Muscle-specific base editor for correcting mutations causing dilated cardiomyopathy

Abstract

Provided herein are compositions and methods for treating or preventing dilated cardiomyopathy (DCM) in a subject in need thereof, for example, through correcting one or more point mutations in the RBM20 locus using CRISPR-associated base editing. TIFF2025538153000011.tif56128

Description

[Technical Field]

[0001] CROSS-REFERENCE TO RELATED APPLICATIONS This application claims priority to U.S. Provisional Application No. 63 / 423,716, filed November 8, 2022, the disclosure of which is incorporated herein by reference in its entirety for all purposes. [Background technology]

[0002] background Dilated cardiomyopathy (DCM) is a heterogeneous disease with multiple causes and nonspecific phenotypes that ultimately lead to left ventricular dilation and systolic dysfunction. DCM is the second leading cause of heart failure, with an estimated prevalence of 1:250. Current therapies for DCM, including those widely used to manage heart failure, can at best only slow the progression of DCM. Therefore, developing strategies to maintain or even restore normal cardiac function would be a breakthrough for patients with limited treatment options.

[0003] In addition to environmental causes of DCM (e.g., viral infection, toxins, inflammation), approximately 30% of all cases are due to inherited mutations in several structural components of the heart. A curation of DCM-associated genes has identified 19 genes that show a strong association with DCM based on genetic and experimental evidence, and the RBM20 gene is one of them. The RBM20 gene encodes a cardiac splice factor that regulates alternative splicing of genes important for cardiomyocyte function. Approximately 2–6% of patients with highly penetrant and aggressive forms of familial DCM have RBM20 mutations. Summary of the Invention

[0004] overview The present disclosure provides methods and compositions for treating or preventing dilated cardiomyopathy (DCM) in a subject in need thereof, for example, by correcting one or more point mutations in the RBM20 locus in cells taken from the subject and then reintroducing the genetically modified cells back into the subject. As a non-limiting example, the methods and compositions involve CRISPR-associated base editing in combination with target-specific viral delivery using myotropic adeno-associated virus (AAVMYO) to achieve gene repair for the treatment and prevention of DCM and other inherited heart diseases.

[0005]

[0010] Accordingly, in one aspect, the present disclosure provides a method for correcting a point mutation at the RBM20 locus in a cell, the method comprising introducing into the cell (i) a single guide RNA (sgRNA) targeting a sequence containing the point mutation and (ii) a base editor (BE), wherein the sgRNA binds to and guides the base editor to the target sequence, such that the base editor corrects the point mutation at the RBM20 locus in the cell. In some embodiments, the method further comprises isolating the cell from a subject prior to introducing the sgRNA and the BE.

[0006] The methods disclosed herein can correct any point mutation in the RBM20 locus or other loci associated with DCM in a cell. In some embodiments, the point mutation in the RBM20 locus comprises a substitution at an amino acid position comprising 83, 455, 535, 633, 634, 635, 636, 637, 638, 703, 716, 783, 831, 888, 913, 914, 1031, 1081, 1182, 1206, or a combination thereof, wherein the substitution and position are determined with reference to SEQ ID NO: 10. In some embodiments, the point mutation in the RBM20 locus comprises a substitution at an amino acid position comprising 633, 634, 635, 636, 637, 638, or a combination thereof. In some embodiments, the point mutation in the RBM20 locus comprises a substitution at amino acid position 633, 634, or a combination thereof. In certain embodiments, the substitutions include P633L, R634Q, or a combination thereof.

[0007] Any sgRNA that targets a sequence containing a point mutation of interest can be used in the claimed methods. In some embodiments, the sgRNA comprises a sequence having about 80% or greater identity to any one of SEQ ID NOs: 1-9. In certain embodiments, the sgRNA comprises the sequence of any one of SEQ ID NOs: 1-9.

[0008] In some embodiments, the BE is an adenine base editor (ABE) or a cytidine base editor (CBE). In some embodiments, the BE comprises an RNA-guided catalytically impaired nuclease fused to a nucleobase deaminase enzyme. In some embodiments, the RNA-guided catalytically impaired nuclease is dead Cas9 (dCas9), dCas12, or Cas9 nickase (Cas9n), or a derivative thereof. In some embodiments, the RNA-guided catalytically impaired nuclease is an engineered Cas9n. In some embodiments, the engineered Cas9n is Cas9n-NRNH, Cas9n-NRTH, Cas9n-NRCH, CP-1041, or SpRY.

[0009] In some embodiments, the nucleobase deaminase enzyme is a single-stranded DNA (ssDNA)-specific deaminase enzyme. In some embodiments, the deaminase enzyme is an adenine deaminase or a cytidine deaminase. In some embodiments, the BE is BE1, BE2, BE3, BE4, ABE6.3, ABE7.8, ABE7.9, ABE7.10, BE4max, AncBE4max, ABEmax, ABE8e, ABE-SpRY, CBE-SpRY, ABE-CP-1041, or CBE-CP-1041.

[0010] In some embodiments, the sgRNA and the BE are introduced into the cell in one or more expression cassettes. In some embodiments, the BE is present in one expression cassette. In some embodiments, the BE is present in two expression cassettes, and the active BE is packaged in the cell via intein-mediated trans-splicing. In some embodiments, the expression cassette comprises a promoter. In some embodiments, the promoter is a CAG promoter. In some embodiments, the promoter is a human cardiac troponin T (hTNNT2) promoter.

[0011] In some embodiments, the sgRNA and BE are introduced into cells using a recombinant adeno-associated virus (rAAV) vector. In some embodiments, the rAAV vector is an AAVMYO vector. In some embodiments, the sgRNA and BE are introduced into cells as ribonucleoproteins (RNPs). In some embodiments, the RNPs are introduced into cells by electroporation.

[0012] In some embodiments, the cells are induced pluripotent stem cells (iPSCs) or iPSC-derived cardiomyocytes (CM-iPSCs).

[0013] In another aspect, the present disclosure provides a method for treating or preventing dilated cardiomyopathy (DCM) in a subject, the method comprising: (i) genetically modifying cells from the subject using a method according to any one of claims 1 to 27; and (ii) reintroducing the cells into the subject, wherein the reintroduction is effective to treat or prevent the subject from having DCM.

[0014] In some embodiments, the subject has a point mutation at the RBM20 locus. In some embodiments, the point mutation at the RBM20 locus comprises a substitution at an amino acid position comprising 83, 455, 535, 633, 634, 635, 636, 637, 638, 703, 716, 783, 831, 888, 913, 914, 1031, 1081, 1182, 1206, or a combination thereof, wherein the substitution and position are determined with reference to SEQ ID NO: 10. In some embodiments, the point mutation at the RBM20 locus comprises a substitution at an amino acid position comprising 633, 634, 635, 636, 637, 638, or a combination thereof. In some embodiments, the point mutation at the RBM20 locus comprises a substitution at amino acid position 633, 634, or a combination thereof. In some embodiments, the substitution comprises P633L, R634Q, or a combination thereof.

[0015] In some instances, the cells are reintroduced into the subject by systemic delivery. In other instances, the cells are reintroduced into the subject by local delivery. In some embodiments, the local delivery is intrafemoral or intrahepatic.

[0016] In some embodiments, the cells are cultured, expanded, selected, and / or induced to differentiate in vitro before being reintroduced into the subject.

[0017] In another aspect, the present disclosure provides an sgRNA that specifically targets the RBM20 gene, comprising a sequence having about 80% or more identity to any one of SEQ ID NOs: 1 to 9.

[0018] In another aspect, the present disclosure provides iPSCs comprising such an sgRNA and a base editor (BE) comprising an RNA-guided catalytically impaired nuclease fused to a nucleobase deaminase enzyme.

[0019] In another aspect, the present disclosure further provides cardiomyocytes derived from such iPSCs.

[0020] In another aspect, the present disclosure provides a pharmaceutical composition comprising a plurality of iPSCs disclosed herein or a plurality of cardiomyocytes disclosed herein. [Brief explanation of the drawings]

[0021] [Figure 1]Molecular and physiological characterization of P635L and R636Q mouse strains. a) Confocal images of isolated adult mouse cardiomyocytes. Scale bar: 20 μm. ACTN1 was used as a cardiomyocyte marker. b, c) Size (b) and quantity (c) of RBM20 granules in adult mouse cardiomyocytes. N = 21 (WT), 16 (P635L HET), 28 (P635L HOM), 16 (R636Q HET), and 39 (R636Q HOM) images, each representing one to four cells from three mice per genotype. Boxplots show median values, with boxes including the 25th and 75th percentiles and whiskers ranging from minimum to maximum. d) Number of DEGs in bulk RNA-seq of Rbm20 mutant mice compared to WT mice (Padjust < 0.05). N = 5 mice per genotype. e) GO analysis (biological function) of overlapping DEGs in both P635L and R636Q HOM mice using a strict cutoff of Padjust < 1e-10 to reduce the number of DEGs displayed in Figure 7. f) Number of differential splice events detected and categorized by rMATS compared to WT: alternative 5' or 3' splice sites (A5SS or A3SS), mutually exclusive exons (MXE), retained introns (RI), and skipped exons (SE). g) Mean ΔPSI (percent spliced-in) values ​​compared to WT for significantly overlapping differential splice events (Padjust < 0.01, ΔPSI > 0.1) in both HOM Rbm20 mutant mice. Multiple splice events per gene are indicated if consistent with the selection cutoff. Genes in red were validated by RT-PCR or qPCR. Gray boxes indicate that no splice events were detected by rMATS. h) RT-PCR of RBM20 target genes Ttn, Ryr2, and Ldb3 and the housekeeping gene Gapdh. i) Kaplan-Meier survival curves of mutant mice monitored for 120 days. P values ​​obtained by log-rank test between each mutant and WT are shown next to the curves. The percentage of survival is shown for HOM mice. j) Percentage of ejection fraction determined by narcosis echocardiography in mutant mice.N = 13 (WT), 5 (P635L HET), 6 (P635L HOM), 11 (R636Q HET), and 11 (R636Q HOM) mice. P values ​​in (b), (c), and (j) were obtained from one-way ANOVA with Tukey's multiple comparison test: ****P < 0.0001, ***P < 0.001, **P < 0.01; ns = not significant. All data were obtained from 16-week-old mice, except for (j), where data from 24-week-old mice are shown. Error bars indicate the standard error of the mean (SEM) in all panels. [Figure 2]Base editing of RBM20 in human iPSC-CMs and mouse. a-c) Transient expression of base editors and gRNAs in human iPSCs and iPSC-CMs. (a) Experimental overview, (b) P633L editing efficiency, and (c) R634Q editing efficiency. "CP" indicates the circularly permuted base editor CP-1041. The purple line indicates the average repair efficiency in iPSC-CMs. d) Generation of stable base editor expression in R634Q iPSCs. The repair efficiency, as determined by amplicon-seq, was 34.26 ± 2.36%. N = 3 independent differentiations. e) Expression of spliced ​​and unspliced ​​isoforms of TTN and IMMT in parental, R634Q, and edited R634Q iPSC-CMs differentiated for 15 and 32 days. Where present, significant changes compared to R634Q are indicated and analyzed by unpaired two-tailed t-test. *P < 0.05, **P < 0.01, ****P < 0.0001. f) Experimental overview of AAVMYO-mediated base editing in mice. g) Percentage of editing for P635L HOM mice injected with AAVMYO with different gRNA-base editor combinations or PBS as an empty control. For NRCH-gRNA1, AAV9 was also used as the vector. Significance was assessed using an unpaired two-tailed t-test. ***P < 0.001, **P < 0.01, *P < 0.05. The sequence shows the position of the on-target edit in blue and the positions of the two bystander edits in red. Numbers indicate the nucleotide positions within the targeting gRNA (gRNA2 was used as a reference) when the PAM sequence is located at positions 21-23. h) Allele frequencies of repaired DNA in the heart, diaphragm, and quadriceps (f.) muscle tissues, as well as in the liver, plotted for the 10 mice with the most editing events in (g). i) Percentage of Rbm20 mRNA editing in mice treated with AAVMYO-SpRY for 6 or 12 weeks. Editing was assessed by amplicon-seq of cDNA isolated from whole hearts.Most base editors contain the deaminase variant Abemax, except where indicated by "8e," which is a base editor containing the Abe8e version. The "repair" percentages in (b, c, g-i) are determined by NGS reads from amplicon-seq containing only wild-type sequences. The number of biological replicates, i.e., the number of independent differentiations in (b, c, e) or the number of mice in (g, i), is indicated in parentheses above the bars. Error bars indicate SEM in all panels. [Figure 3]Phenotypic characterization of mice after AAVMYO-ABE treatment. a) Allele frequency of restored Rbm20 mRNA in mice treated with AAVMYO-ABE, as determined by RNA-seq. N = 3 (R636Q), 4 (P635L), and 8 (WT) mice per condition. b, c) RBM20 staining in whole heart tissue sections from WT and Rbm20 mutant mice treated with PBS or AAVMYO-ABE. (b) Representative image, and (c) quantification of nuclear and cytoplasmic RBM20 localization. Scale bar: 20 μm. Arrows highlight restored nuclear RBM20 (magenta) and cytoplasmic RBM20 (white) in base-edited mice. Manual quantification of >200 nuclei in two mice per condition. d) Isoform expression of the RBM20 target genes Ttn, Ryr2, and Ldb3, and the housekeeping gene Gapdh, as determined by RT-PCR. N = 2–4 mice per condition. e, f) Vertical agarose gel (e) and quantification (f) of titin protein isoforms in WT and Rbm20 mutant mice treated with PBS or AAVMYO-ABE. Based on the gel image in Figure 9d. N = 3 mice per condition, except for WT and P635L / SpRY, where four mice were analyzed. g) RNA-seq data showing changes in ΔPSI values ​​compared to WT in saline- or base editor-injected P635L or R636Q HOM mice. See the Methods section (Bulk RNA Sequencing and Analysis) for definitions of the three categories. Rescue splice events are labeled in red, and all Ttn splice events are labeled in blue. N = number of splice events / category. R636Q is sequenced more deeply compared to P635L, which explains the difference in the number of DSGs detected. N = 3–5 mice per condition. h–j) Percentage of ejection fraction (h), LVID (i), and heart volume (j) determined by narcosis echocardiography in mutant mice treated with PBS or AAVMYO-ABE. N = 5 mice per condition. Same WT cohort as Figure 6l–n (at 16 weeks).P values ​​obtained from one-way ANOVA with Tukey's multiple comparison test: *** P < 0.001, ** P < 0.01, * P < 0.05, ns = not significant. All data were obtained 12 weeks after AAVMYO-ABE injection. Only homozygous P635L or R636Q mice were treated. Error bars indicate SEM in all panels. [Figure 4] Cell-type-specific profiling of cells after base editing by snRNA-seq. a) UMAP projection of all datasets integrated and annotated based on their gene expression profiles. b) Expression of known marker genes defining major cell types. c) UMAP projection of ventricular cardiomyocytes from WT, P635L HOM, and base-edited mice. d) Histogram showing the distribution of pairwise Euclidean distances between ventricular cardiomyocytes from P635L HOM and base-edited mice compared to WT using the two largest principal components (PCs). e) UMAP projection showing cardiomyocyte activity scores (see snRNA-seq analysis in the Methods section) using a subset of genes up- or down-regulated in P635L HOM compared to WT. A maximum of 15 significantly up- or down-regulated genes were used. f) Activity score thresholds based on (e) for the percentage of cells exceeding the threshold for genes up-regulated (upper panel) or down-regulated (lower panel) in P635L HOM cardiomyocytes compared to WT. g) Percentage of cells above the critical threshold for genes up- or down-regulated in P635L HOM cells compared to WT. VCM = ventricular cardiomyocytes, ACM = atrial cardiomyocytes, SMC = smooth muscle cells. Data were generated by snRNA-seq of isolated nuclei from two mice per condition. [Figure 5]WGS of mouse tissues before and after AAVMYO-ABE treatment. a) Mean number of variants (common) detected in tissue-specific and all tissues. b) Mean relative distribution of distinct nucleotide changes for tissue-specific and common SNVs. Tissue overlap represents variants that were common to all three tissues. c) Allele frequency of tissue-specific T > C / A > G variants. N = 33 (heart), 32 (liver), 146 (tail). d) Mean number of mismatches to the gRNA and PAM sequences in the ±30 base region surrounding the variant start site. Error bars indicate SEM in all panels. N = 3 mice for (a, b, d). [Figure 6]Further characterization of the P635L and R636Q mouse strains. a) Sanger sequencing traces of the first-generation homozygous mutant mice used for subsequent breeding and experiments. Red (P635L): C>T mutation; gray (R636Q): GT>AG mutation. Note that with subsequent base editing in R636Q, the CAG codon is converted to CGG, which is synonymous with the WT CGT codon. b, c) Fold change in expression of Rbm20 (b) and Nppa and Nppb (c) mice compared to WT. d) GO analysis (biological function) of DSGs overlapping both P635L and R636Q HOM mice with cutoffs of Padjust <0.01 and ΔPSI >0.1. e) Venn diagram of significant DEGs (upper, Padjust <0.05) and DSGs (lower, Padjust <0.01 and ΔPSI >0.1). f) P values ​​(cutoff Padjust <0.05) for unique or shared DEGs between P635L (red) and R636Q HOM (gray). Significant changes were analyzed by unpaired two-tailed t-test. ****P < 0.0001. Boxplots show median values, with boxes encompassing the 25th and 75th percentiles and whiskers ranging from minimum to maximum. Gene numbers are indicated above the plots. g, h) Fold changes in expression of spliced ​​and unspliced ​​Ttn isoforms and Camk2d isoform A (g), or fibrosis marker genes (h), compared to WT, as determined by qPCR. Significant changes were analyzed by unpaired two-tailed t-test. *P < 0.05, **P < 0.01. i, j) Representative cardiac tissue sections stained with Sirius Red (i) and quantification of Sirius Red-positive areas (j). Scale bar: 500 μm. k, l) Heart volume (k) and LVID (l) determined by narcosis echocardiography. Only significant differences are labeled. P values ​​obtained from one-way ANOVA with Tukey's multiple comparison test: **P < 0.01. Mice were 24 weeks old. m-o) Percentage of ejection fraction (m), LVID (n), and heart volume (o) determined by serial narcosis echocardiography in Rbm20 mutant mice.Asterisks indicate statistical significance compared to WT, as determined by two-way ANOVA with Tukey's multiple comparison test: ****P < 0.0001, ***P < 0.001, **P < 0.01, *P < 0.05. Data at 24 weeks are the same as in Figure 1j and Figure 6j, k. Data for R636Q HET and HOM at 12 and 16 weeks have been previously published by the present inventors. (p) Heart-to-body weight ratios at 24 and 52 weeks in P635L mutant mice. N = 5 for HOM mice and N = 6 for HET mice, unless otherwise indicated in parentheses above the bars. No significant changes were observed in (j), (l), and (p). All data were obtained from 16-week-old mice unless otherwise indicated. Error bars indicate SEM in all panels. Gene expression analysis was performed using RNA isolated from the left ventricle. [Figure 7] Heatmap of DEGs in P635L and R636Q mice derived from RNAseq. Gray bars indicate that the gene was not detected in the RNA-seq data. N = 5 mice per genotype. All DEGs overlapping in P635L and R636Q HOM mice are shown using a strict cutoff of Padjust <1e-10. [Figure 8]Analysis of base editing in iPSC-CMs and mice. a-c) Percentage of indels (a) and bystander edits at P633L (b) and R634Q (c) in iPSCs and iPSC-CMs. Indel formation is the sum of the frequency of insertions or deletions in a window 10 bp upstream to 10 bp downstream of the gRNA binding site. Indel formation is shown combined for both the mutation and the gRNA and is separated only by distinct base editors. Bystander edits are sequences containing an additional A>G conversion within the gRNA window. The observed location of bystander edits is shown in red, and the on-target site is shown in blue. d, e) Representative tissue sections (d) and RNA expression data (e) measuring the fluorescent transgene YFP delivered by AAVMYO and injected at different concentrations into WT mice. Scale bar: 100 μm. One mouse was injected per concentration. f) Percentage of repaired reads relative to viral copy number per diploid genome (left) or RNA expression (right) in muscle tissues of the heart, diaphragm, and quadriceps femoris (quadriceps f.), as well as in the liver, as determined by ddPCR. For the left panel, DNA was used as input with a primer for the CMV promoter. For the right panel, RNA reverse-transcribed to cDNA was used with a primer for the transcribed WPRE element common to all base editor constructs. Only the SpRY-gRNA2 combination was analyzed. Each data point represents one mouse. g) Editing efficiency of NRCH / gRNA2 driven by the hTNNT2 or CAG promoter. Concentrations shown are the combined amounts of N- and C-terminal base editor-containing AAV. N = 2 mice (CAG 1e12 and hTNNT2 2e12, no error bars) or N = 3 mice (hTNNT2 1e12). h) Editing efficacy of SpRY and 8e-NRCH in the heart and liver 6 and 12 weeks after injection in 4-week-old mice. N = 3 (8e-NRCH-gRNA2) or 7 (SpRY-gRNA2) mice.i) Normalized RNA expression of RBM20 derived from single-nucleus RNA-seq of human hearts. SMC = smooth muscle cells, Vent. CM = ventricular cardiomyocytes. Error bars indicate SEM in all panels. Only P635L HOM was treated. [Figure 9]Further phenotypic characterization of mice after AAVMYOABE treatment. a, b) Allele frequencies of repaired DNA (a) and bystander edits (b) in mice treated with AAVMYO-ABE. The sequences show the positions of on-target edits in blue and bystander edits in red. Numbers indicate the nucleotide positions within the targeting gRNA when the PAM sequence is at positions 21-23. c) Example allele frequencies determined by Crispresso2 for one mouse treated with AAVMYO harboring base editor 8e-NRCH and gRNA2. The on-target positions are shown in blue, and the positions of observed bystander edits are shown in red. d) Expression of spliced ​​and unspliced ​​Ttn isoforms and Camk2d isoform A in WT and mutant mice treated with PBS or AAVMYO-ABE. e) Vertical TTN agarose gel showing the G-N2BA, N2BA N2B, and T2 protein isoforms of TTN, as well as MHC as a loading control. f) Percentage of ejection fraction in WT or mutant mice 8 weeks after injection with PBS or AAVMYO-ABE. P values ​​obtained by one-way ANOVA with Tukey's multiple comparison test are shown for AAVMYO versus PBS. N = 3–5 mice per condition. g) Expression of the heart failure biomarkers Nppa and Nppb in WT and mutant mice treated with PBS or AAVMYO-ABE. h) Ejection fraction of AAVMYO-ABE with an hTNNT2-driven NRCH base editor before injection (week 4) and at weeks 8 and 12 after injection. Twice the dose of AAVMYO-ABE was used compared with CAG-driven ABE. In contrast to other experiments, we also performed echocardiography before injection and showed that PBS- and ABE-treated mice had similar physiological parameters. WT data were obtained separately and are the same as in Figure 3h. Statistical significance was assessed by unpaired, two-tailed t-test between PBS-treated and ABE-treated mice. **P < 0.01, ns = not significant. Only P635L or R636Q HOM mice were treated. Error bars indicate SEM in all panels.The number of mice per condition is indicated on the graph. All data except (e) and (g) were obtained 12 weeks after AAVMYO-ABE injection. DNA, RNA, and protein were isolated from the left ventricle. [Figure 10] Further snRNA-seq analysis. a-c) Number of active genes (a), total transcript counts (b), and percentage of mitochondrial gene counts per cell (c) for nuclei from each condition. Two independent snRNA-seq experiments were performed for each condition. d) Relative cell type distribution in WT, P635L HOM, and base-edited mice. e, f) UMAP (e) and quantification of the percentage of cells expressing the base-edited construct delivered by AAVMYO (f). N- and C-terminal base editor expression values ​​were summed. Values ​​are either 0 (not expressed, gray) or 1 (expressed, red). g) Histogram showing the distribution of pairwise Euclidean distances of the depicted cell types from P635L HOM and base-edited mice compared to WT when mapping using two principal components (PCs). h) Activity score thresholds for the indicated cell types versus the percentage of cells above threshold for genes that are upregulated (top panel) or downregulated (bottom panel) in P635L HOM compared to WT (see Methods for calculations). [Figure 11]AAV coverage and editing events detected by WGS. a) Normalized read coverage across autosomes for C- and N-terminal base editor sequences delivered by AAVMYO. b) Allele frequency of the P635L A>G ​​nucleotide transversion. c, d) All (left) and novel (right) SNVs (c) or indels (d) called by four variant callers (HC: HaplotypeCaller, MT: Mutect2, LF: Lofreq, SC: Scalpel). e) Average relative distribution of tissue-specific and common variants within coding or noncoding regions of the genome. Tissue overlap represents variants shared by all three tissues. f) Allele frequency of 16 candidate loci analyzed by amplicon-seq. Seven loci were determined by in silico prediction of off-target editing based on gRNA sequence similarity. Nine loci were obtained from WGS, yielding A>G / T>C SNVs in coverage. N = 5 mice treated with PBS and N = 5 mice treated with AAVMYO-SpRY. Error bars indicate SEM in all panels. [Figure 12]On - target and off - target editing of RNA. a) Expression of base editors in heart and liver tissues 12 weeks after AAVMYO - ABE or PBS treatment in P635L or P636Q HOM mice. RPKM = reads per kilobase million. b) On - target editing (blue arrows) and bystander editing (red arrows) within the gRNA region targeting Rbm20 on chromosome 19. For R636Q and P635L HOM mice, the percentage of each base aligned at the positions plotted on the x - axis is shown. In each row, the three replicates were summed before calculating the percentages. c) Number of heart (H), liver (L), and common variants after filtering as described in the Methods section. d) Average relative amount of distinct types of SNVs identified as heart - specific. A significance test was performed using a logistic regression model from the python package statsmodels, testing the difference in the A<G ratio and non - A<G ratio between pairs of ABE - treated samples and PBS - treated samples. Error bars indicate standard deviation. e) Number of mismatches to the gRNA and PAM sequences in the region ±30 bases around the variant start site. Variants from three replicates were summed. Variants on the X or Y chromosome were excluded. Error bars indicate SEM in all panels. Box - and - whisker plots in (a) and (c) show the median, which includes the 25th to 75th percentiles, along with whiskers extending to the rest of the distribution.

Mode for Carrying Out the Invention

[0022] Detailed Description Introduction Provided herein are compositions and methods for treating or preventing dilated cardiomyopathy (DCM) in a subject in need thereof, for example, by correcting one or more point mutations in the genomic DNA of cells harvested from the subject and then reintroducing the genetically modified cells back into the subject. In one particular aspect, the present disclosure describes the use of CRISPR-associated base editing to correct a point mutation in the RBM20 locus in cells. For the first time, the inventors employed a combination of a viral vector, AAVMYO, with cardiac tissue targeting specificity and a CRISPR base editor (BE) to repair a DCM patient's mutation in the cardiac splice factor RBM20, demonstrating the potential of the base editor combined with AAVMYO to achieve gene repair for the treatment of DCM and other genetic heart diseases. Furthermore, the inventors used an intein-mediated trans-splicing strategy to package the gRNA / BE complex into dual recombinant adeno-associated viruses (rAAVs), optimizing the viral dosage inside the cell and improving base editing efficiency.

[0023] definition Before the present invention is further described, it is to be understood that this invention is not intended to be strictly limited to particular embodiments described, as such may, of course, vary. It is also to be understood that the terminology used herein is for the purpose of describing particular embodiments only, and is not intended to be limiting, since the scope of the present invention will be limited only by the claims.

[0024] As used herein, the terms "a," "an," or "the" include not only aspects with one element, but also aspects with two or more elements. For example, the singular forms "a," "an," and "the" include plural referents unless the context clearly dictates otherwise. Thus, for example, reference to "a cell" includes a plurality of such cells, and so forth.

[0025] As used herein, the terms "about" and "approximately" generally refer to an acceptable degree of error for the quantity measured given the nature or precision of the measurement. Typically, exemplary degrees of error are within 20 percent (%), preferably within 10%, and more preferably within 5% of a given value or range of values. Any reference to "about X" specifically refers to at least X, 0.8X, 0.81X, 0.82X, 0.83X, 0.84X, 0.85X, 0.86X, 0.87X, 0.88X, 0.89X, 0.9X, 0.91X, 0.92X, 0.93X, 0.94X, 0.95X, 0.96X, 0.97X, 0.98X, 0.99X, 0.91X, 0.92X, 0.93X, 0.94X, 0.95X, 0.96X, 0.97X, 0.9 ... The values ​​are 98X, 0.99X, 1.01X, 1.02X, 1.03X, 1.04X, 1.05X, 1.06X, 1.07X, 1.08X, 1.09X, 1.1X, 1.11X, 1.12X, 1.13X, 1.14X, 1.15X, 1.16X, 1.17X, 1.18X, 1.19X, and 1.2X. Thus, "about X" is intended to teach and provide descriptive support for a claim limitation, for example, "0.98X."

[0026] The term "nucleic acid" or "polynucleotide" refers to deoxyribonucleic acid (DNA) or ribonucleic acid (RNA) and polymers thereof, in either single-stranded or double-stranded form. Unless otherwise specified, the term encompasses nucleic acids containing known analogs of natural nucleotides that have similar binding properties as the reference nucleic acid and are metabolized in a manner similar to naturally occurring nucleotides. Unless otherwise specified, a particular nucleic acid sequence also implicitly encompasses its conservatively modified variants (e.g., degenerate codon substitutions), alleles, orthologs, SNPs, and complementary sequences in addition to the sequence explicitly indicated. Specifically, degenerate codon substitutions can be achieved by generating sequences in which the third position of one or more (or all) selected codons is substituted with mixed-base and / or deoxyinosine residues (Batzer et al., Nucleic Acid Res. 19:5081 (1991); Ohtsuka et al., J. Biol. Chem. 260:2605-2608 (1985); and Rossolini et al., Mol. Cell. Probes 8:91-98 (1994)).

[0027] The term "gene" refers to a segment of DNA involved in producing a polypeptide chain. It can include regions preceding and following the coding region (leader and trailer) as well as intervening sequences (introns) between individual coding segments (exons).

[0028] A "promoter" is defined as a set of nucleic acid control sequences that direct the transcription of a nucleic acid. As used herein, a promoter includes necessary nucleic acid sequences near the transcription start site, such as a TATA element in the case of a polymerase II type promoter. A promoter also optionally includes distal enhancer or repressor elements, which may be located several thousand base pairs from the transcription start site.

[0029] An "expression cassette" is a recombinantly or synthetically produced nucleic acid construct that contains a set of specific nucleic acid elements that enable transcription of a specific polynucleotide sequence in a host cell. An expression cassette can be part of a plasmid, a viral genome, or a nucleic acid fragment. Typically, an expression cassette contains a polynucleotide to be transcribed operably linked to a promoter. The promoter can be a constitutive promoter, such as a CAG promoter, that is active in cells under all circumstances. The promoter can also be a regulatable or inducible promoter that is activated under specific circumstances, such as a chemically inducible promoter, a temperature-inducible promoter, a light-inducible promoter, etc. In some embodiments, the promoter can be activated only in specific organs / tissues; for example, the human cardiac troponin T (hTNNT2) promoter is specifically activated in the heart but not in other tissues / organs.

[0030] "Polypeptide," "peptide," and "protein" are used interchangeably herein to refer to polymers of amino acid residues. All three terms apply to amino acid polymers in which one or more amino acid residues are artificial chemical mimics of corresponding naturally occurring amino acids, as well as to naturally occurring amino acid polymers and non-naturally occurring amino acid polymers. As used herein, these terms encompass amino acid chains of any length, including full-length proteins in which amino acid residues are linked by covalent peptide bonds.

[0031] Amino acids may be referred to herein by either their commonly known three letter symbols or by the one-letter symbols recommended by the IUPAC-IUB Biochemical Nomenclature Commission. Similarly, nucleotides may be referred to by their commonly accepted single-letter codes.

[0032] As used herein, the terms "identical" or percent "identity," in the context of describing two or more polynucleotide or amino acid sequences, refer to two or more sequences or specific subsequences that are the same. Two sequences that are "substantially identical" have at least 60% identity, preferably 65%, 70%, 75%, 80%, 85%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, 99%, or 100% identity, when compared and aligned for maximum correspondence over a comparison window (a designated region adjusted using a sequence comparison algorithm, or, if no specific region is designated, by manual alignment and visual inspection). With respect to polynucleotide sequences, this definition also applies to the complement of a test sequence. With respect to amino acid sequences, in some cases, identity exists over a region of at least about 50 amino acids or nucleotides in length, or more preferably over a region of 75-100 amino acids or nucleotides in length.

[0033] For sequence comparison, typically, one sequence serves as reference sequence, and test sequence is compared to it.When using sequence comparison algorithm, test sequence and reference sequence are input into computer, and if necessary, partial sequence coordinate is designated, and program parameters of sequence algorithm are designated.Default program parameters can be used, or alternative parameters can be designated.Then, sequence comparison algorithm calculates the percent sequence identity of test sequence with reference sequence based on program parameters.For the sequence comparison of nucleic acid and protein, the BLAST 2.0 algorithm and default parameters described below are used.

[0034] As used herein, a "comparison window" includes reference to any segment of a number of contiguous positions selected from the group consisting of 20 to 600, usually about 50 to about 200, and more usually about 100 to about 150, within which two sequences may be compared after optimally aligning one sequence with a reference sequence of the same number of contiguous positions.

[0035] Algorithms for determining percent sequence identity and sequence similarity include the BLAST 2.0 algorithm described in Altschul et al., (1990) J. Mol. Biol. 215: 403-410. Software for performing BLAST analysis is publicly available at the National Center for Biotechnology Information website at ncbi.nlm.nih.gov. This algorithm involves first identifying high-scoring sequence pairs (HSPs) by identifying short words of length W in the query sequence that match or meet some positive threshold score T when aligned with words of the same length in database sequences. T is referred to as the neighborhood word score threshold (Altschul et al., supra). These initial neighborhood word hits serve as seeds for initiating searches to find longer HSPs containing them. The word hits are then extended in both directions along each sequence as far as the cumulative alignment score can be increased. Cumulative scores are calculated using, for nucleotide sequences, the parameters M (reward score for a pair of matching residues; always >0) and N (penalty score for mismatching residues; always <0). For amino acid sequences, a scoring matrix is ​​used to calculate the cumulative score. Extension of the word hits in each direction is halted when: the cumulative alignment score falls by an amount X from its maximum achieved value; the accumulation of one or more negative-scoring residue alignments causes the cumulative score to fall below 0; or the end of either sequence is reached. The BLAST algorithm parameters W, T, and X determine the sensitivity and speed of the alignment. The BLASTN program (for nucleotide sequences) uses as defaults a word size (W) of 28, an expectation (E) of 10, M=1, N=-2, and a comparison of both strands. For amino acid sequences, the BLASTP program uses as defaults a word size (W) of 3, an expectation (E) of 10, and the BLOSUM62 scoring matrix (see Henikoff & Henikoff, Proc. Natl. Acad. Sci. USA 89:10915 (1989)).

[0036] The BLAST algorithm also performs a statistical analysis of the similarity between two sequences (see, for example, Karlin & Altschul, Proc. Nat'l. Acad. Sci. USA 90:5873-5787 (1993)). One of the measures of similarity provided by the BLAST algorithm is the minimum sum probability (P(N)), which provides an indication of the probability that a match between two nucleotide or amino acid sequences would occur by chance. For example, if the minimum sum probability in the comparison of the test nucleic acid to the reference nucleic acid is less than about 0.2, more preferably less than about 0.01, and most preferably less than about 0.001, the nucleic acid is considered to be similar to the reference sequence.

[0037] The "CRISPR-Cas" system refers to a class of bacterial systems for defense against foreign nucleic acids. CRISPR-Cas systems are found in a wide range of bacteria and archaea. CRISPR-Cas systems are divided into two classes, including six types (I, II, III, IV, V, and VI) and many subtypes: Class 1 includes type I and III CRISPR systems, Class 2 includes type II, IV, V, and VI, and subtypes of Class 1 include, for example, subtypes IA to IF. See, for example, Fonfara et al., Nature 532, 7600 (2016); Zetsche et al., Cell 163, 759-771 (2015); Adli et al. (2018). Endogenous CRISPR-Cas systems contain a CRISPR locus containing clusters of repeats separated by unique spacer sequences that correspond to sequences from viruses and other mobile genetic elements, and Cas proteins that perform multiple functions, including spacer acquisition, RNA processing from the CRISPR locus, target recognition, and cleavage. In class 1 systems, these activities are carried out by multiple Cas proteins, with Cas3 providing endonuclease activity, whereas in class 2 systems, all of these are performed by a single Cas (Cas9).

[0038] The term "treat" or "treatment" refers to any one of the following: ameliorating one or more symptoms of a disease or condition (e.g., dilated cardiomyopathy); slowing or completely terminating the progression of a disease or condition (which may be manifested by longer periods between recurrent episodes, slowing or preventing worsening of symptoms, etc.); promoting the onset of a remission phase; delaying irreversible damage caused in the progressive chronic phase (both primary and secondary) of a disease or condition; delaying the onset of said progressive phase; or any combination thereof.

[0039] The term "prevention" or "preventing" refers to protecting a subject at risk of a disease or condition (e.g., dilated cardiomyopathy) from developing the disease or condition, or reducing the risk that a subject may develop the disease or condition.

[0040] As used herein, the terms "subject," "individual," or "patient" interchangeably refer to warm-blooded animals such as mammals. In certain embodiments, this term refers to humans. A subject may have, be suspected of having, or be predisposed to a lysosomal storage disorder as described herein. This term also includes livestock, pet animals, or animals kept for research, including horses, cows, sheep, poultry, pigs, cats, dogs, zoo animals, goats, primates (e.g., chimpanzees), and rodents. A "subject in need thereof" refers to a subject who has one or more symptoms of dilated cardiomyopathy (DCM), a subject who has been diagnosed with DCM, a subject who is suspected of having or being predisposed to DCM, and / or a subject who has been identified with one or more point mutations in DCM-related genes.

[0041] The term "administering" as used herein includes oral administration to a subject, topical contact, administration as a suppository, intravenous, intraperitoneal, intramuscular, intralesional, intratumoral, intradermal, intralymphatic, intrathecal, intranasal, or subcutaneous administration.Administration can be by any route, including parenteral and transmucosal (e.g., buccal, sublingual, palate, gingival, nasal, vaginal, rectal, or transdermal).Parenteral administration can include, for example, intravenous, intramuscular, intraarteriolar, intradermal, subcutaneous, intraperitoneal, intraventricular, and intracranial administration.Other delivery modes include, but are not limited to, the use of liposome formulations, intravenous infusion, transdermal patches, etc.

[0042] The term "pharmaceutically acceptable carrier" refers to a substance that aids in the administration of an active agent to a cell, organism, or subject. A "pharmaceutically acceptable carrier" refers to a carrier or excipient that can be included in the compositions of the present disclosure and that does not cause significant adverse toxic effects to the subject. Non-limiting examples of pharmaceutically acceptable carriers include water, sodium chloride, normal saline, lactated Ringer's solution, normal sucrose, normal glucose, binders, fillers, disintegrants, lubricants, coatings, sweeteners, flavorings and colorings, liposomes, dispersion media, microcapsules, cationic lipid carriers, isotonicity and absorption delaying agents, and the like. Carriers can also be substances that provide stability, sterility, and isotonicity to the formulation (e.g., antimicrobial preservatives, antioxidants, chelating agents, and buffers), substances that prevent the action of microorganisms (e.g., antibacterial and antifungal agents such as parabens, chlorobutanol, sorbic acid, etc.), or substances that provide edible flavors to the formulation. In some cases, a carrier is a substance that facilitates delivery of the modified cells to a target cell or tissue. Those of skill in the art will recognize that other pharmaceutical carriers are useful in the present disclosure.

[0043] Dilated cardiomyopathy The present disclosure provides methods and compositions for treating or preventing dilated cardiomyopathy (DCM), a heterogeneous disease with multiple causes and non-specific phenotypes that ultimately lead to left ventricular dilation and systolic dysfunction. Approximately 30% of DCM is due to inherited mutations in several structural components of the heart, and 19 genes have been classified for their strong association with DCM: RBM20 (encoding RNA-binding motif protein 20), BAG3 (BLC2-associated athanogene 3), DES (desmin), FLNC (filamin C), LMNA (lamin A / C), MYH7 (myosin heavy chain 7), PLN (phospholamban), SCN5A (sodium channel α subunit), TNNC1 (troponin C), TNNT2 (troponin T), TTN (titin), DSP (desmoplakin), ACTC1 (cardiac α-actin), ACTN2 (α-actinin-2), JPH2 (junctophilin 2), NEXN (nexilin), TNNI3 (troponin I), TPM1 (α-tropomyosin), and VCL (vinculin). As disclosed herein, the methods and compositions can be used to correct point mutations in any of these genes. In some embodiments, the methods disclosed herein are used to correct one or more point mutations in a gene selected from the group consisting of RBM20, BAG3, DES, FLNC, LMNA, MYH7, PLN, SCN5A, TNNC1, TNNT2, TTN, DSP, ACTC1, ACTN2, JPH2, NEXN, TNNI3, TPM1, and VCL.

[0044] RBM20 In certain embodiments, the methods and compositions disclosed herein can be used to correct point mutations in the RBM20 locus. The RBM20 gene encodes RNA-binding motif protein 20, which regulates RNA splicing of genes important for cardiomyocyte function. The human RBM20 gene is located on the long arm of chromosome 10 and contains 14 exons. It encodes a 1,227-amino acid protein containing two zinc finger domains, a glutamic acid-rich region, a leucine-rich region, an RNA recognition motif (RRM)-type RNA-binding domain, and an arginine / serine-rich region (RS domain). The human RBM20 protein comprises the sequence of SEQ ID NO: 10.

[0045] Pathogenic variants in RBM20 account for approximately 2-6% of familial DCM cases, which are associated with significantly earlier disease onset and more severe clinical manifestations. Three protein regions have been identified with high confidence as harboring pathogenic variants: c.1601-1640 (exon 7, encoding the RRM domain), c.1881-1920 (exon 9, encoding the highly conserved RS domain), and c.2721-2760 (exon 11). Table 1 lists the reported variants along with their corresponding domains. In particular, BM20 mutations concentrated in a narrow six-amino acid stretch, proline-arginine-serine-arginine-serine-proline (PRSRSP), within the RS domain lead to abnormal cytoplasmic granule formation, amplifying the DCM-specific disease phenotype.

[0046] As disclosed herein, methods and compositions for treating or preventing DCM can be used to correct any point mutation in the RBM20 gene. In some embodiments, the RBM20 point mutation comprises a substitution at an amino acid position comprising 83, 455, 535, 633, 634, 635, 636, 637, 638, 703, 716, 783, 831, 888, 913, 914, 1031, 1081, 1182, 1206, or a combination thereof, wherein the substitution and position are determined with reference to SEQ ID NO: 10. In some embodiments, the point mutation in the RBM20 locus comprises a substitution at an amino acid position within a 6 amino acid stretch (PRSRSP) comprising 633, 634, 635, 636, 637, 638, or a combination thereof. In certain embodiments, the point mutation in the RBM20 locus comprises a substitution at amino acid position 633, 634, or a combination thereof.

[0047] In some embodiments, the RBM20 point mutation that is corrected is selected from the group consisting of L83I, S455L, V535I, P633L, R634Q, R634W, S635A, R636C, R636H, R636S, S637G, P638L, R703S, R716Q, R783G, L831I, D888N, E913K, V914A, G1031X, P1081R, R1182H, E1206K, or a combination thereof. In some embodiments, the RBM20 point mutations of interest are concentrated in the 6 amino acid section (PRSRSP) within the RS domain, and include P633L, R634Q, R634W, S635A, R636C, R636H, R636S, S637G, and P638L. In certain embodiments, the RBM20 point mutations include P633L, R634Q, or a combination thereof.

[0048] Table 1. RBM20 variants and corresponding exons and protein domains TIFF2025538153000002.tif147128 * Nonsense mutation; all others are missense mutations.

[0049] Base Editor (BE) As disclosed herein, the method and composition for correcting point mutation involves CRISPR-related base editing.Base editing is a CRISPR-based genome editing technology that can introduce or correct point mutation in DNA without generating double-strand breaks (DSBs).Base editor (BE) can be cytidine base editor (CBE) that allows C>T conversion or adenine base editor (ABE) that allows A>G conversion.

[0050] The BE used to correct the point mutation can be any type of BE known in the art. In some cases, the BE is an adenine base editor (ABE). In other cases, the BE is a cytidine base editor (CBE). In some embodiments, the BE is selected from the group consisting of BE1, BE2, BE3, BE4, BE4max, AncBE4max, ABE6.3, ABE7.8, ABE7.9, ABE7.10, ABEmax, ABE8e, ABE-SpRY, CBE-SpRY, ABE-CP-1041, and CBE-CP-1041.

[0051] Cytosine Base Editor (CBE) All of the early-generation BEs, including BE1, BE2, BE3, and BE4, are CBEs that convert G:C bp to T:A bp. BE1 is the first-generation BE, containing a catalytic dCas9 derived from Streptococcus pyogenes (Sp) fused to a rat deaminase (rAPOBEC1). dCas9 contains the D10A and H840A amino acid substitutions in Cas9, which disable its nuclease activity and avoid DSB generation without interfering with its DNA-binding ability. BE2 is based on BE1 and further fused to a uracil glycosylase inhibitor (UGI) to prevent the excision of newly formed U. BE3 is the third-generation BE, replacing the dCas9 in BE2 with a Cas9 nickase (Cas9n containing the D10A amino acid substitution) that cleaves unedited G-containing DNA strands without generating DSBs. BE4 differs from BE3 as it possesses a second UGI, resulting in higher editing efficiency and improved product purity.

[0052] CBEs can be further optimized by modifying codon usage and nuclear localization sequences to enhance base editing in mammalian cells (e.g., BE4max and AncBE4max). For example, BE4 can be improved by adding bipartite NLSs to both the N- and C-termini and optimizing codons to generate BE4max. Replacing rAPOBEC1 with an optimized ancestral rAPOBEC1 homolog (Anc689, which contains 36 amino acid substitutions compared to rAPOBEC1) resulted in the generation of AncBE4max. Both BE4max and AncBE4max exhibit higher editing efficiency compared to BE4.

[0053] Adenine base editor (ABE) As disclosed herein, the ABE enables the conversion of an A:T bp to a G:C bp at a target locus (e.g., RBM20). In some embodiments, the ABE comprises a catalytically impaired nuclease and an adenine deaminase. In some cases, the adenosine deaminase is a dimeric adenine deaminase. In some embodiments, the dimeric adenine deaminase is a heterodimer comprising a wild-type tRNA adenosine deaminase (TadA) and a genetically modified TadA*. As disclosed herein, the genetically modified TadA* comprises one or more (e.g., 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, or more) substitutions in the wild-type TadA amino acid sequence. As disclosed herein, the genetically modified TadA* can deaminate adenine in a DNA sequence. In some embodiments, the dimeric adenine deaminase is a homodimer containing two genetically modified TadA*. In some embodiments, the adenosine deaminase is a monomeric adenine deaminase containing one genetically modified TadA*. ABE6.3, ABE7.8, ABE7.9, and ABE7.10, and ABE8e contain different mutations in TadA*. * Examples of ABEs having a domain: In some embodiments, the ABE is selected from the group consisting of ABE6.3, ABE7.8, ABE7.9, ABE7.10, ABEmax, ABE8e, ABE-SpRY, and ABE-CP-1041.

[0054] ABE can be optimized by using orthologous or engineered Cas9n to expand the range of adenine base editing targets. For example, Cas9n variants can be introduced into ABE to generate A>G conversions at genomic sites containing non-NGG PAMs. In some cases, SpCas9n is replaced with SaKKHn or SpCas9n-VQR in ABE7.10 to generate SaKKH-ABE and VQR-ABE, respectively. In some cases, xCas9 is introduced into ABE7.10 to generate xCas9-ABE. The ABEmax version includes Cas9 variants that recognize NG PAMs (xCas9 in xABEmax or SpCas9n-NG in NG-ABEmax) or NR PAMs (SpCas9n-NRCH, SpCas9n-NRTH, and SpCas9n-NRRH). ABEmax can be further improved by replacing SpCas9n with SaCas9n or the engineered SaKKHn, SpCas9n-VRER, and SpCas9n-VRQR, enabling targeting of loci containing non-NGG PAMs. SpCas9n-VRER and SpCas9n-VRQR induce A-to-G conversions at many target sites containing PAMs other than NGG. Sa-ABEmax and SaKKH-ABEmax exhibit a wide editing window (positions 4-14 of the protospacer). CP-ABEmax can target bases located outside the canonical editing window.

[0055] In addition, ABE interacts with TadA to enhance adenine base editing in cells. * ABE8e can be optimized by modifying TadA. *The deaminase domain contains eight additional mutations that confer higher processing activity. When combined with SpCas9n or different Cas9 variants (e.g., SaCas9n, SaKKHn, SpCas9n-NG, and LbCas12a), ABE8e further improves editing efficiency compared to the corresponding ABEmax-based enzyme. Furthermore, removal of wild-type TadA did not affect the editing activity of ABE8e, suggesting that the optimized TadA * It has been shown that it can function efficiently as a monomer.

[0056] As disclosed herein, base editors (BEs) typically include two components: a nucleobase deaminase enzyme and an RNA-guided catalytically impaired nuclease. These two components can be linked covalently (e.g., as a fusion protein) or non-covalently (e.g., via an RNA aptamer). Guided by a single guide RNA (sgRNA), the catalytically impaired nuclease recognizes a specific sequence called a protospacer adjacent motif (PAM) and unwinds the DNA sequence upstream of the PAM (the "protospacer"). The deaminase enzyme then converts the base located within a specific DNA section of the protospacer "editing window."

[0057] Nucleobase deaminase enzyme Nucleobase deaminase enzymes can chemically modify specific DNA bases. They can convert one nucleotide into another by catalyzing the removal of amino groups from the base. For example, they can convert cytosine to uracil, ultimately resulting in a base pair change from CG to TA. In some embodiments, the nucleobase deaminase enzyme is a single-stranded DNA (ssDNA)-specific nucleobase deaminase enzyme. In some cases, the deaminase enzyme is an adenine deaminase. In other cases, the deaminase enzyme is a cytidine deaminase.

[0058] The cytosine deaminase used in CBE can be rat deaminase (rAPOBEC1), a rAPOBEC1 variant (evoAPOBEC1), an ancestor of rAPOBEC1 (EvoFERNY), an optimized ancestral rAPOBEC1 homolog (Anc689), lamprey (P. marinus) activation-induced cytidine deaminase (AID or PmCDA1), a PmCDA1 variant (evoCDA1), human APOBEC3A (hA3A), or any variant thereof.

[0059] The adenine deaminase used in ABE comprises a genetically modified TadA*. As disclosed herein, the genetically modified TadA* can deaminate adenine in a DNA sequence. As disclosed herein, the genetically modified TadA* comprises one or more (e.g., 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, or more) substitutions of the wild-type TadA amino acid sequence. In some cases, the adenosine deaminase is a dimeric adenine deaminase. In some embodiments, the dimeric adenine deaminase is a heterodimer comprising a wild-type tRNA adenosine deaminase (TadA) and a genetically modified TadA*. In some embodiments, the dimeric adenine deaminase is a homodimer comprising two genetically modified TadA*. In some embodiments, the adenosine deaminase is a monomeric adenine deaminase that comprises one genetically modified TadA*.

[0060] RNA-guided catalytically impaired nucleases The RNA-guided catalytically impaired nuclease guides the base editor to a specific location in DNA where the desired base change should occur. In some embodiments, the RNA-guided catalytically impaired nuclease is a catalytically impaired CRISPR-associated (Cas) nuclease, such as dead Cas9 (dCas9), dead Cas12 (dCas12), Cas9 nickase (Cas9n), or a derivative thereof.

[0061] The Cas9n or other catalytically impaired nuclease used in the present methods can be derived from any source, as long as it can bind to the sgRNA of the present invention and be guided to the specific sequence (e.g., the RBM20 locus) targeted by the targeting sequence of the sgRNA. In some embodiments, the catalytically impaired Cas nuclease is derived from Streptococcus pyogenes (Sp), Staphylococcus aureus (Sa), Staphylococcus auricularis (Sauri), Acidaminococcus sp. (As), Streptococcus macacae (Spy mac), or other bacteria. In certain embodiments, the Cas9n or other catalytically impaired nuclease is derived from Streptococcus pyogenes.

[0062] In some embodiments, the catalytically impaired Cas nuclease recognizes a non-NGG PAM. In some embodiments, the catalytically impaired Cas nuclease is SpCas9n-NRNH (NRNH PAM), SpCas9n-NRTH (NRTH PAM), SpCas9n-NRCH (NRCH PAM), SpRY (NRN and NYN PAM), CP-1041, SpCas9n-VQR (NGA PAM), SpCas9n-VRQR (NGA PAM), SpCas9n-EQR (NGAG PAM), SpCas9n-VRER(NGCG PAM), SaCas9n(NNGRRT PAM), SaCas9n-KKH(SaKKHn)(NNNRRT PAM), SauriCas9n(NNGG PAM), Spy-macCas9n(TAAA PAM), xCas9(NG, GAA, and GAT PAM), SpCas9n-NG(NG PAM), dLbCas12a (dLbCpf1) (TTTV PAM), and enAsCas12a (TTYN, VTTV, TRTV, or TTTV PAM).

[0063] In certain embodiments, the catalytically impaired Cas nuclease is Cas9n or a derivative thereof. In some embodiments, the Cas9n is an engineered Cas9n. In some embodiments, the engineered Cas9n is selected from the group consisting of Cas9n-NRNH, Cas9n-NRTH, Cas9n-NRCH, SpRY, CP-1041, Cas9n-VQR, Cas9n-VRQR, Cas9n-EQR, Cas9n-VRER, SaCas9n, SaCas9n-KKH (SaKKHn), SauriCas9n, Spy-macCas9n, xCas9, and Cas9n-NG. In certain embodiments, the engineered Cas9n is Cas9n-NRNH, Cas9n-NRTH, Cas9n-NRCH, CP-1041, or SpRY.

[0064] sgRNA Single guide RNA (sgRNA) can target RBM20 gene or any other gene associated with DCM. sgRNA interacts with catalytically impaired nuclease such as Cas9n, and specifically binds or hybridizes to the target nucleic acid in the genome of the cell, so that the sgRNA and the site-specific catalytically impaired nuclease are co-localized to the target nucleic acid in the genome of the cell. As used herein, sgRNA comprises a targeting sequence that comprises homology (or complementarity) to the target DNA sequence and a constant region that mediates binding to the RNA-guided catalytically impaired nuclease.

[0065] In one aspect, the sgRNA targets the RBM20 locus. In some embodiments, the sgRNA targets within exon 7, exon 9, or exon 11 of RBM20. In some embodiments, the sgRNA targets within the RS domain of exon 9 of RBM20. In some embodiments, the sgRNA targets a 6 amino acid stretch (PRSRSP) within the RS domain. In some embodiments, the sgRNA specifically targets the RBM20 gene comprising a sequence having about 80% or more identity to any one of SEQ ID NOs: 1-9. In some embodiments, the sgRNA comprises a sequence having, e.g., at least 80%, 81%, 82%, 83%, 84%, 85%, 86%, 87%, 88%, 89%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, 99% or higher identity to any one of SEQ ID NOs: 1-9, or a sequence that contains, e.g., one, two, three or more nucleotide substitutions in any one of SEQ ID NOs: 1-9. In certain embodiments, the sgRNA comprises the sequence of any one of SEQ ID NOs: 1-9.

[0066] The targeting sequence of the sgRNA can be, for example, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, or 50 nucleotides in length, or 15-25, 18-22, or 19-21 nucleotides in length, and shares homology with the target genomic sequence, particularly at positions adjacent to the CRISPR PAM sequence. The sgDNA targeting sequence is designed to be homologous to the target DNA, i.e., to share the same sequence as the non-binding strand of the DNA template, or to be complementary to the template DNA strand to which the sgRNA binds. The homology or complementarity of the targeting sequence can be perfect (i.e., having 100% homology or complementarity with the target DNA sequence), or the targeting sequence can be substantially homologous (i.e., having less than 100% homology or complementarity with the target DNA sequence, e.g., with 1 to 4 mismatches).

[0067] Each sgRNA also comprises a constant region that interacts with or binds to a site-specific nuclease (e.g., Cas9n). In the nucleic acid constructs provided herein, the constant region of the sgRNA can be about 70 to 250 nucleotides in length, or about 75 to 100 nucleotides in length, 75 to 85 nucleotides in length, or about 80 to 90 nucleotides in length, or 70, 71, 72, 73, 74, 75, 76, 77, 78, 79, 80, 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95, 96, 97, 98, 99, 100 or more nucleotides in length. The overall length of the sgRNA can be, for example, about 80 to 300 nucleotides, or about 80 to 150 nucleotides, or about 80 to 120 nucleotides, or about 90 to 110 nucleotides, or for example, 90, 91, 92, 93, 94, 95, 96, 97, 98, 99, 100, 101, 102, 103, 104, 105, 106, 107, 108, 109, or 110 nucleotides in length.

[0068] It is understood that the method can also use a two-piece gRNA (cr:tracrRNA), i.e., separate crRNA and tracrRNA molecules, where the target sequence is defined by the crispr RNA (crRNA) and the tracrRNA provides a binding scaffold for the Cas nuclease.

[0069] sgRNA can be obtained in any of a number of ways. For sgRNA, primers can be synthesized in the laboratory using an oligo synthesizer, such as those sold by Applied Biosystems, Biolytic Lab Performance, Sierra Biosystems, or others. Alternatively, primers and probes with any desired sequence and / or modification can be easily ordered from any of a number of suppliers, such as ThermoFisher, Biolytic, IDT, Sigma-Aldritch, GeneScript, etc.

[0070] Intein-mediated trans-splicing As disclosed herein, the sgRNA and BE can be introduced into a cell in one or more expression cassettes. In some embodiments, the sgRNA and BE are present together in one expression cassette. In some embodiments, the sgRNA and BE are present separately in two expression cassettes. In some embodiments, the BE is present in one expression cassette. In some embodiments, the BE is present in two or more expression cassettes, and the active BE is packaged in the cell via intein-mediated trans-splicing.

[0071] The term "intein-mediated trans-splicing" refers to the autocatalytic process of joining two protein fragments together to create a functional protein using a natural or engineered intein. The term "intern" or "protein intron" refers to a segment of a protein that can excise itself during protein splicing and join the remaining portion (extein) via a peptide bond. In some embodiments, a BE is present in two or more expression cassettes and is packaged through intein-mediated trans-splicing. For example, in one expression cassette, the N-terminal half of the BE protein is fused to the N-terminal half of an intein, and in another expression cassette, the C-terminal half of the BE protein is fused to the C-terminal half of an intein. After intein-mediated trans-splicing, the N-terminal half of the BE is linked to the C-terminal half of the BE, resulting in a functional BE useful in the present invention. The expression cassettes disclosed herein are typically driven by a promoter. In some cases, the promoter is a constitutive promoter, such as a CAG promoter. In other cases, the promoter is a muscle-specific promoter, such as the human cardiac troponin T (hTNNT2) promoter or the SPc5-12 promoter.

[0072] Delivery method The sgRNA and BE can be introduced into a cell using any suitable method, for example, by introducing one or more polynucleotides encoding the sgRNA and BE into the cell so that the sgRNA and BE are expressed in the cell, for example, by using a vector such as a viral vector, or by delivery as naked DNA or RNA. In some cases, the sgRNA and / or BE are introduced into a cell using a recombinant adeno-associated virus (rAAV) vector. In other cases, the sgRNA and / or BE are introduced into a cell as a ribonucleoprotein (RNP).

[0073] rAAV The rAAV can be derived from serotype 1 (e.g., an rAAV1 vector), 2 (e.g., an rAAV2 vector), 3 (e.g., an rAAV3 vector), 4 (e.g., an rAAV4 vector), 5 (e.g., an rAAV5 vector), 6 (e.g., an rAAV6 vector), 7 (e.g., an rAAV7 vector), 8 (e.g., an rAAV8 vector), 9 (e.g., an rAAV9 vector), 10 (e.g., an rAAV10 vector), or 11 (e.g., an rAAV11 vector). In some embodiments, the vector is an rAAV9 vector or a derivative thereof. In certain embodiments, the vector is AAVMYO, an rAAV9 variant that specifically targets muscle cells such as cardiomyocytes.

[0074] Ribonucleoproteins (RNPs) In some embodiments, the sgRNA and BE are assembled into a ribonucleoprotein (RNP) before delivery to the cell, and the RNP is introduced into the cell, for example, by electroporation. RNP is a complex of RNA and RNA-binding protein. In the context of the present method, the RNP comprises a BE (e.g., ABE) assembled with a guide RNA (e.g., sgRNA) such that the RNP can bind to and modify target DNA (through the sgRNA component of the RNP) (via the BE component of the RNP). As used herein, an RNP for use in the present method can comprise any of the guide RNAs described herein and any of the base editors described herein.

[0075] cell Ex vivo, in vitro, or in vivo modified animal cells, mammalian cells, preferably human cells, are contemplated in this disclosure, as are cells from other primates; mammals, including commercially relevant mammals such as cows, pigs, horses, sheep, cats, dogs, mice, rats, and the like; and avian cells, including commercially relevant birds such as poultry, chickens, ducks, geese, and / or turkeys.

[0076] In some embodiments, the cell is embryonic stem cell, stem cell, progenitor cell, pluripotent stem cell, induced pluripotent stem (iPS) cell, somatic stem cell, differentiated cell, mesenchymal stem cell or mesenchymal stromal cell, neural stem cell, hematopoietic stem cell or hematopoietic progenitor cell, adipose stem cell, keratinocyte, skeletal stem cell, muscle stem cell, fibroblast, NK cell, B cell, T cell, peripheral blood mononuclear cell (PBMC), or any derivative thereof.In some embodiments, the cell is iPSC.In some embodiments, the cell is iPSC-derived cardiomyocyte (CM-iPSC).

[0077] In some embodiments, the disclosure herein further includes a modified cell comprising an sgRNA and a base editor (BE) comprising an RNA-guided catalytically impaired nuclease fused to a nucleic acid base deaminase enzyme. In some embodiments, the modified cell is an iPSC. In some embodiments, the modified cell is an iPSC-derived cardiomyocyte.

[0078] Methods for Treating or Preventing DCM After delivering the sgRNA and BE to cells, e.g., iPSCs, and confirming correct modification of the target gene in the cells, multiple modified cells can be reintroduced into the subject so that they can repopulate and differentiate, e.g., into cardiomyocytes, and so that correction of the targeted point mutation can treat or prevent one or more abnormalities or symptoms in a subject with dilated cardiomyopathy (DCM). In some embodiments, the cells are cultured, expanded, selected, or induced to differentiate in vitro before being reintroduced into the subject.

[0079] In some embodiments, methods for treating or preventing DCM in an individual in need thereof are disclosed herein, the methods comprising correcting one or more point mutations (PM) that cause DCM in the individual using the genome modification methods disclosed herein. In some cases, the methods comprise reintroducing modified cells comprising a BE and an sgRNA that specifically targets a sequence comprising the point mutation, i.e., a sequence in the RBM20 locus, wherein the modified cells comprise the correct nucleotide and amino acid sequence, thereby treating or preventing DCM in the individual.

[0080] In some embodiments, the subject has a point mutation at the RBM20 locus. In some cases, the point mutation at the RBM20 locus comprises a substitution at an amino acid position including 83, 455, 535, 633, 634, 635, 636, 637, 638, 703, 716, 783, 831, 888, 913, 914, 1031, 1081, 1182, 1206, or a combination thereof, wherein the substitution and position are determined with reference to SEQ ID NO: 10. In some cases, the point mutation at the RBM20 locus comprises a substitution at an amino acid position including 633, 634, 635, 636, 637, 638, or a combination thereof. In some cases, the point mutation at the RBM20 locus comprises a substitution at amino acid position 633, 634, or a combination thereof. In particular examples, the substitutions include P633L, R634Q, or a combination thereof.

[0081] The modified cells of the present disclosure can be administered by any delivery route, including systemic or local delivery. These include, but are not limited to, enteral, gastrointestinal, epidural, oral, transdermal, intracerebral, intraventricular, epidermal, intradermal, subcutaneous, nasal, intravenous, intraarterial, intramuscular, intracardiac, intraosseous, intrathecal, intraparenchymal, intraperitoneal, intravesical, intravitreal, intracavernous, intrainterstitial, intraperitoneal, intralymphatic, intramedullary, intrapulmonary, intraspinal, intrasynovial, intrathecal, intraductal, parenteral, percutaneous, periarticular, peridural, perineural, periodontal, rectal, soft tissue, and topical. In some cases, the modified cells are reintroduced into the subject by systemic delivery. In other cases, the modified cells are reintroduced into the subject by local delivery. In some embodiments, the local delivery is intrafemoral or intrahepatic.

[0082] Pharmaceutical Compositions In some embodiments, disclosed herein are pharmaceutical compositions comprising a plurality of cells genetically modified through base editing.

[0083] In some embodiments, the pharmaceutical composition comprises a plurality of genetically modified iPSCs or CM-iPSCs disclosed herein. The pharmaceutical composition can further comprise a pharmaceutically acceptable carrier. In some embodiments, the modified cells can be formulated with one or more excipients, for example, to (1) increase stability, (2) modify biodistribution (e.g., to target a cell line to a specific tissue or cell type), or (3) modify the release profile of the encoded therapeutic factor.

[0084] The formulations of the present disclosure can include, but are not limited to, saline, liposomes, lipid nanoparticles, polymers, peptides, proteins, and combinations thereof.The formulations of the pharmaceutical compositions described herein can be prepared by any method known in the field of pharmacology or that will be developed in the future.As used herein, the term "pharmaceutical composition" refers to a composition comprising at least one active ingredient (e.g., modified cells) and, optionally, one or more pharmaceutically acceptable excipients.The pharmaceutical compositions of the present disclosure can be sterile.

[0085] The relative amounts of the active ingredient (e.g., modified cells), pharmaceutically acceptable excipient, and / or any additional ingredients in a pharmaceutical composition according to the present disclosure may vary depending on the identity, size, and / or condition of the subject being treated, as well as the route by which the composition is administered. For example, the composition may contain 0.1% to 99% (w / w) active ingredient. By way of example, the composition may contain 0.1% to 100%, e.g., 0.5 to 50%, 1 to 30%, 5 to 80%, or at least 80% (w / w) active ingredient.

[0086] The excipients used herein include, but are not limited to, any solvent, dispersion medium, diluent, or other liquid vehicle, dispersion or suspension aid, surfactant, isotonicity agent, thickener or emulsifier, preservative, etc., suitable for the desired specific dosage form. Various excipients for formulating pharmaceutical compositions and techniques for preparing compositions are known in the art (see Remington: The Science and Practice of Pharmacy, 21st Edition, AR Gennaro, Lippincott, Williams & Wilkins, Baltimore, MD, 2006; the entire contents of which are incorporated herein by reference). The use of any conventional excipient medium is contemplated within the scope of the present disclosure, except where the conventional excipient medium may be incompatible with the substance or its derivatives, such as causing any undesirable biological effects or interacting in a harmful manner with any other components of the pharmaceutical composition.

[0087] Exemplary diluents include, but are not limited to, calcium carbonate, sodium carbonate, calcium phosphate, dicalcium phosphate, calcium sulfate, calcium hydrogen phosphate, sodium phosphate, lactose, sucrose, cellulose, microcrystalline cellulose, kaolin, mannitol, sorbitol, inositol, sodium chloride, dry starch, corn starch, powdered sugar, and the like, and / or combinations thereof.

[0088] Injectable preparations can be sterilized, for example, by filtration through a bacterial-retaining filter, and / or by incorporating sterilizing agents in the form of sterile solid compositions which can be dissolved or dispersed in sterile water or other sterile injectable medium prior to use. [Example]

[0089] The following examples are offered to illustrate, but not to limit, the claimed invention.

[0090] Example 1: Striated muscle-specific base editing enables correction of mutations that cause dilated cardiomyopathy summary Dilated cardiomyopathy is the second most common cause of heart failure and has no cure other than high-risk heart transplantation. Approximately 30% of patients carry inherited mutations that may be amenable to CRISPR-based gene therapy. However, challenges associated with delivery of the editing complex and off-target concerns have prevented widespread application of CRISPR agents in the heart. We employ a combination of the viral vector AAVMYO, which has excellent targeting specificity in myocardial tissue, and a CRISPR base editor to repair a patient mutation in the cardiac splice factor Rbm20 that causes aggressive dilated cardiomyopathy. Using optimized conditions, we repair >70% of cardiomyocytes in two Rbm20 knock-in mouse models that we generated to serve as in vivo platforms for our editing strategy. Treatment of young mice restores RBM20 mislocalization and splicing of RBM20 targets, including TTN, in 75% of cells. Three months after injection, cardiac dilation and ejection fraction reach wild-type levels. Single-nucleus RNA sequencing revealed the recovery of transcriptional profiles across all major cardiac cell types, and whole-genome sequencing showed no evidence of aberrant off-target editing. Our study highlights the potential of base editors combined with AAVMYO to achieve gene repair for the treatment of inherited heart diseases.

[0091] Introduction Next-generation CRISPR tools enable genetic repair of disease-associated mutations in situ in organs of interest, thereby achieving complete prevention or cure of disease. 1 To date, a few clinical trials have been initiated applying CRISPR in vivo to treat mutations that cause blindness, high cholesterol, or protein aggregation. 2,3 Hundreds of pathogenic single nucleotide variants (SNVs) have been associated with heart disease, making the heart an attractive target for gene therapy. 4 However, few attempts have been made to correct inherited cardiac disorders in vivo. 5,6,7 and human embryos8 Pathogenic cardiac mutations have been corrected in mice, but this has important ethical considerations and requires prior knowledge of the inherited mutation. 9 , dog 10 , and pigs 11 In this study, exons were disrupted by Cas9-mediated non-homologous end joining, which carries the risk of erroneous DNA repair and can impair gene expression.

[0092] Cardiomyocytes are non-proliferating cells that are not susceptible to homology-directed gene repair, the method of choice for introducing precise genome edits. Recently, CRISPR base editors have been developed that allow efficient nucleotide conversion in vivo in post-mitotic cells. 1 Therefore, the inventors evaluated the use of base editors for the treatment of familial dilated cardiomyopathy (DCM), a severe form of heart disease and the second most common cause of heart failure. 12 Treatment options for patients with DCM include drugs that lower blood pressure or block neurohormonal systems. However, the 15-year survival rate is only 34%. 13 , the mortality rate in DCM patients undergoing this procedure is very high. 14 We focused on mutations in RBM20, which are found in 3% of patients with aggressive, early-onset DCM. 15 Patients with familial RBM20-DCM usually have a disease-causing single-nucleotide mutation. 16 This makes it a prime target for base editors to introduce single-base changes. RBM20 encodes a cardiac splice factor that regulates alternative splicing of genes important for cardiomyocyte function. 16 RBM20 mutations are concentrated in a narrow 6-amino acid stretch within the RS domain and have recently been shown to cause abnormal formation of cytoplasmic granules, which likely amplifies the disease phenotype. 17,18,19 .

[0093] In addition to correcting mutations, a major goal of any CRISPR-related gene therapy is to achieve organ-specific gene delivery to reduce the chance of potentially harmful off-target editing. 20 and embedded 21 Due to its low risk of infection and its high suitability for gene retargeting to desired organs, adeno-associated viruses (AAVs) are one of the safest and most versatile options for gene delivery. Previous cardiac gene transfer has been performed using the AAV9 serotype, despite the fact that it primarily targets the liver when injected intravenously. 22 We recently identified a synthetic variant of AAV9, named AAVMYO, that exhibits high targeting affinity for muscle cells, including cardiomyocytes, and low affinity for other organs, such as the liver. 22 Here, we leverage AAVMYO for systemic delivery of base editors to cardiomyocytes, the primary cell type expressing Rbm20. We optimized a strategy to selectively repair two pathogenic mutations in the RS domain of Rbm20, resulting in near-complete prevention of disease phenotypes in mice without evidence of guide RNA (gRNA)-dependent off-target activity.

[0094] result P635L and R636Q Rbm20 knock-in mice exhibit a DCM phenotype Adenine base editor (ABE) converts adenine (A) to guanine (G) and is being used successfully in new clinical trials 23 Because none of the existing Rbm20 animal models are suitable for ABE-mediated nucleotide conversion, we generated two mouse models harboring a G > A mutation. Specifically, we established two Rbm20 knock-in mouse models harboring the amino acid substitutions P635L and R636Q, respectively, which are orthologous to the human RBM20 mutations P633L and R634Q, previously identified in DCM patients (Figure 6a). 24No significant changes in Rbm20 mRNA expression were observed in these mice (Figure 6b). We performed deep phenotyping to identify abnormal molecular signatures and physiological properties that could be rescued by base editing. Because RBM20 localization, gene expression, and cardiac function are dysregulated in mice, humans, and pigs with RBM20 RS domain mutations, we focused on these parameters. 17,18,19 .

[0095] Immunostaining of isolated cardiomyocytes demonstrated that homozygous (HOM) P635L and R636Q mutant mice possessed cytoplasmic RBM20 granules, indicating mislocalization of mutant RBM20 protein from its normal nuclear localization (Figure 1a). Heterozygous (HET) mutants differed significantly in the degree of RBM20 mislocalization. RBM20 was primarily nuclear in P635L HET mice, but formed small cytoplasmic granules in R636Q HET mice (Figure 1a-c). RNA sequencing (RNA-seq) revealed that the number of differentially expressed genes (DEGs) compared to wild-type (WT) mice was 6-fold higher in R636Q HET compared to P635L HET, but lower than in P635L and R636Q HOM mice (Figure 1d). The DEGs common to both P635L and R636Q showed dose-dependence between HET and HOM (Figure 7). Gene Ontology (GO) analysis of the common DEGs revealed dysregulation of genes involved in muscle function and metabolic genes (Figure 1e). Expression of natriuretic peptide precursors A and B (Nppa and Nppb), biomarkers of heart failure, was also significantly increased. 25was substantially elevated in HOM mice (Figure 6c). We identified 58 differentially spliced ​​genes (DSGs) common to both P635L and R636Q HOM mice, with the majority of splicing events being exon skipping (Figure 1f). These DSGs were associated with muscle and cytoskeletal functions (Figure 6d). Notably, approximately half of all DEGs and DSGs were not overlapped between P635L and R636Q HOM (Figure 6e). While these specific genes may suggest the presence of mutation-specific downstream processes, their P values ​​were, on average, higher than those of overlapped genes (Figure 6f). This is consistent with the detection of subtle changes in transcript abundance due to biological variation, such as between mice, or other confounding factors detected by our deep RNA-seq analysis, which averaged 100 Mio. reads per genotype. No differences in the abundance of splice events were observed between P635L and R636Q HET (Figure 1f), but a subset of key RBM20 targets, including Ttn, Camk2d, and Tpm2, were more dysregulated in R636Q HET compared with P635L HET (Figure 1g). We performed RT-PCR and qPCR to verify the differentially expressed isoforms of Ttn, Camk2d, Ryr2, and Ldb3 in mutant mice and observed stronger dysregulation of Ttn and Ldb3 in R636Q HET compared with P635L HET mice (Figures 1h and 6g). P635L and R636Q HET mice showed similar levels of aberrant splicing, more intense than HET mice (Figures 1g, h, 6g).

[0096] Next, we investigated whether these molecular differences in both mouse models affected cardiac phenotypes. Survival curves showed that both P635L and R636Q HOM mice died prematurely within the first 120 days, albeit to a lesser extent than other Rbm20 RS domain mutations (survival rates: 78% P635L, 81% R636Q, 66% S637A). 26 , 51% S639G 27) (Fig. 1i). Of note, C57BL / 6N strains have been used elsewhere. 28 In contrast, we backcrossed our mutant mice to C57BL / 6J, which may explain the difference in survival, as C57BL / 6N mice are more susceptible to cardiac deterioration during pressure overload. 29 Both histological analysis and gene expression analysis revealed no major signs of fibrosis in 16-week-old mutant mice, except for upregulation of fibrotic markers Col1a2 and Mmp2 in R636Q HOM mice (Figure 6h-j). The clinical definition of DCM is based on an ejection fraction of <45% and left ventricular dilation. 12 We performed echocardiography of narcosis and confirmed that both mouse models exhibited a DCM phenotype with a severely reduced ejection fraction (Figure 1j). However, they showed only a slight increase in heart volume (except for P635L HOM), with no significant changes in left ventricular internal diameter (LVID) (Figure 6k, l). Supporting the RNA-seq results and correlating with cytoplasmic granule formation, the ejection fraction was more reduced in R636Q HET mice compared with P635L HET mice. One year later, no significant deterioration of the DCM-related phenotype was observed in ejection fraction and heart volume for P635L HET and HOM mice, whereas LVID and heart volume were significantly increased in R636Q HET and HOM mice (Figure 6m-p). All mutant mice consistently exhibited higher LVID and heart volume compared with WT mice. We conclude that P635L and R636Q Rbm20 mutant mice exhibit features of DCM seen in animals and patients with other RS ​​domain mutations. 17,18,19 For the subsequent rescue strategy, we focused on P635L and R636Q HOM mice, as they exhibited more pronounced molecular and physiological defects, allowing for better quantification of the efficacy of base editor treatment.

[0097] Base editors repair pathogenic Rbm20 mutations in vitro and in mice To test the feasibility of base editor treatment to repair the pathogenic P635L and R636Q mutations, we first transfected ABE in combination with compatible gRNAs in proliferating human iPSCs and non-proliferating cardiomyocytes derived from induced pluripotent stem cells (iPSC-CMs) harboring the orthologous RBM20 mutations P633L and R634Q (Figure 2a). Due to sequence constraints, we used ABE containing Cas9, which recognizes non-canonical PAMs such as "NRN," in combination with ABE SpRY. 30 , or NRTH / NRCH, which are ABEs named after their PAM priority. 31 Furthermore, we used a circularly permuted ABE (CP-1041) that exhibits a broader editing window for targeting the canonical PAM in P633L. 32 We observed comparable RBM20 mutation editing efficiencies between iPSCs and iPSC-CMs, with an average of up to 30% (Fig. 2b, c). No base editor was clearly superior to the others. 30,31 Indel formation was less than 2.5% and showed no significant bias among different ABEs (Figure 8a). Similarly, bystander editing (i.e., undesired A > G conversions within the gRNA window) showed no significant trend among different base editors, typically less than 1%, except for circular substitution editors. Circular substitution editors resulted in more bystander edits for P633L, likely due to their wider editing window (Figure 8b, c). Next, we analyzed whether RBM20-mediated splicing could be restored by editing iPSC-CMs. We used lentiviral transduction to generate R634Q iPSCs with stable expression of the base editor SpRY together with the targeting gRNA, resulting in a repair efficiency of 34% (Figure 2d). After differentiation into iPSC-CMs, the primary RNA target of RBM20, 33The expression levels of spliced ​​isoforms of TTN and IMMT were increased, whereas the unspliced ​​isoforms were decreased in base-edited cells, suggesting that base editing restored the RBM20-associated splicing defect (Figure 2e).

[0098] Encouraged by these results, we tested the performance of base editors in editing the heart in vivo. We used a split-intein strategy. 34 Utilizing the AAVMYO-ABE gene, we packaged both portions of the ABE (controlled by a constitutive CAG promoter) along with a gRNA expression cassette into a synthetic AAV9 variant, AAVMYO (termed AAVMYO-ABE) (Figure 2f). To determine the optimal viral concentration for systemic delivery, we used a YFP reporter transgene and observed that 1e12 vector genomes (vg) (equivalent to a total viral concentration of 8.33e13 vg / kg) ensured high viral targeting in the heart without overt transgene expression in the liver (Figure 8d, e). To identify the optimal base editor-gRNA combination, we tested the in vivo editing performance of the ABEs NRTH, NRCH, and SpRY. We performed tail vein injections of AAVMYO-ABE combined with two different gRNAs into P635L HOM mice and analyzed mutation editing in the heart, diaphragm, quadriceps, and liver 6 weeks later. Experiments were performed on 4-week-old infant mice, which resemble young DCM patients and may prevent disease progression. We found clear performance differences between both tested gRNAs. gRNA2 showed higher on-target editing efficiency than gRNA1, with less bystander editing (Figure 2g). The base editors NRCH and NRTH outperformed SpRY in terms of editing efficiency (Figure 2g). This is in contrast to in vitro editing, where no clear differences were observed between different ABEs. We also used the latest and most efficient version of adenine deaminase, Abe8e. 35We tested NRCH conjugated with gRNA1 (termed 8e-NRCH) and observed the highest degree of editing, averaging 21.4% (Figure 2g). However, this editor also exhibited 2.7% bystander edits, of which the most common bystander edit (T2; 2.64%) introduced synonymous codon changes and was likely insignificant. Notably, due to a different positioning of the base editor, a second nonsynonymous bystander edit (T1) was observed for gRNA1 in up to 1.31% of reads, resulting in a TCT (serine) to CCT (proline) codon change. Therefore, the use of gRNA1 for subsequent long-range editing and phenotypic analysis was discontinued. T1 was also detected in 8e-NRCH combined with gRNA1, but on average, in only 0.09% of reads. No indels were observed in either condition. For NRTH, we also generated an AAV9 vector, which showed less than half the editing efficiency of its AAVMYO-ABE counterpart, confirming the superiority of AAVMYO for cardiac gene delivery (Figure 2g). Notably, no significant editing was observed in the liver. The highest degree of editing occurred in the heart, followed by the diaphragm and quadriceps. This suggests that the liver and possibly other non-muscle tissues are protected from not only on-target but also off-target base editing activity (Figure 2h). Viral DNA copy number and relative RNA expression broadly correlated with editing efficiency (Figure 8f). AAVMYO primarily infects cardiomyocytes, which comprise only 30–50% of all cardiac cells. 36,37 Therefore, viral expression measured in the heart is likely an underestimate.

[0099] We also engineered a version of ABE driven by the human cardiac troponin T (hTNNT2) promoter, which resulted in highly specific editing in the heart and a lack of editing in other tissues (Figure 8g). However, hTNNT2-driven ABE was only comparable in editing efficiency to the constitutive CAG promoter-driven ABE when the viral concentration was doubled. Because we sought to eliminate potential side effects from high viral loads in the mouse model, we primarily continued with the CAG-driven ABE. To assess the potential for long-term base editing, we harvested hearts at 12 weeks rather than 6 weeks post-injection. While 8e-NRCH outperformed SpRY at 6 weeks, both showed similar editing levels at 12 weeks (Figure 8h). Liver editing did not exceed 2% even at 12 weeks post-injection (Figure 8h).

[0100] Finally, we quantified the extent of edited Rbm20 mRNA, as this would allow us to estimate editing efficacy in cardiomyocytes, the cell type that primarily expresses Rbm20 (Figure 8i). 37 Sequencing of cardiac cDNA 6 weeks after editing revealed that SpRY had edited an average of 35% of mRNA molecules, compared to 8% at the DNA level (Fig. 2g, i). Surprisingly, 12 weeks after editing, an average of 71% of Rbm20 mRNA had been edited, compared to 18% at the DNA level (Fig. 2g, i). As with copy number measurements, the discrepancy in the extent of DNA and RNA editing is likely due to AAVMYO only infecting cardiomyocytes, which represent a smaller proportion of DNA extracted from the heart overall. We conclude that the base editor delivered by AAVMYO enables highly efficient, muscle-specific repair of Rbm20 mutations in mice.

[0101] A base editor corrects the Rbm20-DCM phenotype in mice To measure long-term effects, we performed tail vein injections of AAVMYO-ABE with our best-performing editor-gRNA combination, 8e-NRCH and SpRY (the latter showing less bystander editing). These injections were performed in 4-week-old P635L and R636Q HOM mice. Long-term base editing and physiological effects were analyzed 12 weeks after injection. Amplicon-seq of whole-heart gDNA revealed average repair efficiencies of 18-20% in the heart and less than 2% in the liver across the two Rbm20 mutations tested (Figure 9a). As previously observed with gRNA2, bystander editing was detected in mice treated with 8e-NRCH but not in mice injected with SpRY (Figure 9b). Overall, bystander editing was more prevalent in P635L mice treated with 8e-NRCH compared with 6 weeks of editing. The predominant synonymous bystander edit, T2, occurred in an average of 4.09% of reads, followed by the missense mutation, T1 (0.33%). -2 (0.20%) and T 17 (0.42%) was observed (Fig. 9b). R636Q mice treated with 8e-NRCH showed one bystander edit A2 (0.69%). Notably, we observed bystander editing only in reads that also underwent correct editing, indicating that only repaired alleles are susceptible to bystander editing, effectively reducing editing efficacy (Fig. 9c). No indels were detected around the gRNA window. Furthermore, the level of Rbm20 mRNA editing was substantially higher: 68% for SpRY and >85% for 8e-NRCH (Fig. 3a).

[0102] To assess the extent of phenotypic rescue by base editing, we performed RBM20 localization and gene expression assays, as well as cardiac pump function analysis. Immunostaining for RBM20 in cardiac tissue sections revealed that cytoplasmic RBM20 granules were absent in AAVMYO-ABE-treated mice compared with saline-injected control mice, and characteristic nuclear RBM20 foci were restored in 75% of cells (Figures 3b and 3c). Next, we analyzed Ttn splicing and observed increased expression of the spliced ​​form and decreased expression of the unspliced ​​isoform. Furthermore, the splicing profiles of other RBM20 targets, Camk2d, Ldb3, and Ryr2, approached the levels of WT controls (Figures 3d and 9d). Because Ttn missplicing likely contributes to abnormal cardiomyocyte function in RBM20-DCM, we hypothesized that Ttn missplicing may contribute to abnormal cardiomyocyte function in RBM20-DCM. 38 We then examined TTN expression at the protein level. AAVMYO-ABE-treated mice showed a reduction in expression of the giant TTN isoform (G-TTN) from 83% to 17%, while the levels of the constitutive N2A and N2BA isoforms approached those of WT (Figures 3e, f and ​and9e). 3e, ​f, ​and ​Figure ​Seq​ of PBS- or ABE-treated P635L and R636Q HOM mice revealed that approximately 50% of the misspliced ​​exons in PBS-treated mice were rescued after base editing; notably, Ttn exons were involved in the most strongly reverted splice events (Figure 3g).

[0103] Finally, we performed narcosis echocardiography 8 and 12 weeks after injection. After 8 weeks, there was a clear, but nonsignificant, trend toward an increase in ejection fraction (Figure 9f). However, after 12 weeks, ejection fraction returned to near WT levels (Figure 3h). Consistent with the recovery of cardiac function, base editing reduced LVID and heart volume, although this did not reach statistical significance (Figure 3i, j). Furthermore, the expression of the heart failure biomarkers Nppa and Nppb was reduced after base editing compared with PBS-injected samples (Figure 9g). Notably, we also performed echocardiography after injection of hTNNT2-driven ABE and observed a significant improvement in ejection fraction after 12 weeks (Figure 9h). We conclude that delivery of AAVMYO-ABE significantly ameliorates the molecular and physiological defects associated with the Rbm20 mutation in mice.

[0104] Rescue of cell-type-specific gene expression by base editing To investigate whether base editing restores the cardiac transcriptional landscape, we performed single-nucleus RNA sequencing (snRNA-seq) on 40,235 nuclei isolated from the hearts of 16-week-old mice without and with base editor treatment. We analyzed nuclei from the hearts of WT (n = 7,867), P635L HOM (n = 16,218), and P635L HOM mice 12 weeks after injection of AAVMYO with NRCH (n = 6,246), 8e-NRCH (n = 2,286), or SpRY (n = 7,618) (Figure 10a-c). UMAP projections based on transcriptional similarity and clustering identified 11 major cell types expressing known cell type markers discovered in previous studies. 37(Figure 4a, b). Subclustering within ventricular cardiomyocytes revealed that cells from base-edited mice had a transcriptional profile between those of WT and P635L HOM mice (Figure 4c). AAVMYO treatment only slightly increased the proportion of immune cells (lymphoid and myeloid) (2.4-4% in WT and P635L HOM, 3.6-5.6% in base-edited mice), indicating the absence of an obvious immune response (Figure 10d). Using our snRNA-seq data, we also analyzed the expression of the base editor complex itself and confirmed the predominant targeting of cardiomyocytes by AAVMYO (Figure 10e, f). Next, we compared the transcriptome similarity between WT, P635L HOM, and P635L HOM after base editor treatment. In ventricular cardiomyocytes, the transcriptional profile of cells after base editor treatment approached that of WT, but no clear trends were observed in other major cell types (Figures 4d and 10g). Because transcriptomic effects may be masked by genes unrelated to the Rbm20 mutation, we analyzed the transcriptomic profile for genes that were significantly dysregulated in P635L HOM mice (based on snRNA-seq; see snRNA-seq analysis in the Methods section). Ventricular cardiomyocytes from base-edited mice exhibited gene expression profiles between WT and P635L HOM, indicating at least partial recovery of gene expression (Figures 4e-g). Surprisingly, we also observed major gene expression changes in other cell types, with levels reaching WT levels for atrial cardiomyocytes, pericytes, endothelial cells, myeloid cells, and fibroblasts (Figures 4g and 10h). This indicates that although AAVMYO-ABE treatment specifically targeted cardiomyocytes, downstream effects associated with RBM20-DCM, such as altering gene expression profiles in non-cardiomyocytes, were repaired.

[0105] No evidence of off-target editing induced by AAVMYO-ABE Finally, we sought to identify off-target mutations induced by the base editor. We performed whole-genome sequencing (WGS) on three P635L HOM mice treated with AAVMYO and the base editor SpRY for 12 weeks. For each mouse, we sequenced the tail (harvested before injection), liver, and heart tissues at an average genome coverage of 47x. WGS confirmed a high viral load in the heart, low levels in the liver, and background signal in the tail (Fig. 11a), with an average on-target allele editing frequency of 27% in the heart and no editing in other tissues (Fig. 11b). We compared this with our previous strategy. 39We adapted the genomic analysis (see "Methods" section for whole-genome sequencing and analysis) to identify novel variants for each tissue by overlapping three variant callers that identify SNVs and indels (Figure 11c, d). We focused on variants detected by at least two variant callers. After applying additional filtering steps, we found an average of 208–650 tissue-specific variants in the heart, liver, and tail (Figure 5a). The relative contribution of A > G / T > C nucleotide transitions was not increased in the heart compared with other tissue-specific variants or variants overlapping in all three tissues (Figure 5b). Furthermore, the allele frequency of A > G / T > C mutations was similar in all three tissues (Figure 5c), indicating the absence of systematic off-target mutations introduced by this ABE. The genomic distribution of tissue-specific variants was similar to that of common variants, with only a small proportion of SNVs in exonic regions (Figure 11e). No cardiac-specific variants were shared among the three replicates, with seven identified in two mice. Only one SNV altering the amino acid sequence of a coding gene was found. Compared to the gRNA sequence, no sequence homology was detected in the genomic region surrounding the novel variant, suggesting the absence of gRNA-dependent editing (Figure 5d). We further analyzed 16 selected sites by amplicon-seq: the seven loci with the highest sequence similarity to the gRNA used and the nine candidate A / T > G / C variants determined by WGS with the highest sequencing coverage (Figure 11f). No SNVs were detected at in silico predicted off-target sites. In addition, four of the nine candidate loci from WGS were >90% mutated in both PBS- and AAVMYO-ABE-treated mice, suggesting they are likely germline variants. For the remaining five SNVs, no difference in the percentage of editing was observed after AAVMYO-ABE injection compared to PBS. Overall, the data do not indicate the presence of DNA off-target editing induced by ABE.

[0106] RNA editing has been reported as a by-product of ABE 40,41 Therefore, we also analyzed bulk RNA-seq data obtained 12 weeks after AAVMYO-ABE treatment of P635L and R636Q HOM mice. We confirmed high expression of base editors in the heart and their absence in the liver (Figure 12a), leading to high on-target editing of 8e-NRCH and minimal bystander editing in P635L HOM mice (Figure 12b). We performed unbiased variant detection on the RNA-seq data (Figure 12c; details in Methods) and observed a slight but significant increase in the proportion of A > G mutations (17% to 19%) only in 8e-NRCH compared with PBS-treated R636Q HOM mice (Figure 12d). No significant differences were observed in other AAVMYO-ABE-treated samples. Similar to WGS, we failed to detect significant sequence homology between the region surrounding the variant and the gRNA (Figure 12e), indicating that the increased frequency of A > G mutations is not due to a gRNA-dependent effect. In summary, although our analysis does not allow for the detection of random SNVs that occur in a subset of cells, we find no evidence that our base editing strategy induces systemic off-target edits.

[0107] Consideration The potential success of any gene repair strategy depends on the degree of editing achieved in the target cell type. This efficiency depends on the degree of editor delivery in the target cell type and the efficacy of that gene editing within each cell. Because Rbm20 is primarily expressed in cardiomyocytes, we were able to estimate the overall editing efficiency by analyzing Rbm20 mRNA. Remarkably, 70-87% of cardiomyocytes were repaired, demonstrating both the excellent targeting ability of AAVMYO and the efficacy of the base editor, specifically ABE. In a recent study, Nishiyama and colleagues employed a base editor delivered by AAV9 in 5-day-old mice and achieved an average Rbm20 mRNA editing efficiency of 66%. 28However, three times more virus was used, at a total concentration of 2.5e14 vg / kg, than in our study, which treated larger 4-week-old infant mice. The difference in AAV dosage may be due to the choice of promoter and AAV serotype. Nishiyama et al. employed the TNNT2 (cTnT) promoter to confer cardiac-specific base editor expression, but in our implementation, twice the AAV dosage was required to reach editing efficiencies similar to expression controlled by the constitutive CAG promoter. Using a similar strategy in combination with thoracic injection, another study recently achieved 81% editing efficiency of cDNA in the left ventricle. 42 Notably, AAVMYO selectively targets cardiomyocytes rather than the liver, requiring less virus for delivery to the heart compared with AAV9. In our study, the base editor delivered by AAV9 was only half as efficient at correcting Rbm20 mutations as AAVMYO-ABE when injected at the same concentration. The benefits of this higher targeting specificity are twofold. First, high AAV concentrations have been associated with toxicity due to AAV-induced adaptive immune responses. 43 Second, AAV production is a time-consuming and costly process, especially for human therapy, where up to 1.5e17 vg of virus is applicable. 44 Using more efficient AAVs helps reduce the required viral concentration, further enabling treatment of older (and heavier) specimens, thereby enabling therapy in adults, a stage at which subjects typically develop their first symptoms or genetic testing is initiated. 45 .

[0108] Off-targeting is a major concern for in vivo gene therapy in patients. To uncover potential off-target editing associated with CRISPR treatment, the method of choice is WGS, as it provides unbiased analysis of genome-wide SNVs in vivo. 39We showed that hearts receiving high doses of AAVMYO-ABE did not experience an increase in A > G / T > C nucleotide transversions after treatment, which is consistent with previous versions of base editors, particularly the cytosine base editor. 46 This is contrary to illicit base editing of DNA in other parts of the genome, as observed in

[10] . Additionally, we did not identify sequences similar to gRNAs near cardiac-specific SNVs, indicating that the detected SNVs are not associated with guide-specific activity. Notably, we observed only one missense mutation among 768 cardiac-specific SNVs. However, due to the insensitivity of WGS, it is unclear whether the base editor introduced low-frequency random mutations into a subset of cells. Revealing such edits would require clonal amplification of the target cell type prior to WGS. Such a strategy has been implemented in hepatocytes. 47 However, it would not function in cardiomyocytes because they cannot proliferate. At the RNA level, we showed a significant increase in A > G edits from 17% to 19% in one mouse strain treated with 8e-NRCH, which may indicate illegitimate off-target editing by the base editor. We observed this effect only in R636Q HOM mice treated with 8e-NRCH, without any obvious gRNA-dependent effects, suggesting a low risk of introducing permanent changes in gene expression. In addition to off-target editing, base editors tend to generate bystander edits near the targeting window. 1The percentage of bystander editing varied depending on the deaminase. SpRY and NRCH conjugated to Abemax resulted in no bystander editing, whereas NRCH conjugated to the hyperactive Abe8e deaminase resulted in substantial bystander editing of >4%, even outside the predicted editing window. Depending on the targeted mutation, it may be reasonable to utilize less active Abemax-ABE versions to reduce the risk of bystander editing. In summary, it can be argued that the benefits of dramatically reducing the risk of heart failure due to the low abundance of random DNA and RNA off-target editing and the reduction in bystander editing by choosing the best ABE outweigh the risk of deleterious off-target editing.

[0109] We demonstrated that the extent of Rbm20 mRNA correction correlated well with the number of cardiomyocyte nuclei (75%) showing RBM20 protein relocalization and the percentage of G-TTN reduction (from 83% to 17%), indicating molecular recovery of cardiomyocytes. Additionally, we also observed a transcriptional shift to wild-type in other cardiac cell types, such as fibroblasts and epithelial cells, due to base editing. SnRNA-seq in RBM20 patients showed significant changes in gene expression and abundance in other cell types besides cardiomyocytes. 48 It is therefore encouraging that non-cardiomyocytes also benefit from treatment. Overall, our data suggest that the base editors prevent permanent deterioration of cardiac function. Therefore, we speculate that the positive effects on all cell types are due to the lack of structural changes in the heart throughout the animal's lifespan. AAVMYO-ABE treatment could serve as a preventative measure for genetically predisposed carriers who have not yet developed DCM, as well as a curative strategy for adult patients with DCM symptoms.

[0110] method Mouse strain generation Rbm20-P635L and Rbm20-R636Q knock-in mice were generated by microinjection of recombinant Cas9 (IDT), in vitro reconstituted crRNA:trcrRNA (IDT) targeting Rbm20, and single-stranded donor DNA as a template into zygotes. The hybrid mouse strain B6C3F1 was used for the experiments and backcrossed to C57BL / 6J.

[0111] Cell culture and differentiation Parental iPSCs and iPSCs harboring homozygous P633L or R634Q mutations in RBM20 were previously generated and characterized 24Cells were maintained on vitronectin (A31804, ThermoFisher)-coated plates in Essential 8™ Flex (A2858501, ThermoFisher) medium and passaged using Versene (15040066, ThermoFisher). Cardiomyocyte differentiation was initiated by adding 8 μM CHIR99021 (72054, STEMCELL Technologies) to RPMI-1640 medium supplemented with insulin-free B27 (RPMI-insulin, A1895601, ThermoFisher). After 24 h, 1 volume of RPMI-insulin was added, and after 72 h, the medium was changed to RPMI-insulin containing 2 μM Wnt-C59 (5148, Tocris). On days 5 and 7, the medium was changed to RPMI with insulin. On day 9, the medium was changed to RPMI with complete B27 supplement (RPMI + insulin, 17504044, ThermoFisher). On day 11, the medium was changed to RPMI + insulin without glucose and 5 mM DL-lactate. On day 14, RPMI + insulin was added, and on day 16, cells were passaged using TrypLE10x (A1217701, ThermoFisher) and RPMI + insulin supplemented with 10% knockout serum replacement (10828028, ThermoFisher) and 1.66 μM thiazovivin (72252, StemCell Technologies). The medium was changed to RPMI + insulin on day 1 after passaging, and medium changes were performed every 3 days thereafter. Passages were performed every 2–3 weeks.

[0112] ABE Plasmid Cloning The following plasmids were used for transient transfection of base editors in human iPSCs and iPSC-CMs: ABEmax-NRTH (Addgene ID: 136922), ABEmax-NRCH (Addgene ID: 136923), ABEmax-SpRY (Addgene ID: 140003), ABEmax-CP-1041 (Addgene ID: 119808), and ABE8e-CP-1041 (Addgene ID: 138493). Complementary forward and reverse gRNA sequences with compatible overhangs were annealed and ligated with the gRNA expression plasmid (Addgene ID: 53188) that had been digested with BbsI (R0539S, NEB) prior to ligation. For stable base editor expression, the coding region of Cas9 from the lentiCRISPRv2 plasmid (Addgene: 52961) was replaced with SpRY following a similar strategy previously described. 49The resulting plasmid was digested with BsmBI (R0580S, NEB) and ligated with the annealed gRNA. To generate the split-intein ABE plasmid, we used Cbh_v5 AAV-ABE N-terminus (Addgene: 137177) and Cbh_v5 AAV-ABE C-terminus (Addgene: 137178) as templates to replace the SpCas9 coding sequence with the N- or C-terminal portions of Cas9-NRTH, Cas9-NRCH, or Cas9-SpRY. Furthermore, we generated the Abe8e-NRCH version by replacing ABEmax with Abe8e in the N-terminal portion (common to Cas9-NRTH and Cas9-NRCH) using Addgene plasmid 138489 as a template. The C-terminal AAV plasmid was digested with BsmBI and ligated with the annealed oligonucleotides encoding the gRNA. Gibson assembly (E2611L, NEB) was used for all cloning and assembly steps except for the gRNA oligos, which were ligated to the backbone using T4-DNA ligase (M0202L, NEB). Sanger sequencing was performed to verify the plasmid assembly, and SmaI (R0141S, NEB) digestion was performed to monitor the integrity of the ITRs.

[0113] Lentivirus production Third-generation lentiviruses were produced in Lenti-X 293T cells (632180, Takara) by transfecting the four plasmids with linear PEI (polyethyleneimine, 25 kDa). Virus was harvested after 72 hours (stored at 4°C), fresh medium was added, and the virus was harvested again after 48 hours. All collected virus was filtered through a 0.45 μm low-protein-binding / fast-flow filter unit. Virus was precipitated using a Lenti-X concentrator (631232, Takara) according to the manufacturer's recommendations. Virus was further concentrated by ultracentrifugation at 50,000 × g for 2 hours at 4°C using 20% ​​sucrose buffer and resuspended in sterile 1× HBSS. Titers were estimated using a Lenti-X GoStix Plus (631280, Takara).

[0114] Base editing in iPSCs and iPSC-CMs iPSCs and iPSC-CMs were dissociated as single cells using StemPro™ Accutase™ Cell Dissociation Reagent (A1110501, ThermoFisher) one day before plasmid transfection and replated onto vitronectin-coated 24-well plates with RevitaCell™ Supplement (A2644501, ThermoFisher). After 24 hours, cells were transfected with 375 ng of base editor plasmid, 125 ng of U6 gRNA plasmid, and 100 ng of pmax-GFP (Lonza). Lipofectamine™ 3000 or Lipofectamine™ Stem Transfection Reagent (L3000008 or STEM00008, ThermoFisher) was used for transfection according to the manufacturer's instructions. The medium was changed on days 1 and 3 post-transfection, and GFP-positive cells were sorted by flow cytometry and analyzed by amplicon sequencing 25–35 days later. Stable SpRY-expressing R636Q iPSCs were generated by transducing cells with lentivirus expressing SpRY and R636Q gRNA2. Cells were subjected to 14 days of puromycin (A1113802, ThermoFisher) selection at 2 μg / ml prior to expansion and differentiation into cardiomyocytes. Amplicon sequencing to measure RBM20 editing efficacy was performed in three independent replicates prior to the initiation of cardiomyocyte differentiation.

[0115] Recombinant AAV production, purification, and quantification Recombinant AAVMYO was produced as previously described 50 Briefly, HEK-293T cells (Stratagene / Agilent) plated in 150 mm dishes were transfected with a three-plasmid system (pAdH - encoding the adenovirus helper functions, rep gene, and cap gene, pRep2cap9myo) 22The cells were transfected with 1000kJ / ml IgG (as well as the transgene plasmid) and PEI. After 3 days, cells were harvested, and virus was extracted from the cells by four freeze-thaw cycles. Cell lysates were treated with benzonase for 1 hour to remove non-encapsidated DNA. To remove cellular debris, the samples were centrifuged at 4000 × g, and the supernatants were collected. The supernatants were loaded onto four layers of iodixanol gradient solutions (15, 25, 40, and 60%) and then centrifuged at 183,400 × g (average) in a 70Ti rotor for 2.5 hours. Fractions were collected, and those corresponding to the 40% and 60% interfaces were pooled and concentrated by buffer exchange. Viral genome concentrations (including those in mouse tissues) were determined by ddPCR using a QX200 Droplet Digital PCR System (BioRad) with Taqman primers / probes against the CMV enhancer. Purity was determined by silver staining of SDS-PAGE gels.

[0116] Recombinant AAV9 was produced in HEK-293T / 17 cells (ATCC; CRL-11268) using a triple transfection method (using linear PEI 25 kDa) in a Corning CellSTACK 5 (CS5). After 72 hours, the supernatant (600 ml) was collected and stored at 4°C, and 600 ml of fresh medium was added. After an additional 48 hours, the initial collection was returned to the CS5, the cells were lysed, and DNA was degraded by adding Triton X-100 (1% final concentration) and 94 μl of benzonase (25–35 U / μl) at 37°C for 1 hour with shaking at 100 rpm. The cell debris / virus mixture was removed, and the CS5 was washed with 200 ml of PBS. The wash solution and cell suspension were centrifuged at 4000 × g for 20 minutes. The supernatant was filtered through a 0.45 μm PES filter and then concentrated to 30 ml using tangential flow filtration. The concentrated virus was then purified by iodixanol gradient and titered by qPCR using primers within the CMV promoter.

[0117] Mouse AAV injection Mice were injected with a mixture of AAVs expressing N- and C-terminal base editors or YFP reporters. Unless otherwise specified, 5e11 vg of each AAV was injected into the tail vein of 4-week-old mice. The average mouse weight was 12 g, resulting in a total injected virus concentration of 8.33e13 vg / kg. Mice were sacrificed after 6 or 12 weeks, and organs were collected for subsequent analysis.

[0118] DNA isolation and amplicon sequencing DNA from human cells was isolated using the Monarch® Genomic DNA Purification Kit (T3010L, NEB) according to the manufacturer's instructions, including the recommended RNase A digestion step. For DNA isolation from mice, tissue was immersed in 600 μl of PBS in a tube containing metal beads and then processed using a Fastprep homogenizer for two 30-second runs at maximum speed. One-third of the homogenized tissue was used for DNA isolation using the Monarch® Genomic DNA Purification Kit. No additional tissue lysis buffer was added, and the sample was incubated with 10 μl of proteinase K for 1 hour.

[0119] Purified DNA was amplified with human- or mouse-specific primers covering the RBM20 RS domain mutation hotspot, equipped with Nextera-compatible adapters, using Q5® Hot Start High-Fidelity 2x Master Mix (M0494L, NEB). One microliter of a 1:100 dilution was used for a second PCR to add sample-specific index barcodes (Nextera XT Index Kit v2 Set A, FC-131-2001, Illumina). Libraries were pooled and cleaned with 1x AMPure XP beads (A63881, Beckman Coulter) before sequencing on a MiSeq instrument (Illumina) using a 150 bp paired-end run.

[0120] To obtain the frequency of on-target editing by the extended gRNA binding sequence as well as indel and bystander editing, the demultiplexed amplicons were analyzed using Crispresso2. 51 was analyzed using

[0121] RNA isolation, RT-PCR, qPCR RNA from human cells was isolated using the Monarch® Total RNA Miniprep Kit (T2010S, NEB) according to the manufacturer's instructions, including on-column Dnase I digestion. For RNA isolation from mice, 1 ml of TRIzol™ (15596026, ThermoFisher) was added to 200 μl of homogenized tissue (see DNA extraction), and the tissue was processed using the Direct-zol RNA Miniprep Kit (R2052, Zymo Research) with on-column Dnase I digestion. For heart, RNA was isolated from the left ventricle. 200–500 ng of RNA was used as input for reverse transcription with SuperScript™ IV (18090010, ThermoFisher). For Amplicon-seq, RNA was further treated with ezDNase™ (11766051, ThermoFisher) before reverse transcription. RT-PCR or qPCR was performed with 1 μl of cDNA diluted 1:2 using Q5® Hot Start High-Fidelity 2X Master Mix (M0494L, NEB) or SYBR™ Green PCR Master Mix (4309155, ThermoFisher), respectively, and gene-specific primers. Samples were normalized after qPCR using the Delta Ct method with Gapdh. Additional normalization to average wild-type RNA expression was performed to calculate fold changes. RNA copy number was determined by ddPCR (see AAV virus quantification) using Taqman primers for the WPRE element and Rpp30 (Biorad, Assay ID: dMmuCPE5097025) as a housekeeping gene.

[0122] Bulk RNA sequencing and analysis 500 ng of RNA isolated from the left ventricle was processed using the NEBNext® Ultra II Directional RNA Library Prep Kit for Illumina® (E7760L, NEB) with pre-enriched mRNA using Oligo dT beads in the NEBNext Poly(A) mRNA Magnetic Isolation Module (E7490L, NEB). After library preparation, samples were multiplexed (five samples per lane) and sequenced on an Illumina NextSeq 2000. For bulk RNA-seq in Figure 3, samples were sequenced using the Smart-seq2 library preparation protocol. 52 was processed according to

[0123] Subsequent analyses were performed using Snakemake, available at https: / / github.com / FerreiraAM / dcm_lgreads_mouse_bulkRNA 53 The alignment of different samples was performed using the pipeline constructed in STAR 54 The GENCODE mouse annotation version vM29 with the primary assembly GRCm39 genome was used. We created an index and then aligned the reads for each sample using the default options of the STAR aligner. DESeq2 55 Differential expression analysis was performed using R. P values ​​in Figures 1d, e, 6e, f, and 7, which include differential gene expression analysis, were derived from one-tailed Wald tests with adjustment for multiple comparisons. 56 For each mutation associated with each experiment in the Rsubread R package 57Comparisons were performed using count matrices created from BAM files in Metascape. Individual-wise Log2 fold change (log2FC) was calculated for each mutation using the gene-wise mean value from all WT samples of a mutation in one experiment: log2FC for gene A = log2(value for gene A / WT mean for gene A). 58 was used for gene ontology analysis.

[0124] We used rMATS to detect differential alternative splicing events. 59 Pairwise comparisons were performed, and the results from the Junction Counts (JC) files were analyzed in R. We identified overlapping splice junction events between different conditions and filtered out significant events. P values ​​in Figures 1f and 1g and 6d and 6e, which involve differential splice analysis, were derived from likelihood ratio tests. Events were considered significant if the false discovery rate (FDR) was less than 0.01 and the mean delta PSI value (compared to WT) was higher than 0.1 or lower than -0.1. In Figure 3g, significant splice events were classified into three categories: rescued, misspliced, or unchanged. We calculated the mean delta PSI difference (compared to WT) between PBS-treated and ABE-treated mice and examined the absolute difference (ΔΔPSI). If the delta PSI was less than 0.1, the event was classified as unchanged. The remaining events were classified as either rescued or misspliced. In addition, we defined the PSI value of the PBS-treated sample as the original value ΔPSI_original and the PSI value of the base-edited sample as the edited value ΔPSI_edited, and used the following criteria: Rescued: ΔPSI_original > 0 and ΔPSI_edited >= 0 or -0.2 <= ΔPSI_edited <= 0.2 and ΔPSI_original > ΔPSI_edited ΔPSI_original < 0 and ΔPSI_edited <= 0 or -0.2 <= ΔPSI_edited <= 0.2 and ΔPSI_original < ΔPSI_edited Misspliced: ΔPSI_original > 0 and ΔPSI_edited >= 0 and ΔPSI_original < x_edited ΔPSI_original < 0 and ΔPSI_edited <= 0 and ΔPSI_original > x_edited ΔPSI_original < 0 and ΔPSI_edited >= 0.2 ΔPSI_original > 0 and ΔPSI_edited <= -0.2

[0125] Mouse cardiomyocyte isolation For immunostaining of RBM20 granules (Fig. 1a-c), we performed Langendorff-free cardiomyocyte isolation as previously described. 60Briefly, mice were sacrificed, and the right ventricle was immediately flushed with 7 ml of EDTA-containing buffer. After clamping the ascending aorta, the heart was transferred to a Petri dish containing EDTA buffer. An additional 10 ml of EDTA buffer was injected into the left ventricle. After injection of 3 ml of perfusion buffer, the heart was transferred to a Petri dish containing collagenase. Subsequently, 50 ml of collagenase buffer was injected into the left ventricle, which was then transferred to a plate containing 3 ml of collagenase buffer and cut into small pieces. Five ml of stop solution was added to the whole, and the cells were filtered through a 100 μm cell strainer and allowed to settle by gravity to enrich for cardiomyocytes. Gravity settling was performed twice before plating the cells onto laminin (5 μg / ml in PBS, 23017015, ThermoFisher)-coated plates supplemented with GlutaMAX™ (10565018, ThermoFisher) and DMEM / F12 with 10% FBS. After 2 h, cells were fixed with 4% paraformaldehyde (PFA, methanol-free, 28906, ThermoFisher) for 10 min at room temperature and stored in PBS for subsequent imaging.

[0126] RBM20 immunostaining and granule quantification Immunostaining was performed on either isolated adult mouse cardiomyocytes or tissue sections. Isolated and fixed cardiomyocytes were incubated in 0.5% Triton X-100 in PBS for 5 minutes, then washed with PBS and incubated in blocking solution (2% BSA in PBS) for 1 hour at room temperature. Cells were incubated overnight in blocking solution containing anti-Rbm20 (PA5-58068, Invitrogen) and anti-sarcomeric α-actinin (ab9465, Abcam), both diluted 1:250. Cells were then washed three times with blocking solution and stained with secondary antibodies: Alexa Fluor 488 goat anti-mouse IgG (A11001, Invitrogen) and Alexa Fluor 568 goat anti-rabbit IgG (A110011, Invitrogen), both diluted 1:1000. After incubation for 1 h at room temperature, slides were washed three times with blocking solution and mounted with ProLong Gold antifade reagent containing DAPI (D1306, Invitrogen). Images were acquired using an LSM 980 AIRY confocal microscope (Zeiss). RBM20 granule quantification in isolated cardiomyocytes was performed using the ImageJ plugin AggreCount (v1.13) according to published procedures. 61 The analysis was performed using the MRI software. The number and average size of granules per whole cell were used for the figures. At least 10 cells were analyzed for each of the three mice per genotype.

[0127] For histological staining, 8 μm transverse sections of heart samples were deparaffinized in xylene and rehydrated in water through decreasing concentrations of ethanol. Slide sections were subjected to heat-induced antigen retrieval in 10 mM Tris-EDTA pH 9 buffer for 20 minutes. These sections were then permeabilized with 0.3% Triton X-100, blocked with 5% donkey serum, and incubated overnight at 4°C with 0.5 μg / ml rabbit anti-Rbm20 antibody (PA5-53068, Invitrogen). Immunofluorescence detection was performed by tyramide signal amplification using anti-rabbit HRP (12-348, Sigma), biotinyl-tyramide (SML2135, Sigma), and streptavidin-Alexa 488 (S11223, Molecular Probes). Images were acquired by wide-field microscopy with an automated whole-slide scanner. Nuclear RBM20 localization relative to cytoplasmic RBM20 localization was manually quantified in three to four slices per mouse heart per two mice per condition, totaling 250–500 cells.

[0128] Pico Sirius Red staining Hearts were processed for standard paraffin embedding. Sagittal sections around the midpoint of each sample at 8 μm were collected onto Superfrost Plus slides. After deparaffinization and hydration with alcohol and water, sections were stained with a solution of picrosirius red (0.5 g / 500 ml saturated picric acid; Sigma) for 1 h at room temperature. Sections were then washed in two changes of acidified water (5 ml glacial acetic acid / 1 ml later distilled water), dehydrated in 100% ethanol, and mounted with Permount. Images were acquired with an automated whole-slide scanner.

[0129] Echocardiography Mice were anesthetized with 2-2.5% isoflurane (HDG9623V, Baxter Deutschland GmbH, Germany), and heart and respiratory rates were continuously monitored. Echocardiography was performed using a Vevo 2100 imaging system with an MS400 transducer (both FUJIFILM VisualSonics, Inc., Canada) to capture short- and long-axis B-mode and M-mode images. During echocardiography, mice were placed on a heating pad to prevent hypothermia. Echocardiography parameters were then analyzed using VisualSonics VevoLab software.

[0130] Vertical SDS agarose gel electrophoresis (VAGE) VAGE for detection of TTN protein isoforms was performed as previously described. 62 The method was performed as previously described with minor modifications. Five to ten mg of cardiac tissue fragments were lysed in a microtube using a pestle in 40 volumes (w / v) of VAGE sample buffer (8 M urea, 2 M thiourea, 3% SDS, 0.03% bromophenol blue, 0.05 M Tris-HCl, 75 mM DTT, pH 6.8) at 60°C for 2 to 3 minutes. Then, 50% glycerol buffer (50 ml H2O, 50 ml Ultrapure Glycerol, 1 tablet of protease inhibitor cocktail (Product No. 11697498001, Roche)) was added (final concentration: 12%), and the sample was treated at room temperature for an additional 3 to 5 minutes. After a 5-minute cooling period on ice, the sample was centrifuged at 16,000 × g for 5 minutes. The supernatant was collected and stored at 80°C. The sample was thawed by heating to 60°C for 2 minutes and analyzed using VAGE. After running the gel, the gel was fixed in 50% methanol, 12% acetic acid, 5% glycerol in ddH2O for 1 hour and then dried overnight. The gel was rehydrated in H2O, stained with Coomassie, and scanned for quantification. Analysis and quantification were performed using AIDA software.

[0131] Whole genome sequencing and analysis DNA for WGS was prepared by PCR-free library preparation using the NEBNext Ultra II DNA PCR-free Library Prep kit (E7410L, NEB). Sequencing was performed on an Illumina NextSeq 2000 P3 150PE.

[0132] Analysis of sequencing data was performed using a customized Snakemake workflow. Raw FASTQ files were first analyzed using cutadapt v.3.5 63 The raw FASTQ files were processed for 3' adapter trimming using bwa-mem v.0.7.17. The trimmed reads were then processed using bwa-mem v.0.7.17. 64 The BAM files were aligned to a hybrid reference sequence of the mouse genome mm10 concatenated with the AAV vector backbone sequence containing the N- and C-terminal components of SpRY-Cas9-ABE using the RT-PCR tool. BAM files were sorted, marked for PCR duplicates, and recalibrated for base quality scores using GATK v4.1.9.0. Three variant callers, each with their own SNP (GATK Mutect2 v4.1.9.0 (MU)) and RT-PCR tool (RT-PCR tool). 65 GATK HaplotypeCaller v.4.1.9.0 (HC) 65 , Lofreq v.2.1.5 (LF) 66 ) and indels (MU, HC, Scalpel v.0.5.4 (SC) 67 ) was applied for variant calling. For MU and HC, variant calling was performed in cohort mode using BAM files from all three tissue samples from the same individual. Therefore, allelic depth and frequency (AF) of the reference and alternative alleles were recorded even for variants not present across all tissue types. For LF and SC, variants were called from one sample per call using default parameters. All variants were left-justified and analyzed using bcftools (v.1.9) to allow comparison between variant callers. 68For further analysis, the allele depth called by MU was used if present, otherwise it was replaced by the value determined by HC. To add functional annotations to the detected variants, ANNOVAR (v.2020-06-08) was used. 69 was used. To identify variants with high confidence, we required that variants (1) be called by at least two variant callers, (2) be covered by at least five reads per tissue type, and (3) have at least two alternative allele reads across all tissues. After this quality filtering step, tissue-overlapping variants were defined as variants present in all three tissues. To identify novel mutations with high confidence, we used the Mouse Genome Project (MJD) algorithm. 70 or dbSNP 71 Variants that overlapped with any known variants annotated with were excluded. From this pool of variants, tissue-specific variants were defined as variants with AF > 0 in the tissue of interest and AF = 0 or not measured in other tissues. To further characterize tissue-specific variants, they were tested for potential causation by CRISPR base editor treatment. For each variant, a ±30 base section around its start site was examined for sequence homology to the gRNA and PAM sequences. A custom script was developed for sequence alignment and calculation of the minimum edit distance, allowing 1-bp indels or mismatches only in the seed region but not at the PAM site.

[0133] RNA-seq variant analysis Sequencing data were analyzed using a customized Snakemake workflow. Raw FASTQ files were analyzed using STAR v2.7.9a in two-pass mode. 54The hybrid reference sequences were aligned against the same hybrid reference sequence used in the WGS analysis by qPCR. BAM files were sorted and marked for PCR duplicates using GATK v4.1.9.064. Variant calling was performed using three variant callers: GATK HaplotypeCaller v4.1.9.0 (HC) 65 , Strelka v.2.9.10 (ST) 72 and Platypus v.0.8.1 (PL) 73 For HC, the selected and marked reads were analyzed according to the GATK Best Practices for RNA-seq variant calling. 74,75 For PL, the selected and marked reads were preprocessed using Opossum v.0.2 76 The samples were processed using the sparsely-coded reads, and ST was run in RNA mode using the culled and marked reads as input. All algorithms called variants in cohort mode using BAM files from all three tissue samples from the same individual. All variants were normalized, left-aligned, and annotated as described in the WGS analysis section. For further analysis, the allele depth called by PL was used if present, otherwise it was replaced with the value determined by HC. To identify variants with high confidence, we required that the variants (1) be called by at least two of the three variant callers, (2) be covered by at least five reads (tissues examined individually), (3) have at least two alternative allele reads across all tissues, and (4) be located in exons, introns, or UTR3 / 5 regions. Known variants annotated by dbSNP or MGP were excluded, and the remaining variants were grouped into tissue-specific variants or tissue-overlapping variants.

[0134] In Figure 12b, REDItools2 77All reads were extracted from the target region using the RT-PCR tool. Triplicate reads were combined and base and position-specific percentages were calculated. Sequence similarity between gRNA regions surrounding SNVs was examined as described for WGS analysis.

[0135] Nuclei isolation and snRNA-seq Nuclei from mouse hearts were isolated using a previous protocol with some adaptations. Briefly, hearts were washed three times with PBS, minced, and incubated in 5 ml of 1x Red Blood Cell Lysis Buffer (Z3141, Promega) for 5 minutes with manual shaking. In total, 5 ml of PBS was added, followed by centrifugation at 500 x g for 2 minutes and an additional washing step with 10 ml of PBS. The pellet was then homogenized in 1 ml of homogenization buffer. 78 The nuclei were resuspended in 2 ml of homogenization buffer and homogenized in a Dounce homogenizer on ice, 8 times with pestle "A" and 20 times with pestle "B." The nuclei were filtered through a 70 μM strainer, followed by a 40 μM strainer and a 20 μM strainer. The nuclei were centrifuged at 1000 × g for 5 minutes, and the pellet was resuspended in 2 ml of homogenization buffer. The nuclei solution was then diluted with 10 ml of sucrose buffer. 79 The cells were layered onto a 1000×g plate and centrifuged at 1000×g for 5 minutes. The pellet was washed with 2 ml of homogenization buffer and resuspended in 0.2 ml of calcium- and magnesium-free PBS containing 2% BSA and 0.2 U / μl RNasin® Plus RNase inhibitor (N2615, Promega). Dapi-stained nuclei were sorted by flow cytometry in FACS buffer. The gating strategy is shown in the source data. Sequencing libraries were prepared using the Single Cell 5' Reagent Kit v2 Dual Index (1000265, 10xGenomics) and sequenced using a NextSeq550 Mid 75 PE.

[0136] SnRNA-seq analysis SnRNA-seq data were aligned to the mouse reference mm10 (GENCODE vM23 / Ensembl 98) using 10x Genomics' Cell Ranger 7.0. Downstream analysis on the gene count matrix was performed in R v4.2.1 and Seurat v4. In the preprocessing step, cells were filtered so that each cell had 100–2500 active genes with non-zero counts. Cells showing counts greater than 1% belonging to mitochondrial genes were not included. Counts for each cell were log-normalized, and the 2000 most variable features were identified for each run separately. Preprocessed data from different runs were aligned using the FindIntegrationAnchors method in Seurat. Principal component analysis (PCA) was performed on the integrated data to identify the 30 largest contributors to the variation in gene expression profiles. Clusters were generated using the Louvain algorithm with a resolution parameter of 0.5. 80 For visualization of cell clusters, the Uniform Manifold Approximation and Projection (UMAP) reduction method was used for 30 PCs. 81 Each cluster was analyzed using Heart Cell Atlas 37 The clusters were mapped to specific cell types using markers provided by

[14] . Within each cluster, an additional PCA was performed on the unintegrated gene expression data to identify subcluster variation. UMAP visualizations were obtained from the five largest PCs, and pairwise cell distances were calculated from the first two PCs using a Euclidean metric.

[0137] To compare cell-type-specific genotype variation, differential expression analysis was performed between WT and P635L HOM cells within each cluster using the Wilcoxon Rank Sum test with a maximum P value of 0.05. Up to 15 up- and down-regulated genes were identified per cell type. Their expression values ​​were rescaled across cells so that their values ​​fell between 0 and 1, where 1 indicates maximum activation of the gene in the cell and 0 indicates no expression. An overall activation score was then calculated for each cell by averaging the rescaled expression values ​​across the list of up- and down-regulated genes. By scanning for a threshold between 0 and 1, the percentage of cells with activity scores above the threshold was used as a proxy for the expression of that population's gene set. The critical threshold for comparison was selected as the value below which genotypes with down-regulated activity fell below 50% active cells.

[0138] Statistics and reproducibility GraphPad Prism software (v9.3.1) was used for statistical analysis, except for RNA-seq data, which were analyzed in R. Data were analyzed by unpaired t-test, one-way ANOVA, or two-way ANOVA with Tukey's multiple comparison post hoc test or log-rank test. The legend for each figure indicates the name of the test, P value, and number of biological replicates. Data are presented as mean ± SEM. No statistical methods were used to predetermine sample size. No data were excluded from the analysis. Experiments were not randomized, and in most cases, investigators were not blinded to allocation during the experiment and outcome assessment.

[0139] References TIFF2025538153000003.tif85146TIFF2025538153000004.tif220146TIFF2025538153000005.tif215146TIFF20255381530 00006.tif218146TIFF2025538153000007.tif209146TIFF2025538153000008.tif213146TIFF2025538153000009.tif246147

[0140] Exemplary Embodiments Exemplary embodiments provided in accordance with the subject matter of this disclosure include, but are not limited to, the following embodiments and the claims. Aspect 1. 1. A method for correcting a point mutation at the RBM20 locus in a cell, comprising: introducing into the cell (i) a single guide RNA (sgRNA) that targets the sequence containing the point mutation, and (ii) a base editor (BE); the sgRNA binds to and guides the base editor to the target sequence, such that the base editor corrects the point mutation at the RBM20 locus in the cell. The method. Aspect 2. The method of embodiment 1, further comprising the step of isolating said cell from a subject prior to introduction of said sgRNA and said BE. Aspect 3. 3. The method of embodiment 1 or 2, wherein the point mutation at the RBM20 locus comprises a substitution at an amino acid position comprising: 83, 455, 535, 633, 634, 635, 636, 637, 638, 703, 716, 783, 831, 888, 913, 914, 1031, 1081, 1182, 1206, or a combination thereof, wherein the substitution and the position are determined with reference to SEQ ID NO: 10. Aspect 4. The method of embodiment 3, wherein the point mutation in the RBM20 locus comprises a substitution at an amino acid position comprising 633, 634, 635, 636, 637, 638, or a combination thereof. Aspect 5. The method of embodiment 4, wherein said point mutation in the RBM20 locus comprises a substitution at amino acid position 633, 634, or a combination thereof. Aspect 6. The method of embodiment 5, wherein said substitutions comprise P633L, R634Q, or a combination thereof. Aspect 7. 7. The method of any one of embodiments 1-6, wherein the sgRNA comprises a sequence having about 80% or greater identity to any one of SEQ ID NOs: 1-9. Aspect 8. 8. The method of any one of embodiments 1 to 7, wherein the sgRNA comprises the sequence of any one of SEQ ID NOs: 1 to 9. Aspect 9. The method of any one of embodiments 1-8, wherein said BE is an adenine base editor (ABE) or a cytidine base editor (CBE). Aspect 10.

[0023] Aspect 1. The method of any one of aspects 1-9, wherein the BE comprises an RNA-guided catalytically impaired nuclease fused to a nucleobase deaminase enzyme. Aspect 11. 11. The method of embodiment 10, wherein said RNA-guided catalytically impaired nuclease is dead Cas9 (dCas9), dCas12, or Cas9 nickase (Cas9n), or a derivative thereof. Aspect 12. 12. The method of embodiment 10 or 11, wherein said RNA-guided catalytically impaired nuclease is an engineered Cas9n. Aspect 13. 13. The method of embodiment 12, wherein said engineered Cas9n is Cas9n-NRNH, Cas9n-NRTH, Cas9n-NRCH, CP-1041, or SpRY. Aspect 14. 11. The method of embodiment 10, wherein said nucleobase deaminase enzyme is a single-stranded DNA (ssDNA)-specific deaminase enzyme. Aspect 15. 15. The method of embodiment 14, wherein the deaminase enzyme is an adenine deaminase or a cytidine deaminase. Aspect 16. 16. The method of any one of embodiments 1-15, wherein said BE is BE1, BE2, BE3, BE4, ABE6.3, ABE7.8, ABE7.9, ABE7.10, BE4max, AncBE4max, ABEmax, ABE8e, ABE-SpRY, CBE-SpRY, ABE-CP-1041, and CBE-CP-1041. Aspect 17. 17. The method of any one of embodiments 1 to 16, wherein the sgRNA and the BE are introduced into the cell in one or more expression cassettes. Aspect 18. 20. The method of embodiment 17, wherein the BEs are present in one expression cassette. Aspect 19. 20. The method of embodiment 17, wherein the BEs are present in two expression cassettes, and an active BE is packaged in the cell through intein-mediated trans-splicing. Aspect 20. 20. The method of any one of aspects 17 to 19, wherein the expression cassette comprises a promoter. Aspect 21. 21. The method of embodiment 20, wherein the promoter is a CAG promoter. Aspect 22. 21. The method of embodiment 20, wherein said promoter is a human cardiac troponin T (hTNNT2) promoter. Aspect 23. 23. The method of any one of aspects 1 to 22, wherein the sgRNA and the BE are introduced into the cell using a recombinant adeno-associated virus (rAAV) vector. Aspect 24. The method of embodiment 23, wherein the rAAV vector is an AAVMYO vector. Aspect 25. 23. The method of any one of embodiments 1 to 22, wherein the sgRNA and the BE are introduced into the cell as a ribonucleoprotein (RNP). Aspect 26. 28. The method of embodiment 27, wherein the RNP is introduced into the cell by electroporation. Aspect 27. Aspect 27. The method of any one of aspects 1 to 26, wherein the cells are induced pluripotent stem cells (iPSCs) or iPSC-derived cardiomyocytes (CM-iPSCs). Aspect 28. 10. A method for treating or preventing dilated cardiomyopathy (DCM) in a subject, comprising: (i) genetically modifying cells from the subject using the method of any one of embodiments 1 to 27; and (ii) reintroducing the cells into the subject, wherein the reintroduction is effective to treat or prevent the subject from having DCM. Aspect 29. 29. The method of embodiment 28, wherein the subject has a point mutation in the RBM20 locus. Aspect 30. 30. The method of embodiment 29, wherein the point mutation in the RBM20 locus comprises a substitution at an amino acid position comprising: 83, 455, 535, 633, 634, 635, 636, 637, 638, 703, 716, 783, 831, 888, 913, 914, 1031, 1081, 1182, 1206, or a combination thereof, wherein the substitution and the position are determined with reference to SEQ ID NO: 10. Aspect 31. 31. The method of embodiment 29 or 30, wherein the point mutation in the RBM20 locus comprises a substitution at an amino acid position comprising 633, 634, 635, 636, 637, 638, or a combination thereof. Aspect 32. 32. The method of embodiment 31, wherein the point mutation in the RBM20 locus comprises a substitution at amino acid position 633, 634, or a combination thereof. Aspect 33. 33. The method of embodiment 32, wherein said substitutions comprise P633L, R634Q, or a combination thereof. Aspect 34. The method of any one of aspects 28 to 33, wherein the cells are reintroduced into the subject by systemic delivery. Aspect 35. The method of any one of aspects 28-33, wherein the cells are reintroduced into the subject by local delivery. Aspect 36. 36. The method of embodiment 35, wherein said local delivery is intrafemoral or intrahepatic. Aspect 37. The method of any one of embodiments 28 to 36, wherein the cells are cultured, expanded, selected, and / or induced to differentiate in vitro before being reintroduced into the subject. Aspect 38. An sgRNA that specifically targets the RBM20 gene, comprising a sequence having about 80% or more identity to any one of SEQ ID NOs: 1 to 9. Aspect 39. 39. An iPSC comprising the sgRNA of embodiment 38 and a base editor (BE) comprising an RNA-guided catalytically impaired nuclease fused to a nucleobase deaminase enzyme. Aspect 40. A cardiomyocyte derived from the iPSC of embodiment 39. Aspect 41. A pharmaceutical composition comprising a plurality of iPSCs of embodiment 39 or a plurality of cardiomyocytes of embodiment 40.

[0141] It is understood that the examples and embodiments described herein are for illustrative purposes only, and that various modifications or changes will be suggested to those skilled in the art in light thereof, which are to be included within the spirit and scope of this application and the appended claims. All publications, patents, and patent applications cited herein are incorporated herein by reference in their entirety for all purposes.

[0142] Unofficial sequence listing TIFF2025538153000010.tif218146

Claims

1. 1. A method for correcting a point mutation at the RBM20 locus in a cell, comprising: introducing into the cell (i) a single guide RNA (sgRNA) that targets the sequence containing the point mutation, and (ii) a base editor (BE); the sgRNA binds to and guides the base editor to the target sequence, such that the base editor corrects the point mutation at the RBM20 locus in the cell. The method.

2. The method of claim 1, further comprising isolating the cells from a subject prior to introduction of the sgRNA and the BE.

3. 3. The method of claim 1 or 2, wherein the point mutation in the RBM20 locus comprises a substitution at an amino acid position comprising 83, 455, 535, 633, 634, 635, 636, 637, 638, 703, 716, 783, 831, 888, 913, 914, 1031, 1081, 1182, 1206, or a combination thereof, wherein the substitution and the position are determined with reference to SEQ ID NO:

10.

4. 4. The method of claim 3, wherein the point mutation in the RBM20 locus comprises a substitution at an amino acid position comprising 633, 634, 635, 636, 637, 638, or a combination thereof.

5. 5. The method of claim 4, wherein the point mutation in the RBM20 locus comprises a substitution at amino acid position 633, 634, or a combination thereof.

6. 6. The method of claim 5, wherein the substitution comprises P633L, R634Q, or a combination thereof.

7. The method of claim 1, wherein the sgRNA comprises a sequence having about 80% or greater identity to any one of SEQ ID NOs: 1-9.

8. The method of claim 1, wherein the sgRNA comprises the sequence of any one of SEQ ID NOs: 1-9.

9. 2. The method of claim 1, wherein the BE is an adenine base editor (ABE) or a cytidine base editor (CBE).

10. 10. The method of claim 1, wherein the BE comprises an RNA-guided catalytically impaired nuclease fused to a nucleobase deaminase enzyme.

11. 11. The method of claim 10, wherein the RNA-guided catalytically impaired nuclease is dead Cas9 (dCas9), dCas12, or Cas9 nickase (Cas9n), or a derivative thereof.

12. 11. The method of claim 10, wherein the RNA-guided catalytically impaired nuclease is an engineered Cas9n.

13. 13. The method of claim 12, wherein the engineered Cas9n is Cas9n-NRNH, Cas9n-NRTH, Cas9n-NRCH, CP-1041, or SpRY.

14. 11. The method of claim 10, wherein the nucleobase deaminase enzyme is a single-stranded DNA (ssDNA)-specific deaminase enzyme.

15. 15. The method of claim 14, wherein the deaminase enzyme is adenine deaminase or cytidine deaminase.

16. 2. The method of claim 1, wherein the BE is BE1, BE2, BE3, BE4, ABE6.3, ABE7.8, ABE7.9, ABE7.10, BE4max, AncBE4max, ABEmax, ABE8e, ABE-SpRY, CBE-SpRY, ABE-CP-1041, and CBE-CP-1041.

17. The method of claim 1, wherein the sgRNA and the BE are introduced into the cell in the form of one or more expression cassettes.

18. 18. The method of claim 17, wherein the BEs are present in one expression cassette.

19. 18. The method of claim 17, wherein the BEs are present in two expression cassettes and the active BE is packaged in the cell through intein-mediated trans-splicing.

20. 18. The method of claim 17, wherein the expression cassette comprises a promoter.

21. 21. The method of claim 20, wherein the promoter is a CAG promoter.

22. 21. The method of claim 20, wherein the promoter is the human cardiac troponin T (hTNNT2) promoter.

23. The method of claim 1, wherein the sgRNA and the BE are introduced into the cell using a recombinant adeno-associated virus (rAAV) vector.

24. The method of claim 23, wherein the rAAV vector is an AAVMYO vector.

25. The method of claim 1, wherein the sgRNA and the BE are introduced into the cell as a ribonucleoprotein (RNP).

26. 28. The method of claim 27, wherein the RNP is introduced into the cell by electroporation.

27. 2. The method of claim 1, wherein the cells are induced pluripotent stem cells (iPSCs) or iPSC-derived cardiomyocytes (CM-iPSCs).

28. 1. A method for treating or preventing dilated cardiomyopathy (DCM) in a subject, comprising: (i) genetically modifying cells from the subject using the method of claim 1; and (ii) reintroducing the cells into the subject, wherein the reintroduction is effective to treat or prevent the subject from having DCM.

29. 29. The method of claim 28, wherein the subject has a point mutation in the RBM20 locus.

30. 30. The method of claim 29, wherein the point mutation in the RBM20 locus comprises a substitution at an amino acid position comprising 83, 455, 535, 633, 634, 635, 636, 637, 638, 703, 716, 783, 831, 888, 913, 914, 1031, 1081, 1182, 1206, or a combination thereof, wherein the substitution and the position are determined with reference to SEQ ID NO:

10.

31. 31. The method of claim 29 or 30, wherein the point mutation in the RBM20 locus comprises a substitution at an amino acid position comprising 633, 634, 635, 636, 637, 638, or a combination thereof.

32. 32. The method of claim 31, wherein the point mutation in the RBM20 locus comprises a substitution at amino acid position 633, 634, or a combination thereof.

33. 33. The method of claim 32, wherein the substitution comprises P633L, R634Q, or a combination thereof.

34. 30. The method of claim 28, wherein the cells are reintroduced into the subject by systemic delivery.

35. 30. The method of claim 28, wherein the cells are reintroduced into the subject by local delivery.

36. 36. The method of claim 35, wherein the local delivery is intrafemoral or intrahepatic.

37. 29. The method of claim 28, wherein the cells are cultured, expanded, selected, and / or induced to differentiate in vitro before being reintroduced into the subject.

38. An sgRNA that specifically targets the RBM20 gene, comprising a sequence having about 80% or more identity to any one of SEQ ID NOs: 1-9.

39. 39. An iPSC comprising the sgRNA of claim 38 and a base editor (BE) comprising an RNA-guided catalytically impaired nuclease fused to a nucleobase deaminase enzyme.

40. A cardiomyocyte derived from the iPSC of claim 39.

41. 41. A pharmaceutical composition comprising a plurality of iPSCs according to claim 39 or a plurality of cardiomyocytes according to claim 40.

Images