High-Frequency Ultrasound Echocardiography to Assess Zebrafish Cardiac Function

Summary

We describe a protocol to assess heart morphology and function in adult zebrafish using high-frequency echocardiography. The method allows visualization of the heart and subsequent quantification of functional parameters, such as heart rate (HR), cardiac output (CO), fractional area change (FAC), ejection fraction (EF), and blood inflow and outflow velocities.

Abstract

The zebrafish (Danio rerio) has become a very popular model organism in cardiovascular research, including human cardiac diseases, largely due to its embryonic transparency, genetic tractability, and amenity to rapid, high-throughput studies. However, the loss of transparency limits heart function analysis at the adult stage, which complicates modeling of age-related heart conditions. To overcome such limitations, high-frequency ultrasound echocardiography in zebrafish is emerging as a viable option. Here, we present a detailed protocol to assess cardiac function in adult zebrafish by non-invasive echocardiography using high-frequency ultrasound. The method allows visualization and analysis of zebrafish heart dimension and quantification of important functional parameters, including heart rate, stroke volume, cardiac output, and ejection fraction. In this method, the fish are anesthetized and kept underwater and can be recovered after the procedure. Although high-frequency ultrasound is an expensive technology, the same imaging platform can be used for different species (e.g., murine and zebrafish) by adapting different transducers. Zebrafish echocardiography is a robust method for cardiac phenotyping, useful in the validation and characterization of disease models, particularly late-onset diseases; drug screens; and studies of heart injury, recovery, and regenerative capacity.

Introduction

The zebrafish (Danio rerio) is a well-established vertebrate model for studies of developmental processes and human diseases¹. Zebrafish have high genetic similarity to humans (70%), genetic tractability, high fecundity, and optical transparency during embryonic development, which allows direct visual analysis of organs and tissues, including the heart. Despite having just one atrium and one ventricle, the zebrafish heart (Figure 1) is physiologically similar to mammalian four-chambered hearts. Importantly, the zebrafish heart rate, electrocardiogram morphology, and action potential shape resemble those of humans more than murine species². These features have made zebrafish an excellent model for cardiovascular research and have provided major insights into cardiac development^3,4, regeneration⁵, and pathologic conditions^1,3,4, including arteriosclerosis, cardiomyopathies, arrhythmias, congenital heart diseases, and amyloid light chain cardiotoxicity^1,4,6. Assessment of cardiac function has been possible during the embryonic stage (1-days post fertilization) through direct video analysis using high-speed video microscopy^7,8. However, zebrafish lose their transparency beyond the embryonic stage, limiting functional evaluations of normal mature hearts and late-onset heart conditions. To overcome this limitation, echocardiography has been successfully employed as a high-resolution, real-time, noninvasive imaging alternative to evaluate adult zebrafish heart function^{9,10,11,12,13,14,15}.

In zebrafish, the heart is located ventrally in the thoracic cavity immediately posterior to the gills with the atrium located dorsal to the ventricle. The atrium collects venous blood from the sinus venosus and transfers it to the ventricle where it is further pumped to the bulbus arteriosus (Figure 1). Here, we describe a physiological, underwater, protocol to assess cardiac function in adult zebrafish by non-invasive echocardiography using a linear array ultrasound probe with a center frequency of 50 MHz for B-mode imaging at a resolution of 30 µm. Since ultrasound waves can easily travel through water, keeping close proximity between the fish and the scanning probe underwater provides enough contact surface for heart detection with no need for ultrasound gel and is overall less stressful for the fish. Although alternative zebrafish echocardiography systems were reported by several authors^9,12,13, here we present the general and most commonly used setup that applies to high-frequency ultrasound in animals.

The method allows high resolution imaging of the adult zebrafish heart, tracing of cardiac structures, and quantification of peak-velocities from Doppler blood flow measurements. We show reliable in vivo quantification of important systolic and diastolic parameters, such as ejection fraction (EF), fractional area change (FAC), ventricular blood inflow and outflow velocities, heart rate (HR), and cardiac output (CO). We contribute to establishing a reliable range of normal healthy adult zebrafish cardiac functional and dimensional parameters to allow a more precise evaluation of pathologic states. Overall, we provide a robust method to assess cardiac function in zebrafish, which has proven extremely useful in establishing and validating zebrafish heart disease models^6,16, heart injury and recovery^10,13, and regeneration^11,12, and can be further used to evaluate potential drugs.

Protocol

All procedures involving zebrafish were approved by our Institutional Animal Care and Use Committee and are in compliance with the USDA Animal Welfare Act.

1. Experimental set-up

1. Setting up the platform for image acquisition
   1. Using small scissors or a scalpel make an incision on a sponge at the 12 o'clock position to hold the fish during scanning. Place the sponge in a glass container (Figure 2A).
      NOTE: The position of the incision should allow enough room to move transducer and also to keep the fish bellow the water line when the platform is tilted for scanning (Figure 2). The incision can vary depending on the size of the fish; however, for a standard size and weight, the incision should be approximately 2.5 cm x 0.7 cm x 0.5 cm (length, width, and depth, respectively). The glass container should be at least 6 cm deep to avoid water leakage while imaging the fish.
   2. Affix the glass box containing the sponge on the ultrasound platform, for instance using double-sided tape. Ensure the glass box is at the center of the platform and firmly attached (Figure 2B).
   3. Tilt the platform forward about 30° using the knob on the left side of the platform holder (Figure 2B). Fill the glass square with 200-250 mL of fish system water containing 0.2 mg/mL tricaine methanesulfonate (MS222).
      NOTE: Tricaine can be prepared as a 4 mg/mL stock solution in Tris 40 mM pH 7 and further diluted to the desired concentration in fish system water; 0.2 mg/mL was found to be the best concentration¹⁶. The 4 mg/mL tricaine stock solution can be stored for a long period of time at -20 °C or at 4 °C for one month.
   4. Insert the transducer within the micromanipulator holder on the working rail station, turning the notch of the transducer towards the operator. Keep the array parallel to the ground with the working side longitudinal with respect to the stage (see Figure 2B). Leave enough room (10 cm on both sides) for the now connected transducer-rail system to move along the x- and y-axes.
   5. Log in to the control software and choose Mouse (Small) Vascular. Create a new study as well as a new series for each animal included in the study. Find the new study button located on the bottom left side of the screen on the browser page (the view starts in B-mode).

2. Handling the Fish

NOTE: Zebrafish used in this study were adult, 11-month-old males of the wild-type strain AB/Tuebingen (AB/TU). Zebrafish were maintained in a stand-alone flow-through aquarium system at 28 °C in a constant light cycle set as 14 h light/10 h dark. Zebrafish were fed twice daily with brine shrimp (Artemia nauplii) and dry food flakes.

1. Using a fish net, transfer the fish into a small tank containing system water with 0.2 mg/mL tricaine. Wait until the fish is fully anesthetized (no movement and no response to touch).
2. Using a plastic teaspoon, gently and quickly transfer the fish into the glass box containing the sponge into the previously made incision with ventral side of the fish facing up.
   NOTE: Make sure the head of the fish is positioned towards the operator (same direction as the notch of the transducer) and at a slightly higher level compared to the rest of the body to achieve better heart visualization.
3. Gently lower the transducer (keeping its original position) using the handle on the rail system, placing it longitudinally and close to the ventral side of the fish with the notch of the transducer facing the operator. Leave 2-3 mm (no more than 1 cm) clearance from the fish. Adjust the platform in respect to the transducer using the micromanipulator in all 3 axes until the fish heart is visualized and then start image acquisition. The angle of the transducer should not be changed during the entire image acquisition (Figure 2C).
   NOTE: As long as there is enough proximity (up to 1 cm), the water on top of the fish will provide a contact surface via liquid surface tension that allows transmission of the ultrasound waves between the probe and the fish. Therefore, there is no need to push the transducer against the fish. Try to complete this step and finish the scan in less than 3 minutes to prevent fish death or a decrease of the heart rate during image acquisition. If needed, use a timer. The heart can be found on the upper side of the screen towards the left side of the eye, which can be easily visualized if moving the x-axis all the way to the right. If there is continued difficulty in finding the heart while in B-Mode, switch to color Doppler mode, which will allow for tracking blood flow (red indicates blood flowing towards the operator) and locating the heart.

3. Image acquisition

NOTE: See Table of Materials for imaging system and image analysis software.

1. Longitudinal View B-Mode
   1. After localizing the heart, select or stay in B-Mode (found at the bottom left side of the touchscreen after having initiated a new series) and reduce the field in order to zoom in and have a closer look at the heart for easier tracing during analysis.
   2. In order to have a closer and clearer view of the heart in B-Mode image acquisition, reduce the field by zooming in. Use the touchscreen to manually narrow the field on both the x- and y-axes.
   3. If needed, enhance the quality/contrast of the image by setting the dynamic range to 45-50 dB. Go to the B-mode controls in the More Controls option and subsequently save the change to Mode Presets. Tap Mode Presets to select the optimized image acquisition setting every time before starting to image a new series.
   4. Take as many images as desired in the long axis plane by selecting Save Image.
      NOTE: More detailed information and training resources on image acquisition can be found at https://www.visualsonics.com/product/software/vevo-lab and https://www.visualsonics.com/Learning-hub-online-video-training-our-users
2. Longitudinal View Pulse Wave
   1. Switch to Color Doppler for blood flow detection (select Color button) and acquisition (found at the bottom left side of the touchscreen after having initiated a new series).
   2. Using the touch screen position the quadrant on top of the atrioventricular valve and localize the inflow, which will be distinguished by the red color signal (Figure 3A). Reduce the quadrant area as much as possible to increase the frame rate.
      NOTE: Lower the Color pulse-repetition-frequency (Color PRF) (velocity range) to ensure yellow color can be seen in the velocity profile of the Color Doppler image. This will increase the range of velocities that can be seen and will help to create a mosaic of color that will allow to visualize more clearly the peak velocities.
   3. Activate pulse wave (select PW) Doppler Mode to sample ventricular blood inflow velocity. Position the sample volume gate at the center of the atrioventricular valve (where the red color signal becomes more yellowish) to detect the maximum flow velocity. Adjust the PW angle on the screen using your fingers so it aligns with the direction of the blood inflow. Press start or update to start sampling the velocity of blood flowing into the ventricle.
      NOTE: Make sure the angle correct line is parallel to the blood flow in order to provide consistent and reproducible results. Placing the angle correct line so it matches the direction of blood flow will ensure that velocities are accurately captured.
   4. Repeat step 3.2.3 to determine the outflow velocity by placing the Color Doppler quadrant at the junction between the ventricle and the bulbus (bulbuventricular valve) and localize the flow, which will be distinguished by a blue color signal (Figure 3B). Position the sample volume gate right before the ventricle-bulbus junction and adjust the angle correction line to match the direction of the blood flow.
      NOTE: As mentioned before, to achieve accurate velocity values, make sure the PW angle is aligned with the blood flow.
   5. Adjust the baseline (bar), lowering or raising it in the flow velocity panel, in order to detect and trace completely the signal peaks (Figure 3C,D). Identify the inflow peaks in the upper/positive quadrant (signal going towards the probe) and the outflow peaks in the lower/negative quadrant (signal going away from the probe).

4. Fish recovery

1. As soon as image acquisition is complete, using a teaspoon, transfer the fish into regular system aerated water free of tricaine and let the fish recover (usually takes 30 s to 2 min to resume gill movement and swimming).
2. To help recovery, squirt water repeatedly over the gills using a transfer pipette to promote aeration of the water and oxygen transfer.

5. Image analysis

1. Open the image analysis software.
2. Select an image and click on the image processing icon (Figure 4). Using the available scale (Figure 4), adjust brightness and contrast of the image to allow clear visualization of ventricular walls or blood flow pattern.
3. Using the B-mode image, open the drop-down list from the PSLAX (parasternal long axis) option on the cardiac package/measurements (Figure 4). Select LV trace and trace the ventricular inner wall at systole and diastole to obtain the ventricular area (VA) in systole (VAs) and diastole (VAd), end diastolic volume (EDV), and end systolic volume (ESV) (Figure 5A,B).
   NOTE: Volume values are extrapolated from 2D image tracings and might deviate from the 3D entity. For all measurements, average at least 3 representative cardiac cycles per animal.
4. Note the stroke volume and ejection fraction that will be automatically calculated and displayed by the software.
   NOTE: Stroke volume, and ejection fraction can also be manually calculated using the formulas
   SV = EDV-ESV
   EF = (EDV-ESV)/EDV
   where SV is stroke volume, EDV is end diastolic volume, ESV is end systolic volume, and EF is ejection fraction
5. Calculate fractional area change using the formula
   FAC = (VAd - VAs)/ VAd
   where FAC is fractional area change, VAd is ventricular area in diastole, and VAs is ventricular area in systole.
6. Calculate the cardiac output using the formula
   CO = HR x SV
   where CO is cardiac output, HR is heart rate, and SV is stroke volume
7. Using the Pulsed Wave Doppler Mode image, measure the inflow blood velocity by selecting the MV Flow option under the cardiac package (Figure 4). Select E or A for early diastole and late diastole, respectively, and determine the peak-velocities on the graph (Figure 3C).
8. Measure the outflow blood velocity by selecting AoV Flow and determine the peaks on the tracing (Figure 3D).
9. Measure the heart rate using 2 different methodologies for a more reliable assessment:
   1. When the heart is visualized on the screen during image acquisition, count the beats within 10 s and multiply it by 6.
   2. Using the Pulse Wave Doppler image on the Vevo LAB software, choose the heart rate button and trace intervals between 3 consecutive aortic flow peaks (Figure 4 and Figure 6).
   3. To export data to a spreadsheet after having traced the LV and the peaks of the blood flow, click on report | export | save as | excel.

Representative Results

The described protocol allows for measurement of important cardiac dimensional and functional parameters, analogous to the technique used in human and animal echocardiography. The B-Mode images allow for tracing of ventricular inner wall in systole and diastole (Figure 5) and obtaining of dimensional data, such as chamber and wall dimensions, and functional data, such as heart rate, stroke volume, and cardiac output as well as parameters of ventricular systolic function, such as fractional area change and ejection fraction (Table 1). Measurements at the level of the atrioventricular valve using color Doppler Mode images also provide ventricular inflow and outflow blood velocities (velocity at which blood fills and exits the ventricle, respectively) (Figure 3 and Table 1).

The parameters obtained in this study were comparable with the ones reported in previous studies using similar experimental conditions^6,16,17 (Table 1), further demonstrating the reproducibility of the method. Overall, we show that using this detailed protocol one can effectively and consistently assess zebrafish cardiac function, which is critical when comparing different cardiac phenotypes during a study.

Figure 1: Illustration of adult zebrafish heart. Blood flow circulation is represented by arrows: the blood flows from the sinus venosus to the atrium and is further transferred to the ventricle, where it is pumped to the bulbus arteriosus. Please click here to view a larger version of this figure.

Figure 2: Fish-imaging chamber. (A) To prepare a fish-imaging "chamber", a sponge with an incision towards one end in a vertical orientation is placed in a glass container. (B) The glass container is then firmly taped on the inclined imaging platform. (C) The transducer is mounted on the manipulator and placed parallel to the incision for correct imaging positioning (the transducer notch is pointing towards the operator). Please click here to view a larger version of this figure.

Figure 3: Atrioventricular inflow (A) and outflow (B) in Color Doppler mode and corresponding Pulsed Wave Doppler to assess velocities of the respective ventricular diastolic wave peaks (C) and ventricular outflow (D). Please click here to view a larger version of this figure.

Figure 4: Image analysis. After image processing (to achieve desired contrast and brightness of the image), measurements can be performed in the PW Doppler mode (left) and B-mode (right) images. To trace the LV wall in the B-mode image, select Cardiac Package from the drop-down menu, go to PSLAX, and select LV Trace. To measure peak velocities in the PW Doppler mode image, select Cardiac Package from the drop-down menu. To measure the ventricular blood inflow velocity, select the MV Flow option and select E or A for early diastole and late diastole, respectively. For determination of the outflow blood velocity, select AoV Flow and AV peak velocity. Please click here to view a larger version of this figure.

Figure 5: B-mode images. (A) B-Mode image of the ventricle (V) in diastole, filled with blood coming from the atrium (A). (B) B-Mode image of the ventricle in systole, ejecting blood through the bulbus arteriosus (B, green tracing). Please click here to view a larger version of this figure.

Figure 6: Pulse Wave Doppler image. A heart rate value can be generated by tracing 3 consecutive aortic flow peaks. The aortic flow peaks can be displayed by selecting the heart rate button in measurements tab in the analysis software. Please click here to view a larger version of this figure.

Parameters, units ± sd	This study	Wang, L. et al, 2017; Lee, L. et al, 2016 & Mishra, S. et al, 2019	Comments/Description
Heart rate (HR), bpm	133 ± 7	118 ± 14 - 162 ± 32	Wild-types AB/ABTU males and females between 3-12 months anesthetized in tricaine 0.2 mg/mL
Fractional area change (FAC)	0.38 ± 0.03	0.29 ± 0.07 - 0.39 ± 0.05
Ejection fraction (EF), [%]	42 ± 7	34 ± 0.04 - 48 ± 0.03
Stroke volume (SV), µL	0.21 ± 0.01	0.18 ± 0.06 - 0.28 ± 0.08
Cardiac output (CO), µL min-1	27.3 ± 1.69	19 ± 9.5 - 36.1 ± 7.8
E peak velocity (early ventricular inflow), mm/s	30 ± 6.8	25 ± 7 - 51 ± 16
A peak velocity (late ventricular inflow), mm/s	152 ± 32	144 ± 36 - 288 ± 54
Ventricular outflow, mm/s	86.6 ± 19	n/a

Table 1: Echocardiographic parameters in adult zebrafish. Values obtained for the cardiac function parameters evaluated in the current study for adult male or female zebrafish between 3 and 12 months anesthetized in a 0.2 mg/mL tricaine solution. A range of the values obtained for the same parameters in previous studies^6,16,17 performed in similar conditions is presented for validation and to help standardize the method.

Discussion

We describe a systematic method for echocardiographic imaging and assessment of cardiac function in adult zebrafish. Echocardiography is the only available non-invasive and most robust method for live adult fish cardiac imaging and functional analysis, and it is becoming increasingly popular in zebrafish cardiovascular research. The amount of time needed is short and allows for high-throughput and longitudinal studies. However, there is considerable variation in the methodology employed and data analysis. Standardization of zebrafish echocardiography is very difficult when so many variables can influence the outcoming parameters. When conducting experimental studies, one should consider conditions that can produce variability, including anesthesia, body weight, age, sex, and background strain. Wang, L et al.¹⁶ assessed the variability introduced by these factors and compiled the available data on zebrafish cardiac function in order to help standardize the method. Their study is a very useful resource to design experimental studies involving zebrafish echocardiographic assessment. Based on the information provided by Wang, L et al.¹⁶ and references within and our own observations⁶, we provide an outline of critical steps and conditions we considered important for protocol optimization and reproducibility:

Choice of specimen: Previous studies suggest that while systolic function parameters (EF, FAC) are not significantly affected by sex differences, diastolic function (namely peak wave E/A ratio) can be considerably lower in females older than 6 months. It was also observed that ventricular areas and volumes significantly increase with fish age (3 months and older) and are considerably higher in females due to their higher body weight and size. Indexing diastolic volumes to body-mass index (BMI) and body surface area (BSA) can help abolish differences between age-matched females and males, and indexing to BSA and weight can help overcome age related diastolic volume differences¹⁶. There were also reports of different diastolic functions between fish with different background strains¹⁶. Overall, when choosing experimental design, it is advisable to use age- and strain-matched controls and avoid mixing different sexes. Using males is recommended, as image quality was lower in gravid females.

Scanning position: In this setup two scanning positions are possible: longitudinal axis and short axis. We found that in short axis mode it was very hard to identify the cardiac chambers. Therefore, we used only longitudinal axis and recommend the latter for delineation of the cardiac chambers in B-mode and derivation of ventricular size and function.

Anesthesia: Adequate sedation is critical to avoid significant bradycardia during measurement. Heart rate will affect cardiac functional measurement, compromising the accuracy of the study. Tricaine is the most common anesthetic agent and a dose of 0.2 mg/mL was found to provide adequate sedation. However, measurement time is critical since heart rate starts to decrease after 3-4 min under sedation¹⁶. To avoid introducing variability, it is critical to keep measurements under 3 min.

Critical parameters: Heart rate can be considered as a critical parameter when aiming for consistency and accuracy. Heart rate should be comparable between experimental groups tested and within the range of values reported for the conditions used. We found that a range of 118 ± 14 to 162 ± 32 bpm can represent the normal values for wild type zebrafish 3-12 months old adults anesthetized with 0.2mg/mL of tricaine for less than 3 min.

Result accuracy: To ensure accuracy, measurements should be taken over a minimum of 3 cardiac cycles. To obtain more accurate manual image tracings, the analysis should be done in a blinded manner.

Besides choosing the most appropriate conditions, several aspects are critical to ensure accurate measurement. Ideally, conditions should be kept as close to the normal fish physiologic state as possible. Performing the scan under water has the advantage of keeping the fish in their natural environment and close to normal conditions for gas exchange, water composition, hydrostatic pressure, and temperature. These are clear advantages over previous studies, where during the scanning fish are placed in a wet sponge exposed to room air and conductivity is enabled by ultrasound gel instead of water^9,10. Underwater scanning also allows for recovery of the fish after the procedure, provided that the time between anesthesia and recovery is kept under 3 min and the fish is returned to recovery water immediately after measurement. To ensure the procedure is performed as quickly and effectively as possible, a considerable amount of time spent on training is advisable before performing experiments.

Echocardiography is a very well established method to evaluate cardiac function in clinical practice as well as in murine (or other mammalians) animal models. However, unlike murine or human echocardiography, performing fish ultrasound underwater does not allow connection of the specimen to the electrodes. Therefore, direct measurement of heart and respiratory rates is not possible. In that case, heart rate can be measured by counting the beats per min in a 10 or 15 min interval or by manually tracing 3 consecutive aortic flow peaks (Figure 6). Heart rate also affects determination of other parameters, such as cardiac output, that have to be calculated manually once parameters such as stroke volume have been obtained through ventricular inner wall tracing. Another aspect to consider is that fish heart morphology is quite different from mammals. In the two-chambered zebrafish heart, ventricular filling is mostly determined by atrial contraction, and fish typically present a much lower early to late ventricular filling ratio when compared to mammals¹⁸. This explains the different profile obtained by pulse wave Doppler in A and E peaks between zebrafish and healthy mammalian hearts.

Echocardiography enables a thorough characterization of the fish cardiac profile and quantification of multiple functional parameters. The values obtained for ejection fraction, fractional area change, blood inflow and outflow velocities, heart rate, and cardiac output are in the range reported by previous studies (Table 1), highlighting the reproducibility of the method. Taken together, our data shows that high-frequency ultrasound echocardiography is a robust and reproducible method to measure zebrafish cardiac morphology and function when evaluating disease models or drug testing.

Materials

Name	Company	Catalog Number	Comments
Double sided tape
Fish net
Glass container - 100 inch high
High frequency transducer	Fujifilm/VisualSonics	MX700	Band width 29-71 MHz, Centre transmit 50 MHz, Axial resolution 30 µm
Plastic teaspoon
Scalpel or scissors
Small fish tanks
Sponge (kitchen sponge)
Transfer pipets (graduated 3 mL)	Samco Scientific	212
Tricaine (MS-222)	Sigma-Aldrich	A5040
Vevo 3100 Imaging system and imaging station	Fujifilm/VisualSonics
Vevo LAB sofware v 1.7.1	Fujifilm/VisualSonics