All animal operations were in accordance with the guidelines on the use and care of laboratory animals for biomedical research (National Institutes of Health; no. 85-23; revised 1996). The experimental protocol was approved by the Ethics Committee of The First Affiliated Hospital of Nanjing Medical University (Nanjing, China; approval no. 2019-041).

Mouse HL-1 cardiomyocytes (The Cell Bank of Type Culture Collection of The Chinese Academy of Sciences) were cultured at 37°C in DMEM supplemented with 10% FBS (Gibco; Thermo Fisher Scientific, Inc.) in a humidified atmosphere containing 5% CO₂. For induction of apoptosis, different concentrations of doxorubicin (2.5, 5 and 10 µM; Teva UK Ltd.) were added to the medium for 6 h of culture at 37°C. HL-1 cardiomyocytes in the control group (Con) were treated with the same dose of saline (Gibco; Thermo Fisher Scientific, Inc.).

Full-length polypyrimidine tract binding protein 1 (PTBP1) or transgelin 2 (TAGLN2) was subcloned into pcDNA3.1 to overexpress PTBP1 or TAGLN2, respectively, and empty pcDNA3.1 was used as a control. Similarly, Bax was subcloned into pcDNA3.1 to overexpress Bax to detect if fibrotic events were independent of apoptosis. miR-133b mimics (miR-133b) and a negative control (miR-NC) were used to overexpress miR-133b. The sequence of the miR-133b mimics was 5′-UUU GGU CCC CUU CAA CCA GCU A-3′, and the sequence of the scrambled miR-NC was 5′-UCA UUG CUA UGC ACU GUC CAC C-3′. All pcDNA3.1 vectors (1 µg) and miR-133b mimics (50 nM) were purchased from Shanghai GenePharma Co., Ltd., and transfected in HL-1 cardiomyocytes using Lipofectamine^® 3000 (Invitrogen; Thermo Fisher Scientific, Inc.) at room temperature for 48 h. After 48 h from transfection, subsequent experiments were conducted.

HL-1 cardiomyocyte apoptosis was tested using a PI/FITC Annexin V Apoptosis Detection kit (BD Pharmingen; BD Biosciences). Briefly, HL-1 cardiomyocytes (5×10⁵ cells/ml) were washed with cold PBS (Gibco; Thermo Fisher Scientific, Inc.) twice and then resuspended in 1X binding buffer. Subsequently, cell solution (3×10⁴ cells; 100 µl) was added to a culture tube. Then, 5 µl FITC Annexin V and 5 µl of PI were added at room temperature in the dark for 15 min. Finally, HL-1 cardiomyocytes were analyzed using an Attune NxT flow cytometer (Invitrogen; Thermo Fisher Scientific, Inc.), and apoptotic cardiomyocytes were calculated using FlowJo software (version 10.0; FlowJo LLC). The apoptotic rate was calculated as follows: (Cell number in Q3/cell number in all quadrants) ×100%, with cells in Q3 representing early apoptotic cells.

Following the manufacturer's instructions, a miRNeasy Mini kit (Qiagen, Inc.) was used to extract RNA from HL-1 cardiomyocyte and mouse heart samples. The iScript™ cDNA Synthesis kit (Bio-Rad Laboratories, Inc.) was used to reverse transcribe RNA (400 ng) to cDNA according to the manufacturer's protocol. The relative mRNA expression was determined using SYBR qPCR (Bio-Rad Laboratories, Inc.) with an ABI-7900 Real-Time PCR Detection System (7900HT; Applied Biosystems; Thermo Fisher Scientific, Inc.). The relative miR-133b expression was analyzed using Bulge-Loop™ miRNA qPCR Primer Set (Guangzhou RiboBio Co., Ltd.) with an ABI-7900 Real-Time PCR Detection System. qPCR was performed with the following thermocycling conditions: 95°C for 30 sec, followed by 40 cycles at 95°C for 5 sec and 60°C for 30 sec, and a melting curve analysis protocol (60-95°C with temperature increments of 0.2°C every 10 sec). GAPDH and U6 were used as internal controls for mRNA and miRNA, respectively. The relative gene expression level was calculated using the 2^−ΔΔCq method (29). The primer sequences were as follows: PTBP1 forward, 5′-GAT CCG ACG AGC TCT TCT C-3′ and reverse, 5′-CTA TCG TTT CCA TTG GCT GC-3′; TAGLN2 forward, 5′-CCA ACT GGT TTC CTA AGA AAT CC-3′ and reverse, 5′-CAA TCA CGT TCT TGC CCT C-3′; LHFP forward, 5′-TTT GAC TTG GGC AAG TGT G-3′ and reverse, 5′-CAT GTA CAC AGC AGC ATG G-3′; SEC61B forward, 5′-TTG CTC TTC CCA GCT TCT C-3′ and reverse, 5′-TGT AGA ATC GCC ACA TCC C-3′; CETN3 forward, 5′-TTA GCT CTG AGA GGT GAG C-3′ and reverse, 5′-TTC TTG CTT CTG TTC TTC AGA G-3′; FTL forward, 5′-CCG TGC ACT CTT CCA GGA TGT-3′ and reverse, 5′-CCT TAT CCA GAT AGT GGC TTT CCA G-3′; LDLRAP1 forward, 5′-TGA CAC TCA AGG TGT CAC C-3′ and reverse, 5′-TAG GAG ATC CTG TAA ATG GAC AC-3′; CLTA forward, 5′-ATA CTA CCAG GAG AGC AAT GG-3′ and reverse, 5′-TAC TTT CAG GCT CTG ACT GC-3′; GAPDH forward, 5′-CAT CTTC TTGT GCAG TGC C-3′ and reverse, 5′-CAA ATC CGT TCA CACC GAC-3′; miR-133b forward, 5′-TTG GTC CCC TTCA ACCA GCT A-3′ and reverse, 5′-CAG TGC GTG TCG TGG AGT-3′; and U6 forward, 5′-CTC GCT TCG GCA GCA CA-3′ and reverse, 5′-ACG CTT CAC GAA TTT GCG T-3′.

HL-1 cardiomyocytes and mouse cardiac tissues were lysed using RIPA lysis buffer (Beyotime Institute of Biotechnology). A Bicinchoninic Acid Protein Assay kit (Thermo Fisher Scientific, Inc.) was used to evaluate the concentration of protein samples (1 µl). Subsequently, equal amounts of protein samples (50 µg/lane) were subjected to 12% SDS-PAGE and transferred to PVDF membranes, which were then blocked with 5% skimmed milk at 4°C overnight. Primary antibodies (1:1,000) against cleaved caspase-3 (cat. no. ab2302), Bcl-2 (cat. no. ab32124), Bax (cat. no. ab32503), collagen I (cat. no. ab34710), collagen III (cat. no. ab7778), fibronectin (cat. no. ab2413), collagen IV (cat. no. ab214417), PTBP1 (cat. no. ab133734), TAGLN2 (cat. no. ab121146) and GAPDH (cat. no. ab181602) were then incubated with the membranes at 4°C overnight. Collagen I, III and IV, and fibronectin are extracellular matrix (ECM) proteins. Finally, HRP-conjugated secondary antibodies (cat. no. ab6721; 1:5,000) were added to the membranes in the dark at room temperature for 1 h. All antibodies were obtained from Abcam. Proteins were visualized using an ECL Chemiluminescence kit (Thermo Fisher Scientific, Inc.) and quantified using ImageJ software (version 1.46; National Institutes of Health). GAPDH served as a loading control. The gray ratio of the target bands to the internal reference bands was the relative expression of the target proteins.

The RIP assay was performed using a Magna RIP™ RNA-Binding Protein Immunoprecipitation kit (EMD Millipore) according to the manufacturer's protocol. After centrifugation at 1,000 × g at 4°C for 5 min, HL-1 cardiomyocytes were lysed in complete RIPA lysis buffer (cat. no. P0013B; Beyotime Institute of Biotechnology). Subsequently, 100 µl supernatant was collected and incubated with magnetic beads (1 mg) conjugated with PTBP1 antibody (cat. no. ab133734; 1:10,000; Abcam) or control IgG (cat. no. 12-370; 1:5,000; EMD Millipore) at 4°C for 6 h. After the beads were washed, the complexes were treated with Proteinase K to remove proteins. Beads were isolated from the supernatant after centrifugation at 2,500 × g for 5 min at 4°C and washed with washing buffer (10 mM Tris-HCl pH 7.5, 1 mM EDTA, 2 M NaCl and 0.1% Tween-20), followed by another centrifugation step at 2,500 × g for 5 min at 4°C. Finally, the purified RNA level was measured by RT-qPCR, as aforementioned.

Based on the TargetScan version 7.2 website (http://www.targetscan.org/vert_72/), 703 mRNAs were predicted to harbor binding site(s) for miR-133b (data not shown). According to the starBase v2.0 website (http://starbase.sysu.edu.cn/), PTBP1 was predicted to bind to the 3′-untranslated region (UTR) of TAGLN2 (data not shown). According to the RNA-Protein Interaction Prediction (http://pridb.gdcb.iastate.edu) based on the random forest (RF) and support vector machine (SVM) algorithm, scores of the combination between PTBP1 and TAGLN2 3′-UTR, and between PTBP1 and TAGLN2 were evaluated.

A fragment of the PTBP1 or TAGLN2 3′-UTR containing the predicted binding site of miR-133b was cloned into the pmirGLO vector (Promega Corporation) to construct the wild-type pmirGLO-PTBP1-3′-UTR or pmirGLO-TAGLN2-3′-UTR vectors. The mutant pmirGLO-PTBP1-3′-UTR or pmirGLO-TAGLN2-3′-UTR vectors were also generated by subcloning the mutated PTBP1 or TAGLN2 sequences in the 3′-UTR complementary to miR-133b into the pmirGLO vectors. The vectors were commercially obtained from Shanghai GenePharma Co., Ltd., and then separately transfected with miR-NC or miR-133b mimics using Lipofectamine 3000, as aforementioned. The luciferase activity was evaluated using a dual luciferase reporter assay system (Promega Corporation) 48 h after transfection, and was normalized to Renilla luciferase activity.

A total of 6×10⁶ cells were cultured in a 100-mm culture dish. After discarding the culture medium, cells were washed twice with PBS. Next, cells were digested using 0.5 ml trypsin followed by addition of 2 ml PBS. Cells were transferred to EP tubes followed by centrifugation at 1,000 × g at 4°C for 2 min and were subsequently treated with 100 µl buffer solution for 15 min. After another centrifugation at 1,000 × g at 4°C for 5 min, the supernatant was collected. The TAGLN2 sense or antisense sequence was transcribed using T7 RNA polymerase (cat. no. AM2718; Ambion; Thermo Fisher Scientific) and then purified using the RNeasy Plus Mini kit (cat. no. 74134; Qiagen GmbH) according to the manufacturer's protocol. Subsequently, the TAGLN2 sense or antisense sequence was labeled with Biotin RNA Labeling Mix (Ambio Life). Subsequently, a Magnetic RNA-Protein Pull-Down kit (cat. no. 20164; Pierce; Thermo Fisher Scientific, Inc.) was used for RNA pulldown assays according to the manufacturer's instructions. Beads were isolated from the supernatant after boiling, and the PTBP1 enrichment was analyzed by western blotting.

A total of 50 C57BL/6 mice (20-25 g; 8 weeks old; male; Beijing Vital River Laboratory Animal Technology Co., Ltd.) were used in the present study and were grouped as follows: Sham group (n=15), DOX+AAV-NC group (n=15) and DOX+AAV-miR-133b group (n=20). Mice were raised with ad libitum access to water and food in a temperature-controlled room (22±2°C) with a 12-h light/dark cycle and a relative humidity of 40-60%. A cardiac injury mouse model was generated by chronic intraperitoneal injections (on days 0, 2, 4 and 6) of 4 mg/kg doxorubicin. The mice in the sham group were treated with the same dose of PBS (Invitrogen; Thermo Fisher Scientific, Inc.). All mice were euthanized 4 weeks after the first injection of doxorubicin or PBS. For AAV (serotype 9) injection, mice received a single-bolus injection of AAV-miR-133b or empty AAV (AAV-NC) resuspended in PBS via the tail vein at 1×10¹¹ viral genomes per animal. AAV vectors were purchased from Han Heng Biotechnology (Shanghai) Co., Ltd. One week after injection of AAVs, the mice were treated with doxorubicin or PBS. The mice were anesthetized with 1.5% pentobarbital sodium (60 mg/kg) by intraperitoneal injection and then sacrificed by cervical dislocation under anesthesia. The whole experiment lasted 5 weeks, beginning from AAV injection and ending up with the euthanasia of mice.

Four weeks after the first injection of doxorubicin or PBS, the mice were anesthetized with 2% isoflurane and placed in the supine position. The left ventricular end-systolic diameter (LVESD) and left ventricular end-diastolic diameter (LVEDD) values were analyzed using a VEVO 770 high-resolution system (VisualSonics, Inc.) with a RMV 704 probe at a frequency of 40 MHz. Left ventricular ejection fraction (LVEF) and left ventricular ejection shorting (LVES) were automatically calculated using an echocardiographic instrument.

Cardiac tissue samples were immobilized in 4% paraformaldehyde for 48 h at 4°C, embedded in paraffin and sliced into sections (5-µm-thick). Subsequently, sections were subjected to Masson's trichrome staining as previously described (30). The collagen fibers were blue, the muscles and elastic fibers were red, cellulose was purple-red, and the nucleus was blue-brown. Images were taken using a light microscope Nikon model with a Spot Insight camera (magnification, ×200). The fibrotic region was quantified using ImageJ software (version 1.46; National Institutes of Health). The percentage of fibrosis was calculated as fibrosis areas/total left ventricular areas ×100%.

The collagen content was determined using a quantitative dye-binding method, as previously described (30). A Sircol assay (Biocolor Ltd.) was used to analyze cardiac tissues following the manufacturer's instructions. The mouse heart was weighed and homogenized with pepsin, and collagen content in each sample was quantified using the Gen5™ BioTek software (Agilent Technologies, Inc.).

The data are displayed as the mean ± SD of 3 independent experiments. Statistical analysis was performed using SPSS software (version 21.0; IBM Corp.). All data were assessed for normal distribution (Shapiro-Wilk test) and homogeneity of variance (Bartlett's test). Statistical analysis was performed using unpaired Student's t-test for comparisons between two groups or one-way ANOVA followed by Dunnett's or Tukey's post-hoc tests for comparisons among multiple groups. P<0.05 was considered to indicate a statistically significant difference.

As shown in Fig. 1A, the apoptosis rate of HL-1 cardiomyocytes was significantly increased by treatment with doxorubicin (2.5, 5 and 10 µM). Considering that 10 µM doxorubicin resulted in the highest apoptosis rate of HL-1 cardiomyocytes, 10 µM doxorubicin was selected for subsequent assays. However, the role of miR-133b in doxorubicin-induced cardiac injury remains unclear. According to the RT-qPCR results, doxorubicin treatment significantly decreased miR-133b expression in HL-1 cells (Fig. 1B). A mouse model of cardiac injury was then established by injection of doxorubicin. Data from echocardiography analysis demonstrated that the LVEDD and LVESD were significantly increased by doxorubicin injection (Fig. 1C and D), while LVEF and LVES were significantly decreased by doxorubicin injection (Fig. 1E and F). Additionally, miR-133b expression was significantly downregulated in the DOX group compared with in the sham group (Fig. 1G). All these experimental results suggested that miR-133b expression was decreased by doxorubicin treatment in vitro and in vivo.

The biological function of miR-133b in doxorubicin-treated HL-1 cardiomyocytes was then explored. First, miR-133b was overexpressed by transfection using miR-133b mimics in doxorubicin-treated HL-1 cardiomyocytes (Fig. 2A). miR-133b overexpression significantly decreased the apoptosis rate of doxorubicin-treated HL-1 cardiomyocytes (Fig. 2B). Moreover, western blot assays revealed that miR-133b overexpression triggered a significant increase in Bcl-2 protein expression, but a significant decrease in Bax and cleaved caspase-3 protein expression (Fig. 2C and D). As shown in Fig. 2E and F, overexpression of miR-133b significantly decreased the protein expression levels of collagen I, III and IV, and fibronectin in the doxorubicin-treated HL-1 cardiomyocytes. Moreover, it was revealed that BAX was successfully overexpressed and that BAX overexpression had no significant effects on the expression levels of ECM proteins (Fig. S1A and B). Overall, miR-133b overexpression inhibited the apoptosis and collagen accumulation in the doxorubicin-treated HL-1 cardiomyocytes.

To explore the role of miR-133b in vivo, a mouse model of doxorubicin-induced heart failure was established. First, it was observed that miR-133b expression in the AAV-miR-NC group was significantly decreased compared with in the sham group, and miR-133b expression was significantly increased in cardiac tissues by AAV-miR-133b injection (Fig. 3A). In addition, the injection of doxorubicin induced a significant decrease in Bcl-2 protein expression, but a significant increase in Bax and cleaved caspase-3 protein expression, and miR-133b overexpression reversed these effects (Fig. 3B). Furthermore, Masson's trichrome staining indicated that miR-133b overexpression significantly attenuated doxorubicin-induced cardiac fibrosis and collagen deposition (Fig. 3C-F). Similarly, western blot assays suggested that the doxorubicin-induced augmentation of the protein expression levels of collagen I, III and IV, and fibronectin were also antagonized by miR-133b overexpression (Fig. 3G). In conclusion, overexpression of miR-133b alleviated apoptosis and cardiac fibrosis in vivo.

Previously, miR-133b has been widely reported to bind to the 3′-UTR of mRNAs to exert its biological functions (26,31). Hence, it was hypothesized that miR-133b functioned in the same way in HL-1 cardiomyocytes. According to the TargetScan website, 703 mRNAs were predicted to harbor binding site(s) for miR-133b (data not shown). The top 8 mRNAs (LHFP, SEC61B, CETN3, PTBP1, FTL, LDLRAP1, CLTA and TAGLN2) with the highest total context score were chosen for subsequent experiments. According to RT-qPCR, PTBP1 and TAGLN2 expression was significantly upregulated in the doxorubicin-treated HL-1 cardiomyocytes compared with in control cells, suggesting that PTBP1 and TAGLN2 may be target mRNAs of miR-133b (Fig. 4A). Subsequently, the binding site between miR-133b and PTBP1 or TAGLN2 was predicted using the TargetScan website (Fig. 4B). Luciferase reporter assays demonstrated that the luciferase activity of wild-type pmirGLO-PTBP1-3′-UTR or pmirGLO-TAGLN2-3′-UTR was significantly inhibited by miR-133b mimics, while the mutant pmirGLO-PTBP1-3′-UTR or pmirGLO-TAGLN2-3′-UTR displayed no changes (Fig. 4C). Additionally, the expression levels of PTBP1 and TAGLN2 were significantly upregulated in the DOX group compared with in the sham group (Fig. 4D). In addition, miR-133b overexpression significantly decreased TAGLN2 and PTBP1 protein expression in HL-1 cardiomyocytes (Fig. 4E). Coincidently, according to the starBase website, PTBP1 was predicted to bind to the 3′-UTR of TAGLN2 (data not shown). The interaction between the PTBP1 protein and the TAGLN2 3′-UTR was then evaluated. Data from RNA-Protein Interaction Prediction revealed that the RF classifier score and SVM classifier score of the combination between PTBP1 and TAGLN2 3′-UTR were 0.7 and 0.29, respectively, while the scores of the combination between PTBP1 and TAGLN2 were 0.8 and 0.26, respectively (Fig. 4F). Subsequently, an RNA pulldown assay indicated that PTBP1 was enriched only in the input group (Fig. 4G), and a RIP assay demonstrated that neither anti-IgG nor anti-PTBP1 immunoprecipitated TAGLN2 (Fig. 4H), indicating that PTBP1 was unable to bind to the TAGLN2 3′-UTR. Thus, PTBP1 and TAGLN2 served as targets of miR-133b.

To validate whether miR-133b modulated apoptosis and collagen deposition by targeting PTBP1 and TAGLN2 in the doxorubicin-treated HL-1 cardiomyocytes, rescue assays were performed via PTBP1 and TAGLN2 overexpression. PTBP1 and TAGLN2 overexpression efficiency was verified by RT-qPCR (Fig. S1C and D). The protein expression levels of PTBP1 and TAGLN2 were significantly increased by overexpression of PTBP1 and TAGLN2, respectively, following miR-133b overexpression in the doxorubicin-treated HL-1 cardiomyocytes (Fig. 5A and B). In addition, overexpression of either PTBP1 or TAGLN2 rescued the miR-133b mimic-induced decreased apoptosis rate of the doxorubicin-treated HL-1 cardiomyocytes (Fig. 5C). Moreover, the increase in Bcl-2 expression and the decrease in Bax and cleaved caspase-3 expression resulting from miR-133b overexpression were reversed by the overexpression of PTBP1 or TAGLN2 (Fig. 5D). Additionally, the suppressive effects of miR-133b on the protein expression levels of collagen I, III and IV, and fibronectin were abrogated by PTBP1 or TAGLN2 overexpression (Fig. 5E). All these results confirmed that overexpression of PTBP1 or TAGLN2 reversed the effects of miR-133b on apoptosis and collagen accumulation.