Association Between a CCL17 Genetic Variant and Risk of Coronary Artery Disease in a Chinese Han Population

Abstract

Background: In the present study we investigated the effects of genetic variations in the C-C motif chemokine ligand 17 (CCL17) gene on serum CCL17 levels and risk of coronary artery disease (CAD).

Methods and Results: A case-control study was conducted to determine causal inferences among CCL17 single-nucleotide polymorphisms (SNPs), serum CCL17 levels, and risk of CAD. Luciferase assays, electrophoretic mobility shift assays (EMSA), and allele-specific quantitative chromatin immunoprecipitation (ChIP) assays were used to assess the function of the SNPs. In all, 947 participants (794 with CAD, 153 without CAD) were included in the study. The T allele in rs223828, located in intron of the CCL17 gene, was associated with increased serum CCL17 levels as well as increased CAD risk. A causal inference test using mediation analysis suggested that rs223828 had a significant indirect casual effect on the increased risk of CAD mediated via serum CCL17 levels. Luciferase assays confirmed that the rs223828T allele enhances CCL17 promoter activity. Protein-DNA binding studies using EMSA and allele-specific quantitative ChIP assays indicated preferential activator protein-1 (AP-1) complex formation and recruitment with the rs223828 T allele compared with the C allele.

Conclusions: We propose that the CCL17 SNP rs223828 is associated with increased risk of CAD, and that this site may be a potential AP-1 binding site.

It is well-established that atherosclerosis is an inflammatory process, and there is increasing evidence from experimental studies of the role of inflammation, and the underlying cellular and molecular mechanisms, in atherogenesis.¹ The immigration and infiltration of activated macrophages and T cells into atherosclerotic lesions guided by chemokines is one of the major processes in atherosclerotic disease.² Chemokines are considered to be of great importance for maintaining leukocyte homeostasis in cell mobilization, differentiation, influx, proliferation, and apoptosis or survival to contribute to atherogenesis and neointima formation rather than solely effecting leukocyte recruitment.²

Chemokine CC-motif ligand 17 (CCL17), also known as the thymus and activation-regulated chemokine (TARC), is an important regulator of atherosclerosis that has been shown to drive atherosclerosis in an animal model by restraining regulatory T cell homeostasis.³ In previous studies, we found a link between serum CCL17 concentrations and coronary artery disease (CAD).^4,5 In 971 consecutive patients who underwent coronary angiography, serum CCL17 levels remained associated with CAD even after adjusting for other traditional cardiovascular risk factors.⁵ Moreover, serum CCL17 concentrations were correlated with the subtype and severity of CAD.⁵

In order to determine the causal inference between CCL17 and CAD, we searched for and genotyped single-nucleotide polymorphisms (SNPs) in CCL17 among patients in our previous study.⁵ We then used a mediation model to explore the effects of these genetic variants on CAD risk and whether their effects are causally mediated through serum CCL17. Furthermore, we aimed to determine the effect and potential mechanisms underlying the effects of the SNPs on CCL17 expression in vitro.

Methods

Study Population

Patients visiting our hospital between January and December 2013 for coronary angiography were consecutively recruited to the study (Table 1). The inclusion and exclusion criteria for this study population have been described previously.⁵ Briefly, the exclusion criteria included dermatitis, current infection, malignancy, autoimmune disease, vasculitis, asthma, cirrhosis, severe renal failure (estimated glomerular filtration rate (eGFR) <30 mL/min/1.73 m²) and shock.

Table 1. Baseline Characteristics and Serum CCL17 Concentrations in Patients With or Without CAD

	Non-CAD (n=153)	CAD (n=794)	P value
Age (years)	59.6±10.4	62.2±10.3	0.005
No. males (%)	74 (48.4)	574 (72.3)	<0.001
BMI (kg/m2)	25.7±3.9	25.5±3.6	0.571
Hypertension	84 (54.9)	575 (72.4)	<0.001
Diabetes	48 (31.4)	330 (41.6)	0.019
Current smoker	55 (35.9)	449 (56.5)	<0.001
Family history of CAD	33 (21.6)	165 (20.8)	0.828
TC (mmol/L)	4.24±0.97	4.18±1.57	0.656
TG (mmol/L)	1.31 [0.92–1.96]	1.39 [1.02–2.01]	0.069
LDL-C (mmol/L)	2.39±0.91	2.31±0.88	0.267
HDL-C (mmol/L)	1.27±0.58	1.15±0.57	0.017
Serum CCL17 (pg/mL)	218.8 [144.5–293.9]	254.9 [165.20–363.0]	<0.001

Data are presented as the mean±SD, n (%) or as the median [interquartile range]. BMI, body mass index; CAD, coronary artery disease; CCL17, C-C motif chemokine ligand 17; HDL-C, high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol; TC, total cholesterol; TG, triglycerides.

Subjects with ≥50% stenosis in 1 or more of their main coronary arteries were defined as having CAD. Those with <50% stenosis in any main coronary artery were considered as non-CAD patients. Body mass index (BMI) was calculated as weight (kg) divided by height (m) squared. Hypertension and diabetes were diagnosed on the basis of current guidelines.^6,7 Peripheral venous blood drawn was from the antecubital vein after a 12-h fast and was used to determine total cholesterol (TC), low-density lipoprotein cholesterol (LDL-C), high-density lipoprotein cholesterol (HDL-C), and triglyceride (TG) concentrations.

Determination of Serum CCL17 Concentrations

CCL17 concentrations were determined from blood samples collected from the radial or femoral artery after insertion of the sheath for coronary angiography. Serum CCL17 concentrations were measured using a commercially available ELISA for human CCL17/TARC (R&D Systems, Minneapolis, MN, USA) according to the manufacturer’s instructions. Absorbance was assessed using a Labsystems Multiskan MS spectrophotometer (Thermo Labsystems, Helsinki, Finland) and calculations were performed using Ascent software v2.6 (Thermo Labsystems). The lower limit of detection of the assay was 7pg/mL CCL17. The intra- and interassay coefficients of variation (CV) were approximately 2.7% and 8.2%, respectively.⁵

Tag SNP Selection

Taking into consideration the fact that the study population was of Chinese ancestry, genotype data from the HCB panel (Han Chinese in Beijing, China) of the Phase I, II, and III HapMap Project were used to determine patterns of linkage disequilibrium (LD) and select tag SNPs in CCL17. Tag SNPs capturing common genetic variation in CCL17 were chosen using HAPLOVIEW (v.4.2) software. The criterion for tag SNPs selection was that every SNP in the HCB panel of the HapMap project Phases I, II, and III with 5% allele frequency was captured with a pairwise r²>0.8 by at least 1 tag SNP. Five tag SNPs (rs223895, rs4784805, rs9302690, rs223899, and rs223828) were selected. Characteristics of these SNPs are given in Table 2 and their positions on the structure of the CCL17 gene are shown in Figure S1.

Table 2. Characteristics of Selected Single Nucleotide Polymorphisms Found in the CAD Population

SNP	Position	Minor allele	Major allele	Function	MAF1	MAF2
rs223895	Chr.16:57440896	T	C	Intron	0.49	0.49
rs4784805	Chr.16:57441407	A	C	Intron	0.07	0.10
rs9302690	Chr.16:57442968	A	G	Intron	0.01	0.02
rs223899	Chr.16:57444576	T	G	Intron	0.42	0.47
rs223828	Chr.16:57447414	T	C	Intron	0.34	0.39

CAD, coronary artery disease; Chr., chromosome; MAF1, minor allele frequency in the Chinese Han population; MAF2, minor allele frequency in the present study population; SNP, single nucleotide polymorphism.

SNPs Genotyping

Genomic DNA was extracted from peripheral blood leukocytes using a Tiangen (Beijing, China) Genomic DNA Purification Kit, according to the manufacturer’s instructions. Genotyping was performed using TaqMan on an Applied Biosystems real-time Prism 7900HT Sequence Detection System (ABI, Foster City, CA, USA). Primers and the TaqMan fluorogenic probes bearing a suitable reporter dye on the 5'-end and a quencher dye on the 3'-end were designed using Primer Express software V2.0 (ABI) and procured from Applied Biosystems (Warrington, UK). One probe (for Allele 1) was labeled with the VIC dye and the other (for Allele 2), with a FAM (fluorescein) dye at the 5'-end, and serial dilutions were run to determine the optimal working concentrations. For each reaction, a 25-µL mixture was prepared by mixing 5 µL containing 50 ng DNA, 12.5 µL of 2× Universal mix (Eurogentec, Seraing, Belgium), 1.25 µL of 20× probe assay mix, and 6.25 µL DNase-free distilled water. Three no-template controls were included in each plate to normalize the emission signal. The thermal profile for amplification for the first cycle was as follows: 50℃ for 2 min and 95℃ for 10 min, followed by 50 cycles of 94℃ for 15 s and 60℃ for 30 s. The plates were then scanned for a fluorescence resonance energy transfer (FRET) signal using a 7900HT sequence detection system. Data were analyzed using sodium dodecyl sulfate (SDS) software version 2.0 (ABI).

Luciferase Assay

Genomic DNA samples from 2 patients with a C/C or T/T genotype at rs223828 were amplified using the forward primer 5'-GGCGCGGGTACCAAATAGCGGCTGTTGCAGTTATC-3' (containing the KpnI restriction site, denoted by the underline) and the reverse primer 5'-AATATAaagcttGGTGCCAGGAGCCCAGGAGGGAGTCTCTGTGTGCAG-3' (containing the HindIII restriction site, denoted by the underline), resulting in a 1,432-bp polymerase chain reaction (PCR) product containing the different polymorphisms. Two different amplicons corresponding to the rs223828 variant were cloned into a pGM-T vector (Tiangen) and sequence fidelity was verified by DNA sequencing. Promoter inserts were excised by restriction digestion using KpnI and HindIII, and the resulting fragments were cloned into the pGL3-Basic vector (Promega, Mannheim, Germany) containing a luciferase reporter gene. To analyze the effects of SNPs on gene expression, a dual-luciferase reporter assay (Promega) was performed according to the manufacturer’s instructions. Briefly, cell lysates were obtained 48 h after Lipofectamine 3000 (Invitrogen, Carlsbad, CA, USA) transfection of the pGL3 vector constructs into the THP-1 cell line, using a passive lysis buffer. Then, 100 µL luciferase assay reagent was added to 20 µL lysate and samples were incubated with Stop&Glo reagent, after which luciferase activity was determined using a luminometer (LB952; Berthold, Bad Wildbad, Germany). An empty pGL3 vector was used as a negative control. The activity of cotransfected Renilla luciferase was used to normalize values (firefly luciferase values were divided by Renilla luciferase values).

Electrophoretic Mobility Shift Assay

Nuclear extracts were prepared from harvested THP-1 cells using NE-PER Nuclear and Cytoplasmic Extraction Reagents. THP-1 cells were harvested using ice-cold phosphate-buffered saline (PBS). After centrifugation, the pellet was resuspended in ice-cold sodium dodecyl sulfate (SDS) sample buffer supplemented with 100 mmol/L dithiothreitol. Cytoplasmic and nuclear extracts were isolated with NE-PER Nuclear and Cytoplasmic Extraction Reagents containing 1× Halt Protease Inhibitor Cocktail according to the manufacturer’s instructions. Protein concentrations were determined using a BCA protein assay kit (Bio-Rad Laboratories, Berkeley, CA, USA). The DNA oligonucleotide probes used in the electrophoretic mobility shift assays (EMSA) were as follows: rs223828-C, 5'-CTGCCTGGTCGACTCCCCGACACTTC-3'; rs223828-T, 5'-CTGCCTGGTTGACTCCCCGACACTTC-3'. The underlined bases indicate the position of the SNP. For each oligonucleotide, a complementary oligonucleotide was used. The oligonucleotides were labeled with biotin using the Biotin Light Chemiluminescent EMSA kit (Exprogen Biotechnology, Beijing, China). Then, 10 µg purified protein (in 3.2 µL) was incubated with the probe at 30℃ for 20 min in a 15-µL binding reaction containing 1.5 µL of 10× binding buffer, 1.5 µg poly(deoxyinosinic-deoxycytidylic) acid (poly(dI-dC), and 500 fmol biotin-labeled probe. To demonstrate the specificity of probe-protein binding, competitor experiments with an 100-fold excess of unlabeled probe as a specific competitor were performed. For super-shift assays, 4 µg specific antibodies against c-Jun or c-Fos, or non-immune IgG, was added to the binding reaction and the reaction mixture was incubated on ice for 10 min before labeled probe was added at 120 V in a 1% agarose gel in 0.5× Tris-borate-EDTA for 1.5 h. Then, the gel was then electrophoretically transferred to a nylon membrane at 380 mA for 60 min, and cross-linked DNA was transferred to the membrane using an ultraviolet light cross-linker (UVP, Upland, CA, USA). After cross-linking, biotin-labeled DNA was detected by chemiluminescence and exposed to X-ray film for 5–10 min.

Allele-Specific Quantitative Chromatin Immunoprecipitation (ChIP) PCR

Allele-specific quantitative chromatin ChIP PCR was performed as described previousl.^8,9 Briefly, human peripheral blood mononuclear cells (PBMCs) were isolated by Ficoll density gradient separation. PBMCs were washed with PBS and incubated for 10 min with 1% (v/v) formaldehyde. Cross-linking reactions were terminated by incubating the cells with 0.1 mol/L glycine for 5 min. Cells were then washed twice with PBS and lysed for 1 h at 4℃ in a lysis buffer. Cell lysates were sonicated to obtain chromatin fragments with an average length of 500–800 bp, as assessed by agarose gel electrophoresis. The samples were precleared with protein A-agarose (Roche) for 1 h at 4℃ on a rocking platform, after which 10 μg specific antibody against c-Jun (sc-1694; Santa Cruz Biotechnology, Santa Cruz, CA, USA) was added and samples were incubated overnight at 4℃ with shaking. The immunoprecipitates were captured with 10% (v/v) protein A-agarose for 4 h. Protein A-agarose was blocked overnight at 4℃ with 1 µg/mL bovine serum albumin (BSA) and 1 µg/mL salmon sperm DNA that had been sheared to 500-bp fragments. Quantitative (q) PCR was performed by the SYBR green method using allele-specific primers for rs4648889. The primers for the T and C alleles were 5'-CTGCTTGGTTTCTGCCTGGTT-3' and 5'-CTGCTTGGTTTCTGCCTGGTC-3' respectively, and the common reverse primer was 5'-ACGCACAGCTGGGTTCATCTC-3'.

Statistical Analysis

Baseline characteristics, including demographics and baseline measurements, are given as the mean±SD for normally distributed continuous variables and as the median with interquartile range (IQR) for continuous variables with a skewed distributed. Continuous variables between 2 groups were analyzed by Student’s t-test or the Mann-Whitney U-test depending on data distribution. Categorical variables are summarized as counts and percentages and were compared using the χ² test. Analysis of variance (ANOVA; for SNP with 3 genotypes) or Student’s t-test (for SNP with 2 genotypes) were used for comparisons of serum CCL17 concentrations among different genotypes for 5 SNPs.

All SNPs were encoded using an additive genetic model (0, wild-type; 1, heterozygosity; 2, homozygosity; minor allele as effect allele). The analysis in the present study included 3 steps: (1) association between SNPs and risk of CAD; (2) association between SNPs and serum CCL17 concentrations; and (3) causal inference analysis. The causal inference analysis was conducted only if the previous 2 analyses reached statistical significance. All models were adjusted for covariates of age, gender, BMI, hypertension, diabetes, lipid profile, smoking status, and family history of CAD. Linear regression analysis was used to investigate associations between the 5 SNPs and serum CCL17 concentrations. The results are reported as β coefficients and 95% confidence intervals (CI). The association between SNPs and CAD risk was evaluated by logistic regression analysis and is described by odds ratios (OR) and 95% CI. In causal inference analyses, Vanderweele’s mediation model was used to assess the indirect causal effect of an SNP that was mediated through serum CCL17, which has been used in previous studies.^10,11 The results are described by indirect OR (OR_Indirect) and 95% CI, calculated as follows:

where θ_CCL17 and θ_interaction were estimated from Model 2 and β_SNP was estimated from Model 3. Bootstrap analysis with 1,000 replicates was used to estimate the 95% CI of OR_Indirect and statistical significance as follows:

where θ_SNP, θ_CCL17, and θ_interaction represent the coefficients of SNP, serum CCL17, and the SNP-CCL17 interaction, respectively, and B is a vector of coefficients of covariates. β_SNP was calculated using the following equation:

In Model 3, the weights (w) were calculated for cases or controls as follows:

where the prevalence represents the prevalence of CAD among the population that underwent coronary angiography and r represents the proportion of CAD cases in the analytical dataset.

Statistical significance was set at P<0.05. Bonferroni correction was used to adjust the P value for multiple comparisons. All analyses for the clinical studies were performed using SAS v.9.3.

For in vitro studies, 1-way ANOVA and 2-tailed Student’s t-test were used to determine statistical significance using GraphPad Prism v.5.03 (GraphPad Software, San Diego, CA, USA).

The sample size was calculated using Power and Sample Size Program 11.0 (NCSS, Kaysville, UT, USA.), assuming an α value of 0.05, a 2-sided test, and minor allele frequencies (MAF) for the Chinese population from the HapMap database. For the association between rs223828 and CAD, a sample size of 947 patients (794 in the CAD group, 153 in the non-CAD group) was calculated to achieve 90% power to detect an OR of 1.8 using an MAF of 0.34.

Ethics Approval

This study was approved by the Human Research Ethics Committee of Peking Union Medical College Hospital (Beijing, China), and was performed strictly in accordance with the approved guidelines. The study protocol conforms to the ethical guidelines of the 1975 Declaration of Helsinki. Informed consent was obtained from all subjects.

Results

In all, 971 patients who fulfilled the inclusion and exclusion criteria were selected for the present study. Of these patients, DNA samples were not available for 24 (2.5%). Thus, 947 patients (794 with CAD, 153 without CAD) were included in the study. The baseline characteristics of the CAD and non-CAD groups are given in Table 1. All study participants were from the Chinese Han population. Most of the patients in the CAD group were male and older than those in the non-CAD group. The proportion patients with hypertension, diabetes, and current smoker status was higher in the CAD group, whereas HDL-C concentrations were lower in the CAD group. In addition, serum CCL17 concentrations were significantly higher in the CAD than non-CAD group (Table 1).

Association of CCL17 SNPs With Risk of CAD

Five SNPs, namely rs223895, rs4784805, rs9302690, rs223899, and rs223828, were identified and included in the study. The minor alleles in each SNP were defined as the effect alleles, and the allele–frequency distribution in this study population was similar to that in general Han Chinese population (Table 2). Results from the Hardy-Weinberg equilibrium test revealed no significant differences between the present study population and the general Han Chinese population (P>0.05). Of the 5 SNPs, rs223899 and rs223828 were associated with a significantly increased risk of CAD, even after adjusting for other confounding factors (for rs223899, adjusted [a] OR per effect allele=1.591, 95% CI 1.209–2.094, P=0.001, adjusted P=0.005; for rs223828, aOR per effect allele=1.644, 95% CI 1.233–2.191, P=0.001, adjusted P=0.005). No association was found between CAD and the other 3 alleles (i.e., rs223895, rs4784805, and rs9302690; Table 3).

Table 3. Association Between CCL17 Single Nucleotide Polymorphisms and Risk of CAD

SNP / Genotype	Non-CAD group	CAD group	Unadjusted OR (95% CI) per effect allele	P value	Adjusted OR (95% CI) per effect allele	P value	Adjusted P valueA
rs223895
CC	49 (32.0)	200 (25.2)	1.10 (0.861–1.406)	0.445	1.095 (0.841–1.424)	0.501	1.000
CT	65 (42.5)	408 (51.4)
TT	39 (25.5)	186 (23.4)
rs4784805
CC	127 (83.0)	636 (80.1)	1.269 (0.821–1.959)	0.284	1.318 (0.826–2.105)	0.247	1.000
AC	26 (17.0)	149 (18.8)
AA	0 (0)	9 (1.1)
rs9302690
GG	149 (97.4)	755 (95.1)	1.924 (0.677–5.465)	0.219	1.789 (0.607–5.275)	0.291	1.000
AG	4 (2.6)	39 (4.9)
AA	0 (0)	0 (0)
rs223899
GG	62 (40.5)	201 (25.3)	1.569 (1.215–2.025)	0.001	1.591 (1.209–2.094)	0.001	0.005
GT	66 (43.1)	413 (52.0)
TT	25 (16.3)	180 (22.7)
rs223828
CC	73 (47.7)	271 (34.1)	1.620 (1.239–2.117)	<0.001	1.644 (1.233–2.191)	0.001	0.005
CT	67 (43.8)	392 (49.3)
TT	13 (8.5)	131 (16.5)

Unless indicated otherwise, data show n (%). ^AAdjusted by Bonferroni correction. CI, confidence interval; OR, odds ratio. Other abbreviations as in Tables 1,2.

These results were confirmed by introducing another age- and sex-matched control group (2 : 1) from a community in Shunyi district, Beijing. The characteristics of the control population are presented in Table S1. After adjusting for other confounding factors, rs223899 and rs223828 were still associated with the risk of CAD (for rs223899, aOR per effect allele=1.317, 95% CI 1.076–1.612, P=0.007, adjusted P=0.035; for rs223828, aOR per effect allele=1.445, 95% CI 1.109–1.883, P=0.006, adjusted P=0.030; Table S2).

LD Analysis and Haplotype Association Study

Pairwise D' and r² values are shown in Figure S1. Haplotype blocks were constructed using the CI method described by Gabriel et al¹² and Zhu et al¹³ No haplotype block was detected using the CI method of Gabriel et al¹² Conversely, when the block was defined by a solid spine of LD D' >0.8,¹³ a block consisting of 2 SNPs (rs223899 and rs223828) was detected. A haplotype association study was also performed using HAPLOVIEW software based on the method of Zhu et al¹³ (Table S3), which revealed that the G-C haplotype was linked to a reduced risk of CAD (OR 0.658, 95% CI 0.513–0.844, adjusted P=0.010). In contrast, the T-T haplotype was associated with an increased risk of CAD (OR 1.661, 95% CI, 1.269–2.176, adjusted P=0.004).

Association of CCL17 SNPs With Serum CCL17 Concentrations

Serum CCL17 concentrations were compared among subjects with different genotypes. Across the entire study population, serum CCL17 concentrations were significantly different among the different genotypes for the rs223899, rs223895, and rs223828 SNPs (Figure 1). Using the multiple linear analysis model, only the effect allele T of rs223828 was associated with increased serum CCL17 concentrations (Table 4).

Figure 1.

Effect of selected C-C motif chemokine ligand 17 (CCL17) single nucleotide polymorphisms (SNPs) on circulating CCL17 levels in the study population. Serum CCL17 levels were assessed in patients who were homozygous for major alleles (light gray bars), heterozygous for major alleles (dark gray bars), and homozygous for minor alleles (black bars) of selected SNPs found in the study population. Data are the mean±SD. CAD, coronary artery disease.

Table 4. Associations Between CCL17 Single Nucleotide Polymorphisms and Serum CCL17 Concentrations

SNP	B for effect allele	95% CI for B	P value
rs223895	14.742	−1.512, 30.996	0.075
rs4784805	−13.577	−40.705, 13.551	0.326
rs9302690	4.704	−50.528, 59.936	0.867
rs223899	12.020	−4.485, 28.526	0.153
rs223828	18.921	2.164, 35.678	0.027

The linear regression model was adjusted for age, gender, BMI, hypertension, diabetes, lipid profile, smoking status, and family history of CAD. q value, false discovery rate (FDR)-adjusted P value. Other abbreviations as in Tables 1–3.

Serum CCL17, Increased CAD Risk, and CCL17 (rs223828) SNP

Because rs223828 was independently associated with both increased CAD risk and serum CCL17 concentrations, we hypothesized that serum CCL17 concentrations may be affected by the genetic effects of rs223828 and be an important causal mediator of CAD. To verify this hypothesis, a causal inference test using mediation analysis was conducted. Based on data in the present study, the prevalence of CAD among the at-risk population was found to be approximately 80%, which was used to calculate the weights for mediation analysis. The results revealed that the CCL17 SNP rs223828 had a significant indirect casual effect on the increased risk of CAD mediated via serum CCL17 concentrations (OR_Indirect=1.077, 95% CI 1.015–1.161, P=0.047).

Effects of CCL17 SNP rs223828 on CCL17 Promoter Activity

Luciferase reporter assays were used to assess the effects of different rs223828 genotypes on promoter activity. Luciferase reporter assays were performed in human THP-1 cells transfected with pGL3-basic promoterless plasmid and the 1.4-kb PRE sequence with either the effect allele T or the major allele C. The effect allele T significantly increased luciferase activity compared with the major allele C (P<0.001; Figure 2A).

Figure 2.

Effect of the CCL17 SNPs rs223828 on CCL17 promoter activity and transcription factor binding affinity. (A) Transcriptional activity of the CCL17 major allele rs223828 (C allele) and the CCL17 minor allele rs223828 (T allele) was assessed in THP-1 cells using the luciferase assay (n=6). Relative luciferase activity is expressed as the mean±SEM. The relative luciferase activity of the minor T allele was significantly higher than that of the major C allele. (B) The rs223828 SNP is located in the activator protein-1 (AP-1) DNA-binding motif. Data from the JASPAR database. (C) Prediction of the affinity of AP-1 binding to the T and C alleles of rs223828 according to JASPAR score. Abbreviations as in Figure 1.

Potential Activator Protein-1 Binding Site on CCL17 SNP rs223828

Bioinformatics analysis using Transfac software and the JASPAR database indicated that rs223828 was located at the binding site of the transcription factor activator protein (AP)-1 (containing fos/jun subunits; Figure 2B). Furthermore, prediction programs revealed that the binding affinity of AP-1 for the effect allele T of rs223828 was 2-fold higher than that of the allele C (Figure 2C). EMSA was performed using THP-1 cell nuclear protein extracts and fluorescently labeled rs223828 C and T allele probes. The results showed that THP-1 nuclear protein extracts bound more highly to the effect allele T probe than the allele C probe (Figure 3, Lanes 3 and 4). As shown in Figure 3, the specificity of the assay was also verified by the addition of 100-fold unlabeled competitors to alleles C and T (Lanes 5 and 6). Quantitative competition assays were performed by adding 10-, 20-, 50-, and 100-fold molar excess of unlabeled competitor allele T probe (Figure S2, Lanes 5–8) to labeled allele C probe mixture or 10-, 20-, 50-, and 100-fold molar excess of unlabeled competitor allele C probe (Figure S2, Lanes 9–12) to labeled allele T probe mixture. The band of allele C probe and THP-1 nuclear protein extracts became weaker with increasing amounts of competitor allele T probe and almost vanished when 100-fold competitor probe was added (Figure S2, Lanes 5–8), whereas excess unlabeled allele C probe did not compete with the labeled allele T probe (Figure S2, Lanes 9–12). The results of the quantitative competition assay support our interpretation that THP-1 nuclear protein extracts bound more highly to the effect allele T probe than the allele C probe. By adding anti-AP-1 antibodies (c-jun and c-fos; Figure 3, Lanes 8–11), we also showed the presence of a higher mobility complex (super shift) associated with the rs223828 C allele (due to association with c-fos or c-jun antibodies) that competes with the endogenous complex (rs223828 C allele and THP-1 nuclear extract). Furthermore, ChIP with allele-specific qPCR was used to determine the DNA-binding capacity of c-fos and c-jun in peripheral mononuclear cells freshly isolated from 6 patients harboring heterozygous CCL17 (rs223828) SNP and CAD. The results indicated that both c-jun and c-fos were preferentially recruited to the effect T allele compared with the C allele (Figure 4).

Figure 3.

Effects of CCL17 SNPs rs223828 on protein-DNA complex formation and involvement of AP-1. An electrophoretic mobility gel shift assay (EMSA) was performed using biotin-labeled oligonucleotides corresponding to major (C) and minor (T) CCL17 rs223828 alleles, followed by incubation with nuclear protein extracts (NE) from THP-1 cells. Lanes 1, 2, biotin-labeled oligonucleotides in the absence of THP-1 NE; Lanes 3, 4, a specific binding event (Band) is seen after the addition of THP-1 NE; Lanes 5, 6, disappearance of the binding event after the addition of 100-fold excess of unlabeled competitor oligonucleotides; Lanes 8–11, a super-shifted band (super shift band) is seen in the presence of AP-1 antibodies (c-jun and c-fos Ab) only in the minor T allele lanes. Abbreviations as in Figures 1,2.

Figure 4.

Allele-specific binding of c-jun and c-fos at the CCL17 SNPs rs223828 locus. Chromatin immunoprecipitation (ChIP) was performed with c-jun and c-fos antibodies to investigate changes in binding at the rs223828 (heterozygous) locus in peripheral mononuclear cells from 6 patients with CAD. IgG antibodies were used as a negative control. Note, both c-jun and c-fos were preferentially recruited to the effect minor T allele compared with the major C allele. Relative enrichment is expressed as the mean±SEM. P<0.05 was considered significant. Abbreviations as in Figure 1.

Discussion

In the present study, we demonstrated that the T allele of the CCL17 SNP rs223828 is associated with increased risk of CAD as well as higher serum CCL17 concentrations. Causal inference analysis further showed serum CCL17 concentrations mediate the increased CAD risk, which is associated with the T allele of the CCL17 (rs223828) gene. Mechanistic studies highlighted that the T allele of the CCL17 (rs223828) gene: (1) had a relatively higher promoter activity than the C allele; and (2) preferentially binds to AP-1 and thus has a potential AP-1-binding site, which, together, may provide an explanation as to how the rs223828 SNP affects the risk of CAD.

Increased circulating CCL17 concentrations have been found in several allergy-related diseases, such as atopic dermatitis.^14,15 Recently, Lee et al found that CCL17 gene polymorphisms and their expression were associated with susceptibility and the formation of coronary artery aneurysms in Kawasaki disease.¹⁶ Several previous studies have reported that the CCL17 SNP rs223828 (previously known as −431C>T) is associated with increased CCL17 expression. For example Sekiya et al found that subjects carrying the −431T allele had significantly higher serum CCL17 concentrations than those without this allele, indicating that the −431T allele may have a major effect on CCL17 production.¹⁷ Similarly, another study showed that serum CCL17 concentrations in patients with asthma were significantly associated with the −431C/T allele.¹⁸ Moreover, −431C>T has been shown to enhance CCL17 promoter activity in vitro.¹⁹ Together, these findings indicate that rs223828 is a functional SNP, which is consistent with the findings of the present study.

Several animal studies have proven atherogenic effects of CCL17. For example, CCL17 deficiency has been found to significantly inhibit the formation of atherosclerotic lesions in ApoE^−/− mice.³ Furthermore, CCL17 expression in dendritic cells was shown to limit the expansion of regulatory T cells by restricting their maintenance, and thus precipitated T cell-mediated atherosclerosis.³ In addition, platelets may provide another explanation for the effects of CCL17 on CAD. For example, platelets have been reported to be a major source of CCL17.²⁰ During the clotting process, activated platelets release high concentrations of CCL17 in an autocrine manner, which could further enhance platelet aggregation.^20,21

AP-1 (Fos/Jun) is a transcriptional regulator comprising members of the Fos and Jun families of DNA-binding proteins, which play a vital role in the inflammatory process.²² It has been reported that AP-1 is increased in human carotid plaques and is associated with cholesteryl esters, indicating that AP-1 could be a marker of plaque vulnerability.²³ Upregulation of AP-1 was also associated with increased age-related vascular smooth muscle cell proliferation in a rabbit model.²⁴ Conversely, Jun proteins need to be activated by c-Jun N-terminal kinases (JNKs), and its inhibition protects against endothelial dysfunction and oxidative stress,²⁵ as well as against low shear stress-induced atherogenesis.²⁶ The present study established a potential link between AP-1 and the CCL17 pathway, which may be a key component of the inflammatory network of atherosclerosis.

The present study has certain limitations. Causal inference was verified by a case-control study; however, a prospective cohort study is needed to confirm the findings of this study. Although our in vitro studies evaluated the potential mechanism of the SNP, additional investigations are warranted to fully elucidate the mechanisms that underlie the association between the rs223828 SNP and CAD risk.

In summary, we conclude that the CCL17 SNP rs223828 is associated with both elevated serum CCL17 concentrations and increased risk of CAD. Results from our in vitro studies indicate that rs223828 directly affects CCL17 promoter activity and may be a potential AP-1 transcription factor binding site.

Acknowledgments

This study was supported by a Peking Union Medical College Hospital Young Investigator Grant (PUMCH-2013-015, PUMCH-2016-1.10), the National Natural Science Foundation of China (81441013), and the CAMS Major Collaborative Innovation Project (No. 2016-I2M-1-011)

Conflict of Interest

The authors declare no conflicts of interest.

Supplementary Files

Supplementary File 1

Figure S1. Positions of single nucleotide polymorphisms (SNPs) on the structure of the C-C motif chemokine ligand 17 (CCL17) gene and linkage disequilibrium (LD) among the SNPs in the tested population including both (A) D’ and (B,C) r^2 data.

Figure S2. Results of electrophoretic mobility gel shift assays.

Table S1. Baseline characteristics of the CAD group and additional community control population from Shunyi District, Beijing

Table S2. Association between CCL17 SNPs and risk of CAD

Table S3. Haplotype association study of CCL17

Please find supplementary file(s);

http://dx.doi.org/10.1253/circj.CJ-17-0190

References

• 1.   Libby P, Ridker PM, Maseri A. Inflammation and atherosclerosis. Circulation 2002; 105: 1135–1143.
• 2.   Zernecke A, Shagdarsuren E, Weber C. Chemokines in atherosclerosis: An update. Arterioscler Thromb Vasc Biol 2008; 28: 1897–1908.
• 3.   Weber C, Meiler S, Doring Y, Koch M, Drechsler M, Megens RT, et al. CCL17-expressing dendritic cells drive atherosclerosis by restraining regulatory T cell homeostasis in mice. J Clin Invest 2011; 121: 2898–2910.
• 4.   Ye Y, Zhao X, Chen L, Zhang T, Zhang S. Association between serum chemokine CC-motif ligand 17 and coronary artery disease. Int J Cardiol 2014; 176: 541–542.
• 5.   Ye Y, Yang X, Zhao X, Chen L, Xie H, Zeng Y, et al. Serum chemokine CCL17/thymus activation and regulated chemokine is correlated with coronary artery diseases. Atherosclerosis 2015; 238: 365–369.
• 6.   Mancia G, De Backer G, Dominiczak A, Cifkova R, Fagard R, Germano G, et al. 2007 Guidelines for the management of arterial hypertension: The Task Force for the Management of Arterial Hypertension of the European Society of Hypertension (ESH) and of the European Society of Cardiology (ESC). Eur Heart J 2007; 28: 1462–1536.
• 7.   American Diabetes Association. Diagnosis and classification of diabetes mellitus. Diabetes Care 2014; 37(Suppl 1): S81–S90.
• 8.   Vecellio M, Roberts AR, Cohen CJ, Cortes A, Knight JC, Bowness P, et al. The genetic association of RUNX3 with ankylosing spondylitis can be explained by allele-specific effects on IRF4 recruitment that alter gene expression. Ann Rheum Dis 2016; 75: 1534–1540.
• 9.   Cowper-Sallari R, Zhang X, Wright JB, Bailey SD, Cole MD, Eeckhoute J, et al. Breast cancer risk-associated SNPs modulate the affinity of chromatin for FOXA1 and alter gene expression. Nat Genet 2012; 44: 1191–1198.
• 10.   VanderWeele TJ, Asomaning K, Tchetgen EJ, Han Y, Spitz MR, Shete S, et al. Genetic variants on 15q25.1, smoking, and lung cancer: An assessment of mediation and interaction. Am J Epidemiol 2012; 175: 1013–1020.
• 11.   Wei Y, Wang Z, Su L, Chen F, Tejera P, Bajwa EK, et al. Platelet count mediates the contribution of a genetic variant in LRRC16A to ARDS risk. Chest 2015; 147: 607–617.
• 12.   Gabriel SB, Schaffner SF, Nguyen H, Moore JM, Roy J, Blumenstiel B, et al. The structure of haplotype blocks in the human genome. Science 2002; 296: 2225–2229.
• 13.   Zhu X, Yan D, Cooper RS, Luke A, Ikeda MA, Chang YP, et al. Linkage disequilibrium and haplotype diversity in the genes of the renin-angiotensin system: Findings from the family blood pressure program. Genome Res 2003; 13: 173–181.
• 14.   Takeuchi S, Esaki H, Furusyo N, Hayashida S, Yamamura K, Tsuji G, et al. Incidence, serum IgE and TARC/CCL17 levels in atopic dermatitis associated with other allergic diseases: An update from the Ishigaki cohort. Acta Derm Venereol 2015; 95: 480–484.
• 15.   Kataoka Y. Thymus and activation-regulated chemokine as a clinical biomarker in atopic dermatitis. J Dermatol 2014; 41: 221–229.
• 16.   Lee CP, Huang YH, Hsu YW, Yang KD, Chien HC, Yu HR, et al. TARC/CCL17 gene polymorphisms and expression associated with susceptibility and coronary artery aneurysm formation in Kawasaki disease. Pediatr Res 2013; 74: 545–551.
• 17.   Sekiya T, Tsunemi Y, Miyamasu M, Ohta K, Morita A, Saeki H, et al. Variations in the human Th2-specific chemokine TARC gene. Immunogenetics 2003; 54: 742–745.
• 18.   Leung TF, Tang NL, Li CY, Lam CW, Wong GW, Fok TF. Association between TARC C-431T and atopy and asthma in children. J Allergy Clin Immunol 2004; 114: 199–202.
• 19.   Tsunemi Y, Komine M, Sekiya T, Saeki H, Nakamura K, Hirai K, et al. The −431C>T polymorphism of thymus and activation-regulated chemokine increases the promoter activity but is not associated with susceptibility to atopic dermatitis in Japanese patients. Exp Dermatol 2004; 13: 715–719.
• 20.   Fujisawa T, Fujisawa R, Kato Y, Nakayama T, Morita A, Katsumata H, et al. Presence of high contents of thymus and activation-regulated chemokine in platelets and elevated plasma levels of thymus and activation-regulated chemokine and macrophage-derived chemokine in patients with atopic dermatitis. J Allergy Clin Immunol 2002; 110: 139–146.
• 21.   Abi-Younes S, Si-Tahar M, Luster AD. The CC chemokines MDC and TARC induce platelet activation via CCR4. Thromb Res 2001; 101: 279–289.
• 22.   Zenz R, Eferl R, Scheinecker C, Redlich K, Smolen J, Schonthaler HB, et al. Activator protein 1 (Fos/Jun) functions in inflammatory bone and skin disease. Arthritis Res Ther 2008; 10: 201.
• 23.   Goncalves I, Stollenwerk MM, Lindholm MW, Dias N, Pedro LM, Fernandes JF, et al. Activator protein-1 in carotid plaques is related to cerebrovascular symptoms and cholesteryl ester content. Cardiovasc Pathol 2011; 20: 36–43.
• 24.   Rivard A, Principe N, Andres V. Age-dependent increase in c-fos activity and cyclin A expression in vascular smooth muscle cells. A potential link between aging, smooth muscle cell proliferation and atherosclerosis. Cardiovasc Res 2000; 45: 1026–1034.
• 25.   Osto E, Matter CM, Kouroedov A, Malinski T, Bachschmid M, Camici GG, et al. c-Jun N-terminal kinase 2 deficiency protects against hypercholesterolemia-induced endothelial dysfunction and oxidative stress. Circulation 2008; 118: 2073–2080.
• 26.   Wang J, An FS, Zhang W, Gong L, Wei SJ, Qin WD, et al. Inhibition of c-Jun N-terminal kinase attenuates low shear stress-induced atherogenesis in apolipoprotein E-deficient mice. Mol Med 2011; 17: 990–999.