Formalin-fixed-paraffin-embedded (FFPE) tissue specimens were obtained from 100 lung adenocarcinoma patients who underwent surgical operation from 2014 to 2017 at the First Affiliated Hospital of Anhui Medical University. Patients included in this study had not received any preoperative adjuvant therapy. All specimens and protocols used in this study were approved by the Ethics Committee of Anhui Medical University (No. 2021H004). The protocols used to be under the ethical standards of the Biomedical Ethics Committee of Anhui Medical University Committee and the 1964 Helsinki Declaration.

FFPE samples sliced into 4 μm sections were deparaffinized in xylene and graded alcohol, followed by hydration. Antigen retrieval was performed in a pressure cooker containing 0.01 M sodium citrate buffer (pH 6.0) for 10 min. Tissue sections were incubated with either anti-CD66b (Cat.No.555723, BD Pharmingen) or anti-Notch3 (Cat.No.5276S, Cell Signaling Technology) overnight at 4°C. After blocking by endogenous peroxidase for 30 min, samples were incubated with the HRP-labeled sheep anti-mouse/rabbit IgG polymer (Cat.No.PV-6000, ZSGB-BIO) for 20 min at room temperature. Finally, 3,3′-Diaminobenzidine (DAB) was used as chromogen, and the sections were counterstained with hematoxylin.

The density of CD66b + neutrophils infiltrating in tumor parenchyma was evaluated as previously reported (Wang et al., 2019). Briefly, CD66b + neutrophils were counted under ten random microscopic high power fields (×40 objective lenses). According to the median value of all samples, CD66b + neutrophils were divided into high-density and low-density groups. Notch3 immunostaining intensity was graded on 0–3 scales: 0, absence of staining; 1, weak staining; 2, moderate staining; 3, strong staining. Notch3 positive cancer cells were similarly scored on a scale of 0–4 as follows: 0, 0–5% positive tumor cells; 1, 6–25% positive tumor cells; 2, 26–50% positive tumor cells; 3, 51–75% positive tumor cells; 4, >75% tumor cells. The IHC scores (0–12) were obtained by multiplying the staining intensity by the percentage scores; the score <4 was negative, ≥4 was positive.

Blood obtained from healthy adults was centrifuged at 250 g for 15–20 min. After discarding the plasma, normal saline was added. The diluted blood was carefully added onto a Ficoll solution and centrifuged at 450 g for 25 min. After discarding the supernatant, RBC lysing buffer was added twice each for 10 min. After centrifuging at 1500 rpm and discarding the supernatant, the cells were washed twice with PBS, and the cells were resuspended in a serum-free medium.

TCCS (tumor cells culture supernatant) was obtained from the supernatant of A549 and PC-9 lung cancer cells cultured for 48 h in a serum-free medium. NeuCS (Neutrophils culture supernatant) was obtained from the supernatant of neutrophils cultured in the serum-free medium for 10 h. TANCS (tumor-associated neutrophils culture supernatant) was obtained from the supernatant of neutrophils stimulated by TCCS for 10 h.

The viability of TANs and neutrophils was determined by Annexin-V/PI apoptosis detection kits (Biobest). The data were analyzed by FlowJo (Version. 7.6.1).

Migration and invasion were performed in an 8 μm 24-well culture insert (Corning Company). For invasion assay, Matrigel (BD Company) diluted with the serum-free medium was precoated in the chambers and solidified for 30 min at 37°C. Lung cancer cells were added to the upper chamber, whereas TANCS, NeuCS, TCCS, and medium were added to the lower chamber, respectively. The cancer cells were cultured for 7 and 12 h for migration and invasion assay, respectively. Cancer cells attached to the underside of the chambers were fixed and stained. Quantification was performed by evaluating the mean number of cells in five microscopic fields per chamber.

The cells were seeded in a 6-well plate at 3.6 × 105 cells per well and were cultured overnight. After the next day, 75 pmol si-Notch3 or negative control was transfected by using Lipofectamine3000. The transfected cells and untransfected cells were cultured in opti-mem. After 6 h, each group of cells was treated with TANCS, NeuCS, TCCS, and medium, respectively. At the same time, the cells were grown to confluence and wounded by dragging a 0.2-ml pipette tip through the monolayer. The extent of cell migration was observed. The rate of wound closure was expressed as a percentage of the initial scraped gap.

RNA was extracted from A549 and PC9 cells using TRIzol reagent (ThermoFisher Scientific), and cDNA was made by reverse transcription with (Evo M-MLV RT Master Mix) according to the manufacturer’s instructions. qRT-PCR was performed using (SYBR^® Green Pro Taq HS Premix) in 20 μL reactions. Primers were designed using Primer3 software and purchased from Sangon Biotech: GAPDH (Sangon Biotech, B662104-0001), Notch3 Fw 5′-GCT​ACA​CTG​GAC​CTC​GCT​GT-3′, Notch3 Rv 5′-AGA​CCC​CAC​CGT​TGA​CAC​AG-3’. Reactions were carried out in LightCycler^® 96 Application Software (Roche). The cycling program used was 95°C for 30 s, followed by 40 cycles of 95°C for 5 s, 60°C for 30 s. Data were analyzed using GAPDH as a reference gene.

Proteins extracted using RIPA buffer were loaded onto SDS–polyacrylamide gels and then transferred to PVDF membranes. Membranes were blocked with 5% nonfat milk in TBS-T for 1 h at room temperature. Membranes were incubated overnight at 4°C with the following primary antibodies: NOTCH3 (Cat.No. 5276S, CellSignaling Technology) and GAPDH (Cat.No. abs132004, Absin). After incubating the membranes with peroxidase-conjugated secondary antibodies (Cat.No. ZB-2301, ZSGB-BIO) for 2 h, the reaction was visualized by an enhanced chemiluminescence assay.

The Kaplan–Meier plotter ofers a means of readily exploring the impact of a wide array of genes on patient survival in 21 diferent types of cancer, with large sample sizes for the lung adenocarcionma (n = 865) cohorts. We therefore used this database to explore the association between Notch3 expression and outcome in patients with lung cancers (http://kmplot. com/analysis/).

TIMER (https://cistrome.shinyapps.io/timer/) is a database designed for the analysis of immune cell infltrates in multiple cancers. This database employs pathological examination-validated statistical methodology to estimate infltration by neutrophils, macrophages, dendritic cells, B cells and CD4+/CD8+ T cells into tumors. We therefore explore the efect of neutrophils infltration on the survival rate of lung adenocarcinoma patients (n = 496).

All experiments were carried out three times. The data were expressed as a mean ± standard deviation (SD). Comparison between groups was analyzed by chi-square test. Kaplan-meier was used for survival analysis and p-value < 0.05 was considered statistically significant.

TANs were observed in both parenchyma and stroma of lung adenocarcinoma (Figure 1A). Only those infiltrating in tumor parenchyma were counted. The density of TANs infiltrating in solid and micropapillary adenocarcinoma was significantly higher than that in lepidic adenocarcinoma. The density of TANs showed a positive correlation to the TNM stage but correlated negatively to tumor differentiation (p < 0.05, Supplementary Table S1). Survival analysis showed that the density of TANs was negatively associated with disease-free survival (DFS) (p < 0.05, Figure 1B).

Notch3 was mainly expressed in the cytoplasm of cancer cells as detected by immunohistochemistry (Figure 1A), and the expression of Notch3 in solid adenocarcinoma was significantly higher than that in lepidic adenocarcinoma. The expression of Notch3 was negatively correlated to tumor differentiation (p < 0.05), but showed no significant correlation with gender, age, tumor size, TNM stage, or lymph node metastasis (Supplementary Table S1). Survival analysis showed that the expression of Notch3 was negatively correlated to DFS (p < 0.05, Figure 1C). In lung adenocarcinoma tissues, the infiltration of TANS was positively correlated with the expression of Notch3 (p < 0.05, Figure 1D).

The purity of neutrophils isolated from peripheral blood of healthy adults is 94.0% (Figure 2A). The supernatant of cultured A549 and PC-9 lung adenocarcinoma cell lines was used to stimulate neutrophils isolated from the peripheral blood of healthy adults for 10 h respectively. Compared with unstimulated neutrophils, a significant increase was observed in the lifespan of neutrophils treated with supernatant of cultured PC-9 (87.3% v. s 38.5%) and A549 (82.0% v. s.38.5%) cell lines (Figure 2B).

Transwell assays were performed to investigate the effect of TANs on migration and invasiveness of lung adenocarcinoma cells. Compared with lung adenocarcinoma cells that were stimulated by medium, NeuCS, TCCS, the migration and invasion of lung adenocarcinoma cells stimulated by TANCS were significantly higher (Figures 3A,B). The same result was observed on the wound healing assay (Supplementary Figure S1).

Notch3 has been widely reported to be associated with cancer development. Therefore, whether Notch3 is involved in the migration and invasion of lung adenocarcinoma cells was explored. Our results showed a significant increase in the expression of Notch3 in lung adenocarcinoma cells treated with TANCS than that treated with medium, NeuCS, and TCCS respectively (Figures 3C,D). However, after knocking down of Notch3, the migration and invasion of lung adenocarcinoma cells stimulated with TANCS were not enhanced (Figures 3E–J and Supplementary Figure S2). The results suggested that TANs promoted the migration and invasion of lung adenocarcinoma cells via Notch3.

The results from the Kaplan-Meier plotter showed that the expression of Notch3 correlated negatively with OS (overall survival), FPS (first-progression, survival), and PPS (post-progression survival) of patients (Figure 4A). From the TIMER (http://cistrome.dfci.harvard \.edu/TIMER/, Figure 4B) analysis, the infiltration of TANs was found to correlate negatively to the OS of patients. Our findings indicated that both TANs and Notch3 were associated with poor prognostic outcomes in patients.