Patients and tissue specimens
NAG-NOR (CXCL17, n = 30; GPR35, n = 30), AG-IM (CXCL17, n = 74; GPR35, n = 60), GC-IM (CXCL17, n = 88; GPR35, n = 91), and GC (CXCL17: n = 260; GPR35, n=227) tissues were collected from October 2012 to June 2019 in the First Affiliated Hospital of China Medical University. Samples were obtained from surgical operations or endoscopic biopsies. Patients with unavailable tissue samples or uncertain diagnosis were excluded. Clinical characteristics of patients were obtained from electronic medical record. Survival information was obtained from 111 patients in the CXCL17 group and 103 patients in the GPR35 group, respectively. The last follow-up date was November 25, 2019.
The study was approved by the Institute Research Medical Ethics Committee of the First Affiliated Hospital of China Medical University and informed consent was waived by the committee because of the difficulty for contacting the patients and that the patient's privacy was fully protected.
Formalin-fixed and paraffin-embedded gastric tissues were cut into 4-μm-thick sections and subjected for immunohistochemistry using biotin-avidin complex method at room temperature. Briefly, after deparaffinizing and rehydrating, the sections were heated in boiled EDTA buffer for antigen retrieval. Then successive incubation with primary and secondary antibodies was performed and DAB (DAB-0031, maxim Inc., Fujian, China) was used for staining. Anti-CXCL17 antibody (catalog number: MAB4207; R&D Systems, Inc.) was used at a concentration of 10 ug/ml and anti-GPR35 antibody (cytoplasmic domain ab 188949; Abcam) was used at a concentration of 7 ug/ml, respectively. For quality control, negative controls were run in parallel, and PBS was used instead of the primary antibody.
Evaluation of immunohistochemistry
Semi-quantitative scoring method was utilized to assess the expression level of CXCL17 and GPR35. The percentage of stained cells (S) and immunostaining intensity (I) were evaluated. The percentage score ranged from 0 to 4 (0% - 5% scores 0; 5%–25% scores 1; 25%–50% scores 2; 50%–75% scores 3; 75%–100% scores 4) and the intensity score ranged from 0 to 3 (0, no staining; 1, weak; 2, moderate; 3, strong). All fields under the microscope were viewed. Different pathological lesions were distinguished and scored by two experienced and highly qualified pathologists independently. The final IS score was obtained by multiplying the percentage of stained cells (S) with intensity (I), which ranged from 0 to 12.
TCGA data mining
GEPIA2 (http://gepia2.cancer-pku.cn/#index) website was used to compare the mRNA expression of CXCL17 and GPR35 in stomach adenocarcinoma (STAD) vs. corresponding normal tissues by matching TCGA normal and GTEx data21. R software (R 4.0.3) was applied for co-expression (volcano plots) and correlation analysis (correlation plot) utilizing TGCA data normalized by the log2 [TPM (Transcripts per million) +1] transformation. R package “limma” was used to identify the differential expression genes (DEGs) after dividing the TCGA samples into two groups according to the expression level of CXCL17 and GPR35, respectively. The DEGs shared by CXCL17 and GPR35 were used for PPI network construction using the “multiple protein” module of STRING, with the edges indicating both functional and physical protein associations (https://string-db.org/cgi/input?sessionId=bJl0ekoEqcJ1&input_page_show_search=on). Then the PPI result was import into Cytoscape software for visualization and disconnected nodes in the network were hidden.
Cell culture and transfection
HGC-27 (RRID: CVCL_1279) cell line was purchased from the cell bank of Chinese Academy of Medical Science (Beijing, China). It had been authenticated using STR profiling and were mycoplasma-free. Cells were cultured in RPMI 1640 medium (Solarbio, China) containing 10% foetal bovine serum (Biological industries, Israel) and placed in a 37℃ incubator with 5% CO2. For transfection, 60 to 80% confluent cells in a 6-well plate were transfected with CXCL17 overexpressing or empty plasmid (Gene, China) according to the jetPRIME® DNA transfection protocol. Twenty-four hours later, cells were lysed with TRIzol reagent (Tiangen, China) and the total RNA was extracted, reversed into cDNA and subjected to real-time fluorescence qPCR as indicated. β-Actin was applied for normalization. Program is set as 95 ℃ for 3min, and 45 cycles of 95 ° C for 10s, 60 ° C for 20s and 72 ° C for 30s. The 2− ΔΔCT method was adopted for calculation. The primer sequence of CXCL17 was as follows: forward 5’-TGCTGCCACTAATGCTGATGT-3’ and reverse 5’-CTCAGGAACCAATCTTTGCACT-3’.
Total proteins were isolated from cells 48 hours after transfection using RIPA Lysis Buffer (Beyotime) supplemented with protease inhibitors on ice for 30 min, following by centrifugation, quantification with a BCA protein assay kit (Beyotime) and denaturation by boiling. Then the proteins were separated by SDS-PAGE and transferred onto a PVDF membrane. The membrane was blocked with QuickBlock™ Blocking Buffer (Beyotime) for 15 minutes followed by incubation with primary antibodies overnight at 4°C. After washing with TBST for 3 times, the membrane was incubated with goat anti-rabbit or goat anti-mouse secondary antibodies (Abcam) at room temperature. The chemiluminescent signal was detected by an ECL method. The antibodies and dilutions applied in this study were CXCL17 (MAB4207, R&D): 1:1000, β-Tubulin (100109-MM05T, SinoBiological): 1:8000, CDX1 (204876-T08, SinoBiological): 1:2000, SFRP2 (ab92667, Abcam): 1:500 and CCL20 (10485-T24, SinoBiological): 1:2000.
All statistical analyses were conducted using SPSS software (version 26.0, Chicago, IL) or R software (R 4.0.3). Between-group comparisons of protein expression levels were analyzed by non-parametric test and quantitative data were presented as mean ± SD. The relationship between CXCL17 and GPR35 protein expressions was assessed by Spearman correlation analysis. Chi-square test was used to assess the associations of these two proteins’ expressions with clinical parameters. Univariate survival analysis was performed by log-rank test and cox proportional hazards model was used for multivariate survival analysis adjusted by variables that had P < 0.05 in the univariate analysis (Supplemental Table S1). The median of IS score was used to identify the high and low expression groups. An IS score less than or equal to the median score was evaluated as low expression and higher than that score was identified as high expression. In cancerous cases, the median score was 3 and 6 in CXCL17 and GPR35 cohort, respectively. Reported P values were two-sided, and the significance level was set at 0.05 for all analyses.