Kidney International, Vol. 57 (2000), 2502–2510

Lipopolysaccharide impairs endothelial nitric oxide synthesisin rat renal arteries

HARRO A. PIEPOT, CHRISTA BOER, A.B. JOHAN GROENEVELD, ANTONIE A. VAN LAMBALGEN,and PIETER SIPKEMA

Departments of Physiology, Anesthesiology, and Internal Medicine (Intensive Care Unit), Institute for Cardiovascular Research,Vrije Universiteit, Amsterdam, The Netherlands

Lipopolysaccharide impairs endothelial nitric oxide synthesis Acute renal failure is characterized by a decreased glo-in rat renal arteries. merular filtration rate following decreased renal perfu-

Background. Impaired endothelium-dependent vasodilation sion, which indicates that the renal circulation does notmay contribute to hypoperfusion and failure of abdominal or-contribute to the systemic vasodilation that characterizesgans, including the kidneys during endotoxin or septic shock.human septic shock [1]. The pathophysiology of acuteIn this study, the short-term (2 h) effects of bacterial lipopoly-

saccharide (LPS) on endothelium-dependent vasodilation in renal failure remains incompletely understood. Vaso-rat renal and superior mesenteric arteries were documented. pressor-resistant systemic hypotension during human

Methods. Rat renal and mesenteric arteries were dissected septic shock and in experimental models of shock isand exposed in vitro to LPS for two hours. The effects of LPSbelieved to be mediated by excessive production of theon vascular reactivity were determined and compared withendogenous vasodilator nitric oxide (NO) derived fromtime-matched controls. Endothelial nitric oxide (NO) release

was determined using an NO microsensor in adjacent vessel the conversion of l-arginine to l-citrulline and NO bysegments. the inducible, calcium/calmodulin-independent NO syn-

Results. LPS impaired maximal acetylcholine (ACh)-inducedthase (iNOS) [2–4]. However, it is unlikely that expres-endothelium-dependent vasodilation in renal arteries (62.5 6sion of iNOS and the resulting excessive NO production8.8% vs. 34.4 6 7.5% in controls and LPS-exposed arteries),

but not in mesenteric arteries. LPS did not alter the sensitivity are primarily involved in the development of acute renalof renal arteries to exogenous NO. ACh-dependent vasodila- failure, since acute renal failure is caused by early renaltion was abolished after blocking NO synthesis with 1024 mol/L vasoconstriction rather than vasodilation [1]. In addition,L-NA in control and LPS-incubated renal arteries. When com-

renal vasoconstriction, hypoperfusion, and a decreasedpared with controls, NO release induced by ACh and the recep-glomerular filtration rate have been observed in ratstor-independent calcium ionophore A23187 was significantly

decreased (P , 0.05) in LPS-exposed renal segments and was within two hours after an injection of bacterial lipopoly-fully abolished in endothelium-denuded segments, indicating saccharide (LPS), the pathogenic agent of septic shockthat LPS attenuated receptor-dependent as well as receptor-

[5–8]. Therefore, renal failure may precede the inductionindependent endothelial NO release. In contrast, ACh- andof iNOS, which requires at least four hours to becomeA23187-induced NO release was normal in LPS-exposed mes-

enteric arteries. expressed [9]. Moreover, LPS-induced renal failure is notConclusions. These results indicate that LPS-induced selec- ameliorated by specific inhibition of the iNOS enzyme

tive impairment of ACh-induced endothelium-dependent re- [10, 11]. This may relate to findings that renal NO duringlaxation in rat renal arteries is caused by decreased endothelialendotoxemia protects against intraglomerular thrombo-NO release. This may contribute to the propensity for acutesis, renal vasoconstriction, and renal dysfunction [12, 13].renal failure during septic shock.

Direct vasoactive effects of LPS in the renal circulationmay provide a possible mechanism for the development

Acute renal failure is a frequent complication of hu- of renal vasoconstriction. Indeed, local infusion of LPSman septic shock and carries a high mortality rate [1]. to an in vitro perfused juxtamedullary nephron induced

vasoconstriction in preglomerular arterioles within onehour [14]. Local production of thromboxane A2, prosta-Key words: renal failure, inflammation, vascular smooth muscle, endo-

thelium, hypoperfusion, septic shock. glandin F2a, and endothelins have been implicated as localmediators of renal vasoconstriction during endotoxemiaReceived for publication June 29, 1999[1]. Furthermore, endothelial dysfunction may also con-and in revised form November 17, 1999

Accepted for publication January 17, 2000 tribute to renal vasoconstriction during acute renal failure.Under normal conditions, small (picomolar) quantities of 2000 by the International Society of Nephrology

2502

Piepot et al: LPS inhibits renal endothelial NO release 2503

NO, released by the constitutive endothelial calcium/ Tissue culturecalmodulin-dependent type of NOS (eNOS), relax the The tissue culture method used for endothelium-intactunderlying vascular smooth muscle cells and contribute arterial segments was adapted from that previously de-to endothelium-dependent regulation of renal vascular scribed for isolated arterial segments [27]. Culture platestone [13, 15]. Impaired endothelium-dependent vasodi- (6 wells) were filled with 3.0 mL sterile Dulbecco’s modi-lation has been found within one to two hours after in fied Eagle’s medium (DMEM) containing 4 mmol/Lvivo endotoxemia in humans [16]. Furthermore, in ani- l-glutamine and 22.7 mmol/L glucose. Shortly before themal models of septic shock as well as after prolonged in start of the incubations, the DMEM was supplementedvitro exposure of isolated aorta, impaired endothelium- with 100 IU · mL21 penicillin, 100 mg · mL21 streptomy-dependent vascular relaxation was demonstrated as well cin, and 4% heat inactivated fetal calf serum (FCS).[17–21]. Mechanisms implicated in LPS-mediated impair- Segments of both the renal artery and the mesentericment in endothelium-dependent vasodilation include in- artery were incubated in DMEM containing Escherichiahibition of eNOS activity and/or down-regulation of coli LPS (50 mg · mL21, serotype O127:B8), which waseNOS expression [22–25]. Acute LPS-mediated vasocon- present throughout the incubation period; time-matchedstriction and organ hypoperfusion are not restricted to control segments were incubated in the same medium

without LPS. Arterial segments were maintained duringthe renal circulation. Another target during endotoxintwo hours at 378C in an atmosphere of 95% humidifiedshock is the splanchnic circulation. Both animal modelsair and 5% CO2.of shock and human septic shock are characterized by

mesenteric hypoperfusion, partly caused by acute vaso-Isometric force measurementsconstriction in the splanchnic circulation [26]. Therefore,

Paired control and LPS-incubated arterial segmentswe hypothesized that impaired endothelium-mediatedfrom renal and mesenteric arteries were suspended hori-vascular relaxation may contribute to acute hemody-zontally in organ chambers of two dual-wire myographsnamic alterations and vasoconstriction in the renal circu-between a displacement device and an isometric forcelation as well as in the splanchnic circulation.transducer (Kistler Morse no. 46-1003-01) coupled to aTherefore, the goal of the present study was to investi-computer to visualize isometric force responses usinggate the effects of short-term (2 h) exposure to LPS ondata sampling software that was adapted at our technicalendothelium-dependent vasodilation in isolated rat renaldepartment (NewXLC). The organ chambers were filledand superior mesenteric arteries. The effects of LPS werewith gassed (95% air 1 5% CO2) Krebs solution com-studied using an in vitro approach to avoid the confoundingposed of (in mmol/L): NaCl 110, KCl 5.0, CaCl2 · 2H2Oinfluences of systemic hemodynamic effects of LPS. Direct2.5, MgSO4 · 7H2O 1.0, KH2PO4 1.0, NaHCO3 24.0, glu-measurements of agonist-stimulated NO release from vas-cose 10.0, l-arginine 1.0, and EDTA 0.02. After a 30-cular endothelial cells, using a NO-sensitive microsensor,minute equilibration period (378C), preparations wereparalleled these measurements of vascular reactivity.stretched to the optimal length for mechanical perfor-mance using stepwise stretching and intermittent activa-

METHODS tion with 125 mmol/L KCl solution.Vessel segments were preconstricted with norepine-Animals and preparations of isolated arteries

phrine (NE) to 50% of maximal NE-induced contraction.The Institutional Animal Care and Use Committee ap- Endothelium-dependent relaxation was tested by stimu-

proved the experiments. Male Wistar rats (Harlan, Zeist, lating control- and LPS-incubated arteries simultane-The Netherlands) were housed under standard conditions. ously with cumulative doses (1029 to 1025 mol/L) of theRats (250 to 275 g, N 5 14) were anesthetized with endothelium-dependent vasodilator acetylcholine (ACh;sodium pentobarbital (Nembutalt, 60 mg/kg intraperito- receptor dependent). The sensitivity of vascular smoothneally), and the left kidney and mesentery were removed. muscle cells to exogenous NO was tested by giving in-Under sterile conditions, the kidney and mesentery were creasing doses of saturated NO solution (discussed laterpinned in a dissecting dish containing sterile MOPS in this article) to NE-preconstricted segments. To testbuffer (48C), which consisted of (in mmol/L) NaCl 145, the dependency of ACh-mediated relaxations on the re-KCl 5.0, CaCl2 · 2H2O 2.0, MgSO4 · 7H2O 1.0, pyruvate lease of NO by endothelial cells and to verify whether2.0, MOPS [3-(N-morpholino)-propanesulfonic acid] 3.0, the dependency on NO was altered after exposure toNaH2PO4 · H2O 1.0, glucose 11.2, and ethylenediamine- LPS, responses to ACh were repeated in the presencetetraacetic acid (EDTA) 0.02. Segments (approximately of 1024 mol/L Nv-nitro-l-arginine (L-NA), a competitive1.5 mm length) of third-order branches (passive external inhibitor of eNOS.diameter: 325 to 350 mm) from the renal artery and

Nitric oxide measurementsthe superior mesenteric artery were dissected free fromsurrounding tissue and incubated in cell culture medium The electrochemical monitoring of NO was performed

using the three-electrode system EMS 100, consisting ofaccording to the protocol described later in this article.

Piepot et al: LPS inhibits renal endothelial NO release2504

a NO microsensor (sensitivity approximately 1 nmol/L · Krebs solution was stopped, and the pH was kept at 7.4.Subsequently, three concentrations of ACh (0.32, 1.0,pA21), a reference electrode, and counter electrode (Bio-

logic, Claix, France). The NO microsensor was produced and 3.2 mmol/L) were mixed in a cumulative manner,with the buffer solution in the organ chamber ensuedby threading a single carbon fiber (Ten Cate Advanced

Composites, Amsterdam, The Netherlands) through a by washout and restabilization. To investigate LPS-medi-ated effects on receptor-independent signal transduc-pulled end of a glass capillary, which is 2 mm in length,

according to the method previously described [28]. The tion, three doses of the receptor-independent calciumionophore A23187 (0.32, 1.0, and 3.2 mmol/L) were usedcarbon fibers were coated with a polymeric layer of

nickel (II)tetrakis(3-methoxy-4-hydroxyphenyl)porph- also. Local NO production after stimulation with AChand the calcium ionophore A23187 was monitored con-yrin (Interchim, Montlucon, France) and two 1.25%

Nafion layers (Sigma-Aldrich, Zwyndrecht, The Nether- tinuously, and steady-state levels were determined.lands). The porphyrinic microsensor was free from inter-

Drugs and chemicalsference from all reagents used in these experiments. Po-tential differences were expressed with reference to the Norepinephrine, ACh-chloride, Nv-nitro-l-arginine,

l-arginine, and the calcium ionophore A23187 were ob-standard Ag/AgCl reference electrode. A platinum wirewas used as counter electrode. The oxidation potential tained from Sigma Chemical Co. (St. Louis, MO, USA).

DMEM, FCS, and penicillin/streptomycin solution werefor NO was determined by a voltammogram and was setto 680 mV during all experiments. obtained from GIBCO (GIBCO BRL, Breda, The Neth-

erlands). Escherichia coli LPS (serotype O127:B8) wasA linear calibration curve was obtained for each NOmicrosensor after every experiment using aliquots of sat- obtained from Difco Laboratories (Detroit, MI, USA)

All other chemicals were obtained from Merck (Darm-urated NO solution. To produce a standard NO solution,deionized water was bubbled with argon for 20 minutes. stadt, Germany).This deoxygenated water was then bubbled with pure

StatisticsNO gas (AGA Gas BV, Amsterdam, The Netherlands)for 20 minutes and kept under NO atmosphere until use. All contractile responses were expressed as force di-

vided by the axial segment length. Relaxations to AChThe saturated solution was diluted with deoxygenatedwater. Diluted NO solutions were also used to test the were expressed as the percentage of the preconstriction

response to NE. The amount of NO release is expressedsensitivity of vascular smooth muscle cells to exogenousNO during isometric force measurements. in nmol/L. Student’s t-tests were used to test for differ-

ences between control and LPS-incubated groups afterMeasurements of NO were performed in parallel withthe isometric force measurements on separate control contraction with 125 mmol/L KCl. Repeated-measures

analysis of variance (ANOVA) was performed to testand LPS-exposed vessel segments from the same rat.Two renal artery segments (control 1 LPS exposed) and for differences between concentration-response relation-

ships. P values , 0.05 were considered statistically sig-two mesenteric artery segments (control 1 LPS exposed)were tested in a randomized order. In addition, in a nificant. All values are expressed as mean 6 SEM of N

experiments.separate set of experiments, measurements of NO re-lease were also performed in de-endothelialized seg-ments of the renal artery. These arteries served as nega-

RESULTStive controls. Following incubations with or without LPS,

Effects of LPS on vascular smooth muscle andthe vascular endothelium was mechanically removed byendothelial functiongently rubbing the endothelial surface. Mechanical re-

moval of the endothelium was previously shown the saf- Vascular smooth muscle contractile responses to thereceptor-independent vasoconstrictor 125 mmol/L KClest method for endothelium removal without damaging

vascular smooth muscle cells [29]. The vessel segments were not affected by exposure to LPS in renal arteries(948.0 6 78.3 mg/mm and 934.0 6 114.9 mg/mm in con-(with or without endothelium) were cut open longitudi-

nally and pinned on a silicon layer in an organ chamber, trol and LPS-exposed segments, respectively). LPS simi-larly did not affect vascular smooth muscle responses towith the endothelial side facing upward. The NO-sensi-

tive microsensor was placed gently on the endothelial 125 mmol/L KCl in superior mesenteric arteries (836.0 614.9 and 789.0 6 38.2 mg/mm in control and LPS-exposedside of the vessel segment. The organ chamber was filled

with 1.5 mL Krebs solution (378C). The vessel segment segments). Figure 1 shows cumulative concentration-response relationships to ACh in NE-preconstricted con-was superfused at a flow rate of 50 mL/h using a perfusion

pump (Perfusor Secura, B. Braun, Germany). After 15 trol segments and LPS-exposed segments of renal andsuperior mesenteric arteries. Endothelium-dependentminutes of equilibration, the vessel segments were con-

tinuously exposed to NE (0.5 mmol/L). During measure- vascular relaxation elicited by ACh was significantly at-tenuated (ANOVA, P , 0.05) following LPS exposurement of NO production, the superfusion flow of the

Piepot et al: LPS inhibits renal endothelial NO release 2505

Fig. 1. Concentration-response relationshipselicited by acetylcholine (ACh) in renal arter-ies (A) and superior mesenteric arteries (B)incubated during two hours in cell culture me-dium in the absence (j) and presence (h) ofE. coli lipopolysaccharide (LPS; O127:B8, 50mg · mL21). Vascular relaxation is expressedas a percentage reversal of norepinephrine(NE) preconstriction. Data represent means 6SEM of six experiments.

in renal arteries (Fig. 1A), but not in superior mesenteric release elicited by the receptor-dependent agonist ACharteries (Fig. 1B). was significantly higher (P , 0.05; Fig. 2 C, D) in control

segments from superior mesenteric arteries than in con-Role of NO in attenuated endothelium-dependent trol segments from renal arteries.vascular relaxation Inducing receptor-independent Ca21 influx in endo-

The next series of experiments focused on the role of thelial cells with the calcium ionophore A23187 also re-NO in the ACh-mediated vascular responses. The three sulted in a dose-dependent NO release in control asdoses of ACh at which the effects of LPS were most well as in LPS-exposed renal and superior mesentericapparent in the previously mentioned experiments arteries. LPS caused a significant attenuation (ANOVA,were now used to test endothelium-dependent vascular P , 0.05; Figs. 2E and 3) in the amount of NO releasesmooth muscle function and endothelial NO release in elicited by the calcium ionophore A23187 in renal arter-parallel. Figure 2 again shows that ACh-induced vascular ies when compared with time-matched control segments.relaxation was significantly attenuated by exposure to The calcium ionophore A23187-induced NO release wasLPS in rat renal arteries (ANOVA, P , 0.05; Fig. 2A), fully abolished in de-endothelialized control and LPS-but not in superior mesenteric arteries (Fig. 2B). How- exposed renal segments at all three doses (N 5 3, dataever, in the presence of L-NA, a competitive inhibitor not shown). In contrast, LPS did not influence A23187-of NOS, ACh-mediated vascular relaxation was nearly induced NO release in superior mesenteric arteries (Fig.abolished at all three doses in control, as well as LPS- 2F). Furthermore, no difference in the total amount ofexposed renal arterial segments. Endothelium-depen-

released NO was found between control segments ofdent vascular relaxation in response to 0.32, 1.0, and 3.2

renal and superior mesenteric arteries after the additionmmol/L ACh in the presence of 1024 mol/L L-NA were

of the receptor-independent agonist A23187 (Fig. 2 E, F).4.3 6 3.4%, 5.9 6 3.9%, and 5.2 6 4.0% in controlTaken together, both receptor-dependent and receptor-segments and 8.0 6 2.5%, 1.9 6 4.4%, and 4.5 6 6.8%independent activation of NO release by ACh andin LPS-exposed renal segments, respectively.A23187 were significantly reduced in LPS-exposed renalIn all cases, agonist-induced NO release was dose de-arteries (Fig. 2 C, E), while no effects of LPS were ob-pendent in control as well as LPS-exposed arterial seg-served on NO release by superior mesenteric arteriesments (Fig. 2 C–F). In contrast, endothelium-dependent(Fig. 2 D, F).relaxation induced by ACh showed no dose dependency

To test whether the sensitivity of vascular smooth mus-at equal final concentrations (Fig. 2 A, B).cle cells to exogenous NO was altered after exposure toIn comparison to control segments, exposure to LPSLPS in rat renal arteries, increasing doses of NO solutionled to a statistically significant (ANOVA, P , 0.05; Fig.were given to NE-preconstricted renal arteries (Fig. 4).2C) decrease in ACh-induced NO release in renal arter-Repeated-measures ANOVA of the overall concentra-ies. ACh-induced NO release was fully abolished in de-tion-response relationships indicated no significant dif-endothelialized control and LPS-exposed renal segmentsference in the sensitivity of renal vascular smooth muscleat all three doses (N 5 3, data not shown). The exposurecells to exogenous NO between control and LPS-exposedof superior mesenteric arteries to LPS had no effect on

ACh-induced NO release (Fig. 2D). The amount of NO arteries.

Piepot et al: LPS inhibits renal endothelial NO release2506

Fig. 2. Endothelium-dependent vascular re-laxation elicited by acetylcholine (ACh) incontrol ( ) and LPS-exposed (2 h, E. coli,O127:B8, 50 mg mL21) segments (h) fromrenal arteries (A) and superior mesentericarteries (B). Vascular relaxation is expressedas a percentage reversal of NE preconstric-tion. Agonist-induced endothelial nitric oxide(NO) release in control ( ) and LPS-exposed(h) segments from renal (C and E) and supe-rior mesenteric arteries (D and F). Responseswere obtained after stimulation with ACh (Cand D) and the calcium ionophore A23187(E and F). Agonist-stimulated NO release isexpressed in nmol/L. Data represent means 6SEM of seven to eight experiments.

DISCUSSION rat renal arteries, after two hours of exposure to LPS,coincided with reduced agonist-induced NO release.The objective of the present study was to study theSince agonist-induced NO release was abolished in endo-direct effects of a short-term (2 h) exposure of isolatedthelium-denuded renal arteries, these results indicaterat renal and superior mesenteric arteries to E. coli LPSthat LPS selectively impaired agonist-induced endothe-on ACh-mediated endothelium-dependent vascular re-lial NO release. In contrast, in rat superior mesentericlaxation. Our findings demonstrate that defective ACh-

mediated endothelium-dependent vascular relaxation in arteries, which did not exhibit impaired endothelium-

Piepot et al: LPS inhibits renal endothelial NO release 2507

Fig. 4. Endothelium-independent vascular smooth muscle relaxationFig. 3. Tracing from a nitric oxide (NO) measurement in a rat renalin response to cumulative doses of saturated NO solution in controlartery after stimulation with the calcium ionophore A23187 (0.32 mmol/L).( ) and LPS-exposed (2 h, E. coli, O127:B8, 50 mg · mL21) renalVessels were incubated in cell culture medium in the absence (blackarterial segments (h). Vascular relaxation is expressed as a percentageline) and presence (gray line) of E. coli LPS (O127:B8, 50 mg · mL21)reversal of NE preconstriction. These data represent the means 6 SEMduring two hours. Endothelial NO release was measured using porphy-of eight experiments.rin coated NO-sensitive microsensors.

dependent vascular relaxation after two hours of expo- ported in superior mesenteric arteries isolated from ratsfour hours after intravenous infusion of LPS [33], oursure to LPS, no effect of LPS on agonist-induced NO

release was found. results indicate that in vitro exposure of isolated superiormesenteric arteries during a shorter period (2 h) to E.

Methodological considerations coli LPS alone does not provoke endothelial dysfunction.However, we cannot exclude the possibility that this maySeveral studies have shown that the expression of the

inducible calcium/calmodulin-independent iNOS leads be caused by the differences in experimental conditionsrather than by differences in exposure time to bacterialto decreased endothelium-dependent vascular relaxation

caused by the inhibition of eNOS activity by iNOS-derived LPS. Nevertheless, the fact that endothelial dysfunctionin rat renal arteries was found under the same experi-NO [30, 31]. Strengthened by findings that indicate that

acute renal failure precedes the induction of iNOS, we mental conditions as in the experiments with the superiormesenteric arteries strongly argues in favor of the latter.have chosen to avoid this confounding influence by ex-

posing isolated arteries to LPS for a short period (2 h).Dependency of ACh response on NOUnder these conditions, it is unlikely that iNOS was

expressed because iNOS expression has been found to Endothelium-dependent vascular relaxation, elicitedby ACh or bradykinin, is not always fully dependent onbe significantly increased after four hours or more of

exposure to LPS [9]. Furthermore, after two hours of endothelial NO release. Several studies have shown thathyperpolarization of vascular smooth muscle cells by theexposure to LPS, the contractile responses to 125 mmol/L

KCl in this study were not different from time-matched so-called endothelium-derived hyperpolarizing factor(EDHF) may play a role as well. EDHF, as well as NO,controls in renal as well as superior mesenteric arteries.

Considering that other studies have shown a strong cor- mediates endothelium-dependent vascular relaxation[34, 35]. Therefore, for further characterization of therelation between the expression of iNOS in isolated ar-

teries and attenuated contractile responses to KCl and mechanism involved in LPS-induced defective ACh-induced vascular relaxation in rat renal arteries, weother vasoconstrictors, our results strongly suggest that

iNOS expression was absent in the renal and superior tested whether the response to ACh was dependent onNO alone, and if so, whether LPS might change themesenteric arteries used in our study [32].

Although diminished responses to endothelium-depen- relative contribution of NO and other vasodilating fac-tors such as EDHF in ACh-mediated vascular relaxationdent vasodilator challenges have previously been re-

Piepot et al: LPS inhibits renal endothelial NO release2508

in renal arteries. Responses to ACh were nearly abol- tion ([Ca21]i) [36, 37]. Graier et al have shown in culturedbovine aortic endothelial cells that LPS exposure for oneished in the presence of L-NA, a competitive antagonist

for eNOS in control as well as LPS-exposed renal arter- hour significantly attenuated the agonist-induced in-crease in endothelial [Ca21]i [23]. Nevertheless, releaseies. This indicates first that ACh-induced vascular relax-

ation in the isolated renal arteries was fully dependent of NO in response to the calcium ionophore A23187,which induces a direct receptor-independent increase inon NO release. Second, these results indicate that LPS

exposure does not modify the dependency of ACh-medi- endothelial [Ca21]i, was unaffected in these experiments.Decreased endothelial NO release was associated withated vascular relaxation on NO release.decreased ACh-mediated, but not A23187-mediated, en-

Effect of LPS on sensitivity of vascular smooth muscle dothelium-dependent vascular relaxation in a study fromto exogenous NO Myers et al in isolated aorta from LPS-infused guinea

pigs as well [24]. A principal difference with our study,Nitric oxide produced in endothelial cells rapidly dif-fuses to adjacent vascular smooth muscle cells, where it however, was that aorta were tested after 16 hours of

E. coli endotoxemia. Thus, an effect of iNOS in theseinduces vascular smooth muscle relaxation via activationof soluble guanylate cyclase and subsequent increases in findings cannot be excluded. These studies suggested a

selective impairment of the signal transduction mecha-cGMP levels [15, 36]. Our observed decreases in endo-thelium-dependent vascular relaxation, therefore, may nisms that mediate depletion of endothelial intracellular

Ca21 stores. As a result, agonist-induced increases inbe caused by decreased NO-mediated activation of solu-ble guanylate cyclase. The sensitivity of vascular smooth endothelial [Ca21]i are reduced and subsequently lead

to declined activation of the calcium/calmodulin-depen-muscle cells to exogenous NO was tested in NE-precon-stricted control and LPS-exposed renal arteries. LPS was dent eNOS.

Our experiments do not confirm these findings, sincefound not to alter the sensitivity for exogenous NO in ourexperiments. This result confirms studies from Bhagat et endothelial responses to direct, receptor-independent

Ca21 recruitment by the calcium ionophore A23187 wereal, who found in human saphenous veins incubated withLPS for one hour, that impaired endothelium-dependent similarly decreased as receptor-dependent activation of

eNOS. Our data do suggest that instead of disorderedvascular relaxation in response to bradykinin was notassociated with decreased sensitivity to the NO donor agonist-mediated increases in endothelial [Ca21]i, the cal-

cium-signaling mechanism involving the Ca21/calmodu-glyceryltrinitrate [16]. A disadvantage of the conven-tional NO donors used in many studies such as glyceryl- lin complex that mediates the activation of the eNOS

enzyme in endothelial cells was deficient in LPS-exposedtrinitrate and sodium nitroprusside is that enzymaticalbreakdown is required to release NO. Furthermore, the renal arteries in our experiments. Therefore, we conclude

that decreased endothelium-dependent vascular relax-final concentration of NO, which is released, is not knownand strongly depends on the presence and amount of ation in LPS-exposed renal arteries, elicited by ACh, is

due to reduced NO production in endothelial cells.tissue. The application of NO, using pure NO gas-satur-ized deionized water, therefore, is a good alternative to The measurements of endothelial NO release showed

that the effects of ACh, as well as the calcium ionophorebypass these problems. Based on our findings, we con-clude that LPS did not decrease the NO-mediated activa- A23187, were dose dependent in renal as well as superior

mesenteric arteries, whereas endothelium-dependent re-tion of soluble guanylate cyclase in vascular smooth mus-cle cells. Thus, decreased sensitivity of vascular smooth laxations were not dose dependent at equal doses of

these agonists. This may be explained by the fact thatmuscle cells to NO does not underlie the impairmentsof endothelium-dependent vascular relaxation in renal ACh and the calcium ionophore not only induce rises

in endothelial [Ca21]i, providing the trigger for endothe-arteries provoked by LPS.lial NO production, but also induce increased [Ca21]i in

Effect of LPS on endothelial NO production vascular smooth muscle cells. Therefore, the responseto ACh and the calcium ionophore A23187 is the resultIn cultured endothelial cells, it has been demonstrated

that LPS and inflammatory cytokines are able to de- of two opposing effects: vasodilatory effects of (dosedependent) endothelial NO release opposed by a vaso-crease agonist-mediated endothelial NO release [22, 23].

Mechanisms involved in the inhibitory effect of LPS on contractile effect of ACh and A23187 on smooth musclecells [38]. This idea is substantiated by the finding thatACh-mediated activation of eNOS and NO release could

hypothetically involve several targets, including changes stimulation of preconstricted renal arteries with NO(1027 mol/L) alone induced full relaxation of the vascularin muscarine receptor expression on endothelial cells

and disrupted signal transduction mechanisms coupling smooth muscle.Acetylcholine-stimulated NO release was found lowerreceptor binding to Ca21 increases in endothelial cells

since agonist-induced activation of eNOS is dependent in renal arteries than in superior mesenteric arteries.This may be due to differences in receptor density onon increases in endothelial intracellular Ca21-concentra-

Piepot et al: LPS inhibits renal endothelial NO release 2509

the endothelial surface of both arteries. The finding that nic circulations [26]. In rat renal arteries, endothelium-receptor-independent NO production induced by the cal- dependent vascular relaxation was fully dependent oncium ionophore A23187 was equal in both arterial types endothelial NO release. Moreover, the sensitivity of vas-supports this concept. cular smooth muscle to NO was unchanged after expo-

A selective reduction of endothelium-derived NO re- sure to LPS. Considering these findings, the results ob-lease in rat renal arteries may be the cause of the ob- tained in this study indicate that the high sensitivity ofserved decreases in endothelium-dependent vascular re- rat renal arteries toward the inhibitory effects of short-laxation in rat renal arteries after short-term treatment term exposure to LPS on endothelial NO release resultswith LPS. Decreased expression of eNOS, altered intra- in defective agonist-mediated vascular relaxation. Thiscellular localization, and inactivation of eNOS might pro- may result in renal endothelial dysfunction and contrib-vide possible explanations for the observed decreases in ute to the propensity for acute renal failure during exper-agonist-stimulated NO release and impairment of endo- imental septic shock.thelium-dependent vascular relaxation. Several groups

Reprint requests to Dr. Harro A. Piepot, Laboratory for Physiology,have shown down-regulation of eNOS mRNA and pro-Vrije Universiteit, Van der Boechorststraat 7, 1081 BT Amsterdam, The

tein levels after long-term treatment with LPS or in- Netherlands.E-mail: piepot@physiol.med.vu.nlflammatory cytokines [25, 39, 40]. Although this mecha-

nism may have an impact after long-term treatment withREFERENCESinflammatory mediators or in vivo shock, decreased

eNOS expression does not seem to be a conceivable 1. Thijs A, Thijs LG: Pathogenesis of renal failure in sepsis. KidneyInt 53(Suppl 66):S34–S37, 1998explanation for our findings, since down-regulation of

2. Xie QW, Cho HJ, Calaycay J, Mumford RA, Swiderik KM, LeeeNOS mRNA and protein levels caused by LPS andTD, Ding A, Troso T, Nathan C: Cloning and characterization of

inflammatory cytokines requires four hours or more inducible nitric oxide synthase from mouse macrophages. Science[39, 40]. Therefore, acute inactivation of eNOS rather 256:225–228, 1992

3. Thiemermann C, Vane J: Inhibition of nitric oxide synthesis re-than decreased eNOS-expression may explain the de-duces the hypotension induced by bacterial lipopolysaccharides increase in agonist-stimulated NO release. Several studies the rat in vivo. Eur J Pharmacol 182:591–595, 1990

have shown that in response to receptor-dependent and 4. Kilbourn RG, Jubran A, Gross SS, Griffith OW, Levi R, AdamsJ, Lodato RF: Reversal of endotoxin-mediated shock by NG-receptor-independent agonists (for example, followingmethyl-l-arginine, an inhibitor of nitric oxide synthesis. Biochemincreases in endothelial [Ca21]i), eNOS undergoes dy- Biophys Res Commun 172:1132–1138, 1990

namic alterations in bindings to plasmalemmal proteins 5. Van Lambalgen AA, van Kraats AA, van den Bos GC, Stel HV,Straub J, Donker AJM, Thijs LG: Renal function and metabolismand intracellular localization, which are associated withduring endotoxemia in rats: Role of hypoperfusion. Circ Shockchanges in the activity of eNOS [41, 42]. Furthermore, 35:164–173, 1991

alterations in phosphorylation levels of eNOS, induced 6. Heemskerk AEJ, Huisman E, van Lambalgen AA, van den BosGC, Thijs LG, Tangelder GJ: Laparotomy and renal functionby inflammatory mediators, were shown to dramaticallyduring endotoxin shock in rats. Shock 6:410–417, 1996decrease eNOS activity [43]. Whether these processes

7. Lugon JR, Boim MA, Ramos OL, Ajzen H, Schor N: Renalhave contributed to reduced endothelial NO release in function and glomerular hemodynamics in male endotoxemic rats.

Kidney Int 36:570–575, 1989our experimental setup is uncertain and requires further8. Millar CGM, Thiemermann C: Intrarenal haemodynamics andstudy.

renal dysfunction in endotoxaemia: Effects of nitric oxide synthaseIn summary, the present work shows that short-term inhibition. Br J Pharmacol 121:1824–1830, 1997

9. Wallis G, Brackett D, Lerner M, Kotake Y, Bolli R, McCayexposure to E. coli LPS attenuates endothelium-depen-PB: In vivo spin trapping of nitric oxide generated in the smalldent vascular relaxation in renal arteries but not in supe-intestine, liver, and kidney during the development of endotoxe-

rior mesenteric arteries. In parallel with the effects, ago- mia: A time course study. Shock 6:274–278, 199610. Ruetten H, Southan GJ, Abate A, Thiemermann C: Attenuationnist-stimulated NO release from endothelial cells was

of endotoxine-induced multiple organ dysfunction by 1-amino-significantly reduced in renal arteries but, again, not in2-hydroxy-guanidine, a potent inhibitor of inducible nitric oxide

superior mesenteric arteries. Therefore, it seems unlikely synthase. Br J Pharmacol 118:261–270, 1996that defective agonist-mediated endothelium-dependent 11. Wray GM, Millar CG, Hinds CJ, Thiemermann C: Selective

inhibition of the activity of inducible nitric oxide synthase preventsvascular relaxation might contribute to the abnormalitiesthe circulatory failure, but not the organ injury/dysfunction, causedin splanchnic vascular tone and splanchnic hypoperfu- by endotoxin. Shock 9:329–335, 1998

sion seen in rats within two hours after infusion of LPS 12. Shultz PJ, Raij L: Endogenously synthesized nitric oxide preventsendotoxin-induced glomerular thrombosis. J Clin Invest 90:1718–[26]. Moreover, the lack of effect of short-term exposure1725, 1992to LPS on endothelial-dependent vascular relaxation and

13. Spain DA, Wilson MA, Garrison RN: Nitric oxide synthase inhi-endothelial NO release in superior mesenteric arteries bition exacerbates sepsis-induced renal hypoperfusion. Surgery

116:322–331, 1994versus renal arteries clearly demonstrates the existence14. Van Lambalgen AA, Bouriquet N, Casellas D: Effects of endo-of regional variability in the vascular responses to bacte-

toxin on tone and pressure-responsiveness of preglomerular jux-rial LPS, despite similar hemodynamic characteristics tamedullary vessels. Pflugers Arch 432:574–577, 1996

15. Moncada S, Palmer RMJ, Higgs EA: Nitric oxide: Physiology,during in vivo endotoxin shock in the renal and splanch-

Piepot et al: LPS inhibits renal endothelial NO release2510

pathophysiology, and pharmacology. Pharmacol Rev 43:109–142, 29. Kuo L, Chilian WM, Davis MJ: Coronary arteriolar myogenicresponse is independent of endothelium. Circ Res 66:860–866, 19901991

30. Schwartz D, Mendonca M, Schwartz I, Xia Y, Satriano J,16. Bhagat K, Moss R, Collier J, Vallance P: Endothelial “stun-Wilson CB, Blantz RC: Inhibition of constitutive nitric oxidening” following a brief exposure to endotoxin: A mechanism tosynthase (NOS) by nitric oxide generated by inducible NOS afterlink infection and infarction? Cardiovasc Res 32:822–829, 1996lipopolysaccharide administration provokes renal dysfunction in17. Peters TS, Lewis SJ: Lipopolysaccharide inhibits acetylcholine-rats. J Clin Invest 100:439–448, 1997and nitric oxide-mediated vasodilation in vivo. J Pharmacol Exp

31. Kessler P, Bauersachs J, Busse R, Schini-Kerth VB: InhibitionTher 279:918–925, 1996of inducible nitric oxide synthase restores endothelium-dependent18. Beasley D, Cohen RA, Levinsky NG: Endotoxin inhibits contrac-relaxations in proinflammatory mediator-induced blood vessels.tion of vascular smooth muscle in vitro. Am J Physiol 258(4 PtArterioscler Thromb Vasc Biol 17:1746–1755, 19972):H1187–H1192, 1990

32. Scott JA, Machoun M, McCormack DG: Inducible nitric oxide19. Parker JL, Adams HR: Selective inhibition of endothelium-depen-synthase and vascular reactivity in rat thoracic aorta: Effect ofdent vasodilator capacity by Escherichia coli endotoxemia. Circaminoguanidine. J Appl Physiol 80:271–277, 1996Res 72:539–551, 1993

33. Siegfried MR, Ma XL, Lefer AM: Splanchnic vascular endothelial20. Wylam ME, Samsel RW, Umans JG, Mitcell RW, Leff AR,dysfunction in rat endotoxemia: Role of superoxide radicals. EurSchumacker PT: Endotoxin in vivo impairs endothelium-depen-J Pharmacol 212:171–176, 1992dent relaxation of canine arteries in vitro. Am Rev Respir Dis

34. Chen G, Suzuki H, Weston AH: Acetylcholine releases endothe-142:1263–1267, 1990lium-derived hyperpolarizing factor and EDRF from rat blood21. Young JS, Headrick JP, Berne RM: Endothelial-dependent andvessels. Br J Pharmacol 95:1165–1174, 1988-independent responses in the thoracic aorta during endotoxic

35. Bakker ENTP, Sipkema P: Components of acetylcholine-inducedshock. Circ Shock 35:25–30, 1991dilation in isolated rat arterioles. Am J Physiol 273(4 Pt 2):H1848–22. Myers PR, Wright TF, Tanner MA, Adams HR: EDRF and nitricH1853, 1997oxide production in cultured endothelial cells: Direct inhibition by

36. Ignarro L, Kadowitz P: The pharmacological and physiologicalE. coli endotoxin. Am J Physiol 262(3 Pt 2):H710–H718, 1992 role of cyclic GMP in vascular smooth muscle. Annu Rev Pharma-23. Graier WF, Myers PR, Rubin LJ, Adams HR, Parker JL: Esche- col Toxicol 25:171–191, 1985richia coli endotoxin inhibits agonist-mediated cytosolic Ca21 mo- 37. Fleming I, Bauersachs J, Busse R: Calcium-dependent and cal-bilization and nitric oxide biosynthesis in cultured endothelial cells. cium-independent activation of the endothelial NO synthase.Circ Res 75:659–668, 1994 J Vasc Res 34:165–174, 1997

24. Myers PR, Zhong Q, Jones JJ, Tanner MA, Adams HR, Parker 38. Furchgott RF, Zawadzki JV: The obligatory role of endothelialJL: Release of EDRF and NO in ex vivo perfused aorta: Inhibition cells in the relaxation of arterial smooth muscle by acetylcholine.by in vivo E. coli endotoxemia. Am J Physiol 268(3 Pt 2):H955– Nature 288:373–376, 1980H961, 1995 39. Zhang J, Patel JM, Li YD, Block ER: Proinflammatory cytokines

25. Chen K, Inoue M, Wasa M, Fukuzawa M, Kamata S, Okada A: downregulate gene expression and activity of constitutive nitricExpression of endothelial constitutive nitric oxide synthase mRNA oxide synthase in porcine pulmonary artery endothelial cells. Resin gastrointestinal mucosa and its downregulation by endotoxin. Commun Mol Pathol Pharmacol 96:71–87, 1997Life Sci 61:1323–1329, 1997 40. Lu JL, Schmiege LM, Kuo L, Liao JC: Downregulation of endothe-

26. Mulder MF, van Lambalgen AA, van den Bos GC, Thijs LG: lial constitutive nitric oxide synthase expression by lipopolysaccha-The fall of cardiac output in endotoxemic rats cannot explain all ride. Biochem Biophys Res Commun 225:1–5, 1996changes in organ blood flow: A comparison between endotoxin 41. Prabhakar P, Thatte HS, Goetz RM, Cho MR, Golan DE,and low venous return shock. Shock 5:135–140, 1996 Michel T: Receptor-regulated translocation of endothelial nitric-

27. De Mey JGR, Uitendaal MP, Boonen HCM, Vrijdag MJJF, oxide synthase. J Biol Chem 273:27383–27388, 1998Daemen MJAP, Struyker-Boudier HAJ: Acute and long-term 42. Feron O, Saldana F, Michel JB, Michel T: The endothelialeffects of tissue culture on contractile reactivity in renal arteries nitric-oxide synthase-caveolin regulatory cycle. J Biol Chemof the rat. Circ Res 65:1125–1135, 1989 273:3125–3128, 1998

28. Malinski T, Taha Z: Nitric oxide release from a single cell mea- 43. Garcia-Cardena G, Fan R, Stern DF, Liu J, Sessa WC: Endothe-sured in situ by a porphyrinic-based microsensor. Nature 358:676– lial nitric oxide is regulated by tyrosine phosphorylation and inter-

acts with caveolin-1. J Biol Chem 271:27237–27240, 1996678, 1992