|Home | About | Journals | Submit | Contact Us | Français|
The cytochrome P450 epoxygenase (CYP)-derived metabolites of arachidonic acid the epoxyeicosatrienoic acids (EETs) and hydrogen peroxide (H2O2) both function as endothelium-derived hyperpolarizing factors (EDHFs) in the human coronary microcirculation. However, the relative importance of and potential interactions between these 2 vasodilators remain unexplored. We identified a novel inhibitory interaction between CYPs and H2O2 in human coronary arterioles, where EDHF-mediated vasodilatory mechanisms are prominent. Bradykinin induced vascular superoxide and H2O2 production in an endothelium-dependent manner and elicited a concentration-dependent dilation that was reduced by catalase but not by 14,15-epoxyeicosa-5(Z)-enoic acid (EEZE), 6-(2-propargyloxyphenyl)hexanoic acid, sulfaphenazole, or iberiotoxin. However, in the presence of catalase, an inhibitory effect of these compounds was unmasked. In a tandem-bioassay preparation, application of bradykinin to endothelium-intact donor vessels elicited dilation of downstream endothelium-denuded detectors that was partially inhibited by donor-applied catalase but not by detector-applied EEZE; however, EEZE significantly inhibited dilation in the presence of catalase. EET production by human recombinant CYP 2C9 and 2J2, 2 major epoxygenase isozymes expressed in human coronary arterioles, was directly inhibited in a concentration-dependent fashion by H2O2 in vitro, as observed by high-performance liquid chromatography (HPLC); however, EETs were not directly sensitive to oxidative modification. H2O2 inhibited dilation to arachidonic acid but not to 11,12-EET. These findings suggest that an inhibitory interaction exists between 2 EDHFs in the human coronary microcirculation. CYP epoxygenases are directly inhibited by H2O2, and this interaction may modulate vascular EET bioavailability.
The vascular endothelium releases numerous vasodilatory substances, including nitric oxide (NO), prostacyclin, and endothelium-derived hyperpolarizing factor (EDHF). EDHF is an important modulator of vasomotor tone, particularly in the microcirculation.1,2 Despite more than a decade of investigation, the chemical identity of EDHF remains controversial, and multiple EDHFs have been proposed. Among these, substantial evidence now points to epoxyeicosatrienoic acids (EETs),3–5 cytochrome P450 epoxygenase (CYP)-derived metabolites of arachidonic acid (AA), as EDHFs with another prominent candidate being the reactive oxygen species (ROS) hydrogen peroxide (H2O2)6,7
It has been reported that a significant inhibitory interaction occurs between NO and EET, in which NO inhibits CYP-mediated production of EET from AA.8 Such an interaction presumably exists to conserve vasodilator substances, while at the same time providing a compensatory mechanism of dilation when one is impaired. It is not known whether interactions occur among substances proposed as EDHFs, or whether multiple EDHFs can contribute to vasomotor regulation in a singular vascular bed. Published studies from our laboratory suggest that both EETs and H2O2 mediate dilation of human coronary arterioles (HCAs) in response to shear stress9,10; however, the relative importance of these EDHFs in receptor-mediated endothelium-dependent vasodilation and potential interactions between these EDHFs remain unexplored.
Because of the critical role of EDHFs in maintaining vasodilatory capacity when NO is impaired by oxidative stress,11 it is essential to determine whether and to what extent EDHF-mediated dilation is altered by ROS. It is also essential to explore whether interactions occur between EDHFs that modulate the bioavailability of these vasodilators. The present study addresses both of these issues by focusing on H2O2, a compound that is both a ROS and a putative EDHF. We provide initial evidence of a predominant role of H2O2 in receptor-mediated EDHF-dependent vasodilation and of an inhibitory interaction between 2 EDHFs (inhibition of CYP-mediated EET production by H2O2) in the human coronary microcirculation, where EDHF-mediated vasodilation is prominent. Specifically, we examined (1) whether EETs and/or H2O2 contributes to bradykinin (BK)-induced dilation, (2) whether H2O2 modulates the vascular effects of EET production, and (3) whether H2O2 directly inhibits EET synthesis by CYPs.
Right atrial appendages were obtained as freshly discarded surgical specimens from 74 patients undergoing cardiopulmonary bypass procedures,10 and 86 HCAs with a mean maximum internal diameter of 158±6 μm were used. All protocols were approved by the local Institutional Review Boards. Demographic data and diagnoses were obtained from hospital records at the time of surgery, as summarized in the Table.
Fluorescence detection of superoxide and H2O2 was performed using dihydroethidium (DHE) and 2′,7′-dichlorodihydrofluorescein diacetate (DCFH), respectively.9 Additional details are in the online data supplement.
Bioassay of transferable vasodilator factors was performed using pairs of isolated HCAs from a single patient sample that were cannulated in tandem in a heated (37°C) bioassay dual-chamber perfusion apparatus, as diagrammed in Figure 3B.12 Endothelium-intact “donor” arterioles were positioned upstream of endothelium-denuded “detector” arterioles in the presence of Nω-nitro-l-arginine methyl ester (L-NAME) (10−4 mol/L, NO synthase inhibitor) and indomethacin (10−5 mol/L, cyclooxygenase inhibitor).10 Inflow pipettes were connected to a gravity-feed reservoir that maintained perfusion pressure at 80 mm Hg, and arterioles were equilibrated for 60 minutes. In some experiments, catalase (1000 U/mL) was present in the donor bath and/or 14,15-epoxyeicosa-5(Z)-enoic acid (EEZE) (10−5 mol/L, EET antagonist5) was present in the detector bath. Endothelin-1 (5×10−10 to 10−9 mol/L) was added to the detector bath to adjust tone to 50% to 70% of passive diameter. To ensure adequate endothelial denudation of the detector vessel (see the online data supplement for denudation procedure), BK (10−6 mol/L) was added to the detector bath. Only vessels that dilated <5% were used for subsequent experiments. BK was then added to the donor bath, and internal diameter of the detector vessel was measured by videomicroscopy after reaching a new steady state (usually within 5 minutes following addition of BK). At the end of the experiments, maximum diameter was obtained by incubating detector vessels with papaverine (10−4 mol/L).9
Microsomal or in vitro metabolism of 14C-labeled AA or 3H-labeled 14,15-EET was analyzed by reverse-phase high-performance liquid chromatography (HPLC).5,13 Additional details are in the online data supplement.
Statistical analysis was performed as described in the online data supplement.
Previous studies indicate that both EETs and H2O2 function as EDHFs in HCAs in response to shear stress9,10; however, their contribution to agonist-induced vasodilation, if any, remains unexplored. To determine the relative contribution of EETs and H2O2 to BK-induced dilation, internal diameter was measured in isolated HCAs using videomicroscopy. As shown in Figure 1, BK elicited a concentration-dependent dilation that was not inhibited by EEZE (percentage of maximum dilation [%MD], 80±11 versus 94±1; –logEC50, 7.2±0.2 versus 7.6±0.1 with vehicle; n=5 and 31, respectively; P=NS), 6-(2-propargyloxyphenyl)hexanoic acid (PPOH) (selective CYP inhibitor14; %MD, 90±4; –logEC50, 7.4±0.2; n=7; P=NS), or sulfaphenazole (selective CYP2C inhibitor15; %MD, 87±5; –logEC50, 7.3±0.2; n=6; P=NS). Iberiotoxin, a specific inhibitor of large-conductance calcium-activated (BKCa) channels that reduces HCA dilation to exogenously applied EETs,16 also had no effect on BK-induced dilation (%MD, 86±8; –logEC50 7.7±0.1; n=6; P=NS). In contrast, catalase significantly reduced maximal dilation to BK (%MD, 61±4 [P<0.05]; –logEC50, 7.3±0.2 [P=NS]; n=21). BK-induced dilation was not influenced by sex, age, surgical procedure, or underlying disease. Taken together, these results suggest that H2O2, but not EETs, significantly contributes to EDHF-mediated dilation of HCAs to BK.
The inhibitory effect of catalase indirectly suggests that H2O2 mediates BK-induced dilation. To directly assess whether BK induces intracellular superoxide and H2O2 generation, semi-quantitative analysis of DHE and DCFH histofluorescence was performed. In HCAs, superoxide and H2O2 production was observed after incubation with BK (fluorescence ratio [versus control]: 2.66±1.17 and 3.21±0.97; n=7 for both, respectively; P<0.05), as shown in Figure 2, that was abolished by endothelial denudation (0.81±0.16 and 0.83±0.13; n=4, respectively; P<0.05 versus intact HCAs), suggesting that the endothelium is essential in this response. To determine whether the increase in DCFH fluorescence was attributable to H2O2 and not other peroxide species, additional studies were performed in the presence of catalase. Importantly, catalase completely blocked the increase in DCFH but not DHE fluorescence (0.79±0.11 and 1.99±0.85; n=4; P<0.05 and P=NS versus BK alone, respectively), indicating specificity for H2O2. These results are consistent with earlier reports indicating that catalase penetrates microvessels sufficiently to reduce intracellular DCFH fluorescence.9,17 Taken together, these results suggest that BK directly induces ROS production in an endothelium-dependent manner.
In the presence of catalase, a prominent residual dilation to BK occurs, indicating the presence of an additional vasodilator mechanism. To determine whether EETs contribute to residual BK-induced dilation when H2O2 is reduced, inhibitors were coadministered to HCAs with catalase (Figure 3A). Interestingly, the remaining dilation to BK was inhibited by EEZE (%MD, 18±7 versus 61±4; –logEC50, 6.5±0.1 versus 7.3±0.2 with catalase alone; n=5 and 21, respectively; P<0.05), sulfaphenazole (%MD, 9±2[P<0.05]; –logEC50, 6.8±0.3 [P=NS]; n=5), or iberiotoxin (%MD, 37±9; –logEC50, 6.7±0.1; n=5; P<0.05). Although each of these inhibitors had minimal effects on BK-induced dilation in the absence of catalase, a prominent inhibitory effect was unmasked in the presence of catalase. These results suggest that EETs mediate BK-induced dilation when H2O2 is removed and that H2O2 may modulate the bioavailability and/or action of EETs.
To determine whether vascular bioavailability of EETs is modulated by H2O2, bioassay of transferable vasodilator factors was performed in HCAs. BK produced detector vessel dilation when added to donor vessels (%MD, 56±4; n=5; P<0.05 versus vehicle), as shown in Figure 3B and 3C, but not when added directly to detectors (%MD, 3±1; n=5; P=NS versus vehicle; data not shown). Detector dilation to donor-applied BK was partially inhibited when catalase was added to the donor (%MD, 31±3; n=5; P<0.05 versus BK alone), and the residual dilation was abolished when EEZE was added to the detector (%MD, 7±6; n=5; P<0.05 versus BK+donor catalase). However, in the absence of donor-applied catalase, detector-applied EEZE had no effect (%MD, 53±8; n=5; P=NS versus BK alone). These results suggest that EETs represent transferable endothelium-derived vasodilators that act directly on vascular smooth muscle cells (VSMCs) and that their release is inhibited by endogenous H2O2.
To determine the mechanism underlying the interaction between endogenous H2O2 and EET-mediated dilation, we next examined whether H2O2 interferes with EET signaling by inhibiting CYP-mediated EET production, by directly oxidizing EETs, and/or by interfering with the action of EETs on VSMCs.
To determine whether H2O2 inhibits production of EETs by human CYPs, 14C-AA metabolism was assessed in microsomes overexpressing CYP2C9 or CYP2J2, 2 isoforms that are expressed in HCA endothelium.16 As shown in Figure 4, CYP2C9 epoxygenase activity was inhibited in a concentration-dependent fashion by H2O2 (IC50=13±3 μmol/L). Importantly, synthesis of dihydroxyeicosatrienoic acids was not enhanced by H2O2, indicating that H2O2 decreases EETs by interfering with their synthesis and not by enhancing their degradation to dihydroxyeicosatrienoic acids. Interestingly, CYP2J2 exhibited even greater sensitivity to H2O2 (IC50=0.3±0.05 μmol/L) than CYP2C9. These results indicate that H2O2 inhibits EET production by human CYPs in the low- to mid-micromolar range. To further explore the effect of H2O2 on CYPs, 14C-AA metabolism was assessed in microsomes from rat brain homogenates, a rich source of CYPs.18 Exogenous ROS generation with xanthine plus xanthine oxidase similarly inhibited EET production from rat CYPs (data not shown).
To determine whether H2O2 and/or superoxide directly oxidize EETs, in vitro metabolism of 3H-14,15-EET was assessed by HPLC. As shown in Figure 5, no leftward shift in the 14,15-EET peak occurred in the presence of H2O2 or the superoxide-generating system xanthine plus xanthine oxidase, indicating that 14,15-EET was not oxidized to 14,15-dihydroxyeicosatrienoic acid. 14C-AA was also not oxidized by H2O2 or superoxide (online data supplement), indicating that neither EETs nor AA are redox sensitive in vitro.
To indirectly determine whether H2O2 interferes with EET production, vasomotor responses to AA (an endothelium-dependent dilator of HCAs19) were measured using videomicroscopy. As shown in Figure 6, transient exposure to H2O2 significantly reduced maximal dilation to AA (%MD, 44±13 versus 86±9 [P<0.05[; –logEC50, 6.5±0.3 versus 7.0±0.3 [P=NS]; n=7). No change was observed in vehicle control studies on repeated application of AA (%MD, 84±10 versus 78±14; –logEC50 6.8±0.4 versus 6.8±0.3; n=4; P=NS; data not shown), indicating that tachyphylaxis was not responsible for the inhibitory effect of H2O2. To determine whether H2O2 interferes with the action of EETs on VSMCs, dilation to 11,12-EET was measured in endothelium-denuded HCAs. Importantly, 11,12-EET–induced dilation was not inhibited in these vessels by H2O2 (%MD, 68±7 versus 66±8; –logEC50 6.2±0.3 versus 6.0±0.2 with vehicle; n=6; P=NS), suggesting that H2O2 interferes with EETs at the level of endothelial EET production and not at the level of EET action on VSMCs.
This study is the first to investigate an inhibitory interaction between 2 EDHFs. The novel findings of the present study are three-fold. First, BK-induced vasodilation is predominantly mediated by endothelial production of H2O2 and not by EETs. Second, when H2O2 is reduced by catalase, a prominent residual dilation to BK is unmasked that is mediated by EETs. Third, production of EETs by CYP2C9 and CYP2J2 in vitro is directly inhibited by H2O2. Taken together, these data suggest that redox control of CYPs by H2O2 may modulate vascular bioavailability of EETs. This interaction between 2 EDHFs may play an important role in the modulation of vasomotor tone in the human coronary microcirculation.
Although a large body of evidence points to EETs as EDHFs,3–5 more recent work now implicates H2O2 as an EDHF as well.6,7,9,20 Although the existence of multiple EDHFs may be partially explained by differences between species or vascular beds, previous reports indicate that both EETs and H2O2 function as EDHFs in vasomotor regulation of HCAs by shear stress.9,10 The present results provide additional evidence that these EDHFs both play a physiological role in a singular vascular bed. These results also suggest an explanation for the seemingly disparate reports, implicating 1 or the other as the dominant EDHF; ie, the predominance of EETs or H2O2 as the mediator of the EDHF phenomenon may depend on the redox state of the endothelium.
Endothelial H2O2 likely arises by dismutation of superoxide, which may originate from numerous sources, including xanthine oxidase,21 NADPH oxidase,22 uncoupled endothelial NO synthase,21 mitochondria,23 and CYPs.24 In cardiovascular disease, several of these enzyme systems may be overly active21,22 or overexpressed,25 leading to elevated vascular ROS generation. Both flow-induced9 and BK-induced dilation of HCA are sensitive to inhibition by catalase, implicating a role for H2O2 in these responses. Inhibitors of the mitochondrial electron transport chain reduce flow-induced dilation of HCAs, indicating that H2O2 formation by shear stress requires mitochondria.23 In contrast, studies in mouse26 and human27 mesenteric arteries indicate that BK-induced H2O2 production requires Cu, Zn-superoxide dismutase, suggesting a cytosolic source of ROS in this context. Additional studies will therefore be necessary to identify the source of H2O2 in HCAs in response to BK.
Interestingly, flow-induced vasodilation is blocked by CYP inhibitors,10 unlike the results of the present study with BK, suggesting that differences in signaling and/or ROS generation exist between shear stress- and agonist-induced vasodilation. The effect of CYP inhibitors on flow-induced dilation is likely attributable to diminished production of EETs and not H2O2, as miconazole does not significantly reduce shear-induced H2O2 production in HCAs (H.M., unpublished observation). It is not known whether CYP is a functionally relevant source of H2O2 in response to BK, although the effects of CYP inhibitors on BK-induced dilation in the presence and absence of catalase indirectly suggest that CYP is a source of EETs and not H2O2 in this model. However, future studies will be necessary to directly determine the source(s) of BK-induced ROS production.
Interestingly, iberiotoxin did not inhibit BK-induced vasodilation in the absence of catalase. We observed previously that H2O2-induced dilation of HCAs is reduced by a combination of charybdotoxin and apamin,9 suggesting a role for K+ channels in this response; however, these inhibitors block not only BKCa channels, but also intermediate- and small-conductance KCa channels and voltage-dependent Kv channels.28 Although the present results with iberiotoxin indicate that H2O2-induced dilation does not require BKCa channels, other KCa channels and/or Kv channels may contribute to this response.
H2O2 is an important nonspecific redox modulator of cellular components, and elaborate antioxidant systems have evolved to minimize ambient H2O2, including catalase and glutathione peroxidase. Although less reactive than superoxide or other ROS, H2O2 is capable of inactivating enzymes directly by oxidizing essential thiol groups, as occurs with glyceraldehyde-3-phosphate dehydrogenase.29 H2O2 also reduces the activity of some heme-containing enzymes, including CYP2B4 (a rabbit P450 monooxygenase),30 by directly oxidizing and degrading the heme prosthetic group into monopyrrole and dipyrrole fragments that irreversibly bind to the protein.31 Interestingly, P450BM-3 (a bacterial CYP) is highly unstable in the presence of H2O2.32 Although the mechanism underlying this effect is uncertain, the redox sensitivity of the enzyme depends on a phenylalanine residue (Phe87) that extends into the heme pocket,33 suggesting an interaction at or near the heme moiety. The present study indicates that H2O2 directly inhibits human CYPs, but the specific underlying mechanisms and redox sites remain to be determined.
It has been reported that BK induces endothelial release of 1.5 μmol/L H2O2 from porcine coronary microvessels,20 a concentration that inhibits CYP2J2 but not CYP2C9 in vitro. However, measurement of extracellular H2O2 may underestimate acute H2O2 production in endothelial cells, as intracellular antioxidant systems and extracellular dilution of H2O2 by the perfusate may reduce the measurable extracellular concentration of H2O2. Indeed, intracellular antioxidants consume approximately 85% of H2O2 given to Jurkat T-cells within a few seconds,34 an antioxidant capacity similar to that of human vascular endothelial cells.35 Therefore, intracellular H2O2 may acutely rise as high as 10 μmol/L in endothelial cells stimulated with BK, a concentration within the physiological range.36 In addition, H2O2 production may be greater in our model than in porcine coronary arterioles, because HCAs were obtained from patients with established cardiovascular disease, in whom ROS production may be enhanced.37 Therefore, it is likely that intracellular H2O2 is sufficiently high in HCAs to inhibit both CYP2J2 and CYP2C9. This hypothesis is indirectly supported by our results obtained with sulfaphenazole, where CYP2C-mediated dilation is seen only in the presence of catalase. However, determination of the physiologically relevant concentration range of intracellular H2O2 in endothelial cells of HCAs will require development of new experimental techniques, because accurate quantification of H2O2 within these cells is difficult using currently available assays.
An inherent limitation of the current study is the lack of tissue from subjects who are completely free of disease, because such tissue is rarely obtainable. However, our model also presents the unique advantage of being able to study EETs and H2O2 in the setting of chronic cardiovascular disease that cannot be adequately mimicked in animal models. In HCAs from subjects with coronary artery disease, endothelium-dependent dilation is mediated predominantly by EDHF, with little contribution by NO or prostacyclin10; therefore, such tissue facilitates evaluation of EET and H2O2 as EDHFs in a clinically relevant setting, with minimal confounding influence of other vasodilator factors. Although all experiments were performed in the presence of L-NAME and indomethacin to minimize NO and prostacyclin, respectively, it is possible that residual NO and/or prostacyclin may influence EDHF interactions in vivo.
Although the present study suggests that H2O2 directly inhibits CYPs, the highly reactive hydroxyl radical, formed from H2O2 by the iron-dependent Fenton reaction, may mediate this effect. Although iron-chelating agents are commonly used to assess the contribution of hydroxyl radical by preventing the Fenton reaction, deferoxamine abolished baseline CYP activity in our model (data not shown), thus preventing us from exploring this possibility. In addition, whereas EETs and AA are resistant to oxidation by H2O2 in vitro, hydroxyl radical may oxidize these lipids in vivo, where iron is present. Alternative methods will therefore be necessary to evaluate the contribution of hydroxyl radical, if any, in this model.
An important limitation of the present study is the lack of direct EET measurements. Although numerous attempts were made to quantify EETs released from isolated HCAs using liquid chromatography electrospray–ionization mass spectrometry or HPLC, we were unable to quantify EETs in a consistent and reproducible manner. This may be attributable to an insufficient number of endothelial cells in a single arteriole (5-mm average length, 100- to 200-μm internal diameter) and the limited number of arterioles that can be isolated from a single patient sample. Detection of EETs may also be hindered in these vessels by ambient ROS, which may be elevated in cardiovascular disease.37 EET production has been demonstrated from animal microvessels12,38; however, these vessels were obtained from animals without cardiovascular disease. Not surprisingly, we were unable to demonstrate EET production from human microvascular or coronary artery endothelial cell cultures, a finding that is consistent with earlier reports that endothelial cells rapidly lose CYP expression and activity in culture.39
An alternative explanation is that HCAs may not, in fact, produce EETs. Although this possibility cannot be ruled out, we believe this possibility is less likely, because HCAs produce a transferable vasodilator factor in a bioassay setup that is blocked by EEZE, a structural analog and antagonist of EET. Previous studies indicate that CYP2C9 and CYP2J2, as well as the EET-metabolizing soluble epoxide hydrolase (sEH) enzyme, are robustly expressed in HCA endothelium,16 suggesting that HCAs are capable of not only producing EETs but also modulating their bioavailability. In addition, production of 11,12-EET has been demonstrated from diseased human left internal mammary arteries,40 indicating that the human vasculature is capable of producing EETs in the milieu of cardiovascular disease.
With the expanding body of evidence implicating EETs as endogenous vasculoprotective agents,41 considerable efforts have been made to identify the factors that determine EET bioavailability. Much has been learned about EET metabolism by sEH, β-oxidation, or esterification into the phospholipid membrane41,42; however, less is known about the factors that regulate EET production. The present study demonstrates that EET generation can be modulated by H2O2, an interaction that may increase in importance during cardiovascular disease, when ROS production is elevated. Although sEH inhibitors are now recognized as a potential therapeutic approach to enhance EET-mediated vasculoprotection,43–45 the present study suggests that the efficacy of these agents may be limited by the redox status of the vasculature. Targeted antioxidant therapy, in conjunction with sEH inhibitors, may be beneficial to the diseased vasculature not only by limiting the proinflammatory effects of H2O2 but also by enhancing the bioavailability of vasculoprotective EETs.
The present study supports a role for EETs and H2O2 as EDHFs in the human coronary microcirculation and suggests that an inhibitory interaction exists between them. Human CYPs are directly inhibited by H2O2 in a concentration-dependent manner. This effect of H2O2 may modulate vascular EET bioavailability.
We thank the Division of Cardiothoracic Surgery at the Medical College of Wisconsin, the Cardiothoracic Surgery Group of Milwaukee, the Cardiovascular Surgery Associates of Milwaukee, the Midwest Heart Surgery Institute, and the Wisconsin Heart Group for providing surgical specimens.
Sources of Funding
This work was supported by Predoctoral and Postdoctoral Fellowship Awards (to B.T.L. and A.S., respectively) and a Beginning Grant-in-Aid (to H.M.) from the American Heart Association; a Veterans Administration merit award (to D.D.G.); NIH grants HL80173 (to H.M.), HL68769 (to D.D.G.), HL51055 (to W.B.C.), and DK38266 (to J.R.F.); a grant from the Advancing a Healthier Wisconsin Endowment Fund (to H.M.); and the Robert A. Welch Foundation (to J.R.F.). In addition, this work was supported by intramural funds from the National Institute of Environmental Health Sciences (to D.C.Z.).