Cardiac Pacemaker Cells Harness Stochastic Resonance to Ensure Fail-Safe Operation at Low Rates Bordering on Sinus Arrest

Figures

• 

  Figure 1.. Presence of biological noise in intact SAN tissue. A, Low-zoom microscopic view of the entire mouse SAN preparation loaded with the Ca indicator Fluo-4, with anatomical benchmarks: crista terminalis, superior vena cava (SVC), hepatic vein (HV), and inferior vena cava (IVC). B, Higher-magnification view of a region of interest within the central SAN area. Scattered pink spots indicate background local Ca release (LCR) signals (Video S1). C, LCR signals (in red) detected in another SAN tissue preparation in a 2D image (XY) and in the corresponding XT and YT orthogonal projections (line-scan images along the yellow lines) to illustrate the spatiotemporal extent of the LCRs (Video S2). Our novel LCR-detection algorithm based on the PMD/Trefide/PCA signal-enhancement pipeline is described in detail in the Supplemental Material and Figure S1.

• 

  Figure 2.. Resonance spectra of SAN cells revealed by sine-wave protocols. A-D, Representative examples of the effect of sine-wave currents (0.5 to 6 Hz, 25 pA; protocol shown at the bottom) on different cell populations. Original Vm traces and their corresponding intervalograms (in Hz) are shown. Resonance spectra are indicated by orange bands. E-H, Effects of sine-wave amplitude (color bars) on average AP firing rate (y-axis) at each sine-wave frequency (x-axis). Average AP firing rate in response to each sine-wave frequency was compared with that before sine-wave application using repeated-measures ANOVA or the Friedman test (*p < 0.05, **p < 0.01, ***p < 0.001). In addition, the Jonckheere-Terpstra trend test revealed significant trends in average rate change as wave frequency increased (arrows; tested separately and significant at p < 0.05 for 25 and 50 pA amplitudes). I-L, Percentage of cells exhibiting one-to-one capture (y-axis) in response to each sine-wave frequency (x-axis) at each amplitude (colored bars). The horizontal dashed line indicates the 50% level and illustrates the stronger (broader) responsiveness of fast- and moderate-firing cells, especially at higher frequencies.

• 

  Figure 3.. Stochastic resonance in SAN cell populations that spontaneously fired APs at different average rates. A-C, Examples of the effect of noise with different amplitudes (a) on AP firing in representative cells from each population. In all firing groups, noise increased AP firing rate, but it substantially perturbed the rhythmicity of fast-firing cells. D-F, Average AP firing rate was compared statistically before, during, and after noise application at each amplitude, and also among amplitudes (25, 31.25, 62.5, and 125 pA). In all cell populations, noise significantly increased AP firing rate (P < 0.05, repeated-measures ANOVA or Friedman test), except at 25 and 31.25 pA in the fast-firing group. There was no significant difference in AP firing rate between before noise application and after noise application was completed for any amplitude in any cell population. In the fast- and moderate-firing populations (D and E), 125 pA noise significantly increased the average AP firing rate compared with 25, 31.25, or 62.5 pA noise (p < 0.05, one-way ANOVA or Kruskal-Wallis test). In the slow-firing population (F), AP firing rate increased stepwise as noise amplitude increased. G-I, Similar…

• 

  Figure 4.. Stochastic resonance in dormant SAN cells. A, Representative examples of the effect of noise of different amplitudes on the AP firing rate. Before noise application dormant cells had subthreshold membrane potential oscillations (zoom-in, red squares) which were substantially increased in the presence of noise, reaching AP threshold and triggering AP firing, with the rate increasing as the noise amplitude increased. Moreover, the amplitude of noise-triggered APs substantially varied in each AP cycle, especially at larger amplitudes of noise. B, Distributions of the stochastic resonance effect for each noise amplitude tested. The noise awakened all dormant cells at noise amplitudes exceeding 25 pA. C, Average rate of AP firing (y-axis) caused by stochastic resonance increased as the noise amplitude increased (Kruskal-Wallis test). * p < 0.05, ** p < 0.01, *** p < 0.001.

• 

  Figure 5.. Resonance spectrum (functional range) and stochastic resonance in slow-firing cells during cholinergic receptor stimulation with carbachol. A, Original Vm recordings from the same cell in the basal state and in carbachol during the sine-wave protocol. The resonance spectrum shifted toward lower rates in the presence of carbachol. B, Intervalograms of the recordings in panel A. C, Statistical analysis of the effect of 25- and 50-pA sine waves on average AP firing rate (*p < 0.05, **p < 0.01, paired t-test for during-wave versus before-wave). D, Original Vm recordings from the same cell in the basal state and in carbachol during the noise protocol. Stochastic resonance increased AP firing rate in a representative slow-firing cell in the presence of carbachol. E, Corresponding intervalograms of the recordings in panel D. F, Statistical analysis of the effect of different noise amplitudes on average AP firing rate in cells under carbachol. The rate increased significantly at all amplitudes tested (P values are shown above the bars; paired t-test for during-noise versus before-noise).

• 

  Figure 6.. Resonance spectrum and stochastic resonance in dormant cells during cholinergic receptor stimulation with carbachol. A, Effects of a 20-pA sine-wave protocol in a representative cell in the basal state and after it became dormant in the presence of carbachol. The maximum resonance frequency shifted (red boxes) toward lower rates under carbachol (1 Hz) versus the basal state (2 Hz). B, intervalograms of recordings in A. C, Statistical analysis of the effect of sine waves of various frequencies and amplitudes (25 and 50 pA) on average AP firing rate (*p < 0.05, **p < 0.01, paired t-test for before-wave versus during-wave). D, Example of the noise effect on a dormant cell in the basal state and in the same cell during cholinergic receptor stimulation. The inset shows noisy intrinsic cell signals present under carbachol before noise was applied. E, Statistics for average AP firing frequency during noise at different amplitudes in the dormant-cell population. Statistical significance was determined by unequal-variance t-tests between pairs of different noise amplitudes (P values are indicated above each bar). F, Percentage of dormant cells awakened by a given noise amplitude.

• 

  Figure 7.. Stochastic resonance is amplified by the coupled-clock system in silico and in real cells. A, Responses of a dormant-cell model to noise of different amplitudes. B, Detailed comparison of cell signaling before and during 20 pA noise application before the first AP occurred (outlined by the box in panel A; see also Video S7). Synchronized CRU firing (green, downward black arrows), INCX (blue), ICaL (magenta), and membrane-potential (Vm) oscillations (orange) of increased amplitude, culminating in AP generation. C, Representative example of stochastic resonance in a representative dormant cell (see more examples in Figure S8) obtained in experimental simultaneous recordings of Vm and Ca signals (top panel), with the corresponding cross-correlation function for subthreshold Vm and Ca signaling before and during noise application (before the first AP fired). The time window chosen for cross-correlation analysis is shown by the orange square. D, Cross-correlation of subthreshold signals increased significantly during noise application, culminating in subsequent AP generation. E, Log-scale histogram revealing the architectural shift in LCR path areas between pre-noise (blue) and…

• 

  Figure 8.. Insights from single-cell numerical modeling: stochastic resonance expands the parametric space of AP firing in the basal state and during cholinergic receptor stimulation. A-D, Each panel represents the result of numerical simulations of 5,917 (97 × 61 in xy) models with different coupled-clock parameters. ICaL conductance (gCaL) was distributed evenly along the x-axis from 0.28 to 0.52 nS/pF in 0.0025-nS/pF increments. The sarcoplasmic reticulum Ca pumping rate (Pup) was distributed evenly along the y-axis from 0 to 12 mM/s in 0.2-mM/s increments. Each pixel in the diagrams represents a single-cell model, with average AP firing frequency indicated by red shades. Blue areas indicate non-firing models. The blue area decreased substantially in the presence of noise. See Video S9 for Vm dynamics in all models. E-G, Representative examples of white-noise effects in different cell populations, with gCaL and Pup values shown above the Vm traces in each panel. Noise disturbed the rhythmicity of AP firing in fast-firing cells but enhanced the function of slow-firing cells and awakened dormant cells to begin firing APs. The magenta double-headed arrow in panel A indicates the parameter…