Antisense-mediated exon skipping: a therapeutic strategy for titin-based dilated cardiomyopathy

Frameshift mutations in the TTN gene encoding titin are a major cause for inherited forms of dilated cardiomyopathy (DCM), a heart disease characterized by ventricular dilatation, systolic dysfunction, and progressive heart failure. To date, there are no specific treatment options for DCM patients but heart transplantation. Here, we show the beneficial potential of reframing titin transcripts by antisense oligonucleotide (AON)-mediated exon skipping in human and murine models of DCM carrying a previously identified autosomal-dominant frameshift mutation in titin exon 326. Correction of TTN reading frame in patient-specific cardiomyocytes derived from induced pluripotent stem cells rescued defective myofibril assembly and stability and normalized the sarcomeric protein expression. AON treatment in Ttn knock-in mice improved sarcomere formation and contractile performance in homozygous embryos and prevented the development of the DCM phenotype in heterozygous animals. These results demonstrate that disruption of the titin reading frame due to a truncating DCM mutation can be restored by exon skipping in both patient cardiomyocytes in vitro and mouse heart in vivo, indicating RNA-based strategies as a potential treatment option for DCM.

. AON sequences. Supplementary Table S2. List of primers for nested PCR.

Supplementary Figures
Supplementary Figure S1. Assessment of AON expression and stability after transfection in HL-1 cardiomyocytes.
Control and patient iPSCs were differentiated as embryoid bodies (EBs) by dissociation of iPSC colonies with dispase and maintaining them for 3 days in MEF-conditioned hESC medium in low attachment plates coated with 5% (w/vol) poly-HEMA (Sigma-Aldrich). For spontaneous differentiation, medium was then replaced with DMEM/F12 supplemented with 20% FBS, 2mM L-glutamine, 0.1mM nonessential amino acids, 0.1mM β-mercaptoethanol, 50U/ml penicillin, and 50mg/ml streptomycin. To improve cardiac differentiation, ascorbic acid (50mg/ml) was added to the medium and EBs were plated on day 7 on gelatin-coated dishes for better detection of beating foci. At day 20-30 of EB differentiation, contracting areas were manually micro-dissected and plated on fibronectin-coated plates. These myocytic explants were maintained in culture in the same medium described for EB differentiation, but supplemented with 2% FBS.
For single-cell analysis, myocytic explants were dissociated with several rounds of 15 min incubation at 37 °C with 480U/ml collagenase type II (Worthington) under gentle shaking.
Dissociated cells were plated on fibronectin-coated plates and maintained in EB differentiation medium supplemented with 2% FBS. For Isoprotenerol treatment, 10µM Isoprotenerol (ISO, Sigma-Aldrich) was added to medium. Medium was changed every second day and after one week cells were processed for molecular and immunocytochemical analyses.
Before performing experiments, all the cell lines were tested for mycoplasma contamination.
Microscopy was performed using imaging systems (DMI6000-AF6000), filter cubes and software from Leica microsystems. Images were assigned with pseudo-colors. Morphological analyses were performed by investigators blinded to the genotype of the cells.

Generation of the U7snRNA lentiviral vectors
For the generation of the lentiviral vectors we used as template a previously described lentiviral vector that encodes a modified U7 small nuclear RNA targeting exon 51 of the dystrophin gene (Goyenvalle et al, 2009). We exchanged the antisense sequences targeting exon 51 of dystrophin, with antisense sequences targeting two Exonic Splicing Enhancers in exon 326 of human TTN (U7snRNA-TTNAONs-IRES-GFP vector) or mouse Ttn (U7snRNA-mTtnAONs-IRES-GFP vector) by PCR mutagenesis, as previously described (Goyenvalle, 2012). Scrambled sequences corresponding to the human and mouse titin AONs were also exchanged by PCR mutagenesis (for the U7snRNA-ScrAONs-IRES-GFP and U7snRNA-mScrAONs-IRES-GFP vectors). Then the four modified U7snRNA constructs were separately cloned into the lentiviral transfer vector plasmid pRRLsin18.PPT.PGK.IRES.GFP (IRES-GFP vector). The primers used for PCR mutagenesis are listed in Table S3.

Mass Spectrometry sample preparation and data processing
Mouse heart tissue was homogenized in liquid nitrogen. Both iPSC-derived cardiomyocytes and mouse heart tissue were resuspended in a lysis buffer containing 6M guanidinium chloride (GCl) and subsequently reduced and alkylated prior to overnight digestion with LysC and Trypsin, as previously described (Kulak et al, 2014). Experiments were performed in technical triplicates using a nanoflow HPLC system (Proxeon/Thermo Fisher Scientific) coupled via a nanoelectrospray ion source (Thermo Fischer Scientific) to a Q Exactive mass spectrometer (Thermo Fisher Scientific), as previously described (Michalski et al, 2012;Kulak et al, 2014). Raw MS files were processed with MaxQuant (version. 1.4.3.19, http://www.maxquant.org) (Cox & Mann, 2008). Peak list files were searched by the Andromeda search engine, incorporated into the MaxQuant framework (Cox et al, 2011) against the UniProt database containing forward and reverse sequences. To match identifications across different replicates the 'match between runs' option in MaxQuant was enabled in a time window of two minutes. Processed data were imported into the MaxQB database (Schaab et al, 2012) and identified titin peptides were aligned to the sequence of isoform Q8WZ42 (34350 aminoacids) for human and A2ASS6 (35213 aminoacids) for mouse samples. 298 peptides were annotated as mapping in the exon 326 in human samples (corresponding to aminoacids 21598-27300) and 401 in mouse samples (corresponding to aminoacids 22449-28161). All peptides downstream of the exon 326 were annotated as Cterminal. Bioinformatic analyses were performed with the Perseus software (www.perseusframework.org). Before hierarchical clustering, peptide intensities were logarithmized and normalized by subtracting the median value. Data were filtered for at least 50% valid values.
To determine the in vivo skipping efficiency, the intensities of C-terminal peptides were summed and relative intensity was calculated as percent of total intensity of all mouse titin peptides (from 1 to 35213). The number of peptides used for this analysis was 2129 for total titin and 490 for the C-terminal portion. In vPMO-mScrAON-treated mice, the relative amount of the C-terminal portion resulted as 18.7% of total titin intensity while in vPMO-mTtnAON animals resulted as 20.15% (n=3, P=6.23E-3, two-sided Student´s T-test). Using this C-terminal increase as a proxy to quantitate exon skipping at the protein level, we thus estimated, based on the simple proportion [18.70:100=20.15:x], that the in vivo skipping efficiency was around 8%.

Live-cell imaging in iPSC-derived cardiomyocytes
Control myocytic explants were dissociated, plated on fibronectin-coated plates and infected with the CellLight® Actin-RFP baculovirus (Life Technologies, 10 MOI). The medium was changed after 24 hours and then every second day. Cells were monitored every day for two weeks. Microscopy was performed using imaging systems (DMI6000-AF6000), filter cubes and software from Leica microsystems. Images were assigned with a pseudo-color.

Immunohystological and electron microscopy analyses in Ttn knock-in embryos
Mouse embryos were fixed in 3.7% (vol/vol) formaldehyde and subjected to immunostaining by using the following primary antibodies: α-actinin (Sigma-Aldrich) and myosin heavy chain (Chemicon). Alexa-Fluor-488 and -594 conjugated secondary antibodies specific to the appropriate species were used (Life Technologies, 1:500). Nuclei were detected with DAPI.
Semithin sections of the heart region were stained with toluidine blue. Ultrathin sections (70 nm) were contrasted with uranyl acetate and lead citrate and examined with a Zeiss 910 electron microscope. Digital images were taken with a 1kx1k high-speed slow scan CCD camera (Proscan).

Echocardiographic analysis and vPMO treatment in adult Ttn knock-in mice
After anesthesia induction with 2% isoflurane in an anesthesia chamber, 3-4 months old male mice (B6.129P2-Ttn tm1Mdc ; strain: C57BL/6; genetic modification: Ttn c.43628insAT) were placed on a heated, tilt platform for echocardiography. Body temperature was maintained at 37°C, and anesthesia was continued with 0.5-1.5% isoflurane. An echocardiograph system In the adult mice study, the central hypothesis was that in the Angiotensin-treated animals, vPMO-mTtn-AON admnistration results in an improved LV-EF compared to vPMO-mScr injection. Assuming a standard deviation of 12% of the echocardiographically-determined LV-EF, in order to demonstrate 15% LV-EF increase with a type I error rate of 5% and with 90% power, we estimated that we would need a sample size of ~11 animals per group. The mice were matched by age, body weight and genotype and then randomly allocated to the different treatment groups. 14 mice received the vPMO-mTtn-AON and 11 the vPMO-mScr; one vPMO-mTtn-AON-treated animal died and was excluded from the study.
Housing and husbandry conditions were in accordance to the PHS Policy on Humane Care and Use of Laboratory Animals and were monitored and enforced by the Institutional Animal Care and Use Committee of Eberhard Karls University, Tübingen. According to the ARRIVE guidelines all relevant aspects of the study were reported. All animal investigations were done in compliance with the NIH and MRC recommendations.

Heart morphological and molecular analyses in adult Ttn knock-in mice
At the end of two weeks angiotensin II treatment mice were sacrified and heart muscle tissue was fixed in 4% paraformaldehyde overnight, dehydrated through a gradient of alcohol and embedded in paraffin. Sections (8 µm) were prepared and stained with Masson's trichrome.
For expression analysis, RNA was extracted, reverse transcribed and analyzed as described.
Successful penetration of vPMOs into the cardiomyocytes was demonstrated by fluorescence in situ hybridization (FISH). Slides were fixed in 4% PFA for 10 minutes and permeabilized in 0.2mg/ml Saponin (47036, Sigma-Alidrich) for 20 minutes at room temperature. Slides were postfixed in 4% PFA for 10 minutes at room temperature and treated in acetic anhydride/triethanolamine, after rinsed with PBS, followed by 3% H 2 O 2 /PBS for 1 hour at room temperature. Hybridization buffer (50% Formamide, 5x SSC, 0.5mg/ml yeast tRNA, 0.5mg/ml Herring sperm DNA, 1x Denhardt's solution) was heated at 80°C for 10 minutes and the slides were pre-hybridized at 45°C for 1 hour. 100 nM locked nucleic acid (LNA) Animals´ care was in accordance with institutional guidelines. Mice were randomly allocated to the different treatment groups and no animals were excluded from the analysis. Analyses were performed by investigators blinded to the genotype of the mice.

Characterization of control and DCM induced pluripotent stem cells
All the induced pluripotent stem cell (iPSC) lines showed human embryonic stem cell morphology, alkaline phosphatase activity (Fig S3B), loss of Sendai viral transgenes ( Fig   S3C), immunoreactivity for the embryonic stem-cell markers NANOG and TRA1-81 ( Fig   S3D) and normal karyotype ( Fig S3E); moreover they passed the so called PluriTest with a high "pluripotency score" and a low "novelty score" resembling normal human pluripotent stem cells (Fig S3F). All the iPSC lines showed reactivation of endogenous pluripotency genes (OCT4, SOX2, NANOG, REX1 and TDGF1, Fig S4A) Fig S4B). Pluripotency of each cell line was further confirmed by upregulation of genes specific of the three germ layers during in vitro spontaneous differentiation into embryoid bodies (EBs, Figure S4C).

Sarcomere remodelling in actin-RFP transfected iPSC-derived cardiomyocytes
During enzymatic dissociation into single cells, iPSC-derived cardiomyocytes, differently from adult cardiomyocytes, undergo a disarray of the myofibrils and round up. As soon as the cells are plated and attach to the culture dish, they start to reorganize the myofibrils, similarly to neonatal cardiomyocytes (Atherton et al, 1986). However, with the time in culture as single cells, a certain disorganization of the newly built sarcomeres can occur in some cardiomyocytes that undergo a de-differentiation process. In order to monitor myofibril remodeling in single iPSC-derived cardiomyocyte after dissociation, we overexpressed actin as a fusion with a red fluorescent protein (RFP) in freshly dissociated cells by Baculovirus technology (Fig S7A) and followed over time the distribution pattern of the fluorescent actin molecules in single cardiomyocytes. This live cell imaging analysis of many single control iPSC-derived cardiomyocytes up to 2 weeks after dissociation revealed that, similarly to neonatal cardiomyocytes, there is a considerable heterogeneity in the degree of myofibril organization at day 1-2, but most of the cells show organized RFP+ sarcomeres mainly around the nucleus (Fig S7B), suggesting this as starting site of myofibril reassembly. Within 4-5 days, the myofibril organization extends to the whole cell ( Fig S7B) and this associates with the onset of rhythmic contraction, supporting previously reported data on sarcomerogenesis in rodent embryonic cardiomyocytes (Turnacioglu et al, 1997;Sparrow & further we noticed that, around 7-10 days of culture as single cells, few of the "fully organized" cardiomycytes began to lose their sarcomeric organization, starting at the perinuclear region and eventually throughout the whole cytoplasm till the cell periphery ( Fig   S7B), resembling the pattern observed during sarcomere disassembly in dividing mammalian cardiomyocytes (Ahuja et al, 2004). We saw this phenomenon in hundreds of cells through many independent experiments. Based on these observations, we hypothesized that sarcomere remodeling in iPSC-derived cardiomyocytes after dissociation is a radially control process that mimics the one earlier observed in neonatal cardiomyocyte (Atherton et al, 1986). Thus, we assumed that cells with organized myofibrils in the perinuclear region are somehow in the process of sarcomere assembly, while cells with a striated pattern only in the cell periphery are in the process of sarcomere disassembly.  Figure S1. Assessment of AON expression and stability after transfection in HL-1 cardiomyocytes.

Supplementary Tables Supplementary
Representative images of HL-1 cardiomyocytes after transfection of a 5'-fluorescein-labeled 2OMePS AON (green) at the indicated time points. Fluorescence signal in the nucleus peaks approximately after 3h transfection. Scale bar, 10µm.

Supplementary Figure S2. Location of primers for the nested PCR.
This figure shows a schematic of the primers used for detection of successful removal of titin (Ttn) exon 326 from the transcript. An external long range PCR (using the primer pairs depicted in blue) followed by an internal nested PCR (using the red primer pairs) discriminate between unskipped (wild-type and mutated) and skipped Ttn mRNA. Due to the size of exon 326, exon 325-326 and exon 325-327 were assessed by separate primers pairs, ex325f-ex326r and ex325f-ex327r, respectively. Exact primer sequences are listed in Supplementary Table S2.