Plasma expression of the intestinal metaplasia markers CDH17 and TFF3 in patients with gastric cancer

Abstract

BACKGROUND:

Few reliable blood markers are available for detecting gastric cancer, mainly owing to the heterogeneity of the cancer.

OBJECTIVE:

To establish gastric cancer diagnostic markers, we evaluated the levels of plasma cadherin 17 (CDH17) and trefoil factor 3 (TFF3), which are secretory proteins and known markers for intestinal metaplasia (IM), in patients with gastric cancer.

METHOD:

The protein expression level was analyzed in blood plasma samples from 111 gastric cancer patients and 44 healthy individuals, using a sandwich ELISA kit, followed by statistical analyses.

RESULT:

Overall, the plasma levels of CDH17 and TFF3 were not significantly different between groups ( $p=$ 0.160 and $p=$ 0.113, respectively). However, CDH17 expression was significantly elevated in patients with stage II and III gastric cancers compared to that in healthy controls ( $p=$ 0.023 and $p=$ 0.037, respectively). In contrast, TFF3 levels were significantly elevated in patients with stage I ( $p=$ 0.001) and T1 gastric cancer ( $p=$ 0.013). The sensitivity and specificity of CDH17 were 66.7 and 61.4%, respectively (cutoff point: 0.189 ng/mL); for TFF3, these values were 62.2 and 56.8%, respectively (cutoff point: 5.215 ng/mL).

CONCLUSIONS:

These findings indicate that secretory protein markers for metaplastic lineages can be used as blood markers for gastric cancer.

Keywords

CDH17 TFF3 ELISA gastric cancer marker

1. Introduction

Despite decreasing incidence, gastric cancer (GC) still remains one of the most common cancers in Asia, especially Korea, Japan, and China [1, 2]. Diagnosis of early gastric cancer (EGC) is crucial for treatment and favorable prognosis [3]. Endoscopic examination has allowed for earlier detection but is still regarded as expensive, invasive, and time-consuming [4, 5].

The US National Cancer Institute defines a biomar-ker as “a biological molecule found in blood, other body fluid, or tissues that is a sign of a normal or abnormal process or of a condition or disease.” Generally, cancer biomarkers are classified into four categories: predictive, diagnostic, prognostic, and pharmacodynamic [6]. Among these, diagnostic markers are used for detecting any stage of cancer during cancer development [6, 7]. If there was a reliable GC diagnostic marker, it would help in the diagnosis of early stages of GC. However, until date, the traditional cancer markers carcinoembryonic antigen (CEA) and carbohydrate antigen 19-9 (CA19-9) had been used for GC, but there is no specific and independent marker for the early detection of GC [8].

In addition, GC is very heterogeneous, whereas precancerous lesions are homogeneous [9]. Owing to this, precancerous lesions seem to be a good target for the development of a GC biomarker. Therefore, a first step towards finding a GC biomarker could be the identification of possible biomarkers for gastric precancerous lesions, such as chronic atrophic gastritis (CAG), intestinal metaplasia (IM), and/or spasmolytic polypeptide expressing metaplasia (SPEM). These precancerous markers could then be studied in GC tissue or serum as a second step [10]. For example, serum pepsinogen (PG) I levels or the PG1/II ratio were studied as representative precancerous markers in gastric inflammations and atrophic gastritis, because serum levels of PGI and the PG I/II ratio decreased during advanced mucosal atrophy [11, 12]. Additionally, olfactomedin 4 (OLFM4) expression has been reported as an IM marker that is involved in early-stage gastric carcinogenesis [13].

Our previous study identified 13 IM or SPEM markers by complementary DNA (cDNA) microarray and immunostaining analysis in tissue samples. According to this study, eight of the IM markers were confirmed to be GC markers, including mucin 13 (MUC13), cadherin 17 (CDH17), OLFM4, mucin 5AC (MUC5AC), keratin 20 (KRT20), lectin galactosidase-binding soluble 4 (LGALS4), aldo-keto reductase family 1 member B10 (AKR1B10), and regenerating islet-derived family member 4 (REG4) [10]. In addition, most of the identified GC markers were evaluated for their prognostic impact, either by themselves or in combination [14].

In this study, we selected CDH17, a tissue marker for IM and a potent prognostic marker for GC, as a possible serum biomarker, considering its extracellular location. CDH17 contains a signal peptide that is conventionally found in proteins that are located extracellularly as transmembrane proteins or secreted from the cell to body fluids, which makes these proteins promising candidates for the development of new biomarkers [15]. TFF3, a well-known marker for IM, is already known to be detected in human blood [16].

In this study, we evaluated CDH17 and TFF3 levels in the plasma from patients with GC to evaluate their role as diagnostic serum biomarkers for GC.

2. Materials and methods

2.1 Patient and clinical data

Blood plasma samples from 111 patients with GC and 44 healthy individuals were analyzed. Blood samples from patients were collected directly before surgery, and none of the patients included in this study had received radiation therapy or preoperative chemotherapy. Characteristics of the patients from the GC group are shown in Table 1.

Table 1
Characteristics of patient with gastric cancer

Characteristic Gastric cancer ( $n=$ 111) Percentage

Sex

Male 76 68.47%

Female 35 31.53%

Age, years

$\leqslant$ 60 73 65.77%

$>$ 60 38 34.23%

T classification

T1 42 37.84%

T2 16 14.41%

T3 34 30.63%

T4 17 15.32%

Unknown 2 1.80%

N classification

N0 57 51.35%

N1 23 20.72%

N2 12 10.81%

N3 19 17.12%

Unknown 2 1.80%

Distance metastasis

Absent 108 97.30%

Present 3 2.70%

TNM stage

I 42 37.84%

II 39 35.14%

III 27 24.32%

IV 3 2.70%

Lauren classification

Intestinal 42 37.84%

Diffuse 48 43.24%

Mixed 15 13.51%

Unknown 3 2.70%

WHO classification

Pap 1 0.90%

W/D 8 7.21%

M/D 26 23.42%

P/D 16 14.41%

Mucinous 1 0.90%

SRC 41 36.94%

Mixed 16 14.41%

Unknown 2 1.80%

Characteristic	Gastric cancer ( $n=$ 111)	Percentage
Sex
Male	76	68.47%
Female	35	31.53%
Age, years
$\leqslant$ 60	73	65.77%
$>$ 60	38	34.23%
T classification
T1	42	37.84%
T2	16	14.41%
T3	34	30.63%
T4	17	15.32%
Unknown	2	1.80%
N classification
N0	57	51.35%
N1	23	20.72%
N2	12	10.81%
N3	19	17.12%
Unknown	2	1.80%
Distance metastasis
Absent	108	97.30%
Present	3	2.70%
TNM stage
I	42	37.84%
II	39	35.14%
III	27	24.32%
IV	3	2.70%
Lauren classification
Intestinal	42	37.84%
Diffuse	48	43.24%
Mixed	15	13.51%
Unknown	3	2.70%
WHO classification
Pap	1	0.90%
W/D	8	7.21%
M/D	26	23.42%
P/D	16	14.41%
Mucinous	1	0.90%
SRC	41	36.94%
Mixed	16	14.41%
Unknown	2	1.80%

W/D $=$ well-differentiated adenocarcinoma, M/D $=$ moderately differentiated adenocarcinoma, P/D $=$ poorly differentiated adenocarcinoma, SRC $=$ signet ring cell carcinoma.

Blood samples from subjects without any evidence of malignancy formed the control group. Gastroscopy was performed in all healthy control subjects to exclude the presence of GC. As normal controls, including within normal limits (WNL, $n=$ 26), the patients who had gastritis ( $n=$ 18) were included.

Normal, control blood samples were collected before endoscopy examination and samples from patients with cancer were collected before surgery. Therefore, all the blood samples were collected under fasting conditions.

Blood samples were collected in 8-mL EDTA bottles, centrifuged within 12 h of collection, and the plasma was stored at $-$ 80 ${}^{\circ}$ C until use. This study was approved by the Ethics Committee of Seoul National University Hospital and written consent was obtained from all participants (IRB # 0602-036-168).

2.2 Determination of plasma CDH17 and TFF3 levels by ELISA

Plasma TFF3 and CDH17 levels were analyzed using a human-specific ELISA kit (Biomatik, Cambridge, Canada) following the manufacturer’s protocol. Briefly, a serial dilution of the standard solution was prepared, and plasma samples were diluted 10-fold in PBS. Samples and standards (100 $\mu$ L) were added to a microplate pre-coated with capture antibody and incubated for 2 h at room temperature. The liquid was then aspirated, 100 $\mu$ L detection reagent was added, and the plate was incubated for 1 h at room temperature. After repeated aspiration and washing, 100 $\mu$ L reagent B was added and incubated for 30 min at room temperature. After aspirating and washing five times, 90 $\mu$ L of substrate solution was added and incubated for 20 min, followed by the addition of 50 $\mu$ L stop solution. The absorbance at 450 nm was measured, and the concentration of samples was calculated from the standard curves. All samples and standards were assayed in duplicate. Measurement between normal control sample, patient samples, and TNM stage were considered to be located in each plate evenly.

2.3 Statistical analysis

Unpaired two-tailed $t$ tests were performed to compare the control and patient groups. A P-value below 0.05 was considered statistically significant. Receiver operating characteristic (ROC) curve analysis and statistical analyses were performed using SPSS v13.0 (SPSS, Chicago, IL), and graphs were generated using GraphPad Prism 5.0 (GraphPad Software Inc., CA).

3. Results

3.1 Plasma CDH17 and TFF3 levels in patients with GC and normal control subjects

The mean levels of CDH17 in patients with GC and healthy controls were 0.441 $\pm$ 0.043 ng/mL ( $n=$ 111) and 0.329 $\pm$ 0.060 ng/mL ( $n=$ 44), respectively. The mean levels of TFF3 in GC patients and controls were 7.735 $\pm$ 0.702 ng/mL ( $n=$ 111) and 6.195 $\pm$ 0.702 ng/mL ( $n=$ 44), respectively. CDH17 and TFF3 level were increased in the GC group compared to that in the control group; however, neither of these changes were significant ( $p=$ 0.160 and $p=$ 0.113, respectively) (Figs 1A and 1B) There were no significant differences in CDH17 and TFF3 levels in terms of sex and age.

Figure 1.

Plasma concentration of CDH17 and TFF3. (A) CDH17 level was not significantly different between gastric cancer (GC) patients and normal ( $p=$ 0.160). The mean levels of CDH17 expression in patients with GC and healthy controls were 0.441 $\pm$ 0.043 ng/mL ( $n=$ 111) and 0.329 $\pm$ 0.060 ng/mL ( $n=$ 44) ng/mL, respectively. (B) TFF3 level was not significantly different between GC patients and normal ( $p=$ 0.113). The mean levels of TFF3 expression in patients with GC and controls were 7.735 $\pm$ 0.702 ng/mL ( $n=$ 111) and 6.195 $\pm$ 0.702 ng/( $n=$ 44), respectively.

3.2 Relationship between CDH17 and TFF3 expression and GC pathology

CDH17 levels were significantly increased in patients with advanced GC (AGC, pT2 or higher; 0.553 $\pm$ 0.060 ng/mL) compared to that in normal controls (0.329 $\pm$ 0.060 ng/mL $p=$ 0.015), but the difference was not significant between early GC (EGC, pT1; 0.329 $\pm$ 0.060 ng/mL, $p=$ 0.248) and normal controls (Fig. 2A). When subdividing the patient group according to the N classification, CDH17 levels were found to be significantly increased in the N2 GC group (0.329 $\pm$ 0.060 ng/mL, $p=$ 0.008, $n=$ 12) compared to that in the normal controls; however, CDH17 levels were not significantly different between the N0, N1, and N3 groups, and normal controls (Fig. 2C). In addition, significant differences were found in stage II (0.578 $\pm$ 0.091 ng/mL, $p=$ 0.023, $n=$ 39) and stage III (0.549 $\pm$ 0.088 ng/mL, $p=$ 0.037, $n=$ 27) GC and control samples (Fig. 2E).

Figure 2.

Characteristics of patient with GC according to plasma level of CDH17 and TFF3. (A) CDH17 level was significantly increased in advanced GC (AGC, pT2 or higher) compared to normal control (0.553 $\pm$ 0.060 ng/ml, $p=$ 0.015). (B) TFF3 level showed significantly elevated in early GC (EGC, pT1) compared to normal control (8.932 $\pm$ 0.836 ng/ml, $p=$ 0.013). (C) CDH17 expression was significantly elevated in samples from N2 GC patients as compared with that observed in control samples (0.329 $\pm$ 0.060 ng/mL, $p=$ 0.008, $n=$ 12). (D) No significant difference was observed in TFF3 according to N classification. (E) CDH17 expression was significantly elevated in patients with stage II (0.578 $\pm$ 0.091 ng/mL, $p=$ 0.023, $n=$ 39) and III (0.549 $\pm$ 0.088 ng/mL, $p=$ 0.037, $n=$ 27) compared with that in healthy controls. (F) Significant differences were observed in TFF3 expression in patients with stage I GC (9.931 $\pm$ 0.841 ng/mL, $p=$ 0.001, $n=$ 42) and the control group.

TFF3 levels were significantly elevated in EGC samples (8.932 $\pm$ 0.836 ng/mL) compared to that in normal controls (6.195 $\pm$ 0.702 ng/mL, $p=$ 0.013), but they were not significantly different between AGC (7.060 $\pm$ 0.693 ng/mL, $p=$ 0.407) and normal controls (Fig. 2B). When samples where categorized according to the N classification, no significant differences were observed (Fig. 2D). In addition, significant differences were found between stage I (9.913 $\pm$ 0.841 ng/mL, $p=$ 0.001, $n=$ 71) GC samples and controls (Fig. 2F). When dividing the patient group according to GC differentiation, we found that TFF3 expression was increased in the differentiated, compared to that in the undifferentiated GC group ( $p=$ 0.014). No significant difference were observed when the patient group was subdivided according to the Lauren classification both CDH17 and TFF3 (Fig. 3B).

Figure 3.

CDH17 and TFF3 expression level according to Lauren classification. No significant difference in CDH17 and TFF3 expression was observed in Lauren classification between the two groups.

3.3 ROC curve analysis

ROC curve analysis was performed to analyze the sensitivity and specificity of CDH17 and TFF3 levels as biomarkers. The area under the curve (AUC) for CDH17 levels in patients with GC was 0.608 ( $p=$ 0.036). The sensitivity and specificity was 66.7 and 61.4%, respectively (cutoff point: 0.189 ng/mL) (Fig. 4A). The AUC for TFF3 levels in patients with GC was 0.566 ( $p=$ 0.203), with a sensitivity and specificity of 62.2 and 56.8%, respectively (cutoff point: 5.215 ng/mL) (Fig. 4B).

Figure 4.

Receiver operating characteristic (ROC) analysis of CDH17 and TFF3. (A) The area under the curve (AUC) for CDH17 expression was 0.608 ( $p=$ 0.036), and the sensitivity and specificity were 66.7% and 61.4%, respectively (cut-off point: 0.189 ng/mL). (B) The AUC for TFF3 expression was 0.556 ( $p=$ 0.203), and the sensitivity and specificity were 62.2% and 56.8%, respectively (cutoff point: 5.215 ng/mL). (C) The AUC for CDH17 expression in samples from patients with stage II or III GC was 0.667 ( $p=$ 0.003), with specificity and sensitivity of 77.3% and 61.4%, respectively (cutoff point: 0.189 ng/mL). (D) The AUC for TFF3 expression in samples from patients with stage I GC was 0.703 ( $p=$ 0.001), and the sensitivity and specificity were 83.3% and 54.5%, respectively (cutoff point: 4.886 ng/mL). (E) Combined-marker analysis of CDH17 and TFF3 revealed a sensitivity of 60.4% and a specificity of 61.2%, with an AUC of 0.618 ( $p=$ 0.018).

The AUC for CDH17 levels in the stage II and stage III GC subgroup was 0.667 ( $p=$ 0.003), and the specificity and sensitivity were 77.3 and 61.4%, respectively (cutoff point: 0.189 ng/mL) (Fig. 4C). The AUC for TFF3 levels in the stage I subgroup was 0.703 ( $p=$ 0.001). The sensitivity and specificity were 83.3 and 54.5%, respectively (cutoff point: 4.886 ng/mL) (Fig. 4D).

To analyze the combination of the two biomarkers, regression binary logistic analysis was performed with CDH17 and TFF3 expression as independent variables before ROC curve analysis. Combined marker analysis of CDH17/TFF3 levels showed a sensitivity of 60.4% and a specificity of 61.2% with an AUC of 0.618 ( $p=$ 0.018) (Fig. 4E).

4. Discussion

GC endoscopy screening is a cost-effective diagnostic tool that is commonly performed in Korea and Japan, where GC incidence is very high [17]. This method enables the detection of early stages of cancer and leads to improvements in cure rates and survival [17, 18, 19, 20]. Nevertheless, it is not recommended in many other countries owing to its high cost and invasiveness [21, 22]. Therefore, development of non-invasive and cost-effective screening strategies with high accuracy is needed to detect early stages of GC.

In this study, levels of the potential biomarkers CDH17 and TFF3 were analyzed. CDH17 belongs to a subclass of the 7D-cadherin superfamily [23, 24]. Dysregulation of its gene expression can lead to cancer development and progression because it plays an important role in the transition from the epithelial to the mesenchymal stage [25]. The trefoil proteins, TFF1, TFF2, and TFF3, are expressed in the mucosa of the gastrointestinal tract [26]. TFF1 is expressed in normal gastric mucosa but not in about 50% of gastric tumors. TFF2 is expressed in some GCs and is related to cancer invasion, metastasis, and poor prognosis [27]. TFF3 is a secretory protein expressed in the goblet cells of the intestine and is involved in the maintenance of mucosal integrity [26, 28]. Further, TFF3 is overexpressed in a variety of cancers such as breast, lung, prostate, colon, and stomach cancer [29, 30, 31, 32]. Importantly, TFF3 has been shown to be a marker of IM [26, 33].

In our previous study, which included about 450 samples, we identified CDH17 upregulation in GC tissue and found that it is an independent prognostic biomarker in stage I or node-negative GC [10]. Several other studies have also reported CDH17 upregulation in GC and its association with poor prognosis by tissue analysis, such as immunohistochemistry and micro-arrays; however, blood-level analysis had not been reported until date [14, 34, 35]. CDH17 levels were increased in the blood plasma from the GC group compared to that from the control group, but this difference was not statistically significant (Fig. 1A). However, this finding is meaningful in that we determined the CDH17 expression levels in blood plasma samples for the first time. Furthermore, CDH17 expression was elevated in patients with AGC and significantly different between groups when we analyzed CDH17 levels according to TNM stage and pathological diagnosis (Figs 2A and 2E). From these results and the work of others, we conclude that the membrane protein CDH17 is overexpressed in the early stages of GC but is only released to the bloodstream in the late stages.

Roa et al reported that TFF3 expression is regulated by insulin and glucose [36]. In this study, TFF3 levels increased either upon insulin treatment or after breakfast. Therefore, we analyzed the diabetes status of each patient. The patients with diabetes (7238 $\pm$ 1489 pg/ml, $n=$ 16) showed increased TFF3 expression, but this was not significantly different from that in patients without diabetes (7238 $\pm$ 556.8 pg/ml, $n=$ 93; $p=$ 0.210).

To the best of our knowledge, this is the first report to show that CDH17 can be detected in human serum by ELISA. TFF3 is a secretory protein that has been detected in a variety of cancers, such as colorectal, prostate and endometrial cancer as well as GC [16, 37, 38, 39, 40]. For example, TFF3 was reported as a predictive biomarker of endocrine response in breast cancer metastasis [41]. Its expression was also significantly higher in the sera of patients with lung cancer than in healthy individuals; TFF1 and TFF2 expression was either similar between the two groups or slightly higher in the former [42].

In addition, TFF3 is an IM marker that is associated with gastric carcinogenesis [32]. Based on these findings, we hypothesized that the plasma TFF3 levels, together with CDH17 levels, could be used as a novel combination biomarker to detect early stages of cancer. In our study, we detected TFF3 levels to be between 0 to 25 ng/mL. This corroborates the results of other GC studies, but is very different from the levels detected in other cancers, such as endometrial carcinoma and colorectal cancer, where levels up to 4000 ng/mL have been reported [37, 38, 39]. Our results show that the ELISA kit can detect the target molecules relatively accurately. Although TFF3 levels were not significantly different between the normal and GC group, we found that it was increased in the EGC subgroup compared to that in the control group and the AGC subgroup (Figs 2B and 2F). These results indicate that TFF3 could potentially be used as an EGC marker.

Figure 5.

Correlation of CDH17 and TFF3 levels in patient serum The expression was inversely correlated in patient serum, but we could not find any reference for this correlation (Spearman $r=$ $-$ 0.2688, $p<$ 0.001).

In addition, we analyzed the correlation between CDH17 and TFF3 expression. Several studies reported that CDH17 and TFF3 expression is associated with lymph mode metastasis in GC [32, 43]. In our study, these genes were highly expressed in the patients with lymph node metastasis. Notably, their expression was inversely correlated in patient serum, but we could not find any reference for this correlation (Spearman $r=$ $-$ 0.2688, $p<$ 0.001, Fig. 5). The underlying mechanism needs to be investigated in future studies. The sensitivity and specificity were 65.8 and 61.2% (cutoff point: 0.191 ng/mL), for CDH17 and 62.2 and 55.2%, respectively (cutoff point: 5.215 ng/mL), for TFF3 (Figs 4A and 4B). Similar trends were observed for the combination of CDH17/TFF3. Combined marker analysis of CDH17/TFF3 showed a sensitivity of 60.4% and a specificity of 61.2%, with an AUC of 0.618 ( $p=$ 0.018, Fig. 4E). However, our research has the following limitations: the patient cohort was relatively small compared with those in other similar studies [16, 44] and our data was not has not been corroborated by western analysis.

In conclusion, CDH17 levels were elevated in advanced stages of GC whereas TFF3 levels were increased in the early stages.

Footnotes

Acknowledgment

This study was supported by a grant from Doosan Yonkang Foundation (# 30-2011-0080) and Cancer Research Institute, Seoul National University (2012).

Supplementary data

References

Siegel

Naishadham

and Jemal

, Cancer Statistics, CA: A Cancer Journal for Clinicians 62(1) (2012), 10–29.

Ferlay

et al., Estimates of worldwide burden of cancer in 2008: GLOBOCAN 2008, International Journal of Cancer, Journal International du Cancer 127(12) (2010), 2893–2917.

El-Sedfy

Brar

S.S.

and Coburn

N.G.

, Current role of minimally invasive approaches in the treatment of early gastric cancer, World Journal of Gastroenterology: WJG 20(14) (2014), 3880–3888.

Jung

K.W.

et al., Cancer statistics in Korea: incidence, mortality, survival, and prevalence in 2012, Cancer Research and Treatment: Official Journal of Korean Cancer Association 47(2) (2015), 127–141.

Whiting

J.L.

et al., The long term results of endoscopic surveillance of premalignant gastric lesions, Gut 50(3) (2002), 378–381.

Mishra

and Verma

, Cancer biomarkers: are we ready for the prime time? Cancers (Basel) 2(1) (2010), 190–208.

Aronson

J.K.

, Biomarkers and surrogate endpoints, Br J Clin Pharmacol 59(5) (2005), 491–494.

Shimada

et al., Clinical significance of serum tumor markers for gastric cancer: a systematic review of literature by the Task Force of the Japanese Gastric Cancer Association, Gastric Cancer: Official Journal of the International Gastric Cancer Association and the Japanese Gastric Cancer Association 17(1) (2014), 26–33.

Saito

et al., Histologic heterogeneity and mucin phenotypic expression in early gastric cancer, Pathol Int 51(3) (2001), 165–171.

10.

Lee

H.J.

et al., Gene expression profiling of metaplastic lineages identifies CDH17 as a prognostic marker in early stage gastric cancer, Gastroenterology 139(1) (2010), 213–225. e3.

11.

Sipponen

and Graham

D.Y.

, Importance of atrophic gastritis in diagnostics and prevention of gastric cancer: application of plasma biomarkers, Scand J Gastroenterol 42(1) (2007), 2–10.

12.

Mukoubayashi

et al., Serum pepsinogen and gastric cancer screening, Intern Med 46(6) (2007), 261–266.

13.

Jang

B.G.

Lee

B.L.

and Kim

W.H.

, Olfactomedin-related proteins 4 (OLFM4) expression is involved in early gastric carcinogenesis and of prognostic significance in advanced gastric cancer, Virchows Arch, 2015.

14.

Suh

Y.S.

et al., The combined expression of metaplasia biomarkers predicts the prognosis of gastric cancer, Annals of Surgical Oncology 19(4) (2012), 1240–1249.

15.

Inal

J.M.

et al., Blood/plasma secretome and microvesicles, Biochimica et Biophysica Acta 1834(11) (2013), 2317–2325.

16.

Aikou

et al., Tests for serum levels of trefoil factor family proteins can improve gastric cancer screening, Gastroenterology 141(3) (2011), 837–845. e1-7.

17.

Choi

K.S.

and Suh

, Screening for gastric cancer: the usefulness of endoscopy, Clinical Endoscopy 47(6) (2014), 490–496.

18.

Fukao

et al., The evaluation of screening for gastric cancer in Miyagi Prefecture, Japan: a population-based case-control study, International Journal of Journal International du Cancer 60(1) (1995), 45–48.

19.

Mizoue

et al., Prospective study of screening for stomach cancer in Japan, International Journal of Cancer, Journal International du Cancer 106(1) (2003), 103–107.

20.

Hisamichi

Sugawara

and Fukao

, Effectiveness of gastric mass screening in Japan, Cancer Detection and Prevention 11(3–6) (1988), 323–329.

21.

et al., Endothelial cell-specific molecule-1: a potential serum marker for gastric cancer, Tumour Biology: The Journal of the International Society for Oncodevelopmental Biology and Medicine 35(10) (2014), 10497–10502.

22.

Wanebo

H.J.

et al., Cancer of the stomach. A patient care study by the American College of Surgeons, Annals of Surgery 218(5) (1993), 583–592.

23.

Nollet

Kools

and van Roy

, Phylogenetic analysis of the cadherin superfamily allows identification of six major subfamilies besides several solitary members, Journal of Molecular Biology 299(3) (2000), 551–572.

24.

Angst

B.D.

Marcozzi

and Magee

A.I.

, The cadherin superfamily: diversity in form and function, Journal of Cell Science 114(Pt 4) (2001), 629–641.

25.

Lee

N.P.

et al., Role of cadherin-17 in oncogenesis and potential therapeutic implications in hepatocellular carcinoma, Biochimica et Biophysica Acta 1806(2) (2010), 138–145.

26.

Taupin

and Podolsky

D.K.

, Trefoil factors: initiators of mucosal healing, Nature Reviews. Molecular Cell Biology 4(9) (2003), 721–732.

27.

Dhar

D.K.

et al., Expression of cytoplasmic TFF2 is a marker of tumor metastasis and negative prognostic factor in gastric cancer, Laboratory Investigation; A Journal of Technical Methods and Pathology 83(9) (2003), 1343–1352.

28.

Zhou

et al., MAL facilitates the incorporation of exocytic uroplakin-delivering vesicles into the apical membrane of urothelial umbrella cells, Mol Biol Cell 23(7) (2012), 1354–1366.

29.

Davidson

et al., Gene expression signatures differentiate adenocarcinoma of lung and breast origin in effusions, Human Pathology 43(5) (2012), 684–694.

30.

Perera

et al., Trefoil factor 3 (TFF3) enhances the oncogenic characteristics of prostate carcinoma cells and reduces sensitivity to ionising radiation, Cancer Letters 361(1) (2015), 104–111.

31.

John

et al., Expression of TFF3 during multistep colon carcinogenesis, Histology and Histopathology 22(7) (2007), 743–751.

32.

et al., Reduced expression of TFF1 and increased expression of TFF3 in gastric cancer: correlation with clinicopathological parameters and prognosis, International Journal of Medical Sciences 10(2) (2013), 133–140.

33.

Gutierrez-Gonzalez

and Wright

N.A.

, Biology of intestinal metaplasia in 2008: more than a simple phenotypic alteration, Digestive and Liver Disease: Official Journal of the Italian Society of Gastroenterology and the Italian Association for the Study of the Liver 40(7) (2008), 510–522.

34.

Grotzinger

et al., LI-cadherin: a marker of gastric metaplasia and neoplasia, Gut 49(1) (2001), 73–81.

35.

Ito

et al., Clinicopathological significant and prognostic influence of cadherin-17 expression in gastric cancer, Virchows Archiv: An International Journal of Pathology 447(4) (2005), 717–722.

36.

Roa

G.J.B.

Tortolero

G.S.

and Gonzalez

J.E.

, Trefoil factor 3 (TFF3) expression is regulated by insulin and glucose, Journal of Health Sciences 3(1) (2013).

37.

Huang

Y.G.

et al., Aberrant expression of trefoil factor 3 is associated with colorectal carcinoma metastasis, Journal of Cancer Research and Therapeutics 9(3) (2013), 376–380.

38.

Bignotti

et al., Trefoil factor 3: a novel serum marker identified by gene expression profiling in high-grade endometrial carcinomas, British Journal of Cancer 99(5) (2008), 768–773.

39.

Xiao

et al., Serum TFF3 may be a pharamcodynamic marker of responses to chemotherapy in gastrointestinal cancers, BMC Clinical Pathology 14 (2014), 26.

40.

Vestergaard

E.M.

et al., Plasma levels of trefoil factors are increased in patients with advanced prostate cancer, Clinical Cancer Research: An Official Journal of the American Association for Cancer Research 12(3 Pt 1) (2006), 807–812.

41.

May

F.E.

and Westley

B.R.

, TFF3 is a valuable predictive biomarker of endocrine response in metastatic breast cancer, Endocrine-Related Cancer 22(3) (2015), 465–479.

42.

et al., Increased trefoil factor 3 levels in the serum of patients with three major histological subtypes of lung cancer, Oncology Reports 27(4) (2012), 1277–1283.

43.

et al., Overexpression of LI-cadherin in gastric cancer is associated with lymph node metastasis, Biochemical and Biophysical Research Communications 319(2) (2004), 562–568.

44.

Kaise

et al., The combination of serum trefoil factor 3 and pepsinogen testing is a valid non-endoscopic biomarker for predicting the presence of gastric cancer: a new marker for gastric cancer risk, Journal of Gastroenterology 46(6) (2011), 736–745.