B7-H3 promotes gastric cancer cell migration and invasion

B7-H3 (B7 homologue 3, CD276) is a member of the B7 immunoregulatory family and promotes tumor progression. The present study demonstrated that B7-H3 promotes gastric cancer cell migration and invasion. shRNA-mediated B7-H3 silencing in the N87 gastric cancer cell line suppressed cell migration and invasion in vitro and in vivo; downregulated metastasis-associated CXCR4; and inhibited AKT, ERK, and Jak2/Stat3 phosphorylation. B7-H3-silenced cells injected into the tail veins of 4-week-old female BALB/c nude mice produced fewer metastases than control cells, and resulted in longer survival times. Immunofluorescence analyses confirmed B7-H3/CXCR4 colocalization in N87 cells, and co-immunoprecipitation assays showed a direct interaction between the two proteins. Our analysis of 120 tissue samples from gastric cancer patients showed that increased B7-H3 expression correlated positively with both tumor infiltration depth and CXCR4 expression. These findings suggest that B7-H3 and CXCR4 may be novel targets for anti-gastric cancer therapeutics.


INTRODUCTION
Gastric cancer is the fourth most frequently diagnosed cancer and the second most common cause of cancer-related deaths worldwide [1]. Improved diagnostic and therapeutic strategies have improved early-stage gastric cancer detection and decreased patient mortality [2]. Still, although new anticancer agents, such as S-1, tananes, capecitabine, oxaliplatin, and irinotecan [3][4][5][6][7], have improved gastric cancer patient prognoses, survival rates remain unsatisfactory [8,9]. Novel biomarkers are needed to improve early tumor metastasis predictions, and effective anti-metastasis/invasion agents are required to enhance patient outcomes.
Metastasis is a complex process involving cell intravasation to vessels, extravasation, invasion into a distant organ, and establishment and proliferation of secondary tumors in a new microenvironment [24,25]. Many molecules are implicated in the process of cancer metastasis. CXCR4 is a stromal cell-derived factor-1 (SDF-1) receptor that reportedly promotes cancer progression, including cell migration, invasion, and seeding to distal tissues [25][26][27], and is abnormally overexpressed in gastric cancer tissues [28]. In addition, AKT,
Here, we showed that B7-H3 promotes gastric cancer cell migration and invasion, and its upregulation enhances tumor infiltration depth. B7-H3 silencing downregulated CXCR4 expression and inhibited phosphorylation of AKT, ERK, and Jak2/Stat3 pathway members. Finally, we confirmed that B7-H3 and CXCR4 colocalize in gastric cancer cells and can interact directly.
Mouse survival was analyzed using the Kaplan-Meier method ( Figure 4B). Five of six mice injected with LV-NC-infected cells were sacrificed due to apparent symptoms of metastatic disease, as compared to only two of six in the LV-B7-H3 group (Gehan-Breslow-Wilcoxon test, P < 0.05).   Mice were euthanized at five weeks, and tissues were observed under a dissecting microscope. The lung was the predominant organ with metastases (Table 3). Mice injected with LV-B7-H3-infected N87 cells had more lung metastases than mice injected with LV-NC-infected cells (mean 4.500±0.4282 versus 2.333±0.5578, P < 0.05; Figure 4C-4D).
Mouse lung and spleen tissues were analyzed via H&E staining to further assess metastases. No major differences in lung metastasis morphology were observed between B7-H3-silenced and control cell-treated mice ( Figure 4E). H&E staining of spleen metastases is shown in Supplementary Figure 1C.

Effect of B7-H3 silencing on metastasisassociated molecules
B7-H3 silencing downregulated the metastasisassociated molecule, CXCR4, in N87 cells, as confirmed by flow cytometry, western blotting, and real-time PCR  and LV-NC-infected N87 cells via immunofluorescence (IF) confocal microscopy. Colocalization (yellow fluorescence) was assessed by superimposing B7-H3 (red fluorescence) and CXCR4 (green fluorescence) confocal images, and suppressing all colors except yellow. Nuclei were counterstained with DAPI. Fluorescence imaging demonstrated a high degree of B7-H3 localization with CXCR4 in the cell membrane and cytoplasm ( Figure 5D). In B7-H3-silenced N87 cells, CXCR4 fluorescence intensity was reduced.
To assess the possibility of a direct interaction between B7-H3 and CXCR4 in N87 cells, total protein extracted from LV-NC-infected N87 cells was immunoprecipitated with an anti-B7-H3 antibody (4H7). Precipitated proteins were analyzed by immunoblotting with either B7-H3 or CXCR4 antibodies. CXCR4 was pulled down with B7-H3 by the anti-B7-H3 antibody, whereas neither protein was recovered when a control antibody (IgG) or no antibody was used. Co-immunoprecipitation with each specific antibody showed an association between B7-H3 and CXCR4 in N87 cells ( Figure 5E).
The present study retrospectively studied B7-H3 in 120 human gastric cancer cases. Our results showed that higher B7-H3 levels in gastric cancer tissues was associated with tumor infiltration depth, in agreement with the findings of Wu, et al [23]. We also investigated the impacts of shRNA-mediated B7-H3 silencing in the gastric cancer cell line, N87. CCK-8 assay results showed no difference between B7-H3-silenced N87 cells and controls (P>0.05). However, wound healing and transwell assays showed that B7-H3 silencing reduced N87 cell migration and invasion. B7-H3-silenced cells injected into nude mice produced fewer metastases than control cells, and led to longer survival times than controls.
CXCR4 is a stromal cell derived factor-1 (SDF-1) receptor involved in cancer cell migration and invasion [25][26][27]. CXCR4 was downregulated in B7-H3-silenced N87 cells, as shown by flow cytometry, western blotting, and real-time PCR. We confirmed B7-H3 and CXCR4 colocalization via IF imaging, and direct interaction through a co-immunoprecipitation study. B7-H3 silencing in N87 cells also inhibited AKT, ERK, and Jak2/Stat3 phosphorylation. B7-H3 likely works in concert with these molecules and pathways to promote gastric cancer cell invasion and metastasis.
In summary, we investigated the role of B7-H3 in gastric cancer cell migration and invasion in vitro and in vivo. We showed that B7-H3 silencing downregulated CXCR4 and inhibited AKT, ERK, and Jak2/Stat3 phosphorylation. We confirmed B7-H3/CXCR4 colocalization and direct interaction. Our findings suggest that B7-H3 and/or its associated molecules, including CXCR4, may be novel targets for anti-gastric cancer therapeutics.

Patients and clinical specimens
Tissue samples from 120 gastric cancer patients were analyzed immunohistochemically. Patients were randomly selected from all those undergoing radical gastric cancer resection between 2014 and 2015 at the Department of Pathology of the Second Affiliated Hospital of Soochow University. None had received chemotherapy or radiation therapy before surgery. Gastric carcinoma diagnoses were confirmed via hematoxylin and eosin (H&E) staining after surgical resection. Average patient age was 61.19±1.098 years (range: 28-79 years). The tumor was located in mucosa in 9 cases, in shallow muscularis in 18 cases, and in deep muscularis in 93 cases. Three tumors were well-differentiated, 50 were moderately-differentiated, and 67 were poorly-differentiated (Table 1). This study was approved by the ethics committee of our hospital, and all patients provided written informed consent prior to enrolment.

Immunohistochemical staining
All surgically resected specimens and biopsy samples were fixed with 10% neutral buffered formalin, embedded in paraffin, and serially sectioned at 4 μm. IHC was performed on selected slides using the ChemMate TM Envision/HRP technique [41]. Briefly, sections were deparaffinized and dehydrated, and endogenous peroxidase activity was blocked using H 2 O 2 . Sections were incubated with B7-H3 and CXCR4 primary antibodies followed by secondary antibody, and visualized with diaminobenzydine (DAB). Finally, slides were counterstained with hematoxylin. B7-H3 or CXCR4 immunopositivity was evaluated according to overall staining intensity and area (percentage of positively stained tissue). Staining intensity was scored as follows: weak: 1; moderate: 2; strong: 3. Staining area was scored as follows: < 33%: 1; > 33% to <66%: 2; > 66%: 3. Final scores were calculated by multiplying a given case's intensity and area scores. A final grade of < 3 was considered low expression, while > 3 was classified as high. All slides were evaluated independently by two investigators blinded to patient identities and clinical outcomes.

Gastric cancer cell lines and cultures
The human gastric cancer line, N87, was purchased from ATCC and cultured in RPMI 1640 medium (HyClone, USA) containing 10% heat-inactivated fetal bovine serum (FBS) (Biological Industries, Israel) and 1% penicillin/ streptomycin (Biological Industries, Israel). Cells were incubated at 37°C in a humidified chamber with 5% CO 2 .

Flow cytometry
Isolated cells were washed in phosphate-buffered saline (PBS) containing 2% FBS and incubated with fluorochrome-conjugated antibodies (B7-H3-APC or CXCR4-PE-cy7) for 30 min at 4°C. Labeled cells were re-suspended in 0.5 ml cell staining buffer and analyzed using flow cytometry (FlowJo software v 7.6.2, USA). Isotype controls were performed for each staining.

Western blotting
LV-NC-or LV-B7-H3-infected N87 cells were washed twice with PBS. Cells were incubated in protein Oncotarget 71732 www.impactjournals.com/oncotarget lysis buffer containing protease inhibitors for 30 min on ice, and then centrifuged for 30 min at 12,000 rpm at 4°C. Equal amounts (10-30 μg) of total protein extracts were subjected to SDS-PAGE and transferred to PVDF membranes. Membranes were blocked with 5% BSA for 1 h and incubated overnight with primary antibody at 4°C, followed by incubation with the appropriate HRPconjugated secondary antibody for 2 h at 20°C. Blots were visualized with an ECL detection kit (BIO-RAD, CA, USA) and ChemiScope (Model No. 6300). GAPDH was used as a loading control. Band intensities were calculated via densitometric analysis in Image J (Rawak Software, Inc. Germany).

Cell viability assay
Cell proliferation was evaluated using the Cell Counting Kit-8 (CCK-8, Dojindo, Japan) according to the manufacturer's instructions. 0.2×10 4 LV-NC-or LV-B7-H3-infected N87 cells in 100 μl of RPIM-1640 media supplemented with 10% FBS were seeded into each well of a 96-well plate. At the indicated time points, medium was exchanged for 90 μl RPMI-1640 and 10 μl CCK-8, and cells were incubated for 2 h. Absorbance was measured for each well at a wavelength of 450 nm. Cell growth was monitored every 24 h over 5 d. All experiments were repeated three times.

Wound healing assay
LV-B7-H3-or LV-NC-infected N87 cells were incubated in 6-well plates and small straight-line wounds were made in confluent monolayers using a 200-μl pipette tip. Cells were washed twice with sterile PBS and incubated in RPMI 1640 medium containing 2% FBS at 37°C in a humidified chamber with 5% CO 2 for 24 h. Wound images were captured at 0 and 24 h at 100× magnification using an inverted microscope (Leica DM IL LED, Wetzlar, Germany). This experiment was conducted in triplicate.

Migration and invasion assays
For migration and invasion assays, 0.5×10 4 cells in 100 μl of serum-free RPMI media were seeded onto the tops of 8-μm pore size transwell chambers or transwell matrigel invasion chambers (BD Biosciences, San Jose, CA, USA). 50 μl matrigel was diluted 1:7 in serum-free media and incubated for 4 h at 37°C in preparation for the transwell matrigel invasion chambers. Lower compartments contained RPMI 1640 medium with 10% FBS. After 24 h incubation at 37°C in a humidified chamber with 5% CO 2 , non-invading cells and gel were removed from the upper chamber using cotton tipped swabs. Cells were fixed with methanol for 30 min and stained with crystal violet. Invading cells were counted in three random fields per filter at 200× magnification for triplicate wells. This experiment was conducted in triplicate.

Immunofluorescence analyses
IF analysis was used to assess B7-H3 and CXCR4 colocalization in LV-B7-H3-or LV-NC-infected N87 cells. 1 × 10 5 cells were seeded on glass coverslips in 24-well plates. After 1 d, cells were washed three times in PBS for 5 min/wash, fixed in 4% paraformaldehyde for 5 min, and washed again in PBS three times. Cells were treated with 0.25% Triton X-100 with 0.2% BSA for 5 min and then washed three times in PBS. Cells were blocked with 1% normal horse serum,and incubated with specific primary antibodies (B7-H3 or CXCR4) for 5 h at 4°C. After three PBS washes, cells were incubated with Cy3-conjugated anti-goat IgG (1:400; Jackson ImmunoResearch, USA) or Alexa Fluor 633-conjugated anti-rabbit IgG (1:400; Jackson ImmunoResearch, USA) for 1 h at 37°C, and then rinsed three times with PBS for 5 min/was. Nuclei were counterstained with DAPI (Sigma-Aldrich, USA) for 8 min, and slides were coverslipped. Cells were observed under a laser confocal microscope IX71 (Olympus, Japan) with a digital camera (Olympus, Japan).

Immunoprecipitation
LV-NC-infected N87 cells were lysed in polysome lysis buffer containing 100mM KCl, 5mM MgCl 2 , 10mM HEPES (pH 7.0), 0.5% NP-40, 1mM DTT, and protease inhibitor cocktail. Lysates were incubated on ice for 30 min, and then centrifuged for 30 min at 12,000 rpm at 4°C to remove cellular debris. Supernatants were added to protein A/G beads (Santa Cruz Biotechnology, CA, USA) with 4H7 (anti-B7-H3 antibody), which was previously incubated overnight at 4°C and was washed five times in wash buffer (50 mM Tris-HCl pH 7.4, 150 mM NaCl, 1 mM MgCl 2 , and www.impactjournals.com/oncotarget 0.05% NP-40). Beads were also washed five times in wash buffer before the bead mixture was incubated at 4°C for 6 h. Proteins were eluted with 2×SDS sample buffer, and then subjected to western blotting analysis.

Experimental metastasis model in athymic rodents
LV-B7-H3-or LV-NC-infected N87 cells (1×10 6 in 100 μl PBS) were injected into the tail veins of 4-week-old female BALB/c nude mice. Each group consisted of six mice. Mice were checked daily for symptoms of metastasis, including non-normal physiological conditions suggesting cachexia with weight reduction and weakness. Mice were imaged using an in vivo imaging instrument at four weeks to detect GFP + cells in organs. After anesthetization, mice were fixed in position on their backs (PerkinElmer Ivis Spectrum Imaging System). After five weeks, all mice were euthanized. Tissues were removed and fixed in 10% formalin for H&E staining, and metastasized colonies were counted using a dissecting microscope.

Statistical analysis
The χ 2 test was performed to compare groups and to evaluate histological and clinical parameter correlations (age, gender, infiltration depth of tumor, lymph nodes metastasis, and degree of differentiation). For in vitro experiments, all data were shown as means ± standard deviation (S.D.) of 3 independent experiments, and were analyzed using an unpaired Student's t-test. In animal experiments, Kaplan-Meier analysis was used to estimate survival and Gehan-Breslow-Wilcoxon test was used assess differences between the two groups. All statistical analyses were performed using GraphPad Prism 6.0 software (GraphPad Software Inc., San Diego, USA). P < 0.05 was considered significant.