Peritoneal fluid collection
This study was approved by the institutional review board at Gangnam Severance Hospital, Yonsei University College of Medicine, Seoul, Korea (3-2015-0102) and written informed consent was obtained from each patient. Thirty patients diagnosed with pancreatic cancer in the Gangnam Severance Hospital from 2018 to 2019 received pancreatic resection (Supplementary Table 1). Among them, 17 patients were treated with IORT (irradiated with a single dose of 10 Gy at a depth of 5 mm into the tumour bed) immediately after surgical resection, as previously described.  Patients were divided into the group that received IORT (IORT PF, n=17) and the group that did not receive IORT (no IORT, n=13). Peritoneal fluid (PF) was collected from the usual surgical drain 24 h after the surgery. Sterile filtered PF samples were centrifuged at 2100 rpm for 15 min and then stored at -80 ℃.
Blood sample collection and isolation of peripheral blood mononuclear cells
Blood was collected three times from patients (post-operation day (POD) 1, 7, and 14) using vacutainer EDTA tubes (BD Biosciences, San Jose, CA, USA). Whole blood was diluted with Dulbecco’s phosphate-buffered saline (DPBS), gently layered over an equal volume of Ficoll-Paque solution (GE Healthcare, Logan, UT, USA), and centrifuged at 400 g for 30–40 min. Peripheral blood mononuclear cells (PBMCs) were collected from the second layer, and any remaining platelets were gently washed off using DPBS. Cells were resuspended in freezing medium containing 10% DMSO and 90% foetal bovine serum (FBS; Biowest, Riverside, MO, USA) and stored at -80 ℃.
Mia PaCa-2, Panc1, and Aspc1 cells were obtained from American Type Culture Collection (Manassas, VA, USA). Cells were cultured in DMEM or RPMI medium (Biowest) supplemented with 10% FBS (Biowest) and 1% antibiotic–anti-mycotic reagent (Gibco, Waltham, MA, USA). Cells were incubated at 37 ℃ in a humidified atmosphere under 5% CO2.
PF samples were assayed using the Human Cytokine Antibody Array kit (Abcam, Cambridge, UK) according to the manufacturer’s instructions. Briefly, each sample was incubated for 24 h at 4℃ and then paired with biotinylated detector antibodies and horseradish peroxidase (HRP)-conjugated streptavidin. Each cytokine was analysed using chemiluminescence and levels were quantified using ImageJ software (National Institutes of Health, Bethesda, MD, USA).
Isolated PBMCs were stained with the following fluorochrome-conjugated monoclonal antibodies: anti-human APC-CD4 (eBioscience, San Diego, CA, USA), APC-cy7 CD8 (BD Biosciences, San Jose, CA, USA), PE-cy5 CD25 (Biolegend, San Diego, CA, USA), PE-Cy7 Foxp3 (eBioscience), BV421 CD56 (BD Biosciences, San Jose, CA, USA), and APC-CD19 (Biolegend). Live cells were classified using propidium iodide (BD Biosciences, San Jose, CA, USA) staining. Lymphocytes were further subtyped by their staining properties as T cells (CD3+), Th cells (CD3+CD4+), Tc cells (CD3+CD8+), NK cells (CD3-CD56+), or Treg cells (CD4+CD25+FOXP3+). Treg cells were fixed and underwent permeabilization using the Fix & Perm Buffer (eBioscience), and stained cells were analysed using BD FACSCanto II Cell Analyzer (BD Biosciences, San Jose, CA, USA). FlowJo software (BD Biosciences, San Jose, CA, USA) was used for compensation and data analysis.
For the 2D proliferation assay, 1.5–3×103 cells per well were seeded into a 96-well cell culture plate. Cells were allowed to grow in serum-free medium supplemented with 5% FBS or PF. After incubation for 4 days, reduced medium with 10% WST-1 reagent (EZ-cytox, Dogen, Korea) was placed into the wells after aspiration of the growth medium. The absorbance of each well was measured at 450 nm using a VersaMax microplate reader (Molecular Devices, San Jose, CA, USA).
For the colony formation assay, 100–500 cells/well were seeded into six-well culture plates containing reduced serum medium supplemented with 5% PF and antibiotics. The medium was replaced every 3 days until colonies formed. The rate of colony formation was different for each cell, usually taking 7–14 days. Cells were fixed in cold 100% methanol for 30 min, stained using 2.5% crystal violet, and then washed several times with phosphate-buffered saline (PBS). Colony formation was calculated manually and digitally using ImageJ software, where colony area versus total area was calculated based on staining intensity.
Migration and invasion assays
For the invasion assay, the 8 µm pore size Transwell system (Corning, Midland, MI, USA) was coated with Matrigel (1:50, Corning) at 37 °C for 1 h. Then, 2×104 cells were seeded on the apical side of the Transwell chamber (24-well insert) in serum-free media. DMEM with 7% PF and 1% antibiotic–anti-mycotic agent was added to the basal compartment. The cells were allowed to invade for at 37 °C 24 h. The cells that remained in the top chamber were gently scraped off using wet cotton swabs. The cells that invaded the basal side were fixed in methanol for 10 min, stained with 0.2% crystal violet, and then washed multiple times with distilled water. The invasion assays were performed in triplicate.
For the scratch wound migration assay, 2×104 cells were seeded into a 96-well plate (Image lock, IncuCyte™; Essen Bioscience, Ann Arbor, MI, USA), and wound scratches were made using a wound maker tool (Essen Bioscience) 18 h after plating. The media in each well was supplemented with 5% PF and 1% antibiotic–anti-mycotic agent, and FBS was used as positive control. Images of the migrated cells were captured automatically every 4 h, and the relative wound density was analysed using IncuCyte™ Chemotaxis Cell Migration Software (Essen Bioscience).
After the indicated treatment, cells were harvested and washed with ice-cold PBS and lysed using RIPA lysis buffer. Proteins (30 μg sample) were separated using SDS-PAGE and transferred onto nitrocellulose membranes, blocked in 5% skim milk, and incubated with the following primary antibodies (1:1000): Anti-phospho-smad2, phospho-smad3, smad2, and smad3 (Cell Signaling Technology, Danvers, MA, USA), anti-Snail+slug (Abcam), anti-vimentin (Cell Signaling Technology), N-cadherin (BD Biosciences, San Jose, CA, USA), E-cadherin (BD Biosciences, San Jose, CA, USA), and γ-tubulin (Sigma-Aldrich, St. Louis, MO, USA). The membranes were then washed thrice with Tris-buffered saline and Tween 20 (TBST), followed by incubation with HRP-conjugated secondary antibody (1:7000, Cell Signaling Technology) in TBST with 3% skim milk. Bound antibody was probed using ECL solution.(Biorad,USA) Chemiluminescent signals were captured using X-ray films. All experiments were performed in triplicate.
RNA isolation and qPCR
After the indicated treatment, cells were collected, and their RNA was isolated using TRIZOL Reagent® (Sigma-Aldrich) according to the manufacturer’s instructions. Then, 0.2 µg total isolate RNA was analysed via reverse transcriptase PCR using the One-Step RT-PCR Kit (iNtRON Biotechnology, Seongnam-Si, Korea). First-strand cDNA synthesis was performed with 1 µg RNA as a template using the RT-qPCR cDNA Synthesis Kit (iNtRON Biotechnology), according to the manufacturer’s instructions. RT-qPCR was performed using the SYBR qPCR reaction mix (Applied Biosystems, Foster City, CA, USA). The primer sequences used in this study are listed in Table 1. Relative mRNA expression level was calculated using the 2−ΔΔCT method, using GAPDH as the reference gene.
Statistical analysis was performed using GraphPad Prism version 8.01 software (GraphPad Software, La Jolla, CA, USA). Unpaired t-test was performed for statistical analysis of cytokine array, cell proliferation, western blotting, and qPCR data. Wound healing assay and PBMC phenotyping data were analysed by two-way ANOVA. Differences were considered statistically significant at * p < 0.05, ** p ≤ 0.01, and *** p ≤ 0.001.