Rat fibroblast Rat2 cells were purchased from American Type Tissue Collection (ATCC, Manassas, VA, USA). Cells were maintained in RPMI supplemented with 10% fetal bovine serum (FBS, HyClone, Logan, UT, USA), 50 U/mL penicillin, and 50 µg/mL streptomycin in a humidified atmosphere of 5% CO₂ at 37℃.

Rat2 cells were plated at a density of 10⁵ cells/mL in complete medium. After 24 hours, the cells were serum-starved with 0.5% FBS for 12 hours and treated with TGFβ (Abcam, Cambridge, MA, USA) for 24 hours before the TGFβ was washed out with complete medium for 24 hours (Fig. 1). The cells were homogenized in lysis buffer and separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). Then, the proteins were transferred to nitrocellulose membranes and the signals were detected using antibodies for collagen I, α-smooth muscle actin (αSMA, Abcam) and glyceraldehyde 3-phosphate dehydrogenase (GAPDH, Cell Signaling Biotechnology, Danvers, MA, USA).

Transfection with miRNA-26b mimics was performed in Rat2 cells using Lipofectamine 2000 (Thermo Fisher Scientific, Waltham, MA, USA) for 48 hours. The cells were then treated with TGFβ for 24 hours.

Seven-week-old male Sprague-Dawley rats were anesthetized with ketamine (80 mg/kg, intraperitoneally) and xylazine (8 mg/kg, intraperitoneally). Next, orotracheal intubation was performed with a 20-gauge intravenous catheter. Mechanical ventilation was maintained with a Mini-Vent ventilator (Harvard Apparatus, Holliston, MA, USA) containing 2% isoflurane and 98% O₂ at a tidal volume of 350 µL and a respiratory frequency of 160/min. The banding operation on the ascending aorta was performed with upper hemi-sternotomy and careful separation of the thymus. The ascending aorta was then freely dissected. Banding was performed with a Teflon (DuPont Pharmaceuticals, Wilmington, DE, USA) felt-supported 5–0 silk ligation around the ascending aorta with an 18-G blunted needle to prevent aortic rupture. The sternum was fixed, and the muscle layers and skin were closed with three layers of 4–0 silk sutures. The sham operation was performed using the same hemi-thoracotomy procedure without ligation. Ketoprofen (5 mg/kg, intramuscular, once daily) and gentamycin (5 mg/kg, intramuscular, once daily) were administered postoperatively over a 3-day period.

Rats in the banding group were euthanized 6 weeks after banding. Rats in the debanding group underwent a debanding operation 28 days after the banding operation and were sacrificed 14 days after the debanding procedure. The debanding procedure was performed via right-sided, muscle-saving thoracotomy in the second intercostal space, with the banded silk on the Teflon felt being cut to avoid aortic rupture. The thymus was repositioned, and the chest was closed in three layers. The animals in the sham group underwent a similar thoracotomy procedure and were sacrificed postoperatively at 6 weeks. Each group included eight rats, and the hearts of all subjects were extracted at 13 weeks.

The study protocol was approved by the Ethics Committee of Daejeon St. Mary’s Hospital, College of Medicine, The Catholic University of Korea, Daejeon, and by the Republic of Korea (DMDDJ-AP-2015-09).

After euthanasia, the rat hearts were removed, fixed in 4% formaldehyde, and then embedded in paraffin. Three mid-ventricular sections were stained with hematoxylin-eosin for histological analyses, with periodic acid–Schiff (PAS) stain for cardiomyocyte size determination, and with Masson’s trichrome for fibrosis measurement (Fig. 2). The histopathological evaluation of each slide was performed by a blinded pathologist. The degree of fibrosis was quantified using ImageJ software version 1.44 (U.S. National Institutes of Health, Bethesda, MD, USA) by determining the blue-stained area on the Masson’s trichrome-stained slides. Additionally, the cross-sectional areas of at least 10 myocytes were measured from 200-fold-enlarged images of the PAS-stained slides using the same software. Nucleus-centered myocytes in the endocardial layer were assessed.

Echocardiographic examinations were performed twice on all subjects, at baseline and 6 weeks after banding (except for the debanding group). In the debanding group, additional echocardiographic examinations were performed 4 weeks after banding and 2 weeks after the debanding procedure. Animals were anesthetized with ketamine (80 mg/kg, intraperitoneally) and xylazine (8 mg/kg, intraperitoneally) during the echocardiographic procedures. To achieve a physiologically relevant depth of sedation during echocardiography, the heart rate was maintained at approximately 270 to 320 beats per minute. Conventional echocardiography and tissue Doppler image acquisition were performed using a 13 MHz linear array transducer (GE i13 L, Vivid E9, General Electric, Waukesha, WI, USA). Images were acquired at 35 to 45 frames/beat (110 to 190 frames/sec) using the shallowest possible depth and the smallest possible sector size [12]. Effective banding procedures were confirmed by an echocardiographic finding of maximal peak velocity at the ligation site (i.e., greater than 3 m/sec). Relieved constriction of the ascending aorta after the debanding procedure was also documented through Doppler findings. Echocardiographic imaging findings, including M-mode images and global longitudinal strain (GLS) measurements from the parasternal long-axis view, were used to measure cardiac remodeling and systolic and diastolic left ventricular function (Supplementary Material) [13-16]. A single investigator who was blinded to the animal groups performed all image acquisitions and offline measurements.

LV myocardium samples were stored in RNAlater solution (Ambion, Austin, TX, USA). Frozen tissues were homogenized, followed by centrifugation for 10 minutes at 12,000 rpm at 4℃. CTGF, TGFβ1, and MMP2 levels in LV myocardial tissues were quantified via ELISA according to the manufacturer’s instructions: CTGF (Aviva Systems Biology, San Diego, CA, USA), MMP2 (Abnova, Taipei, Taiwan), and TGFβ1 (Abnova). The optical density of the samples was determined at 450 nm in an automated microplate reader (Bio-Rad Laboratories, Hercules, CA, USA).

To analyze the expression of CTGF in Rat2 cells, total RNA was isolated using the easy-BLUE Total RNA Extraction Kit (iNtRON, Seongnam, Korea). The cDNA was synthesized using the Reverse Transcriptase Premix (Elpis Biotech., Daejeon, Korea) and was amplified in a Power SYBR^® Green PCR Master Mix (Applied Biosystems, Warrington, UK) with CTGF-specific primer pairs (F: 5’-AAGACACATTTGGCCCTGAC-3’, R: 5’-GAGTGTCTGGGGGCAGAGTA-3’) and GAPDH-specific primer pairs (F: 5’-CCCCTTCATTGACCTCAACTACA-3’, R: 5’-TCCCATTCTCAGCCTTGACTGT-3’) as the endogenous control. miRNAs were extracted from the myocardial tissues and Rat2 cells using the miRCURY RNA Isolation Kit (Exiqon, Woburn, MA, USA). Reverse transcription was performed with the miScript II RT Kit (QIAGEN, Hilden, Germany). A miRNA-specific miScript Primer Assay was purchased from QIAGEN, and real-time polymerase chain reaction was performed on a 7500 Fast Real-Time PCR System (Applied Biosystems, Foster City, CA, USA). The fold change for each miRNA was calculated using the comparative Ct (2−ΔΔCt) method, and RNU48 (a small nuclear RNA) served as an endogenous control. All reactions were performed in triplicate for each sample.

Statistical analysis was performed using commercially available software SPSS version 20 (IBM Co., Armonk, NY, USA). Differences between groups were assessed using 2-tailed Student’s t tests. Differences among three groups were assessed using one-way analysis of variance with post hoc Tukey tests. The data are presented as the mean value ± standard deviation. A p < 0.05 was considered to indicate statistical significance.

Online miRNA databases such as www.targetscan.org were used to predict several miRNAs that regulate the expression of CTGF. Using the miRNA databases, we found several conserved miRNAs and compared their expression to that of CTGF; ultimately, we chose miRNA-26b, miRNA-133a, and miRNA-19b with a probability of preferentially conserved targeting greater than 0.69 [17].

Based on echocardiography, concentric LVH resulted from the ascending aortic banding operation and was relieved after the debanding operation (relative wall thickness [RWT]: 0.41 ± 0.01 vs. 0.81 ± 0.13 vs. 0.64 ± 0.09 in the sham vs. banding vs. debanding group, respectively, p < 0.001) (Table 1, Fig. 2). Heart weight/tibia length did not differ significantly between the debanding and sham groups, indicating that the debanding operation resulted in LVH regression. Compared with the LV mass of the banding group, the LV mass, which was calculated using the cubed formula based on echocardiography, of the debanding group was not significantly decreased, as the LV chamber tended to be larger because of relieved concentric LVH after the debanding procedure (Table 1).

Reactive myocardial fibrosis was significantly increased in the banding group but was not significantly relieved by the debanding operation until 2 weeks after the operation (Table 2). Compared with the other groups, the banding group exhibited hyperdynamic LV systolic function. Additionally, E/E’ was significantly increased in the banding group (Table 1).

The three-layered global circumferential strain and radial strain did not differ significantly among the groups. However, compared with the GLS in the sham group, the GLS in the banding and debanding groups was significantly increased (Table 2). In addition, compared with that in the banding group, the GLS in the debanding group was not significantly improved (GLS: –18.7 ± 1.0 vs. –15.2 ± 2.3 vs. –15.5 ± 2.7 for the sham vs. banding vs. debanding group, p = 0.008), possibly because myocardial fibrosis was not significantly improved in the debanding group (Fig. 3).

To confirm the results from the in vivo test, CTGF, miRNA-26b, miRNA-133a, and miRNA-19b were assessed in both the banding model (a pressure remodeling model) and the debanding model (a reverse remodeling model).

CTGF levels were significantly higher in the banding group than in the sham and debanding groups (75.8 ± 7.6 pg/mg vs. 98.2 ± 19.4 pg/mg vs. 64.5 ± 14.1 pg/mg in the sham vs. banding vs. debanding group, respectively, p = 0.002). MMP2 and TGFβ1 levels tended to increase in the banding group and were significantly decreased in the debanding group (MMP2: 426.0 ± 11.8 pg/mg vs. 573.3 ± 122.2 pg/mg vs. 259.6 ± 67.1 pg/mg in the sham vs. banding vs. debanding group, respectively, p < 0.001; TGFβ1: 64.7 ± 9.5 pg/mg vs. 89.8 ± 42.2 pg/mg vs. 47.0 ± 13.2 pg/mL, respectively, p = 0.002) (Fig. 3).

We analyzed the levels of miRNA-26b, miRNA-133a, and miRNA-19b using RT-PCR. The relative expression level of miRNA-26b was higher in the debanding group than that in the banding group (miRNA-26b: 1.0 ± 0.6 vs. 0.6 ± 0.2 vs. 6.0 ± 3.01 in the sham vs. banding vs. debanding group, respectively, p < 0.001) (Fig. 4).

We observed elevation of CTGF after 24 hours of treatment with TGFβ in Rat2 fibroblasts. CTGF levels were observed after 3, 6, 12 and 24 hours of TGFβ washing. After 24 hours of TGFβ washing, CTGF was significantly decreased.

Markers of fibrosis, specifically collagen I, αSMA and GAPDH, were significantly increased after 18 hours of treatment with TGFβ and decreased after 24 hours of TGFβ washing (Fig. 1).

Based on the results, we divided the cells into three groups: the control group, the remodeling group with elevated CTGF levels following 24 hours of TGFβ treatment, and the reverse remodeling group with decreased CTGF levels following 24 hours of TGFβ washing. miRNA-26b was significantly increased in the TGFβ-washed group compared with the control and TGFβ-stimulated groups. However, miRNA-19b was significantly decreased in the TGFβ-stimulated group and TGFβ-washed group compared with the control group (Fig. 1).

Transfection of fibroblasts with miRNA-26b mimic decreased the relative expression of CTGF. Stimulation of fibroblasts with TGFβ1 increased the relative expression of CTGF and decreased that of miRNA-26b. TGFβ1 washing decreased CTGF expression and increased miRNA-26b expression (Fig. 5).