Epcam

Multiplex immunofluorescence staining of tissues from the discovery and validation cohorts were performed using TSAPLus multiple fluorescence staining kit (Cat# G1236, Servicebio) according to manufacturer's instructions. The slides were used for triple staining of AQP1 (Cat# 66 805, Proteintech), AGTR2 (Cat# bs‐2133r, Bioss), and EpCAM (Cat# GB12274, Servicebio) antibodies, double staining of GZMK (Cat# 67 272, Proteintech) and CD8 (Cat# 66 868, Proteintech) antibodies, and double staining of CTSB (Cat# 12 216, Proteintech) and CD68 (Cat# YM3050, ImmunoWay) antibodies. Nuclei were stained with 4′,6‐diamidino‐2‐phenylin‐ dole (DAPI) solution (Cat# G1012, Servicebio) after all the antigens above had been labeled. The complete list of the antibodies is shown in Table S6 (Supporting Information). The stained slides were scanned to obtain multispectral images using the Pannoramic MIDI (3DHISTECH), which captured the fluorescent spectra at 420, 520, 620, and 700 nm with identical exposure times, respectively. For each slide, five fields of tumoral area were selected for image capture and quantification. The fluorescence intensity of each image was measured by using ImageJ 1.52 v (National Institutes of Health, Bethesda, MD, USA).

We established a tissue microarray (TMA, 3 mm) consisting of LUAD samples: 14 AIS samples, 18 MIA samples, and 17 IAC samples (Supplementary Table 4). After antigen repair and blocking, the TMA was incubated with specific primary antibodies (EPCAM, Servicebio, GB14078, 1:200; FOXP3, Servicebio, GB11093, 1:200; TPSB2, Novus, NBP2-33551, 1:300; CD79A, Novus, NB100-64347ss, 1:200; CLDN5, Servicebio, GB11290, 1:200; COL1A1, Affinity, AF7001, 1:200; UBE2C, Abcam, ab12290, 1:50; p-SMAD2, Affinity, AF8314, 1:200; SCGB1A1, Servicebio, GB111412, 1:200; PDPN, Affinity, AF3670-SP, 1:200; FCN1, Novus, NBP1-84706, 1:200; GHD, Proteintech, 67538 1:250; PCNA, Servicebio, GB11010, 1:300), incubated with horseradish peroxidase (HRP)-labeled secondary antibodies or fluorophore-labeled secondary antibodies, and finally stained with diaminobenzidine and counterstained with hematoxylin or stained with DAPI. Three independent pathologists distinguished the pathological type for the TMA.

After antigen retrieval and blocking, tissue sections were incubated with specific primary antibodies (EPCAM, Servicebio, GB14078, 1:200; FOXP3, Servicebio, GB11093, 1:200; CD3, Servicebio, GB11014, 1:200), which were mixed with horseradish oxidase-labeled secondary antibodies or fluorophore-labeled secondary antibodies were incubated and finally stained with diaminobenzidine and counterstained with hematoxylin. This study was approved by the ethics committee of the Shaanxi Provincial People’s Hospital (N0.20220424).