CCL2 Polymorphisms Are Associated With Serum Monocyte Chemoattractant Protein-1 Levels and Myocardial Infarction in the Framingham Heart Stu
http://www.100md.com
《循环学杂志》
the Laboratory of Molecular Immunology, National Institute of Allergy and Infectious Diseases, National Institutes of Health, Bethesda, Md (D.H.M., P.M.M.)
Department of Biostatistics, School of Public Health (Q.Y., L.A.C.), Department of Mathematics (J.M.M., M.G.L.), and School of Medicine (J.F.K., R.S.V., E.J.B.), Boston University, Boston, Mass
Cardiology Division, Massachusetts General Hospital (S.K., C.J.O.) and Department of Genetics (J.N.H.), Harvard Medical School, Boston, Mass
Broad Institute, Cambridge, Mass (S.K., C.J.O.)
National Heart, Lung and Blood Institute, National Institutes of Health, Bethesda, Md (C.J.O.)
National Heart, Lung and Blood Institute’s Framingham Heart Study, Framingham, Mass (S.K., R.S.V., C.J.O., E.J.B.).
Abstract
Background— Monocyte chemoattractant protein-1 (MCP-1) is a chemokine strongly implicated in promoting atherosclerosis in animal models, but human genetic evidence is contradictory.
Methods and Results— We analyzed the association of genetic variation in the MCP-1 gene (CCL2) with prevalent myocardial infarction and serum MCP-1 levels in the community-based Framingham Heart Study Offspring Cohort (50% women; mean age, 62 years). MCP-1 levels and CCL2 genotypes were determined in 3236 and 1797 individuals, respectively. Significant clinical correlates of MCP-1 levels were age, cigarette smoking, triglycerides, body mass index, and waist-to-hip ratio. The MCP-1-2578G allele located in the CCL2 regulatory region was significantly associated with both higher serum MCP-1 levels in a recessive genetic model (358±10 versus 328±3 pg/mL; P=0.002) and higher prevalence of myocardial infarction in a dominant genetic model (adjusted odds ratio, 2.0; 95% CI, 1.2 to 3.3; P=0.005). We also defined the linkage disequilibrium structure at the CCL2 locus and observed 6 common haplotypes in whites. We performed haplotype-based association analysis and found that only the most frequent haplotype, defined by the MCP-1-2578G allele, was associated with prevalent MI.
Conclusions— Our data are consistent with the hypothesis that MCP-1 is involved in the pathogenesis of human atherosclerosis and myocardial infarction.
Key Words: epidemiology ; genetics ; inflammation ; myocardial infarction ; risk factors
Introduction
Atherosclerosis is now recognized as a disease of arterial inflammation that arises from the interactions of migratory leukocytes with resident vascular endothelial cells, smooth muscle cells, and fibroblasts.1 At the molecular level, interactions among these cell types are regulated by cytokines, adhesion molecules, and chemoattractants. Chemokines are a large family of chemoattractants that direct migration of leukocytes from the blood to sites of inflammation.2
Several lines of evidence suggest that the chemokine, monocyte chemoattractant protein-1 (MCP-1; gene name, CCL2, previously SCYA2) and its receptor CCR2 (gene name, CCR2) are involved in atherosclerosis.3 In various murine models of atherosclerosis (apolipoprotein E and LDL receptor knockouts, apolipoprotein B transgenics), deletion of CCL2 or CCR2 results in large (50% to 80%) reductions in atherosclerotic plaque size.4–6 Conversely, overexpression of MCP-1 in the leukocytes of susceptible mice results in increased plaque size.7 Furthermore, in animal models, temporary blockade of CCR2 by antibodies or gene therapy blocks restenosis after balloon angioplasty and stabilizes and reduces atherosclerotic lesion size.8–10
MCP-1 is expressed in human atherosclerotic lesions, and CCR2 is expressed on leukocytes.11–13 In addition, MCP-1 induces arrest and transmigration from the circulation of CCR2+ monocytes under conditions of physiological shear force and promotes monocyte differentiation to lipid-laden macrophages.14,15 MCP-1 also contributes to the proliferation of arterial smooth muscle cells,16 which, along with macrophages, constitute the key cellular components of atherosclerotic plaques.
A growing number of human epidemiological studies have suggested links between circulating MCP-1 levels and atherosclerosis. Higher MCP-1 levels have been associated with increased risks of myocardial infarction (MI), sudden death, coronary angioplasty, and stent restenosis.17–20 However, little is known about the role of MCP-1 levels in cardiovascular disease (CVD) in the general population.
Genetic variation of CCL2 and CCR2 could modulate MCP-1 production and function. In fact, 3 studies have associated the CCL2 distal regulatory region single nucleotide polymorphism (SNP) MCP-1-2578G with increased circulating levels of MCP-1.21–23 However, genetic associations reported between MCP-1-2578G and CVD end points have not been consistent. Two case-control studies have reported an association between MCP-1-2578G and atherosclerosis in whites. The first reported an increased frequency of MCP-1-2578G homozygotes (odds ratio [OR], 2.2; P<0.005) among individuals referred for CABG compared with age- and sex-matched controls (n=638).24 The second found an increased frequency of the MCP-1-2578G allele among HIV patients with subclinical atherosclerosis identified by ultrasound (OR, 5.7; P=0.004) compared with those lacking the allele (n=183).25 However, a third study was unable to find an association between the same polymorphism and MI in a Japanese cohort (n=909),26 and a fourth study found no significant association of this allele with angiographically determined coronary artery disease (CAD) in whites (n=3224).27 The inconsistency of these results could be due in part to the case-control design of these studies or to reliance on the testing of a single polymorphism. To address these issues, we have conducted an analysis of CCL2 genetic variation and its relation to serum levels of MCP-1 and prevalent MI in the Framingham Heart Study (FHS) Offspring Cohort, a large, longitudinal, observational, community-based study of the epidemiology of CVD.
Methods
FHS Offspring Cohort Subjects
The design of the FHS has previously been described.28,29 Briefly, the FHS enrolled 5124 children of the original participants and their spouses, referred to as the Offspring Cohort, beginning in 1971. Participants undergo a routine comprehensive medical examination every 4 years. At the cohort’s seventh examination cycle, 3236 attendees had serum MCP-1 levels (henceforth referred to as MCP-1 levels) measured. The FHS Offspring Cohort contained 1888 unrelated participants for whom DNA was available in 1995 to 1998. Of these, 1797 gave informed consent for genotyping, had clinical and genotype information available, and were included in the genotype-phenotype analyses; MCP-1 levels were performed in 1602 genotyped participants. The study was approved by the National Institute of Allergy and Infectious Diseases and Boston University School of Medicine Institutional Review boards, and all participants signed informed consent.
Clinical Risk Factors and Cardiovascular Outcomes
Outcomes were evaluated up to the time of the seventh examination (2002, the last year for which data were complete) by investigators blinded to genotypic data. MI was defined as either hospitalization for acute MI (95%) or as new ECG evidence consistent with a clinically unrecognized MI (5%). The occurrence of MI was determined by a 3-physician end-point committee that reviewed records from routine study examinations outside physicians’ offices and hospitalizations. Of the subjects whose CCL2 genotypes had been determined, 107 (6%) had prevalent MI at the time of the seventh examination. CVD risk factors, including serum total cholesterol, HDL, and triglycerides, were classified at the seventh examination cycle. Smoking was defined as cigarette use within the prior year. Diabetes was defined as fasting blood glucose level 7.0 nmol/L (126 mg/dL) or treatment with hypoglycemic agents. Hypertension was defined as systolic or diastolic blood pressure 140/90 mm Hg or antihypertensive medication treatment. Use of medications in the statin or ;-blocker classes was also evaluated as 2 separate dichotomous variables (yes/no) at the time of the seventh examination cycle.
Measurement of Serum MCP-1 Levels
Fasting serum samples were collected at the seventh examination and stored at –70°C. For analysis, samples were thawed and vortexed vigorously, and MCP-1 levels (pg/mL) were measured in duplicate according to the manufacturer’s instructions with a commercially available ELISA (R&D Systems). All samples with an intra-assay coefficient of variation above the 95th percentile were repeated; the mean of repeated duplicate values was used. Initially, samples had an average coefficient of variation of 6.1±3.3%; samples with a coefficient of variation >10.9% were rerun. However, we improved our pipetting technique. For the final 43% of samples, we achieved an average coefficient of variation of 1.9±1.6%; samples with a coefficient of variation >5.3% were rerun.
Haplotype Block Definition and Selection of Haplotype Tag SNPs
From the public dbSNP database (http://www.ncbi.nlm.nih.gov/SNP), we selected 24 evenly spaced markers within a 10.6-kb region spanning the CCL2 gene (GenBank NM_002982). The SNPs were genotyped in a panel of 12 multigenerational white family pedigrees from the Centre d’Etude du Polymorphisme Humain (CEPH) Utah panel (Coriell Institute for Medical Research, Camden, NJ).30 These reference pedigrees included 93 individuals representing 96 independent chromosomes of European ancestry. Assays were considered successful if they met the following criteria: at least 75% success for genotyping calls, Hardy-Weinberg equilibrium P>0.01, and Mendelian transmission errors 1. In addition, we imposed a minor allele frequency threshold and defined "common" for the present study as a minor allele frequency 5%. Overall, we developed successful assays for 9 SNPs.
Haplotype blocks were defined through the use of the criteria of Gabriel et al,30 and tag SNPs were selected in the publicly available Haploview software package, version 2.03.31 For each pair of markers, we calculated D', an estimate of the strength of linkage disequilibrium [LD], and a logarithm of the odds (LOD) score, which is an estimate of LD significance. From these 2 measures, each pairwise marker comparison was categorized into 1 of 3 groups: (1) no or minimal evidence of historical recombination (D'=1/LOD >2 or 0.52), (2) strong evidence of historical recombination (D'<1/LOD <2 or D'<0.5/any LOD), and (3) uninformative (D'=1.0/LOD <2).
Genotyping
CCL2 genotyping was performed on coded DNA samples by laboratory personnel blinded to clinical information. The genotyped variants were named to reflect the distance in nucleotides from the translational start point in GenBank D26087 and their unique identifiers in the NCBI SNP database (dbSNP; http://www.ncbi.nlm.nih.gov/SNP). We used previously published methods for the MCP-1-2578 and MCP-1-2136 polymorphisms (also previously called MCP-1-2518 and MCP-1-2076).21 In addition to these 2 previously studied SNPs, 5 tag SNPs were genotyped with commercially available (Sequenom) matrix-associated laser desorption ionization time-of-flight mass spectrometry technology and previously published protocols.30
Statistical Analysis
MCP-1 levels were logarithmically transformed before analysis because they were not normally distributed. Covariate selection for logarithmically transformed MCP-1 levels was conducted with a SOLAR regression model to account for correlation among members of the same family. In the SOLAR regression model, the correlation between any pair of relatives was assumed to be proportional to their kinship coefficient.32 The significance of each covariate was evaluated after adjustment for all the other factors in a likelihood ratio test. Covariates with a value of P<0.05 were kept in the final model (age and sex were forced in), selecting from the following: ratio of total to HDL cholesterol, triglycerides, systolic and diastolic blood pressures, body mass index, waist-to-hip ratio, fasting glucose, diabetes, hypertension treatment, lipid treatment, smoking status, and alcohol consumption. Regression coefficients were expressed by presence (versus absence) for dichotomous variables and per-unit increase for continuous variables. Unadjusted analyses of individual CCL2 genotypes with prevalent MI were performed by cross-tabulation and calculation of ORs. Multivariable-adjusted logistic regression analyses were conducted to examine the partial contribution of CCL2 genotype to risk of prevalent MI (SAS, version 8, SAS Institute Inc).33 The OR and 95% confidence limits were calculated with this model; significance was evaluated based on Wald 2 tests. The association between established risk factors and CCL2 genotypes was examined by a 2 test with 1 df for dichotomous measurements and by testing the differences in means of the risk factors between CCL2 genotype groups with a 2-tailed Student t test for continuous measurements. Tests of interaction between CCL2 genotypes and the established risk factors were also performed using logistic regression, in which 2-df Wald 2 tests were performed for the interactions between the CCL2 genotype and age, sex, body mass index, hypertension, diabetes, and smoking. Haplotype-based association analyses were conducted using a weighted regression approach as implemented in the haplo.score program.34,35 Haplotype frequency was estimated with the expectation-maximization algorithm.36 All compatible haplotype configurations of a multimarker genotype were used in the regression, with weights being the corresponding likelihood of such configuration. A global score statistic tested all haplotypes simultaneously to detect any departure from the null hypothesis of no association.
Results
We assessed MCP-1 levels in 3236 participants and genotyped 7 common CCL2 SNPs spanning the gene (Figure 1) in 1797 unrelated participants of the FHS. Table 1 displays the clinical characteristics of participants with MCP-1 levels and those with genotypes.
Clinical Correlates of MCP-1 Levels
In stepwise linear regression models, the clinical correlates of MCP-1 levels were age, cigarette smoking, triglycerides, body mass index, and waist-to-hip ratio (Table 2). Together, these covariates explained only 6% of the variability in MCP-1 levels. MCP-1 levels were not associated with prevalent MI in sex-specific or sex-pooled multivariable models (OR, 1.0; 95% CI, 0.99 to 1.00; P=0.50).
CCL2 Genotypes in the FHS Offspring Cohort
We found the MCP-1-2578 and MCP-1-2136 allele frequencies to be similar to those reported previously (30% and 21%, respectively, in a previous study versus 27% and 21% in the present study).21 Also consistent with previous studies, in our subjects, these 2 polymorphisms were in complete LD with each other. The genotype frequencies for each allele considered separately were consistent with Hardy-Weinberg predictions.
MCP-1 Genotypes and Serum MCP-1 Levels
A previous association has been demonstrated between MCP-1-2578G and unstimulated MCP-1 levels.21–23 Therefore, we examined the association in participants with available MCP-1 levels and genotypes (n=1602, 89% of those genotyped). We found a significant positive correlation of MCP-1 levels with both the MCP-1-2578G allele in a recessive genetic model and the MCP-1-2136T allele in a dominant genetic model (Figure 2). After multivariable regression adjusting for variables found to be significantly associated with MCP-1 level (Table 2), both associations remained significant (P=0.002 and P0.0001, respectively).
CCL2 Genotypes and Prevalent MI
In a dominant genetic model, possession of an MCP-1-2578G allele was marginally associated with increased MI prevalence in the cohort as a whole (unadjusted OR, 1.5; 95% CI, 1.0 to 2.2; P=0.06; Table 3). In men, the association appeared stronger (unadjusted OR, 1.8; 95% CI, 1.2 to 2.9; P=0.009). There were no significant differences in established risk factors in those with or without an MCP-1-2578G allele. After multivariable adjustment for the covariates listed in Table 3, the MCP-1-2578G allele was associated with prevalent MI in the sex-pooled cohort (OR, 2.0; 95% CI, 1.2 to 3.3; P=0.005). In tests for effect modification, there were no significant interactions between age, sex, diabetes mellitus, cigarette smoking, or hypertension and either of the 2 polymorphisms with regard to MI prevalence. The multivariable-adjusted OR was strong in men (OR, 2.6; 95% CI, 1.5 to 4.5; P=0.0006), whereas the association with increased prevalent MI was not found in women (OR, 0.7; 95% CI, 0.2 to 2.2; P=0.50). However, very few women had prevalent MI (n=23), and there was insufficient power (0.60) to exclude a significant association in women compared with men.
In multivariable models incorporating MCP-1 levels as a covariate, possession of the MCP-1-2578G allele continued to be significantly associated with increased MI prevalence (OR, 1.9; 95% CI, 1.2 to 3.2; P=0.009; Table 3). Hence, measurement of unstimulated MCP-1 level at 1 time point did not fully account for the association between MCP-1-2578G genotype and MI prevalence.
CCL2 Haplotype Analysis
Discussion
This study provides evidence from a large, community-based cohort supporting the hypothesis that the chemokine MCP-1 is a pathogenic factor in human CVD. First, we found that homozygotes for the MCP-1-2578G allele had higher basal unstimulated MCP-1 levels. This result is consistent with previous findings in much smaller studies that the MCP-1-2578G allele is associated with increased circulating levels of MCP-1.21–23 Second, we made the novel observation that individuals possessing the MCP-1-2136T allele also had higher MCP-1 levels. Both of these genetic associations remained highly significant after adjustment for other clinical factors associated with MCP-1 levels. Third, possession of the MCP-1-2578G allele was associated with an increased risk of prevalent MI. This result is consistent with those of 2 previously published case-control studies that reported an association of the MCP-1-2578G allele with an increased risk of atherosclerosis.24,25 Furthermore, we have performed the first comprehensive analysis of the common genetic variation of CCL2 in a large community-based population and identified novel variants that are in strong LD with MCP-1-2578G (MCP-1-2835A) and MCP-1-2136T (MCP-1+764G).
The MCP-1-2578G association with MI persisted after adjustment for other well-established risk factors and was similar in magnitude to that previously reported for angiographically proven CAD.24 Moreover, the magnitude of increased risk associated with possession of the MCP-1-2578G allele was similar to the magnitude of risk augmentation attributable to established MI risk factors. For example, in the FHS Offspring Cohort, the multivariable-adjusted ORs for prevalent MI were 2.5 for diabetes, 2.2 for recent smoking, and 4.9 for male sex compared with 2.0, the OR for possession of the MCP-1-2578G allele compared with all others in this cohort (Table 3). Together, these data suggest a pathogenetic role for MCP-1 in human atherosclerotic CVD.
The exact mechanism by which the MCP-1-2578G allele might increase MI risk remains uncertain; however, presumably locally increased MCP-1 levels in the coronary arteries would promote transmigration of CCR2+ leukocytes from the blood and their organization and activation in plaque. Consistent with this model, MCP-1, CCR2, and activated macrophages are found in atherosclerotic lesions.11–13 In this model of pathogenesis, the interaction between MCP-1 and CCR2 may play several roles, but it is probably not the only factor involved in lesion formation because lesions are able to form in the absence of either of these molecules in susceptible animal models.4–6 It should be noted that CCR2 is also polymorphic in humans, and in a previous study, we investigated the association in the Framingham Offspring Cohort of an SNP of unknown functional significance in the CCR2 receptor, CCR2-64I. This SNP results in the conservative replacement of a valine at position number 64 with an isoleucine. We found a nonsignificant association of this SNP with CVD using a dominant genetic model (OR, 0.7; P=0.12) after multivariable adjustment.37 In the present study, the association of MCP-1-2578G with prevalent MI was not substantially altered by incorporating the baseline MCP-1 level as a covariate. One potential explanation is that baseline unstimulated MCP-1 levels may not reflect local tissue levels in the coronary arteries where inflammatory cytokines like interleukin-1 and tumor necrosis factor are known to be expressed.38 Previous in vitro studies have found these cytokines to have marked effects on the regulatory region where the CCL2 SNPs associated with CVD are located.39,40 In addition, the measured level of MCP-1 was done only at 1 time point, whereas longitudinal levels may be more important.
We noted that the clinical correlates of MCP-1 levels in the community were age, cigarette smoking, triglycerides, body mass index, and waist-to-hip ratio but that these factors explained only 6% of the variability in MCP-1 levels. Genetic factors may play a more important role in MCP-1 level variability. In a separate study, we have recently noted that the adjusted heritability of MCP-1 levels is 44% and found significant linkage on chromosome 1 at 186 cM (LOD, 4.27; genome-wide P=0.005).41 In the Oral Glycoprotein IIb/IIIa Inhibition with Orbofiban in Patients With Unstable Coronary Syndromes (OPUS-TIMI) 16 trial, MCP-1 levels were related to advancing age, female sex, hypertension, diabetes, and renal insufficiency.17 We suspect that de Lemos and colleagues17 observed different correlates because of the referral nature of their cohort. The relation of MCP-1 levels with 2 measures of adiposity in our cohort is of interest because of the observations by Christiansen et al42 that adipocyte MCP-1 mRNA levels correlated with circulating MCP-1 and body mass index and that weight loss was associated with a decline in circulating MCP-1 levels.
Study Strengths and Limitations
FHS is a large community-based cohort, diminishing referral biases, with routine ascertainment of CVD risk factors and events, enabling multivariable analyses and genotyping and phenotyping blinded to one another. The observational, nonrandomized nature of cardiovascular medications indicates that caution should be observed with regard to the lack of association between statins and other treatments and MCP-1 levels. Initially, we performed relatively few statistical tests to confirm the 2 previously reported SNP associations.24,25 However, we augmented our analyses with haplotypes and marker levels to examine more comprehensively the relation of MCP-1 and CCL2 to risk factors and disease. Hence, we have performed multiple statistical tests that increase the possibility of false-positive associations. Furthermore, our study examined prevalent MI; an apparent association for the MCP-1-2578G allele might represent a survival bias if individuals with that allele had a lower case fatality rate than other genotypes. Because we had few prevalent MI cases in women, we were underpowered to find an association between the CCL2 genotypes and MI in women. Additional follow-up of our sample or investigation of larger samples of women will help to clarify whether the risk of MI associated with CCL2 genotype varies by gender. Our study sample was largely middle-aged to elderly and white; the findings may not be generalizable to younger individuals or other racial groups. However, the single-race and community-based structure of the cohort minimizes the possibility that the associations reported here were due to population stratification. Although our results are consistent with 2 previous studies relating MCP-1-2578G to atherosclerosis phenotypes (CAD and carotid intima-media thickness),24,25 our report is the first to relate MCP-1-2578G to the MI phenotype.
Because of the tight LD between MCP-1-2835A and MCP-1-2578G on haplotype H1 and similarly between MCP-1-2136T and MCP-1+764G on haplotype H2, it is very difficult to separate their influence in our genetic association study. Association studies in populations with greater haplotypic diversity (eg, populations of African ancestry) and functional studies involving the newly identified sites (MCP-1-2835A for MCP-1 levels/MI and MCP-1+764G for MCP-1 levels) may help to clarify which, if any, affects production of MCP-1. In addition, more comprehensive resequencing is needed to look for other functional sites in the relevant haplotypes. Whereas it remains possible that the association is due to linkage with another as-yet unknown polymorphism, the evidence that MCP-1-2578G is a functional polymorphism using in vitro promoter assays supports the hypothesis that there is a functional role for this SNP.39 Electromobility gel shift assays have also shown allele-specific nuclear factor binding to oligonucleotide probes corresponding to the MCP-1-2578G site from TNF-stimulated hepatic primary cells and MG-63 osteosarcoma cells.21,40 The identity of these induced nuclear factors is not yet completely known but at least in MG-63 cells appeared to contain interferon regulatory factor-1.21
Research and Clinical Implications
Our data suggest several future research directions such as the investigation of the identity of the functional SNP(s) in these haplotypes, whether these SNPs act alone or instead require interaction with other SNPs, and whether transcription factors bind at these polymorphic sites in leukocytes and endothelial cells. In addition, the CCL2 polymorphisms we examined may be generally useful as genetic probes to evaluate the potential role of MCP-1 in the pathogenesis of other inflammatory diseases in humans.
We observed that the MCP-1-2578G allele is dominantly associated with increased risk of MI in humans after adjustment for other CVD risk factors. Our results provide additional support for the inflammation hypothesis of atherosclerosis pathogenesis and for further study to determine the role of MCP-1 as a proinflammatory risk factor. Identifying inflammatory polymorphisms contributing to atherosclerosis will improve our understanding of the pathogenesis of CVD and may suggest targets for improved prevention and treatment strategies.
Acknowledgments
This work was supported by the Framingham Study (N01-HC 25195) and CardioGenomics Program in Genomics Applications (U01 HL-66582, RO1 HL-076784, and RO1 HL-64753).
References
Ross R. Atherosclerosis: an inflammatory disease. N Engl J Med. 1999; 340: 115–126.
Dong VM, McDermott DH, Abdi R. Chemokines and diseases. Eur J Dermatol. 2003; 13: 224–230.
Ikeda U, Matsui K, Murakami Y, Shimada K. Monocyte chemoattractant protein-1 and coronary artery disease. Clin Cardiol. 2002; 25: 143–147.
Boring L, Gosling J, Cleary M, Charo IF. Decreased lesion formation in CCR2–/– mice reveals a role for chemokines in the initiation of atherosclerosis. Nature. 1998; 394: 894–897.
Gu L, Okada Y, Clinton SK, Gerard C, Sukhova GK, Libby P, Rollins BJ. Absence of monocyte chemoattractant protein-1 reduces atherosclerosis in low density lipoprotein receptor-deficient mice. Mol Cell. 1998; 2: 275–281.
Dawson TC, Kuziel WA, Osahar TA, Maeda N. Absence of CC chemokine receptor-2 reduces atherosclerosis in apolipoprotein E-deficient mice. Atherosclerosis. 1999; 143: 205–211.
Aiello RJ, Bourassa PA, Lindsey S, Weng W, Natoli E, Rollins BJ, Milos PM. Monocyte chemoattractant protein-1 accelerates atherosclerosis in apolipoprotein E-deficient mice. Arterioscler Thromb Vasc Biol. 1999; 19: 1518–1525.
Horvath C, Welt FG, Nedelman M, Rao P, Rogers C. Targeting CCR2 or CD18 inhibits experimental in-stent restenosis in primates: inhibitory potential depends on type of injury and leukocytes targeted. Circ Res. 2002; 90: 488–494.
Egashira K, Zhao Q, Kataoka C, Ohtani K, Usui M, Charo IF, Nishida K, Inoue S, Katoh M, Ichiki T, Takeshita A. Importance of monocyte chemoattractant protein-1 pathway in neointimal hyperplasia after periarterial injury in mice and monkeys. Circ Res. 2002; 90: 1167–1172.
Inoue S, Egashira K, Ni W, Kitamoto S, Usui M, Otani K, Ishibashi M, Hiasa K, Nishida K, Takeshita A. Anti-monocyte chemoattractant protein-1 gene therapy limits progression and destabilization of established atherosclerosis in apolipoprotein E–knockout mice. Circulation. 2002; 106: 2700–2706.
Yla-Herttuala S, Lipton BA, Rosenfeld ME, Sarkioja T, Yoshimura T, Leonard EJ, Witztum JL, Steinberg D. Expression of monocyte chemoattractant protein 1 in macrophage-rich areas of human and rabbit atherosclerotic lesions. Proc Natl Acad Sci U S A. 1991; 88: 5252–5256.
Nelken NA, Coughlin SR, Gordon D, Wilcox JN. Monocyte chemoattractant protein-1 in human atheromatous plaques. J Clin Invest. 1991; 88: 1121–1127.
Takeya M, Yoshimura T, Leonard EJ, Takahashi K. Detection of monocyte chemoattractant protein-1 in human atherosclerotic lesions by an anti-monocyte chemoattractant protein-1 monoclonal antibody. Hum Pathol. 1993; 24: 534–539.
Gerszten RE, Garcia-Zepeda EA, Lim YC, Yoshida M, Ding HA, Gimbrone MA Jr, Luster AD, Luscinskas FW, Rosenzweig A. MCP-1 and IL-8 trigger firm adhesion of monocytes to vascular endothelium under flow conditions. Nature. 1999; 398: 718–723.
Tabata T, Mine S, Kawahara C, Okada Y, Tanaka Y. Monocyte chemoattractant protein-1 induces scavenger receptor expression and monocyte differentiation into foam cells. Biochem Biophys Res Commun. 2003; 305: 380–385.
Viedt C, Vogel J, Athanasiou T, Shen W, Orth SR, Kubler W, Kreuzer J. Monocyte chemoattractant protein-1 induces proliferation and interleukin-6 production in human smooth muscle cells by differential activation of nuclear factor-B and activator protein-1. Arterioscler Thromb Vasc Biol. 2002; 22: 914–920.
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.
Oshima S, Ogawa H, Hokimoto S, Nakamura S, Noda K, Saito T, Soejima H, Takazoe K, Ishibashi F, Yasue H. Plasma monocyte chemoattractant protein-1 antigen levels and the risk of restenosis after coronary stent implantation. Jpn Circ J. 2001; 65: 261–264.
Deo R, Khera A, McGuire DK, Murphy SA, de PMNJ, Morrow DA, de Lemos JA. Association among plasma levels of monocyte chemoattractant protein-1, traditional cardiovascular risk factors, and subclinical atherosclerosis. J Am Coll Cardiol. 2004; 44: 1812–1818.
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.
Tabara Y, Kohara K, Yamamoto Y, Igase M, Nakura J, Kondo I, Miki T. Polymorphism of the monocyte chemoattractant protein (MCP-1) gene is associated with the plasma level of MCP-1 but not with carotid intima-media thickness. Hypertens Res. 2003; 26: 677–683.
Cho ML, Kim JY, Ko HJ, Kim YH, Kim WU, Cho CS, Kim HY, Hwang SY. The MCP-1 promoter-2518 polymorphism in Behcet’s disease: correlation between allele types, MCP-1 production and clinical symptoms among Korean patients. Autoimmunity. 2004; 37: 77–80.
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.
Yamada Y, Izawa H, Ichihara S, Takatsu F, Ishihara H, Hirayama H, Sone T, Tanaka M, Yokota M. Prediction of the risk of myocardial infarction from polymorphisms in candidate genes. N Engl J Med. 2002; 347: 1916–1923.
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.
Dawber TR, Kannel WB. The Framingham study: an epidemiological approach to coronary heart disease. Circulation. 1966; 34: 553–555.
Kannel WB, Feinleib M, McNamara PM, Garrison RJ, Castelli WP. An investigation of coronary heart disease in families: the Framingham Offspring Study. Am J Epidemiol. 1979; 110: 281–290.
Gabriel SB, Schaffner SF, Nguyen H, Moore JM, Roy J, Blumenstiel B, Higgins J, DeFelice M, Lochner A, Faggart M, Liu-Cordero SN, Rotimi C, Adeyemo A, Cooper R, Ward R, Lander ES, Daly MJ, Altshuler D. The structure of haplotype blocks in the human genome. Science. 2002; 296: 2225–2229.
Barrett JC, Fry B, Maller J, Daly MJ. Haploview: analysis and visualization of LD and haplotype maps. Bioinformatics. 2005; 21: 263–265.
Almasy L, Blangero J. Multipoint quantitative-trait linkage analysis in general pedigrees. Am J Hum Genet. 1998; 62: 1198–1211.
Lemeshow S, Hosmer DW Jr. Estimating odds ratios with categorically scaled covariates in multiple logistic regression analysis. Am J Epidemiol. 1984; 119: 147–151.
Schaid DJ, Rowland CM, Tines DE, Jacobson RM, Poland GA. Score tests for association between traits and haplotypes when linkage phase is ambiguous. Am J Hum Genet. 2002; 70: 425–434.
Zaykin DV, Westfall PH, Young SS, Karnoub MA, Wagner MJ, Ehm MG. Testing association of statistically inferred haplotypes with discrete and continuous traits in samples of unrelated individuals. Hum Hered. 2002; 53: 79–91.
Excoffier L, Slatkin M. Maximum-likelihood estimation of molecular haplotype frequencies in a diploid population. Mol Biol Evol. 1995; 12: 921–927.
McDermott DH, Fong AM, Yang Q, Sechler JM, Cupples LA, Merrell MN, Wilson PW, D’Agostino RB, O’Donnell CJ, Patel DD, Murphy PM. Chemokine receptor mutant CX3CR1-M280 has impaired adhesive function and correlates with protection from cardiovascular disease in humans. J Clin Invest. 2003; 111: 1241–1250.
Clinton SK, Libby P. Cytokines and growth factors in atherogenesis. Arch Pathol Lab Med. 1992; 116: 1292–1300.
Rovin BH, Lu L, Saxena R. A novel polymorphism in the MCP-1 gene regulatory region that influences MCP-1 expression. Biochem Biophys Res Commun. 1999; 259: 344–348.
Muhlbauer M, Bosserhoff AK, Hartmann A, Thasler WE, Weiss TS, Herfarth H, Lock G, Scholmerich J, Hellerbrand C. A novel MCP-1 gene polymorphism is associated with hepatic MCP-1 expression and severity of HCV-related liver disease. Gastroenterology. 2003; 125: 1085–1093.
Dupuis J, Larson MG, Vasan RS, Massaro JM, Wilson PWF, Lipinska I, Corey D, Vita JA, Keaney JF, Benjamin EJ Genome scan of systemic biomarkers of vascular inflammation in the Framingham Heart Study: evidence for susceptibility loci on 1q. Atherosclerosis. March 24, 2005. DOI: 10.1016/j.atherosclerosis.2005.02.015. Available at: http://www.sciencedirect.com. Accessed August 10, 2005.
Christiansen T, Richelsen B, Bruun JM. Monocyte chemoattractant protein-1 is produced in isolated adipocytes, associated with adiposity and reduced after weight loss in morbid obese subjects. Int J Obes Relat Metab Disord. 2005; 29: 146–150.(David H. McDermott, MD; Q)
Department of Biostatistics, School of Public Health (Q.Y., L.A.C.), Department of Mathematics (J.M.M., M.G.L.), and School of Medicine (J.F.K., R.S.V., E.J.B.), Boston University, Boston, Mass
Cardiology Division, Massachusetts General Hospital (S.K., C.J.O.) and Department of Genetics (J.N.H.), Harvard Medical School, Boston, Mass
Broad Institute, Cambridge, Mass (S.K., C.J.O.)
National Heart, Lung and Blood Institute, National Institutes of Health, Bethesda, Md (C.J.O.)
National Heart, Lung and Blood Institute’s Framingham Heart Study, Framingham, Mass (S.K., R.S.V., C.J.O., E.J.B.).
Abstract
Background— Monocyte chemoattractant protein-1 (MCP-1) is a chemokine strongly implicated in promoting atherosclerosis in animal models, but human genetic evidence is contradictory.
Methods and Results— We analyzed the association of genetic variation in the MCP-1 gene (CCL2) with prevalent myocardial infarction and serum MCP-1 levels in the community-based Framingham Heart Study Offspring Cohort (50% women; mean age, 62 years). MCP-1 levels and CCL2 genotypes were determined in 3236 and 1797 individuals, respectively. Significant clinical correlates of MCP-1 levels were age, cigarette smoking, triglycerides, body mass index, and waist-to-hip ratio. The MCP-1-2578G allele located in the CCL2 regulatory region was significantly associated with both higher serum MCP-1 levels in a recessive genetic model (358±10 versus 328±3 pg/mL; P=0.002) and higher prevalence of myocardial infarction in a dominant genetic model (adjusted odds ratio, 2.0; 95% CI, 1.2 to 3.3; P=0.005). We also defined the linkage disequilibrium structure at the CCL2 locus and observed 6 common haplotypes in whites. We performed haplotype-based association analysis and found that only the most frequent haplotype, defined by the MCP-1-2578G allele, was associated with prevalent MI.
Conclusions— Our data are consistent with the hypothesis that MCP-1 is involved in the pathogenesis of human atherosclerosis and myocardial infarction.
Key Words: epidemiology ; genetics ; inflammation ; myocardial infarction ; risk factors
Introduction
Atherosclerosis is now recognized as a disease of arterial inflammation that arises from the interactions of migratory leukocytes with resident vascular endothelial cells, smooth muscle cells, and fibroblasts.1 At the molecular level, interactions among these cell types are regulated by cytokines, adhesion molecules, and chemoattractants. Chemokines are a large family of chemoattractants that direct migration of leukocytes from the blood to sites of inflammation.2
Several lines of evidence suggest that the chemokine, monocyte chemoattractant protein-1 (MCP-1; gene name, CCL2, previously SCYA2) and its receptor CCR2 (gene name, CCR2) are involved in atherosclerosis.3 In various murine models of atherosclerosis (apolipoprotein E and LDL receptor knockouts, apolipoprotein B transgenics), deletion of CCL2 or CCR2 results in large (50% to 80%) reductions in atherosclerotic plaque size.4–6 Conversely, overexpression of MCP-1 in the leukocytes of susceptible mice results in increased plaque size.7 Furthermore, in animal models, temporary blockade of CCR2 by antibodies or gene therapy blocks restenosis after balloon angioplasty and stabilizes and reduces atherosclerotic lesion size.8–10
MCP-1 is expressed in human atherosclerotic lesions, and CCR2 is expressed on leukocytes.11–13 In addition, MCP-1 induces arrest and transmigration from the circulation of CCR2+ monocytes under conditions of physiological shear force and promotes monocyte differentiation to lipid-laden macrophages.14,15 MCP-1 also contributes to the proliferation of arterial smooth muscle cells,16 which, along with macrophages, constitute the key cellular components of atherosclerotic plaques.
A growing number of human epidemiological studies have suggested links between circulating MCP-1 levels and atherosclerosis. Higher MCP-1 levels have been associated with increased risks of myocardial infarction (MI), sudden death, coronary angioplasty, and stent restenosis.17–20 However, little is known about the role of MCP-1 levels in cardiovascular disease (CVD) in the general population.
Genetic variation of CCL2 and CCR2 could modulate MCP-1 production and function. In fact, 3 studies have associated the CCL2 distal regulatory region single nucleotide polymorphism (SNP) MCP-1-2578G with increased circulating levels of MCP-1.21–23 However, genetic associations reported between MCP-1-2578G and CVD end points have not been consistent. Two case-control studies have reported an association between MCP-1-2578G and atherosclerosis in whites. The first reported an increased frequency of MCP-1-2578G homozygotes (odds ratio [OR], 2.2; P<0.005) among individuals referred for CABG compared with age- and sex-matched controls (n=638).24 The second found an increased frequency of the MCP-1-2578G allele among HIV patients with subclinical atherosclerosis identified by ultrasound (OR, 5.7; P=0.004) compared with those lacking the allele (n=183).25 However, a third study was unable to find an association between the same polymorphism and MI in a Japanese cohort (n=909),26 and a fourth study found no significant association of this allele with angiographically determined coronary artery disease (CAD) in whites (n=3224).27 The inconsistency of these results could be due in part to the case-control design of these studies or to reliance on the testing of a single polymorphism. To address these issues, we have conducted an analysis of CCL2 genetic variation and its relation to serum levels of MCP-1 and prevalent MI in the Framingham Heart Study (FHS) Offspring Cohort, a large, longitudinal, observational, community-based study of the epidemiology of CVD.
Methods
FHS Offspring Cohort Subjects
The design of the FHS has previously been described.28,29 Briefly, the FHS enrolled 5124 children of the original participants and their spouses, referred to as the Offspring Cohort, beginning in 1971. Participants undergo a routine comprehensive medical examination every 4 years. At the cohort’s seventh examination cycle, 3236 attendees had serum MCP-1 levels (henceforth referred to as MCP-1 levels) measured. The FHS Offspring Cohort contained 1888 unrelated participants for whom DNA was available in 1995 to 1998. Of these, 1797 gave informed consent for genotyping, had clinical and genotype information available, and were included in the genotype-phenotype analyses; MCP-1 levels were performed in 1602 genotyped participants. The study was approved by the National Institute of Allergy and Infectious Diseases and Boston University School of Medicine Institutional Review boards, and all participants signed informed consent.
Clinical Risk Factors and Cardiovascular Outcomes
Outcomes were evaluated up to the time of the seventh examination (2002, the last year for which data were complete) by investigators blinded to genotypic data. MI was defined as either hospitalization for acute MI (95%) or as new ECG evidence consistent with a clinically unrecognized MI (5%). The occurrence of MI was determined by a 3-physician end-point committee that reviewed records from routine study examinations outside physicians’ offices and hospitalizations. Of the subjects whose CCL2 genotypes had been determined, 107 (6%) had prevalent MI at the time of the seventh examination. CVD risk factors, including serum total cholesterol, HDL, and triglycerides, were classified at the seventh examination cycle. Smoking was defined as cigarette use within the prior year. Diabetes was defined as fasting blood glucose level 7.0 nmol/L (126 mg/dL) or treatment with hypoglycemic agents. Hypertension was defined as systolic or diastolic blood pressure 140/90 mm Hg or antihypertensive medication treatment. Use of medications in the statin or ;-blocker classes was also evaluated as 2 separate dichotomous variables (yes/no) at the time of the seventh examination cycle.
Measurement of Serum MCP-1 Levels
Fasting serum samples were collected at the seventh examination and stored at –70°C. For analysis, samples were thawed and vortexed vigorously, and MCP-1 levels (pg/mL) were measured in duplicate according to the manufacturer’s instructions with a commercially available ELISA (R&D Systems). All samples with an intra-assay coefficient of variation above the 95th percentile were repeated; the mean of repeated duplicate values was used. Initially, samples had an average coefficient of variation of 6.1±3.3%; samples with a coefficient of variation >10.9% were rerun. However, we improved our pipetting technique. For the final 43% of samples, we achieved an average coefficient of variation of 1.9±1.6%; samples with a coefficient of variation >5.3% were rerun.
Haplotype Block Definition and Selection of Haplotype Tag SNPs
From the public dbSNP database (http://www.ncbi.nlm.nih.gov/SNP), we selected 24 evenly spaced markers within a 10.6-kb region spanning the CCL2 gene (GenBank NM_002982). The SNPs were genotyped in a panel of 12 multigenerational white family pedigrees from the Centre d’Etude du Polymorphisme Humain (CEPH) Utah panel (Coriell Institute for Medical Research, Camden, NJ).30 These reference pedigrees included 93 individuals representing 96 independent chromosomes of European ancestry. Assays were considered successful if they met the following criteria: at least 75% success for genotyping calls, Hardy-Weinberg equilibrium P>0.01, and Mendelian transmission errors 1. In addition, we imposed a minor allele frequency threshold and defined "common" for the present study as a minor allele frequency 5%. Overall, we developed successful assays for 9 SNPs.
Haplotype blocks were defined through the use of the criteria of Gabriel et al,30 and tag SNPs were selected in the publicly available Haploview software package, version 2.03.31 For each pair of markers, we calculated D', an estimate of the strength of linkage disequilibrium [LD], and a logarithm of the odds (LOD) score, which is an estimate of LD significance. From these 2 measures, each pairwise marker comparison was categorized into 1 of 3 groups: (1) no or minimal evidence of historical recombination (D'=1/LOD >2 or 0.5
Genotyping
CCL2 genotyping was performed on coded DNA samples by laboratory personnel blinded to clinical information. The genotyped variants were named to reflect the distance in nucleotides from the translational start point in GenBank D26087 and their unique identifiers in the NCBI SNP database (dbSNP; http://www.ncbi.nlm.nih.gov/SNP). We used previously published methods for the MCP-1-2578 and MCP-1-2136 polymorphisms (also previously called MCP-1-2518 and MCP-1-2076).21 In addition to these 2 previously studied SNPs, 5 tag SNPs were genotyped with commercially available (Sequenom) matrix-associated laser desorption ionization time-of-flight mass spectrometry technology and previously published protocols.30
Statistical Analysis
MCP-1 levels were logarithmically transformed before analysis because they were not normally distributed. Covariate selection for logarithmically transformed MCP-1 levels was conducted with a SOLAR regression model to account for correlation among members of the same family. In the SOLAR regression model, the correlation between any pair of relatives was assumed to be proportional to their kinship coefficient.32 The significance of each covariate was evaluated after adjustment for all the other factors in a likelihood ratio test. Covariates with a value of P<0.05 were kept in the final model (age and sex were forced in), selecting from the following: ratio of total to HDL cholesterol, triglycerides, systolic and diastolic blood pressures, body mass index, waist-to-hip ratio, fasting glucose, diabetes, hypertension treatment, lipid treatment, smoking status, and alcohol consumption. Regression coefficients were expressed by presence (versus absence) for dichotomous variables and per-unit increase for continuous variables. Unadjusted analyses of individual CCL2 genotypes with prevalent MI were performed by cross-tabulation and calculation of ORs. Multivariable-adjusted logistic regression analyses were conducted to examine the partial contribution of CCL2 genotype to risk of prevalent MI (SAS, version 8, SAS Institute Inc).33 The OR and 95% confidence limits were calculated with this model; significance was evaluated based on Wald 2 tests. The association between established risk factors and CCL2 genotypes was examined by a 2 test with 1 df for dichotomous measurements and by testing the differences in means of the risk factors between CCL2 genotype groups with a 2-tailed Student t test for continuous measurements. Tests of interaction between CCL2 genotypes and the established risk factors were also performed using logistic regression, in which 2-df Wald 2 tests were performed for the interactions between the CCL2 genotype and age, sex, body mass index, hypertension, diabetes, and smoking. Haplotype-based association analyses were conducted using a weighted regression approach as implemented in the haplo.score program.34,35 Haplotype frequency was estimated with the expectation-maximization algorithm.36 All compatible haplotype configurations of a multimarker genotype were used in the regression, with weights being the corresponding likelihood of such configuration. A global score statistic tested all haplotypes simultaneously to detect any departure from the null hypothesis of no association.
Results
We assessed MCP-1 levels in 3236 participants and genotyped 7 common CCL2 SNPs spanning the gene (Figure 1) in 1797 unrelated participants of the FHS. Table 1 displays the clinical characteristics of participants with MCP-1 levels and those with genotypes.
Clinical Correlates of MCP-1 Levels
In stepwise linear regression models, the clinical correlates of MCP-1 levels were age, cigarette smoking, triglycerides, body mass index, and waist-to-hip ratio (Table 2). Together, these covariates explained only 6% of the variability in MCP-1 levels. MCP-1 levels were not associated with prevalent MI in sex-specific or sex-pooled multivariable models (OR, 1.0; 95% CI, 0.99 to 1.00; P=0.50).
CCL2 Genotypes in the FHS Offspring Cohort
We found the MCP-1-2578 and MCP-1-2136 allele frequencies to be similar to those reported previously (30% and 21%, respectively, in a previous study versus 27% and 21% in the present study).21 Also consistent with previous studies, in our subjects, these 2 polymorphisms were in complete LD with each other. The genotype frequencies for each allele considered separately were consistent with Hardy-Weinberg predictions.
MCP-1 Genotypes and Serum MCP-1 Levels
A previous association has been demonstrated between MCP-1-2578G and unstimulated MCP-1 levels.21–23 Therefore, we examined the association in participants with available MCP-1 levels and genotypes (n=1602, 89% of those genotyped). We found a significant positive correlation of MCP-1 levels with both the MCP-1-2578G allele in a recessive genetic model and the MCP-1-2136T allele in a dominant genetic model (Figure 2). After multivariable regression adjusting for variables found to be significantly associated with MCP-1 level (Table 2), both associations remained significant (P=0.002 and P0.0001, respectively).
CCL2 Genotypes and Prevalent MI
In a dominant genetic model, possession of an MCP-1-2578G allele was marginally associated with increased MI prevalence in the cohort as a whole (unadjusted OR, 1.5; 95% CI, 1.0 to 2.2; P=0.06; Table 3). In men, the association appeared stronger (unadjusted OR, 1.8; 95% CI, 1.2 to 2.9; P=0.009). There were no significant differences in established risk factors in those with or without an MCP-1-2578G allele. After multivariable adjustment for the covariates listed in Table 3, the MCP-1-2578G allele was associated with prevalent MI in the sex-pooled cohort (OR, 2.0; 95% CI, 1.2 to 3.3; P=0.005). In tests for effect modification, there were no significant interactions between age, sex, diabetes mellitus, cigarette smoking, or hypertension and either of the 2 polymorphisms with regard to MI prevalence. The multivariable-adjusted OR was strong in men (OR, 2.6; 95% CI, 1.5 to 4.5; P=0.0006), whereas the association with increased prevalent MI was not found in women (OR, 0.7; 95% CI, 0.2 to 2.2; P=0.50). However, very few women had prevalent MI (n=23), and there was insufficient power (0.60) to exclude a significant association in women compared with men.
In multivariable models incorporating MCP-1 levels as a covariate, possession of the MCP-1-2578G allele continued to be significantly associated with increased MI prevalence (OR, 1.9; 95% CI, 1.2 to 3.2; P=0.009; Table 3). Hence, measurement of unstimulated MCP-1 level at 1 time point did not fully account for the association between MCP-1-2578G genotype and MI prevalence.
CCL2 Haplotype Analysis
Discussion
This study provides evidence from a large, community-based cohort supporting the hypothesis that the chemokine MCP-1 is a pathogenic factor in human CVD. First, we found that homozygotes for the MCP-1-2578G allele had higher basal unstimulated MCP-1 levels. This result is consistent with previous findings in much smaller studies that the MCP-1-2578G allele is associated with increased circulating levels of MCP-1.21–23 Second, we made the novel observation that individuals possessing the MCP-1-2136T allele also had higher MCP-1 levels. Both of these genetic associations remained highly significant after adjustment for other clinical factors associated with MCP-1 levels. Third, possession of the MCP-1-2578G allele was associated with an increased risk of prevalent MI. This result is consistent with those of 2 previously published case-control studies that reported an association of the MCP-1-2578G allele with an increased risk of atherosclerosis.24,25 Furthermore, we have performed the first comprehensive analysis of the common genetic variation of CCL2 in a large community-based population and identified novel variants that are in strong LD with MCP-1-2578G (MCP-1-2835A) and MCP-1-2136T (MCP-1+764G).
The MCP-1-2578G association with MI persisted after adjustment for other well-established risk factors and was similar in magnitude to that previously reported for angiographically proven CAD.24 Moreover, the magnitude of increased risk associated with possession of the MCP-1-2578G allele was similar to the magnitude of risk augmentation attributable to established MI risk factors. For example, in the FHS Offspring Cohort, the multivariable-adjusted ORs for prevalent MI were 2.5 for diabetes, 2.2 for recent smoking, and 4.9 for male sex compared with 2.0, the OR for possession of the MCP-1-2578G allele compared with all others in this cohort (Table 3). Together, these data suggest a pathogenetic role for MCP-1 in human atherosclerotic CVD.
The exact mechanism by which the MCP-1-2578G allele might increase MI risk remains uncertain; however, presumably locally increased MCP-1 levels in the coronary arteries would promote transmigration of CCR2+ leukocytes from the blood and their organization and activation in plaque. Consistent with this model, MCP-1, CCR2, and activated macrophages are found in atherosclerotic lesions.11–13 In this model of pathogenesis, the interaction between MCP-1 and CCR2 may play several roles, but it is probably not the only factor involved in lesion formation because lesions are able to form in the absence of either of these molecules in susceptible animal models.4–6 It should be noted that CCR2 is also polymorphic in humans, and in a previous study, we investigated the association in the Framingham Offspring Cohort of an SNP of unknown functional significance in the CCR2 receptor, CCR2-64I. This SNP results in the conservative replacement of a valine at position number 64 with an isoleucine. We found a nonsignificant association of this SNP with CVD using a dominant genetic model (OR, 0.7; P=0.12) after multivariable adjustment.37 In the present study, the association of MCP-1-2578G with prevalent MI was not substantially altered by incorporating the baseline MCP-1 level as a covariate. One potential explanation is that baseline unstimulated MCP-1 levels may not reflect local tissue levels in the coronary arteries where inflammatory cytokines like interleukin-1 and tumor necrosis factor are known to be expressed.38 Previous in vitro studies have found these cytokines to have marked effects on the regulatory region where the CCL2 SNPs associated with CVD are located.39,40 In addition, the measured level of MCP-1 was done only at 1 time point, whereas longitudinal levels may be more important.
We noted that the clinical correlates of MCP-1 levels in the community were age, cigarette smoking, triglycerides, body mass index, and waist-to-hip ratio but that these factors explained only 6% of the variability in MCP-1 levels. Genetic factors may play a more important role in MCP-1 level variability. In a separate study, we have recently noted that the adjusted heritability of MCP-1 levels is 44% and found significant linkage on chromosome 1 at 186 cM (LOD, 4.27; genome-wide P=0.005).41 In the Oral Glycoprotein IIb/IIIa Inhibition with Orbofiban in Patients With Unstable Coronary Syndromes (OPUS-TIMI) 16 trial, MCP-1 levels were related to advancing age, female sex, hypertension, diabetes, and renal insufficiency.17 We suspect that de Lemos and colleagues17 observed different correlates because of the referral nature of their cohort. The relation of MCP-1 levels with 2 measures of adiposity in our cohort is of interest because of the observations by Christiansen et al42 that adipocyte MCP-1 mRNA levels correlated with circulating MCP-1 and body mass index and that weight loss was associated with a decline in circulating MCP-1 levels.
Study Strengths and Limitations
FHS is a large community-based cohort, diminishing referral biases, with routine ascertainment of CVD risk factors and events, enabling multivariable analyses and genotyping and phenotyping blinded to one another. The observational, nonrandomized nature of cardiovascular medications indicates that caution should be observed with regard to the lack of association between statins and other treatments and MCP-1 levels. Initially, we performed relatively few statistical tests to confirm the 2 previously reported SNP associations.24,25 However, we augmented our analyses with haplotypes and marker levels to examine more comprehensively the relation of MCP-1 and CCL2 to risk factors and disease. Hence, we have performed multiple statistical tests that increase the possibility of false-positive associations. Furthermore, our study examined prevalent MI; an apparent association for the MCP-1-2578G allele might represent a survival bias if individuals with that allele had a lower case fatality rate than other genotypes. Because we had few prevalent MI cases in women, we were underpowered to find an association between the CCL2 genotypes and MI in women. Additional follow-up of our sample or investigation of larger samples of women will help to clarify whether the risk of MI associated with CCL2 genotype varies by gender. Our study sample was largely middle-aged to elderly and white; the findings may not be generalizable to younger individuals or other racial groups. However, the single-race and community-based structure of the cohort minimizes the possibility that the associations reported here were due to population stratification. Although our results are consistent with 2 previous studies relating MCP-1-2578G to atherosclerosis phenotypes (CAD and carotid intima-media thickness),24,25 our report is the first to relate MCP-1-2578G to the MI phenotype.
Because of the tight LD between MCP-1-2835A and MCP-1-2578G on haplotype H1 and similarly between MCP-1-2136T and MCP-1+764G on haplotype H2, it is very difficult to separate their influence in our genetic association study. Association studies in populations with greater haplotypic diversity (eg, populations of African ancestry) and functional studies involving the newly identified sites (MCP-1-2835A for MCP-1 levels/MI and MCP-1+764G for MCP-1 levels) may help to clarify which, if any, affects production of MCP-1. In addition, more comprehensive resequencing is needed to look for other functional sites in the relevant haplotypes. Whereas it remains possible that the association is due to linkage with another as-yet unknown polymorphism, the evidence that MCP-1-2578G is a functional polymorphism using in vitro promoter assays supports the hypothesis that there is a functional role for this SNP.39 Electromobility gel shift assays have also shown allele-specific nuclear factor binding to oligonucleotide probes corresponding to the MCP-1-2578G site from TNF-stimulated hepatic primary cells and MG-63 osteosarcoma cells.21,40 The identity of these induced nuclear factors is not yet completely known but at least in MG-63 cells appeared to contain interferon regulatory factor-1.21
Research and Clinical Implications
Our data suggest several future research directions such as the investigation of the identity of the functional SNP(s) in these haplotypes, whether these SNPs act alone or instead require interaction with other SNPs, and whether transcription factors bind at these polymorphic sites in leukocytes and endothelial cells. In addition, the CCL2 polymorphisms we examined may be generally useful as genetic probes to evaluate the potential role of MCP-1 in the pathogenesis of other inflammatory diseases in humans.
We observed that the MCP-1-2578G allele is dominantly associated with increased risk of MI in humans after adjustment for other CVD risk factors. Our results provide additional support for the inflammation hypothesis of atherosclerosis pathogenesis and for further study to determine the role of MCP-1 as a proinflammatory risk factor. Identifying inflammatory polymorphisms contributing to atherosclerosis will improve our understanding of the pathogenesis of CVD and may suggest targets for improved prevention and treatment strategies.
Acknowledgments
This work was supported by the Framingham Study (N01-HC 25195) and CardioGenomics Program in Genomics Applications (U01 HL-66582, RO1 HL-076784, and RO1 HL-64753).
References
Ross R. Atherosclerosis: an inflammatory disease. N Engl J Med. 1999; 340: 115–126.
Dong VM, McDermott DH, Abdi R. Chemokines and diseases. Eur J Dermatol. 2003; 13: 224–230.
Ikeda U, Matsui K, Murakami Y, Shimada K. Monocyte chemoattractant protein-1 and coronary artery disease. Clin Cardiol. 2002; 25: 143–147.
Boring L, Gosling J, Cleary M, Charo IF. Decreased lesion formation in CCR2–/– mice reveals a role for chemokines in the initiation of atherosclerosis. Nature. 1998; 394: 894–897.
Gu L, Okada Y, Clinton SK, Gerard C, Sukhova GK, Libby P, Rollins BJ. Absence of monocyte chemoattractant protein-1 reduces atherosclerosis in low density lipoprotein receptor-deficient mice. Mol Cell. 1998; 2: 275–281.
Dawson TC, Kuziel WA, Osahar TA, Maeda N. Absence of CC chemokine receptor-2 reduces atherosclerosis in apolipoprotein E-deficient mice. Atherosclerosis. 1999; 143: 205–211.
Aiello RJ, Bourassa PA, Lindsey S, Weng W, Natoli E, Rollins BJ, Milos PM. Monocyte chemoattractant protein-1 accelerates atherosclerosis in apolipoprotein E-deficient mice. Arterioscler Thromb Vasc Biol. 1999; 19: 1518–1525.
Horvath C, Welt FG, Nedelman M, Rao P, Rogers C. Targeting CCR2 or CD18 inhibits experimental in-stent restenosis in primates: inhibitory potential depends on type of injury and leukocytes targeted. Circ Res. 2002; 90: 488–494.
Egashira K, Zhao Q, Kataoka C, Ohtani K, Usui M, Charo IF, Nishida K, Inoue S, Katoh M, Ichiki T, Takeshita A. Importance of monocyte chemoattractant protein-1 pathway in neointimal hyperplasia after periarterial injury in mice and monkeys. Circ Res. 2002; 90: 1167–1172.
Inoue S, Egashira K, Ni W, Kitamoto S, Usui M, Otani K, Ishibashi M, Hiasa K, Nishida K, Takeshita A. Anti-monocyte chemoattractant protein-1 gene therapy limits progression and destabilization of established atherosclerosis in apolipoprotein E–knockout mice. Circulation. 2002; 106: 2700–2706.
Yla-Herttuala S, Lipton BA, Rosenfeld ME, Sarkioja T, Yoshimura T, Leonard EJ, Witztum JL, Steinberg D. Expression of monocyte chemoattractant protein 1 in macrophage-rich areas of human and rabbit atherosclerotic lesions. Proc Natl Acad Sci U S A. 1991; 88: 5252–5256.
Nelken NA, Coughlin SR, Gordon D, Wilcox JN. Monocyte chemoattractant protein-1 in human atheromatous plaques. J Clin Invest. 1991; 88: 1121–1127.
Takeya M, Yoshimura T, Leonard EJ, Takahashi K. Detection of monocyte chemoattractant protein-1 in human atherosclerotic lesions by an anti-monocyte chemoattractant protein-1 monoclonal antibody. Hum Pathol. 1993; 24: 534–539.
Gerszten RE, Garcia-Zepeda EA, Lim YC, Yoshida M, Ding HA, Gimbrone MA Jr, Luster AD, Luscinskas FW, Rosenzweig A. MCP-1 and IL-8 trigger firm adhesion of monocytes to vascular endothelium under flow conditions. Nature. 1999; 398: 718–723.
Tabata T, Mine S, Kawahara C, Okada Y, Tanaka Y. Monocyte chemoattractant protein-1 induces scavenger receptor expression and monocyte differentiation into foam cells. Biochem Biophys Res Commun. 2003; 305: 380–385.
Viedt C, Vogel J, Athanasiou T, Shen W, Orth SR, Kubler W, Kreuzer J. Monocyte chemoattractant protein-1 induces proliferation and interleukin-6 production in human smooth muscle cells by differential activation of nuclear factor-B and activator protein-1. Arterioscler Thromb Vasc Biol. 2002; 22: 914–920.
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.
Oshima S, Ogawa H, Hokimoto S, Nakamura S, Noda K, Saito T, Soejima H, Takazoe K, Ishibashi F, Yasue H. Plasma monocyte chemoattractant protein-1 antigen levels and the risk of restenosis after coronary stent implantation. Jpn Circ J. 2001; 65: 261–264.
Deo R, Khera A, McGuire DK, Murphy SA, de PMNJ, Morrow DA, de Lemos JA. Association among plasma levels of monocyte chemoattractant protein-1, traditional cardiovascular risk factors, and subclinical atherosclerosis. J Am Coll Cardiol. 2004; 44: 1812–1818.
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.
Tabara Y, Kohara K, Yamamoto Y, Igase M, Nakura J, Kondo I, Miki T. Polymorphism of the monocyte chemoattractant protein (MCP-1) gene is associated with the plasma level of MCP-1 but not with carotid intima-media thickness. Hypertens Res. 2003; 26: 677–683.
Cho ML, Kim JY, Ko HJ, Kim YH, Kim WU, Cho CS, Kim HY, Hwang SY. The MCP-1 promoter-2518 polymorphism in Behcet’s disease: correlation between allele types, MCP-1 production and clinical symptoms among Korean patients. Autoimmunity. 2004; 37: 77–80.
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.
Yamada Y, Izawa H, Ichihara S, Takatsu F, Ishihara H, Hirayama H, Sone T, Tanaka M, Yokota M. Prediction of the risk of myocardial infarction from polymorphisms in candidate genes. N Engl J Med. 2002; 347: 1916–1923.
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.
Dawber TR, Kannel WB. The Framingham study: an epidemiological approach to coronary heart disease. Circulation. 1966; 34: 553–555.
Kannel WB, Feinleib M, McNamara PM, Garrison RJ, Castelli WP. An investigation of coronary heart disease in families: the Framingham Offspring Study. Am J Epidemiol. 1979; 110: 281–290.
Gabriel SB, Schaffner SF, Nguyen H, Moore JM, Roy J, Blumenstiel B, Higgins J, DeFelice M, Lochner A, Faggart M, Liu-Cordero SN, Rotimi C, Adeyemo A, Cooper R, Ward R, Lander ES, Daly MJ, Altshuler D. The structure of haplotype blocks in the human genome. Science. 2002; 296: 2225–2229.
Barrett JC, Fry B, Maller J, Daly MJ. Haploview: analysis and visualization of LD and haplotype maps. Bioinformatics. 2005; 21: 263–265.
Almasy L, Blangero J. Multipoint quantitative-trait linkage analysis in general pedigrees. Am J Hum Genet. 1998; 62: 1198–1211.
Lemeshow S, Hosmer DW Jr. Estimating odds ratios with categorically scaled covariates in multiple logistic regression analysis. Am J Epidemiol. 1984; 119: 147–151.
Schaid DJ, Rowland CM, Tines DE, Jacobson RM, Poland GA. Score tests for association between traits and haplotypes when linkage phase is ambiguous. Am J Hum Genet. 2002; 70: 425–434.
Zaykin DV, Westfall PH, Young SS, Karnoub MA, Wagner MJ, Ehm MG. Testing association of statistically inferred haplotypes with discrete and continuous traits in samples of unrelated individuals. Hum Hered. 2002; 53: 79–91.
Excoffier L, Slatkin M. Maximum-likelihood estimation of molecular haplotype frequencies in a diploid population. Mol Biol Evol. 1995; 12: 921–927.
McDermott DH, Fong AM, Yang Q, Sechler JM, Cupples LA, Merrell MN, Wilson PW, D’Agostino RB, O’Donnell CJ, Patel DD, Murphy PM. Chemokine receptor mutant CX3CR1-M280 has impaired adhesive function and correlates with protection from cardiovascular disease in humans. J Clin Invest. 2003; 111: 1241–1250.
Clinton SK, Libby P. Cytokines and growth factors in atherogenesis. Arch Pathol Lab Med. 1992; 116: 1292–1300.
Rovin BH, Lu L, Saxena R. A novel polymorphism in the MCP-1 gene regulatory region that influences MCP-1 expression. Biochem Biophys Res Commun. 1999; 259: 344–348.
Muhlbauer M, Bosserhoff AK, Hartmann A, Thasler WE, Weiss TS, Herfarth H, Lock G, Scholmerich J, Hellerbrand C. A novel MCP-1 gene polymorphism is associated with hepatic MCP-1 expression and severity of HCV-related liver disease. Gastroenterology. 2003; 125: 1085–1093.
Dupuis J, Larson MG, Vasan RS, Massaro JM, Wilson PWF, Lipinska I, Corey D, Vita JA, Keaney JF, Benjamin EJ Genome scan of systemic biomarkers of vascular inflammation in the Framingham Heart Study: evidence for susceptibility loci on 1q. Atherosclerosis. March 24, 2005. DOI: 10.1016/j.atherosclerosis.2005.02.015. Available at: http://www.sciencedirect.com. Accessed August 10, 2005.
Christiansen T, Richelsen B, Bruun JM. Monocyte chemoattractant protein-1 is produced in isolated adipocytes, associated with adiposity and reduced after weight loss in morbid obese subjects. Int J Obes Relat Metab Disord. 2005; 29: 146–150.(David H. McDermott, MD; Q)