Received: 19 March 2021 | Revised: 20 October 2021 | Accepted: 25 October 2021 DOI: 10.1111/apha.13736 SIGNALLING β2-­adrenergic stimulation potentiates spontaneous calcium release by increasing signal mass and co-­activation of ryanodine receptor clusters Carme Nolla-­Colomer1 | Sergi Casabella-­Ramon2,3 | Veronica Jimenez-­Sabado3,4 | Alexander Vallmitjana1 | Carmen Tarifa2,3 | Adela Herraiz-­Martínez2,3 | Anna Llach3 | Manel Tauron5 | Jose Montiel5 | Juan Cinca3,6 | S. R. Wayne Chen7 | Raul Benitez1 | Leif Hove-­Madsen2,3 1 Department Automatic Control, Univ. Politècnica de Catalunya, Barcelona, Spain 2 Biomedical Research Institute Barcelona, IIBB-­CSIC, Barcelona, Spain 3 IIB Sant Pau, Barcelona, Spain 4 CIBERCV, Madrid, Spain 5 Department of Cardiac Surgery, Hospital de la Santa Creu i Sant Pau, Barcelona, Spain 6 Universitat Autònoma de Barcelona, Barcelona, Spain 7 Department of Physiology and Pharmacology, University of Calgary, Alberta, Canada Correspondence Leif Hove-­Madsen, Cardiac Rhythm Abstract and Contraction, Biomedical Research Aims: It is unknown how β-­adrenergic stimulation affects calcium dynamics Institute Barcelona (IIBB-­CSIC), Hospital in individual RyR2 clusters and leads to the induction of spontaneous calcium de la Santa Creu i Sant Pau, St Antoni Mª Claret 167, 08025 Barcelona, Spain. waves. To address this, we analysed spontaneous calcium release events in green Email: leif.hove@iibb.csic.es fluorescent protein (GFP)-­tagged RyR2 clusters. Methods: Cardiomyocytes from mice with GFP-­tagged RyR2 or human right Funding information This work was supported by grants from atrial tissue were subjected to immunofluorescent labelling or confocal calcium the Spanish Ministry of Science and imaging. Innovation PID2020-­116927RB-­C21 Results: Spontaneous calcium release from single RyR2 clusters induced and SAF2017-­88019-­C3-­1R (to LHM) and SAF2017-­88019-­C3-­3R (to RB); 91.4% ± 2.0% of all calcium sparks while 8.0% ± 1.6% were caused by release from from Fundació Marató TV3, Marato-­ two neighbouring clusters. Sparks with two RyR2 clusters had 40% bigger ampli- 2015-­2030 (to LHM); from Generalitat de Catalunya SGR2017-­1769 (to tude, were 26% wider, and lasted 35% longer at half maximum. Consequently, LHM); from the Natural Sciences and the spark mass was larger in two-­ than one-­cluster sparks with a median and Engineering Research Council of Canada interquartile range for the cumulative distribution of 15.7 ± 20.1 vs 7.6 ± 5.7 a.u. (to SRWC); from the Canadian Institutes of Health Research (to SRWC); from (P < .01). β2-­adrenergic stimulation increased RyR2 phosphorylation at s2809 the Heart and Stroke Foundation Chair and s2815, tripled the fraction of two-­ and three-­cluster sparks, and significantly in Cardiovascular Research (to SRWC), increased the spark mass. Interestingly, the amplitude and mass of the calcium and from Spanish Ministry of Health See related editorial: Roderick HL. 2022. The cardiomyocyte firestarter— RyR clusters ignite their neighbours after augmentation of Ca 2+ release by β- stimulation Acta Physiol (Oxf). e13798. Raul Benitez, Leif Hove-­Madsen co-­senior authors. This is an open access article under the terms of the Creat​ive Commo​ns Attri​butio​n-­NonCo​mmerc​ial-­NoDerivs License, which permits use and distribution in any medium, provided the original work is properly cited, the use is non-­commercial and no modifications or adaptations are made. © 2021 The Authors. Acta Physiologica published by John Wiley & Sons Ltd on behalf of Scandinavian Physiological Society. Acta Physiologica. 2022;234:e13736.  wileyonlinelibrary.com/journal/apha   |  1 of 17 https://doi.org/10.1111/apha.13736 |    17481716, 2022, 4, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/apha.13736 by Readcube (Labtiva Inc.), Wiley Online Library on [13/11/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License 2 of 17 NOLLA-­COLOMER et al. and Consume CB16/11/00276 (to JC). SC was the recipient of a predoctoral released from a RyR2 cluster were proportional to the SR calcium load, but the grant (FPU18/01250) from the Spanish firing rate was not. The spark mass was also higher in 33 patients with atrial fi- Ministry of Science and Innovation, and brillation than in 36 without (22.9 ± 23.4 a.u. vs 10.7 ± 10.9; P = .015). AL received a PERIS SALUT-­16 grant from Generalitat de Catalunya. Conclusions: Most sparks are caused by activation of a single RyR2 cluster at baseline while β-­adrenergic stimulation doubles the mass and the number of clusters per spark. This mimics the shift in the cumulative spark mass distribu- tion observed in myocytes from patients with atrial fibrillation. KEYWORDS calcium spark, cardiac myocyte, confocal imaging, ryanodine receptor, sarcoplasmic reticulum, β-­adrenergic 1 | I N T RO DU CT ION clusters and the formation of sparks or their fusion into calcium waves capable of triggering spontaneous mem- Ventricular tachycardia and atrial fibrillation (AF) have brane depolarizations. been associated with an elevation of the incidence of We hypothesized that the ability of a local calcium re- calcium sparks and calcium waves,1 which in turn have lease event to trigger neighbouring RyR2 clusters would been shown to induce afterdepolarizations2 capable of depend on its calcium signal mass, which integrates the initiating triggered activity and arrhythmia.3 Among the amplitude, width and duration of the calcium signal. To mechanisms thought to underlie a higher incidence of test this hypothesis, we used transgenic mice with GFP-­ spontaneous calcium release, RyR2 phosphorylation at tagged RyR2s that allowed us to measure calcium release both s28084,5 and s28146,7 that increases the RyR2 open in individual RyR2 clusters and determine how the signal probability4 has been proposed to facilitate the fusion of mass or other features of the calcium release event affects calcium sparks into calcium waves.2 However, it has also the activation of neighbouring RyR2 clusters at baseline been shown that overexpression of phospolamban in trans- and upon beta-­adrenergic stimulation, which is known to genic mice with the R4496C mutation splits large calcium induce cardiac arrhythmia and arrhythmic responses in waves in the R4496C mutants into a large number of mini-­ isolated cardiomyocytes. Moreover, because atrial fibril- waves and calcium sparks, which prevents arrhythmogen- lation (AF) has been associated with changes in the fre- esis in the double mutant mice.8 Thus, the incidence of quency and properties of calcium sparks and attributed calcium sparks or the magnitude of passive calcium leak to increased RyR2 phosphorylation, we also wanted to alone cannot be used unequivocally to estimate the risk of explore whether AF was associated with changes in the arrhythmia. Instead, the ability of individual RyR2 clus- calcium spark mass comparable to those observed under ters to propagate the calcium signal to neighbouring clus- beta-­adrenergic stimulation. ters or to break the propagation of larger calcium release events might be determinant. Because sparks are expected to occur when some or 2 | RESULTS all of the RyR2s in a RyR2 cluster opens, the properties of a spark, and its ability to induce calcium release from 2.1 | Properties, mass and clustering neighbouring RyR2 clusters, is expected to depend on the of calcium signals from individual RyR2 number of RyR2s in a cluster and the fraction of RyR2 that clusters opens.9,10 Presumably, the ability of a spark to propagate its calcium signal will also depend on the number of RyR2 We analysed a total of 3335 RyR2 clusters in 64 cells from clusters that contribute to the spark transient. In this re- 16 mice. Overall, 228 of the 3335 clusters were activated at gard, local non-­propagating calcium release events have least once, giving rise to 467 sparks caused by the activa- been classified according to the magnitude of the resulting tion of a single RyR2 cluster (Figure 1A top) correspond- calcium transient as calcium quarks,11 calcium sparks12 or ing to 91.4% ± 2.0% of all sparks. Moreover, 53 sparks were macrosparks.13,14 Moreover, in a sheep model of AF this caused by the simultaneous activation of two RyR2 clus- arrhythmia was associated with RyR2 cluster fragmenta- ters (Figure 1A bottom) corresponding to 8.0% ± 1.6% of tion and redistribution.14 However, little is known about all sparks. From the 53 double-­cluster sparks, 45 (84.9%) the relationship between the activation of individual RyR2 were caused by RyR2 activations on the same Z-­line while | 17481716, 2022, 4, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/apha.13736 by Readcube (Labtiva Inc.), Wiley Online Library on [13/11/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License NOLLA-­COLOMER et al.     3 of 17 only 8 (15.1%) sparks involved the activation of RyR2 we found only 11 sparks resulting from the activation of clusters on different Z-­lines. Figure 1B illustrates the re- three clusters. As expected, contour plots of the single-­ lationship between single-­cluster signals measured in a and double-­cluster sparks had one and two local maxima 0.5 × 0.5 µm2 region of interest (ROI) and the resulting (Figure 1C-­D). Figure 1E shows the fraction of sparks spark signal measured in a 4 × 4 µm2 ROI. At baseline, with one and two RyR2 clusters. Figure 1F shows the |    17481716, 2022, 4, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/apha.13736 by Readcube (Labtiva Inc.), Wiley Online Library on [13/11/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License 4 of 17 NOLLA-­COLOMER et al. F I G U R E 1 Calcium sparks at baseline are caused by activation of a single or two RyR2 clusters. A, Superimposed image frames of GFP-­ tagged RyR2 clusters (green) and calcium spark signal (red). The colour-­code map below images show the isolated calcium spark signal. Consecutive image frames are shown for a single (top panels) and a double RyR2 cluster spark (bottom panels). Image frames are separated by 8.32ms, white scale bars are 2 µm. B, Calcium signal recordings from individual RyR2 clusters in 0.5 x 0.5 µm2 ROIs (left) and the resulting calcium spark signal in a 4 x 4 µm2 ROI (right) signals are shown for a single (black traces) and a double-­cluster spark (blue traces). C-­D, Two-­ and 3-­D contour plots of single and double-­cluster sparks. E, Fraction of sparks with 1 and 2 RyR2 clusters. F, Kinetics of calcium signals in individual RyR2 clusters from a single (top) and a double RyR2 cluster spark (bottom) Images show cluster location (green) and calcium signal (red). G, Distribution of distances to nearest RyR2 cluster. Dashed line indicates the median distance for all sparks. ‡Indicate P < .001 (rank-­sum test). Number of cells from 16 mice is indicated in parenthesis kinetics of calcium signals in clusters from a single and calcium signal in RyR2 clusters that triggered the activa- double-­cluster spark. The mean time between the first and tion of a second RyR2 cluster (31.72 ± 5.46 a.u.), suggest- the second RyR2 cluster activation was 12.47 ± 1.78 ms. ing that the occurrence of two-­cluster sparks at baseline The distance between clusters in a two-­cluster spark is not determined by the calcium discharged by the initi- was 0.78 ± 0.03 µm for those on the same Z-­line and ating cluster, but more likely depends on the proximity of 1.88 ± 0.06 µm for those on different Z-­lines (Figure 1G). the second cluster to its threshold for releasing calcium When comparing spatial and kinetic properties of spontaneously. single-­ and double-­cluster sparks, we found that double-­ cluster sparks were wider (Figure 2A-­D), lasted sig- nificantly longer (Figure 2E,F), and had larger spark 2.2 | β2-­adrenergic stimulation induce amplitude than those from a single cluster (Figure 2G,H). multi-­cluster activation and larger calcium We did not observe significant differences in the rate of spark mass rise (RoR), 0.036 ± 0.003 (ΔF/Fo)/ms vs 0.040 ± 0.006 (ΔF/Fo)/ms (P = .50), or in the decay of the calcium Because beta-­adrenergic stimulation is known to induce sparks with time constants of 67 ± 10 ms for single-­cluster arrhythmia and arrhythmic responses in cardiomyocytes, sparks and 74 ± 7 ms for double-­cluster sparks (P = .07). we investigated how the β2-­adrenergic agonist fenoterol Interestingly, a comparison of calcium signals in individ- (FENO) affected RyR2 phosphorylation and calcium sig- ual RyR2 clusters during the activation of a single or a nals in individual RyR2 clusters. As shown in Figure 4A, two-­cluster spark revealed no significant difference in the FENO doubled RyR2 phosphorylation at s2809 in mouse amplitude or duration of the calcium signals during the ventricular myocytes with GFP-­tagged RyR2 (Figure 4B). activation of a single-­ or a two-­cluster event. Moreover, FENO also increased RyR2 phosphorylation at s2815 sig- the event frequency was similar in RyR2 clusters that only nificantly (Figure 4C). As shown in Figure 4D, the fraction gave rise to single-­cluster sparks and in those clusters that of de-­phosphorylated RyR2s at s2809 was also halved when also gave rise to two-­cluster events (Figure S1). myocytes were exposed to FENO. Moreover, the selective Since the arrhythmogenic potential of spontaneous protein kinase A (PKA)-­inhibitor KT5720 reverted the ef- calcium release events depends on their ability to trigger fect of FENO on s2809 de-­phosphorylation (Figure 4D,E), neighbouring clusters, ie on the combination of their am- suggesting that β2-­adrenergic stimulation increases PKA-­ plitude, width, and duration; we used the time integral of dependent phosphorylation of the RyR2 at s2809. Analysis the calcium signal in a spark over the baseline contain- of calcium signals in individual RyR2 clusters and calcium ing ROI, namely the spark’s mass, to achieve a measure spark dynamics after exposure to FENO (Figure 5A-­D) of its arrhythmogenic potential (Figure 3A). Figure 3B revealed a four-­fold increase in the frequency of sparks shows that the mass of the calcium spark was 9.1 ± 1.0 triggered by the activation of two RyR2 clusters and a dou- a.u. for single-­cluster sparks and 16.9 ± 3.4 a.u. for double-­ bling of three-­cluster sparks (Figure 5E). Furthermore, cluster sparks (P < .01). Figure 3C shows the probability FENO significantly increased the duration of both 1-­ and distribution of the calcium spark mass for all sparks from 2-­cluster sparks in the presence of FENO (Figure 6A,B). single-­ or double-­cluster events and Figure 3D shows the This, in turn, increased the calcium spark mass of one and cumulative probability distribution of these sparks. The two-­cluster sparks (Figure 6C), shifting both the probabil- median and interquartile range of the distribution was ity distribution and the cumulative probability distribu- 7.6 ± 5.7 a.u. for single-­cluster sparks and 15.7 ± 20.1 a.u. tion towards higher spark masses (Figure 6D). Thus, the for double-­cluster sparks (P < .01). Interestingly, analysis median and interquartile range of single-­cluster distribu- of the mass of the calcium signal for individual RyR2 clus- tion was 13.6 ± 16.6 a.u. in comparison with two-­cluster ters showed no difference between the signal mass of a sparks 20.8 ± 17.8 a.u. For three-­cluster sparks median single-­cluster event (29.77 ± 5.47 a.u.) and the mass of the and interquartile range was 26.8 ± 16.8 a.u. In support of | 17481716, 2022, 4, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/apha.13736 by Readcube (Labtiva Inc.), Wiley Online Library on [13/11/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License NOLLA-­COLOMER et al.     5 of 17 F I G U R E 2 Properties of calcium sparks resulting from the activation of one or two RyR2 clusters. A, Full width of a calcium spark at half maximum (FWHM) was measured from four profiles measured at 45° angles. B, 2D projection of a calcium spark with indication of the lines where calcium spark profiles in A were measured. C, 3D projection of the spark in B. D, FWHM for 1-­ and 2-­cluster sparks. E, Normalized calcium transients for 1-­ and 2-­cluster sparks with indication of the full duration at half maximum (FDHM). F, Average FDHM for 1-­ and 2-­cluster sparks. G, Representative calcium transients for 1-­ and 2-­cluster sparks. H, Average amplitude of 1-­ and 2-­cluster sparks. *P < .05; ‡P < .001 (rank-­sum test). Number of cells from 16 mice is indicated in parenthesis |    17481716, 2022, 4, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/apha.13736 by Readcube (Labtiva Inc.), Wiley Online Library on [13/11/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License 6 of 17 NOLLA-­COLOMER et al. F I G U R E 3 Calcium spark signal mass and the probability distribution for one-­ and two-­cluster sparks. A, Calcium spark traces (left) and their time integral (right) corresponding to the calcium signal mass. Signals are shown for 1-­ and 2-­cluster sparks. B, Average spark signal mass for 1-­ and 2-­cluster sparks. C, Probability distribution of the spark signal masses for 1-­ and 2-­cluster sparks. D, Cumulative probability distribution of the signal masses in C. measurements were from 41 myocytes with one-­cluster sparks and 16 myocytes with two-­ cluster sparks from 16 mice. ‡P < .001 (rank-­sum test). Dotted lines represent smoothing of the data with a kernel density estimate the notion that these effects of FENO on signal mass fa- in individual RyR2 clusters before and after exposure cilitated the fusion of calcium signals from neighbouring to FENO. As shown in Figure 7A and Figure S3, FENO RyR2 clusters, the fraction of clusters that were activated increased both the caffeine-­induced calcium transient during a spontaneous calcium release event (maximal and the number of times a cluster is activated, suggest- path length) increased from 2.9% ± 1.1% in control to ing that the firing frequency of an RyR2 cluster might be 25.5% ± 4.9% with FENO. determined by the SR calcium load. However, analysis of Comparison of calcium signals in individual RyR2 calcium signals in individual RyR2 clusters revealed that clusters during the activation of a single or two-­cluster clusters with spontaneous activity had lower SR calcium spark showed that after exposure to FENO the amplitude load than the clusters without spontaneous activity, and (P < .05) and mass (P < .01) of the calcium signal was the relationship between SR calcium load and RyR2 clus- significantly larger in RyR2 clusters that triggered the ac- ter activity revealed that the activity was not correlated tivation of a second RyR2 cluster than in RyR2 clusters with the load. As shown in Figure 7B, this was true both from single-­cluster sparks. Moreover, the calcium spark for CON and FENO. By contrast, there was a weak posi- frequency was higher in individual RyR2 clusters contrib- tive correlation (r = 0.28, n = 538) between SR calcium uting to two-­cluster events than in RyR2 clusters that only load and spark mass (Figure 7C) and a strong correlation contributed to single-­cluster events (Figure S2). (R = 0.66, n = 538) between SR calcium load and ampli- To test the influence of the SR calcium load on the in- tude (Figure 7D), suggesting that the SR calcium load de- cidence and properties of individual RyR2 cluster activa- termines the amount of calcium released upon activation tions, we measured the caffeine-­induced calcium transient of a RyR2 cluster. Figure 7E shows that the decay of the | 17481716, 2022, 4, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/apha.13736 by Readcube (Labtiva Inc.), Wiley Online Library on [13/11/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License NOLLA-­COLOMER et al.     7 of 17 F I G U R E 4 Beta-­adrenergic stimulation increases RyR2 phosphorylation at s2809 and s2815. A, Confocal images of myocytes with GFP-­ tagged RyR2 (green channel) and immuno-­labelling of s2809 phosphorylated RyR2 (s2809; red channel). Images were recorded in myocytes incubated with control (CON) or 3 µM fenoterol (FENO) solution for 5 minutes. The distribution of S2809-­P/GFP ratios in myocytes exposed to CON and FENO is shown on the right. B, Mean s2809-­P fluorescence intensity normalized to the GFP fluorescence emission for CON and FENO. Number of cells from 4 mice is given for each bar. C, Mean s2815-­P fluorescence intensity normalized to the GFP fluorescence emission for CON and FENO. Number of cells from 5 mice is given for each bar. D, Confocal images of myocytes with GFP-­tagged RyR2 (green channel) and immuno-­labelling of s2809 dephosphorylated RyR2 (s2809-­dP; red channel). Images were recorded in myocytes incubated with CON, FENO or FENO +3 µM KT5720 solution for 5 minutes. E, Mean s2809-­dP fluorescence intensity normalized to the GFP fluorescence emission for CON, FENO and FENO+KT5720. Number of cells from 6 mice is given for each bar. *P < .05 (unpaired t test) calcium transient after activation of a RyR2 cluster was phosphorylation at s2808, similar to the observed effects not correlated with the SR calcium load. Beta-­adrenergic of FENO, we determined the spark mass and its probabil- stimulation with the non-­selective agonist isoproterenol ity distribution in human atrial myocytes from 36 patients (ISO) yielded qualitatively similar results (Figure S4). without and 33 with atrial fibrillation (AF). To determine Neither FENO nor ISO changed baseline calcium signifi- if there were species and tissue dependent differences in cantly (∆F0 was −1.7% ± 8.1% and 3.2% ± 4.7% for FENO RyR2 cluster activation, we also measured the features and ISO respectively). of single-­ and double-­cluster sparks in mouse right and left atrial myocytes. As shown in Table S1 single and double-­cluster events in the atrial myocytes were com- 2.3 | Calcium spark mass in atrial parable among those from left and right atrial chambers fibrillation and similar to those observed in ventricular myocytes. In line with the previous findings,14 Figure 8A,B shows that Because atrial fibrillation is associated with in- sparks were wider, lasted longer, and were more frequent creased spark frequency linked to changes in RyR2 in atrial myocytes from patients with AF (see Figure S5 |    17481716, 2022, 4, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/apha.13736 by Readcube (Labtiva Inc.), Wiley Online Library on [13/11/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License 8 of 17 NOLLA-­COLOMER et al. for details). Consequently, AF was associated with an in- masses larger than 20 a.u. were much more frequent in crease in the spark mass from 16.9 ± 1.9 a.u. in patients AF, being the median and interquartile range of the cu- without AF to 25.4 ± 2.6 a.u. (P < .05) in patients with mulative spark mass distribution 10.7 ± 10.9 a.u. for noAF AF (Figure 8C). The probability distribution of spark and 22.9 ± 23.4 a.u. for AF (P < .05). Comparison of the masses shown in Figure 8D,E revealed that sparks with cumulative probability distributions for human atrial | 17481716, 2022, 4, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/apha.13736 by Readcube (Labtiva Inc.), Wiley Online Library on [13/11/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License NOLLA-­COLOMER et al.     9 of 17 F I G U R E 5 β-­adrenergic stimulation favours induction of sparks with co-­activation of multiple RyR2 clusters. A, Superimposed image frames of GFP-­tagged RyR2 clusters (green) and calcium spark signal (red). The colour-­code map below images show the isolated calcium spark signal. Consecutive image frames are shown for a single (top panels), double (middle panels) and triple RyR2 cluster spark (bottom panels) recorded in the presence of FENO. Image frames are separated by 8.32 ms, scale bar length is 2 µm. B, 3D contour plots of the sparks shown above. C, 2D representation of the same sparks with indication of their area (dashed circles) and individual RyR2 clusters (crosses). D, Calcium signal recordings from individual RyR2 clusters in 0.5 x 0.5 µm ROIs for the single, double and triple-­cluster spark. E, Fraction of all sparks resulting from the activation of 1, 2 or 3 RyR2 clusters. ‡P < .001 (rank-­sum t test). Number of cells from 16 mice is indicated in parenthesis myocytes with the corresponding distributions for sparks with neighbouring clusters, facilitating the initiation of with one or two GFP-­tagged RyR2 clusters revealed that calcium waves. Finally, the spark mass probability distri- the cumulative probability distribution for no AF patients bution in human atrial myocytes from patients without and one-­cluster sparks were similar (Figure 8D). On the and with AF was similar to that observed before and after other hand, the cumulative probability distribution for beta-­adrenergic stimulation respectively, suggesting that AF myocytes reached its maximum at an even slower rate similar mechanisms might underlie the induction of ar- than two-­cluster sparks, suggesting that more clusters rhythmic responses in AF. are recruited in sparks from patients with AF and/or that larger amounts of calcium are released from each cluster in myocytes from these patients (Figure 8E). 3.2 | Calcium sparks and calcium signals Finally, to determine how the distribution of calcium from individual RyR2 clusters spark masses in myocytes without and with AF were com- parable to that of myocytes from mice with GFP-­tagged The frequency and kinetic properties of calcium sparks RyR2s at baseline or with FENO, we compared the cumu- have been widely studied in both mammalian and human lative probability distribution of calcium spark masses in cardiomyocytes to address the influence of pharmaco- the human atrial myocytes with different combinations logical treatments and pathological conditions on RyR2 of 1-­ and 2-­cluster spark mass distributions. As shown in activity and spontaneous calcium release from the SR. Figure 8D bottom, the cumulative probability distribution In particular, heart failure and cardiac arrhythmias have for myocytes from patients without AF was very similar to been associated with an increase in calcium spark fre- the linear combination of 1-­ and 2-­ cluster sparks observed quency.1,2,7,15,16 Different pharmacological treatments, for GFP-­tagged RyR2s at baseline. On the other hand, the such as stimulation of Gs-­protein coupled receptors5,17 bottom panel of Figure 8E shows that the cumulative or angiotensin receptors18 have also been shown to alter probability distribution for myocytes from patients with the amplitude, duration or decay of sparks. Here, we AF was most similar to that of 2-­cluster sparks recorded in show that the properties of the calcium sparks are differ- the presence of FENO. ent in sparks resulting from the activation of one and two RyR2 clusters, with the latter giving rise to macrosparks with significantly larger amplitude, width, and duration. 3 | DI S C USSION Presumably, the combination of the cluster features deter- mines the ability of a calcium spark to trigger larger cal- 3.1 | Main findings cium release events and arrhytmia,10,19 suggesting that the mass of the calcium signal might be a good measure of the Using ventricular myocytes with GFP-­tagged RyR2s, we arrhythmogenic potential of a spark. demonstrate that the majority of calcium sparks at base- The calcium spark kinetics and signal mass, in turn, line occur when a single RyR2 cluster is activated (91.6% are expected to depend on both the number of RyR2s in of all sparks). 8% of the remaining sparks occurred when a RyR2 cluster that is activated and the number of RyR2 two RyR2 clusters were activated. While the signal mass clusters that fire. Indeed, several studies addressing the was doubled in these sparks, the triggering of the second number of RyR2s in a RyR2 cluster, using high resolution RyR2 cluster was not correlated with the magnitude of imaging techniques in small rodents, have shown that the calcium signal in the first cluster, suggesting that two-­ the duration and amplitude of calcium sparks are propor- cluster sparks occur randomly at baseline. By contrast, β2-­ tional to the size of the RyR2 cluster.10 On the other hand adrenergic stimulation increased the fraction of two-­ and in a sheep model of AF, there was a fractioning of RyR2 three-­cluster sparks and activation of a second RyR2 clus- clusters and a reduction in the distance to neighbouring ter concurred with a bigger amplitude of the calcium sig- clusters.14 Here we show that, under control conditions, nal in the first cluster; demonstrating that β2-­adrenergic more than 90% of the sparks occur when a single RyR2 stimulation induces propagation of calcium release along cluster releases calcium spontaneously while the rest of |    17481716, 2022, 4, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/apha.13736 by Readcube (Labtiva Inc.), Wiley Online Library on [13/11/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License 10 of 17 NOLLA-­COLOMER et al. the sparks occurred when spontaneous release from one RyR2 cluster that are activated. Moreover, the mass of the RyR2 triggered release from its neighbour, suggesting that calcium signal was similar in the cluster that triggers cal- at baseline the calcium spark properties and signal mass cium release and the cluster where calcium release was is primarily determined by the number of RyR2s in the triggered, suggesting that the occurrence of two-­cluster | 17481716, 2022, 4, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/apha.13736 by Readcube (Labtiva Inc.), Wiley Online Library on [13/11/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License NOLLA-­COLOMER et al.     11 of 17 F I G U R E 6 β-­adrenergic stimulation increases the signal mass of calcium sparks. A, Calcium signals recorded in the presence of 3 µM FENO from individual RyR2 clusters in 0.5 x 0.5 µm2 ROIs (left) and the resulting calcium spark signal in a 4 x 4 µm2 ROI (right). Signals are shown for a single (black trace), a double (blue traces) or a triple RyR2 cluster spark (grey traces). B, Superimposed traces of sparks with 1 or 2 RyR2 clusters in 41 control cells from 16 mice (black) and 14 cells from 7 mice with FENO (blue). Shown below are the amplitude (amp) and duration of the calcium spark signal at half maximum (FDHM) for sparks with 1 or 2 RyR2 clusters in CON (black) and with FENO (blue). C, Average time integral of the calcium spark signals in CON and with FENO. D, Probability (left) and cumulative probability distributions (right) of the signal masses for sparks with 1, 2 or 3 RyR2 clusters. Dotted lines represent smoothing of the data with a kernel density estimate. *P < .05; †P < .01; ‡P < .001 (rank-­sum test) sparks at baseline is not determined by the calcium dis- the activation frequency of a RyR2 cluster while the SR charged by the initiating cluster, but more likely de- calcium load modulates the amplitude of the calcium re- pends on how close the second cluster is to its threshold leased upon activation of the RyR2 cluster. for releasing calcium spontaneously, also referred to as Interestingly, when subjected to β2-­adrenergic stimula- the threshold for store overload induced calcium release tion, the RyR2 cluster that triggered the activation of a neigh- (SOICR).20 bouring RyR2 cluster had a significantly larger mass than the cluster that was triggered. This pattern indicates a shift from random activation of neighbouring clusters at baseline 3.3 | Neuro-­hormonal modulation of to triggered activation of RyR2 clusters under β2-­adrenergic spark mass and properties stimulation that may facilitate the formation and propaga- tion of calcium waves. In line with this notion, FENO and β-­adrenergic stress has been shown to increase spontane- ISO increased the maximal path length for a spontaneous ous calcium release in mice harbouring mutations in the calcium release event 8.7-­ and 9.9-­fold respectively. RyR215 associated with increased RyR2 activity20,21 and with CPVT in humans.22,23 Hence, β-­adrenergic stress might increase the calcium spark mass by increasing the 3.4 | Relationship between calcium RyR2 activity, the number of active RyR2s in a cluster, or spark mass and atrial fibrillation the activation of multiple RyR2 clusters in each spark. In line with the former, we found that β2-­adrenergic stimu- Because cardiac arrhythmia and heart failure have been lation increases the signal mass of sparks arising from the associated with an increase in RyR2 phosphorylation and activation of a single RyR2 cluster. Moreover, it increased spontaneous calcium release events,1,4,7,15,25,26 the presence PKA-­dependent phosphorylation of the RyR2 at s2809 of these features is often considered to indicate risk of ar- and phosphorylation at s2815 that increases RyR2 open rhythmia. However, overexpression of constitutively active probability,2,4,5 suggesting that the increased spark mass calmodulin kinase, expected to increase RyR2 phosphoryla- in single-­cluster sparks may be caused by increased RyR2 tion at s2815 in mice, was found to suppress calcium spark activity in the cluster. On the other hand, β2-­adrenergic and waves.27 Similarly, overexpression of phospholamban in stimulation also increased SR calcium load, and there was mice harbouring the RyR2 mutation R4496C dramatically a correlation between calcium load and the amplitude of increased spontaneous calcium release events but prevented the released calcium, suggesting increased signal mass in arrhythmogenesis because large calcium waves observed a RyR2 cluster may be caused by both higher SR calcium in R4496C mice were split into smaller but more frequent load and increased RyR2 opening. Moreover, there was a release events, ie with a smaller mass, unable to induce three-­fold increase in double-­ and triple-­cluster sparks in spontaneous electrical activity in the double mutant mice.8 myocytes subjected to β2-­adrenergic stimulation, suggest- Similar to this, the present analysis of spontaneous calcium ing that it doubles the calcium spark mass by increasing release events at the level of individual RyR2 clusters shows both the RyR2 activity in a cluster and by the activation that the cumulative probability distribution of the calcium of multiple RyR2 clusters. In line with previous studies, spark masses in a myocyte population may be more robust suggesting that SR calcium load regulates the calcium than the spark frequency to assess the risk of arrhythmia. In spark frequency,24 we found that beta-­adrenergic stimu- support of these hypotheses the cumulative probability dis- lation increased the global SR calcium load and RyR2 tribution of the spark mass was shifted towards higher spark cluster firing. However, analysis of spontaneous activity masses in atrial myocytes from patients with AF and upon and SR calcium load in individual RyR2 clusters, shown β-­adrenergic stimulation, which also induces calcium waves in Figure 7B, revealed that here was no correlation be- and transient arrhythmia at the cellular level28 and is used to tween SR calcium load and RyR2 cluster activity, sup- identify ectopic foci in patients undergoing ablation of AF.29 porting the notion that RyR2 phosphorylation determines Further support for an association between calcium spark |    17481716, 2022, 4, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/apha.13736 by Readcube (Labtiva Inc.), Wiley Online Library on [13/11/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License 12 of 17 NOLLA-­COLOMER et al. F I G U R E 7 Increased SR calcium load does not increase RyR2 cluster activation. A, Representative caffeine (CAF) induced calcium transients recorded before (CON) and after exposure to fenoterol (FENO). Mean SR calcium load, measured as the amplitude of the transients and mean number of activations per cluster is shown on the right. Number of cells from four mice are given for each bar. B, Calcium signals in single RyR2 clusters in CON and FENO at rest and upon exposure to caffeine. Notice that the amplitude of the caffeine-­ induced calcium transient was smaller in the clusters with spontaneous activity than in those without activity. The relationship between amplitude (SR calcium load) and cluster activity is shown on the right for 269 RyR2 clusters. C-­E, Relationship between SR calcium load and the mass (C), amplitude (D) and decay (E) of calcium signals from single RyR2 cluster activations. ‡P < .001 using paired t test for load and rank-­sum test for activations/cluster | 17481716, 2022, 4, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/apha.13736 by Readcube (Labtiva Inc.), Wiley Online Library on [13/11/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License NOLLA-­COLOMER et al.     13 of 17 mass and arrhythmogenic calcium release is afforded by the contrast, tetracaine, which reduces the frequency but not finding that flecainide, which prevents catecholaminergic the mass of calcium sparks, is unable to prevent arrhythmo- polymorphic ventricular tachycardia in a transgenic mouse genic calcium release events in the same mouse model.30,31 model and in humans,30 increases the frequency of calcium Analysis of the calcium spark mass in human atrial sparks but reduces their mass in the mouse model.31 By myocytes revealed that myocytes from patients without a |    17481716, 2022, 4, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/apha.13736 by Readcube (Labtiva Inc.), Wiley Online Library on [13/11/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License 14 of 17 NOLLA-­COLOMER et al. F I G U R E 8 Calcium spark mass is bigger and more diverse in patients with than without AF. A, Consecutive image frames recorded during a calcium spark in myocytes from a patient without AF (No AF). Colour-­coded calcium sparks are shown in the middle and calcium trace with indication (dashed rectangle) of the spark where images were collected is shown below. Image frames are separated by 11.63 ms, scale bar is 2 µm. B, Same representation as in A from a patient with AF. C, Superimposed spark signals (left) and their time integral (middle) from a patient without and with AF. The average time integral of sparks (mass) is shown on the right for 36 No AF and 33 AF patients that had sparks. *P < .05 (rank-­sum test) D-­E, Cumulative probability distribution of spark masses in patients without (D) and with AF (E). The mass distribution for sparks from mice with 1 (black trace), 2 GFP-­tagged RyR2 clusters (blue trace) or all sparks (brown dotted trace) are superimposed on the no AF distribution. The mass distribution for sparks in mice treated with 3 µM FENO that had 1 (black trace), 2 (blue trace), 3 GFP-­tagged RyR2 clusters (grey trace), or all sparks (purple dotted trace) are superimposed on the AF distribution. Corresponding Q-­Q plots are shown at the bottom for control mice vs patients without AF (D) and for FENO treated mice vs patients with AF (E) history of atrial fibrillation had a calcium spark mass and of the cumulative probability distribution of the spark a cumulative probability distribution of the spark mass masses requires a considerable number of sparks from that was similar to the linear combination of one and two-­ a reasonable number of patients in order to avoid that a cluster sparks recorded at baseline in mouse myocytes with large number of sparks in a few cells or patients biases the GFP-­tagged RyR2s. By contrast, the calcium spark mass spark mass distribution. To minimize this issue, we used a was twice as big in patients with atrial fibrillation and the larger number of myocytes from patients without a history cumulative probability distribution of the spark mass was of AF where calcium sparks were less abundant. shifted from a median of 12a.u. in patients without AF to 23 a.u. in those with the arrhythmia, which is even larger than the linear combination of single, double and three-­cluster 4 | CONC LUSIONS sparks recorded in mouse myocytes with GFP-­tagged RyR2s under β-­adrenergic stimulation. These findings document In conclusion, our findings show that calcium sparks at that β-­adrenergic stimulation mimics the higher incidence baseline are largely caused by the activation of a single and mass of calcium sparks in myocytes from patients with RyR2 cluster in GFP-­tagged mouse ventricular myocytes. AF. However, our findings also suggest that additional Moreover, our results suggest that the calcium spark mass pathological alterations in the calcium homeostasis are re- and the cumulative probability distribution of the spark quired to achieve the accumulated probability distribution mass may be better measures of the arrhythmogenic po- and median spark mass observed in AF. tential of calcium sparks than their frequency. In agree- ment with this notion, calcium spark masses and their cumulative probability distribution in atrial myocytes 3.5 | Study limitations from patients without AF was similar to that observed at baseline in myocytes with GFP-­tagged RyR2s, while β-­ Because of ethical considerations and limitations in the adrenergic stimulation, which is commonly used to induce availability of human atrial tissue samples, isolation of arrhythmogenic calcium release, favoured co-­activation of human atrial myocytes is only feasible in right atrial tissue multiple RyR2 clusters in a spark, increased the maximal samples from patients undergoing surgical interventions path length of propagating calcium release events and in- that require extracorporeal circulation; and some of the creased the spark mass as well as its probability distribution observations from right atrial myocytes might differ from towards values observed in atrial myocytes from patients those of left atrial myocytes. Moreover, we cannot exclude with AF. These findings also suggest that antiarrhythmic that concurrent cardiovascular disease and pharmaco- strategies for the treatment of AF may benefit from identi- logical treatments of the tissue donors might influence fying compounds that shift the cumulative probability dis- our findings in human atrial myocytes. Furthermore, the tribution towards lower calcium spark masses. measurements of calcium signals with confocal imaging are limited by the position of the spark relative to the con- focal plane and presumably out of focus sparks introduce 5 | MATERIALS AND METHO D S some noise in the measurements of cluster activations and spark features. Similarly, our confocal microscope does 5.1 | Myocyte preparation not allow distinguishing between multiple small RyR2 clusters and a supercluster and we cannot exclude that Mouse ventricular myocytes (MVM) were isolated from some clusters with large signal masses could contain sev- transgenic mice overexpressing GFP-­tagged RyR232 using eral small RyR2 clusters with inter-­cluster distances below Langendorff perfusion of the whole heart with an en- the spatial resolution. Finally, a reliable characterization zymatic solution as previously described.32 Mice were | 17481716, 2022, 4, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/apha.13736 by Readcube (Labtiva Inc.), Wiley Online Library on [13/11/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License NOLLA-­COLOMER et al.     15 of 17 sacrificed using cervical dislocation after general anaesthe- Leica Hybrid detector. MVM experiments were performed sia with intraperitoneal administration of ketamine (75 mg/ at 37°C. Beta-­adrenergic stimulation was induced by expos- kg) and Medotomidine (1 mg/kg). The study was approved ing myocytes to the β2-­adrenergic agonist fenoterol (3 µM) by the Bioethical Committee at Hospital de Sant Pau (AF-­ or the non-­selective agonist isoproterenol (ISO). RISK_2015) and authorized by Generalitat de Catalunya. The study conforms to the Guide for the Care and Use of Laboratory Animals published by the US National Institutes 5.3 | Immunofluorescent labelling of Health. Only elongated cardiomyocytes with clear cross striations and without abnormal granulation were used. Isolated GFP-­tagged mouse ventricular myocytes were Human atrial myocytes (HAM). Right atrial myocardial fixed with paraformaldehyde for 10 minutes at room samples were collected from 86 patients undergoing car- temperature. Subsequently, cells were labelled with the diac surgery with extracorporeal circulation. Each patient primary antibodies rabbit anti-­ser2809-­P, rabbit anti-­ gave written consent to obtain a tissue sample that would ser2809-­deP or anti-­ser2815-­P as previously described.34 otherwise have been discarded during the surgical inter- After labelling, cells were washed again and stored at vention. The study was approved by the ethical committee 4ºC until imaging (see supplementary methods for fur- of Hospital de la Santa Creu i Sant Pau, Barcelona (AF-­ ther details). GFP-­tagged RyR2 clusters were visualized RISK_2015) and the investigation conforms to the princi- using excitation at 488nm (attenuated to 5%) and collec- ples outlined in the Declaration of Helsinki. After excision tion of fluorescence emission between 495 and 538 nm. of the atrial sample, it was cut into small fragments and S2809 phosphorylated clusters were visualized using ex- atrial myocytes were isolated as previously described.33 citation at 543 nm (attenuated to 10%) and collection of fluorescence emission between 580 and 750 nm. Total and phosphorylated RyR2 clusters were detected using a 5.2 | Confocal calcium imaging custom-­made algorithm as previously described35 and the degree of RyR2 phosphorylation at s2809 was calculated Human atrial myocytes and mouse ventricular myocytes as the fluorescence intensity ratios s2809/GFP for each were imaged using a resonance-­scanning confocal micro- RyR2 cluster (see supplementary methods). scope (Leica SP5 AOBS) with a 63x glycerol-­immersion objective. Fluorophores were excited with a HeNe laser (100 mW) or an Argon laser (1 mW), and the laser power 5.4 | Image processing was set to 20% of the maximum. HAMs were loaded with 2.5 µM fluo-­4 AM. The fluoro- Images were exported from Leica Image File format using phore was excited at 488 nm and emission was collected be- the Bio-­Formats libraries.36 A custom-­made image pro- tween 500 and 650 nm using a Leica Hybrid Detector. Laser cessing pipeline was designed to automatically detect cal- attenuation was set to 4%. Image sequences of 2497 frames cium release events at different spatial and temporal scales with a 1.7x digital zoom factor were obtained at dimensions and to characterize them using a series of morphological of 512 x 140 pixels with a pixel size of 0.28 x 0.28 μm and and kinetic attributes. The detection algorithms were de- frame time of 11.63 ms yielding a total duration of 29.04 s. veloped using MATLAB (The MathWorks, Boston). The experiments were performed at room temperature. Different event detection and characterization algo- In order to establish the relationship between the ac- rithms were developed to analyse MVM and HAM ex- tivation of individual RyR2 clusters and calcium spark periments. Next, is a simplified itemization of the main frequency or properties, we used the fluorescent calcium processing pipeline (see supplementary methods for fur- indicator Rhod-­2 to monitor the activation of individual ther details). RyR2 clusters in isolated MVM from mice with GFP-­tagged RyR2 clusters. Cells were loaded with 5 µM Rhod-­2 AM for 1. Spatial localization of RyR2 clusters (amplitude thresh- 40 minutes. The digital zoom factor was increased to 8X and olding and detection of local maxima). the image dimension was reduced to 40 x 256 pixels in order 2. Extraction of the average fluorescence signal in a to obtain a good compromise between temporal resolution 0.5 µm x 0.5 µm region of each RyR2 cluster. and pixel size, ie 4.16 ms and 0.1 x 0.1 µm respectively. 3. Detection of RyR2 cluster activations using a continu- GFP was excited with a 488 nm Argon laser attenuated at ous wavelet transform with temporal support of 500 ms 5% and emission was collected between 495 and 535 nm 4. Characterization of individual RyR2 cluster activations with a Leica photomultiplier tube (PMT). Rhod-­2 was ex- (amplitude, kinetics and mass of events). cited with a HeNe laser at 543 nm attenuated at 8% and its 5. Hierarchical clustering in order to group simultaneous emission was collected between 600 and 750 nm with the activations of multiple RyR2 cluster, within a spatial |    17481716, 2022, 4, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/apha.13736 by Readcube (Labtiva Inc.), Wiley Online Library on [13/11/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License 16 of 17 NOLLA-­COLOMER et al. and temporal window of 2 µm and 50 ms, into spark ACKNOWLEDGEMENTS events. The collaboration of the Cardiac Surgery Team at Hospital 6. Measure the average fluorescence signal of spark de la Santa Creu i Sant Pau and the assistance of Mr Sóren events in a 4µm x 4µm region centred at the centroid of M. Almendros with calcium image analysis is greatly the activated RyR2 clusters. appreciated. 7. Characterization of individual sparks (size, amplitude, kinetics and mass). CONFLICTS OF INTEREST Authors have no conflicts of interest to declare. The signal mass of both RyR2 activation and spark events was obtained by calculating the integral of normal- DATA AVAILABILITY STATEMENT ized fluorescence signal above the baseline level, The human data that support the findings of this study are not publicly available Because of privacy or ethical restric- ti tions. Other data are available from the corresponding au- ∫t1 signal mass = F (t) − bl (t) dt. thor upon reasonable request. The baseline signal bl(t) was determined using a lin- ORCID ear interpolation between the initial and final baseline Leif Hove-­Madsen https://orcid. levels of the event, and the temporal window to inte- org/0000-0001-5493-3998 grate the mass starts at the calcium up rise (t1) and lasts four times the FDHM of the spark (ti). The spatial ex- REFERENCES tent of the mass is taken into account in the region of 1. Hove-­Madsen L, Llach A, Bayes-­Genis A, et al. Atrial fibrilla- interest (ROI) used to obtain the average event signal tion is associated with increased spontaneous calcium release F(t), ie, a 0.5 µm x 0.5 µm ROI for the RyR2 cluster ac- from the sarcoplasmic reticulum in human atrial myocytes. Circulation. 2004;110(11):1358-­1363. tivation events and a 4 µm x 4 µm ROI for the spark 2. Voigt N, Li N, Wang Q, et al. Enhanced sarcoplasmic reticulum events. Ca2+ leak and increased Na+-­Ca2+ exchanger function under- Image artefacts, spontaneous transients and calcium lie delayed afterdepolarizations in patients with chronic atrial waves were discarded from the detected RyR2 cluster fibrillation. Circulation. 2012;125(17):2059-­2070. event candidates by requiring a set of predefined mor- 3. Burashnikov A, Antzelevitch C. Reinduction of atrial fibrilla- phological properties (see supplementary methods). In tion immediately after termination of the arrhythmia is medi- particular, the signal amplitude was a critical parameter ated by late phase 3 early afterdepolarization-­induced triggered for separating RyR2 cluster activations from noise and activity. Circulation. 2003;107(18):2355-­2360. 4. Vest JA, Wehrens XH, Reiken SR, et al. Defective cardiac ryan- candidate events were required to have an amplitude that odine receptor regulation during atrial fibrillation. Circulation. was at least twice the standard deviation of the baseline 2005;111(16):2025-­2032. noise. Moreover, events that had a full duration at half 5. Llach A, Molina CE, Prat-­Vidal C, et al. Abnormal calcium han- maximum larger than 250 ms or decay times longer than dling in atrial fibrillation is linked to up-­regulation of adenosine 200 ms were discarded. A2A receptors. Eur Heart J. 2011;32(6):721-­729. 6. Neef S, Dybkova N, Sossalla S, et al. CaMKII-­dependent dia- stolic SR Ca2+ leak and elevated diastolic Ca2+ levels in right 5.5 | Statistical analysis atrial myocardium of patients with atrial fibrillation. Circ Res. 2010;106(6):1134-­1144. 7. Ai X, Curran JW, Shannon TR, Bers DM, Pogwizd SM. Ca2+/ Data were analysed in accordance with Acta Physiologica calmodulin-­dependent protein kinase modulates cardiac ryan- guidelines for good publishing practice37 and data from odine receptor phosphorylation and sarcoplasmic reticulum the RyR2s activations and sparks analysis was analysed Ca2+ leak in heart failure. Circ Res. 2005;97(12):1314-­1322. in MATLAB. In the paper, the reported values of the pa- 8. Bai Y, Jones PP, Guo J, et al. Phospholamban knockout breaks rameters are the average of each cell ± standard error. arrhythmogenic Ca2+ waves and suppresses catecholamin- Distributions are characterized by the median and inter- ergic polymorphic ventricular tachycardia in mice. Circ Res. 2013;113(5):517-­526. doi:10.1161/CIRCR​ESAHA.113.301678 quartile range. For cluster activation and calcium sparks, 9. Soeller C, Crossman D, Gilbert R, Cannell M b. Analysis of ryan- all comparisons between groups were performed using odine receptor clusters in rat and human cardiac myocytes. Proc a non-­parametric Wilcoxon rank-­sum test since the ob- Natl Acad Sci USA. 2007;104(38):14958-­14963. doi:10.1073/ servations were not normally distributed (Kolmogorov-­ pnas.07030​16104 Smirnov normality test). Statistical tests and significance 10. Galice S, Xie Y, Yang Y, Sato D, Bers DM. Size matters: ryano- levels are given in each figure and expressed as P-­values or dine receptor cluster size affects arrhythmogenic sarcoplasmic *, meaning P < .05, † for P < .01 and ‡ for P < .001. reticulum calcium release. J Am Heart Assoc. 2018;7(13):8724. | 17481716, 2022, 4, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/apha.13736 by Readcube (Labtiva Inc.), Wiley Online Library on [13/11/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License NOLLA-­COLOMER et al.     17 of 17 11. Lipp P, Niggli E. Submicroscopic calcium signals as fundamen- 27. Yang D, Zhu WZ, Xiao B, et al. Ca2+/calmodulin kinase II-­ tal events of excitation-­contraction coupling in guinea-­pig car- dependent phosphorylation of ryanodine receptors suppresses diac myocytes. J Physiol. 1996;492(1):31-­38. doi:10.1113/jphys​ Ca2+ sparks and Ca2+ waves in cardiac myocytes. Circ Res. iol.1996.sp021286 2007;100(3):399-­407. doi:10.1161/01.RES.00002​58022.13090.55 12. Cheng H, Lederer WJ, Cannell MB. Calcium sparks: elemen- 28. Fujiwara K, Tanaka H, Mani H, Nakagami T, Takamatsu T. Burst tary events underlying excitation-­contraction coupling in heart emergence of intracellular ca2+ waves evokes arrhythmogenic muscle. Science. 1993;262(5134):740-­744. oscillatory depolarization via the Na+-­Ca2+ exchanger: simul- 13. Bolton TB, Imaizumi Y. Spontaneous transient outward cur- taneous confocal recording of membrane potential and intracel- rents in smooth muscle cells. Cell Calcium. 1996;20(2):141-­152. lular Ca2+ in the heart. Circ Res. 2008;103(5):509-­518. doi:10.1016/S0143​-­4160(96)90103​-­7 29. Isa-­Param R, Pérez-­Castellano N, Villacastín J, et al. Inducibility 14. Macquaide N, Tuan HTM, Hotta JI, et al. Ryanodine recep- of atrial arrhythmias after adenosine and isoproterenol infu- tor cluster fragmentation and redistribution in persistent sion in patients referred for atrial fibrillation ablation. Revista atrial fibrillation enhance calcium release. Cardiovasc Res. Española De Cardiología (English Edition). 2006;59(6):559-­566. 2015;108(3):387-­398. doi:10.1093/cvr/cvv231 doi:10.1016/S1885​-­5857(07)60007​-­3 15. Lehnart SE, Wehrens XH, Laitinen PJ, et al. Sudden death in 30. Watanabe H, Chopra N, Laver D, et al. Flecainide prevents cate- familial polymorphic ventricular tachycardia associated with cholaminergic polymorphic ventricular tachycardia in mice and calcium release channel (ryanodine receptor) leak. Circulation. humans. Nat Med. 2009;15(4):380-­383. doi:10.1038/nm.1942 2004;109(25):3208-­3214. 31. Hilliard FA, Steele DS, Laver D, et al. Flecainide inhibits ar- 16. Voigt N, Heijman J, Wang Q, et al. Cellular and molecular mech- rhythmogenic Ca2+ waves by open state block of ryanodine anisms of atrial arrhythmogenesis in patients with paroxysmal receptor Ca2+ release channels and reduction of Ca2+ spark atrial fibrillation. Circulation. 2014;129(2):145-­156. mass. J Mol Cell Cardiol. 2010;48(2):293-­301. doi:10.1016/j. 17. Ghigo A, Perino A, Mehel H, et al. Phosphoinositide 3-­kinase yjmcc.2009.10.005 γ protects against catecholamine-­induced ventricular arrhyth- 32. Hiess F, Detampel P, Nolla-­Colomer C, et al. Dynamic and irreg- mia through protein kinase A-­mediated regulation of distinct ular distribution of RyR2 clusters in the periphery of live ven- phosphodiesterases. Circulation. 2012;126(17):2073-­2083. tricular myocytes. Biophys J. 2018;114(2):343-­354. doi:10.1016/j. doi:10.1161/CIRCU​LATIO​NAHA.112.114074 bpj.2017.11.026 18. Gassanov N, Brandt MC, Michels G, Lindner M, Er F, Hoppe UC. 33. Hove-­Madsen L, Prat-­Vidal C, Llach A, et al. Adenosine A2A Angiotensin II-­induced changes of calcium sparks and ionic receptors are expressed in human atrial myocytes and mod- currents in human atrial myocytes: potential role for early re- ulate spontaneous sarcoplasmic reticulum calcium release. modeling in atrial fibrillation. Cell Calcium. 2006;39(2):175-­186. Cardiovasc Res. 2006;72(2):292-­302. 19. Kolstad TR, van den Brink J, MacQuaide N, et al. Ryanodine 34. Herraiz-­Martínez A, Llach A, Tarifa C, et al. The 4q25 vari- receptor dispersion disrupts Ca2+ release in failing cardiac my- ant rs13143308T links risk of atrial fibrillation to defective ocytes. eLife. 2018;7:39427. doi:10.7554/eLife.39427 calcium homoeostasis. Cardiovasc Res. 2019;115(3):578-­589. 20. Jiang D, Xiao B, Yang D, et al. RyR2 mutations linked to ven- doi:10.1093/cvr/cvy215 tricular tachycardia and sudden death reduce the threshold for 35. Vallmitjana A, Nolla C, Herraiz-­Martínez A, Hove-­Madsen L, store-­overload-­induced Ca2+ release (SOICR). Proc Natl Acad Benítez R. Spatial localization of ryanodine receptors in human Sci USA. 2004;101(35):13062-­13067. cardiac cells. In: Proceedings of the Annual International 21. Jiang D, Xiao B, Zhang L, Chen SR. Enhanced basal activity of Conference of the IEEE Engineering in Medicine and Biology a cardiac Ca2+ release channel (ryanodine receptor) mutant Society, EMBS. 2015. doi:10.1109/EMBC.2015.7319832 associated with ventricular tachycardia and sudden death. Circ 36. Linkert M, Rueden CT, Allan C, et al. Metadata matters: access Res. 2002;91(3):218-­225. to image data in the real world. J Cell Biol. 2010;189(5):777-­782. 22. Priori SG, Napolitano C, Tiso N, et al. Mutataions in the car- doi:10.1083/jcb.20100​4104 diac ryanodine receptor gene (hRyR2) underlie catechol- 37. Persson PB. Good publication practice in physiology 2019. Acta aminergic polymorphic ventricular tachycardia. Circulation. Physiol. 2019;227(4):13405. doi:10.1111/apha.13405 2001;103(2):196-­200. doi:10.1161/01.CIR.103.2.196 23. Laitinen PJ, Brown KM, Piippo K, et al. Mutations of the cardiac SUPPORTING INFORMATION ryanodine receptor (RyR2) gene in familial polymorphic ven- tricular tachycardia. Circulation. 2001;103(4):485-­490. Additional supporting information may be found in the 24. Cheng H, Lederer MR, Lederer WJ, Cannell MB. Calcium sparks online version of the article at the publisher’s website. and [Ca2+]i waves in cardiac myocytes. Am J Physiol Cell Physiol. 1996;270(1):C148-­C159. doi:10.1152/ajpce​ll.1996.270.1.C148 25. Sossalla S, Fluschnik N, Schotola H, et al. Inhibition of el- How to cite this article: Nolla-­Colomer C, evated Ca2+/calmodulin-­dependent protein kinase II im- Casabella-­Ramon S, Jimenez-­Sabado V, et al. proves contractility in human failing myocardium. Circ Res. β2-­adrenergic stimulation potentiates spontaneous 2010;107(9):1150-­1161. calcium release by increasing signal mass and 26. Reiken S, Wehrens XH, Vest JA, et al. Beta-­blockers re- co-­activation of ryanodine receptor clusters. Acta store calcium release channel function and improve cardiac Physiol. 2022;234:e13736. doi:10.1111/apha.13736 muscle performance in human heart failure. Circulation. 2003;107(19):2459-­2466.