Sepsis and septic shock are the major causes of morbidity and mortality in critically ill patients. In general, two distinct cardiovascular stages of septic shock have been recognized. The initial phase (circulatory hyperdynamic phase) is characterized by an increased cardiac output and a low peripheral resistance, whereas the later phase (circulatory hypodynamic phase) is characterized by a diminished cardiac performance with low cardiac output associated with hypotension. Although the host mobilized a vast array of vasoconstricting and inotropic substances, they are ineffective in compensating the cardiovascular derangement. The loss in the vascular resistance takes place even with greatly increased endogenous production of vasoconstrictors. To make matters worse, the pressor response of septic patients to exogenously administered catecholamines and other vasoconstrictors is greatly reduced, characterizing the so-called hyporeactivity to vasoconstrictors that contributes to hypotension (1). Evidence from animal studies suggests that such situation results from over production of NO due to iNOS induction, which has been proposed as an important pathway leading to vascular hyporeactivity, hypotension, organ injury, and mortality in the bacterial LPS-induced shock models (2, 3). In addition, isolated blood vessels exposed to LPS in vitro or obtained from LPS-treated animals ex vivo also show diminished responses to vasoconstrictor agents, which is independent of NO (4, 5).
On the other hand, abnormal activation of the vascular K+ channel is now recognized as another major cause of hypotension and vascular hyporesponsiveness to catecholamines in septic shock (6). There are four types of K+ channels in vascular smooth muscle cells (7): (i) the voltage-dependent K+ (KV) channel opens with membrane depolarization and is preferentially inhibited by low doses of 4-aminopyridine (4-AP); (ii) the Ca2+-activated K+ (KCa) channel opens with membrane depolarization and with increases in the intracellular Ca2+ and is specifically inhibited by nanomolar amounts of the scorpion toxins charybdotoxin and iberiotoxin (IbTX) or apamin; (iii) the inward rectifier channel (KIR) is opened by membrane hyperpolarization and, in contrast with the other K+ channel, by increases in the extracellular K+ and is inhibited by low concentration of divalent cations such as Ba2+ or Mg2+; and (iv) the adenosine triphosphate (ATP)-sensitive K+ (KATP) channel opens when adenosine diphosphate levels rise and closes when ATP levels rise and is blocked by sulfonylureas (e.g., glibenclamide [GB]). Given the potassium gradient across the plasma membrane, activation of any of these channels allows potassium efflux, thus hyperpolarizing the plasma membrane and preventing calcium entry, causing vasodilatation or preventing the vasoconstrictor action of catecholamines.
Most NO physiological actions are mediated through activation of the soluble guanylyl cyclase (sGC) with the consequent increase in cyclic guanosine monophosphate (cGMP) levels, which activates cGMP-dependent protein kinase (PKG) (8). In addition, NO has been shown to open the large-conductance KCa (BKCa) channel directly (9), or indirectly by PKG, which phosphorylates to activate BKCa channels (10). Concerning pathological effects, abnormal activation of K+ channel through the NO pathway seems to be involved in the hyporesponsiveness to vasoconstrictors in isolated blood vessels obtained from endotoxemic animals (11). Evidence of membrane potential recording has also been reported in isolated rat aortas from the LPS animal model (12).
However, there are distinct differences in the pathophysiology of septic shock induced by LPS versus shock triggered by live polymicrobial sepsis because the actual factors causing human septic conditions are much more complicated than those just simply induced by LPS. One such model that is frequently used is cecal ligation and puncture (CLP) in rats. This model resembles human sepsis in several important aspects such as an early phase of hyperdynamic/hypermetabolic sepsis, followed by a late hypodynamic/hypometabolic phase (13, 14). There is still no directly electrophysiological evidence in vessels from the CLP animal model. Thus, the aim of the present study was to investigate the role of K+ channel in the changes of vascular tone in septic shock elicited by CLP. We examined the role of K+ channel in aortas from controls and CLP-treated rats by simultaneously measuring the changes of membrane potential and functional tension.
MATERIALS AND METHODS
CLP model of sepsis
Male Wistar-Kyoto (WKY) rats (10-12 weeks old; 250-350 g), whose stock originated from the Charles River Breeding Laboratories in Japan, were purchased from the National Laboratory Animal Center in Taiwan. This study was approved by the Institutional Animal Care and Use Committee of National Defense Medical Center according to the recommendations from Helsinki and the internationally accepted principles for the care and the use of experimental animals. Rats were anesthetized by intraperitoneal injection of sodium pentobarbital (40-50 mg/kg). The temperature was maintained at 24°C with an air conditioning system. The left carotid artery was cannulated (PE-50) and exteriorized to the back of the neck and connected to a pressure transducer (P231D; Statham, Oxnard, Calif) for the measurement of phasic blood pressure and MAP as well as heart rate (HR), which were displayed on a Gould model TA5000 polygraph recorder (Gould Inc., Valley View, Ohio). The left jugular vein was cannulated for the administration of drugs.
After the vascular cannulations, intraperitoneal sepsis was induced by a CLP surgery in part of rats as previous described (13, 14). Briefly, a small midabdominal incision was made, and the cecum was exposed. A distended portion of the cecum just distal to the ileocecal valve was isolated, filled with fecal content, and tied with a 3-0 silk suture in a manner not to disrupt bowel continuity. The ligated portion of the cecum was punctured twice with an 18-gauge needle, and a small amount of stool was expelled from the punctures to ensure leakage of the intestinal content. The cecum was then replaced in its original position within the abdomen, and the abdomen was then closed with a 3-0 suture in two layers, and the animals were allowed to recover and kept awake. In the sham-operated (SOP) rat, the cecum was exposed, manipulated, and returned to the peritoneal cavity without being punctured. Saline (4 mL/100 g body weight) was given subcutaneously to all rats at this time to prevent dehydration. On completion of the surgical procedure, cardiovascular parameters were allowed to stabilize for 15 to 20 min. Then, blood pressure and HR were monitored at 0, 9, and 18 h. In addition, after recording of baseline hemodynamic parameters, animals were given norepinephrine (NE; 1 μg/kg, i.v.) to examine the vascular reactivity to catecholamine. A pressor response profile was recorded when NE was administrated at each time point. The pressor responses to NE were reassessed at time 0 and 18 h after vehicle or CLP treatment and which were calculated as the ratio of integral area under the contraction curve (AUC; the area surrounded by contraction curve and baseline). The AUC at time 0 h in both groups was taken as 100%.
Membrane potential recording and tension measurement
At 18 h after the sham or CLP operation, animals were anesthetized with sodium pentobarbital (40-50 mg/kg), and thoracic aortas were obtained from these rats. The vessels were cleared of adhering periadventitial fat, and the thoracic aortas were cut in lengths of 3 mm and opened longitudinally. The endothelium was removed by gently rubbing the intimal surface of the vessel with a moistened cotton ball. Later, the lack of a relaxation to acetylcholine (1 μM) of aortic segments precontracted with NE (1 μM) was considered as evidence that the endothelium had been removed. The tissue was pinned down, intimal side upward, on the bottom of an organ chamber (capacity, 6.5 mL) and superfused at a constant flow rate of 3 mL/min with warmed (37°C), oxygenated (95% O2 and 5% CO2) Krebs solution (pH 7.4) consisting of 118 mM NaCl, 4.7 mM KCl, 1.2 mM KH2PO4, 1.17 mM MgSO4·7H2O, 2.5 mM CaCl2·2 H2O, 25 mM NaHCO3, and 11 mM glucose. One end of the segment was fixed to the organ bath chamber, whereas the other end was connected to a Grass FT03 transducer (Grass Instrument Co., Quincy, Mass). After the preparations had equilibrated for at least 60 min, glass microelectrodes filled with 3 M KCl (tip resistance, 10-30 MΩ) were inserted into the aortic smooth muscle from the intimal side. Electrical signals were detected by electrometer (Duo 773; World Precision Instruments Inc., Sarasota, Fla), monitored, and recorded continuously on a computer monitor oscilloscope (with Clampex 7 software). The effects of the following inhibitors were examined on the membrane potential and the vascular tension simultaneously: (1) NOS (Nω-nitro-l-arginine methyl ester [l-NAME], 0.3 mM), (2) sGC (1H-[1,2,4]oxadiazolo[4,3-a]quinoxalin-1-one [ODQ], 3 μM); GC (methylene blue [MB], 10 μM), (3) K+ channels [IbTX, 100 nM; 4-AP, 1 mM; apamin, 100 nM; Ba2+, 50 μM; GB, 10 μM; N-(1-adamantyl)-N′-cyclohexyl-4-morpholinecarboxamidine hydrochloride [PNU-37883A], 1 μM].
Each preparation was used to test only one inhibitor. An impalement was considered to be successful only if an electrical signal was maintained continuously before, during, and after the drug application. This continuous recording usually took 2 h (12).
All values in the text and figures are expressed as mean ± SEM of n observations, where n represents the number of animals studied. Statistical evaluation of membrane potential and vascular tension recording was performed by using paired and unpaired Student t test. A P value of less than 0.05 was considered to be statistically significant.
CLP-induced hypotension and vascular hyporeactivity in vivo
The mean baseline values for MAP, HR, and pressor responses to NE were not significantly different between CLP-treated and sham-treated (i.e., control) rats (Table 1). The CLP caused a significant fall of MAP at 9 and 18 h (Fig. 1) and was associated with a severe vascular hyporeactivity to NE in vivo. In addition, CLP induced a significant increase in HR during the experimental period.
Vascular hyporeactivity induced by CLP ex vivo
The NE-induced contraction in endothelium-denuded aortic strips from CLP rats was significantly suppressed when compared with that in the SOP group (Fig. 2). This reveals that the vascular hyporeactivity to NE (10−9-10−5 M) can last in aortas obtained from CLP rats ex vivo.
Effects of IbTX and apamin on the CLP-induced changes of membrane potential and vessel tension
A nanomolar concentration of IbTX (100 nM), which is considered to be selective to BKCa channels (15), reversed the hyperpolarization seen in the aortas obtained from CLP rats. However, this was not concomitant with an increase in tension in these preparations simultaneously. In addition, IbTX had no significant effect on both membrane potential and basal tension in aortas obtained from SOP rats (Fig. 3, A and C).
In the presence of apamin, a selective inhibitor of small-conductance KCa (SKCa) channels, it had no significant effect on membrane potential and basal tension in aortas obtained from both CLP and SOP rats (Fig. 3, B and D).
Effects of 4-AP and Ba2+ on the CLP-induced changes of membrane potential and vessel tension
In aortas from CLP-treated rats, but not from SOP-treated rats, incubation with the 4-AP (1 mM, a KV inhibitor) or Ba2+ (50 μM, a KIR inhibitor) reversed the hyperpolarization induced by CLP with an increase in the basal tension concomitantly (Fig. 4).
Effects of GB and PNU-37883A on the CLP-induced changes of membrane potential and vessel tension
GB (10 μM), a selective inhibitor of KATP channels, had no significant effect on membrane potential and basal tension in aortas obtained from CLP or SOP rats (Fig. 5, A and C). However, PNU-37883A (1 μM), a selective inhibitor of KATP channels by blocking the pore portion of the channels, had significant effect on membrane potential and basal tension in aortas obtained from CLP, but not from SOP rats. That is, PNU-37883A reversed the hyperpolarization induced by CLP with an increase in the basal tension simultaneously (Fig. 5, B and D).
Effects of l-NAME, ODQ, and MB on the CLP-induced changes of membrane potential and vessel tension
The results showed that neither l-NAME (0.3 mM) nor ODQ (3 μM) had significant effects on membrane potential and basal tension in aortas obtained from CLP or SOP rats (Fig. 6, A and D; B and E). In contrast, MB (10 μM), a nonspecific GC inhibitor, had significant effects on membrane potential and basal tension in aortas obtained from CLP, but not from SOP rats. That is, MB reversed the hyperpolarization induced by CLP with an increase in the basal tension simultaneously (Fig. 6, C and F).
In the present study, we are the first to provide direct evidence that vascular hyperpolarization recorded by membrane potential ex vivo involves abnormal activation of vascular K+ channels in septic rats induced by CLP. This is associated with functional changes in vascular tension. Hall et al. (4) demonstrate that tetraethylammonium chloride, a nonselective inhibitor of K+ channels, is able to fully reverse the LPS-induced impairment of phenylephrine contractions in isolated endothelium-denuded muscle strips of aorta. Similar findings were also observed in this peritonitis-induced sepsis model (data not shown). In addition, our previous studies have indicated that abnormal activation of K+ channels is associated with the vascular hyporeactivity-to-vasoconstrictor agents seen in LPS-induced septic shock (11, 12). Thus, we suggest that activation of vascular K+ channels can contribute to hypotension and hyporeactivity to vasoconstrictor agents in vascular smooth muscle tissues obtained from animals with peritonitis-induced septic shock.
It is unclear what kinds of K+ channels are involved in hypotension elicited by CLP-induced sepsis. Results in isolated blood vessels either exposed to LPS (4) or removed from endotoxemic animals (11, 12) reveal that the hyperpolarization (via activation of K+ channels) is partially reversed by l-NAME, ODQ, MB, charybdotoxin, GB, or tetraethylammonium chloride, whereas selective inhibitors of SKCa, KV, and KIR (i.e., apamin, 4-AP, and Ba2+, respectively) have no significant effects on this hyperpolarization, indicating that the hypotension caused by LPS involves activation of BKCa and KATP channels, but not of SKCa, KV, and KIR channels. In addition, this hyperpolarization seems to be associated with the NO/sGC pathway. In this CLP model, we demonstrated that the hyperpolarization caused by CLP was reversed by 4-AP, Ba2+, PNU-37883A, MB, and IbTX, but not by l-NAME, ODQ, apamin, and GB, suggesting that the hyperpolarization effect involves activation of KV, KIR, KATP, and BKCa channels. In addition, this hyperpolarization effect seems to be associated with the non-NO/non-sGC pathway.
There is one interesting result we would like to point out, which is that KATP channels are likely to play an important role in sepsis-induced vascular hyporeactivity and the development of septic shock induced either by LPS or CLP. The KATP channel is a complex of at least two proteins: (i) the sulfonylurea receptor and (ii) a pore-forming subunit belonging to Kir6.0 (16). In rat in vivo models of LPS-induced shock, the KATP channel inhibitor GB restores (or partially reverses) blood pressure without having any effect in control animals (17, 18). NO causes hyperpolarization via the activation of KATP channels in mesenteric arteries, which is blocked by GB (19). Our results showed that PNU-37883A, but not GB, reversed the hyperpolarization induced by CLP. Similar results have been demonstrated by the Buckley et al. (20) and O'Brien et al., showing that KATP channels play an important role in vascular hyporeactivity induced by LPS in vitro (21). It seems that only inhibitors of the pore-forming subunit, but not the sulfonylurea receptor subunit, are effective in reversing this hyporeactivity, suggesting that CLP induces rat aortic cytoskeletal disruption, resulting in a loss of high-affinity GB binding (22) and facilitating pore-forming subunit function. More importantly, a randomized, double-blind, placebo-controlled crossover pilot study of a potassium channel blocker in patients with septic shock showed that the KATP channel inhibitor GB failed to achieve a greater reduction in NE dose than placebo in septic shock patients, and in such patients, the blockade of KATP channels does not have a potent effect on vasomotor tone (23). However, it has been shown that endotoxemia for 6 h is associated with aortic (endothelium-denuded) membrane hyperpolarization, and this hyperpolarization is partially reversed by GB, which is also accompanied with an increase in the basal tension (12). Thus, we suggest that effects of GB ex vivo in the CLP model do not affect the vascular tone.
Enhanced formation of NO by iNOS has been shown to contribute to hypotension and vascular hyporeactivity to endogenous and exogenous vasoconstrictor agents in septic shock (24). NO and NO donors relax smooth muscle primarily by activating sGC and elevating intracellular cGMP level (25). In addition, NO and cGMP-elevating agents are also able to hyperpolarize some arteries by activating K+ channels. Evidence from patch-clamp studies in isolated smooth muscle cells demonstrates that NO activates KCa channels directly in rabbit aortas (9) and indirectly via cGMP-dependent kinase in rat pulmonary arteries (10). However, our study showed that the vascular hyperpolarization induced by CLP was not due to the NO/sGC pathway as it was in the LPS model. This is because neither l-NAME nor ODQ had significant effect on membrane potential and basal tension in aortas obtained from CLP rats. Fox et al. (26) have showed that after l-NAME infusion, the attenuated pressor response to phenylephrine in the pulmonary or the systemic circulation in CLP model was unchanged, suggesting that in rats with CLP-induced septic shock, excess NO is not an important mediator of the attenuated vascular reactivity observed in sepsis. A rat model of sepsis induced by CLP for 48 h manifests no beneficial effects on mortality with either NG-monomethyl l-arginine or the potent selective iNOS inhibitor S-(2-aminoethyl)-isothiourea (27). In addition, Vromen et al. (28) also report that mercaptoethylguanidine (a novel inhibitor with selectivity toward iNOS) or the non-isoform-specific inhibitor l-NAME is unable to restore contractility in vascular rings obtained at 18 h after CLP. It reveals that iNOS does not seem to play an important role in the delayed vascular hyporeactivity or mortality associated with CLP-induced sepsis. According to their suggestion, although iNOS is expressed, NO is produced at a relatively low level in this model of shock. Thus, iNOS may play little role in the vascular hyporeactivity induced by CLP.
In addition, our results showed that MB (a nonspecific inhibitor of GC) had significant effect on membrane potential and basal tension in aortas obtained from the CLP rat, suggesting that the hyperpolarization induced by CLP could be via the non-NO-sensitive GC pathway (e.g., particulate guanylyl cyclase) (29) because of the exclusion of NO-sensitive sGC pathway by ODQ. However, some studies report that MB also inhibits prostacyclin production (30), generates superoxide anions (31), and directly inhibits NOS (32). Indeed, we also tested effects of indomethacin (an inhibitor of cyclooxygenase; 10 μM) and tempol (an inhibitor of superoxide anions; 100 μM) on membrane potential and basal tension in aortas obtained from CLP or SOP rats and found that neither of these compounds had significant effects on membrane potential and basal tension in both groups (data not shown). Thus, prostacyclin and superoxide anions do not seem to involve this hyperpolarization. In addition, NO also does not contribute to this hyperpolarization induced by septic shock either because l-NAME has no effect on this vascular reactivity and hyperpolarization. However, our data illustrated that KV and KIR channels were activated in the CLP model. The KV channels may participate in the mechanism of action of both vasodilators and vasoconstrictors. Vasodilators that act via cyclic adenosine monophosphate signaling cascade may open these channels, and vasoconstrictors may close KV channels by mechanisms that involve elevated intracellular Ca2+ and protein kinase C (33). The role of KIR channel in sepsis is even less well understood. We suggest that KIR channels are activated by small increases in extracellular potassium, and indeed, mild hyperkalemia frequently happened in patients with septic shock (34). However, the role of KV and KIR channels in the pathogenesis of the hypotension of sepsis deserves scrutiny.
In summary, we have provided direct electrophysiological evidence showing that abnormal activation of the BKCa, KATP, KV, and KIR channels, but not SKCa, are likely to be involved in and play an important role in the hyperpolarization vessels, and this hyperpolarization can contribute to hypotension in rats with CLP-induced septic shock. In addition, our results suggest that this hyperpolarization induced by CLP may be via the non-NO-sensitive GC pathway.
The authors thank James Culpepper for critical reading of the article.
1. Assreuy J: Nitric oxide and cardiovascular dysfunction in sepsis. Endocr Metab Immune Disord Drug Targets
2. Parratt JR: Nitric oxide in sepsis and endotoxaemia. J Antimicrob Chemother
3. Wu CC: Possible therapies of septic shock: based on animal studies and clinical trials. Curr Pharm Des
4. Hall S, Turcato S, Clapp L: Abnormal activation of K+
channels underlies relaxation to bacterial lipopolysaccharide in rat aorta. Biochem Biophys Res Commun
5. Wu CC, Chen SJ, Yen MH: Nitric oxide-independent activation of soluble guanylyl cyclase contributes to endotoxin shock in rats. Am J Physiol
6. Oliver JA, Landry DW: Potassium channels
and septic shock. Crit Care Med
7. Nelson MT, Quayle JM: Physiological roles and properties of potassium channels
in arterial smooth muscle. Am J Physiol
8. Ignarro LJ: Signal transduction mechanisms involving nitric oxide. Biochem Pharmacol
9. Bolotina VM, Najibi S, Palacino JJ, Pagano PJ, Cohen R: Nitric oxide directly activates calcium-dependent potassium channels
in vascular smooth muscle. Nature
10. Archer SL, Huang JM, Hampl V, Neslon DP, Schultz PJ, Weir EK: Nitric oxide and cGMP cause vasorelaxation by activation of a charybdotoxin-sensitive K channel by cGMP-dependent protein kinase. Proc Natl Acad Sci U S A
11. Chen SJ, Wu CC, Yen MH: Role of nitric oxide and K+
-channels in vascular hyporeactivity induced by endotoxin. Naunyn Schmiedebergs Arch Pharmacol
12. Chen SJ, Wu CC, Yang SN, Lin CI, Yen MH: Abnormal activation of K+
channels in aortic smooth muscle of rats with endotoxic shock: electrophysiological and functional evidence. Br J Pharmacol
13. Wichterman KA, Baue A, Chaudry IH: Sepsis and septic shock: a review of laboratory models and a proposal. J Surg Res
14. Chen SJ, Wu CC, Yen MH: Alterations in vascular reactivity in intraperitoneal sepsis. J Cardiovasc Pharmacol
15. Suarez-Kurtz G, Garcia ML, Kaczorowski GJ: Effects of charybdotoxin and iberiotoxin on the spontaneous motility and tonus of different guinea pig smooth muscle tissues. J Pharmacol Exp Ther
16. Rodrigo GC, Standen NB: ATP-Sensitive potassium channels
. Curr Pharm Des
17. Wu CC, Thiemermann C, Vane JR: Glibenclamide-induced inhibition of the expression of inducible nitric oxide synthase in cultured macrophages and in the anaesthetized rat. Br J Pharmacol
18. Gardiner SM, Kemp PA, March JE, Bennett T: Regional haemodynamic responses to infusion of lipopolysaccharide in conscious rats: effects of pre- or post-treatment with glibenclamide. Br J Pharmacol
19. Murphy ME, Brayden JE: Nitric oxide hyperpolarizes rabbit mesenteric arteries via ATP-sensitive potassium channels
. J Physiol
20. Buckley JF, Singer M, Clapp LH: Role of KATP
channels in sepsis. Cardiovasc Res
21. O'Brien AJ, Thakur G, Buckley JF, Singer M, Clapp LH: The pore-forming subunit of the KATP
channel is an important molecular target for LPS-induced vascular hyporeactivity in vitro. Br J Pharmacol
22. Loffler-Walz C, Quast U: Disruption of the actin cytoskeleton abolishes high affinity 3
H-glibenclamide binding in rat aortic rings. Naunyn Schmiedebergs Arch Pharmacol
23. Warrillow S, Egi M, Bellomo R: Randomized, double-blind, placebo-controlled crossover pilot study of a potassium channel blocker in patients with septic shock. Crit Care Med
24. Thiemermann C: The role of the l-arginine: nitric oxide pathway in circulatory shock. Adv Pharmacol
25. Moncada S, Palmer RMJ, Higg EA: Nitric oxide: physiology, pathophysiology, and pharmacology. Pharmacol Rev
26. Fox GA, Paterson NA, McCormack DG: Effect of inhibition of NO synthase on vascular reactivity in a rat model of hyperdynamic sepsis. Am J Physiol
27. Strand OA, Kirkeboen KA, Giercksky KE: Inhibition of nitric oxide synthase does not affect survival in a rat model of abdominal sepsis. Scand J Clin Lab Invest
28. Vromen A, Arkovitz MS, Zingarelli B, Salzman AL, Garcia VF, Szabo C: Low-level expression and limited role for the inducible isoform of nitric oxide synthase in the vascular hyporeactivity and mortality associated with cecal ligation and puncture
in the rat. Shock
29. Hamad AM, Range S, Holland E, Knox AJ: Regulation of cGMP by soluble and particulate guanylyl cyclases in cultured human airway smooth muscle. Am J Physiol
30. Okamura T, Yoshida K, Toda N: Suppression by methylene blue of prostaglandin I2 synthesis in isolated dog renal arteries. J Pharmacol Exp Ther
31. Wolin MS, Cherry PD, Rodenburg JM, Messina EJ, Kaley G: Methylene blue inhibits vasodilation of skeletal muscle arterioles to acetylcholine and nitric oxide via the extracellular generation of superoxide anion. J Pharmacol Exp Ther
32. Mayer B, Brunner F, Schmidt K: Inhibition of nitric oxide synthesis by methylene blue. Biochem Pharmacol
33. Korovkina VP, England SK: Molecular diversity of vascular potassium channel isoforms. Clin Exp Pharmacol Physiol
34. Zaritsky JJ, Eckman DM, Wellman GC: Targeted disruption of Kir2.1
genes reveals the essential role of the inwardly rectifying K+
current in K+
-mediated vasodilatation. Circ Res