Therapeutic Cardiac‐Targeted Delivery of miR‐1 Reverses Pressure Overload–Induced Cardiac Hypertrophy and Attenuates Pathological Remodeling

Hypertrophy Model

Male Sprague‐Dawley rats (180 to 200 g) underwent ascending aortic banding (AAB) after induction of anesthesia with intraperitoneal ketamine (up to 85 mg/kg) and xylazine (up to 10 mg/kg), as described previously.²⁰ Additional animals underwent a left thoracotomy without aortic banding to serve as age‐matched controls (sham).

Recombinant AAV9.miR‐1 Construct

A genomic fragment containing the miR‐1 precursor was polymerase chain reaction (PCR) amplified from the rat miR‐1 precursor sequence, cloned into self‐complementary AAV genome vector, and pseudotyped into rAAV9 capsids (AAV9.miR‐1). The green fluorescent protein (GFP)‐expressing vector (AAV9.GFP) was used as a control. Recombinant AAV‐9 viruses were produced as previously described.²¹

Viral Delivery Protocol

Two weeks after AAB, the rats with evident LV hypertrophy, as assessed by echocardiography, were randomly chosen to receive a single‐bolus tail‐vein injection of either AAV9.miR‐1 (n=6) or AAV9.GFP (n=9) at 5×10¹¹ vg (viral genomes) per animal. Echocardiographic measurements were performed at baseline and 2 and 9 weeks post AAB. Invasive hemodynamics measurements were also obtained at 9 weeks post AAB, and the animals were killed (protocol summarized in Figure 1A).

Echocardiographic Analysis

Transthoracic echocardiography was performed using a Vivid 7 (GE Healthcare) echocardiography apparatus with a 14‐MHz probe (i13L probe; General Electric). Animals were sedated with ketamine (up to 80 mg/kg) injected intraperitoneally. Long‐axis parasternal views and short‐axis parasternal 2‐dimensional (2D) views, at the mid‐papillary level, of the left ventricle were obtained to calculate the LV end‐diastolic (LVEDV) and end‐systolic (LVESV) volumes as well as the ejection fraction of the left ventricle. Volumes were calculated by using the formulae of the area‐length method (V=5/6×A×L, where V is the volume in mL, A is the cross‐sectional area of the LV cavity in cm², obtained from the mid‐papillary short parasternal image in diastole and in systole, and L is the length of the LV cavity in cm, measured from the long‐axis parasternal image as the distance from the endocardial LV apex to the mitral‐aortic junction in diastole and in systole). M‐mode images were obtained by 2‐dimensional guidance from the parasternal short‐axis view for the measurements of LV wall thickness of the septum (cm) and of the posterior wall (cm), LV end‐diastolic diameter (cm), and LV end‐systolic diameter (cm), as well as to calculate the LV fractional shortening (FS, %).

Hemodynamics Analysis

At end point, LV pressure–volume loops (P‐V) measurements were obtained as previously described.²² Briefly, rats were anesthetized with inhaled isoflurane (5% v/v) for induction and subsequently intubated and mechanically ventilated as noted in the surgery section. Isoflurane was lowered (2% to 3% v/v) for surgical incision. The chest was opened through a median sternotomy, and a 1.9F rat P‐V catheter (Scisense) was inserted into the LV apex through an apical stab performed with a 25‐gauge needle. The animals were kept sedated with 0.75% to 1% isoflurane maintaining a stable heart rate (≈350 beats/min). Hemodynamic recordings were performed after 5 minutes of stable heart rate. The intrathoracic inferior vena cava was transiently occluded to decrease venous return during the recording to obtain load‐independent P‐V relationships. Linear fits were obtained for end‐systolic P‐V relationships (ESPVR) and end‐diastolic P‐V relationships (EDPVR). At the end of the experiment, 30% NaCl 50 μL was slowly injected into the external jugular vein for ventricular parallel conductance measurement as previously described.^22–23 Blood resistivity was measured using a special probe (Scisense). Volume measurements were initially obtained as blood conductance and calibrated using Baan equation 3, and pressure sensors were calibrated according to manufacturer's instructions (Scisense).

Quantification of Mature miRNAs

Total RNA was isolated with use of an mirVana miRNA Isolation Kit (Ambion) followed by a DNase treatment to eliminate contaminating genomic DNA (Invitrogen). Mature miR‐1 expression was quantified by real‐time quantitative PCR (qPCR) using the Taqman MicroRNA Assays according to the manufacturer's instructions (Applied Biosystems). Gene expression levels were normalized to U6 rRNA endogenous control, and fold changes were calculated using the ΔΔCt method.

qRT‐PCR Gene Expression Analysis

Relative gene expression was determined using 2‐step qRT‐PCR. Quantitative PCRs were performed with Power SYBR Green Master Mix (Applied Biosystems) on an ABI Prism 7500 Real Time PCR System. Fold changes were calculated using the ΔΔCt method with normalization to 18S rRNA housekeeping gene.

Western Blotting Analysis

Protein expression was evaluated in LV lysates by Western blot analysis according to standard procedures with antibodies against phospholamban (Pln) (Badrilla, UK), Serca2a (custom made in our laboratory), extracellular signal‐regulated protein kinase (ERK1/2), phospho‐ERK1/2 (Thr202/Tyr204), phospho‐p38 (Thr180/Tyr182), p38, phospho–c‐jun NH₂‐terminal kinase (JNK) (Thr183/Tyr185), JNK, Bcl‐2, and Bax (Cell Signaling Technology), Fbln2 (Genetex), Ncx1 (Abcam), and Gapdh (Sigma‐Aldrich). The signals were detected with ECL‐Plus chemiluminescence detection kit (Pierce) or with the ODYSSEY Infrared Imaging System (Li‐CoR).

Histological Assessment of Fibrosis

LV cryosections (≈10 μm) were fixed in 10% buffered formalin and stained with picrosirius solution (0.1% Sirius Red in picric acid; Sigma‐Aldrich). Images were acquired at ×20 magnification under circular polarized illumination using an Olympus BX50 microscope. The relative amount of collagen area to total tissue area was measured in each image by using a color threshold technique with National Institutes of Health ImageJ software.

Histological Assessment of Myocyte Cross‐sectional Area

Heart LV cryosections (10 μm) cut at the papillary muscle level were fixed in 4% paraformaldehyde and incubated with Texas Red‐X conjugate of wheat germ agglutinin (5 μg/mL; Invitrogen) for 45 minutes at room temperature. The cross‐sectional area of approximately 100 myocytes with circular shape of the cell membrane was measured in the LV free walls of each animal using ImageJ software.

Dual Luciferase Assays

The 3′‐UTR of Fbln2 and Ncx1 containing the predicted miR‐1 target sequences were amplified by PCR and cloned into the multiple cloning region located downstream of the Renilla translational stop codon of the psiCHECK‐2 dual luciferase vector (Promega). Using the DharmaFECT Duo transfection reagent (Thermo Fisher Scientific), 50 ng of psiCHECK–3′‐UTR constructs and 10 pmol of pre–miR‐1 miRNA precursors or 10 pmol of pre‐miR negative control (Life Technologies) were cotransfected in HEK293FT (Invitrogen) cells. Forty‐eight hours after transfection, the normalized Renilla luciferase activity (Renilla luciferase/firefly) was measured using the Dual‐Glo Luciferase Assay System according to the manufacturer's instructions (Promega).

Isolation and Culture of Rat Ventricular Myocytes

Neonatal rat ventricular myocytes were isolated by enzymatic dissociation of cardiac ventricle from 1‐ to 2‐day‐old Sprague‐Dawley pups using the neonatal cardiomyocyte isolation system according to the manufacturer's instructions (Worthington). Adult rat ventricular myocytes were isolated from male Sprague‐Dawley rats (280 to 300 g) using a modified Langendorff perfusion system as previously described.²⁴

Terminal Deoxynucleotidyl Transferase–Mediated dUTP Nick End Labeling Assay

The assessment of apoptosis was performed by the indirect terminal deoxynucleotidyl transferase–mediated dUTP nick end labeling assay (TUNEL) method, using an antidigoxigenin antibody with a rhodamine fluorochrome with the Apoptag Red in Situ Apoptosis Detection kit (Millipore) according to the manufacturer's protocol.

Human Heart Specimens

Human heart tissue specimens were obtained from the National Disease Research Interchange through the Human Tissues and Organs for Research program.

Statistical Analysis

The underlying assumption of normal distribution was investigated by performing a Kolmogorov–Smirnoff normality test and normal probability plot test. When the distribution was found to be normal, statistical significance between 2 groups was examined by t test and by 1‐way ANOVA for multigroup comparisons. When the ANOVA results were significant, the differences among individual groups were determined with the Bonferroni post hoc test. The echocardiography data were analyzed with repeated measures ANOVA test. The luciferase reporter assay data were analyzed with 2‐factor ANOVA test. P<0.05 was considered significant.

Systemic Administration of AAV9.miR‐1 Restores miR‐1 Gene Expression In Vivo

A genomic fragment containing the miR‐1 precursor was PCR amplified from the rat miR‐1 precursor sequence and cloned into the self‐complementary AAV genome vector under the control of cytomegalovirus promoter (Figure 1B). The expression cassette was comprised of the miR‐1 stem loop sequence flanked by its native intron sequence, which preserves the putative hairpin structure and proper endogenous processing. The expression of mature miR‐1 was validated in vitro in neonatal (Figure 1C) and adult cardiac myocytes (Figure 1D) by qRT‐PCR with stem‐loop primers (Figure 1E). AAV9‐mediated miR‐1 gene transfer in vivo in the setting of pressure‐overload hypertrophy (Figure 1A) restored the expression of the mature miR‐1 in the hypertrophic hearts (Figure 1F).

miR‐1 Gene Transfer Reverses Hypertrophy and Prevents Functional Deterioration

LV function and dimensions were measured by serial echocardiography at baseline and 2 weeks and 9 weeks post‐AAB (Figure 2A). After 2 weeks, the AAB hearts exhibited concentric hypertrophy characterized by marked increase in LV thickness (LV wall thickness of the septum and of the posterior wall) and a significant increase in fractional shortening (FS) compared with sham‐operated animals (Table 1). Seven weeks after gene transfer, the LV wall thickness of the AAV9.miR‐1 group was significantly decreased compared with the AAV9.GFP group (LV wall thickness of the posterior wall: 2.32±0.08 versus 2.75±0.07 mm, P<0.001; LV wall thickness of the septum: 2.23±0.06 versus 2.54±0.10 mm, P<0.05) (Figure 2B and Table 1). In addition, the AAV9.GFP group displayed substantial increase in LV chamber end‐systolic (LVEDs) and end‐diastolic (LVEDd) dimensions compared with the sham group, indicating ventricular dilatation, which has been significantly attenuated by AAV9.miR‐1 gene transfer (Table 1). The AAV9.GFP group exhibited significantly reduced contractility compared with the AAV9.miR‐1–treated animals, as assessed by FS (37.60±5.01% versus 70.68±2.93%, P<0.05) (Figure 2C) and the ejection fraction (53.86±7.12% versus 87.99±1.43%, P<0.05) (Table 1).

Additionally, we further examined the effects of miR‐1 restoration on cardiac function in vivo by using the PV‐loop analysis. Catheterization and hemodynamic analysis after 7 weeks of gene transfer showed a significant increase in LV chamber dimensions in the AAV9.GFP– compared with the AAV9.miR‐1–treated hearts (EDV: 1316.30±79.15 μL versus 683.37±79.14 μL, P<0.05; ESV: 869.40±150 μL versus 312.86±83.85 μL, P<0.05) (Table 2), which is in agreement with the echocardiographic measurements. The analysis of LV function also revealed a significant improvement in the AAV9.miR‐1 systolic parameters compared with the AAV9.GFP group as measured by the end‐systolic PV relation normalized to the effective arterial elastance (Ea) (1.87±0.46 versus 0.96±0.38, P<0.05) (Figure 2D). Moreover, diastolic parameters were also normalized in the AAV9.miR‐1 group as evidenced by the decrease in the time constant of LV pressure decay during the isovolumic relaxation phase tau (7.94±0.7 versus 9.69±1.42, P<0.05) (Figure 2E). Taken together, these data suggest that miR‐1 gene transfer preserved cardiac function and prevented the transition to heart failure induced by pressure overload.

Postmortem analysis revealed that the heart weight–to–body weight ratio was significantly lower in AAV9.miR‐1– compared with AAV9.GFP–treated hearts (3.72±0.10 versus 5.9±0.69 mg/g, P=0.02) (Figure 3A). We confirmed these observations at the cellular level by evaluating the cross‐sectional area of cardiac myocytes in histological sections of LV tissues at 7 weeks post gene transfer (Figure 3B). We observed that cardiac myocytes from AAV9.GFP were significantly larger than those from the sham‐operated hearts (689.33±14.53 versus 379.68±9.10 μm², P<0.05). In contrast, AAV9.miR‐1 treatment significantly reduced the cardiac myocyte cross‐sectional area to levels similar to those of the sham‐operated hearts (400.92±12.62 versus 379.68±9.10 μm², P<0.05) (Figure 3B).

miR‐1 Modulates the Expression of Molecular Markers of Cardiac Hypertrophy

Next we investigated the effect of miR‐1 treatment on molecular abnormalities associated with pathological hypertrophy. We tested the hypothesis that the reversal of the pathological remodeling observed in the AAV9.miR‐1 group in vivo is paralleled by a reversal in the reinduction of the maladaptive fetal cardiac gene program. We assessed the expression of the hypertrophic fetal genes, atrial natriuretic factor (Anf), skeletal muscle α‐actin (Acta1), cardiac alpha‐actinin (Actc1), as well as the contractile protein isoforms α‐ and β‐myosin heavy chain (Myh6 and Myh7, respectively), at 7 weeks post gene transfer. Pressure‐overload–induced hypertrophy in the AAV9.GFP–treated group was associated with reinduction of the “fetal gene program” characterized by a significant increase in the mRNA expression of Anf (21‐fold), Myh7 (4.3‐fold), and Acta1 (1.8‐fold) and a decrease in Myh6 (21.6‐fold) and Actc1 (1.7‐fold) compared with sham‐operated animals (Figure 3C). In contrast, we observed a reversal of the MHC shift that was characterized by the significant increase in the “adult” Myh6 (17.84‐fold) and decrease in the “fetal” Myh7 (1.55‐fold) as well as the significant decrease in Anf expression (4.37‐fold) and the significant increase of Actc1 (1.8‐fold) in miR‐1‐treated hearts compared with AAV9.GFP–treated hearts (Figure 3C).

miR‐1 Regulates the Expression and Activity of Key Calcium Homeostasis Genes

Ca²⁺ uptake into the sarcoplasmic reticulum (SR) is mediated by the cardiac SR Ca²⁺‐ATPase (Serca2a) and regulated by Pln.²⁵ Serca2a is an important regulator of intracellular Ca²⁺ signaling in the heart, and its decreased expression and activity are implicated in heart failure.²⁶ We performed immunoblotting analyses to assess the protein levels of Serca2a and Pln and the phosphorylation state of Pln (Figure 4A). We observed a significant decrease in Serca2a protein expression levels (50%) but no change in Pln expression in ventricular tissue from AAV9.GFP compared with sham‐operated animals (Figure 4B). Phosphorylation of Pln at Ser16 was also decreased (Figure 4B). In contrast, AAV9.miR‐1 significantly upregulated Serca2a protein levels and restored the Serca2a/Pln ratio to sham levels compared with AAV9.GFP; it also increased Pln phosphorylation at Ser16 and the phosphoPln/Pln ratio (Figure 4B). Taken together, the data suggest that the changes in Serca2a content and activity (increased Pln phosphorylation and Serca2a/Pln ratio) are indicative of enhanced SR Ca²⁺ cycling, which correlates with the improvement in cardiac contractility observed following miR‐1 gene transfer.

Activity of the Mitogen‐Activated Protein Kinase Signaling Pathways Is Altered by miR‐1

It is well established that pressure‐overload–induced biomechanical stresses stimulate various signaling pathways essential for induction of the hypertrophic response.³ These signaling pathways include the mitogen‐activated protein kinase superfamily with its 3 terminal effector kinase subfamilies: ERK1/2, JNK, and p38.²⁷ We sought to determine whether miR‐1 restoration had an effect on these pathways. The activity of both ERK1/2 and p38 kinase proteins were significantly increased in AAV9.GFP hearts compared with sham‐operated animals, as assessed by their phosphorylation levels. In contrast, the p38 and the ERK1/2 kinase activation was significantly decreased in AAV9.miR‐1–compared with AAV9.GFP treated animals. We did not observe significant changes in JNK or phospho‐JNK levels among the groups (Figure 5).

miR‐1 Inhibits Cardiac Fibrosis and Apoptosis

Since cardiac fibrosis and apoptosis are prominent features in the transition from compensatory hypertrophy to heart failure, we sought to examine the potential involvement of miR‐1 restoration in the regulation of cardiac ECM remodeling and apoptosis. Fibrosis is a pathological feature of cardiac adaptation to stress, where the proliferation of fibroblasts and increased deposition of ECM components results in myocardial stiffness and diastolic dysfunction,²⁸ and recently it has been demonstrated that miRNAs play a central role in the control of cardiac fibrosis and pathological LV remodeling.^29–31 Histological examination of LV sections by sirius‐red staining and subsequent quantification of the fibrotic area revealed that AAB induced a profound increase in interstitial fibrosis in the AAV9.GFP–treated hearts compared with sham‐operated hearts. In contrast, AAV9.miR‐1 treatment significantly decreased fibrosis (Figure 6A and 6B). In addition, we observed a significant reduction in mRNA expression of the profibrotic genes, the transforming growth factor beta‐1 (Tgfb1), and the connective tissue growth factor (Ctgf), in the AAV9.miR‐1– compared with AAV9.GFP–treated hearts (Figure 6C and 6D, respectively).

In addition, in response to long‐term pressure overload, cardiomyocyte apoptosis may further contribute to the transition from LV hypertrophy to heart failure.³² By Western blot analysis, we quantified the protein expression of the antiapoptotic gene, Bcl‐2, and the proapoptotic gene, Bax, which are involved in the apoptotic pathway (Figure 7A). Results showed that there was a significant increase in Bcl‐2 and a decrease in Bax expression in AAV9.miR‐1– compared with AAV9.GFP–treated hearts. Consequently, the Bcl‐2/Bax ratio, an important marker of myocardial cell survival probability,³² was significantly increased in the AAV9.miR‐1– compared with AAV9.GFP–treated hearts (Figure 7B). These findings were further corroborated in situ apoptosis detection by TUNEL assay, which demonstrated a marked decrease in DNA fragmentation of nuclei detected in AAV9.miR‐1– compared with AAV9.GFP–treated hearts (Figure 7C through 7D).

Identification of Fbln2 as a Direct Target of miR‐1

Previous studies have identified several direct targets of miR‐1, including calmodulin (Calm),¹⁸ insulin growth factor 1 (Igf1),¹⁹ Ncx1,³³ and twinfilin (Twf1),³⁴ which play a significant role in hypertrophy and heart failure. To gain additional insights into the molecular function of the miR‐1 in the setting of pressure‐overload–induced cardiac hypertrophy, we used computational and experimental approaches for the identification of novel, biologically relevant direct target genes. Using a target prediction algorithm,³⁵ we identified Fbln2 as putative target gene of miR‐1. The mRNA sequence of Fbln2 is predicted to contain a conserved “seed” sequence complementary to miR‐1 in the 3′‐UTR (Figure 8A). Fbln2 is an extracellular matrix protein that is highly expressed in the developing heart³⁶ and plays an important role during embryonic development³⁷ and cardiac remodeling.³⁸

We performed a luciferase‐based expression assays in vitro to verify that Fbln2 is a bona fide miR‐1 target. A short (≈250 bp) fragment from the 3′‐UTR of Fbln2 was cloned downstream of the stop codon of the Renilla luciferase in a dual‐luciferase reporter vector. In parallel, a fragment of the Ncx1 3′‐UTRs predicted to contain a conserved “seed” sequence complementary to miR‐1 was also cloned, serving a positive control for the luciferase assay. Each construct was cotransfected in HEK293 cells with synthetic miR‐1 mimics (pre–miR‐1) or control mimics (control pre‐miR). In the presence of miR‐1, we observed a significant decrease in luciferase activity of Fbln2 and Ncx1 constructs (Figure 8B). In contrast, cotransfection of a control miRNA did not result in a decrease in luciferase activity (Figure 8B). In addition, cotransfection of pre‐miR‐1 or control pre‐miR mimics with constructs containing the deleted 3′‐UTR sequences had no significant effect on the luciferase activity (Figure 8B).

In addition, we evaluated whether Fbln2 and Ncx1 expression was regulated by miR‐1 in vivo in the setting of pressure‐overload–induced hypertrophy. Western blot analysis demonstrated that at the protein level, the expression of both Fbln2 and Ncx1 was significantly decreased in the LV tissue of the AAV9.miR‐1– compared with AAV9.GFP–treated animals (Figure 8C through 8E). In addition to translational repression, miRNAs can lead to mRNA degradation of their targets.^39–40 We observed a significant decrease in the mRNA expression levels of Fbln2 and Ncx1 in the AAV9.miR1 animals compared with AAV9.GFP‐treated, as assessed by qRT‐PCR (Figure 8F). Furthermore, we verified that the expression of Igf1, Twf1, and Calm2 was decreased in AAV9.miR‐1– compared with AAV.9.GFP–treated animals (Figure 8G through 8I). Taken together, these results suggest that the restoration of miR‐1 levels in the setting of pressure‐overload–induced hypertrophy in vivo was paralleled by the modulation of a novel direct target gene, Fbln2, as well as the previously identified targets, Ncx1, Igf1, Twf1, and Calm2.