and P > 0.99, respectively; Fig. 5A), with most mice individually demonstrating similar velocities at 4 and 8 wk (Fig. 5B). One mouse with a TAV with moderately thickened cusps had a peak systolic velocity of ∼1,200 mm/s at 4 wk of age. However, for all the other mice with abnormal valves, at both 4 and 8 wk, there was a threshold of ∼1,250 mm/s that distinguished normal TAV from abnormal valve mice. Only two mice at 4 wk and three mice at 8 wk of the 43 normal TAV mice had transvalvular velocities exceeding this threshold. Consistent with elevated transvalvular velocities in mice with abnormal aortic valves, the average effective AVA was smaller in abnormal cases relative to normal TAVs at both 4 wk (0.50 ± 0.1 vs. 0.65 ± 0.2 mm², P < 0.05; Fig. 5C) and 8 wk (0.47 ± 0.1 vs. 0.65 ± 0.2 mm², P < 0.005; Fig. 5C), but differences were not statistically different between time points. No statistical differences for transvalvular velocities between male and female mice with normal TAVs were observed at 4 wk (1,057 ± 121 vs. 991 ± 70 mm/s; P > 0.1) and 8 wk of age (1,028 ± 123 vs. 995 ± 120 mm/s; P > 0.99) or when comparing 4- versus 8-wk-old mice (Supplemental Fig. S1A; Supplemental figures may be found at https://doi.org/10.5683/SP3/Q6E2AY). There were also no statistical differences in effective AVA between male and female mice with normal TAVs at 4 wk (0.69 ± 0.3 vs. 0.60 ± 0.2 mm²; P > 0.99) or 8 wk of age (0.66 ± 0.22 vs. 0.64 ± 0.17 mm²; P > 0.99) or when comparing 4- versus 8-wk-old mice (Supplemental Fig. S1C). The small incidence of abnormal female valves prevented a statistical comparison based on sex, but qualitatively, there were no obvious differences in transvalvular velocities or effective AVA between male and female mice with abnormal valves at 4 or 8 wk of age or when comparing 4- versus 8-wk-old mice (Supplemental Fig. S1, B and D).

Figure 5.

Pulsed-wave Doppler on the parasternal long-axis view of the aorta was used to obtain velocities distal to the aortic valve. A: transvalvular velocities of mice with tricuspid aortic valves (TAVs) (n = 43) were significantly lower than those with abnormal valves (n = 7) at both 4 and 8 wk of age but did not change with age (P > 0.99 for valve types). Data are presented as means ± SD. Mean value for a single mouse calculated from 3 technical replicates, in which each replicate was a velocity from a different cardiac cycle. Statistical analyses by the Scheirer–Ray–Hare test with Dunn’s test for pairwise comparisons were used. ****P < 0.0001. B: peak systolic velocities measured for most individual mice were comparable at the 4- and 8-wk time points. C: effective aortic valve area in mice with abnormal valves was significantly lower than in mice with normal TAVs at both 4 and 8 wk of age but did not change with age (P = 0.38). Data are presented as means ± SD. Mean value for a single mouse calculated from three technical replicates, in which each replicate was a velocity from a different cardiac cycle. Statistical analyses by the Scheirer–Ray–Hare test with Dunn’s test for pairwise comparisons were used. **P < 0.05; ***P < 0.005.

EKV on the right and left short-axis view of the aortic valve identified commissures between the RC, LC, and NC cusps. In most of the 43 cases with normal TAV, a distinctive “Y-shaped” echogenic pattern was observed (Fig. 6A). However, in the four mice with the abnormal TAVs with thickened cusps and putative asymmetric BAVs, a thick echogenic line between a pair of cusps was evident (Fig. 6, B and C). In the three mice with a Sievers type-0 BAV, an elongated echogenic line spanning the diameter of the aortic opening was noted (Fig. 6D).

Figure 6.

Electrocardiogram-gated kilohertz visualization (EKV) on both right and left short-axis views of the aortic valve. Commissures (yellow arrows) between the right-coronary (RC), left-coronary (LC), and noncoronary (NC) cusps were visualized in 4-wk-old mice. A: representative normal tricuspid aortic valve (TAV) with a “Y-shaped” echogenicity pattern. B: representative abnormal TAV with thickened cusps, evidenced by thickened line of echogenicity between LC-NC cusps in left short-axis EKV view. C: representative partially fused BAV with thickened cusps evidenced by a longer, thickened line of echogenicity between RC and NC cusps in the right short-axis EKV view. D: representative Sievers type-0 bicuspid aortic valve (BAV) with a line of echogenicity spanning the diameter of the aortic orifice in the right short-axis view. Commissure was located laterally relative to left coronary artery (LCA) and right coronary artery (RCA).

Npr2^+/−;Ldlr^−/− Mice with Abnormal Aortic Valves Did Not Demonstrate Aortic Remodeling by 8 wk of Age

Aortic diameter measured distal to the sinotubular junction (Fig. 7A) increased with growth from 4 to 8 wk in both mice with normal TAVs (P < 0.05) and abnormal valves (P < 0.0001) (Fig. 7B). However, there were no statistical differences in diameters between mice with normal TAVs or abnormal valves at either 4 wk (1.20 ± 0.1 vs. 1.18 ± 0.1 mm, P = 0.9; Fig. 7B) or 8 wk (1.33 ± 0.1 vs. 1.35 ± 0.1 mm, P > 0.99; Fig. 7B) of age. In addition, there were no differences in aortic diameter between male and female mice with normal TAVs observed at 4 wk (1.21 ± 0.1 vs. 1.19 ± 0.1 mm; P > 0.1) or 8 wk of age (1.33 ± 0.09 vs. 1.33 ± 0.09 mm; P > 0.99) (Supplemental Fig. S2A). There were, however, statistical differences in aortic diameter between 4- versus 8-wk-old males (P < 0.0001) and 4- versus 8-wk-old females (P < 0.005) with normal TAVs (Supplemental Fig. S2A). Qualitatively, there were no sex differences in aortic diameter between male and female mice with abnormal valves at 4 or 8 wk of age (Supplemental Fig. S2B).

Figure 7.

There was no congenital aortic or ventricular remodeling in the Npr2^+/−;Ldlr^−/− mice with TAVs (n = 43) or abnormal valves (n = 7). Data are presented as means ± SD. A: ascending aortic diameters distal (white arrow) to the left ventricle (LV), left ventricle outflow tract (LVOT), and sinotubular junction (STJ) were measured in the left long-axis view. B: ascending aorta diameters of mice with normal TAVs and abnormal valves increased with age but were not statistically different between mice with normal TAVs vs. abnormal valves at 4 or 8 wk. Analyses by 2-way ANOVA with the Tukey test for pairwise comparisons. **P < 0.05; ****P < 0.0001. C: cardiac output (CO) was derived from velocity time integral measurements made using pulsed-wave Doppler on the parasternal long-axis view of the ascending aorta. CO was not statistically different between mice with normal TAVs vs. abnormal valves at 4 or 8 wk of age. CO also did not statistically differ in 4- vs. 8-wk-old mice with normal TAVs (P = 0.53) or abnormal valves (P = 0.82). Analyses by the Scheirer–Ray–Hare test with the Dunn’s test for pairwise comparisons. D: long-axis view of the left ventricle quantified left ventricular fraction shortening (LVFS) (%). LVFS was not statistically different between mice with normal TAVs vs. abnormal valves at 4 (P = 0.09) or 8 (P = 0.07) wk of age and was not statistically different in 4- vs. 8-wk-old mice with normal TAVs (P = 0.4) or abnormal valves (P = 0.6). Analyses by the Scheirer–Ray–Hare test with the Dunn’s test for pairwise comparisons. TAV, tricuspid aortic valve.

Npr2^+/−;Ldlr^−/− Mice with Abnormal Aortic Valves Did Not Demonstrate Left Ventricular Remodeling by 8 wk of Age

Average CO did not differ between mice with normal TAVs and those with abnormal valves at either 4 wk (9.73 ± 3.0 vs. 10.72 ± 1.9 mL/min, P = 0.21; Fig. 7C) or 8 wk (10.02 ± 2.87 vs. 10.95 ± 3.5 mL/min, P = 0.54; Fig. 7C) of age. There were also no statistical differences in CO between male and female mice with normal TAVs at 4 wk (10.65 ± 3.1 vs. 8.18 ± 2.1 mL/min; P > 0.05) and 8 wk of age (10.39 ± 2.7 vs. 9.39 ± 3.0 mL/min; P > 0.99) or when comparing 4- versus 8-wk-old mice (Supplemental Fig. S2C). There were also no qualitative differences in CO between male and female mice with abnormal valves at 4 or 8 wk of age (Supplemental Fig. S2D). Similarly, left ventricular fractional shortening (LVFS) was not statistically different between valve morphotypes in mice at either 4 wk (35.5 ± 10.0 vs. 40.2 ± 6.7%, P = 0.43; Fig. 7D) or 8 wk (33.9 ± 9.3 vs. 3.38 ± 7.0%, P = 0.53; Fig. 7D) of age. No statistical differences were observed in LVFS between male and female mice with normal TAVs at 4 wk (36.8 ± 10.0 vs. 33.5 ± 10.0%; P > 0.1) and 8 wk of age (35.6 ± 10.8 vs. 30.7 ± 4.7%; P > 0.99) or when comparing 4- versus 8-wk-old mice (Supplemental Fig. S2E). In addition, there were no qualitative sex differences noted in LVFS between male and female mice with abnormal valves at 4 or 8 wk of age.

DISCUSSION

High-frequency ultrasound is a mainstay for mouse cardiovascular imaging (20, 27), particularly for adult disease models. Its application to diagnose congenital valvular defects, such as BAVs, in juvenile mice had not been previously explored. As mouse models of BAVs have low penetrance, efficient, economical, and ethical longitudinal studies of BAV disease require methods to identify BAVs early. However, direct imaging of cardiac valve morphology is challenging in young mice for several reasons: 1) it is difficult to target the aortic valve using a relatively large transducer with a wide, flat footprint within the constraints of the limited transthoracic acoustic windows; 2) the aortic valvular cusps, being thin and exhibiting minimal ultrasound signal reflection, are difficult to visualize, particularly in juvenile and younger mice; and 3) heart rates ranging from ∼300 to 550 beats/min render the real-time frame rate of standard high-frequency ultrasound imaging inadequate for capturing aortic valve movement (28, 29).

In the present study, we identified a set of echocardiographic features associated with valve malformations and disrupted flow in 4-wk-old Npr2^+/−;Ldlr^−/− mice with confirmed congenital aortic valve abnormalities: 1) recirculation and regurgitation on color Doppler; 2) peak systolic velocities reaching or surpassing ∼1,250 mm/s quantified on pulsed-wave Doppler; and 3) echogenicity at commissures by EKV, signifying congenital valve thickening and/or partial/full fusions (Table 2). The ultrasound-based protocol established here therefore provides a multifaceted approach to diagnose valvular abnormalities in the form of BAVs (including a partially fused BAV), leaflet thickening, and/or flow disturbances in juvenile mice.

Table 2.

Echocardiographic criteria associated with valvular abnormalities in 4-wk-old Npr2^+/−;Ldlr^−/− mice

	Observation(s) Met
	Color Doppler: Mapping blood flow through the ascending aorta		Pulsed-wave Doppler: Quantification of peak systolic velocity distal to the aortic valve		Electrocardiogram-gated kilohertz visualization (EKV): Visualizing commissures between the NC, RC, and LC cusps
Diagnoses Mice, N, (Males/Females)	Recirculation and regurgitation	Absence of recirculation and regurgitation	≥1,250 mm/s	<1,250 mm/s	Line of echogenicity on EKV extends diameter of the aortic orifice	Line of echogenicity on EKV is thickened at the commissures	Y-shaped pattern of echogenicity
Normal TAV, 43 (27/16)	0/43	43/43	2/43	41/43	0/43	0/43	43/43
Abnormal TAV with thickened cusps, 2 (1/1)	2/2	0/2	1/2	1/2	0/2	2/2	0/2
Putative partially fused BAV with thickened cusps, 2 (2/0)	2/2	0/2	2/2	0/2	0/2	2/2	0/2
Sievers type-0 BAV, 3 (2/1)	3/3	0/3	3/3	0/3	3/3	0/3	0/3

Values are number of mice within category that showed echogenic variable listed; N, number of mice (males/females). Mice with abnormal valves showed 1) recirculation and regurgitation observed on color Doppler (mapping blood flow through the ascending aorta); 2) peak systolic velocities reaching or surpassing 1,250 mm/s, as measured using pulsed-wave Doppler (quantification of peak systolic velocity distal to the aortic valve); and 3) putative fusion of cusps along commissures or thickened lines of echogenicity at commissures, signifying congenital valve thickening and/or partial/full fusions between right coronary (RC)-noncoronary (NC), RC-left coronary (LC), or LC-NC cusps, as observed on electrocardiogram-gated kilohertz visualization (visualizing commissures between the NC, RC, and LC cusps). In this study, mice were classified into 4 categories: Sievers type-0 bicuspid aortic valve (BAV), abnormal tricuspid aortic valve (TAV) with thickened cusps, putative partially fused BAV, or a normal TAV.

No single criterion of the three considered was sufficient in isolation to distinguish normal TAVs from congenitally abnormal valves, emphasizing the multicriteria approach (Table 2). For example, two mice with normal TAVs exhibited peak systolic velocities ≥1,250 mm/s at 4 wk of age but displayed undisturbed flow on color Doppler coupled with a typical Y-shaped echogenicity pattern on the short-axis EKV. One mouse with moderately thickened TAV cusps had a peak systolic velocity just under 1,250 mm/s and no recirculation or regurgitation on color Doppler at 4 wk of age. However, the combination of peak systolic velocity on the higher end of the spectrum, aliasing, and thickened lines of echogenicity on EKV enabled an accurate diagnosis of this mouse as one with valvular abnormalities. Such a gradient in valve abnormalities recapitulates the complexity of malformations observed in human BAV disease (4, 5, 30).

Histological examination of longitudinal cross sections revealed increases in cellularity and collagen and proteoglycan content in the thickened cusps of four of the seven mice with abnormal valves relative to normal TAVs (Fig. 3), consistent with other congenital valvular abnormalities (12, 17, 31). The cusps of Sievers type-0 BAVs were also somewhat thicker than those of TAVs, but with similar cellularity and extracellular matrix protein content (Fig. 3). Aberrant tissue stresses and hemodynamic forces experienced by abnormal valves likely drive maladaptive tissue remodeling within the valve cusps to exacerbate cusp thickening, and consequently stenosis as the mice age (28, 31, 32). Effective average AVA was significantly lower in juvenile Npr2^+/−;Ldlr^−/− mice with valvular abnormalities compared with those with normal TAVs, indicating mild congenital stenosis. However, there was sizeable overlap in the distributions of AVAs between normal TAVs and valves with congenital abnormalities, and thus an individual mouse’s AVA cannot serve as a distinguishing criterion of valve morphotype.

Abnormal hemodynamics due to valve malformations can also instigate progressive remodeling in both the aorta and the left ventricle (29, 33,34). For example, the irregular shape of the BAV causes the redirection of blood flow toward the right or left aortic wall, which associates with regions of aortic dilatation or aneurysm formation in adults. Valve abnormalities can also induce regurgitant blood flow back into the left ventricle, which over time leads to left ventricular remodeling and reduced CO and LVFS (34, 35). Thus, BAV-associated aortic and ventricular remodeling is typically acquired in adulthood as opposed to being congenital (5, 34). In our study, ascending aortic diameters were not statistically different between 4- or 8-wk-old mice with normal TAVs versus abnormal valves (Fig. 7B). There were also no statistical differences in CO and LVFS between normal TAVs versus abnormal valves (Fig. 7, C and D). These data indicate that differences in valve morphotypes do not induce pathological changes in the aorta or ventricles by 8 wk of age, and that valve type-specific changes observed in adult Npr2^+/−;Ldlr^−/− mice (18) are acquired, not congenital. These findings collectively support the use of the Npr2^+/−;Ldlr^−/− mouse model for future studies concentrating on acquired aortopathy or remodeling due to BAVs.

This study has some limitations to consider. Npr2^+/−;Ldlr^−/− mice were the only strain used for assessing the feasibility of the echocardiogram protocol. However, the valvular phenotypes and prevalence rates in the Npr2^+/−;Ldlr^−/− mice resemble those reported for other mouse models (27, 28). Consequently, the echocardiography protocol established in this study should be adaptable to other models with BAVs or other congenital valvular malformations. Clinically, BAVs can present with various morphologies (4, 36), including undeveloped cusps, acquired commissural fusion, cusp fusion by raphe, and different fusion patterns from Sievers type-0 BAVs and putative partially fused BAVs observed here. Other valvular abnormalities including normally functioning BAVs (37), potentially discovered with more screening of Npr2^+/−;Ldlr^−/− mice or in other models, may necessitate tuning of the criteria established here. Nonetheless, the current echocardiography protocol distinguishes a range of morphologically abnormal valves with disturbed hemodynamics that differ functionally from normal TAVs, which will enable future studies with this focus. Although the ultrasound-based criteria did not differentiate between TAVs with thickened cusps and partially fused BAVs (Table 2 and Supplemental Fig. S3), the primary goal of this study was to identify congenital valvular abnormalities, not differentiate between them. Finally, we only identified two female mice with abnormal valves (of 18 screened), one with a Sievers type-0 BAV and one with a TAV with thickened cusps. Additional Npr2^+/−;Ldlr^−/− female mice would have to be screened to confirm the 11% (2/18) female BAV prevalence rate and to confirm that female Npr2^+/−;Ldlr^−/− mice can develop putative partially fused BAVs. Additional screening of female mice with abnormal valves would also strengthen our observations of no differences in the metrics analyzed between male and female mice with abnormal valves up to 8 wk of age (Supplemental Figs. S1 and S2).

In conclusion, this study establishes a robust protocol to classify and characterize valve morphologies in 4-wk-old mice. This is an essential first step to studying associations between congenital valve abnormalities and acquired valvulopathies and aortopathies, such as the role of hemodynamics in BAV-associated AAD. The ultrasound protocol can also be adapted by other investigators for investigation of other congenital cardiovascular diseases in small animal models.

DATA AVAILABILITY

Data will be made available upon reasonable request.

SUPPLEMENTAL DATA

Supplemental Figs. S1–S3: https://doi.org/10.5683/SP3/Q6E2AY.

GRANTS

This project was supported by Canadian Institutes of Health Research (CIHR) Project Grant PJT-165817 (to C.A.S.).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

V.G. and C.A.S. conceived and designed research; V.G., Y.-Q.Z., M.T., Z.M., Y.D., and N.L. performed experiments; V.G. and Y.D. analyzed data; V.G., Y.-Q.Z., M.T., M.E., and C.A.S. interpreted results of experiments; V.G. and Y.-Q.Z. prepared figures; V.G., Y.-Q.Z., and C.A.S. drafted manuscript; V.G., Y.-Q.Z., M.T., Z.M., Y.D., M.E., and C.A.S. edited and revised manuscript; C.A.S. approved final version of manuscript.