Human cell lines and reagents
AGS (RRID CVCL_0139) and NCI-N87 (RRID CVCL_1603) subsequently referred to as N87 are Lauren intestinal-type GA cell lines, and KATOIII (RRID CVCL_0371) and SNU-668 (RRID CVCL_5081) are Lauren diffuse-type GA cell line. AGS and N87 cells were obtained from the America Type Culture Collection (ATCC). KATOIII and SNU-668 cells were obtained from the Korean Cell Line Bank (KCLB). Cancer cell lines were actively passaged for less than 6 months from the time that they were received from the ATCC or KCLB, and United Kingdom Co-ordinating Committee on Cancer Research (UKCCCR) guidelines were followed [12]. KATOIII cells were maintained in DMEM, and and the other GA cell lines were maintained in RPMI 1640. All media were supplemented with 10% FBS, 100 U/mL penicillin and 100 mg/mL streptomycin, and L-glutamine 2 mmol/L ("regular media"). Human umbilical vein endothelial cells (HUVEC) were obtained from Lonza (Basel, Switzerland) and used within 8 passages. All endothelial cells were grown in EGM-2-MV media (Lonza). Mouse Primary Vein Endothelial Cells (MVEC; Cell biologics, C57-6009) were maintained with Endothelial Cell Medium /w Kit (Cell Biologics, M1168).
Reagents were purchased from the following sources: XenoLight D-Luciferin, PerkinElmer Inc. (#122799); MEK inhibitor, Santa Cruz (sc-364412A); Matrigel, BD Bioscience (Cat. 354248); B27, Sigma-Aldrich (0080085SA) ; N2, Thermo Fisher Scientific (A13707-01); N-acetylcysteine, Sigma-Aldrich (A9145); Epidermal growth factor (EGF), Sigma Aldrich (E9644); Fibroblast growth factor-basic (FGF-Basic), Sigma Aldrich (341583); Gastrin I, Sigma Aldrich (G9145); Nicotinamide, Sigma Aldrich (N0636), Y-27632, Sigma Aldrich (Y0503); SB202190, Sigma Aldrich (S7067); Prostaglandin E2, Tocris Bioscience (#2296); Recombinant Rspondin 1, PeproTech (120 − 38); mNoggin, PeproTech (250 − 38); Wnt3A, R&D Systems (5036-WIN); FGF-10, PeproTech (100 − 26); A83-01, R&D Systems (#2939); Recombinant human-VEGF-A, R&D Systems (293-VE-010), FGF-2, R&D Systems (233-FB-010).
Spheroid generation
Cells were resuspended in spheroid media comprised of DMEM/F12 containing epidermal growth factor (EGF), basic fibroblast growth factor (bFGF), N-2 supplement, and B27 and then plated on ultra-low attachment culture dishes (Corning Life Sciences). Spheroids were collected after 5–7 days except when noted otherwise. Proteins were extracted for analysis, or cells were dissociated with Accutase and used for other experiments. Spheroid growth was quantified as the average number of spheroids > 50–100 µm in diameter among 5 fields after image processing using Imaris 7.6 (Bitplane).
Mouse tumor-derived organoids and cell lines
The Tcon3077 and Tcon3944 gastric tumors were harvested from Tcon mice (see below), and organoids were generated and maintained as previously described [13]. Tcon3077 and Tcon3944 cell lines were also generated from Tcon tumors. Tcon cell lines were maintained in DMEM supplemented with 10% FBS, 100 U/mL penicillin and 100 mg/mL streptomycin, and L-glutamine 2 mmol/L ("regular media").
Western blot analysis
Proteins were extracted by collecting cells in RIPA buffer (Sigma-Aldrich) containing Complete Protease Inhibitor Cocktail (Roche), and protein concentration was determined using the Bio-Rad Protein Assay (Bio-Rad). Proteins were detected using the following antibodies: KRAS (sc-30) from Santa Cruz.; VEGF-A (ab46154, sc-7269) from Abcam and Santa Cruz; and β-actin (A5441) from Sigma.
Lentiviral transduction
VEGF-A was silenced via lentiviral transduction of mouse VEGF shRNA (sc-36815-V; Santa Cruz Biotechnology). KRAS was silenced via lentiviral transduction of human or mouse KRAS shRNA (sc-36815-V, sc-43876-V; Santa Cruz Biotechnology). Lentiviral transduction of a scramble shRNA (sc-108080; Santa Cruz) was used as a control. Luciferase (firefly) lentivirus (FCT005, Kerafast, Inc.) was transduced following the manufacturer’s protocol. Maximal knockdown of genes occurred 72–96 h after transduction.
Angiogenesis antibody arrays
The relative levels of human and mouse angiogenesis-related proteins in GA cells grown as monolayers and as spheroids were measured using the Proteome Profiler Human Angiogenesis Array Kit and Proteome Profiler Mouse Angiogenesis Array Kit (ARY007, ARY015, R&D Systems Inc.) following the manufacturer’s protocol. The results were analyzed with ImageJ software.
Human and mouse endothelial cell tube formation assay
Tube formation assays were performed in Matrigel-coated (BD Biosciences) 24-well plates. Conditioned medium in GA cells treated with control, KRAS and/or VEGF-A shRNA were collected after 3 days. After HUVEC (3×104) and MVEC (3×104) were seeded on the 24 wells in 12 hours, 300 µL of conditioned medium from cancer cells were replaced. Images were taken using a bright-field microscope at 100× magnification, and the total length of completed tubule structures was quantified.
Gastric cancer mouse models
All mouse protocols were approved by the Institutional Animal Care and Use Committee. Tcon mice were generated by cross breeding as previously described [5]. Treatment of Tcon mice with a MEK inhibitor (PD0325901, APExBIO) was initiated in 4-week-old mice (n = 7 per group). The drug was administered ad libitum in the mouse chow (Purina 5010) at 7 mg/kg (incorporation by Research Diets Inc). The same mouse chow without drug was used as the control. DC101 (50-562-188, Bio X Cell) 20 mg/kg or control IgG antibody 20 mg/kg was also initiated in 4-week-old mice. At 10 weeks, 2 mice from each group were sacrificed and tumors were harvested. Stomachs were fixed in 10% buffered formalin for 24 h, embedded in paraffin, and processed into 5 µm sections.
For orthotropic intra-gastric injection mouse models, anesthesized mice were injected with 1 x 105 CD44(+) cells in 100 ul /HBSS into the gastric wall via a laparotomy incision. At 8, 10, 12, and 15 weeks after tumor cell injection, mice (n = 7 per group) were intraperitoneally injected with 100 µL D-luciferin solution (150 mg/kg, #122799, Perkin Elmer), anesthesized with 2% isoflurane, and imaged in a Xenogen IVIS 200 (PerkinElmer) after 5 min according to the manufacture's protocol. To harvest the tumors, 2 mice in each group were sacrificed at 12 weeks, stomachs were harvested and processed as above.
Lung, liver, and lymph node metastasis models
To generate experimental lung metastases, mice were injected via the tail vein with 5 x 104 CD44(+), CD44(-), or unsorted Tcon3077 cells (n = 6 per group). Mince were sacrificed at 3.5 weeks. Lungs of all mice were fixed in 10% buffered formalin for 24 h, embedded in paraffin, and processed into 10 µm sections. For each mouse lung, 10 sections were examined by H&E staining for lungs metastases. In a subsequent experiment, 5 X104 CD44(+) cell stably transduced with control shRNA, KRAS shRNA, VEGF-A shRNA, or both KRAS and VEGF-A shRNA were injected into the tail vein of mice (n = 7 per group). At 3.5 weeks after tumor cell injection, mice were intraperitoneally injected with 100 µL D-luciferin solution (150 mg/kg, #122799, Perkin Elmer), anesthetized with 2% isoflurane, and imaged in a Xenogen IVIS 200 (PerkinElmer) after 5 min according to the manufacture's protocol. Mice were then sacrificed, and lungs were harvested.
To generate experimental liver metastases, 5 x 104 Tcon3077 CD44(+) cells were injected into the spleen (n = 6 per group). At 2, 3, and 4 weeks after tumor cell injection, mice were intraperitoneally injected with 100 µL D-luciferin solution (150 mg/kg, #122799, Perkin Elmer), anesthetized with 2% isoflurane, and imaged in a Xenogen IVIS 200 (PerkinElmer) after 5 min according to the manufacture's protocol. Mice were sacrificed at 4 weeks, and livers and spleens were harvested. Livers and spleens of all mice were fixed in 10% buffered formalin for 24 h, embedded in paraffin, and processed into 10 µm sections. For each mouse liver and lung, 10 sections were examined.
To generate lymph node metastases, 5 x 104 Tcon3077 CD44(+) cells grown from cells were injected into the foot-pads of mice on day 0 (n = 6 per group). Following sacrifice at 4 weeks, enlarged inguinal lymph nodes were harvested, fixed in 10% buffered formalin for 24 h, embedded in paraffin, and processed into 10 µm sections.
Immunohistochemistry (IHC)
Immunohistochemistry was performed for patient and mouse tumor samples using the following primary antibodies: CD31 (MA5-13188; Thermo Fisher Scientific), CD44 (NBP1-31488; NOVUS Biological USA), p-MEK1/2 (#9154; Cell signaling), YFP (MBS833304; MYBIOSource), p-ERK1/2 (#9101; Cell signaling), VEGF-A (sc-7269; Santa Cruz), and LYVE-1 (AF2089, AF2125; R&D system).
Validation in independent patient cohorts
RNA-seq data of paired normal and tumor samples for 83 GAs from Jun et al. were used for independent validation [14]. We performed hierarchical clustering using ComplexHeatmap v2.9.3 [15] with log2 fold-change values of three genes (KRAS, VEGFA, and CD44), and divided patients into three subgroups based on CD44 expression. Fisher’s exact and Chi-squared tests were performed to evaluate differences between subgroups. Gene set enrichment analysis was used to identify overrepresented biological functions using Gene Set Enrichment Analysis (GSEA) [16]. We calculated the enrichment scores using gene sets from the MSigDB Hallmark collection (n = 50) [17].
Patient samples
Institutional review board approval was obtained from FMUUH. Gastric tumors from surgically resected specimens were fixed in formalin, embedded in paraffin, and sectioned. GC sections or FMUUH patient tumor TMAs were deparaffinized prior to incubation with antibodies against human CD44 (NBP1-31488; NOVUS Biological USA) and VEGF-A (sc-7269; Santa Cruz) in a solution of PBS with 1% BSA and 0.1% Triton X-100 at 4°C overnight according to standard protocols as previously described. The intensity of staining for CD44 and VEGF-A was scored as 0 to 3. The proportion of positively stained cells was scored as follows: ≤ 5% positive cells, 0; 6 to 25% positive cells, 1; 26 to 50% positive cells, 2; ≥ 51% positive cells, 3. To obtain an IHC score that considers the IHC signal intensity and the frequency of positive cells, the intensity score was multiplied by the percentage score. Composite scores less than 3 were defined as low expression and scores of 4 or higher as high expression.
Statistical analysis
Data are represented as mean ± standard deviation unless otherwise noted. Statistical analyses were performed using GraphPad9. P values were calculated using Student’s t-test, except for comparisons between more than two groups, in which case treatment groups were compared to controls using one-way ANOVA with Bonferroni adjustment for multiple comparisons.
For analyses of patient data, categorical variables were analyzed using χ2 or Fisher’s exact test. Overall survival curves were plotted using Kaplan–Meier methods and compared by the log-rank test. Cox proportional hazards regression modeling was used to examine the relationship between CD44 and VEGF-A expression and survival, while controlling for confounding covariates. Analyses were performed using IBM SPSS software for Windows version 21 (IBM). A p value < 0.05 was considered statistically significant.