- Email: [email protected]
Relation of polymorphism within the C-reactive protein gene and plasma CRP levels
Atherosclerosis 178 (2005) 139–145
Relation of polymorphism within the C-reactive protein gene and plasma CRP levels H. Jacqueline Suka,b,c,∗ , Paul M. Ridkera,b,c , Nancy R. Cooka,b , Robert Y.L. Zeea,b a
Center for Cardiovascular Disease Prevention, Brigham and Women’s Hospital, 75 Francis Street, Boston, MA 02215, USA b Donald W. Reynolds Center for Cardiovascular Research, Harvard Medical School, Boston, MA, USA c Division of Cardiology, Brigham and Women’s Hospital, Harvard Medical School, Boston, MA, USA Received 19 December 2003; received in revised form 15 July 2004; accepted 16 July 2004 Available online 12 October 2004
Abstract We sought to investigate whether genetic variability within the CRP gene affects CRP levels. C-reactive protein (CRP), an acute-phase reactant in inflammation, is an important predictor for cardiovascular disease. The genetic determinants of plasma CRP level remain unknown. We assessed the genotypes of two common polymorphisms within the CRP gene, an exonic 1059G > C and an intronic T > A base substitution, among 2397 participants of a community-based study; 1334 had no prior cardiovascular history, while 1063 had a prior cardiovascular history. Univariable and multivariable-adjusted analyses were performed to examine the association of CRP polymorphisms with CRP levels, modeling different modes of inheritance. The genetic polymorphisms were significantly associated with baseline CRP levels in univariable analysis (1059G > C: GG versus GC or CC genotypes, median CRP levels 0.22 versus 0.15 mg/L, P < 0.0001; intronic T > A: TT versus AT versus AA genotypes, median CRP levels 0.19 versus 0.23 versus 0.24 mg/L, P = 0.003). This relationship persisted after adjusting for age, sex, body mass index, ethnicity, hypertension, smoking, diabetes, hyperlipidemia, and aspirin use. Furthermore, these effects were present in subgroup analyses limited to those with and without prevalent coronary disease, and when assessed within Caucasians only. The present data provide evidence of a genetic component of CRP levels, independently of traditional risk factors for cardiovascular disease. Whether genetic markers can add to information yielded by high sensitivity CRP (hsCRP) in assessing cardiovascular risk needs further evaluation. © 2004 Elsevier Ireland Ltd. All rights reserved. Keywords: C-reactive protein; Inflammation; Genetics; Risk factors
1. Introduction There is strong evidence that C-reactive protein (CRP) is a powerful predictor of incident cardiovascular events at all levels of LDL-C, all levels of the Framingham risk score and at all levels of the metabolic syndrome [1–9]. Basal levels in individuals free of acute illness are reproducible [10]. What drives baseline human CRP levels is unknown, although there is evidence of substantial heritability (35–40%) in familial aggregation studies [11]. The human CRP gene lies on chromosome 1q23, within a conserved genetic region that encodes for proteins criti∗
Corresponding author. Tel.: +1 617 278 0808; fax: +1 617 232 3541. E-mail address: [email protected] (H.J. Suk).
0021-9150/$ – see front matter © 2004 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.atherosclerosis.2004.07.033
cal to the immune system and to intercellular communication [12]. Small studies have reported that polymorphisms within the CRP gene are associated with plasma CRP levels. As in all gene-association studies, such findings must be validated within large and independent sample populations to safeguard against spurious associations [13]. Many smaller studies are later disproved because of systemic differences in genetic composition within study groups. Recently, a polymorphism within the untranslated 3 region of the CRP gene [14], but not one within the promoter region of the gene was associated with plasma CRP levels in 250 study participants. A dinucleotide repeat polymorphism [15] in the intronic region of the CRP gene studied within 546 participants had high correlation with plasma CRP levels.
140
H.J. Suk et al. / Atherosclerosis 178 (2005) 139–145
We hypothesized that two genetic markers within the CRP gene would be associated with high sensitivity CRP (hsCRP) levels [16] and performed genetic analysis of two single nucleotide polymorphisms (SNPs) within the CRP gene: (1) a polymorphism within exon 2 of the CRP gene (1059G > C, NCBI SNP accession no. rs1800947 [17]) and (2) a neighboring polymorphism (intronic T > A, NCBI SNP accession no. rs1417938) among 2743 people from around the United States who provided consent for genetic analysis.
2. Methods 2.1. Study population The 2884 men and women from all states of the United States participated in a multicenter community-based study which has been previously described in detail [18]. Of these, 2743 participants provided blood samples and consent for genetic testing. 2397 participants were free of hormone replacement therapy (HRT), a therapy that elevates CRP levels [19,20]; these participants were chosen for our genetic investigation. 1334 of these participants had no history of prior myocardial infarction, stroke or coronary revascularization (primary prevention group) and 1063 had such history (secondary prevention group). Clinical data such as age, gender, race, systolic and diastolic blood pressures, body mass index, smoking status and diabetes mellitus were collected. Plasma lipid levels and high sensitivity plasma CRP levels were measured. Genomic DNA was extracted from whole blood. 2.2. Genotype determination of the 1059G > C exonic polymorphism Genotypes were determined by polymerase chain reaction-restriction fragment length polymorphism (PCRRFLP), with forward primer 5 -GCCCAGGGTGAGGAAGAGTCT-3 and reverse primer 5 -CCCGCCAGTTCAGGACATTAG-3 . The amplified product was digested with MaeIII enzyme (New England Biolabs) and size-fractionated by electrophoretic separation in 3% agarose gels. 2.3. Genotype determination of the intronic T > A polymorphism Genotyping was performed using the kinetic thermocycling approach, a fluorescent-based assay, as described by Germer et al. [21], on an ABI Prism 7900 Sequence Detection System. The forward allele specific primers for the detection of the A allele was 5 -ACCCCCATACCTC AGATCAAAA of the T allele, 5 -ACCCCC- -3 ; for detection ATACCTCAGATCAAAT--3 . The common reverse primer was 5 -AGGTAAGGGCCACCCCAG-3 . To confirm genotype assignment, two independent observers performed scoring. Disagreements (<2%) were resolved by joint review, and a repeat genotyping. PCR results
were scored blinded to all clinical and biochemical characteristics. 2.4. Statistical analysis Differences in proportions of gender, smoking status, diabetes, ethnicity, aspirin use, and prior vascular disease among different genotypes were analyzed by the χ2 -test for two groups and the χ2 trend test for multiple groups. Continuous variables with normal distribution were analyzed by t-test and ANOVA; variables with non-normal distributions were analyzed by nonparametric tests: the Wilcoxon rank sum tests for two categories, and Kruskal–Wallis tests for multiple groups. Hardy–Weinberg equilibrium for the study population was assessed by the general χ2 -test with one degree of freedom. Linkage disequilibrium was assessed by the general χ2 -test for independence. The association between baseline log-normalized hsCRP levels and CRP genotypes was further investigated by multivariable linear regression models. These models were evaluated for the relationship between hsCRP levels and genotypes after adjustment for age, gender, body mass index (BMI), ethnicity, smoking status, systolic hypertension, current aspirin use, hyperlipidemia, and prior vascular disease. Because the distribution of hsCRP was left-skewed, hsCRP was log transformed for use as the outcome variable in multivariable analysis. Effect estimates were calculated by re-exponentiation of the beta coefficients, which is equivalent to the ratio of CRP levels in the given genotype versus the homozygous wild type. For the 1059G > C genotype, the relative rarity of the CC genotype made the dominant genetic models most statistically robust. Both additive and dominant genetic modeling was considered for the intronic T > A polymorphism. SAS version 8e (SAS Institute, Cary, NC) was used for all statistical analyses and all probability values were assessed in a two-tailed fashion.
3. Results Table 1 shows the baseline characteristics of the study participants by their genotypes. The GG genotype occurred most frequently (88%) in our population for the exotic polymorphism; the CC genotype was rare (<1%) and was considered in conjunction with the heterozygous genotype. The TT genotype (49%) was most common for the intronic T > A polymorphism followed by the AT (42%) and AA (9%) genotypes. Our population had no significant deviations of genotype distributions from expected Hardy–Weinberg equilibrium, which was assessed by the general χ2 -test with one degree of freedom, within both the population with and without history of prior vascular disease. There were no significant differences in the baseline characteristics across the different genotypes in terms of sex, presence of hyperlipidemia, diabetes, known prior vascular disease or aspirin use. Furthermore, the mean values for body
Table 1 Baseline clinical characteristics and plasma CRP levels by genotypes Overall data
1059G > C exonic SNP GG (n = 2100)
GC and CC (n = 295)
TT (n = 1174)
AT (n = 988)
AA (n = 208)
Age (y) Women (%)
62.2 ± 13.3 29.6
62.3 ± 13.3 29.9
61.2 ± 12.9 27.7
0.17 0.45
61.8 ± 13.5 29.6
62.8 ± 13.1 30.2
61.9 ± 13.2 27.3
0.30 0.76
Smoking status (%) Never Former Current
40.1 45.0 14.9
40.0 45.2 14.8
40.5 43.6 15.9
0.82
42.5 43.4 14.1
37.7 46.9 15.4
36.9 45.5 17.7
0.02
Body mass index (kg/m2 ) Hyperlipidemia > 220 (mg/dL, %)
29.2 ± 5.4 47.0
29.2 ± 5.3 46.6
29.5 ± 6.1 49.5
0.32 0.35
29.3 ± 5.0 47.7
29.2 ± 5.3 45.0
29.1 ± 5.3 50.0
0.54 0.80
220.4 ± 38.2 135.6 ± 29.5 37.6 ± 10.3
220.2 ± 38.8 135.3 ± 29.7 37.7 ± 10.4
221.7 ± 33.2 137.6 ± 27.7 36.8 ± 9.4
0.49 0.22 0.12
221.4 ± 37.9 136.8 ± 29.9 37.8 ± 10.3
219.2 ± 38.2 134.4 ± 29.0 37.3 ± 10.2
218.4 ± 39.1 133.4 ± 29.1 37.5 ± 10.6
0.15 0.03 0.38
132.6 ± 16.9 79.3 ± 9.8
132.7 ± 16.8 79.3 ± 9.6
132.2 ± 18.0 79.3 ± 11.0
0.64 0.99
133.0 ± 17.4 79.6 ± 9.9
132.3 ± 16.6 79.0 ± 9.7
131.1 ± 15.4 78.8 ± 9.4
0.14 0.12
Total cholesterol (mg/dL) LDL HDL Blood pressure (mmHg) Systolic Diastolic
p
Intronic T > A SNP
p
Diabetes (%)
18.6
19.1
14.8
0.08
18.4
18.4
20.1
0.69
Ethnicity (%) White Black Hispanic Asian Other
87.1 6.9 3.7 1.7 0.6
86.3 7.7 3.8 1.7 0.5
93.1 0.7 3.1 2.4 0.7
0.0003
81.5 10.9 4.2 2.9 0.6
91.7 3.4 3.5 0.8 0.5
96.0 0.5 3.0 – 0.5
<0.0001
44.4 46.4
44.4 46.4
44.1 46.5
0.91 0.97
43.9 45.3
45.1 48.0
45.2 46.2
0.56 0.39
Known prior vascular disease (%) Aspirin use (%) CRP level (mg/L) Median (25th, 75th %)
0.21 [0.09, 0.43]
0.22 [0.10, 0.45]
0.15 [0.07, 0.32]
<0.0001
0.19 [0.09, 0.41]
0.23 [0.10, 0.46]
0.24 [0.12, 0.45]
H.J. Suk et al. / Atherosclerosis 178 (2005) 139–145
Covariates
0.003
Values are mean ± S.D., except as noted. When applicable, tests for trend are used for multiple group comparisons. General chi-square tests are used for two-group comparisons.
141
142
H.J. Suk et al. / Atherosclerosis 178 (2005) 139–145
Fig. 1. CRP levels by genotype. Median levels and 25th and 75th percentiles. P-values are from Wilcoxon rank sum and Kruskal–Wallis tests.
mass index, total cholesterol, HDL levels, systolic and diastolic blood pressures were not significantly different by genotype. Differences in smoking status as well as a slight trend toward higher LDL levels with the presence of the T allele were seen for the intronic T > A base substitution. Significant differences in genotype representation were seen across different ethnicities for both SNPs, and analysis was later repeated within the Caucasian population only. Baseline levels of CRP were significantly different among subjects with different genotypes (0.22 versus 0.15 mg/L median levels, P < 0.0001; 0.19 versus 0.23 versus 0.24 median levels, P = 0.003) as shown in Table 1 and Fig. 1. With the presence of the less frequent C allele, there is a 29% difference in CRP levels (geometric mean CRP levels 0.21 versus 0.15 mg/L, P < 0.0001). Likewise, there is a highly statistically significant incremental increase in CRP levels with the presence of the less frequent A allele (P = 0.001). The alleles of these two SNPs, furthermore, are in linkage disequilibrium, consistent with a conserved relationship between the G and A alleles (χ2 = 51.9; P < 0.0001 by the general χ2 -test for independence).
Table 2 shows the prevalence of the participants’ genotypes by quartiles of baseline CRP values. As expected, 91% of the participants with CRP levels within the highest quartile of CRP levels had the GG genotype, compared with 83% in the lowest quartile (P-trend < 0.0001). The 47% within the highest quartile of CRP levels had the TT genotype, compared with 55% in the lowest quartile (P-trend = 0.001). These genetic effects on CRP levels persisted in multivariable analysis, after controlling for clinical characteristics, chosen because of their importance as traditional risk factors for vascular disease. For the 1059G > C SNP, the effect estimates of multivariable linear regression, which in linear regression correspond to residual ratios of CRP levels after control for other clinical variables, were determined. These effect estimates will be referred to as “CRP ratios”. After control for age, gender, body mass index, race, smoking status, diabetes, hypertension, aspirin use, hyperlipidemia and history of vascular disease, the CRP ratio of the GC or CC genotype compared with the GG referent genotype was 0.71 (P < 0.0001, Table 3). Similarly, for the intronic T > A SNP, the effect estimates for the AT genotype and the AA genotype
Table 2 Prevalence of genotypes in patients by quartiles of baseline CRP levels Quartile 1 of CRP levels (n, %)
Quartile 2 of CRP levels (n, %)
Quartile 3 of CRP levels (n, %)
Quartile 4 of CRP levels (n, %)
1059G > C exonic SNP GC GC or CC
497 (83%) 102 (17%)
518 (87%) 80 (13%)
539 (90%) 62 (10%)
546 (91%) 51 (9%)
Intronic T >A SNP TT AT AA
324 (55%) 229 (39%) 38 (6%)
295 (50%) 240 (41%) 53 (9%)
277 (46%) 261 (44%) 60 (10%)
278 (47%) 258 (44%) 57 (10%)
∗
p for trends* <0.0001
0.001
Chi-square test for trend was used to compare the prevalence of genotypes among quartile of CRP values.
H.J. Suk et al. / Atherosclerosis 178 (2005) 139–145
143
Table 3 CRP levels within each genotype and effect estimates in multivariable analysis after adjustment for vascular risk factorsa CRP levels geometric means (mg/L)
Confidence intervals
1059G > C exonic SNP Crude GG GC or CC
0.21 0.15
[0.20, 0.22] [0.13, 0.17]
Adjusted Dominant model GG GC or CC
0.21 0.15
[0.20, 0.22] [0.13, 0.17]
0.18 0.22
[0.17, 0.20] [0.20, 0.23]
0.18 0.21 0.24
[0.17, 0.19] [0.20, 0.23] [0.21, 0.28]
Intronic T > A SNP Crude TT AT or AA Adjusted Additive model TT AT AA
p
Effect estimatesb
Confidence interval
p
1 0.72
[0.62, 0.83]
0.0001
1 0.71
[0.63, 0.81]
0.0001
1 1.18
[1.08, 1.30]
0.0004
1 1.17 1.31
[1.07, 1.28] [1.11, 1.53]
0.0008 0.001
<0.0001
<0.0001
0.001
0.0002
a Values are adjusted for age, gender, BMI, ethnicity, smoking status, systolic hypertension, current aspirin use, diabetes, hyperlipidemia, history of prior MI, stroke and any other prior vascular disease. b Effect estimates are equivalent to CRP ratios after adjustment for clinical variables. The CRP ratios and 95% CIs were calculated by multivariable linear regression analyses using log of CRP values as outcome, and then re-exponentiating the resulting beta coefficient values.
compared to the referent TT genotype were 1.17 and 1.31 (P = 0.0008, 0.001), respectively. The effect estimates increase in incremental fashion. By dominant genetic modeling, the CRP ratio for the AT or AA genotype compared to the referent AA genotype was 1.19 (P = 0.0001). CRP geometric means are also shown in Table 3 for each genotype, assessed by genetic models and are concordant with the trends presented in Table 1. In addition, because of a difference seen in genotype distributions among participants of different ethnic descent (Table 1), and of potential unexplained confounding by ethnicity, we evaluated the relationships within Caucasian participants only, again free of HRT. The originally described relationships are again seen. For the 1059G > C SNP, the CRP ratios for the GC or CC genotype compared with the GG referent genotype was 0.70 (P < 0.0001), and for the intronic T > A SNP, the ratios for the AT genotype and the AA genotype compared to the referent TT genotype were 1.15 and 1.28 (P = 0.005, 0.003) by additive genetic modeling, after adjusting for the same traditional risk factors for vascular disease. By dominant genetic modeling, the effect estimates of the AT or AA genotype compared to the referent TT genotype was 1.17 (P = 0.0008). We sought to ensure that this relationship persisted both within the primary prevention and secondary prevention cohorts. Within the primary prevention cohort, for the 1059G > C SNP, the CRP ratio for the GC or CC genotype compared with the referent GG genotype was 0.76 (P = 0.001), and for the intronic T > A SNP, the effect estimates for the AT
genotype and the AA genotype compared to the referent TT genotype were 1.18 and 1.26 (P = 0.005, 0.03) by additive genetic modeling, after adjusting for the same traditional risk factors for vascular disease. Similar trends are seen within the secondary prevention cohort. The CRP ratio of the GC or CC genotype compared with the referent GG genotype was 0.66 (P = 0.0001), and for the intronic T > A SNP, the CRP ratios of the AT genotype and the AA genotype compared to the referent TT genotype were 1.15 and 1.36 (P = 0.06, 0.02) by additive genetic modeling, after adjusting for the same traditional risk factors for vascular disease. In summary, whether the CRP polymorphisms were modeled assuming dominant inheritance for the 1059G > C polymorphism or additive and dominant inheritance for intronic T > A polymorphism, both were strongly and independently associated with CRP levels.
4. Discussion CRP is an acute-phase pentraxin protein produced primarily by the liver in response to production of cytokines such as interleukin-6, interleukin-1 and tumor necrosis factor-␣ during tissue injury or inflammation. The reproducibility and measurability of basal hsCRP levels in people free of acute illness have spawned great clinical interest. Whether genetic determinants of basal CRP levels would provide even more stable risk markers is an intriguing even if yet unproven hypothesis.
144
H.J. Suk et al. / Atherosclerosis 178 (2005) 139–145
The two polymorphisms of the CRP gene studied in this large-scale study were highly associated with CRP levels. CRP levels were significantly higher in participants with the G allele than those with the C allele for the 1059G > C polymorphism, in concordance with the trends previously observed within men only in the Physicians’ Health Study [22]. In addition, we found that hsCRP levels were incrementally higher with additional presence of the A allele for the intronic T > A polymorphism: highest for the AA genotype, followed by the AT genotype and then the TT genotype for the intronic T > A polymorphism. These associations persisted strongly even after controlling for traditional cardiovascular risk factors. The association of these polymorphisms with hsCRP levels also persists when reassessed among Caucasians only, the most represented group in the study. The CRP gene is lies on chromosome 1q23 and is composed of two exons and an intron. While many polymorphisms have been recently reported within the CRP gene, none, to our knowledge, are known to encode for nonsynonymous amino acid change. Several points about the genetic variants that we studied merit mention. The intronic SNP is located only 29 base pairs away from a 5 intron/exon splice site. The 1059G > C SNP in exon 2 results in a codon change from CTC to CTG, both of which encode for leucine. While this is a synonymous change, the CTG codon occurs with 20% less frequency in humans. Particularly low frequency codons do raise questions about DNA stability during translation, but we caution about drawing mechanistic conclusions from primary DNA sequences. The intronic and exonic SNP described may simply be in linkage disequilibrium with other functional variants or have subtle, yet to be discovered, genetic effects. Until further work clarifies potential mechanisms, furthermore, it is prudent to interpret such polymorphisms as genetic markers, potentially in high linkage disequilibrium with causative genetic mutations. Debate persists regarding the precise biological role of CRP, although the prognostic value of CRP has a marker of cardiovascular risk has been firmly established. Intriguing and hypothesis-generating data has recently emerged, adding to discussion about whether CRP is a surrogate risk marker or a mediator of atherothrombosis [23,24]. CRP may mediate opsonization of low-density lipoprotein by macrophages, a critical step in the formation of fatty streaks in the endothelium and plaque formation [25,26]. It may also help induce increased expression of intercellular adhesion molecule1 and vascular adhesion molecule-1, facilitating the migration and tethering of cells to sites of injury in the endothelium [27,28]. CRP has been found to localize with the complement membrane complex in early atherosclerotic tissue [29]. A haplotype of polymorphic variants [30] within this region may ultimately control differential expression. Further investigation of this gene and chromosomal region, as a region of potential high association with basal CRP levels, is warranted.
Acknowledgements This work was supported by grants from the Doris Duke Charitable Foundation and the Donald W. Reynolds Foundation (PMR). There are no foreseen conflicts of interest.
References [1] Ridker PM, Cushman M, Stampfer MJ, Tracy RP, Hennekens CH. Inflammation, aspirin, and the risk of cardiovascular disease in apparently healthy men. N Engl J Med 1997;336:973–9. [2] Ridker PM, Hennekens CH, Buring JE, Rifai N. C-reactive protein and other markers of inflammation in the prediction of cardiovascular disease in women. N Engl J Med 2000;342:836–43. [3] Ridker PM, Rifai N, Rose L, Cook NR. Comparison of C-reactive protein and low-density lipoprotein cholesterol levels in the prediction of first cardiovascular events. N Engl J Med 2002;347:1557– 65. [4] Ridker PM, Buring JE, Cook NR, Rifai N. C-reactive protein, the metabolic syndrome and risk of incident cardiovascular events. Circulation 2003;107:391–7. [5] Tracy RP, Lemaitre RN, Psaty BM, et al. Relationship of C-reactive protein to risk of cardiovascular disease in the elderly: results from the Cardiovascular Health Study and the Rural Health Promotion Project. Arterioscler Thromb Vasc Biol 1997;17:1121–7. [6] Danesh J, Whincup P, Walker M, et al. Low grade inflammation and coronary heart disease: prospective study and updated meta-analysis. BMJ 2000;321:199–204. [7] Koenig W, Sund M, Frohlich M, et al. C-reactive protein, a sensitive marker of inflammation predicts future risk of coronary heart disease in initially healthy middle-aged men: results from the MONICA (monitoring trends and determinants in cardiovascular disease) Augsburg Cohort Study, 1984 to 1992. Circulation 1999;99:237– 42. [8] Kuller LH, Tracy RP, Shaten J, Meilahn EN. Relation of C-reactive protein and coronary artery heart disease in the MRFIT nested casecontrol study: multiple risk factor intervention trial. Am J Epidemiol 1996;144:537–47. [9] Albert C, Ma J, Rifai N, Stampfer MJ, Ridker PM. Prospective study of C-reactive protein, homocysteine, and plasma lipid levels as predictors of sudden cardiac death. Circulation 2002;105:2595– 9. [10] Pepys MB, Hirschfield GM. C-reactive protein: a critical update. J Clin Invest 2003;111:1805–12. [11] Pankow JS, Folsom AR, Cushman M, et al. Familial and genetic determinants of systemic markers of inflammation: the NHLBI family heart study. Atherosclerosis 2001;154:681–9. [12] Watson ML, Kinsmore SF, Johnston GI, et al. Genomic organization of the selectin family of leukocyte adhesion molecules on human and mouse chromosome 1. J Exp Med 1990;172:263–72. [13] Cooper DN, Nussbaum RL, Krawczak M. Proposed guidelines for papers describing DNA polymorphism-disease associations. Hum Genet 2002;110:207–8. [14] Brull DJ, Serrano N, Zito F, et al. Human CRP gene polymorphism influences CRP levels. Arterioscler Thromb Vasc Biol 2003;23:2063–9. [15] Szalai AJ, McCrory MA, Cooper GS, Wu J, Kimberly RP. Association between baseline levels of C-reactive protein (CRP) and a dinucleotide repeat polymorphism in the intron of the CRP gene. Genes Immun 2002;3:14–9. [16] Roberts WL, Molton L, Law TC, et al. Evaluation of nine automated high sensitivity C-reactive protein methods: implications for clinical and epidemiological application, II. Clin Chem 2001;47:418– 25.
H.J. Suk et al. / Atherosclerosis 178 (2005) 139–145 [17] NCBI SNP database. http://www.ncbi.nlm.nih.gov/SNP. [18] Albert MA, Danielson E, Rifai N, Ridker PM. Effect of statin therapy on C-reactive protein levels, the pravastatin inflammation/CRP evaluation (PRINCE): a randomized trial and cohort study. JAMA 2001;286:64–71. [19] Cushman M, Legault C, Barrett-Connor E, et al. Effect of postmenopausal hormones on inflammation-sensitive proteins: the postmenopausal estrogen/progestin interventions (PEPI) study. Circulation 1999;100:717–22. [20] Ridker PM, Hennekens CH, Rifai N, Buring JE, Manson JE. Hormone replacement therapy and increased plasma concentration of C-reactive protein. Circulation 1999;100:713–6. [21] Germer S, Hollan HJ, Higuchi R. High-throughput SNP allelefrequency determination in pooled DNA samples by kinetic-PCR. Genome Res 2000;10:258–66. [22] Zee YL, Ridker PM. Polymorphism in the human C-reactive protein (CRP) gene, plasma concentrations of CRP, and the risk of future arterial thrombosis. Atherosclerosis 2002;162:217–9. [23] Pasceri V, Cheng JS, Willerson JT, Yeh ET, Chang J. Modulation of C-reactive protein-mediated monocyte chemoattractant protein-1 induction in human endothelial cells by anti-atherosclerosis drugs. Circulation 2001;103:2531–4.
145
[24] Verma S, Wang CH, Li SH, et al. A self-fulfilling prophecy: Creactive protein attenuates nitric oxide production and inhibits angiogenesis. Circulation 2002;106:913–9. [25] Zwaka TP, Hombach V, Torzewski J. C-reactive protein-mediated low density lipoprotein uptake by macrophages: implications for atherosclerosis. Circulation 2001;103:1194–7. [26] Tashinen S, Konanven PT, Jarva H, Meri S, Pentikainen MO. Binding of C-reactive protein to modified low-density lipoprotein particles: identification of cholesterol as a novel ligand for C-reactive protein. Biochem J 2002;367:403–12. [27] Pasceri V, Willerson JT, Yeh ET. Direct proinflammatory effect of C-reactive protein on human endothelial cells. Circulation 2000;102:2165–8. [28] Verma S, Li SH, Badiwala MV, et al. Endothelin antagonism and interleukin-6 inhibition attenuate the proatherogenic effects of Creactive protein. Circulation 2002;105:1890–6. [29] Torzewski J, Torzewski M, Bowyer DR, et al. C-reactive protein frequently colocalizes with the terminal complement complex in the intima of early atherosclerotic lesions of human coronary arteries. Arterioscler Thromb Vasc Biol 1998;18:1386–92. [30] Lander E. The new genomics: global views of biology. Science 1996;274:536–9.
Relation of polymorphism within the C-reactive protein gene and plasma CRP levels H. Jacqueline Suka,b,c,∗ , Paul M. Ridkera,b,c , Nancy R. Cooka,b , Robert Y.L. Zeea,b a
Center for Cardiovascular Disease Prevention, Brigham and Women’s Hospital, 75 Francis Street, Boston, MA 02215, USA b Donald W. Reynolds Center for Cardiovascular Research, Harvard Medical School, Boston, MA, USA c Division of Cardiology, Brigham and Women’s Hospital, Harvard Medical School, Boston, MA, USA Received 19 December 2003; received in revised form 15 July 2004; accepted 16 July 2004 Available online 12 October 2004
Abstract We sought to investigate whether genetic variability within the CRP gene affects CRP levels. C-reactive protein (CRP), an acute-phase reactant in inflammation, is an important predictor for cardiovascular disease. The genetic determinants of plasma CRP level remain unknown. We assessed the genotypes of two common polymorphisms within the CRP gene, an exonic 1059G > C and an intronic T > A base substitution, among 2397 participants of a community-based study; 1334 had no prior cardiovascular history, while 1063 had a prior cardiovascular history. Univariable and multivariable-adjusted analyses were performed to examine the association of CRP polymorphisms with CRP levels, modeling different modes of inheritance. The genetic polymorphisms were significantly associated with baseline CRP levels in univariable analysis (1059G > C: GG versus GC or CC genotypes, median CRP levels 0.22 versus 0.15 mg/L, P < 0.0001; intronic T > A: TT versus AT versus AA genotypes, median CRP levels 0.19 versus 0.23 versus 0.24 mg/L, P = 0.003). This relationship persisted after adjusting for age, sex, body mass index, ethnicity, hypertension, smoking, diabetes, hyperlipidemia, and aspirin use. Furthermore, these effects were present in subgroup analyses limited to those with and without prevalent coronary disease, and when assessed within Caucasians only. The present data provide evidence of a genetic component of CRP levels, independently of traditional risk factors for cardiovascular disease. Whether genetic markers can add to information yielded by high sensitivity CRP (hsCRP) in assessing cardiovascular risk needs further evaluation. © 2004 Elsevier Ireland Ltd. All rights reserved. Keywords: C-reactive protein; Inflammation; Genetics; Risk factors
1. Introduction There is strong evidence that C-reactive protein (CRP) is a powerful predictor of incident cardiovascular events at all levels of LDL-C, all levels of the Framingham risk score and at all levels of the metabolic syndrome [1–9]. Basal levels in individuals free of acute illness are reproducible [10]. What drives baseline human CRP levels is unknown, although there is evidence of substantial heritability (35–40%) in familial aggregation studies [11]. The human CRP gene lies on chromosome 1q23, within a conserved genetic region that encodes for proteins criti∗
Corresponding author. Tel.: +1 617 278 0808; fax: +1 617 232 3541. E-mail address: [email protected] (H.J. Suk).
0021-9150/$ – see front matter © 2004 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.atherosclerosis.2004.07.033
cal to the immune system and to intercellular communication [12]. Small studies have reported that polymorphisms within the CRP gene are associated with plasma CRP levels. As in all gene-association studies, such findings must be validated within large and independent sample populations to safeguard against spurious associations [13]. Many smaller studies are later disproved because of systemic differences in genetic composition within study groups. Recently, a polymorphism within the untranslated 3 region of the CRP gene [14], but not one within the promoter region of the gene was associated with plasma CRP levels in 250 study participants. A dinucleotide repeat polymorphism [15] in the intronic region of the CRP gene studied within 546 participants had high correlation with plasma CRP levels.
140
H.J. Suk et al. / Atherosclerosis 178 (2005) 139–145
We hypothesized that two genetic markers within the CRP gene would be associated with high sensitivity CRP (hsCRP) levels [16] and performed genetic analysis of two single nucleotide polymorphisms (SNPs) within the CRP gene: (1) a polymorphism within exon 2 of the CRP gene (1059G > C, NCBI SNP accession no. rs1800947 [17]) and (2) a neighboring polymorphism (intronic T > A, NCBI SNP accession no. rs1417938) among 2743 people from around the United States who provided consent for genetic analysis.
2. Methods 2.1. Study population The 2884 men and women from all states of the United States participated in a multicenter community-based study which has been previously described in detail [18]. Of these, 2743 participants provided blood samples and consent for genetic testing. 2397 participants were free of hormone replacement therapy (HRT), a therapy that elevates CRP levels [19,20]; these participants were chosen for our genetic investigation. 1334 of these participants had no history of prior myocardial infarction, stroke or coronary revascularization (primary prevention group) and 1063 had such history (secondary prevention group). Clinical data such as age, gender, race, systolic and diastolic blood pressures, body mass index, smoking status and diabetes mellitus were collected. Plasma lipid levels and high sensitivity plasma CRP levels were measured. Genomic DNA was extracted from whole blood. 2.2. Genotype determination of the 1059G > C exonic polymorphism Genotypes were determined by polymerase chain reaction-restriction fragment length polymorphism (PCRRFLP), with forward primer 5 -GCCCAGGGTGAGGAAGAGTCT-3 and reverse primer 5 -CCCGCCAGTTCAGGACATTAG-3 . The amplified product was digested with MaeIII enzyme (New England Biolabs) and size-fractionated by electrophoretic separation in 3% agarose gels. 2.3. Genotype determination of the intronic T > A polymorphism Genotyping was performed using the kinetic thermocycling approach, a fluorescent-based assay, as described by Germer et al. [21], on an ABI Prism 7900 Sequence Detection System. The forward allele specific primers for the detection of the A allele was 5 -ACCCCCATACCTC AGATCAAAA of the T allele, 5 -ACCCCC- -3 ; for detection ATACCTCAGATCAAAT--3 . The common reverse primer was 5 -AGGTAAGGGCCACCCCAG-3 . To confirm genotype assignment, two independent observers performed scoring. Disagreements (<2%) were resolved by joint review, and a repeat genotyping. PCR results
were scored blinded to all clinical and biochemical characteristics. 2.4. Statistical analysis Differences in proportions of gender, smoking status, diabetes, ethnicity, aspirin use, and prior vascular disease among different genotypes were analyzed by the χ2 -test for two groups and the χ2 trend test for multiple groups. Continuous variables with normal distribution were analyzed by t-test and ANOVA; variables with non-normal distributions were analyzed by nonparametric tests: the Wilcoxon rank sum tests for two categories, and Kruskal–Wallis tests for multiple groups. Hardy–Weinberg equilibrium for the study population was assessed by the general χ2 -test with one degree of freedom. Linkage disequilibrium was assessed by the general χ2 -test for independence. The association between baseline log-normalized hsCRP levels and CRP genotypes was further investigated by multivariable linear regression models. These models were evaluated for the relationship between hsCRP levels and genotypes after adjustment for age, gender, body mass index (BMI), ethnicity, smoking status, systolic hypertension, current aspirin use, hyperlipidemia, and prior vascular disease. Because the distribution of hsCRP was left-skewed, hsCRP was log transformed for use as the outcome variable in multivariable analysis. Effect estimates were calculated by re-exponentiation of the beta coefficients, which is equivalent to the ratio of CRP levels in the given genotype versus the homozygous wild type. For the 1059G > C genotype, the relative rarity of the CC genotype made the dominant genetic models most statistically robust. Both additive and dominant genetic modeling was considered for the intronic T > A polymorphism. SAS version 8e (SAS Institute, Cary, NC) was used for all statistical analyses and all probability values were assessed in a two-tailed fashion.
3. Results Table 1 shows the baseline characteristics of the study participants by their genotypes. The GG genotype occurred most frequently (88%) in our population for the exotic polymorphism; the CC genotype was rare (<1%) and was considered in conjunction with the heterozygous genotype. The TT genotype (49%) was most common for the intronic T > A polymorphism followed by the AT (42%) and AA (9%) genotypes. Our population had no significant deviations of genotype distributions from expected Hardy–Weinberg equilibrium, which was assessed by the general χ2 -test with one degree of freedom, within both the population with and without history of prior vascular disease. There were no significant differences in the baseline characteristics across the different genotypes in terms of sex, presence of hyperlipidemia, diabetes, known prior vascular disease or aspirin use. Furthermore, the mean values for body
Table 1 Baseline clinical characteristics and plasma CRP levels by genotypes Overall data
1059G > C exonic SNP GG (n = 2100)
GC and CC (n = 295)
TT (n = 1174)
AT (n = 988)
AA (n = 208)
Age (y) Women (%)
62.2 ± 13.3 29.6
62.3 ± 13.3 29.9
61.2 ± 12.9 27.7
0.17 0.45
61.8 ± 13.5 29.6
62.8 ± 13.1 30.2
61.9 ± 13.2 27.3
0.30 0.76
Smoking status (%) Never Former Current
40.1 45.0 14.9
40.0 45.2 14.8
40.5 43.6 15.9
0.82
42.5 43.4 14.1
37.7 46.9 15.4
36.9 45.5 17.7
0.02
Body mass index (kg/m2 ) Hyperlipidemia > 220 (mg/dL, %)
29.2 ± 5.4 47.0
29.2 ± 5.3 46.6
29.5 ± 6.1 49.5
0.32 0.35
29.3 ± 5.0 47.7
29.2 ± 5.3 45.0
29.1 ± 5.3 50.0
0.54 0.80
220.4 ± 38.2 135.6 ± 29.5 37.6 ± 10.3
220.2 ± 38.8 135.3 ± 29.7 37.7 ± 10.4
221.7 ± 33.2 137.6 ± 27.7 36.8 ± 9.4
0.49 0.22 0.12
221.4 ± 37.9 136.8 ± 29.9 37.8 ± 10.3
219.2 ± 38.2 134.4 ± 29.0 37.3 ± 10.2
218.4 ± 39.1 133.4 ± 29.1 37.5 ± 10.6
0.15 0.03 0.38
132.6 ± 16.9 79.3 ± 9.8
132.7 ± 16.8 79.3 ± 9.6
132.2 ± 18.0 79.3 ± 11.0
0.64 0.99
133.0 ± 17.4 79.6 ± 9.9
132.3 ± 16.6 79.0 ± 9.7
131.1 ± 15.4 78.8 ± 9.4
0.14 0.12
Total cholesterol (mg/dL) LDL HDL Blood pressure (mmHg) Systolic Diastolic
p
Intronic T > A SNP
p
Diabetes (%)
18.6
19.1
14.8
0.08
18.4
18.4
20.1
0.69
Ethnicity (%) White Black Hispanic Asian Other
87.1 6.9 3.7 1.7 0.6
86.3 7.7 3.8 1.7 0.5
93.1 0.7 3.1 2.4 0.7
0.0003
81.5 10.9 4.2 2.9 0.6
91.7 3.4 3.5 0.8 0.5
96.0 0.5 3.0 – 0.5
<0.0001
44.4 46.4
44.4 46.4
44.1 46.5
0.91 0.97
43.9 45.3
45.1 48.0
45.2 46.2
0.56 0.39
Known prior vascular disease (%) Aspirin use (%) CRP level (mg/L) Median (25th, 75th %)
0.21 [0.09, 0.43]
0.22 [0.10, 0.45]
0.15 [0.07, 0.32]
<0.0001
0.19 [0.09, 0.41]
0.23 [0.10, 0.46]
0.24 [0.12, 0.45]
H.J. Suk et al. / Atherosclerosis 178 (2005) 139–145
Covariates
0.003
Values are mean ± S.D., except as noted. When applicable, tests for trend are used for multiple group comparisons. General chi-square tests are used for two-group comparisons.
141
142
H.J. Suk et al. / Atherosclerosis 178 (2005) 139–145
Fig. 1. CRP levels by genotype. Median levels and 25th and 75th percentiles. P-values are from Wilcoxon rank sum and Kruskal–Wallis tests.
mass index, total cholesterol, HDL levels, systolic and diastolic blood pressures were not significantly different by genotype. Differences in smoking status as well as a slight trend toward higher LDL levels with the presence of the T allele were seen for the intronic T > A base substitution. Significant differences in genotype representation were seen across different ethnicities for both SNPs, and analysis was later repeated within the Caucasian population only. Baseline levels of CRP were significantly different among subjects with different genotypes (0.22 versus 0.15 mg/L median levels, P < 0.0001; 0.19 versus 0.23 versus 0.24 median levels, P = 0.003) as shown in Table 1 and Fig. 1. With the presence of the less frequent C allele, there is a 29% difference in CRP levels (geometric mean CRP levels 0.21 versus 0.15 mg/L, P < 0.0001). Likewise, there is a highly statistically significant incremental increase in CRP levels with the presence of the less frequent A allele (P = 0.001). The alleles of these two SNPs, furthermore, are in linkage disequilibrium, consistent with a conserved relationship between the G and A alleles (χ2 = 51.9; P < 0.0001 by the general χ2 -test for independence).
Table 2 shows the prevalence of the participants’ genotypes by quartiles of baseline CRP values. As expected, 91% of the participants with CRP levels within the highest quartile of CRP levels had the GG genotype, compared with 83% in the lowest quartile (P-trend < 0.0001). The 47% within the highest quartile of CRP levels had the TT genotype, compared with 55% in the lowest quartile (P-trend = 0.001). These genetic effects on CRP levels persisted in multivariable analysis, after controlling for clinical characteristics, chosen because of their importance as traditional risk factors for vascular disease. For the 1059G > C SNP, the effect estimates of multivariable linear regression, which in linear regression correspond to residual ratios of CRP levels after control for other clinical variables, were determined. These effect estimates will be referred to as “CRP ratios”. After control for age, gender, body mass index, race, smoking status, diabetes, hypertension, aspirin use, hyperlipidemia and history of vascular disease, the CRP ratio of the GC or CC genotype compared with the GG referent genotype was 0.71 (P < 0.0001, Table 3). Similarly, for the intronic T > A SNP, the effect estimates for the AT genotype and the AA genotype
Table 2 Prevalence of genotypes in patients by quartiles of baseline CRP levels Quartile 1 of CRP levels (n, %)
Quartile 2 of CRP levels (n, %)
Quartile 3 of CRP levels (n, %)
Quartile 4 of CRP levels (n, %)
1059G > C exonic SNP GC GC or CC
497 (83%) 102 (17%)
518 (87%) 80 (13%)
539 (90%) 62 (10%)
546 (91%) 51 (9%)
Intronic T >A SNP TT AT AA
324 (55%) 229 (39%) 38 (6%)
295 (50%) 240 (41%) 53 (9%)
277 (46%) 261 (44%) 60 (10%)
278 (47%) 258 (44%) 57 (10%)
∗
p for trends* <0.0001
0.001
Chi-square test for trend was used to compare the prevalence of genotypes among quartile of CRP values.
H.J. Suk et al. / Atherosclerosis 178 (2005) 139–145
143
Table 3 CRP levels within each genotype and effect estimates in multivariable analysis after adjustment for vascular risk factorsa CRP levels geometric means (mg/L)
Confidence intervals
1059G > C exonic SNP Crude GG GC or CC
0.21 0.15
[0.20, 0.22] [0.13, 0.17]
Adjusted Dominant model GG GC or CC
0.21 0.15
[0.20, 0.22] [0.13, 0.17]
0.18 0.22
[0.17, 0.20] [0.20, 0.23]
0.18 0.21 0.24
[0.17, 0.19] [0.20, 0.23] [0.21, 0.28]
Intronic T > A SNP Crude TT AT or AA Adjusted Additive model TT AT AA
p
Effect estimatesb
Confidence interval
p
1 0.72
[0.62, 0.83]
0.0001
1 0.71
[0.63, 0.81]
0.0001
1 1.18
[1.08, 1.30]
0.0004
1 1.17 1.31
[1.07, 1.28] [1.11, 1.53]
0.0008 0.001
<0.0001
<0.0001
0.001
0.0002
a Values are adjusted for age, gender, BMI, ethnicity, smoking status, systolic hypertension, current aspirin use, diabetes, hyperlipidemia, history of prior MI, stroke and any other prior vascular disease. b Effect estimates are equivalent to CRP ratios after adjustment for clinical variables. The CRP ratios and 95% CIs were calculated by multivariable linear regression analyses using log of CRP values as outcome, and then re-exponentiating the resulting beta coefficient values.
compared to the referent TT genotype were 1.17 and 1.31 (P = 0.0008, 0.001), respectively. The effect estimates increase in incremental fashion. By dominant genetic modeling, the CRP ratio for the AT or AA genotype compared to the referent AA genotype was 1.19 (P = 0.0001). CRP geometric means are also shown in Table 3 for each genotype, assessed by genetic models and are concordant with the trends presented in Table 1. In addition, because of a difference seen in genotype distributions among participants of different ethnic descent (Table 1), and of potential unexplained confounding by ethnicity, we evaluated the relationships within Caucasian participants only, again free of HRT. The originally described relationships are again seen. For the 1059G > C SNP, the CRP ratios for the GC or CC genotype compared with the GG referent genotype was 0.70 (P < 0.0001), and for the intronic T > A SNP, the ratios for the AT genotype and the AA genotype compared to the referent TT genotype were 1.15 and 1.28 (P = 0.005, 0.003) by additive genetic modeling, after adjusting for the same traditional risk factors for vascular disease. By dominant genetic modeling, the effect estimates of the AT or AA genotype compared to the referent TT genotype was 1.17 (P = 0.0008). We sought to ensure that this relationship persisted both within the primary prevention and secondary prevention cohorts. Within the primary prevention cohort, for the 1059G > C SNP, the CRP ratio for the GC or CC genotype compared with the referent GG genotype was 0.76 (P = 0.001), and for the intronic T > A SNP, the effect estimates for the AT
genotype and the AA genotype compared to the referent TT genotype were 1.18 and 1.26 (P = 0.005, 0.03) by additive genetic modeling, after adjusting for the same traditional risk factors for vascular disease. Similar trends are seen within the secondary prevention cohort. The CRP ratio of the GC or CC genotype compared with the referent GG genotype was 0.66 (P = 0.0001), and for the intronic T > A SNP, the CRP ratios of the AT genotype and the AA genotype compared to the referent TT genotype were 1.15 and 1.36 (P = 0.06, 0.02) by additive genetic modeling, after adjusting for the same traditional risk factors for vascular disease. In summary, whether the CRP polymorphisms were modeled assuming dominant inheritance for the 1059G > C polymorphism or additive and dominant inheritance for intronic T > A polymorphism, both were strongly and independently associated with CRP levels.
4. Discussion CRP is an acute-phase pentraxin protein produced primarily by the liver in response to production of cytokines such as interleukin-6, interleukin-1 and tumor necrosis factor-␣ during tissue injury or inflammation. The reproducibility and measurability of basal hsCRP levels in people free of acute illness have spawned great clinical interest. Whether genetic determinants of basal CRP levels would provide even more stable risk markers is an intriguing even if yet unproven hypothesis.
144
H.J. Suk et al. / Atherosclerosis 178 (2005) 139–145
The two polymorphisms of the CRP gene studied in this large-scale study were highly associated with CRP levels. CRP levels were significantly higher in participants with the G allele than those with the C allele for the 1059G > C polymorphism, in concordance with the trends previously observed within men only in the Physicians’ Health Study [22]. In addition, we found that hsCRP levels were incrementally higher with additional presence of the A allele for the intronic T > A polymorphism: highest for the AA genotype, followed by the AT genotype and then the TT genotype for the intronic T > A polymorphism. These associations persisted strongly even after controlling for traditional cardiovascular risk factors. The association of these polymorphisms with hsCRP levels also persists when reassessed among Caucasians only, the most represented group in the study. The CRP gene is lies on chromosome 1q23 and is composed of two exons and an intron. While many polymorphisms have been recently reported within the CRP gene, none, to our knowledge, are known to encode for nonsynonymous amino acid change. Several points about the genetic variants that we studied merit mention. The intronic SNP is located only 29 base pairs away from a 5 intron/exon splice site. The 1059G > C SNP in exon 2 results in a codon change from CTC to CTG, both of which encode for leucine. While this is a synonymous change, the CTG codon occurs with 20% less frequency in humans. Particularly low frequency codons do raise questions about DNA stability during translation, but we caution about drawing mechanistic conclusions from primary DNA sequences. The intronic and exonic SNP described may simply be in linkage disequilibrium with other functional variants or have subtle, yet to be discovered, genetic effects. Until further work clarifies potential mechanisms, furthermore, it is prudent to interpret such polymorphisms as genetic markers, potentially in high linkage disequilibrium with causative genetic mutations. Debate persists regarding the precise biological role of CRP, although the prognostic value of CRP has a marker of cardiovascular risk has been firmly established. Intriguing and hypothesis-generating data has recently emerged, adding to discussion about whether CRP is a surrogate risk marker or a mediator of atherothrombosis [23,24]. CRP may mediate opsonization of low-density lipoprotein by macrophages, a critical step in the formation of fatty streaks in the endothelium and plaque formation [25,26]. It may also help induce increased expression of intercellular adhesion molecule1 and vascular adhesion molecule-1, facilitating the migration and tethering of cells to sites of injury in the endothelium [27,28]. CRP has been found to localize with the complement membrane complex in early atherosclerotic tissue [29]. A haplotype of polymorphic variants [30] within this region may ultimately control differential expression. Further investigation of this gene and chromosomal region, as a region of potential high association with basal CRP levels, is warranted.
Acknowledgements This work was supported by grants from the Doris Duke Charitable Foundation and the Donald W. Reynolds Foundation (PMR). There are no foreseen conflicts of interest.
References [1] Ridker PM, Cushman M, Stampfer MJ, Tracy RP, Hennekens CH. Inflammation, aspirin, and the risk of cardiovascular disease in apparently healthy men. N Engl J Med 1997;336:973–9. [2] Ridker PM, Hennekens CH, Buring JE, Rifai N. C-reactive protein and other markers of inflammation in the prediction of cardiovascular disease in women. N Engl J Med 2000;342:836–43. [3] Ridker PM, Rifai N, Rose L, Cook NR. Comparison of C-reactive protein and low-density lipoprotein cholesterol levels in the prediction of first cardiovascular events. N Engl J Med 2002;347:1557– 65. [4] Ridker PM, Buring JE, Cook NR, Rifai N. C-reactive protein, the metabolic syndrome and risk of incident cardiovascular events. Circulation 2003;107:391–7. [5] Tracy RP, Lemaitre RN, Psaty BM, et al. Relationship of C-reactive protein to risk of cardiovascular disease in the elderly: results from the Cardiovascular Health Study and the Rural Health Promotion Project. Arterioscler Thromb Vasc Biol 1997;17:1121–7. [6] Danesh J, Whincup P, Walker M, et al. Low grade inflammation and coronary heart disease: prospective study and updated meta-analysis. BMJ 2000;321:199–204. [7] Koenig W, Sund M, Frohlich M, et al. C-reactive protein, a sensitive marker of inflammation predicts future risk of coronary heart disease in initially healthy middle-aged men: results from the MONICA (monitoring trends and determinants in cardiovascular disease) Augsburg Cohort Study, 1984 to 1992. Circulation 1999;99:237– 42. [8] Kuller LH, Tracy RP, Shaten J, Meilahn EN. Relation of C-reactive protein and coronary artery heart disease in the MRFIT nested casecontrol study: multiple risk factor intervention trial. Am J Epidemiol 1996;144:537–47. [9] Albert C, Ma J, Rifai N, Stampfer MJ, Ridker PM. Prospective study of C-reactive protein, homocysteine, and plasma lipid levels as predictors of sudden cardiac death. Circulation 2002;105:2595– 9. [10] Pepys MB, Hirschfield GM. C-reactive protein: a critical update. J Clin Invest 2003;111:1805–12. [11] Pankow JS, Folsom AR, Cushman M, et al. Familial and genetic determinants of systemic markers of inflammation: the NHLBI family heart study. Atherosclerosis 2001;154:681–9. [12] Watson ML, Kinsmore SF, Johnston GI, et al. Genomic organization of the selectin family of leukocyte adhesion molecules on human and mouse chromosome 1. J Exp Med 1990;172:263–72. [13] Cooper DN, Nussbaum RL, Krawczak M. Proposed guidelines for papers describing DNA polymorphism-disease associations. Hum Genet 2002;110:207–8. [14] Brull DJ, Serrano N, Zito F, et al. Human CRP gene polymorphism influences CRP levels. Arterioscler Thromb Vasc Biol 2003;23:2063–9. [15] Szalai AJ, McCrory MA, Cooper GS, Wu J, Kimberly RP. Association between baseline levels of C-reactive protein (CRP) and a dinucleotide repeat polymorphism in the intron of the CRP gene. Genes Immun 2002;3:14–9. [16] Roberts WL, Molton L, Law TC, et al. Evaluation of nine automated high sensitivity C-reactive protein methods: implications for clinical and epidemiological application, II. Clin Chem 2001;47:418– 25.
H.J. Suk et al. / Atherosclerosis 178 (2005) 139–145 [17] NCBI SNP database. http://www.ncbi.nlm.nih.gov/SNP. [18] Albert MA, Danielson E, Rifai N, Ridker PM. Effect of statin therapy on C-reactive protein levels, the pravastatin inflammation/CRP evaluation (PRINCE): a randomized trial and cohort study. JAMA 2001;286:64–71. [19] Cushman M, Legault C, Barrett-Connor E, et al. Effect of postmenopausal hormones on inflammation-sensitive proteins: the postmenopausal estrogen/progestin interventions (PEPI) study. Circulation 1999;100:717–22. [20] Ridker PM, Hennekens CH, Rifai N, Buring JE, Manson JE. Hormone replacement therapy and increased plasma concentration of C-reactive protein. Circulation 1999;100:713–6. [21] Germer S, Hollan HJ, Higuchi R. High-throughput SNP allelefrequency determination in pooled DNA samples by kinetic-PCR. Genome Res 2000;10:258–66. [22] Zee YL, Ridker PM. Polymorphism in the human C-reactive protein (CRP) gene, plasma concentrations of CRP, and the risk of future arterial thrombosis. Atherosclerosis 2002;162:217–9. [23] Pasceri V, Cheng JS, Willerson JT, Yeh ET, Chang J. Modulation of C-reactive protein-mediated monocyte chemoattractant protein-1 induction in human endothelial cells by anti-atherosclerosis drugs. Circulation 2001;103:2531–4.
145
[24] Verma S, Wang CH, Li SH, et al. A self-fulfilling prophecy: Creactive protein attenuates nitric oxide production and inhibits angiogenesis. Circulation 2002;106:913–9. [25] Zwaka TP, Hombach V, Torzewski J. C-reactive protein-mediated low density lipoprotein uptake by macrophages: implications for atherosclerosis. Circulation 2001;103:1194–7. [26] Tashinen S, Konanven PT, Jarva H, Meri S, Pentikainen MO. Binding of C-reactive protein to modified low-density lipoprotein particles: identification of cholesterol as a novel ligand for C-reactive protein. Biochem J 2002;367:403–12. [27] Pasceri V, Willerson JT, Yeh ET. Direct proinflammatory effect of C-reactive protein on human endothelial cells. Circulation 2000;102:2165–8. [28] Verma S, Li SH, Badiwala MV, et al. Endothelin antagonism and interleukin-6 inhibition attenuate the proatherogenic effects of Creactive protein. Circulation 2002;105:1890–6. [29] Torzewski J, Torzewski M, Bowyer DR, et al. C-reactive protein frequently colocalizes with the terminal complement complex in the intima of early atherosclerotic lesions of human coronary arteries. Arterioscler Thromb Vasc Biol 1998;18:1386–92. [30] Lander E. The new genomics: global views of biology. Science 1996;274:536–9.