Estrogen Receptor α Gene Variation and the Risk of Stroke
Background— Estrogen receptor α (ESR1) gene variation is associated with a range of important estrogen-dependent characteristics, including responses of lipid profile and atherosclerotic severity to hormone replacement therapy, coronary heart disease risk, and migraine. The roles that reproductive steroids play in cerebrovascular pathophysiology and ischemia are an important area of investigation. Given that there is a significantly higher risk of myocardial infarction among men with the CC genotype (PP of PvuII) of c.454-397T>C (rs2234693), we asked whether this genotype is associated with a higher risk of stroke.
Methods— Relative risk of stroke by genotype was determined in 2709 participants of the Second Northwick Park Heart Study, white males with a mean baseline age 56 years and follow up 10.5 years.
Results— Compared with participants with the ESR1 c.454-397CT or TT genotype, those with the CC genotype had a relative risk of stroke of 1.92 (95% confidence interval, 1.06 to 3.48, P=0.03) after adjustment for age, primary care practice; additional adjustment for body mass index, serum cholesterol and triglyceride levels, hypertension, diabetes, and smoking status. Exclusion of stroke cases with coronary heart disease gave results that were essentially unchanged.
Conclusions— In this study, subjects with the common ESR1 c.454-397CC genotype have a substantial increase in risk of stroke. In another publication, other ESR1 variation was associated with migraine. We thus hypothesize that estrogen receptor variation may provide a basis for the established relationship among estrogens, migraine, and stroke.
There is recent evidence for a role of estrogen receptor α (ESR1) gene variation in determining a range of important estrogen-dependent characteristics, including responses of lipid profile and atherosclerotic severity to hormone replacement therapy, coronary reactivity, and coronary heart disease (CHD) risk.1–6 The roles that reproductive steroids play in cerebrovascular pathophysiology and ischemia are an important area of ongoing investigation. Given that there is a significantly higher risk of myocardial infarction among men with the CC genotype (PP of PvuII)4,5 of c.454-397T>C (rs2234693), we tested the hypothesis that the CC genotype is associated with a higher risk of stroke.
Materials and Methods
We studied white men from the prospective population-based Second Northwick Park Heart Study in the United Kingdom.7 Follow up is ongoing with a current median duration of 10.5 years. Strokes were categorized according to the International Classification of Diseases, Ninth Edition (ICD-9): cerebral artery occlusion (434.9, n=31), unspecific cerebrovascular accident (436.0, n=18), intracerebral hemorrhage (431.0, n=6), subarachnoid hemorrhage (430.0, n=0), and cerebral embolism (434.1, n=0) on the basis of clinical presentation, computed tomography (CT), lumbar puncture, or autopsy findings. Genotype data from 2709 men were included in the current study. (Please see additional information in the online only Appendix to this article available at http://www.strokeaha.org.)
Individuals with stroke had higher baseline means or frequencies of the nonlipid-established cardiovascular disease risk factors (Table). In a model with adjustment for age and primary care practice, compared with men with the ESR1 c.454-397CT or TT genotype, those with the CC genotype had a relative risk of 1.92 (95% confidence interval [CI], 1.06 to 3.48, P=0.03). This result was essentially unchanged, at 1.84 (95% CI, 1.01 to 3.35, P=0.045), after additional adjustment for body mass index, serum cholesterol and triglyceride levels, hypertension, diabetes, and smoking status. Further adjustment of the model for CHD, or exclusion of participants with CHD, gave similar results. Survival curves were calculated for stroke by genotype (Figure).
Apart from diabetes, baseline characteristics (Table), high-density lipoprotein cholesterol, and Apo-A1 levels gave no evidence of association with genotype.
Estrogen receptors are required for normal vascular physiology in males,8 and in animal models, estrogen receptor α plays a role in estradiol-mediated cerebral injury protection.9 The mechanisms that underlie the association seen here with stroke are not clear, although there is evidence from 2 groups that the ESR1 c.454-397C allele results in a relatively high level of ESR1 transcription.6,10 Other polymorphisms in linkage disequilibrium with ESR1 c.454-397T>C may nevertheless be responsible for some or all of the observed association.11,12 Our coincidental support for an association of genotype with diabetes is consistent with a report of impaired glucose tolerance and hyperinsulinemia in a man with a premature stop codon in both copies of ESR1.13 However, this result may be because of multiple testing, and a much larger study found no such evidence of association with diabetes.6
Importantly, the association of genotype with risk of stroke was independent of established cardiovascular risk factors, including diabetes, and was essentially unchanged after exclusion of stroke cases with CHD.
These results require replication in other studies of stroke among both ethnically similar and disparate subjects. We do not currently have access to such data, but our findings are strengthened by the fact that they resemble previous results for myocardial infarction in a Finnish Study and the Framingham Heart Study.4,5 A common mechanism based on ESR1 variation may determine risk of both CHD and stroke. A recent report of association between another ESR1 polymorphism and migraine susceptibility14 makes variation in this gene a potential basis for the close but complex relationship among estrogens, migraine, and stroke.15 In men, the common ESR1 c.454-397CC genotype, present in approximately 20% of individuals, may be a risk factor for stroke and, if confirmed in replicate studies, may be useful for identifying at-risk subjects.
Materials and Methods
Participants were registered with 9 primary care general practices in the United Kingdom. To be eligible for initial inclusion subjects had to be free from a history of unstable angina, myocardial infarction or evidence of silent infarction, coronary surgery, aspirin or anticoagulant therapy, cerebrovascular disease, malignancy (except skin cancer other than melanoma), or any condition precluding informed consent. Participants answered a health and life-style questionnaire at enrolment. Weight, height and systolic blood pressure were recorded and venous blood samples were collected for plasma and DNA analysis. Diabetes status was self-reported. Cholesterol and triglyceride concentrations were measured using automated enzyme procedures. Serum apolipoprotein-A1 (Apo-A1) was determined by immunoprecipitation, and high-density lipoprotein (HDL) cholesterol after precipitation of apolipoprotein-B containing particles with polyethylene glycol. Blood pressure was recorded twice per examination, with a random zero sphygmomanometer, after subjects had been seated for at least 5 min, and the results averaged for statistical analysis. Hypertension was defined as a systolic blood pressure ≥140 mm Hg and/or a diastolic blood pressure [mtqu]90 mm Hg, or the use of antihypertensive medication.
Strokes were defined as a rapid onset of focal or global loss of cerebral function lasting >24 hours or causing death beforehand, with no apparent cause other than cerebral ischemia or haemorrhage. Diagnosis was supported by CT in 30 of 31 coded 434.9, 5 of 18 coded 436.0, and 4 of 6 coded 431.0. Genotyping was carried out, with inclusion of positive and negative genotypic control samples on each plate, using automated Amplifluor-based methods. All subjects gave informed consent, and the examination protocols were approved by the appropriate Institutional Review Boards.
Statistical analysis was conducted using Intercooled STATA version 8.2. The association of genotype with risk of stroke was assessed using Cox proportional hazards model, with significance assessed by the likelihood ratio test. The threshold for statistical significance was set at P=0.05.
The study is powered to detect a relative risk of 2.25 (for CC versus CT or TT genotypes) with 80% power at the 5% level of significance.
Genotype and allele frequencies were similar to those previously observed, with no significant deviation from Hardy-Weinberg expectations.
A covariate adjusted genotype model gave hazard ratios of 1.00 for TT, 1.10 (95% CI, 0.55-2.21) for TC, and 2.04 (95% CI, 0.96-4.34) for CC genotypes, P=0.12. From an additive model the hazard ratio for each C-allele was 1.46 (95% CI, 0.98-2.17) P=0.06. After additional adjustment for CHD the result was 1.0 for TT, 1.09 (0.54-2.21) for TC and 1.83 (0.86-3.91) for CC, P=0.22. For each C-allele the hazard ratio was 1.40 (0.94-2.08), P=0.10.
At baseline 8.9% (70) of TTs, 8.0% (109) of CTs and 8.8% (48) of CCs were on cholesterol lowering treatments (p=0.72). Of those with stroke 23.1% (3) of TTs, 16.0% (4) of CTs and 17.7% (3) of CCs were on cholesterol lowering treatment (p=0.90).
Of the 55 stroke patients 12 had CHD: 5 with TT, 4 with CT, and 3 with CC genotype. Additional adjustment of the model for CHD gave a relative risk for the CC genotype of 1.73 (95% CI, 0.95-3.15, P=0.07), while exclusion of participants with CHD gave a relative risk of 1.97 (95% CI, 1.01-3.84, P=0.048).
Analysis of baseline characteristics by genotype provided evidence for an association with diabetes, present in 1.9% (15) of TT, 3.0% (41) of TC, and 1.3% (7) of CC individuals. Logistic regression models of diabetes adjusted for age and primary care practice, and additionally for alcohol consumption, BMI, smoking status, hypertension, and high serum triglyceride levels provided similar, significant results for the genotype model (ORs for diabetes by genotype: TT 1.00, TC 1.58 [0.87-2.88], CC 0.66 [0.27-1.63], P=0.04; and TT 1.00, TC 1.66 [0.91-3.05], CC 0.62 [0.25-1.56], P=0.02 respectively). Among subjects with diabetes there was deviation from Hardy-Weinberg expectations (P=0.01).
As we only studied Caucasian men from the United Kingdom and included adjustment for primary care practice at which the men were recruited it is unlikely that the results will have been affected by stratification. Limitations of our study are the small number and mixed etiologies of strokes. The study did not provide power to test separately for associations with type of stroke. Thus our study may underestimate the true size of effect that might be seen in a homogenous group of stroke cases. However, the fact that we are reporting a prospective study reduces the likelihood of bias that is inherent in case-control studies.
We are indebted to all those who participated in the Study. This work was supported by a Specialized Center of Research in Ischemic Heart Disease from the National Institutes of Health’s National Heart, Lung and Blood Institute (NHLBI, P50 HL63494). The Second Northwick Park Heart Study was supported by the UK Medical Research Council, the British Heart Foundation (RG 2000/015), NHLBI 33014, RO1-HL65230, and Du Pont Pharma, Wilmington, De.
- Received May 11, 2005.
- Accepted May 12, 2005.
Herrington DM, Howard TD, Hawkins GA, Reboussin DM, Xu J, Zheng SL, Brosnihan KB, Meyers DA, Bleecker ER. Estrogen-receptor polymorphisms and effects of estrogen replacement on high-density lipoprotein cholesterol in women with coronary disease. N Engl J Med. 2002; 346: 967–974.
Cooper JA, Miller GJ, Bauer KA, Morrissey JH, Meade TW, Howarth DJ, Barzegar S, Mitchell JP, Rosenberg RD. Comparison of novel hemostatic factors and conventional risk factors for prediction of coronary heart disease. Circulation. 2000; 102: 2816–2822.
Mendelsohn ME, Rosano GM. Hormonal regulation of normal vascular tone in males. Circ Res. 2003; 93: 1142–1145.
Dubal DB, Zhu H, Yu J, Rau SW, Shughrue PJ, Merchenthaler I, et al. Estrogen receptor alpha, not beta, is a critical link in estradiol-mediated protection against brain injury. Proc Natl Acad Sci U S A. 2001; 98: 1952–1957.
Herrington DM, Howard TD, Brosnihan KB, McDonnell DP, Li X, Hawkins GA, Reboussin DM, Xu J, Zheng SL, Meyers DA, Bleecker ER. Common estrogen receptor polymorphism augments effects of hormone replacement therapy on E-selectin but not C-reactive protein. Circulation. 2002; 105: 1879–1882.
Kunnas TA, Laippala P, Penttila A, Lehtimaki T, Karhunen PJ. Association of polymorphism of human alpha oestrogen receptor gene with coronary artery disease in men: a necropsy study. BMJ. 2000; 321: 273–274.
Bousser MG. Estrogens, migraine, and stroke. Stroke. 2004; 35: 2652–2656.