Fig 1.
Differences between a healthy cardiomyocyte and a cardiomyocyte suffering from heart failure.
Ca2+ -induced Ca2+ release occurs at units called dyads, where t-tubules and SR are in close proximity. In comparison with healthy cardiomyocytes (left), diseases such as HF have been linked to marked subcellular remodeling (right). Reported changes in failing myocytes include loss of T-tubule density, dispersion of RyR clusters, and changes in the spatial pattern of RyR phosphorylation. Created with BioRender.
Fig 2.
Cluster geometries and cytosolic domain.
(A) The five different RyR geometries studied in the model. The red dots show the single RyRs and the tan coloured area represents the junctional SR membrane. All five geometries contain 50 RyRs. The first two geometries (G1 and G2) contain a single cluster to simulate the healthy case. The latter 3 geometries (G3, G4, and G5) differ in the number of CRU sub-clusters, to simulate the disrupted RyR clusters observed during HF. G3, G4 and G5 are organized into 2, 3 and 12 sub-clusters respectively. (B) The computational domain of the cytosol is presented. The space corresponds 1.008 μm x 1.008 μm x 1.008 μm. The grey bars represent the non-junctional SR.
Fig 3.
Schematic of the different phosphorylation setups used in the model demonstrated using the geometry G1 as an example case.
For the three phosphorylation patterns (outer, inner and uniform) two configurations were assumed: 20% (upper row) and 50% (lower row) of the RyRs being phosphorylated.
Table 1.
and
used for the simulations.
Fig 4.
Spark properties of unphosphorylated RyR in geometries G1 and G2.
For each geometry 200 simulations were conducted. (A) Intensity over time for G1 calcium sparks. Dark gray: successful sparks (n = 63), light dotted gray lines: failed sparks. (B) Intensity over time for G2 calcium sparks. Gray: successful sparks (n = 29), light dotted gray lines: failed sparks. (C) The probability of spark generation (spark fidelity) for G1 and G2. Error bars indicate the 95% Agresti-Coull confidence interval. (D) Spark amplitude (mean ± standard error) for successful sparks in G1 and G2. (E) Average time to peak (TTP) (mean ± standard error) for G1 and G2, in ms. (F) Scatter plot of spark amplitude versus TTP. The crosses represent the mean values across all simulations for G1 and G2. (G) Histogram tracking total number of opened RyRs across individual simulations for G1 and G2. For example, for geometry G1 all 50 RyRs are opened in 18 simulations, whereas for G2 no simulations had 50 open RyRs. ***=p<0.001. The p values were calculated from a t-test, except for the fidelity p value which was calculated from Fisher’s exact test (due to binary data).
Fig 5.
Properties of sparks obtained from uniformly distributed phosphorylation pattern versus blanket phosphorylation for G2 and G3.
Red bars represent simulations with a uniformly distributed phosphorylation pattern of 20% (light red) or 50% (dark red) of the receptors. The gray bars represent simulations with a blanket phosphorylation pattern. (A) Fidelity for G2 and G3 at different phosphorylation levels and patterns; the black bars indicate the 95% Agresti-Coull confidence interval. (B) Amplitude (mean ± standard error) is shown using a bar chart (for G2 uniform 20% n = 84 ( 25 μm and
), for G2 uniform 50% n = 117 (
and
), for G2 blanket 20% n = 77 (
45 μm), for G2 blanket 50% n = 108 (
), for G3 uniform 20% n = 50 (
and
), for G3 uniform 50% n = 100 (
and
), for G3 blanket n = 56 (
), and for G2 blanket n = 102(
)). (C) TTP (mean ± standard error) is shown using a bar chart. (D) Spark duration (mean ± standard error) is shown using a bar chart. No statistically significant difference was measured between uniform and blanket phosphorylation.
Fig 6.
Effect of phosphorylation pattern on Ca2+ spark properties in G1.
For each pattern, 200 simulations were conducted. green: inner, red: uniform, blue: outer. Lighter shades: 20% phosphorylation, darker shades: 50% phosphorylation. (A) Spark fidelity for inner, uniform, and outer phosphorylation patterns in G1. The black bars indicate the 95% Agresti-Coull confidence interval. (B) Scatter plot comparing spark duration and amplitude. Circles: single simulations, crosses: mean values. (C) Spark amplitude (mean ± standard error) for inner, uniform, and outer phosphorylation patterns in G1. (For no phosphorylation n = 63, for inner 20% n = 97, for inner 50% n = 125, for uniform 20% n = 94, for uniform 50% n = 121, for outer 20% n = 63, and for outer 50% n = 118). (D) TTP (mean ± standard error) for inner, uniform, and outer phosphorylation patterns in G1. (E) Spark duration (mean ± standard error) for inner, uniform, and outer phosphorylation patterns in G1. (F) Intensity timecourse for successful sparks with inner phosphorylation. (G) Intensity timecourse for successful sparks with uniform phosphorylation. (H) Intensity timecourse for successful sparks with outer phosphorylation. (I) Mean open time for each RyR throughout 200 simulations—inner 50% phosphorylation. (J) Mean open time for each RyR throughout 200 simulations—uniform 50% phosphorylation. (K) Mean open time for each RyR throughout 200 simulations—outer 50% phosphorylation. **=p<0.01 and ***=p<0.001. The p values were calculated with respect to the non phosphorylated case using a t-test, except for the fidelity p value which was calculated from Fisher’s exact test (due to binary data).
Fig 7.
Effect of dispersed and phosphorylated RyR nanoclusters on spark properties.
Green: inner; red: uniform; blue: outer; gray: no phosphorylation. Opaque colours: G2; translucent colours: G3. (A) Spark fidelity for none, inner, uniform, and outer phosphorylation in G2 and G3. The black bars indicate the 95% Agresti-Coull confidence interval. (B) Spark duration (mean ± standard error) for none, inner, uniform, and outer phosphorylation in G2 and G3. (C) Kernel density estimate plot of the amplitudes for the three phosphorylation patterns in G2. (D) Kernel density estimate plot of the amplitudes for the three phosphorylation patterns in G3. (E) Histogram of opened RyRs per simulation for G2 and G3 with no phosphorylation. (F) Geometric visualization of mean open times for each RyR throughout all 200 simulations for G2 and G3 assuming no phosphorylation. (G) Histogram of opened RyRs per simulation for G3 with inner and outer phosphorylation. (H) Geometric visualization of mean open times for G3 with inner and outer phosphorylation. *=p<0.05, **=p<0.01 and ***=p<0.001. The p values were calculated with respect to the non phosphorylated case using a t-test, except for the fidelity p value which was calculated from Fisher’s exact test (due to binary data).
Fig 8.
Spark properties for disrupted geometries G4 and G5.
For each setup 200 single simulations were conducted. Green: inner; red: uniform; blue: outer; gray: no phosphorylation. Light colours: 20% phosphorylation; dark colours: 50% phosphorylation. (A) Activation map with the first opening time of each RyR for a representative simulation for uniform phosphorylation, G4 (left) and G5 (right). ‘x’ indicates that the RyR did not open during the simulation. red circle: initial open RyR. (B) Spark fidelity for none, inner, uniform, and outer phosphorylation in G4. The black bars indicate the 95% Agresti-Coull confidence interval. (C) Spark fidelity for none, inner, uniform, and outer phosphorylation in G5. The black bars indicate the 95% Agresti-Coull confidence interval. (D) Amplitude (mean ± standard error) for none, inner, uniform, and outer phosphorylation in G4 (no phosphorylation n = 3, inner 20% n = 37, inner 50% n = 77, uniform 20% n = 15, uniform 50% n = 60, outer 20% n = 5, outer 50% n = 24). (E) TTP (mean ± standard error) for none, inner, uniform, and outer phosphorylation in G4. (F) Spark duration (mean ± standard error) for none, inner, uniform, and outer phosphorylation in G4. *=p<0.05, **=p<0.01 and ***=p<0.001. The p values were calculated with respect to the non phosphorylated case using a t-test, except for the fidelity p value which was calculated from Fisher’s exact test (due to binary data).
Fig 9.
Heat maps depicting a summary of the Ca2+ spark properties for different phosphorylation patterns.
The rows show the different spark properties (fidelity, amplitude, TTP and spark duration) that were analyzed. The columns represent the different phosphorylation patterns studied in this work. Ca2+ spark properties depend on RyR cluster geometry, cluster integrity, and spatial organization of phosphorylation.