LncRNA- Safe contributes to cardiac fibrosis through Safe- Sfrp2-HuR complex in mouse myocardial infarction

Abstract

Rationale: As a hallmark of various heart diseases, cardiac fibrosis ultimately leads to end-stage heart failure. Anti-fibrosis is a potential therapeutic strategy for heart failure. Long noncoding RNAs (lncRNAs) have emerged as critical regulators of heart diseases that promise to serve as therapeutic targets. However, few lncRNAs have been directly implicated in cardiac fibrosis. Methods: The lncRNA expression profiles were assessed by microarray in cardiac fibrotic and remote ventricular tissues in mice with myocardial infarction. The mechanisms and functional significance of lncRNA-AK137033 in cardiac fibrosis were further investigated with both in vitro and in vivo models. Results: We identified 389 differentially expressed lncRNAs in cardiac fibrotic and remote ventricular tissues in mice with myocardial infarction. Among them, a lncRNA (AK137033) we named Safe was enriched in the nuclei of fibroblasts, and elevated in both myocardial infarction and TGF-β-induced cardiac fibrosis. Knockdown of Safe prevented TGF-β-induced fibroblast-myofibroblast transition, aberrant cell proliferation and secretion of extracellular matrix proteins in vitro, and mended the impaired cardiac function in mice suffering myocardial infarction. In vitro studies indicated that knockdown of Safe significantly inhibited the expression of its neighboring gene Sfrp2, and vice versa. The Sfrp2 overexpression obviously disturbed the regulatory effects of Safe shRNAs in both the in vitro cultured cardiac fibroblasts and myocardial infarction-induced fibrosis. Dual-Luciferase assay demonstrated that Safe and Sfrp2 mRNA stabilized each other via their complementary binding at the 3'-end. RNA electrophoretic mobility shift assay and RNA immunoprecipitation assay indicated that RNA binding protein HuR could bind to Safe-Sfrp2 RNA duplex, whereas the knockdown of HuR dramatically reduced the stabilization of Safe and Sfrp2 mRNAs, down-regulated their expression in cardiac fibroblasts, and thus inhibited TGF-β-induced fibrosis. The Safe overexpression partially restrained the phenotype change of cardiac fibroblasts induced by Sfrp2 shRNAs, but not that induced by HuR shRNAs. Conclusions: Our study identifies Safe as a critical regulator of cardiac fibrosis, and demonstrates Safe-Sfrp2-HuR complex-mediated Sfrp2 mRNA stability is the underlying mechanism of Safe-regulated cardiac fibrosis. Fibroblast-enriched Safe could represent a novel target for anti-fibrotic therapy in heart diseases.

Keywords: cardiac remodeling; fibrosis; non-coding RNA.

Figure 1

Safe is a fibroblast-enriched lncRNA and associated with cardiac fibrosis. (A) Heatmap showing 389 lncRNAs differentially expressed in tissues from the infarction zone (IZ) and non-infarcted remote zone (RZ) at day 14 post MI (fold ≥ 2, p<0.05). (B) qRT-PCR validation of four upregulated lncRNAs and four downregulated lncRNAs in cardiac samples from IZ and RZ of infarcted hearts (n=3). (C) qRT-PCR analysis of Safe expression in healthy (sham) or infarcted myocardium at indicated days post MI (n=3). (D) qRT-PCR detection of Safe expression in adult cardiomyocytes (aCM), adult cardiac fibroblasts (aCF) and adult endothelial cells (aEC). Myh7 (myosin heavy chain 7), FSp1 (fibroblast specific protein-1) and Pecam1 (Platelet endothelial cell adhesion molecule-1) were used as markers of aCM, aCF and aEC, respectively (n=3). (E) qRT-PCR detection of Safe, Col1a1 and α-SMA expression in TGF-β-treated cardiac fibroblasts (n=3). (F) Representative western blot analysis and relative densitometric quantification of COL1A1 and α-SMA protein levels in cardiac fibroblasts with TGF-β treatment (n=3). (G) Agarose gel electrophoresis of the 5' and 3' RACE amplification products. (H) Schematic presentation of full-length Safe showing the extended regions identified by RACE (blue). The full-length of Safe was 1517 base pairs with two exon regions, which was different from the sequence information provided by UCSC mm10 showing 1048 base pairs in length and three exon regions. (I) In vitro translation of Safe sense or antisense transcript. Luciferase (Luc) was used as a positive control. (J) Subcellular localization of Safe in cytoplasm and nucleus of cardiac fibroblasts (n=3). The gene U6 and Gapdh were respectively used as nuclear and cytoplasmic RNA markers. (K) Representative images of RNA FISH showing nuclear localization of Safe (green) in cardiac fibroblasts. The nucleuses were counterstained with DAPI (blue). Scale bar indicates 20 μm. Data are presented as mean ± SEM and analyzed using Student's t-test or one-way ANOVA; *p < 0.05, and ns, not significant.

Figure 2

Suppression of Safe prevents TGF-β-induced cardiac fibrosis in vitro. (A) mRNA expression of Safe, Col1a1, α-SMA and Bmp1 in TGF-β-untreated or TGF-β-treated cardiac fibroblasts after Safe knockdown (n=3). (B) Western blot analysis and relative densitometric quantification of COL1A1 and α-SMA protein levels in TGF-β-untreated or TGF-β-treated cardiac fibroblasts after Safe knockdown (n=3). (C) Representative images of immunofluorescence staining for α-SMA (n=5) and quantification of the relative cell areas of cardiac fibroblasts after treatments as indicated (n=40). The nuclei were counterstained with DAPI (blue). Scale bar indicates 50 μm. (D) Representative images of collagen gel contraction for 24 hours and quantification of collagen area inside the dashed circles (n=3). Scale bar indicates 0.5 cm. (E) BMP1 protein enzyme activity in the supernatant of cultured fibroblasts after indicated treatments (n=3). The excitation wavelength is 320 nm, and the emission wavelength is 405 nm. (F) ELISA assay of COL1A1 protein in the supernatant of cultured fibroblasts after indicated treatments (n=3). (G) CCK-8 assay of cardiac fibroblasts with TGF-β-untreated or TGF-β-treated showing repressed cell proliferation by Safe knockdown. The cell proliferation rate was expressed as optical density value at 450 nm (OD₄₅₀) wavelength (n=3). (H) Flow cytometry analysis showing decreased ratios of pH3-positive fibroblasts and myofibroblasts in the group of Safe knockdown (n=3). (I) Representative images of immunofluorescence staining for mitosis marker 5-ethynyl-2'-deoxyuridine (EdU, magenta) and DAPI (blue), PDGFR-α (red) was stained as a marker of fibroblasts. Scale bar indicates 50 μm. Right panel: Percent of EdU⁺ cells in PDGFR-α⁺ cells (n=25). Data are presented as mean ± SEM; Student's t-test or one-way ANOVA; *p < 0.05.

Figure 3

Inhibition of Safe ameliorates MI-induced cardiac fibrosis. (A) Representative M-mode echocardiographic images obtained from MI mice injected with scramble control shRNA (Ctr) or shSafe lentiviral particles at day 28 after surgery (n=10 in each group). (B) Left ventricular ejection fraction (EF) of MI mice as assessed via echocardiography in indicated time points (n=10). (C) Left ventricular fractional shortening (FS) of MI hearts in indicated time points (n=10). (D) Schematic diagram of the slice, starting from left ventricular exposure, each level is 500 μm apart. Representative images of Masson's trichrome-stained MI hearts, and quantification of fibrosis (% area) showing a significant decrease of fibrosis areas in MI hearts injected with shSafe lentiviral particles (n=8). Scale bar indicates 2 mm. (E) qRT-PCR analysis of Col1a1 and α-SMA in the infarction zone of MI hearts (n=3). (F) Western blot analysis and relative densitometric quantification of COL1A1 and α-SMA protein levels in MI hearts after Safe inhibition (n=3). Data are presented as mean ± SEM; Student's t-test or two-way repeated-measures ANOVA; *p < 0.05.

Figure 4

Safe and Sfrp2 mutually regulates each other's RNA stability in fibroblasts. (A) Schematic diagram of RNA-RNA interaction between Safe and Sfrp2 at their reverse complementary regions (462 nucleotides in length). Thick blocks and thin blocks represent the coding sequence (CDS) and untranslated regions (UTRs), respectively. (B) qRT-PCR detection of Sfrp2 expression in left ventricles of sham and MI mice at indicated days after surgery (n=3). (C) qRT-PCR detection of Sfrp2 expression in the infarct zone (IZ), border zone (BZ), and remote zone (RZ) of infarcted hearts after 14 days following MI surgery (n=3). (D) qRT-PCR detection of Sfrp2 expression in aCM, aCF and aEC (n=3). (E) qRT-PCR analysis showing increased expression of Sfrp2 in cardiac fibroblasts after TGF-β treatment (n=3). (F) mRNA expression of Sfrp2, Col1a1, α-SMA and Bmp1 in TGF-β-untreated or TGF-β-treated cardiac fibroblasts after Sfrp2 knockdown (n=3). (G) Representative western blot analysis and relative densitometric quantification of SFRP2, COL1A1 and α-SMA protein levels in TGF-β-untreated or TGF-β-treated cardiac fibroblasts after Sfrp2 knockdown (n=3). (H) qRT-PCR analysis showing decreased expression of both Safe and Sfrp2 mRNA in cardiac fibroblasts transfected with shSafe (n=3). (I) Representative western blot analysis and relative densitometric quantification of SFRP2 protein with Safe inhibition (n=)3. (J) qRT-PCR analysis showing decreased expression of both Sfrp2 mRNA and Safe in cardiac fibroblasts transfected with shSfrp2 (n=3). (K) Subcellular localization of Sfrp2 in cytoplasm and nuclei of cardiac fibroblasts (n=3). (L) qRT-PCR analysis showing preferentially inhibition of Sfrp2 mRNA level in the nucleus of cardiac fibroblasts after Safe knockdown (n=3). (M) Representative images of RNA FISH showing co-localization of Safe (green), Sfrp2 mRNA (magenta) in the nuclei (DAPI, blue) of cardiac fibroblasts. Scale bar: 20 μm. (N) Dual luciferase assay showing shSafe inhibited Firefly luciferase activities of pGL3-control vectors carrying Sfrp2 3'-UTR (n=3). (O) Dual luciferase assay showing shSfrp2 inhibited Firefly luciferase activities of pGL3-control vectors carrying Safe RNA fragments complementary to Sfrp2 mRNAs (n=3). All data are presented as mean ± SEM; Student's t-test or one-way ANOVA; *p < 0.05, and ns, not significant.

Figure 5

Sfrp2 overexpression disturbed the regulatory effects of shSafe in cardiac fibroblasts. (A) qRT-PCR analysis showing mRNA expression of Safe, Sfrp2, Col1a1, α-SMA and Bmp1 by Sfrp2 overexpression in Safe-silenced cardiac fibroblasts (n=3). (B) Representative western blot analysis and relative densitometric quantification showing up-regulated expression of SFRP2, COL1A1, and α-SMA proteins by Sfrp2 overexpression in Safe-silenced cardiac fibroblasts (n=3). (C) Representative images of immunofluorescence staining for α-SMA (red, n=5) and quantification of the relative cell area of cardiac fibroblasts after treatments as indicated (n=40). Scale bar indicates 50 μm. (D) Representative images of collagen gel contraction for 24 hours and quantification of collagen area inside the dashed circles (n=3). Scale bar indicates 0.5 cm. (E) BMP1 protein enzyme activity in the supernatant of cultured fibroblasts after indicated treatments (n=3). The excitation wavelength is 320 nm, and the emission wavelength is 405 nm. (F) ELISA assay of COL1A1 protein in the supernatant of cultured fibroblasts after indicated treatments (n=3). (G) CCK-8 assay of cardiac fibroblasts showing restored cell proliferation by Sfrp2 overexpression in Safe-deficient cardiac fibroblasts (n=3). (H) Flow cytometry analysis of pH3 incorporation in Sfrp2 overexpressed-stimulated fibroblasts with indicated treatments. (I) Representative images of immunofluorescence staining for EdU (magenta), PDGFR-α (red) and DAPI (blue). Scale bar indicates 50 μm. Right panel: Percent of EdU⁺ cells in PDGFR-α⁺ cells (n=25). All data are presented as mean ± SEM; Student's t-test or one-way ANOVA; *p < 0.05.

Figure 6

Sfrp2 overexpression disturbed the protective effect of shSafe on cardiac function post MI. (A) Representative M-mode echocardiographic images obtained from MI mice after Safe knockdown, in combination with Sfrp2 overexpression (n=10 in each group). (B) Sfrp2 overexpression led to a reduced improvement of EF in shSafe-injected MI mice at day 28 post-surgery (n=10). (C) shSafe-mediated increase in FS in MI hearts was inhibited by Sfrp2 overexpression (n=10). (D) Representative images of Masson's trichrome-stained MI hearts, and quantification of relative fibrosis areas (n=8). Scale bar indicates 2 mm. (E) qRT-PCR analysis showing increased expression of Safe, Sfrp2, Col1a1 and α-SMA in left ventricles of shSafe-injected MI hearts after Sfrp2 overexpression (n=3). (F) Western blot analysis and relative densitometric quantification of SFRP2, COL1A1 and α-SMA protein levels in shSafe-injected MI hearts after Sfrp2 overexpression (n=3). All data are presented as mean ± SEM; Student's t-test, one-way ANOVA or two-way repeated-measures ANOVA, *p < 0.05.

Figure 7

Binding of HuR to the Safe-Sfrp2 duplex accelerates RNA stabilization of both Safe and Sfrp2. (A) Schematic presentation of predicated HuR binding sites and corresponding RNA probes for EMSA assay in the complementary region of Safe and Sfrp2 RNA. (B) EMSA assay indicating specific binding of nuclear proteins of cardiac fibroblasts to the 26-nucleotide RNA duplexes corresponding to nucleotide at the 1408 nt protein binding site of Safe. (C) EMSA supershift assay revealing the in vitro interaction between HuR protein and the 26-nucleotide RNA duplexes in nuclear extracts. RNA immunoprecipitation (RIP) assay using HuR antibody and IgG (isotype control) showing enrichment of Safe (D) and Sfrp2 (E) RNAs to HuR protein. (F) qRT-PCR analysis showing decreased expression of HuR, Safe and Sfrp2 in cardiac fibroblasts after HuR knockdown (n=3). (G) Representative western blot analysis and relative densitometric quantification of HuR and SRRP2 protein levels in cardiac fibroblasts with or without HuR inhibition (n=3). (H) Dual luciferase assay showing shHuR inhibited Firefly luciferase activities of pGL3-control vectors carrying the 3'-end of Safe (462-nucleotide in length) (n=3). (I) Dual luciferase assay showing shHuR inhibited Firefly luciferase activities of pGL3-control vectors carrying Sfrp2 3'-UTR (n=3). (J) qRT-PCR detection of HuR, Col1a1, α-SMA and Bmp1 in TGF-β-untreated and TGF-β-treated cardiac fibroblasts after HuR knockdown (n=3). (K) Representative western blot analysis and relative densitometric quantification of HuR, COL1A1 and α-SMA protein levels in sh-Scr or HuR-silenced cardiac fibroblasts after TGF-β treatment (n=3). All data are presented as mean ± SEM; Student's t-test or one-way ANOVA, *p < 0.05, and ns, not significant.

Figure 8

Mechanism of lncRNA-Safe-mediated cardiac fibrosis. Pathological stimuli such as TGF-β and myocardial infarction facilitate the expression of lncRNA-Safe in cardiac fibroblasts. In cooperation with HuR protein, Safe can complementarily binds to the 3'-UTR of Sfrp2 mRNA, and promote Sfrp2 RNA stabilization and protein expression, which in turn accelerates the transformation and proliferation of myofibroblasts, eventually leading to excessive secretion of extracellular matrix proteins.