Rbm20 antisense oligonucleotides alleviate diastolic dysfunction in a mouse model of cardiometabolic heart failure (HFpEF)

Abstract

Aims: Heart failure with preserved ejection fraction (HFpEF) is prevalent, deadly, and difficult to treat. Risk factors such as obesity and hypertension contribute to cardiac inflammation, metabolic defects, and pathological remodelling that impair ventricular filling in diastole. Titin based stiffness is a main determinant of diastolic function and can be adjusted by the splicing regulator RNA binding motif protein 20 (RBM20). Inhibition of RBM20 using antisense oligonucleotides (ASOs) induces expression of compliant titin isoforms, which reduce stiffness. However, dose finding and documenting utility in primarily cardiometabolic disease remains challenging.

Methods and results: Here, we optimized RBM20-ASO dosing in a HFpEF mouse model that closely mimics human disease, characterized by metabolic syndrome and comorbidities, but without primary defects in titin or RBM20. Partial inhibition of RBM20 (∼50%) selectively increased compliant titin isoforms, improving diastolic function while preserving systolic performance. This intervention reduced left ventricular stiffness, enhanced relaxation, and mitigated cardiac hypertrophy, despite ongoing systemic comorbidities.

Conclusion: Our findings demonstrate that targeting titin stiffness with Rbm20-ASOs can serve as an alternative or adjunctive therapeutic strategy for HFpEF to restore cardiac function and prevent further organ damage. The approach may offer benefits even in the presence of phenotypic heterogeneity and unresolved systemic comorbidities.

Keywords: ASO; HFpEF; RBM20; Therapy; Titin.

Graphical Abstract

Figure 1

Titin isoform expression in LV myocardium of control C57BL/6n mice after weekly ASO injections. (A) Rbm20-ASO was administered subcutaneously at 25 mg/kg weekly to control mice. (B) Mice were sacrificed at the indicated time points for titin protein isoform analysis. (C) A significant amount of N2BA-N titins was expressed starting after 4 weeks. After six injections, 50% of the total titin is N2BA-N titin. The data are presented as means ± SD, **** P ≤ 0.0001 indicates a significant increase in N2BA-N compared to no injections (n = 7, 3, 3, 2, 2, 2 mice for weeks 0, 2, 4, 5, 6, 8, respectively)—comparison using 2-way ANOVA with Dunnett’s multiple comparisons.

Figure 2

Rbm20-ASO treatment in a 2-hit model down-regulates Rbm20 expression and restores cardiac function. (A) Schematic of HFpEF experimental protocol. Adult male C57BL/6N mice underwent a 2-hit regimen to induce HFpEF conditions. Following this, Rbm20-ASOs were administered once a week for 6 weeks. Subsequently, these mice were employed in endpoint studies. (B) Western blot analysis of RBM20 with β-tubulin normalization revealed a down-regulation of RBM20 in the LVs following ASO treatment. (C) Titin isoform analysis indicated an up-regulation of compliant N2BA-N titins (represented by striped bars) in the ASO-treated groups. (D) Examples of mitral inflow pattern (upper panel) and mitral annular velocity (D, lower panel) from echocardiogram after 6 weeks of ASO/PBS treatment. 2-hit-PBS mice showed HFpEF-like conditions characterized by an increased E/A ratio (E), shortened E wave deceleration time (F), preserved LV ejection fraction (G), increased concentricity (H), elevated E/e′ ratio (I), and left atrial enlargement (J). Rbm20-ASO treatment in 2-hit mice normalized the E/A ratio (E), E deceleration time (F), and E/e′ ratio (I). The data are presented as means ± SD. * P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001, ****P ≤ 0.0001, ns indicates not statistically significant. The sample sizes of Ctrl-PBS, Ctrl-ASO, 2-hit-PBS, and 2-hit-ASO were 23, 29, 26, 27 mice for echocardiogram, 5, 4, 5, 5 mice for RBM20 western blot, and 10, 10, 10, and 10 mice for titin isoform analysis. Statistical analysis was performed using 2-way ANOVA with Tukey’s multiple comparisons. Two-way ANOVA analysis results are shown above each figure. Additional echocardiographic parameters are provided in Supplementary material online, Table S1.

Figure 3

PV analysis of 2-hit mice after Rbm20-ASO treatment. Representative PV analysis of Ctrl-PBS (A), 2-hit-PBS (B), and 2-hit-ASO (C) mice. The 2-hit-PBS mice had increased End Diastolic Pressure–Volume Relation (EDPVR (β); D), increased relaxation constant (E), elevated LV End-Diastolic Pressure (LVEDP; F), heightened LV End-Systolic Pressure (LV ESP; G), increased Effective Arterial Elastance (Ea; H), and preserved End-Systolic Pressure–Volume Relation (ESPVR; I). Treatment with Rbm20-ASOs normalized EDPVR (β) (D), relaxation constant (E), and LVEDP (F). (J, K) EDPVR and LVEDP are inversely correlated with the ratio of compliant titin isoforms (N2BA and N2BA-N titins). The data are presented as means ± SD. The sample size included n = 11, 11, 11, and 10 mice for Ctrl-PBS, Ctrl-ASO, 2-hit-PBS, and 2-hit-ASO, respectively. * P ≤ 0.05, ** P ≤ 0.01 ***, P ≤ 0.001, ****P ≤ 0.0001, ns indicates not statistically significant. The analysis employed Two-way ANOVA with Tukey’s multiple comparisons and Spearman’s rank for correlation analysis. Two-way ANOVA analysis results are shown above each figure. Additional PV analysis parameters are available in Supplementary material online, Table S2.

Figure 4

Cardiomyocyte contributions to LV diastolic stiffness. The intact cardiomyocyte stress-sarcomere length (SL) analysis is shown in (A-F). (A-C) shows examples of cellular work loops of intact cardiomyocytes from Ctrl-PBS (A), 2-hit-PBS (B), and 2-hit-ASO (C). Intact cardiomyocytes from 2-hit-PBS hearts demonstrate increased slopes of End-Diastolic Stress-Sarcomere Length Relation (ED-SSLR; D) and End-Systolic Stress-Sarcomere Length Relation (ES-SSLR; E). The ASO treatment normalized the ED-SSLR (D) and ES-SSLR (E) in the 2-hit mice. ASO treatment did increase stroke length (F; P < 0.05 for ASO effect). (G) The permeabilized cardiomyocytes’ passive stiffness is elevated in 2-hit-PBS and reduced in Ctrl-ASO and 2hit-ASO mice. (H) Maximal active stress is not different between groups. (I and J) Myocardial fibrosis was quantified by Picrosirius red staining of the LV section. The data are presented as means ± SD, except G, which showed mean ± SE, for simplicity. *** P ≤ 0.001, **** P ≤ 0.0001, ns indicates not statistically significant. For (D-F), the sample sizes were n = 7, 7, 8, and 9 mice for Ctrl-PBS, Ctrl-ASO, 2-hit-PBS, and 2-hit-ASO, respectively. Each data point represents the average value from one mouse, with 4–8 cells analysed per mouse. For (G and H), the sample sizes were n = 6, 7, 6, 6 mice, with 7–8 cells analysed per mouse. For (J), the sample sizes were n = 8, 8, 7, 7 mice. Two-way ANOVA with Tukey’s multiple comparisons was used for D–F, H, and J). Two-way ANOVA analysis results are shown above each figure. Non-linear regression analysis with a least squares fitting was performed for (G). Additional mechanical parameters for intact cardiomyocytes are provided in Supplementary material online, Table S4.

Figure 5

Rbm20-ASOs modify splicing of Camk2d, Ldb3, and titin. Transcriptomic analysis of LV myocardium shows differential gene splicing of Ctrl-PBS, Ctrl-ASO, 2-hit-PBS, and 2-hit-ASO after 6 weeks of ASO/PBS treatment. (A) Intersections of differentially spliced skipped exon (SE) events (FDR < 0.05, |dPSI| > 0.1) across the individual comparisons. (B, C) Percent spliced in (PSI) values for titin, Camk2d, and Ldb3. RNAseq, n = 4 mice per group.

Figure 6

Rbm20-ASOs partially revert the differential expression caused by the 2-hit intervention and interleukin and cytokine signalling. Transcriptomic analysis of LV myocardium shows differential gene expression of Ctrl-PBS, Ctrl-ASO, 2-hit-PBS, 2-hit-ASO after 6 weeks of ASO/PBS treatment. (A) Principal component analysis (PCA) illustrating distinct clustering by condition. (B) Intersection plot showing the overlap of differentially expressed genes (adjusted P-value < 0.05, |log2FC| > 0.5) across the individual comparisons. (C) Matrix of differentially expressed genes (adjusted P-value < 0.05, |log2FC| > 0.5). Number of differentially expressed genes per comparison is provided in each square. The greyscale reflects the gene count as indicated by the heat map. (D) Genes differentially expressed between 2-hit-PBS and Ctrl-PBS but not between Ctrl-PBS and 2-hit-ASO were normalized with reduced RBM20 activity (reverted genes). Genes differentially expressed between Ctrl- PBS and 2-hit-ASO but not between Ctrl-PBS and 2-hit-PBS were additionally regulated by reduced RBM20 levels (mis-regulated genes). (E) Gene enrichment analysis of the mis-regulated (Wikipathways), co-regulated (KEGG), and reverted genes (Wikipathways). (F) Expression levels of fibrosis-associated genes, Col3a1 and Loxl2, are increased upon ASO treatment. RNAseq, n = 4 mice per group.

Figure 7

Ca²⁺ release-reuptake analysis in intact cardiomyocytes after ASO treatment. Ca²⁺ transients were measured in unloaded intact cardiomyocytes using the Fura-2 340/380 ratio at a 4 Hz stimulation rate (A). Diastolic cytoplasmic Ca²⁺ levels are increased in the Ctrl-ASO group (B). However, no significant differences are observed among groups in terms of Ca²⁺ transient amplitude (C), time to 50% peak (D), time to peak (E), time to 50% transient decay (F), and time to 90% transient decay (G). The data are presented as means ± SD. * P ≤ 0.05, ns indicates not statistically significant. Two-way ANOVA with Tukey’s multiple comparisons. Two-way ANOVA analysis results are shown above each figure. Each data point represents an average from a single animal, with 40–50 cells analysed per mouse. The study involved n = 7, 7, 8, and 10 mice for Ctrl-PBS, Ctrl-ASO, 2-hit-PBS, and 2-hit-ASO, respectively.