We assessed mechanisms of acetylcholine- and bradykinin-induced relaxations in human omental resistance vessels. Ring segments (approximately 200 μm normalized ID) were dissected from omental biopsies obtained from women at laparotomy (nonpregnant) or at cesarean delivery (pregnant) and were studied under isometric conditions in a Mulvany-Halpern myograph. All arginine vasopressin-preconstricted vessels relaxed in a strictly endothelium-dependent manner to acetylcholine and bradykinin; maximal relaxations were not decreased by either NG-nitro-l-arginine or indomethacin. By contrast, bradykinin failed to relax vessels that had been preconstricted with potassium gluconate. In the combined presence of NG-nitro-l-arginine and indomethacin, addition of charybdotoxin, a selective antagonist of some calcium-sensitive potassium channels, did not inhibit maximal bradykinin-induced relaxation. By contrast, addition of 10 mmol/L tetraethylammonium chloride abolished relaxation in vessels from nonpregnant women but not in vessels from gravidas. We conclude that bradykinin relaxes these human resistance arteries in an endothelium-dependent but predominantly nitric oxide- and prostanoid-independent manner; relaxation likely depends on the action of an endothelium-derived hyperpolarizing vasodilator. Furthermore, in striking contrast to mechanistic insights from animal studies, human pregnancy appears to augment a mechanism of endothelium-dependent relaxation in these vessels that is insensitive to the inhibitors noted above. Whether a similar novel vasodilator mechanism in vivo contributes to the physiological vasodilation that characterizes human gestation or whether failure of such a mechanism might lead to preeclampsia remains the subject of future study.
Many pharmacological and physiological vasodilator stimuli relax systemic arteries in an endothelium-dependent manner through the release of an NO-containing endothelium-derived relaxing factor.1 More recently, it has been shown that endothelium-dependent relaxation in a variety of in vitro vascular preparations depends importantly on contributions of an endogenous hyperpolarizing vasodilator acting at charybdotoxin-sensitive vascular smooth muscle potassium channels.2 3 Much of these mechanistic data derive from animal models; there is very little information with regard to the human vasculature, particularly resistance-sized arteries. The need for such mechanistic information is especially important given the known variability in endothelium-dependent vasodilation among species, arteries from different vascular beds, or even vessels of differing caliber within the same animal model.1 2
We explored endothelium-dependent vasodilation in human omental resistance arteries, focusing on the action of NO-independent relaxing factors. We performed experiments in both nonpregnant and pregnant subjects because of our prior observation that acetylcholine-induced relaxation of rat mesenteric microvessels was augmented during pregnancy because of a selective increase in NO but not in possible hyperpolarizing vasodilator-dependent mechanisms.4 We hypothesized that vasodilator-stimulated relaxation of human omental small arteries would be similarly augmented in normal pregnancy because of increased contributions of endothelium-derived NO to maximal relaxation. Therefore, we assessed acetylcholine- and bradykinin-induced relaxations of these vessels in the absence and presence of inhibitors of NO synthesis, of prostaglandin synthesis, and of potassium channel-mediated hyperpolarization.
All protocols were in accord with institutional guidelines and approved by the local institutional review board. Normotensive gravidas (n=12; age, 27±6 years; gestational age, 37.3±2.3 weeks; blood pressure, 115±9/67±7 mm Hg) at the completion of uncomplicated pregnancies and without significant past medical histories, granted written informed consent for omental biopsy at the time of planned cesarean delivery. Likewise, premenopausal women (n=12; age, 37±4 years; blood pressure, 126±9/69±5 mm Hg) without hypertension, diabetes, known cardiovascular disease, or prior history of pregnancy complications, granted similar consent for omental biopsy during gynecologic laparotomy.
Microvessel Preparation and Tension Measurements
After resection, a portion of omentum was placed immediately into cold, oxygenated, physiologic salt solution (PSS) consisting of (mmol/L) NaCl 118.2, NaHCO3 24.8, KCl 4.6, KH2PO4 1.2, MgSO4 1.2, CaCl2 2.5, and dextrose 10.0. Vessels were dissected sharply under a microscope and cleaned of adherent adipose and connective tissues. Arterial segments were cut into two 0.5-mm-long rings that were then mounted on 16-μm wires in a Mulvany-Halpern myograph (Living Systems Instrumentation). The myograph bath contained PSS at 37°C bubbled with 5% CO2 in oxygen to maintain a pH of 7.4. After 1 hour of equilibration, including six bath changes, each ring underwent two conditioning stretches of 0.6 mN/mm. Passive tension-internal circumference characteristics were then determined and each ring set to a normalized internal diameter equivalent to that which would have resulted at a transmural pressure of 100 mm Hg (mean normalized diameters of 213±30 and 210±20 μm, nonpregnant and pregnant, respectively). Preliminary studies in vessels from both groups demonstrated maximal contractile responses to AVP with complete preservation of endothelium-dependent relaxation when resting vessel dimensions were set in this manner.
Drugs and Chemicals
Acetylcholine chloride, AVP, bradykinin, SNP, indomethacin, TEA, and NNLA were obtained from Sigma Chemical Co. Charybdotoxin was obtained from Accurate Chemical. Stock solutions of reagents were prepared fresh daily in distilled water and diluted serially for use in myograph baths. All concentrations refer to final bath concentrations.
Vessels were contracted twice by exposure to PSS in which 60 mmol/L potassium gluconate was substituted for an equimolar concentration of NaCl. Then, on the basis of preliminary dose-response studies, rings were submaximally preconstricted with either 10 nmol/L AVP or 125 mmol/L potassium gluconate-substituted PSS. Endothelium-dependent relaxation was then assessed after application of maximally effective concentrations of acetylcholine (10 μmol/L) or bradykinin (10 μmol/L); these doses were likewise chosen from preliminary dose-response studies. Relaxation of AVP- or potassium gluconate-preconstricted rings was also assessed in the presence of NNLA (100 μmol/L) and/or indomethacin (10 μmol/L), inhibitors of arginine-dependent NO biosynthesis and cyclooxygenase, respectively. Likewise, bradykinin-induced relaxation in the combined presence of NNLA and indomethacin was reassessed in the presence of 10 nmol/L charybdotoxin, 1 mmol/L TEA, or 10 mmol/L TEA; these inhibitor concentrations all fully inhibit the NNLA-insensitive component of acetylcholine-induced relaxation in rat mesenteric microvessels. Endothelium-independent relaxation was assessed in some rings by addition of 10 μmol/L SNP. Finally, some rings were mechanically denuded of endothelium by passage of a human hair through the vessel lumen, and the relaxation protocols were repeated.
In preliminary studies, some arterial rings were studied in the presence of higher concentrations of NNLA (1 mmol/L, n=3) or methylene blue (10 μmol/L, n=3); neither maneuver altered either contraction or relaxation compared with results obtained in the presence of 100 μmol/L NNLA. Also, time controls were carried out in paired vessel rings to confirm the stability and reproducibility of AVP-induced preconstriction as well as the reproducibility of repeated acetylcholine- and bradykinin-induced relaxations (in the absence of inhibitors).
Data are presented as mean±SD. Data from multiple rings from the same subject were averaged; n refers to the number of subjects studied. Relaxation was calculated as the fractional decrement in contraction induced by the preconstrictor agent before addition of a vasodilator. Group data were compared by paired and unpaired t tests, with correction for multiple comparisons, as appropriate. A value of P<.05 was considered significant.
All AVP-preconstricted vessels from either pregnant or nonpregnant women relaxed completely and in a strictly endothelium-dependent manner (Fig 1⇓) in response to either acetylcholine or bradykinin (Table 1⇓). By contrast, bradykinin failed to relax vessels that had been preconstricted with potassium gluconate rather than with AVP (Table 1,⇓ Fig 2⇓). Although there was an apparent tendency toward greater SNP-induced relaxation of vessels from gravidas after potassium gluconate preconstriction (Fig 2⇓), conclusions regarding this observation were limited by the small number of studies including this combination of agents (n=3 nonpregnant, n=9 pregnant).
Inhibition of NO synthase with NNLA, inhibition of cyclooxygenase with indomethacin, or the combination of NNLA plus indomethacin failed to significantly inhibit maximal relaxation induced by either acetylcholine or bradykinin (Table 1⇑). In the continued presence of NNLA plus indomethacin, the further addition of either charybdotoxin (10 nmol/L, Table 2⇓) or in a few additional studies, of 1 mmol/L TEA (maximal relaxation: 92%, pregnant, n=1; 94.7±3.6% nonpregnant, n=3), both selective inhibitors of Ca2+-activated potassium channels, had no effect on bradykinin-induced relaxation. However, 10 mmol/L TEA, a less-selective potassium channel blocker, significantly antagonized bradykinin-induced relaxation as well as reversed already established relaxation in vessels from nonpregnant women, although it was also without effect in vessels from normal gravidas (Fig 3,⇓ Table 2⇓).
On the basis of previous studies in rat mesenteric resistance vessels, we had hypothesized both NO-mediated and NO-independent contributions to maximal endothelium-dependent relaxation of small human omental arteries. We further anticipated that the latter NO-independent mechanism would depend on a factor that opens charybdotoxin-sensitive potassium channels and that the former NO-dependent mechanism would be augmented in normotensive human pregnancy. Indeed, several other lines of evidence had pointed to the notion of augmented NO synthesis during rat pregnancy, including increments in urinary excretion of NO metabolites and cGMP and the apparent ability of NO synthase inhibition to reverse the expected gestational refractoriness to the pressor effect of infused vasoconstrictors.5 6 By contrast with results in the rat, our current results suggested no important contributions by either endothelium-derived NO or a factor acting at charybdotoxin-sensitive potassium channels to either maximal acetylcholine- or bradykinin-induced relaxation in these human vessels. We cannot yet comment on the role of NO in the response to intermediate concentrations of either agonist, as these were not evaluated. However, previous studies in human arteries from other vascular beds, in which NO plays an obvious contributory role in the response to endothelium-dependent vasodilators (see below), have suggested that reduced responses to maximally effective concentrations of vasodilators accompany any decrements in response to intermediate concentrations of vasodilators after NO synthase inhibition.7 8 Therefore, we consider it unlikely that NO plays a significant role in mediating endothelium-dependent relaxation of these small omental arteries.
Our results also revealed an unexpected effect of gestation, in that it rendered bradykinin-induced relaxation insensitive to apparent inhibition by nonselective potassium channel blockade with 10 mmol/L TEA. Vessels all failed to relax when preconstricted with depolarizing concentrations of potassium gluconate, suggesting important contributions by a hyperpolarizing vasodilator. However, the actual mediator and mechanism of endothelium-dependent vasodilation in these human omental microvessels remain to be identified.
Most previous studies have suggested that endothelium-dependent relaxation depends on the action of an NO-containing vasodilator that acts predominantly via stimulation of smooth muscle guanylyl cyclase.1 Although endothelium-derived hyperpolarizing factors, some apparently acting at (glyburide-sensitive) ATP-gated potassium channels, have been well described in many conduit artery preparations, they have not been thought to contribute importantly to vascular relaxation.9 Recent work has demonstrated important contributions of an NO-independent endothelium-derived vasodilator, mediating agonist-induced relaxation in rat mesenteric microvessels,3 rat femoral and intrarenal microvessels,10 11 rabbit abdominal but not thoracic aorta,12 porcine and bovine coronary arteries,13 and others.2 On the basis of the use of selective antagonists, many of these latter hyperpolarizing vasodilators appear to act by facilitating potassium efflux through Ca2+-activated potassium channels in a manner similar to the effect of 11,12-epoxyeicosatetraenoic acid, a cytochrome P-450-derived arachidonic acid metabolite.2 11 14 Whether this metabolite is identical with the hyperpolarizing vasodilator in all of the above vascular preparations remains to be determined.
Given the variations in mechanism of endothelium-dependent arterial relaxation among species, in vascular beds, and in vessel caliber,3 15 there has been a striking paucity of studies focused on the mechanisms of endothelium-dependent vasodilation of human vessels. Deng et al16 reported that contributions by potassium conductance-dependent hyperpolarizing vasodilator(s) exceeded those by NO in mediating acetylcholine-induced relaxation of subcutaneous resistance vessels obtained by gluteal biopsy. There was significant interindividual heterogeneity in the contributions of NO and of hyperpolarizing factors to the substance P-induced relaxation of human pial arteries.17 Likewise, bradykinin-induced relaxation of human coronary arteries (in vitro) and of dorsal hand veins or forearm resistance vessels (in vivo) was only partly mediated by NO- or cGMP-dependent mechanisms, the hand vein preparation exhibiting contributions by a hyperpolarizing factor.18 19 20 To our knowledge, our data are the first to explore mechanisms of endothelium-dependent relaxation in the otherwise well-characterized21 omental resistance artery preparation.
Aalkjaer and colleagues22 previously observed diminished angiotensin-induced contraction of omental vessels obtained from normotensive gravidas compared with those obtained from nonpregnant women. However, they did not examine any aspect of endothelial function. Recently, McCarthy and colleagues compared acetylcholine-induced relaxation of subcutaneous resistance vessels from normal gravidas and preeclamptic women7 and from nonpregnant women and normal gravidas,8 noting no significant alteration of acetylcholine-induced relaxation due to pregnancy. They further reported that 57% and 32% to 36% of maximal acetylcholine-induced relaxation was insensitive to inhibition by NG-nitro-l-arginine methyl ester plus indomethacin in nonpregnant women and normal gravidas, respectively.7 8 Our data demonstrated no effect of NO synthase inhibition on maximal acetylcholine- and bradykinin-induced relaxations in omental resistance vessels from both gravid and nonpregnant women; however, all relaxations were inhibited by depolarization with potassium gluconate. Endothelial depolarization would be expected to impair NO synthesis, along with any other process depending on calcium influx into endothelial cells. Since our results were obtained in the presence of both indomethacin and NNLA, we believe that they are more consistent with the effect of a hyperpolarizing vasodilator acting at vascular smooth muscle. Since 10 mmol/L TEA inhibited bradykinin-induced relaxations only in vessels from nonpregnant women, we infer that pregnancy augmented an alternative mechanism of endothelium-dependent relaxation in these vessels. Whether similar augmentation of a TEA-insensitive vasodilator mechanism in vivo contributes to the physiological vasodilation that characterizes human gestation or whether failure of such a mechanism might lead to preeclampsia, a uniquely human hypertensive disorder, remains the subject of future study.
Selected Abbreviations and Acronyms
This work was supported in part by grants from the American Heart Association of Metropolitan Chicago, Baxter Healthcare, National Institutes of Health (HD-31939, HL-48302), and the Mother's Aid Research Fund of Chicago Lying-In Hospital. I.F.P. was the recipient of a postdoctoral fellowship from the Conselho Nacional de Desenvolvimento Cientifico e Tecnologico, Brazil (201635/92-1), and of a Merck Young Investigator award for presentation of this work at the 49th Annual Scientific Sessions of the Council for High Blood Pressure Research. J.G.U. is the recipient of a career development award in clinical pharmacology from the Pharmaceutical Research and Manufacturers of America Foundation. We thank Marshall D. Lindheimer for his support, encouragement, and helpful comments.
Reprint requests to Jason G. Umans, MD, PhD, Section of Nephrology, Department of Medicine, University of Chicago, 5841 S Maryland Ave, MC-5100, Chicago, IL 60637.
Previously published in abstract form (Hypertension. 1995;26:568).
- Received September 21, 1995.
- Revision received November 14, 1995.
- Revision received March 12, 1996.
Hwa JJ, Ghibaudi L, Williams P, Chatterjee M. Comparison of acetylcholine-dependent relaxation in large and small arteries of rat mesenteric vascular bed. Am J Physiol.. 1994;266:H952-H958.
Pascoal IF, Lindheimer MD, Nalbantian-Brandt C, Umans JG. Contraction and endothelium-dependent relaxation in mesenteric microvessels from pregnant rats. Am J Physiol. 1995;269:H1899-H1904.
Conrad KP, Joffe GM, Kruszyna H, Kruszyna R, Rochelle LG, Smith RP, Chavez JE, Mosher MD. Identification of increased nitric oxide biosynthesis during pregnancy in rats. FASEB J.. 1993;7:566-571.
McCarthy AL, Taylor P, Graves J, Raju SK, Poston L. Endothelium-dependent relaxation of human resistance arteries in pregnancy. Am J Obstet Gynecol. 1994;171;1309-1315.
Umans JG, Lindheimer MD, Yamasaki M. Vasoconstriction and endothelium-dependent vasodilation in femoral microvessels from gravid and virgin rats. J Am Soc Nephrol. 1992;3:556. Abstract.
Cowan CL, Palacino JJ, Najibi S, Cohen RA. Potassium channel-mediated relaxation to acetylcholine in rabbit arteries. J Pharmacol Exp Ther.. 1993;266:1482-1489.
Nagao T, Illiano S, Vanhoutte PM. Heterogeneous distribution of endothelium-dependent relaxations resistant to NG-nitro-L-arginine in rats. Am J Physiol. 1992;263:H1090-H1094.
Nakashima M, Mombouli JV, Taylor AA, Vanhoutte PM. Endothelium-dependent hyperpolarization caused by bradykinin in human coronary arteries. J Clin Invest. 1993;92:2867-2871.
Taddei S, Mattei P, Virdis A, Sudano I, Ghiadoni L, Salvetti A. Effect of potassium on vasodilation to acetylcholine in essential hypertension. Hypertension. 1994;23:485-490.