Allele-Specific Silencing of Mutant mRNA Rescues Ultrastructural and Arrhythmic Phenotype in Mice Carriers of the R4496C Mutation in the Ryanodine Receptor Gene (RYR2)Novelty and Significance
Rationale: Mutations in the cardiac Ryanodine Receptor gene (RYR2) cause dominant catecholaminergic polymorphic ventricular tachycardia (CPVT), a leading cause of sudden death in apparently healthy individuals exposed to emotions or physical exercise.
Objective: We investigated the efficacy of allele-specific silencing by RNA interference to prevent CPVT phenotypic manifestations in our dominant CPVT mice model carriers of the heterozygous mutation R4496C in RYR2.
Methods and Results: We developed an in vitro mRNA and protein-based assays to screen multiple siRNAs for their ability to selectively silence mutant RYR2-R4496C mRNA over the corresponding wild-type allele. For the most performant of these siRNAs (siRYR2-U10), we evaluated the efficacy of an adeno-associated serotype 9 viral vector (AAV9) expressing miRYR2-U10 in correcting RyR2 (Ryanodine Receptor type 2 protein) function after in vivo delivery by intraperitoneal injection in neonatal and adult RyR2R4496C/+ (mice heterozygous for the R4496C mutation in the RyR2) heterozygous CPVT mice. Transcriptional analysis showed that after treatment with miRYR2-U10, the ratio between wild-type and mutant RYR2 mRNA was doubled (from 1:1 to 2:1) confirming the ability of miRYR2-U10 to selectively inhibit RYR2-R4496C mRNA, whereas protein quantification showed that total RyR2 was reduced by 15% in the heart of treated mice. Furthermore, AAV9-miRYR2-U10 effectively (1) reduced isoproterenol-induced delayed afterdepolarizations and triggered activity in infected cells, (2) reduced adrenergically mediated ventricular tachycardia in treated mice, (3) reverted ultrastructural abnormalities of junctional sarcoplasmic reticulum and transverse tubules, and (4) attenuated mitochondrial abnormalities.
Conclusions: The study demonstrates that allele-specific silencing with miRYR2-U10 prevents life-threatening arrhythmias in CPVT mice, suggesting that the reduction of mutant RyR2 may be a novel therapeutic approach for CPVT.
- cardiac arrhythmia
- cardiac sudden death
- cathecholaminergic polymorphic ventricular tachycardia
- gene therapy
- Ryanodine receptor calcium channel
- transgenic mice
Catecholaminergic polymorphic ventricular tachycardia (CPVT) is a genetic disease characterized by stress-induced life-threatening arrhythmias occurring in otherwise healthy individuals. CPVT usually manifests in early childhood with unexplained syncopal spells in patients with normal heart.1
Editorial, see p 480
Meet the First Author, see p 470
Dominant CPVT is caused by mutations in the Ryanodine Receptor (RYR2) gene2 that encodes a large 560 KDa protein that forms a homotetrameric ion channel localized in the membrane of the junctional sarcoplasmic reticulum (jSR). RyR2 (Ryanodine Receptor type 2 protein) channels face the L-type Ca2+ channels (Cav1.2), which are positioned in specialized invaginations of the cellular membrane called transverse tubules (TT): influx of Ca2+ through these channels triggers RyR2-mediated Ca2+ release from the sarcoplasmic reticulum (SR), eliciting contraction of the heart.3
In normal myocytes, sympathetic activation promotes phosphorylation of RyR2 and increases Ca2+ release from the SR, whereas, in the presence of RYR2 gain-of-function mutations, in addition to the physiological response, adrenergic activation also elicits spontaneous diastolic Ca2+ release. In response to the increase in cytosolic Ca2+, the Na+/Ca2+ exchanger extrudes 1 Ca2+ ion for 3 Na+ ions: this electrogenic stoichiometry generates delayed afterdepolarizations (DADs) that depolarize cardiac cells and precipitate triggered arrhythmias.4
The first-line therapy to prevent adrenergically mediated arrhythmias in CPVT is β-blockers. However, follow-up data in different populations2,5 showed recurrence of arrhythmic events in about 25% of patients, thus highlighting the incomplete protection of this treatment. Patients who show recurrent arrhythmias on β-blockers are treated with flecainide. In high-risk patients with arrhythmic outbreaks despite optimal medical therapy, the implant of a defibrillator is recommended.6 The burden of life-long treatment with partially effective drugs and the complications related to the implant of a defibrillator in young patients poses the rationale for the development of more effective therapeutic approaches for CPVT.
We previously demonstrated that adeno-associated virus (AAV)–mediated delivery of the cDNA of the wild-type (WT) sequence of the Calsequestrin 2 (CASQ2) gene to knock-in mice affected by the recessive form of CPVT is highly effective in preventing life-threatening arrhythmias.7 Our data also showed that AAV-mediated delivery of WT cDNA of CASQ2 reverts all the biomarkers of the disease such as reduction of levels of CASQ2, triadin, and junctin, the abnormal intracellular calcium physiology, and the ultrastructural abnormalities of the jSR and the TT.7 In the attempt to devise a gene therapy strategy for the dominant form of CPVT, we hypothesized that the silencing of the mutant allele of the RYR2 gene using RNA interference could represent a viable strategy to test in our RyR2R4496C/+ (mice heterozygous for the R4496C mutation in the Ryanodine Receptor 2 gene) heterozygous CPVT mice that we extensively characterized in the past 10 years.8–10 This approach has to face several possible challenges because it is unpredictable whether it will be possible to obtain in vivo a highly selective silencing of the mutant allele devoid of silencing of the WT protein. Furthermore, it is uncertain whether RNA interference will be able to reduce a significant amount of the mutant protein sufficient to reduce the arrhythmic burden.
For a more detailed description of Methods, the readers are encouraged to access the Online Data Supplement.
Cloning of Reporter Alleles, siRNA Design, and Selection
We developed an in vitro mRNA and protein-based assay suitable to screen multiple siRNAs designed to selectively target the murine R4496C mutant allele over the WT RYR2 allele. We used 2 mini-constructs containing the WT and the mutant portion of the RYR2 gene encompassing exons 91 to 96 linked with a protein tag and a fluorescent reporter gene (Online Figure IA) to screen a set of siRNA duplexes. These were modeled on the murine RYR2 mRNA sequence spanning the R4496C point mutation site and differing for the position of the mutation recognition site, the key parameter that determines the thermodynamic properties of the pairing between the siRNA and the target mRNA and influences the specificity of the silencing activity (Online Figure IB). Heterozygous conditions were generated by cotransfecting the 2 reporter alleles and siRNA duplexes into cultured HEK-293 (human embryonic kidney cells 293) cells.
Among the tested molecules designed to target RYR2 mRNA containing the R4496C mutation, we selected siRYR2-U10 siRNA duplex sequence (siRYR2: siRNA targeted to RYR2 sequence), meaning that the mutation recognition site is in position 10 of the sense strand, as the most selective molecule able to reduce R4496C-RYR2 mRNA without inducing significant knockdown of the corresponding WT mRNA.
By ligation of annealed oligonucleotides, siRYR2-U10 was cloned into a commercially available artificial microRNA (miRNA) expression vector (BLOCK-iT Pol II miR RNAi Expression vector; Life Technologies) that allows the continuous and long-term transcription of the silencing molecule by a cytomegalovirus (CMV) promoter and the coexpression with the GFP (green fluorescent protein) reporter gene.
Subsequently, we transferred the CMV-miRYR2-U10-GFP cassette (miRYR2: miRNA targeted to RYR2 sequence) and a similar cassette encoding a scramble miRNA, unable to target any known vertebrate gene, into an AAV backbone plasmid (pAAV2.1_CMV-GFP-miRYR2-U10-TkpolyA and pAAV2.1_CMV-GFP-miRNA-Scramble-TkpolyA). We will refer to the 2 viral constructs as AAV9-miRYR2-U10 and AAV9-miRNA-Scramble (Online Figure II). The AAV production was performed at the Tigem Institute (http://www.tigem.it/core-facilities/vector-core).
Cell Culture and Transfection
HEK-293 cells were cultured in Dulbecco modified Eagle medium supplemented with 10% fetal bovine serum and 1% penicillin/streptomycin, 1% of l-glutamine, 1 mmol/L sodium pyruvate, and 1% nonessential amino acid solution at 37°C in 5% CO2. Lipofectamine 2000 (Life Technologies) was used for transient transfection according to the manufacturer’s protocol. siRNAs were diluted in OptiMEM together with pGFP_RYR2ex91-96/R4496C_3xFLAG and pRFP_RYR2ex91-96/WT_3xHA. siRNA final concentration was 50 nmol/L unless otherwise indicated. Analyses were performed after 48 hours.
RNA Extraction, Retrotranscription, and Real-Time Polymerase Chain Reaction
HEK-293 cells transiently transfected with equal amount of the reporter alleles and RNAi molecules were lysed, and total RNA was purified with RNeasy mini kit (Qiagen). RNeasy Fibrous Tissue mini kit (Qiagen) was used for RNA extraction from isolated hearts and other organs derived from infected and control mice. A total amount of 1 µg template RNA per reaction was used for a 20 µL retrotranscription, performed with iScript cDNA Synthesis kit (Bio-Rad) according to the manufacturer’s instruction.
Protein Extraction and Immunoblotting
Protein expression analysis was performed using the following antibodies: anti-FLAG (F3165; Sigma-Aldrich), anti-HA (H3663; Sigma-Aldrich), anti-RyR2 (MA3-916; Thermo Scientific). Pan Cadherin was used as reference protein using an anti-Pan Cadherin antibody (C1821; Sigma-Aldrich). As secondary antibody we used a horseradish peroxidase–conjugated antimouse antibody (W402B; Promega). Immunoblotting membranes were revealed using the Clarity Western ECL substrate (Bio-Rad) and detected using ChemiDoc MP Imaging System (Bio-Rad). Quantification of protein expression has been performed by densitometry normalization on the reference protein Pan Cadherin with Image Laboratory Software (Bio-Rad).
EKG Monitoring, Echocardiography, and Drug Testing
All animal studies were conducted in compliance of institutional guidelines and according to the Committee for animal well-being of the University of Pavia. Animal studies were approved by the Italian Ministry of Health (Approval no. 879/2015-PR). In this study, we used Het (heterozygous RyR2R4496C/+) knock-in mice (patent US7741529 B1) previously characterized in our laboratory.8–10 In vivo studies were performed delivering miRYR2-U10 and the miRNA-Scramble packaged in an AAV serotype 9 (AAV9) which was selected for its cardiac tropism.11 We injected intraperitoneally 1.3×1012 AAV9 viral particles in neonates at day 8 and 3.9×1012 AAV9 viral particles at day 30 (p30) after birth in Het. A dose–response curve was performed in Het mice injecting 1×1012 AAV9, 2×1012 AAV9, and 3×1012 AAV9 viral p30 after birth to identify the threshold for the antiarrhythmic efficacy. Animals were euthanized after 8 weeks, and cardiac myocytes were isolated for characterization. Fluorescence microscopy imaging of RyR2R4496C/+ myocytes from infected hearts showed a transduction rate of about 80±5% in mice infected at particles at day 8 (Online Figure IIIA and IIIB) and of about 66±12% in mice infected at p30; the cardioselectivity of transduction was confirmed by real-time polymerase chain reaction (PCR) comparing GFP expression in different organs (Online Figure IIIC).
EKG recording was performed using subcutaneous implantable EKG transmitters with the DSI PhysioTel Implantable Telemetry system (Data Sciences International). Baseline EKGs were recorded for 10 minutes in resting conditions in conscious animals followed by intraperitoneal delivery of 2 mg/kg epinephrine and 120 mg/kg caffeine injection and observation for additional 15 minutes. Quantification of arrhythmic episodes was performed offline by 2 blinded investigators. Echocardiography was performed at 8 weeks in mice sedated with isoflurane 1% with a Visual Sonics Vevo 2100 Ultrasound (Visual Sonics Inc) equipment using a 40-MHz linear array transducer. M-mode tracings in parasternal long-axis view were used to measure end-diastolic ventricular diameter and ejection fraction with Teicholz formula.
Action Potential Recordings in Isolated Ventricular Myocytes
Ventricular myocytes from Het, Het-SCR (heterozygous RyR2R4496C/+ infected with AAV9-miRNA-Scramble), and Het-U10 (heterozygous RyR2R4496C/+ infected with AAV9-miRYR2-U10) mice were isolated with the Langendorff method, using retrograde perfusion through the aorta with enzyme-containing solutions as previously described.4,7
Freshly isolated myocytes were kept in a plastic well surrounded by a heating ring to maintain the perfusion solution at 37°C. Action potentials were recorded with borosilicate glass pipettes with a resistance of 1 to 3 MΩ using the whole-cell patch-clamp technique in the current-clamp mode. Isoproterenol (30 nmol/L) was administered to elicit DADs and triggered activity (TA). Data were analyzed offline with pCLAMP 9.2 (Molecular devices).
Allele-Specific Real-Time PCR
Samples obtained from Het, Het-SCR, and Het-U10 mice were evaluated for RYR2 allelic expression using quantitative real-time PCR and allele-specific TaqMan probes (Applied Biosystems). RNA extraction and retrotranscription were performed as described above. All reactions were performed using an Applied Biosystems ViiA7 Real-Time PCR System (Applied Biosystems). A standard curve constructed by mixing the normal and mutant cDNA samples at known ratios and fitted with the equation log2 (WT: R4496C ratio)=αΔCT+β (α and β are fit parameters) was used to interpolate the ratio of allele expression of samples of animals infected with different doses of viral particles containing miRYR2-U10 and of control samples.
Immunofluorescence on Heart Sections
Cryosections of heart tissue were prepared for immunofluorescence performed with the following antibodies: anti-α-actinin (A7811; Sigma-Aldrich), anti-GFP 488–conjugated (600-141-215; Rockland) or Alexa Fluor 594–conjugated goat antimouse secondary antibodies (A31624; Life Technologies). Dako mounting medium was applied to all slides. Fluorescence images were obtained with a Leica TCS-SP5 II inverted confocal microscope.
Electron Microscopy and Quantitative Analysis of Electron Microscopy Images
Hearts isolated from WT, Het, Het-SCR, and Het-U10 mice were fixed by retrograde aortic perfusion. Ultrathin sections were cut in a Leica Ultracut R microtome (Leica Microsystem, Austria) using a Diatome diamond knife (Diatome Ltd, Biel, Switzerland) and double stained with uranyl acetate and lead citrate. All sections were examined with an FP 505 Morgagni Series 268D electron microscope (FEI Company, Brno, Czech Republic), equipped with Megaview III digital camera and Soft Imaging System (Munster, Germany).
The percentage of cardiac cells exhibiting severe structural alterations was quantified. Measurements of relevant parameters of jSR, TT, and mitochondria were performed applying the stereology point-counting technique12,13 in micrographs taken at 22.000 of magnification from cross-sections of papillary cardiac myocytes (Online Figure IV). In the assessment of the efficacy of gene therapy in reverting ultrastructural abnormalities in calcium release units (CRUs) and mitochondria, we used the following definitions: we defined as fully recovered those changes induced by therapy with AAV9-miRYR2-U10 that induce a statistically significant difference in the monitored parameters between Het and Het-U10, while also showing no statistically significant differences in the comparison between WT and Het-U10, thus confirming the return to normal values. We defined as partially recovered those changes induced by therapy with AAV9-miRYR2-U10 that document a statistically significant difference in the monitored parameters between Het and Het-U10; however, they do not show normalization because data present a statistically significant difference between WT and Het-U10.
Analyses were performed with SPSS Version 21.0 (IBM Corp, Armonk, NY) and GraphPad Prism (version 6.00 for Windows; La Jolla, CA). Sample sizes were calculated on the basis of published studies on the same animal model.8 Data are reported as mean±SD or percentage. Continuous variables were compared by ANOVA with Tukey–Kramer post hoc test or Mann–Whitney test, as appropriate. Categorical variables were compared with cross-tabulation applying the Fisher–Freeman–Halton exact test when >2 groups were compared. This was followed by pairwise comparisons with the Fisher exact test with Bonferroni correction. The probabilities of the formation of tetramers with different compositions in Figure 3 were modeled using binomial distribution. The probabilities of forming different tetramers are calculated from the mutant (dark grey)/WT (light grey) ratio.14 All tests were 2-tailed. P values <0.05 were considered significant unless Bonferroni correction was used.
Identification of the siRNA Sequence Suitable for Targeted Silencing of the R4496C-RYR2 Allele
We tested 13 siRNA duplexes in an in vitro assay for their ability to silence the mutant RYR2 mRNA over the corresponding WT allele. Real-time data showed that the siRYR2-U10 reduced R4496C-RYR2 mRNA by 80% without significantly affecting levels of WT mRNA (Figure 1A). Silencing of mutant RYR2 was confirmed by fluorescence microscopy taking advantage of the fluorescent tags in the reporter alleles (Figure 1B). We then evaluated, in the same in vitro system, the specificity and efficacy of a viral vector plasmid expressing an artificial miRNA, designed to replicate, when transcribed, sense and antisense strands of siRYR2-U10 cassette under a CMV promoter and linked to a GFP reporter gene. The miRYR2-U10 molecule induced about 50% reduction of the expression of R4496C allele both at the mRNA (Figure 1C; *P<0.05) and the protein level (Figure 1D and 1E; *P<0.05) without affecting WT expression.
Analysis of the Effect of miRYR2-U10 and miRNA-Scramble on levels of RYR2 Transcripts and RyR2 Protein After In Vivo Administration
We quantified the ratio of WT:R4496C-RYR2 mRNA transcripts from the heart of mice infected with either the AAV9-miRNA-Scramble (Het-SCR) or the AAV9-miRNA-U10 (Het-U10) particles at 8 or 30 days after birth. Transcript analysis showed that in both group of animals, the ratio between WT and mutant RYR2 transcripts changes from 1:1 seen in untreated heterozygous (Het) and Het-SCR mice and increases to 2:1 in both groups of mice treated at the dose of 3.9×1012 AAV9 Figure 2A and 2B), whereas total RYR2 mRNA showed no significant changes among groups (Figure 2C and 2D).
Interestingly, the dose–response curve of the AAV9-miRYR2-U10 (1×1012 Genome Copies [GC], 2×1012 GC, and 3×1012 GC) showed a linear increase of the ratio between WT and mutant transcript at progressively higher dose of the viral construct (Figure 2E).
We therefore modeled the combination of mutant and WT RyR2 monomers to form the tetrameric channel mimicking the setting of untreated Het and Het exposed to AAV9-miRYR2 U10 therapy at the dose of 3.9×1012 GC assuming random assembly and applying a binomial distribution to assess the impact of mutant mRNA knockdown on the channel. We showed that the change in the ratio between WT and Mutant RYR2 transcripts enhances the likelihood of completely functional tetramer formation from 6.25% to 19.75% and reduces the probability of totally dysfunctional tetramer formation from 6.25% to 1.23% (Figure 3A). When the combination of WT and mutant RyR2 monomers to form the tetrameric channel was calculated for each of the doses of AAV9-miRYR2-U10 therapy tested, we observed that each increase in the dose corresponded to a linear increase of the percentage of WT monomers and a decrease of the mutant monomers predicted to be incorporated in the RyR2 channels (Figure 3B).
Western blot analysis demonstrated that, when compared with Het mice, the amount of total RyR2 protein was reduced by ≈15% in the heart of Het-U10 mice (Online Figure V).
Functional Evaluation of Allele-Specific Silencing in RyR2R4496C/+ Mice-Derived Cardiac Myocytes
Experiments were performed 8 weeks after infection. Patch-clamp experiments were performed to compare the electrophysiological behavior of untreated RyR2R4496C/+ myocytes (Het) exposed to isoproterenol (30 nmol/L) to that of infected (GFP positive) and noninfected (GFP negative) myocytes isolated from the heart of RyR2R4496C/+ mice infected at day 8 (neonates) with either AAV9-miRYR2-U10 (Het-U10) or AAV9-miRNA-Scramble (Het-SCR). A striking reduction of DADs (Figure 4A and 4B) and TA (Figure 4A and 4C) was observed in GFP-positive cardiac myocytes derived from Het-U10 but not in those from Het-SCR mice or in cells derived from the heart of untreated Het mice.
Similar results were obtained in cells isolated from mice infected with miRYR2-U10 or miRNA-Scramble at day 30 (young adults; Figure 4A, 4D, and 4E). Results of statistical analysis are reported in the legend of Figure 4.
Evaluation of the Incidence of Ventricular Arrhythmias after Allele-Specific Silencing Administration in RyR2R4496C/+ Mice
In vivo experiments were performed 8 weeks after infection using an established protocol to elicit arrhythmias in RyR2R4496C/+ mice through the intraperitoneal delivery of epinephrine and caffeine (2 and 120 mg/kg, respectively)4,10 and to compare arrhythmic events occurring in Het versus Het-U10 and Het-SCR mice (Figure 5A). Data showed that 52% of Het mice (11/21) and 65% of Het-SCR (15/23) mice exhibited the typical bidirectional ventricular tachycardia,8 although treatment with miRYR2-U10 completely prevented the development of arrhythmias (0/25; Het-U10 versus Het-SCR, ***P<0.001; Het-U10 versus Het, ***P<0.001; Figure 5B).
Similar results were observed in mice infected at 30 days after birth and studied 8 weeks after infection: a remarkable reduction of the ventricular tachycardia occurred in Het-U10 (2/24, 8%) in comparison with the Het-SCR (13/21, 62%) and Het mice (10/20, 50%; Het-U10 versus Het-SCR, ***P<0.001; Het-U10 versus Het, *P<0.016 [P=0.0051]; Figure 5C).
We performed a dose–response curve to determine the threshold for antiarrhythmic efficacy of AAV9-miRYR2-U10 in mice infected at 30 days after birth: as shown in Figure 5D, the antiarrhythmic efficacy appears at 3×1012 GC and it moderately further small improve at the dose we used in the in vivo study: 3.9×1012 GC.
Echocardiography showed no significant differences in ejection fraction and left ventricular diastolic diameter among heterozygous mice and mice treated with AAV9-miRYR2-U10 and AAV9-miRNA-Scramble. Data are presented in the Online Table I and Online Movie I.
Comparative Electron Microscopy Analysis of Heart Samples From WT, RyR2R4496C/+, and RyR2R4496C/+ Mice Treated With Allele-Specific Silencing
We performed electron microscopy on cardiac tissue of WT and Het mice to investigate whether, in analogy with mice with recessive CPVT,15 mice with the dominant form of CPVT present ultrastructural abnormalities. We observed structural alterations of the CRUs that, as shown in Figure 6, represent the contact area between the membranes of the jSR and the TT. On the surface of the jSR, the Ryanodine Receptor channels can be visualized (Figure 6A, small arrows). In WT cardiac myocytes, the jSR cisternae are usually narrow and flat. CASQ2 is clearly visible as a chain-like electron-dense line that runs parallel to the SR membrane (Figure 6A, single black arrow). In Het cardiac myocytes, the shape of jSR is more variable and slightly wider and does not always contain the chain-like electron-dense polymer of CASQ2 (Figure 6B): we think that the apparent reduction of CASQ2 is most likely related to the increased volume of the abnormally enlarged jSR. In Het-SCR, cardiac myocytes CRUs appear as in Het cardiac myocytes (Figure 6C), although viral infection in Het-U10 rescues and restores the CRUs profile (Figure 6D).
When we compared the number of CRUs, the number of couplons, their length, and the jSR width, we observed that the number of CRUs, the number of couplons, and the average length of individual couplon are reduced in Het mice when compared with WT (Online Table II). These abnormalities were superimposable to those previously reported in the CASQ2-R33Q mice,7,15 suggesting that these ultrastructural abnormalities are a common feature of CPVT models.16
When we analyzed Het-U10, we observed that most ultrastructural parameters were restored to normal values after the infection, whereas Het-SCR presented similar abnormalities as the nontreated mice (Het; Online Tables II and III).
Interestingly, we observed that although cardiac samples from WT mice have contractile elements well aligned laterally with each other and mitochondria distributed longitudinally between myofibrils that exhibit an electron-dense matrix with parallel and tightly packed internal cristae (Figure 7A), ≈46% of myocytes from heart of Het mice presented damaged mitochondria with increased empty cytoplasmic spaces and alterations of the contractile elements (WT versus Het, *P<0.001; Figure 7B and 7C; Online Figure VI). Of relevance, hearts treated with miRYR2-U10 (Het-U10; Figure 7E), but not those treated with miRNA-Scramble (Het-SCR; Figure 7D; Het-U10 versus Het-SCR, ‡P<0.005), showed a reduction in the percentage of cardiac cells with severe mitochondrial abnormalities (from 46% in Het to 28% in Het-U10; Het-U10 versus Het, †P<0.05; Figure 7F).
The advancement of techniques to induce in vivo overexpression or silencing of selected genes has opened doors to the development of gene therapies to rescue the WT phenotype in carriers of inherited diseases. The applicability of gene therapy to inherited cardiac arrhythmias has not been explored mainly because of the concern that gene delivery could disrupt the highly regulated pattern of expression of proteins that regulate electric properties of the heart, thus aggravating, rather than improving, the arrhythmogenic substrate.17
We recently broke this conceptual barrier demonstrating that overexpression of cardiac CASQ2 is beneficial in 2 animal models of recessive CPVT that manifests high propensity to arrhythmias caused by reduced levels of CASQ2. Our data showed that systemic intravenous delivery of the AAV9 construct containing WT CASQ2 cDNA dramatically reduced arrhythmias in CPVT mice.7,18
Here, we undertook a bigger challenge and attempted to rescue the WT phenotype in knock-in mice carriers of the pathogenic gain-of-function mutation R4496C in the RYR2 gene associated with the CPVT phenotype. As we previously demonstrated, when exposed to β adrenergic activation, the frequency of calcium sparks in RyR2R4496C/+ mice is much higher than in WT animals.9 The same animal model is also prone to develop DADs and triggered arrhythmias that cannot be elicited in WT mice.4 We therefore speculated that the reduction of the amount of mutant protein induced by RNA interference targeted to the mutant allele could attenuate the development of arrhythmias.
We anticipated that even a partial reduction of mutant RyR2 protein would be sufficient to attenuate arrhythmogenesis. The rationale behind our expectation was based on data from Xie et al19 who demonstrated that TA develops only when a sufficient number of adjacent cells develop DADs. When this synchronization does not occur, afterdepolarizations are suppressed. In light of these data, we anticipated that even a partial reduction of mutant RyR2 would be able to exert an antiarrhythmic effect while limiting the risk of inducing an adverse reaction because of the loss of total RyR2 levels. This hypothesis is aligned with the results that we obtained when overexpressing the WT CASQ27 in our homozygous recessive CPVT mouse model in which the infection of 40% of cardiac myocytes was sufficient to completely prevent arrhythmias.
The key challenge for our experimental plan is represented by the need to identify a siRNA able to selectively silence the mutant cDNA fragment containing the R4496C mutation while preserving unaltered the expression of the WT cDNA. Out of 13 screened siRNA, the best performing one (named siRYR2-U10) was able to silence the mutant transcript while leaving almost unaffected the expression of WT RYR2 in HEK-293 cells. Analogously, its homologous artificial miRNA (miRYR2-U10) induced reduction of the expression of R4496C allele both at the mRNA and at the protein level without affecting WT expression in the same in vitro system.
The second challenge that we faced is represented by the unpredictable efficacy of the selected miRNA in vivo. We therefore treated the RyR2R4496C/+ mice with AAV9-miRYR2-U10 and 8 weeks later assessed whether this treatment would protect the animals from the arrhythmogenic effect of the injection of caffeine and epinephrine. We observed a dramatic reduction of the inducibility of bidirectional and polymorphic ventricular tachycardia in both group of animals, that is, mice infected in the perinatal period (at day 8) and those infected at reproductive age (p30) when compared with untreated heterozygous mice and mice treated with AAV9-miRNA-Scramble. Aligned with the suppression of arrhythmias, experiments performed in isolated cells demonstrated that myocytes derived from the heart of mice infected with AAV9-miRYR2-U10, but not those infected with AAV9-miRNA-Scramble, showed abolition of DADs in fluorescent infected cells, whereas nonfluorescent cells responded to isoproterenol administration developing DADs and TA. These data support the concept that silencing of the mutant allele results in a powerful antiarrhythmic effect irrespective of the age of treatment.
Applying the results of allele-specific real-time PCR, we simulated a binomial distribution of the WT and mutant RyR2 monomers in the formation of the tetrameric RyR2 structure and we derived that myocytes infected with the AAV9-miRYR2-U10, thanks to the selective inhibition of the transcription of the mutant allele, present an increase from 6.25% to 19.75% in the homomeric WT RyR2 channels (ie, RyR2 channels composed of 4 WT subunits) and a reduction from 6.25% to 1.23% of the homomeric mutant RyR2. Interestingly, when we performed a dose–response curve of the AAV9-miRYR2-U10, we observed that the antiarrhythmic efficacy parallels the predicted reduction of mutant versus WT RyR2 subunits (Figures 2E and 5D). The evidence that partial selective reduction of RYR2 mutant transcript attenuates arrhythmias in mice raises the novel concept that the variability in the severity of clinical manifestations observed among CPVT patients within the same family may be related to the ratio between WT and mutant RyR2. Whether the relative amount of the WT RyR2 and the mutant RyR2 may also be modulated by posttranscriptional modifications and influenced by environmental factors is an intriguing hypothesis that may justify the paroxysmal nature of clinical manifestations in CPVT patients. Our interpretation of results is supported by data from Li and Chen20 who showed that coexpression of the loss-of-function mutant E3987A and WT RyR2 protein in HEK-293 cells produces channels in which the reduction of the open probability of the channel is proportional to the number of the mutant monomers. It is therefore likely to assume that, in line with our observation, tetramers composed of a gain-of-function mutant and WT RyR2 protein would show a larger increase in the open probability as the number of mutant subunits in the tetrameric channel augments.
An open question that we anticipated in designing our experimental protocol was represented by the unpredictable consequences of reducing the amount of total RyR2 on cardiac contractility. It is therefore reassuring to see that assessment of cardiac function by echocardiography, performed before testing inducibility of arrhythmias in mice, did not send signals of changes in the end-diastolic diameters of the heart and of reduction of the ejection fraction.
On the basis of the findings obtained in our recessive CPVT model in which AAV9-mediated expression of WT CASQ2 not only rescued in vitro and in vivo phenotype but it was also able to recover the ultrastructural abnormalities of the jSR and the TT, we investigated whether targeted silencing of mutant RyR2 would normalize ultrastructural changes observed in Het mice and as expected both dilatation of jSR and fragmentation of TT were rescued by the treatment.
In this study, we also reported for the first time the presence of mitochondrial alterations in the heart of CPVT mice: this observation is relevant when placed in the context of the emerging evidence that mitochondrial alterations may occur in association with the presence of dysfunctional RyR channels in different tissues. The initial observation was made by 2 coauthors of this study (S.B. and F.P.) who reported back in 2009 the observation that a gain-of-function mutation in RyR1 (Y522S), causative for malignant hyperthermia, was associated with the development at young age of mitochondrial swelling and disruption that progressed with aging to cause local disruption of nearby SR and TT resulting in extreme sarcomere shortening and lack of mitochondria. The authors proposed that the Y522S mutation enhances SR Ca2+ leak, which, in turn, increases reactive oxygen species/reactive nitrogen species production and subsequent RYR1 S-nitrosylation and glutathionylation, further enhancing SR Ca2+ leak and release channel heat sensitivity. Both Ca2+ overload and increased redox stress promote activation of the mitochondrial permeability transition pore, the opening of which leads to mitochondrial depolarization, Ca2+ overload, and swelling.21 This hypothesis has not been tested, nor the presence of mitochondrial degeneration has been confirmed in humans; however, the observation has opened the field to further dwelling into the relationship between RyR2 dysfunction and mitochondrial disease.
Recently, Santulli et al22 showed that in the presence of a RyR2 mutation, functional and structural mitochondrial abnormalities develop in pancreatic β cells that express RyR2. The authors attribute the development of mitochondrial damage to the activation of the stress response in the endoplasmic reticulum in response to abnormal calcium leakage from the endoplasmic reticulum.
Finally, few months ago, Lavorato et al23 published an intriguing study that shows that the A4860G RYR2 loss-of-function mutation causes ultrastructural abnormalities in the heart of the mice characterized by the formation of nanotunnels in cardiac mitochondria. The authors suggested that the reduced release of calcium from the SR as a consequence of the primary mutation leads to random bursts of Ca2+ release because of Ca2+ overload in the SR leading to the formation of nanotubules: in the study, the mechanisms linking Ca2+ overload and development of nanotubules remain unknown.
In this context, our observation that approximately half of myocytes from heart of Het mice present altered mitochondria with degeneration of crista and increased empty cytoplasmic spaces demonstrates for the first time that gain-of-function RyR2 mutations are associated with mitochondrial damage in the heart. The evidence that reducing the mutant RYR2 transcripts through the delivery of our RNA-interference–based gene therapy rescues ultrastructural phenotype, supports the view of the existence of a causative link between RyR2 dysfunction and mitochondrial abnormalities. Although, the mechanisms linking RyR2 gating anomalies and degeneration of mitochondria in the heart remain to date undefined. Whether similar structural changes are present also in the heart of CPVT patients remains to be defined.
The ability of RNA interference to rescue the murine CPVT phenotype by preventing the development of TA in isolated cells, inhibiting the development of ventricular tachycardia in vivo and reverting ultrastructural abnormalities without causing adverse consequences, provides the proof of concept that mutant allele-specific silencing may evolve into a therapeutic strategy for CPVT patients.
We thank Patrizia Vaghi (Centro Grandi Strumenti of the University of Pavia) for technical assistance on Confocal Microscopy. We thank Vittorio Bellotti and Sofia Giorgetti for helpful suggestions and for the revision of the article.
Sources of Funding
We acknowledge the support of the following grants to S.G. Priori: European Research Council (ERC) 2015 to 2020 EU-rhythmy no. 669387, the Intramural Research Grant 2015 to 2017 from the Fondazione Salvatore Maugeri.
S.G. Priori, C. Napolitano, and M. Denegri own shares of Audentes Therapeutics. The other authors report no conflicts.
In May 2017, the average time from submission to first decision for all original research papers submitted to Circulation Research was 12.28 days.
This manuscript was sent to Masao Endoh, Consulting Editor, for review by expert referees, editorial decision, and final disposition.
The online-only Data Supplement is available with this article at http://circres.ahajournals.org/lookup/suppl/doi:10.1161/CIRCRESAHA.117.310882/-/DC1.
- Nonstandard Abbreviations and Acronyms
- adeno-associated virus serotype 9
- calsequestrin 2
- catecholaminergic polymorphic ventricular tachycardia
- calcium release unit
- delayed afterdepolarizations
- Genome Copies
- green fluorescent protein
- heterozygous RyR2R4496C/+
- heterozygous RyR2R4496C/+ infected with AAV9-miRNA-Scramble
- heterozygous RyR2R4496C/+ infected with AAV9-miRYR2-U10
- junctional sarcoplasmic reticulum
- polymerase chain reaction
- Ryanodine Receptor type 2 gene
- Ryanodine Receptor type 2 protein
- mice heterozygous for the R4496C mutation in the Ryanodine Receptor 2 gene
- sarcoplasmic reticulum
- triggered activity
- transverse tubules
- wild type
- Received February 22, 2017.
- Revision received June 7, 2017.
- Accepted June 14, 2017.
- © 2017 American Heart Association, Inc.
- Leenhardt A,
- Lucet V,
- Denjoy I,
- Grau F,
- Ngoc DD,
- Coumel P
- Priori SG,
- Napolitano C,
- Tiso N,
- Memmi M,
- Vignati G,
- Bloise R,
- Sorrentino V,
- Danieli GA
- Liu N,
- Colombi B,
- Memmi M,
- Zissimopoulos S,
- Rizzi N,
- Negri S,
- Imbriani M,
- Napolitano C,
- Lai FA,
- Priori SG
- Hayashi M,
- Denjoy I,
- Hayashi M,
- Extramiana F,
- Maltret A,
- Roux-Buisson N,
- Lupoglazoff JM,
- Klug D,
- Maury P,
- Messali A,
- Guicheney P,
- Leenhardt A
- Priori SG,
- Blomström-Lundqvist C,
- Mazzanti A
- Denegri M,
- Bongianino R,
- Lodola F,
- et al
- Cerrone M,
- Colombi B,
- Santoro M,
- di Barletta MR,
- Scelsi M,
- Villani L,
- Napolitano C,
- Priori SG
- Cerrone M,
- Noujaim SF,
- Tolkacheva EG,
- Talkachou A,
- O’Connell R,
- Berenfeld O,
- Anumonwo J,
- Pandit SV,
- Vikstrom K,
- Napolitano C,
- Priori SG,
- Jalife J
- Liu N,
- Ruan Y,
- Denegri M,
- Bachetti T,
- Li Y,
- Colombi B,
- Napolitano C,
- Coetzee WA,
- Priori SG
- Pacak CA,
- Mah CS,
- Thattaliyath BD,
- Conlon TJ,
- Lewis MA,
- Cloutier DE,
- Zolotukhin I,
- Tarantal AF,
- Byrne BJ
- Mobley BA,
- Eisenberg BR
- Chan WM,
- Siu WY,
- Lau A,
- Poon RY
- Rizzi N,
- Liu N,
- Napolitano C,
- et al
- Denegri M,
- Avelino-Cruz JE,
- Boncompagni S,
- De Simone SA,
- Auricchio A,
- Villani L,
- Volpe P,
- Protasi F,
- Napolitano C,
- Priori SG
- Li P,
- Chen SR
- Boncompagni S,
- Rossi AE,
- Micaroni M,
- Hamilton SL,
- Dirksen RT,
- Franzini-Armstrong C,
- Protasi F
- Lavorato M,
- Iyer VR,
- Dewight W,
- Cupo RR,
- Debattisti V,
- Gomez L,
- De la Fuente S,
- Zhao YT,
- Valdivia HH,
- Hajnóczky G,
- Franzini-Armstrong C
Novelty and Significance
What Is Known?
RyR2 (Ryanodine Receptor type 2 protein) release Ca2+ from the sarcoplasmic reticulum during excitation–contraction coupling.
Mutations in Ryanodine Receptor (RYR2) gene cause catecholaminergic polymorphic ventricular tachycardia (CPVT).
In RYR2 mutant cardiac myocytes, adrenergic stimulation elicits diastolic Ca2+ release causing triggered activity and life-threatening arrhythmias.
What New Information Does This Article Contribute?
Specific inhibition of the mutant allele of RYR2 by adeno-associated virus–mediated RNA interference dramatically reduces arrhythmias in RyR2R4496C/+ (mice heterozygous for the R4496C mutation in the Ryanodine Receptor 2 gene) knock-in mice exposed to adrenergic stimulation and suppresses triggered activity in infected cardiac myocytes.
Ultrastructural abnormalities of junctional sarcoplasmic reticulum and mitochondria are present in the heart of RyR2R4496C/+ CPVT mice.
Allele-specific targeted RNA interference substantially reduces ultrastructural abnormalities, suggesting that diastolic calcium release may promote mitochondria and sarcoplasmic reticulum alterations.
CPVT is an inherited arrhythmic disorder linked to mutations in the cardiac RYR2 gene. Current therapies provide an incomplete protection from sudden death; thus, the search for a therapeutic approach to revert molecular and functional abnormalities of the disease is a priority. In this study, we show that in a mouse model of CPVT specific inhibition of mutant RYR2 expression by viral vector-mediated RNA interference decreased the incidence of adrenergic induced ventricular arrhythmias in vitro and in vivo. Phenotypic rescue derives from the suppression of triggered action potentials that is observed in transduced cardiac myocytes of treated animals. Transcriptional analysis demonstrates that such antiarrhythmic efficacy parallels the predicted reduction of mutant versus wild-type subunits, suggesting that the ratio between wild-type and mutant RyR2 may be a determinant of the severity of clinical manifestations. Furthermore, we found that cardiac myocytes of CPVT mice present ultrastructural abnormalities of mitochondria and calcium release units that are mainly recovered by the mutant allele-specific silencing therapy. These data provide the proof of concept that selective inhibition of mutant RYR2 allele may evolve into a therapeutic strategy for CPVT patients.