Cell lines and culture conditions

Taiwanese ESCC cell lines CE81T and CE48T were obtained from Dr. Han-Suei Hsu (Division of Thoracic Surgery, Taipei Veterans General Hospital, Taiwan). Japanese ESCC cell line, KYSE510, was purchased from DSMZ-German Collection of Microorganisms and Cell Cultures (Braunschweig, Lower Saxony, Germany), where they were characterized by DNA-fingerprinting and isozyme detection. Taiwanese lung patient-derived CAF cells were kindly provided by Dr. Wu-Chou Su (Department of Internal Medicine, National Cheng Kung University Hospital, Taiwan). Mouse Tongue Carcinoma MTCQ1 cell line was obtained from Dr. Shu-Chun Lin (National Yang-Ming University, Taipei, Taiwan). Mouse embryo fibroblast cell line NIH/3T3 was purchased from the American Type Culture Collection (ATCC, Manassas, VA, USA). CE81T, CE48T, MTCQ1, and NIH/3T3 cells were cultured in Dulbecco's Modified Eagle Medium (DMEM, Gibco, Waltham, MA, USA) containing 10% fetal bovine serum (Gibco) and 1% penicillin/streptomycin (Gibco). KYSE510 and CAF cells were maintained in RPMI-1640 Medium (Gibco) with 10% fetal bovine serum (Gibco) and 1% penicillin/streptomycin (Gibco). All cells were cultured at 37 °C with 5% CO2 in air.

Radiation treatment of cell lines

To establish the CCRT-resistant cell line, the parental cells (P) KYSE510 (KYSE510-P), CE81T (CE81T-P), and CE48T (CE48T-P) were continuously exposed to 5 Gy radiation per treatment in media containing 2 μM cisplatin (Sigma-Aldrich, Saint Louis, MO, USA). The chemoradiation-treated cells were given a chance to rest and recover until the next treatment, and received a total cumulative dose of 70 Gy radiation. The CCRT-resistant cells (R) KYSE510 (KYSE510-R), CE81T (CE81T-R), and CE48T (CE48T-R) were used for comparison studies.

Clinical samples of ESCC patients

A total of 172 plasma samples were collected from ESCC patients in the endoscopy operation room, National Cheng Kung University Hospital, Tainan. Tumor samples used for tissue microarray were from 77 ESCC patients at the Cancer Center of National Cheng Kung University Hospital, Tainan. All patients were recruited in this study with appropriate institutional review board permission (permit #A-BR-108-069) and informed consent were obtained from the patients. The follow-up of patients was performed at 3-month intervals. The last day of the follow-up period at National Cheng Kung University Hospital was defined as March 2021. Endoscopy tumor biopsy and corresponding normal biopsy samples were collected. Patient’s response to CCRT was evaluated using endoscopic ultrasound and computed tomographic scans after completion of 36 Gy radiotherapy. The evaluation was carried out before and after CCRT therapies at the clinics. Good CCRT responders were patients with post-CCRT esophageal wall thickness < 8 mm by endoscopic ultrasound and without enlargement or newly developed distant metastatic foci on computed tomographic scan, while poor CCRT responders were patients with post-CCRT esophageal wall thickness ≥ 8 mm or with enlargement or newly developed distant metastatic foci on computed tomographic scan [21]. These patient samples were examined for GSN protein expression using enzyme-linked immunosorbent assay and immunohistochemistry staining.

Enzyme-linked immunosorbent assay (ELISA)

The secretion level of human pGSN was detected by Gelsolin (Plasma) (Soluble) (Human) ELISA Kit according to the manufacturer’s instructions (AVISCERA BIOSCIENCE, Santa Clara, CA, USA). Briefly, 10,000 dilutions of plasma samples was prepared and analyzed concurrently with a range of known pGSN standards. The optical absorbance at 450 nm was measured.

The cancer genome atlas (TCGA) and genotype-tissue expression (GTEx) datasets analysis

The gene expression data of 62 ESCC samples from TCGA and 649 normal esophageal tissue samples from GTEx were derived from UCSC Xena (https://xena.ucsc.edu/). The HTseq-Counts were converted to FPKM (Fragments Per Kilobase per Million mapped reads) normalized data. The Log2 transformation of FPKM + 1 was used to present the differential expression of the GSN gene in normal and tumor samples.

Expression vectors and transfection

The entire coding region of pGSN PCR product was restricted using Bamh I and XhoI enzymes and then subcloned in frame into the pcDNA3.1/myc-His A vector. The pGSN-expressing vector was introduced into target cells using Turbofect reagent (Invitrogen, Waltham, MA, USA) according to manufacturer instructions. After 24 h incubation, the transfected cells were harvested for further assays. The plasmids used in the study are listed in Table S1.

Conditioned media (CM) collection, protein extraction, and western blot analysis

CM were harvested from the supernatant of cancer cells after transfection with pGSN-expressing vector for 24 h. 0.1% v/v protease inhibitors cocktail (Sigma-Aldrich) and 0.1% v/v 1 M dithiothreitol were added to the collected CM (pGSN-CM), and centrifuged (Z 216 MK, HERMLE, Germany) at 13,200 rpm at 4 °C for 15 min. The supernatant was collected for further assays.

The transfected cells were lysed on ice using the RIPA buffer containing protease inhibitors cocktail (Sigma-Aldrich), and the lysates were then centrifuged at 13,200 rpm for 15 min. An equal amount (50 µg) of protein extract was separated on a 8% SDS-PAGE and transferred onto a polyvinyl difluoride membrane. The membranes were blocked with 5% skim milk in Tris-buffered saline with 0.1% Tween-20 for 1 h at room temperature and subsequently incubated with primary antibodies at 4 °C overnight, followed by incubation with horseradish peroxidase-conjugated secondary antibodies. The antibody conditions are described in Table S2.

RNA extraction and quantitative reverse transcription polymerase chain reaction (RT-qPCR)

Total RNA was extracted using TRIzol reagent (Invitrogen), and purified RNA was reversely transcribed into cDNA using the High Capacity cDNA Reverse Transcription Kit (Applied Biosystems, Carlsbad, CA, USA). RT-qPCR was performed with SYBR Green Master Mix (Invitrogen) using the StepOnePlus™ Real-Time PCR system (Applied Biosystems). Expression levels were normalized with internal control β-actin. The primer sequences are listed in Table S3.

DNA demethylation treatment

ESCC cells were seeded (1 × 106 cells) in 10 cm dish and were treated with 100 µmol/L 5-aza-2’-deoxycitidine (5-aza, Sigma-Aldrich) or nutlin-3 (Selleck Chemicals, Houston, TX, USA) for 3 days, with replacement of fresh medium containing demethylation reagent 5-aza or nutlin-3 every 24 h. After treatment, total genomic DNA, RNA, and protein were extracted for the analysis of pGSN gene methylation status, and mRNA and protein expression.

DNA bisulfite conversion and methylation-specific polymerase chain reaction (MSP)

Genomic DNA was extracted using the Quick-DNA™ Miniprep Plus Kit (Zymo Research, Orange, CA, USA), and then an equal amount of DNA samples was treated with EZ DNA methylation-Gold kit (Zymo Research). The methylation level was determined by PCR analysis using the primers specific for either methylated (M) or unmethylated (U) DNA. Primers are listed in Table S3.

Immunohistochemistry (IHC) staining

All slides were dewaxed with xylene and antigens were retrieved by incubating the slides for 10 min at 100 °C. Then IHC staining was performed using Novolink Max Polymer kit (Leica Biosystems, Wetzlar, Germany) according to the manufacturer’s instructions. Evaluation of IHC was conducted blindly without knowledge of the clinical and pathologic characteristics of the patients. The staining was scored with brown-staining intensity and expression area percentage translated into a four-tier system, including strong positive (100–75%) as 4, moderate (75–50%) as 3, weak (50–25%) as 2 or negative (25–0%) as 1. Tumor spots of each patient on tissue microarray were scored and then averaged as the score of expression level. The antibody conditions are listed and described in Table S2.

Cell-based IP

ESCC cells were seeded (3 × 106 cells/well) in 10 cm dish. To perform IP analysis of cell surface protein, cells were washed with PBS and incubated with medium containing normal mouse IgG or integrin αvβ3 antibody on a shaker at 4 °C for 1 h. After hybridization, cells were rinsed with ice-cold PBS twice and lysed with 1 × IP buffer (50 mM Tris–HCl, pH 7.5, 150 mM NaCl, 20 mM α-glycerol-phosphate, 1% NP-40, 5 mM EDTA, 0.01% protease inhibitor cocktail) at 4 °C for 15 min. Cell lysate was sonicated, and centrifuged at 13,200 rpm at 4 °C for 15 min to isolate the supernatant. The protein extracts were quantified (500 μg) and incubated with 15 μL Protein G/Protein A agarose beads (Calbiochem, San Diego, CA, USA) at 4 °C overnight. Complexes were then washed with 1 × IP buffer 3 times. Proteins were eluted by boiling in 2 × SDS loading buffer, separated by 8% SDS-PAGE, then blotted with αvβ3 integrin, pGSN, or TNC antibodies. The antibodies conditions are listed and described in Table S2.

Cycloheximide (CHX) chase assay

For CHX chase assay, CE81T-R and 3T3 cells were treated with CM from empty vector (EV) control or pGSN-overexpressing (OE) ESCC cells along with cycloheximide (CHX, 20 μg/mL), a potent inhibitor of protein synthesis. Protein lysates were harvested at the indicated times and analyzed by immunoblotting analysis.

Immunofluorescence (IF)

For immunofluorescence staining, Opal stain kit (PerkinElmer, Waltham, MA, USA) was utilized. Cells were fixed with 4% formaldehyde and then antigen retrieval was performed with citrate buffer (pH 6.0) at 100 °C for 20 min. After blocking, the slides were incubated with primary antibody of p-FAK or p-paxillin, followed by incubation with secondary antibody polymer HRP and subsequently with Opal fluorophore for 10 min at room temperature. Finally, DAPI was applied for nuclear staining, and images were visualized using a fluorescence microscope (Olympus, Tokyo, Japan). The antibody conditions are listed in Table S2.

Collagen gel contraction assay

Collagen gel matrices were prepared by adding 250 μL of a neutralized collagen (Vitrogen, Angiotech, Palo Alto, CA, USA) solution [8 mL collagen (3 mg/mL), 1 mL 10X DMEM, 1 mL 200 mmol/L HEPES, pH 8.5], to a 24-well tissue culture well. After incubating CAFs on the restrained gels, gels were released and their two longest diameters were measured at the indicated time points. Gel size was defined as the sum of the two gel diameters and gel contraction expressed as a percentage of the original gel size.

Animal studies

All animal experiments were performed in compliance with institutional guidelines for use and care of animals (permit #109,060). To investigate the role of pGSN in tumor growth, 5–6-week-old BALB/c nude mice were subcutaneously implanted with 5 × 106 CE81T-R cells. To evaluate the anti-CAF effects of pGSN, 5–6-week-old BALB/c nude mice were subcutaneously co-implanted with 5 × 106 CE81T-R and 5 × 105 3T3 fibroblasts.

Mice were weighed, and the volumes of the xenografts were measured and quantified during experiment. At the endpoint of the experiments, tumor tissues were excised and then subjected to RNA extraction or fixed with 4% formaldehyde (Sigma-Aldrich).

Statistical analysis

The statistical analyses of pGSN expression level, overall survival, and cancer risk factors were performed using Statistical Package for the Social Sciences version 26.0 (SPSS Inc., Chicago, IL, USA). The chi-square test and multivariate logistic regression analyses were conducted. Correlations were examined using Pearson’s correlation test. Overall survival curves were calculated according to the Kaplan–Meier method using the log-rank test. Three independent experiments for cell studies and six mice per group for animal studies were analyzed. Two-tailed Student’s t-test and one-way ANOVA test were used in cell and animal studies. Data represented mean ± s.e.m. The levels of statistical significance were expressed as p-values, *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001; ns: non-significant.