Mol Genet Genomics DOI 10.1007/s00438-014-0873-x
Original Paper
MicroRNAs related polymorphisms and genetic susceptibility to esophageal squamous cell carcinoma Yanhong Qu · Honghong Qu · Manli Luo · Peng Wang · Chunhua Song · Kaijuan Wang · Jianying Zhang · Liping Dai
Received: 30 March 2014 / Accepted: 30 May 2014 © Springer-Verlag Berlin Heidelberg 2014
Abstract Esophageal cancer (EC) is the sixth leading cause of cancer-associated death worldwide and the incidence and mortality in China are the highest. The single nucleotide polymorphisms (SNPs) related to microRNAs could lead to alteration in microRNA expression and contribute to the susceptibility of cancer. To evaluate the association between microRNA-related SNPs and EC, a case–control study including 381 patients with esophageal squamous cell carcinoma (ESCC) and 426 gender, agematched controls was carried out to investigate the genetic susceptibility of five microRNA-related SNPs (rs2910164 in microRNA-146a, rs11614913 in microRNA-196a-2, rs7813 in GEMIN4, rs1595066 and rs16845990 in ErbB4) as well as the interactions of gene–gene and gene–environment in the development of ESCC. Variant homozygote genotype of rs11614913 in microRNA-196a-2 and
rs1595066 in ErbB4 were significantly associated with reduced ESCC risk (ORadjusted: 0.62, 95 % CI: 0.39–0.99 and ORadjusted: 0.38, 95 % CI: 0.24–0.61). The analysis of haplotypes in ErbB4 gene showed significant increased ESCC risk in Grs1595066Crs16845990 and Grs1595066Trs16845990 haplotypes (ORadjusted: 1.46, 95 % CI: 1.08–1.99 and ORadjusted: 1.33, 95 % CI: 1.10–1.62), and inversely reduced ESCC risk in Ars1595066Crs16845990 and Ars1595066Trs16845990 haplotypes with OR (95 % CI) of 0.75 (0.60–0.94) and 0.65 (0.49–0.86), respectively. These findings suggest that the polymorphisms in the microRNA-related genes may affect susceptibility of ESCC in Chinese Han population and the gene–gene interactions play vital roles in the progression on esophageal cancer. Future studies with larger sample and different ethnic populations are required to support and validate our findings.
Communicated by S. Hohmann.
Keywords Esophageal squamous cell carcinoma · MicroRNA · Polymorphism · Susceptibility
Y. Qu and H. Qu have the same contribution to this work. Y. Qu · P. Wang · C. Song · K. Wang · J. Zhang · L. Dai (*) Department of Epidemiology and Biostatistics, College of Public Health, Zhengzhou University, 100 Science Avenue, 450001 Zhengzhou, China e-mail: [email protected] Y. Qu · P. Wang · C. Song · K. Wang · J. Zhang · L. Dai Key Laboratory of Tumor Epidemiology, Zhengzhou 450052, China H. Qu Information Center of Shenzhen Longgang District Maternity and Child Healthcare Hospital, Shenzhen 518172, China M. Luo Department of Interventional Radiology, The First Affiliate Hospital in Zhengzhou University, Zhengzhou 450052, China
Introduction Esophageal cancer (EC) is the sixth leading cause of cancer-associated death worldwide with increasing incidence rate (Brown et al. 2008). And the majority of EC patients are diagnosed at advanced stage with poor prognosis and the overall 5-year survival rate is only 5–10 % (Xing et al. 2003). Esophageal squamous cell carcinoma (ESCC) and esophageal adenocarcinoma (EAC) are the two main subtypes of EC. And in Chinese, ESCC is the dominant subtype. Some environmental factors such as smoking, drinking, nitrite, lack of fruits and vegetables, as well as tumor family history are risk factors of EC according to the studies of epidemiology (Yang et al. 2009; Lin et al. 2002; Sun
13
et al. 2010). Also some ECs are family aggregation (Zhang et al. 2000, 2010), which indicate an important role of inheritance in ECs. MicroRNAs are tiny noncoding RNAs that act as posttranscriptional gene regulatory elements (Sand et al. 2009). Emerging evidence indicates that microRNAs are involved in important biological processes related to differentiation, proliferation, apoptosis, metabolism and immunomodulation (Alvarez-Garcia and Miska 2005; Doench et al. 2003; He and Hannon 2004; Sonkoly and Pivarcsi 2009). MicroRNAs downregulate the target genes by binding to the 3′-untranslated region (Perrotti 2013). Previous studies showed that 52 % of the microRNAs located in the tumor related genomes or gene fragile sites (Calin et al. 2004), indicating that microRNAs may play critical roles in tumorigenesis. Single nucleotide polymorphisms (SNPs) are the most common type of gene variation, which were associated with disease susceptibility and individual response to medicine (Shastry 2009). SNPs in the microRNA-related genes contain three parts which include SNPs in microRNA genes, SNPs in microRNA biogenesis genes and SNPs in microRNA target genes. MicroRNA-related SNPs affect tumor susceptibilities by generation of new microRNAs, changes of mature microRNAs and effect of combination to the target genes (Ryan et al. 2010). Since microRNAs are important in the initiation and development of tumors (He et al. 2012), and have the potential to regulate the expression of hundreds of target mRNAs, SNPs in microRNAs may produce more significant functional consequences and represent an ideal candidate for disease prediction. To our knowledge, the associations between polymorphisms in microRNA-related genes and EC risk were remained largely unknown and the published results had inconsistencies. In the present study, molecular epidemiology and case–control study were conducted to identify association between polymorphisms in microRNA-related genes and susceptibility of ESCC. We took five SNPs including two SNPs of microRNA genes (microRNA146a: rs2910164, microRNA-196a-2: rs11614913), one SNP of gene in the canonical microRNA biogenesis pathway (GEMIN4: rs7813) and SNPs in microRNA target gene (ErbB4: rs1595066 and rs16845990). Simultaneously, these five SNPs were combined with environmental factors to assess the interaction of gene–gene and gene–environment during ESCC progression.
Mol Genet Genomics
ethnic Han Chinese population. Patients with newly diagnosed and histopathologically confirmed ESCC were consecutively recruited from The First and Second Affiliated Hospital of Zhengzhou University from May 2008 to January 2010 without chem- or radiotherapy. The controls frequency matched to cases on age (±5 years) and sex were individuals without cancer history selected from a digestive disease census in Xin’an City of Henan province. A guided questionnaire on demographic (name, age, sex, race, family history of tumor, etc.) and lifestyle factors was administered through face-to-face interviews by trained interviewers. People who smoked no less than one cigarette daily for more than 1 year were defined as smokers. Individuals who consumed more than 100 g of liquor once and more than once a week for at least 6 months were considered as drinkers. Additionally, a 5 ml of venous blood sample was collected with a coded tube with ethylenediamine tetraacetic acid from each subject. This study was approved by the Institutional Review Board of Zhengzhou University. Informed consent was obtained from each study participant. SNPs genotyping Genomic DNA was extracted using the Blood Genome DNA Extraction kit of TIANGEN BIOTECH (Beijing). The genotyping of miR-146a rs2910164 and miR-196a-2 rs11614913 were performed with allele-specific PCR (AS-PCR), while GEMIN4 rs7813, ErbB4 rs1595066 and rs16845990 polymorphisms were determined by polymerase chain reaction–restriction fragment length polymorphism (PCR–RFLP). Table 1 shows the information of the five SNPs. The sequencing of the 5 SNPs were obtained from dbSNP database of NCBI and sequencing comparison was conducted in CHIP website (http://snpper.chip.org/bio/snpper-enter-snp). After confirmation, PCR primers were designed using software of Primer 5.0. For quality control, all analyses were performed blindly without knowledge of the case or control status. In addition, 10 % of all samples were randomly selected to assess the reproducibility, and the results showed 100 % concordant. Furthermore, five PCR products of three genotypes in each SNP were selected and sent to BGI Sequencing (Beijing) for sequencing validation. Statistical analysis
Materials and methods Study subjects 381 ESCC cases and 426 normal controls were recruited in this study. And all subjects were genetically unrelated
13
Differences in the distributions of demographic characteristics, frequencies of genotypes and haplotypes between cases and controls were evaluated by χ2 test. Unconditional logistic regression was used to estimate odds ratios (ORs) and 95 % confidence intervals (CIs) for ESCC risk
Mol Genet Genomics Table 1 The information of the five SNPs
Gene
SNP ID
Function
Alleles
Primers
miR-146a
rs2910164
3-UTR
C/G
miR-196a-2
rs11614913
3-UTR
T/C
GEMIN4
rs7813
Missense
T/C
ErbB4
rs1595066
3-UTR
A/G
ErbB4
rs16845990
3-UTR
C/T
Forward 1: GGGTTGTGTCAGTGTCAGACTTC Forward 2: AGGTTGTGTCAGTGTCAGACGTG Reverse: CTGCCTTCTGTCTCCAGTCTTC Forward 1: CGGCAACAAGAAACGGC Forward 2: CTCGGCAACAAGAAACGGT Reverse: GACCCTCTTTGTCTGTCTCCAT Forward: CCTCACCTGCTATGAGACTTTG Reverse: CACCATTATTGCCATGACTTTT Forward: TAGCACTCCAAGAACCCTTTTGGGA Reverse: CCCTCACACAGCTGCTCCGTTTAAT Forward: GAAAGGAACAGGACAGCATCG Reverse: TGACATTTTGGGCAGAAGTAA
in relation to these five SNPs, with the adjustment for sex, age, smoking, alcohol use and family history of tumor. The online Hardy–Weinberg equilibrium for genotypic frequencies (http://ihg2.helmholtz-muenchen.de/cgi-bin/hw/hwa1. pl) was used to test the demographic representative of controls. The value of Linkage disequilibrium (LD) between rs1595066 and rs16845990 in ErbB4 gene was calculated by Haploview4.2. And online SHEsis (http://analysis.biox.cn/myAnalysis.php) was used for haplotype prediction, construction and analysis. The χ2 test by trend was conducted to analyze dose-dependent effects between ESCC risk and number of mutant sites. The multifactor dimensionality reduction (MDR) was performed to analyze gene– environment interaction. All statistical analyses were carried out using SPSS18.0.
Results Characteristics of the study population As shown in Table 2, a total of 381 ESCC cases and 426 controls were included in the study. As expected, no significant differences were seen between cases and controls with regard to age (χ2 = 4.44, P = 0.11) and gender (χ2 = 0.74, P = 0.39), which suggested that our frequency matching on age and sex was satisfactory. Similarly, smoking status and alcohol use among cases and controls were no significant differences. However, the ESCC cases were more like to have a family history of tumor than the healthy controls (χ2 = 14.23, P = 0.00). Association between single SNP and susceptibility of ESCC All individuals in cases and controls were successfully genotyped for the five SNPs. For quality control, 5 % of the
Table 2 Characteristics of 381 ESCC cases and 426 controls Characteristics Age 60 Gender Men Women Smoking status Non-smoker Smoker Alcohol use Non-drinker Drinker Family history of tumor No Yes
Cases (%) (n = 381)
Controls (%) (n = 426)
χ2
P
5 (1.3) 139 (36.5) 237 (62.2)
10 (2.3) 180 (42.3) 236 (55.4)
4.44
0.11
256 (67.2) 125 (32.8)
274 (64.3) 152 (35.7)
0.74
0.39
189 (59.6) 128 (40.4)
260 (61.0) 166 (39.0)
0.15
0.70
255 (80.7) 61 (19.3)
350 (82.4) 75 (17.6)
0.33
0.56
233 (73.7)
351 (85.0)
83 (26.3)
62 (15.0)
14.23
0.00
samples (5 % cases, 5 % controls) were randomly selected and genotyped in duplicate, and the results showed 100 % accordance. The genotype distributions and allele frequencies of the SNPs are summarized in Table 3. Compared with CC genotype, the individuals with TT genotype of rs11614913 in miR-196a-2 showed decreased risk to ESCC (the adjusted OR: 0.62; 95 % CI: 0.39–0.99). No significant differences were observed in frequencies of C and T alleles between cases and controls (χ 2 = 2.98, P = 0.08). As to ErbB4 rs1595066, genotype AA and GA/AA showed significantly reduction risk of ESCC with adjusted OR (95 % CI) as 0.38 (0.24–0.61) and 0.62 (0.45–0.85), respectively, when compared with
13
Mol Genet Genomics
Table 3 Association of five SNPs with ESCC risk χ2
OR (95 % CI)
P
OR (95 % CI)a
Pa
0.72 0.54 0.63
1.00 1.08 (0.76–1.53) 0.91 (0.58–1.44) 1.04 (0.74–1.45)
0.66 0.69 0.83
1.00 1.09 (0.78–4.53) 0.62 (0.39–0.99) 0.96 (0.69–1.32)
0.61 0.04 0.78
1.00 1.01 (0.74–1.39) 1.10 (0.64–1.90) 1.03 (0.76–1.39)
0.95 0.73 0.88
1.00 0.72 (0.52–1.02) 0.38 (0.24–0.61) 0.62 (0.45–0.85)
0.06 0.00 0.00
1.00 0.93 (0.67–1.29) 0.72 (0.44–1.18) 0.89 (0.65–1.21)
0.67 0.19 0.44
Genotype
Cases (%) (n = 381)
Controls (%) (n = 426)
rs2910164 CC CG GG CG/GG C G
116 (30.4) 203 (53.3) 62 (16.3) 265 (69.6) 435 (57.1) 327 (42.9)
123 (28.9) 228 (53.5) 75 (176) 303 (71.1) 474 (55.6) 378 (44.4)
rs11614913 CC CT TT CT/TT C T
126 (33.1) 207 (54.3) 48 (12.6) 255 (66.9) 459 (60.2) 303 (39.8)
133 (31.2) 211 (49.5) 82 (19.2) 293 (68.7) 477 (56.0) 375 (44.0)
TT CT CC CT/CC T C
163 (42.8) 181 (47.5) 37 (9.7) 218 (57.2) 507 (66.5) 255 (33.5)
186 (43.7) 203 (47.7) 37 (8.7) 240 (56.4) 575 (67.5) 277 (32.5)
rs1595066 GG AG AA AG/AA G A
144 (37.8) 192 (50.4) 45 (11.8) 237 (62.2) 480 (63.0) 282 (37.0)
114 (26.8) 215 (50.5) 97 (22.8) 312 (73.3) 443 (34.3) 409 (59.2)
rs16845990 TT CT CC CT/CC T
137 (36.0) 195 (51.2) 49 (12.9) 244 (64.1) 469 (61.5)
146 (34.3) 217 (50.9) 63 (14.8) 280 (65.7) 509 (59.7)
0.08 0.70 0.25
1.00 0.96 (0.71–1.30) 0.83 (0.53–1.29) 0.93 (0.70–1.24) 1.00
0.78 0.40 0.62
293 (38.5)
343 (40.3)
0.55
0.93 (0.76–1.13)
0.46
0.13 0.38 0.24 0.35
0.05 4.81 0.32 2.98
1.00 0.94 (0.69–1.03) 0.88 (0.58–1.34) 0.93 (0.69–1.25) 1.00 0.94 (0.77–1.15) 1.00 1.04 (0.76–1.41) 0.62 (0.40–0.95) 0.92 (0.68–1.23) 1.00 0.84 (0.69–1.02)
0.56
0.83 0.03 0.57 0.08
rs7813
C
0.01 0.27 0.06 0.17
4.71 21.39 11.26 19.87
1.00 1.02 (0.76–1.36) 1.14 (0.69–1.88) 1.04 (0.78–1.37) 1.00 1.04 (0.85–1.29) 1.00 0.71 (0.52–0.97) 0.37 (0.24–0.56) 0.60 (0.45–0.81) 1.00 0.64 (0.52–0.78)
0.91 0.61 0.80 0.68
0.03 0.00 0.00 0.00
a
Adjusted for sex, age, smoking, alcohol use and family history of tumor
Table 4 Haplotype analysis of two polymorphism sites in Erbb4
a
SNPs sequence: rs1595066, rs16845990
Haplotypea
Cases (%)
Controls (%)
OR (95 % CI)
χ2
P
AC AT GC
189.87 (24.9) 92.13 (12.1) 103.13 (13.5)
260.67 (30.6) 148.33 (17.4) 82.33 (9.7)
0.75 (0.60–0.94) 0.65 (0.49–0.86) 1.46 (1.08–1.99)
6.44 8.98 5.93
0.01 0.00 0.01
GT
376.87 (49.5)
360.67 (42.3)
1.33 (1.10–1.62)
8.23
0.00
GG genotype. However, the other three SNPs (miR-146a rs2910164, GEMIN4 rs7813 and ErbB4 rs16845990) did not show any significant difference in either genotype or allelic analysis.
13
Analysis of haplotype Haplotype analysis was performed with two SNPs (rs1595066 and rs16845990) across the ErbB4 gene
Mol Genet Genomics Table 5 Analysis of combined effect of the five SNPs
a
SNPs: rs2910164, rs11614913, rs7813, rs1595066 and rs16845990 b Values of trend Chi-square and Ptrend
Combined SNPsa
Cases (%) (n = 381)
Controls (%) (n = 426)
0–1 2 3 4 5
26 (6.8) 76 (20.0) 123 (32.3) 103 (27.0) 53 (13.9)
22 (5.1) 83 (19.5) 117 (27.5) 121 (28.4) 83 (19.5)
Total
381
426
Table 6 MDR models of selected polymorphisms and environment factors Model
TBAa
CVCb OR (95 % CI)
FH FH, smoking FH, smoking, rs7813
0.6062 10/10 2.97 (2.15–4.12)
Original Paper
MicroRNAs related polymorphisms and genetic susceptibility to esophageal squamous cell carcinoma Yanhong Qu · Honghong Qu · Manli Luo · Peng Wang · Chunhua Song · Kaijuan Wang · Jianying Zhang · Liping Dai
Received: 30 March 2014 / Accepted: 30 May 2014 © Springer-Verlag Berlin Heidelberg 2014
Abstract Esophageal cancer (EC) is the sixth leading cause of cancer-associated death worldwide and the incidence and mortality in China are the highest. The single nucleotide polymorphisms (SNPs) related to microRNAs could lead to alteration in microRNA expression and contribute to the susceptibility of cancer. To evaluate the association between microRNA-related SNPs and EC, a case–control study including 381 patients with esophageal squamous cell carcinoma (ESCC) and 426 gender, agematched controls was carried out to investigate the genetic susceptibility of five microRNA-related SNPs (rs2910164 in microRNA-146a, rs11614913 in microRNA-196a-2, rs7813 in GEMIN4, rs1595066 and rs16845990 in ErbB4) as well as the interactions of gene–gene and gene–environment in the development of ESCC. Variant homozygote genotype of rs11614913 in microRNA-196a-2 and
rs1595066 in ErbB4 were significantly associated with reduced ESCC risk (ORadjusted: 0.62, 95 % CI: 0.39–0.99 and ORadjusted: 0.38, 95 % CI: 0.24–0.61). The analysis of haplotypes in ErbB4 gene showed significant increased ESCC risk in Grs1595066Crs16845990 and Grs1595066Trs16845990 haplotypes (ORadjusted: 1.46, 95 % CI: 1.08–1.99 and ORadjusted: 1.33, 95 % CI: 1.10–1.62), and inversely reduced ESCC risk in Ars1595066Crs16845990 and Ars1595066Trs16845990 haplotypes with OR (95 % CI) of 0.75 (0.60–0.94) and 0.65 (0.49–0.86), respectively. These findings suggest that the polymorphisms in the microRNA-related genes may affect susceptibility of ESCC in Chinese Han population and the gene–gene interactions play vital roles in the progression on esophageal cancer. Future studies with larger sample and different ethnic populations are required to support and validate our findings.
Communicated by S. Hohmann.
Keywords Esophageal squamous cell carcinoma · MicroRNA · Polymorphism · Susceptibility
Y. Qu and H. Qu have the same contribution to this work. Y. Qu · P. Wang · C. Song · K. Wang · J. Zhang · L. Dai (*) Department of Epidemiology and Biostatistics, College of Public Health, Zhengzhou University, 100 Science Avenue, 450001 Zhengzhou, China e-mail: [email protected] Y. Qu · P. Wang · C. Song · K. Wang · J. Zhang · L. Dai Key Laboratory of Tumor Epidemiology, Zhengzhou 450052, China H. Qu Information Center of Shenzhen Longgang District Maternity and Child Healthcare Hospital, Shenzhen 518172, China M. Luo Department of Interventional Radiology, The First Affiliate Hospital in Zhengzhou University, Zhengzhou 450052, China
Introduction Esophageal cancer (EC) is the sixth leading cause of cancer-associated death worldwide with increasing incidence rate (Brown et al. 2008). And the majority of EC patients are diagnosed at advanced stage with poor prognosis and the overall 5-year survival rate is only 5–10 % (Xing et al. 2003). Esophageal squamous cell carcinoma (ESCC) and esophageal adenocarcinoma (EAC) are the two main subtypes of EC. And in Chinese, ESCC is the dominant subtype. Some environmental factors such as smoking, drinking, nitrite, lack of fruits and vegetables, as well as tumor family history are risk factors of EC according to the studies of epidemiology (Yang et al. 2009; Lin et al. 2002; Sun
13
et al. 2010). Also some ECs are family aggregation (Zhang et al. 2000, 2010), which indicate an important role of inheritance in ECs. MicroRNAs are tiny noncoding RNAs that act as posttranscriptional gene regulatory elements (Sand et al. 2009). Emerging evidence indicates that microRNAs are involved in important biological processes related to differentiation, proliferation, apoptosis, metabolism and immunomodulation (Alvarez-Garcia and Miska 2005; Doench et al. 2003; He and Hannon 2004; Sonkoly and Pivarcsi 2009). MicroRNAs downregulate the target genes by binding to the 3′-untranslated region (Perrotti 2013). Previous studies showed that 52 % of the microRNAs located in the tumor related genomes or gene fragile sites (Calin et al. 2004), indicating that microRNAs may play critical roles in tumorigenesis. Single nucleotide polymorphisms (SNPs) are the most common type of gene variation, which were associated with disease susceptibility and individual response to medicine (Shastry 2009). SNPs in the microRNA-related genes contain three parts which include SNPs in microRNA genes, SNPs in microRNA biogenesis genes and SNPs in microRNA target genes. MicroRNA-related SNPs affect tumor susceptibilities by generation of new microRNAs, changes of mature microRNAs and effect of combination to the target genes (Ryan et al. 2010). Since microRNAs are important in the initiation and development of tumors (He et al. 2012), and have the potential to regulate the expression of hundreds of target mRNAs, SNPs in microRNAs may produce more significant functional consequences and represent an ideal candidate for disease prediction. To our knowledge, the associations between polymorphisms in microRNA-related genes and EC risk were remained largely unknown and the published results had inconsistencies. In the present study, molecular epidemiology and case–control study were conducted to identify association between polymorphisms in microRNA-related genes and susceptibility of ESCC. We took five SNPs including two SNPs of microRNA genes (microRNA146a: rs2910164, microRNA-196a-2: rs11614913), one SNP of gene in the canonical microRNA biogenesis pathway (GEMIN4: rs7813) and SNPs in microRNA target gene (ErbB4: rs1595066 and rs16845990). Simultaneously, these five SNPs were combined with environmental factors to assess the interaction of gene–gene and gene–environment during ESCC progression.
Mol Genet Genomics
ethnic Han Chinese population. Patients with newly diagnosed and histopathologically confirmed ESCC were consecutively recruited from The First and Second Affiliated Hospital of Zhengzhou University from May 2008 to January 2010 without chem- or radiotherapy. The controls frequency matched to cases on age (±5 years) and sex were individuals without cancer history selected from a digestive disease census in Xin’an City of Henan province. A guided questionnaire on demographic (name, age, sex, race, family history of tumor, etc.) and lifestyle factors was administered through face-to-face interviews by trained interviewers. People who smoked no less than one cigarette daily for more than 1 year were defined as smokers. Individuals who consumed more than 100 g of liquor once and more than once a week for at least 6 months were considered as drinkers. Additionally, a 5 ml of venous blood sample was collected with a coded tube with ethylenediamine tetraacetic acid from each subject. This study was approved by the Institutional Review Board of Zhengzhou University. Informed consent was obtained from each study participant. SNPs genotyping Genomic DNA was extracted using the Blood Genome DNA Extraction kit of TIANGEN BIOTECH (Beijing). The genotyping of miR-146a rs2910164 and miR-196a-2 rs11614913 were performed with allele-specific PCR (AS-PCR), while GEMIN4 rs7813, ErbB4 rs1595066 and rs16845990 polymorphisms were determined by polymerase chain reaction–restriction fragment length polymorphism (PCR–RFLP). Table 1 shows the information of the five SNPs. The sequencing of the 5 SNPs were obtained from dbSNP database of NCBI and sequencing comparison was conducted in CHIP website (http://snpper.chip.org/bio/snpper-enter-snp). After confirmation, PCR primers were designed using software of Primer 5.0. For quality control, all analyses were performed blindly without knowledge of the case or control status. In addition, 10 % of all samples were randomly selected to assess the reproducibility, and the results showed 100 % concordant. Furthermore, five PCR products of three genotypes in each SNP were selected and sent to BGI Sequencing (Beijing) for sequencing validation. Statistical analysis
Materials and methods Study subjects 381 ESCC cases and 426 normal controls were recruited in this study. And all subjects were genetically unrelated
13
Differences in the distributions of demographic characteristics, frequencies of genotypes and haplotypes between cases and controls were evaluated by χ2 test. Unconditional logistic regression was used to estimate odds ratios (ORs) and 95 % confidence intervals (CIs) for ESCC risk
Mol Genet Genomics Table 1 The information of the five SNPs
Gene
SNP ID
Function
Alleles
Primers
miR-146a
rs2910164
3-UTR
C/G
miR-196a-2
rs11614913
3-UTR
T/C
GEMIN4
rs7813
Missense
T/C
ErbB4
rs1595066
3-UTR
A/G
ErbB4
rs16845990
3-UTR
C/T
Forward 1: GGGTTGTGTCAGTGTCAGACTTC Forward 2: AGGTTGTGTCAGTGTCAGACGTG Reverse: CTGCCTTCTGTCTCCAGTCTTC Forward 1: CGGCAACAAGAAACGGC Forward 2: CTCGGCAACAAGAAACGGT Reverse: GACCCTCTTTGTCTGTCTCCAT Forward: CCTCACCTGCTATGAGACTTTG Reverse: CACCATTATTGCCATGACTTTT Forward: TAGCACTCCAAGAACCCTTTTGGGA Reverse: CCCTCACACAGCTGCTCCGTTTAAT Forward: GAAAGGAACAGGACAGCATCG Reverse: TGACATTTTGGGCAGAAGTAA
in relation to these five SNPs, with the adjustment for sex, age, smoking, alcohol use and family history of tumor. The online Hardy–Weinberg equilibrium for genotypic frequencies (http://ihg2.helmholtz-muenchen.de/cgi-bin/hw/hwa1. pl) was used to test the demographic representative of controls. The value of Linkage disequilibrium (LD) between rs1595066 and rs16845990 in ErbB4 gene was calculated by Haploview4.2. And online SHEsis (http://analysis.biox.cn/myAnalysis.php) was used for haplotype prediction, construction and analysis. The χ2 test by trend was conducted to analyze dose-dependent effects between ESCC risk and number of mutant sites. The multifactor dimensionality reduction (MDR) was performed to analyze gene– environment interaction. All statistical analyses were carried out using SPSS18.0.
Results Characteristics of the study population As shown in Table 2, a total of 381 ESCC cases and 426 controls were included in the study. As expected, no significant differences were seen between cases and controls with regard to age (χ2 = 4.44, P = 0.11) and gender (χ2 = 0.74, P = 0.39), which suggested that our frequency matching on age and sex was satisfactory. Similarly, smoking status and alcohol use among cases and controls were no significant differences. However, the ESCC cases were more like to have a family history of tumor than the healthy controls (χ2 = 14.23, P = 0.00). Association between single SNP and susceptibility of ESCC All individuals in cases and controls were successfully genotyped for the five SNPs. For quality control, 5 % of the
Table 2 Characteristics of 381 ESCC cases and 426 controls Characteristics Age 60 Gender Men Women Smoking status Non-smoker Smoker Alcohol use Non-drinker Drinker Family history of tumor No Yes
Cases (%) (n = 381)
Controls (%) (n = 426)
χ2
P
5 (1.3) 139 (36.5) 237 (62.2)
10 (2.3) 180 (42.3) 236 (55.4)
4.44
0.11
256 (67.2) 125 (32.8)
274 (64.3) 152 (35.7)
0.74
0.39
189 (59.6) 128 (40.4)
260 (61.0) 166 (39.0)
0.15
0.70
255 (80.7) 61 (19.3)
350 (82.4) 75 (17.6)
0.33
0.56
233 (73.7)
351 (85.0)
83 (26.3)
62 (15.0)
14.23
0.00
samples (5 % cases, 5 % controls) were randomly selected and genotyped in duplicate, and the results showed 100 % accordance. The genotype distributions and allele frequencies of the SNPs are summarized in Table 3. Compared with CC genotype, the individuals with TT genotype of rs11614913 in miR-196a-2 showed decreased risk to ESCC (the adjusted OR: 0.62; 95 % CI: 0.39–0.99). No significant differences were observed in frequencies of C and T alleles between cases and controls (χ 2 = 2.98, P = 0.08). As to ErbB4 rs1595066, genotype AA and GA/AA showed significantly reduction risk of ESCC with adjusted OR (95 % CI) as 0.38 (0.24–0.61) and 0.62 (0.45–0.85), respectively, when compared with
13
Mol Genet Genomics
Table 3 Association of five SNPs with ESCC risk χ2
OR (95 % CI)
P
OR (95 % CI)a
Pa
0.72 0.54 0.63
1.00 1.08 (0.76–1.53) 0.91 (0.58–1.44) 1.04 (0.74–1.45)
0.66 0.69 0.83
1.00 1.09 (0.78–4.53) 0.62 (0.39–0.99) 0.96 (0.69–1.32)
0.61 0.04 0.78
1.00 1.01 (0.74–1.39) 1.10 (0.64–1.90) 1.03 (0.76–1.39)
0.95 0.73 0.88
1.00 0.72 (0.52–1.02) 0.38 (0.24–0.61) 0.62 (0.45–0.85)
0.06 0.00 0.00
1.00 0.93 (0.67–1.29) 0.72 (0.44–1.18) 0.89 (0.65–1.21)
0.67 0.19 0.44
Genotype
Cases (%) (n = 381)
Controls (%) (n = 426)
rs2910164 CC CG GG CG/GG C G
116 (30.4) 203 (53.3) 62 (16.3) 265 (69.6) 435 (57.1) 327 (42.9)
123 (28.9) 228 (53.5) 75 (176) 303 (71.1) 474 (55.6) 378 (44.4)
rs11614913 CC CT TT CT/TT C T
126 (33.1) 207 (54.3) 48 (12.6) 255 (66.9) 459 (60.2) 303 (39.8)
133 (31.2) 211 (49.5) 82 (19.2) 293 (68.7) 477 (56.0) 375 (44.0)
TT CT CC CT/CC T C
163 (42.8) 181 (47.5) 37 (9.7) 218 (57.2) 507 (66.5) 255 (33.5)
186 (43.7) 203 (47.7) 37 (8.7) 240 (56.4) 575 (67.5) 277 (32.5)
rs1595066 GG AG AA AG/AA G A
144 (37.8) 192 (50.4) 45 (11.8) 237 (62.2) 480 (63.0) 282 (37.0)
114 (26.8) 215 (50.5) 97 (22.8) 312 (73.3) 443 (34.3) 409 (59.2)
rs16845990 TT CT CC CT/CC T
137 (36.0) 195 (51.2) 49 (12.9) 244 (64.1) 469 (61.5)
146 (34.3) 217 (50.9) 63 (14.8) 280 (65.7) 509 (59.7)
0.08 0.70 0.25
1.00 0.96 (0.71–1.30) 0.83 (0.53–1.29) 0.93 (0.70–1.24) 1.00
0.78 0.40 0.62
293 (38.5)
343 (40.3)
0.55
0.93 (0.76–1.13)
0.46
0.13 0.38 0.24 0.35
0.05 4.81 0.32 2.98
1.00 0.94 (0.69–1.03) 0.88 (0.58–1.34) 0.93 (0.69–1.25) 1.00 0.94 (0.77–1.15) 1.00 1.04 (0.76–1.41) 0.62 (0.40–0.95) 0.92 (0.68–1.23) 1.00 0.84 (0.69–1.02)
0.56
0.83 0.03 0.57 0.08
rs7813
C
0.01 0.27 0.06 0.17
4.71 21.39 11.26 19.87
1.00 1.02 (0.76–1.36) 1.14 (0.69–1.88) 1.04 (0.78–1.37) 1.00 1.04 (0.85–1.29) 1.00 0.71 (0.52–0.97) 0.37 (0.24–0.56) 0.60 (0.45–0.81) 1.00 0.64 (0.52–0.78)
0.91 0.61 0.80 0.68
0.03 0.00 0.00 0.00
a
Adjusted for sex, age, smoking, alcohol use and family history of tumor
Table 4 Haplotype analysis of two polymorphism sites in Erbb4
a
SNPs sequence: rs1595066, rs16845990
Haplotypea
Cases (%)
Controls (%)
OR (95 % CI)
χ2
P
AC AT GC
189.87 (24.9) 92.13 (12.1) 103.13 (13.5)
260.67 (30.6) 148.33 (17.4) 82.33 (9.7)
0.75 (0.60–0.94) 0.65 (0.49–0.86) 1.46 (1.08–1.99)
6.44 8.98 5.93
0.01 0.00 0.01
GT
376.87 (49.5)
360.67 (42.3)
1.33 (1.10–1.62)
8.23
0.00
GG genotype. However, the other three SNPs (miR-146a rs2910164, GEMIN4 rs7813 and ErbB4 rs16845990) did not show any significant difference in either genotype or allelic analysis.
13
Analysis of haplotype Haplotype analysis was performed with two SNPs (rs1595066 and rs16845990) across the ErbB4 gene
Mol Genet Genomics Table 5 Analysis of combined effect of the five SNPs
a
SNPs: rs2910164, rs11614913, rs7813, rs1595066 and rs16845990 b Values of trend Chi-square and Ptrend
Combined SNPsa
Cases (%) (n = 381)
Controls (%) (n = 426)
0–1 2 3 4 5
26 (6.8) 76 (20.0) 123 (32.3) 103 (27.0) 53 (13.9)
22 (5.1) 83 (19.5) 117 (27.5) 121 (28.4) 83 (19.5)
Total
381
426
Table 6 MDR models of selected polymorphisms and environment factors Model
TBAa
CVCb OR (95 % CI)
FH FH, smoking FH, smoking, rs7813
0.6062 10/10 2.97 (2.15–4.12)
