Association of Genetic Variants in and Promoter
Hypermethylation of CDH1 With Gastric Cancer
A Meta-Analysis
Huiquan Jing, PhD, Fei Dai, MD, PhD, Chuntao Zhao, PhD, Juan Yang, MD, Lizhuo Li, MD,
Pravina Kota, MS, Lijuan Mao, MD, Kaimin Xiang, MS, Changqing Zheng, MD,
and Jingyun Yang, PhD
Abstract: Gastric cancer (GC) is a common cause of cancer-related
death. The etiology and pathogenesis of GC remain unclear, with
genetic and epigenetic factors playing an important role. Previous
studies investigated the association of GC with many genetic variants
in and promoter hypermethylation of E-cadherin gene (CDH1), with
conflicting results reported.
To clarify this inconsistency, we conducted updated metaanalyses to assess the association of genetic variants in and the
promoter hypermethylation of CDH1 with GC, including C-160A
(rs16260) and other less-studied genetic variants,
Data sources were PubMed, Cochrane Library, Google Scholar, Web
of Knowledge, and HuGE, a navigator for human genome epidemiology.
Study eligibility criteria and participant details are as follows:
studies were conducted on human subjects; outcomes of interest
include GC; report of genotype data of individual genetic variants in
(or methylation status of) CDH1 in participants with and without GC
(or providing odds ratios [OR] and their variances).
Study appraisal and synthesis methods included the use of OR as
a measure of the association, calculated from random effects models
Editor: Lu Wang.
Received: May 3, 2014; revised: July 18, 2014; accepted: August 10,
2014.
From the Institute of Social Science Survey (HJ), Peking University,
Beijing; Department of Social Science (HJ), Shenyang Medical College;
Emergency Department (LL); Department of Gastroenterology (CZ),
Shengjing Hospital, China Medical University, Shenyang, Liaoning;
Division of Gastroenterology (FD, JY, LM), Second Affiliated
Hospital, Medical College of Xi’an Jiaotong University, Xi’an, Shaanxi;
Department of General Surgery (KX), Third Xiangya Hospital, Central
South University, Changsha, Hunan, China; Brain Tumor Center (CZ),
Cancer and Blood Diseases Institute, Cincinnati Children’s Hospital
Medical Center, Cincinnati, OH; Department of Biostatistics and
Epidemiology (PK), University of Oklahoma Health Sciences Center,
Oklahoma City, OK; Rush Alzheimer’s Disease Center (JYY); and
Department of Neurological Sciences (JYY), Rush University Medical
Center, Chicago, IL.
Correspondence: Huiquan Jing, Department of Social Science, Shenyang
Medical College, 146 Huanghe North Street, Shenyang, Liaoning
110034, China (e-mail: hqjing@hotmail.com).
Dr JYY’s research was supported by National Institutes of Health/
National Institute on Aging R01AG036042 and the Illinois Department of Public Health.
The authors have no conflicts of interest to disclose.
Supplemental digital content is available for this article. Direct URL
citation appears in the printed text and is provided in the HTML and
PDF versions of this article on the journal’s Web site (www.mdjournal.com).
Copyright © 2014 Wolters Kluwer Health | Lippincott Williams &
Wilkins. This is an open access article distributed under the Creative
Commons Attribution License 4.0, which permits unrestricted use,
distribution, and reproduction in any medium, provided the original work
is properly cited.
ISSN: 0025-7974
DOI: 10.1097/MD.0000000000000107
Medicine
•
Volume 93, Number 19, October 2014
in meta-analyses. We used I2 for the assessment of between-study
heterogeneity, and publication bias was assessed using funnel plot
and Egger test.
A total of 33 studies from 30 published articles met the eligibility
criteria and were included in our analyses. We found no association
between C-160A and GC (OR ¼ 0.88; 95% confidence interval [CI],
0.71–1.08; P ¼ 0.215), assuming an additive model (reference
allele C). C-160A was associated with cardia (OR ¼ 0.21; 95% CI,
0.11–0.41; P ¼ 2.60 10 6), intestinal (OR ¼ 0.66; 95% CI, 0.49–
0.90; P ¼ 0.008), and diffuse GC (OR ¼ 0.57; 95% CI, 0.40–0.82;
P ¼ 0.002). The association of C-160A with noncardia GC is of bottom
line significance (OR ¼ 0.65; 95% CI, 0.42–1.01; P ¼ 0.054). Multiple
other less-studied genetic variants in CDH1 also exhibited association
with GC. Gene-based analysis indicated a significant cumulative
association of genetic variants in CDH1 with GC (all Ps <10 5).
Sensitivity analysis excluding studies not meeting Hardy–Weinberg
equilibrium (HWE) yielded similar results. Analysis by ethnic groups
revealed significant association of C-160A with cardia GC in both
Asian and whites, significant association with noncardia GC only in
Asians, and no significant association with intestinal GC in both ethnic
groups. There was significant association of C160-A with diffuse GC
in Asians (P ¼ 0.011) but not in whites (P ¼ 0.081). However, after
excluding studies that violate HWE, this observed association is no
longer significant (P ¼ 0.126). We observed strong association of
promoter hypermethylation of CDH1 with GC (OR ¼ 12.23; 95% CI,
8.80–17.00; P ¼ 1.42 10 50), suggesting that epigenetic regulation of
CDH1 could play a critical role in the etiology of GC.
Limitations of this study are as follows: we could not adjust for
confounding factors; some meta-analyses were based on a small
number of studies; sensitivity analysis was limited due to unavailability of data; we could not test publication bias for some metaanalyses due to small number of included studies.
We found no significant association of the widely studied genetic
variant C-160A, but identified some other genetic variants showing
significant association with GC. Future studies with large sample sizes
that control for confounding risk factors and/or intensively interrogate
CpG sites in CDH1 are needed to validate the results found in this
study and to explore additional epigenetic loci that affect GC risk.
(Medicine 93(19):e107)
Abbreviations: CDH1 = E-cadherin gene, GC = gastric cancer,
HWE = Hardy–Weinberg equilibrium, OR = odds ratio, SNP =
single-nucleotide polymorphism.
INTRODUCTION
G
astric cancer (GC) is one of the most common gastrointestinal malignancies throughout the world. Over the past half
www.md-journal.com
| 1
Medicine
Jing et al
century, the incidence of GC has gradually decreased. However,
GC remains to be the second most common cause of cancerrelated death, with >700,000 deaths/y.1 Lauren2,3 proposed a
histological classification of gastric adenocarcinoma into an
intestinal type, including papillary adenocarcinomas and welldifferentiated tubular adenocarcinomas, and a diffuse type,
including signet ring cell carcinomas and poorly differentiated
adenocarcinomas. Based on anatomic conditions, GC can also be
divided into 2 subtypes: gastric cardia cancer and noncardia GC,
with the former referring to cancers of the top portion of the
stomach and the latter referring to cancers in the other areas of
the stomach. Noncardia cancer is commonly associated with the
Helicobacter pylori infection. There was no overall association
between gastric cardia cancer and H pylori infection, whereas
a positive association was observed in high-risk populations.4
The etiology and pathophysiology of GC is not fully
understood. It is well established that gastric carcinogenesis
is a complex multifactorial and multistage process. Previous
studies have identified several risk factors that might
contribute to gastric carcinogenesis including H pylori
infection,5 inadequate vitamin C uptake,6 smoking,7 high salt
intake,8 and low vegetable intake.9 Meanwhile, multiple
genetic variants and different genetic pathways have been
identified to contribute to GC risk,10 suggesting that genetic
factors play important roles in GC susceptibility. Many
studies have been conducted to search for susceptibility
genes for GC, such as Interleukin-1, Interleukin-8, Glutathione S-Transferase, and Cytochrome P450 2E1.11
E-cadherin glycoprotein, encoded by E-cadherin gene
(CDH1), is involved in the establishment and maintenance of
intercellular adhesion.12 In vitro studies found that the A allele
of C-160A could decrease the transcriptional efficiency of
CDH1 by approximately 70%, suggesting that the A allele
could potentially increase susceptibility to GC.13 Many previous studies investigated the association of the genetic variants,
C-160A (rs16260) in CDH1 with GC risk, with conflicting
results reported. Several meta-analyses have also been conducted to examine the association of C-160A with GC.
Although all of them found no significant association of
C-160A with GC, subgroup analysis by ethnic groups reported
inconsistent findings (Table 1). In addition to the widely
studied genetic variant C-160A, the association between GC
and many other less-studied genetic variants in CHD1 has also
been explored in many studies, with inconsistent results
reported. Meanwhile, promoter hypermethylation of CDH1
has also been studied for its effect on GC susceptibility, with
inconsistent results found. Therefore, in this study we
performed updated meta-analyses to assess the genetic and
epigenetic effect of CDH1 on GC risk. Since GC is a complex
disease, a single-nucleotide polymorphism (SNP) may only
confer a small or marginal individual effect on GC susceptibility. Studies focused on individual genetic variant may be
less powerful in detecting small genetic effect and fail to
capture the joint contribution from multiple genetic variants.
We therefore conducted a gene-based analysis to examine the
cumulative effect of multiple genetic variants in CDH1 on GC
risk.
METHODS
Search Strategy and Study Selection
From January to May 2014, we did an extensive
literature search in PubMed, Cochrane Library, Google
2 |
www.md-journal.com
•
Volume 93, Number 19, October 2014
Scholar, Web of Knowledge, and HuGE, a navigator for
human genome epidemiology, for candidate gene studies on
the association of GC with genetic variants in and promoter
hypermethylation of CDH1. Details of keywords used in the
literature search can be found in the supplementary file
(http://links.lww.com/MD/A54, Key words used in the literature search). We used the following inclusion criteria in
determining study eligibility: studies on human subjects,
outcomes of interest include GC, and report of genotype data
of individual genetic variants in (or methylation status of)
CDH1 in participants with and without GC (or providing
odds ratios [ORs] and their variances). All potentially
relevant publications were retrieved and further evaluated for
inclusion. We also hand-searched references of all relevant
publications for additional studies missed by the database
search. Only studies published in the English language were
included in our analysis. Two authors (H.J. and J.Y.Y.)
performed the search independently. Disagreement over
eligibility of a study was resolved by discussion until a
consensus was reached.
Data Extraction
Two reviewers (J.Y. and L.M.) independently extracted
the following data according to a prespecified protocol: first
author’s name, year of publication, characteristics of the
study participants (sample size, number of GC patients, and
number of participants in the control group, race/country of
participants), genotype or methylation status data for subjects
with and without GC (or OR and the corresponding
variances), and the genetic model used (additive, allelic,
dominant, or recessive). Discrepancies were resolved by
discussion, and extracted data were entered into a computerized spreadsheet for analysis.
Statistical Analysis
We used the OR as a measure of the association between
the genetic variants in and methylation status of CDH1 and
GC. We used random effects models to calculate ORs and the
corresponding 95% confidence intervals (CIs). The inverse of
the variance of each study was used as the weight for that
study. We used forest plots to graphically represent the
calculated pooled ORs and their 95% CIs. Each study was
represented by a square in the plot, the area of which is
proportional to the weight of the study. The overall effect
from the meta-analysis is represented by a diamond, with its
width representing the 95% CI for the estimate. We used I2
for assessment of between-study heterogeneity, and publication bias was assessed using funnel plot and Egger test, and a
P value <0.20 was considered statistically significant.
We performed an updated meta-analysis for the association of C-160A with GC, and also conducted meta-analysis
for association of other genetic variants in CDH1 with GC,
when there are multiple eligible studies for the genetic
variants. Otherwise, we compiled the results of the association with GC for genetic variants that appear in single
studies. We also analyzed the association between C-160A
and subtypes of GC (cardia and noncardia GCs and intestinal
and diffuse GCs). Meta-analyses were conducted when there
were multiple studies for the analysis of each subtype.
In order to assess the cumulative association of CDH1
with GC, we conducted a gene-based analysis using the P
values for the association of individual genetic variants in
CDH1 with GC, calculated from our meta-analyses and/or
ã 2014 Lippincott Williams & Wilkins
Medicine
•
Volume 93, Number 19, October 2014
Association Between CDH1 and Gastric Cancer
TABLE 1. Summary of Previous Meta-Analyses on the Association of CDH1 C-160A Polymorphism With Risk of GC
Authors
No. of
Studies
Main Genetic
Model
Result
GC
Wang et al14
Li et al15
19*
16†
Multiple
Multiple
Cui et al16
14
Recessive
Chen et al17
17
Multiple
Loh et al18
14
AA vs AB+BB‡
Gao et al19
Wang et al20
10
11
Dominant
Dominant
4
NA
Recessive
Multiple
No significant association observed
No significant association observed
3
NA
Recessive
Multiple
No significant association observed
No significant association observed
6
NA
4
Recessive
Multiple
Dominant
No significant association observed
No significant association observed
No significant association in the overall population, Asians or whites
6
NA
4
Recessive
Multiple
Dominant
No significant association observed
No significant association observed
No significant association in the overall population, Asians or whites
Cardia GC
Cui et al16
Chen et al17
Noncardia GC
Cui et al16
Chen et al17
Intestinal GC
Cui et al16
Chen et al17
Gao et al19
Diffuse GC
Cui et al16
Chen et al17
Gao et al19
No significant association in the overall population, Asians or whites
No significant association in the overall population for all genetic
models used
In Asians, A-allele conferred a decreased risk
In whites, no significant association
No significant association in the overall population
Significant association found in Asians but not in whites
No significant association in the overall population for all genetic
models used
In whites, A-allele conferred an increased risk
In Asians, no significant association
No significant association in the overall population
Significant association found in Asians but not in whites
No significant association in the overall population, Asians or whites
No significant association in the overall population
Significant associations were found in both Asians and whites, but they
are in opposite direction
GC ¼ gastric cancer.
*One article has 2 studies21 and another article has 3 studies.22 Therefore, the 19 studies are from 16 published articles.
One article has 3 studies.22 Therefore, these 16 studies are from 14 published articles.
‡
A is the major allele and B is the minor allele.
†
from published literature. We used 4 popular P value
combination methods to assess this cumulative association:
the Fisher method,23 the Simes method,24 the modified
inverse normal method,25 and the truncated product method
(TPM).26,27 A detailed description of the 4 methods has been
reported elsewhere.26,28 We used 100,000 simulations to
estimate the combined P value for TPM because the
individual P values are most likely to be dependent.
Finally, we performed meta-analysis to examine the
effect of promoter hypermethylation of CDH1 on susceptibility of GC.
Sensitivity Analysis
We performed separate meta-analyses by excluding
studies in which genotype in the control group did not meet
Hardy–Weinberg equilibrium (HWE). We also performed
meta-analysis separately for individual ethnic groups/countries of origin (Asian and whites).
As a research using systematic review and metaanalysis, ethical approval of this study is not required. This
ã 2014 Lippincott Williams & Wilkins
work was reported according to the PRISMA guidelines.29
Meta-analysis was performed using Stata 11.2 (StataCorp
LP, College Station, TX). All other analyses were performed
using SAS version 9.3 (SAS Institute Inc, Cary, NC), R
(www.R-project.org), and Matlab 8.1.0.604 (The MathWorks,
Inc, Natick, MA). A P value <0.05 was considered statistically significant.
RESULTS
Literature Search and Eligible Studies
Figure 1 is the flow diagram showing the selection of
studies to be included in our analysis. Using our predefined
search strategy, we identified a total of 311 potential publications
through our initial search. After screening the abstracts of these
studies, 221 were excluded either because they were irrelevant,
not about human subjects, not genetic studies, or not published in
English. The remaining 90 studies were retrieved for more
detailed evaluations, which excluded an additional 62 studies
because they were irrelevant, there were not sufficient data, the
www.md-journal.com
| 3
Medicine
Jing et al
outcome of interest was not GC, or they were meta-analyses or
review studies. This left 28 potentially relevant publications (with
31 studies) to be included in our analysis. A further review of the
references of these studies and review articles identified 3 more
studies. Further exploration of the data from these studies
excluded 1 more study because of insufficient data. A total of 33
studies from 30 published articles met the eligibility criteria and
were included in our analyses.21–57
All qualified publications were published since 2002 and
had sample sizes ranging from 14 to 1197 (Table 2). Prevalence of GC ranged from 16% to 88%. Of these 33 studies, 22
studies reported association results for C-160A, 4 studies for
rs1801552, rs3743674, and rs5030625, and 3 studies for
rs1801026. Two studies investigated the association of GC
•
Volume 93, Number 19, October 2014
with rs2010724, 2296-616G>C, and rs33964119. The combined study population included 9593 participants in the metaanalysis of C-160A, 1563 of rs1801552, 1993 of rs3743674,
2048 of rs5030625, 1373 of rs1801026, 783 of rs2010724, 771
of 2296-616G>C, and 447 of rs33964119. In addition to the 8
genetic variants included in the respective meta-analyses, the
association between GC and 17 additional genetic variants in
CDH1 was reported in individual studies (or calculated based
on individual studies). These results, together with results
obtained from our mea-analyses, were included in our genebased analysis. Moreover, the association of the promoter
hypermethylation of CDH1 with GC has been examined in 8
studies, and meta-analysis was performed to explore the effect
of promoter hypermethylation on GC risk. DNA methylation
PubMed
Cochrane Library
Google Scholar
Web of Science
HuGE
219
2
111
30
65
Filtering doubles
311
Screening titles and abstracts
Papers retrieved for more detailed evaluation (n = 90)
Potential studies to be included (n = 28, studies = 31)
Critical appraisal
Additional studies
included by reviewing
of references (3)
Papers excluded (n = 221)
- Not genetic studies
- Not in English
- Not human studies
- Irrelevant
Papers excluded (n = 62)
Reasons
- Not about GC
- Review/meta-analysis
- Not on CDH1
- Not relevant data
Studies excluded (1)
Reasons
- No sufficient data
Final studies included (n = 30, studies = 33)
FIGURE 1. Flow diagram of the selection process of the studies included in the meta-analyses. Please see the Methods section for
additional details. CDH1 ¼ E-cadherin gene, GC ¼ gastric cancer.
4 |
www.md-journal.com
ã 2014 Lippincott Williams & Wilkins
Medicine
•
Volume 93, Number 19, October 2014
Association Between CDH1 and Gastric Cancer
was measured similarly across studies (ie, bisulfate treatment
followed by methylation-specific polymerase chain reaction).
of other SNPs is not meaningful due to the low number of
studies included in the corresponding meta-analysis.
Assessment of Publication Bias
Association of C-160A With GC
Both funnel plot and Egger test were used to assess
publication bias. There was no evidence of publication bias for
the meta-analysis of C-160A (P ¼ 0.380, Figure 2). We found
no evidence of publication bias for the meta-analysis of
rs1801552, rs3743674, rs5030625, rs1801026 (all Ps >0.38).
There was some evidence of publication bias for the metaanalysis for promoter hypermethylation of CDH1 (P ¼ 0.128,
Figure 3). Assessment of publication bias for the meta-analysis
We calculated the association between the C-160A in
CDH1 and GC assuming 4 different genetic models (additive, recessive, dominant, and allelic). Due to space limitations, we only present the results using an additive model.
Results obtained using other models can be found in the
supplementary materials.
Of the 22 studies included in our meta-analysis, 10
showed significant association between C-160A and GC
TABLE 2. Characteristics of Studies Included in the Meta-Analyses
Study
C-160A (rs16260)
Humar et al22
Wu et al30
Pharoah et al31 (a)
Pharoah et al31 (b)
Pharoah et al31 (c)
Kuraoka et al32
Park et al33
Shin et al34
Lu et al35
Cattaneo et al36
Medina-Franco et al37
Yamada et al21
Jenab et al38
Zhang et al39 (a)
Zhang et al39 (b)
Zhang et al40
Corso et al41
Al-Moundhri et al42
Borges et al43
Zhan et al44
Menbari et al45
Chu et al57
Other genetic variants in CDH1
Humar et al22
Shin et al34
Zhang et al56
Yamada et al21
Jenab et al38
Nasri et al46
Zhang et al39
Zhang et al40
Al-Moundhri et al42
Borges et al43
Li et al47
Zhan et al44
Promoter hypermethylation of CDH1
To et al48
Oue et al49
Tan et al50
Zhang et al51
Poplawski et al52
Tahara et al53
Yu et al54
Nomura et al55
ã 2014 Lippincott Williams & Wilkins
Ethnicity
Total Sample Size
No. of Cases
No. of Controls
Whites
Asian
Whites
Whites
Whites
Asian
Asian
Asian
Asian
Whites
Mixed
Asian
Whites
Asian
Asian
Asian
Whites
Asian
Mixed
Asian
Whites
Asian
123
397
241
174
484
196
438
170
467
353
117
440
1195
489
1197
582
820
362
112
779
306
241
53
201
148
132
153
106
292
28
206
107
39
148
245
96
572
239
412
174
58
354
144
107
70
196
93
42
331
90
146
142
261
246
78
292
949
393
625
343
408
166
51
361
162
134
Asian
Asian
Asian
Asian
Whites
Whites
Asian
Asian
Asian
Mixed
Asian
Asian
123
170
206
440
1195
234
1197
582
362
112
460
779
53
28
101
148
245
134
572
239
192
58
230
387
70
142
105
292
950
100
625
343
170
54
230
392
Asian
Asian
Asian
Asian
Whites
Asian
Asian
Asian
62
85
14
78
52
305
180
527
52
75
4
47
27
125
92
115
10
10
10
31
25
180
88
412
www.md-journal.com
| 5
Medicine
Jing et al
•
Volume 93, Number 19, October 2014
Begg funnel plot with pseudo 95% confidence limits
1
Log (OR)
0
–1
–2
0
0.2
0.4
Standard error of log (OR)
FIGURE 2. Funnel plot for meta-analysis of C-160A in CDH1. The x-axis is the standard error of the log-transformed OR (log [OR]),
and the y-axis is the log (OR). The horizontal line in the figure represents the overall estimated log (OR). The 2 diagonal lines
represent the pseudo 95% confidence limits of the effect estimate. CDH1 ¼ E-cadherin gene, log (OR) ¼ log-transformed OR,
OR ¼ odds ratio.
(Table 3). Specifically, 3 studies31,33 indicated that compared
with CC carriers, those carrying each additional copy of the
A allele had increased risk of GC, whereas the other 7
studies reported decreased risk. Our meta-analysis indicates
no significant association of C-160A with GC (OR ¼ 0.88;
95% CI, 0.71–1.08; P ¼ 0.215; Table 3,Figure 4). We found
no significant association using different genetic models
(Figures S1–S3 [http://links.lww.com/MD/A44, http://links.
lww.com/MD/A45, and http://links.lww.com/MD/A46], Forest plots for meta-analysis of C-160A in CDH1; Tables S1–
S3 [http://links.lww.com/MD/A47, http://links.lww.com/MD/
A48, and http://links.lww.com/MD/A49], Meta-analysis of
the association of C-160A in CDH1 with GC).
Association of C-160A With Subtypes of GC
We found significant association of C-160A with cardia
(OR ¼ 0.21; 95% CI, 0.11–0.41; 2.60 10 6), intestinal (OR
¼ 0.66; 95% CI, 0.49–0.90; P ¼ 0.008), and diffuse GCs
(OR ¼ 0.57; 95% CI, 0.40–0.82; P ¼ 0.002). The association
Begg funnel plot with pseudo 95% confidence limits
6
Log (OR)
4
2
0
–2
0
0.5
1
1.5
2
Standard error of log (OR)
FIGURE 3. Funnel plot for meta-analysis of promoter hypermethylation of CDH1. The x-axis is the standard error of the logtransformed OR (log [OR]), and the y-axis is the log (OR). The horizontal line in the figure represents the overall estimated log (OR).
The 2 diagonal lines represent the pseudo 95% confidence limits of the effect estimate. CDH1 ¼ E-cadherin gene, log (OR) ¼ logtransformed OR, OR ¼ odds ratio.
6 |
www.md-journal.com
ã 2014 Lippincott Williams & Wilkins
Medicine
•
Volume 93, Number 19, October 2014
Association Between CDH1 and Gastric Cancer
TABLE 3. Meta-Analysis of the Association of GC and C-160A in CDH1*
Study
Case
Control
Humar et al22
Wu et al30
Pharoah et al31 (a)
Pharoah et al31 (b)
Pharoah et al31 (c)
Kuraoka et al32
Park et al33
Shin et al34
Lu et al35
Cattaneo et al36
Medina-Franco et al37
Yamada et al21
Jenab et al38
Zhang et al39 (a)
Zhang et al39 (b)
Zhang X et al40
Corso et al41
Al-Moundhri et al42
Borges et al43
Zhan et al44
Menbari et al45
Chu et al57
Total
53
201
148
132
153
106
292
28
206
107
39
148
245
96
572
239
412
174
58
354
144
107
4014
70
196
93
42
331
90
146
142
261
246
78
292
949
393
625
343
408
166
51
361
162
134
5579
OR (95% CI)
0.86
1.29
1.65
2.59
0.57
0.90
1.62
0.26
0.82
0.58
0.88
0.56
0.34
0.36
0.97
0.63
0.97
1.19
1.46
1.05
1.21
1.24
0.88
P
(0.68–1.08)
(0.87–1.90)
(1.23–2.22)
(1.72–3.91)
(0.46–0.70)
(0.66–1.24)
(1.23–2.14)
(0.12–0.57)
(0.65–1.04)
(0.44–0.76)
(0.59–1.32)
(0.42–0.75)
(0.28–0.40)
(0.26–0.49)
(0.83–1.14)
(0.49–0.82)
(0.83–1.12)
(0.92–1.54)
(0.96–2.22)
(0.87–1.27)
(0.85–1.71)
(0.91–1.67)
(0.71–1.08)
0.196
0.208
0.001
5.50 10
3.00 10
0.519
0.001
0.001
0.100
8.46 10
0.538
8.82 10
1.33 10
6.97 10
0.716
4.34 10
0.645
0.177
0.074
0.602
0.290
0.172
0.215
6
7
5
5
37
11
4
P value in bold indicates statistical significance. CI ¼ confidence interval, GC ¼ gastric cancer, OR ¼ odds ratio.
*Assuming an additive model with reference genotype CC.
of C-160A with noncardia GC is of bottom line significance
(OR ¼ 0.65; 95% CI, 0.42–1.01; P ¼ 0.054; Table 4).
individual studies because, due to inadequate number of
studies, it is not possible to determine whether there is
selective reporting that can lead to inflation of the P values.
Association of Other Genetic Variants With GC
Our meta-analysis of the less-studied genetic variants in
CDH1 found no significant association with GC (Table 5).
However, several genetic variants that appeared in single
studies showed significant association with GC. Specifically,
7 genetic variants from a single study38 showed strong
association with GC, whereas 1 other genetic variant
(rs1125557) from another individual study46 also exhibited
significant association with GC (P ¼ 7.53 10 5).
Gene-Based Analysis
To examine the cumulative association of multiple
genetic variants in CDH1 with GC, we performed a genebased analysis using all the P values we obtained for each
individual genetic variant in CDH1. Additionally, we examined whether the association varies in meta-studies only
(including only results for genetic variants covered in metaanalyses) and in individual studies only (including only
results for genetic variants that appeared in single studies).
Our gene-based analysis indicated a significant association
between the genetic variants in CDH1 and GC (all Ps
<10 5). The association held when pooling results from only
individual-studies, but disappeared when only results from
meta-studies were included, indicating that the observed
gene-based association was driven mainly by results from the
less-studied genetic variants (Table S4 [http://links.lww.com/
MD/A50], Gene-based analysis with GC). We would like to
caution against over interpretation of the results from
ã 2014 Lippincott Williams & Wilkins
Association of Promoter Hypermethylation of
CDH1 With GC
Our meta-analysis of 8 studies showed very strong and
significant association of promoter hypermethylation of CDH1
with GC (OR ¼ 12.23; 95% CI, 8.80–17.00; P ¼ 1.42 10 50;
Table 6, Figure 5). More specifically, of the 8 studies, 1
study50 showed no association of promoter hypermethylation
of CDH1 with GC, probably due to insufficient statistical
power resulting from limited sample size (n ¼ 14), and
another study48 showed marginal association (P ¼ 0.080). All
the other 6 studies indicated that promoter hypermethylation
of CDH1 is significantly associated with increased risk of GC.
Sensitivity Analysis
There were 4 studies in which the genotype in the
control group did not meet HWE.32,40,44,45 After excluding
these 4 studies, our meta-analysis again indicated no significant association of C-160A with GC (OR ¼ 0.87; 95% CI,
0.68–1.11; P ¼ 0.261). There were 12 studies from 11
articles and 8 studies from 6 articles examining the association of C-160A with GC in Asian and white participants,
respectively. Our meta-analysis based on these studies found
no significant association of C-160A with GC in Asian
(OR ¼ 0.82; 95% CI, 0.66–1.02; P ¼ 0.075) and white
participants (OR ¼ 0.94; 95% CI, 0.60–1.47; P ¼ 0.793;
Table S5 [http://links.lww.com/MD/A51], Meta-analysis of
www.md-journal.com
| 7
Medicine
Jing et al
Study
•
Volume 93, Number 19, October 2014
ES (95% CI)
Weight, %
Humar et al22
1.29 (0.87, 1.90)
4.30
Wu et al30
0.86 (0.68, 1.08)
4.77
Pharoah et al31 (a)
1.65 (1.23, 2.22)
4.60
Pharoah et al31 (b)
2.59 (1.72, 3.91)
4.23
Pharoah et al31 (c)
0.57 (0.46, 0.70)
4.80
Kuraoka et al32
0.90 (0.66, 1.24)
4.53
Park et al33
1.62 (1.23, 2.14)
4.65
Shin et al34
0.26 (0.12, 0.57)
3.01
Lu et al35
0.82 (0.65, 1.04)
4.76
Cattaneo et al36
0.57 (0.44, 0.76)
4.65
Medina-Franco et al37
0.88 (0.59, 1.32)
4.24
Yamada et al21
0.56 (0.41, 0.75)
4.60
Jenab et al38
0.34 (0.28, 0.40)
4.91
Zhang et al39 (a)
0.36 (0.26, 0.49)
4.56
Zhang et al39 (b)
0.97 (0.83, 1.14)
4.93
Zhang et al40
0.63 (0.49, 0.82)
4.71
Corso et al41
0.96 (0.83, 1.12)
4.94
Al-Moundhri et al42
1.19 (0.92, 1.54)
4.71
Borges et al43
1.46 (0.96, 2.22)
4.21
Zhan et al44
1.05 (0.87, 1.27)
4.87
Menbari et al45
1.21 (0.85, 1.71)
4.44
Chu et al57
1.24 (0.91, 1.67)
4.58
Overall (I2) = 93.0%, P = 0.000
0.22 (0.71, 1.08)
100.00
Note: Weights are from random effects analysis
0.2
1
3
FIGURE 4. Forest plot for meta-analysis of C-160A in CDH1. Each study is represented by a square, whose area is proportional to the
weight of the study. The overall effect from meta-analysis is represented by a diamond whose width represents the 95% CI for the
estimated OR. CDH1 ¼ E-cadherin gene, CI ¼ confidence interval, ES ¼ effect size, OR ¼ odds ratio.
C-160A by ethnic groups). These results do not change after
excluding studies that did not meet HWE.
There were 1 study30 and 2 studies38,41 examining the
association of C-160A with cardia GC in Asian and white
participants, respectively. A significant association was found
between C-160A and GC in both Asian (OR ¼ 0.12; 95% CI,
0.07–0.22; P ¼ 9.71 10 13) and white participants (OR
¼ 0.27; 95% CI, 0.14–0.55; P ¼ 2.50 10 4; Table S6 [http://
links.lww.com/MD/A52], Association with cardia and noncardia GC by ethnic groups). There were 2 studies examining
the association of C-160A with noncardia GC in Asian30,35 and
white participants.38,42 Our meta-analysis found significant
association in Asian (OR ¼ 0.80; 95% CI, 0.68–0.95; P ¼ 0.011)
but not in white participants (OR ¼ 0.53; 95% CI, 0.20–1.37;
P ¼ 0.190). There were 3 studies30,33,44 and 2 studies38,41
examining the association of C-160A with intestinal GC in
Asian and white participants, respectively. Our meta-analysis
based on these studies found no significant association in
Asian (OR ¼ 0.72; 95% CI, 0.50–1.02; P ¼ 0.066) and white
participants (OR ¼ 0.52; 95% CI, 0.23–1.14; P ¼ 0.103;
Table S7 [http://links.lww.com/MD/A53], Association with
intestinal and diffuse GC by ethnic groups). There were 6
studies and 3 studies examining the association of C-160A
with diffuse GC in Asian and white participants, respectively.
Our meta-analysis based on these studies indicated significant
association of C-160A with diffuse GC in Asian participants
(OR ¼ 0.55; 95% CI, 0.35–0.87; P ¼ 0.011) but not in white
participants (OR ¼ 0.49; 95% CI, 0.22–1.09; P ¼ 0.081;
Table S7 [http://links.lww.com/MD/A53], Association with
8 |
www.md-journal.com
intestinal and diffuse GC by ethnic groups). However, after
excluding the studies in which genotype in the control group
did not meet HWE (n ¼ 2), the association in Asian participants is no longer statistically significant (OR ¼ 0.64; 95%
CI, 0.36–1.13; P ¼ 0.126).
DISCUSSION
In this study, we conducted an extensive literature search
for publications on the association of GC with genetic variants
in and promoter hypermethylation of CDH1. We provided an
updated meta-analysis on the widely studied genetic variant
C-160A. Our analysis showed that C-160A is not associated
with GC, either in the overall population, or in Asian or white
participants. However, within a very limited set of articles that
evaluated subtypes of GC, we found significant association of
C-160A with cardiac, intestinal, and diffuse GC. We found
that the promoter hypermethylation of CDH1 is strongly
associated with GC, indicating potential epigenetic influences
in the carcinogenesis and development of GC. To the best of
our knowledge, this is the most comprehensive meta-analysis
on the association of GC with a number of genetic variants in
CDH1, and with promoter methylation of CDH1.
In the meta-analysis of C-160A, we identified significant
heterogeneity between the studies included for analysis
(I2 ¼ 93.0%; 95% CI, 90.6%–94.7%). Identifying the source of
heterogeneity is challenging with limited information provided
in many studies. Variation in patient characteristics might be
an important source of heterogeneity. Some studies used
ã 2014 Lippincott Williams & Wilkins
Medicine
•
Volume 93, Number 19, October 2014
Association Between CDH1 and Gastric Cancer
TABLE 4. Association of C-160A With Subtypes of GC
Study
Cardia GC
Wu et al30
Jenab et al38
Corso et al41
Total
Noncardia GC
Wu et al30
Lu et al35
Jenab et al38
Corso et al41
Total
Intestinal GC
Wu et al30
Park et al33
Jenab et al38
Corso et al41
Borges et al43
Zhan et al44
Total
Diffuse GC
Humar et al22
Wu et al30
Park et al33
Shin et al34
Medina-Franco et al37
Jenab et al38
Zhang et al40
Corso et al41
Borges et al43
Zhan et al44
Chu et al57
Total
Case
Control
OR (95% CI)
24
69
62
155
196
257
408
861
0.12
0.39
0.19
0.21
(0.07–0.22)
(0.29–0.52)
(0.14–0.27)
(0.11–0.41)
9.71 10
4.92 10
3.56 10
2.60 10
177
206
128
350
861
196
261
506
408
1371
0.79
0.82
0.32
0.86
0.65
(0.62–1.00)
(0.65–1.04)
(0.26–0.41)
(0.73–1.00)
(0.42–1.01)
0.048
0.100
6.58 10
0.054
0.054
99
165
96
285
33
257
935
196
146
372
408
51
361
1534
0.49
0.95
0.34
0.77
0.94
0.79
0.66
(0.37–0.65)
(0.70–1.30)
(0.26–0.45)
(0.65–0.91)
(0.59–1.51)
(0.64–0.97)
(0.49–0.90)
1.42 10
0.751
1.70 10
0.002
0.810
0.024
0.008
53
92
127
28
39
93
239
127
25
52
107
982
70
196
146
142
78
368
343
408
51
361
134
2297
1.29
0.46
0.87
0.26
0.88
0.29
0.63
0.32
0.88
0.26
1.24
0.57
(0.87–1.90)
(0.34–0.62)
(0.63–1.21)
(0.12–0.57)
(0.59–1.32)
(0.22–0.39)
(0.49–0.82)
(0.25–0.41)
(0.53–1.45)
(0.19–0.37)
(0.91–1.67)
(0.40–0.82)
0.208
3.75 10
0.410
0.0001
0.538
4.49 10
4.43 10
2.13 10
0.614
1.02 10
0.172
0.002
P
13
10
22
6
21
6
15
7
17
4
18
13
CI ¼ confidence interval, GC ¼ gastric cancer, OR ¼ odds ratio.
matched controls (eg, age and sex matched),21,22,35,37,38,43,46
whereas most other studies did not perform matching. Other
patient characteristics, such as smoking behavior, H pylori
infection, and tumor location, can also contribute to the
heterogeneity of the included studies in the meta-analyses.
Of the 24 less-studied genetic variants in CDH1, our
analysis found multiple genetic variants showing significant
association with GC. Specifically, 1 study by Jenab et al38
reported findings for 7 less-studied SNPs and all of them
showed significant association with GC. Another study46
showed that rs1125557 was significantly associated with GC
(Table 5). The gene-based analysis indicated that these lessstudied genetic variants other than C-160A cumulatively
confer significant genetic susceptibility of GC (Table S4
[http://links.lww.com/MD/A50], Gene-based analysis with
GC). Realizing that the observed gene-based association might
be driven by the results reported in the study by Jenab et al,38
in sensitivity analysis we dropped that study from the genebased analysis and still observed significant gene-based
association (all Ps <0.003). The SNP rs1125557 is in high
linkage disequilibrium (LD) with C-160A (D0 ¼ 1, SNP
annotation and proxy search, http://www.broadinstitute.org/
mpg/snap/ldsearchpw.php). Given the high LD, we feel that
the significant finding was probably because of the small
ã 2014 Lippincott Williams & Wilkins
sample size based on which the result was reported.46 Studies
on functional outcomes of these less-studied genetic variants
in CDH1 are scarce, and further studies are needed to
elucidate whether and how they function in influencing
disease susceptibility.
DNA methylation is the most extensively studied epigenetic modification, and plays an important role in regulating
gene expression and cell differentiation. Aberrant DNA
methylation leads to silencing of tumor suppressor genes or
loss of oncogene repression, and therefore is an important
mechanism in the initiation and development of GC.58 The
precise molecular mechanism underlying the association of
promoter hypermethylation of CDH1 with GC remains to be
understood. A key challenge remains whether changes in
methylation are a cause or an effect of the pathological
process. Although some studies suggest that altered methylation in CDH1 might be involved in carcinogenesis of GC but
not development of GC,55 others indicate that accumulation of
aberrant methylation might be an important mechanism for
GC development.48 There are also studies indicating that the
accumulation of DNA methylation might be caused by
proliferative changes during tumor progression.49 Moreover,
CDH1 methylation seems to be age related,51 making it more
complicated to disentangle the exact role of methylation in
www.md-journal.com
| 9
Medicine
Jing et al
•
Volume 93, Number 19, October 2014
TABLE 5. Association of Other Genetic Variants in CDH1 With GC
Genetic Variant
Case
Control
rs5030625*
rs3743674*
rs2010724*
rs1801026*
rs1801552*
2296-616G>C*
rs33964119*
rs1078621
rs2276329
rs2276330
rs3785076
rs4076177
rs7188750
rs7203904
rs10673765
rs1125557
rs9282650
rs9931853
rs9932686
rs28372783
rs3833051
rs13689
rs17690554
1937-13T>C
885
766
325
643
574
320
208
241
245
244
243
244
245
245
134
134
134
134
134
572
572
373
354
107
1163
986
458
730
748
451
239
934
949
940
938
939
950
946
100
100
100
100
100
625
625
371
370
134
OR (95% CI)
0.81
0.73
0.83
1.05
0.78
0.76
1.26
0.39
0.23
0.30
0.14
0.38
0.28
0.32
1.38
1.58
1.25
1.32
1.52
0.90
0.93
1.06
1.02
1.29
(0.62–1.06)
(0.52–1.02)
(0.50–1.39)
(0.97–1.13)
(0.59–1.03)
(0.42–1.38)
(0.52–3.04)
(0.35–0.45)
(0.15–0.35)
(0.23–0.39)
(0.07–0.27)
(0.33–0.44)
(0.22–0.35)
(0.26–0.39)
(0.54–3.51)
(1.26–1.98)
(0.90–1.74)
(1.03–1.70)
(0.51–4.51)
(0.74–1.09)
(0.81–1.07)
(0.85–1.33)
(0.82–1.27)
(0.64–2.59)
P
0.128
0.065
0.484
0.254
0.083
0.368
0.610
9.32 10
7.96 10
1.26 10
5.49 10
1.24 10
4.41 10
1.15 10
0.497
7.53 10
0.184
0.027
0.450
0.283
0.321
0.575
0.868
0.481
46
12
18
9
41
24
29
5
P value in bold indicates statistical significance. CI ¼ confidence interval, GC ¼ gastric cancer, OR ¼ odds ratio.
*Indicates results obtained from meta-analysis; other results were from single studies.
the initiation and development of GC. More future large-scale
studies are needed that examine subjects at risk of developing
GC as well as subjects with GC to better elucidate whether
and how CDH1 promoter hypermethylation is implicated in
GC initiation and development.
Our study has some limitations. Since relevant data were
not available, our meta-analysis could not adjust for confounding
factors such as age, sex, smoking behavior, or H pylori infection.
First, future studies are needed to validate our results—
especially large consortium studies that provide control for
such confounding factors. Second, some meta-analyses were
based on few studies, and the gene-based analysis used some
results from individual studies. Third, sensitivity analyses by
ethnicity are limited because race information was not
available in all studies. Fourth, due to the limited number of
studies included in some of the meta-analyses, we could not
test publication bias for them. This might lead to bias in the
resulting data, and subsequently influence the validity of the
gene-based analysis. Finally, there are other types of genetic
variations that are not included in our study, such as copy
number variation that was recently reported to be associated
with GC.59,60
In summary, in this study, we performed meta-analyses to
analyze the genetic and epigenetic effect of CDH1 on GC risk.
We found no significant association of the widely studied
genetic variant C-160A with GC. However, a limited number
of studies suggest that C-160A may be associated with
subtypes of GC in different ethnic groups, and we identified
some other genetic variants showing significant association
with GC. Gene-based analysis indicated that the previously
studied variants cumulatively influence GC susceptibility.
Meta-analysis on the promoter hypermethylation of CDH1
TABLE 6. Association of Promoter Hypermethylation of CDH1 With GC
Study
Case
Control
To et al48
Oue et al49
Tan et al50
Zhang et al51
Poplawski et al52
Tahara et al53
Yu et al54
Nomura et al55
Total
52
75
4
47
27
125
92
115
412
10
10
10
31
25
180
88
412
586
OR (95% CI)
13.25
25.26
9.00
8.94
7.35
9.29
763.63
8.83
12.23
(0.74–238.44)
(1.43–446.75)
(0.29–275.56)
(2.39–33.51)
(2.16–25.04)
(5.27–16.37)
(45.16–13000.00)
(5.19–15.02)
(8.80–17.00)
P
0.080
0.028
0.208
0.001
0.001
6.40 10
0.004
8.88 10
1.42 10
15
16
50
P value in bold indicates statistical significance. CI ¼ confidence interval, GC ¼ gastric cancer, OR ¼ odds ratio.
10 |
www.md-journal.com
ã 2014 Lippincott Williams & Wilkins
Medicine
•
Volume 93, Number 19, October 2014
Association Between CDH1 and Gastric Cancer
Study
OR (95% CI)
Weight, %
To et al48
13.25 (0.74, 238.44)
2.09
Oue et al49
25.26 (1.43, 446.75)
1.63
Tan et al50
9.00 (0.29, 275.56)
0.90
Zhang et al51
8.94 (2.39, 33.51)
7.60
Poplawski et al52
7.35 (2.16, 25.03)
7.76
Tahara et al53
9.29 (5.27, 16.37)
35.09
Yu et al54
763.63 (45.16, 12912.20)
Nomura et al55
Overall (I 2 = 40.3%, P = 0.110)
0.1
1
0.40
8.83 (5.19, 15.02)
44.53
12.23 (8.80, 17.00)
100.00
10
FIGURE 5. Forest plot for meta-analysis of promoter hypermethylation of CDH1. Each study is represented by a square whose area is
proportional to the weight of the study. The overall effect from meta-analysis is represented by a diamond whose width represents
the 95% CI for the estimated OR. CDH1 ¼ E-cadherin gene, CI ¼ confidence interval, OR ¼ odds ratio.
suggests that epigenetics also plays a critical role in the
carcinogenesis of GC. Future studies with large sample sizes
that control confounding risk factors and/or intensively
interrogate CpG sites in CDH1 are needed to validate the
results found in this study and explore additional epigenetic
loci that affect GC risk.
REFERENCES
1. Ferlay J, Shin HR, Bray F, et al. Estimates of worldwide burden of cancer
in 2008: GLOBOCAN 2008. Int J Cancer. 2010;127:2893–2917.
2. Lauren P. Histogenesis of intestinal and diffuse types of gastric
carcinoma. Scand J Gastroenterol Suppl. 1991;180:160–164.
3. Lauren P. The two histological main types of gastric carcinoma:
diffuse and so-called intestinal-type carcinoma. An attempt at a histoclinical classification. Acta Pathol Microbiol Scand. 1965;64:31–49.
4. Cavaleiro-Pinto M, Peleteiro B, Lunet N, et al. Helicobacter pylori
infection and gastric cardia cancer: systematic review and metaanalysis. Cancer Causes Control. 2011;22:375–387.
5. Schistosomes, liver flukes and Helicobacter pylori. IARC Working
Group on the Evaluation of Carcinogenic Risks to Humans. Lyon,
7–14 June 1994. IARC Monogr Eval Carcinog Risks Hum.
1994;61:1–241.
10. Hamilton JP, Meltzer SJ. A review of the genomics of gastric
cancer. Clin Gastroenterol Hepatol. 2006;4:416–425.
11. Gianfagna F, De Feo E, van Duijn CM, et al. A systematic review
of meta-analyses on gene polymorphisms and gastric cancer risk.
Curr Genomics. 2008;9:361–374.
12. Graziano F, Humar B, Guilford P. The role of the E-cadherin gene
(CDH1) in diffuse gastric cancer susceptibility: from the laboratory
to clinical practice. Ann Oncol. 2003;14:1705–1713.
13. Li LC, Chui RM, Sasaki M, et al. A single nucleotide polymorphism
in the E-cadherin gene promoter alters transcriptional activities.
Cancer Res. 2000;60:873–876.
14. Wang L, Wang G, Lu C, et al. Contribution of the -160C/A
polymorphism in the E-cadherin promoter to cancer risk: a metaanalysis of 47 case-control studies. PLoS One. 2012;7:e40219.
15. Li YL, Tian Z, Zhang JB, et al. CDH1 promoter polymorphism and
stomach cancer susceptibility. Mol Biol Rep. 2012;39:1283–1286.
16. Cui Y, Xue H, Lin B, et al. A meta-analysis of CDH1 C-160A
genetic polymorphism and gastric cancer risk. DNA Cell Biol.
2011;30:937–945.
17. Chen B, Zhou Y, Yang P, et al. CDH1 -160C> a gene
polymorphism is an ethnicity-dependent risk factor for gastric
cancer. Cytokine. 2011;55:266–273.
6. You WC, Zhang L, Gail MH, et al. Gastric dysplasia and gastric
cancer: Helicobacter pylori, serum vitamin C, and other risk factors.
J Natl Cancer Inst. 2000;92:1607–1612.
18. Loh M, Koh KX, Yeo BH, et al. Meta-analysis of genetic
polymorphisms and gastric cancer risk: variability in associations
according to race. Eur J Cancer. 2009;45:2562–2568.
7. Ji BT, Chow WH, Yang G, et al. The influence of cigarette smoking,
alcohol, and green tea consumption on the risk of carcinoma of the cardia
and distal stomach in Shanghai, China. Cancer. 1996;77:2449–2457.
19. Gao L, Nieters A, Brenner H. Meta-analysis: tumour invasionrelated genetic polymorphisms and gastric cancer susceptibility.
Aliment Pharmacol Ther. 2008;28:565–573.
8. Joossens JV, Hill MJ, Elliott P, et al. Dietary salt, nitrate and
stomach cancer mortality in 24 countries. European Cancer
Prevention (ECP) and the INTERSALT Cooperative Research
Group. Int J Epidemiol. 1996;25:494–504.
20. Wang GY, Lu CQ, Zhang RM, et al. The E-cadherin gene
polymorphism 160C->A and cancer risk: a HuGE review and
meta-analysis of 26 case-control studies. Am J Epidemiol.
2008;167:7–14.
9. Shimazu T, Wakai K, Tamakoshi A, et al. Association of vegetable
and fruit intake with gastric cancer risk among Japanese: a pooled
analysis of four cohort studies. Ann Oncol. 2014;25:1228–1233.
21. Yamada H, Shinmura K, Ikeda S, et al. Association between CDH1
haplotypes and gastric cancer risk in a Japanese population. Scand J
Gastroenterol. 2007;42:1479–1485.
ã 2014 Lippincott Williams & Wilkins
www.md-journal.com
| 11
Jing et al
Medicine
•
Volume 93, Number 19, October 2014
22. Humar B, Graziano F, Cascinu S, et al. Association of CDH1
haplotypes with susceptibility to sporadic diffuse gastric cancer.
Oncogene. 2002;21:8192–8195.
42. Al-Moundhri MS, Al-Khanbashi M, Al-Kindi M, et al. Association
of E-cadherin (CDH1) gene polymorphisms and gastric cancer risk.
World J Gastroenterol. 2010;16:3432–3436.
23. Fisher RA. Statistical Methods for Research Workers. Edinburgh:
Oliver and Boyd; 1932.
43. Borges Bdo N, Santos Eda S, Bastos CE, et al. Promoter polymorphisms
and methylation of E-cadherin (CDH1) and KIT in gastric cancer
patients from northern Brazil. Anticancer Res. 2010;30:2225–2233.
24. Simes RJ. An improved Bonferroni procedure for multiple tests of
significance. Biometrika. 1986;73:751–754.
25. Hartung J. A note on combining dependent tests of significance.
Biometrical J. 1999;41:849–856.
26. Sheng X, Yang J. Truncated product methods for panel unit root
tests. Oxford B Econ Stat. 2013;75:624–636.
27. Zaykin DV, Zhivotovsky LA, Westfall PH, et al. Truncated product
method for combining P-values. Genetic Epidemiol. 2002;22:
170–185.
28. Wang H, Dong S, Xu H, et al. Genetic variants in FTO associated
with metabolic syndrome: a meta- and gene-based analysis. Mol Biol
Rep. 2012;39:5691–5698.
29. Moher D, Liberati A, Tetzlaff J, et al. Preferred reporting items for
systematic reviews and meta-analyses: the PRISMA statement. PLoS
Med. 2009;6:e1000097.
30. Wu MS, Huang SP, Chang YT, et al. Association of the -160C—a
promoter polymorphism of E-cadherin gene with gastric carcinoma
risk. Cancer. 2002;94:1443–1448.
31. Pharoah PD, Oliveira C, Machado JC, et al. CDH1 c-160a promotor
polymorphism is not associated with risk of stomach cancer. Int J
Cancer. 2002;101:196–197.
32. Kuraoka K, Oue N, Yokozaki H, et al. Correlation of a single
nucleotide polymorphism in the E-cadherin gene promoter with
tumorigenesis and progression of gastric carcinoma in Japan. Int J
Oncol. 2003;23:421–427.
33. Park WS, Cho YG, Park JY, et al. A single nucleotide polymorphism
in the E-cadherin gene promoter-160 is not associated with risk of
Korean gastric cancer. J Korean Med Sci. 2003;18:501–504.
34. Shin Y, Kim IJ, Kang HC, et al. The E-cadherin -347G->GA
promoter polymorphism and its effect on transcriptional regulation.
Carcinogenesis. 2004;25:895–899.
35. Lu Y, Xu YC, Shen J, et al. E-cadherin gene C-160A promoter
polymorphism and risk of non-cardia gastric cancer in a Chinese
population. World J Gastroenterol. 2005;11:56–60.
36. Cattaneo F, Venesio T, Molatore S, et al. Functional analysis and
case–control study of -160C/A polymorphism in the E-cadherin gene
promoter: association with cancer risk. Anticancer Res.
2006;26:4627–4632.
37. Medina-Franco H, Ramos-De la Medina A, Vizcaino G, et al. Single
nucleotide polymorphisms in the promoter region of the E-cadherin
gene in gastric cancer: case–control study in a young Mexican
population. Ann Surg Oncol. 2007;14:2246–2249.
38. Jenab M, McKay JD, Ferrari P, et al. CDH1 gene polymorphisms,
smoking, Helicobacter pylori infection and the risk of gastric cancer
in the European Prospective Investigation into Cancer and Nutrition
(EPIC-EURGAST). Eur J Cancer. 2008;44:774–780.
39. Zhang B, Pan K, Liu Z, et al. Genetic polymorphisms of the Ecadherin promoter and risk of sporadic gastric carcinoma in Chinese
populations. Cancer Epidemiol Biomarkers Prev. 2008;17:
2402–2408.
40. Zhang XF, Wang YM, Ge H. et al. Association of CDH1 single
nucleotide polymorphisms with susceptibility to esophageal
squamous cell carcinomas and gastric cardia carcinomas. Dis
Esophagus. 2008;21:21–29.
41. Corso G, Berardi A, Marrelli D, et al. CDH1 C-160A promoter
polymorphism and gastric cancer risk. Eur J Cancer Prev.
2009;18:46–49.
12 |
www.md-journal.com
44. Zhan Z, Wu J, Zhang JF, et al. CDH1 gene polymorphisms, plasma
CDH1 levels and risk of gastric cancer in a Chinese population. Mol
Biol Rep. 2012;39:8107–8113.
45. Menbari MN, Rahmani SA, Zandi F, et al. Evaluation of E-cadherin
(CDH1) gene polymorphism related to gastric cancer in Kurdish
population. Life Sci J. 2013;10:212–216.
46. Nasri S, More H, Graziano F, et al. A novel diffuse gastric cancer
susceptibility variant in E-cadherin (CDH1) intron 2: a case control
study in an Italian population. BMC Cancer. 2008;8:138.
47. Li Y, Tang Y, Zhou R, et al. Genetic polymorphism in the 30 untranslated region of the E-cadherin gene is associated with risk of
different cancers. Mol Carcinog. 2011;50:857–862.
48. To KF, Leung WK, Lee TL, et al. Promoter hypermethylation of
tumor-related genes in gastric intestinal metaplasia of patients with
and without gastric cancer. Int J Cancer. 2002;102:623–628.
49. Oue N, Mitani Y, Motoshita J, et al. Accumulation of DNA
methylation is associated with tumor stage in gastric cancer. Cancer.
2006;106:1250–1259.
50. Tan SH, Ida H, Lau QC, et al. Detection of promoter hypermethylation
in serum samples of cancer patients by methylation-specific
polymerase chain reaction for tumour suppressor genes including
RUNX3. Oncol Rep. 2007;18:1225–1230.
51. Zhang KL, Sun Y, Li Y, et al. Increased frequency of CpG island
methylator phenotype and CDH1 methylation in a gastric cancer
high-risk region of China. Transl Oncol. 2008;1:28–35.
52. Poplawski T, Tomaszewska K, Galicki M, et al. Promoter
methylation of cancer-related genes in gastric carcinoma. Exp Oncol.
2008;30:112–116.
53. Tahara T, Shibata T, Nakamura M, et al. Increased number of CpG
island hypermethylation in tumor suppressor genes of non-neoplastic
gastric mucosa correlates with higher risk of gastric cancer.
Digestion. 2010;82:27–36.
54. Yu QM, Wang XB, Luo J, et al. CDH1 methylation in preoperative
peritoneal washes is an independent prognostic factor for gastric
cancer. J Surg Oncol. 2012;106:765–771.
55. Nomura T, Tahara T, Shiroeda H, et al. Influence of HRH2
promoter polymorphism on aberrant DNA methylation of DAPK and
CDH1 in the gastric epithelium. BMC Gastroenterol. 2013;13:1.
56. Zhang Y, Liu X, Fan Y, et al. Germline mutations and polymorphic
variants in MMR, E-cadherin and MYH genes associated with
familial gastric cancer in Jiangsu of China. Int J Cancer.
2006;119:2592–2596.
57. Chu CM, Chen CJ, Chan DC, et al. CDH1 polymorphisms and
haplotypes in sporadic diffuse and intestinal gastric cancer: a case–
control study based on direct sequencing analysis. World J Surg
Oncol. 2014;12:80.
58. Kim J, Min SY, Lee HE, et al. Aberrant DNA methylation and
tumor suppressive activity of the EBF3 gene in gastric carcinoma.
Int J Cancer. 2012;130:817–826.
59. Deng N, Goh LK, Wang H, et al. A comprehensive survey of
genomic alterations in gastric cancer reveals systematic patterns of
molecular exclusivity and co-occurrence among distinct therapeutic
targets. Gut. 2012;61:673–684.
60. Fan B, Dachrut S, Coral H, et al. Integration of DNA copy number
alterations and transcriptional expression analysis in human gastric
cancer. PLoS One. 2012;7:e29824.
ã 2014 Lippincott Williams & Wilkins