In the heart, where ATP consumption is highest of any tissue, Ca²⁺ homeostasis in the mitochondrial matrix and ATP synthesis in the inner mitochondrial membrane are tightly controlled and interacting in a tissue (muscle)-specific manner regulated by the mitochondrial membrane potential (Wescott et al., 2019; Boyman et al., 2020; Boyman et al., 2021). Pathologically altered intracellular Ca²⁺ signaling and impaired bioenergetics due to mitochondrial dysfunction contribute to both ventricular tachyarrhythmia (VT) and heart failure development, at least in part through increased production of reactive oxygen species (ROS) (Zhou and Tian, 2018; Santulli et al., 2021). Under physiological conditions at rest, mitochondrial oxidative phosphorylation and ATP synthesis are accompanied by a low level of cardiac ROS generation mainly through byproducts of the electron transport chain (Murphy, 2009). However, increased stress and enhanced cardiac performance strongly augment mitochondrial respiratory activity significantly increasing ROS production (Bertero and Maack, 2018). Although several enzymatic systems protect cardiomyocyte function through ROS neutralization (Munzel et al., 2017), chronically compromised mitochondrial bioenergetics in heart disease may shift the balance toward increased oxidation of the glutathione (GSH) redox potential (E_GSH) (Bertero and Maack, 2018), especially since an increased energy demand during sustained β-adrenergic stimulation was proposed to weaken mitochondrial antioxidant defense (Bovo et al., 2015). Notably, increased mitochondrial E_GSH oxidation itself is a quantitative live-cell indicator of a strongly increased ROS production, and hence may identify an early key disease process and mark the beginning of metabolic remodeling in the failing heart (Doenst et al., 2013).

Under physiological conditions of excitation-contraction (E-C) coupling, Ca²⁺ influx through voltage-gated L-type Ca²⁺ channels (Ca_V1.2) activates mass intracellular Ca²⁺ release via ryanodine receptor type 2 (RyR2) large-conductance channels in the sarcoplasmic reticulum (SR). This Ca²⁺-induced Ca²⁺ release increases the intracellular Ca²⁺ concentration throughout the myofilaments activating myocardial contraction (Bers, 2002), while it accounts for up to two-thirds of cardiac ATP consumption (Suga, 1990; Meyer et al., 1998). Vice versa, for cardiac relaxation intracellular Ca²⁺ is decreased mainly via the SR Ca²⁺-ATPase (SERCA2a) and extruded by electrogenic trans-sarcolemmal Na⁺/Ca²⁺ exchange (NCX). Hence, diastolic Ca²⁺ cycling accounts for an additional ∼30% of total energy consumption at baseline in healthy cardiomyocytes and in human heart tissue (Gibbs et al., 1988; Meyer et al., 1998).

Interestingly, pathophysiological changes in cytosolic Ca²⁺ cycling are widely accepted as a central cause of arrhythmias and contractile dysfunction in atrial and ventricular tissue (Bers, 2002; Alsina et al., 2019; Brandenburg et al., 2019; Dridi et al., 2020; Berisha et al., 2021; Boyman et al., 2021; Sadredini et al., 2021; Santulli et al., 2021). In this context, increased mitochondrial ROS generation can contribute to altered Ca²⁺ handling and vice versa, which may furthermore constitute a vicious cycle in cardiomyocytes (Zhou and Tian, 2018; Santulli et al., 2021). However, the quantitative and molecular mechanisms if and how acutely increased SR Ca²⁺ leak causes excess mitochondrial Ca²⁺ uptake and ROS generation have been and continue to be controversially discussed (Mallilankaraman et al., 2012; Antony et al., 2016; Kohlhaas et al., 2017; Wescott et al., 2019; Boyman et al., 2020; Rossini and Filadi, 2020; Boyman et al., 2021; Santulli et al., 2021). Nonetheless, the largest known ion channel, RyR2, contains a relatively high number of redox-reactive cysteine residues and hence provides an important molecular substrate for increased ROS generation in cardiomyocytes (Santulli and Marks, 2015). For instance, RyR2 is oxidized in atrial tissue in vivo in a genetic model of catecholaminergic polymorphic ventricular tachycardia (CPVT), RyR2-R2474S^+/− knockin mice, with increased susceptibility for pacing-induced atrial fibrillation (Shan et al., 2012).

Interestingly, while CPVT has been initially characterized solely as an arrhythmogenic cardiomyopathy (Yano et al., 2009), subsequently seizures originating in the brain (Lehnart et al., 2008; Lieve et al., 2019) and glucose intolerance in the Langerhans cells of the pancreas have been demonstrated in humans and mouse models (Santulli and Marks, 2015; Santulli et al., 2015). Although the majority of biophysically investigated RyR2 missense mutations point to a gain-of-function defect following protein kinase A phosphorylation both at the recombinant and native single-channel level (Wehrens et al., 2003; Lehnart et al., 2004; Tester et al., 2007; Lehnart et al., 2008), the role of PKA-dependent RyR2 phosphorylation remains controversial (Leenhardt et al., 2012; Wleklinski et al., 2020) since, e.g., inhibition of cardiac CaMKII by gene therapeutically applied CaMKII inhibitory peptide, but not PKA inhibitory peptide, effectively suppressed ventricular arrhythmias in a murine and an iPSC-CM model of CPTV (Bezzerides et al., 2019). However, a general consensus exists that adult cardiomyocytes and intact hearts from knockin mice expressing different gain-of-function RyR2 mutations are susceptible for significantly increased diastolic Ca²⁺ leak and Ca²⁺ waves following β-adrenergic stimulation (Lehnart et al., 2008; Fernandez-Velasco et al., 2009; Liu et al., 2009; Uchinoumi et al., 2010; Leenhardt et al., 2012; Borile et al., 2021). Strikingly, consistent with the human phenotype of a normal left-ventricular structure and function at rest, morphological and contractile changes have been excluded in a distinct independent RyR2-R2474S knockin mouse with a clear β-adrenergic driven phenotype of Ca²⁺ triggered cardiac VT (Uchinoumi et al., 2010). Given that the altered SR Ca²⁺ leak/load balance needs to be acutely compensated at the cost of increased ATP consumption during catecholaminergic stimulation (Eisner et al., 2020), CPVT-affected hearts may face a challenging metabolic state during episodes of transient intracellular Ca²⁺ overload and β-adrenergic stimulation, which may lead to increased mitochondrial Ca²⁺ load with excessive ROS generation and contribute to electrical and contractile cardiomyocyte dysfunction (Bers, 2014; Wescott et al., 2019; Boyman et al., 2021).

In the failing heart, ROS signaling has been established as a central process of the organic disease progression (Heymes et al., 2003; Yano et al., 2005; Guo et al., 2007; Akki et al., 2009). Increasing evidence suggests that excessive ROS production can lead to chronic cardiac pathologies, such as apoptosis, fibrosis, and hypertrophy jointly promoting the maladaptive ventricular remodeling (von Harsdorf et al., 1999; Cucoranu et al., 2005; Takimoto et al., 2005; Hori and Nishida, 2009; Santulli et al., 2021). Recently, the mechanisms that temporally and spatially link excess ROS and Ca²⁺ as two principal detrimental factors have been addressed in the context of diabetic cardiomyopathy. Acute elevation of the extracellular glucose concentration induces cardiomyocyte ROS production through O-GlcNAcylation of CaMKIIδ and activation of NOX2 (NADPH oxidase type 2) resulting in increased ROS production in the cytosol but not in the mitochondria (Lu et al., 2020). On the other hand, RyR2-R2474S^+/– knockin mice exhibiting catecholamine-induced intracellular Ca²⁺ leak subsequently revealed mitochondrial dysfunction, increased ROS production, increased atrial RyR2 channel oxidation, and arrhythmia susceptibility (Xie et al., 2015). Moreover, inhibitory pharmacological treatment selectively targeting the RyR2 Ca²⁺ leak or genetic inhibition of mitochondrial ROS production both prevented atrial fibrillation, indicating that RyR2 leakage and mitochondrial ROS generation may indeed constitute a self-enforcing subcellular cycle that can set off arrhythmias (Xie et al., 2015). However, the acute mechanism and hence direct but transient link of the catecholamine-stimulated RyR2-R2474S^+/– Ca²⁺ leak on mitochondrial ROS production have not been investigated previously.

Novel transgenic reporter mouse lines expressing the Grx1-roGFP GSH redox biosensor in a cardiomyocyte-restricted manner targeted to either the cytoplasm or the mitochondrial matrix have been established previously (Swain et al., 2016). Thus, cardiomyocytes from Grx1-roGFP biosensor mice allow the direct live-cell monitoring of changes in compartmentalized - cytosolic vs. mitochondrial - E_GSH (Swain et al., 2016), as demonstrated acutely in response to high extracellular glucose concentrations recently (Lu et al., 2020).

The objective of the present study is to investigate mitochondrial and cytosolic E_GSH changes during pacing-controlled contractions in RyR2-R2474S cardiomyocytes and intact hearts in a knockin mouse model of CPVT. To investigate the interplay between acute catecholamine-stimulated increased SR Ca²⁺ leak and mitochondrial oxidation changes, we crossed the mitochondrial matrix-targeted Grx1-roGFP (mito-Grx1-roGFP) redox biosensor line with mitochondrial redox-competent C57BL/6N background RyR2-R2474S^+/– knockin mouse strain by heterozygous breeding (Lehnart et al., 2008; Nickel et al., 2015; Swain et al., 2016). While double transgenic mito-Grx1-roGFP:RyR2-R2474S^+/– (RyR2-R2474S dTG) cardiomyocytes showed a significant E_GSH increase during pacing-evoked contractions following β-adrenergic stimulation compared to RyR2-WT control cardiomyocytes, ex vivo perfusion of isoproterenol-stimulated RyR2-R2474S dTG hearts significantly extended and confirmed the cellular findings. Importantly, acute treatment of RyR2-R2474S dTG-cardiomyocytes with the RyR channel blocker dantrolene diminished this E_GSH increase.

This work has examined how mitochondrial E_GSH oxidation is increased by essential factors in cardiomyocytes with important physiological and pathophysiological consequences that converge at the dynamic interface between RyR2 channel function, sympathetic catecholaminergic stimulation, and contractile activity. Overall, we find that a significant increase following β-adrenergic stimulation significantly changes mitochondrial E_GSH in dTG-cardiomyocytes and hearts, which was further dependent on both the presence of the RyR2-R2474S missense mutation and faster beating rates. Importantly, dantrolene treatment of RyR2-R2474S dTG-cardiomyocytes results in a pronounced reduction of this increased mitochondrial oxidation. We discuss our findings on these interrelated RyR2-R2474S regulatory and E_GSH driven mitochondrial mechanisms in the context of previously published reports.

Recent quantitative imaging studies established that an increase of mitochondrial Ca²⁺ uptake was necessary to drive inner mitochondrial membrane depolarization and enhanced ATP synthesis in a heart-specific and previously unknown voltage-dependent fashion (Wescott et al., 2019; Boyman et al., 2020; Boyman et al., 2021). This contrasts with heart disease where alterations of mitochondrial E_GSH may profoundly shift organelle metabolism and thus cardiac bioenergetics (Bertero and Maack, 2018). Beyond bioenergetics, increased mitochondrial E_GSH oxidation revealed a strongly increased ROS production, which might be of immediate relevance for heart disease progression and metabolic remodeling (Doenst et al., 2013). In analogy, increased mitochondrial oxidation and E_GSH depolarization in RyR2-R2474S cardiomyocytes and hearts indicate a strongly boosted ROS generation during faster beating rates concurrent with β-adrenergic stimulation.

Interestingly, increased SR Ca²⁺ leak and mitochondrial dysfunction have been demonstrated to enhance ROS production in heart disease previously (Zhou and Tian, 2018; Santulli et al., 2021). Remarkably, mitochondrial ROS scavenging attenuated the effects of augmented RyR2 activity leading to proarrhythmic Ca²⁺ waves in Calsequestrin2-null mice serving as a model for CPVT (Hamilton et al., 2020). In line with this view, ROS scavenging by mito-TEMPO was shown to reduce the number of Ca²⁺ waves also in cardiomyocytes from rabbit hearts (Bovo et al., 2012; Cooper et al., 2013). As isoproterenol-stimulated RyR2-R2474S cardiomyocytes exhibit an over threefold increased Ca²⁺ spark frequency that additionally sustains regenerative intracellular Ca²⁺ waves, here the observed robust increase in mitochondrial oxidation will depend both on acute catecholamine stimulation and the beating rate-dependent extent and complex spatiotemporal dynamics of the cell-wide diastolic Ca²⁺ leak reported previously (Lehnart et al., 2008; Xie et al., 2015; Sadredini et al., 2021). However, even under physiological conditions, an increased mitochondrial oxidation was evidenced at higher beating rates in sTG-cardiomyocytes albeit much more pronounced in RyR2-R2474S dTG-cardiomyocytes. For the latter, the faster beating rate clearly led to mitochondrial E_GSH oxidation and this was most strongly increased after β-adrenergic stimulation.

Tight control of nanomolar diastolic Ca²⁺ levels is essential for normal cardiac function depending on the precise balance between cytosolic Ca²⁺ influx and efflux (Eisner et al., 2020). If the local control mechanisms become compromised, cardiomyocytes will develop arrhythmia and hypercontraction from cytosolic Ca²⁺ overload. This balance is disturbed when the diastolic Ca²⁺ spark frequency is significantly increased, a phenomenon linked molecularly to RyR2 channel gain-of-function defects (Cheng and Lederer, 2008; Bers, 2014). Interestingly, more than a dozen RyR2 mutations originally identified in CPVT patients have been shown to increase Ca²⁺ leak (Jiang et al., 2002; Wehrens et al., 2003; Lehnart et al., 2004; Kontula et al., 2005; Tester et al., 2007; Lehnart et al., 2008; Marjamaa et al., 2011; Shan et al., 2012). However, only two knockin models were confirmed to exhibit exercise-triggered VTs, RyR2-R2474S^+/– (Lehnart et al., 2008) and RyR2-R4496C^+/– (Priori and Napolitano, 2005; Liu et al., 2006), and are hence reminiscent of the human arrhythmia phenotype diagnosed by exercise stress testing. In line with the in vivo pathology, RyR2-R2474S^+/– cardiomyocytes exhibited an overall normal diastolic Ca²⁺ spark behavior at rest, but a strong spark frequency increase occurred after isoproterenol treatment (Lehnart et al., 2008). In this context, increased diastolic Ca²⁺ extrusion mainly by SERCA2a and NCX will necessarily consume additional ATP at rest at the cost of concomitantly increased ROS generation (Bers, 2002, 2014). Thus, our data further link increased catecholaminergic RyR2 channel dysfunction in a murine CPVT model to acute mitochondrial oxidation.

In the present study, we monitored the effect of increased contractile activity on E_GSH in the mitochondrial matrix, while the former process has been long known to correlate with mitochondrial metabolic activity and ROS generation (Gudbjarnason et al., 1962). Since a similar extent of E_GSH change was excluded in the cytoplasm, NOX2 seems unlikely as ROS source here (Santos et al., 2011). In another context, high extracellular glucose concentrations increased E_GSH in the cytosol through NOX2 but not in the mitochondrial matrix (Lu et al., 2020). In contrast, we found that mitochondrial oxidization depended on the duration and frequency of dTG-cardiomyocyte beating. This finding is in line with reported changes of NADH in response to a stepwise increase of the beating frequency from 0.1 to 3.3 Hz in guinea-pig cardiomyocytes (Jo et al., 2006). As the increase in the mitochondrial E_GSH following beating was observed both in RyR2-WT and RyR2-R2474S dTG-cardiomyocytes, a relation with increased SR Ca²⁺ leak seems plausible after pacing-evoked contractile activity reported previously (Satoh et al., 1997). Finally, blocking RyR2 Ca²⁺ leak with dantrolene diminished E_GSH oxidation as hypothesized previously (Hartmann et al., 2017).

A recent study demonstrated that Ca²⁺ leak increased mitochondrial ROS production in cultured WT rat cardiomyocytes transfected for 48 h with the OMM-HyPer mito-ROS biosensor (Hamilton et al., 2020). An increased ROS production has also been demonstrated during increased beating in healthy cardiomyocytes with the fluorescent probe dichlorodihydrofluorescein, which is confirmed by our sTG-cardiomyocyte data (Heinzel et al., 2006). In rat trabeculae, NADH oxidation was observed at 0.25–2 Hz after 60 s with pronounced ATP hydrolysis (Brandes and Bers, 2002). A similar time frame in response to contractile activity was observed in the present study, where the maximal change of the biosensor signal occurred within 2–3 min after beating. This further agrees with the reported change of the redox signal observed in cardiomyocytes in the same biosensor cardiomyocytes after a high glucose or pH challenge (Lu et al., 2020; Kreitmeier et al., 2021). Since the observed redox changes occurred only after completion of the given beating period, it is suggested to correlate further with the aftercontraction related metabolic activity observed frequently in isoproterenol-stimulated RyR2-R2474S dTG-cardiomyocytes.

Importantly, our report provides the first evidence that E_GSH changes happen under quasi physiological conditions in the intact beating heart as well. Using the genetically encoded reporter targeted to mitochondria allowed to monitor the average E_GSH from the whole ventricular, anterior heart surface (Supplementary Videos 1, 2) and to perform a ratiometric imaging excluding any major interference due to cardiac contractions. In comparison to the first description of the glutathione redox sensor mouse lines, we were able to increase the ratio differences from maximal oxidization to reduction to comparable signal-to-noise values to single-cell measurements (see Figures 5, 9). This increase in signal amplitude was the key for the detection of the comparable relatively small steps in oxidization during β-adrenergic stimulation, which we could only detect in the intact hearts of dTG mice expressing the RyR2-R2474S mutated channels. However, in contrast to isolated cardiomyocytes, we could not detect a change in oxidation by additional faster pacing in the spontaneously beating Langendorff hearts. One possible explanation may be that these changes are in comparison subtle and thus still not detectable in this setting. The other is that we focused on the effects of isoproterenol and such changes would occur only after prolonged phases of pacing which we could not implement in these experiments.

In summary, we have shown that contractile activity induced by EFS-pacing in isolated cardiomyocytes increased the mitochondrial E_GSH to more oxidized levels in a time- and frequency-dependent manner. The increase was most pronounced after β-adrenergic stimulation and in cardiomyocytes expressing the human RyR2-R2474S mutated tetrameric channel. The catecholaminergic E_GSH increase, however, was significantly attenuated by the pharmacological RyR2 channel antagonist dantrolene. Likewise, we observe a larger increase in the mitochondrial E_GSH toward more oxidized levels in spontaneously beating hearts expressing the RyR2-R2474S mutated channels after β-adrenergic stimulation as compared to control hearts. Hence, we suggest that the RyR2-R2474S mutation induces a metabolic burden in cardiomyocytes through excess ROS generation, especially after β-adrenergic stimulation, tipping the balance from a compensated to a high energetic cost and acutely increased diastolic Ca²⁺ leak, a multifactorial process, which may acutely contribute to the development of the arrhythmogenic CPVT phenotype as proposed for calsequestrin2-null mice previously (Hamilton et al., 2020).

The original contributions presented in the study are included in the article/Supplementary Material, further inquiries can be directed to the corresponding author/s.

The animal study was reviewed and approved by the Institutional IACUC. This study has not used in vivo examination of mice but ex vivo investigations, which are not subject to an ethics protocol at our institution.

JW and AW performed the research and analyzed the data. JW, TB, DK, and SL wrote the manuscript. DK contributed to the biosensor mouse models and analyzed the data. JW, SL, TB, and GH designed the study. All authors approved the manuscript.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.