GATA4-overexpressing BMSCs-derived exosome regulation of myocardial infarction in mice by key miRNA and apoptosis gene CLU

Abstract

Background

Myocardial infarction (MI) is the most common cardiovascular disease that has a serious impact on human health and is one of the most common causes of death in the world. Apoptosis and myocardial fibrosis after MI are the key pathological features of poor myocardial remodelling, which further lead to heart failure, and are also the main reason for the high mortality from MI.

Methods

Exosomes (Exos) from GATA4-overexpressing bone marrow mesenchymal stem cells (BMSCs) were extracted and identified using transmission electron microscopy (TEM), nanoscale tracking analysis (NTA), and marker detection. Tandem mass tags (TMT) quantitative proteomics, Agilent miRNA microarray, and GO/KEGG function analysis were used to obtain differentially expressed proteins/miRNAs. The expression levels using the Wald test and RT-qPCR to identify and validate. Their related mRNA–miRNA–circRNA regulatory networks were constructed. Hypoxic mouse cardiomyocytes were cultured, and MI mice were used for subsequent experiments. Cell differentiation, apoptosis, and marker gene expression were detected using RT-qPCR, IF, flow cytometry, and Western blot.

Results

Four key apoptosis proteins, CXCL12, CLU, CD44, and IGF1, and 20 key differentially expressed miRNAs. Among them, CLU and IGF1 were upregulated, but CXCL12 and CD44 were downregulated in Exos from the GATA4-overexpressing BMSCs group. The expression of mmu-miR-467 g and mmu-miR-5127 was downregulated in Exos from the GATA4-overexpressing BMSCs group, whereas the other 18 key miRNAs were upregulated. Exos from GATA4-overexpressing BMSCs inhibit apoptosis, activate the LXR/RXR pathway, and improve MI. Knockout of CLU reversed this effect. Our research further discovered that GATA4-overexpressing BMSCs-derived Exos improved MI through downregulation of CLU.

Conclusions

GATA4-overexpressing BMSCs-derived Exos may regulate MI via the aforementioned four key proteins and may be related to the LXR/RXR signalling pathway.

Supplementary Information

The online version contains supplementary material available at 10.1186/s40001-026-03872-5.

Introduction

Cardiovascular disease has a serious impact on human health and is one of the most frequent causes of death in the world. Myocardial infarction (MI) is the most common cardiovascular disease, whose yearly morbidity and mortality rates have been steadily increasing [1, 2]. MI is characterised by myocardial cell ischaemia and hypoxia caused by coronary artery occlusion, which leads to inflammation and apoptosis. Apoptosis and myocardial fibrosis after MI are the key pathological features of poor myocardial remodelling, which further leads to heart failure and is also the main reason for the high mortality of this life-threatening condition [3, 4]. Evidence has confirmed that addressing physiological and pathological features, such as inflammatory response, apoptosis, and fibrosis, is critical to the effective treatment and prognosis of patients with MI [5]. Since cardiomyocytes are terminally differentiated cells that lose their ability to regenerate, excessive apoptosis significantly affects the recovery of cardiac function. Thus, the early regulation of cardiomyocyte apoptosis is of substantial significance to the recovery of cardiac function [6]. Therefore, elucidating the potential mechanisms of this fatal disease and developing effective treatment strategies are of critical importance.

GATA4 is a cardiomyocyte-specific zinc finger transcription factor that regulates the differentiation, growth, and survival of cardiomyocytes [7]. The overexpression of cardiac GATA4 protein can enhance angiogenesis and reduce fibrosis, thereby preserving cardiac function after cardiac injury. GATA4 plays a key role in myocardial differentiation and function [8]. Exosomes (Exos) are 50–200 nm vesicles secreted into the outer space of cells, shuttling between various microRNAs (miRNA), long non-coding RNA (lncRNA), and proteins to regulate cell–cell communication and participate in intercellular signal transduction [9, 10]. Exos from bone marrow mesenchymal stem cells (BMSCs) have been shown to provide cardioprotective effects similar to those of BMSCs, such as reduced cardiomyocyte apoptosis, increased angiogenesis, increased cardiac function, and diminished infarct size after ischaemia/reperfusion injury [11]. Exos secreted by CXCR4-overexpressing mesenchymal stem cells (MSCs) promote cardiac function recovery through the Akt signalling pathway after MI, reducing infarct size and improving cardiac remodeling [11]. Zheng et al. reported that BMSCs-derived Exos played an important role in regulating cardiomyocyte apoptosis in MI rats, and the upregulation of miRNA improved myocardial angiogenesis and ventricular remodelling and reduced myocardial fibrosis and apoptosis [12]. The results of our previous study confirmed that Exos from GATA4-overexpressing BMSCs can improve the differentiation of BMSCs into cardiomyocyte-like cells; as a result, hypoxia-induced cardiomyocyte apoptosis was reduced, and post-infarction myocardial function was recovered [13]. GATA4 exhibits specific anti-apoptosis properties and promotes cardiomyocyte differentiation. In our earlier investigation, we confirmed that Exos from GATA4-overexpressing BMSCs are important for recovery from MI [13, 14]. However, the specific molecular mechanisms by which Exos released from GATA4-overexpressing BMSCs influence MI require further investigation. Therefore, to further elucidate the specific molecular mechanisms by which GATA4-overexpressing BMSCs-derived Exos exert their effects in cardiomyocytes, this study integrated proteomics and miRNA sequencing analysis to clarify how these Exos modulate MI through apoptosis-related proteins and microRNAs. Furthermore, our findings can serve as a further scientific and clinical basis for the clinical improvement of cardiac function after MI.

Material and methods

Establishment of mouse BMSCs

BMSCs were extracted from the femur and tibia of C57BL/6 mice as previously reported [15] and co-incubated in BMSCs growth medium (Cyagen Biosciences, Sunnyvale, CA, USA) containing 10% fetal bovine serum (FBS) (Gibco, Carlsbad, CA, USA). The specific procedures were described in our previous publication [13]. P9 cells were selected for follow-up experiments.

Cell transfection

GATA4-overexpressing lentivirus [13], si-CLU, oe-CLU and si-IGF1 Lipofectamine 2000 reagent (Invitrogen Life Technologies, Carlsbad, CA, USA) were separately used to transduce BMSCs and cardiomyocytes. After 24 h, we measured the expression levels using RT-qPCR.

Extraction and identification of Exos

Collecting transfected BMSCs, culture medium with BMSCs growth medium without FBS for 48 h, collected and centrifuged 1 h at 110,000 × g and 4 ℃. After removing the supernatant, the phosphate-buffered saline (PBS) was resuspended through centrifugation for 1 h at 110,000 × g and 4 ℃. Transmission electron microscopy (TEM) was employed to observe the Exos morphology, nanoparticle tracking analysis (NTA) was used to analyse, and Western blot analysis of exosome markers.

Exosome co-culture analysis

Using the Exos labelling and tracer kit (PKH26) (C3637S, Beyotime, Shanghai, China), PKH26-labelled Exos (10 μg) were resuspended in 100 μL of PBS and then incubated with 1 × 10⁵ C57BL/6 mouse ventricular cardiomyocytes (ScienCell, San Diego, CA, USA) in a hypoxic, serum-free environment. After 24 h, the cells were harvested for immunofluorescence analysis. In the co-cultured cells, the Exos from GATA4-overexpressing BMSCs were referred to as the Exos^GATA4 group (n = 3), whereas the Exos from control (normal) BMSCs were the Exos group (n = 3).

Real-time fluorescence quantitative PCR (RT-qPCR)

Total RNA was extracted with TRIzol reagent (Invitrogen, 15,596,026) and reverse transcribed to cDNA with the One Step Prime Script miRNA cDNA Synthesis Kit (Takara, Kyoto, Japan). We used SYBR Green PCR Master Mix (Life Technologies, USA), following the manufacturer’s instructions for the RT-qPCR procedures. The RT-qPCR primers are presented in Table 1. The value was calculated using the 2^−ΔΔCt method, with GAPDH or U6 as the internal reference.

Table 1.

PCR primer sequences

Genes	Primers	Sequence (5′−3′)
CXCL12	Forward	5′- AAGGTCGTCGCCGTGCTG −3′
	Reverse	5′- CTCTTCTTCTGTCGCTTCT −3′
CLU	Forward	5′- TGGCATCATAGACACGCTCTT −3′
	Reverse	5′- GGGCAGGATTGTTGGTTG −3′
CD44	Forward	5′- GTCCAACACCTCCCACTA −3′
	Reverse	5′- TCCGTTCTGAAACCACGT −3′
IGF1	Forward	5′- CTCTTCTACCTGGCGCTCTG −3′
	Reverse	5′- TCCTTTGCAGCTTCGTTT −3′
ABCA1	Forward	5′- CAGTGGCGGCAACAAACG −3′
	Reverse	5′- TTCCCGGAAACGCAAGTC −3′
ABCG1	Forward	5′- CTGAACTGCCCTACCTACCA −3′
	Reverse	5′- GGCTTCGTTCCCAATCCC −3′
LXR	Forward	5′- GGGTTGCTTTAGGGATAGG −3′
	Reverse	5′- TTCTTGCCGCTTCAGTTT −3′
RXRα	Forward	5′- CACATTGGGCTTCGGGACT −3′
	Reverse	5′- CTCGTTGGCACTGCTGGT −3′
SREBP-1c	Forward	5′- GATCAAAGAGGAGCCAGTGC −3′
	Reverse	5′- TAGATGGTGGCTGCTGAGTG −3′
GAPDH	Forward	5′-TGACCACAGTCCATGCCATCAC-3′
	Reverse	5′-CGCCTGCTTCACCACCTTCTT-3′
mmu-miR-103-3p	Forward	5′-ACACTCCAGCTGGGAGCAGCATTGTAC-3′
	Reverse	5′- TGGTGTCGTGGAGTCG −3′
mmu-miR-107-3p	Forward	5′- GAGCAGCATTGTACAGG −3′
	Reverse	5′- AGTGCAGGGTCCGAGGTATT −3′
mmu-miR-130a-3p	Forward	5′-CGGCAGTGCAATGTTAAAAGGGCAT-3′
	Reverse	5′- GATCGCCCTTCTACGTCGTAT −3′
mmu-miR-140-3p	Forward	5′- TACCACAGGGTAGAACCACGG −3′
	Reverse	5′- GTGCGTGTCGTGGAGTC −3′
mmu-miR-15a-5p	Forward	5′- CGCGTAGCAGCACATAATGG −3′
	Reverse	5′- AGTGCAGGGTCCGAGGTATT −3′
mmu-miR-15b-5p	Forward	5′- TAGCAGCACATCATGGTTTACA −3′
	Reverse	5′- TGCGTGTCGTGGAGTC −3′
mmu-miR-16-5p	Forward	5′- CGCGTAGCAGCACGTAAATA −3′
	Reverse	5′- AGTGCAGGGTCCGAGGTATT-3′
mmu-miR-199a-3p	Forward	5′- GTCACAGTAGTCTGCACAT −3′
	Reverse	5′- GTGCAGGGTCCGAGGT −3′
mmu-miR-21a-5p	Forward	5′-ACGTTGTGTAGCTTATCAGACTG-3′
	Reverse	5′-AATGGTTGTTCTCCACACTCTC-3′
mmu-miR-23a-3p	Forward	5′- CCAATTGCGCCTTCAGGCTA −3′
	Reverse	5′- CGGCAGAGTCCTTACCCACA −3′
mmu-miR-23b-3p	Forward	5′-GGAAATCCCTGGCAATGT-3′
	Reverse	5′-TAATCCCTGGCAATGTGA-3′
mmu-miR-25-3p	Forward	5′-ACACTCCAGCTGGGCATTGCACTTGTCTCG-3′
	Reverse	5′-ACACTCCAGCTGGGCATTGCACTTGTCTCG-3′
mmu-miR-27a-3p	Forward	5′- TCACAGTGGCTAAGTTCCGC −3′
	Reverse	5′- CTCAACTGGTGTCGTGGAGTC −3′
mmu-miR-29b-3p	Forward	5′- CGCGTAGCACCATTTGAAATC −3′
	Reverse	5′- AGTGCAGGGTCCGAGGTATT −3′
mmu-miR-301a-3p	Forward	5′- CGTGCGAAGCTCAGGAGGG −3′
	Reverse	5′- TGGCTGTCGTGGACTGCG −3′
mmu-miR-322-5p	Forward	5′- AGCGTGCTGTGCGTGTGAC −3′
	Reverse	5′- CAGTGCAGGGTCCGAGGTATT −3′
mmu-miR-467 g	Forward	5′-GCGCGATATACATACACACACCAACAC-3′
	Reverse	5′-AACGCTTCACGAATTTGCGT-3′
mmu-miR-497a-5p	Forward	5′- GGTATGACAGCAGCACACTGT −3′
	Reverse	5′- CTCAACTGGTGTCGTGGAGTC −3′
mmu-miR-5127	Forward	5′- CGCGTCTCCCAACCCTT-3′
	Reverse	5′- AGTGCAGGGTCCGAGGTATT-3′
mmu-miR-92a-3p	Forward	5′- GGGGCAGTTATTGCACTTGTC −3′
	Reverse	5′- CCAGTGCAGGGTCCGAGGTA −3′
U6	Forward	5′-CAAATTCGTGAAGCGTTCCA-3′
	Reverse	5′-AGTGCAGGGTCCGAGGTATT-3′

Western blot

The total proteins were extracted using RIPA buffer (Sigma-Aldrich, USA) with 1% protease and phosphatase inhibitor. Next, the proteins were transferred from a 10% SDS-PAGE gel onto a polyvinylidene fluoride (PVDF) membrane (Millipore, MA, USA) for separation, followed by the addition of 5% skim milk for sealing and incubation at 25 ℃ for 2 h. Then, the first antibody was incubated overnight at 4 ℃. The next day, the samples were co-incubated with a secondary HRP-conjugated antibody (1:2000, ab205718, Abcam, UK) at 25 ℃ for 1 h. Finally, the protein was detected using chemiluminescence immunoassay. An anti-GAPDH antibody (1:1000, ab181602, Abcam, UK) was used in the control group. Protein band analysis was performed using Image J. Primary antibodies were as follows: anti-cytochrome C (1:5000, ab133504, Abcam), anti-cleaved-caspase-3 (1:5000, ab214430, Abcam), anti-Bax (1:1000, ab32503, Abcam), anti-Bcl-2 (1:10,000, ab182858, Abcam), anti-ABCA1 (1:1000, ab18180, Abcam), anti-ABCG1 (1:1000, CSB-PA004679, Cusabio), LXR-α (1:2000, CSB-PA619753LA01HU, Cusabio), RXRα (1:1000, A15242, abclonal), and SREBP-1c (1:1000, ab28481, Abcam).

Immunofluorescence (IF) assay

Cells were cultured in a 24-well plate at a cell density of 2 × 10⁴/well. After 24 h, washing with PBS was performed, and 4% paraformaldehyde was added, followed by sealing for 1 h with bovine serum albumin. Further, the cells were incubated overnight at 4 ℃ with anti-cTnT (1:200, ab8295, Abcam, UK), anti-connexin-43 (1:8000, ab312836, Abcam, UK), anti-desmin (1:500, ab32362, Abcam, UK), and anti-α-actin (1:100, ab8226, Abcam, UK) antibodies. The next day, we added secondary antibodies for 1 h, followed by DAPI staining. Finally, the discoloured cells were examined, and images were captured with a fluorescence microscope (400,857, Nikon, Japan).

Flow cytometry

Cells were collected and washed twice with PBS, followed by the addition of 200 μL of PBS. Then, the rate of programmed cell death was determined using the Annexin-V-FITC/PI apoptosis kit manufactured by Absin Bioscience Inc. (Shanghai, China) following the guidelines of the producer. Next, 5 μL of Annexin V-FITC and 5 μL of PI were added to each well. The mixture was incubated for 15 min in a dimly lit chamber, after which apoptosis was evaluated using a FACScan flow cytometer (Becton Dickinson, San José, CA, USA).

MI mouse model establishment and experiment

Male C57BL/6 mice (24–28 g) were purchased from SPF Biotechnology Co., Ltd (China, Beijing). C57BL/6 mice were used to establish the MI model as described in our previous research [16]. Anaesthesia was administered using 1% sodium pentobarbital at a concentration of 6 μL/g. Before ligating the left anterior descending coronary artery to establish the MI mouse model, the Exos from GATA4-overexpressing BMSCs and the sh-CLU were used for tail vein injection and heart gene delivery as referred in previous research [17]. To achieve cardiac-specific gene delivery, adeno-associated virus serotype 9 (AAV9) carrying short hairpin RNA targeting mouse CLU (AAV9-sh-CLU) was purchased from Genomeditech Co., Ltd (Shanghai, China). The sequence for sh-CLU was: 5'-GGAATAAAGTTCAA CCGTAAT-3'. We administered 30 μL of the AAV-sh-CLU vector solution via local intramyocardial injection into the mouse heart[18]. Mice were randomly assigned to the following groups: [1] MI group, [2] MI + Exos^GATA4 group, [3] MI + Exos^GATA4 + sh-CLU group (n = 6). Further, we determined the left ventricular fractional shortening (FS) and ejection fraction (EF) by M-mode echocardiography after 48 h. The TTC (T8170, Beijing Solebao Technology Co., Ltd., Beijing, China) staining method was implemented to determine the myocardial infarct size.

Histological staining

Each group of paraffin slice sections of heart tissue samples was stained with a hematoxylin–eosin (HE) Stain Kit (G1120, Beijing Solarbio Science&Technology Co., Ltd), and immunohistochemistry (IHC) analysis was performed using the anti-c-Kit antibody (1:100, ab231780, Abcam) and the anti-CLU (1/1000, ab184099, Abcam) for pathological and cardiac injury evaluation. Frozen heart sections were fixed in a 4% formaldehyde solution and stained with Oil Red O to observe lipid droplet aggregation.

TUNEL staining

The tissue specimens were washed with PBS twice, and 50 μL of TUNEL detection solution (2 μL of TDT enzyme + 48 μL of fluorescein-labelled dUTP solution) was added to the tissue specimens. The tissue samples were rinsed three times in a dark wet box at 37 ℃ and sealed with glycerol. Then, they were observed and photographed under a laser confocal microscope.

Enzyme-linked immunosorbent assay (ELISA)

Mouse tumour necrosis factor alpha (TNF-α) ELISA Kit (JL10484), mouse interleukin-1β (IL-1β) ELISA Kit (JL18442), mouse IL-10 ELISA Kit (JL20242), and mouse IL-6 ELISA Kit (JL20268) were purchased from JONLNBIO Co., Ltd (China). The experiment was completed in strict accordance with the kit instructions.

Extraction of total protein and trypsin digestion

After extracting the protein according to the above method, we determined the protein concentration. 100 mg protein solution was reduced with 90 μL of TCEP (10 mmol/L, Thermo Fisher, Waltham, MA, USA) at 37 ℃ for 30 min and alkylated with 40 mmol/L iodoacetamide (final concentration) for 40 min at room temperature in the dark. Then, the alkylated samples were incubated with cooled acetone at – 20 ℃ for 4 h, followed by centrifugation at 10,000 g/min and 4 ℃ for 20 min. Next, 100 μL of 50 mmol/L TEAB was used to dilute the protein samples, and trypsin/protein was used for digestion twice[19].

Tandem mass tags (TMT) peptide labelling and high-pH reversed-phase peptide fractionation

Peptides were dissolved in 0.5 M TEAB and labelled according to the operating instructions of the 6-label TMT kit (Thermo Fisher Scientific, MA, USA). High-pH reverse HPLC was implemented to fractionate the peptides using an Agilent 300 Extend C18 column (5-μm particles, 4.6 mm ID, 250 mm length). The specific operating steps have been previously described by Stryiński et al.[20].

LC–MS analysis and database search

After recombination, separation was done using an Easy-NLC 1000 high-performance liquid chromatography system, connected to a Q Exactive Plus mass spectrometer (Thermo Fisher, USA), and the samples were analysed by MS/MS at 2.0 kV. The complete peptide was detected in the Orbitrap with a resolution of 70,000; the entire scanning range was 350–1800. Ion fragments were detected in the Orbitrap with a resolution of 17,500 and a fixed first mass of 100 m/z. After investigation and scanning, the first 20 ions with a threshold ion count of more than 5 × 10⁴ were automatically changed by data-dependent mode, and 30 s were dynamically excluded. The automatic gain control was turned on to prevent overfilling of the track trap. The secondary MS data were searched by MaxQuant-integrated Andromeda search engine (v.1.5.2.8).

Obtaining differentially expressed miRNAs using a miRNA chip

After extracting the total RNA according to the above method, the quality of the total RNA was tested by formaldehyde denatured gel electrophoretic assay, and purified with the mirVana purification miRNA Isolation Kit (AM1561). The miRNA Complete Labelling and Hyb Kit, produced by Agilent, was used for dephosphorylation and labelling. After hybridisation, washing was done with 0.2 SDS, 2 × SSC. After the slides had dried, we used the Agilent chip scanner (G2565CA) for scanning and obtained the hybrid image. Finally, Agilent Feature Extraction (v10.7) software analyses were conducted on the hybrid images, and the data were extracted.

Differential gene/protein acquisition

After standardising the expression data using the DEqMS package of R language (x) R script, the differences in protein expression and miRNA expression between groups were analysed using the limma package. The significantly differentially expressed proteins and miRNAs were screened according to the criteria of | log₂FC |≥ 0.5 and adj. P < 0.05. A total of 166 differentially expressed proteins and 161 differentially expressed miRNAs were screened and visualised using the pheatmap R package.

Acquisition of apoptosis-related genes

A total of 680 apoptosis-related genes were downloaded from GO (Gene Ontology) and KEGG (Kyoto Encyclopedia of Genes and Genomes) data sets in the GSEA database, as earlier reported by Yang et al.[21] (Supplementary Table 1).

GO/KEGG enrichment analysis

GO annotation and KEGG pathway enrichment analysis were performed using the cluster Profiler R package with a false discovery rate (FDR) < 0.05. GO annotation included biological process (BP), molecular function (MF), and cellular component (CC) terms.

Protein interaction network

A protein–protein interaction network (PPI) was constructed using the STRING (https://cn.string-db.org/) database, and the proteins were further screened.

Gene set enrichment analysis (GSEA)

To determine the differences in the biological functions and signalling pathways involved in the four key proteins, their correlations with other proteins were calculated and sorted. The C2: KEGG gene set for mouse (Mus musculus) was downloaded with the minder package of the R language as the background set, and the sequenced genes were enriched and analysed in the background gene set using the GSEA function of the R language (P < 0.05).

CeRNA network and ingenuity pathway analysis (IPA)

In StarBase (https://starbase.sysu.edu.cn/), under the condition that clipExpNum ≥ 10, 20 key miRNA-related lncRNAs and circRNAs were predicted. Then, network visualisation was carried out with Cytoscape software. IPA pathway analysis was also conducted.

Statistical analysis

The experimental data in this paper are expressed as mean ± standard deviation (mean ± SD). The data were analysed and plotted using GraphPad Prism 8. The comparison of two groups was performed using a t-test; for multiple comparisons, we employed one-way and two-way ANOVA followed by Tukey’s HSD post hoc test. All differential genes and miRNAs were defined according to the criteria of |log₂FC|≥ 0.5 and adj. P < 0.05. In the differential analysis, Benjamini–Hochberg correction was applied; in the enrichment analysis, Benjamini–Hochberg correction was applied. The miRNA-related mRNA screening condition was clipExpNum ≥ 10. P < 0.05 was considered to indicate a statistically significant difference.

Data source

Proteomic data and miRNA sequencing data were collected following the aforementioned experimental steps. Exos from mouse BMSCs were extracted and divided into two groups, with three samples in each group: group G (Exos from GATA4-overexpressing BMSCs) and group N (Exos from unloaded BMSCs). Figure 1 shows a complete analysis flow of this study.

Fig. 1.

Bioinformatics analysis flowchart

Results

Effects of Exos from GATA4-overexpressing BMSCs on hypoxic cardiomyocytes

To explore the effect of Exos on hypoxic cardiomyocytes, we obtained Exos from GATA4-overexpressing BMSCs (Exos^GATA4 group) and Exos from BMSCs (Exos group). Using TEM to determine the morphology (Fig. 2A), NTA results showed a median of 143.5 nm and 145.5 nm (Fig. 2B). Western blot was used to identify exosome markers (Fig. 2C). Subsequently, results from PKH26-labelled exosomes demonstrated that both Exos from GATA4-overexpressing BMSCs and Exos from BMSCs were taken up by cardiomyocytes (Fig. 2D). After Exos from GATA4-overexpressing BMSCs were co-cultured with hypoxic cardiomyocytes, the results showed that in Exos from GATA4-overexpressing BMSCs, not only was GATA4 highly expressed (Fig. 2E), but cTnT, connexin-43, desmin, and α-actin were also highly expressed, as confirmed by RT-qPCR (Fig. 2F). Similarly, the IF results revealed that the expression of cTnT, connexin-43, desmin, and α-actin was significantly increased in the Exos^GATA4 group (Fig. 2G). The apoptosis rate was low (Fig. 2H). Additionally, the protein expression of cytochrome C, cleaved-caspase-3, and Bax was downregulated, whereas that of Bcl-2 was upregulated in the Exos^GATA4 group vs the Exos group (Fig. 2I). Our results showed that the Exos from GATA4-overexpressing BMSCs stimulated hypoxic cardiomyocyte differentiation and inhibited their apoptosis.

Fig. 2.

Effects of Exos from GATA4-overexpressing BMSCs on hypoxic cardiomyocytes. A Exosome morphology identified by TEM (scale bar = 100 nm); B NTA analysis; C Western blot was used to identify exosome markers; D IF detected PKH26-labelled exosomes (scale bar = 10 μm); E GATA4 detected by RT-qPCR (t-test); F cTnT, connexin-43, desmin, and α-actin expression detected by RT-qPCR (Two-way ANOVA); G cTnT, connexin-43, desmin, and α-actin fluorescence values detected by IF (scale bar = 20 μm); H cell apoptosis rate detected by flow cytometry (t-test); I protein expression detected by Western blot (two-way ANOVA). *P < 0.05, **P < 0.01, and ***P < 0.001 (n = 3)

Analysis of differentially expressed proteins in the Exos of mouse BMSCs

A total of 921 proteins were identified in Exos derived from mouse BMSCs (Supplementary Table 2). A total of 166 significantly differentially expressed proteins were screened in the Exos of GATA4-overexpressing BMSCs (group G) and Exos of empty BMSCs (group N) (Supplementary Table 3), including 94 upregulated proteins and 72 downregulated proteins (Fig. 3A–B). To understand the biological functions and pathways, the results of GO and KEGG enrichment analysis are depicted in Fig. 3C–D and Figure S1 (only the top 5 are displayed; Supplementary Tables 4–5). GO analysis results showed 669 enriched categories, including 56 CC terms (collagen-containing extracellular matrix, proteasome core complex, endopeptidase complex, peptidase complex, and proteasome complex), 44 MF terms (extracellular matrix structural constituent, glycosaminoglycan binding, heparin-binding, growth factor binding, and sulfur compound binding) and 569 BP terms (wound healing, renal system development, morphogenesis of a branching structure, extracellular matrix organisation, and extracellular structure organisation). In KEGG, 20 functional pathways were enriched (including proteasome, ECM–receptor interaction, coronavirus disease-COVID-19, spinocerebellar ataxia, and prion disease).

Fig. 3.

Analysis of the differentially expressed proteins in the Exos of mouse BMSCs. A Volcano plot of protein differential expression: the abscissa represents the differential expression multiple log₂FC, the ordinate represents the significant − log10 (adj. p); each dot represents a protein; red indicates a significantly upregulated protein, green indicates a significantly downregulated protein; grey indicates a non-significant protein; B heat map of differential protein expression; sample in Abscissa direction and differential expression protein in ordinate order; red indicates high expression; blue indicates low expression; C KEGG enrichment results: top 5 differentially expressed proteins,); D GO enrichment results (differentially expressed proteins): the top 5 pathways (in their order of significance) and the proteins they contain; the circle size of the pathway indicates the number of proteins contained

Key protein screening

Next, we screened the differentially expressed cell apoptosis proteins. 680 apoptosis-related human genes were converted into mouse gene format (671 remaining) and were intersected with 166 differentially expressed proteins and with the apoptosis-related genes mentioned above. As can be seen in Figs. 4A, 11 differentially expressed apoptosis proteins (DEARPs) were detected: COL2A1, CXCL12, MMP9, IL1RAP, VNN1, CLU, CD44, FGG, EPHA2, RPS7, and IGF1. The DEARPs protein–protein interaction was established using the STRING database (interaction score_min = 0.4). The results are illustrated in Fig. 4B, including 11 nodes and 15 edges. The GO results are depicted in Fig. 4C (Supplementary Table 6). The DEARPs found 707 entries in GO, including 25 in CC terms, 42 in MF terms, and 640 in BP terms. The top 5 of them included regulation of the apoptotic signalling pathway, negative regulation of the apoptotic signalling pathway, the intrinsic apoptotic signalling pathway, regulation of the intrinsic apoptotic signalling pathway, and negative regulation of the intrinsic apoptotic signalling pathway.

Fig. 4.

Key protein screening. A Venn diagram of the differentially expressed apoptosis proteins; B PPI network; C GO enrichment results (DEARPs) displaying the top 10 pathways of each classification (ranked by significance); the square size indicates the number of genes contained, and the colour indicates their significance

Fig. 11.

Exos from GATA4-overexpressing BMSCs attenuated cardiomyocyte infarction through CLU. RT-qPCR detection of transfection efficiency (A) and the expression of CLU (B) (one-way ANOVA). C RT-qPCR detected cTnT, connexin-43, desmin, and α-actin expression (two-way ANOVA); D detection of cTnT, connexin-43, desmin, and α-actin fluorescence values by IF (scale bar = 20 μm); E detection of the cell apoptosis rate by flow cytometry (one-way ANOVA); F detection of protein expression by Western blot (two-way ANOVA). *P < 0.05, **P < 0.01, and ***P < 0.001 vs the Exos group. ^#P < 0.05, ^##P < 0.01, and ^###P < 0.001 vs the Exos^GATA4 group

Variance analysis of miRNA

We further analysed the differentially expressed miRNAs. In the Exos miRNA data, a total of 1881 miRNAs were detected; the miRNA expressed in at least two samples in each group was retained, and the remaining 262 miRNAs (Supplementary Table 7). A total number of 161 differentially expressed miRNAs (Supplementary Table 8) were screened in group G and group N, including 86 upregulated miRNAs and 75 downregulated miRNAs (Fig. 5A–B). In the StarBase, 161 differentially expressed miRNAs-related mRNAs were predicted (Supplementary Table 9). A total of 5503 mRNAs were obtained as GATA4 overexpression-related genes (GATA4-OEGs) in this study. The 5503 genes related to GATA4 overexpression were analysed by GO enrichment analysis. The results are presented in Fig. 5C (Supplementary Table 10). In the GO analysis, a total number of 5308 items were enriched, including 484 CC, 580 MF, and 4244 BP terms. The top 5 of them included transcription coregulator activity, proteasome-mediated ubiquitin-dependent protein catabolic process, ubiquitin-like protein transferase activity, ubiquitin-like protein ligase binding, and ubiquitin-protein transferase activity.

Fig. 5.

Variance analysis of miRNA. A Volcano plot of miRNA differential expression. The abscissa represents the differential expression of multiple log₂FC; the ordinate represents significant-log10 (adj. p); each point in the map represents a miRNA; red indicates significantly upregulated miRNAs, green indicates significantly downregulated miRNAs; gray indicates non-significant miRNAs; B Differential protein expression heat map. At the top is the miRNA distribution heat map; at the bottom is the miRNA expression heat map. The abscissa represents the sample, and the ordinate represents the differential expression of miRNA. Different colours indicate the standardised miRNA expression: red indicates high expression, and blue indicates low expression. C GO enrichment analysis results (GATA4 overexpression-related genes)

Key protein/miRNA

The DEARPs in 3.3 and the GATA4-OEGs in 3.4 were intersected to obtain the functions in which both participated. A total of 553 common GO functions were obtained (Fig. 6A; Supplementary Table 11). The 553 DEARPs in the common GO functions and their related differentially expressed miRNAs were used to construct the regulatory network displayed in Fig. 6B. As can be seen in the figure, CXCL12, CLU, CD44, and IGF1 are the four key proteins; the 20 key miRNAs include mmu-miR-103-3p, mmu-miR-107-3p, mmu-miR-130a-3p, mmu-miR-140-3p, mmu-miR-15a-5p, mmu-miR-15b-5p, mmu-miR-16-5p, mmu-miR-199a-3p, mmu-miR-21a-5p, mmu-miR-23a-3p, mmu-miR-23b-3p, mmu-miR-25-3p, mmu-miR-27a-3p, mmu-miR-29b-3p, mmu-miR-301a-3p, mmu-miR-322-5p, mmu-miR-467 g, mmu-miR-497a-5p, mmu-miR-5127, and mmu-miR-92a-3p.

Fig. 6.

Key protein/miRNAs. A Venn of common GO functions; B gene–protein–miRNA regulatory network: the purple circle represents the gene, the prism represents the key protein, red represents upregulation (G group is higher than N group), green represents downregulation (G group is lower than N group); rectangle represents key miRNA, red represents upregulation (higher in the G group than in the N group), green represents downregulation (lower in the G group than in the N group); the dotted frame represents the key proteins involved in GO functions (the top 10 ranked by significance)

GSEA

Further, we explored the biological functions and signalling pathways involved in the four key proteins. The mouse (Mus musculus) KEGG gene set was downloaded, and the sequenced genes were enriched and analysed in the background gene set. The results of the four key proteins are displayed in Fig. 7 (Supplementary Tables 13–16). We found that CXCL12 was enriched mainly in the proteasome, hypertrophic cardiomyopathy (HCM), dilated cardiomyopathy, cytokine–cytokine receptor interaction, and focal adhesion pathways. Additionally, CLU was enriched mainly in the proteasome, hypertrophic cardiomyopathy (HCM), cytokine–cytokine receptor interaction, dilated cardiomyopathy, and focal adhesion pathways. We also found that CD44 was enriched predominantly in the proteasome, cytokine–cytokine receptor interaction, hypertrophic cardiomyopathy (HCM), focal adhesion, and dilated cardiomyopathy pathways. IGF1 was enriched mainly in small-cell lung cancer, the proteasome, focal adhesion, pathogenic Escherichia coli infection, and the cytokine–cytokine receptor interaction pathway.

Fig. 7.

Gene set enrichment analysis (GSEA). A–D The GSEA enrichment trend map displays the enrichment pathways of the top 5 pathways (ranked by their significance). The upper part depicts the process of calculating the enrichment fraction (ES, enrichment score), value, calculated as an ES value and connected into a line. The particularly obvious peaks on the far left and the far right are the ES values for the phenotypes of the gene sets. Each line in the following section represents a gene in the gene set and its sequencing position in the gene list

LncRNA/circRNA prediction

Twenty key miRNA-related lncRNAs and circRNAs were predicted using StarBase. A total of 41 lncRNAs and 95 circRNAs (Supplementary Table 17) were obtained. The mRNA–miRNA–lncRNA regulatory network is presented in Fig. 8A; the mRNA–miRNA–circRNA regulatory network is depicted in Fig. 8B.

Fig. 8.

LncRNA/circRNA prediction. A mRNA–miRNA–lncRNA regulatory network. The red represents mRNAs, the yellow miRNAs, and the green lncRNAs; B mRNA–miRNA–circRNA regulatory network. The red represents mRNAs, the yellow miRNAs, and the blue circRNAs

IPA

All differentially expressed genes were analysed using IPA, and the pathways of the key proteins (Z-score > 1) were screened out. Eight pathways are displayed in Fig. 9A. The highest Z-score pathway was obtained for LXR/RXR activation, and only CLU was found to participate in this pathway. Its regulatory relationship is illustrated in Fig. 9B. The interaction network between key proteins and their associated molecules is presented in Fig. 9C.

Fig. 9.

Ingenuity pathway analysis (IPA). A Key targets involved in the pathway; B regulatory relationships in the top 1 pathway; C intermolecular interactions network

Expression of key proteins and miRNAs

We performed a differential analysis of the expression of key proteins and miRNAs in groups G and N. The results showed that the expression of CLU and IGF1 in the G group was higher, and the expression of CXCL12 and CD44 was lower than that in the N group (Fig. 10A, Supplementary Table 18). In the key miRNAs, except that the expression of mmu-miR-467 g and mmu-miR-5127 was low, the expression of the others was high (Fig. 10C, Supplementary Table 18). Additionally, we found that CXCL12, CD44, mmu-miR-467 g, and mmu-miR-5127 were downregulated in the Exos from the GATA4-overexpressing BMSCs treatment group, whereas CLU, IGF1, and the other 18 key miRNAs were upregulated in the Exos from the GATA4-overexpressing BMSCs treatment group (Fig. 10B–D). The above results indicate that CLU, IGF1, CXCL12, CD44, and the 20 key miRNAs have an important role in MI.

Fig. 10.

Expression of key proteins and miRNA. A The expression of key proteins vs the N group, *P < 0.05, **P < 0.01, and ***P < 0.001; B the expression of key proteins detected by RT-qPCR vs the Exos group (one-way ANOVA), *P < 0.05, **P < 0.01, and ***P < 0.001; C the expression of the key miRNAs vs the N group, *P < 0.05, **P < 0.01, and ***P < 0.001; D the expression of the key miRNAs by RT-qPCR vs the Exos group (one-way ANOVA), *P < 0.05, **P < 0.01, and ***P < 0.001

Exos from GATA4-overexpressing BMSCs attenuated cardiomyocyte infarction through CLU

To elucidate the role of Exos from GATA4-overexpressing BMSCs in cardiomyocyte infarction, we transfected with si-CLU. Using RT-qPCR detection of transfection efficiency. Results indicate successful transfection; si-CLU#3 was selected for subsequent studies (A). Exos from GATA4-overexpressing BMSCs were co-cultured with hypoxic cardiomyocytes. The results showed that CLU expression in the Exos^GATA4 group was higher than that in the Exos^GATA4 + si-CLU group (Fig. 11B). CLU knockdown significantly restored the effect of the Exos^GATA4 group and decreased the mRNA expression and fluorescence value of cTnT, connexin-43, desmin, and α-actin (Fig. 11C–D). The knockdown of CLU increased the cardiomyocyte cell apoptosis, which was detected by flow cytometry (Fig. 11E) and Western blot (Fig. 11F). Next, we transfected cardiomyocytes with oe-CLU or si-IGF1 and detected the transfection efficiency (Figure S2A-D). Rescue experiments and parallel IGF1 perturbation studies similarly demonstrated that CLU overexpression potentiated the effects observed in the Exos^GATA4 group, increasing mRNA expression and fluorescence values for cTnT, desmin, actin, and α-actinin (Figure S2E-F). CLU overexpression further inhibited cardiomyocyte apoptosis (Figure S2G), reduced the expression of Cytochrome C, Cleaved-caspase-3, and Bax, and increased Bcl-2 (Figure S2H). While co-transfection with si-IGF1 further suppressed the effects observed in the CLU overexpression and Exos^GATA4 groups. Therefore, our results indicate that CLU is involved in the attenuation of cardiomyocyte infarction by Exos GATA4-overexpressing BMSCs.

Exos from GATA4-overexpressing BMSCs improved the cardiac function of MI mice via CLU regulation

In the MI mouse model, Exos from GATA4-overexpressing BMSCs increased EF and FS, but this effect was reversed by sh-CLU (Fig. 12A). The results of TTC and HE staining showed that Exos from GATA4-overexpressing BMSCs limited the size of MI (Fig. 12B) and alleviated myocardial damage in MI mice (Fig. 12C). However, this function was reversed by sh-CLU. After MI, c-kit reduced infarct size, protected cardiac function, and improved cardiomyocyte differentiation [22]. Our study found that Exos^GATA4 increased c-kit expression, but after CLU knockdown, the c-kit expression was reduced (Fig. 12D). Finally, our results indicated that the cardiomyocyte apoptosis levels in the MI and Exos^GATA4 + sh-CLU groups increased significantly, as confirmed by TUNEL staining (Fig. 12E). RT-qPCR and IHC results showed that Exos from GATA4-overexpressing BMSCs upregulated the expression of CLU, and was downregulated in the Exos^GATA4 + sh-CLU group (Fig. 12F–G). Our findings further confirm CLU’s involvement in the LXR/RXR pathway. Exos from GATA4-overexpressing BMSCs activated the LXR/RXR signalling, increased the expression of ABCA1, ABCG1, LXR, RXRα, and SREBP-1c (Fig. 13A–B), inhibited TNF-α, IL-6, and IL-1β secretion, promoted IL-10 secretion (Fig. 13C), and reduced lipid accumulation (Fig. 13D). Knockout of CLU inhibits activation of this pathway, leading to increased lipid accumulation and cellular inflammation within myocardial tissue. This effect thereby reverses the beneficial impact of Exos from GATA4-overexpressing BMSCs on MI.

Fig. 12.

Exos from GATA4-overexpressing BMSCs improve cardiac function in MI mice by CLU regulation. A Left ventricular FS and EF of MI mice (two-way ANOVA); B TTC staining; C HE staining (scale bar = 50 μm); D detection of c-kit expression by IHC assay (scale bar = 50 μm, one-way ANOVA); E detection of cell apoptosis by TUNEL staining (scale bar = 20 μm); RT-qPCR (F) and IHC (G) were used to detect the expression of CLU (scale bar = 50 μm) (one-way ANOVA). ***P < 0.001 vs the MI group, and ^###P < 0.001 vs the Exos.^GATA4 group (n = 6)

Fig. 13.

Exos from GATA4-overexpressing BMSCs activated the LXR/RXR signalling. RT-qPCR (A) and Western blot (B) were used to detect the expression of ABCA1, ABCG1, LXR, RXRα, and SREBP-1c (one-way/two-way ANOVA); C ELISA detection of TNF-α, IL-6, IL-1β and IL-10 (one-way ANOVA); D Oil Red O staining (scale bar = 20 μm); ***P < 0.001 vs the MI group, and ^###P < 0.001 vs the Exos^GATA4 group (n = 6)

Discussion

Myocardial infarction (MI) is primarily the result of myocardial necrosis caused by ischaemia and hypoxia, and is one of the leading causes of death globally. It could cause irreversible damage to myocardial tissue [23, 24]. Abnormal apoptosis plays an important role in the pathogenesis of many diseases, including the occurrence and development of heart failure after MI [25]. miRNA has been confirmed to be a biomarker for the diagnosis, treatment, and prognosis of several diseases, including MI [26]. Many in vitro and in vivo studies have been conducted to determine the potential roles of apoptosis and miRNA in the pathogenesis of MI and their potential protective effect against MI. In this study, we used bioinformatics to identify novel key apoptosis-related genes and miRNAs that may be significant in the treatment and prognosis of MI.

Maharsy et al. reported that cardiomyocyte death-induced GATA4 activated the transcription of the anti-apoptosis genes Bcl-2 and Bcl-xL, and exerted a protective effect against DOX-induced cardiomyocyte death [27]. Our study confirmed that exosomes (Exos) from GATA4-overexpressing BMSCs stimulate hypoxic cardiomyocyte differentiation and inhibit their apoptosis. In agreement with the results of previous studies, we found that Exos isolated from GATA4-overexpressing BMSCs effectively attenuated MI and preserved cardiac function [28]. Apoptosis was finely regulated by genes involved in programmed cell death. Apoptosis cardiomyocyte death plays a key role in ventricular remodelling after infarction, further protecting cardiac function [29]. In cardiovascular diseases, inhibition of apoptosis is important for improving cardiac function. Therefore, the regulation of apoptosis protein genes plays a key role in the treatment and prognosis of MI. We speculated that Exos from GATA4-overexpressing BMSCs could inhibit cardiomyocyte apoptosis by regulating apoptosis genes. In this study, 11 differentially expressed apoptosis proteins (DEARPs) were obtained and screened through proteomic data analysis: COL2A1, CXCL12, MMP9, IL1RAP, VNN1, CLU, CD44, FGG, EPHA2, RPS7, and IGF1. Our results showed that these proteins were highly associated with the apoptosis signalling pathway.

Exos isolated from mesenchymal stem cells (MSCs) have bioactive components, such as proteins, nucleic acids (DNA, mRNA, miRNA), lipids, and enzymes. They perform different key functions in cell-to-cell communication in various cell types [30]. miRNA is well known as the most abundant molecule in the Exos. Accumulating evidence has shown that Exos play a central role in miRNA-mediated intercellular communication, serving as information carriers facilitating the increase in the self-renewal of stem cells and promoting their differentiation and pluripotency [31, 32]. Wang et al. found that ADSC-Exos contributed to the improvement of cardiac function in MI mice; this process was realised via miRNA-205 [33]. Interestingly, our results also indicate that Exos secreted by GATA4-overexpressing BMSCs contribute to improvement in MI progression, although the specific molecular mechanisms remain unclear.

Therefore, this study screened miRNA sequencing data and detected a total number of 161 differentially expressed miRNAs, including 86 upregulated and 75 downregulated miRNAs. Further, 5503 miRNA-related mRNAs were obtained as GATA4 overexpression-related genes (GATA4-OEGs). We established that Exos from GATA4-overexpressing BMSCs inhibited apoptosis and attenuated MI. We obtained 553 common GO functions from DEARPs and GATA4-OEGs and identified the genes encoding DEARPs in common GO functions and their related differentially expressed miRNAs. Furthermore, we detected four key proteins: CXCL12, CLU, CD44, and IGF1. Twenty key miRNAs were identified, including mmu-miR-103-3p, mmu-miR-107-3p, mmu-miR-130a-3p, mmu-miR-140-3p, mmu-miR-15a-5p, mmu-miR-15b-5p, mmu-miR-16-5p, mmu-miR-199a-3p, mmu-miR-21a-5p, mmu-miR-23a-3p, mmu-miR-23b-3p, mmu-miR-25-3p, mmu-miR-27a-3p, mmu-miR-29b-3p, mmu-miR-301a-3p, mmu-miR-322-5p, mmu-miR-467 g, mmu-miR-497a-5p, mmu-miR-5127, and mmu-miR-92a-3p. Our results showed that the expression of 18 key miRNAs was significantly upregulated except for that of mmu-miR-467 g and mmu-miR-5127 in the Exos from GATA4-overexpressing BMSCs. miR-130a-3p was previously found to promote cell proliferation, inhibit apoptosis, and improve MI induced by PDE4D [34]. Furthermore, earlier evidence has revealed that ellagic acid (EA) improves ventricular remodelling after MI by upregulating the expression of miR-140-3p [35]. In another study, the exocrine miR-25-3p derived from MSC promoted cardiomyocyte survival and targeted EZH2, inhibiting post-MI inflammation [36]. The upregulation of miR-29b-3p was established to protect cardiomyocytes from hypoxia-induced injury by downregulating TRAF5. Hence, targeting TRAF3 with miR-5b-29p may be a potential treatment for AMI [37]. However, our miRNA sequencing revealed that Exos from GATA4-overexpressing BMSCs contain a complex mixture of regulatory RNAs, such as miR-92a, the miR-15 family, and miR-130a. These miRNAs may exert both detrimental effects and beneficial roles under specific conditions following cardiac injury. For instance, miR-92a exhibits high expression in cardiomyocyte-derived Exos and fibroblasts isolated post-MI [38]. miR-92a-2-5p mitigates oxidative stress damage in cardiomyocytes whilst simultaneously promoting cardiomyocyte apoptosis [39–41]. Downregulation of miR-15b-5p exerts a protective effect on cardiomyocytes [42]. miR-130a-3p can mitigate cardiomyocyte hypertrophy [43] and is downregulated in cardiomyocytes following MI [34, 44]. Inhibition of miR-130a enhances cardiomyocyte viability after hydrogen peroxide/hypoxia (H/R) treatment and reduces apoptosis rates [45]. Exos contain multiple molecules, and their functional outcomes depend on the combined effects of their entire cargo rather than the action of any single miRNA [46]. Although miR-199a-3p typically exerts detrimental effects, it has been reported to promote myocardial cell repair following hydrogen peroxide treatment in MI [47]. Furthermore, miR-199a-3p has been documented to prevent MI [48]. Consequently, it is challenging to fully comprehend the function of Exos’ miRNAs. In subsequent studies, we shall further investigate the mechanisms by which these 20 miRNAs interact with cardiomyocytes in MI. Additionally, 41 core miRNA-related lncRNAs and 95 circRNAs were detected. Our IPA analysis results showed that LXR/RXR signal transduction regulates inflammation, cholesterol homeostasis, and lipid and glucose metabolism. Finally, we verified the expression levels of differential proteins and miRNAs in hypoxic cardiomyocytes co-cultured with Exos from GATA4-overexpressing BMSCs by RT-qPCR and obtained consistent results.

Our research has revealed that CXCL12 and CD44 are highly expressed in MI, whereas their expression is reduced following treatment for cardiomyocyte hypoxia or MI with Exos from GATA4-overexpressing BMSCs. This finding appears to contradict a substantial body of published literature. Multiple studies show the beneficial effects of CXCL12 (SDF-1) [49]; SDF‑1 delivery (protein, protease‑resistant analogue, or overexpression) improves function by recruiting reparative cells and supporting the endothelium. The CXCR4/CXCL12 pathway promotes the viability, proliferation, migration, adhesion, and tubule formation of endothelial progenitor cells (EPCs) [50, 51]. However, other studies have found that following myocardial cell hypoxia, CXCL12 expression increases. Knocking out CXCL12 reduces the infarct size and decreases myocardial cell apoptosis, yielding beneficial effects [52]. Similarly, Silke et al. also discovered that CXCL12-overexpressing transgenic (Tg) rats revealed impaired cardiac function post-MI, accompanied by enhanced fibrosis. Collectively, their findings demonstrate that cardiomyocyte-derived CXCL12 is not involved in cardiac development but has adverse effects on the heart after injury by promoting inflammation and fibrosis [53]. This finding may be related to the temporal dynamics of CXCL12 signalling. The expression of CD44 is increased in MI and infarcted hearts, and is related to ischaemic angiogenesis and plasma exocrine secretion, particularly in the early stages of MI. The high expression of CD44 lays the groundwork for the subsequent progression of the condition [54]. CD44 expression is significantly elevated in blood vessels within vulnerable zones, with high CD44 expression being closely associated with pro-inflammatory cytokines. Following acute MI, IL-6 further enhances CD44 expression in cardiac fibroblasts [55, 56]. However, the expression of CLU and IGF1 was significantly upregulated in the Exos from GATA4-overexpressing BMSCs. Cell protective sex chaperone protein CLU is produced and secreted under the action of a stress signal. Studies have shown that CLU is related to the survival of patients with heart failure [57]. Numerous studies have demonstrated that CLU can regulate apoptosis [58, 59]. The overexpression of CLU can result in secretion of isomers, which further reduces the apoptosis induced by MG132 treatment [60]. In ischaemic cardiomyocytes, CLU can prevent apoptosis by regulating the expression of matrix metalloproteinases and stimulating angiogenesis. CLU plays an important role in the recovery of cell function and heart protection [61]. CLU has also been found to be associated with autophagy and the PI3K/AKT/mTOR signalling pathway in the regulation of apoptosis. Overexpression of CLU protects against high-temperature-induced cardiomyocyte apoptosis by mediating the PI3K/Akt signalling pathway [62]. Similarly, our study further confirms that Exos from GATA4-overexpressing BMSCs exert a protective effect on cardiomyocytes during MI progression by upregulating CLU. Many studies have found that insulin-like growth factor 1 (IGF1) affects vascular function and atherosclerosis by inducing cytochrome C release and apoptosis through anti-inflammation and anti-apoptosis activities [63]. Heinen et al. have confirmed that IGF1 improved the function of the ischaemic area and the overall cardiac function and promoted angiogenesis after MI [64]. Meanwhile, downregulating IGF1 also reverses the protective effect exerted by Exos from GATA4-overexpressing BMSCs on cardiomyocytes. Therefore, our findings confirm that Exos from GATA4-overexpressing BMSCs exert beneficial effects on MI progression by regulating CLU and IGF1. Previous studies have already reported that the LXR/RXR activation pathway is closely associated with CLU [65], whilst IGF1 also appears to be linked to the LXR/RXR signalling pathway [66]. Interestingly, our findings also indicate that the protective role of Exos from GATA4-overexpressing BMSCs in MI progression is closely associated with the LXR/RXR signalling pathway. The LXR/RXR pathway is closely implicated in the regulation of lipid metabolism, cholesterol transport, and inflammatory responses. Inhibition of the LXR/RXR pathway activation leads to intracellular lipid accumulation, ultimately resulting in cell death and the inflammation phenomenon termed lipotoxicity [67]. After ischaemic injury, the LXR/RXR pathway reduces cardiomyocyte hypertrophy and promotes cell survival, regulates myocardial metabolism, enhances angiogenesis, and reduces fibrosis, thus playing an important role in MI remodeling [68]. Therefore, we further investigated the mechanism of action of CLU in the treatment of MI with Exos from GATA4-overexpressing BMSCs. The findings further confirm CLU's involvement in the LXR/RXR pathway. Knockout of CLU inhibits the activation of this pathway, leading to increased lipid accumulation and cellular inflammation within myocardial tissue. This effect thereby reverses the beneficial impact of Exos from GATA4-overexpressing BMSCs on MI. This study not only links the therapeutic efficacy of Exos from GATA4-overexpressing BMSCs to the specific activation of the LXR/RXR pathway, but also reveals the central role of this pathway in stem cell exosome therapy.

Conclusion

Through systematic screening and validation, this study demonstrated that Exos derived from GATA4-overexpressing BMSCs improve cardiac function after MI by activating the LXR/RXR pathway. This occurs by regulating key apoptotic proteins such as CLU and miRNAs. These findings provide novel evidence that GATA4-expressing BMSCs improve post-MI cardiac function and establish a new molecular mechanism for stem cell-derived Exos therapy for MI. Clinically, CLU emerges as a potential therapeutic target due to its pivotal role in regulating cardiomyocyte apoptosis and inflammatory responses. Future development of CLU-targeting drugs or gene therapy strategies may enhance myocardial protection. GATA4-overexpressing BMSC-derived Exos can serve as “natural nanomedicines” for acellular therapy for MI. LXR/RXR pathway modulators may function as adjunctive therapeutic agents to enhance cardiac function recovery. The 20 key miRNAs and 4 key proteins identified in this study show promise as biomarkers for MI prognosis assessment or treatment response prediction. Future research should validate the safety and efficacy of Exos in large animal models to advance their clinical translation.

Limitations of the study

This study also has certain limitations: the mechanism of action of miRNAs in Exos remains unclear, and miR-92a, the miR-15 family, and miR-130a may exert dual effects in MI. Further follow-up studies need to clarify the interaction between these 20 key miRNAs and their lncRNAs/circRNAs and related pathways. Moreover, among the four key proteins, there are also controversial ones, such as CXCL12. Therefore, CLU was selected for functional experiments in this study. However, we must not overlook the crucial role of other key proteins in MI. Future research will expand the sample size to thoroughly investigate the mechanism of action of Exos from GATA4-overexpressing BMSCs in MI. This will involve further exploration of the molecular mechanisms involving key proteins and miRNAs. Furthermore, this study has only been validated in mouse models and cellular experiments and remains distant from clinical application. Subsequent clinical validation of these findings is warranted to explore their potential applications and clinical value in MI treatment. Future research should conduct in-depth mechanistic validation of these results and assess their long-term efficacy and safety.

Abbreviations

MI

Myocardial infarction

Exos

Exosomes

BMSCs

Bone marrow mesenchymal stem cells

miRNA

MicroRNA

lncRNA

Long non-coding RNA

MSCs

Mesenchymal stem cells

TMT

Tandem mass tags

FBS

Fetal bovine serum

TEM

Transmission electron microscopy

PVDF

Polyvinylidene fluoride

IF

Immunofluorescence

ELISA

Enzyme-linked immunosorbent assay

TNF-α

Mouse tumour necrosis factor alpha

IL-1β

Interleukin-1β

IL-10

Interleukin-10

IL-6

Interleukin-6

PBS

Phosphate-buffered saline

FS

Fractional shortening

EF

Ejection fraction

IHC

Immunohistochemistry

HE

Hematoxylin–eosin

GO

Gene Ontology

KEGG

Kyoto Encyclopedia of Genes and Genomes

FDR

False discovery rate

BP

Biological process

MF

Molecular functions

CC

Cellular components

PPI

Protein–protein interaction network

GSEA

Gene set enrichment analysis

IPA

Ingenuity pathway analysis

DEARPs

Differentially expressed apoptosis proteins

GATA4-OEGs

GATA4 overexpression-related gene

Author contributions

Conceptualization, Writing—review and editing: Tao Ma. Conceptualization, Writing -original draft: Shunfu Yang. Validation, Resources: Yan Gao. Validation: Xinxin Wu. Resources: Si Li. Formal analysis: Dan Yan. Conceptualization, Writing—review and editing, Writing—original draft, Validation: Jigang He.

Funding

The National Natural Science Foundation of China (NO: 82060299). Medical discipline Leader Project of Yunnan Provincial Health Commission (NO: D-2019020). Yunnan Provincial Government’s Ten Thousand Talents Plan-Youth Top Talents Project (NO: KH-SWR-QNBJ-2019-002). Clinical Medical Center of the first People’s Hospital of Yunnan Province (NO: 2022LCZXKF-HX05). Kunming Medical Joint Project-Outstanding Youth Training Project (NO: 202101AY070001-034). Yunnan First People’s Hospital Clinical Medicine Center Open Project (NO: 2022LCZXKF-HX05). Kunming Medical Union Special Project-General Project (NO: 202101AY070001-272). Kunming Medical Joint Special Project-General Project Yunnan Province “Xingdian Talent Support Plan” Famous Medical Special Project (NO: XDYC-MY-2022-0037). Yunnan Provincial Research Project on Undergraduate Education Teaching Reform in 2025 (NO: JGYF251042). All fund managers are Jigang He.

Data availability

Data is applicable after the approval of co-authors.

Declarations

Ethics approval and consent to participate

Our study was approved by the Medical Ethics Committee of The First People’s Hospital of Yunnan Province. All datasets in the present study were downloaded from public databases, including GSEA, GEO, GO, KEGG, STRING, MSigDB and StarBase database. These public databases allowed researchers to download and analyse public datasets for scientific purposes. Thus, the use of public database data in this study was reviewed and exempted by the Medical Ethics Committee of The First People’s Hospital of Yunnan Province. The current research follows the GSEA, GEO, GO, KEGG, STRING, MSigDB and StarBase data access policies and publication guidelines. GSEA, GEO, GO, KEGG, STRING, MSigDB and StarBase belong to public databases. The patients involved in the database have obtained ethical approval. Users can download relevant data for free for research and publish relevant articles. Our study is based on open-source data, so there are no ethical issues and other conflicts of interest. And the animal experiments with the assistance of Yunnan Labreal Biotech Co., Ltd., the animal experiment was conducted and received approval from the Experimental Animal Ethics Committee of Yunnan Labreal Biotech Co., Ltd (IACUC Issue No. PZ20221003) and complied with ARRIVE Guidelines 2.0.

Competing interests

The authors declare no competing interests.