Calycosin was purchased from the Nature Standard Technical Service Co., Ltd. (Shanghai, China). Fosinopril, DMEM, FBS, trypsin, penicillin, streptomycin, sodium cacodylate buffer, and DAPI were purchased from the Beijing BioDee Biotechnology Co., Ltd. (Beijing, China). Paraformaldehyde (4%) and saline (0.9%) were from Applygen Technology Inc. (Beijing, China). Dimethyl sulfoxide (DMSO) was acquired from Sigma-Aldrich LLC (Shanghai, China). LY294002 was purchased from Abmole China Branch. All other chemicals were purchased from commercial sources.

After 1 week of acclimation, Sprague–Dawley (SD) male rats (220 g) obtained from the Beijing Vital River Laboratory Animal Technology Co., Ltd. were randomly divided into four groups (number/each group = 8): sham group, model group, calycosin (Cal) treatment group, and fosinopril treatment group. Rats in the sham group underwent sham surgery, while HF was induced in other rats by direct left anterior descending (LAD) artery ligation as described in our previous study (Zhang et al., 2018). Based on our previous literature (Zhang et al., 2019; Wang et al., 2020a), 24 h after surgery, the acute myocardial infarction model was established, and the drug treatment was started. Rats in the Cal group were treated with Cal at a dosage of 80 mg/kg per day (Li et al., 2020). Rats in the fosinopril group were treated with fosinopril at a dosage of 4.67 mg/kg per day (Wang et al., 2020b). Rats in the sham group and model group were treated with the same volume of distilled water. All the drugs and distilled water were orally administrated with an amount of l ml/100 g for 28 days. It is reported that based on available clinical evidence, fosinopril is an effective and well-tolerated option for the management of patients with heart failure (Davis et al., 1997). For this reason, we set the fosinopril group as a positive control. During the whole procedure, the total mortality of the rats was 30%; most deaths occurred during surgery or after surgery, possibly owing to acute pump failure or fatal arrhythmia. This study conforms to the Guide for the Care and Use of Laboratory Animals published by the US National Institutes of Health (NIH publication no. 85–23, revised 1996), and it was approved by the Institutional Animal Care and Use Committee at the Beijing University of Chinese Medicine (consent number: BUCM20200914-YW).

Transthoracic echocardiography was performed by using a Vevo 2100 instrument (VisualSonics, Canada) equipped with an MS-400 imaging transducer. The echocardiographic measurements were performed under general anesthesia with 1% pentobarbital sodium. M-mode tracings were recorded through the anterior and posterior left ventricular (LV) walls at the papillary muscle level. Heart functions were assessed by related parameters including left ventricular internal dimension—systole (LVID; s), left ventricular internal dimension—diastole (LVID; d), ejection fraction (EF), and fractional shortening (FS). Three cardiac cycles were recorded for the measurements.

Hearts were cut horizontally through the mid region to create cross sections of both the left and right ventricles, and the apex part was fixed with 4% paraformaldehyde. Then the tissues were embedded in paraffin and cut into 5-μm sections. After deparaffinized by xylene and rehydrated via different grades of ethanol, the sections were stained with hematoxylin–eosin (HE) staining and Masson staining to assess overall pathological changes. Digital images were observed under a microscope at ×400 magnification (Leica Biosystems Richmond, Inc.). Inflammatory cell rate is evaluated by the area of inflammatory cell infiltration using HE staining images. In brief, we selected the region of inflammatory cell infiltration in Image Pro Plus software, then calculated the area of this region as the area of inflammatory cell infiltration of this picture. Then we calculated the area of inflammatory cell infiltration of eight HE views (magnification = original × 400) in the same group. The intensity of inflammatory cell infiltration in other groups was calculated and compared. The collagen volume fraction was also analyzed by Image Pro Plus software in the infarcted border zone. Eight separate images (magnification = original × 400) of Masson staining sections were selected, and collagen volume fraction (CVF) was calculated using the following formula: CVF = collagen area / total visual area × 100%, to assess the degree of cardiac fibrosis.

The serum was collected from fresh blood and centrifuged at 3,000 × g for 10 min at 4°C. The concentrations of serum NTpro-BNP, malondialdehyde (MDA), interleukin-1 (IL-1), and TNFα were detected by enzyme-linked immunosorbent assay.

Heart sections were deparaffinized and blocked with 5% goat serum for 1.5 h at room temperature. After washing three times with PBS, the sections were incubated with primary antibodies: rabbit polyclonal anti-TNFα antibody (1:500; Abcam; ab220210), rabbit polyclonal anti-collagen I antibody (1:500; Abcam; ab34710), and rabbit polyclonal anti-collagen III antibody (1:500; Abcam; ab7778) at 4°C overnight, and then were incubated with a secondary antibody. Finally, the sections were stained with diaminobenzidine (DAB).

In our previous report, RNA-seq was applied in HF rats, the same model as this research (Gao et al., 2020). According to the evaluation results of cardiac function, the difference of intergroup in each group was not significant by comparing with pair-matching t test, so we chose three animals per group by random sampling. Total RNA of the cardiac tissues was extracted using TRIzol Reagent^® (Invitrogen, Carlsbad, CA, USA). Extracted RNA was digested with dnase to remove contaminating genomic DNA, and the quality of RNA was evaluated by RNA Nano 6000 Assay Kit of the Agilent Bioanalyzer 2100 system (Agilent Technologies, CA, USA). RNA was purified using poly-T oligo-attached magnetic beads, then reverse transcribed into cDNA. After cDNA was ligated with adaptors, PCR amplification was applied with Phusion High-Fidelity DNA polymerase, universal PCR primers, and index (X) primer, and build the library of each sample. The library was sequenced on Illumina Hiseq4000 platform and 150-bp paired-end reads. Quality control and alignment were performed with rat reference sequences. Subsequently, read counts of each gene were computed as raw gene expression.

Using the R package “DESeq2,” differentially expressed genes (DEG) were identified with a p-value <0.05 and |log2 (foldchange)| >1 in the R version 3.5.1 software. Furthermore, GO, reactome, and KEGG pathway enrichment were performed using the R package “clusterProfiler” with p < 0.05 (R version 3.5.1 software).

Cell used in the present study were purchased from the China Infrastructure of Cell Line Resources (Institute of Basic Medical Sciences, Chinese Academy of Medical Sciences) and cultured in Dulbecco’s modified Eagle medium (DMEM, Hyclone, USA) supplemented with 10% fetal bovine serum (FBS, Corning, USA), as well as a mixture of penicillin (100 U/ml, Corning, USA) and streptomycin (100 μg/ml, Gibco, USA) at 37°C in a humidified atmosphere of 5% CO₂. To evaluate the cytotoxicity of Cal in H9C2 cells, cells were cultured in 96-well plates at a density of 6 × 10³ cells/well and subjected to different concentrations (2.5, 5, 10, and 20 μM) of Cal treatments. Referring to the inflammatory cell model in our previous study (Li et al., 2016), RAW264.7 cells were subjected to lipopolysaccharide (LPS) (1 μg/ml) for 24 h. Then cell supernatants for conditioned media (CM) were collected for the next experiments. To investigate the effects of Cal on CM-stimulated cardiomyocytes, H9C2 cells were precultured with Cal for 6 h, then stimulated with CM (with/without Cal) for 24 h. CCK-8 was applied to detect the cell viability at 450 nm under a microplate reader.

Cardiac fibroblasts were isolated from neonatal SD rats using mixed enzymatic digestion (0.06% trypsin/0.04% collagenase type Ⅱ in D-PBS without Ca²⁺/Mg²⁺). Then fibroblasts were separated from cardiomyocytes by differential adhesion for about 90 min. The adherent cells are fibroblasts. Cells at passages 2–6 were cultured with serum-free media for about 24 h and subsequently stimulated with 20 ng/ml of TGF-β1 (PeproTech, USA) and Cal or LY294002 for 24 h.

Levels of tumor necrosis factor-α (TNF-α) and IL-1 in cell supernatant were assessed by following the protocols of commercially available kits (NanjingJiancheng, China). The content was expressed as pg/ml.

Mitochondrial membrane potential (MMP) in different groups were evaluated by the JC-1 staining kit (Beyotime Biotechnology, China). H9C2 cells were cultured on laser confocal dishes and then induced by CM. Different groups of H9C2 cells were incubated with JC-1 probe for 30 min at 37°C in the dark. After washing three times with PBS, the images were scanned by laser confocal microscopy (Leica Microsystems GmbH). Image ProPlus (IPP) software was applied to calculate and analyze the ratio of aggregates/monomers fluorescence intensity.

ROS was assessed by a commercial assay using fluorescent probe DCFH-DA. H9C2 cells were induced by CM in the presence or absence of Cal and LY294002, and then observed at excitation and emission wavelengths of 488 and 525 nm under a fluorescence microscope (Leica Microsystems GmbH).

H9C2 cells were grown onto confocal dishes for the specified experiment time, fixed with 4% paraformaldehyde for 12 min followed by 0.5% Triton X-100 for 20 min, and blocked with normal goat serum for 1.5 h. Then cells were incubated with NFκB antibody overnight at 4°C. After three times of washing, cells were incubated with the secondary antibody at room temperature for 1.5 h in the dark. After being washed three times, cells were counterstained with 5 μg/ml of DAPI for 20 min. Images were then obtained under a confocal microscope. For α-SMA immunofluorescence, anti-α-SMA antibody (1A4, Santa Cruz Biotech) and an Alexa–Fluor 488-labeled secondary antibody (Molecular Probes) were used.

Heart tissues and H9C2 cells were homogenized in RIPA lysis buffer and quantified by the bicinchoninic acid (BCA) method. A total of 50 μg of protein was separated by 10% sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) gel and then transferred to a PVDF membrane. After being blocked in a solution of TBST with 5% skimmed milk for 1.5 h at room temperature, the membranes were incubated overnight at 4°C with the following primary antibodies: anti-p-NFκB (ab97726; Abcam, USA), NFκB (CST8242, Cell Signaling Technology, Germany), antiTNFα antibody (ab183218; Abcam, USA), anti-p-STAT3 (ab76315; Abcam, USA), STAT3 (ab68153; Abcam, USA), MMP-9 (ab38898; Abcam, USA), PI3K (ab182651; Abcam, USA), AKT (ab182729; Abcam, USA), p-AKT (CST4060, Cell Signaling Technology, Germany), p-IKKα/β (CST2697T, Cell Signaling Technology, Germany), IKKα (CST2682, Cell Signaling Technology, Germany), IKKβ (CST8943, Cell Signaling Technology, Germany), and anti-GAPDH (ab8245, Abcam, 1:5,000) at 4°C overnight. Afterward, membranes were washed and incubated with specific horseradish peroxidase (HRP)-conjugated secondary antibodies (goat antirabbit IgG 1:12,000 and goat antimouse IgG 1:5,000) for 1 h. The blots were visualized with enhanced chemiluminescent (ECL) plus Western blotting detection reagent (GE Healthcare, UK) for 1 min at room temperature without light, and then captured and analyzed by UVP BioImaging Systems (Bio-Rad, Hercules, CA, USA). Furthermore, protein expressions were normalized based on GAPDH level, and grayscale analysis was performed by the Image-Lab software.

Statistical analysis was performed with the SPSS software (SPSS version 22.0) or GraphPad Prism 7. All data were presented as the mean ± standard deviation (SD). Data were carried out by Dunnett’s test and one-way analysis of variance (ANOVA) to compare differences among multiple groups. p-Values less than 0.05 were considered statistically significant.

After 28 days of treatments, echocardiography was implemented to examine cardiac function. The results indicated that rats in the model group had significantly lower values of EF and FS, while the diameters on LVID; s and LVID; d were longer than that in the sham group, revealing that the HF model was successfully induced. After treatment with Cal and fosinopril, both EF and FS were increased significantly compared with the model group, suggesting that the left ventricular function was improved by Cal and fosinopril treatment. Additionally, compared with the model group, LVID; d in Cal and fosinopril groups were reduced significantly, while LVID; s in Cal and fosinopril groups were of no statistical significance (Figures 1A, B). The pathological changes in heart tissue were detected by H&E staining. As shown in Figure 1C, the left ventricular areas of the rats in the sham group were clearly visible, with the myocardial fibers arranged neatly and no inflammatory cell infiltration exerted in the myocardial interstitial space. However, in the heart tissue of rats in the model group, a significantly increased infiltration of inflammatory cell was present, myocardial fibers in the infarcted area were almost dissolved, and the myocardial stripes disappeared. Treatment with Cal and fosinopril rescued hearts from inflammatory cell infiltration and improved these pathological changes obviously (Figure 1C). Besides, Masson staining showed that there was obvious extensive collagen deposition in the border zone of the infarction in HF rats, whereas treatment with Cal and fosinopril significantly reduced contents of collagen deposition (Figure 1D). In addition, compared with the model group, Cal and fosinopril significantly reduced the serum levels of NTpro-BNP and MDA, which are the biomarkers of cardiac injury in HF (Figure 1E). These data indicated the protective effects of Cal and fosinopril on cardiac function.

We utilized RNA-Seq to determine the effect of Cal on the cardiac transcriptome. Principal component analysis (PCA) was used for visualizing RNA-seq between different groups (Figure 2A). A hierarchical cluster heatmap between different groups illustrated the differentially expressed genes (DEGs) (Figure 2B). Compared with the model group, a total of 3,104 DEGs (1,473 upregulated and 1,631 downregulated genes) satisfied |log2 (foldchange)| > 1 and the adjusted p-value less than 0.05 in the Cal-treated group (Figure 2C). Besides, the 986 key targets of Cal treatment against the HF model were obtained by overlapping the 4,288 targets of Cal treatment and 2,458 targets of HF model with a Venn diagram (Figure 2D). Furthermore, the identified 3,104 DEGs (Cal group vs. model group) were further used for Gene Ontology and functional pathway enrichment analysis. Intriguingly, enrichment analyses of GO annotations and reactome pathways revealed that the DEGs are related to inflammatory response and fibrosis in the Cal treatment group compared with the model group (Figures 2E, F). Besides, KEGG enrichment analysis of 3,104 DEGs showed that the involved pathway of Cal treatment was mainly composed of the PI3K–Akt signaling pathway, TNF signaling pathway, cytokine–cytokine receptor interaction, and MAPK signaling (Figure 2G). In conclusion, the integrated RNA-seq results suggested that the cardioprotective effects of Cal may owe to anti-inflammation and antifibrosis.

The IKKs-NFκB pathway is considered as the initiation of inflammatory cascade. WB results implied that expressions of IKKα/β and NFκB remained relatively unaltered among different groups, while the expression levels of phosphorylated IKKα/β (p-IKKα/β) and NFκB (p-NFκB) increased significantly in the model group. After treatment with Cal, the expressions of p-IKKα/β and p-NFκB were reduced significantly, while only p-NFκB was decreased markedly in the fosinopril group (Figure 3A). The expression of downstream target activated by NFκB was further detected. Protein level of TNFα in the model group was increased, whereas Cal and fosinopril treatment suppressed the expression of TNFα (Figure 3A). Additionally, immunohistochemistry results showed that IOD of TNFα in the model group was upregulated significantly. After treatment with Cal and fosinopril, IOD of TNF was decreased (Figure 3B). Besides, Cal and fosinopril significantly reduced the serum levels of TNFα (Figure 3C) and IL-1 (Figure 3D). Collectively, these data indicated that Cal and fosinopril could alleviate inflammatory responses in HF rats.

To investigate the effects of Cal and fosinopril on antifibrosis, the contents of collagen I and Ⅲ in the cardiac tissue were determined by IHC. Results showed that IOD of collagen Ⅰ and Ⅲ in the model group were increased, compared with that in the sham group. After treatment with Cal and fosinopril, IODs of collagens Ⅰ (Figure 4A) and Ⅲ (Figure 4B) were both reduced. These results demonstrated that Cal and fosinopril could effectively attenuate cardiac fibrosis through the inhibition of expressions and depositions of collagen I and collagen III. Recent evidence indicates that the sustained activation of STAT3 signaling after MI may contribute to adverse remodeling and progression to heart failure (Nural-Guvener et al., 2015). MMP-9 is involved in post-MI repair and remodeling by regulating ECM metabolism and processing inflammatory mediators (Frangogiannis, 2017). The results showed that expressions of phosphorylated STAT3 (p-STAT3) and MMP-9 were upregulated in the model group, and Cal and fosinopril impressively inhibited the expressions of p-STAT3 and MMP-9 compared with the model group (Figure 4C).

PI3Ks are kinases that are responses for different types of membrane receptors, which have been observed to be activated in many cardiovascular diseases such as heart failure, hypertension, and atherosclerosis (Ghigo et al., 2013). In vivo results implied that the expressions of PI3K and p-AKT in the model group were decreased, while after treatment with Cal and fosinopril, the expressions were both upregulated remarkably (Figure 4D). Besides, transcriptome analysis revealed that the mRNA level of Pik3cd was significantly increased under Cal treatment, compared with the model group (Figure 4E). The Cal treatment dramatically downregulated the mRNA levels of mmp9, tnf, and nfap (NFκB activating protein), respectively (Figure 4E).

To further confirm the regulatory mechanism of Cal on inflammation and to better simulate the pathological environment of cardiomyocytes under inflammatory condition in HF, we applied a macrophage conditioned media (CM)-stimulated cardiomyocyte model described in our previous study (Li et al., 2016). Cell viability was reduced dramatically, and cell injury had occurred as characterized by induction of ROS and mitochondria damage. H9C2 cell, an embryonic cardiomyocyte cell line, was selected here owing to its robust and fast reaction to various stimuli. As shown in Figures 5A, B, treatment with 1–25 μM Cal proved to be effective, and 5 μM Cal showed the best protective effect on cell viability. So, 5 μM Cal was the optimal concentration applied in the subsequent experiments. During the inflammatory phase of heart failure, myocardial cell death and hypoxia trigger the overgeneration of reactive oxygen species (ROS) and the damage of the mitochondria. ROS staining showed that Cal dramatically reduced the level of ROS (Figure 5C). The evaluation of mitochondrial transmembrane potential (MMP) was conducted by JC-1 probe (Zhang et al., 2021). The results suggested that the ratio of aggregates/monomers increased in response to Cal, suggesting that the MMP returned to normal (Figure 5D). Besides, Cal significantly reduced the release of TNFα and IL-1 (Figure 5E). Intriguingly, LY294002, an inhibitor of PI3K, compromised these protective effects of Cal in CM-induced H9C2 cells.

As the PI3K–AKT pathway plays a vital role in regulating the inflammation, we compared the expression levels of p-AKT, p-IKKα/β, and p-NFκB in CM-induced H9C2 cells with or without Cal treatment. WB results implied that expression of p-AKT was impressively reduced in the model group, while treatment with Cal could promote the expression of p-AKT (Figure 5F). Besides, the expressions of p-IKKα/β and p-NFκB were both markedly increased in the model group compared with the control group, while treatment with Cal downregulated their expressions, respectively (Figure 5F). Immunofluorescence results also showed that Cal treatment inhibited nuclear translocation of NFκB (Figure 5G). To further explore the anti-inflammation effect of Cal on the PI3K–AKT pathway, LY294002, an inhibitor of PI3K, was added together with Cal. Intriguingly, LY294002 suppressed the expression of p-AKT. Furthermore, the regulation on p-IKKα/β, NFκB activation, ROS level, MMP, and proinflammatory cytokine level by Cal was also eliminated by LY294002 (Figures 5D–G), indicating that Cal protected against CM-induced injury in H9C2 cells partly by targeting on the PI3K–AKT pathway. Collectively, these results demonstrated that Cal could alleviate inflammation by activating PI3K–AKT pathway in cardiomyocytes.

The effects of Cal on TGFβ-stimulated cardiac fibroblasts were further investigated. Cardiac fibroblasts were incubated with TGFβ at a concentration of 20 ng/ml for 24 h. TGFβ induces transformation of fibroblast to myofibroblast, which is the main producer of collagens and is characterized by the presence of α-smooth muscle actin (α-SMA). Our results showed that α-SMA expression was increased in TGFβ-stimulated cells, suggesting that fibroblasts were phenotypically transformed into myofibroblasts (Figure 6A). Cal treatment suppressed TGFβ-induced expression of α-SMA, dramatically reduced the expressions of p-STAT3 and MMP-9, and upregulated the expressions of PI3K and p-AKT compared with the model group (Figures 6A, B). While cotreatment with LY294002 abolished these effects. These data suggested that Cal could block the transformation of fibroblast to myofibroblast, thereby suppressing cardiac fibrosis via the PI3K–AKT pathway.