Proinflammatory Mediators Chronically Downregulate the Formation of the Endothelium-Derived Hyperpolarizing Factor in Arteries Via a Nitric Oxide/Cyclic GMP–Dependent Mechanism
Background—Endothelium-dependent dilator responses mediated by NO and endothelium-derived hyperpolarizing factor (EDHF) are altered in arteriosclerosis and sepsis. The possibility that proinflammatory mediators that stimulate the expression of inducible NO synthase (NOS II) affect the generation of EDHF was examined in isolated arteries.
Methods and Results—Under combined blockade of NOS and cyclooxygenase, EDHF-mediated relaxation elicited by several agonists was significantly attenuated in rabbit carotid and porcine coronary arteries exposed to cytokines and lipopolysaccharide. The blunted relaxation was coincident with NOS II expression and was prevented by inhibition of NOS II as well as of global protein synthesis. The NO donor CAS 1609 and 8-bromo-cGMP mimicked the proinflammatory mediator effect. In contrast, long-term blockade of endothelial NO generation increased the relaxation in carotid but not in coronary arteries. Proinflammatory mediators reduced the synthesis of EDHF assessed as hyperpolarization of vascular smooth muscle cells elicited by the effluent from bradykinin-stimulated coronary arteries. Proinflammatory mediators induced NOS II expression in cultured endothelial cells and decreased the expression of cytochrome P450 enzymes, which are the most probable candidates for the synthesis of EDHF.
Conclusions—Proinflammatory mediators inhibit the formation of EDHF in isolated arteries. This impairment is coincident with NOS II expression in the arterial wall and seems to be mediated through the induced generation of NO, which downregulates the putative EDHF-forming enzyme. Thus, a decreased formation of EDHF may contribute to the endothelial dysfunction in arteriosclerosis and sepsis.
Endothelial cells play a pivotal role in the local control of vascular tone by the release of potent short-lived vasoactive autacoids. Among the best-characterized autacoids is NO generated by the activation of NO synthase in endothelial cells (NOS III), which mediates endothelium-dependent responses in most isolated arteries and vascular beds.1 However, inhibition of NOS III only slightly attenuates endothelium-dependent relaxations in some arteries.2 These relaxations are mediated by an endothelial factor(s) distinct from NO, called endothelium-derived hyperpolarizing factor (EDHF), that hyperpolarizes the vascular smooth muscle, most likely by opening calcium-sensitive potassium channels.2 Although the chemical nature of EDHF is still unclear, the findings that inhibitors of phospholipase A2 and of cytochrome P450 diminished EDHF-mediated relaxation and hyperpolarization in the coronary microcirculation and macrocirculation and also in carotid arteries suggest that in these arteries, EDHF is a cytochrome P450 monooxygenase–derived arachidonic acid metabolite.2 3 4 5 6
Endothelial dysfunction is considered to be an important factor in the pathogenesis of atherosclerosis and sepsis.7 Lipopolysaccharide (LPS), interleukin-1β (IL-1β), and tumor necrosis factor-α (TNF-α), the generation of which is increased during endotoxemia and arteriosclerosis, severely blunted endothelium-dependent relaxations in both in vitro and in vivo studies,8 9 10 suggesting that these proinflammatory mediators may contribute to the development of endothelial dysfunction. Bioassay studies have indicated that a decreased generation of endothelium-derived NO is responsible for the impaired endothelium-dependent relaxation in atherosclerotic and cytokine-treated arteries.10 11 Among the potential mechanisms proposed to account for the reduced formation of NO are a downregulation of NOS III by proinflammatory mediators12 and oxidized LDLs,13 a decreased NOS III activity due to the formation of an endogenous inhibitor such as NG,NG′-dimethyl-l-arginine14 or due to the cytokine-induced sustained generation of NO that affects the signal transduction cascade in endothelial cells,10 and an excessive inactivation of NO by oxygen-derived radicals such as superoxide anions.15
Although cytokines and LPS can severely attenuate the generation of NO, the effect of these proinflammatory mediators on the endothelial generation of EDHF remains to be clarified. Therefore, the present study was designed to test whether proinflammatory mediators affect EDHF-mediated relaxation and hyperpolarization in isolated arteries, and if so, to elucidate the underlying mechanisms.
IL-1β was obtained from Collaborative Research; oxyhemoglobin from Calzyme Laboratories; bradykinin from Bachem Biochemica; recombinant TNF-α from Boehringer-Ingelheim; pentobarbital sodium from Sanofi; diclofenac from Novartis; NG-nitro-l-arginine (L-NNA) from Serva; minimum essential medium containing Earle’s salts and FCS from PAN Systems; penicillin, streptomycin, medium M199, dispase, and bovine albumin fraction V from Gibco BRL; 11,12-epoxyeicosatrienoic acid (11,12-EET) from Biomol; and all other chemicals from Sigma Chemical Co. U46619 (9,11-dideoxy-11α,9α-epoxymethano-prostaglandin F2α) was provided by Upjohn; S-methylisothiourea (SMT) by Dr Garry Southan (Frederick Cancer Research, Frederick, Md); and CAS 1609 (4-hydroxymethyl-furoxan-3-carboxamide) by Hoechst Marion Roussel.
Preparation of Blood Vessels
Carotid arteries were obtained from New Zealand White rabbits (anesthetized with sodium pentobarbital 60 mg/kg IV) and coronary arteries from pig hearts, placed into ice-cold Krebs-Henseleit solution (composition in mmol/L: NaCl 144.0, KCl 5.9, CaCl2 1.6, MgSO4 1.2, KH2PO4 1.2, NaHCO3 25.0, and d-glucose 11.1) containing diclofenac 1 μmol/L, and prepared for experimentation as described previously.4 10 The expression of NOS II was elicited by incubating carotid artery rings in culture medium [minimum essential medium containing 2 mmol/L glutamine, 5 mmol/L TES, 5 mmol/L HEPES (the latter 2 both at pH 7.3), 50 U/mL penicillin, 50 μg/mL streptomycin, 0.1% BSA, and 1 μg/mL polymyxin B] in the presence of IL-1β 100 U/mL for 7 hours and coronary artery rings in culture medium without polymyxin B in the presence of a combination of TNF-α 1000 U/mL, interferon-γ (IFN-γ) 500 U/mL, and LPS 10 μg/mL for 15 hours in a cell culture incubator.
Porcine Aortic Endothelial Cells
Endothelial cells were isolated as reported previously,3 seeded onto Petri dishes coated with fibronectin, and grown in medium M199 containing 10% FCS with antibiotics. The expression of NOS II was elicited by incubating confluent primary cultures of endothelial cells in medium M199 containing antibiotics and 0.1% FCS in the presence of TNF-α 1000 U/mL, IFN-γ 500 U/mL, and LPS 10 μg/mL for 24 hours in a cell culture incubator.
Rat Aortic Smooth Muscle Cells
Smooth muscle cells (SMCs) were isolated and cultured as described previously.3 All experiments were performed on cells between passages 12 and 16 seeded on glass coverslips.
Organ Chamber Studies
Arterial rings were set up in organ chambers containing warm (37°C) and oxygenated (95% O2/5% CO2) Krebs-Henseleit solution and prepared for experimentation as described previously.4 10 Carotid arteries were constricted with phenylephrine 1 to 3 μmol/L and coronary arteries with the thromboxane A2 mimetic U46619, 0.1 to 0.3 μmol/L. When tension stabilized, the presence of a functional endothelium was demonstrated by the relaxation to acetylcholine 1 μmol/L in carotid arteries and to bradykinin 0.1 μmol/L in coronary arteries. After washout, both preparations were allowed to equilibrate for 30 minutes in the presence of either L-NNA 0.1 mmol/L or oxyhemoglobin 0.1 μmol/L. Thereafter, the rings were constricted again with phenylephrine and U46619, respectively, before a cumulative concentration-relaxation curve to the test compound.
Detection of EDHF Release
The release of EDHF was detected by recording changes in membrane potential of cultured rat aortic SMCs exposed to the effluent of a perfused endothelium-intact porcine coronary artery segment as described previously.3
Assay of Cytochrome P450 Monooxygenase Activity
Cytochrome P450 monooxygenase–dependent metabolic activity was assayed as the dealkylation of 7-ethoxyresorufin in cultured endothelial cells. Endothelial cells were incubated with 7-ethoxyresorufin 3.4 μmol/L for 15 minutes at 37°C. Thereafter, the resorufin formed in the supernatant was measured at an excitation of 522 nm and emission at 586 nm. Preliminary experiments demonstrated that the cell-mediated conversion of 7-ethoxyresorufin to resorufin was linear over a period of ≥30 minutes and was abolished by the P450 inhibitor clotrimazole 3 μmol/L.6
Expression of Cytochrome P450 Monooxygenase and NOS II Protein
Endothelial cells were lysed in double-distilled water by 5 cycles of freeze/thaw before the addition of homogenization buffer [Tris-HCl 50 mmol/L (pH 7.4), KCl 1.15%, EDTA 1 mmol/L, glucose 5 mmol/L, phenylmethylsulfonyl fluoride 4.4 mg/L, and 1 mg/L each of leupeptin, pepstatin A, trypsin inhibitor, antipain, chymostatin, and aprotinin]. Cell homogenates were centrifuged at 5000g for 10 minutes. The supernatant was removed and subjected to a 1-hour centrifugation at 100 000g. The partially purified microsomal fraction was resuspended in a buffer containing Tris-HCl 50 mmol/L (pH 7.4), glycerin 10%, EDTA 0.1 mmol/L, and the protease inhibitors and was used for the detection of cytochrome P450 proteins, whereas the supernatant was used for the detection of NOS II protein as described previously.10 Cytochrome P450 immunoreactivity was detected by use of a polyclonal rabbit antibody directed against cytochrome P450 2C11 (which recognizes most members of the cytochrome P450 2C family; dilution 1:5000, provided by Dr E. Morgan, Atlanta, Ga) and NOS II immunoreactivity by use of a polyclonal rabbit antibody directed against NOS II (dilution 1:1000, provided by Dr J. Pfeilschifter, Frankfurt am Main, Germany). The autoradiographs were analyzed by scanning densitometry (ImageMaster, Pharmacia).
Tissue Content of cGMP
Isobutylmethyl xanthine 500 μmol/L was added to the coronary artery rings during the last 30 minutes of the 15-hour incubation period. Thereafter, arterial rings were homogenized in trichloroacetic acid (6%) at 4°C. After centrifugation (10 minutes at 5000g), the supernatant was extracted with water-saturated ethylether, and the cGMP content in each sample was determined with a radioimmunoassay including an acetylation step.
Results are shown as mean±SEM of n experiments. The negative logarithm of the concentration of U46619 causing 50% contraction (EC50) was calculated for each concentration-response curve. Statistical analyses were performed with Student’s paired t test (2-tailed) or an ANOVA followed by Fisher’s protected least significant difference test. A value of P<0.05 was considered to be statistically significant.
Proinflammatory Mediators Impair EDHF-Mediated Relaxation
In rabbit carotid arteries, the L-NNA/diclofenac–resistant EDHF-mediated relaxations to acetylcholine, substance P, and A23187 were markedly attenuated by the IL-1β treatment (100 U/mL for 7 hours; Figure 1⇓). Similar findings were obtained with acetylcholine when the generation of NO was abolished by oxyhemoglobin 0.1 μmol/L instead of L-NNA (maximal relaxations were reduced from 45.2±3.6% to 25.2±3.5%, n=7). The treatment with proinflammatory mediators (TNF-α 1000 U/mL, IFN-γ 500 U/mL, and LPS 10 μg/mL for 15 hours) also strongly attenuated the EDHF-mediated relaxation to bradykinin and, to a smaller but significant extent, that to A23187, whereas the relaxation to 11,12-EET, an activator of calcium-sensitive potassium channels,5 was unaffected in porcine coronary arteries (Figures 2⇓ and 3⇓).
Both functional and biochemical studies were next performed to demonstrate that the proinflammatory mediator treatment induced the generation of NO in coronary arteries. Exposure of endothelium-intact coronary arteries to TNF-α, IFN-γ, and LPS for 15 hours was accompanied by an attenuation of the contractions to U46619 (EC50 changed from 0.15±0.03 to 0.56±0.04 μmol/L, and the maximal contraction was reduced from 18.9±0.8 to 8.1±0.9 g, n=7) and by an increase in the tissue content of cGMP (from 11.5±3.5 to 162.4±21.0 pmol/mg protein, n=5). Because both of these effects were prevented by inhibitors of NOS (EC50 was 0.36±0.07 μmol/L and the maximal contraction was 17.6±0.9 g in the presence of proinflammatory mediators and 10 μmol/L SMT, and the cGMP levels were 10.1±3.0 pmol/mg protein in the presence of proinflammatory mediators and 600 μmol/L L-NNA), they can reasonably be attributed to an induced generation of NO. The expression of NOS II in carotid artery rings exposed to IL-1β for 7 hours has been confirmed in a previous study.10
Next, the role of the proinflammatory mediator–induced generation of NO in the attenuated EDHF-mediated relaxation was examined. EDHF-mediated relaxations to acetylcholine were unaffected in carotid arteries exposed to IL-1β for only 15 minutes (Figure 4a⇓) and also in those that had been incubated with IL-1β for 7 hours in combination with either cycloheximide (Figure 4b⇓) or N-α-tosyl-l-lysine chloromethylketone (TLCK, Figure 4c⇓); these treatments have been shown to prevent the induced generation of NO.16 17 The cycloheximide treatment alone slightly but significantly increased, whereas that of TLCK alone significantly attenuated, the EDHF-mediated relaxation (Figure 4b⇓ and 4c⇓). Moreover, EDHF-mediated relaxations to acetylcholine were slightly but significantly restored in carotid arteries exposed to IL-1β in combination with SMT (used at a concentration that abolished the activity of NOS II, whereas that of NOS III was only minimally affected)10 for 7 hours (Figure 5⇓).
To further assess the possibility that chronic exposure of arteries to NO impairs the EDHF-mediated relaxation, the effects of both the basal generation of NO by the endothelium and exogenous activators of the NO-cGMP effector pathway were investigated. Blockade of basal endothelial NO generation by L-NNA for several hours increased EDHF-mediated relaxations in carotid but not in coronary arteries (Figure 6⇓). In contrast, long-term exposure of carotid and coronary arteries to either CAS 1609 or 8-bromo-cGMP significantly attenuated EDHF-mediated relaxation (Figures 6⇓ and 7⇓).
Proinflammatory Mediators Reduce the Synthesis of EDHF
Addition of bradykinin to the superfusate from control coronary arteries stimulated the release of EDHF, as indicated by a pronounced but transient hyperpolarization of the SMCs (Figures 8⇓ and 9⇓). This response to bradykinin was significantly reduced by the treatment of coronary arteries with either a combination of TNF-α, IFN-γ, and LPS; CAS 1609; or 8-bromo-cGMP for 15 hours (Figures 8⇓ and 9⇓). Direct application of bradykinin to the SMCs resulted in only a transient membrane depolarization (13±2 mV, n=9).
Proinflammatory Mediators Decrease Cytochrome P450 Monooxygenase Expression in Endothelial Cells
Exposure of cultured endothelial cells to a combination of TNF-α, IFN-γ, and LPS for 24 hours decreased cytochrome P450 dealkylase activity significantly, by 39.6±5.1% (n=7). L-NNA 300 μmol/L slightly but significantly attenuated the inhibitory effect of the proinflammatory mediator treatment (to 29.7±4.6%), whereas the NOS inhibitor alone had only minor effects. Both CAS 1609 (100 μmol/L, n=7) and S-nitrosoglutathione (100 μmol/L, which was added 3 times during a 24-hour incubation period, n=2) also reduced cytochrome P450 dealkylase activity by 23.9±3.8% and by 38.9±8.7%, respectively. Expression of cytochrome P450 2C proteins, a major family of cytochrome P450 monooxygenases in endothelial cells,18 was decreased by the proinflammatory mediator treatment (Figure 10A⇓), and this effect was associated with the expression of NOS II protein (Figure 10B⇓). A decreased cytochrome P450 2C protein level was also found in CAS 1609–treated endothelial cells (Figure 10A⇓).
Proinflammatory mediators have been shown to severely attenuate endothelium-dependent relaxations both in vitro and in vivo.8 9 10 This altered endothelial cell function has been shown to reflect a reduced ability of endothelial cells to generate NO10 11 and, as indicated by the present study, probably also EDHF. Indeed, the proinflammatory mediator treatment evoked in isolated carotid and coronary arteries a pronounced decrease of the component of the endothelium-dependent relaxation that is resistant to NOS and cyclooxygenase inhibitors and that has been attributed to a diffusible vascular smooth muscle hyperpolarizing factor. The impaired relaxation cannot be explained by a decreased smooth muscle response to EDHF because relaxations to 11,12-EET, an activator of calcium-sensitive potassium channels,5 were unaffected by the proinflammatory mediator treatment. Moreover, the blunted release of EDHF from endothelium-intact donor coronary arteries, as demonstrated under bioassay conditions, suggests that proinflammatory mediators most likely interfere with the synthesis and/or release of EDHF rather than its mechanism of action at the vascular smooth muscle.
The mechanism underlying the inhibitory effect of proinflammatory mediator treatment appears to be dependent on the synthesis of a peptide/protein in the arterial wall, which in turn dampens the biosynthesis of EDHF in endothelial cells. Indeed, short-term exposure of arteries to proinflammatory mediators failed to attenuate EDHF-mediated relaxations, and inhibition of de novo protein synthesis with cycloheximide fully prevented the inhibitory effect of the proinflammatory mediator treatment. In addition, because the cycloheximide treatment increased EDHF-mediated relaxations in arteries not exposed to proinflammatory mediators, the biosynthesis of EDHF is likely to be constantly subjected to an intrinsic inhibitory effect that requires de novo protein synthesis.
In parallel to the attenuation of endothelium-dependent relaxations, the proinflammatory mediator treatment caused the induction of NOS II in the arterial wall.8 10 The NOS II–derived NO appears to mediate the inhibitory effect of proinflammatory mediators on the biosynthesis of EDHF partly via a cGMP-dependent mechanism, because inhibition of either NOS II expression or activity significantly restored EDHF-mediated relaxations, and the proinflammatory mediator effect was mimicked by the chronic activation of the NO-cGMP effector pathway in endothelium-intact arteries.19 Moreover, because blunted EDHF-mediated responses were still observed in arteries examined 2 hours after the washout of the NO donor, NO may not only directly inhibit EDHF formation19 20 but also probably reduce the level of the EDHF-producing enzyme and/or induce the generation of an endogenous inhibitor. Such an inhibitory effect seems to be effective even in normal arteries, because interruption of the basal release of NO significantly increased EDHF-mediated relaxations in carotid arteries not exposed to proinflammatory mediators. The absence of such an increase in coronary arteries is probably due to the relatively small basal NO release in endothelium-intact coronary arteries,21 which may not be sufficient to affect the biosynthesis of EDHF.
Although the chemical identity of EDHF is still unclear, a number of studies support the hypothesis that EDHF is a cytochrome P450 monooxygenase–derived metabolite of arachidonic acid in the coronary microcirculation and macrocirculation as well as in rabbit carotid arteries but not in the guinea pig carotid and rat hepatic artery.2 3 4 5 6 22 23 Endothelial cells are capable of metabolizing arachidonic acid via the cytochrome P450 monooxygenase pathway to 4 major epoxyeicosatrienoic acids (5,6-EET, 8,9-EET, 11,12-EET, and 14,15-EET) that have been shown to elicit both hyperpolarization and relaxation of vascular SMCs.4 5 Thus, cytochrome P450 monooxygenases are a potential target for the inhibitory effect of the proinflammatory mediator treatment on the biosynthesis of EDHF. Consistent with such a hypothesis, chronic exposure of cultured endothelial cells to proinflammatory mediators decreased cytochrome P450 dealkylase activity that was coincident with NOS II expression and was mimicked by NO donors. Because inhibition of NOS partially but significantly prevented the proinflammatory mediator effect, the induced formation of NO, as well as other mechanisms, seems to be implicated. Such an inhibitory effect of NO could reflect a direct inhibition of the heme-containing cytochrome P450 monooxygenases24 but also a downregulation of their protein level.25
In conclusion, proinflammatory mediators cause a chronic attenuation of EDHF-mediated relaxation via a decreased formation of this hyperpolarizing factor. This attenuation is coincident with the expression of NOS II in the arterial wall and appears to be due to a downregulation of the EDHF-forming enzyme, partly by the induced formation of NO. Such a mechanism may contribute to the development of endothelial dysfunction in pathological states such as sepsis and atherosclerosis.
This study was supported by grants from the Deutsche Forschungsgemeinschaft (Schi 389/2-3 and Po 521/1).
Reprint requests to V.B. Schini-Kerth, PhD, Institut für Kardiovaskuläre Physiologie, Klinikum der J.W.G.-Universität, Theodor-Stern-Kai-7, D-60590 Frankfurt/Main, Germany.
- Received September 16, 1998.
- Revision received November 18, 1998.
- Accepted December 7, 1998.
- Copyright © 1999 by American Heart Association
Hecker M, Bara A, Bauersachs J, Busse R. Characterization of endothelium-derived hyperpolarizing factor as a cytochrome P450-derived arachidonic acid metabolite in mammals. J Physiol (Lond). 1994;48:407–414.
Campbell WB, Gebredhin D, Pratt PF, Harder DR. Identification of epoxyeicosatrienoic acids as endothelium-derived hyperpolarizing factors. Circ Res. 1996;78:415–423.
Vanhoutte PM. Endothelial dysfunction and atherosclerosis. Eur Heart J. 1997;18:E19–E29.
Julou-Schaeffer G, Gray GA, Fleming I, Schott C, Parratt JR, Stoclet JC. Loss of vascular responsiveness induced by endotoxin involves the L-arginine pathway. Am J Physiol. 1990;259:H1038–H1043.
Aoki N, Siegfried M, Lefer AM. Anti-EDRF effect of tumor necrosis factor in isolated, perfused cat carotid arteries. Am J Physiol. 1989;256:H1509–H1512.
Kessler P, Bauersachs J, Busse R, Schini-Kerth VB. Inhibition of inducible NO synthase restores the impaired endothelium-dependent relaxations in isolated arteries exposed to proinflammatory mediators. Arterioscler Thromb Vasc Biol. 1997;17:1746–1755.
Yoshizumi M, Perella MA, Burnett JC, Lee M. Tumor necrosis factor downregulates an endothelial nitric oxide synthase mRNA by shortening its half-life. Circ Res. 1993;73:205–209.
Liao JK, Shin WS, Lee WY, Clark SL. Oxidised low-density lipoprotein decreases the expression of endothelial nitric oxide synthase. J Biol Chem. 1995;270:319–324.
Mügge A, Elwell JH, Peterson TE, Hofmeyer TG, Heistad DD, Harrison DG. Chronic treatment with polyethylene-glycolated superoxide dismutase partially restores endothelium-dependent vascular relaxations in cholesterol-fed rabbits. Circ Res. 1991;69:1293–1300.
Schini-Kerth VB, Boese M, Busse R, Fisslthaler B, Mülsch A. N-α-Tosyl-l-lysine chloromethylketone prevents the expression of the iNOS in vascular smooth muscle by blocking activation of NF-κB. Arterioscler Thromb Vasc Biol. 1997;17:672–679.
Durante W, Schini VB, Scott-Burden T, Junquero DC, Kroll MH, Vanhoutte PM, Schafer AI. Platelet inhibition by an L-arginine-derived substance released by IL-1β-treated vascular smooth muscle cells. Am J Physiol. 1991;261:H2024–H2030.
Olmos L, Mombouli J-V, Illiano S, Vanhoutte PM. cGMP mediates the desensitization to bradykinin in isolated canine coronary arteries. Am J Physiol. 1991;268:H865–H870.
Bauersachs J, Popp R, Hecker M, Sauer E, Fleming I, Busse R. Nitric oxide attenuates the release of endothelium-derived hyperpolarizing factor. Circulation. 1996;94:3341–3347.
Khatsenko OG, Gross SS, Rifkind AB, Vane JR. Nitric oxide is a mediator of the decrease in cytochrome P450-dependent metabolism caused by immunostimulants. Proc Natl Acad Sci U S A. 1993;90:11147–11151.
Carlson TJ, Billings RE. Role of nitric oxide in the cytokine-mediated regulation of cytochrome P-450. Mol Pharmacol. 1996;49:796–801.