Immunofluorescence microscopy

The 29 NPC samples in our institutional cohort were divided into low and high expression groups based on the median value of IHC score of KIF18B. Eight samples were randomly selected from each group and conducted immunofluorescence staining. Briefly, the steps before incubation of secondary antibody were same as described in the above IHC protocol. The following primary antibodies were used: anti-FOXP3 (1:200, 98377S, Cell Signaling Technology), anti-CD8 (1:100, ZA-0508, OriGene), and anti-IBA-1 (1:200, 17198, Cell Signaling Technology). After incubation of primary antibodies, the slides were incubated with DyLight 488-labeled secondary antibody (1:500, ab96899, Abcam) in the dark at 37°C for 1h. Images were captured using immunofluorescence microscopy (Olympus, Tokyo, Japan). The quantification of cell numbers was carried out manually by counting the positive cells, while the immunofluorescence area was automatically identified using Image Pro Plus software (version 6.0.0.260).

TUNEL staining was conducted with TUNEL assay kit (Roche) following antigen retrieval and permeabilization. Nucleus labeled with DAPI (Sigma‐Aldrich) appears blue, and TUNEL‐positive apoptosis cells labeled with FITC are green. All the slides were visualized using immunofluorescence microscopy (Olympus, Tokyo, Japan).

For tumor metastases assessment, both lungs were gently dissected and photographed for quantification of lung tumor burden. For histological examination, formalin-fixed paraffin-embedded lung tissue was cut into 4- μM sections, placed on glass slides, dewaxed with xylene, and rehydrated with 30–100% ethanol. The sections were microwave-boiled in extract buffer (10 mM citric acid buffer, pH 6.0) for 5 min and stained with H&E, and a histopathological examination was performed. The stained sections were analyzed under a light microscope (Carl Zeiss). For immunofluorescence experiment, pulmonary specimens were fixed with 4% paraformaldehyde and embedded in optimal cutting temperature compound before sectioned. Frozen sections were incubated with DAPI (Invitrogen). Immunofluorescence was assessed by immunofluorescence microscopy (Olympus, Tokyo, Japan), and data were collected with ImageJ (Bethesda).

U87 cells were trypsinized, then plated onto 6-well culture plates with suitable culture medium (high-glucose DMEM, 10% fetal bovine serum, 100 U/ml penicillin, 100 μg/ml streptomycin) containing 10³ cells. When 80% of the U87 cells were confluent, the complete culture media was aspirated and replaced with fresh EBM-2 medium containing 5% fetal bovine serum. EC monolayers grown on glass coverslips were placed into culture plates and co-cultured for 4 days. The GEC monolayers grown on glass coverslips were fixed with 4% paraformaldehyde and permeabilized with 0.5% Triton X-100. After blocking with 2% bovine serum albumin (BSA) in PBS, the cells were incubated with mouse anti-occludin (diluted 1:50; Life Technologies Corporation, Frederick, MD, USA) and rabbit anti-claudin-5 (diluted 1:50; Life Technologies Corporation, Frederick, MD, USA) to visualize the distribution of occludin and claudin-5, respectively. The cells were incubated with rhodamine-labeled phalloidin (diluted 1:250; Sigma-Aldrich) to assess actin filaments. The cells were incubated with mouse anti-RhoA (diluted 1:150; Abcam, Cambridge, UK) and rabbit anti-ROCK II (diluted 1:50; GeneTex, Irvine, California, USA) to analyze their expression. The glass slides were then evaluated using immunofluorescence microscopy (Olympus, Japan).