The Oil Red O staining kit (#MAK194), hematoxylin (#MHS32), eosin solution (#HT110116), sivelestat (#S7198), and lipopolysaccharides (#L2630) were purchased from Sigma (Sigma Chemicals, St. Louis). Anti-Ly6G antibody (#561105) was purchased from BD (BD Biosciences). Anti-moma-2 antibody (#ab33451) was purchased from Abcam (Cambridge, United Kingdom). Anti-ZO-1 (#5406), anti-Tubulin (#5568), anti-phosphorylated IκB (#2859), and anti-IκB (#9242) antibodies were purchased from Cell Signaling Technology, Inc. (Danvers, MA). DX-4000-FITC (#46944) was purchased from Sigma (Sigma chemicals, St. Louis), and the endotoxin kit (#88282) was from Pierce (Thermo Fisher, CA). Intestinal protein levels of TNF-α (#BMS607-3), IL-1β (#BMS6002), and MCP-1 (#BMS6005) were measured by using ELISA kits from Invitrogen (Thermo Fisher, CA). The intestinal neutrophil elastase activity was measured with a commercial kit (# ab204730) purchased from Abcam (Cambridge, United Kingdom).

All mouse experimental procedures were approved by the Animal Research and Teaching Committee at Southern Medical University (Guangzhou, China). The mice were housed in 21 ± 2 ^oC condition and were given free access to diet and tap water. In detail, male aged 6-week ApoE-deficient mice (ApoE^-/-) were purchased from the Shanghai Model Organisms Center (Shanghai, China), and every group included six mice. Mice were fed with high-fat high-cholesterol diet (HFHC, 40% fat and 1.25% cholesterol, Cat#D12108C, research diets). After being fed with HFHC for 4-week, the ApoE^-/- mice were administered with sivelestat (50 mg/kg per day, intraperitoneal injection) or phosphate-buffered saline (Veh) for another 8 weeks. For lipopolysaccharide stimulation, the ApoE^-/- mice were fed with HFHC for 4 weeks and then treated with sivelestat (50 mg/kg per day, intraperitoneal injection) and LPS (25 μg/day, subcutaneous injection) or Veh for 8 weeks. Then, mouse tissues, including the aorta, small intestine, and serum, were collected for further measurement.

For the analysis, 500 mg/kg FITC-labeled dextran was orally gavaged into mice, and the serum samples were collected for further analysis. The serum concentration of DX-4000-FITC was measured by using a fluorescence spectrophotometer (Synergy H1, BioTek, VT) with an excitation wavelength of 485 nm and an emission wavelength of 535 nm.

For lipid staining of the en face aorta and aorta root, whole aorta or 10-µm frozen sections of the aorta root were stained with the Oil Red O staining kit. For hematoxylin and eosin staining, the aorta and colon were fixed in 4% paraformaldehyde and embedded in paraffin. Then, 5-µm paraffin sections were dehydrated and stained with hematoxylin and eosin solution. For immunofluorescence staining, 5-µm aorta sections were hydrated, blocked with 3% BSA solution, and incubated with the anti-moma-2 antibody at 1:100 dilution. After washing in PBS, the sections were stained with the fluorescence-labeled secondary antibody at 1:500 dilution. The nuclei were stained with DAPI. The images were viewed and captured by using a fluorescence microscope (Nikon, Tokyo), and relative lipid contents were measured in 50 images/per mouse by ImageJ software.

Intestinal resident cells were isolated from 5 to 8 cm of the intestine. Briefly, intestinal segments were excised, cleaned, and cut into small pieces. Samples were then washed with sterile bovine serum-free RPMI 1640 medium three times and digested in 10 ml RPMI containing 1 mg/ml collagenase A (Roche #10103578001) for 30 min at 37°C and 200 rpm in a shaking incubator. Digestion was quenched with FACS buffer and centrifuged for 5 min at 800 g. After washing with PBS t three times, the samples were passed through a 100-μm cell strainer to obtain a single cell suspension. For neutrophil staining, the cells were stained with anti-Cd11b and anti-Ly6G antibodies at 1:100 dilution. The cell percentage was analyzed by using a flow cytometer (BD FACSAria™ III).

The rat intestinal epithelial cell line (#IEC-18, ATCC) was gifted from Dr. Ning Zhu (Zhejiang University). Cells were cultured in DMEM containing 10% FBS and 1% penicillin/streptomycin antibiotics in a humidified chamber (37°C, 21% O₂, and 5% CO₂). For cell experiments, 1 × 10 6 cells were pretreated with sivelestat (100 μM) or DMSO and stimulated with 0.1 μg/ml human neutrophil elastase recombinant protein for 24 h.

Total RNA from the whole aorta or intestine tissues was extracted by using TRIzol (Invitrogen), according to the manufacturer’s instructions. Complementary DNA was reverse-transcribed using the PrimeScript™ RT reagent kit with gDNA Eraser (Takara Biotech, Dalian, China), and qPCR was performed with the SYBR Green quantitative kit (Applied Biosystems, CA). The primer sequences used in this study were listed as follows: mouse Zo-1: F-5′-CAACATACAGTGACGCTTCACA-3’; R-5′- CAC​TAT​TGA​CGT​TTC​CCC​ACT​C-3′, mouse Tnf-α: F-5′-ACGGCATGGATCTCAAAGAC-3’ R-5′-AGATAGCAAATCGGCTGACG-3′, mouse Il-1β: F-5′-CTGGTGTGTGACGTTCCCATTA-3’; R-5′-CCGACAGCACGAGGCTTT-3′, mouse Il-6: F-5′-CCACGGCCTTCCCTAC-3’; R-5′- AAG​TGC​ATC​ATC​GTT​GT-3′, mouse Mcp-1: F-5′-CCACTCACCTGCTGCTACTCA-3’; R-5′-TGGTGATCCTCTTGTAGCTCTCC-3′, mouse Vcam-1: F-5′-CCGGCATATACGAGTGTGAA-3’; R-5′-TAGAGTGCAAGGAGTTCGGG-3′, and mouse Gapdh: F-5′-AGGAGCGAGACCCCACTAAC-3’; R-5′-GATGACCCTTTTGGCTCCAC-3’. The relative gene expression was calculated by normalizing to the Gapdh level.

Protein lysates were extracted from the intestinal tissues or epithelial cells, and the protein concentration was measured by using a BCA assay kit (Thermo Fisher, CA). 50 μg protein was subjected to 10% SDS–PAGE electrophoresis and electro-transferred to polyvinylidene difluoride membranes (Amersham Biosciences). Then, the membranes were blocked with 10% non-fat milk and incubated with anti-ZO-1, anti-phos-IκB, IκB, or anti-Tubulin antibody at 1:1,000 dilution and relative secondary antibodies at 1:5,000 dilution. The relative protein expression was visualized by using enhanced chemiluminescence reagents (Bio-Rad, CA) and quantitatively analyzed by ImageJ software.

This study included 26 individuals, including nine atherosclerotic subjects with carotid intima-medial thickness (IMT) ≥ 0.85 mm and 17 healthy subjects with IMT <0.85 mm, who were recruited from July 2019 to May 2020 at the Shenzhen Hospital of Southern Medical University (NYSZYYEC20190005). The basic clinical parameters of these subjects are showed in Supplementary Table S1. Furthermore, among these subjects, 12 patients were under intestinal polyp surgery (3 patients with ≥0.85 mm and nine subjects with IMT <0.85 mm). The subjects with IMT <0.85 mm were characterized with no history of angina and other heart diseases, a normal resting ECG, and normal exercise ECG stress testing. In patients under the intestinal polyp surgery, the intestinal activity was measured by using the commercial kit. Before statistical analysis, the Shapiro–Wilk test was conducted to identify the distribution for normality. Then, Pearson analysis was conducted for clinical correlation. All participants have been informed clinical consent, and the related analysis protocol was approved by the Human Ethics Committee of Southern Medical University.

Data were shown as mean ± SEM. The Student’s t-test was used for comparing two groups, and ANOVA was used for multiple groups (GraphPad, San Diego, CA). Pearson analysis was used for analyzing the statistical correlation. p < 0.05 was considered to be significant.

Multiple studies have addressed high-fat diet severely injured intestinal homeostasis, such as intestinal shortening, disruption of the intestinal barrier, and inflammatory response (Araújo et al., 2017; Rohr et al., 2020). To identify the pathological changes of the intestine in atherosclerotic mice, at first, we measured the tissue remodeling of colon. As shown in Figure 1A, the length of the colon was significantly decreased after the mice were fed with high-fat high-cholesterol (HFHC) diet for 4 and 12 weeks, and HFHC feeding also induced colic structural disorders. For an inflammatory response, flow cytometry analysis indicated Cd11b⁺Ly6G⁺ neutrophils obviously infiltrated into the intestine (Figure 1B), especially in the HFHC-fed mice for 12 weeks (p < 0.01). Furthermore, neutrophil elastase (NE), one of the neutrophil serine proteases, was also highly expressed in the intestines of HFHC-fed mice (Figure 1C), and the intestinal NE activity was increased time-dependently during HFHC feeding (Figure 1D). Meanwhile, as shown in Figures 1E,F, mice with HFHC feeding had remarkable lipid plaque formation in the aorta. To testify the possible role of intestinal NE in atherosclerotic development, Pearson analysis showed the intestinal NE levels had a positive correlation of the aorta lipid deposit (Figure 1G).

Neutrophils play critical roles in the inflammatory response (Kolaczkowska and Kubes, 2013) and participated in the progress of cardiovascular diseases (Gaul et al., 2017). Sivelestat, as a selective inhibitor of neutrophil elastase, could suppress inflammation in several inflammatory diseases (Endo et al., 2006; Shimoda et al., 2008). In the present study, we aimed to explore the pharmacological effects of sivelestat on atherosclerosis in mice. Consistent with previous studies (Jin Li et al., 2016; Zhu et al., 2018), high-fat high-cholesterol (HFHC) diet accelerated lipid accumulation in the aorta from Apo E^-/- mice (Figures 2A–D). But the administration of sivelestat significantly decreased lipid contents in en-face aorta (Figures 2A,B, p < 0.01) and the sections of the aorta root (Figures 2C,D, p < 0.001). Infiltration of immune cells, including macrophages and other monocytes, was a key phenotype of atherosclerosis (Hansson, 2005). As shown in Figures 2E,F, treatment of sivelestat inhibited the infiltration of monocytes and macrophages (p < 0.01). Meanwhile, sivelestat also decreased the gene levels of inflammatory factors, including pro-inflammatory cytokines (Figure 2G, p < 0.01) and chemokines (Figure 2H, p < 0.001). However, sivelestat had no significant effects on the other basic parameters of a mouse, including body weight, the fasting blood glucose level, and serum lipid profiles (Supplementary Table S2).

The abnormal induction of intestinal permeability and subsequent endotoxemia could initiate the progress of atherosclerotic plaque formation (Zhu et al., 2018). In the present study, we found the treatment of sivelestat led to a significant reduction in the intestinal NE activity (Figure 3A) and the circulating level of endotoxin in the HFHC-fed ApoE^-/- mice (Figure 3B). Next, we measured intestinal permeability by treating mice with fluorescent-labeled dextran (DX-4000-FITC). As shown in Figure 3C, the circulating concentration of DX-4000-FITC was decreased in sivelestat-treated Apo E^-/- mice (p < 0.01). H&E staining of the colon also showed the HFHC diet loosened the structure of mucosa, which facilitated the leakage of toxic substances into circulation, whereas sivelestat improved the intestinal structure (Figure 3D). Intestinal permeability was mainly controlled by zonula occludens (ZO)-1, one of the key epithelial tight junction proteins (Van Itallie et al., 2009). As shown in Figure 3E, the intestinal gene level of ZO-1 was upregulated in sivelestat-treated Apo E^-/- mice (p < 0.001). The protein expression of ZO-1 was also significantly upregulated in sivelestat-treated Apo E^-/- mice, as compared with Veh-treated Apo E^-/- mice (Figures 3F,G). To determine the direct pharmacological effects of sivelestat on the ZO-1 protein expression, intestinal epithelial cells were treated with NE recombinant protein with or without sivelestat. Figures 3H–J showed NE recombinant protein decreased the gene and protein expression of ZO-1 (p < 0.001), whereas sivelestat attenuated NE-induced ZO-1 reduction in intestinal epithelial cells (p < 0.05).

Previous studies have showed there were abnormal immune responses in the gastrointestinal homeostasis of patients with atherosclerosis (Cani et al., 2007; Pendyala et al., 2012). Therefore, we investigated the effects of sivelestat on the intestinal inflammatory response in the HFHC-fed Apo E^-/- mice. As shown in Figures 4A–C, the protein levels of inflammatory cytokines, including TNF-α (Figure 4A, p < 0.001), IL-1β (Figure 4B, p < 0.01), and MCP-1 (Figure 4C, p < 0.01), were significantly decreased in intestinal lysates from sivelestat-treated ApoE^-/- mice. NF-κB signaling is one of the key transcriptional factors in regulating the inflammatory response (Pan et al., 2012; Pan et al., 2014). To this end, we measured the expression of phosphorylated (phos-) IκB and IκB in intestinal tissues. Figures 4D,E showed sivelestat inhibited NF-κB activation by decreasing phos-IκB and increasing the IκB level (p < 0.01).

To clarify whether the pharmacological benefits of sivelestat is dependent on lowering endotoxemia, we subcutaneously injected lipopolysaccharides (LPS) into sivelestat-treated ApoE^-/- mice. Administration of sivelestat decreased lipid accumulation in the aorta of Apo E^-/- mice (Figures 1A–D), but co-treatment with LPS reversely increased lipid contents in the en-face aorta (Figures 5A,B, p < 0.05) and the sections of the aorta root (Figures 5C,D, p < 0.01). Real-time PCR analysis of the mouse aorta further showed LPS eliminated the benefits of sivelestat on suppression of vascular inflammation, characterized by reduction of pro-inflammatory cytokines and chemokines (Figures 5E,F, p < 0.05).

To address the clinical applications, 26 volunteers were recruited for atherosclerotic analysis, including 12 patients with intestinal polyp surgery. The circulating endotoxin level and carotid intima-medial thickness (IMT) were measured in all subjects (Supplementary Table S1). As shown in Pearson analysis, the carotid IMT value was positively associated with the circulating level of endotoxin (Figure 6A) and intestinal NE activity (Figure 6B). Furthermore, there were structural disorders and higher expression of NE in atherosclerotic patients, as compared with healthy subjects (Figure 6C).