Leaky Ca2+ release channel/ryanodine receptor 2 causes seizures and sudden cardiac death in mice (original) (raw)
Heterozygous RYR2 missense mutations have been found in patients with CPVT and cohorts with sudden unexpected death (15). CPVT-linked RYR2 mutations are clustered in 3 disease-susceptible regions of the channel (24, 25). One mutation cluster occurs in the central RYR2 region and includes an _RYR2_-R2474S missense mutation originally described in identical twin brothers with syncope and exercise-induced ventricular tachycardia (VT) (13). Previously, we reported a gain-of-function defect in PKA-phosphorylated _Ryr2_-R2474S channels coexpressed with calstabin2 under conditions of low activating [Ca2+] to simulate cardiac diastole (the period in the cardiac cycle when arrhythmic triggers known as delayed afterdepolarizations [DADs] occur) during exercise (22). We showed that the _Ryr2_-R2474S mutant channels and 3 other CPVT-linked mutant RyR2 channels all exhibited a leaky phenotype under conditions that simulate cardiac diastole during exercise. This defective channel function was due to decreased binding of calstabin2 resulting in a shift to the left of the Ca2+ dependence of RyR2 activation such that these channels, unlike WT controls, were capable of being activated at extremely low (i.e., ~150 nM) cytosolic [Ca2+]. We interpreted these findings as being consistent with the generation of diastolic SR Ca2+ leak via mutant RyR2 channels in the heart that could trigger arrhythmias during exercise.
To examine the role of leaky RyR2 channels in vivo, we generated knock-in mice carrying the missense mutation R2474S in the endogenous Ryr2 gene (Ryr2RS/WT; see Figure 1A and Methods). The targeting vector for homologous recombination consisted of genomic DNA spanning a region that includes introns 46–54, with the codon change resulting in the R2474S variant occurring in exon 49. Homologous recombination was confirmed in ES cells using BamHI digestion (Figure 1B). Following germline transmission and mating of chimeras with 129Sv females, offspring were backcrossed into the C57BL/6 background for at least 5 generations, in agreement with previous arrhythmia studies in mice (26). Importantly, RyR2 expression was not different in the heart and brains of Ryr2-R2474S mice compared with WT littermate controls.
RyR2-homozygous R2474S knock-in mice exhibit increased embryonic lethality that is reduced by a novel RyR2 stabilizing drug (S107) that inhibits Ca2+ leak. (A) Generation of _Ryr2_-R2474S knock-in mouse. Top: targeted mutagenesis of mouse RyR2 exon 49; middle: homologous ES cell recombination of the mutant _Ryr2_-R2474S allele; bottom: Cre-mediated excision of the floxed neo cassette results in _Ryr2_-R2474S knock-in. (B) Confirmation of homologous recombination of mutant _Ryr2_-R2474S allele by Southern blot (left); PCR detects mutant R2474S (RS) allele in progeny (right). (C) Lethality of the homozygous _Ryr2_-R2474S (RS/RS) mice at day 28 after birth as evidenced by significant non-Mendelian inheritance with underrepresentation of the homozygous genotype. (D) Normal embryonic development and cardiac maturation up to day E13.5 as shown by representative histological sections. Original magnification: ×3 (left, longitudinal section), ×5 (right, cross-sections). (E) Rescue of homozygous _Ryr2_-R2474S embryos by treatment of the pregnant mothers with S107, a small compound that binds specifically to RyR2, enhances calstabin2 binding and inhibits Ca2+ leak from mutant RyR2 channels. Up to day E13.5, there was normal Mendelian inheritance (left). However, embryonic lethality was evidenced by abnormal Mendelian inheritance at day E16.5 (middle), and this was prevented by S107 treatment, which resulted in normal Mendelian inheritance due to improved survival of the homozygous _Ryr2_-R2474S embryos (right). P values represent statistical comparison of the observed genotypes to expected genotypes based on Mendelian inheritance patterns. Asterisks indicate significant difference between the observed and expected genotype ratios.
Heterozygous _Ryr2_-R2474S/WT (Ryr2RS/WT) mice were viable and survived into adulthood. However, _Ryr2_-R2474S knock-in mice were not born at the expected Mendelian ratios: observed/expected genotype ratios at weaning age were Ryr2WT/WT, 149/113.5 (32.8%); Ryr2RS/WT, 289/227 (63.6%); Ryr2RS/RS, 16/113.5 (3.5%); P < 0.0001 (χ2 = 111.7; Figure 1C). This result is in agreement with the fact that all previously identified RYR2 mutation carriers are heterozygous (individuals homozygous for CPVT-linked RyR2 mutations have not been reported) (15). The selective loss of homozygous Ryr2RS/RS embryos occurred due to increased intrauterine lethality. Indeed, homozygous Ryr2RS/RS mating did not produce any offspring from confirmed impregnated female mice, indicating intrauterine lethality of the Ryr2RS/RS genotype (n = 6).
RyR2 channels are first detected in the heart at day E8.5 in mice (27). Histology confirmed normal intrauterine development of homozygous Ryr2RS/RS mice until E13.5. In particular, cardiac maturation, chamber development, and circulation appeared to be normal at E13.5, as evidenced by transverse embryonic sections through the heart region (Figure 1D). While the expected Mendelian distribution was observed at E13.5, at E16.5 embryonic genotypes deviated significantly from the expected Mendelian distribution, consistent with an increased number of growth-retarded or digested Ryr2RS/RS E16.5 embryos due to intrauterine death (Figure 1E). Homotetrameric RyR2 channels comprising 4 CPVT mutant subunits exhibited a significantly greater gain-of-function defect as compared with heterotetrameric RyR2 channels (22), which is consistent with the fact that heterozygous Ryr2RS/WT mice did not show increased intrauterine mortality. To test the hypothesis that the intrauterine death observed in Ryr2RS/RS mice was due to defective RyR2 channel function (i.e., Ca2+ leak), we used a novel, RyR-selective 1,4-benzothiazepine derivative, S107 (23), that stabilizes RyR2 channels by enhancing the binding affinity of calstabin2 to mutant and/or PKA-phosphorylated channels. S107 is a small (MW, 245.7) compound that enhances calstabin2 binding to RyR2 at low nanomolar concentrations and failed to interact with over 400 receptors, enzymes, and ion channels in screens using up to 10 μM of the compound (see supplemental data). Specifically, S107 had no effect on cardiac ion channels including the voltage-gated Na+, K+, and Ca2+ channels at concentrations up to 10 μM, and S107 had no effect on normal Ca2+ signaling in cells. We reasoned that if the defective RyR2 function was causing late intrauterine death in Ryr2RS/RS mice, then S107 might prevent the excess Ryr2RS/RS mortality observed at E16.5. Indeed, we found that treatment of Ryr2RS/WT mothers with S107 (5 mg/kg s.c. continuously by osmotic minipump for 14 days) resulted in rescue of the expected Mendelian distribution at E16.5 (Figure 1E). On the basis of these data, we concluded that the likely cause of intrauterine mortality at E16.5 is excess Ca2+ leak via mutant defective RyR2 channels in Ryr2RS/RS mice.
Ryr2RS/WT mice began to exhibit spontaneous seizures during the weaning period between P21 and P28. There were no obvious developmental abnormalities, and the brains were histologically normal in _Ryr2_-R2474S mice (see supplemental data). The seizures were spontaneous and recurrent, and their frequency (1–3/week) did not change during a 3-month observation period (e.g., Supplemental Video 1). Generalized tonic-clonic seizures in Ryr2RS/WT heterozygous mice occurred following mouse placement in a new cage and/or arousal from sleep; however, seizures also occurred in the absence of these environmental changes (see Supplemental Video 1). Seizures exhibited a progression through several stages. Following an initial state of reduced activity and lying flat on the cage floor, the mice typically exhibited several behavioral signs of seizure activity (grimacing, ear deflection, nose hair stiffening, Straub tail, head bobbing), which quickly progressed from partial, isolated limb jerking to sustained and violent generalized myoclonic jerking. Following a generalized myoclonic phase, Ryr2RS/WT mice developed outstretched hind limbs consistent with tonic-clonic seizures, bilateral forelimb clonus, and head bobbing. The entire sequence of seizure activity typically lasted for 20–120 seconds, and both clonic and tonic-clonic seizures were recorded; this was followed by the mouse freezing and then resting for several minutes (corresponding to the postictal state). We were able to capture several spontaneous, generalized seizures by EEG recording, which confirmed the generalized nature of the seizures. Overall, spontaneous seizures were directly observed in 23 of 50 heterozygous Ryr2RS/WT and 2 of the very few homozygous Ryr2RS/RS mice surviving into adulthood but never in WT mice; however, it is likely that the actual incidence of seizures was higher than observed, because it was not possible to observe all of the mice constantly, and 100% of the mutant mice exhibited lowered thresholds for pharmacologic-induced seizures compared with WT littermate controls (see below).
Heterozygous Ryr2RS/WT mice with confirmed recurrent generalized seizures were implanted with cortical EEG electrodes and ECG telemeters and monitored during a 4-week period. Following electrode placement, spontaneous epileptiform activity was never observed in freely moving WT mice, which showed regular cardiac sinus rhythm (SR) (n = 3; e.g., Figure 2A). Similarly, heterozygous Ryr2RS/WT mice exhibited normal periods of low-amplitude baseline cortical EEG and regular SR activity (n = 4; e.g., Figure 2B). EEG monitoring from 2 heterozygous Ryr2RS/WT mice confirmed the occurrence of spontaneous electrographic and behavioral seizures, consistent with epileptiform ictal activity (Figure 2C). In agreement with video-documented behavioral abnormalities (see Supplemental Video 1), EEG rapid spike discharges during generalized tonic-clonic seizures typically lasted for 20–30 seconds. Moreover, slower spike discharges were observed in the postictal period (Figure 2D).
Epileptic seizures in heterozygous Ryr2RS/WT mice are inhibited by prevention of the leak via RyR2 channels. (A) Representative EEG and simultaneous ECG (ECG-EEG) recordings from a WT littermate mouse, showing normal brain activity and a normal cardiac SR; and (B) a heterozygous Ryr2RS/WT mouse during an interictal period without epileptiform activity, showing normal brain activity and a normal sinus heart rhythm. (C) Recordings from an Ryr2RS/WT mouse during an ictal episode, showing abnormal EEG activity but a normal cardiac rhythm (sinus tachycardia). The recorded seizure began with the Straub tail sign (single asterisk), followed by generalized tonic-clonic seizure (double asterisks), then spiking activity that gradually increased in amplitude and rhythmicity and ended with a depressed electric postictal period characterized by relaxed muscle tone, indicating end of seizure (triple asterisks). Consecutive 8-second traces from one EEG channel (left hemisphere) are shown; there was no difference in the spiking patterns between the left and right hemispheric activity; and (D) EEG recording during a postictal period with epileptiform activity (arrows). (E) Average latency to development of generalized tonic-clonic seizures and overall seizure score comparison in WT and Ryr2RS/WT mice following 4-AP and CAF susceptibility testing. *P < 0.05. (F) Average latency to development of generalized tonic-clonic seizures and overall seizure score comparison in Ryr2RS/WT mice treated with S107 subjected to a seizure protocol identical to that in E, showing that S107 partially protects against seizures. *P < 0.03 (left); *P < 0.04 (right). (G) Morphology of seizure-like events recorded from CA3 region of hippocampal slices incubated in 10 μM 4-AP from WT (n = 4) and Ryr2RS/WT (n = 4) mice shown at low and high time resolution. Short bars above traces indicate portion of recording that is expanded in the lower traces. (H) Histogram analysis of number of spikes per burst, showing longer bursts in the Ryr2RS/WT compared with WT mice. (I) Bar graphs showing analyses of complex burst activity only (defined as more than 10 spikes per burst, consistent with seizure-like activity). Left: Frequency per minute of complex bursts; middle: frequency of spikes per minute per complex burst; right: average number of spikes per complex burst, all of which were significantly elevated in Ryr2RS/WT compared with WT mice. *P < 0.005; †P < 0.0001, #P < 0.001.
Since seizures in CPVT mutation carriers have been attributed to malignant arrhythmias causing rapid compromise of cerebral blood supply (Stokes-Adams attacks) (18), we performed simultaneous ECG-EEG recording. During spontaneous tonic-clonic seizures with ictal EEG spike activity, mice displayed regular fast SR (sinus tachycardia, ~690/min); however, no peri-ictal arrhythmias were ever observed (Figure 2C, ECG-EEG). Our data show that spontaneously recurring seizures in heterozygous Ryr2RS/WT mice are due to abnormal brain excitability and are not associated with loss of consciousness (syncope) occurring from arrhythmias (i.e., the seizures are not Stokes-Adams attacks).
Since it was not possible to accurately assess the incidence of spontaneous seizures in _Ryr2_-R2474S mice, we sought to compare seizure susceptibility in WT and Ryr2RS/WT mice using pharmacological testing. Intraperitoneal administration of 4-aminopyridine (4-AP) has been reported to cause generalized tonic-clonic seizures and a prototypical progression of epileptic seizure activity in rodents (28). To enable quantitative examination of seizure susceptibility in the mice, we administered 4-AP (2.5 mg/kg i.p.) and the RyR agonist caffeine (CAF; 250 mg/kg i.p., administered 25 minutes after 4-AP), following which generalized tonic-clonic seizures (stage 3) occurred significantly earlier in Ryr2RS/WT mice (n = 5) compared with WT (Figure 2E, left; n = 5, P < 0.05). Moreover, in Ryr2RS/WT mice, combined 4-AP and CAF treatment resulted in faster progression to a significantly higher seizure severity score, including death (Figure 2E, right). To further determine whether the observed seizure activity was in fact due to defective RyR2 channel function, we tested whether treatment with S107, which inhibits Ca2+ leak via mutant RyR2 channels, could protect against seizures in the _Ryr2_-R2474S heterozygous mice. Pretreatment for 1 week with S107 (5 mg/kg/h via s.c. osmotic pump) significantly increased the latency time for developing seizures and reduced the seizure score in the _Ryr2_-R2474S heterozygous mice (vehicle-treated Ryr2RS/WT, n = 7; S107-treated Ryr2RS/WT, n = 10; P < 0.03 for latency time and P < 0.04 for seizure score; Figure 2F).
Certain behavioral aspects of the characteristic epileptiform activity in Ryr2RS/WT mice are indicative of a hippocampal seizure origin in rodent epilepsy models. To explore this possibility, we performed extracellular local field potential (LFP) recordings in the CA3 region of hippocampal brain slices that were exposed to 4-AP (10 μM). Ryr2RS/WT hippocampal slices (n = 4) exhibited significantly increased excitability in response to 4-AP, as evidenced by a significantly higher frequency of more-complex interictal burst discharges (>10 spikes/burst), higher spike rates per burst, and higher spike numbers per complex burst (n = 4; Figure 2, G–I). When population burst discharge activity following 4-AP treatment (10 μM) over 10 minutes was compared in WT (n = 488 events analyzed) and Ryr2RS/WT slices (n = 322 events analyzed), a significant shift toward longer and faster burst activity consistent with a second peak in the event distribution was evident (Figure 2H). The observed increase in 4-AP–induced complex burst activity in the Ryr2RS/WT compared with WT hippocampal CA3 region suggests that the seizure activity in the Ryr2RS/WT mice may originate in the hippocampus.
To determine whether the mechanism of the seizures observed in Ryr2RS/WT mice involves aberrant Ca2+ signaling in the hippocampal CA3 region, we performed confocal Ca2+ imaging experiments using hippocampal brain slices (n = 3). During perfusion with a low-Mg2+ (0.5 mM) and high-K+ (8.5 mM) solution, cells in the principal hippocampal CA3 layer (Figure 3B) exhibited profound burst activity compared with quiescent cells under control conditions (Figure 3A, left vs. middle) or cells from WT (n = 4) mice hippocampal brain slices examined under the same conditions (data not shown). Moreover, addition of the specific RyR channel blocker ryanodine (10 μM) reduced the amplitude and frequency of the burst activity, suggesting that leaky Ryr2RS/WT activity likely contributes to Ca2+-dependent burst activity (Figure 3A, right). Histological examination of the WT and Ryr2RS/WT mouse brains using H&E staining (e.g., Figure 3C, left; n = 5 Ryr2RS/WT brains, n = 5 WT brains) and trichrome staining (data not shown) did not reveal any abnormalities in the hippocampal formation or other brain areas of Ryr2RS/WT mice, and RyR immunostaining confirmed normal RyR2 expression, particularly in the CA3 region (e.g., Figure 3C, right; see also Supplemental Figure 3). These data indicate that the increased susceptibility to seizure-like activity in vitro and generalized seizures in vivo are likely associated with high-frequency Ca2+ signaling and/or burst activity.
Hippocampal Ryr2RS/WT brain slices and channels exhibit burst activity, which can be inhibited by treatment with ryanodine or the RyR-stabilizing drug S107, respectively. (A and B) Continuous confocal Ca2+ fluorescence imaging of the Ryr2RS/WT CA3 principal cell layer under control conditions (left), and seizure activity induced by low Mg2+ (0.5 mM) plus high K+ (8.5 mM) (middle) and following ryanodine (10 μM) treatment (right). Fluorescence (F) signals 1–3 in A correspond to regions of interest indicated by white circles in the CA3 layer in B. Data are representative of 3 experiments using Ryr2RS/WT hippocampal slices; dimensions are as indicated. (C) H&E histology (left) and RyR2 immunohistochemistry (right) of the hippocampal CA3 region in Ryr2RS/WT brain slices show increased RyR2 expression in the preserved principal cell layer. There were no histological abnormalities compared with WT (data not shown). (D) Representative Ryr2RS/WT single-channel traces from vesicles of isolated hippocampus from sedentary mice (left), after injection of NE (5 mg/kg twice over 3 hours; middle) or after 1 week treatment with S107 (5 mg/kg/h) followed by NE treatment (5 mg/kg twice over 3 hours; right). _P_o, mean open (To) and mean closed (Tc) times, closed state (c), and the fully open level (4 pA) are indicated. Thick bars above the 5-second traces indicate area shown at higher resolution in the 0.5-second traces. All-point histograms corresponding to the single-channel traces show increased numbers of partial openings (subconductance states) and overall increased activity of the brain channels from NE-treated Ryr2RS/WT mice (middle histogram). The histogram on the right shows more channels in the closed state (0 pA), consistent with the channel-stabilizing properties of the drug S107 that enhances binding of the calstabin2 subunit to the channel. (E) Average P_o of WT and Ryr2RS/WT brain channels under different treatment conditions as indicated. Single-channel measurements were performed at a cis (cytosolic) Ca2+ concentration of 150 nM. *P < 0.05 versus NE-untreated; †_P < 0.05, NE-treated WT versus Ryr2RS/WT. Each bar represents the average of 7–9 channels. Equivalent amounts of RyR2 were immunoprecipitated from brain homogenates with an RyR2 isoform–specific antibody followed by immunoblotting; bar graphs show the amount of PKA phosphorylation of RyR2 at Ser2808 (F) and the amount of calstabin2 bound to RyR2 (G) under the indicated conditions. Animals were treated with S107 via implantable osmotic pumps (5 mg/kg/h) for 7 days before NE stimulation. *P < 0.05 versus NE-untreated.
To further characterize the molecular mechanisms that may contribute to abnormal cellular Ca2+ signals, we studied native RyR channels harvested from the isolated mouse hippocampus. RyR2 Ca2+ release channels comprise 4 RyR2 monomers, each of which binds a single calstabin2 subunit (also known as FKBP12.6), which stabilizes the channel closed state (22, 29). We have previously found significantly increased channel activity and calstabin2 depletion in recombinant homotetrameric _Ryr2_-R2474S as well as in heterotetrameric _Ryr2_-WT/R2474S channels following in vitro PKA phosphorylation (22). Since activation of noradrenergic stimulation can result in epileptic synchronization and altered hippocampal activity (30), we treated Ryr2RS/WT and WT littermates with norepinephrine (NE; 5 mg/kg/h s.c. by osmotic pump for 48 hours) and measured RyR single-channel activity from brain vesicles fused to planar lipid bilayers. The open probability (_P_o) of hippocampal RyR channels was significantly increased in NE-treated compared with untreated Ryr2RS/WT mice (Figure 3, D and E). Current amplitude histograms at 150 nM cis Ca2+ were consistent with increased _P_o including multiple subconductance states, suggesting a leaky channel phenotype (Figure 3D, middle). Given that 3 RyR isoforms are expressed throughout the brain, we could not be sure that the channels in these single-channel recordings were RyR2; however, RyR2 is the most abundant isoform in the hippocampus (the exception being cerebellar Purkinje neurons expressing primarily the skeletal muscle RyR1 isoform; see Supplemental Figure 3). As compared with channels from NE-treated WT mouse brains, Ryr2RS/WT channels showed a significantly higher average _P_o, consistent with a gain-of-function defect (Figure 3E; n = 9, P < 0.05). Moreover, Ryr2RS/WT mice pretreated for 1 week with the RyR2-stabilizing drug S107 showed brain RyR channel activity comparable to that in animals not treated with NE and consistent with normalized function (Figure 3, D and E; n = 9, P < 0.05).
RyR2 immunoprecipitation (using an antibody specific for the RyR2 isoform) from Ryr2RS/WT brain homogenates followed by immunoblotting showed no change in RyR2 levels compared with WT (data not shown), and NE treatment resulted in a similar increase in RyR2 PKA phosphorylation in WT and Ryr2RS/WT brains (Figure 3F). Calstabin2 depletion from PKA-phosphorylated RyR2 complexes was significantly increased in both WT and Ryr2RS/WT brains following NE stimulation (Figure 3G). When Ryr2RS/WT mice were pretreated with the blood brain barrier–permeable S107 compound (5 mg/kg/h s.c. for 1 week) prior to NE treatment, calstabin2 binding to PKA-phosphorylated Ryr2RS/WT brain channels was restored to normal levels observed in nonphosphorylated channels (Figure 3G), consistent with single-channel function being normalized (Figure 3D, right). In these experiments, WT RyR1 and RyR3 channels were likely also PKA phosphorylated in mice treated with NE. However, previous studies have shown that despite transient PKA phosphorylation, WT RyR channels do not develop pathologic Ca2+ leak that is observed in PKA-phosphorylated CPVT-linked mutant RyR2 channels (16, 22). This is likely due to the finding that the CPVT-linked mutant RyR2 channels bind the stabilizing protein calstabin2 at a lower affinity and are, therefore, more prone to developing a significant, persistent Ca2+ leak.
Since telemetric ECG recording of cage-habituated Ryr2RS/WT mice did not show any sustained VT leading to syncope or seizures, we performed strenuous exercise stress testing followed by catecholamine injection using previously established protocols (22, 31). WT mice did not exhibit any arrhythmias when this protocol was used (Figure 4A; n = 6). Exercise at incrementally faster treadmill speeds for at least 45 minutes followed by epinephrine injection (0.5 mg/kg i.p.) resulted in sustained polymorphic VT (sVT) and SCD in Ryr2RS/WT mice (Figure 4B; n = 8 of 9). The cardiac arrhythmias in the _Ryr2_-R2474S mice typically started with premature ventricular ectopic beats, progressing into bidirectional and polymorphic VTs (Figure 4B) similar to the stress-induced arrhythmias observed in human RyR2 mutation carriers. These arrhythmias typically degenerated into polymorphic VT and/or fibrillation and finally sudden death. SCD in Ryr2RS/WT mice was associated with occasional limb jerking and rapid breathing. However, during or after sVT, none (n = 0 of 9) of the Ryr2RS/WT mice displayed the sustained generalized tonic-tonic seizures described above (Figure 2; see also Supplemental Video 1). These data further suggest that spontaneous seizures in Ryr2RS/WT mice are likely due to epilepsy and are not associated with sustained cardiac arrhythmias.
Fixing the leak in mutant channels from Ryr2RS/WT mice with S107 protects against fatal cardiac arrhythmias. (A–C) Representative telemetric ECG recordings of WT (n = 6), heterozygous Ryr2RS/WT (n = 9), and Ryr2RS/WT mice treated with S107 (n = 9). (A) ECGs of a WT mouse, sedentary (rest) and after 45 minutes of treadmill running immediately followed by catecholamine injection (ex + EPI; epinephrine, 0.5 mg/kg i.p.). (B) ECGs of a Ryr2RS/WT mouse, sedentary and following arrhythmia provocation stress testing, which resulted in rapid sVT and SCD. *Bidirectional VT; **polymorphic VT; ***rapid polymorphic VT. (C) Ryr2RS/WT mice treated with S107 under sedentary housing conditions (rest) and following stress testing (S107, 5 mg/kg/h s.c. for 7 days by osmotic pump). S107 prevented stress-induced arrhythmias. (D) Occurrence of sVT (left) and SCD (right). (E) Example of Langendorff-perfused Ryr2RS/WT hearts (n = 6) that exhibited regular SR recorded by volumetric ECG (vECG) with 2 epicardial breakthroughs (arrows) and homogenous apical-to-basal voltage activation. (F) Epicardial voltage activation map of the same Ryr2RS/WT heart showing multiple activation foci (arrows) and abnormal activation wavefront propagation during rapid polymorphic VT. The black dot marks the last regular sinus beat of vECG, and the red dot the first abnormal beat occurring at a short coupling interval. (G) Ryr2RS/WT heart showing abnormal focal activation (arrows) rapidly moving toward apex and left ventricle during sVT. (H) Occurrence of sVT in ex vivo perfused WT and Ryr2RS/WT hearts (each n = 6). Perfusion in the absence or presence of either high extracellular Ca2+ (9 mM) or ISO (100 nM) resulted in sVT in 9 of 12 Ryr2RS/WT but only in 1 of 12 WT hearts. Orientation of the heart is as indicated; a, apex; b, base; l, left; r, right of anterior activation maps.
Using a different CPVT mouse model (calstabin2-deficient mice), we previously showed that VT and SCD could be prevented by pretreatment with JTV519, a 1,4-benzothiazepine (32, 33). Since JTV519 has off-target activity including potent HERG K+ channel block that can cause drug-induced long QT potentially predisposing to polymorphic VT, we have developed a novel orally available derivative (S107) with higher RyR2 activity and no significant off-target effects (supplemental data, S107 Characterization). Using previously established protocols (32), Ryr2RS/WT mice were treated for 1 week with S107 (5 mg/kg/h s.c. by osmotic pump) prior to exercise stress testing. In contrast to untreated Ryr2RS/WT mice, of which 8 of 9 exhibited sVT and 7 of 9 had SCD, S107-treated Ryr2RS/WT mice subjected to exercise stress testing and low-dose epinephrine injection were less likely to develop sVT (2 of 9) and SCD (n = 1 of 9; Figure 4, C and D). Since S107 has high RyR2 specificity and efficacy in vitro, we hypothesized that it may prevent arrhythmias and SCD though an RyR2-targeted mechanism in vivo (see below).
Since a variety of ventricular arrhythmia types have been documented in RYR2 mutation carriers (34), we sought to determine which form of arrhythmia is associated with SCD (see supplemental data for details). Optical mapping of the ventricular arrhythmias (Figure 4, E–H) indicated a catecholaminergic and/or intracellular Ca2+ overload–dependent mechanism of arrhythmia induction in CPVT hearts, consistent with previous studies (35, 36).
Isolated ventricular cardiomyocytes from Ryr2RS/WT hearts were patch-clamped for action potential (AP) recording under current clamp conditions and continuously paced at 1 Hz in control extracellular solution, followed by stimulation with isoproterenol (ISO; 1 μM). In contrast to WT cells (data not shown), ISO-treated Ryr2RS/WT cardiomyocytes exhibited DADs between pacing cycles (Figure 5A, asterisks). DADs progressively increased the membrane depolarization rate from 1 Hz (pacing frequency) to approximately 3 Hz (pacing and DADs), and coupling intervals were decreased, resulting in AP fusion, incomplete repolarization, and spontaneous pacemaker activity (automaticity). In many cells, automaticity was self-sustained, e.g., trains of DADs would continue spontaneously after cessation of pacing (data not shown), confirming earlier results from _Ryr2_-R4496C knock-in cells (37). Similarly, ISO-treated Ryr2RS/WT cardiomyocytes exhibited aberrant intracellular Ca2+ transients (asterisks) between regular field-stimulated Ca2+ transients (Figure 5B) that became self-sustained when pacing was stopped, consistent with regenerative intracellular Ca2+ release.
Electrical and Ca2+ cycling abnormalities in Ryr2RS/WT cardiomyocytes are consistent with Ca2+-triggered afterdepolarizations and are reduced by the RyR2-stabilizing drug S107. (A) Representative examples of whole-cell current (_E_m) recordings from Ryr2RS/WT cardiomyocytes paced continuously at 1 Hz (recording started after 2 minutes; pacing stimuli as indicated) under control (left) and ISO-stimulated conditions (1 μM; right). Following ISO, aberrant membrane depolarizations (asterisks) increased in frequency and became self-sustained when pacing was stopped (not shown). (B) Intracellular Ca2+ transients from Ryr2RS/WT cardiomyocytes field-paced continuously at 1 Hz under control (left) and ISO-stimulated conditions (1 μM, right; recording started after 2 minutes continuous pacing). Aberrant spontaneous Ca2+ waves (asterisks) occurred irregularly and became self-sustained when pacing was stopped (not shown). (C) Representative current traces from Ryr2RS/WT cardiomyocytes recorded during an 0.5-Hz depolarization train in the absence (left) and presence of ISO (1 μM; right). ISO-treated Ryr2RS/WT cardiomyocytes showed frequent _I_TI (asterisks) during and after pacing. The figure illustrates currents recorded during and following the first 2 and last 5 pulses from a continuous 10-pulse (1 Hz) conditioning train. A 5.5-second-long period of the recording during the train was omitted for display purposes, and the time scale is expanded in the ISO-treated examples in the rights panels of A–C. (D) Normalized _I_TI density in ISO-treated cells. ISO-treated Ryr2RS/WT cardiomyocytes (n = 7) showed significantly increased _I_TI densities (*P < 0.05 vs. control; n = 6); in vivo S107 treatment significantly decreased _I_TI density in ISO-treated Ryr2RS/WT cells (#P < 0.05; n = 4). (E) Simultaneous confocal Ca2+ imaging of a small region of interest and _I_TI recording from an ISO-stimulated Ryr2RS/WT cardiomyocyte. Following regular pacing-induced Ca2+ release (last cycle shown on left side), intracellular Ca2+ and membrane current rapidly normalized to the resting (diastolic) state. Ryr2RS/WT cardiomyocytes showed abnormal intracellular Ca2+ release events, which became organized as Ca2+ waves, coinciding with secondary arrhythmogenic _I_TI. Scale bar: 10 μm; time and normalized F/F0 fluorescence signal are as indicated. (F) Representative confocal Ca2+ spark images from WT and Ryr2RS/WT cardiomyocytes before (–ISO) and after (+ISO; 1 μM) treatment. Bar graphs show significant differences (*P < 0.05) in average spark frequencies and average spatial area dimensions, indicating increased intracellular Ca2+ leak in Ryr2RS/WT cardiomyocytes following ISO stimulation as the subcellular origin of arrhythmogenic diastolic DADs (A), Ca2+ waves (B and E), and _I_TI (C and E) events.
We have previously shown that abnormal intracellular Ca2+ release from RyR2s in a mouse model of catecholaminergic arrhythmias activates arrhythmogenic transient inward current (_I_TI) as a mechanism of triggered activity and DADs (33). Accordingly, in ISO-treated Ryr2RS/WT cardiomyocytes, we observed _I_TI events during and immediately following cessation of pacing (e.g., Figure 5C). Control Ryr2RS/WT cells without ISO treatment exhibited significantly fewer _I_TI events, consistent with earlier studies (33) (n = 6, P < 0.05), and β-adrenergic stimulation of Ryr2RS/WT cardiomyocytes resulted in significantly increased cellular _I_TI density (Figure 5D; n = 7, P < 0.05). Moreover, treatment of Ryr2RS/WT mice with S107 for 1 week significantly decreased _I_TI density (Figure 5D; n = 4, P < 0.05). Simultaneous intracellular Ca2+ and _I_TI recording in Ryr2RS/WT cardiomyocytes showed that regular pacing-induced Ca2+ transients were followed by abnormal intracellular Ca2+ release events, which became organized into cell-wide Ca2+ waves (Figure 5E, top). Interestingly, arrhythmogenic _I_TI formation coincided with early aberrant Ca2+ release (starting at ~1.1 seconds) and became progressively amplified during regenerative cell-wide Ca2+ waves (Figure 5E, bottom). We further characterized the spontaneous Ca2+ release from RyR2 by measuring sparks in resting cells in the presence or absence of ISO. While 1 μM ISO had a small effect on the WT cardiomyocytes, the frequency and area of Ca2+ sparks were dramatically increased in the ISO-treated Ryr2RS/WT cardiomyocytes (Figure 5F: WT, n = 6; Ryr2RS/WT, n = 9; P < 0.05). The increased spark areas and frequencies in Ryr2RS/WT cells are consistent with our observation of Ca2+ spark fusions and repetitive activation of regenerative waves as the cause of arrhythmogenic _I_TI (Figure 5, C and E). We conclude that an abnormal temporal and spatial increase in microscopic Ca2+ release events causes arrhythmogenic _I_TI, DADs, and automaticity in Ryr2RS/WT cardiomyocytes.
Following exercise stress testing consisting of treadmill running and EPI injection, which resulted in polymorphic VT in Ryr2RS/WT mice (n = 8 of 9), hearts from WT and Ryr2RS/WT mice were flash frozen for SR vesicle isolation for single-channel measurements in lipid bilayers. Single channels from sedentary Ryr2RS/WT mice exhibited multiple partial opening events (subconductance states) and a low _P_o at 150 nM [Ca2+]cis simulating diastolic conditions (see also Methods; Figure 6A, left). In contrast, following stress testing, Ryr2RS/WT single channels exhibited a significant increase in _P_o (Figure 6A, center). One week of S107 treatment followed by stress testing resulted in redistribution of open events toward the 0-pA (closed) level (Figure 6A, right). The average RyR2 _P_o was significantly increased in both EPI-treated WT and Ryr2RS/WT hearts. However, the _P_o increase following stress testing was significantly higher in channels from Ryr2RS/WT hearts, indicating a gain-of-function defect (Figure 6B; n = 7, P < 0.05). Compared with those from untreated Ryr2RS/WT mice undergoing stress testing, RyR2 channels from S107-treated hearts showed a significantly lower _P_o, comparable to that observed in RyR2 channels from sedentary mice (Figure 6B; n = 7, P < 0.05).
RyR2 channels from heterozygous Ryr2RS/WT hearts show a gain-of-function defect that is rescued by S107 treatment. (A) Representative single-channel traces from isolated from the hearts of sedentary Ryr2RS/WT mice (RS/WT), after maximal exercise followed by injection of 0.5 mg/kg epinephrine (RS/WT + EPI), or after 1 week treatment with S107 (5 mg/kg/h) followed by maximal exercise and EPI injection (RS/WT + EPI + S107). Thick horizontal bars below 5-second traces indicate area shown in 0.5-second traces. _P_o, mean open (To) and mean (Tc) closed times, closed state (c), and fully open level (4 pA) are as indicated. Corresponding all-point histograms demonstrate altered current amplitude distribution, including multiple subconductance states and full open events in the EPI group, consistent with a gain-of-function defect. In contrast, S107-treated group histograms show redistribution toward closed states. (B) Average P_o of single cardiac WT and Ryr2RS/WT channels under different treatment conditions. Single-channel measurements were performed in low activating Ca2+ concentrations (150 nM) to mimic diastolic conditions. *P < 0.05 versus hearts from sedentary mice; #P < 0.05 versus hearts from exercise- and EPI-treated mice; †_P < 0.05 between sedentary and EPI-treated Ryr2RS/WT mice. Each bar represents the average of 7 channels. (C–E) Equivalent amounts of RyR2 were immunoprecipitated with an RyR2-specific antibody followed by immunoblotting: amounts of relative PKA phosphorylation of RyR2 at Ser2808 (D) and of calstabin2 bound to RyR2 (E) under the indicated conditions. Data refer to the same cardiac vesicles used in A and B. *P < 0.05 versus hearts from sedentary mice; #P < 0.05 versus hearts from exercise- and EPI-treated mice.
We have previously found significantly increased activity and calstabin2 depletion in recombinant _Ryr2_-R2474S channels following in vitro PKA phosphorylation (22). RyR2 immunoprecipitation from cardiac homogenates followed by immunoblotting showed no change in RyR2 protein levels in WT and Ryr2RS/WT hearts. Exercise followed by EPI treatment resulted in similar increases in RyR2 PKA phosphorylation in WT and Ryr2RS/WT hearts (Figure 6, C and D). However, following stress testing, Ryr2RS/WT hearts revealed significant calstabin2 depletion from the PKA-phosphorylated RyR2 complex (Figure 6E; n = 3, P < 0.05). These data are consistent with earlier in vitro experiments showing significantly decreased calstabin2 binding to recombinant homo- and heterotetrameric _Ryr2_-R2474S channels (22). When Ryr2RS/WT mice were pretreated with S107 (5 mg/kg/h s.c. for 1 week) prior to stress testing, calstabin2 binding to PKA-phosphorylated Ryr2RS/WT cardiac channels was normalized (Figure 6E), consistent with rescue of single-channel function (Figure 6, A and B).