Adeno-associated virus-mediated CASQ2 delivery rescues phenotypic alterations in a patient-specific model of recessive catecholaminergic polymorphic ventricular tachycardia

Catecholaminergic Polymorphic Ventricular Tachycardia type 2 (CPVT2) is a highly lethal recessive arrhythmogenic disease caused by mutations in the calsequestrin-2 (CASQ2) gene. We have previously demonstrated that viral transfer of the wild-type (WT) CASQ2 gene prevents the development of CPVT2 in a genetically induced mouse model of the disease homozygous carrier of the R33Q mutation. In the present study, we investigated the efficacy of the virally mediated gene therapy in cardiomyocytes (CMs) differentiated from induced pluripotent stem cells (iPSCs) obtained from a patient carrying the homozygous CASQ2-G112+5X mutation. To this end, we infected cells with an Adeno-Associated Viral vector serotype 9 (AAV9) encoding the human CASQ2 gene (AAV9-hCASQ2). Administration of the human WT CASQ2 gene was capable and sufficient to restore the physiological expression of calsequestrin-2 protein and to rescue functional defects of the patient-specific iPSC-derived CMs. Indeed, after viral gene transfer, we observed a remarkable decrease in the percentage of delayed afterdepolarizations (DADs) developed by the diseased CMs upon adrenergic stimulation, the calcium transient amplitude was re-established and the density and duration of calcium sparks were normalized. We therefore demonstrate the efficacy of the AAV9-mediated gene replacement therapy for CPVT2 in a human cardiac-specific model system, supporting the view that the gene-therapy tested is curative in models with different human mutations of CPVT.

On this subject, our group has demonstrated that viralmediated administration of the wild-type (WT) CASQ2 gene was able to prevent the development of the disease in a CASQ2 knockout mouse model and in mice knocked-in for the homozygous CASQ2-R33Q mutation. [5][6][7] In this early study, however, we did not study whether the strategy was effective also in an experimental model based on human cells. Indeed, a critical step in devising a clinically usable gene therapy for CPVT is to demonstrate that the vector is effective in human cells and independently of the specific mutation. Thus, we decided to test whether AAV9-based CASQ2 delivery reverts the disease phenotype in the human setting, studying a CASQ2 homozygous G112+5X nonsense mutation, we had previously found in a family suffering from CPVT2 and extensively characterized in rat cardiomyocytes (CMs). 8 To this end, we employed induced Pluripotent Stem Cells (iPSCs), an increasingly used model system to study human inherited cardiovascular diseases, in particular primary cardiomyopathies and arrhythmogenic diseases, including Long QT and Brugada syndromes as well as CPVT. [9][10][11][12][13][14][15][16][17] We have previously generated an iPSC-based model for the autosomal dominant form of the disease, CPVT1 (which is linked to mutation of the cardiac ryanodine receptor, RyR2), demonstrating that inhibition of Calcium-Calmodulin-depen-dent Kinase II (CAMKII) restored the normal electrophysiological phenotype. 4 Here, we took advantage of the iPSC technology to generate a cardiac model of CPVT2 from a patient carrying the G112+5X nonsense mutation and tested the efficacy of AAV9-based CASQ2 delivery on cells phenotype. The results of the present study support the notion that AAV9-mediated gene therapy can be used on human cells to reinstate CASQ2 functionality.

Results
Generation of the human cardiac model of CPVT2. To assess the efficacy of biological therapy for CPVT2, we developed a human model of the disease by applying iPSC technology to a family in which we identified a deletion of 16 nucleotides (c.339-354, p.G112+5X) in the CASQ2 gene 8 (Figure 1a). This mutation generates a premature stop codon and, hence, a lack of protein expression. The clinical counterpart is a life-threatening phenotype of bidirectional ventricular tachycardia in response to catecholaminergic stress ( Figure 1b).

Figure 1
Pedigree and clinical phenotype of the CPVT family. (a) Pedigree of the recessive CPVT family investigated in this study. B05 is the proband, who is homozygous (HO) for the mutation and clinically affected, whereas heterozygous kin are not clinically affected; square = male; circle = female. (b) Bidirectional ventricular tachycardia recorded off therapy in the proband (paper speed, 25 mm/s) Figure 2 Generation of iPSCs from skin biopsy of a CPVT2 patient. (a) Phase-contrast images of iPSC colonies, from both clones (#1 and #20) reprogrammed from the proband B05 (HO) and subsequently used for the experiments. Scale bar: 400 μm (b) Representative images of iPSC colonies (one each iPSC line) showing positive staining for alkaline phosphatase activity. (c) Immunostaining of CPVT-iPSC lines (clones #1 and #20) showing expression of stemness-specific markers OCT-4 (top), SSEA-4 (middle) and TRA1-60 (bottom). (d) Semiquantitative real-time PCR showing upregulation of specific markers of pluripotency (Rex-1, DNMT3B and Oct4) in two CPVT-iPSC clones (#1 and #20). The data are presented relative to parental fibroblasts and were normalized to HGPRT expression; RUES2 embryonic stem cell line has been used as positive control reference. Values are mean ± S.E. Diagram shows results from one representative experiment (out of three). (e-f) Evaluation of the developmental competence of CPVT-iPSC lines by embryoid bodies aggregation (e) and teratoma formation assay (f). Panel E show the results of semiquantitative real-time PCR of EBs from two CPVT-iPSC lines (#1 and #20) at d30 of differentiation and indicate upregulation of expression of markers of the three germ layers in EBs obtained from both lines. The data are relative to undifferentiated iPSC and were normalized to HGPRT and 18S housekeeping genes expression. Values are mean ± S.E. Diagram shows results of one representative experiment (out of three). (f) Hematoxylin-eosin staining of teratomas formed by CPVT-iPSC lines injection into immunocompromised mice, showing presence of tissues that derive for all the three germ layers: neural rosettes and retinal epithelium are indicative of ectoderm formation, cartilage and adipose tissue are from mesoderm, and gut and respiratory epithelium indicate presence of endodermal differentiation. (g) RT-PCR against the SeV genome indicating loss-of-expression of the SeV exogenous genes in both the iPSC clones (#1 and #20) selected for the study. Parental fibroblasts and those infected with SeV genes for reprogramming have been respectively used as negative and positive controls. Detection of HGPRT gene expression has been used as loading control. (h) Representative image of the analysis of the karyotype of the iPSC lines generated from the proband, showing the reprogramming did not induce any major chromosomal abnormality iPSC lines were generated from skin fibroblasts of both the proband who carries the mutation in homozigosity (HO) and from the healthy father (HE-heterozigous) carrier of a single copy of the same mutation, using a Sendai virus (SeV)-based  Figure 1F), proving the generated models possess a complete developmental potential and therefore are fully pluripotent. Direct sequencing analysis also confirmed that the generated models were genetically matched to the donors and carried the CASQ2 mutation (Supplementary Figure 2). We also verified that the generated iPSC lines were free of the exogenous SeV genes used for the reprogramming ( With the aim to develop the cardiac model of CPVT2, we next generated CMs through direct differentiation of iPSCs from HO (Supplementary Figure 3A); as a control, CMs have been also obtained from HE-iPSC lines and from WT iPSC lines we previously generated from an unrelated healthy subject. 9 In a first instance, we verified that CMs generated from iPSCs expressed cardiac-specific proteins by checking the presence of α-sarcomeric actinin (Supplementary Figure 3B), as well as calsequestrin-2 (Supplementary Figure 3C)which is normally induced at late stages during differentiation of pluripotent stem cells, 10,18 indicating that this model could be used as a surrogate of human CMs.
Electrophysiological studies indicated CMs derived from CPVT2-iPSC lines recapitulate the phenotypic features typical of the disease. Although iPSCs are recognized as valuable tools for reproducing diseases' traits in vitro, 12,15 the demonstration that this concept applies also to our CPVT2 model is an essential requirement for the next investigations. To this purpose, we first determined the electrophysiological properties of generated CMs using the patch clamp technique. As expected, analysis of general (i.e., non CPVTspecific) action potential properties, including overshoot, amplitude, maximal diastolic potential (MDP), maximal upstroke velocity, maximal repolarization velocity and action potential duration (APD) at 30, 50 and 90% of repolarization, did not show any difference between HO-CMs and those differentiated from WT and HE lines (Figure 3a; Supplementary Figures 4A and B). Parameters were also comparable to those previously recorded in CMs from CPVT1 iPSC lines generated by us from a patient carrying the E2311D heterozygous mutation in the cardiac ryanodine receptor gene (RyR2). 9 Instead, when exposed to β-adrenergic stimulation by isoproterenol, HO-CMs displayed delayed afterdepolarizations (DADs) and, less frequently, triggered activity (TA) during the diastolic depolarization phase ( Figure 3b) that represent hallmarks of CPVT-CMs. [9][10][11][12]19 These phenomena were absent in both WT-and HE-CMs, used as controls (Supplementary Figures 4C,D).
We also determined whether the absence of CASQ2 was affecting induction toward specific cardiac cell populations during differentiation (i.e., pacemaker cells, ventricular or atrial cells). To this purpose, we clustered CMs obtained from iPSC differentiation into two distinct populations, the nodal-like (cells from the atrio-ventricular node) and working-like myocardial cells (i.e., cells from the atrial and ventricular chamber), based on their electrophysiological properties (maximal upstroke velocity, action potential amplitude and ratio between APD 90 and APD 50 ), as previously described for the CPVT1 model. 9 Results from these analyses revealed no significant differences between HO-, HE-and WT-iPSCs in differentiation potential and show a proportion of nodallike cells of~25-30% in all conditions (Supplementary Figure 5).
Altogether, these results indicate that iPCS-derived CMs replicated key features of recessive CPVT and therefore provide an in vitro model of the disease suitable to experiment the gene therapy approach that we have previously reported on a different human CPVT mutant. [3][4][5] AAV9-mediated CASQ2 delivery re-establish correct expression of calsequestrin-2 in CPVT2-CMs. We then aimed to determine whether AAV9-based gene delivery that effectively prevented arrhythmias in geneticallymodified mice (R33Q-CASQ2 knock-in and CASQ2 knockout) 5,6 is also effective in patient-derived myocytes. Our idea was to test a hypothesis that the observations obtained in the CPVT mouse also apply to a human-derived CPVT model caused by a different mutation.
To this aim, we took advantage of the AAV9 vector previously developed in our laboratory 5   Spontaneous intracellular Ca 2+ transients and Ca 2+ sparks were also assessed after hCASQ2 gene delivery. We measured these properties before and after adrenergic activation with 1 μM isoproterenol. A summary of the measurements for all parameters and conditions is provided in the Supplementary Figure 7.
These experiments revealed an amelioration of Ca 2+ transients amplitude (F/F 0ratio between peak and base-     20 Intracellular Ca 2+ sparks density and duration (FDHM: full duration at half maximum) also significantly recovered in HO-CMs after hCASQ2 administration (Figures 6b-d); other parameters, such as Ca 2+ sparks amplitude and size (FWHM : full width at half maximum) (see Supplementary  Figure 7) were not significantly affected in mutated HO-CMs, a consequence most probably linked to the developmental heterogeneity of different iPSC lines and to the immaturity of CMs derived from pluripotent cells. 14,18

Discussion
Here we demonstrate the therapeutic efficacy of CASQ2 gene therapy in human cardiac myocytes derived from a CPVT patient presenting with a severe clinical phenotype and a radical mutation (frameshift leading to premature truncation). We do also provide the proof of concept that iPSC-CM is a robust model that allows to explore the efficacy of gene therapy on different mutations within a target gene. This is an important concept that overcomes a major limitation of preclinical studies on gene therapy that are usually performed in the few available knockin mice models. Gene replacement is ideal in homozygous conditions in which the protein is absent or non-functioning, such as described here. Our study thus provides a further rational basis for clinical gene therapy studies on CPVT2 and opens to the systematic assessment of this therapeutic strategy to multiple CASQ2 mutations.
Clinical implications and clinical needs. CPVT patients currently receive lifelong therapy with β-blockers; however, this therapy has been demonstrated to be only partially effective in preventing the development fatal arrhythmias. As much as 25-30% of patients indeed experience recurrent cardiac arrests or die suddenly while on β-blocker therapy over an observation time of 5 years. 21,22 An additional subgroup of subjects representing 5-10% of the population does not tolerate the required dose of β-blockers and therefore is only partially protected. For all these patients either addition of flecainide to β-blockers or implantation of ICDs are the recommended treatment. 23 Although ICDs can prevent deaths, they cannot prevent the onset of arrhythmias, and their use is accompanied by adverse events, such as infections, lead fractures and inappropriate discharges. Overall, ICDs significantly affect the quality of life, especially in young children and teenagers. 24 On the basis of strong experimental evidence, gene therapy is emerging as an attractive strategy to treat CPVT2; 4 preclinical studies have shown that delivery of CASQ2 gene using AAV9 vectors is sufficient to restore the normal function of calsequestrin-2 in the heart, and to prevent the onset of the disease in CPVT2 murine models. 5,6 Whether this effect is replicated in humans and whether it is extendable to other CASQ2 mutations remains uncertain.
In order to answer these questions, we therefore developed a human model of CPVT2 derived from a patient carrying the G112+5X mutation in the CASQ2 gene through reprogramming to iPSCs, a technology that has served as a tool model for evaluating potential therapies for various cardiovascular diseases, including CPVT. 9,10,25,26 Similar to mouse models, AAV9-based administration of the human CASQ2 gene to CPVT2-CMs was sufficient to normalize the functional defect detected in those cells, demonstrating the therapeutic value of viral-based therapy of CPVT also in a human setting. These results, together with recent clinical studies in support of the safety of AAV for cardiac delivery, 27-29 may accelerate the route toward the translational application of iPSC-based experimental models for customized preventive tests, efforts which may eventually lead to the application of gene therapy strategies in the clinical setting for patients with CPVT or other cardiovascular recessive disorders.
Materials and Methods iPSC generation and differentiation into CMs. iPSCs have been generated from patients' skin fibroblasts using the CytotuneiPS-2.0 Sendai Reprogramming kit (Thermo Scientific). Reprogrammed clones were selected based on their morphology and subjected to complete validation, that included analysis expression of pluripotency markers, assessment of developmental potential in vitro and in vivo, and karyotype analysis, as described. 9 Loss-of-expression of exogenous Sendai viral genes has been verified by RT-PCR using the following primers that specifically amplify the SeV genome (SeV_Forward primer: GGATCAC TAGGTGATATCGAGC; SeV_Reverse primer: ACCAGACAAGAGTTTAAGAGATATG TATC; product size: 189 bp).
All the experiments have been conducted on two fully characterized iPSC lines from each individual.
Cardiac differentiation has been achieved using a chemically-defined serum free protocol, based on activation (CHIR99021) and inhibition (IWR1) of the Wnt pathway in RPMI-B27 medium, as previously described. 30,31 CMs were used for experiments in a stage of differentiation comprised between d25 and d30 (25-30 days after spontaneous contracting activity started), a differentiation stage that we showed to be sufficient to induce expression of calsequestrin-2 (as shown in the Supplementary Figure 3C). High differentiation efficiency of each iPSC line has been also verified by flow cytometry using α-sarcomeric actinin to specifically detect cardiac cells (Supplementary Figure 3B).
Flow cytometry analysis. Differentiated iPSC-derived CMs were harvested and dissociated in single cell as previously described. 9 Intracellular α-sarcomeric actinin staining was performed after fixation in paraformaldehyde 1% and cell permeabilization, using the appropriate saturating concentration of the unconjugated antihuman α-sarcomeric actinin antibody (mouse monoclonal, 1 : 400 from Abcam, Cambridge, UK). Detection was carried out using a goat anti-mouse Alexa-647conjugated antibody (1 : 500 from Molecular Probes, Thermo Scientific). Dead cells were excluded from the analysis using LIVE/DEAD fixable aqua stain kit (Molecular Probes, Thermo Scientific). Analysis of stained cells was performed on FACS LSRFortessa flow cytometer (BD Bioscience, San Jose, CA, USA). DIVA software (BD Pharmingen, San Diego, CA, USA) was used for the data acquisition and analysis.
Viral construct. The complete cDNA of the human cardiac CASQ2 was cloned into a AAV9 vector in frame with the red fluorescent protein (RFP) gene (AAV9-hCASQ2-RFP) through a highly efficient self-cleaving 2A peptide derived from porcine Teschovirus-1, as previously described. 6,32 T2A sequence: GGAAGCGGAGCTACTAACTTCACGCTGCTGAAGCAGGCTGGAGAC GTGGAGGAGAACCCTGGACCT.
As a negative control, an empty vector carrying the only RFP has been also generated.
Cells have been infected twice at MOI = 2 × 10 5 as previously described: 33 two rounds of infection (6 h and overnight) have been performed on differentiated CMs immediately after beating started.
Electrophysiological studies. iPSC-CMs were enzymatically dissociated and seeded at low confluence on Permanox slides (Lab-Tek II from Sigma-Aldrich) previously coated with fibronectin (5 μg/cm 2 , from Sigma-Aldrich). Action potentials (APs) were recorded using the patch clamp technique in the whole-cell configuration with a MultiClamp 700B (Axon Instruments, Sunnyvale, CA, USA). The experiments were performed at 37°C under continuous perfusion of extracellular solution containing (in mM) 140 NaCl, 4 KCl, 2 CaCl 2 , 1 MgCl 2 , 10 HEPES, and 5 glucose (pH adjusted to 7.40 with NaOH). Patch clamp pipettes, formed from borosilicate glass with a P-97 horizontal puller (Sutter Instruments, Novato, CA, USA), and had a resistance of 2-3 MΩ when filled with an intracellular solution containing (in mM) 20 KCl, 120 K-aspartate, 1 MgCl 2 , 4 Na 2 -ATP, 0.1 GTP, 10 glucose, and 10 HEPES (pH adjusted to 7, 20 with KOH). Some experiments were carried out with intracellular electrophysiology recordings. In this case, spontaneously beating clusters were impaled using sharp glass microelectrodes with resistances ⩾ 10 MΩ. Electrode capacitance was nulled and the recordings were made using the previously described MultiClamp 700B amplifier in gap-free mode.
Solutions containing isoproterenol (Iso, 1 μM) were prepared fresh before the experiments and applied using a gravitational flow system for 2-3 min prior to the data collection. All signals were acquired at 10 KHz, digitized (Digidata 1332A, Axon Instruments), and analysed with pCLAMP 9.2 software (Axon Instruments). We defined delayed DADs as low-amplitude depolarizations following completion of repolarization, and have an amplitude ⩾ 5% of the preceding AP. Triggered activity (TA) was defined as an AP developing from a DAD rather than from an external stimulus.
Calcium imaging measurements. Ca 2+ transients and Ca 2+ sparks were acquired using a TrimScope II upright two-photon microscope (LaVision Biotec, Germany) with a 60 × LUMPFL NA 1.1 (Olympus, Japan) water immersion objective, as previously described. 34 iPSC-CMs were seeded onto 35 mm dishes and the next day loaded with 5 μM Fluo-4 AM dye (Thermo Scientific) for 20 min. During acquisition, cells were kept under physiological condition with a perfusion pump in a buffer heated at 37°C. Ti : Sa infrared laser was tuned at 810 nm, and fluorescence light was separated from excitation light with a SP700 filter, then collected with a 525/50 emission filter and a non-descanned GaAsP detector (H7422-40, Hamamatsu Photonics, Japan) . After a first 512-pixel-wide XY image for each cell, we drew line ROIs and acquired 5000 XT line scans per ROI with a 0.375 um pixel size and 300 Hz time resolution, during which some calcium transients and some sparks occurred. Calcium transients were processed with Fiji/ ImageJ (http://fiji.sc/) to generate table intensity data over time, then R statistical software (R Foundation for Statistical Computing, Vienna, Austria) was used to isolate single transients and measure normalized transient amplitudes. Minimum five transients per condition were analyzed. Calcium sparks were analyzed with SparkMaster ImageJ plugin (https://sites.google.com/site/sparkmasterhome/). Statistical analysis of both transient and sparks measurements was conducted in Graphpad Prism 4.03 (Graphpad Software Inc, La Jolla, CA, USA), One-Way ANOVA and Tukey multiple post-test (*Po0.05, **Po0.01, ***Po0.001).
200-580 sparks were analyzed over a minimum of 25 cells per condition.
Statistical analysis. The data are represented as mean ± M.S.E (or mean ± S.D. where indicated). The significance of differences between the two groups was evaluated with unpaired Student's t-test. Po0.05 was considered statistically significant. (*) indicates Po0.05, (**) refers to Po0.01, while ns is for not significant.
Human subjects. Approval for the use of human samples in agreement with the protocols described here has been obtained by Fondazione Maugeri Review Board (ID. 921CEC 06/14/2013). Skin biopsies have been carried out using routine surgical techniques, without excision of excess tissue. Tissue specimens were collected in sterile saline solution and processed for isolation of fibroblast within the following 24 h.

Conflict of Interest
Dr Priori, Dr Napolitano and Dr Denegri own stocks of Audentes Therapeutics. Dr Lodola, Dr Morone, Dr Bongianino, Dr Nakahama, Dr Rutigliano, Dr Gosetti, Dr Rizzo, Dr Vollero, Dr Buonocore, Dr Condorelli and Dr Di Pasquale declare no conflict of interest.