Correlations of IFN-γ genetic polymorphisms with susceptibility to breast cancer: a meta-analysis

RESEARCH ARTICLE

Correlations of IFN-γ genetic polymorphisms with susceptibilityto breast cancer: a meta-analysis

Chun-Jiang Li & Yue Dai & Yan-Jun Fu & Jia-Ming Tian &

Jin-Lun Li & Hong-Jun Lu & Feng Duan & Qing-Wang Li

Received: 24 January 2014 /Accepted: 17 March 2014# International Society of Oncology and BioMarkers (ISOBM) 2014

Abstract The meta-analysis was conducted to evaluate thecorrelations between common genetic polymorphisms in theIFN-γ gene and susceptibility to breast cancer. The followingelectronic databases were searched without language restric-tions: MEDLINE (1966~2013), the Cochrane Library Data-base (issue 12, 2013), EMBASE (1980~2013), CINAHL(1982~2013), Web of Science (1945~2013), and the ChineseBiomedical Database (CBM) (1982~2013). Meta-analysiswas performed with the use of the STATA statistical software.Odds ratios (OR) with their 95 % confidence intervals (95 %CIs) were calculated. Nine clinical case-control studies met allthe inclusion criteria and were included in this meta-analysis.A total of 1,182 breast cancer patients and 1,525 healthycontrols were involved in this meta-analysis. Three functionalpolymorphisms were assessed, including rs2069705 C>T,rs2430561 T>A, and CA repeats 2/X. Our meta-analysisresults indicated that IFN-γ genetic polymorphisms might besignificantly associated with an increased risk of breast cancer

(allele model: OR=1.37, 95 % CI=1.03~1.83, P=0.031;dominant model: OR=1.55, 95 % CI=1.01~2.37, P=0.046;homozygous model: OR=2.23, 95 % CI=1.30~3.82, P=0.004; respectively), especially the rs2430561 T>A polymor-phism. Subgroup analysis based on ethnicity suggested thatgenetic polymorphisms in the IFN-γ gene were closely corre-lated with increased breast cancer risk among Asians (allelemodel: OR=1.21, 95 % CI=1.02~1.58, P=0.017; dominantmodel: OR=3.44, 95 % CI=2.07~5.71, P<0.001; recessivemodel: OR=1.58, 95 % CI=1.06~2.37, P=0.025; ho-mozygous model: OR=1.83, 95 % CI=1.19~2.80, P=0.006; respectively), but not among Caucasians (allP>0.05). Our meta-analysis supported the hypothesisthat IFN-γ genetic polymorphisms may contribute toan increased risk of breast cancer, especially thers2430561 T>A polymorphism among Asians.

Keywords IFN-γ . Genetic polymorphism . Susceptibility .

Breast cancer .Meta-analysis

Introduction

Breast cancer, a type of cancer originating from breast tissue,was the most commonly diagnosed malignancy and the sec-ond leading cause of cancer death among women worldwide,accounting for 23 % of total cancer cases and 14 % of cancerdeaths in 2008 [1]. It has been predicted that an estimated230,480 new invasive cases of breast cancer were diagnosedin the USA in 2011 and that about half of breast cancer casesand 60 % of related deaths worldwide are reported to occur ineconomically developing countries [1, 2]. Generally, breastcancer is a complex and multifactorial disease, with interac-tion between environmental and genetic factors potentiallyplaying a crucial role in the development of breast cancer [3,4]. A large number of epidemiological studies have shown

C.<J. Li :Q.<W. Li (*)College of Environment and Chemical Engineering, YanshanUniversity, Hebei Street No. 438 Qinhuangdao 066004, People’sRepublic of Chinae-mail: [email protected]

C.<J. Li :Y. DaiCollege of Basic Medicine, Jiamusi University, Jiamusi 154007,People’s Republic of China

Y.<J. Fu : J.<M. Tian :H.<J. LuThe First Affiliated Hospital of Jiamusi University, Jiamusi 154007,People’s Republic of China

J.<L. Li : F. DuanThe Second Affiliated Hospital of Jiamusi University,Jiamusi 154007, People’s Republic of China

Q.<W. LiCollege of Animal Science, Northwest A&F University, XinongStreet No. 22, Yangling 712100, People’s Republic of China

Tumor Biol.DOI 10.1007/s13277-014-1856-6

that factors such as age, diet and alcohol, weight and height,breastfeeding, smoking, and physical activity may be stronglyassociated with changes in breast cancer risk [5–8]. Familyhistory has also been suggested to be involved in the patho-genesis of breast cancer, and various identified genetic sus-ceptibility loci may contribute to a higher risk of breast cancer[9, 10]. Recently, a growing body of evidence has demonstrat-ed that the production levels of certain cytokines, includinginterferon-γ (IFN-γ), which inhibits the growth of many celllines originating from tumors, may be implicated in the de-velopment of breast cancer [11–13].

IFN-γ, also known as type II interferon, is tradition-ally regarded as a T helper type 1 (Th1) cytokineprimarily produced by activated T lymphocytes andnatural killer cells in response to mitogen or antigenstimulation [14]. To the best of our knowledge, IFN-γplays an important role in defending against viruses andintracellular pathogens and induces immune and inflam-matory responses by activating macrophages, NK cells,and T cells [15]. Therefore, IFN-γ has been documentedto be a pleiotropic macrophage-activating cytokine thatpossesses a number of biological functions, includingantiviral, antiproliferative, antitumor, antiangiogenesis,and immunomodulatory properties [16]. In this regard,it is generally believed that IFN-γ may inhibit thegrowth of many cell lines that originated from tumors,including breast cancer, and that a high level of IFN-γexpression may indicate strong antitumor activity [17].However, accumulating epidemiological studies haveshown that IFN-γ may serve as a risk factor significant-ly implicated in the development and progression ofbreast cancer, which is largely attributed to IFN-γ ge-netic polymorphisms [12, 18]. The human immuneIFN-γ gene, located on the long arm of chromosome12q24.1, contains four exons and three introns andencodes 166 amino acids of the protein, spanning ap-proximately 6 kb in length [19]. It is interesting to notethat the level of IFN-γ expression might be influencedby genetic polymorphisms in the IFN-γ gene, leading tothe impaired function or decreased activity of IFN-γ;subsequent low expression levels of IFN-γ would pro-mote tumor growth, which may contribute to increasingsusceptibility to breast cancer [15]. To be more specific,several common polymorphisms in the IFN-γ gene, suchas rs2069705 (−1615C/T) and rs2430561 (+874T/A),have been found to increase the risk of breast cancer[12, 20]. Recently, while several studies have also indicatedthat IFN-γ genetic polymorphisms may play a critical role inthe pathogenesis of breast cancer, contradictory results havealso been reported [12, 15]. Given the conflicting evidence onthis issue, we performed a meta-analysis of all available datato evaluate the correlations between IFN-γ genetic polymor-phisms and susceptibility to breast cancer.

Methods

Publication search

The following electronic databases were searched withoutlanguage restrictions: MEDLINE (1966~2013), the CochraneLibrary Database (Issue 12, 2013), EMBASE (1980~2013),CINAHL (1982~2013), Web of Science (1945~2013), andthe Chinese Biomedical Database (CBM) (1982~2013). Weused the following keywords and MeSH terms in conjunctionwith a highly sensitive search strategy: [“single nucleotidepolymorphism” or “SNP” or “polymorphism” or “mutation”or “mutant” or “variation” or “variant”] and [“breast cancer”or “breast tumor” or “breast carcinoma” or “breast neo-plasms”] and [“interferons” or “interferon-γ” or “IFN-γ” or“interferon-gamma”]. A manual search of reference lists fromrelevant articles was conducted to find further potentialarticles.

Inclusion and exclusion criteria

The following criteria were used to determine inclusion eligi-bility: (1) must be a clinical case-control study focused on therelationships between IFN-γ genetic polymorphisms and sus-ceptibility to breast cancer; (2) all patients diagnosed withbreast cancer must be confirmed by histopathologic examina-tions; (3) the genotype frequencies of healthy controls shouldfollow the Hardy-Weinberg equilibrium (HWE); and (4) thestudy must provide sufficient information about genotypefrequencies. Articles that did not meet our inclusion criteriawere excluded. The most recent or largest sample size publi-cation was included when authors published several studiesusing the same subjects.

Data extraction and methodological assessment

Data were systematically extracted by two authors from eachincluded study using a standardized form, including languageof publication, publication year of article, the first author’ssurname, geographical location, design of study, sample size,the source of the subjects, genotype frequencies, source ofsamples, genotyping method, evidence of HWE, etc.

Methodological quality was evaluated separately by twoobservers using the Newcastle-Ottawa Scale (NOS) criteria[21]. The NOS criteria is scored based on three aspects: (1)subject selection, 0~4; (2) comparability of subject, 0~2; and(3) clinical outcome, 0~3. Total NOS scores range from 0 to9, with scores ≥7 indicating good quality.

Statistical analysis

Meta-analysis was performed using the STATA statisticalsoftware (Version 12.0, Stata Corporation, College Station,

Tumor Biol.

TX, USA). Odds ratios (OR) and 95 % confidence interval(95 % CI) were calculated as estimates of relative risk forbreast cancer under five genetic models: allele model (mutant[M] allele vs. wild [W] allele), dominant model (heterozygote

[WM]+mutant homozygote [MM] vs. wild homozygote[WW]), recessive model (MM vs. WW+WM), homozygousmodel (MM vs. WW), and heterozygous model (MM vs.WM). The Z test was used to estimate the statistical

Iden

tific

atio

nS

cree

ning

Elig

ibili

tyIn

clud

ed

Studies were excluded, due to:(N = 12) Letters, reviews, meta-analysis(N = 20) Not human studies(N = 34) Not related to research topics

Studies were excluded, due to:(N = 9) Not case-control or cohort study(N = 14) Not relevant to IFN- geneγ(N = 20) Not relevant to breast cancer

Articles identified through electronic database searching

Articles reviewed for duplicates

(N = 121)

Articles after duplicates removed(N = 122)

Full-text articles assessed for eligibility

(N = 56)

Studies included in qualitative synthesis

(N = 13)

Studies included in quantitative synthesis (meta-analysis)

(N = 9)

Additional articles identified through a manual search

(N = 3)

(N = 124)

Fig. 1 Flow chart shows studyselection procedure. Nine case-control studies were included inthis meta-analysis

0

5

10

15

20

25

30

35

2012~20132010~20112008~20092006~20072004~20052002~20032000~2001

Publication year

Num

ber

of a

rtic

les

Pubmed database

All database

Fig. 2 The distribution of thenumber of topic-related literaturein electronic databases over thelast decade

Tumor Biol.

significance of pooled ORs. Heterogeneity among studies wasestimated by the Cochran’sQ statistic and I2 tests [22]. If theQtest showed a P<0.05 or I2 test exhibited >50 %, indicatingsignificant heterogeneity, the random-effects model was con-ducted; otherwise, the fixed-effects model was used. We alsoexplored potential sources of heterogeneity using meta-regression and subgroup analyses. In order to evaluate theinfluence of single studies on overall estimates, a sensitivityanalysis was performed. Funnel plots and Egger’s linear re-gression test were applied to investigate publication bias [23].

Results

Study selection and characteristics of included studies

Initially, our highly sensitive search strategy identified 124articles. After screening the titles and abstracts of all retrievedarticles, 66 articles were excluded. Full texts were thenreviewed, and 43 articles were further exclude. Another fourstudies were excluded due to lack of data integrity (Fig. 1).Finally, nine clinical case-control studies with a total of 1,182breast cancer patients and 1,525 healthy controls met ourinclusion criteria for qualitative data analysis [12, 15, 18, 20,24–28]. The publication years of eligible studies ranged from2005 to 2012. Figure 2 shows the distribution of the number oftopic-related literatures in electronic databases over the lastdecade. Overall, seven studies were conducted among Asiansand two studies among Caucasians. There are five studies withsmall sample sizes and four studies with larger sample sizes.Genotyping was performed using the PCR-RFLP, TaqMan,Sequenom, AS-PCR, and PCR-SSP subgroupmethods. Threefunctional polymorphisms including rs2069705 C>T,rs2430561 T>A, and CA repeats 2/X in the IFN-γ gene wereassessed. We summarized the study characteristics and meth-odological quality in Table 1.

Quantitative data synthesis

A summary of our meta-analysis findings on the correlationsbetween IFN-γ polymorphisms and the pathogenesis of breastcancer is shown in Table 2. The random-effects model wasperformed due to obvious heterogeneity between studies. Ourmeta-analysis results indicated that IFN-γ genetic polymor-phisms might be significantly associated with an increasedrisk of breast cancer (allele model: OR=1.37, 95% CI=1.03~1.83, P=0.031; dominant model: OR=1.55, 95 % CI=1.01~2.37, P=0.046; homozygous model: OR=2.23, 95 % CI=1.30~3.82, P=0.004; respectively).

A subgroup analysis was carried out to evaluate the impactof the IFN-γ promoter polymorphism on the risk of breastcancer. SNP-stratified analysis indicated that the IFN-γrs2430561 T>A polymorphism was associated with an Ta

ble1

Baselinecharacteristicsandmethodologicalq

ualityof

allincludedstudies

Firstauthor

Year

Country

Ethnicity

Num

ber

Age

(years)

Genotypingmethod

SNPtype

HWEtest

NOSscore

Case

Control

Case

Control

HeJR

[15]

2012

China

Asians

354

504

48.3±11.2

48.6±11.8

Directsequencing

rs2069705C>T

0.354

8

Karakus

N[12]

2011

Turkey

Asians

204

204

52(28~82)

48(35~86)

PCR-RFL

Prs2430561T>A

0.052

8

ErdeiE[20]

2010

USA

Caucasians

4040

49.2±13.6

50.2±10.8

TaqM

anassay

rs2069705C>T

0.354

7

WuGH[24]

2008

China

Asians

9496

51(30~67)

AS-PC

Rrs2430561T>A

0.521

7

Gonullu

G[18]

2007

Turkey

Asians

3824

46(35~70)

39(20~57)

PCR-SSP

rs2430561T>A

0.377

7

WuJM

[26]

2005

China

Asians

87144

42.7±7.9

AS-PC

RCArepeats2>

X0.726

7

SkerrettDL[25]

2005

USA

Caucasians

88102

49.2±12.4

PCR-SSP

rs2430561T>A

0.460

7

SahaA[27]

2005

India

Asians

54144

–PC

R-SSP

CArepeats2>

X0.726

6

Kam

ali-SarvestaniE

[28]

2005

Iran

Asians

223

267

50(27~85)

ASO

-PCR

rs2430561T>A

0.460

7

PCR-RFLPpolymerasechainreactio

n-restrictionfragmentlengthpolymorphism,A

Sallele-specific,SSPspecificprim

ers,ASO

allele-specificoligonucleotide,SN

Psinglenucleotid

epolymorphism,H

WE

Hardy-W

einbergequilib

rium

,NOSNew

castle-O

ttawaScale

Tumor Biol.

Table2

Meta-analysisof

therelatio

nships

ofIFN-γ

genetic

polymorphismswith

therisk

ofbreastcancer

Mallelevs.W

allele(allele

model)

WM+MM

vs.W

W(dom

inant

model)

MM

vs.W

W+WM

(recessive

model)

MM

vs.W

W(hom

ozygous

model)

MM

vs.W

M(heterozygous

model)

OR

95%

CI

POR

95%

CI

POR

95%

CI

POR

95%

CI

POR

95%

CI

P

Overall

1.37

1.03–1.83

0.031

1.55

1.01–2.37

0.046

1.70

1.00–2.89

0.052

2.23

1.30–3.82

0.004

1.51

0.83–2.73

0.174

SNPtype

rs2069705C>T

2.04

0.59–7.02

0.257

2.09

0.52–8.40

0.297

3.41

0.54–21.50

0.192

4.50

0.42–48.75

0.216

2.67

0.64–11.11

0.177

rs2430561T>A

1.27

1.01–1.60

0.044

1.52

0.80–2.90

0.200

1.40

0.83–2.37

0.205

1.77

1.30–2.41

<0.001

1.42

0.62–3.25

0.407

CArepeats2>

X1.04

0.29–3.64

0.957

1.29

0.32–5.17

0.718

0.43

0.01–17.34

0.656

0.51

0.01–37.84

0.762

0.40

0.01–11.01

0.587

Ethnicity

Asians

1.21

1.02–1.58

0.017

3.44

2.07–5.71

<0.001

1.58

1.06–2.37

0.025

1.83

1.19–2.80

0.006

1.52

0.86–2.70

0.151

Caucasians

2.19

0.72–6.65

0.168

1.24

0.81–1.90

0.319

2.31

0.16–32.37

0.535

4.79

0.47–48.63

0.185

1.58

0.13–19.35

0.721

Genotypingmethod

Directsequencing

1.11

0.89–1.37

0.362

1.08

0.82–1.43

0.571

1.35

0.80– 2.29

0.258

1.38

0.80–2.37

0.247

1.33

0.77–2.30

0.314

PCR-RFL

P1.36

1.03–1.80

0.031

2.23

1.42–3.50

0.001

0.97

0.62–1.53

0.905

1.75

1.01–3.04

0.046

0.69

0.42–1.12

0.134

TaqM

anassay

3.90

2.45–6.23

<0.001

4.48

2.03–9.93

<0.001

8.88

4.25–18.55

<0.001

15.67

6.15–39.92

<0.001

5.69

2.58–12.55

<0.001

AS-PCR

1.29

0.84–1.96

0.242

1.11

0.50–2.45

0.794

2.07

1.49–2.87

<0.001

2.08

1.27–3.40

0.004

2.42

1.43–4.09

0.001

PCR-SSP

1.17

0.54–2.51

0.694

1.72

0.56–5.23

0.340

0.68

0.12–3.70

0.652

1.07

0.17–6.90

0.944

0.53

0.11–2.55

0.431

Sam

plesize

Large

(N≥3

00)

1.16

0.98–1.39

0.092

1.08

0.67–1.75

0.739

1.48

1.02–2.14

0.039

1.61

1.21–2.14

0.001

1.67

0.81–3.43

0.162

Small(N<300)

1.66

0.89–3.10

0.109

2.25

1.10–4.63

0.027

1.63

0.48–5.50

0.433

2.77

0.83–9.30

0.099

1.21

0.38–3.86

0.747

Wwild

allele,M

mutantallele,W

Wwild

homozygote,WM

heterozygote,M

Mmutanth

omozygote,ORodds

ratio

,95%

CI95

%confidence

interval,SNPsinglenucleotid

epolymorphism,P

CR-RFLP

polymerasechainreactio

n-restrictionfragmentlengthpolymorphism,A

Sallele-specific,SSPspecificprim

ers

Tumor Biol.

increased risk of breast cancer (allele model: OR=1.27, 95 %CI=1.01~1.60, P=0.044; dominant model: OR=1.52, 95 %CI=1.08~2.20, P=0.020; homozygous model: OR=1.77,95 % CI=1.30~2.41, P<0.001; respectively), whilers2069705 C>T and CA repeats polymorphisms showed nocorrelation with breast cancer susceptibility (all P>0.05)(Fig. 3). The results of our subgroup analysis based on eth-nicity suggest that genetic polymorphisms in the IFN-γ genewere closely linked to the pathogenesis of breast canceramong Asian populations (allele model: OR=1.21, 95 %CI=0.92~1.58, P=0.172; dominant model: OR=1.24, 95 %

CI=0.81~1.90, P=0.319; recessive model: OR=1.58, 95 %CI=1.06~2.37, P=0.025; homozygous model: OR=1.83,95 % CI=1.19~2.80, P=0.006; respectively), but not amongCaucasians (allP>0.05) (Fig. 4).We also performed subgroupanalyses based on genotyping method and sample size. Theirresults indicate that IFN-γ genetic polymorphisms might beimplicated in the risk of breast cancer in the majority of

Heterogeneity test (I2 = 83.5%, P < 0.001)


Kamali−Sarvestani E (2005)

Wu JM (2005)


Karakus N (2011)

Heterogeneity test (I2 = 52.7%, P = 0.076)

Saha A (2005)

Gonullu G (2007)

Skerrett DL (2005)

CA repeats

Wu GH (2008)

rs2069705

rs2430561

He JR (2012)

Erdei E (2010)

1.37 (1.03, 1.83)

1.04 (0.29, 3.64)

1.30 (1.00, 1.68)

1.95 (1.33, 2.85)

2.04 (0.59, 7.02)

1.36 (1.03, 1.80)

1.27 (1.01, 1.60)

0.54 (0.33, 0.89)

2.65 (1.25, 5.59)

1.25 (0.89, 1.76)

0.84 (0.57, 1.23)

1.11 (0.89, 1.37)

3.90 (2.45, 6.23)

100.00

21.24

12.66

11.30

23.35

12.45

55.41

9.94

7.22

11.77

11.31

13.06

10.28

)

10.142 7.02

Included studies OR (95% CI) Weight%Allele model

(M allele versus W allele)

Included studies OR (95% CI) Weight%

Domiant model(WM+MM versus WW)

Z test (Z = 2.15, P = 0.031)

Z test (Z = 0.05, P = 0.957)

Z test (Z = 2.01, P = 0.044)

Z test (Z = 1.13, P = 0.257)

Random effects analysis

Z test (Z = 1.99, P = 0.046)

Z test (Z = 0.36, P = 0.718)

Z test (Z = 1.28, P = 0.200)

Z test (Z = 1.04, P = 0.297)


Heterogeneity test (I2 = 83.3%,P < 0.001)

Erdei E (2010)

Wu JM (2005)

Skerrett DL (2005)

Saha A (2005)

Wu GH (2008)

Karakus N (2011)

Gonullu G (2007)



Heterogeneity test (I2 = 85.3%,P < 0.001)

He JR (2012)

rs2430561


rs2069705

CA repeats

1.55 (1.01, 2.37)

4.48 (2.03, 9.93)

2.62 (1.40, 4.91)

2.86 (1.47, 5.54)

0.64 (0.34, 1.19)

0.54 (0.33, 0.87)

2.23 (1.42, 3.50)

3.17 (1.04, 9.72)

1.04 (0.70, 1.55)

1.29 (0.32, 5.17)

1.52 (0.80, 2.90)

1.08 (0.82, 1.43)

2.09 (0.52, 8.40)

100.00

9.58

10.94

10.66

10.92

12.10

12.34

7.23

12.74

21.86

55.06

13.50

23.08

10.101 9.93

Fig. 3 Forest plots for therelationships between IFN-γgenetic polymorphisms andsusceptibility to breast cancer

�Fig. 4 Subgroup analyses by ethnicity, genotyping method, and samplesize of the relationships between IFN-γ genetic polymorphisms andsusceptibility to breast cancer

Tumor Biol.


Wu JM (2005)

Asians

Saha A (2005)

He JR (2012)

Karakus N (2011)

Erdei E (2010)



Wu GH (2008)


Gonullu G (2007)

Skerrett DL (2005)

Caucasians

1.37 (1.03, 1.83)

1.95 (1.33, 2.85)

0.54 (0.33, 0.89)

1.11 (0.89, 1.37)

1.36 (1.03, 1.80)

3.90 (2.45, 6.23)

1.30 (1.00, 1.68)

2.19 (0.72, 6.65)

0.84 (0.57, 1.23)

1.21 (1.02, 1.58)

2.65 (1.25, 5.59)

1.25 (0.89, 1.76)

100.00

11.30

9.94

13.06

12.45

10.28

12.66

22.05

11.31

77.95

7.22

11.77

10.15 6.65

Included study OR (95% CI) Weight%Ethnicity


Included study OR (95% CI) Weight%Genotyping method


Included study OR (95% CI) Weight%Sample size


Z test (Z = 2.15, P = 0.031)

Z test (Z = 1.38, P = 0.168)

Z test (Z = 2.36, P = 0.017)


Z test (Z = 2.15, P = 0.031)

Z test (Z = 0.39, P = 0.694)

Z test (Z = 1.17, P = 0.242)

Z test (Z = 5.71, P < 0.001)

Z test (Z = 2.15, P = 0.031)

Z test (Z = 0.91, P = 0.362)


Z test (Z = 2.15, P = 0.031)

Z test (Z = 1.60, P = 0.109)

Z test (Z = 1.69, P = 0.092)



PCR-SSP

Sequenom

Wu JM (2005)

AS-PCR


Skerrett DL (2005)

Erdei E (2010)

Wu GH (2008)

PCR-RFLP


TaqMan assay

Saha A (2005)

Karakus N (2011)


He JR (2012)

Gonullu G (2007)

1.37 (1.03, 1.83)

1.95 (1.33, 2.85)1.30 (1.00, 1.68

1.25 (0.89, 1.76)

3.90 (2.45, 6.23)

1.36 (1.03, 1.80)

0.84 (0.57, 1.23)

1.29 (0.84, 1.96)

1.11 (0.89, 1.37)

0.54 (0.33, 0.89)

1.36 (1.03, 1.80)

3.90 (2.45, 6.23)

1.17 (0.54, 2.51)

1.11 (0.89, 1.37)

2.65 (1.25, 5.59)

100.00

11.3012.66

11.77

10.28

12.45

11.31

35.27

13.06

9.94

12.45

10.28

28.93

13.06

7.22

10.161 6.23


Saha A (2005)


Erdei E (2010)

Karakus N (2011)

Skerrett DL (2005)

Small sample-size

Gonullu G (2007)

Wu GH (2008)



He JR (2012)

Large sample-size

Wu JM (2005)

1.37 (1.03, 1.83)

0.54 (0.33, 0.89)

1.66 (0.89, 3.10)

3.90 (2.45, 6.23)

1.36 (1.03, 1.80)

1.25 (0.89, 1.76)

2.65 (1.25, 5.59)

0.84 (0.57, 1.23)

1.16 (0.98, 1.39)

1.30 (1.00, 1.68)

1.11 (0.89, 1.37)

1.95 (1.33, 2.85)

100.00

9.94

50.52

10.28

12.45

11.77

7.22

11.31

49.48

12.66

13.06

11.30

10.161 6.23

Tumor Biol.

subgroups, except the Sequenom and PCR-SSP subgroups (asshown in Table 2).

Meta-regression analysis also indicated that none ofthe considered factors was the main sources of hetero-geneity (as shown in Table 3). Sensitivity analysis sug-gested that no single study could influence the pooledORs (Fig. 5). Funnel plots demonstrated evidence ofobvious asymmetry (Fig. 6). Egger’s test displayedstrong statistical evidence of publication bias (allelemodel: t=0.79, P=0.457; dominant model: t=1.46,P=0.187; respectively).

Discussion

In the present meta-analysis, we evaluated the hypoth-esis that IFN-γ genetic polymorphisms contribute tosusceptibility to breast cancer. The findings of ourmeta-analysis indicate that genetic polymorphisms inthe IFN-γ gene were significantly correlated with anincreased risk of breast cancer, especially thers2430561 T>A polymorphism. However, we foundno associations between rs2069705 C>T and CA re-peats polymorphisms and susceptibility to breast can-cer, revealing that linkage disequilibrium (LD) plays animportant role in the effects of IFN-γ genetic polymor-phisms on susceptibility to breast cancer. We hypothe-size that IFN-γ gene variants may reduce its expressionlevel and decrease its functions, which are suspected to

p l ay a v i t a l ro l e in immune regu la t i on andantiproliferation [12]. It has been reported that IFN-γpossesses the ability to inhibit the growth of severaltumor-derived cell lines including breast cancer cells[28]. As the most common cancer in women world-wide, breast cancer has been suggested to be caused byviral infection, and Epstein-Barr virus (EBV) has re-ceived considerable interest as the candidate virus [29,30]. It is noteworthy that EBV infection results in bothhumeral and cellular immune responses [31]. As one ofthe major immune mediators, IFN-γ has been reportedto inhibit the outgrowth of EBV-transformed B cellsin vitro and thus has a significant function inpreventing the carcinogenesis of breast cancer [15,32]. However, IFN-γ gene polymorphisms may alterthe function and expressions of IFN-γ, with low levelsof IFN-γ expression promoting tumor growth [12]. Liuet al. revealed that the rs2430561 A>T polymorphismin the intron 1 of the IFN-γ gene may functionallytransform its transcription and contribute to suscepti-bility to breast cancer [33].

To further investigate the relationship between IFN--γ genetic variations and the pathogenesis of breastcancer, we conducted subgroup analyses based on eth-nicity, genotyping method, and sample size. Our find-ings revealed that there were significant associationsbetween IFN-γ genetic polymorphisms and breast can-cer risk among Asians, but not among Caucasians,indicating that ethnic differences may be a potential

Table 3 Univariate and multivariate meta-regression analyses of potential source of heterogeneity

Heterogeneity factors Coefficient SE Z P 95 % CI

LL UL

Publication year

Univariate 0.039 0.072 0.54 0.589 −0.102 0.179

Multivariate −0.996 1.009 −0.99 0.324 −2.974 0.982

SNP type

Univariate −0.318 0.275 −1.15 0.248 −0.857 0.222

Multivariate −1.729 1.377 −1.26 0.209 −4.427 0.969

Ethnicity

Univariate 0.571 0.421 1.35 0.176 −0.255 1.397

Multivariate −2.852 2.796 −1.02 0.308 −8.333 2.628

Genotyping method

Univariate 0.180 0.131 1.37 0.170 −0.077 0.437

Multivariate 1.476 1.471 1.00 0.316 −1.408 4.360

Sample size

Univariate 0.375 0.371 1.01 0.312 −0.352 1.103

Multivariate −0.571 1.502 −0.38 0.704 −3.516 2.374

SE standard error, 95 % CI 95 % confidence interval, LL lower limit UL upper limit, SNP single nucleotide polymorphism

Tumor Biol.

factor affecting individual’s susceptibility to breast can-cer. Even though the exact mechanisms of ethnicity-related differences are only partially understood, onepossible explanation is that ethnicity may result indifferences in alleles and genotypes between differentethnic populations. The results of our subgroup analy-sis by genotyping method revealed that IFN-γ geneticpolymorphisms were related to an increased risk ofbreast cancer in the PCR-RFLP and TaqMan array

subgroups, while no associations were detected in theSequenom, AS-PCR, and PCR-SSP subgroups. In thefurther stratification analysis by sample size, our resultsillustrated a positive correlation between IFN-γ geneticpolymorphisms and susceptibility to breast cancer in thelarge sample size subgroup, but not in the small samplesize subgroup. In short, our findings were consistentwith previous studies that IFN-γ genetic polymorphismsmay be potential risk factors in the development of

0.96 1.371.03 1.83 2.01

He JR (2012)

Karakus N (2011)

Erdei E (2010)

Wu GH (2008)

Gonullu G (2007)

Wu JM (2005)

Skerrett DL (2005)

Saha A (2005)


Lower CI Limit Estimate Upper CI Limit

Allelel model(M allele versus W allele)

Lower CI Limit Estimate Upper CI Limit

Dominant model(WM+MM versus WW)

0.90 1.551.01 2.37 2.82

He JR (2012)

Karakus N (2011)

Erdei E (2010)

Wu GH (2008)

Gonullu G (2007)

Wu JM (2005)

Skerrett DL (2005)

Saha A (2005)


Fig. 5 Sensitivity analysis of thesummary odds ratio coefficientson the relationships betweenIFN-γ genetic polymorphismsand susceptibility to breast cancer

Tumor Biol.

breast cancer, suggesting that these polymorphisms maybe of great value for the early diagnosis of breastcancer.

The current meta-analysis has several limitations thatshould be pointed out. First, our results were short ofsufficient statistical power to assess the correlationsbetween IFN-γ genetic polymorphisms and the patho-genesis of breast cancer due to the small number ofstudies. Since the sample sizes of some included studieswere small, our meta-analysis might have induced fairlywide confidence intervals, which limit confidence in ourconclusions. Besides, the small number of studies mayconstrain the general applicability of our findings, andso our meta-analysis can only be regarded as prelimi-nary. Second, our meta-analysis failed to obtain originaldata from the included studies, which may have limitedfurther evaluation of the potential roles of IFN-γ geneticpolymorphisms in the development and progression ofbreast cancer. Although our study has several limita-tions, this is the first meta-analysis focusing on therelationships between IFN-γ genetic polymorphismsand the risk of breast cancer. Furthermore, we per-formed a highly sensitive literature search strategy of

electronic databases, a manual search of reference listsfrom relevant articles, and the selection process of eli-gible articles was based on strict inclusion and exclu-sion criteria. Importantly, rigorous statistical analysisprovided a basis for pooling of information from indi-vidual studies.

In conclusion, our meta-analysis supports the notion thatIFN-γ genetic polymorphisms may contribute to an increasedrisk of breast cancer, especially the rs2430561 T>A polymor-phism among Asians. Thus, IFN-γ genetic polymorphismsmay be valuable biomarkers for breast cancer. However, dueto the limitations acknowledged above, more research withlarger sample sizes is required to provide a more representa-tive statistical analysis.

Acknowledgments This work was supported by the Natural ScienceFoundation of Heilongjiang Province (No. D201166). We would like toacknowledge the reviewers for their helpful comments on this paper.

Conflicts of interest None

References

1. Jemal A, Bray F, Center MM, Ferlay J, Ward E, Forman D. Globalcancer statistics. CA Cancer J Clin. 2011;61(2):69–90.

2. Guo LJ, Zhang QY. Decreased serum miR-181a is a potential newtool for breast cancer screening. Int J Mol Med. 2012;30(3):680–6.

3. Brinton LA, Carreon JD, Gierach GL, McGlynn KA, Gridley G.Etiologic factors for male breast cancer in the U.S. Veterans Affairsmedical care system database. Breast Cancer Res Treat. 2010;119(1):185–92.

4. Nickels S, Truong T, Hein R, Stevens K, Buck K, Behrens S, et al.Evidence of gene-environment interactions between common breastcancer susceptibility loci and established environmental risk factors.PLoS Genet. 2013;9(3):e1003284.

5. La Vecchia C, Giordano SH, Hortobagyi GN, Chabner B.Overweight, obesity, diabetes, and risk of breast cancer: interlockingpieces of the puzzle. Oncologist. 2011;16(6):726–9.

6. Xue F, Willett WC, Rosner BA, Hankinson SE, Michels KB.Cigarette smoking and the incidence of breast cancer. Arch InternMed. 2011;171(2):125–33.

7. Park SK, Kim Y, Kang D, Jung EJ, Yoo KY. Risk factors and controlstrategies for the rapidly rising rate of breast cancer in Korea. J BreastCancer. 2011;14(2):79–87.

8. Fontein DB, de Glas NA, Duijm M, Bastiaannet E, Portielje JE, Vande Velde CJ, et al. Age and the effect of physical activity on breastcancer survival: a systematic review. Cancer Treat Rev. 2013;39(8):958–65.

9. Reiner AS, John EM, Brooks JD, Lynch CF, Bernstein L,Mellemkjaer L, et al. Risk of asynchronous contralateral breastcancer in noncarriers of BRCA1 and BRCA2 mutations with afamily history of breast cancer: a report from the Women’sEnvironmental Cancer and Radiation Epidemiology Study. J ClinOncol. 2013;31(4):433–9.

10. Bouchardy C, Rapiti E, Fioretta G, Schubert H, Chappuis P, VlastosG, et al. Impact of family history of breast cancer on tumour charac-teristics, treatment, risk of second cancer and survival among menwith breast cancer. Swiss Med Wkly. 2013;143:w13879.

0 0.2 0.4

−0.5

0

0.5

1

1.5

Log[

OR

]

SE Log[OR]

(Egger’s test: t = 0.79, P = 0.457)

0 0.2 0.4 0.6

−1

0

1

2

Log[

OR

]

SE Log[OR]

(Egger’s test: t = 1.46, P = 0.187)

Allelel model(M allele versus W allele)

Dominant model(WM+MM versus WW)

Fig. 6 Funnel plot of publication biases on the relationships betweenIFN-γ genetic polymorphisms and susceptibility to breast cancer

Tumor Biol.

11. Su Y, Tang LY, Chen LJ, He JR, Su FX, Lin Y, et al. Joint effects offebrile acute infection and an interferon-gamma polymorphism onbreast cancer risk. PLoS One. 2012;7(5):e37275.

12. Karakus N, Kara N, Ulusoy AN, Ozaslan C, Bek Y. Tumornecrosis factor alpha and beta and interferon gamma genepolymorphisms in Turkish breast cancer patients. DNA CellBiol. 2011;30(6):371–7.

13. Kim K, Cho SK, Sestak A, Namjou B, Kang C, Bae SC. Interferon-gamma gene polymorphisms associated with susceptibility to sys-temic lupus erythematosus. Ann Rheum Dis. 2010;69(6):1247–50.

14. Horras CJ, Lamb CL, Mitchell KA. Regulation of hepatocyte fate byinterferon-gamma. Cytokine Growth Factor Rev. 2011;22(1):35–43.

15. He JR, Chen LJ, SuY, CenYL, Tang LY, YuDD, et al. Joint effects ofEpstein-Barr virus and polymorphisms in interleukin-10 andinterferon-gamma on breast cancer risk. J Infect Dis. 2012;205(1):64–71.

16. Chou SF. Development of a manual self-assembled colloidal goldnanoparticle-immunochromatographic strip for rapid determinationof human interferon-gamma. Analyst. 2013;138(9):2620–3.

17. Pluddemann A, Mukhopadhyay S, Gordon S. Innate immunity tointracellular pathogens: macrophage receptors and responses to mi-crobial entry. Immunol Rev. 2011;240(1):11–24.

18. Gonullu G, Basturk B, Evrensel T, Oral B, Gozkaman A,ManavogluO. Association of breast cancer and cytokine gene polymorphism inTurkish women. Saudi Med J. 2007;28(11):1728–33.

19. Naylor SL, Gray PW, Lalley PA. Mouse immune interferon (IFN-gamma) gene is on chromosome 10. Somat Cell Mol Genet.1984;10(5):531–4.

20. Erdei E, Kang H, Meisner A, White K, Pickett G, Baca C, et al.Polymorphisms in cytokine genes and serum cytokine levels amongNew Mexican women with and without breast cancer. Cytokine.2010;51(1):18–24.

21. Stang A. Critical evaluation of the Newcastle-Ottawa scale for theassessment of the quality of nonrandomized studies in meta-analyses.Eur J Epidemiol. 2010;25(9):603–5.

22. Zintzaras E, Ioannidis JP. HEGESMA: genome search meta-analysisand heterogeneity testing. Bioinformatics. 2005;21(18):3672–3.

23. Peters JL, Sutton AJ, Jones DR, Abrams KR, Rushton L.Comparison of two methods to detect publication bias in meta-analysis. JAMA. 2006;295(6):676–80.

24. Wu GH, Zhang JY, Lu PX. Association of single nucleotide poly-morphism of interferon-ganuna gene+874 site and breast cancer.Cancer Res Prev Treat. 2008;35(9).

25. Skerrett DL, Moore EM, Bernstein DS, Vahdat L. Cytokine genotypepolymorphisms in breast carcinoma: associations of TGF-beta1 withrelapse. Cancer Invest. 2005;23(3):208–14.

26. Wu JM, Bensen-Kennedy D, Miura Y, Thoburn CJ, Armstrong D,Vogelsang GB, et al. The effects of interleukin 10 and interferongamma cytokine gene polymorphisms on survival after autologousbone marrow transplantation for patients with breast cancer. BiolBlood Marrow Transplant. 2005;11(6):455–64.

27. Saha A, Dhir A, Ranjan A, Gupta V, Bairwa N, Bamezai R.Functional IFNG polymorphism in intron 1 in association with anincreased risk to promote sporadic breast cancer. Immunogenetics.2005;57(3–4):165–71.

28. Kamali-Sarvestani E, Merat A, Talei AR. Polymorphism in the genesof alpha and beta tumor necrosis factors (TNF-alpha and TNF-beta)and gamma interferon (IFN-gamma) among Iranian women withbreast cancer. Cancer Lett. 2005;223(1):113–9.

29. Zur Hausen H. The search for infectious causes of human cancers:where and why. Virology. 2009;392(1):1–10.

30. Joshi D, Quadri M, Gangane N, Joshi R. Association of Epstein Barrvirus infection (EBV) with breast cancer in rural Indian women.PLoS One. 2009;4(12):e8180.

31. Farrell RA, AntonyD,Wall GR, Clark DA, Fisniku L, Swanton J, et al.Humoral immune response to EBV in multiple sclerosis is associatedwith disease activity on MRI. Neurology. 2009;73(1):32–8.

32. Wingate PJ, McAulay KA, Anthony IC, Crawford DH. Regulatory Tcell activity in primary and persistent Epstein-Barr virus infection. JMed Virol. 2009;81(5):870–7.

33. Mi YY, Yu QQ, Xu B, Zhang LF, Min ZC, Hua LX, et al. Interferongamma +874 T/A polymorphism contributes to cancer susceptibility:a meta-analysis based on 17 case-control studies. Mol Biol Rep.2011;38(7):4461–7.

Tumor Biol.

Documents

Correlations of IFN-γ genetic polymorphisms with susceptibility to breast cancer: a meta-analysis