Bioinformatics prediction and experimental verification of a novel microRNA for myocardial fibrosis after myocardial infarction in rats

Establishment of experimental animal models

SPF male Sprague–Dawley (SD) rats weighing 200 ± 20 g were purchased from Beijing Vital River Laboratory Animal Technology Co. Ltd., Raised in the Barrier Animal Laboratory of Dongzhimen Hospital, Beijing University of Chinese Medicine, with a 12 h light/dark cycle and ad libitum food and water. All protocols and applications of animals in this study were approved by The Animal Care & Welfare Committee of Dongzhimen Hospital, Beijing University of Chinese Medicine (No: 16–25). The rats were divided into model group and sham-operated group by random number table. A model of AMI was established by ligating the left anterior descending coronary artery (LAD) of the rat heart. The rat was anesthetized intraperitoneally with 1% sodium pentobarbital (40 mg/kg). The skin was cut laterally at the third and fourth ribs of the left chest, approximately 1.5 cm in length. The musculature was bluntly separated, the rib was braced with a thoracotomy and the thymus was fixed upward to fully enlarge the surgical field, and the pericardium was carefully torn open to expose the ligation site of the rat heart. The anterior descending branch of the left coronary artery was ligated with a 5/0 line approximately 2 mm below the left atrial appendage and the arterial cone. The anterior myocardial wall was quickly ischemic and whitening, and the electrocardiogram (ECG) revealed saddle-back-like ST-segment elevations. After 24 h, the ECG was conducted again, and pathological Q waves appeared in the chest leads and lead I and aVL, indicating that the model was successful (Fig. 2A). The sham-operated rats were only threaded without ligation as a control group. There were five rats in each group and ten rats in total. The rats were fed adequate food and deionized water for 4 successive weeks. Finally, the rats were euthanized with 1% sodium pentobarbital (40 mg/kg), and myocardial tissue was collected at the infarct margin.

Histopathological staining

Dehydrated, paraffin-embedded myocardial tissues were cut into 3-μm-thick sections. The sections were dewaxed in xylene and rehydrated stepwise in a descending ethanol series. For hematoxylin-eosin (HE) staining, the sections were stained with hematoxylin staining solution and eosin staining solution (BLB-03 and YH-102, Beijing Kyushu Berlin Biotechnology Co. Ltd., Beijing, China). For Masson trichrome staining, the sections were dyed with nuclear staining, cytoplasmic staining, color separation solution, and counterstaining solution (D026-20, Nanjing Jiancheng Bioengineering Institute, Jiangsu, China) in sequence. The samples were dehydrated with ethanol in a series of concentrations and cleared with xylene. Coverslips were mounted with neutral balsam and sections were photographed by light microscopy. With the help of Image-Pro Plus 6.0, the area of myocardial collagen could be quantified. Collagen volume fraction (CVF) = Collagen area/(Collagen area + Myocardial area) × 100%. Five nonoverlapping fields were randomly selected for each section (×200), and the results are presented as the mean.

MiRNA gene chip

Due to the cost of microarray, three myocardial tissue samples were randomly selected from each group for microarray detection. Total RNA was extracted from myocardial tissue according to the mirVanaTM RNA Isolation Kit (AM1561, Applied Biosystems, Waltham, MA, USA). The Agilent Rat miRNA gene chip was used in this research. The hybridization and scanning of the miRNA gene chip were carried out. The standardized data were filtered, and at least one set of 100% probes marked as “Detected” in the two sets of samples used for comparison was left for subsequent analysis. The fold change value and p value (FC > 2, P < 0.05) based on a t test were used to screen differentially expressed miRNAs. The differentially expressed miRNAs were clustered in an unsupervised hierarchy, and the expression patterns of the differentially expressed miRNAs among different samples were displayed in the form of heatmaps. The microarray data discussed in this study have been deposited in the National Center for Biotechnology Information (NCBI) Gene Expression Omnibus (GEO) and are accessible through (GEO) Series accession number GSE208159.

MiRNAs bioinformatics analysis

TargetScan and microRNAorg databases were used to predict target genes of differentially expressed miRNAs, and the intersection was taken for subsequent analysis. Database for Annotation, Visualization and Integrated Discovery (DAVID) (https://david.ncifcrf.gov/) is a common software for target genes. The Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) databases were used to perform enrichment analyses to elucidate biological functions and pathways regulated by miRNAs (P < 0.05). The protein-protein interaction (PPI) network was constructed using the STRING database and Cytoscape software.

MiRNAs expression analysis

Total RNA was extracted from myocardial tissue using the mirVanaTM RNA Isolation Kit (AM1561; Applied Biosystems, Waltham, MA, USA). The RNA to be tested was reverse transcribed into cDNA using the miScript II Reverse Transcription Kit (Qiagen, Hilden, Germany). Using cDNA as a template, real-time PCR was performed with QuantiFast® SYBR® Green PCR Master Mix (Qiagen, Hilden, Germany). Reactions were incubated at 95 °C for 5 min, followed by 40 cycles of 95 °C for 10 s and 60 °C for 30 s. The expression levels of microRNAs were normalized to U6 and calculated using the 2^−ΔΔCT method. The microRNA-specific primer sequences were as follows: sense CCAGTGTTCAGACTACCTGTTC for rno-miR-199a-5p, sense CAAGGATGACACGCAAATTCG for U6.

Extraction and culture of primary cardiac fibroblasts

The 3-day-old Wistar suckling rats were purchased from Beijing Vital River Laboratory Animal Technology Co. Ltd. All protocols and applications of animals in this study were approved by The Animal Care & Welfare Committee of Dongzhimen Hospital, Beijing University of Chinese Medicine (No: 21–06). The 3-day-old Wistar suckling rats were sacrifice by cervical dislocation and disinfected with alcohol three times. The skin on the chest was exposed, and cut perpendicular to the ribs along the left margin of the breastbone of the infant rats at 0.5 cm with sterile ophthalmic straight scissors. The heart was extruded by gently squeezing the incision, and the apex of heart was cut off with a sterile ophthalmic curved scissors and quickly placed in precooled PBS (SH30256.01, Hyclone, Logan, UT, USA). Using the Neonatal Heart Dissociation Kit (130-098-373; Miltenyi Biotechnology, Bergisch Gladbach, Germany), the myocardial tissue was digested under aseptic conditions. The digested tissue was seeded in culture flasks with DMEM (C11995500BT; Gibco, WA, Australia) containing 15% fetal bovine serum (FBS) (10099141; Gibco, Hackett Drive Crawley, WA, Australia), 1% penicillin, and streptomycin in an incubator at 37 °C and 5% CO₂. After 90 min, the unattached cells were discarded, and the medium was replaced with 10% FBS-DMEM. The experiments used 3–5 generations of cardiac fibroblasts (CFs).

Immunocytochemistry

CFs were inoculated on 24-well polylysine-coated sterile slides with 8 × 10⁴ CFs per well. After 24 h incubation at 37 °C and 5% CO₂ incubator, CFs were sequentially fixed with 4% paraformaldehyde, permeated with 0.5% triton and inactivated endogenous peroxidase with 3% H₂O₂. CFs were incubated with anti-Vimentin (60330-1-Ig, proteintech) and anti-alpha-smooth muscle actin (α-SMA) (55135-1-AP, proteintech) antibodies overnight at 4 °C. After washing with PBS, the CFs were incubated with a secondary antibody. DAB chromogenic reagent was added in the dark environment, claybank was positive expression, and no chromogenic was negative expression. Hematoxylin restained the nucleus.

MiRNA transfection

CFs were digested and inoculated into culture plates. After 24 h, DMEM without FBS, growth factor, penicillin, and streptomycin was replaced and transfected with miR-199a-5p mimics (GenePharma, Suzhou, China). The CFs were cultured in an incubator at 37 °C and 5% CO₂ for 6 h and then the medium was replaced with 10% FBS-DMEM.

Dual-luciferase reporter assay

293T cells were co-transfected with rno-miR-199a-5p mimic or miR-NC and pMIR-REPORT Luciferase-GSK3β 3′UTR (WT) or pMIR-REPORT luciferase-GSK3β3′UTR (MUT), respectively. A total of 48 h after transfection, Luciferase activity was detected by microplate reader.

Western blot

CFs was homogenized with RIPA lysis buffer containing protease and phosphatase inhibitors. After SDS–PAGE, the proteins were transferred to polyvinylidene difluoride (PVDF) membranes. The membranes were blocked with 5% skim milk powder. The PVDF membranes were incubated with anti-GSK-3β (12456, CST) and anti-GAPDH (ab8245; Abcam, Cambridge, UK) antibodies overnight at 4 °C. After washing with Tris-buffered saline with 0.05% Tween-20 (TBST), the membranes were incubated with a secondary antibody. Membranes were placed in an enhanced chemiluminescence (ECL) mixture and then exposed, developed, and fixed in a dark room. Using GAPDH as the internal reference, the immunoreactive bands were analyzed using ImageJ analysis software.

Statistical analysis

The data are presented as the mean ± standard deviation (SD). Statistical analysis was performed with SPSS 20.0 and GraphPad Prism 8. When the data were normally distributed, an unpaired t test was conducted for comparisons between two groups. The nonparametric Kruskal–Wallis test was used when the data did not obey a normal distribution. Statistical significance was defined when P < 0.05.

Evidence of successful LAD ligation

HE staining showed that the structure of cardiomyocytes in the control group was complete, with uniform morphology, clear nuclei, a uniformly stained cytoplasm, and tightly arranged muscle bundles. However, in the model group, the myocardial tissue structure was destroyed, with a large area of cardiomyocyte fragmentation, nuclear pyrosis and deformation, disordered arrangement of muscle fibers, and widening of space (Fig. 2B). To investigate the collagen content, Masson trichrome staining was used. Compared with the control group, a large amount of collagen was interspersed between cardiomyocytes, and CVF was significantly increased in the model group (P < 0.0001, t = 28.21, df = 8, Figs. 2C and 2D). These results suggested that in rats with MI, a large number of necrotic cardiomyocytes were replaced by collagen, indicating that MF was formed.

MiRNAs expression profile and bioinformatics analysis

As shown in Figs. 3A and 3B and Table 1, thirteen differentially expressed miRNAs were screened using the Agilent Rat miRNA gene chip (FC > 2, P < 0.05), and their expression levels in the model group were higher than those in the control group. In the longitudinal sample clustering tree, the model group was clustered into one cluster, while the control group was clustered into another cluster, indicating that the expression levels of miRNAs between the two groups had prominent differences. In the horizontal sample clustering tree, miRNAs with similar biological functions were clustered. Red indicates that the expression of miRNA was upregulated, and blue indicates that the expression of miRNA was downregulated; the brighter the color, the greater the difference.

MiRNAs, small single-stranded noncoding RNAs (ncRNAs), can regulate the expression of target genes. In this study, two databases, microRNAorg and TargetScan, were used to predict target genes of the thirteen differentially expressed miRNAs, and 2155 common target genes were obtained (Fig. 3C). Through GO analysis of common target genes, the function of differential miRNAs could be described. As shown in Fig. 3D, GO included three parts: cellular component, molecular function, and biological process. Enrichment statistics were performed on the target genes in each GO entry (P < 0.01). KEGG enrichment analysis of common target genes was performed to predict the signaling pathways that differentially expressed miRNAs might target. As shown in Fig. 3E, 63 signaling pathways targeted by differentially expressed miRNAs were obtained, including the adipocytokine signaling pathway, chemokine signaling pathway, insulin signaling pathway, insulin resistance, MAPK signaling pathway, HIF-1 signaling pathway, TNF signaling pathway, and PPAR signaling pathway (P < 0.05).

MiRNA expression

Based on the miRNA expression profile, we selected miR-199a-5p that showed significantly upregulated expression for qRT-PCR validation in the myocardial tissue with MF post MI. As shown in Fig. 4A, the expression of miR-199a-5p in the model group was increased, which was consistent with the results of the gene chip (P = 0.01, t = 5.033, df = 4).

CFs were extracted and identified by immunocytochemistry. Vimentin is commonly used to characterize the cells of fibroblast origin, while α-SMA is used to recognize myofibroblasts (Tarbit et al., 2019). Immunohistochemical staining showed that Vimentin, located in the cell cytoplasm, was positively expressed in CFs, and α-SMA was negatively expressed (Fig. 4B).

To further confirm the role of miR-199a-5p in MF, miR-199a-5p mimics were transfected into CFs. After transfection, the expression level of miR-199a-5p was increased in the miR-199a-5p mimics group, indicating that the transfections were successful (P < 0.0001, t = 13.22, df = 8, Fig. 4C). These results suggested that miR-199a-5p was activated and played a crucial role in the progression of MF.

Bioinformatics analysis of miR-199a-5p

Studies have shown that the miR-199a family is closely related to CVD, suggesting that miR-199a-5p may play an important role in MF after MI. To further evaluate the biological function of miR-199a-5p, target genes prediction were performed using the microRNAorg and TargetScan databases, and 234 identical target genes were obtained (Fig. 5A). As shown in Fig. 5B, GO analysis included three parts: cellular component, molecular function, and biological process. Enrichment statistics were performed on the target genes in each GO entry (P < 0.01). KEGG analysis of miR-199a-5p target genes showed that a total of 7 signaling pathways (FC > 2, P < 0.05, Fig. 5C) were screened, such as GABAergic synapse, insulin signaling pathway, Lysosome, HIF-1 signaling pathway, retrograde endocannabinoid signaling, glycine, serine and threonine metabolism, nicotine addiction.

Interestingly, we found that insulin signaling pathway was prominent not only in the KEGG enrichment analysis of 13 miRNAs target genes, but also in the KEGG enrichment analysis of miR-199a-5p target genes. Therefore, insulin signaling pathway became the object of our study.

Target gene of miR-199a-5p

Above results indicated that miR-199a-5p could regulate multiple functions and signaling pathways through multiple target genes and suggested that miR-199a-5p might be involved in the process of MF after MI through the insulin signaling pathway. Although there are many target genes of miR-199a-5p, KEGG analysis showed that only eight target genes of miR-199a-5p were clustered in the insulin signaling pathway, including Ppp1cb, Gsk3β, Slc2a4, Prkag2, Rps6kb1, Rhoq, Crkl, and Eif4e. We performed PPI analysis on these target genes, the results showed that GSK-3β interacted with multiple proteins, such as Slc2a4, Rps6kb1, and Eif4e, which was of great significance in the insulin signaling pathway (Table 2, Fig. 6A). As shown in Fig. 6B, the binding site between rno-miR-199a-5p and GSK-3β was predicted by bioinformatics website (http://www.mirdb.org). By Dual-luciferase reporter assay, we indicated that overexpression of rno-miR-199a-5p inhibited the luciferase activity of wild-type GSK-3β 3′-UTR, but it also suppress luciferase activity of mutated one (P = 0.0001, t = 15.01, df = 4 or P = 0.0013, t = 8.03, df = 4, Fig. 6C). The result indicated that rno-miR-199a-5p could inhibit the expression of GSK-3β, but not by the predicted binding site. Rno-miR-199a-5p also regulated the luciferase expression of mutated GSK-3β 3′-UTR, it may be because in addition to the mutated binding sites, there may be other atypical binding sites or the overall downregulation of target gene expression may be caused by the indirect effect of overexpression of rno-miR-199a-5p. Furthermore, the protein expression of GSK-3β was decreased in CFs transfected with miR-199a-5p mimics (P < 0.0001, t = 28.72, df = 4, Figs. 6D and 6E).