Interleukin-6 −174G>C Polymorphism and Risk of Coronary Heart Disease in West of Scotland Coronary Prevention Study (WOSCOPS)
Interleukin (IL)-6 plays an important role in the pathogenesis of coronary heart disease (CHD). Two functional polymorphisms in the IL-6 promoter have been identified (−174G>C and −572G>C), with both the rare alleles being associated with higher plasma levels of IL-6 after bypass surgery and one of them (−174G>C) associated with CHD risk. We have studied the contribution of these polymorphisms to CHD risk in the West of Scotland Coronary Prevention Study (WOSCOPS), a primary prevention trial that demonstrated the effectiveness of pravastatin in reducing morbidity and mortality from CHD. Four hundred ninety-eight cases (consisting of individuals experiencing a cardiovascular event during 4.8 years of follow-up) and 1109 controls (individuals matched for age and smoking habits) were genotyped. In the placebo group, there was no significant evidence of higher risk associated with the −174CC genotype compared with the GG+GC group. However, in the pravastatin-treated group, CC homozygotes had a significantly lower risk of CHD compared with the GG+GC placebo group (odds ratio 0.46, 95% CI 0.27 to 0.79), and this remained statistically significant after adjustment for classic risk factors. Compared with the GG+GC group, men with the CC genotype had modestly, but not significantly, higher baseline levels of IL-6, C-reactive protein, or fibrinogen but showed a significantly greater fall in LDL cholesterol with statin treatment (P=0.036). The −572G>C polymorphism was not significantly associated with any plasma trait or CHD risk. Thus, in subjects under pravastatin treatment, the −174CC genotype was associated with a lower risk of CHD. These results demonstrate the importance of the inflammatory system in determining the risk of CHD and support the nonlipid effect of statins on risk.
It is now clear that inflammation plays a role in the pathogenesis of coronary heart disease (CHD). Inflammation is characterized by a local reaction, which may be followed by activation of a systemic acute-phase reaction.1 In this context, the acute-phase reactant interleukin (IL)-6 plays a major role by upregulating the synthesis of acute-phase proteins, including fibrinogen and C-reactive protein (CRP), from hepatocytes.2,3⇓
IL-6 is a pleiotropic cytokine involved in regulation of the acute-phase response.4 IL-6 promotes coagulation in animal models2,5⇓ and may thus increase the risk of an arterial thrombotic event. It may also be an indirect indicator of more specific underlying pathologies that could increase risk. Higher IL-6 levels may cause cellular damage, such as oxidative stress, which activates nuclear factor-κ B.6 Furthermore, IL-6 may be counterregulatory for tumor necrosis factor (TNF)-α and IL-1β.7 Elevated IL-6 levels are common in patients with CHD and unstable angina, particularly in smokers8; in the Physicians Health Study, baseline IL-6 levels were higher in those who subsequently suffered a myocardial infarction (MI) than in those remaining healthy, with a graded relationship between IL-6 levels and risk.9 IL-6 is strongly correlated with plasma concentrations of inflammatory or hemostatic factors, such as CRP and fibrinogen,10,11⇓ but risk in the Physicians Health Study remained statistically significant after adjustment for levels of CRP.9
Two common functional polymorphisms have been found in the promoter region of the IL-6 gene, −174G>C and −572G>C.12,13⇓ Initial studies suggested that the −174C allele was associated with lower plasma levels of IL-6,12 but later studies found higher levels in subjects with aneurysmal disease,14 in patients undergoing coronary artery bypass grafting (CABG),15 and in newborns, but not in adults.16 Carriers of the −572C allele had higher levels of IL-6 than did GG subjects in the inflammatory state after CABG but not at baseline (before the operation).15 These data suggest that there is a genetically determined difference in the degree of the IL-6 response to a stressful or inflammatory situation that is influenced by these promoter variants.
The magnitude of these genotype effects is likely to be of biological significance in causing an elevated risk of atherosclerosis and thrombosis and greater predisposition to CHD, and in 3 studies, the −174C allele was associated with higher CHD mortality.14,17,18⇓⇓ In the present study, we investigate the relationship of these polymorphisms with the risk of CHD in the West of Scotland Coronary Prevention Study (WOSCOPS), which demonstrated a significant reduction in CHD events in men treated with pravastatin.19 The prespecified analytical strategy was to compare the risk associated with IL-6 genotype in the placebo and pravastatin groups and to explore the impact of genotypes on pravastatin treatment.
The design of the present study has been described previously.20 The study was composed of 6595 moderately hypercholesterolemic men aged 45 to 64 years with no history of MI and with normal renal and hepatic function who were randomized to receive placebo or pravastatin. Individuals who, during the (average) 4.8 years of follow-up, experienced a definite fatal or nonfatal MI or sudden coronary death or who required CABG or angioplasty were defined as cases.20 Two controls per case were matched by age (2-year bands) and smoking habit (2 categories, ie, current smokers and past/never smokers) and consisted of those individuals free of cardiovascular events during follow-up. The ethics committee of the Glasgow Royal Infirmary approved the present study.
Fasting serum lipids were measured as described.21 CRP and IL-6 were measured by use of commercial assays (R&D Systems).15,21⇓ Interassay and intra-assay coefficients of variation were 6.2% and 1.9%, respectively, and assay sensitivity >0.1 mg/L and assay sensitivity <0.70 pg/m were 5% and 3%, respectively. Plasma fibrinogen was assayed by heat-precipitation nephelometry.22 IL-6 was determined in a randomly selected sample of the genotyped individuals, with 100 controls and 65 cases chosen from each −174G>C genotype group. Actual numbers obtained vary slightly because of the unavailability of some of the selected samples.
DNA extraction and genotyping were as previously described.17
Differences in the number of cases and controls that have previously been reported19,22⇓ and those in the present study are due to missing samples or technical difficulties in the extraction of DNA or in obtaining the IL-6 genotype. The effects of the covariates, including genotype, were examined in univariate analysis and, simultaneously, by conditional logistic regression analysis in the whole group and in subjects randomized to placebo or pravastatin. Significance was calculated by using either the Wald test or likelihood-ratio tests. Plasma levels of triglycerides and CRP were highly skewed and were normalized by logarithmic transformations, and untransformed data and approximate SEs are presented.
The baseline characteristics of the genotyped men with and without a coronary event are presented in Table 1. Because of the matching criteria, age and proportion of current smokers were similar in these 2 groups and similar to those reported previously.19,22⇓ Body mass index (BMI), several lipid traits, and fibrinogen and CRP levels were significantly higher in the cases than in the controls. The distribution of genotypes for both −174G>C and −572G>C polymorphisms did not differ from Hardy-Weinberg proportions in either cases or controls, and the rare allele frequencies in the nonevent group were similar to those reported previously in healthy UK men.17 Overall, there was no significant difference in the genotype frequency for either polymorphism between cases and controls; thus, there was no significant effect of these genotypes on risk either before or after adjustment for classic risk factors measured at baseline (age, BMI, HDL cholesterol, LDL cholesterol, blood pressure, logarithmic triglyceride level, and alcohol intake). However, there was suggestive evidence for heterogeneity of genotype effect in the placebo and pravastatin groups (P=0.06). As shown in Table 2, compared with the men in the placebo group homozygous for the common G allele, neither the GC group nor the CC group of men showed a significantly higher risk, although those with the genotype CC showed the highest risk of CHD (odds ratio [OR] 1.26, genotype effect in placebo P= 0.33). In the pravastatin-treated group, the relative risk in the GG and GC men was significantly lower than in the GG placebo men, reflecting the protective benefit of pravastatin, but there was no significant difference in the risk between these 2 groups. However, for the pravastatin group, the frequency of the −174C homozygotes was 10.2% in the cases versus 16.6% in the controls (P=0.056, Table 1), and those with the CC genotype had a significantly lower risk of CHD than did the GG placebo men (OR 0.46, 95% CI 0.27 to 0.79; genotype effect in pravastatin group P=0.019). This lower risk remained statistically significant (P= 0.02) after adjustment for baseline classic risk factors (not shown). Because the risk estimates in the GG and GC men in the placebo group were similar, data from these groups were combined, and as shown in the Figure, after adjustment for classic risk factors along with fibrinogen and CRP, men in the placebo group with the genotype CC had an OR of 1.19 (95% CI 0.80 to 1.77), whereas in the treatment group, these estimates were 0.75 (95% CI 0.58 to 0.97) for the GG+GC men and 0.42 (95% CI 0.21 to 0.84) for the CC men (genotype-treatment interaction P=0.014). There was no evidence of a significant effect of −572G>C genotype on risk and no evidence of heterogeneity of effect in the placebo and treatment arms (Table 2).
To examine the possible mechanism of the −174G>C genotype risk effect, the relationship between genotype and levels of CHD risk factors was determined. As shown in Table 3, homozygotes for the −174C allele had among the highest mean levels of fibrinogen, CRP, and IL-6, and for IL-6, compared with the levels in the GG+GC group, levels in the CC group were 8% higher in controls and 19% higher in cases, but none of these differences were statistically significant. There was no significant difference in the levels of these proteins according to −572G>C genotype (Table 3). There were no significant differences for either genotype within the case and control groups for BMI, smoking, blood pressure, triglyceride levels, total and HDL cholesterol levels, or alcohol consumption (not shown).
To evaluate the possibility that the IL-6 genotype might be influencing the response to pravastatin, plasma levels of LDL cholesterol, HDL, fibrinogen, and CRP at baseline and on treatment were examined according to 174G>C genotype. As expected for a randomized trial, none of the baseline characteristics were significantly different between the placebo and statin groups, the genotype distribution was as expected for Hardy-Weinberg proportions, and the allele frequencies were not different between groups (not shown). Although at baseline, LDL cholesterol levels were similar by genotype (Table 4), after 1 year of treatment, the reduction in LDL cholesterol was greater in the CC group than in the GG+GC group, and the greater fall in the CC group was statistically significant (28% [CC] versus 24.2% [GG+ GC], P=0.036). By contrast, in the placebo group, both genotype groups showed a modest and similar fall in LDL cholesterol (1.2% [CC] versus 1.6% [GG+GC], P=0.71) by 1 year. Although with treatment there were larger declines in fibrinogen and CRP and a larger elevation in HDL levels in the CC group compared with the GG+GC group, none of these differences were statistically significant.
Inflammation is now strongly implicated in the process of atherosclerosis and clinical cardiovascular disease,23 with inflammation being involved in all stages of atherosclerotic development, including oxidative damage,24 cell proliferation, and plaque development and destabilization,25 as well as coagulation, thrombosis, and fibrinolysis.5 If inflammation is both a cause and a consequence of CHD, the plasma levels of acute-phase proteins, such as IL-6, fibrinogen, and CRP, will be markers of the magnitude of inflammatory response and the severity of cardiovascular disease, and genetic polymorphisms that determine the rate of acute-phase protein production would be important genetic risk factors. WOSCOPS has previously reported that mean plasma levels of fibrinogen and CRP were higher at baseline in men who subsequently had a CHD event than in those who remained event free,22 and we have now examined the impact on CHD traits and CHD risk for 2 promoter polymorphisms in the IL-6 gene that have been shown in vitro to affect transcriptional strength12,13⇓ and are thus functional.
For the −572G>C polymorphism, no association was observed with levels of any trait or risk. The low frequency of the −572C allele means that the sample has adequate power to detect only a 1.85-fold higher risk in carriers, and although it cannot be ruled out that this polymorphism may have an effect on risk in some situations, the WOSCOPS data17 confirm that this polymorphism has negligible or extremely modest effects on determining plasma levels of inflammatory markers, such as CRP and fibrinogen, and risk in middle-aged men.
In 3 previous prospective studies,14,17,18⇓⇓ it has been reported that compared with GG subjects, those carrying ≥1 −174C allele have an elevated risk of CHD. The Northwick Park Heart Study (NPHS), a prospective study of healthy men,17 reported that compared with risk in the GG group, risk in the GC group was 1.55 (1.06 to 2.22), but in the CC group, the risk was not significantly elevated (1.07 [0.65 to 1.77]). In the Etude Cas-Temoins sur l’Infarctus du Myocarde(ECTIM) case-control study,18 the risk estimates in the GC and CC genotype groups were similar (1.31 and 1.35) but only significant in the larger GC group (P<0.02). A study of aneurysm patients14 noted that compared with the GG subjects, risk for cardiovascular mortality was 2.83 in the GC subjects (P=0.06) and 3.41 in the CC subjects (P=0.06) and was statistically significant in the combined group (2.95, 95% CI 1.01 to 8.65; P=0.047). Therefore, it is surprising that the GC genotype was not associated with a risk of CHD in the WOSCOPS placebo group and that only in the CC group was there a suggestion of higher risk, but this effect was smaller than that found in previous studies and was not statistically significant. One possible explanation for this is that the presence of other risk factors in the WOSCOPS men (such as the high plasma cholesterol levels that were part of the recruitment criteria) has reduced the impact of the inflammatory effect of the IL-6 genotype. However, the CI of the risk estimate in men in the WOSCOPS CC placebo group overlaps that reported in published data14,17,18⇓⇓; thus, there is no significant evidence of risk heterogeneity among these 4 studies. In NPHS, there was evidence that the −174C allele risk effect was greater in current smokers, but because the WOSCOPS cases and controls selected for the genetic study were matched for smoking habit, it was not possible to carry out this analysis in the sample.
The relationship between IL-6 genotypes and plasma levels of IL-6 appear to be complex. In the first report of 102 healthy subjects, the − 174CC group had the lowest levels,12 but in 3 subsequent large studies, the −174C allele has been associated with higher levels of IL-614–16⇓⇓ and with IL-6–related proteins, such as CRP,14,17⇓ but this effect was statistically detectable only in subjects with inflammation, such as those with aneurysmal disease14 or after birth trauma.16 In a group of patients undergoing CABG, preoperative IL-6 levels did not differ by − 174G>C or −572G>C genotype, but 6 hours after the inflammatory trauma of the operation, levels were significantly higher in carriers of either of the C alleles.15 The WOSCOPS data showed no statistically significant association between IL-6, fibrinogen, or CRP levels with genotype, although the CC group had among the highest mean plasma levels of all of these proteins. Therefore, the lack of a statistically significant association between these IL-6 genotypes and baseline levels of IL-6 and CRP is in line with recent studies of subjects who did not experience an inflammatory trauma.15–17⇓⇓ The modestly higher baseline levels in subjects who subsequently suffered a CHD (ie, those who probably had inflammatory atherosclerosis) support the exaggeration of the C allele– raising effect in an inflammatory situation.15,16⇓
The major novel finding of the present study is that in the pravastatin-treated men, those with the CC genotype had a significantly lower risk of a subsequent CHD event than did the GG+GC placebo group. All IL-6 genotype groups showed significant benefit from pravastatin treatment; however, after baseline CHD risk factors were taken into account, compared with the GG+GC placebo group, men in the treatment group with the GG+GC genotype had 25% lower risk (OR 1 reduced to OR 0.75), whereas in the CC group, risk was 77% lower than that in the CC placebo men (OR reduced from 1.19 to 0.42). Because the lower risk remained statistically significant after adjustment for baseline classic risk factors, including CRP and fibrinogen, we hypothesized that the protective effect of the CC genotype in the treated group was likely to be due to either a greater pravastatin lipid-lowering effect or a greater inflammation-lowering effect in this genotype group. The analysis of the WOSCOPS data suggests that both of these effects may be involved.
Results from the Physicians Heart Study showed that the benefit of aspirin treatment was greatest in subjects with elevated baseline CRP levels.26 This suggests that the benefit of anti-inflammatory treatment may be greatest in those with highest inflammation, and several studies have now identified an anti-inflammatory effect of statin treatment. Ridker et al27 showed that among MI survivors, pravastatin significantly reduced CRP levels by 17% independently of the effect on serum lipids, with the average fall in CRP in these WOSCOPS men being of similar magnitude (14.5%). A crossover study in hyperlipidemic subjects comparing several statins demonstrated that CRP levels fell by ∼30% over a 6-week period although the IL-6 levels did not,28 with this lack of effect on IL-6 levels being confirmed in a 12-week study using simvastatin.29 In support of these findings, we have reported that in a group of patients undergoing CABG, baseline levels of IL-6 were not significantly different for those taking statins versus those not taking statins, although peak IL-6 levels 6 hours after bypass were significantly lower in the statin group,30 indicating that statins may inhibit the acute inflammatory IL-6 increase. TNF-α levels have been reported to be lower in patients treated with pravastatin,31 and in vitro pravastatin downregulates several inflammatory markers in human monocytes/macrophages32 and also results in lower IL-6 mRNA levels in human umbilical vein endothelial cells.33
The 1-year data from the WOSCOPS men suggests that the lower risk associated with the CC genotype in pravastatin-treated patients is in part associated with a greater reduction of inflammatory markers, with the fall in CRP plasma levels being ∼1.7-fold greater in the CC group than in the GG+GC group, although the difference was not statistically significant. Overall, in WOSCOPS, no effect of pravastatin on fibrinogen levels was observed,22 and although fibrinogen levels were lower in the CC group after 1 year of treatment and rose slightly in the GG+GC group, these differences were not statistically significant. However, although baseline levels of LDL cholesterol were not significantly different by −174 genotype, there was a greater and beneficial difference in the response to treatment in the CC group, with the fall in LDL cholesterol levels being ∼1.16-fold greater than that in the GG+GC group. The very small changes in LDL levels seen in the placebo group were not different in those with different −174G>C genotypes.
Because the IL-6 −174G>C genotype is unlikely to be having a direct effect on lipoprotein metabolism, these data suggest that changes in plasma (or locally produced) IL-6 levels (or possibly CRP levels) on pravastatin treatment may be having an effect on ≥1 component of lipoprotein production from the liver or lipoprotein clearance from the circulation and that there is some evidence that cytokines have effects on lipid metabolism. HepG2 cells cultured with IL-1 and IL-6 show reduced secretion of apoB,34 suggesting that lower IL-6 would lead to higher apoB secretion. IL-6 stimulates LDL receptor gene expression via the activation of sterol-responsive elements and stimulatory protein-1–binding elements,35 suggesting that an inflammatory stimulation would be hypolipidemic. This may explain why cholesterol levels fall after an MI, but lowering IL-6 with statin treatment would be predicted to downregulate LDL receptor expression and lead to slower LDL clearance and higher plasma cholesterol levels, which is opposite to the effect seen in the present study.
Although treatment with pravastatin reduced the risk of CHD in men irrespective of the IL-6 −174 genotype, this benefit was particularly evident in the CC group, who represent 16% of the sample. Although the data in WOSCOPS is at best suggestive that the CC group has high inflammation, previous studies14– 16⇓⇓ strongly support the view that this genotype is associated with higher plasma IL-6 levels in an acute inflammatory situation. This result needs to be confirmed, and the mechanism should be explored in further studies, but if correct, it supports the concept that subjects with high inflammation may benefit most from anti-inflammatory treatments, such as aspirin or statins.26,27⇓
The present work was supported by a British Heart Foundation Project grant (PG97/160). F.B. was supported by a grant from The Commission of the European Communities (HIFMECH study, contract BMH4-CT96-0272). S.E.H., G.D.O.L., and A.R. were supported by the British Heart Foundation (RG97006/RG98002).
Received September 8, 2001; revision accepted January 25, 2002.
- ↵Mestries C, Kruithof EKO, Gascon MP, Herodin F, Agay D, Ythier A. In vivo modulation of coagulation and fibrinolysis by recombinant glycosylated human interleukin-6 in baboons. Eur Cytokine Netw. 1994; 5: 273–281.
- ↵Van der Poll T, Levi M, Hack CE, ten Cate H, van Devenster SJH, Eerenberg AJM, de Groot ER, Jansen J, Gallati H, Buller HR, et al. Elimination of interleukin-6 attenuates coagulation activation in experimental endotoxemia in chimpanzees. J Exp Med. 1994; 179: 1253–1259.
- ↵Biasucci ML, Vitelli A, Liuzzo G, Altamura S, Caligiuri G, Monaco C, Rebuzzi AG, Ciliberto G, Maseri A. Elevated levels of interleukin-6 in unstable angina. Circulation. 1996; 94: 874–877.
- ↵Ridker PM, Rifai N, Stampfer MJ, Hennekens CH. Plasma concentration of interleukin-6 and the risk of future myocardial infarction among apparently healthy men. Circulation. 2000; 101: 1767–1772.
- ↵Fishman D, Faulds G, Jeffrey R, Mohamed-Ali V, Yudkin JS, Humphries SE, Woo P. The effect of novel polymorphisms in the interleukin-6 (IL-6) gene on IL-6 transcription and plasma IL-6 levels, and an association with systemic-onset juvenile chronic arthritis. J Clin Invest. 1998; 102: 1369–1376.
- ↵Terry CF, Loukaci V, Green FR. Cooperative influence of genetic polymorphisms on interleukin 6 transcriptional regulation. J Biol Chem. 2000; 275 (24): 18138–18144.
- ↵Jones KG, Brull DJ, Sian M, Brown LC, Humphries SE, Powell JT. Interleukin-6 and the prognosis of abdominal aortic aneurysm. Circulation. 2001; 103: 2260–2265.
- ↵Brull DJ, Montgomery HE, Sanders J, Dhamrait S, Luong L, Rumley A, Lowe GDO, Humphries SE. Interleukin-6 gene −174G>C and −572G>C promoter polymorphisms are strong predictors of plasma IL-6 levels after coronary artery bypass surgery. Arterioscler Thromb Vasc Biol. 2001; 21: 1458–1463.
- ↵Humphries SE, Luong LA, Ogg MS, Hawe E, Miller G. The interleukin-6 -174C>G promoter polymorphism is associated with risk of coronary artery disease and systolic blood pressure in healthy men. Eur Heart J. 2001; 22: 2243–2252.
- ↵Lipid Research Clinics Program. Lipid and Lipoprotein Analysis: Manual of Laboratory Operations. Vol I. Bethesda, Md: Department of Health, Education, and Welfare, National Institute of Health; 1974.Publication 75-628.
- ↵Berliner J, Navab M, Fogelman A, Frank J, Demer L, Edwards P, Watson A, Lusis A. Atherosclerosis: basic mechanisms: oxidation, inflammation, and genetics. Circulation. 1995; 91: 2488–2496.
- ↵Buja L, Willerson J. Role of inflammation in coronary plaque disruption. Circulation. 1994; 89: 503–505.
- ↵Ridker PM, Rifai N, Pfeffer MA, Sacks F, Braunwald E, for the Cholesterol and Recurrent Events (CARE) Investigators. Long-term effects of pravastatin on plasma concentration of C-reactive protein. Circulation. 1999; 100: 230–235.
- ↵Jialal I, Stein D, Balis D, Grundy SM, Adams-Huet B, Devaraj S. Effect of hydroxymethyl glutaryl coenzyme A reductase inhibitor therapy on high sensitive C-reactive protein levels. Circulation. 2001; 103: 1933–1935.
- ↵Inoue I, Goto S, Mizotani K, Awata T, Mastunaga T, Kawai S, Nakajima T, Hokari S, Komoda T, Katayama S. Lipophilic HMG-CoA reductase inhibitor has an anti-inflammatory effect: reduction of MRNA levels for interleukin-1beta, interleukin-6, cyclooxygenase-2, and p22phox by regulation of peroxisome proliferator-activated receptor alpha (PPARalpha) in primary endothelial cells. Life Sci. 2000; 67: 863–876.
- ↵Yokoyama K, Ishibashi T, Yi-qiang L, Nagayoshi A, Teramoto T, Maruyama Y. Interleukin-1 and interleukin-6 increase levels of apolipoprotein B mRNA and decrease accumulation of its protein in culture medium of HepG2 cells. J Lipid Res. 1988; 39: 103–113.
- ↵Gierens H, Nauck M, Roth M, Schinker R, Schürmann C, Scharnagl H, Neuhaus G, Wieland H, März W. Interleukin-6 stimulates LDL receptor gene expression via activation of sterol-responsive and Sp1 binding elements. Arterioscler Thromb Vasc Biol. 2000; 20: 1777–1783.