Abstract
Calcium sparks in cardiac myocytes are brief, localized calcium releases from the sarcoplasmic reticulum (SR) believed to be caused by locally regenerative calcium-induced calcium release (CICR) via couplons, clusters of ryanodine receptors (RyRs). How such regeneration is terminated is uncertain. We performed numerical simulations of an idealized stochastic model of spark production, assuming a RyR gating scheme with only two states (open and closed). Local depletion of calcium in the SR was inevitable during a spark, and this could terminate sparks by interrupting CICR, with or without assumed modulation of RyR gating by SR lumenal calcium. Spark termination by local SR depletion was not robust: under some conditions, sparks could be greatly and variably prolonged, terminating by stochastic attrition-a phenomenon we dub "spark metastability." Spark fluorescence rise time was not a good surrogate for the duration of calcium release. Using a highly simplified, deterministic model of the dynamics of a couplon, we show that spark metastability depends on the kinetic relationship of RyR gating and junctional SR refilling rates. The conditions for spark metastability resemble those produced by known mutations of RyR2 and CASQ2 that cause life-threatening triggered arrhythmias, and spark metastability may be mitigated by altering the kinetics of the RyR in a manner similar to the effects of drugs known to prevent those arrhythmias. The model was unable to explain the distributions of spark amplitudes and rise times seen in chemically skinned cat atrial myocytes, suggesting that such sparks may be more complex events involving heterogeneity of couplons or local propagation among sub-clusters of RyRs.