Bioinformatics analysis

The dataset used in this study was obtained from the GEO database (https://www.ncbi.nlm.nih.gov/geo/). The mRNA expression profiles related to atherosclerosis in mice (GSE28783) were selected. Four untreated AS mice (GSM712678, GSM712679, GSM712680, and GSM712681) were assigned to the treatment group (referred to as "treat"), while four AS mice that showed symptom relief after miR-33 treatment (GSM712670, GSM712671, GSM712672, and GSM712673) were assigned to the control group (referred to as "control"). Macrophages from the atherosclerotic plaques in the mice's aortas were used for gene expression profiling. Gene ID conversion was performed using the annotation file GPL1261 on the sequencing platform. The Limma package (version: 3.40.2, http://www.bioconductor.org/packages/release/bioc/html/limma.html) in R software was used to analyze the differential expression of mRNA. The threshold for distinguishing differentially express.

The interaction network of target genes was obtained from the STRING database (https://string-db.org) and subsequently imported into Cytoscape software (version: 3.8.2, https://cytoscape.org/). The regulatory relationship network was constructed, and Cytoscape was used to analyze the results and apply conditional filtering to the network diagram. The degree and combined score values were represented by colors and shapes in the network diagram.

Isolation and characterization of macrophage-EVs

Macrophage-EVs were isolated from the supernatant of RAW264.7 cell medium by differential centrifugation. Briefly, after RAW264.7 cells were cultured in EV-free medium containing fetal bovine serum (FBS) for 48 h, the cell supernatant was collected and successively centrifuged at 800g for 5 min to remove dead cells, 1500g for 15 min to remove cell debris, and 15,000g for 30 min to remove large EVs. The obtained supernatant was then ultra-centrifuged at 150,000g for 2 h. The isolated macrophage-EVs were identified and used for the experiments.

Identification of macrophage-EVs was carried out by transmission electronic microscope (TEM, H-7650, HITACHI, Japan). Nanoparticle Tracking Analysis (NTA) was performed to determine macrophage-EVs' size. This analysis automatically tracks and determines particle size based on Brownian motion and diffusion coefficient. The macrophage-EV was resuspended in 1 mL PBS, and the filtered PBS served as a control. The diluted macrophage-EV was then injected into the NanoSight LM10 instrument to measure the particle size at 23.75 ± 0.5 ℃ with the measurement duration as 60 s.

Identification of surface protein markers of EVs was performed through Western blot. The EVs were extracted using Triton X-100 lysis buffer containing a proteinase inhibitor cocktail (Merck, 539134). The lysis buffer consisted of 20 mM Tris–HCl (pH 7.4), 137 mM NaCl, 1% Triton X-100, 2 mM EDTA, and 10% glycerol (Xu et al. 2017). Quantitative analysis was done using the BCA Protein Assay Kit (P0028, Biyuntian, Shanghai). The following antibodies were used for EV protein immunoblotting: CD9 (ab92726, 1:2000, Rabbit, Abcam, UK), CD63 (ab217345, 1:1000, Rabbit, Abcam, UK), and Calnexin (ab133615, 1:2000, Rabbit, Abcam, UK). The experiments were repeated three times.

Preparation of macrophage-EV carrying small interfering RNA against HMGB1 (siHMGB1)

siHMGB1 was loaded in macrophage-EV using Gene Pulser X Cell Electroporation System by electroporation technique. The macrophage-EV with a total protein concentration of 20 μg was mixed with 20 μg of siHMGB1 in 400 μL PBS (pH = 7.3), which was immediately moved into the ice after 400 V electroporation. Unloaded siHMGB1 was removed by ultra-centrifugation at 100,000g for 1 h, followed by removing the supernatant. The precipitated macrophage-EV was resuspended, and the final prepared macrophage-EV loaded with siHMGB1 was called macrophage-EV/siHMGB1, with the sequence of the plasmid as siNC (5′-CAAAGACGACCAGGCGTATCGATCT-3′), siHMGB1 (5′-CAAGAAGAAGCACCCGGATGCTTCT-3′). The plasmids required for the experiment were purchased from GenePharma (Shanghai, China).

Uptake of macrophage-EV/siHMGB1

PKH67 dye was added to macrophage-EV/siHMGB1 according to the kit instructions (PKH67GL, Sigma Aldrich, St. Louis, MO, USA), which was incubated at room temperature for 15 min, and then centrifuged at 1000g for 5 min, followed by removal of the supernatant. The mixture was suspended in a macrophage-EV medium and centrifuged at 1000 for 5 min. Precipitate was obtained after repeating twice, which was PKH67-labeled macrophage-EV. Then, RAW264.7 cells were incubated in the dish pre-coated with the specific cell slides. When cell confluence reached 50%, PKH67-labeled macrophage-EV/siHMGB1 was added for incubation at 37 ℃ for 24 h. The cell slides were removed, washed three times in PBS, soaked with 4% paraformaldehyde for 30 min at room temperature, permeabilized with 2% Triton X-100 for 15 min, and then stained by 4′,6-diamidino-2-phenylindole (DAPI; 2 μg/mL, C1005, Beyotime) for 10 min. Fluorescence expression was observed by confocal microscopy.

With the similar method mentioned above, Cy3-HMGB1 was loaded to the macrophage-EV. Macrophage-EV/Cy3-siHMGB1 was then co-cultured with RAW264.7 cells for 1 h and fixed in 4% paraformaldehyde. The nuclei were stained with DAPI. Internalization of macrophage-EV by RAW264.7 cells was observed under confocal microscopy.

Uptake of Dli-ox-LDL by macrophages

Cell-specific slides were placed in cell culture dishes, in which RAW264.7 cells were seeded and incubated overnight with or without macrophage-EV/siHMGB1. Subsequently, 10 μg/mL of Dli-ox-LDL was incubated with cells at 37 ℃ for 24 h. After that, the cells were slowly washed 3 times, soaked in 4% paraformaldehyde for 30 min at room temperature, permeabilized using 2% Triton X-100 for 15 min, and stained with DAPI (2 μg/mL, C1005, Beyotime) for 10 min. The expression of the fluorescence was observed under confocal microscopy.

Establishment of AS mouse models

Eighteen ApoE−/− male mice (Laboratory Animal Resources, Chinese Academy of Sciences, Beijing, China) aged 8–10 weeks were maintained under specific pathogen-free conditions at 26–28 ℃ and 50–65% humidity. All animal experiments were approved by the Animal Ethics Committee of Panyu Central Hospital (GT-IACUC202208181), Cardiovascular Institute of Panyu District.

A total of 12 ApoE−/− mice were fed a high-fat diet (protein: 20%; fat: 40%; carbohydrate: 40%; cholesterol: 1.25%) for 12 consecutive weeks for the establishment of AS model. The remaining 6 ApoE−/− mice were fed regular food as a control. Six mice subjected to AS modeling were randomly selected and injected with macrophage-EV/siHMGB (100 μg/mouse, once every two days) for 12 weeks, and the remaining mice were injected with PBS as a control.

Cell culture and grouping

Mouse mononuclear macrophage RAW264.7 (CL-0190, Procell, Wuhan, China) were cultured in Dulbecco’s modified eagle medium (PM150210, Procell) with 100 U/mL penicillin and 100 U/mL streptomycin (PB180120, Procell) at 37 ℃, with 5% CO2. The culture medium was renewed every 3 days. When cell confluence reached 80%, cell passage was performed with 0.25% trypsin/ethylenediamine tetraacetic acid.

At 24 h post cell seeding in 6-well plates, RAW264.7 cells were treated with 20 μg/mL macrophage-EV, 20 μg/mL macrophage-EV/siHMGB, or 50 μg/mL ox-LDL for 48 h, or with 1 μg/mL LPS for 24 h. Cells treated with an equal volume of PBS were taken as the control.

Enzyme-linked immunoassay (ELISA)

Serum and cell culture supernatants were collected for determination of the contents of HMGB1, TNF-α, IL-6, IL-1β, and IL-18 using the hum-HMGB1 (SP11733, Procell), mmu-HMGB1 (SP14752, Procell), TNF-α (SP13726, Procell), IL-6 (SP13755, Procell), IL-1β (SP13701, Procell), and IL-18 (SP13710, Procell) kits. The absorbance of each well was measured within 30 min using a microplate reader at 450 nm.

Oil red o staining

Mouse aorta was taken from differently-treated mice, washed with precooled PBS, put into a small box with liquid-embedded gel, and quickly frozen in liquid nitrogen. The frozen tissue blocks were sliced to 5 μm thickness, and sections were stained by Oil red O staining. For the macrophages, cell slides were washed with precooled PBS, fixed with 4% paraformaldehyde for 30 min at room temperature, washed in 50% ethanol after drying, and treated with the Oil Red O Ethanol Dye Solution for 8 min. Then, the cell slides were differentiated by 50% ethanol, counterstained by hematoxylin, and washed with running water to return to blue color. After air-drying, the cell slides were mounted and observed under a light microscope to analyze the vascular atherosclerotic plaque. Plaque area was quantified using Image Pro Plus (IPP) software.

RNA isolation and quantification

Total RNA was extracted using Trizol (16096020, Thermo Fisher Technology, New York, NY, USA). The purity and concentration of the obtained RNA were assessed by measuring the absorbance at 260 and 280 nm. Reverse transcription was performed according to the instructions of the complementary DNA reverse transcription kit (RR047A, Takara, Japan). Polymerase chain reaction (PCR) was performed according to the instructions of LightCycler 480 SYBR Green I Master (04707516001, Roche, Germany) with glyceraldehyde-3-phosphate dehydrogenase (GAPDH) used as the internal reference. The relative mRNA expression was detected with the 2−ΔΔCt method, and the primer sequences are shown in Additional file 1: Table S1.

Western blot analysis

Total protein was extracted from tissues, cells, and EVs using RIPA lysis buffer (P0013B, Beyotime) containing PMSF, and the protein concentration was determined using a BCA kit (P0028, Beyotime). Proteins were denatured by boiling at 100 ℃ for 10 min and then saved at – 80 ℃. According to the size of the target protein band, 8–12% sodium dodecyl-sulfate electrophoresis gels were prepared, and equal volumes of protein samples were added to each lane for electrophoresis separation. The protein from the gel was transferred to a polyvinylidene fluoride membrane (1620177, BIO-RAD, Hercules, CA, USA), which was blocked by 5% bovine serum albumin for 1 h at room temperature. The membrane was incubated with primary antibodies against HMGB1 (ab79823, 1:10,000, Rabbit, Abcam), Caspase-11 (ab180673, 1:1000, Rabbit, Abcam), LOX-1 (11837-1-AP, 1:1000, Rabbit, Proteintech), CD36 (18836-1-AP; 1:1000, Rabbit, Proteintech), SRA (ab151707, 1:1000, Rabbit, Abcam), FL-GSDMD (# 39754, 1:1000, Rabbit, CST, Danvers, MA, USA), and N-GSDMD (#39754, 1:1000, Rabbit, CST) overnight at 4 ℃. The next day, the membranes were incubated with HRP-labeled goat anti-rabbit IgG (ab6721, 1:5000, Abcam) for l h at room temperature. The membrane was immersed in an enhanced chemiluminescence reaction solution (1705062, Bio-Rad) at room temperature for 1 min and then developed on an Image Quant LAS 4000C gel imager. GAPDH (A01021, 1:5000, Rabbit, Abbkine, Wuhan, China) was used as an internal reference to detect the expression level of each protein.

Immunohistochemistry (IHC)

Paraffin-embedded sections of the mouse aorta were dewaxed to water, dehydrated by gradient alcohol, and subjected to antigen retrieval. Sections were blocked with normal goat serum blocking solution (C0265, Beyotime) at room temperature for 20 min. Then, the sections were incubated with primary antibody HMGB1 (ab79823, 1:400, Rabbit, Abcam) at 4 ℃ overnight, followed by another incubation with the secondary antibody goat anti-rabbit IgG (ab6721, 1:1000, Abcam) at 37 ℃ for 20 min. Next, the sections were incubated with HRP-labeled Streptavidin (A0303, Beyotime) at 37 ℃ for 20 min and developed by DAB (P0203, Beyotime). Sections were counterstained by hematoxylin (C0107, Beyotime) for 1 min, treated with 1% ammonia to return to blue color, dehydrated, cleared by xylene and mounted. Subsequently, the sections were observed and photographed by a microscope. An experienced pathologist read the sections, and 5 high-power fields were randomly selected for analysis in each section. Each field counted 100 cells, and the percentage of positively stained cells was calculated.

IHC score = A × B (A: 0 indicates cells without positive staining, 1 indicates 10% positive cells, 2 indicates 11–50% positive cells, 3 indicates 51–80% positive cells and 4 indicates more than 80% positive cells. B: 0 indicates no staining, 1 indicates shallow staining intensity, 2 indicates medium staining intensity, and 3 indicates deep staining intensity).

Immunofluorescence (IF)

Paraffin-embedded sections of the mouse aorta were taken, dewaxed, dehydrated, and subjected to antigen retrieval. Sections were added with 0.03% Triton for 10 min and sealed with normal goat serum blocking solution (C-0005, HaoranBio, Shanghai, China) for 60 min at room temperature. For macrophages, macrophages were fixed with 4% paraformaldehyde for 30 min, washed with precooled PBS And sealed as the above. Next, primary antibodies of HMGB1 (66525-1-Ig, 1:250, Mouse, Proteintech, VA, USA), F4/80 (ab6640, 1:50, Rabbit, Abcam), GSDMD (sc-393581, 1:200, Mouse, SANTA, UC, USA), iNOS (ab210823, 1: 100, Mouse, Abcam), and CD206 (60143-1-Ig, 1:200, Mouse, Proteintech) were selected for sample incubation at 4 ℃ overnight. Then, the samples were incubated with fluorescent secondary antibody, Alexa Fluor® 647-Anti-Rabbit IgG (ab150075, 1:500, Donkey, Abcam) or Alexa Fluor® 488-Anti-Mouse IgG (ab150113, 1:500, Goat, Abcam) at room temperature for 60 min in the dark. A fluorescence microscope was utilized for observation, with fluorescence intensity recorded.

Flow cytometry

Cells were collected after different treatments and then made into single-cell suspensions. Next, cell suspension was incubated with F4/80-FITC (11-4801-82, eBioscience, San Diego, CA, USA), iNOS-PE (61-5920-82, eBioscience), and CD206-APC (17-2061-82, eBioscience). After incubation, cells were collected and analyzed using a CytoFlex flow cytometer (Beckman Coulter, Brea, CA, USA). Data were analyzed using the FlowJo software (version 7.0; FlowJo LLC, Ashland, OR, USA).

Statistical analysis

Statistical analysis of the study data was performed using SPSS 21.0 (IBM, Armonk, NY, USA). Measurement data are expressed as mean ± standard deviation. The data conformed to normal distribution and homogeneity of variance. The t-test was applied for comparing data between groups, while one-way analysis of variance was for comparing data among multiple groups, followed by Tukey's post-hoc test. A p < 0.05 indicates a statistically significant difference.