Loss of conserved long non-coding RNA MIR503HG leads to altered NOTCH pathway signalling and left ventricular non-compaction cardiomyopathy

Abstract

Aims: The highly conserved long non-coding RNA (lncRNA) MIR505HG has been primarily recognized as a precursor for microRNAs (miR)-424 and miR-503. However, studies have since demonstrated that MIR503HG has distinct functions from its associated miRNAs, playing important roles in cell proliferation, invasion, apoptosis, and differentiation. While these miRNAs are known to influence cardiomyocyte differentiation, the specific role of MIR503HG in heart development remains unexplored. We seek to determine how MIR503HG deletion impacts ventricular chamber development and to identify underlying molecular mechanisms.

Methods and results: To study the role of the lncRNA in vivo, we generated a functional MIR503HG knockout mouse model (MIR503HG-/-) using a synthetic polyadenylation signal to terminate MIR503HG transcription without affecting miR-424/503 expression. We performed morphological analyses on embryonic and adult hearts using microCT along with cardiac functional analysis via transthoracic echocardiography. We further apply single-nuclei RNA sequencing (snRNA-seq) on adult hearts to identify potential molecular mechanisms underlying the observed phenotypes. Functional deletion of MIR503HG alone was associated with reduced compact myocardium thickness and increased trabecular myocardium in the left ventricle (LV) at embryonic day 17.5 compared to wild-type mice, indicating a LV non-compaction (LVNC) phenotype. Moreover, adult MIR503HG-/- mutant hearts showed increased trabecular complexity, impaired LV relaxation, and mitral valve regurgitation. SnRNA-seq further revealed altered expression of several genes associated with cardiomyocyte function and LVNC, including Actc1, Mib1, Mybpc3, and Myh7. Lastly, Notch1 activity was also significantly increased in mutant hearts which has been previously associated with LVNC.

Conclusion: MIR503HG plays a role in ventricular chamber development, and its deletion leads to an LVNC phenotype independent of the miRNA cluster within its locus, highlighting its importance in cardiac development and disease. We further suggest that abnormal Notch1 activity may underpin the LVNC phenotype presented.

Keywords: Heart development; Left ventricular non-compaction; Long non-coding RNA.

Figure 1

(A) CRISPR/Cas9 strategy to generate the MIR503HG null allele (MIR503HG^−/−) in mice. To prevent unintended alterations in the gene structure, typical of large deletions, we inserted a SPA signal cassette (pAS-MAZ). This terminates MIR503HG transcription while preserving upstream miRNA precursor expression. The SPA cassette was inserted into Exon 1 of the MIR503HG gene; this includes a compact and highly efficient SPA followed by two MAZ protein binding sites, which further enhances termination of transcription. This approach ensures specific disruption of MIR503HG lncRNA without affecting miR-322(424)/503 expression. (B–B″) Expression of MIR503HG (B), miR-503 (B′), and miR-424 (B″) in heart tissue from 8-week-old MIR503HG^−/− mice compared to WT littermates was assessed by qPCR. Relative quantification value for gene expression was determined following normalization to the levels of 18S (n = 8 mice/group; two-tailed Student’s t-test). (C–F″) Morphological (C–D″) and proliferation (E–F″) analysis. Sections underwent immunofluorescence staining for myocardial marker MF20 (1:500, Thermo Fisher, magenta in C–D″), endocardial marker CD31 (1:200, DIANOVA, cyan in C–D″), proliferation marker Ki67 (1:200, Abcam, green in E–F″), and nuclear marker Hoechst (1:1000, Sigma, blue in C–D″, red in E–F″). Sections show a general view of E17.5 hearts (C–F) and a higher magnification view of the left (LV, C′–F′) and right (RV, C″–F″) ventricles from WT (C–C″ and E–E″) and MIR503HG^−/− mutants (D–D″ and F–F″). Arrows point to the compact layer outlined with a white contour; arrowheads point to the trabecular myocardium. (G–H) Quantitative analysis of LV (G) and RV (H) compact myocardium area (i), trabecular myocardium area (ii), compact myocardium cells (iii), trabecular myocardium cells (iv), compact myocardium proliferation (v), and trabecular myocardium proliferation (vi) in MIR503HG^−/− mutants compared with WT (n = 4 embryos/group). (I–J′) 3D reconstructions from WT (I) and MIR503HG^−/− (J) adult hearts generated in Amira software (Thermo Fisher) from microCT datasets where right (blue) and left (green) auricles, RV (cyan) and LV (orange), interventricular septum (purple), aorta, pulmonary artery, and tricuspid, bicuspid, and semilunar valves were segmented for morphological analysis and volumetric quantifications. (I′ and J′) 3D reconstructions from WT (I′) and MIR503HG^−/− (J′) LV showing the trabecular myocardium and papillary muscles from the luminal side of the LV (arrows). (I″ and J″) MicroCT optical transverse sections in the middle region of the LV showing the luminal perimeter used to perform fractal analysis for trabecular complexity. Arrows point to trabecular myocardium. (K) Volumetric quantifications of the entire heart (i) and LV (ii) and fractal analysis of LV (iii) (n = 3 mice/group). (L) Cardiovascular morphology and functional analysis by transthoracic echocardiography on 13-week-old MIR503HG^−/− mutants compared with WT showing quantifications for LV isovolumetric relaxation time (IVRT, i), LV anterior wall thickness (LVAW, ii), mitral valve regurgitation (MV E, iii), Futon index (iv), and ascending aorta (v) and pulmonary vein (vi) diameter. In addition to the structural adaptations described, we identified a decrease in diameter for the ascending aorta and pulmonary vein (v, vi), which may be the result of further changes in cardiovascular development or physiology in mutants (n > 8 mice/group; two-tailed Student’s t-test). (M) Uniform manifold approximation and projection of snRNA-seq) data identifying 14 different clusters at 0.17 clustering resolution, with respective biological cluster identities as defined by canonical marker genes. (M′) Analogous uniform manifold approximation and projection displaying both MIR503HG^−/− (red) and WT (green) cell identities suggests no major changes in the cell clusters identified in MIR503HG^−/− hearts compared with WT. (M″) Dot plot representation of marker genes for each cluster represented by average expression (colour) and percentage of positive cells (size). (N and N′) Violin plots showing log normalized expression of Actc1, Myh7, Mib1, and Mybpc3 in Cardiomyocyte1 cluster. All nuclei used for snRNA-seq analysis were isolated from snap frozen 13-week-old WT and MIR503HG^−/− mutant hearts (n = 1 pooling three mice/group). (O and O′) Validation of snRNA-seq results for Actc1, Myh7, and Mib1 expression using qPCR on RNA isolated from snap frozen 13-week-old MIR503HG^−/− and WT whole hearts. Relative quantification values for gene expression were quantified by qPCR assay relative to Gapdh (n > 8 mice/group; two-tailed Student’s t-test). (P and P′) GO biological process enrichment analysis of up-regulated and down-regulated DEGs in the Cardiomyocyte1 cluster. (Q–R′ and T–U′) Molecular analysis by immunofluorescence of myocardial marker MF20 (1:500, Thermo Fisher, blue), PTBP1 (Q–R′; 1:200, Abcam, green), or N1ICD (T–U′; 1:100, Cell Signaling, green) and nuclear marker Hoechst (1:1000, Sigma, red) in a mid-ventricular view of a E17.5 heart (Q, R, T, and U) and a higher magnification view of LV (Q′, R′, T′, and U′) from WT (Q, Q′, T, and T′) and MIR503HG^−/− mutants (R, R′, U, and U′). (Q′ and R′) Arrows point to PTBP1-positive endocardial cells. (T′ and U′) Arrows point to N1ICD-positive endocardial cells; arrowhead points to N1ICD negative cells. (S and V) Quantitative analysis of PTBP1 (S) and N1ICD (V) expression in the entire endocardium (i) and after segmentation of the endocardium into endocardial base (ii), middle (iii), and apex (iv) (n = 3 embryos/group). Data in graphs in B′, B″, G, H, K, L, O, S, and V represented as mean ±SEM. Datasets were analysed for normal distribution using Shapiro–Wilk test. Normally distributed datasets were analysed using t-test whereas non-normally distributed datasets were analysed using Mann–Whitney test. Statistical significance was considered when P<0.05. Significant P-values indicated in the graphs; ns, not significant. (W–Z′) Molecular analysis by immunofluorescence of myocardial marker MF20 (1:500, Thermo Fisher, blue), Dll4 (W–X′; 1:200, R&D Systems, yellow), or Jagged1 (Y–Z′; 1:100, Cell Signalling, yellow) and nuclear marker Hoechst (1:1000, Sigma, red) in a mid-ventricular view of a E17.5 heart (W–Z) and a higher magnification view of LV (W′–Z′) from WT (W, W′, Y, and Y′) and MIR503HG^−/− mutants (X, X′, Z, and Z′). Arrows point to Dll4 or Jagged1 positive cells (n = 3 embryos/group). Scale bars: 200 µm in Q-R, T-U and W-Z; 100 µm in C–F; 50 µm in C′–F″, Q′–R′, T′–U′, and W′–Z′; 2 mm in I–J; 1 mm in I′–J′, and I″–J″.