The present study was performed as a joint research project of the Faculty of Medicine, University of Yamanashi (Yamanashi, Japan) and the Division of Digestive Surgery, Department of Surgery, Kyoto Prefectural University of Medicine (Kyoto, Japan). A total of 46 and 84 patients with gastric cancer were respectively included in a development cohort (Table I) and an independent validation cohort (Table SI) to assess the clinicopathological characteristics of plasma PDPN levels. Furthermore, the analysis was performed using the data of all 130 patients from both cohorts (Table II). The development cohort included patients who underwent radical resection at Yamanashi University Hospital (Yamanashi, Japan) between December 2017 and May 2020. The independent validation cohort included patients with stage I–III gastric cancer who underwent radical resection at Kyoto Prefectural University of Medicine Hospital (Kyoto, Japan) between June 2010 and December 2015. The prognostic impact of plasma PDPN expression was also investigated by determining the 5-year overall survival (OS) rate and 5-year the recurrence-free survival (RFS) rate in the validation cohort with stage I–III gastric cancer. Patient demographic data and details of tumor recurrence and subsequent management were recorded. The pathological classification of tumors was determined according to the Union for International Cancer Control classification (12).

From each patient, a 5-ml blood sample was collected in an EDTA tube prior to surgery. Plasma was immediately separated from the cellular fraction by centrifugation as described elsewhere (13) and then stored at −80°C for further processing.

Quantification of PDPN expression in the plasma was performed using the Human Podoplanin ELISA kit (cat. no. ELH-PDPN-1; RayBiotech, Inc.) according to the manufacturer's protocol. In brief, 100 µl of plasma from each patient was pipetted into wells coated with an antibody specific for human PDPN. The plate was incubated for 2.5 h at room temperature to allow binding of the immobilized antibody to PDPN in the plasma. After washing the wells, biotinylated anti-human PDPN antibody was added, followed by incubation for 1 h at room temperature. After washing, bound PDPN was incubated with horseradish peroxidase (HRP)-conjugated streptavidin for 45 min at room temperature; the binding of HRP-conjugated streptavidin was detected with 3,3′,5,5′-tetramethylbenzidine substrate solution. The intensity of the color was measured at a wavelength of 450 nm.

Immunohistochemistry was performed using 32 tissue samples of stages II and III from the Yamanashi University cohort. Formalin-fixed, paraffin-embedded tissue was cut into 4-µm slices that were placed on glass slides. Slides were deparaffinized using xylene and rehydrated using a graded series of ethanol solutions. Antigen retrieval was performed by heating the samples in Dako Target Retrieval Solution (Agilent Technologies, Inc.) for 20 min at 120°C. Endogenous peroxidases were quenched using peroxidase blocking reagent (Dako; Agilent Technologies, Inc.). Sections were incubated overnight at 4°C with D2-40 monoclonal antibody (cat. no. 413451; not diluted; Nichirei Biosciences, Inc.). After washing, immunoperoxidase staining was performed using a Vectastain ABC elite kit (Vector Laboratories) and 3,3′-diaminobenzidine tablet (Fujifilm) according to the manufacturers' instructions, followed by counterstaining with hematoxylin. Lymphatic epithelium stained positive for PDPN was used as a positive control for each slide. The investigators were blinded to the clinicopathological data of the patients. PDPN expression was evaluated using high-power microscopy (magnification, ×100) in five different fields. PDPN expression in the stroma surrounding the cancer cells was evaluated. Samples with the presence of immunoreactivity in >10% of stromal cells were considered positive. The CAFs at the peritumoral and tumor invasion fronts were assessed.

The human gastric cancer cell lines NUGC-3 and MKN74 were used in the present study. Cell lines were obtained from the Japanese Collection of Research Bioresources Cell Bank and were cultured in RPMI 1640 medium (Thermo Fisher Scientific, Inc.) supplemented with 100 U/ml penicillin (Sigma-Aldrich; Merck KGaA), 100 µg/ml streptomycin (Sigma-Aldrich; Merck KGaA) and 10% fetal bovine serum (FBS; Thermo Fisher Scientific, Inc.). Cells were incubated in a 5% carbon dioxide atmosphere at 37°C.

To assess whether soluble PDPN is able to alter the phenotype of gastric cancer cells, cells were treated with PDPN protein and then subjected to the assays specified below. As the PDPN protein, the Podoplanin/Fc Chimera, Human, Recombinant, carrier-free (R&D Systems, Inc.) was used.

A migration assay was performed using Falcon cell culture inserts with 8-µm pore membranes for use with 12-well plates (Corning, Inc.). Furthermore, an invasion assay was performed using Falcon cell culture inserts with 8-µm pore membranes for use in 24-well plates (Corning, Inc.), for which the insert was coated with Biocoat Matrigel^® (BD Biosciences) prior to use.

In the migration and invasion assays, gastric cancer cells (2×10⁵ cells/ml) were seeded in the upper chambers in FBS-free medium with or without PDPN. The medium volume in the migration assay was 1 ml and the medium volume in the invasion assay was 500 µl. PDPN protein was used at a concentration of 2 µg/ml. Medium containing 10% FBS was added to the lower chambers.

After incubation for 24 h, cells that had not migrated or invaded through the pores were removed using cotton swabs. The migrated or invaded cells were fixed and stained with Diff-Quick staining reagent. Cells were counted in four independent fields at a magnification of ×100 using a BZ-X9000 All-in-One fluorescence microscope and BZ-X Analyzer Software Hybrid cell count (Keyence Corporation).

Gastric cancer cells were seeded into 12-well plates at a concentration of 2.0×10⁵/ml and then treated with PDPN at 2 µg/ml. Cells were then incubated for 48 h and subsequently peeled off and counted by the Automated cell counter Countess (Invitrogen; Thermo Fisher Scientific, Inc.).

Continuous variables in each group were compared using the Wilcoxon signed-rank test. The χ² test and Fisher's test were used to compare the categories of each group. The Cox proportional hazards model was used to investigate the clinicopathological factors that were significantly associated with plasma PDPN. Survival curves were constructed using the Kaplan-Meier method and compared using the log-rank test. Differences were assessed using a two-sided test P<0.05 was considered to indicate a statistically significant difference.

The range of the plasma PDPN concentration was 0-678.6 ng/ml. In the present study, the cut-off value of the plasma-PDPN concentration was set to the median plasma-PDPN concentration in the development cohort that was then divided into the high-PDPN and low-PDPN groups. The cut-off value was 0.6 ng/ml.

The associations between the clinicopathological characteristics of patients with gastric cancer and the results of the plasma PDPN status determined with PDPN ELISAs are provided in Tables I and II. In the development cohort, there were no significant differences in clinicopathological parameters between the two groups; however, the high-PDPN group tended to have a higher proportion of diffuse-type gastric cancer according to their Lauren classification (14) (P=0.079; Table I). Multivariate logistic regression analysis of all 130 patients in both cohorts revealed that a high-PDPN status was significantly associated with diffuse-type gastric cancer (P=0.008; Table II).

The prognostic impact of PDPN in the plasma was then investigated in the validation cohort (Fig. 1). Kaplan-Meier survival estimates indicated that the recurrence-free survival (RFS) rate was significantly lower (P=0.029) and the overall survival rate tended to be lower (P=0.126) in the high-PDPN group (Fig. S1). The 5-year RFS rate in the PDPN-positive group was 56.1% and that in the negative group was 76.8%. There was no significant difference in recurrence patterns between the high- and low-PDPN groups. A subgroup analysis of 36 patients with diffuse-type gastric cancer according to the levels of plasma PDPN expression revealed that the RFS rate tended to be lower in the high-PDPN group than in the low-PDPN group, but it was not significant (P=0.166; Fig. S1).

To investigate the origin of soluble plasma PDPN, the association between plasma PDPN expression and tissue PDPN expression was examined through immunohistochemical staining. Representative results of the immunohistochemical detection of the PDPN protein are provided in Fig. 2. Of the 46 patients in the development cohort, the data of 32 patients, for whom both tissue and plasma samples were obtained, were examined. Among these, 15 patients had high plasma PDPN expression. PDPN expression on the tumor invasion front was observed in 12 patients, of whom 9 (75%) had high plasma PDPN expression. By contrast, 14 (70%) of the 20 patients who did not exhibit PDPN expression on the tumor invasion front had low plasma PDPN expression (P=0.027; Table III). These results indicated that plasma PDPN may reflect PDPN expression in CAFs at the tumor invasion front.

To investigate whether soluble PDPN itself is able to affect phenotypic changes in gastric cancer cells, NUGC-3 and MKN74 cells were treated with soluble PDPN.

In the migration and invasion assays, the addition of PDPN did not enhance the migratory ability compared with the control. In addition, there was no significant difference in proliferative ability between the two groups (Figs. S2 and S3).