|Home | About | Journals | Submit | Contact Us | Français|
Abnormal calcium release from sarcoplasmic reticulum (SR) is considered an important trigger of atrial fibrillation (AF). Whereas increased CaMKII activity has been proposed to contribute to SR leak and AF-induction, downstream targets of CaMKII remain controversial.
To test the hypothesis that inhibition of CaMKII-phosphorylated type-2 ryanodine receptors (RyR2) prevents AF initiation in FKBP12.6-deficient (−/−) mice.
Mice lacking RyR2-stabilizing subunit FKBP12.6 had a higher incidence of spontaneous and pacing-induced AF compared to wildtype mice. Atrial myocytes from FKBP12.6−/− mice exhibited spontaneous Ca2+ waves (SCaWs) leading to Na+/Ca2+-exchanger (NCX) activation and delayed afterdepolarizations (DADs). Mutation S2814A in RyR2, which inhibits CaMKII phosphorylation, reduced Ca2+ spark frequency, SR Ca2+ leak and DADs in atrial myocytes from FKBP12.6−/−:S2814A mice compared with FKBP12.6−/− mice. Moreover, FKBP12.6−/−:S2814A mice exhibited a reduced susceptibility to inducible AF, whereas FKBP12.6−/−:S2808A mice were not protected from AF.
FKBP12.6 mice exhibit AF caused by SR Ca2+ leak, NCX activation and DADs, which promote triggered activity. Genetic inhibition of RyR2-S2814 phosphorylation prevents AF induction in FKBP12.6−/− mice by suppressing SR Ca2+ leak and DADs. These results suggest that suppression of RyR2-S2814 phosphorylation as a potential anti-AF therapeutic target.
Atrial fibrillation (AF) is the most prevalent sustained arrhythmia and is associated with extensive morbidity and mortality.1 Defects in intracellular Ca2+ handling may directly contribute to pathogenic mechanisms underlying AF, such as ectopic activity and reentry.2 Previous studies revealed a higher open probability of type-2 ryanodine receptors (RyR2) isolated from dogs with chronic AF.3 This has been attributed to increased S2808 phosphorylation on RyR2 by protein kinase A (PKA),3 decreased levels of FK506-binding protein 12.6 (FKBP12.6) binding to RyR23 and increased Ca2+/calmodulin-dependent protein kinase II (CaMKII) phosphorylation of RyR2 at S2814.4 Whereas CaMKII inhibition might suppress aberrant sarcoplasmic reticulum (SR) Ca2+ release events associated with AF,4,5 it remains unknown whether phosphorylation of S2808 and/or S2814 play a causal role in AF induction. Moreover, it is unknown whether SR Ca2+ leak and spontaneous Ca2+ waves (SCaWs) via RyR2 generate Na+/Ca2+-exchanger (NCX) current sufficiently large to cause delayed afterdepolarizations (DADs) and triggered action potentials (APs). Here, we tested the hypothesis that genetic inhibition of RyR2 phosphorylation at S2808 and/or S2814 phosphorylation sites prevents AF induction in FKBP12.6-deficient (FKBP12.6−/−) mouse model of AF6, and assessed the underlying mechanisms at the cellular, organ and whole-animal level.
An expanded Methods section is available in Supplemental Materials and Methods. All animal studies were performed in accordance with the Guide for the Care and Use of Laboratory Animals published by the U.S. National Institutes of Health. To test the consequences of inhibition of RyR2 phosphorylation at S2814 on Ca2+ release and arrhythmogenesis, we intercrossed FKBP12.6−/− mice with S2814A knock-in mice4. In vivo electrophysiology was performed as previously described 4, 6. Ca2+ imaging and patch clamp studies were conducted in isolated atrial myocytes4, 6.
To determine whether FKBP12.6 deficiency can cause atrial arrhythmias at tissue level, optical mapping of isolated atria was performed. Figure 1 shows representative optical APs recorded at 3 sites in a FKBP12.6−/− mouse atrium during a spontaneous atrial tachyarrhythmia. Voltage mapping revealed fragmented activation as evidenced by an area of rapid activation or potential “driver” (Site 1 – cycle length (CL) = 80 ms) compared to areas of variable activation (Site 2 – variable voltage amplitude) and 2:1 activation (Site 3 – CL = 160 ms) (Fig. 1A). The incidence of spontaneous AF and atrial flutter was higher in FKBP12.6−/− mice compared to wildtype (WT) mice (P<0.05, Fig. 1A). This suggests that atria from FKBP12.6−/− mice are susceptible to spontaneous arrhythmias, which arise from a focal source driving the remaining atrium.
Localized diastolic Ca2+ release may produce DADs and triggered beats in intact tissue7. Therefore, we next measured SCaW-induced DADs and membrane depolarizing currents (INCX) in atrial myocytes. Simultaneous recordings of Ca2+ transients and membrane voltage revealed an increased incidence of SCaWs-induced DADs (Fig. 1B) in FKBP12.6−/− myocytes. Additionally, simultaneous recordings of SCaW and INCX revealed occurrence of SCaWs with corresponding INCX in FKBP12.6−/− myocytes only (Fig. 1C), directly linking SR Ca2+ leak with INCX and potentially-arrhythmogenic DADs. For those INCX-positive FKBP12.6−/− myocytes, there was a strong correlation between the amplitudes of SCaW and INCX.
Previous studies have demonstrated that FKBP12.6 deficiency promotes atrial arrhythmias owing to a destabilizing effect on RyR2.6 Frequency of spontaneous Ca2+ sparks (CaSF) in atrial myocytes from FKBP12.6−/− mice was higher than those from WT mice (7.9±0.9 vs. 2.9±0.51 sparks/100µm/s, P<0.001). Compared to FKBP12.6−/− mice, CaSF was significantly reduced in myocytes from FKBP12.6−/−:S2814A mice (7.9±0.90 vs. 4.0±0.42 sparks/100µm/s, P<0.001) (Fig. 2A). There was no difference in full-duration of half-maximum of Ca2+ sparks among the three genotypes (Fig. 2B), whereas there was a small decrease in full-width of half-maximum in FKBP12.6−/−:S2814A mice (Fig. 2C). Amplitude of the SR Ca2+ transient as an index of SR Ca2+ content was reduced in FKBP12.6−/− mice compared to WT mice (4.6±0.5 vs. 7.5±1.1 F/F0; P<0.05), and was partially restored in FKBP12.6−/−:S2814A mice (5.9±0.9 F/F0; P=0.31 vs. FKBP12.6−/−) (Fig. 2D).
Recent studies suggest that a fraction of SR Ca2+ leak is not detectable as Ca2+ sparks 8. Therefore, total SR Ca2+ leak was also measured using the tetracaine protocol (Online Figure I). The ratio of SR Ca2+ leak to SR Ca2+ load was significantly larger in FKBP12.6−/− mice compared with WT (11.8±0.8 vs. 5.8±0.8 %; P<0.001; Online Figure I).6 Conversely, SR Ca2+ leak was reduced in FKBP12.6−/−:S2814A myocytes to levels seen in WT mice (5.9±0.4%; P<0.05), suggesting that prevention of RyR2 phosphorylation at S2814 was sufficient to reverse abnormal SR Ca2+ leak in FKBP12.6−/− myocytes.
Next, we measured membrane potentials in myocytes from WT, FKBP12.6−/−, and FKBP12.6:S2814A mice at 3-Hz (Fig. 3). Mean AP durations at 30% and 50% repolarization (APD30 and APD50, respectively) were similar in the three genotypes (APD30 and APD50: 3.7±0.9; 9.9±1.9ms in WT; 4.5±0.9; 11.9±2.3ms in FKBP12.6−/−; 4.7±0.5; 10.2±0.9ms in FKBP12.6−/−:S2814A). Whereas WT myocytes typically showed pacing-induced APs only, we observed frequent non-paced (spontaneous) APs (Fig. 3A) and/or DADs (Fig. 3B) in FKBP12.6−/− myocytes. In fact, 73% (8 out of 11) cells from FKBP12.6−/− mice developed DADs, whereas only 16.7% (1 out of 6) of WT myocytes showed this behavior (P<0.05). In contrast, incidence of DADs was significantly reduced to 25% (4 out of 16) in FKBP12.6−/−:S2814A myocytes (P<0.05 vs. FKBP12.6−/−; Fig. 3C). Finally, simultaneous recordings of Ca2+ transients and membrane voltage, or Ca2+ transients and membrane current in FKBP12.6:S2814A myocytes under the same conditions as in Fig. 1B–C revealed significantly reduced incidence of SCaW-induced DADs and SCaW-induced INCX in FKBP12.6−/−:S2814A compared to FKBP12.6−/− myocytes (Online Figure II). Thus, inhibition of S2814 phosphorylation on RyR2 prevents triggered activity in FKBP12.6−/− mice myocytes.
We also determined whether inhibition of CaMKII phosphorylation of RyR2 could prevent AF induction in FKBP12.6−/− mice. We found that 53% (10 out of 19) of FKBP12.6−/− mice developed AF/atrial flutter following atrial-burst pacing, compared to 13% (2 out of 15) of WT mice (Fig. 4A). Incidence of AF decreased to 15% (3 out of 22) in S2814-ablated FKBP12.6−/−:S2814A mice (P<0.05 vs. FKBP12.6−/−), but was not decreased in FKBP12.6−/−:S2808A mice with S2808 ablation (43%; 10 out of 23; P=N.S. vs. FKBP12.6−/−). The number of burst-pacing trials leading to AF/atrial flutter induction (out of 3 attempts) was 1.63±0.27 in FKBP12.6−/− compared to 0.53±0.19 in WT mice (P<0.01; Fig. 4C). Whereas inhibition of S2808 phosphorylation did not significantly reduce the number of episodes (1.35±0.24) in FKBP12.6−/−:S2808A mice, ablation of S2814 in FKBP12.6:S2814A significantly decreased the number of inducible AF episodes to 0.55±0.16 (P<0.01). Inactivation of phosphorylation sites on RyR2 did not alter baseline electrophysiological and conduction properties (Online Table I). Finally, the importance of the S2814 phosphorylation site was confirmed in a second mouse model of RyR2 dysfunction previously associated with an elevated propensity towards AF as a result of heterozygosity for mutation R176Q in RyR2 (R176Q/+ mice)4 (Online Figure III).
To gain more insight into the mechanisms underlying induction of AF following rapid atrial pacing, we determined whether the CaMKII was activated in atria of paced mice (Online Figure IV.). There were no differences in CaMKII expression or the level of CaMKII-T287 auto-phosphorylation at baseline among the different genotypes. Following atrial-burst pacing, activation of CaMKII (evidenced by increased CaMKII-T287 auto-phosphorylation) resulted in enhanced phosphorylation of both RyR2 at S2814 and phospholamban (PLN) at T17 (Online Figure IV). There were no differences in S2808 phosphorylation on RyR2 and S16 on PLN before and after pacing (Online Figure V). Finally, protein expression levels of sarco/endoplasmic reticulum Ca2+-ATPase-2a (SERCA2a) and NCX were similar among all four genotypes (Online Figure V), suggesting that RyR2 mutations at S2808A and S2814A did not cause compensatory remodeling of SR Ca2+-handling proteins.
Studies over the past decade have demonstrated that SR Ca2+ release is abnormal in patients with chronic AF.2 Whereas the amplitude of the L-type Ca2+ current is generally decreased in AF, Ca2+ leak through RyR2 is typically elevated despite similar or decreased SR Ca2+ contents.2,5 It has been proposed that triggered activity due to DADs is caused by an inward depolarizing INCX current, which occurs in response to the removal of excess Ca2+ release from the cytosol.9,10 Here we provide direct experimental evidence for this mechanism in the FKBP12.6−/− mouse model of AF. Our data revealed that FKBP12.6−/− mice exhibit atrial focal activity and AF caused by SR Ca2+ leak, NCX activation and DADs generation. Because recent studies revealed that CaMKII phosphorylation of RyR2 at S2814 is elevated in patients with chronic AF,4,5 we investigated whether inhibition of S2814 phosphorylation of RyR2 affected susceptibility of FKBP12.6−/− to AF. Our results demonstrate that inhibition of S2814 but not S2808 phosphorylation suppressed pacing-induced AF in FKBP12.6−/− mice by preventing spontaneous SCaWs and related DADs. Therefore, our studies suggest that elevated CaMKII phosphorylation on RyR2 might be the primary phosphorylation event associated with triggered activity and AF induction,2 at least in the particular mutant mice examined in this study. This data are consistent with evidence showing that expression levels and activity of cytosolic CaMKII are upregulated in patients with chronic AF.4,5
Elevated CaMKII activity not only leads to increased RyR2 phosphorylation at S2814 but also causes increased PLN phosphorylation at T17, which might help preserve SR Ca2+ content in AF,11 by increasing SR Ca2+ uptake through SERCA2a disinhibition. We demonstrated that (partial) inhibition of S2814 phosphorylation on RyR2 is sufficient to suppress inducibility of AF following atrial-burst pacing in R176Q/+:S2814A/+ mice. Whereas this suggests that CaMKII phosphorylation of RyR2 contributes to SR Ca2+ leak associated with arrhythmogenesis, this did not exclude the possibility that CaMKII phosphorylation of PLN or other ion channels/transporters also contribute under some circumstances.
In conclusion, our data demonstrate that an increase in CaMKII phosphorylation of RyR2 at S2814 contributes to AF initiation in FKBP12.6−/− mice by amplifying SR Ca2+ leak and inducing DADs. Conversely, inhibition of CaMKII phosphorylation of RyR2 prevents AF initiation by decreasing aberrant SR Ca2+ release, NCX activation and DADs generation, whereas inhibition of S2808 phosphorylation of RyR2 failed to prevent AF induction. Together, our findings imply that CaMKII phosphorylation of S2814 on RyR2 might play an important role in enhancing RyR2-mediated SR Ca2+ leak that promotes DADs and atrial triggered activity associated with AF.
Previous studies demonstrated higher open probability of RyR2 in patients with chronic atrial fibrillation. Biochemical studies revealed increased phosphorylation levels of serine 2808 and serine 2814 on RyR2, as well as increased CaMKII activity. Here, we present evidence that inhibition of serine 2814 but not serine 2808 prevents induction of atrial fibrillation in the FKBP12.6-deficient mouse model of AF. Atrial myocytes from FKBP12.6-deficient mice exhibited spontaneous Ca2+ waves (SCaWs) leading to Na+/Ca2+-exchange current activation, and delayed afterdepolarizations (DADs). We therefore conclude that serine 2814 is an important downstream target of CaMKII in atrial fibrillation. We propose that defective sarcoplasmic reticulum calcium release via hyperphosphorylated RyR2 may cause triggered activity in atria, and contribute to the initiation of AF. Our findings suggest that inhibition of CaMKII phosphorylation of RyR2 could be a potential target for AF treatment.
Sources of Funding
This work was supported by grants from NIH (HL089598 and HL091947 to X.H.T.W., HL54807 to D.S.R., NRSA to M.J.C.), American Heart Association (09post2260300 to N.L.), Heart Rhythm Society (to M.J.C.), W.M. Keck foundation (to X.H.T.W.), European Union (EUTRAF Network, 261057, to D.D.), German Center for Cardiovascular Research (to D.D.) and Leducq Foundation (08CVD01 to X.H.T.W., 07CVD03 to D.D.).
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.