Empagliflozin attenuates cardiac microvascular ischemia/reperfusion through activating the AMPKα1/ULK1/FUNDC1/mitophagy pathway

Abstract

Mitophagy preserves microvascular structure and function during myocardial ischemia/reperfusion (I/R) injury. Empagliflozin, an anti-diabetes drug, may also protect mitochondria. We explored whether empagliflozin could reduce cardiac microvascular I/R injury by enhancing mitophagy. In mice, I/R injury induced luminal stenosis, microvessel wall damage, erythrocyte accumulation and perfusion defects in the myocardial microcirculation. Additionally, I/R triggered endothelial hyperpermeability and myocardial neutrophil infiltration, which upregulated adhesive factors and endothelin-1 but downregulated vascular endothelial cadherin and endothelial nitric oxide synthase in heart tissue. In vitro, I/R impaired the endothelial barrier function and integrity of cardiac microvascular endothelial cells (CMECs), while empagliflozin preserved CMEC homeostasis and thus maintained cardiac microvascular structure and function. I/R activated mitochondrial fission, oxidative stress and apoptotic signaling in CMECs, whereas empagliflozin normalized mitochondrial fission and fusion, neutralized supraphysiologic reactive oxygen species concentrations and suppressed mitochondrial apoptosis. Empagliflozin exerted these protective effects by activating FUNDC1-dependent mitophagy through the AMPKα1/ULK1 pathway. Both in vitro and in vivo, genetic ablation of AMPKα1 or FUNDC1 abolished the beneficial effects of empagliflozin on the myocardial microvasculature and CMECs. Taken together, the preservation of mitochondrial function through an activation of the AMPKα1/ULK1/FUNDC1/mitophagy pathway is the working mechanism of empagliflozin in attenuating cardiac microvascular I/R injury.

Graphical abstract

Highlights

• •Empagliflozin reduces I/R-induced microvascular damage.
• •Empagliflozin suppresses I/R-induced endothelial cell damage.
• •Empagliflozin activates FUNDC1-dependent mitophagy through the AMPKα1/ULK1 pathway.
• •Ablation of FUNDC1 or AMPKα1 abolishes the protective effects of empagliflozin against I/R-induced microvascular damage.

1. Introduction

Coronary thrombosis can induce ischemic stress in the heart by restricting the flow of fresh oxygen and nutrients to the myocardium, thus triggering oxidative stress, calcium overload and apoptosis in cardiomyocytes [1]. The pathological course of myocardial ischemia is often termed as myocardial infarction, which has been ranked as the main cause of death in elderly patients. The standard therapeutic approach for myocardial infarction is the rapid and timely restoration of the coronary blood flow to the ischemic heart. Although such reperfusion strategies can alleviate ischemic damage, the re-introduction of fresh blood is always associated with additional damage to the post-ischemic myocardium [2]. The occurrence of ischemia/reperfusion (I/R) injury increases the morbidity and mortality of myocardial infarction, thus limiting the clinical benefits of reperfusion strategies [3]. Compared to cardiomyocyte I/R injury, cardiac microvascular I/R injury is a neglected topic [4,5], so its molecular mechanism is not fully understood and few drugs have been validated or used for its management.

Mitochondria are crucial for coordinating extracellular damage signals and maintaining cardiac microvascular endothelial function [6]. When myocardial perfusion is restricted, energy metabolism in cardiac microvascular endothelial cells is reduced; thus, mitochondrial fission is activated to augment the mitochondrial population and meet the cellular metabolic demand [7]. However, pathological mitochondrial fission damages mitochondrial DNA (mtDNA) during mitosis, thus impairing mitochondrial redox biology, triggering oxidative stress, reducing the mitochondrial membrane potential and inducing the opening of the mitochondrial permeability transition pore (mPTP) [7,8]. Under such conditions, mitochondria switch from electric generators to endothelial apoptosis inducers [9]. Endothelial cell dysfunction or death contributes to microvessel swelling and vascular wall collapse, further interrupting the flow of fresh blood to the post-ischemic myocardium [10,11].

Mitophagy is the evolutionarily conserved process of repairing poorly structured mitochondria with the help of lysosomes [11]. Mitophagy has been reported to preserve microvascular structure and function under various pathophysiological conditions, including myocardial I/R injury [4], diabetes-related microvascular damage [12], sepsis-associated microcirculatory injury [13] and smoke-induced pulmonary endothelial injury [14]. However, it has been difficult to find an effective drug that specifically activates mitophagy and alleviates mitochondrial damage to improve the resistance of microvessels to I/R injury.

Empagliflozin is a novel anti-diabetes drug that promotes the urinary excretion of glucose [15]. In the EMPA-REG OUTCOME clinical trial [16], empagliflozin treatment was found to reduce the risk of death from cardiovascular causes in diabetic patients with cardiovascular disease. Functional analyses indicated that empagliflozin administration improved the ratio of adenosine monophosphate (AMP) to adenosine triphosphate (ATP) in cardiomyocytes, thus accelerating glucose metabolism and increasing cell viability [17]. Animal studies and molecular experiments have illustrated that empagliflozin protects not only cardiomyocytes, but also endothelial cells; for instance, empagliflozin increased phosphorylated endothelial nitric oxide synthase (p-eNOS) expression and nitric oxide production in cardiac endothelial cells in a murine model of heart failure [18]. Empagliflozin also attenuated oxidative stress-induced endothelial barrier dysfunction through an undefined mechanism [19]. Regarding mitochondrial biology, empagliflozin was found to inhibit hyperglycemia-induced mitochondrial fission [20] and lipopolysaccharide-stimulated mitochondrial oxidative stress [21] by activating AMP-activated protein kinase catalytic subunit alpha 1 (AMPKα1).

Based on the above data, we investigated whether empagliflozin could protect against cardiac microvascular I/R injury by sustaining mitochondrial homeostasis. We specifically examined whether empagliflozin could ameliorate I/R-induced endothelial mitochondrial damage by activating mitophagy, and assessed its molecular mechanism in cardiac microvascular I/R injury.

2. Materials and methods

2.1. Mice

AMPKα1^f/f mutant mice (stock no.: 014141, The Jackson Laboratory) possess loxP sites flanking exon 3 of the Prkaa1 gene. FUNDC1^f/f mutant mice were generated as our previously described [22]. Tie2^Cre (Tek-Cre) transgenic mice have the mouse endothelial-specific receptor tyrosine kinase (Tek or Tie2) promoter directing the expression of Cre recombinase (stock no.: 008863, The Jackson Laboratory). Either AMPKα1^f/f mutant mice or FUNDC1^f/f mutant mice were crossed with Tie2^Cre mice to generate endothelial-specific AMPKα1-knockout (AMPKα1^EKO) mice or endothelial-specific FUNDC1-knockout (FUNDC1^EKO) mice. Tie2^Cre mice served as controls. All mice were maintained on a C57BL/6 background and were used in this study at the age of 10 weeks. Male mice were used, as they are the best-validated model for cardiac I/R injury, and male gender is well recognized as a significant risk factor for cardiovascular diseases.

2.2. Mouse model of myocardial I/R injury

All experiments were conducted according to the ethical committee guidelines for animal experimentation of Southern Medical University. The mice were randomly assigned to either the sham operation group or the myocardial ischemia (45 min)/reperfusion (2 h) (I/R) injury group. A single sequence of simple random assignments was used, in accordance with previous studies [23]. Allocation concealment was established using ear tag-based sequentially numbered mice. Mice (10 weeks old) were treated with empagliflozin (10 mg/kg/d) seven days before myocardial I/R injury, in accordance with a previous study [20,24]. Empagliflozin was kindly provided by Boehringer Ingelheim Pharma GmbH & Co. KG, Germany, and was administered via oral gavage with 0.5% hydroxyethylcellulose as the vehicle. Compound C (CC, Selleck, Cat. S7306, 10 mg/kg/d) was used to inhibit empagliflozin-induced AMPK activation in certain mice (the CC group).

2.3. Echocardiography

Two-dimensional echocardiography was performed in anesthetized mice (3.0% isoflurane with O₂ flow at 1.0 L/min) using Vevo 2100 echo with the MS400 transducer. Echocardiographic images were recorded along parasternal short axis, and then analyzed off-line (Vevo 2100 software) by an observer blinded to the treatment groups. Ejection fractions (%EF) and percent fractional shortening (%FS) were calculated [25].

2.4. CMEC isolation

CMECs were isolated from mice after myocardial I/R injury, in accordance with previous research [26,27]. In brief, isolated hearts were washed extensively with Hanks' Balanced Salt Solution with Ca²⁺ and Mg²⁺ in a Petri dish. The atria and right ventricle were removed, while the left ventricle was used for subsequent experiments. Excised pieces of the left ventricle were incubated with 0.2% (w/v) collagenase type I (Gibco, USA) for 10 min and 0.25% (w/v) trypsin (Hyclone, USA) for 5 min at 37 °C in a shaking bath. After the final dissociation step, collagenase was neutralized through the addition of 500 μL of fetal bovine serum, and the mixture was transferred through a filter (200 μM) into a 50-mL Falcon tube. The filtrates were centrifuged at 80×g at 4 °C for 1 min. The sedimented CMECs at the bottom of the tube were resuspended in 5 mL of Hanks’ Balanced Salt Solution I plus bovine serum albumin, and were plated in a 6-cm² dish at 37 °C [28].

2.5. Transmission electron microscopy

Heart tissues were fixed with 2.5% glutaraldehyde, 2.5% polyvidone 25 and 0.1 M sodium cacodylate (pH 7.4). After being washed with 0.1 M sodium cacodylate buffer (pH 7.4), the samples were post-fixed in the same buffer containing 2% osmium tetroxide and 1.5% potassium ferrocyanide for 1 h [8]. The samples were then washed in water, contrasted en bloc with uranyl acetate, dehydrated using an ascending ethanol series and embedded in a Durcupan ACM-based resin [29]. Ultrathin sections were cut with a Reichert Ultracut S ultramicrotome (Science Service, Munich, Germany) and contrasted with lead citrate. Images were viewed and captured with an EM 10 CR electron microscope (Carl Zeiss Gemini, Germany), and were analyzed by an independent blinded investigator [30].

2.6. qRT-PCR

All reagents and containers used for RNA processing were RNase-free grade or treated with 0.1% DEPC (4387937, Thermo Fisher) to eliminate RNase contaminants. Total RNA was extracted in TRIzol (15596018, Invitrogen) according to the manufacturer's instructions [7]. Reverse transcription of 500 ng RNA was performed to synthesize cDNA using random primers (SO142, Thermo Fisher) and RevertAid H Minus Reverse Transcriptase (EP0452, Thermo Fisher) according to the manufacturer's instructions [31]. The qPCR assays were carried out on a StepOnePlus system (Applied Biosystem) using SYBR Green Master Kit (A25918, Thermo Fisher). Three biological replicates per condition were measured. Gene expression was normalized to GAPDH expression, and changes in gene expression were calculated using the 2^-ΔΔCT method. The primers used in our study were as follows: TNFα (Forward, 5′- AGATGGAGCAACCTAAGGTC-3'; Reverse, 5′GCAGACCTCGCTGTTCTAGC-3′), IL6 (Forward, 5′-CAGACTCGCGCCTCTAAGGAGT3'; Reverse, 5′-GATAGCCGATCCGTCGAA-3′), MCP1 (Forward, 5′GGATGGATTGCACAGCCATT-3'; Reverse, 5′-GCGCCGACTCAGAGGTGT-3′).

2.7. Western blotting

Tissues and cells were lysed in RIPA buffer (R0278, Sigma-Aldrich) supplemented with protease inhibitors (36978, Thermo Fisher) and phosphatase inhibitors (P5726, Sigma-Aldrich) on ice for 10 min. The lysates were collected and centrifuged at 14000 rpm for 15 min at 4 °C. Total protein concentrations in the supernatant were determined by a Bio-Rad protein assay kit (5000113–115, Bio Rad, Bradford method). Following the addition of LDS sample buffer (4x, NP0007, Lifetechnologies) [32], the samples were boiled at 90 °C for 10 min. Twenty microgram proteins were separated by 10–15% SDS-PAGE and transferred onto 0.2 μm nitrocellulose membranes (4685.1, Carl Roth). Following blocking with 5% non-fat milk for 1 h at room temperature, the membranes were incubated with primary antibodies at 4 °C overnight. The next day, the membranes were incubated with secondary antibodies for 1 h. After washing by Tris buffered saline with 0.05% Tween 20 (TBST) three times, the membranes were incubated visualized by the Pierce™ ECL Plus Western Blotting Substrate (32132, Lifetechnologies) using the ChemiDoc gel imaging system (BIORAD). Intensity of each protein strips was quantified using Image J software. Primary antibodies for immunoblotting were as follows: ET-1 (1:1000, Abcam, #ab178454), eNOS (1:1000, Abcam, #ab199956), p-eNOS (1:1000, Abcam, #ab215717), Fak (1:1000, Abcam, #ab40794), Src (1:1000, Abcam, #ab133283), Fak (1:1000, Abcam, #ab40794), Drp1 (1:1000, Abcam, #ab184247), p-Drp1 (1:1000, Cell Signaling Technology, #4494), FUNDC1 (1:1000, Abcam, #ab224722), Fis1 (1:1000, Abcam, #ab156865), Mfn2 (1:1000, Abcam, #ab124773), Opa1 (1:1000, Abcam, #ab227868), GAPDH (1:1000, Abcam, #ab8245), Atg5 (1:1000, Abcam, #ab108327), Beclin1 (1:1000, Abcam, #ab207612), p62 (1:1000, Abcam, #ab91526), LC3II (1:1000, Abcam, #ab192890), Tom20 (1:1000, Abcam, #ab186735), AMPK (1:1000, Abcam, #ab186735), p-AMPK (1:1000, Abcam, #ab131357), ULK1 (1:1000, Abcam, #ab167139), p-ULK1 (1:1000, Abcam, #ab229909).

2.8. Histological analysis

Heart tissues were fixed in 4% paraformaldehyde and embedded in paraffin using standard procedures [33]. After deparaffinization and rehydration of the tissue sections, H&E staining was performed to assess the erythrocyte morphology. Images were acquired with a light microscope (Olympus, Japan). Immunohistochemistry was performed as previously described [11].

2.9. Immunofluorescence

For immunofluorescence analyses of colocalization, cells were fixed for 30 min at room temperature with 4% paraformaldehyde, and then were permeabilized with 0.2% Triton X-100. Subsequently, the cells were incubated in a blocking solution of phosphate-buffered saline/Tween-20 with 10% goat serum and 3% bovine serum albumin for 1 h at room temperature [34]. The samples were then incubated with the primary antibodies overnight at 4 °C, and with appropriate fluorescent secondary antibodies for 3 h at room temperature. The nuclei were stained with 4’,6-diamidino-2-phenylindole (DAPI). Images were captured with a laser-scanning confocal microscope (LSM700; Carl Zeiss Inc., Oberkochen, Germany). Primary antibodies for immunoblotting were as follows: Gr1 (1:500, Abcam, #ab25377), troponin T (1:500, Abcam, #ab8295), CD31 (1:500, Abcam, #ab222783), VE-Cadherin (1:500, Abcam, #ab33168), Tom-20 (1:500, Abcam, # ab78547). Mitophagy was determined through mt-Kemia as our previously described [22].

2.10. siRNA transfection

Confluent cells were seeded 24 h before transfection. AMPK small interfering RNAs (siRNA-AMPK) were transfected into cells using Lipofectamine RNAiMAX reagent (13778–075, Thermo Fisher) according to the manufacturer's instruction [35]. The transfected cells were incubated at 37 °C for 48 h, followed by extraction of cellular DNA, RNA and proteins. A non-targeting negative stealth siRNA (scrambled, sc-37007, Santa Cruz) was used as a negative control [36].

2.11. ROS staining, mitochondrial membrane potential measurement and mPTP opening detection

Mitochondrial and cytoplasmic ROS levels were measured in CMECs plated on six-well dishes at equal densities [37]. For ROS measurements, the cells were washed with phosphate-buffered saline and then stained with the mitochondrial ROS indicator MitoSOX™ Red (2 μmol/L, M36008; Invitrogen) and the cytoplasmic ROS probe CM-H2DCFDA (C6827; Invitrogen) at 37 °C for 15 min. The mitochondrial membrane potential was detected using the JC-1 probe according to the manufacturer's instructions (T3168; Invitrogen). Cell images were obtained using a confocal fluorescence microscope (LSM 880 with AiryScan) and ZEN BLUE imaging software. The opening of the mPTP was visualized based on the rapid dissipation of tetramethylrhodamine ethyl ester (TMRE) fluorescence, as reported previously [7].

2.12. ImageJ quantitative analysis

Confocal micrographs of CMEC confocal fluorescence images were analyzed using the ImageJ fluorescence measuring function (NIH Image, Bethesda, MD, USA). The relative fluorescence intensity was determined using randomly selected cross-sections of CMECs. The fluorescent intensities were extracted, and individual images were normalized to the average background intensity [38]. During quantitation, RGB images were converted to 32-bit grayscale images. The grayscale images were adjusted for brightness and contrast to exclude noise pixels. The threshold was also adjusted to highlight all positive pixels (green or red fluorescence) to be measured [39]. The background fluorescence intensity was measured using five different areas of the images. The staining intensity was then calculated as the integrated density minus the background intensity. The values (expressed as Area %) from each measurement were averaged and expressed as the mean ± standard error [40].

2.13. Enzyme-linked immunosorbent assay (ELISA)

The activity levels of intracellular anti-oxidative molecules were determined using commercially available ELISA kits (GSH: cat. no. EIAGSHC, Invitrogen; SOD: cat. no. EIASODC, Invitrogen; GPX: cat. no. MBS776262, MyBioSource, USA) according to the manufacturers' instructions. Caspase-9 activity (Mouse Caspase-9 ELISA kit, cat. no. MBS458138, MyBioSource) and ATP production (ATP/Adenosine Triphosphate ELISA kit, cat. no. LS-F24998, LifeSpan BioSciences, USA) were determined based on the manufacturers’ instructions. After the addition of the stop solution, the absorbance was measured at 450 nm (ELx800, BioTek Instruments) [41].

2.14. Statistics

All values are expressed as the mean ± standard error. The sample numbers are provided in the individual Figure legends and represent biological replicates. All raw data and results were interpreted in a blinded fashion. The representative image for each group was selected based on the mean value after quantitative analysis. Statistical analyses were performed with GraphPad Prism software (version 7.01). For comparisons of two groups, Student's t-test was used. For comparisons of more than two groups, one-way or two-way analysis of variance was used, with Bonferroni's or Dunnett's multiple-comparisons test for normally distributed data or the Mann-Whitney test for non-normally distributed data. Unless otherwise stated, p < 0.05 was considered statistically significant.

3. Results

3.1. Empagliflozin reduces I/R-induced microvascular damage

In this study, mice were subjected to myocardial I/R injury or a sham operation, with or without prior treatment with empagliflozin. To determine the effects of empagliflozin on cardiac microvascular damage, we used an electron microscope to examine the microvascular structure of the reperfused heart. As shown in Fig. 1A, microvascular endothelial cell inflation and luminal stenosis were observed in I/R-treated mice compared with sham-operated control mice. Empagliflozin inhibited these changes in the microvascular ultrastructure, as evidenced by the smooth endothelial cells and regular vascular walls in the drug-treated group.

Fig. 1.

Empagliflozin attenuates I/R-induced cardiac microvascular damage. Mice were assigned to the sham operation group or the myocardial I/R injury group. Empagliflozin (10 mg/kg/d) was administered seven days before myocardial I/R injury. (A) An electron microscope was used to detect changes in the microvascular structure of the reperfused heart. Yellow arrows indicate microvascular endothelial cell inflation and luminal stenosis. (B) Heart tissue was stained using H&E to visualize the morphology of erythrocytes. Black arrows indicate erythrocyte accumulation in microvessels. (C–D) Double immunofluorescence staining of Gr1⁺ neutrophils and troponin T (TnT)⁺ cardiomyocytes. DAPI was used to visualize the nucleus in the myocardium. (E–G) RNA was isolated from heart tissues, and qRT-PCR was used to determine the expression of inflammatory cytokines (TNFα, MCP1 and IL-6). Experiments were repeated at least three times and the data are shown as mean ± SEM (n = 6 mice or three independent cell isolations per group). *p < 0.05. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

Microvascular structural impairments will slow the blood flow or even induce the occlusion of the microcirculation. Reduced blood perfusion can change the local fluid mechanics from laminar to turbulent blood flow, thus promoting erythrocyte accumulation and thrombosis. Hematoxylin and eosin (H&E) staining of heart tissue revealed that erythrocytes gathered into a mass after I/R injury (Fig. 1B). Interestingly, empagliflozin treatment sustained the linear morphology of erythrocytes and prevented their convergence in microvessels, suggesting that empagliflozin reduced the risk of microthrombus formation during I/R injury. In addition to a reduced blood flow and microvascular obstruction, microvascular I/R injury is associated with microcirculatory hyperpermeability, which leads to myocardial inflammatory responses and swelling. Immunofluorescence staining revealed that Gr1⁺ neutrophil levels in the myocardium were elevated after I/R injury (Fig. 1C and D), and quantitative real-time PCR (qRT-PCR) revealed that inflammatory cytokines such as tumor necrosis factor α (TNFα), monocyte chemotactic protein 1 (MCP1) and interleukin 6 (IL-6) were upregulated (Fig. 1E–G). Empagliflozin prevented the permeation of Gr1⁺ neutrophils into the myocardium and therefore suppressed the transcription of these inflammatory cytokines. These data indicated that empagliflozin ameliorated I/R-induced microvascular ultrastructural damage by maintaining normal local fluid mechanics and preventing luminal obstruction, microvessel hyperpermeability and myocardial inflammatory responses.

3.2. Empagliflozin suppresses I/R-induced endothelial cell damage

The cardiac microcirculation is composed of a monolayer of endothelial cells, which are known as cardiac microvascular endothelial cells (CMECs). Ample studies have identified CMEC dysfunction or death as a pathological feature of cardiac microvascular I/R injury [42,43]; thus, we assessed the effects of empagliflozin on endothelial cells. CMECs synthesize and release nitric oxide via eNOS, thereby promoting vascular relaxation. We found that p-eNOS was downregulated in heart tissue following I/R injury, whereas the endothelial vasoconstrictor endothelin-1 (ET-1) was rapidly upregulated (Fig. 2A–C). Empagliflozin supplementation restored the cellular levels of p-eNOS while repressing ET-1 expression in the reperfused heart (Fig. 2A–C).

Fig. 2.

Empagliflozin alleviates I/R-induced CMEC damage.In vivo, mice were assigned to the sham operation group or the myocardial I/R injury group. Empagliflozin (10 mg/kg/d) was administered seven days before myocardial I/R injury. In vitro, CMECs were isolated from I/R- or empagliflozin-treated hearts. The cells were cultured for 24 h and then used for functional analyses. (A–C)In vivo, proteins were isolated from reperfused heart tissues, and Western blots were used to analyze eNOS and ET-1 levels. (D, E)In vivo, immunofluorescence staining of VE-cadherin was performed to observe the changes in endothelial integrity and barrier function. (F, G) Immunohistochemistry was used to assess the expression of ICAM1, an endothelium-specific adhesive factor. (H–J) Proteins were isolated from CMECs, and the levels of Src and Fak were analyzed through Western blots. (K, L)In vitro, FITC-dextran clearance and TER assays were used to analyze the endothelial barrier function and integrity of CMECs. Experiments were repeated at least three times and the data are shown as mean ± SEM (n = 6 mice or three independent cell isolations per group). *p < 0.05.

Reduced eNOS activity has been associated with endothelial dysfunction, which is characterized by impaired endothelial integrity and barrier function. Immunofluorescence analyses of endothelial integrity revealed that endothelial junctional proteins such as vascular endothelial (VE)-cadherin were slightly downregulated in the I/R injury group compared with the sham group (Fig. 2D and E). In contrast, intercellular adhesion molecule 1 (ICAM1, an adhesive factor) was upregulated in CMECs after I/R injury (Fig. 2F and G). Empagliflozin restored VE-cadherin expression and inhibited ICAM1 expression in CMECs after I/R injury. At the molecular level, VE-cadherin activity is mainly regulated by Src and focal adhesion kinase (Fak). In CMECs, Src and Fak levels were significantly reduced following I/R injury, but returned to normal following empagliflozin treatment (Fig. 2H–J).

To further assess endothelial barrier function and integrity, we isolated CMECs from sham-operated or I/R-treated mice. Then, we performed fluorescein isothiocyanate (FITC)-dextran clearance and transendothelial electrical resistance (TER) assays. As shown in Fig. 2K and L, I/R injury elevated the remaining FITC-dextran content and reduced the TER value compared with the control group, reflecting an impaired CMEC adhesion capacity. However, empagliflozin treatment prevented FITC-dextran accumulation and increased the TER value. These data suggested that empagliflozin maintained CMEC function and integrity in the setting of cardiac I/R injury.

To understand whether the improvements in microvascular structure and endothelial function are associated with increased cardiac function, echocardiography was used to analyze the alterations of myocardial performance. The data shown in Table 1 demonstrated that the reduction of ejection fraction and the dilation of left ventricular dimension in the empagliflozin group were not as remarkable as those of the I/R group. These results highlight that empagliflozin-mediated microvascular protection may improve cardiac function during I/R injury.

Table 1.

The difference of echocardiography parameter and cardiac/body weight between mice treated with empagliflozin (EMPA) after ischemia/reperfusion injury. BW, body weight; HW, heart weight; LVDd, diastolic dimension of left ventricle; LVDs, systolic dimension of left ventricle; IVS, thickness of interventricular septum; PW, thickness of posterior wall; PW, thickness of posterior wall; FS, ratio of left ventricular fractional shortening; HR, heart rate; bmp, beats/min; #p < 0.05 vs. sham, *p < 0.05 vs. I/R.

Parameter	Sham + PBS	Sham + EMPA	I/R + PBS	I/R + EMPA
BW, g	26.0 ± 1.0	26.1 ± 0.8	25.9 ± 0.7	25.9 ± 0.9
HW, mg	156.4 ± 7.8	154.2 ± 8.3	237.6 ± 1.2#	186.4 ± 9.6*
HW/BW, mg/g	6.04 ± 0.33	5.94 ± 0.58	9.27 ± 0.75#	7.15 ± 0.66*
HR, bmp	451.7 ± 12.5	462.2 ± 14.1	477.4 ± 18.1	465.3 ± 17.4
LVDd, mm	3.25 ± 0.14	3.31 ± 0.12	3.96 ± 0.16#	3.57 ± 0.13*
LVDs, mm	2.24 ± 0.09	2.20 ± 0.11	3.15 ± 0.18#	2.56 ± 0.13*
IVS, mm	0.82 ± 0.03	0.81 ± 0.02	0.70 ± 0.03#	0.78 ± 0.04*
PW, mm	0.77 ± 0.05	0.76 ± 0.04	0.79 ± 0.03	0.78 ± 0.02
FS, %	32.7 ± 1.5	33.1 ± 1.2	21.3 ± 2.3#	29.4 ± 2.1*
EF, %	61.3 ± 2.7	62.1 ± 2.4	44.6 ± 4.8#	58.2 ± 4.5*

3.3. Empagliflozin alleviates I/R-induced endothelial mitochondrial dysfunction

Mitochondrial damage has been reported to contribute to I/R-induced endothelial damage. We performed Western blot analyses, which demonstrated that I/R injury activated mitochondrial fission but inhibited mitochondrial fusion in CMECs (Fig. 3A–E). To confirm these findings, we subjected CMECs to immunofluorescence analyses, which revealed that I/R injury promoted the formation of fragmented mitochondria with shorter lengths than those of the control group (Fig. 3F–H). Due to increased mitochondrial fission, mitochondrial reactive oxygen species (ROS) levels in CMECs were elevated in the I/R injury group, which was accompanied with an increase in the levels of cytoplasmic ROS levels (Fig. 3I–K). On the other hand, anti-oxidative molecules such as glutathione (GSH), superoxide dismutase (SOD) and glutathione peroxidase (GPX) in CMECs were significantly consumed in the setting of I/R injury (Fig. 3L–N). Due to increased oxidative stress, the mitochondrial membrane potential dissipated (Fig. 3O–P), the mPTP opening rate increased (Fig. 3Q) and caspase-9 was activated in CMECs (Fig. 3R).

Fig. 3.

Empagliflozin reduces I/R-induced mitochondrial fission, oxidative stress and apoptosis in CMECs. In vivo, mice were assigned to the sham operation group or the myocardial I/R injury group. Empagliflozin (10 mg/kg/d) was administered seven days before myocardial I/R injury. In vitro, CMECs were isolated from I/R- or empagliflozin-treated hearts. The cells were cultured for 24 h and then used for functional analyses. (A–E) Proteins were isolated from CMECs, and Western blots were used to assess the expression of phosphorylated dynamin-related protein 1 (p-Drp1), Mff, mitochondrial fission 1 (Fis1), mitofusin 2 (Mfn2) and optic atrophy 1 (Opa1). (F–H) Immunofluorescence assay of the mitochondrial morphology. The average mitochondrial length was recorded to reflect mitochondrial fission. At least 100 mitochondria from 10 CMECs were used to evaluate the number of CMECs with fragmented mitochondria. (I–K) Mitochondrial and cytoplasmic ROS levels in CMECs were determined using immunofluorescence analyses. Mitochondrial ROS were assessed using MitoSOX™ Red, while cytoplasmic ROS were measured using CM-H2DCFDA. (L–N) The activity levels of intracellular anti-oxidative molecules such as GSH, SOD and GPX in CMECs were determined using commercially available ELISA kits. (O–P) JC-1 staining was used to assess the mitochondrial membrane potential in CMECs. A reduced red-to-green immunosignal of JC-1 indicates an abnormal mitochondrial membrane potential. (Q) mPTP opening in CMECs was determined based on the fluorescence intensity of TMRE. (R) An ELISA was used to assess capsase-9 levels in CMECs. Experiments were repeated at least three times and the data are shown as mean ± SEM (n = 6 mice or three independent cell isolations per group). *p < 0.05. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

Interestingly, empagliflozin treatment suppressed mitochondrial fission and enhanced mitochondrial fusion (Fig. 3F–H), thus normalizing the mitochondrial morphology in I/R-treated CMECs. Empagliflozin also inhibited mitochondrial ROS production, which accordingly reduced cytoplasmic ROS levels in CMECs following I/R injury (Fig. 3I–K). The anti-oxidative effects of empagliflozin were associated with elevated GSH, SOD and GPX levels in I/R-treated CMECs (Fig. 3L–N), possibly due to reduced mitochondrial ROS generation. Additionally, empagliflozin stabilized the mitochondrial membrane potential and reduced the mPTP opening rate in CMECs following I/R injury, ultimately reducing caspase-9 activation (Fig. 3O–R). Overall, I/R-induced CMEC damage was characterized by disordered mitochondrial dynamics, increased mitochondrial oxidative stress, a reduced mitochondrial membrane potential and enhanced mitochondrial apoptosis. However, empagliflozin attenuated these pathological alterations in CMEC mitochondria.

3.4. Empagliflozin activates FUN14 domain-containing 1 (FUNDC1)-dependent mitophagy through the AMPKα1/Unc-51-like autophagy activating kinase 1 (ULK1) pathway

Mitophagy is a protective mechanism that sustains mitochondrial function and structure [44]. Unfortunately, mitophagy is inhibited during myocardial I/R injury through a complex mechanism [11,22,45]; however, the activation of mitophagy through genetic modifications has been reported to preserve mitochondrial homeostasis and thus protect the cardiac microvasculature against I/R injury [8,11,27,46]. In view of this, we asked whether empagliflozin protected mitochondria and CMECs during I/R by inducing mitophagy. Western blot analysis of mitophagy markers demonstrated that mitochondrial levels of microtubule-associated protein light chain 3-II (mito-LC3II) were reduced in I/R-treated CMECs, whereas p62 accumulation was elevated (Fig. 4A–F). Empagliflozin treatment enhanced mito-LC3II expression and p62 degradation in CMECs following I/R injury.

Fig. 4.

Empagliflozin activates FUNDC1-dependent mitophagy through the AMPKα1/ULK1 pathway.In vivo, mice were assigned to the sham operation group or the myocardial I/R injury group. Empagliflozin (10 mg/kg/d) was administered seven days before myocardial I/R injury. In vitro, CMECs were isolated from I/R- or empagliflozin-treated hearts. The cells were cultured for 24 h and then used for functional analyses. (A–F) Proteins were isolated from reperfused heart tissues, and the levels of mitophagy-related proteins were analyzed through Western blots. (G–H) The mt-Kemia assay was used to analyze mitophagy in CMECs. (I–L) Proteins were isolated from CMECs, and the levels of p-AMPK, p-ULK1 and p-FUNDC1 were analyzed through Western blots. CC was administered to mice to inhibit empagliflozin-induced AMPK activation. Experiments were repeated at least three times and the data are shown as mean ± SEM (n = 6 mice or three independent cell isolations per group). *p < 0.05.

To confirm these findings, we performed in vitro studies in I/R-treated CMECs. The mt-Kemia assay detects acidic mitochondria as a marker of lysosome-engulfed mitochondria. The results of this assay demonstrated that I/R injury inhibited mitophagy in CMECs (Fig. 4G and H), as evidenced by reduced acidic autolysosome formation. However, Empagliflozin promoted mitochondria-lysosome interactions in I/R-treated CMECs (Fig. 4G and H). These data suggested that empagliflozin can induce mitophagy in cardiac microvascular I/R injury.

FUN14 domain-containing 1 (FUNDC1) has been reported to be a critical regulator of mitophagy in cardiac I/R injury [22,47]. The phosphorylation of FUNDC1 by ULK1 was recently revealed to activate FUNDC1-dependent mitophagy [47]. AMPKα1, a downstream effector of empagliflozin [48], is well accepted as the crucial activator of ULK1 [49]. Thus, we speculated that the AMPKα1/ULK1/FUNDC1 signaling pathway could be the mechanism whereby empagliflozin induces mitophagy in CMECs. Western blotting demonstrated that I/R injury repressed AMPK phosphorylation and ULK1 expression in CMECs, whereas empagliflozin ameliorated these effects (Fig. 4I-L). Interestingly, when we administered AMPK siRNA to prevent AMPK activation in CMECs in the presence of empagliflozin, both ULK1 and p-FUNDC1 expression were inhibited (Fig. 4I-L). These results suggested that empagliflozin activated mitophagy in I/R-treated CMECs by inducing the AMPKα1/ULK1/FUNDC1 pathway.

3.5. Ablation of FUNDC1 or AMPKα1 abolishes the protective effects of empagliflozin against I/R-induced microvascular damage

To determine whether empagliflozin attenuated cardiac microvascular I/R injury by inducing the AMPKα1/ULK1/FUNDC1 pathway in vivo, we performed loss-of-function experiments for AMPKα1 and FUNDC1. We generated endothelial-specific AMPKα1-knockout (AMPKα1^EKO) mice by crossing AMPKα1^f/f mice with Tie2^Cre transgenic mice. Similarly, we generated endothelial-specific FUNDC1-knockout (FUNDC1^EKO) mice by crossing FUNDC1^f/f mice with Tie2^Cre transgenic mice. Tie2^Cre mice were used as the control group for both AMPKα1^EKO and FUNDC1^EKO mice. Empagliflozin was administered to the mice before cardiac I/R injury.

We then used electron microscopy to observe the microvascular structure of reperfused heart tissue from each group. As shown in Fig. 5A, the cardiac microvasculature was more vulnerable to I/R injury in AMPKα1^EKO or FUNDC1^EKO mice than in Tie2^Cre mice, suggesting that AMPKα1 and FUNDC1 are necessary for microvascular protection. Although empagliflozin reduced reperfusion-induced endothelial cell swelling and luminal stenosis, these beneficial effects were negated in AMPKα1^EKO and FUNDC1^EKO mice (Fig. 5A).

Fig. 5.

Ablation of FUNDC1 or AMPKa1 prevents empagliflozin from protecting against I/R-induced microvascular damage.In vivo, AMPKα1^EKO, FUNDC1^EKO or Tie2^Cre (control) mice were assigned to the sham operation group or the myocardial I/R injury group. Empagliflozin (10 mg/kg/d) was administered seven days before myocardial I/R injury. (A) An electron microscope was used to detect changes in the microvascular structure of the reperfused heart. Yellow arrows indicate microvascular endothelial cell inflation and luminal stenosis. (B–D) Proteins were isolated from reperfused heart tissues, and the levels of eNOS and ET-1 were analyzed through Western blots. (E, F) Immunohistochemistry assay of ICAM1, an endothelium-specific adhesive factor. (G, H)In vitro, CMECs were isolated from I/R- or empagliflozin-treated hearts and cultured for 24 h. Then, FITC-dextran clearance and TER assays were used to analyze the endothelial barrier function and integrity of the CMECs. Experiments were repeated at least three times and the data are shown as mean ± SEM (n = 6 mice or three independent cell isolations per group). *p < 0.05. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

We also performed Western blotting and immunohistochemistry on heart tissues from the various groups of mice. In Tie2^Cre mice, I/R injury downregulated p-eNOS and upregulated ET-1 in heart tissue (Fig. 5B–D). Empagliflozin reversed these effects in Tie2^Cre mice, but not in AMPKα1^EKO or FUNDC1^EKO mice (Fig. 5B–D). Similarly, empagliflozin prevented the I/R-induced upregulation of ICAM1 in heart tissue from Tie2^Cre mice, but not from AMPKα1^EKO or FUNDC1^EKO mice (Fig. 5E and F).

We subsequently isolated CMECs from the mice in each group so that endothelial barrier function and integrity could be assessed in vitro. As shown in Fig. 5G and H, empagliflozin suppressed FITC-dextran accumulation and increased the TER value in CMECs isolated from I/R-treated Tie2^Cre mice, but not from I/R-treated AMPKα1^EKO or FUNDC1^EKO mice. These findings confirmed that empagliflozin protected cardiac microvessels and endothelial cells against I/R injury by inducing the AMPKα1/ULK1/FUNDC1 pathway.

3.6. Empagliflozin sustains mitochondrial homeostasis in cardiac microvascular endothelial cells by inducing the AMPKα1/ULK1/FUNDC1 pathway

To verify that empagliflozin maintained endothelial mitochondrial homeostasis by activating the AMPKa1/ULK1/FUNDC1 pathway, we assessed the mitochondrial morphology and structure in freshly isolated CMECs from Tie2^Cre mice (Tie2^Cre-CMECs), AMPKα1^EKO mice (AMPKα1^EKO-CMECs) and FUNDC1^EKO mice (FUNDC1^EKO-CMECs). Immunofluorescence staining (Fig. 6A–C) indicated that I/R injury augmented mitochondrial fission in the Tie2^Cre-CMEC group, and this effect was exacerbated in the AMPKα1^EKO-CMEC and FUNDC1^EKO-CMEC groups. Empagliflozin treatment inhibited I/R-induced pathological mitochondrial division in Tie2^Cre-CMECs, but not in AMPKα1^EKO-CMECs or FUNDC1^EKO-CMECs. Furthermore, empagliflozin attenuated mitochondrial ROS production (Fig. 6D–F) and enhanced anti-oxidative molecule levels (Fig. 6G–I) under I/R injury in Tie2^Cre-CMECs, but not in AMPKα1^EKO-CMECs or FUNDC1^EKO-CMECs. Similarly, empagliflozin reversed the I/R-induced mitochondrial membrane potential reduction (Fig. 6J and K) in CMECs, but these effects were prevented by the deletion of AMPKα1 or FUNDC1. These data indicated that the activation of the AMPKα1/ULK1/FUNDC1 pathway could be the mechanism whereby empagliflozin sustained mitochondrial homeostasis in CMECs during I/R injury.

Fig. 6.

Empagliflozin sustains mitochondrial homeostasis in CMECs through the AMPKα1/ULK1/FUNDC1 pathway.In vivo, AMPKα1^EKO, FUNDC1^EKO or Tie2^Cre (control) mice were assigned to the sham operation group or the myocardial I/R injury group. Empagliflozin (10 mg/kg/d) was administered seven days before myocardial I/R injury. In vitro, CMECs were isolated from I/R- or empagliflozin-treated hearts. The cells were cultured for 24 h and then used for functional analyses. (A–C) Immunofluorescence assays were used to assess the mitochondrial morphology in CMECs. The average mitochondrial length was recorded to reflect mitochondrial fission. At least 100 mitochondria from 10 CMECs were used to evaluate the number of CMECs with fragmented mitochondria. (D–F) Mitochondrial and cytoplasmic ROS levels in CMECs were determined using immunofluorescence assays. Mitochondrial ROS were assessed using MitoSOX™ Red, while cytoplasmic ROS were measured using CM-H2DCFDA. (G–I) The activity levels of intracellular anti-oxidative molecules such as GSH, SOD and GPX in CMECs were determined using commercially available ELISA kits. (J–K) JC-1 staining was used to determine the mitochondrial membrane potential in CMECs. A reduced red-to-green immunosignal of JC-1 indicates an abnormal mitochondrial membrane potential. Experiments were repeated at least three times and the data are shown as mean ± SEM (n = 6 mice or three independent cell isolations per group). *p < 0.05. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.

4. Discussion

Cardiac microvascular I/R injury is a neglected topic in the field of perioperative cardioprotection. The molecular mechanism of I/R-induced microvascular damage is relatively complex, so few drugs are available to protect the cardiac microcirculation during cardiac I/R injury. In the present study, we demonstrated the potential therapeutic value of the AMPKα1/ULK1/FUNDC1 mitophagy pathway for the preservation of endothelial cell homeostasis and microvessel integrity during microvascular I/R injury. We also found that empagliflozin could be considered as an endothelium-specific protective drug that increases the resistance of cardiac microvessels to I/R injury by maintaining endothelial function and structure. Moreover, we discovered that empagliflozin prevented I/R-induced microvascular dysfunction by increasing mitophagy and consequently improving mitochondrial behavior. As far as we know, this is the first study to explore the therapeutic potential of empagliflozin for acute myocardial microvascular impairment. Importantly, these effects of empagliflozin seemed to depend on mitochondrial protection via mitophagy and the AMPK pathway.

The endothelial protective effects of empagliflozin have been described in several in-depth studies. In transverse aortic constriction-induced heart failure, empagliflozin was reported to suppress endothelial apoptosis and maintain capillarization through the Akt/eNOS/nitric oxide pathway, thus increasing heart performance [18,50]. Similarly, empagliflozin was found to inhibit leukocyte-endothelium interactions during diabetes by repressing inflammatory cytokine production [51]. In human coronary artery endothelial cells, empagliflozin exhibited anti-oxidative effects by neutralizing cyclic stretch-induced mitochondrial ROS overproduction, thereby preventing endothelial hyperpermeability [19]. In addition to inhibiting endothelial dysfunction, empagliflozin has been reported to attenuate angiotensin II-induced endothelial senescence by suppressing abnormal activity of the angiotensin II receptor type 1/nicotinamide adenine dinucleotide phosphate oxidase pathway [52,53]. In a murine model of myocardial infarction, empagliflozin was found to reduce the infarct size by increasing the CD31⁺/vascular endothelial growth factor receptor 2⁺ endothelial cell ratio (a critical feature of angiogenesis) in the infarcted zone [54]. These studies suggested that empagliflozin activates specific signaling cascades to improve mitochondrial anti-oxidative activity, enhance the anti-inflammatory capacity and induce pro-survival programs in endothelial cells.

In accordance with these previous findings, our in vitro and in vivo analyses revealed that empagliflozin had pleiotropic effects on endothelial function and structure. Empagliflozin enhanced endothelial integrity, permeability, barrier function and relaxation factor production in the setting of cardiac I/R injury. On the other hand, empagliflozin significantly suppressed endothelial adhesive protein expression and inflammatory responses, thus normalizing local fluid mechanics, maintaining well-organized vessel walls and improving microvascular perfusion. Our study is the first to provide comprehensive evidence that empagliflozin is an effective drug against acute microvascular dysfunction.

Mitochondria serve as sensors of environmental stress and cellular adaptions, in addition to performing their classical functions of glucose metabolism and energy production. Mitochondrial damage contributes to endothelial dysfunction and cardiac microvascular I/R injury [[55], [56], [57]]. Myocardial reperfusion rapidly augments mitochondrial fission factor (Mff)-induced mitochondrial fission, resulting in mtDNA damage and endothelial oxidative stress [7]. Moreover, reperfusion activates mitochondrial apoptosis by inducing mitochondrial membrane hyperpermeability, mPTP opening and cytochrome c leakage [27,58]. Consistently, we found that mitochondrial fission was activated and fusion was suppressed in I/R-treated CMECs. Due to these mitochondrial morphological alterations, mitochondrial oxidative stress was induced, the mitochondrial membrane potential was reduced and apoptosis markers were upregulated.

Mitophagy is regarded as an important repair mechanism for damaged and dysfunctional mitochondria through normalization of mitochondrial fission/fusion [59]. Increased mitochondrial fission and/or decreased mitochondrial fusion promotes the formation of fragmented mitochondria which promotes cellular oxidative stress or release pro-apoptotic factors (such as cytochrome c) into the cytoplasm where the mitochondria-dependent apoptosis is induced [9,[60], [61], [62]]. Mitophagy promotes the removal of fragmented mitochondria and therefore play anti-oxidative and/or anti-apoptotic roles in endothelial cells [11,63]. However, at the stage of myocardial reperfusion, mitophagy is largely suppressed due to multiple mechanisms [11,22,45,64]. Interestingly, we found that empagliflozin restored mitophagy by inducing the AMPKa1/ULK1/FUNDC1 pathway during I/R injury, thus reducing mitochondrial damage in CMECs. Genetic ablation of FUNDC1 or AMPKα1 prevented empagliflozin from inhibiting mitochondrial oxidative stress, stabilizing the mitochondrial membrane potential and suppressing mitochondrial apoptosis in CMECs. These findings demonstrated that empagliflozin maintained endothelial mitochondrial performance primarily by preserving mitophagy under I/R injury.

It is well documented that the intracellular effectors of empagliflozin include extracellular signal-regulated kinase 1/2 [65], phosphoinositide 3-kinase/Akt [18], transforming growth factor β1 [66], nuclear factor κB [67], Janus kinase 2 [67], Sirtuin 1 [68], hypoxia inducible factor 1α [69], protein kinase G Iα [70], heme oxygenase 1 [71], peroxisome proliferator-activated receptor gamma coactivator 1α [72] and AMPKα1 [20]. Among them, AMPK has been identified as an indispensable inducer of mitophagy. Reduced energy production increases AMP levels and thus promotes the phosphorylation of AMPKα1, which directly activates ULK1 by phosphorylating it at Ser317 or Ser777 [73]. Upon its activation, ULK1 translocates to the mitochondrial outer membrane surface to phosphorylate FUNDC1 at Ser17 [47], thus inducing mitophagy. The ULK1/FUNDC1 pathway has been characterized as a protective mechanism under various pathological conditions, including renal I/R injury [47], hypoxia-related nerve cell apoptosis [74] and myocardial I/R injury [75].

In the present study, we demonstrated that empagliflozin stimulated ULK1/FUNDC1-dependent mitophagy by activating AMPKα1. This is the first evidence of a relationship between empagliflozin and the AMPKα1/ULK1/FUNDC1 mitophagy pathway. Our in vivo genetic deletion experiments and biochemical studies solidly confirmed that this pathway improved endothelial cell function and microvascular integrity. Thus, the AMPKα1/ULK1/FUNDC1 mitophagy pathway may be an endothelium-specific orchestrator of the cellular response to empagliflozin in the pathogenesis of cardiac microvascular I/R injury. However, there are several limitations in the present study. First, although we demonstrated the protective effects of empagliflozin on microvascular structure, there is no direct evidence to show the influence of empagliflozin on microvascular function in vivo. Coronary flow reserve (CFR) is an important evaluator for assessing cardiac microcirculation function in vivo and therefore it requires additional experiments to further validate the impact of empagliflozin on microvascular function. Second, although we reported the roles of empagliflozin in regulating mitochondrial fission/fusion and mitophagy in I/R-induced cardiac microvascular damage, the complex relationships between mitochondrial fusion/fission and mitophagy in the presence of empagliflozin need further investigation.

5. Conclusions

Overall, our study revealed the mechanism whereby empagliflozin inhibits reperfusion-induced cardiac microvascular damage: namely, by activating AMPKα1 and thus augmenting FUNDC1-dependent mitophagy through ULK1 phosphorylation. Increased mitophagy normalizes mitochondrial fission/fusion, reduces endothelial oxidative stress and hampers mitochondrial apoptotic signaling, thereby improving endothelial function and microvascular structure. Our discovery of the links between empagliflozin, the AMPKα1/ULK1/FUNDC1/mitophagy axis and endothelial protection will enable the exploration of new treatment targets and drugs for cardiac microvascular I/R injury. Clinical trials are urgently needed to affirm our observations for the benefit of patients with acute microvascular dysfunction.

Author's contributions

YT, ZZG, XC and ZYL contributed to the study designation. YT, FW JH and ST performed experiments and data analysis. TTC, KRW, NXS and HZ contributed to manuscript writing and editing. DM, CC, XC and YT conducted a critical revision of the manuscript. All authors approved the final manuscript.

Funding

This study is supported by the National Natural Science Foundation of China (NO.82102262, NO.82170241, NO.81900252) and Guangdong Basic and Applied Basic Research Foundation (NO. 2021A1515010977 and NO. 2020A1515110174).

Availability of data and materials

All data generated or analyzed during this study are included in this published article.

Declaration of competing interest

The authors have declared that they have no conflicts of interest.

Abbreviations

ULK1

Unc-51-like autophagy activating kinase 1

AMPKα1

adenosine monophosphate-activated protein kinase catalytic subunit alpha 1

FUNDC1

FUN14 domain-containing 1

I/R

ischemia/reperfusion

mPTP

mitochondrial permeability transition pore

mtDNA

mitochondrial DNA

AMP

adenosine monophosphate

ATP

adenosine triphosphate

TNFα

tumor necrosis factor α

MCP1

monocyte chemotactic protein 1

IL-6

interleukin 6