Excitation–contraction coupling (ECC) relies on temporally synchronized sarcoplasmic reticulum (SR) Ca2+ release via ryanodine receptors (RyRs) at dyadic membrane compartments. Neurohormones, such as endothelin-1 (ET-1), that act via Gαq-associated G protein-coupled receptors (GPCRs) modulate Ca2+ dynamics during ECC and induce SR Ca2+ release events involving Ca2+ release via inositol 1,4,5-trisphosphate (InsP3) receptors (InsP3Rs). How the relatively modest Ca2+ release via InsP3Rs elicits this action is not resolved. Here, we investigated whether the actions of InsP3Rs on Ca2+ handling during ECC were mediated by a direct influence on dyadic Ca2+ levels and whether this mechanism contributes to the effects of ET-1. Using a dyad-targeted genetically encoded Ca2+ reporter, we found that InsP3R activation augmented dyadic Ca2+ fluxes during Ca2+ transients and increased Ca2+ sparks. RyRs were required for these effects. These data provide the first direct demonstration of GPCR and InsP3 effects on dyadic Ca2+, and support the notion that Ca2+ release via InsP3Rs influences Ca2+ transients during ECC by facilitating the activation and recruitment of proximal RyRs. We propose that this mechanism contributes to neurohormonal modulation of cardiac function.
The contraction of cardiomyocytes required for the pumping action of the heart is brought about via excitation–contraction coupling (ECC) (Bers, 2002; Gilbert et al., 2020). Underlying this process is the action potential-mediated opening of voltage-gated L-type Ca2+ channels (LTCCs) on the sarcolemma, which evokes a brief Ca2+ influx into the cell that is amplified by ryanodine receptors (RyRs) on the sarcoplasmic reticulum (SR). This process of Ca2+-induced Ca2+ release (CICR) occurs in cellular microdomains termed dyads that are formed by juxtaposition of LTCCs in the sarcolemma and RyRs in the SR within a narrow 12–15 nm junctional cleft (Sun et al., 1995). In ventricular cardiomyocytes, due to T-tubular invaginations of sarcolemma (TTs), these release sites are distributed across the entire volume of the cardiomyocyte, thereby facilitating a synchronized and cell-wide release of Ca2+ from the SR during cell depolarization (Brette and Orchard, 2003).
Through alteration of Ca2+ dynamics, circulating and local mediators regulate the strength of cardiomyocyte contraction to meet hemodynamic needs. Although catecholamines play a pivotal role in this process, other autocrine/paracrine hormones, including endothelin-1 (ET-1) and angiotensin II (Ang II) also contribute (Mayourian et al., 2018). Under disease conditions, where their circulating levels are often increased, these mediators contribute to pathology, for example stimulation of pro-arrhythmic activity, including induction of spontaneous Ca2+ release events (Hiroe et al., 1991; McMurray et al., 1992; Signore et al., 2013; Stewart et al., 1992; Van De Wal et al., 2006; Yorikane et al., 1993).
Inositol 1,4,5-trisphosphate (InsP3)-induced Ca2+ release (IICR) via SR-localized InsP3 receptors (InsP3R) contributes to the effects of Gαq-associated G protein-coupled receptor (GPCR) activation on cardiomyocyte Ca2+ handling. GPCR engagement produces an increase in intracellular InsP3 via phospholipase C-dependent hydrolysis of phosphatidyl inositol 4,5-bisphosphate (Drawnel et al., 2013). Although expression of InsP3R (predominantly the type 2 isoform, InsP3R2, also known as ITPR2) in the heart of various mammalian species is well described, the interactions between IICR and ECC are less consistent, particularly in healthy cardiomyocytes, and shows variation between species (Blanch i Salvador and Egger, 2018; Domeier et al., 2008; Harzheim et al., 2010, 2009; Ljubojevic et al., 2014; Proven et al., 2006; Signore et al., 2013; Smyrnias et al., 2018; Wu et al., 2006; Zima and Blatter, 2004). Indeed, while in rabbit, inotropic effects are observed (Domeier et al., 2008), IICR does not contribute to the inotropic action of ET-1 in rat (Harzheim et al., 2009; Smyrnias et al., 2018). Elsewhere, in human and mouse cardiomyocytes, GPCR/InsP3/InsP3R axis activation augments pacing-evoked Ca2+ transients and cell contraction (Signore et al., 2013). While these effects of InsP3 on healthy cardiomyocytes may be beneficial, the effects of InsP3R activation are amplified in disease contributing to pathological increases in spontaneous Ca2+ release events and to arrhythmogenic Ca2+ transients (Blanch i Salvador and Egger, 2018; Harzheim et al., 2009; Nakayama et al., 2010; Proven et al., 2006; Signore et al., 2013). These deleterious aspects of InsP3 signaling arise due to an increase in expression of InsP3R and levels of GPCRs and their ligands that promote generation of InsP3 (Go et al., 1995; Harzheim et al., 2010; Regitz-Zagrosek et al., 1995; Tsutsumi et al., 1998; Zolk et al., 1999).
Despite the lack of consistency in the effects of InsP3 on Ca2+ transients between species and studies, a Ca2+-mobilizing activity of InsP3 is universally reported (Blanch i Salvador and Egger, 2018; Domeier et al., 2008; Harzheim et al., 2009; Horn et al., 2013). However, these specific actions of InsP3 often require analysis in the absence of the bulk changes in Ca2+ associated with ECC, precluding a full understanding of the mechanism by which IICR modulates ECC-associated Ca2+ transients. Under these conditions, the actions of InsP3 are manifested as an increase in Ca2+ sparks and/or a depletion of the SR Ca2+ store (Blanch i Salvador and Egger, 2018; Domeier et al., 2008; Harzheim et al., 2009; Horn et al., 2013; Wullschleger et al., 2017; Zima and Blatter, 2004). Notably, RyRs are in general required for this activity of InsP3 to be fully manifested. Together, these data indicate that InsP3Rs are functionally expressed in cardiomyocytes and that the flux of Ca2+ via these receptors is small relative to RyRs. Based on these studies, we and others have proposed that despite their minor capacity to mobilize Ca2+ by themselves, through signaling crosstalk with RyRs, InsP3Rs acquire the capacity to influence ECC, including stimulation of pro-arrhythmic activity (Domeier et al., 2008; Harzheim et al., 2009; Wullschleger et al., 2017).
While dyadic colocalization of InsP3Rs and RyRs has been reported, enabling inter channel crosstalk (Harzheim et al., 2009), InsP3Rs have also been suggested to reside on regions of the SR distinct from those occupied by RyRs (Bare et al., 2005; Ljubojevic et al., 2014; Mohler et al., 2005). Although both scenarios could allow for InsP3R–RyR signaling crosstalk, by generating a greater increase in Ca2+ local to the RyR, colocalization of InsP3Rs with RyRs in the dyad would be significantly more effective in promoting channel crosstalk. This could occur either through a direct activation via CICR or by bringing RyRs closer to threshold for activation by Ca2+ arising from voltage-gated Ca2+ channels and/or from Ca2+ arising from neighboring RyRs. Functional evidence supporting a role for IICR in modulation of dyadic RyR and Ca2+ dynamics is, however, lacking.
In this study, we therefore tested the hypothesis that IICR elicits its effects on Ca2+ transients through elevating dyadic Ca2+, thereby either facilitating recruitment of RyRs or by enhancing Ca2+ fluxes via RyR clusters. To address these questions, we deployed a recently described dyad-targeted genetically encoded Ca2+ reporter GCaMP6f–triadin (targeted to the dyad by fusion with triadin) (Shang et al., 2014). In contrast to techniques employing inorganic dyes, this reporter allows the direct measurement of Ca2+ dynamics during ECC in intact cardiomyocytes at individual dyads. Indeed, while linescan imaging of Ca2+ transients can provide insights into regional regulation of Ca2+ release, including by hormonal agonists, a heavy cytosolic buffering to restrict Ca2+ signal to the site of generation is required to shed light on the activity of individual dyads (Song et al., 1998). Although used for analysis of dyadic Ca2+ dynamics during ECC, the capacity for GCaMP6f–triadin to detect hormonally induced changes in Ca2+ release kinetics in the dyad has not, however, previously been tested.
Here, we demonstrated the utility of GCaMP6f–triadin to detect the changes in the expected dyadic Ca2+ fluxes induced by catecholamines and then took advantage of this property to assess the effect of stimulation with ET-1 and InsP3 on dyadic Ca2+ signals in rat ventricular cardiomyocytes. Using live-cell imaging of dyadic Ca2+ transients, we showed that ET-1 increased spatial recruitment and Ca2+ flux at individual release sites, independently of SR Ca2+ load. Importantly, augmented dyadic Ca2+ release was attenuated by InsP3R inhibition. Furthermore, we demonstrated increased frequency of Ca2+ sparks in response to ET-1, using the dyadic reporter, indicating their localization to this domain. Together, our results suggest that IICR signals to RyRs in the dyad, thereby increasing the likelihood of their activation and SR Ca2+ release. We propose that, through this mechanism, ET-1 elicits its effects to increase the fidelity of release site activation and to augment dyadic SR Ca2+ release fluxes.
GCaMP6f–triadin reports Ca2+ changes at individual dyads during the Ca2+ transient
Appropriate targeting of GCaMP6f–triadin to the dyad in rat ventricular cardiomyocytes was first assessed by confocal imaging. Analysis of deconvolved confocal images revealed a striated distribution of GCaMP6f–triadin throughout the cardiomyocyte volume (Fig. 1i). Through examination of GCaMP6f–triadin distribution in cardiomyocytes immunostained with antibodies against the type 2 RyR (RyR2) and the sarcolemmal and TT proteins [NCX (herein referring to the cardic isoform NCX1, also known as SLC8A1) and Cav3], the reporter was found to be mainly associated with RyRs that colocate with TTs (junctional RyRs) (Fig. 1ii). There is a high degree of colocalization of GCaMP6f–triadin with RyRs as exhibited by the substantial overlap of the fluorescence signal of the two fluorescence channels observed in the images (Fig. 1ii) and quantified through Manders’ colocalization analysis, which showed overlap of 90±1.2% (mean±s.e.m.) of GCaMP6f–triadin with RyRs and 73±4.4% of RyR with GCaMP6f. By distance analysis, 70% of GCaMP6f–triadin was found to be located within 0.2 µm of the nearest RyR cluster center (Fig. 1iii). In addition to proximity to RyRs, ∼56% of the GCaMP6f–triadin fluorescence was within 0.2 µm of the skeletonized TTs (Fig. 1iv) (cf. to 48% of RyR). In agreement with the distance-based analysis, Manders’ colocalization analysis of the two labels showed a 64±1.6% overlap of GCaMP6f–triadin with TTs. Together, the overlap of the majority of GCaMP6f–triadin with TTs and RyRs is consistent with its appropriate targeting to dyadic junctions.
Fluorescence responses of GCaMP6f–triadin were next compared with those of non-targeted GCaMP6f, according to the experimental protocol outlined in Fig. 2A. Consistent with reporting of dyadic Ca2+ changes during ECC, GCaMP6f–triadin displayed rapid increases in fluorescence at punctate sites across the cell that tracked electrical pacing (Fig. 2Bi). In comparison with the fluorescence changes of non-targeted GCaMP6f during the Ca2+ transient, cell-averaged fluorescence changes of GCaMP6f–triadin exhibited a more rapid rate of rise and were 2-fold greater in amplitude (Fig. 2Ci,ii). Given that GCaMP6f–triadin Ca2+ affinity is not substantially affected by fusion to triadin (Shang et al., 2014), these data are consistent with greater proximity of GCaMP6f–triadin to dyadic release sites with higher Ca2+ concentration than the non-targeted GCaMP6f.
GCaMP6f–triadin fluorescence changes at individual dyads during Ca2+ release were next analyzed. The maximal first derivative of the fluorescence increase of the reporter during the upstroke of the Ca2+ transient was used as a measure of Ca2+ release flux (the maximal rate of Ca2+ release). Since this measure preceded substantial contraction, it was not affected by cardiomyocyte contraction and dyad movement. As a measure of synchronicity or temporal dispersion of Ca2+ release between dyads, the mean and standard deviation of the latencies of Ca2+ release between individual sites and the first Ca2+ release site detected in a cell were used. These analyses revealed a substantial temporal dispersion in release site activation and in maximal Ca2+ release flux between dyads (Fig. 2Bii; Fig. S2A,B). To further probe the advantage of the dyad targeting for measurements of Ca2+ responses, responses of non-targeted GCaMP6f at dyadic regions demarcated using the membrane stain Di-8-ANEPPS were also analyzed. While heterogeneity of release site activation was detected using GCaMP6f, it was substantially lower than for GCaMP6f–triadin (Fig. S2A). Ca2+ flux measured with GCaMP6f was also significantly lower and showed less variation between release sites than that measured using GCaMP6f–triadin (Fig. S2B).
To explore the potential for using GCaMP6f–triadin to detect changes in dyadic Ca2+ dynamics, its ability to detect alterations in dyadic Ca2+ flux during ECC in response to robust stimulation with β-adrenergic agonist (osoproterenol; Iso) was first investigated. Cell-averaged measurements of GCaMP6f–triadin showed increases in Ca2+ flux and Ca2+ amplitude of the Ca2+ transients following Iso stimulation (Fig. 2Ci,ii). Effects of Iso stimulation were also detected using the non-targeted GCaMP6f but were substantially smaller (Fig. 2Ci,ii). Ca2+ dynamics at individual dyads during Iso stimulation was next investigated. Using GCaMP6f–triadin, significant increases in the sensitivity of SR Ca2+ release onset, synchrony of release site activation, maximal dyadic Ca2+ release flux and the number of active sites were detected in response to Iso (Fig. 2Di–iv). The magnitude of the Iso effects were significantly smaller for the non-targeted GCaMP6f at Di-8-ANNEPS-labeled sites, and no inter-site heterogeneity in activation was detected (Fig. 2Di,ii).
Together, these data highlight the requirement for proximity of the reporter to the dyadic nanodomain to gain insights into Ca2+ release kinetics, free from contamination by indicator mobility, Ca2+ diffusion and cytosolic Ca2+ buffering. These data further indicate that GCaMP6f–triadin is a reliable tool to analyze the spatial and temporal synchrony of Ca2+ release between dyads and their modulation by hormonal agonists.
The activity and number of dyadic Ca2+ release sites during electrically evoked Ca2+ transients are increased by ET-1 independent of SR Ca2+ load
Whether alterations in dyadic Ca2+ dynamics contribute to the actions of ET-1 on ECC-associated Ca2+ transients was next determined according to the protocol indicated (Fig. 3A). Spatially averaged cell-wide measurements of dyad-targeted (Fig. 3B,Ci,ii) and non-targeted GCaMP6f (Fig. S3A) both reported significant increases in the rate of rise and amplitude of the whole-cell Ca2+ transient, consistent with previous findings using inorganic Ca2+ indicators (Harzheim et al., 2009; Proven et al., 2006). As previously described (Proven et al., 2006; Zima and Blatter, 2004), not all cells responded to ET-1 with an alteration in Ca2+ transient properties. Specifically, increases in Ca2+ transient amplitude and maximal Ca2+ release flux were detected in 11/17 and 15/18 cardiomyocytes using the dyad-targeted (Fig. 3Ci,ii) and non-targeted reporters (Fig. S3Aii,iii). In the absence of ET-1, no changes in kinetics or amplitude of the Ca2+ transient were detected with either the dyad-targeted or non-targeted reporter (Fig. S3A,B).
The enhanced dyadic Ca2+ release kinetics in the ET-1-responding cell population could result from enhanced synchronization of Ca2+ release between sites, increased recruitment of release sites and/or increased maximal Ca2+ release flux at individual sites. To test between these possibilities, we analyzed GCaMP6f–triadin responses at individual dyads before and after ET-1 stimulation. Analysis of the mean±s.d. of the latency of Ca2+ release revealed no significant effect of ET-1 on Ca2+ release site synchrony (Fig. 3Ciii,iv). As indicated by the greater number of lines of high GCaMP6f–triadin fluorescence intensity on the xt plot (kymograph) (Fig. 3Bi), ET-1 exposure for 10 min induced a significant increase in the number of active Ca2+ release sites (Fig. 3Cv). No changes in these parameters were observed in control cardiomyocytes perfused with buffer alone for 10 min (Fig. S3C).
Consistent with its enhancement of SR Ca2+ release, ET-1 application also increased maximal Ca2+ release flux at individual dyads (Fig. 3Di). After ET-1 application, the variability in maximal Ca2+ release flux between dyads was, however, significantly increased, as shown in the frequency distribution of maximal Ca2+ release flux at individual release sites (Fig. 3Dii). Whether the greater heterogeneity of maximum Ca2+ release flux following ET-1 application arose due to alteration in the properties of individual release sites or as a result of recruitment of more sites with heterogeneous properties was next investigated. To this end, the effect of ET-1 on the variability of Ca2+ release between sites (inter-site) during a Ca2+ transient and at individual release sites (intra-site) during consecutive beats was determined. Under control conditions, a substantial variation in maximal Ca2+ release flux at individual dyads and between successive beats was observed. Intra-site variability during consecutive Ca2+ transients was, however, greater than inter-site variability during a single Ca2+ transient (Fig. 3Fi–iii). While ET-1 stimulation did not affect the inter-site variability in the maximal Ca2+ release flux [illustrated by the coefficient of variance (CV)] between different Ca2+ release sites (Fig. 3Fii), it significantly reduced the variability in maximal Ca2+ release flux at individual dyads between Ca2+ transients (Fig. 3Fiii).
Whether an increase in SR Ca2+ load contributed to the augmentation of dyadic Ca2+ release by ET-1 was next tested. To this end, the effect of ET-1 on the maximal dyadic Ca2+ release flux induced by caffeine, which induces opening of all RyRs independently of sarcolemmal depolarization (Smith et al., 1988), was next examined. By exhausting the SR Ca2+ store, this approach informs on SR Ca2+ load and location of SR Ca2+ release sites, including those that are not directly coupled to sarcolemmal depolarization. Consistent with its expected mode of action, rapid local application of 20 mM caffeine increased the fluorescence intensity at all dyads demarcated by expression of the targeted Ca2+ biosensor, indicating activation of these release sites and release of available Ca2+ from the intracellular store (Fig. S3Di,ii). The maximal SR Ca2+ release flux following caffeine application was on average 2-fold greater than during electrical pacing (Fig. S3Diii). ET-1 stimulation did not significantly alter the peak of caffeine-induced Ca2+ release (Fig. 3E), consistent with an effect of ET-1 on dyadic Ca2+ independent of an increase in the SR releasable Ca2+.
Together, these data highlight the plasticity of individual dyads and support the notion that ET-1 augments SR Ca2+ release via a mechanism involving recruitment of normally ‘silent’ release sites and by enhancing Ca2+ flux at them, suggesting that more RyRs and/or RyR clusters are activated synchronously among or within release sites.
InsP3R activation underlies increased recruitment of Ca2+ release sites by ET-1
The contribution of InsP3Rs to the enhancement of dyadic Ca2+ release by ET-1 was next determined. Our previous studies show the presence of populations of InsP3Rs that are proximal to RyRs (Harzheim et al., 2009), but their location relative to the dyad and RyRs within the dyad was not determined. Confocal imaging of InsP3Rs and RyRs in cardiomyocytes immunolabeled with antibodies that recognize the type 2 isoforms of these receptors revealed an overlap of 49.9±1.6% (mean±s.e.m.) of the two channel types as well as a proportion of InsP3Rs located between RyR clusters, albeit along the same Z-line (Fig. 4Ai). Consistent with this observation, similar findings regarding the relative distributions of these channels were obtained using a distance-based analysis (Fig. 4Aii). The location of InsP3Rs relative to TTs, was next analyzed by confocal imaging of InsP3Rs labeled using the same anti-type 2 InsP3R antibody used in Fig. 4Ai and TTs immunolabeled using an NCX and Cav3 antibody cocktail. By this analysis, 32±0.7% of InsP3Rs were found to overlap with TTs (Fig. 4Aiii). In line with these data, distance-based analysis of InsP3R and TT localization revealed ∼30% of InsP3Rs to be located <0.2 µm from TTs (Fig. 4Aiv). This observation suggests that, because of close localization, Ca2+ signals through InsP3R could influence Ca2+ release via RyRs clusters. In line with the overlap of InsP3Rs and RyRs with the TTs, immunofluorescent labeling revealed significant colocalization of GCaMP6f–triadin with the InsP3R2 (Fig. 4Ai,ii).
The influence of InsP3Rs on dyadic Ca2+ was examined. To this end, IICR was either stimulated by ET-1 application or suppressed by co-application of the InsP3R inhibitor 2-aminoethoxydiphenyl borate (2-APB) according to the protocol indicated (Fig. 4B). 2-APB was used at a relatively low concentration of 2 µM previously shown to inhibit InsP3Rs without effects on pacing-induced Ca2+ transients in cardiomyocytes (Peppiatt et al., 2003). Consistent with the hypothesis that dyadic Ca2+ dynamics were influenced by ET-1-stimulated IICR, 2-APB abrogated the increased maximal Ca2+ release flux integrated across all dyadic sites during the Ca2+ transient elicited by its application (Fig. 4C,Di). In control experiments, no effects of 2-APB on these parameters were detected (Fig. S4A,B).
The consequence of InsP3R inhibition on maximal Ca2+ release flux and its heterogeneity at individual Ca2+ release sites in cardiomyocytes exposed to ET-1 was next assessed. To allow direct comparison between conditions, measurements at 10 min after treatment were normalized to those before. ET-1 application increased the maximal Ca2+ release flux at individual dyads, which was inhibited by 2-APB (Fig. 4Dii). 2-APB application also reduced the number of active release sites during ECC in ET-1-stimulated cells (Fig. 4Diii). Application of 2-APB alone did not affect dyadic maximal Ca2+ release flux (Fig. S4C–F). Similarly, 2-APB application did not show any effect on SR releasable Ca2+, as determined by caffeine-induced SR Ca2+ release (Fig. 4Div). Together, these data indicate that Ca2+ release via InsP3Rs in the vicinity of dyadic RyRs mediates the increase in maximal Ca2+ flux and sensitivity of local dyadic SR Ca2+ release as well as the spatial recruitment of active release sites observed in ET-1-stimulated cardiomyocytes.
ET-1 stimulation induced more frequent and larger Ca2+ sparks in quiescent ventricular cardiomyocytes
To further examine InsP3R–RyR interactions in dyadic Ca2+ release, elementary Ca2+ release events were analyzed. As Ca2+ sparks are the building blocks of the Ca2+ transient generated via the coordinated activation of RyRs, their analysis provides information regarding how activation of InsP3Rs influences the overall Ca2+ release process. In these experiments, as the Ca2+ reporter is dyadic, Ca2+ release events are highly localized with dyadic RyR clusters, and have been termed nanosparks (Shang et al., 2014).
Ca2+ nanosparks were recorded in quiescent intact ventricular cardiomyocytes, following 10 min pacing, in the presence of buffer or ET-1 (according to the protocol in Fig. 5Ai). Representative recordings of Ca2+ nanosparks are shown in the xt plot images and F/F0 traces in Fig. 5Aii,iii. As shown in the xt plot images, Ca2+ nanosparks were reported at dyadic sites of GCaMP6f–triadin localization. As would be expected following conditioning with low pacing frequency, Ca2+ nanosparks were relatively rare under control conditions (observed in 11/16 cardiomyocytes) (Fig. 5Aii,Bi). The frequency of these events was, however, significantly increased following ET-1 application (15/16 cardiomyocytes displayed Ca2+ release events) (Fig. 5Aiii,Bi). The spatiotemporal properties of the Ca2+ nanosparks were not affected by ET-1, while their associated maximal Ca2+ release flux was significantly larger (Fig. 5Bii,vi). Although, the means of the amplitudes and full width at half maximum (FWHM) of Ca2+ nanosparks were unaffected by ET-1, a significant increase in the variance of these measures was observed (P=0.0001 and P<0.0001 for amplitude and FWHM, respectively). The distribution of FWHM was also broader than in control, revealing a subpopulation of smaller Ca2+ release events (Fig. 5Biii). Interestingly, a population of Ca2+ nanosparks with FWHM ranging between 0.99 and 1.6 µm (those typically observed under control conditions) were of larger amplitude in ET-1-treated cells (Fig. 5C), suggesting an increased recruitment of RyRs. To uncover whether the limited changes in Ca2+ nanospark parameters observed in ET-1-stimulated cells impacted dyadic Ca2+ signaling, their signal mass (1.206×ΔF/F0×FWHM3) (Chandler et al., 2003), which reflects the quantum of Ca2+ released with each event, and spark-mediated Ca2+ leak (product of mass and frequency) were calculated. ET-1 treatment did not, however, significantly alter Ca2+ nanospark mass or associated leak (Fig. 5D).
The requirement for RyRs for the action of ET-1 was next investigated. Tetracaine inhibition of RyRs resulted in an almost complete inhibition of Ca2+ nanosparks in control cells as well as in cells exposed to ET-1 (Fig. 5Bi). In the few remaining Ca2+ release events detected under control and ET-1-stimulated conditions, RyR inhibition also resulted in significant decreases in the amplitude, size (FWHM), duration (full duration half maximum; FDHM) and rise time (TTP). Maximal Ca2+ release flux of these events was not affected however (Fig. 5Bii–vi). As the analysis of Ca2+ release event properties in RyR inhibited cells is based on very few events in a few cells, effects of ET-1 cannot be ruled out.
Together, these data support the notion that ET-1 augments dyadic Ca2+ nanospark frequency via stimulation of Ca2+ release through InsP3Rs and its activation of RyRs.
Direct InsP3R activation evoked more frequent and heterogeneous Ca2+ nanosparks
To probe the functional interaction between InsP3Rs and RyR and its regulation of dyadic Ca2+ signaling independent of other cellular consequences of ET-1 stimulation, Ca2+ release via InsP3Rs was directly induced by InsP3 liberated from a caged precursor (cag-InsP3) (Fig. 6A). Uncaging of InsP3 resulted in an increase in Ca2+ nanospark frequency in individual cells when compared to controls (Fig. 6Bi). No effect of InsP3 uncaging on the amplitude, duration (FDHM), rise time (time to peak, TTP) and maximal Ca2+ release flux of nanosparks was detected, while their size (FWHM) was decreased (Fig. 6Bii–vi). InsP3 uncaging did, however, induce an increase in the variance in amplitude (P=0.0106) and FWHM (P<0.0001) of the Ca2+ nanosparks (Fig. 6Bii,iii). No effect of InsP3 uncaging on Ca2+ nanospark mass and Ca2+ leak was detected (Fig. 6Bvii,viii). The above results suggest that IICR increases the probability of Ca2+ nanospark generation at dyadic sites by both recruiting RyR clusters and more RyRs within a cluster.
The effects of direct stimulation of InsP3Rs on dyadic Ca2+ in the absence of RyRs activation were next examined. To this end, Ca2+ release events were recorded with RyR inhibition according to the protocol outlined in Fig. 6C. Application of tetracaine alone led to an almost complete cessation of Ca2+ release events. The few remaining events had a lower FWHM than those detected in control cells (Fig. 6D; Table 1). Owing to the low number of observed events and their great variability, we could not reliably conclude whether these Ca2+ release events arose from incompletely inhibited RyRs or through the opening of InsP3R clusters (Table 1). However, InsP3 uncaging under conditions of RyR inhibition with tetracaine produced a gradual but significant increase in the fluorescence of the dyadic reporter that was not detected in control cardiomyocytes in which InsP3 was not uncaged (Fig. 6E). This increase in fluorescence is shown as a small increase in fluorescence at dyadic sites (brighter regions) in the xt plots of raw fluorescence, and more clearly, in the profiles of fluorescence intensity normalized to pre photo-stimulation levels and the histograms of averaged data.
Here, we tested the hypothesis that signaling crosstalk between dyadic InsP3Rs and RyRs contributed to the actions of ET-1 on Ca2+ transients during ECC in ventricular cardiomyocytes. Our investigation was enabled through the use of a dyad-targeted GFP-based Ca2+ indicator, which allowed analysis of dyadic Ca2+ with high temporal and spatial resolution. The data generated provides, for the first time, an analysis of modulation of dyadic Ca2+ by neurohormonal agents and their signaling via InsP3Rs in intact cardiomyocytes. Applying this approach, we showed that ET-1 treatment recruits Ca2+ release sites and augments maximal Ca2+ release flux during ECC without modulating SR Ca2+ load. Furthermore, we demonstrated that activation of dyadic InsP3Rs is necessary for the action of ET-1 on kinetics of local Ca2+ signals. Moreover, we show that Ca2+ release via RyRs is required for the effect of IICR to be manifest.
Application of GCaMP6f–triadin to study dyadic Ca2+ signals
While electrophysiological and modeling approaches have provided quantitative insights into dyadic Ca2+ changes on a cell-wide level, analysis of Ca2+ release during ECC, particularly at individual dyads has been challenging (Acsai et al., 2011; Cannell et al., 2013). Contributing to the inaccessibility of dyadic Ca2+ dynamics to analysis by inorganic Ca2+ indicator dyes is dye diffusion and rapid dissipation of dyadic Ca2+ changes (Sipido and Wier, 1991; Song et al., 1998). To deal with these issues, Ca2+ changes have been measured at individual dyads through the use of a high-affinity Ca2+ dye in the presence of strong but slow buffer to restrict Ca2+ diffusion in the cytosol (Song et al., 1998). This technique is, however, disruptive to ECC and requires low frequency electrical stimulation (∼0.1 Hz) to allow SR Ca2+ loading and buffer equilibration. Fundamental aspects of Ca2+ release can also be uncovered through analysis of Ca2+ sparks, although this is generally performed in the absence of electrical pacing. Genetically engineered Ca2+ probes targeted to the dyadic cleft (Despa et al., 2014; Shang et al., 2014) have enabled measurements of Ca2+ dynamics at individual dyads in intact cardiomyocytes under physiological conditions and have thus provided new insights into their activity during ECC. As we show here, dyad movement during cardiomyocyte contraction makes tracking of individual dyads for the full duration of a Ca2+ transient difficult, precluding accurate measurement of transient amplitude. The onset of Ca2+ release and maximal Ca2+ release flux, which provide important information regarding the sensitivity and magnitude of Ca2+ flux central to our study, occurred prior to the onset of substantial contraction, and could thus be used without interference. This allowed us to avoid using pharmacological inhibitors of contraction and their associated issues (Gwathmey et al., 1991; Kolega, 2004). Analysis of these parameters showed significant variability in the activation time of Ca2+ release and maximal Ca2+ release flux at individual dyads during ECC. These data likely reflect the reported differences in RyR density and distribution per cluster, the number of clusters per dyad and location relative to the TT membrane (Dries et al., 2013; Galice et al., 2018; Kolstad et al., 2018). The rapid equalization of Ca2+ across the contiguous SR store makes it unlikely that the heterogeneity of SR Ca2+ release is due to inhomogeneous SR Ca2+ loading (Picht et al., 2011). Consistent with this notion, heterogeneity of Ca2+ release between dyads was significantly reduced when caffeine was used to activate RyRs.
Using GCaMP6f–triadin, we report the first description of an enhancement and spatial synchronization of SR Ca2+ release specifically in the dyad during ECC in response to stimulation with the β-agonist isoproterenol. While recruitment and augmentation of dyadic Ca2+ release has been previously reported using inorganic dyes, these analyses were not based on measurements at individual dyads during ECC. Rather, these data came from regional analysis of Ca2+ release or were from Ca2+ spikes at low pacing frequency. Our new data and analysis therefore directly demonstrate the potential for using this reporter for analysis of changes in dyadic SR Ca2+ release in response to hormonal or pharmacological interventions. Thus, although the Ca2+-binding properties of the reporter were considered not ideal for detection of the large and rapid changes in Ca2+ at the dyad, the detection of increase in spatiotemporal synchrony of SR Ca2+ release, augmented maximal Ca2+ release flux and recruitment of previously inactive dyads by β-agonists would, however, indicate the capacity of the reporter to sense these changes. Furthermore, increases in SR Ca2+ flux and number of active release sites were also detected after caffeine application, which rapidly and synchronously activates SR Ca2+ release, thus reducing variability due to desynchronized activation by Ca2+ entry across the sarcolemma. Considering the relatively high dissociation constant of GCaMP6f–triadin (Kd=632 nM), although lower than of synthetic dyes, and that Ca2+ reaches 10–100 µM in the dyad during CICR (Acsai et al., 2011; Shannon et al., 2004), rapid saturation of the reporter was expected. The relatively slow ‘turn-on’ kinetics (kon∼20 ms) of GCaMP6f (Chen et al., 2013; Helassa et al., 2016), which likely does not allow equilibration to the steady state at saturating but short transient Ca2+ increases, is suggested to allow it to report the Ca2+ changes occurring in the dyad.
Dyadic localization of the indicator was also sufficient to detect β-adrenergic-mediated enhancement of Ca2+ release during ECC in cell-wide analysis of linescan imaging experiments. Despite measurement of fluorescence changes at TT demarcated with a membrane dye, effects of β-adrenergic agonist were less apparent when a non-targeted GCaMP6f was employed. Specifically, in contrast to the more heterogeneous rise in local Ca2+ release at junctions between the TT and SR reported by the dyad-targeted GCaMP6f, Ca2+ signals among dyads across the cell (highlighted by membrane stain) were homogenous when measured by the non-targeted reporter. This difference is, however, likely to be explained by Ca2+ diffusion and the homogenous distribution of the non-targeted reporter. In particular, after release, Ca2+ rapidly diffuses from its source where it is detected by the homogenously distributed reporter. As our measure of latency represents the time difference between the release site studied and the first Ca2+ release detected, the lower latencies of the non-targeted reporter reflect the homogenization of detection of Ca2+ release when using such a reporter. Together, these data highlight the substantial advantages of using a dyad-targeted reporter to measure the influence of hormonal mediators on SR Ca2+ release and moreover to dissect out mechanisms of regulation of individual Ca2+ release sites in this process.
InsP3-induced Ca2+ release raises dyadic Ca2+ facilitating RyR activation
Although IICR modulates Ca2+ transients during ECC following GPCR engagement, the mechanism by which the relatively small Ca2+ flux via InsP3Rs influences the bulk changes in Ca2+ mediated by RyRs during ECC is not fully determined. Indeed, the low conductance of InsP3Rs (∼3-fold lower than RyRs) (Foskett et al., 2007; Zima and Blatter, 2004), together with a substantially lower expression level than RyRs would suggest little capacity to impact global Ca2+. Consistent with this, InsP3R activation in cells in which RyRs are inhibited leads to the generation of either a low frequency of Ca2+ puffs or ‘silent’ events (Harzheim et al., 2009; Horn et al., 2013; Wullschleger et al., 2017; Zima and Blatter, 2004).
Remarkably, by using the dyadic targeted GCaMP6f to analyze Ca2+ release at individual dyads during the Ca2+ transient, we uncovered that ET-1 acted to both recruit more dyads and to augment Ca2+ flux at individual dyads during the Ca2+ transient. ET-1 did not influence dyadic Ca2+ dynamics in every cardiomyocyte, however, and the magnitude of change varied substantially among the cells. This high degree of variability was previously suggested to result from variation between cardiomyocytes in the expression and/or ratio of ETA and ETB receptors (Domeier et al., 2008; Proven et al., 2006; Zima and Blatter, 2004). In addition to enhanced spatial recruitment of dyads during the Ca2+ transient, ET-1 also reduced the variability in dyadic Ca2+ release between Ca2+ transients. These observations could be explained by increased cooperativity of RyRs or RyR clusters activity within the single release site and/or increased Ca2+ sensitivity of RyRs (Galice et al., 2018).
The capacity for ET-1 to influence dyadic Ca2+ was dependent on Ca2+ release via InsP3Rs. The localization of InsP3Rs to dyadic sites, as reported here and previously, supports this activity of InsP3Rs. As shown previously for effects on global Ca2+ transients and Ca2+ sparks, the augmentation of dyadic Ca2+ signaling by IICR downstream of ET-1 was sensitive to InsP3R inhibition with 2-APB. Furthermore, photoliberation of InsP3 from a caged precursor was sufficient to modulate dyadic Ca2+, ruling out the requirement for other downstream effectors of GPCR signaling. Using this approach, the frequency but no other properties, of dyadic Ca2+ release events was increased, suggesting enhanced recruitment of the machinery underlying Ca2+ sparks. In support of this notion, inhibition of RyRs prevented the action of InsP3. These data are in agreement with conclusions drawn in previous studies that did not directly measure dyadic Ca2+ release. The data presented here, however, now demonstrate the important contribution of dyadic InsP3R–RyR channel crosstalk to the mechanism by which InsP3Rs can influence global Ca2+ changes in healthy ventricular cardiomyocytes but that have little effect when acting alone. Several studies, including from our laboratory, have, however, detected Ca2+ release events (Ca2+ puffs) in the presence of RyR inhibition in atrial cardiomyocytes and ventricular cardiomyocytes of different animal species (Harzheim et al., 2009; Wullschleger et al., 2017; Zima and Blatter, 2004). InsP3R expression is, however, greater in the atrial and diseased ventricular cardiomyocytes in which these observations were made. In the present study, we were unable to directly resolve Ca2+ puffs. The amplitude of Ca2+ puffs is ∼20% of Ca2+ sparks (Zima and Blatter, 2004), which approaches the signal-to-noise detection threshold of a confocal microscope. Although, Ca2+ puffs may be generated at clusters containing as few as two InsP3Rs in mammalian cells, a substantial increase in Ca2+ flux and frequency is observed when the number of InsP3Rs forming a cluster is greater (Dickinson et al., 2012). Considering the low abundance of functionally active InsP3R in healthy ventricular cardiomyocytes, it is therefore not surprising that Ca2+ puffs were only rarely detected in the absence of active RyRs. Instead, and consistent with induction of Ca2+ release via InsP3Rs, we observed an increase in baseline Ca2+ levels when InsP3Rs were directly activated by InsP3 under conditions of RyR blockade. This eventless release is reminiscent of that previously described using a diffusible Ca2+ indicator when cardiomyocytes were stimulated with ET-1 or InsP3 (Blanch i Salvador and Egger, 2018; Horn et al., 2013).
The increase in Ca2+ nanospark frequency induced by InsP3 uncaging indicates that the numerous other cellular targets of ET-1, including L-type Ca2+ channels (He et al., 2000; Lauer et al., 1992; Watanabe and Endoh, 1999), K+ channels (James et al., 2001), the Na+/Ca2+ exchanger (Yang et al., 1999; Zhang et al., 2006) and the Na+/H+ exchanger (Kramer et al., 1991), are not required for ET-1 modulation of dyadic Ca2+. InsP3Rs may, however, influence cardiomyocyte electrophysiology. Indeed, dyadic localization could allow IICR regulation of the L-type Ca2+ current and localization of InsP3Rs in the neighborhood of the electrogenic Na+/Ca2+ exchanger (NCX; located in microdomains along the TTs that are distinct to release sites), is reported to modulate the electrical properties of ventricular cardiomyocytes that may lead to arrhythmia (Mohler et al., 2005; Signore et al., 2013).
To explain how InsP3 elicits its effects, we reasoned that either Ca2+ release via InsP3Rs served to directly activate RyR clusters or facilitated their activation by Ca2+ arising from other sources, including via LTCC and RyRs (Cannell et al., 2013; Kolstad et al., 2018; Walker et al., 2015; Wang et al., 2004; Wescott et al., 2016). The normal cytosolic Ca2+ sensitivity of cardiac RyRs is inherently low, resulting in infrequent openings at cytosolic Ca2+ concentrations <1 µM (Xu et al., 1996). Modeling studies predict that concentrations of upwards of 10 µM are in fact required for RyR activation indicating that gating of RyRs is only achieved when they are in close vicinity of a Ca2+ source (Cannell et al., 2013). Ca2+ release via proximal InsP3Rs could therefore act via CICR to stimulate RyR openings. Furthermore, based on the notion that clusters rapidly shut down when RyRs numbers are low or are disorganized, IICR could also act to maintain cluster activity and increase spark fidelity (Walker et al., 2015).
Through inducing a leak of Ca2+ from the SR, IICR is reported to lead to a reduction in SR Ca2+ load (Blanch i Salvador and Egger, 2018). As lower SR Ca2+ results in a reduction in RyR-mediated Ca2+ release and spark termination (Györke and Györke, 1998), this depletion of the SR Ca2+ store could thus lead to a reduction in Ca2+ flux during the Ca2+ transient. We did not, however, see such an effect of ET-1 activation of IICR under the pacing conditions used for recordings of electrically evoked Ca2+ release, suggesting that in healthy cardiomyocytes, homeostatic mechanisms are sufficient to maintain SR Ca2+ levels in the face of IICR activation. Further supporting the ability of homeostatic mechanisms to counter substantial Ca2+ leak, we also did not see any InsP3-stimulated elevation in diastolic Ca2+ levels in paced cardiomyocytes or in quiescent cardiomyocytes in which RyRs were inhibited. These findings are in agreement with our previous study (Smyrnias et al., 2018) and are in line with findings in rabbit cardiomyocytes, which showed a substantially lower Ca2+ leak via InsP3Rs than via RyRs (Zima et al., 2010).
Through activation of dyadic InsP3R-dependent Ca2+ signaling to RyRs and a consequent increase in the fidelity of Ca2+ spark generation, ET-1, and possibly other hormones that act via Gαq-coupled GPCRs, modulate SR Ca2+ release during ECC in ventricular cardiomyocytes (Fig. 7). Through this recruitment of dyadic RyR clusters and sensitization of the Ca2+ release process, ET-1 stimulation augments Ca2+ transients and/or promotes an increase in potentially arrhythmogenic extra-systolic Ca2+ release events. These effects can be especially detrimental under the circumstances of RyR post-translational modifications, such as by Ca2+/calmodulin-dependent protein kinase II (CaMKII) and reactive oxygen species (ROS) that sensitize them to Ca2+.
MATERIALS AND METHODS
All experimental procedures were approved by the in-house ethical committee (Ethische Commissie Dierproeven, KU Leuven), with permit number P070/2018 and comply with European legislation (European Commission Directive 2010/63/EU) on animal care.
Adult rat ventricular cardiomyocyte isolation, culture and adenoviral infection
Cardiomyocytes were enzymatically isolated from male 6–7-week-old Wistar Kyoto rats (weight 180–210 g) purchased from Harlan Laboratories (Horst, The Netherlands). Briefly, rats were deeply anesthetized by intraperitoneal injection of pentobarbital (80 mg/kg body weight) supplemented with heparin (2500 units/kg body weight) in the same syringe to prevent blood coagulation, and subsequently killed by cervical dislocation. The heart was removed from the rat via bilateral thoracotomy, placed on a Langendorff perfusion apparatus and digested with collagenase II (Worthington Biochemical Corp., Lakewood, HJ, USA) at 37°C as previously described (Smyrnias et al., 2018). Culture and viral infections were performed with modifications from that previously described (Drawnel et al., 2012). Freshly isolated cardiomyocytes were plated on 18 mm coverslips coated with 25 µg/ml laminin (Gibco, #23017015) and thereafter cultured in M199 medium (Sigma-Aldrich, #M7528), supplemented with 0.1% bovine serum albumin (BSA), 1% insulin-transferrin-selenium and penicillin/streptomycin, 5 mM creatine, 2 mM L-carnitine, 5 mM taurine. At 1 hour after plating, unattached cells were removed and remaining cardiomyocytes were infected with adenovirus carrying the target gene at a multiplicity of infection of 100. Cardiomyocytes were used for experiments after 48 h in culture. Cardiomyocytes that responded to electrical stimulation and did not show spontaneous activity were selected for analysis (Fig. S1A).
Molecular biology and preparation of adenoviruses
The recombinant adenoviral plasmid that contained GCaMP6f–triadin gene was kindly provided by Prof. Heping Cheng (Peking University, China; Shang et al., 2014). pGP-CMV-GCaMP6f was from Addgene (#40755; deposited by Douglas S. Kim; Chen et al., 2013). Adenoviruses were generated using the AdEasy Adenoviral vector system (Agilent Technologies, #240009). Briefly, the coding region of the gene to be expressed was cloned into the pShuttle-CMV vector, after which it was recombined into pAdEasy-1 vector by LR recombinase reaction in E. coli. After sequence-verification the adenoviral recombinant plasmid was linearized with PacI and, after purification with SureClean Plus beads (Bioline, #BIO-37047), was transfected into HEK293 cells (Microbix Biosystems Inc., Ontario, Canada). Crude adenovirus was harvested after 10–14 days. This crude viral preparation was used for large-scale amplification of virus in HEK293 cells, which was purified using the Vivapure AdenoPack 100 RT kit (Sartorius, #VS-AVPQ102). Adenoviral stock titer was determined using an end-point dilution assay.
Confocal Ca2+ imaging
On the day of experiment, coverslips were mounted in the imaging chamber (Multichannel systems; model #RC-49MFSH), supplemented with perfusion and aspiration system, and a solenoid-controlled local perfusion system was positioned near the cell to allow continuous perfusion and rapid switching between control and agonist or blocker solutions. Throughout the experiment, cells were constantly perfused with normal Tyrode solution (135 mM NaCl, 5.4 mM KCl, 10 mM HEPES, 10 mM D-glucose, 2 mM MgCl2 and 1 mM CaCl2 at pH 7.35). All experiments were performed at room temperature on cardiomyocytes electrically paced with a pair of platinum electrodes at a stimulation frequency of 0.5 Hz. A voltage was set at the value required to evoke contraction in >90% of cardiomyocytes. To visualize TTs, cardiomyocyte membranes were fluorescently labeled with Di-8-ANEPPS (Thermo Fisher Scientific, #D3167; 5 µM) dye for 5 min. Ca2+ imaging was performed using a Nikon A1R confocal microscope through a Nikon Apo 60× Oil λS DIC N2 (1.4 NA) oil immersion objective, the pinhole was set to 1 A.U. achieving a Z-section thickness of 0.42 µm. Ca2+ transients were recorded at 512 lines/s with a pixel size of 0.19–0.22 µm.
Pharmacological agents were from Sigma-Aldrich (Belgium) and concentrations used were as follows: isoproterenol, 10 nM (Iso; #I6504); endothelin-1, 100 nM (ET-1; #E7764); angiotensin II, 1 µM (Ang II; #A9525); 2-aminoethyl diphenylborinate, 2 µM (2-APB; #D9754); tetracaine, 1 mM (#T7508); caffeine, 20 mM (#C0750); InsP3 AM, 1 µM (cag-InsP3; SiChem GmbH, #cag-iso-2-145-1).
Ca2+ release measurements and protocols
Local and spatially averaged SR Ca2+ release signals were measured using GCaMP6f–triadin (GCaMP6f-T) and non-targeted GCaMP6f. For analysis of dyadic Ca2+ release during Ca2+ transients, cardiomyocytes were paced at 0.5 Hz. At the end of the experiment, to determine the SR Ca2+ load and all recruitable Ca2+ release sites, electrical stimulation was stopped and 20 mM caffeine rapidly applied. To measure the effect of treatment on the frequency and properties of elementary Ca2+ release events, pacing was paused for 30 s to record spontaneous Ca2+ release events. For experiments involving photo-release of InsP3, cardiomyocytes were loaded with cag-InsP3 diluted to 1 µM in Pluronic F-127 (Thermo Fisher Scientific, #P30000MP) for 1 h. InsP3 was photo-released at user-defined cellular regions using the 405 nm laser integrated in the Nikon A1R confocal imaging system. The laser exposure required to achieve uncaging was determined empirically by imaging Ca2+ release in HeLa cells, which express InsP3Rs. Four exposures of 250 ms duration (pixel dwell time is 7.5 µs) with the laser at 100% was found to achieve optimal uncaging.
Analysis of local cytosolic Ca2+ release and elementary Ca2+ release events was performed using Fiji (Schindelin et al., 2012) and Matlab (The MathWorks Inc., Natick, MA). Acquired fluorescence linescan images were processed using custom-written routines in ImageJ (available from the authors upon request). Parameters of Ca2+ signals were calculated from the extracted profiles of fluorescence change using Matlab. The amplitude and maximal rate of SR Ca2+ release of spatially averaged Ca2+ transients were analyzed after normalization to resting fluorescence levels (F/F0) in Matlab. The time course of fluorescence was averaged for 5 pixels around the center of the identified release site. Synchrony of Ca2+ release was assessed as previously described (Song et al., 2001), by measuring the latency of Ca2+ release onset in relation to the earliest release site and measuring the standard deviation of these values. The peak of the first derivative of fluorescence change [max(ΔF0/ms)] during Ca2+ release was used as a measure of maximal Ca2+ release flux from the SR as previously described (Sipido and Wier, 1991). Only active release sites that displayed a fluorescence change exceeding the threshold of 5×SD×F0 were included in the analysis. For each release site, kinetic parameters of Ca2+ release were averaged from the seven consecutive Ca2+ transients to reduce the effect of beat-to-beat variability of stochastic Ca2+ release via RyRs. SR Ca2+ content was assessed by rapidly applying 20 mM caffeine and measuring the amplitude of the elicited Ca2+ transient.
The SparkMaster plugin implemented in ImageJ was used to analyze Ca2+ spark parameters (Picht et al., 2007). The Ca2+ spark event detection was set at six times the standard deviation of the background noise over the mean value of the background. This value was required to accommodate for the non-homogeneous distribution of targeted GCaMP6f. Detected events were further verified by manual analysis. Parameters analyzed included spark frequency (normalized to cell length and time), amplitude (ΔF/F0), size (full width at half-maximal amplitude; FWHM), duration (full duration at half-maximal amplitude; FDHM), rise time (time to peak; TTP) and maximal Ca2+ release rate [max(ΔF0/ms)]. Ca2+ spark mass was calculated according to the formula 1.206×ΔF/F0×FWHM3 (Chandler et al., 2003). Spark-mediated Ca2+ leak was calculated as the product of spark mass and frequency.
Immunofluorescence and imaging
Cardiomyocytes expressing GCaMP6f–triadin were fixed with 1% paraformaldehyde (PFA) in normal Tyrode solution for 5 min at room temperature. The distribution of GCaMP6f–triadin relative to SR Ca2+ release channels was investigated by confocal microscopy. In brief, dual labeling was performed using rabbit anti-RyR2 (gift from Prof. V. Sorrentino, University of Sienna; 1:200; Giannini et al., 1995) or anti-InsP3R2 (Atlas Antibodies, HPA059144; Lot#R82918; 1:100) primary antibodies in combination with a cocktail of mouse anti-NCX1 (Swant, R3F1; Lot#mr05; 1:200) and anti-caveolin-3 (BD Bioscience, N610420; 1:200) primary antibodies designed to stain the sarcolemma and TTs as previously described (Hou et al., 2015; Munro et al., 2016). For simultaneous visualization of InsP3Rs and RyRs, dual labeling was performed with mouse anti-RyR1 (clone 34C, which also recognizes type 2 RyR) (Invitrogen, #MA3-925; Lot#TG268734; 1:200) and rabbit anti-InsP3R2 (Atlas Antibodies, HPA059144; 1:100) primary antibodies. Cardiomyocytes were permeabilized with PBS containing 0.4% Triton X-100 (diluted from Surfact-Amps detergent solution, Thermo Fisher Scientific, #28314) for 15 min at room temperature, followed by blocking with either 5% normal goat serum (NGS) (Sigma-Aldrich, #S26-M) in PBS (for RyR2 antibody) or Image-iT Signal Enhancer (InsP3R antibody; Thermo Fisher Scientific, #I36933) for 1 h at room temperature. Primary antibodies were applied overnight at 4°C in a buffer composed of PBS, 0.1% Triton X-100, 2% BSA and 2% NGS. Appropriate secondary antibodies, conjugated to either Alexa Fluor 568 (goat anti-rabbit, Thermo Fisher Scientific, # A-11036) or Alexa Fluor 647 (goat anti-mouse, Thermo Fisher Scientific, # A-21236), were applied at a dilution of 1:100 or 1:500 for goat anti-rabbit-IgG and 1:200 for goat anti-mouse-IgG to stain InsP3R2, RyR2 and TTs/sarcolemma, respectively. All secondary antibodies were incubated for 2 h at room temperature. Cardiomyocytes were mounted on glass slides in Vectashield containing DAPI (Vector Laboratories, #H-1200-10), and images were acquired using a Nikon A1R confocal microscope configured on a Nikon Ti2 equipped with a Nikon Apo 60×/1.4 NA (MRD71600) oil immersion objective. The pinhole was set to achieve a Z-section thickness of 0.45 µm. Images were acquired with voxel size set at 0.11×0.11×0.2 µm.
Acquired image stacks were deconvolved using Huygens Professional version 19.04 (Scientific Volume Imaging, The Netherlands), using the CMLE algorithm with the signal to noise ratio set at 20 and iterations to 40. A single image of the stack from the cell central plane was used to assess relative expression of proteins. After images for each protein staining were thresholded, the relative distances were measured using a custom-written macro in ImageJ and were calculated as Euclidean distance between the center of two nearest clusters. Additionally, overlap between two proteins of interest was calculated in analogy with Manders’ coefficient.
GCaMP6f–triadin expression analysis of confocal images
Quantitative analysis of cellular GCaMP6f–triadin protein expression relative to RyRs and TTs was based on confocal immunofluorescence imaging as described above. To exclude sub-sarcolemmal cellular regions of the analysis, the cell mask was created by blurring the RyR2 images (Gaussian Blur, sigma 25 pixels) and consecutive dilation (25 pixels). Similarly, nuclei and perinuclear regions were removed from GCaMP6f–triadin, TT and RyR2 images. To create a mask of these regions, DAPI images were consecutively smoothed (mean, radius 1 pixel), blurred (Gaussian Blur, sigma 15 pixels) and binarized. The immunofluorescence signal pattern of GCaMP6f–triadin, TTs and RyR2 was segmented with Fiji operations: (1) GCaMP6f–traidin images were background subtracted (rolling ball radius: 5 pixels), smoothed (mean, radius 2 pixels), and global Moments thresholding was used to binarize the GCaMP6f-triadin signals; (2) TT images were smoothed (median, radius 2 pixels), segmented by local Bernsen thresholding (radius: 15) and consecutively skeletonized; and (3) RyR2 images were background subtracted (rolling ball radius: 5 pixels), smoothed (median, radius 2 pixels) and segmented by local Otsu thresholding (radius: 12). Watershed function was applied to separate adjacent clusters that were segmented together. Binarized images were then used to determine the percentage of GCaMP6f–triadin clusters overlapping with RyR2 signals and TTs. For overlap analysis, we accepted clusters that ranged between partial and mutually complete overlap in binarized images. The skeleton of TTs was used to calculate the Euclidean distance between TTs and GCaMP6f–triadin clusters.
InsP3R expression analysis of confocal images
Quantitative analysis of cellular expression of InsP3Rs was performed as described for GCaMP6f–triadin except for image processing and segmentation methods, which were as follows: (1) RyR2 images were background subtracted (rolling ball radius: 10), smoothed (mean, radius 1 pixel) and segmented by local Bernsen thresholding (radius: 5); (2) InsP3R2 images were background subtracted (rolling ball radius: 15), smoothed (median, radius 1 pixel) and binarized by local Otsu threshold (radius: 15); and (3) TT images were smoothed (median, 1 pixel), segmented by local Otsu threshold (radius: 18) and consequently skeletonized.
All individual data points are shown in dot plots, with each dot representing a cell. Normal distribution of data was assessed with D'Agostino & Pearson omnibus normality testing (α=0.05). Data is reported as mean±s.e.m, Tukey boxplots (median with whiskers equaling 1.5 times the interquartile range), mean±s.d., or as before-and-after measurements, and the calculated effect sizes (Δ) with 95% confidence interval. Statistical comparisons were made using paired or unpaired Student's t-test, Wilcoxon matched-pairs signed rank test (paired data, non-normal distribution), Mann–Whitney U (unpaired data, non-normal distribution) and ordinary one-way ANOVA with Bonferroni post hoc test to compare more than two groups. A repeated measures two-way ANOVA test with Bonferroni post hoc testing was applied for pharmacological interventions. Difference in distributions of maximal dyadic Ca2+ release fluxes and Ca2+ sparks parameters was assessed using an equal variance (F) test. Statistical analysis was performed using GraphPad Prism 9 software (GraphPad Software, La Jolla, California, USA) and statistical significance was set at P<0.05. The number of animals (Nanimals), cells (ncells), release sites (nrelease sites) and Ca2+ release events are given in the figure legends. Sample sizes were determined by power analysis based on data in our previous publication on similar parameters (Harzheim et al., 2009). We assessed the presence of outliers by using ROUT method implemented in Graphpad (with Q=1%).
We thank Roxane Menten for assistance with animal handling and cell isolation, Prof. Heping Cheng for the GCaMP6f-triadin construct and Prof. Douglas S. Kim & Genie project (Howard Hughes Medical Institute, USA) for the pGP-CMV-GCaMP6f plasmid. RyR2 antibody was a gift from Prof. Vincenzo Sorrentino (University of Siena, Italy).
Conceptualization: H.L.R.; Methodology: K.D., K.R.S., H.L.R.; Formal analysis: K.D., H.L.R.; Investigation: K.D., H.L.R.; Resources: K.R.S., H.L.R.; Data curation: K.D., H.L.R.; Writing - original draft: K.D., H.L.R.; Writing - review & editing: K.D., K.R.S., H.L.R.; Visualization: K.D., H.L.R.; Supervision: H.L.R.; Project administration: H.L.R.; Funding acquisition: H.L.R.
This work was supported by the KU Leuven and the Fund for Scientific Research-Flanders [Fonds Wetenschappelijk Onderzoek (FWO) project grant (G08861N, Odysseus Project 90663 to H.L.R.)].
The authors declare no competing or financial interests.