All reagents and materials used in this research are documented in the Supplementary Material.

Male C57BL/6 mice (20 ± 2 g, 8–10 weeks old) were purchased from Beijing SPF Biotechnology (Beijing, China). All procedures were performed following the guidelines of the Institutional Animal Care and Use Committee and approved by the Beijing University of Chinese Medicine Animal Care Committee. The animals were housed in a 12:12-h light–dark cycle and temperature-controlled (22°C) environment and fed with standard rodent chow and tap water. A DIC model was generated as described previously (Wang et al., 2019; Wang et al., 2020). After a week of adjustable feeding, mice randomly assigned in the model, FGL, and ENA groups were injected by tail vein with DOX following a dose of 5 mg/kg once weekly for 4 weeks. Meanwhile, mice in the control group were treated with saline solution (0.9% NaCl) instead. Then, FGL at 20 mg/kg and ENA at 15 mg/kg were administered intragastrically and daily for 4 weeks at 1 week after last injection of DOX, respectively. The control and model groups also received the same volume of ultrapure water. ENA has been repeatedly reported to protect against DIC in clinical and preclinical studies (Hullin et al., 2018); thus, we applied ENA as a positive control for animal experiments.

Samples were prepared according to the protocol of the manufacturer and then underwent RNA extraction, reverse transcription, and DNA amplification, followed by sequencing on an Illumina HiSeq. We identified differentially expressed genes (DEGs) based on the following criteria: p < 0.05 and absolute value of log2fold-change>0. Biological and pathway analysis of DEGs was explored by GO term enrichment analysis and KEGG pathway enrichment analysis, respectively. Protein–protein interaction (PPI) network analysis of DEGs was acquired from the Search Tool for the Retrieval of Interacting Genes (STRING) database (version 11.0; https://www.string-db.org/).

After 4 weeks’ administration, the mice were immobilized and anesthetized with isoflurane gas for echocardiography. The transthoracic echocardiography was performed in M-mode with the VEVO 2100 echocardiography system (Visual Sonics, Toronto, ON, Canada) by the same operator on all mice. Left ventricular end-systolic volume (LVESV), left ventricular end-diastolic volume (LVEDV), left ventricular end-diastolic dimension (LVEDD), and left ventricular end-systolic dimension (LVESD) were measured using computer algorithms for at least three uninterrupted cardiac cycles. The calculated EF and FS values of the mice treated by DOX were uniformly decreased compared to those in the control group.

Cardiac tissue samples were immersed in 4% paraformaldehyde and then embedded in paraffin and cut into 4-μm serial sections. The sections were stained with hematoxylin–eosin (H&E) and Masson trichromatic solution and were then observed under an optical microscope at ×400 magnification.

Blood samples were obtained from abdominal aorta, and the sera were isolated to measure tissue injury markers, lactate dehydrogenase (LDH), and creatine kinase isoenzymes (CK-MB) with standard diagnostic kits (Nanjing Jiancheng Bioengineering Institute, Jiangsu, China).

H9C2, HepG2, and MCF-7 cells were grown in DMEM with 10% fetal bovine serum supplement and incubated in humidified air (5% CO₂) at 37°C. H9C2, HepG2, and MCF-7 cells were grown on 96-well plates at a density of 4 × 104 cells/ml for 24 h and then divided into the control, DOX, and 0.1–50 µM FGL groups. Next, FGL groups were pretreated with FGL, followed by cotreating with 1 µM DOX as previously confirmed (Wang et al., 2020) and FGL (0.1–50 µM) in FGL groups while merely with 1 µM DOX in the DOX group for 24 h. Subsequently, cell viability was detected using 10% CCK-8 (Dojindo, Kumamoto, Japan). H9C2 cells were divided into five groups as follows: control group, DOX group, DOX + FGL, DOX + FGL + Selisistat, and DOX + FGL + SR-18292. Selisistat is an inhibitor of SIRT1, inhibiting the deacetylation activity of SIRT1. SR-18292 is an inhibitor of PGC-1α, promoting the acetylation of PGC-1α and inhibiting its activity as a transcriptional coactivator. Amounts of 1 μM DOX, 0.1 μM FGL, 20 μM Selisistat, and 20 μM SR-18292 were applied, respectively.

The apoptosis of cardiac tissue was determined by terminal deoxynucleotidyl transferase-mediated nick end labeling (TUNEL) according to the instructions of the manufacturer and then staining with DAPI to label the nuclei. In addition, H9C2 cells were incubated with 2 mg/ml Hoechst 33,342 for 30 min at 37°C in the dark for detecting the apoptosis of H9C2 cells. Finally, a fluorescence microscope was used to visualize apoptotic cells.

The process was performed as described previously (Wang et al., 2019). Images were acquired under an electron microscope (Leica, Buffalo Grove, IL, United States). Mitochondria were counted randomly based on six images from various fields of view (×2,500 magnification).

Changes of mitochondrial membrane potential (ΔѰm) were detected using the fluorescent probe JC-1 (Beijing Solaibao Technology Co., Ltd., Beijing, China) according to the user manual. JC-1 is a cationic dye that accumulates in the mitochondria in a potential-sensitive manner. At high ΔΨm, JC-1 accumulates in the mitochondria forming J-aggregates and emitting red fluorescence; at low ΔѰm, the J-monomers generate green fluorescence. After pharmacological treatments, the cells were incubated with 1×JC-1 working solution for 30 min and washed twice with JC-1 dye buffer. Finally, the images were monitored under a fluorescence microscope. ΔΨm was evaluated and calculated by the ratio of red/green fluorescence intensities.

DCFH-DA staining was performed for the detection of ROS in H9C2 cells. Pretreated H9C2 cells were cultured with 10 μM DCFH-DA for 30 min at 37°C in the dark. Then DAPI was used to stain the nuclei. The images were acquired using a laser scanning confocal microscope at ×100 magnification.

ATP levels in H9C2 cells were detected by using an ATP assay kit (Beyotime Biotechnology, Shanghai, China). Briefly, H9C2 cells were lysed and centrifuged, and the supernatant was collected for the determination of ATP and total protein content. The content of ATP was normalized to the cellular protein concentration.

The process was performed as described previously (Wang et al., 2019). The primers used in this study are shown in Supplementary Material.

H9C2 cells were fixed with 4% formaldehyde, washed twice with PBS, and then permeabilized with 0.3% Triton X-100. After being blocked with 5% bovine serum albumin, cells were incubated with primary antibodies against SIRT1 and PGC-1α, and then with the corresponding secondary antibodies. The nuclei were counterstained with DAPI. Images were captured by a laser scanning confocal microscope at 1,000× magnification.

Co-IP was executed according to the indicated procedures of the manufacturer (Cell Signaling Technology, United States). H9C2 cells were lysed in a lysis buffer. Equivalent proteins from each sample were combined with anti-PGC-1α antibody or control IgG antibody overnight at 4°C. Later, the resulting immunocomplex was precipitated employing Protein G magnetic beads, followed by washing five times. Thereafter, precipitates mixed in the SDS buffer solution were analyzed by western blot with either anti-acetylated lysine antibody to determine the extent of PGC-1α acetylation or PGC-1α antibody to determine the total amount of PGC-1α.

To detect protein expression, western blotting was carried out as previously described (Wang et al., 2019; Wang et al., 2020). The antibodies used in the study are shown in Supplementary Material.

Statistical analysis was performed using the GraphPad software 6. Numerical results are displayed as mean ± s.d. Statistical significance was evaluated by applying two-tailed unpaired Student’s t-test. Alternatively, comparisons of more than two groups were performed by one-way analysis of variance (ANOVA) with Bonferroni post hoc tests for multiple groups to calculate the differences. A p-value of <0.05 was considered as statistically significant.

The design of animal experiments, including the generation of a DIC model and subsequent pharmacological treatment, is shown in Figure 1A. The body weight of mice was recorded throughout the animal experiment (Figure 1B). Comparison with the control group indicated an evident decline in the body weight in the DOX-treated group. Meanwhile, the body weight of mice in the FGL and ENA groups was significantly increased during weeks 5–8 compared with that in the model group. The mice underwent echocardiography to evaluate cardiac function after 4 weeks of pharmacological intervention. FGL alleviated the decrease in EF and FS values caused by DOX (Figures 1C,F). Consistently, H&E and Masson staining of heart sections showed that DOX induced myofibrillar loss, disordered arrangement, enlarged intercellular structures, and plasma-dissolved cardiomyocytes and myocardial fibrosis, which could be attenuated by FGL (Figures 1D,E). Additionally, increased serum content of LDH and CK-MB, markers of myocardial damage, was observed in the DOX group, while FGL treatment partially reversed these changes (Figure 1G).

To further explore the pathogenic mechanism of DOX and underlying molecular mechanisms of FGL on cardiac protection, we performed high-throughput RNA-Seq to investigate DEGs in the control, DOX, and DOX + FGL groups of mice. Figure 2A showed distinct hierarchical clustering of the genes and groups. Gene ontology and KEGG analysis between the control and DOX groups revealed that genes related to mitochondria/metabolism, fatty acid oxidation, PPAR signal pathway, etc. were significantly enriched (Figures 2B,C). Then we presented eight DEGs involved in mitochondrial metabolism, biogenesis, and PPAR signaling pathways between the control and DOX groups in the form of heatmap (Figure 2D). Protein–protein interaction (PPI) network analysis (version 11.0; https://www.string-db.org/) indicated that a cluster containing eight nodes (Sirt1, Ppargc1a, Ppara, Nrf1, Tfam, Cd36, Lpl, and Cpt2) possessed a strong connection (Figure 2E).

To validate the results of RNA-Seq and quantify the expression of these key genes, we assessed eight DEGs involved in mitochondrial biogenesis and the PPAR signaling pathway by quantitative RT-PCR on three to five heart samples. The results of the RT-qPCR analysis were in good agreement with the RNA-Seq data that DOX resulted in the downregulation of eight genes compared with the control group, while FGL significantly increased the RNA levels of all genes except Ppara compared with the DOX group (Figure 2F).

DOX-induced apoptosis has long been considered to be one of the leading causes of cardiac function decline. Therefore, we evaluated the apoptosis of cardiomyocytes in vivo and in vitro. As shown in Figure 3C, the minimum concentration of FGL that enhanced cell viability was 0.1 μM. TUNEL staining of mouse heart slices and Hoechst staining of H9C2 cells showed that FGL alleviated DOX-induced apoptosis (Figures 3A,B,D,E). Moreover, mitochondrial membrane potential reduction is a hallmark event of early apoptosis. The results showed that DOX caused the notable drop of the mitochondrial membrane potential, which could be alleviated by FGL. Compared with the FGL-treatment group, selisistat and SR-18292 reduced the mitochondrial membrane potential (Figures 3F,G). We also found that DOX caused a large amount of ROS production, while FGL significantly reduced ROS production. These results suggested that FGL alleviated apoptosis and ROS production by associating with SIRT1 and PGC-1α.

We then hypothesized that FGL exerted cardioprotection by improving mitochondrial biogenesis and fatty acid oxidization. In DIC mice, we found that DOX led to loose arrangement of mitochondria and a large reduction in the number of mitochondria via TEM, while FGL relieved the above adverse changes (Figures 4A,B). Moreover, coinciding with the results of RNA-seq, the expressions of mitochondrial biogenesis and fatty acid oxidization markers were decreased in the DOX group compared to those in the control group, and these changes were relieved by FGL (Figures 4C–F). However, DOX-induced reduction in the protein levels of PPARα was not mitigated by FGL. PGC-1α acts as a transcriptional coactivator to promote the transcriptional regulation of PPARα to downstream proteins (CD36, CPT2, LPL, etc.). We concluded that FGL merely upregulated the expression of PGC1-α without improving the expression of PPARα and thus activated the transcriptional activity of PPARα to promote the expression of downstream genes. In addition, the expression of SIRT1, a deacetylase, was also significantly enhanced by FGL treatment. Combined with reported research that SIRT1 deacetylated PGC-1α and activated its transcriptional activity, we proposed that FGL may promote downstream MB&FAO through the SIRT1–PGC-1α axis.

To further confirm whether FGL facilitated MB&FAO through the SIRT1–PGC-1α pathway, we cotreated FGL with selisistat and SR-18292 in vitro, respectively. First, we found that DOX resulted in a decrease in the number of mitochondria per unit area in vitro consistent with the results of electron microscopy in vivo, mtDNA copy number, and ATP content in vitro. Contrariwise, FGL contributed to a marked increase in these aspects compared to the DOX group, which were significantly reversed by selisistat and SR-18292 (Figures 5A–D). Impressively, we found that FGL inhibited the downregulation of MB&FAO marker except PPARα induced by DOX; however, both selisistat and SR-18292 significantly reversed the effects of FGL on the downstream mitochondrial biogenesis marker (Nrf1 and TFAM) and the fatty acid oxidation marker (CD36, CPT2, and LPL) (Figures 5E–J). These collective data documented that FGL coordinated MB&FAO through SIRT1 and PGC-1α. However, its specific molecular mechanism between SIRT1 and PGC-1α still needs to be further confirmed.

In order to confirm the specific interaction between SIRT1 and PGC-1α and whether FGL could interfere with this interaction, the binding of SIRT1 and PGC-1α and the acetylation level of PGC-1α were detected by immunocoprecipitation and immunofluorescence. As shown in Figure 6A, SIRT1 and PGC-1α were distributed in both the cytoplasm and the nucleus, and both of them showed obvious fluorescent spots in the nucleus. Compared with the control group, in the nucleus, DOX resulted in a significant reduction of SIRT1 (red) and PGC-1α (green) and a decrease in colocalization (white). Compared with the DOX group, FGL increased the number of SIRT1 and PGC-1α fluorescence spots and colocalization of both (Figures 6A,B). These results revealed that DOX treatment reduced the expression and binding of SIRT1 and PGC-1α in the nucleus, while FGL promoted the expression and binding of SIRT1 and PGC-1α in the nucleus. The results of CO-IP showed that FGL inhibited the increased acetylation of PGC-1α induced by DOX (Figures 6C,D), while selisistat increased the acetylation of PGC-1α. Collectively, we concluded that FGL inhibited the acetylation of PGC-1α by promoting the expression of SIRT1 and, concurrently, increased the expression of PGC-1α to facilitate PGC-1α-mediated MB&FAO.

According to current research on DIC, identification of cardioprotective agents without interfering with the anticancer effect of DOX has been of interest to oncologists and cardiologists. Thus, in addition to confirming the protective effect of FGL on cardiomyocytes, we evaluated whether FGL interferes with the anticancer effect of DOX on the human breast carcinoma cell line MCF-7 and human hepatocellular carcinoma cell line HepG2. As shown in Figure 7, FGL within 0.1–1 μM did not affect the inhibition of cancer cell viability by DOX (Figures 7A,B).