High-speed 2D light-sheet fluorescence microscopy enables quantification of spatially varying calcium dynamics in ventricular cardiomyocytes

Figures

• 

  FIGURE 1. T-tubule and distance to nearest tubule maps. (A) Binary tubule map of cell D with the cell outline shown in grey. (B) Pixel distance to nearest tubule (DNT) map of cell D. (C) Tubulated regions of cell D with 0 < DNT ≤ 5 px. (D) De-tubulated regions of cell D with DNT > 5 px. Scalebar: 10 µm.

• 

  FIGURE 2. Calcium spark detection and correlation with cell structure. (A) Centre of mass coordinates (red markers) for a total of 139 detected sparks overlaid on the extracted tubular microstructure of the example Cell X. The cell membrane outline is shown in grey and the manually segmented nuclei are shown in orange. (B) Intensity time profile through the 1,000-frame sequence for the spark indicated by the yellow arrow in the inset shown on the right side of (A). The upper and lower intensity thresholds are indicated by T H and T L respectively (grey lines). (C) A zoom-in of the time region indicated by the purple rectangle in (B). The detected peak is indicated by the blue marker, and the first and last frames above F50 are indicated with yellow. (D) Normalized fluorescence intensity at the time point corresponding to the selected spark’s peak intensity. (E) The 100 by 100 pixel region centred on the spark centroid indicated by the blue square in panel (D) at peak intensity (middle) with the intensity distributions at the timepoints corresponding to F50 before and after the peak shown on the left and right respectively. Scalebar: 10 μm.

• 

  FIGURE 3. Pixel-wise determination of transient T50. (A) Cell average transient (ΔF/F0) for the example cell X as a function of time across the sampled time-sequence including three consecutive transients (T1, T2, T3). The recorded pacing signal at 0.5 Hz is shown in grey. For the third transient T3, cropped to a 525 frame (1.3 s) window (green rectangle), two-dimensional maps for the baseline fluorescence (in digital numbers), peak intensity (in digital numbers) and ΔF max/F o are shown in (B–D) respectively. (E) Intensity time traces through individual pixels indicated with circular markers in panel (G) for T3, and the cell-average time trace for T3. (F) Zoom-in on the rising edge of the transient [indicated by the green rectangle in (E)] with the T HM time points for the selected pixel traces indicated with markers. (G) T50 (ms) maps for the three consecutive transients, with some exemplar repeating features, corresponding to areas of early T50, indicated by white arrows. (H) Average T50 map (ms) and (I) standard deviation in T50 (ms) of the three transients. Scalebar: 10 μm.

• 

  FIGURE 4. Spark parameters compared for epitubular (EpiT, DNT = 0) and paratubular (ParaT, 0 < DNT ≤ 5) regions. (A) Spark parameter distributions, with each datapoint corresponding to the median across epitubular (orange) and paratubular (blue) spark populations within each cell. The horizontal lines indicate the mean ± SD over all cells (n = 60). The levels of significance illustrated in the form of stars are based on paired testing (see Supplementary Table S2). (B) Median spark parameter for epitubular and paratubular spark populations in each cell, averaged for each heart (1–6, see legend). The overall average (“Av”) across the six hearts is shown in black. The error bars represent the standard deviation across the heart averages. When comparing cell-by-cell differences, averaged for each heart using a one-sample t-test, only the spark rate is significantly different between epitubular and paratubular populations.

• 

  FIGURE 5. Comparison of average time-to-half-maximum (T50) for different regions of the cell: the cell (excluding the nucleus), the nucleus, the central IQR and the outer quarters. (A) Transient T50 distributions, with each datapoint corresponding to the median across all pixels within a region of a cell. The horizontal lines indicate the mean ± SD over all cells (n = 40 for cells with identified nuclei for the first category, and all n = 60 cells for the second category. Levels of significance illustrated with stars are based on paired testing (see Supplementary Table S3). (B) Transient T50 in different regions of each cell, averaged over the cells from each heart (1–6, see legend). The overall average (“Av”) across the six hearts is shown in black. The error bars represent the standard errors (SE = SD/√n) across the heart averages. When comparing cell-by-cell differences, averaged for each heart using a one-sample t-test, both comparisons maintained statistical significance, with the nuclei and central regions of the cell having a significantly longer time-to-half-maximum than the rest of the cell and the outer region respectively.

• 

  FIGURE 6. Correlation of the t-tubule structure with transient time-to-half-maximum (A) Time-to-half-maximum (T50, in ms) map for the T50 average over three consecutive transients of cell D with binary t-tubules overlaid in white. (B) T50 correlation with DNT across all pixels in cell (D). Linear fits to tubulated (purple) and detubulated (orange) regions have slopes of 0.99 ms/px, 95% CI [0.94 1.04], and 1.99 ms/px, 95% CI [1.89 2.08], indicating increased positive correlation of T50 with DNT in detubulated regions. (C) Horizontal line profiles (averaged vertically over a 5-pixel width) taken through the T50 (left and top) and DNT (bottom right) maps of cell C. (D) The line profiles through T50 and DNT maps shown in panel (C) are overlaid, indicating earlier transient rise in areas closer to tubules. Scalebar: 10 µm.

• 

  FIGURE 7. Transient T50 and T50/DNT slope for different DNT regions. (A) Transient T50 distributions, with each datapoint corresponding to the median T50 (ms) across all pixels within epitubular (EpiT, DNT = 0), paratubular (ParaT, 0 < DNT ≤ 5) or detubulated (DeT, DNT > 5) regions. The horizontal lines indicate the mean ± SD over all cells (n = 60). None of the differences were statistically significant. (B) Transient T50/DNT slope in different regions of each cell, averaged over the cells from each heart (1–6, see legend). The overall average (“Av”) across the six hearts is shown in black. The error bars represent the standard errors (SE = SD/√n) across the heart averages. When comparing cell-by-cell differences, averaged for each heart using a one-sample t-test, there was no significant difference in T50 and T50/DNT slopes in parts of the cell corresponding to the various DNT categories.

• 

  FIGURE 8. Comparison of spark and transient parameters for LV and RV cardiomyocytes for the restricted n = 24 dataset. The horizontal lines represent the averages across the medians for each cell within each ventricle population (n LV = 10, n RV = 14). The error bars correspond to the standard deviation. The average spark amplitude in left ventricle cardiomyocytes was found to be 25% larger than those in right ventricle cardiomyocytes (p = 0.0016, UT).