Journal of Thrombosis and Thrombolysis

, Volume 36, Issue 4, pp 394–401

Meta-analyses of four eosinophil related gene variants in coronary heart disease

Authors

  • Jiangfang Lian
    • Ningbo Medical CenterLihuili Hospital, Ningbo University
  • Yi Huang
    • The Affiliated Hospital, School of Medicine, Ningbo University
  • R. Stephanie Huang
    • Department of MedicineUniversity of Chicago
  • Limin Xu
    • The Affiliated Hospital, School of Medicine, Ningbo University
  • Yanping Le
    • The Affiliated Hospital, School of Medicine, Ningbo University
  • Xi Yang
    • Ningbo Medical CenterLihuili Hospital, Ningbo University
  • Weifeng Xu
    • Ningbo Medical CenterLihuili Hospital, Ningbo University
  • Xiaoyan Huang
    • Ningbo Medical CenterLihuili Hospital, Ningbo University
  • Meng Ye
    • The Affiliated Hospital, School of Medicine, Ningbo University
    • Ningbo Medical CenterLihuili Hospital, Ningbo University
    • The Affiliated Hospital, School of Medicine, Ningbo University
Article

DOI: 10.1007/s11239-012-0862-z

Cite this article as:
Lian, J., Huang, Y., Huang, R.S. et al. J Thromb Thrombolysis (2013) 36: 394. doi:10.1007/s11239-012-0862-z

Abstract

The goal of our study is to assess the contribution of four eosinophil related gene variants (rs12619285, rs1420101, rs3184504 and rs4143832) to the risk of coronary heart disease (CHD). We conducted four meta-analyses of studies examining the association between four eosinophil related gene variants and the risk of CHD. A systematic search was conducted using MEDLINE, EMBASE, Web of Science and China National Knowledge Infrastructure (CNKI), Wanfang Chinese Periodical. A case–control study was conducted between 162 CHD cases and 119 non-CHD controls to explore their contribution to CHD. For rs3184504 of SH2B3 gene, the meta-analysis was performed among 19 study stages among 94,555 participants. Significant association between rs3184504 and CHD risk was observed in European and South Asian populations (OR = 1.13, 95 % CI = 1.10–1.16, p < 0.0001, fixed-effect method). For the other SNPs (rs12619285, rs1420101, and rs4143832), we combined our case–control data with the previous studies and found no association of them with the risk of CHD. No significant contribution of the four genetic variants to CHD was observed in Han Chinese (p > 0.05). In conclusion, our results supported a significant association between rs3184504 of SH2B3 gene and the risk of CHD in Europeans and South Asians, although we were unable to observe association between the four variants and the risk of CHD in Han Chinese.

Keywords

Coronary heart diseaseGenetic variantSH2B3IKZF2IL1RL

Abbreviations

SNP

Single nucleotide polymorphism

CAC

Coronary artery calcification

CHD

Coronary heart disease

MI

Myocardial infarction

HWE

Hardy–Weinberg equilibrium

Introduction

Coronary heart disease (CHD) is one of the top causes of morbidity and mortality worldwide [1]. Researchers have identified a handful of genes for the susceptibility of CHD [2, 3], however at least 95 % of genes are remained to be discovered to explain the pathogenesis of CHD [4, 5]. Among the proposed mechanisms, regulation of eosinophil counts is likely to become a potential way of reducing blood vessel inflammation that is a major feature in the development of CHD [6, 7].

Four genetic variants affecting eosinophil numbers are shown to be associated with the risk of myocardial infarction (MI) [8]. These variants comprise rs12619285 of IKZF2 gene, rs1420101 of IL1RL1 gene, rs3184504 of SH2B3 gene, and rs4143832 of IL5 gene. IKZF2 is a member of the Ikaros family of zinc-finger proteins and is involved in the regulation of lymphocyte and early hematopoietic development [9]. IL5 encodes interleukin 5 that is a growth and differentiation factor for B cells and mediator for eosinophil activation [10]. DNA sequence variants in IKZF2 (rs12619285) and IL5 (rs4143832) are shown to be associated with blood eosinophil counts in European [8]. IL5 can modulate eosinophil behavior at every stage from maturation to survival. The expression of IL5 is shown to be associated with the eosinophilic diseases [11]. IL5 variant rs4143832 is observed to be associated with the risk of CHD in European and South Asian population [12]. IL1RL1 encodes interleukin 1 receptor-like 1 mediating the inflammatory responses in cardiac disease or injury [13]. IL1RL1 is able to predict the presence and outcome of heart failure [14, 15] and plays an important role in myocardial injury [16]. IL1RL1 gene variants are shown to be associated with angiographic severity of CHD [17]. As an inhibitor of growth factor and cytokine signaling pathways, SH2B adapter protein 3 (SH2B3) is a key regulator of integrin signaling in endothelial cells and control cell adhesion and migration [18]. SH2B3 gene variants are found to be associated with the type 1 diabetes [19], cardiovascular disease [20], and MI [8].

The goal of our study is to summarize the contribution of the polymorphisms of four genes (IKZF2, IL1RL, SH2B3 and IL5) to the risk of CHD by meta-analyses and a case–control study in Eastern Chinese.

Methods and materials

Literature review and data extraction

Meta-analyses were performed for the studies examining the association of the four SNPs with CHD. We did a literature search up to March 21, 2012 in multiple literature databases including PubMed and Embase, Web of Science and China National Knowledge Infrastructure (CNKI), Wanfang Chinese Periodical Database. The keywords for the search included various combinations of terms such as “SH2B3”, “LNK”, “rs3184504” or “IL1RL1”, “ST2”, “rs1420101”, or “IKZF2”, “HELIOS”, “rs12619285”, or “IL5”, “rs4143832”, paired with “coronary artery disease” or “CHD” or “MI”, “allele” or “polymorphism”. The studies in the meta-analysis must be case–control or cohort studies with genotype and allele information to estimate odds ratio (OR) and 95 % confidence interval (95 % CI). Data were separately extracted from the candidate studies by two authors (YH and LX) based on a standard protocol, and any deviation was resolved by consensus. If there were numerous publications from the same study group, the most complete and recent results were extracted. The extracted information included the name of the first author, publication year, study population, counts of genotypes in cases and controls, study design, total number of cases and controls, OR and 95 % CI.

Case–control study sample collection

A total of 275 patients from Lihuili Hospital were included in the case–control study. Among them, 162 CHD patients were defined by the angiographic evidence that had stenosis more than 50 % in at least major coronary artery [21]. Participators with a history of prior angioplasty or coronary artery bypass surgery were also treated as the CHD cases. The 113 non-CHD controls were the patients who had a less than 50 % stenosis in the major coronary artery and did not have any atherosclerotic vascular disease. All the individuals were unrelated Han Chinese and didn’t have congenital heart diseases, primary valvular disease, cardiomyopathy and severe liver or kidney disease. Patients were examined by at least two independent cardiologists according to the standardized protocol of coronary angiography [22]. The blood samples were collected by the same investigators. The study was approved by the Ethical Committee of Lihuili Hospital of Ningbo University, and the informed written consent was obtained from all participants before testing.

SNP Genotyping

Genomic DNA was isolated from peripheral white blood cells using a conventional phenol/chloroform extraction method [23]. SNPs were genotyped by polymerase chain reaction (PCR) with primers as described in Supplemental Table 1. The cycling program consisted of 15 s of initial denaturation at 94 °C, followed by 45 cycles at 94 °C for 20 s, 56 °C for 30 s, primer extension at 72 °C for 1 min and a final extension at 72 °C for 3 min. Genotyping for the amplification stage was performed by PCR on GeneAmp® PCR System 9700 (Applied Biosystems, Foster City, CA). The iPLEX single base primer extension and subsequent tests were performed on a Sequenom® platform (Sequenom, San Diego, CA) [24].

Statistical analyses

The pooled OR values in the meta-analyses were estimated using the REVMAN software (version 5.1, Cochrane Collaboration, Oxford, United Kingdom) and the Stata software (version 11.0, Stata Corporation, College Station, TX). The heterogeneity of studies in the meta-analysis was calculated by a χ2 based Q test [25, 26]. The inconsistency index (I2 statistic) was also examined and a value of I2 > 50 % indicated a significant heterogeneity among the studies [25]. Meta-analysis with significant heterogeneity was performed using the random-effect model, otherwise a fixed-effect model would be applied [27]. Publication bias was evaluated using the funnel plots and Egger regression test [28]. Z test was used to conclude the pooled OR and statistical significance is set at p < 0.05.

For our case–control study, consistency of the genotype frequencies with Hardy–Weinberg equilibrium (HWE) was analyzed using Arlequin program (version 3.5) [29]. Differences in the genotype and allele frequencies between CHD cases and controls were determined by CLUMP22 software with 10,000 Monte Carlo simulations [30]. Power analysis was performed using the Power and Sample Size Calculation software (v3.0.43). A two-sided p < 0.05 was considered as significant.

Results

As shown in the Fig. 1, we retrieved a total of 46 studies initially. Then we excluded two review articles, 12 irrelevant studies, and two studies that were not a case–control study. After reading the full text of these articles, only 19 studies were involved with the association between rs3184504 and CHD. Of them, we filtered out five studies without enough information and three duplicated studies. Finally, 11 studies (including 19 stages) [8, 3140] were selected into the meta-analysis of rs3184504 (Fig. 1). The involved samples comprised 25,743 CHD cases and 68,812 controls from two ethnicities (Europeans and Asians). As shown in Fig. 2, there was a significant association between rs3184504 and the risk of CHD (OR = 1.13, 95 % CI = 1.10–1.16, p < 0.0001, fixed-effect method). Minimal heterogeneity was found in this meta-analysis (I2 = 12.5 %, p = 0.302). There was no visual publication bias in the funnel plot (Fig. 3).
https://static-content.springer.com/image/art%3A10.1007%2Fs11239-012-0862-z/MediaObjects/11239_2012_862_Fig1_HTML.gif
Fig. 1

Selection of studies in the meta-analysis

https://static-content.springer.com/image/art%3A10.1007%2Fs11239-012-0862-z/MediaObjects/11239_2012_862_Fig2_HTML.gif
Fig. 2

Meta-analysis of CHD association studies of rs3184504. GerMIFSI German MI Family Study I, MIGen Myocardial Infarction Genetics Consortium, GerMIFSII German MI Family Study II, COROGENE Corogene study, PennCATH PennCATH, Medstar MedSTAR, OHGS Ottawa Heart Genomics Study, HPS Heart Protection Study, PROMIS Pakistan Risk of Myocardial Infarction Study, LOLIPOP London Life Sciences Prospective Population, CHARGE Cohorts for Heart and Aging Research in Genome Epidemiology

https://static-content.springer.com/image/art%3A10.1007%2Fs11239-012-0862-z/MediaObjects/11239_2012_862_Fig3_HTML.gif
Fig. 3

Publication bias analysis by funnel plots for rs3184504

In order to investigate the contribution of the four SNPs to CHD in Han Chinese, we recruited a case–control cohort in the current study. No departure of HWE was observed for all the four SNPs (p > 0.05, data not shown). As shown in Table 1, the distribution of the genotypes and alleles was not significantly different between CHD cases and non-CHD controls (p > 0.05). A further breakdown comparison by gender was done between cases and controls on the allele and genotype levels. Again, no significant differences were found between CHD cases and non-CHD controls in the females and males subgroups (Table 2). Under the dominant and recessive inheritance models, no significant differences were observed between cases and controls (Supplemental Tables 2, 3, 4).
Table 1

Genotype and allele frequencies in SNPs of each gene in cases and controls

SNP/group

Genotypea

χ2

p (df = 2)

Allele

χ2

p (df = 1)

OR (95 % CI)

rs12619285

GG (%)

GA (%)

AA (%)

  

G (%)

A (%)

   

Case

75 (46.3)

77 (47.5)

10 (6.1)

3.76

0.16

227 (70.1)

97 (29.9)

3.15

0.07

1.38 (0.97–1.98)

Control

42 (37.2)

58 (51.3)

13 (11.5)

142 (62.8)

84 (37.2)

rs1420101

GG (%)

GA (%)

AA (%)

  

G (%)

A (%)

   

Case

62 (38.5)

73 (45.3)

26 (16.1)

1.31

0.52

197 (61.2)

125 (38.8)

0.30

0.58

1.10 (0.78–1.56)

Control

37 (32.7)

59 (52.2)

17 (15.1)

133 (58.8)

93 (41.2)

rs3184504

CC (%)

TC (%)

TT (%)

  

C (%)

T (%)

   

Case

161 (99.4)

1 (0.6)

0 (0.0)

0.7

1.00

323 (99.7)

1 (0.3)

0.70

1.00

NA

Control

113 (100.0)

0 (0.0)

0 (0.0)

226 (100.0)

0 (0.0)

rs4143832

CC (%)

CA (%)

AA (%)

  

C (%)

A (%)

   

Case

116 (71.6)

43 (26.5)

3 (1.9)

1.14

0.57

275 (84.9)

49 (15.1)

0.89

0.34

0.79 (0.47–1.30)

Control

85 (77.3)

23 (20.9)

2 (1.8)

193 (87.7)

27 (12.3)

aAll are in HWE

NA not analyzed

Table 2

Comparison of genotype and allele frequencies between cases and controls and distribution according to gender

SNP/sex

Group

Genotype (n)

χ2

p

Allele (n)

χ2

p

OR (95 % CI)

rs12619285

 

GG

GA

AA

  

G

A

   

Male

Case

51

58

8

1.84

0.40

160

74

1.49

0.22

1.33 (0.84–2.12)

Control

21

31

7

73

45

Female

Case

25

19

2

3.12

0.21

69

23

2.87

0.09

1.70 (0.92–3.13)

Control

21

27

6

69

39

rs1420101

 

GG

GA

AA

  

G

A

   

Male

Case

45

53

18

0.61

0.74

143

89

0.04

0.84

0.96 (0.60–1.51)

Control

22

30

7

74

44

Female

Case

17

20

9

1.20

0.55

54

38

0.33

0.57

1.18 (0.67–2.07)

Control

15

29

10

59

49

rs3184504

 

CC

TC

TT

  

C

T

   

Male

Case

116

1

0

0.51

0.78

233

1

NA

NA

NA

Control

59

0

0

118

0

Female

Case

46

0

0

0.00

1.00

92

0

NA

NA

NA

Control

54

0

0

108

0

rs4143832

 

CC

CA

AA

  

C

A

   

Male

Case

83

32

2

2.47

0.29

198

36

2.16

0.14

0.59 (0.29–1.20)

Control

46

11

0

103

11

Female

Case

34

11

1

0.23

0.89

79

13

0.04

0.84

1.08 (0.49–2.38)

Control

39

12

2

90

16

NA not analyzed

Since only one research [8] reported the association between CHD/MI and the three SNPs (rs1420101, rs12619285 and rs4143832), we combined our data and performed meta-analyses of the three variants. The results indicated that none of the three SNPs were associated with the risk of CHD (Supplemental Figure 1).

Discussions

In the present study, we evaluate the significance of the polymorphisms of the four eosinophil association genes (IKZF2, IL1RL, SH2B3 and IL5) through the meta-analysis and case–control study. SNP rs3184504 of SH2B3 gene is shown to be significantly associated with the risk of CHD in Europeans and South Asians. Our findings support the emerging hypothesis that SH2B3 gene participates in the risk of CHD.

Our meta-analyses show that only rs3184504 of SH2B3 gene is associated with the risk of CHD, while the combined analyses of other three SNPs are unable to produce any convincing evidence for their association with CHD risk. A meta-analysis [35] was performed in 2009 and found that rs3184504 was significantly associated with CHD (OR = 1.18, 95 % CI = 1.12–1.23, p = 4.23E−11). Seven association studies were subsequently performed to confirm the positive result in European ancestry populations [35]. However, inconsistent results were shown for rs3184504. This may be due to a lack of power for some studies or genetic heterogeneity in SH2B3 gene. Our meta-analysis of rs3184504 include 11 publication articles (including 19 study stages) and is involved with 94,555 participants in the European and South Asian populations. Our results establish that rs3184504 was able to predict a 13 % increased risk of CHD in European and South Asian populations (OR = 1.13, 95 % CI = 1.10–1.16, p < 0.0001).

Another goal of our study is to examine the relationship between four SNPs related to eosinophil and the risk of CHD in Han Chinese. Allele rs3184504-T of SH2B3 gene is very rare in Han Chinese (T % = 0.3 %) that is much lower than in European population (38.0 %) [8]. According to the information in the online HapMap dataset, allele rs3184504-T frequency in the HapMap-HCB is 1.1 % in contrast of 40.8 % in the HapMap-CEU. This agrees with our observation and implicates a significant ethnic difference for rs3184504 of SH2B3 gene between Chinese and Europeans (Fst = 0.61, iHS = −2.7559). SNP rs12619285 of IKZF2 gene shows significant association with CHD in the Europeans (G allele frequency = 74 %, p = 5 × 10−10) and the East Asians (G allele frequency = 36 %, p = 0.017) [8]. In the present study, a borderline significance is found in Han Chinese (p = 0.07). In addition, we also observe that G allele has a different frequency in Han Chinese (G allele frequency = 62.8 %). The allele frequency in the present population is consistent with the G allele frequency in HapMap CHB population (64.6 %). Compared with a total of 17,330 participants in the previous study [8], our sample size is relative smaller. Combined with all the evidence, the lack of association may be due to the genetic heterogeneity of this locus or a lack of power in our sample for this minor-effect genetic marker in Chinese. Further investigation in large sample is necessary to evaluate the ethnic difference in the contribution of this SNP to CHD in various populations.

Our case–control study is unable to replicate the positive association between the rest of the three SNPs and CHD in Han Chinese. These may be due to the heterogeneity from different environment of life, gene environment interaction and genetic background. A power estimation shows that rs12619285 of IKZF2 gene had a best power of 63.3 % to detect a susceptibility locus with an odds ratio of 1.50 at the nominal Type I error rate of 0.05 under the additive model. Our sample size may not be the optimal condition, but should be sufficient to describe a tendency that may guide clinical practice. We could not exclude the possibility that the negative findings of the four SNPs may be due to a lack of power. Further works of the three SNPs are warranted with a larger sample size.

As the pleiotropic multifunctional leukocytes, eosinophils are important for the initiation and propagation of diverse inflammatory responses [41] that are implicated in the pathophysiology of CHD [42]. Eosinophilia is able to promote thrombus growth [43, 44] and thus play an important role in the pathogenesis of CHD [45]. Eosinophil count has been found to be associated with coronary artery calcification in participants with clinical suspicion of CHD [46, 47].

Recent studies showed that the eosinophil related genes played an important role in initiating eosinophilic inflammation and activating other immune cells resulting in immune/inflammatory disease including cardiovascular disease [4850]. In addition to the four genes in the present study, there were other eosinophil related genes in susceptibility to CHD. For example, eotaxin as the eosinophil chemoattractant was overexpressed in the atherosclerotic lesions of CHD patients [51, 52]. Eotaxin gene variations were reported to affect the eosinophil count and be associated with an increased risk of CHD [53].

There were several limitations in our study. Firstly, the sample size in our study is moderate and may not be enough to detect susceptible gene with moderate or minor effect. Secondly, the control groups come from individuals with healthy angiographic evidence. However, there is a possibility for some of control subjects to develop CHD in the future. And this may be explained by epigenetic regulatory factors such as DNA methylation. These subjects may attenuate the detection of a significant association between the SNP and CHD [54]. Thirdly, only four SNPs are examined for their association with CHD. Other genetic variants in the tested genes may exist as the real functional markers contributing to the risk of CHD [55]. Fourthly, our study only focuses on a few SNPs of these genes for their association with CHD. The negative results in our study may not help draw conclusion of denying the role of eosinophil related genes in the pathogenesis of CHD.

In conclusion, our meta-analysis supports the significant association between rs3184504 of SH2B3 gene and the risk of CHD, although we are unable to find association between four SNPs (rs12619285, rs1420101, rs3184504 and rs4143832) and the risk of CHD in Han Chinese. Further studies are warranted to test the emerging hypothesis that other eosinophil related genes participate in CHD risk in Han Chinese.

Acknowledgments

The research was supported by the grants from: National Natural Science Foundation of China (31100919 and 30772155), Zhejiang Provincial Program for the Cultivation of High level Innovative Health Talents, Natural Science Foundation of Zhejiang Province (Y206608), K.C. Wong Magna Fund in Ningbo University, and Youth and Doctor Foundation of Ningbo (2005A610016). The authors gratefully acknowledge the support of K.C. Wong Education Foundation, Hong Kong.

Conflict of interest

The authors declare no conflict of interest.

Supplementary material

11239_2012_862_MOESM1_ESM.doc (136 kb)
Supplementary material 1 (DOC 136 kb)

Copyright information

© Springer Science+Business Media New York 2013