Human Cancer Biology
Expression of Trefoil Factor Family Members Correlates
with Patient Prognosis and Neoangiogenesis
Dipok Kumar Dhar,1,4 Timothy C. Wang,5 Hideki Tabara,1 Yasuhito Tonomoto,1 Riruke Maruyama,2
Mitsuo Tachibana,1 Hirofumi Kubota,3 and Naofumi Nagasue1
Abstract
Purpose:Trefoil factor family (TFF) peptides are thought to contribute to epithelial protection and
restitution by virtue of their protease-resistant nature and their strong affinity for mucins. However,
they are often overexpressed in tumors, where they seem to be negative prognostic factors, possibly contributing to tumor spread, although the precise mechanisms have not been defined.
Experimental Design:Tissue sections from 111patients with curatively resected advanced gastric carcinoma were immunohistochemically stained forTFF2, ITF (TFF3), and CD34. Microvessel
density was expressed as number and area of microvessels. Results were correlated with clinicopathological characteristics and patient survival.
Results: Forty-nine (44.1%) and 41 (36.9%) tumors were immunohistochemically positive for
TFF3 and TFF2, respectively. Among the various clinicopathologic variables, overexpression of
TFF3 had a significant correlation with patient age only. In addition, a significantly higher prevalence of positive TFF2 staining was detected in large, diffuse tumors and in tumors with lymph
node metastasis. The number of microvessels had a significant correlation with both TFF3 and
TFF2 staining, whereas the area of microvessels had a significant correlation only with TFF3
staining. Both TFF3 and TFF2 were independent predictors of a worse disease-free survival.
TFF3 had a gender-specific negative survival advantage, with a 91.3% disease-free survival in
female patients with TFF3-negative advanced gastric carcinoma.
Conclusions: Induction of increased tumor vascularity might be one of the mechanisms by which
TFFs confer metastatic phenotype and frequent disease recurrence in gastric carcinomas. Female
patients with TFF3-negative advanced gastric carcinoma have comparable survival as that
reported for patients with early gastric carcinoma.
It is well established that the integrity of the gastrointestinal
mucosa and other epithelium are maintained by a number of
secreted factors, but recent studies suggest that one important
mechanism involves the secretion of the members of a
protease-resistant protein family known as the trefoil factor
family (TFF; ref. 1). Trefoil proteins are characterized by the
presence of at least one 40 – amino acid protein domain with
three conserved disulfide bonds, a domain that has been
labeled as the trefoil motif. TFFs are typically secreted with
mucus and as such contribute to the barrier film that often
coats surface epithelial cells. In addition, TFFs are often rapidly
up-regulated in the setting of mucosal damage or injury, where
Authors’Affiliations: Departments of 1Digestive and General Surgery and 2Central
Pathological Laboratory, Faculty of Medicine; Shimane University, 3Department of
Surgery, Izumo Citizen Hospital, Izumo; 4Naze Tokushukai Hospital, Kagoshima,
Japan; and 5Division of Digestive and Liver Diseases, Columbia University Medical
Center, New York, New York
Received 3/29/05; revised 6/15/05; accepted 6/23/05.
The costs of publication of this article were defrayed in part by the payment of page
charges. This article must therefore be hereby marked advertisement in accordance
with 18 U.S.C. Section 1734 solely to indicate this fact.
Requests for reprints: Mitsuo Tachibana, Department of Digestive and General
Surgery, Faculty of Medicine, Shimane University, Izumo 693-8501, Japan. Phone:
81-853-202235; Fax: 81-853-202229; E-mail: nigeka35@ shimane-med.ac.jp.
F 2005 American Association for Cancer Research.
doi:10.1158/1078-0432.CCR-05-0671
Clin Cancer Res 2005;11(18) September 15, 2005
they contribute to epithelial restitution (2). Whereas the
healing properties of TFFs were initially attributed to
the effects of cellular proliferation, it was later established that
the effect was primarily due to promotion of cellular motility/
migration and possibly to inhibition of apoptosis (3 – 5). In
any case, the luminal activity and stability of trefoil peptides
make these proteins attractive candidates for the treatment of
gastrointestinal disease, such as inflammatory bowel disease
and radiochemotherapy-induced intestinal damage (6 – 8).
In addition to a role in restitution and healing, TFFs have also
been linked to the development of neoplasia. The role of TFFs in
tumorigenesis remains elusive; in general, TFF1 seems to
function as a gastric-specific tumor suppressor gene, whereas
TFF2 and TFF3 have been thought to augment tumor progression
by increasing cellular invasion and metastasis (9, 10). TFF2 is
frequently up-regulated in gastric preneoplastic lesions (11) and
microarray analysis of gastric cancers has revealed alterations in
TFF gene expression (12 – 14). Although the mechanisms are not
understood, expression of TFFs is often dysregulated in cells
during the progression to invasive cancer (15 – 17). Recently, we
showed that the overexpression of TFF2 is common in gastric
carcinoma and is often associated with cancer invasion,
metastasis, and a poor prognosis (18). Similarly, Yamachika
et al. (17) reported that patients with TFF3/ITF-positive gastric
carcinomas also show invasive characteristics and a poor
prognosis. To date, a number of mecha-nisms have been
6472
www.aacrjournals.org
Downloaded from clincancerres.aacrjournals.org on June 18, 2017. © 2005 American Association for Cancer
Research.
Trefoil Factor Family in Angiogenesis
proposed to account for the invasive nature of TFF-expressing
tumors, including the activation of several growth-promoting
genes and the inhibition of apoptosis-related pathways. In an
earlier study, we speculated that the invasive phenotype of
TFF2-expressing tumors might be related to the induction of the
neovascularization (18). Interestingly, Rodrigues et al. (19)
recently provided concrete evidence for the induction of
angiogenesis by the TFFs in an in vitro study. However, no
clinical correlation has, as yet, been shown between the
overexpression of TFFs and angiogenesis. Therefore, in this
study, we investigated the possibility of a relationship between
TFF2 and TFF3 expression and angiogenesis in primary gastric
carcinoma specimens.
Materials and Methods
Patients and samples. Between February 1990 and December 1998,
111 patients with advanced gastric cancer undergoing curative resection
were included in this study. Of these patients, 39 were female and 72
were male; their ages ranged from 32 to 94 years. Routinely, all tumors
and lymph nodes were cut at several levels, embedded in paraffin, and
sections were taken for routine H&E staining. A group of experienced
pathologists in the Central Pathological Laboratory checked all of these
slides and documented the pathologic characteristics of the tumor and
lymph nodes. Tumor stage was defined according to the tumor, node,
and metastasis (TNM 2002) staging system. Histologic classification
was done according to the Lauren classification. The follow-up was
complete in all patients. The data obtained at regular follow-up visits at
the outpatient department were prospectively stored in a database
specially designed for gastric carcinoma patients. An update inquiry
about the present status of all surviving patients was made by telephone
call in March 2004. In case of any recurrence, the exact date of disease
recurrence was collected from the referring physician or from the
physician who attended the patient the first time for the diagnosis of
the disease recurrence. All patients provided consent for use of tumor
tissue for clinical research and the University Ethical Committee
approved the research protocol.
Immunohistochemistry. All sections used in this study contained
grossly normal-appearing, noncancerous gastric mucosa (in most
cases, with intestinal metaplasia or spasmolytic protein-expressing
metaplasia) and tumor tissue. For each patient, we selected one
representative paraffin block, which contained mostly tumor tissue
and a small amount of adjacent normal gastric mucosa. The normal
mucosa areas were always adjacent to the tumor tissue, and were
approximately within 0.5 to 1.5 cm of the invasive front. Briefly,
slides were deparaffinized, rehydrated in graded alcohols, and placed
in PBS solution. Antigen retrieval was done by autoclaving the slides
for 10 minutes in citrate buffer (pH 6.0; Dako ChemMate, Dako,
Denmark) for TFF2 and TFF3 staining. Endogenous peroxidase activity
was quenched by dipping in 3% H2O2 for 30 minutes. Nonspecific
binding was blocked by treating slides with normal serum from the
species in which secondary antibody was developed for 30 minutes.
Slides were treated with mouse anti-human TFF2 monoclonal
antibody (GE16C clone, Novocastra Laboratories Ltd., Newcastle,
United Kingdom, 1:50 dilution), Rabbit polyclonal anti-TFF3 antibody (HM89, a generous gift from Dr. Daniel K. Podolsky,
Massachusetts General Hospital, Boston, MA) overnight at 4jC and
with mouse anti-CD34 antibody for 1 hour at room temperature. The
streptavidin-biotin peroxidase complex method was employed for the
immunohistochemical steps and was done with a commercially
available kit (Histofine SAB-PO kit, Nichirei, Tokyo, Japan) as we
described previously (20). Color development was done with the
peroxidase substrate 3-amino-9-ethylcarbazole. Adjacent normal mucosa served as an internal positive control, and in negative control
slides, the primary antibody was replaced by nonimmune serum of
www.aacrjournals.org
the same species in which the primary antibody was raised.
Counterstaining was done with hematoxylin in TFF2 and TFF3
staining, however, it was omitted in CD34 staining to facilitate
subsequent image analysis.
Evaluation of immunohistochemical staining. Evaluation of all
immunohistochemical staining was done blindly and independently
by two of the authors (D.K. Dhar and Y. Tonomoto). Any discordant
finding between the two observers was settled by using a conference
microscope. The evaluation criteria of TFF2 in gastric carcinoma were
described in detail in our previously published article (20). TFF2 was
stained uniformly in positive cases and there was no staining at all in
negative cases. Tumors having cytoplasmic or both cytoplasmic and
membranous staining were considered to be positive for TFF2. On the
other hand, there were wide variations in intensity and extent of TFF3
staining among the tumors and were evaluated according to a
comprehensive scoring formula described previously (21). Briefly, the
intensity of staining was scored as follows: 1, weak expression; 2,
moderate expression; and 3, strong expression. The extent of staining
was scored as 1, <33% of the tumor cells had positive staining; 2, 33%
to 67% of the tumor cells had positive staining and; 3, >67% tumor
cells had positive staining. The results obtained with the two scales of
staining score were multiplied against each other, yielding a single scale
with scores of 1, 2, 3, 4, 6, and 9. Patients having a TFF3 staining score
of >4 were considered to be positive for TFF3.
Evaluation of angiogenesis. Tumor angiogenesis was evaluated in
two formats, number of microvessels (nMVD) and cross-sectional area
of microvessels (aMVD). In nMVD, the most vascular area of the tumor
(hotspot) was identified under low magnification (40) and later, the
number of microvessels stained by CD34 was counted at higher
magnification in at least five highly vascular fields. Any cluster of
endothelial cells with or without any lumen was considered as a
microvessel. In the aMVD method, the whole tumor was scanned at low
(40) magnification under a light microscope and five of the most
vascular areas were selected (200). Digital images were captured by
using a charge-coupled device video camera (HC-300/OL, Olympus,
Tokyo, Japan) and a color image freezer computer software (Photograb300 SH-3, Fuji Film, Tokyo, Japan), and stored on optical discs.
Quantification of the CD34-stained area was done by a public domain
NIH Image program (developed at the U.S. NIH and available on the
Internet at http://rsb.info.nih.gov/nih-image/) on a Macintosh computer. Low frequency background was removed with the ‘‘background
subtraction’’ tool and the area of the stained vessels was measured as
the number of pixels per unit area of the tumor (2.8 million pixels).
Statistical analysis. The correlation between TFF expression and
various clinicopathologic variables was determined by the Fisher’s exact
test or Mann-Whitney U test as appropriate. The correlation between
two continuous variables was examined by Sperman’s rank test. The
survival curves were plotted using the Kaplan-Meier method and the
statistical significance between groups was determined by the log rank
test. The end points for analysis were the disease-free survival and the
overall survival starting from the day of operation. Independent
variables predicting survival were evaluated by the multivariable
analysis using the Cox proportional hazard model. The Statview 4.5J
(Abacus Concepts, Berkley, CA) software was used for data analysis.
Results
TFF3 and TFF2 immunohistochemistry. There was no
moderate and marked background staining for TFF3, however,
consistent staining was noticed in the smooth muscle layer of
the stomach. Intestinal metaplastic epithelium adjacent to the
tumor was always positive for TFF3, and occasionally, mucosal
cells in the lower half of the fundic glands were positive for
TFF3. Tumor staining was heterogeneous and staining of TFF3
was mostly noticed in the cytoplasm (Fig. 1A) and occasionally
in the cell membrane and nucleus (Fig. 1B). In two cases, tumor
6473
Clin Cancer Res 2005;11(18) September 15, 2005
Downloaded from clincancerres.aacrjournals.org on June 18, 2017. © 2005 American Association for Cancer
Research.
Human Cancer Biology
Fig. 1. Representative sections of gastric
carcinomas immunostained forTFF3. A, strong
cytoplasmic staining is observed in the advancing
border of the tumor (40). B, occasionally,
strong staining in both cytoplasmic and nuclear
compartments was observed. Also endothelial cells
lining the blood vessels were positively stained for
TFF3 (100). Arrows, nuclear staining. C, metastatic
gastric carcinoma tissue inside the metastatic lymph
nodes and blood vessels adjacent to the primary
tumor was strongly stained forTFF3 (10). Arrows,
vascular invasion by tumor cells. D, infiltrating cells,
most likely macrophages, were strongly stained for
TFF3 (40; inset, 400). Arrows inside the inset,
TFF3-stained interstitial cells.
tissues inside metastatic lymph nodes and blood vessels
adjacent to the primary tumor specimen were strongly positive
for TFF3 (Fig. 1C). Strong staining in the infiltrating cells, most
probably macrophages, was noticed but the lymphoid follicles
and lymphocytes were negative for TFF3 staining (Fig. 1D).
Also, vascular endothelial cells were strongly positive for TFF3
(Fig. 1C).
In contrast to TFF3 staining, intestinal metaplastic mucosa
was always negative for TFF2, whereas consistent staining for
TFF2 was noticed in gastric mucosal cells with abundant
mucous granules and in those having morphologic similarity
with both antral glands and Brunner’s gland of the duodenum
(also known as spasmolytic protein-expressing metaplasia; data
not shown). Weak membranous staining was the characteristic
feature for most of the intestinal types of tumors, whereas
strong cytoplasmic staining was observed in diffuse types of
tumors (Fig. 2B). TFF2 staining was occasionally noticed in
vascular endothelial cells. However, there was no staining
observed in the interstitial infiltrating cells and in the
lymphocytes. Strong immunoreactivity for both TFF3 and
TFF2 was noticed in the vicinity of highly vascularized tumor
areas as shown in Fig. 2. A stronger staining was noticed in the
vascular endothelial cells by anti-TFF3 (Fig. 1C) than by antiTFF2 antibody.
TFF3 and TFF2 reactivity and correlation with clinicopathologic features. Among the 111 patients with advanced gastric
carcinoma, 49 (44.1%) patients were positive for TFF3 and 62
(55.9%) patients were negative. The average age of TFF3positive patients was significantly higher (P = 0.0174, MannWhitney U test) than those with TFF3-negative tumors (mean
F SD, 67.88 F 11.76 versus 63.40 F 10.86, respectively). There
was no significant correlation with gender, tumor location,
differentiation status and T stage. Only a trend of higher
prevalence of TFF3 positivity was noticed in larger tumors
Clin Cancer Res 2005;11(18) September 15, 2005
(mean F SD, 6.38 F 2.90 versus 5.52 F 2.94; P = 0.0954) and
in lymph node – positive cases (P = 0.0934); however, a
statistical significance could not be reached (Table 1). Positivity
of interstitial cells for TFF3 had no impact on any of the
clinicopathologic factors evaluated in this study (data not
shown).
Forty-one (36.9%) tumors were positive for TFF2 and 70
(63.1%) were negative. There was a strong correlation with
tumor differentiation status and higher prevalence of TFF2
positivity in diffuse types of tumors (P = 0.0023). A
significantly high (P = 0085) prevalence of TFF2 positivity
was noticed when the tumors were located in the body and
antrum of the stomach. The average tumor size of TFF2-positive
tumors (mean F SD, 6.97 F 3.55 versus 5.27 F 2.32) was
significantly higher (P = 0.0195) than those of negative ones.
Prevalence of expression of TFF2 was significantly (P = 0.0470)
higher in patients with lymph node metastasis than in those
with no lymph node metastasis (Table 2).
TFF3, TFF2, and angiogenesis. The average of nMVD was
68.2 (median, 64.0; range, 24.0-163.0) and that of aMVD was
22.4 103 (median 20.5 103, range 4.12 103-53.5 103).
According to Spearman’s rank correlation test, there was a
significant correlation (P = 0.0016, r = 30.6) between the
nMVD and aMVD (data not shown). As shown in Table 1, the
averages of both nMVD (P = 0.0079, F test and P = 0.1470,
Mann-Whitney U test) and aMVD (P = 0.0280, Mann-Whitney
U test) were significantly higher in TFF3-positive tumors than
in TFF3-negative tumors. Whereas, the average of only nMVD
was significantly high in TFF2-positive tumors but aMVD had
no correlation with TFF2 positivity (Table 2). According to
Spearman’s correlation test, TFF3 had significant correlation
with both nMVD (P = 0.0224) and aMVD (P = 0.0172);
however, TFF2 had significant correlation only with nMVD
(P = 0.0228; data not shown).
6474
www.aacrjournals.org
Downloaded from clincancerres.aacrjournals.org on June 18, 2017. © 2005 American Association for Cancer
Research.
Trefoil Factor Family in Angiogenesis
Fig. 2. Serial sections of a tumor show thatTFF3
(A),TFF2 (B), and CD34 (C) are stained in the
same area in a highly vascularized tumor.
Quantification of CD34-stained areas was done by a
public domain NIH Image program following
background subtraction and keeping only the
CD34-stained areas (D). No counterstaining was
done for CD34 staining.
Survival analysis and multivariable analysis of prognostic
factors. TFF3 expression had a significant impact on diseasefree survival (P = 0.0013; Fig. 3A) but not on overall survival
(P = 0.0855; Fig. 3B). However, when the survival analysis was
repeated separately in male and female patients, TFF3 overexpression had significant impact on both disease-free survival
(P = 0.0029; Fig. 3C) and overall survival (P = 0.0213; Fig. 3D)
only in female patients. A similar significant impact of TFF3
expression was not found in male patients (data not shown).
The 5-year disease-free survival in female patients with TFF3negative tumors was 91.3%, whereas it was only 42.8% in
patients with TFF3-positive tumors. TFF2 had a significant
Table 1. Correlation of clinicopathologic findings with ITF expression
Variables
ITF negative (n = 62)
ITF positive (n = 49)
P
Age (y), mean F SD (range)
Sex
Female
Male
Tumor size (cm), mean F SD (range)
Tumor location
Upper
Middle
Lower
Differentiation
Well
Poor
Lymph node positivity
Negative
Positive
Tstage
T2
T3
T4
nMVD, mean F SD
aMVD, mean F SD
63.40 F10.86 (29-85)
67.88 F11.76 (32-84)
0.0174
24
38
5.52 F 2.94 (1.2-18)
15
34
6.38 F 2.90 (2.2-18)
0.3749
20
19
23
16
11
22
0.5791
30
32
26
23
0.6248
23
11
39
38
0.0934
34
24
4
63.98 F 21.73
20.50 F 9.74
22
26
1
73.77 F 31.60
24.80 F 11.67
0.0954
0.2265
0.1470 (0.0079, F test)
0.0280
Abbreviations: nMVD, number of microvessel density; aMVD, area of microvessel density.
www.aacrjournals.org
6475
Clin Cancer Res 2005;11(18) September 15, 2005
Downloaded from clincancerres.aacrjournals.org on June 18, 2017. © 2005 American Association for Cancer
Research.
Human Cancer Biology
Table 2. Correlation of clinicopathologic findings with TFF2 expression
Variables
Age (y), mean F SD (range)
Sex
Female
Male
Tumor size (cm), mean F SD (range)
Tumor location
Upper
Middle
Lower
Differentiation
Well
Poor
Lymph node positivity
Negative
Positive
Tstage
T2
T3
T4
nMVD, mean F SD
aMVD, mean F SD
TFF2 negative (n = 70)
TFF2 positive (n = 41)
P
66.64 F 11.24 (29-85)
63.22F 11.58 (32-84)
0.1829
27
43
5.27 F 2.32 (1.2-18)
12
29
6.97 F 3.55 (2.2-18)
0.3185
27
12
31
9
18
14
0.0085
43
27
13
28
0.0023
26
44
8
33
0.0470
38
30
2
64.34 F 25.94
22.20 F 11.42
18
20
3
75.42 F 27.18
22.72 F 9.77
0.0195
0.3876
0.0146
0.4823
Abbreviations: nMVD, number of microvessel density; aMVD, area of microvessel density.
impact only on disease-free survival (P = 0.0030, and P =
0.3745 on overall survival) where patients with TFF2 overexpression had frequent disease recurrences. A gender-specific
survival advantage for TFF2 negativity could not be detected
(data not shown).
The median values of nMVD and aMVD were used to
dichotomize patients into high and low angiogenic profiles.
The disease-free survival was significantly longer (P = 0.0393)
in patients with low aMVD in comparison with those with high
aMVD. Whereas, nMVD had no significant survival benefit
either in terms of disease-free survival or overall survival. Only
when the cut-off point for nMVD was set at 95 was there a
significant difference (P = 0.0308) in the disease-free survival
(data not shown). Therefore, in the multivariable analysis, only
aMVD was included as a covariate.
Conventional prognostic factors along with TFF3 expression,
TFF2 expression, and aMVD entered the multivariable analysis
for the disease-free survival. Union Internationale Contra
Cancrum tumor stage had the highest hazard ratio for disease
recurrence and strongest prognostic impact (hazard ratio = 3.19
and P < 0.0001). Both TFF3 and TFF2 became independent
prognostic factors in the multivariable analysis, however, the
TFF3 expression (hazard ratio = 3.05 and P = 0.0090) had
superior predictability of the disease recurrence than TFF2
(hazard ratio = 2.30 and P = 0.0489; Table 3).
apoptosis, motility, and invasion. In this study, we present
data that TFFs may also contribute to tumorigenesis through
promotion of angiogenesis. Both TFF2 and TFF3 expression was
detectable in vascular endothelial cells. Whereas the number of
microvessels correlated with TFF2 and TFF3, only TFF3
Discussion
Although TFFs exhibit a number of beneficial effects in
cytoprotection and restitution, they may have adverse consequences when overexpressed in tumors. Previous studies have
suggested the possible effects of trefoils on proliferation,
Clin Cancer Res 2005;11(18) September 15, 2005
Fig. 3. Disease-free and overall survival according toTFF3 expression. Patients
withTFF3-positive tumors had significantly worse disease-free survival than those
withTFF3-negative tumors (A and B).When female patients were further stratified
according to theTFF3 expression, both disease-free and overall survivals were
significantly shorter in patients withTFF3-positive tumors than inTFF3-negative
tumors (C and D).
6476
www.aacrjournals.org
Downloaded from clincancerres.aacrjournals.org on June 18, 2017. © 2005 American Association for Cancer
Research.
Trefoil Factor Family in Angiogenesis
Table 3. Multivariate analysis of prognostic factors
Disease-free survival
Prognostic factors
Hazard ratio (95% CI)
P
3.19 (1.88-5.39)
3.05 (1.32-7.0 4)
2.30 (1.00-5.27)
1.17 (0.99-1.4 0)
1.53 (0.72-3.2 5)
1.55 (0.66-3.68)
1.34 (0.57-3.18)
0.92 (0.64-1.33)
1.35 (0.54-3.35)
<0.0001
Union Internationale Contra Cancrum (UICC) stage
ITF expression
TFF2 expression
Blood loss
aMVD
Venous invasion
Lymphatic invasion
Age
Differentiation
0.0090
0.0489
0.0714
0.2697
0.3176
0.5052
0.6499
0.5236
Abbreviation: aMVD, area of microvessel density.
correlated with the total area of microvessels, suggesting that
the angiogenic property of TFF3 was greater than that of TFF2.
Finally, overexpression of TFF2 and TFF3 had a significant
negative impact on patient survival after curative resection of
advanced gastric carcinoma, and in a multivariable analysis
along with the tumor, node, and metastasis stage was an
independent predictor of disease recurrence.
TFF3 has previously been suggested by Yamachika et al. (17)
to represent a poor prognostic factor in patients with gastric
carcinoma. Our results confirm this earlier finding that in a
multivariable analysis, TFF3 expression represents, after the
tumor, node, and metastasis stage, the second highest risk for
disease recurrence following surgical resection. Nevertheless,
our results differed in several important respects from this
earlier study. In contrast with the Yamachika study, which
showed that TFF3 expression had no survival impact in
females but significantly impacted survival in male patients,
we found that TFF3 expression was a strong predictor of
survival in female patients. Indeed, in our study, female
patients with advanced gastric carcinoma that were TFF3negative exhibited a 91.3% 5-year survival, comparable to
patients with early gastric cancers. The reason for the stronger
female preference for TFF3’s clinical associations is not entirely
clear, although direct evidence for estrogen-mediated TFF3
expression was reported by May and Westley (21). In this
study, the authors showed that estradiol treatment increased
TFF3 mRNA expression by 3- to 10-fold, and that up-regulated
TFF3 gene expression could also be induced by tamoxifen.
Direct evidence for differences in the overall level of TFF3
expression between men and women is lacking, and although
a few studies have suggested that tamoxifen, in combination
with other agents, may have some therapeutic role in the
treatment of gastric cancer (22 – 24).
Although the reason for the negative impact of TFF3
expression is not entirely clear, circumstantial evidence in this
study supports a role for angiogenesis. Tumor angiogenesis is
a coordinated process that involves the migration of
endothelial progenitors into a region of hypoxia where
exposure to angiogenic growth factors leads to the generation
of new blood vessels (25). At present, it is not clear as to the
mechanism by which TFFs are able to promote tumor
angiogenesis, although a number of previous studies have
www.aacrjournals.org
suggested that TFFs can stimulate migration of intestinal
epithelial cells and possibly tumor cells as well. Recently,
studies by Rodrigues et al. (19) showed in an in vitro model
that TFFs could stimulate angiogenesis in a manner
comparable to vascular endothelial growth factor; the
angiogenic activity of TFF3 was dose-dependent and could
be blocked by epidermal growth factor receptor (ZD1839)
and cyclooxygenase-2 (NS-3998) inhibitors. Given these data,
it seems likely that TFFs contribute directly to angiogenesis,
although further studies will be required to show that this is
a direct effect. An alternative possibility is that hypoxia,
which has been shown to induce TFF3 transcription through
hypoxia-inducible factor-1a, could lead to increased angiogenesis and increased TFF3 expression (26). Indeed, a recent
report by Zhang et al. (27) showed that immature rats
suffering from hypoxia-induced necrotizing enterocolitis
could be rescued by recombinant TFF3 treatment.
Although expression of both TFF2 and TFF3 showed a
significant positive correlation with tumor angiogenesis and
a negative correlation with patient survival, the two trefoil
factors differed in several important respects. As noted
previously, TFF2 was not expressed in intestinal metaplasia,
strongly expressed in spasmolytic protein-expressing metaplasia, and overexpressed in tumors with a diffuse histology, large
tumors and tumors with frequent lymph node metastasis. In
contrast, TFF3 was consistently expressed in intestinal metaplasia, occasionally in spasmolytic protein-expressing metaplasia, was overexpressed in elderly patients, but had no significant
correlation with histology, tumor size, or metastases. In diffuse
gastric cancers, TFF2 was preferentially expressed in cancer cells
invading into the muscularis mucosa and beyond, whereas
TFF3 staining had no such characteristic feature and was more
uniformly expressed in cancer tissue.
In conclusion, the results of this study show that expression
of TFF2 and TFF3 in gastric cancer could be associated with
patient survival after a curative resection and were independent predictors of disease recurrence. Expression of TFF2 and
TFF3 in gastric cancer seemed to correlate with tumor MVD
in gastric carcinoma samples, suggesting a possible role in
tumor angiogenesis. Further studies are needed to more
precisely define the role of trefoil peptide expression in gastric
cancer.
6477
Clin Cancer Res 2005;11(18) September 15, 2005
Downloaded from clincancerres.aacrjournals.org on June 18, 2017. © 2005 American Association for Cancer
Research.
Human Cancer Biology
References
1. Wright NA, Hoffmann W, Otto WR, Rio MC, Thim L.
Rolling in the clover: trefoil factor family (TFF)-domain
peptides, cell migration and cancer. FEBS Lett 1997;
408:121 ^ 3.
2. Alison MR, Chinery R, Poulsom R, Ashwood P,
Longcroft JM, Wright NA. Experimental ulceration
leads to sequential expression of spasmolytic polypeptide, intestinal trefoil factor, epidermal growth
factor and transforming growth factor a mRNAs in rat
stomach. J Pathol 1995;175:405 ^ 14.
3. Chen YH, Lu Y, De Plaen IG, Wang LY, Tan XD.
Transcription factor NF-nB signals antianoikic
function of trefoil factor 3 on intestinal epithelial
cells. Biochem Biophys Res Commun 2000;274:
576 ^ 82.
4. Taupin DR, Kinoshita K, Podolsky DK. Intestinal trefoil
factor confers colonic epithelial resistance to apoptosis. Proc Natl Acad Sci U S A 2000;97:799 ^ 804.
5. Kinoshita K,Taupin DR, Itoh H, Podolsky DK. Distinct
pathways of cell migration and antiapoptotic response
to epithelial injury: structure-function analysis of human intestinal trefoil factor. Mol Cell Biol 2000;20:
4680 ^ 90.
6. Beck PL, Wong JF, Li Y, et al. Chemotherapy- and
radiotherapy-induced intestinal damage is regulated
by intestinal trefoil factor. Gastroenterology 2004;126:
796 ^ 808.
7. Goke MN, Cook JR, Kunert KS, Fini ME, Gipson IK,
Podolsky DK. Trefoil peptides promote restitution of
wounded corneal epithelial cells. Exp Cell Res 2001;
264:337 ^ 44.
8. Chinery R, Playford RJ. Combined intestinal trefoil
factor and epidermal growth factor is prophylactic
against indomethacin-induced gastric damage in the
rat. Clin Sci (Lond) 1995;88:401 ^ 3.
9. Lefebvre O, Chenard MP, Masson R, et al. Gastric
mucosa abnormalities and tumorigenesis in mice
lacking the pS2 trefoil protein. Science 1996;274:
259 ^ 62.
10. Tebbutt NC, Giraud AS, Inglese M, et al. Reciprocal
regulation of gastrointestinal homeostasis by SHP2
and STAT-mediated trefoil gene activation in gp130
mutant mice. Nat Med 2002;8:1089 ^ 97.
11. Houghton J, Stoicov C, Nomura S, et al. Gastric
cancer originating from bone marrow-derived cells.
Science 2004;306:1568 ^ 71.
12. Lee JY, Eom EM, Kim DS, Ha-Lee YM, Lee DH.
Analysis of gene expression profiles of gastric normal
and cancer tissues by SAGE. Genomics 2003;82:
78 ^ 85.
13. Hippo Y, Yashiro M, Ishii M, et al. Differential gene
expression profiles of scirrhous gastric cancer cells
with high metastatic potential to peritoneum or lymph
nodes. Cancer Res 2001;61:889 ^ 95.
14. Missiaglia E, Blaveri E,Terris B, et al. Analysis of gene
expression in cancer cell lines identifies candidate
markers for pancreatic tumorigenesis and metastasis.
Int J Cancer 2004;112:100 ^ 12.
15. Nomura S, Baxter T, Yamaguchi H, et al. Spasmolytic
polypeptide expressing metaplasia to preneoplasia in
H. felis-infected mice. Gastroenterology 2004;127:
582 ^ 94.
16. Garraway IP, Seligson D, Said J, Horvath S, Reiter
RE. Trefoil factor 3 is overexpressed in human prostate
cancer. Prostate 2004;61:209.
17. Yamachika T, Werther JL, Bodian C, et al. Intestinal
trefoil factor: a marker of poor prognosis in gastric
carcinoma. Clin Cancer Res 2002;8:1092 ^ 9.
18. Dhar DK,Wang TC, Maruyama R, et al. Expression of
cytoplasmicTFF2 is a marker of tumor metastasis and
negative prognostic factor in gastric cancer. Lab Invest
2003;83:1343 ^ 52.
Clin Cancer Res 2005;11(18) September 15, 2005
6478
19. Rodrigues S, Van Aken E, Van Bocxlaer S, et al.
Trefoil peptides as proangiogenic factors in vivo and
in vitro : implication of cyclooxygenase-2 and EGF
receptor signaling. FASEB J 2003;17:7 ^ 16.
20. Dhar DK, Udagawa J, Ishihara S, et al. Expression of
regenerating gene I in gastric adenocarcinomas :
correlation with tumor differentiation status and
patient survival. Cancer 2004;100:1130 ^ 6.
21. May FE, Westley BR. Expression of human intestinal
trefoil factor in malignant cells and its regulation by
oestrogen in breast cancer cells. J Pathol 1997;182:
404 ^ 13.
22. Hosoya Y, Kitoh Y, Kobayashi E, Okabe R, Fujimura
A, Kanazawa K. Combination effects of tamoxifen plus
5-fluorouracil on gastric cancer cell lines in vitro.
Cancer Lett 1999;140:139 ^ 43.
23. Ohkura H, Kitaoka H. Chemoendocrine therapy of
gastric ‘‘scirrhous’’ carcinoma. Gan To Kagaku Ryoho
1988;4:1292 ^ 6.
24. Koullias GJ, Pratsinis CI, Korkolis DP, et al. Brief
tamoxifen pretreatment enhances the chemosensitivity of gastric carcinoma cells to 5-fluorouracil
in vitro. Anticancer Res 2003;23:1575 ^ 80.
25. Saaristo A, Karpanen T, Alitalo K. Mechanisms
of angiogenesis and their use in the inhibition of
tumor growth and metastasis. Oncogene 2000;19:
6122 ^ 9.
26. Furuta GT, Turner JR, Taylor CT, et al. Hypoxiainducible factor 1-dependent induction of intestinal
trefoil factor protects barrier function during hypoxia.
J Exp Med 2001;193:1027 ^ 34.
27. Zhang BH, Yu HG, Sheng ZX, Luo HS, Yu JP.
The therapeutic effect of recombinant human
trefoil factor 3 on hypoxia-induced necrotizing
enterocolitis in immature rat. Regul Pept 2003;116:
53 ^ 60.
www.aacrjournals.org
Downloaded from clincancerres.aacrjournals.org on June 18, 2017. © 2005 American Association for Cancer
Research.
Expression of Trefoil Factor Family Members Correlates with
Patient Prognosis and Neoangiogenesis
Dipok Kumar Dhar, Timothy C. Wang, Hideki Tabara, et al.
Clin Cancer Res 2005;11:6472-6478.
Updated version
Cited articles
Citing articles
E-mail alerts
Reprints and
Subscriptions
Permissions
Access the most recent version of this article at:
http://clincancerres.aacrjournals.org/content/11/18/6472
This article cites 27 articles, 8 of which you can access for free at:
http://clincancerres.aacrjournals.org/content/11/18/6472.full#ref-list-1
This article has been cited by 2 HighWire-hosted articles. Access the articles at:
http://clincancerres.aacrjournals.org/content/11/18/6472.full#related-urls
Sign up to receive free email-alerts related to this article or journal.
To order reprints of this article or to subscribe to the journal, contact the AACR Publications
Department at [email protected].
To request permission to re-use all or part of this article, contact the AACR Publications
Department at [email protected].
Downloaded from clincancerres.aacrjournals.org on June 18, 2017. © 2005 American Association for Cancer
Research.