Research ArticleCARDIAC HYPERTROPHY

Long noncoding RNA Chast promotes cardiac remodeling

See allHide authors and affiliations

Science Translational Medicine  17 Feb 2016:
Vol. 8, Issue 326, pp. 326ra22
DOI: 10.1126/scitranslmed.aaf1475

The missing lnc in cardiac hypertrophy

RNA that does not code for a protein comprises a large portion of the human genome. These so-called noncoding RNAs are emerging as important players in disease pathogenesis, yet their functional roles are not always well known. Viereck et al. have discovered a new long noncoding RNA (lncRNA) that promotes cardiac remodeling and hypertrophy in mice, which could one day be targeted with therapeutics to treat human cardiovascular diseases. The identified lncRNA, which the authors named Chast (for “cardiac hypertrophy–associated transcript”), was discovered to be up-regulated in hypertrophic mouse hearts. When mouse and human heart cells expressed Chast, they tended to be larger than their normal counterparts. By silencing Chast with antisense oligonucleotides, mice either did not develop hypertrophy or were rescued from established disease. In a step toward translation, the authors discovered a human homolog, CHAST, that similarly caused cells in a dish to enlarge. Additional investigation in patients will confirm the relevance of this lncRNA in human disease and whether it is indeed a promising target for treating cardiac hypertrophy and heart failure.

Abstract

Recent studies highlighted long noncoding RNAs (lncRNAs) to play an important role in cardiac development. However, understanding of lncRNAs in cardiac diseases is still limited. Global lncRNA expression profiling indicated that several lncRNA transcripts are deregulated during pressure overload–induced cardiac hypertrophy in mice. Using stringent selection criteria, we identified Chast (cardiac hypertrophy–associated transcript) as a potential lncRNA candidate that influences cardiomyocyte hypertrophy. Cell fractionation experiments indicated that Chast is specifically up-regulated in cardiomyocytes in vivo in transverse aortic constriction (TAC)–operated mice. In accordance, CHAST homolog in humans was significantly up-regulated in hypertrophic heart tissue from aortic stenosis patients and in human embryonic stem cell–derived cardiomyocytes upon hypertrophic stimuli. Viral-based overexpression of Chast was sufficient to induce cardiomyocyte hypertrophy in vitro and in vivo. GapmeR-mediated silencing of Chast both prevented and attenuated TAC-induced pathological cardiac remodeling with no early signs on toxicological side effects. Mechanistically, Chast negatively regulated Pleckstrin homology domain–containing protein family M member 1 (opposite strand of Chast), impeding cardiomyocyte autophagy and driving hypertrophy. These results indicate that Chast can be a potential target to prevent cardiac remodeling and highlight a general role of lncRNAs in heart diseases.

INTRODUCTION

Diseases of the cardiovascular system are the most common cause of human morbidity and mortality in the world. Exposure of the heart to different stressors leads to “cardiac remodeling” with detrimental outcomes. At the cellular level, a hallmark of the remodeling process is pathological cardiomyocyte hypertrophy. Prolonged hypertrophic growth of the heart results in heart failure and often death of the affected individual (1). Despite advances in pharmacologic treatment of cardiac remodeling by β blockers, angiotensin-converting enzyme inhibitors, and mineralocorticoid receptor blockers, the prognosis of affected individuals remains poor.

The human genome is tremendously larger than that of more simple organisms such as worms, although it has a comparable number of protein-regulating genes. All protein-coding genes originate from about only 1.5% of the genome, and the larger remaining portion of the genome remains either untranscribed or (more often) transcribed to noncoding RNAs (2), of which functional importance is still less understood. Long noncoding RNAs (lncRNAs) are a class of noncoding RNAs that are larger than 200 nucleotides (nt) and do not generally encode proteins. LncRNAs regulate pathways by diverse mechanisms including chromatin modification (3), transcriptional (4) and posttranscriptional regulation (5), and influencing subcellular trafficking (6, 7). The role of lncRNAs in the heart has just begun to be uncovered. LncRNA Braveheart (Bvht) was recently shown to be crucial for cardiac development in mice (8), whereas another lncRNA, Fendrr, was able to control chromatin modifications and, thus, developmental signaling in the rodent heart (9). LncRNAs have demonstrated functional roles in the adult murine cardiac system and in human cardiovascular cells (1012), and have strong prognostic potential as biomarkers in heart failure patients (13).

Thus, here, we generated a whole-genome lncRNA profile to identify differentially expressed lncRNAs in hypertrophic mouse hearts. Subsequent functional studies revealed that lncRNA ENSMUST00000130556, which we have named cardiac hypertrophy–associated transcript (Chast), plays an important role in the pathogenesis of pressure overload–induced cardiac remodeling. Mechanistically, Chast impedes expression of the autophagy regulator Pleckstrin homology domain–containing protein family M member 1 (Plekhm1), and its silencing results in autophagic inhibition and cardiomyocyte hypertrophy. Suppression of Chast before and after induction of hypertrophy prevented or attenuated established disease, respectively, indicating the potential for lncRNAs as new therapeutic targets in cardiovascular diseases.

RESULTS

LncRNA expression pattern is altered during cardiac pressure overload in mice

To investigate lncRNA transcriptome changes during pressure overload–induced cardiac hypertrophy, we performed a lncRNA microarray analysis in RNA isolated from sham or transverse aortic constriction (TAC)–operated mouse hearts (Fig. 1A). Most lncRNAs represented on the array (19,427 of 31,423 lncRNAs) were detected, and several lncRNAs were deregulated in hypertrophic hearts (Fig. 1B): 1237 lncRNAs were up-regulated after TAC, whereas 1623 lncRNAs were down-regulated. To avoid detection of primary mRNA in quantitative polymerase chain reaction (qPCR), we excluded lncRNA transcripts that were located within a protein-coding gene in sense orientation. Next, we selected 22 highly deregulated (11 up-regulated, 11 down-regulated with fold changes >3) candidates for validation. Of these 22 candidates, transcripts were consistently detected in all tested samples (n = 5 mice per group) for 13 lncRNAs (six up-regulated and seven down-regulated). qPCR analyses indicated that four lncRNAs were significantly deregulated (one up-regulated and three down-regulated) in TAC-operated hearts (Fig. 1C).

Fig. 1. Chast is induced by pressure overload in mice.

(A) Selection strategy of TAC-induced murine lncRNAs derived from microarray-based profiling. (B) MA plot (M, log ratio; A, average mean) of differentially regulated lncRNAs [fold change (FC) > 1.5 indicated by symbols, signs, and dashed line; Chast is highlighted in red]. (C) Validation of deregulated lncRNAs induced by TAC. Chast is highlighted in red. Data are mean FC relative to sham ± SEM (n = 6 animals per group). *P < 0.05; **P < 0.01; ***P = 0.001 versus respective sham, Student’s t test. (D) Time course of Chast expression compared to heart weight–to–tibia length ratio and ejection fraction (EF), each relative to sham controls. Data are means ± SEM (n = 5 to 8 animals). HW, heart weight. (E) Chast expression in heart cell fractions after TAC (n = 6). Data are mean FC relative to sham ± SEM. P values versus sham determined by Student’s t test. (F) Graphical representation of 3′ and 5′ rapid amplification of cDNA (complementary DNA) ends (RACE).

From the three down-regulated, two showed protein-coding potential (ENSMUST00000019980 and ENSMUST00000046048, but not AK083183). We nevertheless focused on the up-regulated lncRNAs, specifically ENSMUST00000130556 (Fig. 1, B and C), because of our interest in developing inhibitory oligonucleotide drugs as a therapeutic approach. Herein, we refer to this lncRNA as Chast. Expression of Chast was temporally regulated between 4 and 13 weeks after left ventricular (LV) pressure overload in mice, with a peak expression level at 6 weeks (Fig. 1D). At this time point (4 to 6 weeks after initiation of TAC-mediated LV pressure overload), compensated cardiac function turns toward decompensation, for example, normal cardiac function started to decrease (Fig. 1D). Cellular fractionation experiments revealed that Chast was specifically up-regulated in cardiomyocytes (rather than in fibroblasts) of 6-week TAC-operated murine hearts (Fig. 1E).

We performed 3′ and 5′ RACE experiments in RNA isolated from mouse hearts and confirmed that Chast is expressed in heart tissue as a single, full-length, 823-nt transcript with two exons (Fig. 1F). This validated sequence of Chast strongly correlates to the sequence information of transcript ENSMUST00000130556 annotated by Ensembl (depicted in red) but harbors a transcription start site located 100 nt downstream of the annotated transcript (Fig. 1F, depicted in blue). Furthermore, the Chast gene partially overlaps in antisense to the protein coding genes Rho GTPase (guanosine triphosphatase) activating protein 27 (Ahrgap27) and Plekhm1.

To confirm that Chast is a noncoding RNA, the protein-coding potential of the full-length transcript was analyzed using the Protein Coding Potential Calculator (14). Similar to other bona fide murine lncRNAs Xist, Bvht, and Fendrr, Chast received low coding-potential scores. Known protein-coding genes Gapdh, β-actin (Actb), α-MHC (Myh6), and troponin T (Tnnt2) served as positive controls (fig. S1A). Further, we applied the sequence frame translator and the NCBI (National Center for Biotechnology Information) open reading frame (ORF) finder and identified four open reading frames that were small (<100 amino acids) (fig. S1B). To rule out the possibility that Chast encodes one of these peptides, we overexpressed the transcript in murine HL-1 atrial cardiomyocytes and analyzed the resulting peptides by high-resolution mass spectrometry. None of the peptides corresponding to potential ORFs were identified (fig. S1B).

Chast induces cardiomyocyte hypertrophy

Expression of Chast was induced upon stimulation of HL-1 cardiomyocytes with the prohypertrophic factors phenylephrine (PE) and isoproterenol (ISO) (Fig. 2A). Lentiviral overexpression of Chast induced hypertrophic growth in HL-1 cells (Fig. 2, B and C). Silencing of Chast with GapmeRs—antisense oligonucleotides containing a central DNA complementary to its target and flanked by blocks of locked nucleic acids, which forms a stable DNA/RNA heteroduplex that provoke ribonuclease H (RNase H)–mediated target degradation—significantly attenuated PE + ISO–induced hypertrophic response in HL-1 cells (Fig. 2, E and F). Consistently, expression of the cardiac stress marker atrial natriuretic peptide (Anp) was increased either by PE + ISO treatment or upon Chast overexpression, whereas GapmeR-mediated suppression of Chast prevented PE + ISO–induced Anp expression (Fig. 2, D and G). These results indicate that Chast is involved in prohypertrophic growth of cultured cardiomyocytes.

Fig. 2. Chast induces cardiomyocyte hypertrophy in vitro.

(A) Expression levels of Chast after stimulation of murine HL-1 cardiomyocytes with ISO and PE (100 μM each, 48 hours). Data are means ± SEM (n ≥ 3 independent experiments). P value determined by one-way analysis of variance (ANOVA) with Bonferroni correction. (B to D) Cardiomyocyte size and cardiac stress marker Anp expression with lentivral overexpression (pLV+) of Chast or empty control before and after prohypertrophic (PE + ISO) treatment. (E to G) Cardiomyocyte size and Anp expression upon GapmeR-mediated silencing of Chast before and after prohypertrophic (PE + ISO) treatment. GapmeR-control was a scrambled sequence. In (C), (D), (F), and (G), data are mean FC relative to control ± SEM (n ≥ 3 independent experiments). *P < 0.05; **P < 0.01; ***P < 0.001, as indicated, one-way ANOVA with Bonferroni correction.

Chast is activated by NFAT signaling and acts as a regulator of hypertrophic genes and pathways

Bioinformatic analysis revealed that the gene encoding Chast has at least one binding site for the prohypertrophic transcription factor NFAT (nuclear factor of activated T cells) (Fig. 3A and table S1). To test if the expression of Chast is regulated by NFAT, we treated murine HL-1 cardiomyocytes with the pharmacological NFAT inhibitor 11R-VIVIT. Upon NFAT inhibition, expression of Chast, along with another known NFAT target Mcip1.4, was significantly down-regulated (Fig. 3B). To determine whether Chast is regulated by NFAT activation, we tested Chast expression in hearts of calcineurin (Myh6-CnA) transgenic (TG) mice (15), where NFAT is constitutively active. Calcineurin TG mice displayed high Mcip1.4 levels and an approximately threefold increase in Chast expression compared with wild-type animals (Fig. 3C). Binding of NFAT to the Chast promoter was verified by chromatin immunoprecipitation (ChIP) assays (Fig. 3D).

Fig. 3. Regulation of Chast and Chast-mediated regulation of Plekhm1.

(A) Transcription factor binding site prediction within the genomic upstream region of the Chast lncRNA. ATF3, activating transcription factor 3; RXRA, retinoid × receptor α; PPARA, peroxisome proliferator-activated receptor α; NF-κB, nuclear factor κ-light-chain-enhancer of activated B cells. (B) Chast and Mcip1.4 gene expression 2 days after treatment of mouse HL-1 cardiac muscle cells with the pharmacological NFAT inhibitor 11R-VIVIT (1 μM) or saline [phosphate-buffered saline (PBS)]. (C) Chast and Mcip1.4 gene expression in heart tissue of calcineurin TG (Myh6-CnA) mice compared to wild-type (WT) littermates. Data in (B) and (C) are mean FC over PBS control ± SEM (n ≥ 3 independent experiments; n = 4 mice per group). (D) Quantitative reverse transcription PCR (qRT-PCR) after ChIP with three different primer pairs surrounding the potential Nfatc3 binding site in comparison to a control primer targeting the Chast intron. Data are mean fold enrichment relative to immunoglobulin G (IgG) control ± SEM (n = 3). (E) Time course of Chast and Plekhm1 expression at different time points after TAC. Data are means ± SEM (n = 5 to 8). (F) Plekhm1 expression in mouse HL-1 cardiomyocytes after Chast repression with GapmeRs. Data are means ± SEM (n = 3). (G to I) Cardiomyocyte size of HL-1 cells calculated from fluorescent microscopy images (G), expression of hypertrophic stress markers (H), and LC3 II/I ratios (I) after small interfering RNA (siRNA)–mediated silencing of Plekhm1 (25 nM, for 48 hours). Data are means ± SEM (n ≥ 3 independent experiments). (J) Atg5 expression levels in hearts from mice treated with a control–AAV9 (adeno-associated virus serotype 9) vector (AAV9-empty) or an AAV9 vector overexpressing Chast (AAV9-Chast). Data are means ± SEM (n = 6 animals per group). In (E), *P < 0.05; **P < 0.01 versus 4-week time point; in all other panels, P value was compared to respective control; Student’s t test.

To delineate a potential mechanism of action, we measured Chast expression in RNA extracted from different cellular compartments of mouse HL-1 cardiomyocytes. As expected, Gapdh and ActB were predominantly found in cytosol, whereas known epigenetic modulators, such as Xist and Neat1 lncRNAs, were predominantly found associated with chromatin. Chast was found to be ubiquitously present in cytosol, nuclear soluble, and chromatin-associated forms (fig. S2), indicating that Chast has the ability to modulate both transcriptional and posttranscriptional processes.

Chast overlaps with two protein-coding genes, Arhgap27 and Plekhm1 (Fig. 1F). Expression of Arhgap27 was not deregulated upon TAC operation or upon in vitro suppression of Chast using GapmeRs in HL-1 cells (fig. S3). However, upon TAC operation, Plekhm1 expression was regulated inversely to that of Chast with strong down-regulation at 6 weeks after operation, corresponding to the highest expression level of Chast (Fig. 3E). Conversely, GapmeR-mediated silencing of Chast in cultured HL-1 cells resulted in a significant up-regulation of Plekhm1 (Fig. 3F). SiRNA-mediated silencing of Plekhm1 resulted in hypertrophic responses of cultured HL-1 cardiomyocytes (Fig. 3, G and H). These results imply that one possible mechanism of Chast activity is regulation of Plekhm1. Plekhm1 has been shown to impede autophagy in various cell lines (16), and inhibition of autophagy in cardiomyocytes drives cardiac remodeling processes (17, 18). Indeed, silencing of Plekhm1 in HL-1 cells inhibited normal cardiac autophagy, as indicated by the reduction in LC3 II/I ratios (Fig. 3I). In line with this observation, hearts overexpressing Chast in vivo showed reduced Atg5 levels, indicating reductions in autophagic capacity (Fig. 3J).

We next performed global microarray mRNA expression analyses after GapmeR-mediated suppression of Chast in murine HL-1 cardiomyocytes to get a broader insight into Chast-mediated transcriptome changes. Chast suppression affected expression of many mRNAs of the HL-1 cardiomyocyte transcriptome (fig. S4A). On the basis of two bioinformatic approaches—correlation of deregulated genes with the KEGG database (fig. S4B) and gene ontology analysis (fig. S4C)—several pathways were identified to be regulated upon Chast silencing: muscle development, metabolic processes, Wnt signaling, and cardiomyopathy. The deregulation of mRNAs involved in pathological cardiac pathways was validated independently by quantitative reverse transcription PRC (qRT-PCR) of HL-1 cells treated with GapmeR-Chast (fig. S4D). Similarly, RNA pulldown assays using biotinylated Chast constructs indicated that Chast interacts with several cardiac-relevant proteins (table S2). Pathway enrichment analysis of the interacting proteins using Enrichr revealed that Chast interacts mainly with proteins involved in cardiomyopathies (table S3). These bioinformatic and experimental screening tools revealed a strong correlation (as quantified by main bioinformatic hits of cardiac disease pathways rather than of others) of the lncRNA Chast with cardiac disease pathways, qualifying Chast as a potential target for therapeutics.

Chast is of translational relevance for humans

We next tested if Chast is conserved and regulated by prohypertrophic stimuli in humans. An alignment of the mouse Chast sequence to human expressed sequence tags (ESTs) revealed existence of a transcript that has a similar sequence (66% homology) and structure compared to mouse Chast originating from chromosome 17 (NC_000017.11|64783199-64783552; Homo sapiens chr17, GRCh38.p2 Primary Assembly) (Fig. 4, A to C). We tested expression of CHAST and other markers of cardiac hypertrophy, as positive controls, in human embryonic stem cell (hESC)–derived cardiomyocytes. Upon treatment with a prohypertrophic stimulus (PE + ISO), human CHAST expression was strongly elevated, along with the other hypertrophy-responsive genes ANP and MCIP1 (Fig. 4D).

Fig. 4. Chast has a functional human homolog.

(A) UCSC (University of California, Santa Cruz) Browser–derived representation of the genomic locus of Chast indicating conserved regions. (B) Pairwise alignment of the human and murine Chast sequence by Clustal Omega. (C) Consensus structure of human CHAST and murine Chast derived from LocARNA (red indicates there is only one possible pair, ochre means two possible pairs). (D) CHAST and other hypertrophic gene expression in hESC (HES3)–derived cardiomyocytes after PE and ISO treatment (10 μM each, 48 hours). Data are mean FC versus control ± SEM (n ≥ 3 independent experiments). (E) HL-1 cardiomyocyte size after lentiviral overexpression of human CHAST. Data are means ± SEM (n ≥ 3 independent experiments). (F) Expression of human CHAST in healthy heart tissues (n = 23) or in heart tissues from patients with aortic stenosis (n = 21). Data are mean FC versus control ± SEM. P values versus respective control in (D) to (F) were determined by Student’s t test. High-resolution files for (A) and (C) are online.

Similar to murine Chast, overexpression of human CHAST in mouse HL-1 cells resulted in a significant enlargement in cardiomyocyte size showing functional conservation (Fig. 6E). CHAST expression was also up-regulated in hearts of patients with aortic stenosis (n = 21), a main direct cause for cardiac hypertrophy, compared to healthy hearts (n = 23) (Fig. 4F). Although preliminary, these results highlight the potential translational relevance of the CHAST lncRNA in human cells and tissues.

Fig. 5. In vivo overexpression of Chast promotes cardiac remodeling in mice.

(A) Map of the cardiomyocyte-specific AAV9 vector. ITR, inverted terminal repeat; cTnT, cardiac troponin T promoter; miR-122 compl, miR-122 binding site (two sites). (B) Elevation of Chast levels in cardiomyocyte fraction 6 weeks after injection compared to the control virus (AAV9-empty). (C) Heart weight–to–tibia length ratio. Data are individual animals with means and SEM (n = 6). (D) Determination of the cardiomyocyte diameters in histological sections. WGA, wheat germ agglutinin. (E) Expression of hypertrophic marker genes Anp, Bnp, and β-Mhc after Chast overexpression. (F and G) Cardiac fibrosis assessed by collagen stain (picrosirius red) and (G) Ctgf expression. (H) Hemodynamic analysis of Chast overexpressing mice using pressure-volume catheter system (LV-ESP, left ventricular end-systolic pressure; LV-EDP, left ventricular end-diastolic pressure). Data in (B) and (D) to (H) are means ± SEM (n = 6 to 14). P values were determined by Student’s t test.

In vivo targeting of Chast ameliorates cardiac remodeling and hypertrophy

To test if prohypertrophic effects of Chast in cultured HL-1 cells and hESC-derived cardiomyocytes translate to therapeutic opportunities in cardiac hypertrophy, we performed in vivo gain- and loss-of-function experiments in mice. For cardiac overexpression of Chast, we created AAV9 viral particles where Chast is expressed under the control of the cardiac troponin T promoter (Fig. 5A). In vivo overexpression of Chast for 6 weeks induced hypertrophic growth in the mouse heart accompanied by increased cardiomyocyte size (Fig. 5, B to D). Expression of cardiac stress markers Bnp and β-Mhc, but not Anp, was also increased upon Chast overexpression (Fig. 4E). Genetic changes were accompanied by increased cardiac fibrosis as evident from picrosirius red staining (Fig. 5F) and elevated Ctgf (connective tissue growth factor) expression (Fig. 5G).

Fig. 6. Pharmacological inhibition of Chast in vivo prevents cardiac remodeling.

(A) Overview of the experimental setup, where six weekly injections of GapmeR-control (negative control, scrambled sequence) or GapmeR-Chast were administered starting day 1 after TAC until removal of the heart at 5 weeks. D0, day 0. (B) Expression of Chast relative to sham. (C) Heart weight–to–tibia length ratio with representative photographs of sham- and TAC-operated mouse hearts. (D and E) Expression of hypertrophic marker genes, Anp and β-Mhc (D), as well as levels of the Chast antisense gene Plekhm1 (E) relative to sham. (F) Cardiac fibrosis analysis and representative collagen stainings (picrosirius red). (G) Ctgf expression. (H) Echocardiographic analysis of cardiac dimensions and function: EF, ESV (end-systolic volume), and EDV (end-diastolic volume). Data in (B) and (D) to (G) are individual animals with means ± SEM (n = 5 to 12). Data in (C) and (H) are means ± SEM (n = 5 to 12). P values were determined by one-way ANOVA with Bonferroni correction.

Consistently, Millar catheter–based measurements identified a significant increase in LV end-systolic and LV end-diastolic pressure upon Chast overexpression (Fig. 5H), indicating that Chast overexpression results in early signs of pathological remodeling processes. This is in line with increased LV pressure values in other models of pathologic cardiac hypertrophy, such as abdominal aortic banding or angiotensin II infusion (19, 20). GapmeR-mediated silencing of Chast 1 day after TAC with weekly follow-up treatments (Fig. 6, A and B) prevented cardiac hypertrophy, keeping animals at sham levels of heart weight–to–tibia length (Fig. 6C) and cardiac stress markers Anp and β-Mhc (Fig. 6D) compared to mice treated with a scrambled control. Consistently, in vivo repression of Chast prevented Plekhm1 down-regulation (Fig. 6E). Furthermore, cardiac pressure–overloaded mice treated with GapmeRs against Chast showed reduced fibrosis and Ctgf expression levels (Fig. 6, F and G), as well as improved EF (ejection fraction) and smaller systolic and diastolic volumes (Fig. 6H). The weekly dosing regimen (Fig. 6A) did not result in significant changes in clinical chemistry parameters, such as aspartate aminotransferase (AST; liver toxicity marker), cholinesterase (ChE; neurotoxicity marker), urea (kidney toxicity marker), and interleukin 6 (IL-6; a proinflammatory marker) (fig. S5).

We next tested the therapeutic potential of Chast inhibition in another model of established cardiac disease. Here, we treated mice 4 weeks after TAC by weekly injections of a GapmeR against a scrambled sequence (control) or against Chast (Fig. 7A). After 8 weeks, Chast inhibition led to a significant reduction in heart weight/tibia length ratios in mice compared to animals treated with control-GapmeRs (Fig. 7, B and C). This is in line with smaller cardiomyocyte sizes in histological sections of TAC animals treated with the Chast-GapmeR when compared with control-GapmeR treatment (Fig. 7C). Echocardiographic assessment of hearts from mice 8 weeks after TAC and treated with the control-GapmeR showed a significant decline in EF, whereas animals treated with the Chast-GapmeR had normal EF close to that of sham-operated animals (Fig. 7D).

Fig. 7. Chast inhibition in established hypertrophy.

(A) Overview of the experimental setup where five weekly injections of GapmeR-control (negative control, scrambled sequence; 20 mg/kg) or GapmeR-Chast (20 mg/kg) were administered starting week 4 after TAC until removal of the heart at 8 weeks. (B) Heart weight–to–tibia length ratio with representative photographs of sham- and TAC-operated mouse hearts. (C) Cardiomyocyte size of sham- and TAC-operated mouse hearts after GapmeR treatment with representative fluorescence images. WGA, wheat germ agglutinin. Data are means ± SEM (n = 6 to 9). (D) Echocardiographic determination of cardiac function at the study endpoint (EF). Data in (B) and (D) are individual animals with mean and SEM (n = 6 to 9). P values were determined by one-way ANOVA with Bonferroni correction.

DISCUSSION

Here, we present the lncRNA Chast as a crucial player mediating cardiac hypertrophy in vivo in mice and in vitro in murine and human cells. Targeting Chast by specific oligonucleotide inhibitors both prevented disease and attenuated established cardiac hypertrophy with improved heart function. Although most of the human genome is transcribed, less than 2% of it encodes for proteins. One class of noncoding RNAs, microRNAs, has been linked to several biological processes, and their translational importance has been well studied in cardiac diseases in particular. In contrast, few reports exist on the role of lncRNAs in cardiac development (21). One recent study identified that the lncRNA Mhrt, which originates from MYH7 locus, is cardioprotective, and restoration of Mhrt levels protects the heart from hypertrophy and failure (22). We have identified a new lncRNA, Chast, that is prohypertrophic in mice, demonstrates 66% homology with human CHAST, and can be inhibited in vivo to prevent or reverse experimental cardiac hypertrophy.

Chast is expressed as a full-length transcript with several “stop” codons and no ORF activity, validating Chast as a noncoding RNA. Chast is not a cardiac-specific lncRNA; however, its cell type–specific expression is very sensitive to pressure overload in vivo or prohypertrophic stimuli in vitro, such as PE and ISO, in both murine and human cardiomyocytes. Consistent with this, we found that Chast expression was induced partly by the prohypertrophic transcription factor NFAT, leading to hypertrophic growth, as evidenced by increases in cardiac stress markers in vitro and in vivo.

LncRNAs are likely powerful treatment targets because they intensively interact with their genetic environment, placing them as crucial regulators of genetic networks. LncRNAs can function as guide strands influencing genes or proteins in trans (on distantly located genes) or in cis (on neighboring genes) (23). Our data reveal a potential cis-regulating role of Chast on the gene located on the opposite strand, named Plekhm1. Although our current data do not explain the exact mechanism of the Chast/Plekhm1 interaction, the reciprocal regulation of both genes in disease and the direct effects of Chast manipulation on Plekhm1 expression make a cis-regulatory action likely. Further studies need to show how Chast leads to Plekhm1 silencing. Plekhm1 has been recently described as a regulator of autophagy (16), which fits with our observation that Plekhm1 silencing in the cardiomyocyte cell line HL-1 inhibited cardiomyocyte autophagy and drove cardiac hypertrophy. Thus, strategies overexpressing Plekhm1 in cardiac hypertrophy–associated diseases may be of additional therapeutic relevance.

Very likely, the lncRNA Chast will have many more targets. For instance, bioinformatic analyses showed that Chast affects the transcriptome of murine HL-1 cardiomyocytes with a strong effect on pathways of cardiac muscle morphogenesis, cardiomyopathies, and Wnt signaling. Indeed, activation of Wnt signaling has been described to be an additional driver of cardiac hypertrophy (24). How Chast may affect Wnt signaling pathways as suggested by the transcriptome analyses is unknown. Additionally, our bioinformatic and mechanistic studies show that Chast interacts with many proteins involved in cardiac diseases, including vinculin (25), calmodulin (26), or laminins (27). Future studies will deepen our understanding of what additional mechanisms are involved in Chast’s mode of action in cardiac hypertrophy and remodeling. These studies should address how Chast regulates the cardiomyoctye transcriptional changes (for instance, by epigenetic processes) and what the cardiac implications of Chast/protein-binding events are.

GapmeR-mediated inhibition of Chast in pressure-overloaded mouse hearts prevented or halted hypertrophic growth, which suggests that Chast could be a new therapeutic target for cardiomyopathies. To translate to patients, we identified a new EST in the human genome with a sequence and secondary structure similar to murine Chast. This homolog, CHAST, was significantly up-regulated in hypertrophic heart tissues from patients with aortic stenosis—but not healthy controls—and in hESC-derived cardiomyocytes after hypertrophic stimuli. This is an intriguing finding because it potentially allows translation into clinical scenarios. However, there are translational hurdles to be cognizant of. Specific human CHAST inhibitors need to be developed and tested in human primary, hESC-derived, and/or induced pluripotent stem cell (iPSC)–derived cardiomyocytes. Because there seems to be a very high functional conservation of CHAST among species, as human CHAST drove hypertrophy in rodent cells (Fig. 4E), the next steps should be pharmacokinetic and toxicological assessments in additional animal species (rodent and nonrodent models) of CHAST inhibitors with the aim to test these in future clinical trials. Indeed, early toxicological evaluation of the murine Chast inhibitor in this study showed promising results. Intensive toxicological testing of leading CHAST inhibitors will be needed to search for potential adverse effects in other organs, especially the liver and kidney, because systemic toxicity affecting these organs is a primary concern in the oligonucleotide therapeutics field (28).

Our results highlight that Chast is a prohypertrophic lncRNA, and in vivo targeting of this lncRNA could be beneficial in cardiac hypertrophy. Our study sets the ground for future preclinical and clinical developments of CHAST inhibitors for the treatment of pathological cardiac remodeling processes, such as cardiac hypertrophy and potentially heart failure.

MATERIALS AND METHODS

Study design

The aim of this study was to identify a lncRNA that played an important role in the development of cardiac hypertrophy and to start translating these findings into a therapeutic approach. By transcriptional profiling of mouse hypertrophic hearts, we derived a lncRNA as a potential therapeutic target and performed loss- and gain-of-function studies to assess the influence of the lncRNA on the growth and genetics of murine cardiomyocytes in vitro. The TAC model using C57BL/6N mice was chosen as a well-established model of pathological cardiac hypertrophy. Real-time qPCR and measurements of the cell surface area were used to quantify cardiac hypertrophy in vitro. Experiments were reproduced in at least three independent experiments as stated in the figure legends.

For in vivo experiments, mice were randomly assigned to treatment groups, and, wherever applicable, treatment conditions were kept blinded until statistical analysis. Group sizes were chosen on the basis of previous experiences with the TAC model, which indicated that group sizes of at least five animals could robustly identify changes in cardiac hypertrophy after at least 4 weeks of TAC. Pharmacological inhibition of the lncRNA by locked nucleic acid long RNA GapmeR injection 1 day and weekly follow-up treatments after TAC was used to illustrate the effect of in vivo repression on cardiac function and the hypertrophic gene program. In the preventive approach (weekly treatments starting at day 1 after TAC), animals were analyzed at 5 weeks, when pathological cardiac hypertrophy can be easily assessed. In the therapeutic approach (weekly treatments starting at week 4 after TAC), animals were evaluated at 8 weeks, because time was needed to establish hypertrophy by TAC before treatment started. Cardiac hypertrophy was quantified by heart weight–to–body weight or by heart weight–to–tibia length ratios, echocardiography, and qRT-PCR measurements of expression changes.

Additional studies of quantification of the human CHAST homolog were performed in hESC-derived cardiomyocytes treated with prohypertrophic stimuli. Although we used hESCs to demonstrate CHAST homology, cell size in hESC-derived cardiomyocytes continuously increases in culture during differentiation. Therefore, to avoid confounding influence of normal cell enlargement, we used murine HL-1 cells for the determination of CHAST-induced phenotype. In addition, by this experiment, we could show functional conservation of the CHAST sequence in other species. The effect of CHAST was also evaluated on cardiac tissue derived from patients with cardiac hypertrophy due to aortic stenosis or healthy human heart tissue. Aortic stenosis can be viewed as the most similar human disease to the TAC mouse model.

Animal experiments

All animal experiments were performed in accordance with the relevant guidelines and regulations with the approval of the responsible local and national authorities. To generate cardiac pressure overload, male C57BL/6N mice (8 to 10 weeks old) were subjected to TAC as described previously (29). Sham-operated animals underwent the same procedure without constriction of the aorta. At different time points after the surgery, as indicated in the figure captions, hearts were excised and frozen in liquid nitrogen or fractionated as described below.

Adult (8- to 10-week-old) male C57BL/6N mice were once treated by tail vein injection with 0.5 × 1012 copies of AAV9 harboring either the Chast sequence or an empty vector. GapmeR-Chast or GapmeR-negative control A (20 mg/kg) (both from Exiqon) in saline (0.9% NaCl) as carriage medium were delivered weekly by intraperitoneal injection. After 6 or 5 weeks, respectively, cardiac function and morphometric changes were assessed by echocardiography and hemodynamic measurements, and the hearts were removed for biochemical analysis. Heart tissues of calcineurin TG mice [MYH6-CnA (15)] and their wild-type littermates were provided by L.J.d.W. (Academic Hospital Maastricht, Maastricht, Netherlands).

Human tissue sampling

This study was performed with the approval of the institutional ethical committees of the University of Würzburg, Germany, and the Hannover Medical School, Hannover, Germany. We analyzed cardiac tissues from patients undergoing aortic valve replacement because of aortic stenosis (n = 21; mean age, 72.81 ± 14.61; male/female 14:7). For comparison, tissue of healthy adult human hearts was used (n = 23, mean age: 38.61 ± 12.96, male/female 14:9; hearts were donor hearts not used for transplantation or RNA from healthy human hearts was purchased from Biochain Institute). Cardiac biopsies were obtained from the left ventricle and frozen in liquid nitrogen until RNA was isolated, and gene expression measurements were performed.

Plasma sampling and biochemical analysis

EDTA-plasma samples were drawn from the mice and immediately centrifuged at 1000g for 10 min. The supernatant was stored at −80°C until analysis. Enzyme activities were assessed on a Modular P800 automatic analyzer using standard methods (Roche Diagnostics): AST and ChE. Urea was measured photometrically [UREAL; kinetic urease test with GLDH (glutamate dehydrogenase)] on a Cobas 6000 analyzer system (Roche Diagnostics). IL-6 levels were determined by applying the Mouse IL-6 Quantikine ELISA (enzyme-linked immunosorbent assay) Kit (R&D Systems Inc.) according to the manufacturer’s protocol.

RNA isolation

Total RNA of tissues and cultured cells was isolated using RNeasy Mini Kit (Qiagen) or TriFast method (Peqlab) according to the manufacturer’s instructions. Quantification and quality control were performed with Synergy HT Reader (BioTek) and Agilent 2100 Bioanalyzer (Agilent Technologies), respectively.

qRT-PCR analysis

For microarray validation and quantitative detection of lncRNAs and mRNAs, cDNAs were synthesized with the iScript Select cDNA synthesis kit (Bio-Rad), and real-time RT-qPCR analysis was performed using specific primers (table S4) and the iQ SYBR Green Mix (Bio-Rad) according to the manufacturer’s protocol. Gene-specific expression levels were normalized to levels of glyceraldehyde-3-phosphate dehydrogenase (Gapdh) or β-actin (ActB).

Cell culture, transfection, and treatment

Murine HL-1 cardiomyocytes were provided by W. Claycomb (Louisiana State University Health Sciences Center, New Orleans, LA) and cultured in Claycomb medium supplemented with 10% fetal bovine serum (Sigma-Aldrich) as described (30). For inhibition of lncRNA function, GapmeRs (Exiqon) against specific lncRNAs were used. Cells were transiently transfected with GapmeRs (Chast: 5′-TGGATTTGGAGGTTGG-3′; negative control A: 5′-AACACGTCTATACGC-3′). Silencing of Plekhm1 was achieved using 25 nM Plekhm1 siRNA (Santa Cruz, sc-37007) and its corresponding control (Santa Cruz, sc-152320). All transfections were performed using Lipofectamine 2000 reagent (Invitrogen) according to the manufacturer’s protocol.

To mimic cardiomyocyte hypertrophy in vitro, HL-1 cells were cultured in standard medium and then shifted to standard medium with low (1%) serum supplemented with 100 μM PE and 100 μM ISO (both from Sigma-Aldrich) or PBS as control. Cell exposure to the hypertrophic stimuli was stopped after 48 hours of incubation at 37°C in 5% CO2.

In NFAT inhibition experiments, the cell-permeable peptide 11R-VIVIT (Calbiochem) at a concentration of 1 μM was added to HL-1 cells in standard medium with low serum for 48 hours.

Human pluripotent stem cells and cardiac differentiation

The hESC reporter line HES3 MESP1mCherry/w/NKX2-5eGFP/w (31) was maintained at standard conditions on mouse embryonic fibroblasts. Cells were accutase-treated (PAA Laboratories) and seeded at the density of 0.5 × 105 cells/cm2 on Geltrex-coated flasks in E8 medium (Stemcell Technologies) (32) supplemented with 10 mM Y-27632 (Tocris Bioscience) for 24 hours. Medium was replaced daily, and cells were passaged twice per week. The differentiation protocol was based on published work (3335). Briefly, single cells were pre-inoculated on Geltrex-coated plates at 3.5 × 105 cells/cm2 with 10 mM Y-27632 in E8. Differentiation of resulting confluent monolayer was induced 4 days after single cell seeding using the chemical Glycogen synthase kinase 3β (GSKβ) antagonist CHIR99021 (synthesized by the Institute of Organic Chemistry, Leibniz University Hannover, Hannover, Germany) at 7.5 μM for 24 hours to up-regulate Wnt-pathway activity. On day 3, the chemical Wnt-pathway inhibitor IWP2 (Tocris Bioscience) was added at 5 mM for 48 hours. During the differentiation, cells were kept in RPMI 1640 medium (Life Technologies) supplemented with B27 (without insulin). Medium was entirely replaced on days 0 (+CHIR), 1, 3 (+IWP2), and 5. Cells were cultured in RPMI 1640 supplemented with B27 from day 7 onward. For long-term maturation of cardiomyocytes, cells were cultured in RPMI 1640 supplemented with B27. Medium was replaced every 48 hours.

Assessment of cardiac differentiation efficiency was performed by flow cytometry using with the endogenous Nkx2.5-GFP reporter signal. Staining was performed as described before (34, 35), applying with monoclonal antibodies targeting human troponin T, α-myosin heavy chain, and β-myosin heavy chain.

Cloning and virus production of AAV9 vectors

Chast (lncRNA ENSMUST00000130556) was synthesized and cloned into a self-complementary AAV backbone plasmid under the HEK293T control of the chicken troponin T promoter and followed by two complementary binding sites for miR-122 to minimize expression in the liver (Fig. 4A). The same plasmid lacking the Chast sequence was used as a control. Human embryonic kidney (HEK) 293 T cells were grown in 10-layer cell stacks (Thermo Scientific, 140400) for 24 hours (Dulbecco’s modified Eagle’s medium, 10% fetal bovine serum) before transfection. The transgene plasmid and the helper plasmid were transfected into HEK293T cells using polyethylenimine (Polysciences, 24765). After 72 hours, the viruses were harvested and purified from benzonase-treated cell lysates over an iodixanol density gradient (OptiPrep, Fresenius Kabi Norge). AAV titers were determined by real-time PCR on vector genomes using the SYBR Green Master Mix (Roche). The chicken troponin T promoter and the pDP9rs plasmid were provided by R. Hajjar (Icahn School of Medicine at Mount Sinai, New York, NY).

Fluorescence microscopy and histology

HL-1 cells were treated with 100 μM PE and 100 μM ISO (PE + ISO) as indicated in Results. For measurement of cell size, cells were fixed with 4% paraformaldehyde and stained with phalloidin–tetramethylrhodamine B isothiocyanate (10 μg/ml) and 4′,6-diamidin-2-phenylindol (DAPI) (5 μg/ml) (both from Sigma Aldrich). Cryosections of the LV myocardium were visualized by wheat germ agglutinin stain coupled to Alexa Fluor 488 (Invitrogen). The cell surface area of cardiomyocytes was calculated using the NIS-Elements BR 3.2 package (Nikon Instruments Inc.). For histological examination of fibrosis, paraffin sections of the left ventricle were stained with Sirius red and piric acid to visualize collagen deposition. The collagen content was calculated as the percentage of the area in each heart section.

Statistical analysis

For statistical analysis, GraphPad 6 (GraphPad Software) was used. Data are displayed as means ± SEM. Statistical comparison among two groups was evaluated by two-tailed unpaired Student’s t test. For comparison of more than two groups, one-way ANOVA corrected with the Bonferroni posttest was performed. In all cases, P < 0.05 was considered statistically significant.

SUPPLEMENTARY MATERIALS

www.sciencetranslationalmedicine.org/cgi/content/full/8/326/326ra22/DC1

Materials and Methods

Fig. S1. Chast is a lncRNA.

Fig. S2. Chast is located in all subcellular fractions of cardiomyocytes.

Fig. S3. Chast does not alter Arhgap27 levels.

Fig. S4. GapmeR-mediated silencing of Chast alters the transcriptome of HL-1 cells.

Fig. S5. Chast repression in vivo has no significant influence on plasma marker of kidney, neurological, or liver damage as well as on circulating proinflammatory IL-6 levels.

Table S1. Transcription factor binding site prediction for the Chast promoter.

Table S2. Cardiac relevant interaction partners enriched in the pulldown fraction of Chast-biotin compared to luciferase-biotin.

Table S3. OMIM enrichment analysis for the Chast-protein pulldown.

Table S4. Oligonucleotide primers used in this study.

References (3643)

REFERENCES AND NOTES

Acknowledgments: We thank A.S. and T.M. for help with the lentiviral lncRNA overexpression plasmid, S.E. and P.A. for the AAV constructs, A. Pich (Hannover Medical School) for the mass spectrometry analysis, as well as F.W. and L.J.d.W. for the human and murine heart tissues, respectively. We thank A. Pfanne (Hannover Medical School) for excellent technical support. Some of the healthy human heart tissues were provided by T. Kraft (Hannover Medical School). The chicken troponin T promoter and the pDP9rs plasmid were provided by R. Hajjar, Icahn School of Medicine at Mount Sinai, New York. Funding: The authors had support from the IFB-Tx (Integrierte Forschungs- und Behandlungszentrum Transplantation) (BMBF 01EO1302; T.T.), DFG (Deutsche Forschungsgemeinschaft) (TH 903/10-1 and RE 3523/1-1; T.T. and R.K.), the REBIRTH Excellence Cluster (T.T.), Fondation Leducq (MIRVAD; T.T.), the European Union funded ERC (European Research Council) Consolidator Grant (LONGHEART; T.T.), the Hannover Biomedical Research School program (Regenerative Sciences; T.T. and J.V.), and the Bayerische Forschungsstiftung (P.A.). R.Z. received project-related funding from the German Research Foundation (Cluster of Excellence REBIRTH DFG EXC62/3; ZW 64/4-1), the German Ministry for Education and Science (13N12606), and StemBANCC [support from the Innovative Medicines Initiative joint undertaking under grant agreement no. 115439-2, whose resources are composed of financial contribution from the European Union (FP7/2007-2013) and EFPIA (European Federation of Pharmaceutical Industries and Associations) companies’ in-kind contribution]. Author contributions: J.V., R.K., and T.T. devised the study, designed the experiments, and wrote the manuscript. J.V. performed and analyzed all experiments with exception of the in vivo mouse studies. In vivo mouse studies were performed by A.F., K.Z., J.R., and S.B. Bioinformatic analysis and microarray evaluation were done by K.X., M.K., M.D., and T.D. Lentiviral overexpression vectors were provided by T.M. and A.S., whereas P.A. and S.E. designed and generated the adeno-associated viral constructs. Tissue samples from calcineurin TG mice and human aortic stenosis and healthy hearts were provided by L.J.d.W. and F.W., respectively. hESC-derived cardiomyocytes were prepared by E.B. and R.Z. A.J., J.F., and K.S. supported in experimental work. Competing interests: J.V., R.K., and T.T. have filed a patent about the diagnostic and therapeutic use of several cardiovascular lncRNAs including Chast/CHAST. Data and materials availability: All data and materials are available.
View Abstract

Stay Connected to Science Translational Medicine

Navigate This Article