Chemokine Ligand 2 Genetic Variants, Serum Monocyte Chemoattractant Protein-1 Levels, and the Risk of Coronary Artery Disease
Objective— In humans, evidence about the association between levels of monocyte chemoattractant protein-1 (MCP-1), its coding gene chemokine (C-C motif) ligand 2 (CCL2), and risk of coronary artery disease (CAD) is contradictory.
Methods and Results— We performed a nested case-control study in the prospective EPIC-Norfolk cohort investigating the relationship between CCL2 single-nucleotide polymorphisms (SNPs), MCP-1 concentrations, and the risk of future CAD. Cases (n=1138) were apparently healthy men and women aged 45 to 79 years who developed fatal or nonfatal CAD during a mean follow-up of 6 years. Controls (n=2237) were matched by age, sex, and enrollment time. Using linear regression analysis no association between CCL2 SNPs and MCP-1 serum concentrations became apparent, nor did we find a significant association between MCP-1 serum levels and risk of future CAD. Finally, Cox regression analysis showed no significant association between CCL2 SNPs and the future CAD risk. In addition, we did not find any robust associations between the CCL2 haplotypes and MCP-1 serum concentration or future CAD risk.
Conclusion— Our data do not support previous publications indicating that MCP-1 is involved in the pathogenesis of CAD.
- coronary artery disease
- monocyte chemoattractant protein-1
- single-nucleotide polymorphism
Chemokines (chemotactic cytokines) are small heparin-binding proteins that direct the movement of circulating leukocytes toward sites of inflammation, such as injury or atherosclerotic plaque. One of the best characterized chemokines is monocyte chemoattractant protein 1 (MCP-1; in the systematic nomenclature the gene is know as chemokine ligand [C-C motif] 2; CCL2).1 CCL2 lies on the long arm of chromosome 17. It has 3 exons extending over ≈2000 bp. The gene has both distal and proximal regulatory elements important for cytokine and constitutive activity, respectively. MCP-1 is a potent chemoattractant for monocytes, dendritic cells, memory T cells, and basophils.2,3 MCP-1 is present in macrophage-rich atherosclerotic plaques,4,5 where its production in endothelial and smooth-muscle cells is induced by oxidized low-density lipoprotein (LDL)–cholesterol. MCP-1 has thus emerged as a potential link between oxidized lipoproteins and the recruitment of monocytes to the arterial wall. Several lines of evidence suggest that MCP-1 is indeed involved in atherosclerosis.
To clarify the role of MCP-1 in the pathophysiology of coronary artery disease (CAD), we conducted an analysis of the associations among CCL2 genetic variants, serum levels of MCP-1, and the risk of future CAD among apparently healthy men and women.
Materials and Methods
For the present nested case-control study in the EPIC-Norfolk prospective cohort (for a description of the cohort, please see supplemental material, available online at http://atvb.ahajournals.org), we identified apparently healthy individuals who developed fatal or nonfatal CAD during follow-up. Apparently healthy individuals were defined as study participants who did not report a history of heart attack or stroke at the baseline clinic visit. Controls were apparently healthy study participants who remained free of cardiovascular disease during follow-up. Controls were matched cases by sex, age (within 5 years), and date of visit (within 3 months). The average follow-up was 6 years.
Nonfasting blood samples were taken by vein puncture into serum tubes. Blood samples were stored at −80°C before analysis. Lipid alevels and C-reactive protein (CRP) levels were measured as described previously.6 Serum MCP-1 levels were determined by a multiplex assay using the Bioplex Suspension Array (Bio-Rad, Veenendaal, the Netherlands) as readout system. All samples above the 95th percentile were repeated. Intraassay coefficient of variation was less than 3%, whereas the interassay coefficient of variation was 3.2%. Samples were analyzed in random order to avoid systematic bias. Researchers and laboratory personnel had no access to identifiable information and could identify samples by number only.
MCP-1 Genotyping and Haplotype Analysis
We selected 7 common CCL2 single-nucleotide polymorphisms (SNPs): −2835A>C (rs2857654), −2578A>G (rs1024611), −2136A>T (rs1024610), −1811A>G (rs3760399), −927G>C (rs3760396), +764C>G (rs2857657), and +3726T>C (rs2530797) spanning the gene based on previously published selection criteria.7 The SNPs −2835, −2578, −2136, and −1811 are located on the distal regulatory region, whereas −927, +764, and +3726 are located on a promoter, intron 1, and 3′ flanking region, respectively. Positions of the 7 SNPs at the CCL2 locus and linkage disequilibrium (LD) structure are depicted in Supplemental Figure I. CCL2 genotyping was performed on coded DNA samples by laboratory personnel blinded to clinical information. Genotyping was conducted by KBioscience (http://www.kbioscience.co.uk) using KASPar technology. Genotyping was carried out on an ABI 7900 system, using Assay by Design assays (Applied Biosystems, Foster City, Calif). Allelic discrimination was performed using FAM and VIC as fluorophore. Polymerase chain reaction conditions were denaturation for 10 minutes at 95°C, followed by 40 cycles of 30 seconds at 92°C and 45 seconds at 60°C. Polymerase chain reaction assay mix was obtained from Applied Biosystems. Assays were considered successful if they met the following criteria: at least 75% for genotyping calls, a Hardy-Weinberg equilibrium with a probability value >0.01, and a minor allele frequency >5%. Haplotype block selection and estimations of the LD were performed with the publicly available Haploview software package, version 4.2 (http://www.broadinstitute.org/mpg/haploview).
Using a logistic regression model, we calculated the power to detect statistically significant differences in CAD risk. With minor allele frequencies ranging from 0.4 to 0.05, our study had 80% power to detect an odds ratio (OR) of 1.3 to 1.65, respectively. Likewise, the study had 80% power to detect 10 to 4.25 pg/mL differences in MCP-1 levels assuming an overall standard deviation of 35 pg/mL and minor allele frequencies ranging from 0.4 to 0.05. For both models, we assumed a (log)additive effect of the SNP and a corrected 2-sided α of 0.0005. Calculations were carried out using Quanto (version 1.2, http://hydra.usc.edu/gxe/).
Baseline characteristics were compared between cases and controls with a mixed-effect model for continuous variables or conditional logistic regression for categorical variables. Because MCP-1, triglycerides, and CRP levels had a skewed distribution, values were natural log transformed before statistical analysis. The associations between MCP-1 quartiles and both cardiovascular risk factors and CAD risk were assessed. For this purpose, quartiles were based on the MCP-1 distribution among controls. The relationships between CCL2 genotype and MCP-1 levels were determined by a linear regression model. Multivariable-adjusted Cox regression analyses were conducted to examine the association between CCL2 genotype and risk of CAD. We tested for interaction between sex and CCL2 polymorphisms because sex differences have been described previously for the association between MCP-1 and CAD risk.7 Because we observed a statistically significant interaction between sex and one of the SNPs for CAD risk, we performed additional subgroup analyses for men and women separately. For all SNP statistical analyses with the 7 typed SNPs, we present uncorrected probability values and consider a multiple testing Bonferroni corrected probability value of <0.0005 significant; otherwise, a probability value of <0.05 was considered statistically significant. Data were analyzed with SPSS, version 16.0 (SPSS Inc, Chicago, Ill), unless otherwise described.
From the unphased SNP genotype data, haplotype frequencies and their association with MCP-1 concentrations and CAD risk were estimated using weighted linear or logistic regression, respectively.8,9 In short, haplotype effects and haplotype frequencies were jointly estimated using an expectation-maximization algorithm in which individual haplotypes were handled as missing data. In the first expectation step, the initial probabilities were calculated using Bayes’ theorem and estimated haplotype frequencies. In the following expectation steps, the posterior probabilities of haplotype pairs compatible with an individual’s genotype were calculated based on the phenotype of the individual. In the maximization steps, the haplotype effects were estimated using a weighted linear or logistic regression model, where the posterior probabilities functioned as weights. The expectation and maximization steps were alternated until convergence. Haplotype analyses were performed with R (GNU project, http://www.r-project.org/).
MCP-1 Serum Levels, Cardiovascular Risk Factors, and Risk of Subsequent CAD
A complete dataset was available for 985 cases and 1778 matched controls. From these individuals, 793 cases were matched to 2 controls each, whereas 192 cases could be matched to 1 control only. Matching ensured that age and sex distributions were comparable between cases and controls. Table 1 shows the distribution of cardiovascular risk factors among cases and controls. As expected individuals who developed CAD during follow-up were more likely than controls to have cardiovascular risk factors. There was no significant difference between circulating MCP-1 level between cases and controls (Table 1). Serum MCP-1 levels were associated with waist circumference and triglycerides (Table 2). There were also weak but significant relationships with body mass index and systolic blood pressure.
Next, we examined the relationship between circulating MCP-1 levels and the risk of future CAD. We found no evidence for an association between serum MCP-1 levels and CAD risk (Table 3). Because we found a significant interaction for MCP-1 to 3725 with sex for CAD risk (P=0.004), we performed an additional sex-specific subgroup analyses showing no indication for an association between MCP-1 serum levels and CAD risk in men or women separately (Supplemental Table Ι).
CCL2 Genotype Variants and Circulating MCP-1 Levels
Supplemental Table II displays characteristics for the CCL2 SNPs that were typed. Slight differences between the minor allele frequencies for cases and controls were found for CCL2 −2835, −2578, −1811, and +764. All polymorphisms in the control population were in complete Hardy-Weinberg equilibrium. The various cardiovascular risk factors were equally distributed among the 7 different SNPs (Supplemental Table ΙΙI). Median levels of MCP-1 serum concentration showed minor variations according to CCL2 genotype, but no significant differences were found (Table 4.) Subgroup analyses for men and women separately showed no evidence for a significant association between CCL2 genotype and MCP-1 serum levels (Supplemental Tables IV and V).
CCL2 Genotype Variants and Risk of CAD
Although CCL2 genotype variation was not associated with serum levels of MCP-1, genotype variations could still affect CAD risk via other mechanisms independent of circulating levels of MCP-1. We therefore assessed the association between CCL2 genotype variants and the risk of CAD. Table 5 shows the association between the typed CCL2 SNPs and risk of CAD. We did not find any robust associations with CAD risk of the specific CCL2 genotype variants. A subgroup analysis among men showed significant associations with CAD risk of both CCL2 −2835 (OR 1.28; 95% CI, 1.05 to 1.57; P=0.017 for CC versus AA+AC) and CCL2 −2578 (OR 1.26; 95% CI, 1.03 to 1.53; P=0.027 for AA versus GG+GA). Adjustment for age, sex, body mass index, smoking status, systolic blood pressure, LDL-cholesterol, high-density lipoprotein (HDL)–cholesterol, CRP, and adjustment for the Framingham Risk Score did not influence these associations. In addition, among women only, CCL2 +3726 was associated with CAD risk in a recessive model (OR, 1.59; 95% CI, 1.11 to 2.27; P=0.011), that was highly robust for multivariable correction and correction for the Framingham Risk Score. However, probability values did not reach significance beyond the multiple testing criterion of 0.0005 (Supplemental Tables VI and VII).
MCP-1 Haplotype Analysis
To better understand the associations among CCL2 genetic variation, circulating MCP-1 concentrations, and the risk of future CAD, we performed a haplotype-based analysis. The CCL2 gene was encompassed in 1 haplotype block. We estimated 6 common haplotypes (H1 to H6) from the 7 typed SNPs (Supplemental Table VIII) that accounted for 99% of all possible CCL2 haplotypes. Using the haplotype with the highest frequency in our study as reference, we did find a trend toward lower concentrations of MCP-1 for individuals with H4 (ratio, −1.81; 95% CI, −3.71 to −0.09; P=0.062) and a lower risk ratio for future CAD in individuals with H5 of 0.77 (95% CI, 0.61 to 0.98; P=0.030). These probability values did not reach significance above our predefined multiple testing criterion (Table 6). After finding a significant interaction between sex and the haplotype H5 for MCP-1 concentration, we performed subgroup analyses for men and women separately. We did not find any significant associations between CCL2 haplotypes and MCP-1 serum levels or future risk of CAD (Supplemental Tables IX and X).
In this large prospective case-control study, we found no evidence for an association between MCP-1 serum levels and the risk of future CAD in apparently healthy men and women. In addition, no significant associations were found between CCL2 genetic variants and either serum MCP-1 levels or the risk of future CAD. We found no robust evidence for any of these associations in a subsequent CCL2 haplotype analysis.
Several studies have suggested that increased levels of MCP-1 are associated with atherosclerosis, myocardial infarction size, as well as with an increased risk of myocardial infarction, sudden death, coronary angioplasty, and stent restenosis,10–14 whereas other studies could not confirm such an association.15 In addition, several studies have reported an association between the CCL2 SNPs investigated in our analysis and MCP-1 serum levels. Increased levels of MCP-1 were found in individuals with the CCL2 −2578G variant,16–18 but this could not be confirmed in other case-cohort studies.19,20 We found similar MCP-1 serum concentrations among CCL2 −2578GG and −2578AA individuals. Likewise, in the community-based Framingham Heart Study Offspring Cohort, McDermott et al7 demonstrated that the CCL2 −2136 and the CCL2 +764 polymorphisms were significantly associated with MCP-1 serum concentrations. Although we found a trend for CCL2 −2136 and for CCL2 +764, the probability values did not reach significance above our multiple testing criteria. In addition, 3 studies have reported associations between the CCL2 −2578G allele and atherosclerosis.7,21,22 We could not demonstrate an association for the CCL2 −2578G allele with the future risk of CAD, but we did find a nonsignificant trend for a higher risk of CAD among CCL2 +3726CC individuals (OR, 1.31; 95% CI, 1.05 to 1.65; P=0.020).
To further clarify the role of CCL2 genotype and the risk of future CAD, we performed a haplotype analysis estimating the effect of CCL2 genotype combinations on MCP-1 serum concentrations and CAD risk. Haplotype frequencies were comparable to previously published studies showing associations between CCL2 genotype combinations and MCP-1 serum levels.7 We found no significant associations between CCL2 haplotypes and MCP-1 serum levels or the risk of future CAD.
In contrast to other studies reporting evidence for MCP-1 in the pathogenesis of CAD, we could not confirm an association between MCP-1 and CAD risk in our cohort. To the best of our knowledge, strong and consistent associations between a single CCL2 SNP, MCP-1 serum level, and the risk of future CAD have not been reported in large prospective studies. Despite the substantial amount of research into the role of MCP-1 in atherogenesis, there is little information with regard to the functionality of the MCP-1 protein affected by any of the known SNPs, and this could explain inconsistent associations between CCL2 genotype, MCP-1 serum levels, and CAD. There are several other considerations that might explain the differences between our observations and previous publications. First, case ascertainment is an issue in the design of every prospective study, including this one. However, a validation study indicated that case ascertainment in our study was at least equivalent to that of other large prospective cohort studies.23 Another possibility to explain our negative findings is an insufficient power to detect differences in MCP-1 serum levels or the risk of CAD. However, our power analysis showed that with the present sample size, the study has 80% power to detect an OR of 1.3 for any of the typed SNPs. This is well below previously published ORs for the typed SNPs in previous publications, where ORs ranged between 1.5 and 2.6, and strong enough to detect clinically relevant differences.7 This observation may point toward selective publication of positive findings in previous studies. Third, we present full and transparent data, in accordance with the current STREGA guidelines.24 We present all tests performed for possible associations of a well described, previously published, large case-control study and specifically define or present the selection criteria and quality controls. Although this is not uncommon, several previous publications lack crucial information to assess the reliability of the presented data and the underreporting of negative associations. Furthermore, many studies do not correct for multiple comparisons, which in our opinion should be taken into account when addressing associations between multiple SNPs and diseases, especially when subgroup analyses are performed. Our results apply to whites only, and conclusions should not be compared with nonwhite populations, especially because CCL2 genotype frequencies have been reported to vary substantially among various populations.
This large community-based prospective study among apparently healthy men and women does not support an association between common variants in the CCL2 gene or MCP-1 serum levels and the risk of future CAD.
We thank the participants, general practitioners, and staff in EPIC-Norfolk.
Sources of Funding
EPIC-Norfolk is supported by program grants from the Medical Research Council UK and Cancer Research UK.
The funding sources had no role in study design, conduct analysis, or decision to submit the manuscript for publication.
Received on: February 25, 2010; final version accepted on: April 19, 2010.
Bacon K, Baggiolini M, Broxmeyer H, Horuk R, Lindley I, Mantovani A, Maysushima K, Murphy P, Nomiyama H, Oppenheim J, Rot A, Schall T, Tsang M, Thorpe R, Van Damme J, Wadhwa M, Yoshie O, Zlotnik A, Zoon K. Chemokine/chemokine receptor nomenclature. J Interferon Cytokine Res. 2002; 22: 1067–1068.
Zernecke A, Shagdarsuren E, Weber C. Chemokines in atherosclerosis: an update. Arterioscler Thromb Vasc Biol. 2008; 28: 1897–1908.
Yu X, Dluz S, Graves DT, Zhang L, Antoniades HN, Hollander W, Prusty S, Valente AJ, Schwartz CJ, Sonenshein GE. Elevated expression of monocyte chemoattractant protein 1 by vascular smooth muscle cells in hypercholesterolemic primates. Proc Natl Acad Sci U S A. 1992; 89: 6953–6957.
Bruins P, te Velthuis H, Yazdanbakhsh AP, Jansen PG, van Hardevelt FW, de Beaumont EM, Wildevuur CR, Eijsman L, Trouwborst A, Hack CE. Activation of the complement system during and after cardiopulmonary bypass surgery: postsurgery activation involves C-reactive protein and is associated with postoperative arrhythmia. Circulation. 1997; 96: 3542–3548.
McDermott DH, Yang Q, Kathiresan S, Cupples LA, Massaro JM, Keaney JF, Jr., Larson MG, Vasan RS, Hirschhorn JN, O'Donnell CJ, Murphy PM, Benjamin EJ. CCL2 polymorphisms are associated with serum monocyte chemoattractant protein-1 levels and myocardial infarction in the Framingham Heart Study. Circulation. 2005; 112: 1113–1120.
Tanck MW, Klerkx AH, Jukema JW, De Knijff P, Kastelein JJ, Zwinderman AH. Estimation of multilocus haplotype effects using weighted penalised log-likelihood: analysis of five sequence variations at the cholesteryl ester transfer protein gene locus. Ann Hum Genet. 2003; 67: 175–184.
de Lemos JA, Morrow DA, Sabatine MS, Murphy SA, Gibson CM, Antman EM, McCabe CH, Cannon CP, Braunwald E. Association between plasma levels of monocyte chemoattractant protein-1 and long-term clinical outcomes in patients with acute coronary syndromes. Circulation. 2003; 107: 690–695.
Cipollone F, Marini M, Fazia M, Pini B, Iezzi A, Reale M, Paloscia L, Materazzo G, D'Annunzio E, Conti P, Chiarelli F, Cuccurullo F, Mezzetti A. Elevated circulating levels of monocyte chemoattractant protein-1 in patients with restenosis after coronary angioplasty. Arterioscler Thromb Vasc Biol. 2001; 21: 327–334.
Mosedale DE, Smith DJ, Aitken S, Schofield PM, Clarke SC, McNab D, Goddard H, Gale CR, Martyn CN, Bethell HW, Barnard C, Hayns S, Nugent C, Panicker A, Grainger DJ. Circulating levels of MCP-1 and eotaxin are not associated with presence of atherosclerosis or previous myocardial infarction. Atherosclerosis. 2005; 183: 268–274.
Gonzalez E, Rovin BH, Sen L, Cooke G, Dhanda R, Mummidi S, Kulkarni H, Bamshad MJ, Telles V, Anderson SA, Walter EA, Stephan KT, Deucher M, Mangano A, Bologna R, Ahuja SS, Dolan MJ, Ahuja SK. HIV-1 infection and AIDS dementia are influenced by a mutant MCP-1 allele linked to increased monocyte infiltration of tissues and MCP-1 levels. Proc Natl Acad Sci U S A. 2002; 99: 13795–13800.
Simeoni E, Winkelmann BR, Hoffmann MM, Fleury S, Ruiz J, Kappenberger L, Marz W, Vassalli G. Association of RANTES G-403A gene polymorphism with increased risk of coronary arteriosclerosis. Eur Heart J. 2004; 25: 1438–1446.
Szalai C, Duba J, Prohaszka Z, Kalina A, Szabo T, Nagy B, Horvath L, Csaszar A. Involvement of polymorphisms in the chemokine system in the susceptibility for coronary artery disease (CAD): coincidence of elevated Lp (a) and MCP-1–2518 G/G genotype in CAD patients. Atherosclerosis. 2001; 158: 233–239.
Alonso-Villaverde C, Coll B, Parra S, Montero M, Calvo N, Tous M, Joven J, Masana L. Atherosclerosis in patients infected with HIV is influenced by a mutant monocyte chemoattractant protein-1 allele. Circulation. 2004; 110: 2204–2209.
Boekholdt SM, Peters RJ, Day NE, Luben R, Bingham SA, Wareham NJ, Hack CE, Reitsma PH, Khaw KT. Macrophage migration inhibitory factor and the risk of myocardial infarction or death due to coronary artery disease in adults without prior myocardial infarction or stroke: the EPIC-Norfolk Prospective Population study. Am J Med. 2004; 117: 390–397.
Little J, Higgins JP, Ioannidis JP, Moher D, Gagnon F, von Elm E, Khoury MJ, Cohen B, Davey-Smith G, Grimshaw J, Scheet P, Gwinn M, Williamson RE, Zou GY, Hutchings K, Johnson CY, Tait V, Wiens M, Golding J, van Duijn C, McLaughlin J, Paterson A, Wells G, Fortier I, Freedman M, Zecevic M, King R, Infante-Rivard C, Stewart A, Birkett N. STrengthening the REporting of Genetic Association Studies (STREGA)—an extension of the STROBE statement. Genet Epidemiol. 2009; 33: 581–598.