Fei et al., 2016 - Cardiac Light-Sheet Fluorescent Microscopy for Multi-Scale and Rapid Imaging of Architecture and Function. Scientific Reports   6:22489 Full text @ Sci. Rep.

Fig. 3

4-D synchronized images to quantify global longitudinal strain rates and volume change of the ventricle at 100 hours post fertilization (hpf).

(a) Changes in global longitudinal strain rates were quantified during the entire cardiac cycle. (b) The ventricular volume was measured in terms of EDV at 95.4 × 105 μm3 and ESV at 1.5 × 105 μm3, respectively. (c) LSFM images captured the zebrafish hearts in the x-y, x-z, and y-z planes during the cardiac cycle. (d) 4-D synchronized LSFM-acquired images revealed endocardial trabeculation in the x-y, x-z and y-z plane during the cardiac cycle. (e,f) 4-D zebrafish cardiac motion was captured during ventricular diastole and systole. A: Atrium, V: Ventricle. Scale bar: 50 μm

Fig. 4

Rapid 3-D images to recapitulate trabeculated network in response to doxorubicin (Dox) treatment in the adult zebrafish.

(a) A representative wild-type zebrafish heart at 120 dpf. The coronal, sagittal and transverse planes of the heart displayed a compact trabecular network. The atrioventricular valve (AV) was identified (yellow arrows). Scale bars are 200 μm in length. In the rightmost column, a 3-D rendering of the “digital heart” was reconstructed by stacking 500 slices of plane images in volume. The 3-D structure of the “digital heart” can be assessed by arbitrary cropping. (b) A representative Dox-injected zebrafish heart at 120 dpf. The endocardial cavity appeared enlarged and the trabecular network was accentuated. (c) The quantified volume ratios of the myocardium (left) and the ventricle cavity (right) in the whole heart.

Fig. 5

Cardiac LSFM (c-LSFM) imaging of a 1-day neonate mouse heart with enhanced cellular resolution.

(a) The coronal, sagittal, and transverse planes at different depths uncover 3-D architecture. Scale bars are 1 mm in length in all of the sub-graphs. (b) The boxes were cropped from the volume rendering of the reconstructed “digital heart” to reveal the endocardial architecture. (c) The cardiac architecture is compared with the (i) 18 μm light-sheet and 4X/0.13 objective, (ii) 4X/0.13 resolution enhanced images, and (iii) 9 μm light-sheet and 10X/0.3 objective. Magnification from left to right reveals the field of view, lateral, and axial resolving power, followed by the volumetric rendering effects of 3 configurations. Myocardial orientation was resolved in detail in the resolution-enhanced c-LSFM group. All scale bars are 500 μm, except for 50 μm in the rightmost column.

Fig. 6

(a) 3-D LSFM revealed the distinct helical organization of individual cardiomyocyte fibers from the right ventricular wall to septum to left ventricular walls (zones 1, 2, and 3), providing insights into the mechanics of ventricular contraction in RV vs. LV. Endocardial structure of the left atrial appendage revealed the muscular ridge and muscular trabeculation (zone 4). The yellow curved arrows indicate the orientation of cardiomyocyte fibers. (b) Ultrastructure in the RV (zone 1) and LV cavity (zone 2) unravel trabeculation/papillary muscle (zone 1). LV: left ventricle; RV: right ventricle; LA: left atrium.

Fig. S3

Post-image processing of adult zebrafish heart with resolution enhancement by different deconvolution techniques. (a) Zebrafish atrium and ventricle are visualized on one section of raw image. (b) The individual cardiomyocytes were unresolvable from the selected region-of-interest from the atrium due to the optics blurring and under-sampling by the camera. (c) The image was first scaled up with 3X b-spline interpolation to partially recover the information loss from incomplete sampling. (d), (e) and (f) illustrate the interpolated images deblurred by iterative MRNSD, WPL and CGLS deconvolution, respectively. The WPL algorithm generates most effective deblurring to recover sharp and high frequency signals. However, it also generates image discontinuity, likely due to the application of wiener filter. The CGLS algorithm appears to be mild, generating the least degree of deblurring. Of the three resolution enhancement algorithms, the MRNSD method provides optimal trade-off between the resolution enhancement and information preservation.

Fig. S5

Comparison between prior to and post 4-D synchronization algorithm. (a) Before synchronization, zebrafish cardiac contractions at different Z positions were not in the same stage. (b) After synchronization, all Z positions were synchronized in the same cardiac contraction stage. Therefore, stacked images for 3-D reconstruction at certain time points were obtained and provided volume information. Scale bar = 50µm.

Fig. S6

Imaging comparison between c-LSFM and confocal microscopy using 120 dpf zebrafish hearts. (a) The original x-y plane image, reconstructed x-z and y-z plane images obtained from confocal data. (b) The original x-y plane image, reconstructed x-z and y-z plane images obtained from c-LSFM data. The zoomed-in images shown in the right columns (2, 4, 6) indicate the lateral and axial resolving powers of confocal and c-LSFM.

Fig. S7

Comparison of 4-D synchronized images with a combination of 3 different parameters. (a & b) Both low frame rate and low z resolution were inadequately synchronized. Both images demonstrated a crinkled pattern in the cardiac wall. (c) Reducing the capturing number to 5 heartbeats (cardiac cycles) revealed similar synchronization with capturing 10 beats. However, negligible artifact appeared behind the wall (yellow circle). (d & e) Increasing the frame rate to 200fps revealed identical image quality to that of 100fps. Therefore, we selected (d) as the optimal combination for 4-D synchronized imaging parameters. Scale bar = 10µm.

Acknowledgments:
ZFIN wishes to thank the journal Scientific Reports for permission to reproduce figures from this article. Please note that this material may be protected by copyright. Full text @ Sci. Rep.