B7-H3 expression in gastric cancer: A novel molecular

blood marker for detecting circulating tumor cells

Takaaki Arigami,1 Yoshikazu Uenosono, Munetsugu Hirata, Shigehiro Yanagita, Sumiya Ishigami and
Shoji Natsugoe

Department of Surgical Oncology and Digestive Surgery, Field of Oncology, Course of Advanced Therapeutics, Kagoshima University Graduate School of
Medical and Dental Sciences, Kagoshima, Japan

(Received October 2, 2010 ⁄ Revised January 11, 2011 ⁄ Accepted January 12, 2011 ⁄ Accepted manuscript online January 19, 2011 ⁄ Article first published online February 17, 2011)

The clinical significance of B7-H3 expression in gastric cancer B7-H3 was recently identified as a new member of the B7
remains unclear, although the B7 ligand family plays a critical role family, the family of which plays a principal role in the T cell-
in the T cell-mediated immune response. We therefore investi- mediated immune response.(22) Immunohistochemistry has con-
gated B7-H3 expression as a blood marker of circulating tumor firmed B7-H3 protein expression in prostate, non-small cell
cells and determined correlations with tumor progression in lung, gastric and clear cell renal cell cancers.(23–26) B7-H3
patients with gastric cancer. B7-H3 expression in gastric cell lines expression in tumor cells is closely correlated with tumor pro-
was initially evaluated by immunocytochemistry. Furthermore, we gression in these malignancies.(23,25,26) However, little is under-
used quantitative RT-PCR to assess B7-H3 mRNA expression in four stood about the clinical significance of B7-H3 mRNA
cell lines and in 95 blood specimens from patients with gastric can- expression in blood specimens from patients with gastric cancer.
cer, as well as in 21 samples of peripheral blood lymphocytes from The present study investigates relationships between B7-H3
healthy volunteers. B7-H3 expression in cell lines was identified by expression in blood specimens and tumor progression, as well as
immunocytochemistry and quantitative RT-PCR. Blood specimens the prognosis of patients with gastric cancer.
from patients with gastric cancer contained significantly more cop-
ies of B7-H3 mRNA than those from healthy volunteers without Materials and Methods
cancer (P < 0.0001). Levels of B7-H3 expression significantly corre-
lated with overall stage (P = 0.013). The 5-year survival rate was Gastric cancer cell lines. The gastric cancer cell lines, MKN-
significantly lower in patients with high B7-H3 expression than 7, MKN-45, MKN-74, and NUGC-4, were cultured in RPMI
with low expression (P = 0.02). Multivariate analysis demonstrated 1640 (Nissui Pharmaceutical Co. Ltd., Tokyo, Japan) supple-
that B7-H3 expression was an independent prognostic factor mented with 10% fetal calf serum (Mitsubishi Kasei, Tokyo,
(P = 0.046). Our results indicate that B7-H3 appears to be a useful Japan), 100 units ⁄ mL penicillin and 100 units ⁄ mL streptomy-
blood marker for predicting tumor progression in gastric cancer. cin. All cell lines were incubated at 37C in a humidified atmo-
(Cancer Sci 2011; 102: 1019–1024) sphere containing 5% CO2, as previously described(27,28) and
then examined by immunocytochemistry and RT-PCR.
Patients. Blood specimens were collected preoperatively

G astric cancer is one of the most common gastrointestinal

tract malignancies in Asia.(1–4) New anticancer agents
such as S-1, taxanes, capecitabine, irinotecan and oxaliplatin
from 95 patients (64 men and 31 women; age range, 35–
87 years; average, 68 years) with gastric cancer who underwent
curative gastrectomy with lymphadenectomy at Kagoshima Uni-
have improved the prognosis of patients with unresectable versity Hospital (Kagoshima, Japan) between 2003 and 2005.
advanced gastric cancer.(5–9) Nevertheless, the 5-year survival Patients who had undergone endoscopic mucosal resection, pal-
rates of patients with International Union Against Cancer liative resection, preoperative chemotherapy and ⁄ or radiation
(UICC) stage IIIA, IIIB, and IV gastric cancers are 30.8–54.0%, therapy were excluded from this study. Furthermore, none of
16.1–36.5% and 9.2–23.9%, respectively.(10,11) Carcinoembry- patients enrolled in this study had synchronous or metachronous
onic antigen (CEA) and carbohydrate antigen 19-9 (CA19-9) are cancer in other organs. Postoperative complications occurred in
commonly used as established tumor markers in blood for 18 patients (18.9%). Within the 18 patients with complications,
patients with gastric cancer. However, the sensitivity and speci- two patients required surgical intervention. These complications
ficity of predicting tumor progression and postoperative recur- consisted of anastomotic leakage in three patients, pancreatic fis-
rence are clinically insufficient and no molecular biomarkers tula in three patients, bowel obstruction in four patients, wound
can yet predict the prognosis of patients with gastric can- infection in eight patients, cholecystitis in two patients, and
cer.(12,13) pneumonia in four patients. None of patients enrolled in this
The RT-PCR is a useful assay for detecting circulating tumor study were died in-hospital 30 days after surgery due to postop-
cells (CTC) within blood specimens from patients with various erative complications. Blood specimens were classified and
malignancies, including gastric cancer.(14–18) We reported that staged based on the UICC criteria of tumor-node-metastasis
the presence of CTC correlates with tumor progression and out- classification for gastric carcinoma.(29) Peripheral blood lym-
comes of patients with esophageal, gastric and biliary-pancreatic phocytes (PBLs) from 21 healthy volunteers without cancers
cancer.(14–18) Several epithelial-specific antigens such as CEA, served as the control group. All patients were followed up every
cytokeratin-19, and cytokeratin-20, have been used generally in 3–6 months by regular clinical diagnostic examinations, such as
many studies of CTC detection in gastric cancer.(15,19) However, tumor marker studies (CEA and CA19-9), radiography, ultra-
recent studies have found false-positive results in RT-PCR sonography and computed tomography at Kagoshima University
assays using these markers for CTC detection.(20,21) Conse- Hospital. The median follow-up period was 24 months (range,
quently, RT-PCR assays using conventional CTC markers have 1–74 months).
low specificity for detecting occult CTC from blood specimens.
Furthermore, few candidate molecular blood markers of gastric 1To whom correspondence should be addressed.
cancer are presently available. E-mail: [email protected]

doi: 10.1111/j.1349-7006.2011.01877.x Cancer Sci | May 2011 | vol. 102 | no. 5 | 1019–1024
ª 2011 Japanese Cancer Association
 To assess B7-H3 protein expression in gastric cancer, 10 par- positive (gastric cancer cell line), negative (H2O) and reagent
affin-embedded archival tissue (PEAT) specimens of primary (without DNA) controls to confirm the qRT-PCR assays. Abso-
gastric tumors from patients enrolled in this study were used for lute copy numbers were calculated in qRT-PCR assays based on
immunohistochemical analysis. All patients provided written, standard curves with six serial dilutions of plasmid templates.
informed consent to specimen collection in accordance with our B7-H3 mRNA copy numbers were normalized by GAPDH
institutional guidelines. mRNA copy numbers (relative B7-H3 mRNA copies: absolute
Immunocytochemistry. Gastric cancer cells were cultured on B7-H3 mRNA copies ⁄ absolute GAPDH mRNA copies).
Lab-Tek II chamber slides (Nalge Nunc International Corp., Immunohistochemistry. The PEAT sections (3 lm thick) with
Naperville, IL, USA), washed with PBS and then fixed with 4% surgical primary gastric tumors were incubated on slides at
paraformaldehyde in PBS for 10 min. The cells were washed in 50C overnight, deparaffinized with xylene, and then rehydrated
PBS again and then stained using an antihuman B7-H3 mono- with a graded series of ethanol. The sections were heated in
clonal antibody (mAb) (NSO-185504; R&D Systems Inc., Min- EDTA buffer (1 mM, pH 8.0) at 121C for 10 min to activate
neapolis, MN, USA) and a PE-conjugated goat antimouse the antigen. After cooling at room temperature, endogenous
secondary antibody (Santa Cruz Biotechnology Inc., Santa Cruz, peroxidase was blocked by Peroxidase Blocking Reagent
CA, USA) at room temperature for 1 h. Slides were mounted (DakoCytomation, Carpinteria, CA, USA) for 10 min. Non-
with Vectashield‘ Mounting Medium containing DAPI (Vector specific binding was blocked at room temperature for 30 min
Laboratories, Burlingame, CA, USA) for nuclear staining and with Protein Block Serum-Free (DakoCytomation). The sections
then cells were analyzed using a fluorescence microscope with were incubated at 4C overnight with an antihuman B7-H3 mAb
Olympus UIS2 (Olympus, Tokyo, Japan). (R&D Systems) diluted 1:50 in PBS. After three 5-min
Determination of CEA and CA19-9 in serum samples. Serum washes in PBS, the reaction for anti-B7-H3 mAb was developed
samples were used for assay of tumor markers CEA and CA19- using the ABC method (Vectastain ABC kit; Vector Laborato-
9. Serum levels of CEA and CA19-9 were determined with a ries) and visualized using diaminobenzidine tetrahydrochloride
commercial enzyme immunoassay kit (Abbott Co., Ltd, Tokyo, (DakoCytomation). Negative controls were treated with PBS
Japan). A serum CEA level of > 5 ng ⁄ mL and a serum CA19-9 without primary antibody under the same conditions. Based
level of >37 ng ⁄ mL were judged as positive. on the immunostainable intensity, B7-H3 immunoreactivity
Blood processing and RNA extraction. Blood (5 mL) was pre- was classified into three groups: negative, weak and strong
operatively withdrawn from each patient into tubes containing immunoreactions.
sodium citrate and these blood specimens were processed within Statistical analysis. Differences in B7-H3 mRNA expression
1 h after drawing. Then, blood cells were collected using lym- between gastric cancer cell lines and normal PBLs from healthy
phocyte separation buffer (Gentra Systems, Inc., Minneapolis, volunteers, and between PBLs from patients with gastric cancer
MN, USA). All blood and cell lines mixed with 1 mL of Isogen and from healthy volunteers, were assessed using the Wilcoxon
(Nippon Gene, Toyama, Japan) were immediately cryopreserved rank sum test. Receiver operating characteristic (ROC) curves
at )80C until RNA extraction. Total RNA was isolated and and the area under the curve were constructed to assess the pre-
purified using phenol-chloroform as previously described.(27,28) dictive utility of B7-H3 mRNA expression for discriminating
The concentration and purity of the total RNA were determined patients with gastric cancer from normal healthy volunteers. To
using a GeneQuant pro UV ⁄ Vis spectrophotometer (Amersham compare categorical clinicopathological factors, group differ-
Pharmacia Biotech, Cambridge, UK). ences were statistically analyzed based on B7-H3 expression
Primers and probes. Primer and probe sequences of B7-H3 using the chi-squared and Fisher’s exact tests. Differences in
and GAPDH were designed to assess the mRNA expression of survival analyzed using the Kaplan–Meier method were exam-
each marker for RT-PCR assays in blood specimens of patients ined by the log-rank test. Prognostic factors were assessed by
with gastric cancer. The forward primers, fluorescence reso- univariate and multivariate analyses (Cox proportional hazard
nance energy transfer probe sequence and reverse primers for regression model). All statistical calculations were performed
B7-H3 and GAPDH were as follows: B7-H3 (forward), 5¢-GAC- using SAS statistical software (SAS Institute, Inc., Cary, NC,
AGCAAAGAAGATGATGGA-3¢; (probe), 5¢-FAM-CCTCCC- USA). A P-value of <0.05 was considered statistically signifi-
TACAGCTCCTACCCTCTGG-TAMRA-1-3¢; (reverse), 5¢-AC- cant.
CTGTCAGAGCAGGATGC-3¢; GAPDH (forward), 5¢-GGG-
TGTGAACCATGAGAAGT-3¢; (probe), 5¢-FAM-CAGCAAT- Results
GCCTCCTGCACCACCAA-TAMRA-1-3¢; (reverse), 5¢-GAC-
TGTGGTCATGAGTCCT-3¢. The integrity of the RNA was Immunocytochemical analysis of B7-H3 expression in cell
confirmed by RT-PCR assays using GAPDH. lines. Morphological B7-H3 protein expression in the gastric
Quantitative RT-PCR. All total RNA samples were reverse cancer cell lines MKN-7 and MKN-45 was immunocytochemi-
transcribed using the Advantage RT-for-PCR kit (Clontech cally assessed. The surface of gastric cancer cell lines expressed
Laboratories, Inc., Palo Alto, CA, USA) as previously high levels of B7-H3 (Fig. 1).
described.(27,28) Samples were analyzed by quantitative RT-PCR B7-H3 mRNA expression in cell lines and clinical blood
(qRT-PCR) using the LightCycler System (Roche Diagnostics, specimens. B7-H3 mRNA expression in four gastric cancer cell
Mannheim, Germany). Reaction mixtures contained cDNA from lines, 95 blood specimens from patients with gastric cancer and
250 ng of RNA, each primer, probe, MgCl2 and LightCycler in 21 normal PBLs from healthy volunteers was assessed by
FastStart DNA Master hybridization probes (Roche). The ampli- qRT-PCR.
fication profile consisted of a precycling hold at 95C for The relative numbers of B7-H3 mRNA copies ranged from
10 min, followed by 45 cycles of denaturation at 95C for 10 s, 9.90 · 10)5 to 1.01 · 10)3 in gastric cancer cell lines, from
annealing for 20 s (63C for B7-H3, and 55C for GAPDH), and 3.09 · 10)7 to 5.68 · 10)5 in blood specimens from patients
extension at 72C for 10 s. Plasmids for each marker were syn- with gastric cancer, and from 2.05 · 10)7 to 8.83 · 10)6 in
thesized using pT7Blue-2 T-Vector (Novagen, Madison, WI, PBLs from healthy volunteers. The mean numbers of relative
USA) according to the manufacturer’s instructions. Standard B7-H3 mRNA copies (±SD) were 3.78 · 10)4 ± 4.25 · 10)4,
curves for each assay were generated using the threshold cycles 1.00 · 10)5 ± 9.00 · 10)6 and 2.00 · 10)6 ± 1.00 · 10)6, res-
of six serial dilutions of plasmid templates (106–101 copies). pectively (Fig. 2). Consequently, the relative numbers of B7-H3
The mRNA copy number was calculated by the LightCycler mRNA copies were significantly higher in gastric cancer cell
software (Roche). Each assay was repeated in duplicate with lines and in blood specimens from patients with gastric cancer

1020 doi: 10.1111/j.1349-7006.2011.01877.x

ª 2011 Japanese Cancer Association
 MKN-7 MKN-45
(a) (e)

Bright field

(b) (f)

DAPI

Fig. 2. RT-PCR analysis for B7-H3 mRNA expression in blood

specimens. B7-H3 mRNA expression in cell lines and in clinical blood
(c) (g) specimens from patients with gastric cancer and healthy volunteers.
Horizontal bars indicate mean B7-H3 mRNA copy number.

B7-H3

(d) (h)

Merged

Fig. 1. Immunocytochemical staining of B7-H3 expression in gastric

cancer cell lines. Bright field (a,e). DAPI staining (b,f). Anti-B7-H3
mAb-phycoerythrin staining (c,g). Merged staining of B7-H3-PE and
DAPI staining (d,h). B7-H3 is expressed on the cell surface of MKN-7
and MKN-45. Scale bars indicate 100 lm (original magnification ·400).

Fig. 3. Receiver operating characteristic curve for discriminating

patients with gastric cancer from healthy volunteers according to B7-
than in PBLs from healthy volunteers (P < 0.0001 and H3 mRNA expression. Area under the curve (AUC) was 0.86.
P < 0.0001, respectively). The area under the curve value of the
ROC curve of B7-H3 mRNA copies was 0.86 (Fig. 3). The sen-
sitivity was 81% when the specificity was set to 90%. between B7-H3 mRNA expression and the number of CTC in
B7-H3 mRNA expression in cell spiking study. To assess the clinical blood specimens.
relationship between B7-H3 mRNA expression and the number B7-H3 expression and clinicopathological factors. Based on
of tumor cells in blood, we did quantitative RT-PCR on serial the median value of the relative number of B7-H3 mRNA cop-
diluted gastric tumor cells mixed with peripheral blood cells ies, all patients were divided into groups with high (n = 48) and
from healthy donors. Serial dilutions of MKN-74 tumor cells low expression (n = 47) to assess the relationship between
(104, 103, 102, 101, and 0 cells) were mixed with 107 donor- B7-H3 expression and clinicopathological factors. This binary
derived peripheral blood cells and assayed for B7-H3 expression system in B7-H3 mRNA expression was used for statistical anal-
by quantitative RT-PCR assay. This assay was done three times yses in this study. B7-H3 expression significantly correlated
to validate its reproducibility. This in vitro model system was with UICC stage (P = 0.013) (Table 1), but not with any other
used to mimic the detection of CTC in clinical blood specimens. clinicopathological factors.
Relative B7-H3 mRNA copies gradually decreased as the gas- B7-H3 expression and prognosis. The 5-year survival rate was
tric tumor cells in normal peripheral blood cells decreased significantly lower in patients with high B7-H3 expression than
(Fig. 4). This result may indicate the positive correlation with low expression (57.1% vs 76.4%, P = 0.02) (Fig. 5).

Arigami et al. Cancer Sci | May 2011 | vol. 102 | no. 5 | 1021
ª 2011 Japanese Cancer Association
Fig. 4. RT-PCR analysis for B7-H3 mRNA expression in cell spiking
assay. Serially diluted MKN-74 tumor cells (104, 103, 102, 101, and 0) Fig. 5. Kaplan–Meier survival curves for patients with gastric cancer
were mixed with 107 normal peripheral blood cells. Bars indicate based on status of B7-H3 expression. All patients were divided into
standard deviation (SD). groups with high (n = 48) and low expression (n = 47) based on the
median value of B7-H3 mRNA copies. Patients with B7-H3 high
expression had a significantly poorer prognosis than those with B7-H3
Table 1. Relationship between B7-H3 expression and clinicopatho- low expression (P = 0.02).
logical factors in patients with gastric cancer

B7-H3 expression (%) Univariate and multivariate analyses of survival. Univariate

Factors P-value
Low (n = 47) High (n = 48) analysis showed that the depth of tumor invasion, lymph node
metastasis and B7-H3 expression were significantly related to
Gender postoperative survival (P = 0.049, P = 0.002, and P = 0.023,
Male 30 (63.8) 34 (70.8) 0.516 respectively) (Table 2). Multivariate regression analysis demon-
Female 17 (36.2) 14 (29.2) strated that lymph node metastasis and B7-H3 expression were
Age (years) independent prognostic factors (P = 0.014 and P = 0.046,
£ 70 24 (51.1) 24 (50.0) 1.000 respectively) (Table 2).
> 70 23 (48.9) 24 (50.0) B7-H3 protein expression in primary tumor specimens. To
Tumor location confirm B7-H3 protein expression in gastric tumors, immunohis-
Upper 15 (31.9) 15 (31.3) 0.989 tochemical analysis was performed on 10 surgical PEAT speci-
Middle 15 (31.9) 16 (33.3) mens of primary gastric tumors. According to the results of
Lower 17 (36.2) 17 (35.4) immunohistochemistry, gastric tumors had a variety of immuno-
Histological type reactivity for B7-H3 and its protein expression was observed in
Differentiated 24 (51.1) 22 (45.8) 0.683 the membrane and ⁄ or cytoplasm of gastric tumor cells (Fig. 6).
Undifferentiated 23 (48.9) 26 (54.2) Although B7-H3 protein expression in primary tumor specimens
Depth of tumor invasion was not significantly correlated with the status of B7-H3 mRNA
pT1-T2 32 (68.1) 26 (54.2) 0.208 expression in blood specimens from same patients, strong
pT3-T4 15 (31.9) 22 (45.8) immunoreaction was identified in 80% (4 ⁄ 5) and 20% (1 ⁄ 5) of
Lymph node metastasis patients with high and low B7-H3 mRNA expression, respec-
Negative 20 (42.5) 16 (33.3) 0.402 tively.
Positive 27 (57.5) 32 (66.7)
Distant metastasis
Discussion
Negative 46 (97.9) 47 (97.9) 1.000
Positive 1 (2.1) 1 (2.1) Here, we assessed B7-H3 mRNA expression by qRT-PCR assay
Stage in blood specimens from patients with gastric cancer and from
I-II 32 (68.1) 20 (41.7) 0.013 healthy volunteers. We also investigated the relationships
III-IV 15 (31.9) 28 (58.3) between B7-H3 expression and tumor progression, and the prog-
Lymphatic invasion nosis of patients with gastric cancer. To our knowledge, the clin-
Negative 14 (29.8) 15 (31.3) 1.000 ical significance of B7-H3 mRNA expression in CTC has not
Positive 33 (70.2) 33 (68.7) been studied.
Venous invasion We initially demonstrated B7-H3 expression in gastric cancer
Negative 15 (31.9) 16 (33.3) 1.000 cell lines and found significantly higher expression in patients
Positive 32 (68.1) 32 (66.7) with gastric cancer than in healthy volunteers. Furthermore, the
Serum CEA levels (>5 ng ⁄ mL) ROC curve of B7-H3 mRNA copies indicated that the qRT-PCR
Negative 36 (76.6) 34 (70.8) 0.642 assays were highly sensitive and specific for B7-H3 expression.
Positive 11 (23.4) 14 (29.2) The immunocytochemical analysis morphologically confirmed
Serum CA 19-9 levels (>37 ng ⁄ mL) B7-H3 protein expression on the surface of gastric cell lines. A
Negative 39 (83.0) 32 (66.7) 0.098 recent report indicates that B7-H3 is a tumor-associated antigen
Positive 8 (17.0) 16 (33.3) that controls tumor cell invasion and migration in melanoma,
pT1, invasion of lamina propria or submucosa; pT2, invasion of breast cancer and osteosarcoma.(30) These findings suggest that
muscularis propria or subserosa; pT3, penetration of serosa without B7-H3 mRNA expression is significantly higher in CTC than in
invading adjacent structures; pT4, invasion of adjacent structures. other cells such as antigen presenting cells and that B7-H3 could

1022 doi: 10.1111/j.1349-7006.2011.01877.x

ª 2011 Japanese Cancer Association
Table 2. Univariate and multivariate analyses of survival among patients with gastric cancer

Univariate analysis Multivariate analysis

Independent factors
Hazard ratio 95% CI P-value Hazard ratio 95% CI P-value

Depth of tumor invasion

pT1-T2 ⁄ pT3-T4 1.53 1.00–2.35 0.049 1.18 0.76–1.86 0.456
Lymph node metastasis
Negative ⁄ positive 2.29 1.34–4.73 0.002 2.05 1.15–4.33 0.014
B7-H3 expression
High ⁄ low 1.66 1.07–2.69 0.023 1.56 1.01–2.54 0.046

CI, confidence interval; pT1, invasion of lamina propria or submucosa; pT2, invasion of muscularis propria or subserosa; pT3, penetration of
serosa without invading adjacent structures; pT4, invasion of adjacent structures.

(a)
Recent studies have demonstrated that the B7-H3 signaling
pathway plays an important role in the T cell-mediated immune
response.(22) However, the biological significance of B7-H3
expression in tumor cells and the immunological role in tumor
progression have not been sufficiently clarified. Furthermore,
B7-H3 is thought to have two opposing properties in T cell-med-
iated immunity.(32) B7-H3 ligand binds to uncharacterized
receptors expressed on activated T cells, and it might act as a
co-inhibitory or co-stimulatory mediator. B7-H3 expression sig-
nificantly correlates with tumor progression in malignancies
such as neuroblastoma, non-small cell lung, clear cell renal cell
and prostate cancers.(23,25,26,31) In the present study, patients
with deeper tumor invasion were more likely to be high B7-H3
expression (45.8 vs 31.9%), although this difference was not sta-
tistically significant. Similarly, the status of B7-H3 expression
in node-positive patients had a tendency to be higher than that in
node-negative patients (66.7% vs 57.5%). Consequently, these
results may lead to a significant correlation between its expres-
sion and UICC stage in statistical analysis (P = 0.013). These
findings suggest that B7-H3 expression is related to tumor
(b) aggressiveness in patients with gastric cancer and that it might
serve as a co-inhibitory regulator in T cell-mediated antitumor
immunity. Accordingly, B7-H3 is a novel immune molecular
marker for predicting tumor progression.
Patients with early gastric cancer have a good 5-year progno-
sis.(10,11) Most of them undergo surgical resection with lympha-
denectomy. Nevertheless, some patients die due to recurrent
disease. To predict patients at high risk of disease recurrence
among postoperative patients with early gastric cancer is diffi-
cult. The present study found that the prognosis was poorer
among patients with high, than with low B7-H3 expression. This
may be due to the strong relationship between B7-H3 expression
and UICC stage (P = 0.013). Furthermore, B7-H3 expression
was an independent prognostic factor in multivariate analysis.
This result indicates that qRT-PCR assay for B7-H3 expression
might be a useful tool with which to predict patients with a poor
prognosis and improve the selection of patients for adjuvant sys-
temic chemotherapy in the postoperative management of gastric
cancer.
Fig. 6. Representative immunohistochemical staining of B7-H3 Cytotoxic T lymphocyte antigen-4 expressed on activated T
expression in primary gastric tumor specimens. (a) Tumor cells with cells is a receptor for B7-1 and B7-2 ligand and this signaling
weak expression of B7-H3. (b) Tumor cells with strong expression of pathway has the function of co-inhibition.(32) Several antago-
B7-H3. Scale bars indicate 200 lm (original magnification ·400).
nists have been developed to block cytotoxic T lymphocyte anti-
gen-4 receptors and clinical trials have examined patients with
melanoma, colon, prostate, renal cell and ovarian cancers.(33–36)
be a promising biomarker for detecting CTC. To discover novel B7-H3 expression significantly correlated with tumor progres-
molecular blood markers that can detect subclinical patients and sion in the present study. Moreover, siRNA-mediated knock-
the treatment response of patients to chemotherapy is important down of B7-H3 expression reduces tumor cell adhesion,
for clinical management. B7-H3 is not unique to gastric tumor migration and invasion in vitro.(30) These results indicate that
cells and it is expressed in other malignancies.(23–26,30,31) There- antagonists of the B7-H3 signaling pathway might serve as new
fore, B7-H3 might serve as a tumor marker of various malignan- targets to inhibit tumor progression in various malignancies,
cies. including gastric cancer.

Arigami et al. Cancer Sci | May 2011 | vol. 102 | no. 5 | 1023
ª 2011 Japanese Cancer Association
 In conclusion, we demonstrated that B7-H3 is expressed in Acknowledgment
circulating gastric tumor cells and that such expression is an
independent prognostic factor in patients with gastric cancer. This work was supported in part by grants-in-aid (no. 22791256) for sci-
Although validation large studies are required to strengthen our entific research from the Ministry of Education, Science, Sports, and
findings, the evaluation of B7-H3 expression in blood specimens Culture Technology, Japan.
could be useful as a tool for predicting tumor aggressiveness in
patients with gastric cancer. Further understanding of its immu- Disclosure Statement
nological role might allow the development of a new targeted
immunotherapy that inhibits the B7-H3 signaling pathway in The authors have no conflict of interest.
patients with gastric cancer.

References 19 Koga T, Tokunaga E, Sumiyoshi Y et al. Detection of circulating gastric

cancer cells in peripheral blood using real time quantitative RT-PCR.
1 Davis PA, Sano T. The difference in gastric cancer between Japan, USA and Hepatogastroenterology 2008; 55: 1131–5.
Europe: what are the facts? what are the suggestions? Crit Rev Oncol Hematol 20 Bostick PJ, Chatterjee S, Chi DD et al. Limitations of specific reverse-
2001; 40: 77–94. transcriptase polymerase chain reaction markers in the detection of metastases
2 Dicken BJ, Bigam DL, Cass C, Mackey JR, Joy AA, Hamilton SM. Gastric in the lymph nodes and blood of breast cancer patients. J Clin Oncol 1998; 16:
adenocarcinoma: review and considerations for future directions. Ann Surg 2632–40.
2005; 241: 27–39. 21 Jung R, Petersen K, Krüger W et al. Detection of micrometastasis by
3 Katanoda K, Yako-Suketomo H. Comparison of time trends in stomach cancer cytokeratin 20 RT-PCR is limited due to stable background transcription in
incidence (1973–2002) in Asia, from cancer incidence in five continents, Vol. granulocytes. Br J Cancer 1999; 81: 870–3.
IV–IX. Jpn J Clin Oncol 2009; 39: 71–2. 22 Chapoval AI, Ni J, Lau JS et al. B7-H3: a costimulatory molecule for T cell
4 Li ZX, Kaminishi M. A comparison of gastric cancer between Japan and activation and IFN-gamma production. Nat Immunol 2001; 2: 269–74.
China. Gastric Cancer 2009; 12: 52–3. 23 Sun Y, Wang Y, Zhao J et al. B7-H3 and B7-H4 expression in non-small-cell
5 Hirata K, Horikoshi N, Aiba K et al. Pharmacokinetic study of S-1, a novel lung cancer. Lung Cancer 2006; 53: 143–51.
oral fluorouracil antitumor drug. Clin Cancer Res 1999; 5: 2000–5. 24 Wu CP, Jiang JT, Tan M et al. Relationship between co-stimulatory molecule
6 Ajani JA, Faust J, Ikeda K et al. Phase I pharmacokinetic study of S-1 plus B7-H3 expression and gastric carcinoma histology and prognosis. World J
cisplatin in patients with advanced gastric carcinoma. J Clin Oncol 2005; 23: Gastroenterol 2006; 12: 457–9.
6957–65. 25 Roth TJ, Sheinin Y, Lohse CM et al. B7-H3 ligand expression by prostate
7 Ohtsu A. Current status and future prospects of chemotherapy for metastatic cancer: a novel marker of prognosis and potential target for therapy. Cancer
gastric cancer: a review. Gastric Cancer 2005; 8: 95–102. Res 2007; 67: 7893–900.
8 Ohtsu A, Yoshida S, Saijo N. Disparities in gastric cancer chemotherapy 26 Crispen PL, Sheinin Y, Roth TJ et al. Tumor cell and tumor vasculature
between the East and West. J Clin Oncol 2006; 24: 2188–96. expression of B7-H3 predict survival in clear cell renal cell carcinoma. Clin
9 Koizumi W, Narahara H, Hara T et al. S-1 plus cisplatin versus S-1 alone for Cancer Res 2008; 14: 5150–7.
first-line treatment of advanced gastric cancer (SPIRITS trial): a phase III 27 Arigami T, Natsugoe S, Uenosono Y et al. Lymphatic invasion using D2-40
trial. Lancet Oncol 2008; 9: 215–21. monoclonal antibody and its relationship to lymph node micrometastasis in
10 Park JM, Kim YH. Current approaches to gastric cancer in Korea. pN0 gastric cancer. Br J Cancer 2005; 93: 688–93.
Gastrointest Cancer Res 2008; 2: 137–44. 28 Arigami T, Natsugoe S, Uenosono Y et al. Evaluation of sentinel node
11 Yamashita K, Sakuramoto S, Kikuchi S, Katada N, Kobayashi N, Watanabe concept in gastric cancer based on lymph node micrometastasis determined by
M. Validation of staging systems for gastric cancer. Gastric Cancer 2008; 11: reverse transcription-polymerase chain reaction. Ann Surg 2006; 243: 341–7.
111–8. 29 Sobin LH, Wittenkind CH. International Union Against Cancer. TNM
12 Webb A, Scott-Mackie P, Cunningham D et al. The prognostic value of serum Classification of Malignant Tumors, 6th edn. New York: John Wiley-Liss,
and immunohistochemical tumour markers in advanced gastric cancer. Eur J 2002.
Cancer 1996; 32: 63–8. 30 Chen YW, Tekle C, Fodstad O. The immunoregulatory protein human B7H3
13 Gaspar MJ, Arribas I, Coca MC, Diez-Alonso M. Prognostic value of is a tumor-associated antigen that regulates tumor cell migration and invasion.
carcinoembryonic antigen, CA 19-9 and CA 72-4 in gastric carcinoma. Curr Cancer Drug Targets 2008; 8: 404–13.
Tumour Biol 2001; 22: 318–22. 31 Castriconi R, Dondero A, Augugliaro R et al. Identification of 4Ig-B7-H3 as a
14 Miyazono F, Takao S, Natsugoe S et al. Molecular detection of circulating neuroblastoma-associated molecule that exerts a protective role from an NK
cancer cells during surgery in patients with biliary-pancreatic cancer. Am J cell-mediated lysis. Proc Natl Acad Sci USA 2004; 101: 12640–5.
Surg 1999; 177: 475–9. 32 Zang X, Allison JP. The B7 family and cancer therapy: costimulation and
15 Miyazono F, Natsugoe S, Takao S et al. Surgical maneuvers enhance coinhibition. Clin Cancer Res 2007; 13: 5271–9.
molecular detection of circulating tumor cells during gastric cancer surgery. 33 Hodi FS, Mihm MC, Soiffer RJ et al. Biologic activity of cytotoxic T
Ann Surg 2001; 233: 189–94. lymphocyte-associated antigen 4 antibody blockade in previously vaccinated
16 Uchikura K, Takao S, Nakajo A et al. Intraoperative molecular detection of metastatic melanoma and ovarian carcinoma patients. Proc Natl Acad Sci USA
circulating tumor cells by reverse transcription-polymerase chain reaction in 2003; 100: 4712–7.
patients with biliary-pancreatic cancer is associated with hematogenous 34 Sanderson K, Scotland R, Lee P et al. Autoimmunity in a phase I trial of a
metastasis. Ann Surg Oncol 2002; 9: 364–70. fully human anti-cytotoxic T-lymphocyte antigen-4 monoclonal antibody with
17 Nakashima S, Natsugoe S, Matsumoto M et al. Clinical significance of multiple melanoma peptides and Montanide ISA 51 for patients with resected
circulating tumor cells in blood by molecular detection and tumor markers in stages III and IV melanoma. J Clin Oncol 2005; 23: 741–50.
esophageal cancer. Surgery 2003; 133: 162–9. 35 Beck KE, Blansfield JA, Tran KQ et al. Enterocolitis in patients with cancer
18 Mataki Y, Takao S, Maemura K et al. Carcinoembryonic antigen messenger after antibody blockade of cytotoxic T-lymphocyte-associated antigen 4.
RNA expression using nested reverse transcription-PCR in the peripheral J Clin Oncol 2006; 24: 2283–9.
blood during follow-up period of patients who underwent curative surgery for 36 O’Mahony D, Morris JC, Quinn C et al. A pilot study of CTLA-4 blockade
biliary-pancreatic cancer: longitudinal analyses. Clin Cancer Res 2004; 10: after cancer vaccine failure in patients with advanced malignancy. Clin
3807–14. Cancer Res 2007; 13: 958–64.

1024 doi: 10.1111/j.1349-7006.2011.01877.x

ª 2011 Japanese Cancer Association