|Home | About | Journals | Submit | Contact Us | Français|
Cytochrome P450-derived epoxyeicosatrienoic acids are potent vasodilators in preclinical models and are hydrolyzed by soluble epoxide hydrolase (EPHX2). Associations between the EPHX2 Lys55Arg and Arg287Gln polymorphisms and cardiovascular disease risk have been reported; however, their impact on vascular function in humans has not been investigated. In 265 volunteers (198 white, 67 black American), forearm blood flow was measured by strain-gauge venous occlusion plethysmography at baseline and in response to bradykinin, methacholine and sodium nitroprusside. Forearm vascular resistance was calculated as mean arterial pressure/forearm blood flow. In white Americans, Lys55Arg genotype was associated with vasodilator response to bradykinin, such that forearm blood flow was significantly lower (P=0.043) and forearm vascular resistance was significantly higher (P=0.013) in Arg55 variant allele carriers compared to wild-type individuals. Significant associations were also observed with methacholine and sodium nitroprusside. In contrast, no relationship was observed in black Americans. In black Americans, Arg287Gln genotype was associated with vasodilator response to bradykinin. Although the difference in forearm blood flow did not reach statistical significance (P=0.058), forearm vascular resistance was significantly lower (P=0.037) in Gln287 variant allele carriers compared to wild-type individuals. Significant associations were also observed with methacholine and sodium nitroprusside. In white Americans, Gln287 variant allele carriers did not exhibit significantly higher forearm blood flow (P=0.128) or lower forearm vascular resistance (P=0.080). Genetic variation in EPHX2 is associated with forearm vasodilator responses in a bradykinin receptor- and endothelium-independent manner, suggesting an important role for soluble epoxide hydrolase in the regulation of vascular function in humans.
Cytochrome P450 (CYP) epoxygenase enzymes from the CYP2J and CYP2C subfamilies catalyze the oxidative metabolism of arachidonic acid to epoxyeicosatrienoic acids (EETs).1 EETs are synthesized in the endothelium and cause vasodilation in numerous vascular beds via activation of calcium-sensitive potassium channels (BKCa) and smooth muscle cell hyperpolarization.2,3 Consequently, CYP-derived EETs are regarded as one of the primary endothelium-derived hyperpolarizing factors (EDHFs).2 Soluble epoxide hydrolase (sEH) rapidly hydrolyzes EETs to less biologically active dihydroxyeicosatrienoic acid (DHET) metabolites.4 Pharmacological inhibition of sEH increases cellular and circulating EET levels, potentiates EET-mediated vasodilation, and lowers blood pressure in preclinical models.4–6 Importantly, sEH inhibitors are currently in clinical development for the treatment of hypertension.
Associations between polymorphisms in the gene encoding sEH (EPHX2) and cardiovascular and cerebrovascular disease risk in humans have been recently reported.7–13 Most notably, the nonsynonymous Arg287Gln and Lys55Arg polymorphisms exhibit lower and higher sEH metabolic activity in vitro, respectively.14–16 The Gln287 variant allele has been associated with a lower prevalence of ischemic stroke in a Chinese population,8 although the presence and strength of this relationship has been inconsistent across studies and racial groups.9,10,17 The Arg55 variant allele has been associated with significantly higher risk of developing coronary artery disease and ischemic stroke events in populations of European ancestry.11,12 Although multiple reports demonstrate presence of a significant relationship between genetic variation in EPHX2 and cardiovascular and cerebrovascular disease risk in humans,8–13 the role of sEH in the regulation of vascular function in humans remains poorly understood. Due to their effects on sEH metabolic function in vitro, we hypothesized that the Arg55 and Gln287 variant alleles, respectively, predispose individuals to higher and lower vascular resistance in vivo. Consequently, we characterized the relationship between genetic variation in EPHX2 and forearm vasodilator responses.
Please see http://hyper.ahajournals.org for an expanded description of the methods.
DNA was extracted from 265 healthy adults who participated in studies that evaluated forearm vasodilator responses.18,19 All subjects gave written informed consent. The protocol was approved by the Vanderbilt University Institutional Review Board and conducted according to the Declaration of Helsinki.
Forearm blood flow (FBF) was measured using strain-gauge venous occlusion plethysmography (D.E. Hokanson, Bellevue, WA) at baseline and in response to incremental doses of bradykinin (100, 200, and 400 ng/min; Clinalfa AG, Läufelfingen, Switzerland), methacholine (3.2, 6.4, and 12.8 μg/min; Pharmaceutical Compounding Center, Nashville, TN) and sodium nitroprusside (1.6, 3.2, and 6.4 μg/min; Gensia Siccor Pharmaceuticals, Irvine, CA), as described previously.18–21 Forearm vascular resistance (FVR) was calculated as the ratio of mean arterial pressure/FBF.
In addition to the nonsynonymous Lys55Arg and Arg287Gln polymorphisms, eight polymorphisms in EPHX2 with >5% frequency and r2 <0.80 were also genotyped for haplotype construction (Table S1).
Due to previously reported racial differences in forearm vasodilator responses,18,21 all analyses were stratified by race. Differences in FBF and FVR were determined using general linear model repeated-measures ANOVA in which the within-subject variable was vasoactive agent dose and the between-subject variables were genotype/haplotype group, cigarette smoking status (yes/no), and body mass index (quartile).18,19 A post-hoc Scheffe’s test was performed to identify between-group differences. Data are presented as least-squares mean ± standard error (SE) unless otherwise indicated. To minimize the impact of the multiple statistical tests conducted, we estimated the false discovery rate (FDR) q-value for each comparison.11,22 Only q-values for statistically significant findings (P<0.05) are presented.
Study subject characteristics and genotype distributions for the EPHX2 Lys55Arg and Arg287Gln polymorphisms are presented in Table 1 according to race. No significant differences in age, gender, cigarette smoking status, body mass index, or resting blood pressure were observed across either genotype in either racial group (Table S2).
Bradykinin infusion increased FBF and decreased FVR in a dose-dependent fashion (Figures 1, ,22 and and3),3), but did not significantly affect mean arterial pressure or heart rate (P=NS). Race, cigarette smoking status and body mass index (ANOVA P<0.05), but not gender, were each significant predictors of bradykinin-stimulated changes in FBF and FVR.
There was no relationship between Lys55Arg genotype and either resting FBF (P=0.932 and P=0.876) or FVR (P=0.663 and P=0.752) in white and black American subjects, respectively (Table S2). In white Americans, vasodilator response to bradykinin was significantly associated with Lys55Arg genotype, such that FBF was significantly lower (Figure 1A, P=0.043, q=0.078) and FVR was significantly higher (Figure 1C, P=0.013, q=0.047) in Arg55 variant allele carriers (Lys/Arg or Arg/Arg) relative to wild-type (Lys/Lys) individuals. In contrast, genotype differences were not observed in black Americans (Figure 1B and 1D).
Resting FVR was significantly lower in black American Gln287 variant allele carriers (Arg/Gln or Gln/Gln) compared to wild-type (Arg/Arg) individuals (P=0.040, Table S2). Although resting FBF also appeared to be higher in white and black Americans (P=0.086 and P=0.078, respectively), and FVR appeared to be lower in white Americans (P=0.053), these differences were not statistically significant (Table S2). Vasodilator response to bradykinin was associated with Arg287Gln genotype in black Americans. Although the difference in FBF did not reach statistical significance (Figure 2B, P=0.058), FVR was significantly lower (Figure 2D, P=0.037, q=0.072) in Gln287 variant allele carriers relative to wild-type individuals. In white Americans, the observed differences in FBF (Figure 2A, P=0.128) and FVR (Figure 2C, P=0.080) were less substantial than those observed in black Americans and were not statistically significant.
Inclusion of both Lys55Arg and Arg287Gln genotype in the model did not alter the observed relationships between each individual polymorphism and FVR in either white (Lys55Arg: P=0.017; Arg287Gln: P=0.171) or black Americans (Lys55Arg: P=0.731; Arg287Gln: P=0.037). An exploratory analysis evaluating the combination of both genotypes yielded consistent results (Figure S1).
Vasodilator responses to methacholine (P=0.021, q=0.063) and sodium nitroprusside (P=0.006, q=0.039) were also significantly associated with Lys55Arg genotype in white Americans (Table 2). As observed with bradykinin, FVR was significantly higher in response to each agent in Arg55 variant allele carriers (Lys/Arg or Arg/Arg) relative to wild-type (Lys/Lys) individuals. Significant genotype differences were not observed in black Americans.
Arg287Gln genotype was associated with vasodilator responses to methacholine (P=0.033, q=0.072; and P=0.025, q=0.063) and sodium nitroprusside (P=0.125; and P=0.006, q=0.039) in both white and black Americans, respectively (Table 2). As observed with bradykinin, FVR was lower at baseline and in response to each agent in Gln287 variant allele carriers (Arg/Gln or Gln/Gln) relative to wild-type (Arg/Arg) individuals.
Nine EPHX2 polymorphisms were included in haplotype construction, which identified six common haplotypes (>5%) in white American subjects (Figure S2). Only one common haplotype was tagged by the Lys55Arg polymorphism variant G allele, which had a frequency of 13.7%. No differences in either resting FBF (P=0.471) or FVR (P=0.301) were observed in subjects carrying the variant Arg55-tagged haplotype. Vasodilator response to bradykinin, however, was significantly associated with this haplotype, such that FBF was significantly lower (Figure 3A, P=0.016, q=0.053) and FVR was significantly higher (Figure 3B, P=0.004, q=0.039) in subjects carrying the Arg55-tagged haplotype. Similar results were observed with Arg55-tagged haplotypes constructed using the sliding-window approach (Figure S3). In response to incremental doses of methacholine and sodium nitroprusside, respectively, FBF was significantly lower (P=0.025, q=0.063; and P=0.001, q=0.016) and FVR was significantly higher (P=0.009, q=0.047; and P=0.001, q=0.016) in subjects carrying the Arg55-tagged haplotype.
A distinct haplotype tagged by the Arg287Gln polymorphism variant A allele had a frequency of 11.1%. Although carriers of the Gln287-tagged haplotype appeared to exhibit higher resting FBF (P=0.127) and lower FVR (P=0.083), these differences were not statistically significant. Similar results were observed with Gln287-tagged haplotypes constructed using the sliding-window approach (Figure S3). No statistically significant associations between the Gln287-tagged haplotype and vasodilator responses to incremental doses of bradykinin (Figure 3C and 3D), methacholine (FBF: P=0.184, FVR: P=0.107) or sodium nitroprusside (FBF: P=0.254, FVR: P=0.233) were observed.
Cytochrome P450 (CYP) epoxygenases catalyze endothelial EET biosynthesis. These potent vasodilators are rapidly hydrolyzed by sEH, and pharmacological inhibition of sEH potentiates the effect of EETs.1,2,4 This study identified a significant association between functional genetic variants in EPHX2, the gene which encodes sEH, and vascular function in humans. Specifically, the Arg55 variant allele was associated with significantly lower FBF and higher FVR in response to vasodilators in white Americans, while the Gln287 variant allele was associated with higher FBF and lower FVR at baseline and in response to vasodilators in black Americans. Similar genotype-phenotype relationships were observed in response to bradykinin, methacholine and sodium nitroprusside, suggesting that the mechanism underlying the functional link between sEH and forearm vascular responsiveness in humans is both bradykinin receptor-and endothelium-independent. Collectively, these data demonstrate a potentially important role for sEH in the regulation of vascular function in humans, and offer important mechanistic insight into previously reported associations between the EPHX2 Lys55Arg and Arg287Gln polymorphisms and cardiovascular and cerebrovascular disease risk.
The nonsynonymous Lys55Arg polymorphism exhibits higher sEH metabolic activity in vitro14,15 and in vivo,11 and has been associated with higher risk of coronary artery disease and ischemic stroke events in individuals of European ancestry.11,12 Consistent with higher sEH metabolic activity and EET hydrolysis, white American carriers of the Arg55 variant allele exhibited lower vasodilation responses compared to wild-type individuals. Lower vasodilation responses were also observed in carriers of the Arg55-tagged haplotype, suggesting that Lys55Arg is the functional allele driving these associations. We cannot rule out, however, that Lys55Arg is simply a marker in linkage disequilibrium with the true causative locus. In contrast, there was no association between Lys55Arg genotype and either resting or agonist-stimulated changes in FBF or FVR in black American subjects. Although the mechanism underlying the racial differences in this genotype-phenotype relationship is not known, these findings are consistent with the prior observation that Lys55Arg genotype was associated with sEH metabolic activity in vivo and cardiovascular disease risk in white, but not black, Americans enrolled in the Atherosclerosis Risk in Communities (ARIC) study.11
Multiple studies have demonstrated that the Arg287Gln polymorphism exhibits significantly lower sEH metabolic activity and EET hydrolysis in vitro,14–16 although its functional impact on sEH metabolic activity in vivo remains unknown. Epidemiological studies provide inconsistent data regarding the presence and direction of a significant relationship between the Arg287Gln polymorphism and risk of cardiovascular and cerebrovascular events.8–13,17 Consistent with the vascular protective effects associated with sEH inhibition in preclinical models,4 the Gln287 variant allele was recently associated with a significantly lower risk of ischemic stroke in a Chinese population.8 In contrast, the Gln287 variant allele was associated with a higher prevalence and extent of coronary artery calcification in black Americans9 and insulin resistance in Japanese type 2 diabetics.23 Similarly, in contrast to the neuronal protective effects conferred by the Gln287 variant allele in vitro,15 an association with significantly higher risk of ischemic stroke was reported in a white European population.10 No association with coronary artery disease and ischemic stroke risk has been observed in other white or black American11,13 and European populations.12,17 Collectively, these conflicting data demonstrate that the functional relevance of the Arg287Gln polymorphism in humans remains unclear and requires further study.
Consistent with the hypothesis that decreased endogenous sEH metabolic activity enhances the vasodilator effects of EETs, Gln287 variant allele carriers demonstrated higher FBF and lower FVR under basal and agonist-stimulated conditions compared to wild-type individuals in the current investigation. These data are also consistent with the observation that administration of pharmacological inhibitors of sEH cause vasodilation in preclinical models and isolated human vessels.4–6 The observed relationship between the Arg287Gln polymorphism and vasodilator responses was more pronounced in black American subjects compared to the white American subjects studied. Although the current study was underpowered to investigate genotype-by-genotype interactions in each racial group, an interaction between the Lys55Arg and Arg287Gln polymorphisms appears unlikely to account for the observed racial differences.
The Lys55Arg and Arg287Gln polymorphisms were associated with altered vasodilation responses to both bradykinin and methacholine. These endothelium-dependent vasodilators cause vascular relaxation in part by stimulating the endothelial formation of EETs, which subsequently hyperpolarize vascular smooth muscle.2 Somewhat unexpectedly, these functional variants in EPHX2 also affected the endothelium-independent vasodilator response to sodium nitroprusside. Although it is well-established that endothelial-derived EETs act as paracrine mediators of vasodilation,2 EPHX2 deletion and pharmacological sEH inhibition also attenuate vascular remodeling in preclinical models by potentiating the anti-proliferative and anti-inflammatory effects of EETs.24,25 Our findings suggest that chronic inter-individual differences in EET exposure based on Lys55Arg and Arg287Gln genotype could contribute to physiological differences in forearm vascular responsiveness via an effect on vascular remodeling. An important limitation of our work, however, is the lack of functional data demonstrating that Arg55 and Gln287 variant allele carriers, respectively, exhibit higher and lower EET hydrolysis in vivo. Further studies will be necessary to characterize the direct effects of exogenous EET administration and sEH inhibition on vasodilation and vascular remodeling in humans, and define the underlying mechanisms. Moreover, although in vitro studies have demonstrated that the Arg55 and Gln287 variant alleles elicit lower and higher sEH phosphatase activity, respectively,26,27 the potential contribution of sEH phosphatase activity to the association between EPHX2 polymorphisms, vascular function, and cardiovascular and cerebrovascular risk remains unknown and requires further investigation.
This study examined the vascular effects of EPHX2 polymorphisms in one of the largest cohorts that directly measured vasodilation responses to infused agonists. Nevertheless, the relatively small sample size represents a limitation since we may have been underpowered to detect an association, particularly in the black Americans studied. Moreover, due to the very low number of homozygous Arg55 and Gln287 carriers, we were only powered to characterize these relationships using a dominant genetic model of inheritance. In the bradykinin studies, there was excellent power to detect a 30% difference in FVR across the Lys55Arg (β=0.97) and Arg287Gln (β=0.94) genotype groups in the white American subset. The power to detect these differences in the black American subset was considerably less (Lys55Arg: β=0.67; Arg287Gln: β=0.52). We also made multiple comparisons and there is a possibility of false positive associations. In order to account for this possibility, we calculated a FDR q-value for each comparison. Although no gold-standard q-value threshold has been established to identify “true” associations, incorporation of this statistical approach into candidate gene association studies has become an increasingly recognized method to account for multiple comparisons and enhance confidence in observed associations.28 Since all q-values were estimated to be <0.08, we have a higher level of confidence in our reported findings. Importantly, validation in a well-powered, independent population will be ultimately necessary to confirm presence of a functional relationship between genetic variation in EPHX2 and vasodilator responses in humans.
Cytochrome P450 epoxygenases catalyze the biosynthesis of EETs, which are potent vasodilators that undergo rapid hydrolysis by sEH. Although associations between polymorphisms in EPHX2 and cardiovascular and cerebrovascular disease risk have been recently reported, their impact on vascular function has remained unknown. The current study demonstrates that genetic variation in EPHX2 is significantly associated with vasodilator responses in humans. Specifically, carriers of the increased function Arg55 and decreased function Gln287 variant alleles were predisposed to attenuated and enhanced forearm vasodilator responses, respectively. Collectively, these findings suggest that sEH has an important role in the regulation of vascular resistance and blood pressure in humans, and offer important mechanistic insight into previously reported associations between the EPHX2 Lys55Arg and Arg287Gln polymorphisms and cardiovascular and cerebrovascular disease risk. Pharmacological inhibition of sEH is currently under clinical development as a novel therapeutic strategy for the treatment of hypertension. These findings may ultimately help identify specific subsets of the population who may be more (Arg55 carriers) or less (Gln287 carriers) likely to respond to the anti-hypertensive effects of sEH inhibitors.
The authors gratefully acknowledge Delia Woods, RN for her nursing assistance and Jeff Petro, BS for his technical assistance.
SOURCES OF FUNDING
This publication was made possible by a Beginning Grant-in-Aid from the American Heart Association to Dr. Lee, grant HL085740 to Dr. Pretorius, funds from the Intramural Research Program of the NIH, National Institute of Environmental Health Sciences to Dr. Zeldin (Z01 ES025034), and grants 2P01 DK38226 and 5R01 HL065193 to Dr. Brown. Its contents are solely the responsibility of the authors and do not necessarily represent the official views of the NIEHS, NHLBI or NIH.
CONFLICT OF INTEREST DISCLOSURES
Dr. Zeldin is a co-inventor on U.S. Patent No. 6,531,506 B1 (issued March 11, 2003) titled “Inhibition of Epoxide Hydrolases for the Treatment of Hypertension,” U.S. Patent No. 6,693,130 B2 (issued February 17, 2004) titled “Inhibition of Epoxide Hydrolases for the Treatment of Hypertension,” and U.S. Patent No.6,916,843 B1 (issued July 12, 2005) titled “Anti-inflammatory Actions of Cytochrome P450 Epoxygenase-Derived Eicosanoids.” No other authors have conflicts of interest to disclose.