Mitochondrial oxidative stress promotes atrial fibrillation

1Atrial fibrillation (AF), the most common cardiac arrhythmia, has high morbidity and mortality in adults. Despite intense research for over 100 years, AF remains incompletely understood and treatment is challenging. AF can be associated with structural and electrical remodeling of the cardiac atria. Structural changes may directly or indirectly induce atrial electric abnormalities leading to atrial ectopic events and AF. However, it has been difficult to distinguish whether structural changes are the cause or the consequence of AF. In fact, AF has been shown to occur in individuals with structurally normal hearts including those with catecholaminergic polymorphic ventricular tachycardia (CPVT) which is linked to inherited mutations in the intracellular Ca2+ release channel/ryanodine receptor (RyR2) that cause intracellular Ca2+ leak. Altered intracellular Ca2+ homeostasis has been associated with the pathogenesis of AF. In atrial myocytes, type 2 RyR (RyR2) is the major intracellular Ca2+ release channel, and RyR2 dysfunction can affect cellular electric activity. Atrial myocytes from both patients and animals with AF display increased diastolic SR Ca2+ leak via RyR2. Moreover, we and others have reported that AF can be induced in knock-in mice harboring specific mutations in RyR2, leading to intracellular Ca2+ leak. Recently, a report revealed that AF occurred in a 2-year-old child with a CPVT linked RyR2 mutation. These reports indicate that RyR2 dysfunction may be a critical contributor to AF.

2Oxidative stress has been associated with the development of AF both in patients and animal models of AF, and antioxidant drugs have shown beneficial effects on AF development. Furthermore, the prevalence of AF increases with age and the age-dependent increase in oxidative damage is widely acknowledged. However, the molecular mechanisms underlying oxidative stress in the development of AF remain essentially unclear. RyR2 is an important molecular target of oxidative stress in cardiac myocytes. While stress-induced oxidation of RyR2 in ventricular myocytes has been associated with cardiovascular disease, little is known about the pathophysiological role of atrial RyR2 oxidation. In a previous report we demonstrated that RyR2 is oxidized in atrial myocytes from a murine model of CPVT, RyR2-R2474S+/– mice, that display substantial increased AF susceptibility.

3In the present study, we explored the mechanistic role of atrial RyR2 oxidation in the pathophysiology of AF in two murine models of RyR2-mediated intracellular Ca2+ leak: mice harboring an RyR2 mutation linked to human CPVT (RyR2-R2474S+/–) and mice expressing a phosphomimetic aspartic acid residue at position 2808 (RyR2-S2808D+/+) leading to constitutively leaky channels. Moreover, we evaluated the role of mitochondrial free radicals on RyR2 oxidation by crossing mice harboring RyR2 mutations associated with Ca2+ leak with mice overexpressing human catalase targeted to mitochondria (mCAT mice).

4Atrial RyR2s from patients with chronic AF were oxidized, phosphorylated and depleted of calstabin 2 (a subunit of the complex that stabilizes the closed state of RyR2 during diastole) compared to subjects in sinus rhythm (Fig. 1a–d). These results are consistent with our previous report in mice harboring CPVT-mutated RyR2 channels, where RyR2 channels were oxidized and DTT treatment reduced SR Ca2+ leak in atrial myocytes.

5To explore the role of intracellular Ca2+ leak via RyR2 in the development of AF, we used a transgenic mouse harboring a constitutively leaky RyR2 channel (RyR2-S2808D+/+), which displays an age-related RyR2 oxidation in ventricular myocytes. WT and RyR2-S2808D+/+ littermates were divided into three different age groups (3-, 6- and 9-month old). Similar to ventricular RyR2, atrial RyR2 from RyR2-S2808D+/+ mice exhibited age-related oxidation and depletion of calstabin 2 compared to RyR2 from WT mice which displayed no PKA phosphorylation of RyR2 at 3-, 6- and 9-months of age (Fig. 1e–g, Supplementary Figure 1). To measure intracellular Ca2+ leak and its effects we examined Ca2+ spark frequency and SR Ca2+ load. At 3 months of age there was no significant difference in Ca2+ spark frequency (Ca2+ sparks/100 μm/s: 3.80 ± 0.47 in S2808D vs. 3.69 ± 0.31 in WT, n.s.) and SR Ca2+ load (∆F/F0: 10.43 ± 0.59 in S2808D vs. 10.78 ± 0.33 in WT, n.s.) in atrial myocytes isolated from WT versus RyR2-S2808D+/+ mice (Fig. 2a–d). In contrast, at 6 months of age atrial myocytes from RyR2-S2808D+/+ mice displayed significantly higher Ca2+ spark frequencies (Ca2+ sparks/100 μm/s: 6.13 ± 0.42 vs. 3.39 ± 0.38 in WT, p < 0.01) and markedly depleted SR Ca2+ stores (∆F/F0: 8.86 ± 0.37 vs. 10.56 ± 0.27 in WT, p < 0.05), consistent with age-related intracellular Ca2+ leak, compared with WT littermates (Fig. 2a–d). These differences were further increased in 9-month-old mice (Ca2+ sparks/100 μm/s: 8.00 ± 0.45 in S2808D vs. 3.93 ± 0.40 in WT, p < 0.01; SR Ca2+ load (∆F/F0: 7.93 ± 0.51 in S2808D vs. 10.93 ± 0.64 in WT, p < 0.01) as shown in Fig. 2A–D. Despite the observed depletion of SR Ca2+ stores, there was no significant change in the paced Ca2+ transient amplitude (Supplementary Figure 2a). This is most likely attributable to the increased fractional release of Ca2+ (Supplementary Figure 2b) due to a leftward shift in the Ca2+–sensitivity to activation of RyR2. Consistent with the increased SR Ca2+ leak in atrial myocytes, ex vivo confocal imaging in intact hearts and in vitro confocal imaging in isolated atrial myocytes revealed increased intracellular arrhythmic Ca2+ activity in the atria of 9-month-old RyR2-S2808D+/+ mice compared to WT littermates (Supplementary Figures 3–4). Similar to RyR2-R2474S+/– mice, the SR Ca2+ leak observed in atrial myocytes isolated from 9-month-old RyR2-S2808D+/+ mice was reduced by either 2-hour treatment with the Rycal S107 (10 μM), which prevented the oxidation-induced dissociation of calstabin 2 from RyR2, or treatment with the reducing agent dithiothreitol (DTT, Supplementary Figure 5a). CaMKII-mediated RyR2 phosphorylation has been recently reported to be involved in AF. However, KN-93 (1 μM), a CaMK inhibitor, did not significantly reduce the SR Ca2+ leak in atrial myocytes from RyR2-S2808D+/+ and RyR2-R2474S+/– mice (Supplementary Figure 5).

6AF can be linked to intracellular arrhythmic Ca2+ activity in atrial myocytes. To investigate AF susceptibility in RyR2-S2808D+/+ mice we induced AF by pacing the left atrium via an endo-esophageal pacing catheter. At baseline, all mice were in normal sinus rhythm (Fig. 2e). Upon pacing, RyR2-S2808D+/+ mice exhibited a progressive, age-related augmented susceptibility to AF in response to atrial burst pacing measured as the percentage of mice with pacing induced AF: 9.7% (3/31) at 3 months; 45.8% (11/24) at 6 months; and 68.4% (13/19) at 9 months). In contrast AF was rarely inducible in WT mice at the same ages: 5.6% (2/36); 5.0% (1/20); and 10.5% (2/19) at 3, 6 and 9 months of age, respectively (Fig. 2f).

7AF has also been associated with structural remodeling of the atria. However, histological analyses of atrial tissue did not show any obvious structural abnormality in 9-month-old RyR2-S2808D+/+ mice (Supplementary Figure 6), suggesting that the altered Ca2+ homeostasis is the main factor contributing to AF in this mouse model. Moreover, a 2-week pharmacological treatment with a stabilizer of the closed state of RyR2 channel (S107, 40 mg/kg/d in drinking water) prevented the development of AF in 9-month-old RyR2-S2808D+/+ mice (Supplementary Figure 7).

8Reactive oxygen species (ROS) production measured using the fluorescent indicator H2DCFDA was significantly increased in freshly isolated atrial myocytes from 6- and 9- month old RyR2-S2808D+/+ mice, indicating increased oxidative stress (Fig. 3a). Mitochondria are the major source of intracellular ROS. MitoSOX Red revealed that atrial myocytes from 6- and 9-month old RyR2-S2808D+/+ mice had significantly higher mitochondrial ROS level (Fig. 3b). Additionally, ultrastructural analyses revealed profound mitochondrial dysmorphology, including lamellar degeneration, outer membrane disruption and swelling, in atria from 9-month old RyR2-S2808D+/+ mice, compared with age-matched WT littermates (Fig. 4a–c).

9To further investigate the role of mitochondrial ROS in atrial RyR2 dysfunction we crossed RyR2-S2808D+/+ mice with transgenic mice (mCAT) overexpressing the human catalase gene targeted to mitochondria to decrease mitochondrial ROS production. RyR2-S2808D+/+/mCAT mice exhibited significantly reduced atrial mitochondrial abnormalities (Fig. 4a–c). In atrial myocytes isolated from 9-month old RyR2-S2808D+/+/mCAT mice, both cellular and mitochondrial oxidative stress were markedly reduced compared with age-matched RyR2-S2808D+/+ littermates (Fig. 4d–e). Strikingly, blunted ROS production was associated with reduced atrial RyR2 oxidation and castabin2 dissociation resulting in decreased atrial diastolic SR Ca2+ leak and AF susceptibility (Fig. 4f–h and Supplementary Figure 8-9).

10To test the hypothesis that atrial mitochondrial dysfunction is attributable to SR Ca2+ leak, we examined the effects of S107 treatment on atrial mitochondria of RyR2-S2808D+/+ mice. Following a 2-month treatment with S107 (40 mg/kg/d in the drinking water), both mitochondrial dysmorphology (Fig. 5a–c) and ROS production (Fig. 5d) were significantly reduced in 9-month-old RyR2-S2808D+/+ mice compared with vehicle group.

11RyR2-R2474S+/– mice display oxidized atrial RyR2 and AF susceptibility and increased mitochondrial ROS level (Supplementary Figure 10). Both mitochondrial ROS levels and intracellular oxidative stress were greatly reduced in RyR2-R2474S+/–/mCAT mice compared with RyR2-R2474S+/– mice (Fig. 6a–b). Similarly, crossing RyR2-R2474S+/– with mCAT mice significantly decreased the intracellular Ca2+ leak in atrial myocytes compared to RyR2-R2474S+/– mice (Fig. 6c). Importantly, AF susceptibility in RyR2-R2474S+/–/mCAT was diminished to 22.2% (4/18) compared with 65.0% (13/20) in RyR2-R2474S+/– mice (Fig. 6d–e).

12Accumulating evidence suggests that oxidative stress plays a pivotal role in the development and perpetuation of AF. Our present findings strongly support previous reports suggesting that AF is associated with myocardial oxidative stress.

13We and others have proposed that atrial diastolic SR Ca2+ leak is an essential contributing factor in the pathogenesis of AF. Diastolic SR Ca2+ leak due to RyR2 dysfunction may be linked to multiple factors including phosphorylation, oxidation or pathological mutations of RyR2, all of which can result in channel dysfunction. We have previously reported PKA hyperphosphorylation and calstabin2 dissociation from RyR2 in several models of chronic AF. We have also demonstrated that chronic PKA phosphorylation of ventricular RyR2 is associated with oxidation of the channel and the combination of PKA phosphorylation and oxidation of RyR2 results in significantly more calstabin 2 dissociation from RyR2 compared to each post-translational modification alone. This observation is consistent with our present results, indicating the importance of oxidation of atrial RyR2 in promoting AF. Furthermore, oxidation of atrial RyR2 has been shown to be a key contributor to diastolic SR Ca2+ leak and AF in animal models of CPVT.

14It has been reported that CaMKII is also a molecular target of oxidative stress, and that oxidized CaMKII can phosphorylate RyR2 at Ser2814 inducing intracellular Ca2+ leak. However, we found that KN-93 treatment did not reduce atrial SR Ca2+ leak in RyR2-S2808D+/+ or RyR2-R2474S+/− atrial cardiomyocytes, suggesting that direct oxidation of atrial RyR2, but not phosphorylation by oxidized CaMKII, is the main factor inducing atrial SR Ca2+ leak and AF.

15The importance of PKA phosphorylation in cardiac disease has been challenged by Valdivia’s group, who concluded that phosphorylation of Ser2808 plays no role in β-adrenergic cardiac response. The fact that our findings have been confirmed by multiple other groups is addressed in a recent review.

16Multiple sources of ROS, including mitochondria, NADPH oxidases and NOS uncoupling, contribute to AF. Intriguingly, recent reports indicate that atrial sources of ROS vary with the duration and the substrate of AF, and mitochondria have been proposed as the major ROS source for long-term AF and age-related functional decline. This view is consistent with our results showing that mitochondrial ROS promote age-related AF.

17In the present study we establish for the first time that atrial RyR2 is a specific molecular target of oxidative stress that is fundamental in the development of AF. We demonstrate the functional importance of RyR2 oxidation in AF pathophysiology, showing that mitochondrial-derived ROS oxidize RyR2 in atrial myocytes leading to increased intracellular Ca2+ leak. Importantly, reducing mitochondrial ROS production attenuates atrial diastolic SR Ca2+ leak and prevents AF.

18Mitochondria have been reported to be abnormal in the atrial tissue of patients with AF. In this study, mice harboring leaky RyR2 also display mitochondrial abnormalities and increased ROS production. In cardiac myocytes, mitochondria and the SR are co-localized in the ‘mitochondrial microdomain’. Since mitochondrial Ca2+ uptake via the mitochondrial Ca2+ uniporter is dependent on SR Ca2+ release, alterations in SR Ca2+ release can also affect mitochondrial function by regulating mitochondrial Ca2+ uptake. In atrial myocytes isolated from RyR2-S2808D+/+ mice, leaky RyR2 channels were associated with an age-related increase in diastolic SR Ca2+ release without any significant change in the systolic Ca2+ transient amplitudes. Pharmacological inhibition of RyR2 Ca2+ leak restored atrial mitochondrial morphology and function suggesting that mitochondrial Ca2+ overload plays a key role in AF pathophysiology.

19Taken together, our data demonstrate that RyR2 oxidation resulting from intracellular oxidative stress in atrial myocytes leads to increased SR Ca2+ leak contributing to the pathogenesis of AF. Alterations of RyR2 inducing intracellular Ca2+ leak, including constitutive PKA phosphorylation or CPVT mutations, trigger a vicious cycle, in which SR Ca2+ leak in atrial myocytes impairs mitochondrial function leading to an increase in ROS production, thereby promoting RyR2 oxidation and further Ca2+ leak. Pharmacological targeting of leaky RyR2 channels or genetically inhibiting mitochondrial ROS production prevents AF providing mechanistic insights that could lead to new therapeutic targets for AF.

20All human analyses were performed in accordance with protocols approved by the Institutional Review Board of the New York Presbyterian Hospital and by the Ethics Committee of Columbia University. Written informed consent was obtained from all participants. RA appendage tissue was obtained at the time of cardiac surgery from patients with chronic AF (>6 months; n = 10), and patients in sinus rhythm (n = 10).

21All animal experiments were performed in accordance with NIH Guidelines for the Care and Use of Laboratory Animals, and animal protocols were approved by the Institutional Animal Care and Use Committee (IACUC) of Columbia University.

22Generation of the mCAT mouse has been described previously. RyR2-R2474S+/- and RyR2-S2808D+/+ mice were generated as described. All mice were backcrossed into the C57BL/6 background for >10 generations. All in vivo and in vitro experiments were conducted by operators who were blinded to the genotypes of the mice.

23For in vitro experiments using isolated atrial myocytes, 10 μM S107 was added to the extracellular solution for 2 hours. For in vivo experiments, S107 was diluted in drinking water at 0.25 mg/ml (40 mg/kg/d). No differences were detected in water consumption between vehicle and S107-treated groups.

24Intra-esophageal pacing was performed by placing in the esophagus, close to the left atrium, a 1.1-Fr octapolar catheter (EPR-800, Millar Instruments, Houston, Texas) connected to an external stimulator (STG-3008, MultiChannel Systems, Reutlingen, Germany). A computerized data acquisition system (EMKA Technologies, Falls Church, VA) was used to record a 3-lead surface ECG, and up to 4 intra-esophageal bipolar electrocardiograms. Inducibility of atrial arrhythmias was tested by applying a series of 2-second bursts. The first 2-second burst had a cycle length (CL) of 40 ms; then the CL was progressively decreased by 2 ms in each successive burst until reaching 10 ms. AF was defined as a period of rapid irregular atrial rhythm lasting at least 1 sec.

25Adult murine atrial myocytes were isolated as follows. The heart was rapidly isolated, cannulated and perfused with AfCS perfusion buffer, comprised of (in mM): NaCl 113, KCl 4.7, KH2PO4 0.6, Na2HPO4 0.6, MgSO4 1.2, NaHCO3 12, KHCO3 10, HEPES 10, taurine 30, glucose 1.5 and 2,3-Butanedione 2-monoxime (BDM) 10, for 5 minutes at a speed of 3 ml/min. Then, perfusion was switched to ‘digestion buffer’ (which includes 0.65 mg/mL type 2 collagenase and 50 μM CaCl2 in AfCS) and perfused for 10–15 minutes. Atria were cut and teased into small pieces in ‘stop 1’ buffer (0.65 mg/mL type 2 collagenase, 0.065 mg/mL Protease XIV, 15 mg/mL BSA and 50 μM CaCl2 in AfCS) and bathed at 37 °C for 10 minutes after which the enzyme was removed by centrifugation (4 min at 200 rpm). The cells were then resuspended in ‘stop 2’ buffer (15 mg/mL bovine serum albumin and 50 μM CaCl2 in AfCS) and gradually recovered to a final [Ca2+] of 1.8 mM.

26To measure intracellular Ca2+, cells were loaded with 5 μM Fluo-4 AM (ThermoFisher Scientific, Waltham, MA) for 20 minutes and washed three times and maintained in the following solution (in mM): NaCl 125, KCl 4.75, MgSO4 1.2, KH2PO4 1.2, HEPES 30, glucose 10, taurine 50, CaCl2 2 and pH = 7.4. A Leica TCS SP2 confocal microscopy with 40x, 1.25 NA oil immersion objectives was used for linescan imaging. The scan zoom was adjusted to fit the cells, and scan line was performed along the long axis of cells. The excitation for Fluo-4 was 488 nm, while emission was collected at 505–530 nm. For Ca2+ sparks recording, cells were scanned at 400 Hz for 20 s following 1 minute of pacing at 3 Hz. Ca2+ sparks detection and analysis was performed as previously described. KN-93 was purchased from Sigma-Aldrich (St. Louis, MO). For simultaneous recording of Ca2+ transient amplitudes and SR Ca2+ contents, cells were exposed to 10 mM caffeine immediately following termination of pacing at 1 Hz for 1 minute. Sampling started 10 s before caffeine treatment.

27For the evaluation of intracellular oxidative stress, cells were pre-incubated with the chloromethyl derivative CM-H2DCFDA (10 μM, ThermoFisher Scientific) for 30 minutes and washed. Fluorescence intensity and images were obtained using a confocal microscope (Zeiss 5 Live, 40 x oil immersion lens). Excitation was at 488 nm, and emission was collected at 505–530 nm. Since CM-H2DCFDA is light sensitive and oxidized progressively, we used the same scanning parameters for all experiments. For each dish, images were rapidly acquired for ~10 randomly selected cells. More than 100 cells per group were examined for intracellular fluorescent intensities.

28For mitochondrial ROS detection, cells were incubated for 20 min with 5 μM MitoSOX Red (Invitrogen/Molecular probes) and washed. Fluorescence intensity and images were obtained using confocal microscopy (Zeiss 5 Live, 40x oil immersion lens). MitoSOX Red was excited at 488 nm and emission was collected at 540–625 nm. The scanning parameters were unchanged for all the scans. For each group, fluorescence intensities of >100 cells randomly selected from several different dishes were examined.

29RyR2 was immunoprecipitated from atrial homogenates (250 μg) using anti-RyR antibody in 0.5 ml of a modified RIPA buffer (50 mM Tris-HCl pH 7.4, 0.9% NaCl, 5.0 mM NaF, 1.0 mM Na3VO4, 0.5% Triton-X100, and protease inhibitors) overnight at 4 °C. The samples were incubated with protein A sepharose beads (Amersham Pharmacia Biotech, Piscataway, NJ) at 4 °C for 1 hr and washed five times with 1.0 ml RIPA. Samples were heated to 95 °C and size fractionated by PAGE (6% for RyR2, 15% for calstabin 2). Levels of RyR2 bound proteins were normalized to the total RyR2 immunoprecipitated. All immunoblots were developed using the Odyssey system (LI-COR, Inc., Lincoln, NE) with IR labeled anti-mouse and anti-Rabbit IgG (1:10000 dilution) secondary antibodies. To detect RyR2 protein oxidation, atrial SR membrane samples (50 ug) were immunoprecipitated as described above. Immunoprecipitates were treated with 2, 4-dinitrophenylhydrazine (DNP) and the derivatized carbonyls were detected using an OxyBlotTM Protein Oxidation Detection Kit (Cat # S7150, Chemicon International, Inc., Temecula, CA). Proteins were size fractionated on 6% SDS-PAGE gels and transferred onto nitrocellulose membranes and immunoblots were developed with an anti-RyR antibody (Affinity Bioreagents, Bolder, CO, 1:2,000). The DNP signal associated with RyR was determined using an anti-DNP antibody (1:2000).

30Atria were fixed in 2.5% glutaraldehyde in 0.1 M Sørensen’s buffer and post-fixed in 1% OsO4. Following dehydration, samples were embedded in Lx-112 (Ladd Research Industries, Williston, VT). After cutting (ultramicrotome MT-7000), the 60 nm sections were stained with uranyl-acetate and lead-citrate and visualized (JEM-1200 EXII, JEOL, Tokyo, Japan), as previously described. For each animal at least twenty randomly selected sections were used for the analysis of mitochondrial morphology.

31All results are presented as mean ± s.e.m. Statistical analyses were performed using the unpaired Student’s t test (for 2 groups) and one-way ANOVA with Bonferroni post hoc correction (for groups of 3 or more) unless otherwise indicated. P < 0.05 was considered to be statistically significant.

32How to cite this article: Xie, W. et al. Mitochondrial oxidative stress promotes atrial fibrillation. Sci. Rep. 5, 11427; doi: 10.1038/srep11427 (2015).