A total of 45 C57BL/6J mice (male; 6-8 weeks old; weight, 20±2 g) were obtained from the Shanghai Jihui Laboratory Animal Care Co., Ltd. All mice were housed individually and were maintained at 22°C with 40-60% humidity and a 12-h light/dark cycle starting at 6:00 am. The mice were fed with normal chow and had free access to food and water throughout the experiment. Nicorandil was purchased from Beijing Sihuan Kebao Pharmaceutical Co., Ltd. FFAs and reactive oxygen species (ROS) assay kits were obtained from Beijing Solarbio Science & Technology Co., Ltd. The anti-GAPDH antibody was obtained from Abcam, whereas the anti-KLF2 antibody was purchased from BIOSS, anti-p-eNOS antibody was purchased from Affinity Biosciences, and the anti-AMPK, anti-phosphorylated (p-)AMPK and anti-eNOS antibodies were obtained from Cell Signaling Technology, Inc. The lipid emulsion was purchased from Fresenius Kabi Huarui Pharmaceutical Co., Ltd. Anti-CD62L and anti-CD11b antibodies were purchased from BD Biosciences. Mouse interferon-γ (IFN-γ) ELISA kit (cat. no. F10660), mouse interleukin-6 (IL-6) ELISA kit (cat. no. F10830), mouse NO ELISA kit (cat. no. F11321), mouse superoxide dismutase (SOD) ELISA kit (cat. no. F11502) and mouse tumor necrosis factor-α (TNF-α) ELISA kit (cat. no. F11630) were obtained from Shanghai Xitang Biotechnology Co., Ltd. The AMPK activator, 5-aminoimidazole-4-carboxamide ribonucleotide (AICAR), was purchased from MedChemExpress. The Cell Counting Kit-8 (CCK-8) assay kit was obtained from Shanghai DODGEN Chemical Technology Co., Ltd. PA was obtained from Sigma-Aldrich; Merck KGaA. The KLF2 overexpression plasmid was obtained from Shanghai GeneChem Co., Ltd. Finally, Lipofectamine^® 3000 transfection reagent was obtained from Thermo Fisher Scientific, Inc.

The entire animal experimental protocol used in the present study was carefully examined and approved by the Animal Experiment Ethics Committee of the Second Affiliated Hospital of Naval Medical University (approval no. 2020SL037; Shanghai, China). All animal care procedures were carried out in accordance with the Animal Research: Reporting of In Vivo Experiments guidelines on animal research (28).

A total of 30 male C57BL/6J mice were randomly assigned to the control, model and nicorandil groups (n=10 mice in each group). Normal saline solution (200 µl once per day) was administered to the mice in the control and model groups via intraperitoneal (i.p.) injection for 3 days, whereas the mice in the nicorandil group were treated with an injection (i.p.) of nicorandil (200 µl nicorandil at a concentration of 1.95 mg/kg once per day for 3 days). Nicorandil, a drug for CMD, is used to stimulate the upregulation of eNOS (29) and was used in the present study to investigate the AMPK/KLF2/eNOS signaling pathway. As shown in Fig. 1A, the mice were subsequently anaesthetized with sodium pentobarbital (40 mg/kg), and the jugular vein was catheterized using a microcatheter. Afterwards, the mice underwent a 6-h infusion with either normal saline or lipid emulsion combined with heparin using a micro-pump (the lipid component was 20% Intralipid^®; Fresenius Kabi Huarui Pharmaceutical Co., Ltd.; heparin was infused at a concentration 20 U/ml, and at a rate of 0.1 ml/h).

Following lipid infusion, the mice were immediately anaesthetized using isoflurane (3% for induction and 1-2.5% for maintenance). Hyperemic flow was achieved by increasing the percentage of isoflurane to 2.5%, whereas the baseline flow was achieved by reducing the isoflurane percentage to 1% (30). Data were measured using a Doppler signal processing workstation. The CFR was calculated according to the following formula: CFR=P/R=V_high/V_low.(where P is the peak blood flow, R is the resting blood flow, V_high is the hyperemic blood flow velocity and V_low is the baseline blood flow velocity).

Mice were anesthetized with sodium pentobarbital (40 mg/kg) immediately after the lipid infusion. Intravital microscopy (Olympus BX51 WI upright microscope; Olympus Corporation) was used to record the cremaster microvascular blood flow and leukocyte adhesion in venules. Leukocyte flux and leukocyte endothelium interactions (rolling and adhesion) analysis were conducted as described previously (31). Briefly, rolling leukocyte flux was measured at the indicated time points by counting the number of rolling leukocytes per 20 sec that passed a reference point in the microvessel and the number was then presented as units of cells/min. Leukocyte rolling velocity was calculated by the velocity of 10 leukocytes rolling along the endothelial cell lining (mm/sec) and leukocyte adhesion (defined as being stationary for 20 sec) was counted in 100-µm-long vascular segments, and presented as the number of adherent cells/mm². Diameters (in mm) were measured perpendicularly to the vessel path. Finally, the venular wall shear rate was calculated based on the Newtonian definition, according to the following formula: Wall shear rate (sec⁻¹)=8 × [red blood cell velocity (µm·sec⁻¹)/venular diameter (µm)] (32).

A total of 15 male mice were randomly assigned to control (n=5), model (n=5) and AICAR (n=5) groups. Normal saline solution (200 µl once per day) was administered to the mice in the control and model groups via i.p. injection for 3 days. Mice in the AICAR group were treated with 200 µl AICAR (concentration, 0.5 mg/g) by i.p. injection for 3 days prior to lipid infusion (once per day for 3 days). Subsequently, microvascular dysfunction was induced in the model and AICAR groups by lipid infusion as aforementioned. Cultured CMECs in vitro were treated with AICAR (500 µM) for 3 h at 37°C, then 400 µM PA was added and cells were incubated at 37°C for 24 h.

Mice were sacrificed by cervical dislocation while under anesthesia induced by an i.p. injection of 40 mg/kg pentobarbital, and respiratory arrest was used to confirm animal death. Fresh heart sections were fixed in 4% paraformaldehyde at room temperature for 24 h, embedded in paraffin and serial 5-µm-thick sections were prepared. Morphological analysis of cardiomyocytes and microvessels was subsequently performed using H&E staining. Then, the sections were stained with hematoxylin for 5 min at room temperature, then washed with running water for 5 min and differentiated in 1% acid alcohol for 10 sec at room temperature. The sections were then stained with 1% eosin for 1 min, washed with running water and dehydrated in increasing concentrations of alcohol (70, 80, 90 and 100%) and cleared in xylene. Finally, H&E staining was observed under a BX43 light microscope (Olympus Corporation).

For the immunohistochemical analysis of KLF2, eNOS and p-eNOS, fresh heart sections were fixed in 4% paraformaldehyde at room temperature for 24 h, embedded in paraffin and serial 5-µm-thick sections were prepared. The sections were dewaxed in xylene, then rehydrated in a descending alcohol series (hydrated in 100, 95 and 80% ethanol and water for 5 min each). Antigen retrieval was performed by pre-treatment of the slides in citrate buffer (pH 6.0) (cat. no. G1202; Wuhan Servicebio Technology Co., Ltd.) in a microwave (720 W heating) oven for 12 min. Thereafter, the slides were cooled to room temperature in deionized water for 5 min. The sections were incubated in 0.3% H₂O₂-methanol solution for 10 min and then washed with PBS three times. To block endogenous peroxidase activity, 50 µl peroxidase blocking solution was added to each section and these were incubated for 10 min at room temperature. Each section was then incubated with 50 µl 10% nonimmune goat serum (OriGene Technologies, Inc.) for 10 min at room temperature, rinsed with PBS and incubated with 50 µl primary antibody [KLF2 (dilution, 1:400; cat. no. bs-2772R; BIOSS), eNOS (dilution, 1:1,000; cat. no. GB12086; Wuhan Servicebio Technology Co., Ltd.) and p-eNOS (diluted 1:100; cat. no. AF3247; Affinity Biosciences)] at 4°C overnight. The sections were then incubated with 50 µl secondary antibody [HRP conjugated goat anti-rabbit IgG (dilution, 1:200; cat. no. GB23303; Wuhan Servicebio Technology Co., Ltd.) or HRP-conjugated goat anti-mouse IgG (diluted 1:200; cat. no. GB23301; Wuhan Servicebio Technology Co., Ltd.)] at room temperature for 10 min. The reaction was developed with freshly prepared 3,3′-diaminobenzidine, and observed under a Leica DM4000 microscope (Leica Microsystems GmbH) for 3-10 min, with brown indicating positive staining. The slides were then counterstained with hematoxylin for 3 min at room temperature, washed with running water, and dehydrated in increasing concentrations of alcohol (75, 85 and 100%) and cleared in xylene. The pathological and immunohistochemical changes were evaluated and photographed under a BX43 light microscope (Olympus Corporation). Data analysis was performed using Multiplex IHC v2.2.0 module analysis software (Indica Labs, Inc.).

Heart samples were fixed in 2.5% glutaraldehyde in 0.1 mol/l cacodylate buffer at 4°C overnight, fixed in 1% osmium tetroxide at 4°C for 2 h and then embedded in Epon using an Embed-812 kit (Electron Microscopy Sciences) at room temperature overnight. Ultrathin sections (80-100 nm) were obtained from selected areas using an ultrathin microtome (Leica EM UC7; Leica Microsystems GmbH). The ultrathin sections of the cubes were stained with 3% uranyl acetate and lead citrate at 25°C for 30 min. Then, the sections were examined using transmission electron microscopy (TEM; Tecnai™ G2 20; FEI; Thermo Fisher Scientific, Inc.).

Blood samples (100 µl) were collected from the tail-veins of the mice in heparin-containing tubes, and plasma was prepared after centrifugation at 4,500 × g for 10 min at room temperature. Subsequently, aliquots were stored at -80°C until use. The levels of plasma biomarkers, including IFN-γ, IL-6, NO, SOD and TNF-α, were determined using ELISA kits (Shanghai Xitang Biotechnology Co., Ltd.). FFAs were measured using an FFA Content Assay Kit (cat. no. BC0595; Beijing Solarbio Science & Technology Co., Ltd.) was conducted according to the manufacturer's user guide. The preparation of reagent1 was accomplished following the ratio of n-heptane: absolute methanol: chloroform=24:1:25 at room temperature. The physical state of diphenylcarbazide was powder, absolute ethanol was added at the time of application and the liquid volume was 13 ml. The standard solution was chloroform-dissolved 5 µmol/ml palmitic acid solution. The microplate reader was preheated for 30 min, the wavelength was adjusted to 550 nm, and absolute ethanol was used as a blank (set to 0). Copper reagent was water-bathed at 37°C for 30 min. Standard solution was diluted to 0.05, 0.1, 0.2, 0.4, 0.6, 0.8 and 1 µmol/ml with chloroform. Each 1.5-ml Eppendorf tube was first supplemented with 10 µl distilled water, sample, chloroform or standard solution. Then 100 µl reagent1 and 40 µl copper reagent were added sequentially, followed by shaking for 10 min at room temperature and finally centrifugation at 850 g for 10 min at room temperature. A total of 50 µl of the supernatant was transferred to a 1.5-ml Eppendorf tube and 200 µl diphenylcarbazide was added, followed by shaking for 2 min and standing for 15 min at room temperature. Absorbance at 550 nm of each well of a 96-well plate was measured using a microplate reader (200 µl liquid volume was added per well). FFA concentrations were calculated for each plasma sample according to the standard curve.

The intracellular levels of ROS were measured by a fluorometric assay using the probe 2′-7′-dichlorodihydrofluorescin diacetate (DCFH-DA). Cells were resuspended in 2 ml PBS (10⁶ cells/ml) and were subsequently incubated with 10 µM DCFH-DA at 37°C for 20 min with continuous shaking. Cells were washed with PBS three times (2 min each at room temperature) and were transferred from a tube to a microplate prior to fluorescence being measured. To minimize the process of photo-oxidation, samples were stored in the dark. Finally, DCF fluorescence was measured using a microplate reader (488 nm excitation and 525 nm emission wavelengths).

Blood (100 µl) was drawn from the great saphenous vein of the mice after lipid infusion. Ammonium chloride-potassium lysis buffer was used for erythrocyte removal. Blood samples were centrifuged at room temperature for 10 min at 850 × g, and the plasma was subsequently removed. Leukocytes were resuspended in 20 µl normal saline and then incubated with the appropriate antibodies as previously described (33). Phycoerythrin rat anti-mouse CD11b (cat. no. 557397) and allophycocyanin rat CD62L (cat. no. 5333152) antibodies (BD Biosciences) were diluted at a ratio of 1:5. Flow cytometry was performed using a Cytek DXP 8 Color upgraded BD FACS Calibur™ flow cytometer (BD Biosciences) by staining with appropriate anti-bodies, and the data were analyzed using FlowJo 10.0 software (Tree Star, Inc.).

Heart tissue (50 mg) and CMECs were collected for quantitative analysis of the proteins. Proteins were extracted using RIPA lysis buffer [Hangzhou Multi Sciences (Lianke) Biotech Co., Ltd.]. Protein concentrations of heart tissues and CMECs were subsequently determined using a Takara Bradford Protein Assay kit (Takara Bio, Inc.). The proteins were diluted to a final protein concentration of 2.5 µg/µl in 1X loading buffer (Wuhan Servicebio Technology Co., Ltd.), and 25 µg protein was loaded onto each lane of the western blot gel (10%). After the proteins were separated, they were transferred to polyvinylidene fluoride membranes (Bio-Rad Laboratories, Inc.), which were blocked with 5% BSA (Shanghai Siding Biotechnology Co., Ltd.) for 2 h at room temperature. The membrane was incubated with the corresponding primary antibody overnight at 4°C. The following day, HRP-labeled secondary antibodies (cat. no. 98164; Cell Signaling Technology, Inc.) were incubated with the membranes at room temperature for 2 h. The dilution of rabbit anti-KLF2 polyclonal antibody (cat. no. bs-2772R; BIOSS) was 1:300, and the dilution of rabbit anti GAPDH monoclonal antibody (cat. no. ab181602; Abcam) was 1:10,000. The dilution of rabbit anti-eNOS (cat. no. 32027; Cell Signaling Technology, Inc.), anti-AMPK (cat. no. 5832; Cell Signaling Technology, Inc.), anti-p-eNOS (Ser1177) (cat. no. AF3247; Affinity Biosciences) and anti-p-AMPK (Thr172) (cat. no. 50081; Cell Signaling Technology, Inc.) polyclonal antibody was 1:1,000. HRP-labeled secondary antibodies were used at a dilution of 1:5,000. All antibodies were diluted using TBS with 0.1% Tween-20 (Wuhan Servicebio Technology Co., Ltd.). Enhanced chemiluminescence western blotting substrate (Wuhan Servicebio Technology Co., Ltd.) was used to detect the immuno-reactive signals. The protein gray values were analyzed using ImageJ 1.6 software (National Institutes of Health).

Primary mice CMECs were purchased from Saibaikang (Shanghai) Biotechnology Co., Ltd. and cultured in Complete™ Endothelial Cell Medium (ECM; Sigma-Aldrich; Merck KGaA) containing 10% FBS (ScienCell Research Laboratories. Inc.), 1% endothelial cell growth supplement and 1% penicillin/streptomycin at 37°C in an atmosphere of 5% CO₂. CMECs were identified by immunofluorescence using an antibody against von Willebrand factor VIII, which is constitutively expressed on the surface of CMECs (34) (Fig. S1). In order to identify the CMECs, cells were seeded in 24-well plates (2.0×10⁴ cells per well). The cells were then washed twice with PBS and fixed in ice-cold 100% methanol for 15 min at -20°C. After rinsing with PBS three times for 5 min each, the samples were incubated with 0.1% Triton X-100 at room temperature for 10 min, blocked with 1% BSA at room temperature for 60 min, and incubated with rabbit anti-vWF polyclonal antibody (1:200; cat. no. 11778-1-AP; Proteintech Group, Inc.) at 4°C overnight. The samples were then washed three times with PBS, followed by incubation with Alexa Fluor 488-conjugated secondary antibodies (1:500; cat. no. SA00006-2; Proteintech Group, Inc.) at room temperature for 120 min after 4′,6-diamidino-2-phenylindole counterstaining at room temperature for 5 min. Images were captured with a fluorescence microscope system (DS-Ri2; Nikon Corporation). The Image-Pro Plus 6.0 image analysis software (Media Cybernetics, Inc.) was used for quantification analysis.

PA was dissolved in sodium hydroxide solution. Then, 10% BSA was added to the solution in order to form a PA and BSA complex. The solution was filter-sterilized, and then diluted with 1% BSA to prepare different concentrations of PA solutions, which were adjusted to pH 7.4. CMECs were seeded in 96-well plates (5,000 cells per well) and subjected to treatment with different concentrations (0, 100, 200, 400 and 800 µM) of PA with or without nicorandil (100 µM) at 37°C for 24 h. Subsequently, 10 µl CCK-8 solution was added to the cell culture and the cells were incubated for 2 h at 37°C in a humidifier (containing 5% CO₂ and 95% O₂). Finally, cell viability was determined at an optical density of 450 nm using a microplate reader.

KLF2 overexpression and empty control pCMV plasmids were purchased from Shanghai GeneChem Co., Ltd. CMECs were seeded in 6-well plates (1.0×10⁵ cells per well) for 24 h prior to transfection. Cells were divided into the control group, control overexpression group and KLF2 overexpression group. Transient transfection was performed using Invitrogen Lipofectamine^® 3000 (Thermo Fisher Scientific, Inc.) according to the standard protocol of the manufacturer. P3000 (Thermo Fisher Scientific, Inc.) combined with the overexpression-DNA (2.5 µg) or the control overexpression-DNA vectors (2.5 µg) was added to serum-free medium and incubated with the cells at 25°C for 10 min, while the control group was treated without vectors. Subsequently, Lipofectamine 3000 was added to each group and cultured with the cells in serum-free ECM. Following 6 h of culture (37°C; 5% CO₂), the medium was replaced with ECM containing 10% FBS. After 48 h, subsequent experiments were performed.

Small interfering RNAs targeting KLF2 (siKLF2) and the corresponding non-specific control (NC) siRNA were obtained from Sigma-Aldrich; Merck KGaA. Cells were divided into the control group, NC group and siKLF2 group. The control group was treated without siRNAs. The transfection procedures were performed according to the protocol described by Wang et al (35). siRNA (100 nM) was used for cell transfection using Lipofectamine^® 3000 (Invitrogen; Thermo Fisher Scientific, Inc.). Cells were incubated at 37°C with 5% CO₂. After a 6-h antibiotic-free medium incubation, the transfection medium was removed, and the cells were incubated in fresh medium for 24 h. Then, subsequent experiments were performed. All siRNA sequences are shown in Table SI. The transfection of KLF2 overexpression vector and siKLF2 was successful and KLF2 expression was increased and decreased, respectively (Fig. S2).

All experiments were performed at least three times, and all data are presented as the mean ± standard deviation. One-way analysis of variance was used to compare the differences among groups, and Tukey's post hoc test was used for multiple comparisons. All the statistical analyses were performed using IBM SPSS 25.0 software (IBM Corp.). P<0.05 (two-sided) was considered to indicate a statistically significant difference.

The CMD mouse model was successfully established by lipid infusion, which was characterized by a reduction of the CFR (model group vs. control group, 1.89±0.37 vs. 2.74±0.30; Fig. 1E). Serum tests revealed that the levels of FFAs in model mice were significantly increased, whereas the concentration of NO was decreased compared with those in control mice (Fig. 1B and C). Leukocyte activation in the cremaster microvascular wall was subsequently investigated (Fig. 1D). Lipid infusion led to a significantly increased level of leukocyte adhesion to the cremaster microvascular wall (model group vs. control group, 4,350±1,057.5 vs. 11.8±5.4 cells/mm²) and a decreased leukocyte rolling velocity. However, the microvascular blood flow velocity, vascular diameter, leukocyte rolling flux and shear stress were found not to be impacted by lipid infusion.

To confirm whether lipid infusion caused structural changes in the mouse heart, the cardiac tissue was examined using H&E staining and TEM. No obvious structural changes were observed under the light microscope (Fig. 1F). However, TEM revealed severely swollen endothelial cells in the model group (Fig. 1G), demonstrating the damage of coronary microvascular ultrastructure.

To further investigate the effects of FFAs on CMECs, different concentrations of PA (0, 100, 200, 400 and 800 µM) were used to treat cultured CMECs in vitro, which revealed that 400 or 800 µM PA suppressed cell viability and increased ROS production in a dose-dependent manner (Fig. S3). Treatment with PA at a concentration of 400 µM for 24 h led to a reduction in cell viability of nearly 40% (Fig. S3A). Collectively, these results suggested that FFAs could induce CMD by damaging CMECs.

Nicorandil was found to alleviate CMD induced by lipid infusion, as demonstrated by an increase in the CFR and leukocyte rolling velocity, a decrease in the level of leukocyte adhesion and the improved endothelial damage (Fig. 1), the underlying mechanism of which was hypothesized to be associated with the upregulation of NO.

Subsequently, the effects of lipid infusion on the inflammatory response were examined. Expression levels of the leukocyte adhesion molecule CD11b (Fig. 2A and B) and intracellular ROS levels (Fig. 2D) were found to be increased in the model mice, while the expression levels of CD62L on leukocytes were not altered (Fig. 2A and C). Subsequently, the levels of serum inflammatory markers were examined using ELISA. Higher serum levels of TNF-α and IL-6 were identified in the model group (Fig. 2F and G), whereas the serum SOD and IFN-γ levels were not significantly different between the model and control groups (Fig. 2E and H). Experiments using the AMPK activator, AICAR, also revealed anti-inflammatory effects with resultant reduced serum levels of TNF-α and IL-6 identified in AICAR-treated model mice (Fig. S4).

To assess the underlying mechanistic basis whereby FFAs could induce CMD, the AMPK/KLF2/eNOS pathway was subsequently analyzed via immunohistochemical and western blot analyses. Activation of AMPK and eNOS was assessed by detecting the levels of AMPK phosphorylation at Thr172 and eNOS phosphorylation at Ser117. As shown in Fig. 3A and B, lipid infusion led to downregulation of the expression level of KLF-2 and a decrease of the p-AMPK/AMPK and p-eNOS/eNOS ratios in the myocardium of model mice. Treatment of CMECs with 200, 400 and 800 µM PA in vitro caused significant suppression of the expression level of KLF2 and a decrease of the p-AMPK/AMPK and p-eNOS/eNOS ratios in a dose-dependent manner (Fig. 3C). Nicorandil increased the ratio of p-eNOS/eNOS and did not influence the expression of KLF2 or the ratio of p-AMPK/AMPK both in vivo (Fig. 3A and B) and in vitro (Fig. 3C). Nicorandil also improved CMEC viability and oxidative stress induced by PA (Fig. S5).

To further demonstrate that inhibition of the AMPK/KLF2/eNOS signaling pathway could account for CMD induced by FFAs, AICAR was utilized to confirm whether activating AMPK could improve the condition of CMD in model mice. As shown in Fig. 4G and H, AICAR intervention in vivo did lead to an improvement in CFR and a decrease in leukocyte adhesion compared with those in model mice. In vitro, AICAR was found to enhance the viability of the CMECs and to inhibit oxidative stress (Fig. 4B and C). AICAR also led to an increase in the expression of both AMPK and the downstream proteins, KLF2 and eNOS, in CMECs (Fig. 4A). Subsequently, KLF2 overexpression plasmid was used to evaluate the effect of KLF2 on the phosphorylation of eNOS. As shown in Fig. S2A, KLF2 was found to be overexpressed. KLF2 overexpression, in itself, led to significant upregulation of eNOS in CMECs treated with PA, although it exerted no significant effects on AMPK (Fig. 4D). Either upregulation of KLF2 with the overexpression plasmid or increasing the concentration of NO with nicorandil led to both an improvement in the viability of CMECs and inhibition of oxidative stress (Figs. 4E and F, and S5). However, when siKLF2 was used to suppress KLF2 expression in CMECs, AICAR did increase the ratio of p-AMPK/AMPK but did not increase the expression of KLF2 or the ratio of p-eNOS/eNOS (Fig. S6). Taken together, the underlying mechanism through which FFAs induce CMD may involve disruption of the AMPK/KLF2/eNOS signaling pathway.