Haplotype-Based Association of Four Lymphotoxin-Alpha Gene Polymorphisms with the Risk of Coronary Artery Disease in Han Chinese

Tohoku J. Exp. Med., 2011, 224, 119-125Lymphotoxin-Alpha Gene SNPs and CAD in Chinese 119

Yan Liu,1 Haihui Sheng,2 Lin Lu,1 Zhijun Wu,1 Qiujin Chen,1 Huasheng Xiao2 and Wei Jin1

1Department of Cardiology, Ruijin Hospital, Shanghai Jiaotong University School of Medicine, Shanghai, P.R. China 2National Engineering Center for Biochip at Shanghai, Shanghai, P.R. China

Lymphotoxin-alpha (LTA), a pro-inflammatory cytokine, has been implicated in the pathogenesis of coronary atherosclerosis. Meanwhile, association of some single nucleotide polymorphisms (SNPs) of LTA gene with coronary artery disease (CAD) has been evaluated; however, the results are irreproducible. We therefore investigated the relationship between four SNPs of LTA gene and CAD in Han Chinese: G+10A (rs1800683, 5′-untranslated region), A+80C (rs2239704, 5′-untranslated region), T+496C (Cys13Arg, rs2229094, exon 2), and C+804A (Thr26Asn, rs1041981, exon 3). Genotyping was performed in 438 CAD patients and 330 healthy controls. Single-locus analysis showed that the genotype and allele frequencies of G+10A polymorphism exhibited marginal differences between CAD patients and controls, although no statistical significance was observed after the Bonferroni correction. Logistic regression analysis revealed that GG genotype of G+10A polymorphism was significantly associated with the risk of CAD under the dominant mode, whereas no significant association was detected between A+80C polymorphism and CAD. In contrast, individuals carrying TT or TC genotype of T+496C polymorphism showed a decreased CAD risk relative to those with CC genotype under the recessive mode. Likewise, CC genotype of C+804A polymorphism was associated with a protective effect on CAD under the dominant mode. Further, in haplotype analysis, the haplotype G-C-T-C (in order of rs1800683, rs2239704, rs2229094 and rs1041981) was significantly associated with a decreased risk of CAD after assigning the most common haplotype A-C-T-A as a reference. In conclusion, we show a protective effect of the haplotype G-C-T-C on the occurrence of CAD, suggesting the involvement of LTA in CAD pathogenesis.

Coronary artery disease (CAD) is a complex multifac- the risk of CAD (Jang et al. 2007; Koch et al. 2007; Palikhe torial disorder to which genetic and environmental factors et al. 2007). Meanwhile, family-based linkage analyses contribute interactively; it has become one of the leading (PROCARDIS Consortium 2004) and genome-wide associ- causes of death in China. Ample evidence exists proving ation studies (Iida et al. 2003; Koch et al. 2005; Ozaki and the involvement of inflammatory processes in all phases of Tanaka 2005; Tanaka and Ozaki 2006) had implicated some atherosclerotic lesion progression linked to various cyto- single nucleotide polymorphisms (SNPs) of LTA gene, such kines (Berliner et al. 1995; Ross 1999). Lymphotoxin- as A allele of C+804A polymorphism and C allele of alpha (LTA), a pro-inflammatory cytokine, plays an impor- T+496C polymorphism, were in predisposition to CAD or tant part in activating inflammatory and immunomodulatory MI, but the results were not often reproducible (Clarke et processes, suggesting its implication in the pathogenesis of al. 2006; Asselbergs et al. 2007; Park et al. 2007; Sedlacek atherosclerosis and CAD (Schreyer et al. 2002; Ozaki et al. et al. 2007; Palikhe et al. 2008; Ryan et al. 2008; Gao et al. 2004). 2010; Li et al. 2010). The LTA gene, firstly identified in 2002, was found to Suna et al. (2008) reported that LTA stimulation might be independently associated with myocardial infarction up-regulate various cytokines’ expression associated with (MI) in Japanese (Ozaki et al. 2002). Subsequently, several signal transduction, cell adhesion and chemoattraction, such studies confirmed the relationship between LTA gene and as vascular cell adhesion molecule 1 (VCAM-1), E-selectin

Received December 27, 2010; revision accepted for publication May 11, 2011. doi: 10.1620/tjem.224.119 Correspondence: Wei Jin, M.D., PH.D., Department of Cardiology, Ruijin Hospital, Shanghai Jiaotong University School of Medicine, 197 Ruijin Er Road, Shanghai 200025, P.R. China. e-mail: [email protected]

119 120 Y. Liu et al. and monocyte chemotactic protein 1 (MCP 1) in human Genotyping analysis endothelial cells via the activation of the transcription Genomic DNA was extracted from peripheral blood leukocytes nuclear factor of kappa light polypeptide gene enhancer in by standard phenol-chloroform extraction. The genotypes of four B-cells (NFkB) as in tumor necrosis factor (TNF) receptor evaluated SNPs were determined by the MassARRAY technology signaling by TNF-α (a family member of LTA). Studies on platform (Sequenom Inc., San Diego, California, USA). The test was the role of LTA in the procedure of monocyte adhesion to based on the principle of allele-specific primer extension reaction and the read-out of the result was realized on the matrix-assisted laser endothelial cells, which was an initial and crucial step of desorption/ionization time-of-flight mass spectrometry (MALDI-TOF atherosclerosis, also raised the contribution of LTA genetic MS) analytic system. All the polymerase chain reaction (PCR) prim- polymorphisms to the regulation of those cytokines (Suna ers and extension primers were designed by MassARRAY Assay et al. 2008, 2009). Asselbergs et al. (2007) revealed that the Design 2.0 software according to reference sequence (NT_007592). A allele of C+804A polymorphism was associated with ele- To provide a significant improvement in overall performance, 10-mer vated plasma level of VCAM-1 in CAD cases. It is thus of tag (5′-ACGTTGGATG-3′) was added to the 5′ ends of each PCR interest to confirm which or how many SNPs of LTA gene primer (Table 2). might have functional potentials to affect the final bioavail- The amplification mixture of total volume 5 μL contained: 1 μL ability of LTA, and further the development of CAD. of genomic DNA (10 ng/μL), 0.325 μL MgCl2 (25 mmol/L), 0.25 μL It has been postulated that some SNPs in the putative dNTP (10 mmol/L), 0.625 μL 10 × buffer, 0.125 μL of each primer (10 promoter, transcriptional-regular regions or coding regions μmol/L) and 0.03 μL Hotstar® Taq DNA polymerase (all from Qiagen could affect the efficacy of transcription or the function of Inc., Valencia, California, USA). The amplification conditions were: protein, while haplotype analysis, which focuses on SNPs 95°C for 15 min, followed by 45 cycles at 95°C for 20 s, 56°C for in their combination simultaneously, has a higher complex- 30 s, 72°C for 1 min, then a final extension at 72°C for 3 min. ity level than single-locus analysis and more power to Preparation for extend reaction included incubation of 5 μL PCR explore the association between candidate genes and com- products with 0.3 μL shrimp alkaline phosphatase (SAP, Sequenom Inc., San Diego, California, USA) and 0.17 μL hME buffer at 37°C plex disease (Rioux et al. 2001; Tabor et al. 2002; Buckland for 20 min, followed by heat inactivation at 85°C for 5 min. et al. 2004). In the present study, to explore the association The extension mixture contained, in a final volume of 9 μL: of SNPs, both individually and as haplotypes, with the risk 7 μL of purified PCR products, 0.2 μL homogeneous MassEXTEND of CAD in Han Chinese, we focused on four SNPs: G+10A Mix (Sequenom Inc., San Diego, California, USA), 9 μM of each (rs1800683) and A+80C (rs2239704) in the 5′-untranslated extend primer and 0.018 μL MassEXTEND enzyme (Sequenom Inc., region encoded by exon 1, T+496C in exon 2 (rs2229094, a San Diego, California, USA). The extension conditions were: 94°C non-synonymous coding SNP that substitutes arginine for for 2 min, followed by 55 cycles at 94°C for 5 min, 52°C for 5 min, cysteine at codon 13, Cys13Arg), and C+804A in exon 3 72°C for 5 min, then a final extension at 72°C for 3 min. PCR (rs1041981, a non-synonymous coding SNP resulting in a products were purified with 3 mg of clean Resin (Sequenom Inc., San threonine to asparagine amino acid substitution at codon Diego, California, USA) and then sequenced using MassARRAY 26, Thr26Asn). Analyzer, version 3.0.1 (Sequenom Inc., San Diego, California, USA) on 384-well SpectroCHIP (Sequenom Inc., San Diego, California, Materials and Methods USA). Study population Our study population comprised 768 unrelated Han Chinese Statistical methods who were admitted to Ruijin Hospital, Shanghai Jiaotong University Continuous variables were expressed as mean ± standard School of Medicine when they were experiencing various symptoms deviation (s.d.) and compared by the unpaired Student's t-test. or for a medical checkup from June 2006 to June 2008. Participants Genotype/allele frequencies were calculated by the counting method were divided into 2 groups. The CAD group contained 438 patients and the differences in genotype/allele frequencies between CAD 2 aged 61.11 ± 9.66 years and the diagnosis of CAD was established group and healthy controls were analyzed by χ test. Each genotype angiographically in the presence of more than 50% stenosis in at least was assessed by logistic regression analysis assuming additive (major one of the three major coronary arteries or major branches. Patients homozygotes vs. heterozygotes vs. minor homozygotes), dominant with simple spasm of coronary arteries, myocardial bridge or other (major homozygotes vs. heterozygotes plus minor homozygotes) and non-coronary atherosclerotic lesions were excluded. The healthy recessive (major homozygotes plus heterozygotes vs. minor homozy- controls contained 330 participants aged 53.16 ± 7.38 years who gotes) modes of inheritance, crude odds ratio (ORcrude) and adjusted received routine health examinations in the hospital. They had odds ratio (ORadjusted) and 95% confidence intervals (95% CI) were normal electrocardiographs and blood pressure, no symptoms of chest calculated as well, respectively without or with considering other pain, and no family history of hypertension or CAD in first-degree covariates, including gender, age, blood pressure, the serum levels of relatives. Participants were not eligible if they had diabetes mellitus, fasting glucose and lipid profiles. To account for multiple testing, the chronic liver diseases, uremia, acute infection, autoimmune diseases Bonferroni method was applied. Significance would be claimed if the or cancers. Written informed consents were obtained from each overall two-sided P value was < 0.05. Statistical analyses were done participant, and the project was approved by the Ethics Committee of with SPSS 13.0 software (SPSS Inc., Chicago, Illinois, USA). Study Ruijin Hospital, Shanghai Jiaotong University School of Medicine. power was calculated by PS (Power and Sample Size Calculations) The baseline and clinical characteristics of study population are software (Version 3.0) (Dupont and Plummer 1990). shown in Table 1. Hardy-Weinberg equilibrium calculations were performed with Lymphotoxin-Alpha Gene SNPs and CAD in Chinese 121 the Arlequin program (http://anthro.unige.ch/software/arlequin). The ence was reached after the Bonferroni correction. There haplo.em program was used to estimate the haplotype frequencies. was no significant association between each of other three The haplo.cc and haplo.glm programs were employed to calculate SNPs and CAD. ORcrude and ORadjusted and 95% CI for each haplotype, respectively. Logistic regression analysis (Table 4) indicated that These two approaches are based on a generalized linear mode, and the GG genotype of G+10A polymorphism was signifi- compute the regression of a trait on haplotypes and above covariates. cantly associated with the risk of suffering CAD under the Only the haplotypes with frequencies greater than 3% were consid- dominant mode, which preserved significance after adjust- ered. The haplo.em, haplo.cc and haplo.glm were implemented in the ment for gender, age, blood pressure, the serum levels of Haplo.stats software (version 1.4.0) developed by the R language (http://www.r-project.org). fasting glucose and lipid profiles (GG vs. GA + AA: ORcrude = 6.554, 95% CI: 2.129-20.174, P = 0.001; ORadjusted = 11.352, 95% CI: 1.843-69.901, P = 0.009). In contrast, Results individuals carrying the TT or TC genotype of T+496C Clinical characteristics of the study population polymorphism had a decreased CAD risk relative to those Compared with healthy controls, CAD patients were with the CC genotype under the recessive mode after cor- older and had higher blood pressure. Also, they had higher rection (TT + TC vs. CC: ORadjusted = 0.202, 95% CI: 0.047- serum levels of fasting glucose and triglycerides (TG), and 0.874, P = 0.032). Similarly, the CC genotype of C+804A lower levels of high density lipoprotein-cholesterol (HDL- polymorphism was associated with a protective effect on C). Serum levels of total cholesterol (TC) and low density CAD under the dominant mode, even after adjusting for lipoprotein cholesterol (LDL-C) were similar between two those confounders (CC vs. AA + CA: ORcrude = 0.240, 95% groups (Table 1). CI: 0.078-0.735, P = 0.012; ORadjusted = 0.162, 95% CI: 0.026-0.997, P = 0.050). For A+80C polymorphism, we Association between LTA gene polymorphisms and CAD failed to observe a significant association with CAD. For each polymorphism, the genotype/allele frequen- Four SNPs were incorporated in the haplotype analy- cies were in equilibrium with the Hardy-Weinberg equation. sis, and they contributed to thirteen haplotypes in our study Single-locus analysis (Table 3) showed that genotype and population. Haplotype A-C-T-A (alleles in order of allele frequencies of G+10A polymorphism exhibited dif- rs1800683, rs2239704, rs2229094 and rs1041981) was ferences between CAD patients and healthy controls (P = observed comprising 40% of the total. Two additional hap- 0.032, 0.047, respectively); however, no significant differ- lotypes, G-A-T-C and G-C-C-C, had frequencies around

Table 1. Baseline and Clinical characteristics of the study population. Characteristics Controls (n = 330) CAD (n = 438) Age (years) 53.16 ± 7.32 61.11 ± 9.66a Systolic blood pressure (mmHg) 117.99 ± 12.04 136.81 ± 20.61a Diastolic blood pressure (mmHg) 77.20 ± 8.12 81.96 ± 11.75a Fasting Glucose (mmol/L) 4.84 ± 0.83 5.83 ± 1.77a Triglyceride (mmol/L) 1.40 ± 0.66 2.01 ± 1.67a Total cholesterol (mmol/L) 4.86 ± 0.80 4.85 ± 1.14 High density lipoprotein cholesterol (mmol/L) 1.24 ± 0.23 1.12 ± 0.27a Low density lipoprotein cholesterol (mmol/L) 2.76 ± 0.76 2.83 ± 0.93

Data are reported as means ± s.d. aP < 0.01 vs. controls.

Table 2. The PCR and EXTEND primers of the four studied polymorphisms in LTA gene.

SNPs PCR primer sequence (5′-3′) EXTEND primer sequence (5′-3′)

F: ACGTTGGATGTATAAAGGGACCTGAGCGTC G+10A R: GAGAGCCTCACCTGCTGTG R: ACGTTGGATGTAGTCCAAAGCACGAAGCAC F: ACGTTGGATGATCCAGGCAGCAGGTGCAG A+80C F: CGTGCTTTGGACTACCGCCC R: ACGTTGGATGGTGCTTCGTGCTTTGGACTA F: ACGTTGGATGAGAAGGAGGAGGTGTAGGGT T+496C F: CGTCTCTTCCTCCCAAGGGTG R: ACGTTGGATGCTCTTTCTCTGCAGGTTCTC F: ACGTTGGATGACCCCAAGATGCATCTTGCC C+804A R: GAGCAGCAGGTTTGAGG R: ACGTTGGATGGAGGTCAGGTGGATGTTTAC F: forward primer; R: reverse primer. 122 Y. Liu et al.

Table 3. Comparisons of genotype and allele frequencies of LTA gene polymorphisms between CAD patients and healthy controls. Genotype frequency (%) χ2 p Allele frequency (%) χ2 p G+10A GG GA AA G A controls 103 (31.2) 151 (45.8) 76 (23.0) 357 (54.1) 303 (45.9) CAD patients 100 (22.8) 229 (52.3) 109 (24.9) 6.890 0.032 429 (49.0) 447 (51.0) 3.946 0.047 A+80C AA AC CC A C controls 38 (11.5) 150 (45.5) 142 (43.0) 226 (34.2) 434 (65.8) CAD patients 35 (8.0) 206 (47.4) 194 (44.6) 2.617 0.270 276 (31.7) 594 (68.3) 1.080 0.299 T+496C TT TC CC T C controls 219 (66.4) 98 (29.7) 13 (3.9) 536 (81.2) 124 (18.8) CAD patients 290 (66.7) 133 (30.6) 12 (2.8) 0.851 0.653 713 (82.0) 157 (18.0) 0.138 0.711 C+804A CC CA AA C A controls 90 (27.3) 165 (50.0) 75 (22.7) 345 (52.3) 315 (47.7) CAD patients 98 (22.5) 233 (53.4) 105 (24.1) 2.335 0.311 429 (49.2) 443 (50.8) 1.421 0.233 Data are expressed as the number of cases (percentage of each genotype/allele).

Table 4. Comparison of three genetic modes of inheritance for LTA polymorphisms associated with CAD.

P; OR Additive model Dominant model Recessive model (95% CI) a a a SNP Crude Adjusted Crude Adjusted Crude Adjusted 0.016; 2.240 0.081; 2.869 0.001; 6.554 0.009; 11.352 0.697; 1.163 0.848; 1.130 G+10A (1.163~4.316) (0.877~9.387) (2.129~20.174) (1.843~69.901) (0.544~2.485) (0.323~3.958) 0.502; 1.208 0.671; 1,233 0.358; 1.174 0.521; 1.215 0.095; 0.655 0.535; 0.766 A+80C (0.696~2.095) (0.469~3.245) (0.834~1.651) (0.670~2.202) (0.399~1.076) (0.330~1.780) 0.410; 1.267 0.896; 1.069 0.257; 1.228 0.693; 1.125 0.339; 0.675 0.032; 0.202 T+496C (0.722~2.224) (0.394~2.897) (0.861~1.751) (0.626~2.023) (0.302~1.510) (0.047~0.874) 0.132; 0.632 0.244; 0.537 0.012; 0.240 0.050; 0.162 0.736; 0.877 0.984; 1.013 C+804A (0.348~1.149) (0.189-1.528) (0.078~0.735) (0.026~0.997) (0.408~1.883) (0.290~3.535) a adjusting for gender, age, blood pressure, fasting glucose, triglyceride, total cholesterol, high density lipoprotein cholesterol and low density lipoprotein cholesterol.

Table 5. Frequencies of estimated haplotypes of LTA in Han Chinese between CAD patients and healthy controls.

Frequencyb (%) Haplotypesa Pc; OR (95%CI) Pd; OR (95%CI) Controls Cases A-C-T-A 42.6 49.1 1.00 (Ref.) 1.00 (Ref.) G-A-T-C 29.3 29.9 0.866; 0.889 (0.696~1.135) 0.934; 0.850 (0.648~1.117) G-C-C-C 17.7 16.5 0.699; 0.803 (0.600~1.073) 0.798; 0.783 (0.568~1.079) G-C-T-C 3.4 0.6 0.0003; 0.190 (0.071~0.505) 0.0002; 0.252 (0.094~0.677) aOnly haplotypes (alleles in order of rs1800683, rs2239704, rs2229094 and rs1041981) with a frequency greater than 3% are listed. bData are expressed as the percentage of each haplotype. cCrude and dadjusted P values, OR and 95% CI are calculated using the most common haplotype A-C-T-A as a reference with the latter correcting gender, age, blood pressure, serum levels of fasting glucose, triglyceride, total cholesterol, high density lipoprotein cholesterol and low density lipoprotein cholesterol.

20-30% and haplotype G-C-T-C around 3%, whereas the adjustment (ORcrude = 0.190, 95% CI: 0.071-0.505, P = remaining haplotypes were rare (frequencies less than 3%). 0.0003; ORadjusted = 0.252, 95% CI: 0.094-0.677, P = 0.0002), After dropping nine rare haplotypes (data not shown), we even after the Bonferroni correction. As for study power, further evaluated the remaining four haplotypes. our sample size was adequately powerful (97.5%) to iden- The frequency of haplotype G-C-T-C was significantly tify this significant haplotype difference between CAD lower in CAD patients compared with controls (0.6% vs. patients and controls when type I error probability for a two 3.4%, Table 5). Using the most common haplotype A-C- sided-test (α) was defined as 0.05 and OR was estimated at T-A as a reference, the haplotype G-C-T-C was significantly 1.5. associated with a decreased risk of CAD before and after Lymphotoxin-Alpha Gene SNPs and CAD in Chinese 123

In addition, we found that T allele of T+496C poly- Discussion morphism was protective against CAD under the recessive In the present study, we provide evidence for an asso- mode, which was reinforced by further haplotype analysis. ciation of LTA gene polymorphisms with the risk of CAD Previous studies suggested that haplotype harboring +496C in Han Chinese, and show a protective effect of the haplo- allele as a tagging SNP increased the mobility of CAD, type G-C-T-C on the occurrence of CAD, leading us to however, +496C allele alone was not associated with CAD believe that LTA gene might be a logical candidate that is (Palikhe et al. 2007; 2008). Further in vitro functional involved in the pathogenesis of CAD. To our knowledge, analysis conducted by the same group indicated that this is the first haplotype-based association study demon- T+496C polymorphism caused a change of the 13th amino strating the combined effect of LTA gene polymorphisms acid cysteine to arginine on the signal peptide region of the on susceptibility to CAD. mature LTA protein and +496C allele might reduce the flex- Lymphotoxin-alpha, a member of the tumor necrosis ibility of the signal peptide and affect its binding activity to factor family, is an important pro-inflammatory cytokine in the TNF receptor. Nevertheless, understanding the exact the inflammation process of atherosclerosis. Schreyer et al. molecular mechanisms of T+496C polymorphism remains (2002) demonstrated that LTA was expressed in atheroscle- an open question. rotic lesions in mice and LTA deficiency resulted in a 62% However, it should be noted that although haplotype reduction in the size of the lesions. Later on, Ozaki et al. G-C-T-C was consistently confirmed as a protective factor (2004) revealed that LTA was expressed in smooth muscle against CAD in our study, it also carried +10G allele, which cells (SMCs) and macrophages in human atherosclerotic acted as a risk-conferring allele in our single-locus analysis. plaques, but were absent in quiescent or normal SMCs. There were discrepant findings as to which SNPs were actu- Suna et al. (2008) also confirmed that LTA increased the ally related to or responsible for this protective effect. expression of various genes (such as E-selectin and Furthermore, other studies were unable to identify a signifi- VCAM1) involving in the process of atherosclerosis or cant association of G+10A polymorphism with CAD inflammation in human endothelial cells. These results (Clarke et al. 2006; Sedlacek et al. 2007), which were in together suggested a direct role of LTA in modulating ath- agreement with the present results after multiple correc- erosclerotic progression. Therefore, researchers sought to tions. We hereby reasoned that polymorphism G+10A validate the potential effect of LTA gene on the pathogene- might not be a major locus or its contribution was diluted or sis of CAD. reversed when annexed with other loci, such as the haplo- By means of a large-scale, case-control association types in the present case. Therefore, we agree that replica- study, Ozaki et al. (2002) firstly confirmed a significant tion/validation of those LTA SNPs in larger, well-designed association between LTA gene polymorphisms and myocar- studies within genetic and functional framework is critical. dial infarction in a Japanese population, which was to some Knight et al. (2004) had identified the A allele of degree consistent with our single-locus results supporting A+80C polymorphism as a main predictive variable of LTA the involvement of A allele of C+804A polymorphism in protein production by a detailed haplotype analysis of TNF/ the pathogenesis of CAD. Moreover, another large-scale LTA locus. Contrarily, we did not find a significant single- study documented an independent role of C+804A poly- locus association between CAD and A+80C polymorphism, morphism in the development of coronary atherosclerosis, which was one of the SNPs harboring the protective haplo- showing that +804AA genotype conferred a 1.98-fold type G-C-T-C. In concordance with our results, a recent increased risk for multiple-vessel disease (Laxton et al. meta-analysis provided reliable evidence that the A+80C 2005). Consistently, our study indicated that +804CC gen- polymorphism was not strongly associated with susceptibil- otype yielded a decreased CAD risk under the dominant ity to coronary disease (Clarke et al. 2006). mode (OR = 0.240), which was further strengthened after CAD is a complex disease, which is partly due to adjustment (OR = 0.162). Functionally, Ozaki et al. (2002) genetic determinants. Gene-gene and gene-environment indicated that C+804A polymorphism influenced mRNA interactions are likely to contribute significantly to patho- expression of adhesion molecules and cytokines in the cul- genesis of atherosclerosis and CAD (Yamada et al. 2008; tured human coronary artery SMCs, and +804A allele had a Roberts 2008). In view of the unclear picture of LTA gene two-fold higher level of transcriptional activity for VCAM1 in CAD, we thus speculated that the association of LTA than did +804C allele. In addition, +804A allele had signif- gene polymorphisms with CAD susceptibility might be the icantly high mRNA expression of E-selectin, whereas result of a direct functional effect of polymorphisms under +804C allele did not affect its expression. Asselbergs et al. study or due to their interactions with: (1) other genes or (2007) demonstrated that +804A allele was directly associ- polymorphisms located in the same or neighboring genomic ated with the plasma levels of the inflammatory biomarkers region, such as HLA-A, HLADRB, C4A gene in the major and cell adhesion molecules (i.e. interleukin-6, sTNF-R2 histocompatibility complex (MHC) region (Palikhe et al. and VCAM1). Therefore, our finding of the protective 2007), BAT1-NFKBIL1-LTA region (Koch et al. 2005, effect of +804C allele against CAD risk is consistent with 2007) and TNF/LTA locus (Iida et al. 2003; Ozaki et al. its postulated function. 2004; Edgel et al. 2010); (2) other inflammatory pathways, 124 Y. Liu et al. such as 5-lipoxygenase pathway (Bäck 2009; Cao et al. Bäck, M. (2009) Inhibitors of the 5-lipoxygenase pathway in 2009) and leukotriene pathway (Topol et al. 2006; Crosslin atherosclerosis. Curr. Pharm. Des., 15, 3116-3132. Berliner, J.A., Navab, M., Fogelman, A.M., Frank, J.S., Demer, et al. 2009); and (3) other proatherogenic pathways, such as L.L., Edwards, P.A., Watson, A.D. & Lusis, A.J. (1995) lipid metabolism, blood pressure regulation, insulin resis- Atherosclerosis: basic mechanisms. Oxidation, inflammation, tance and cellular adhesion (Hamid et al. 2005; Topol et al. and genetics. Circulation, 91, 2488-2496. 2006; Rasouli and Kiasari 2008; Suna et al. 2009; Phillips Buckland, P.R., Hoogendoorn, B., Guy, C.A., Coleman, S.L., Smith, S.K., Buxbaum, J.D., Haroutunian, V. & O’Donovan, et al. 2010), which play an important role in the onset and M.C. (2004) A high proportion of polymorphisms in the development of atherosclerosis. promoters of brain expressed genes influences transcriptional Interpretation of our current study, however, should be activity. Biochim. Biophys. Acta, 1690, 238-249. viewed in light of several limitations. A major limitation Cao, R.Y., St Amand, T., Gräbner, R., Habenicht, A.J. & Funk, C.D. (2009) Genetic and pharmacological inhibition of the 5-lipox- was that only four polymorphisms, whose combinations ygenase/leukotriene pathway in atherosclerotic lesion devel- were inherited as one block in the LTA genomic region, opment in ApoE deficient mice. Atherosclerosis, 203, 395- were analyzed. Others susceptible to CAD, such as 400. A+252G polymorphism in intron 1 (Jang et al. 2007), were Clarke, R., Xu, P., Bennett, D., Lewington, S., Zondervan, K., not under investigation. The jury must refrain from draw- Parish, S., Palmer, A., Clark, S., Cardon, L., Peto, R., Lathrop, M. & Collins, R.; International Study of Infarct Survival (ISIS) ing a conclusion until more comprehensive studies of LTA Collaborators. (2006) Lymphotoxin-alpha gene and risk of gene are completed to confirm whether the present associa- myocardial infarction in 6,928 cases and 2,712 controls in the tion is causal or due to linkage disequilibrium and illustrate ISIS case-control study. PLoS. Genet., 2, e107. whether the LTA genetic polymorphisms truly influence the Crosslin, D.R., Shah, S.H., Nelson, S.C., Haynes, C.S., Connelly, J.J., Gadson, S., Goldschmidt-Clermont, P.J., Vance, J.M., susceptibility to CAD or not. Another limitation was that Rose, J., Granger, C.B., Seo, D., Gregory, S.G., Kraus, W.E. & we had only carried out the genotype-phenotype association Hauser, E.R. (2009) Genetic effects in the leukotriene biosyn- study, which did not adequately reveal the complex patho- thesis pathway and association with atherosclerosis. Hum. genic essence of CAD. Further investigations of some Genet., 125, 217-229. Dupont, W.D. & Plummer, W.D. Jr. (1990) Power and sample size intermediate phenotypes such as biomarkers of inflamma- calculations. A review and computer program. Control. Clin. tion (i.e. C-reactive protein, TNF and LTA) would be help- Trials, 11, 116-128. ful to identify the causality between LTA gene and CAD. Edgel, K.A., Leboeuf, R.C. & Oram, J.F. (2010) Tumor necrosis Finally, molecular functional studies are warranted in the factor-alpha and lymphotoxin-alpha increase macrophage ABCA1 by gene expression and protein stabilization via future to elucidate the crucial role of LTA gene in CAD different receptors. Atherosclerosis, 209, 387-392. pathogenesis. Gao, H., Zhang, Z., Zhang, J., Zhao, N., Li, Q. & Bai, M. (2010) In conclusion, our study suggests that LTA gene may Association of LT-alpha Ala252Gly gene polymorphism and be an important risk factor for the occurrence of CAD and the genetic predisposition of coronary heart disease in Chinese. Mol. Biol. Rep., 37, 47-50. the four SNPs may have a combined effect on the risk of Hamid, Y.H., Urhammer, S.A., Glümer, C., Borch-Johnsen, K., CAD in Han Chinese. Our study provides an anchoring Jørgensen, T., Hansen, T. & Pedersen, O. (2005) The common point for better understanding of the pathogenesis of CAD. T60N polymorphism of the lymphotoxin-alpha gene is associ- Nevertheless, for practical reasons, we hope that this study ated with type 2 diabetes and other phenotypes of the metabolic syndrome. Diabetologia, 48, 445-451. will not remain just another endpoint of research instead of Iida, A., Ozaki, K., Ohnishi, Y., Tanaka, T. & Nakamura, Y. (2003) a beginning to establish the background data for further Identification of 46 novel SNPs in the 130-kb region investigation on interaction of LTA gene and CAD in containing a myocardial infarction susceptibility gene on chro- Chinese. mosomal band 6p21. J. Hum. Genet., 48, 476-479. Jang, Y., Kim, H.J., Koh, S.J., Hyun, Y.J., Chae, J.S., Cho, H., Ha, J.W., Kim, B.K., Kim, J.S. & Lee, J.H. (2007) Lymphotoxin- Acknowledgments alpha gene 252A>G and metabolic syndrome features in This work was supported by the National Natural Science Korean men with coronary artery disease. Clin. Chim. Acta, Foundation of China (No. 81070177), Shanghai Science and 384, 124-128. Technology Development Funds (No. 08411963500) and Knight, J.C., Keating, B.J. & Kwiatkowski, D.P. (2004) Allele- Shanghai “Phosphor” Science Foundation, China (No. specific repression of lymphotoxin-alpha by activated B cell 09QB1403100). factor-1. Nat. Genet., 36, 394-399. Koch, W., Hoppmann, P., Michou, E., Jung, V., Pfeufer, A., Müller, J., Meitinger, T., Schömig, A. & Kastrati, A. (2005) TaqMan Conflict of Interest assays for genotyping of single nucleotide polymorphisms The authors declare that they have no competing interest present at a disease susceptibility locus on chromosome 6. for this manuscript. Clin. Chem. Lab. Med., 43, 167-172. Koch, W., Hoppmann, P., Michou, E., Jung, V., Pfeufer, A., References Mueller, J.C., Gieger, C., Wichmann, H.E., Meitinger, T., Schömig, A. & Kastrati, A. (2007) Association of variants in Asselbergs, F.W., Pai, J.K., Rexrode, K.M., Hunter, D.J. & Rimm, the BAT1-NFKBIL1-LTA genomic region with protection E.B. (2007) Effects of lymphotoxin-alpha gene and galectin-2 against myocardial infarction in Europeans. Hum. Mol. gene polymorphisms on inflammatory biomarkers, cellular Genet., 16, 1821-1827. adhesion molecules and risk of coronary heart disease. Clin. Laxton, R., Pearce, E., Kyriakou, T. & Ye, S. (2005) Association Sci. (Lond), 112, 291-298. of the lymphotoxin-alpha gene Thr26Asn polymorphism with Lymphotoxin-Alpha Gene SNPs and CAD in Chinese 125

severity of coronary atherosclerosis. Genes Immun., 6, H., Cohen, Z., Delmonte, T., Kocher, K., Miller, K., 539-541. Guschwan, S., Kulbokas, E.J., O’Leary, S., Winchester, E., Li, W., Xu, J., Wang, X., Chen, J., Zhang, C., Sun, K. & Hui, R. Dewar, K., Green, T., Stone, V., Chow, C., Cohen, A., (2010) Lack of association between lymphotoxin-alpha, Langelier, D., Lapointe, G., Gaudet, D., Faith, J., Branco, N., galectin-2 polymorphisms and coronary artery disease: a meta- Bull, S.B., McLeod, R.S., Griffiths, A.M., Bitton, A., analysis. Atherosclerosis, 208, 433-436. Greenberg, G.R., Lander, E.S., Siminovitch, K.A. & Hudson, Ozaki, K., Inoue, K., Sato, H., Iida, A., Ohnishi, Y., Sekine, A., T.J. (2001) Genetic variation in the 5q31 cytokine gene Sato, H., Odashiro, K., Nobuyoshi, M., Hori, M., Nakamura, cluster confers susceptibility to Crohn disease. Nat. Genet., Y. & Tanaka, T. (2004) Functional variation in LGALS2 29, 223-228. confers risk of myocardial infarction and regulates lympho- Roberts, R. (2008) Genetics of premature myocardial infarction. toxin-alpha secretion in vitro. Nature, 429, 72-75. Curr. Atheroscler. Rep., 10, 186-193. Ozaki, K., Ohnishi, Y., Iida, A., Sekine, A., Yamada, R., Tsunoda, Ross, R. (1999) Atherosclerosis-an inflammatory disease. N. T., Sato, H., Sato, H., Hori, M., Nakamura, Y. & Tanaka, T. Engl. J. Med., 340, 115-126. (2002) Functional SNPs in the lymphotoxin-alpha gene that Ryan, A.W., O’Brien, E., Shields, D. & McManus, R. (2008) Lack are associated with susceptibility to myocardial infarction. of association between NFKBIL1/LTA polymorphisms and Nat. Genet., 32, 650-654. hypertension, myocardial infarct, unstable angina and stable Ozaki, K. & Tanaka, T. (2005) Genome-wide association study to angina in a large Irish population sample. Atherosclerosis, identify SNPs conferring risk of myocardial infarction and 197, 465-466. their functional analyses. Cell. Mol. Life Sci., 62, 1804-1813. Schreyer, S.A., Vick, C.M. & LeBoeuf, R.C. (2002) Loss of Palikhe, A., Lokki, M.L., Pussinen, P.J., Paju, S., Ahlberg, J., lymphotoxin-alpha but not tumor necrosis factor-alpha reduces Asikainen, S., Seppänen, M., Valtonen, V., Nieminen, M.S. & atherosclerosis in mice. J. Biol. Chem., 277, 12364-12368. Sinisalo, J. (2008) Lymphotoxin alpha LTA+496C allele is a Sedlacek, K., Neureuther, K., Mueller, J.C., Stark, K., Fischer, M., risk factor for periodontitis in patients with coronary artery Baessler, A., Reinhard, W., Broeckel, U., Lieb, W., Erdmann, disease. Tissue Antigens, 71, 530-537. J., Schunkert, H., Riegger, G., Illig, T., Meitinger, T. & Palikhe, A., Sinisalo, J., Seppänen, M., Valtonen, V., Nieminen, Hengstenberg, C. (2007) Lymphotoxin-alpha and galectin-2 M.S. & Lokki, M.L. (2007) Human MHC region harbors both SNPs are not associated with myocardial infarction in two susceptibility and protective haplotypes for coronary artery different German populations. J. Mol. Med., 85, 997-1004. disease. Tissue Antigens, 69, 47-55. Suna, S., Sakata, Y., Sato, H., Mizuno, H., Nakatani, D., Shimizu, Park, S., Youn, J.C., Shin, D.J., Park, C.M., Kim, J.S., Ko, Y.G., M., Usami, M., Takashima, S., Takeda, H. & Hori, M. (2008) Choi, D., Ha, J.W., Jang, Y. & Chung, N. (2007) Genetic Up-regulation of cell adhesion molecule genes in human endo- polymorphism in the pregnancy-associated plasma protein-A thelial cells stimulated by lymphotoxin alpha: DNA micro- associated with acute myocardial infarction. Coron. Artery array analysis. J. Atheroscler. Thromb., 15, 160-165. Dis., 18, 417-422. Suna, S., Sakata, Y., Shimizu, M., Nakatani, D., Usami, M., Phillips, C.M., Goumidi, L., Bertrais, S., Ferguson, J.F., Field, Matsumoto, S., Mizuno, H., Ozaki, K., Takashima, S., Takeda, M.R., Kelly, E.D., Mehegan, J., Peloso, G.M., Cupples, L.A., H., Tanaka, T., Hori, M. & Sato, H. (2009) Lymphotoxin- Shen, J., Ordovas, J.M., McManus, R., Hercberg, S., Portugal, alpha3 mediates monocyte-endothelial interaction by TNFR I/ H., Lairon, D., Planells, R. & Roche, H.M. (2010) Additive NF-kappaB signaling. Biochem. Biophys. Res. Commun., 379, effect of polymorphisms in the IL-6, LTA, and TNF-{alpha} 374-378. genes and plasma fatty acid level modulate risk for the meta- Tabor, H.K., Risch, N.J. & Myers, R.M. (2002) Candidate-gene bolic syndrome and its components. J. Clin. Endocrinol. approaches for studying complex genetic traits: practical Metab., 95, 1386-1394. considerations. Nat. Rev. Genet., 3, 391-397. PROCARDIS Consortium. (2004) A trio family study showing Tanaka, T. & Ozaki, K. (2006) Inflammation as a risk factor for association of the lymphotoxin-alpha N26 (804A) allele with myocardial infarction. J. Hum. Genet., 51, 595-604. coronary artery disease. Eur. J. Hum. Genet., 12, 770-774. Topol, E.J., Smith, J., Plow, E.F. & Wang, Q.K. (2006) Genetic Rasouli, M. & Kiasari, A.M. (2008) Interactions of lipoprotein(a) susceptibility to myocardial infarction and coronary artery with diabetes mellitus, apolipoprotein B and cholesterol disease. Hum. Mol. Genet., 15, R117-123. enhance the prognostic values for coronary artery disease. Yamada, Y., Ichihara, S. & Nishida, T. (2008) Molecular genetics Clin. Chem. Lab. Med., 46, 667-673. of myocardial infarction. Genomic Med., 2, 7-22. Rioux, J.D., Daly, M.J., Silverberg, M.S., Lindblad, K., Steinhart,