Peripheral blood of 90 hyperlipidemic patients were collected from University-Town Hospital of Chongqing Medical University. Hyperlipidemic patients with high blood pressure or diabetes mellitus were excluded. Peripheral blood of 90 healthy donors were used as control. Informed consent was signed by each subject. The study was approved by the Ethics Committee of Chongqing Medical University.

Six-to-eight-week-old male LDLr^-/- mice and C57BL/6 mice (wild-type mice, WT mice) were purchased from Beijing huafukang Biotechnology Co. All mice were bred in specific pathogen-free conditions. All animal experiments were undertaken with review and approval from the Animal Ethical and Experimental Committee of Chongqing Medical University.

LDLr ^-/- mice were fed with high-fat-diet (HFD) containing 0.15% cholesterol (medicine Ltd, China) for 12 weeks to generate lipid-induced atherosclerosis (LDLr^-/- HFD). WT and LDLr^-/- mice fed with normal diet were used as control (WT ND and LDLr^-/- ND). Each group contained eight mice. After feeding for 12 weeks, LDLr^-/- mice and WT mice were anesthetized and sacrificed. Blood in heart and aorta were flushed with sterile phosphate-buffered saline (PBS)-sodium heparin solution via cardiac puncture. Hearts were excised and fixed with 4% paraformaldehyde. Aortas were digested for 1 h at 37°C using an enzyme mixture containing 450 U/ml collagenase Is (Sigma-Aldrich, USA), 125 U/ml collagenase XI (Sigma-Aldrich, USA), 60 U/ml DNase I (Sigma-Aldrich, USA), and 60 U/ml hyaluronidase (Sigma-Aldrich, USA) as previously reported (36). Mouse peripheral blood was collected in the Eppendorf tubes containing Heparin sodium. Digested aorta, bone marrow from femur and tibia and spleen were prepared into a single cell suspension by grinding for further use.

Plasma samples were collected from LDLr^-/- mice and WT mice for lipids measurement. Total cholesterol (TC), total triglyceride (TG), low-density lipoprotein-cholesterol (LDL-C) and high-density lipoprotein-cholesterol (HDL-C) in mouse plasma were assayed with corresponding assay kit (Mindray, China) by using biochemical analyzer (Mindray, China).

Frozen sections of aortic sinuses were stained with Oil-red O (Solarbio, China) to determine the lipids deposition. Total lesion areas defined as intimal atherosclerotic areas and lesion lipid deposition areas were evaluated by Image J. Paraffin-embedded hearts were cut into 6-8-μm thick slides for Hematoxylin-Eosin staining and Masson’s staining to visualize the necrotic core areas and the collagen contents. Necrotic core areas and collagen contents of atherosclerotic lesions were measured by ImageJ.

Peripheral blood was obtained from healthy adult volunteers. Human neutrophils were isolated by density gradient centrifugation using Ficoll (Solarbio, China) according to manufacturer’s instructions. The purity of neutrophils was up to 98% ( Supplementary Figure S1A ). Peripheral blood mononuclear cells (PBMC) from healthy donors were isolated by density gradient centrifugation using Ficoll (Solarbio, China). CD3⁺ T cell from PBMC were sorted by Easysep™ human T cell isolated kit (Stemcell, Canada). The viability of sorted CD3⁺ T cells was higher than 90% and their purity was up to 97% ( Supplementary Figure S1B ).

Freshly isolated neutrophils were treated with ox-LDL, PMA, ox-LDL plus PMA, respectively. As for PMA and ox-LDL plus PMA groups, neutrophils were pretreated with 1 nmol/L PMA (Solarbio, China) for 30 min and the cells were washed with RPMI1640 medium. Then the cells were incubated with 40 μg/ml ox-LDL (Yiyuanbiotech, China) for 12h, 24h, 36h, 48h, separately, and harvested for flow cytometric analysis.

Purified CD3⁺ T cells (3×10⁶ cells/ml) were labelled with carboxyfluorescein succinimidyl ester (CFSE) and co-culture with neutrophils pre-stimulated with ox-LDL, PMA, or ox-LDL plus PMA at a 1:1 ratio in 200 μL RPMI-1640 medium containing rhIL-2 (20IU/mL) (Peprotech), anti-CD3 (2 μg/mL) (Biolegend), and anti-CD28 (1 μg/mL) (Biolegend) antibodies, with or without human CD80 neutralizing antibody (20 μg/mL) (Biolegend), human CD86 neutralizing antibody (20 μg/mL) (Biolegend) and human HLA-DR neutralizing antibody (20 μg/mL) (Biolegend). After 5-day incubation, the cells were collected for intracellular cytokine staining.

Flow cytometric analysis was performed according to standard protocols. Cell surface markers were stained with fluorescence labeled antibodies for 30 min at 4°C. Anti-human CD45, CD66b, CD80, CD86, HLA-DR, CD11c antibodies were applied to analyze the phenotype of neutrophils in hyperlipidemic patients and healthy donors. Anti-mouse CD45, Ly6G, CD80, CD86, MHC-II antibodies were applied for detecting neutrophil phenotype in atherosclerotic mice. For intracellular staining, the cells were stimulated for 6 hours with Cell Activation Cocktail with Brefeldin A (Biolegend, USA) (5). Intracellular cytokine staining was performed after the cells were fixed and permeabilized with fixation/permeabilization buffer (eBioscience, USA) for 20 min (37). Anti-human CD45, CD3, IFN-γ antibodies and anti-mouse CD45, CD3, IFN-γ antibodies were used for intracellular cytokine staining. All flowcytometric antibodies except anti-IFN-γ antibody (Invitrogen, USA)were purchased from Biolegend (USA).

Each experiment was performed at least three times. Data were expressed as mean ± SEM. Differences between hyperlipidemic patients and healthy donors were analyzed by the unpaired, Student’s test. Correlations between parameters were assessed using the Pearson correlation analysis and linear regression analysis as appropriate. The animal data and in vitro data were analyzed by non-parametric analysis. GraphPad Prism 7.0 was used for all statistical analysis. All data were analyzed using two-tailed tests, and p<0.05 was considered statistically significant.

To identify neutrophil subsets in hyperlipidemic patients, we first used flow cytometry to analyze the percentage of CD45⁺CD66b⁺ neutrophils within the total CD45⁺ leukocytes in different samples from 90 hyperlipidemic patients. 90 healthy donors were used as control. It showed that patients with hyperlipidemia had a greater proportion of neutrophils in peripheral blood than healthy donors ( Figure 1A ). Since neutrophils in patients with cancer or infectious diseases have the ability to heterogeneously express some co-stimulatory molecules, we postulated that there might be a subset of neutrophils with characteristics of APCs in atherosclerosis. So, the expression of APC-like markers on neutrophils were examined by flow cytometry. As expected, increased percentages of peripheral neutrophils in hyperlipidemic patients expressed APC signature markers including HLA-DR ( Figure 1B ), CD80 ( Figure 1C ) and CD86 ( Figure 1D ), but few neutrophils expressed dendritic cell (DC) marker CD11c ( Supplementary Figure S2 ). Further analysis of the co-expression of HLA-DR, CD80 and CD86 on neutrophils showed that the percentage of HLA-DR⁺ CD80⁺ CD86⁺ neutrophils in peripheral blood of hyperlipidemic patients is significantly higher than that of healthy donors ( Supplementary Figure S3 ). These results imply that a subset of APC-like neutrophils exists in hyperlipidemic patients. As most of the atherosclerosis cases developed from hyperlipidemia, we speculate that APC-like neutrophils may play pivotal roles in atherosclerosis.

To assess the distribution of APC-like neutrophils in atherosclerosis, we established a murine model of atherosclerosis by feeding LDLr^-/- mice on a high-fat diet (HFD) for 12 weeks (LDLr^-/- HFD cohort). LDLr^-/- mice and WT mice fed with normal diet (ND) were set as control groups (LDLr^-/- ND cohort; WT ND cohort). Oil Red O staining of the whole aorta from aortic arch to abdominal aorta showed that LDLr^-/- mice with HFD had increased atherosclerotic areas than control groups ( Supplementary Figure S4A ). Oil Red O staining of aortic root also showed the same trend ( Supplementary Figure S4B ). Masson trichrome staining and HE staining showed increased collagen contents and enlarged necrotic core areas appeared in aortic root of LDLr^-/- mice with HFD ( Supplementary Figures S4C, D ). Blood lipid levels including TC, TG, LDL-C and HDL-C of LDLr^-/- mice with HFD were significantly higher than control groups ( Supplementary Figure S4E ). Besides, the weight of LDLr^-/- mice fed with HFD was much higher than control ( Supplementary Figure S4F ). These results indicate that the mouse model of atherosclerosis is established successfully.

Then, the distribution of neutrophils in aorta, blood, spleen and bone marrow of atherosclerotic mice were analyzed by flow cytometry. As shown in Figure 2A , neutrophils were enriched in aorta of LDLr^-/- mice feeding with HFD. Similarly, atherosclerotic mice showed a higher neutrophil percentage in peripheral blood ( Figure 2B ) and spleen ( Figure 2C ) than control groups. However, no such differences were observed in the bone marrow ( Figure 2D ). Next, the subsets of neutrophils in different tissues were detected. In accordance with what found in hyperlipidemic patients, neutrophils in aorta of LDLr^-/- mice with HFD showed higher expression of APC-like markers including MHC-II molecules ( Figure 3A ), CD80 ( Figure 3B ) and CD86 ( Figure 3C ). The same trends were also observed in neutrophils from peripheral blood ( Figures 3D–F ) and spleen ( Figures 3G–I ) of atherosclerotic mice, while neutrophils in the bone marrow expressed none such markers ( Figures 3J–L ). Taken together, these results clarify the existence of APC-like neutrophils in atherosclerotic mice. The different distribution of APC-like neutrophils between peripheral tissues and bone marrow suggests that the APC-like neutrophils may be regulated under hyperlipidemic conditions other than generated in the bone marrow.

Since ox-LDL is one of the most important lipids in atherosclerosis, we wonder whether it could modulate the differentiation of APC-like neutrophils. To verify this hypothesis, neutrophils were isolated from peripheral blood of healthy donors and stimulated with ox-LDL, PMA, ox-LDL plus PMA for 12 hours, respectively. Surprisingly, neutrophils treated with ox-LDL alone showed little expression of HLA-DR, CD80, and CD86 ( Figures 4A–C ). However, these markers were dramatically upregulated on neutrophils treated with ox-LDL plus PMA ( Figures 4A–C ). The same results were observed when neutrophils were stimulated for 24h, 36h, 48h, respectively ( Supplementary Figure S5 ). These findings demonstrate that when neutrophils are activated by PMA, they can differentiate into APC-like phenotype by ox-LDL, which implies other factors in atherosclerosis may work together with ox-LDL to regulate the differentiation of APC-like neutrophils.

APCs are the key players in the immune response since they are capable of presenting antigens to T cells thereby initiating T cell responses (38). To determine the role of APC-like neutrophils in T cell function, CFSE-labeled CD3⁺ T cells were co-cultured with conditioned neutrophils which were pretreated with ox-LDL plus PMA. Then, T cell proliferation and IFN-γ production were measured by flow cytometry. Neutrophil/T-cell co-cultures showed that APC-like neutrophils (induced by ox-LDL and PMA) significantly promoted the proliferation and IFN-γ production of T cells, which could be significantly attenuated by blockade of HLA-DR, CD80 and CD86 on neutrophils ( Figure 5 ).

To verify these findings in vivo, we analyzed the distribution of T cells in atherosclerotic mice. Gate strategies of T cells and IFN-γ⁺CD3⁺ T cells were shown in Supplementary Figure S6 . Consistently, we found that more CD3⁺ T cells were infiltrated in aorta of LDLr^-/- mice with HFD than control groups ( Figure 6A ). The percentage of IFN-γ⁺CD3⁺ T cells in the aorta of LDLr^-/- mice with HFD was also much higher than that of control groups ( Figure 6B ). Increased CD3⁺ T cells and IFN-γ⁺CD3⁺ T cells were found in blood, spleen and bone marrow of atherosclerotic mice as well ( Figures 6C–H ). The in vitro and in vivo results together suggest that APC-like neutrophils may play an essential role in T cell activation, but further research is needed to verify it.

The demographic and clinical characteristics of the patients with hyperlipidemia were outlined in Supplementary Table S1 . The distribution of T cells in human peripheral blood was analyzed. The results showed that the percentage of CD3⁺ T cells in peripheral blood of hyperlipidemic patients was much higher than that in healthy donors ( Figure 7A ). Furthermore, a greater proportion of IFN-γ⁺ CD3⁺ T cells was found in blood of hyperlipidemic patients comparing to that of healthy donors ( Figure 7B ). Within the patient cohort, APC-like neutrophils were positively correlated with CD3⁺ T cells and IFN-γ⁺ CD3⁺ T cells, respectively ( Figures 7C, D ). The results were in accordance with that observed in neutrophil/T-cell co-cultures, which implies a stimulatory role of APC-like neutrophils in the early stage of human atherogenesis.