Effects of nitric oxide on blood flow distribution and O2 extraction capabilities during endotoxic shock HAIBO ZHANG,1 PETER ROGIERS,1 NADIA SMAIL,1 ANA CABRAL,1 JEAN-CHARLES PREISER,1 MARIE-ODILE PENY,2 AND JEAN-LOUIS VINCENT1 1Department of Intensive Care, and 2Department of Pathology, Erasme University Hospital, Free University of Brussels, B-1070 Brussels, Belgium Zhang, Haibo, Peter Rogiers, Nadia Smail, Ana Cabral, Jean-Charles Preiser, Marie-Odile Peny, and JeanLouis Vincent. Effects of nitric oxide on blood flow distribution and O2 extraction capabilities during endotoxic shock. J. Appl. Physiol. 83(4): 1164–1173, 1997.—The effects of the nitric oxide (NO) synthase inhibitor NG-monomethyl-Larginine (L-NMMA) and the NO donor 3-morpholinosydnonimine (SIN-1) were tested in 18 endotoxic dogs. L-NMMA infusion (10 mg · kg21 · h21 ) increased arterial and pulmonary artery pressures and systemic and pulmonary vascular resistances but decreased cardiac index, left ventricular stroke work index, and blood flow to the hepatic, portal, mesenteric, and renal beds. SIN-1 infusion (2 µg · kg21 · min21 ) increased cardiac index; left ventricular stroke work index; and hepatic, portal, and mesenteric blood flow. It did not significantly influence arterial and pulmonary artery pressures but decreased renal blood flow. The critical O2 delivery was similar in the L-NMMA group and in the control group (13.3 6 1.6 vs. 12.8 6 3.3 ml · kg21 · min21 ) but lower in the SIN-1 group (9.1 6 1.8 ml · kg21 · min21, both P , 0.05). The critical O2 extraction ratio was also higher in the SIN-1 group than in the other groups (58.7 6 10.6 vs. 42.2 6 7.6% in controls, P , 0.05; 43.0 6 15.5% in L-NMMA group, P 5 not significant). We conclude that NO is not implicated in the alterations in O2 extraction capabilities observed early after endotoxin administration. cardiac output; hypotension; organ perfusion; oxygen delivery; sepsis; endothelium-derived relaxing factor; nitrite; tumor necrosis factor-a

SEPTIC SHOCK, a major clinical problem with mortality rates of up to 70%, is characterized by systemic hypotension, vascular hyporeactivity and myocardial depression. Despite the increased cardiac output, cellular O2 utilization may be inadequate because maldistribution of blood flow can profoundly alter O2 availability. Regional blood flow to the various organs may also be altered nonuniformly. In particular, the hepatosplanchnic blood flow may decrease more than blood flow to other regions to favor blood supply to vital organs, such as the heart and the brain. Because hepatosplanchnic hypoxia may contribute to the development of multiple organ failure (11), the maintenance of sufficient hepatosplanchnic blood flow may be a fundamental goal in septic shock. Nitric oxide (NO) is a paracrine-acting gas enzymatically synthesized from L-arginine. In basal conditions, the release of NO via calcium-dependent constitutive NO synthase (cNOS) plays an essential role in the maintenance of capillary flow and O2 availability to the tissues. The greater release of NO via a calciumindependent inducible form of NO synthase (iNOS) has been implicated in the pathophysiological alterations of

1164

severe sepsis and septic shock. The resulting overproduction of NO may induce deleterious effects, including arterial hypotension (6), vascular hyporeactivity, and myocardial depression (5). The induction of iNOS in various cells, including macrophages, endothelial cells, vascular smooth muscle cells, or even myocardial cells, requires several hours to be expressed after stimulus with endotoxin and cytokines, such as tumor necrosis factor-a (TNF-a) or interferon-g. Nitric oxide synthase (NOS) inhibition by the administration of L-arginine analogs has yielded controversial results in sepsis. Although such intervention does reverse endotoxin- or TNF-a-induced hypotension (6, 38), it also generally decreases blood flow (19, 23), so that it may worsen tissue injury and even increase mortality rate in some experimental models of septic shock (6, 23, 38). Several groups of investigators have also stressed that NOS inhibition may result in hepatosplanchnic ischemia (23). During ischemic or hypoxic hypoxia in dogs, blockade of NOS by NG-nitro-L-arginine methyl ester (L-NAME) does not appear to affect the O2 extraction ratio (ERO2 ) (15, 31, 33). During endotoxemia in dogs, Walker et al. (36) reported that L-NAME administration decreased ˙ O2 ) by 50% but whole body and intestinal O2 delivery (D ˙ O2 uptake (VO2) remained stable, so that whole body and intestinal ERO2 values increased above 50%. However, the effects of NO on tissue O2 extraction capabilities during septic shock have not been well defined. We chose to study the effects of NG-monomethyl-L-arginine (L-NMMA), a NOS inhibitor presently tested in clinical trials (35), on regional blood flow distribution and O2 extraction capabilities during endotoxic shock. Recently, we found (40) that N-acetyl-L-cysteine (NAC), a potent antioxidant substance that enhances endothelium-derived relaxing factor activity, improved the O2 extraction capabilities during endotoxic shock in ˙ O2 dogs, as indicated by a significantly lower critical D ˙ O2 crit ) and higher critical ERO2 (ERO2 crit ) when blood (D flow was progressively reduced by tamponade. NAC also significantly increased cardiac index (CI) by an improvement in myocardial contractility and lowered pulmonary hypertension (40). It is intriguing, therefore, to test the hypothesis that NO-releasing compounds may have beneficial effects in septic shock. As a NO donor we chose 3-morpholinosydnonimine (SIN-1), the vasoactive metabolite of molsidomine currently used as a nitrate compound in the treatment of coronary artery disease. Also, we recently observed (41) that the administration of SIN-1 during endotoxic shock in dogs increased splanchnic blood flow without adverse effects on arterial pressure.

0161-7567/97 $5.00 Copyright r 1997 the American Physiological Society

http://www.jap.org

NITRIC OXIDE ON O2 EXTRACTION IN ENDOTOXEMIA

Hence, the present study investigated the effects of and SIN-1 in a dog model of endotoxic shock. In addition to global hemodynamics, we studied regional blood flow in hepatic, portal, mesenteric, and renal vasculatures. We also studied the effects of these interventions on tissue O2 extraction capabilities by progressively decreasing blood flow induced by cardiac tamponade. L-NMMA

MATERIALS AND METHODS

Surgical preparation. Eighteen mongrel dogs weighing 24.5 6 6.5 kg were anesthetized with pentobarbital sodium, with an initial loading dose of 30 mg/kg followed by a constant intravenous infusion of 4 mg · kg21 · h21 (Infusomat II pump; B. Braun, Melsungen, Germany) through the right forepaw vein. After being endotracheally intubated with a cuffed endotracheal tube, the dog was mechanically ventilated with room air by using a servo ventilator (model 900B, SiemensElema, Solna, Sweden). Controlled ventilation was facilitated with pancuronium bromide given as an initial bolus of 0.15 mg/kg followed by an infusion of 0.075 mg · kg21 · h21. Respiratory rate was set at 12 breaths/min and tidal volume was adjusted to obtain an end-tidal PCO2 (PETCO2) between 28 and 34 Torr. These ventilatory conditions were not changed thereafter. A right femoral arterial catheter was inserted and connected to a pressure transducer for arterial pressure monitoring. The left forepaw vein was cannulated for normal saline and drug infusion. A balloon-tipped pulmonary artery catheter (model 93A-131–7-Fr; Baxter, Irvine, CA) was inserted through the right external jugular vein under guidance of pressure waves, as determined from a four-channel monitor (Sirecust 302A; Siemens, Erlangen, Germany). A left thoracotomy between the fourth and fifth intercostal space was performed; bleeding was controlled by electrocautery. Via a 2- to 3-mm incision in the anterior pericardium, a 16-gauge polyethylene catheter (Intracath; Deseret Medical, Sandy, UT) with multiple side holes was positioned in the pericardial space with its tip adjacent to the diaphragmatic surface of the left ventricle. The catheter was secured with purse-string sutures. The pericardial cavity was drained, with replacement of 30 ml of warm sterile saline before sealing. The thoracic cavity was then carefully closed in three layers, and a chest tube (Trocar catheter A75, 28 Ch–40 cm, Argyle, Ireland) was placed through the seventh intercostal space to allow gentle evacuation of the chest. Through a midline laparotomy, a splenectomy was performed after maximal splenic contraction to 1 mg epinephrine (spread on the surface of the spleen) to prevent autotransfusion of erythrocytes during hypotension. Ultrasonic flow probes (R-series, positioned for optimum resolution at a 45° angle to the vessel under study) were placed around the common hepatic artery (3 mm), the portal vein (12 mm), the mesenteric artery (4–6 mm), and the left renal artery (3–4 mm), respectively, for simultaneous determinations of blood flow in these vessels. The probes were chosen according to the typical flowprobe sizes for acute application in canine vessels (Transonic Systems Catalog, Ithaca, NY), and there was a close fit to the vessels. The space between the vessel and probe was then filled with an acoustical gel. All flow probes destined for acute use were calibrated by the manufacturer to read 110% of bench-top flow with water (100% of bench-top flow with water was calibrated for chronic use). Measurements and calculations. Pressures were determined from a strip-chart recorder (2600S recorder; Gould, Cleveland, OH) at end expiration. CI (ml · kg21 · min21 ) was measured by the thermodilution technique (cardiac output computer, COM-2, Baxter) by using three to five 5-ml injec-

1165

tions of cold 5% dextrose in ice water. Each injection was started at end inspiration. A temperature probe was used on-line to control for variations in injectate temperature. Regional blood flow was estimated simultaneously in the common hepatic artery, portal vein, mesenteric artery, and left renal artery by a previously calibrated blood flowmeter (model T208; Transonic Systems, Ithaca, NY). Because ,90% of common hepatic blood flow goes to the liver, and the other 10% goes to the gastroduodenal artery in control conditions (30), one should consider that the hepatic blood flow we obtained was an estimation rather than an exact measurement. Exhaled gases were directed through a mixing chamber for sampling of expired gases to measure expired O2 fraction (FEO2) (P. K. Morgan, Chatham, UK). PETCO2 was monitored simultaneously (47210A capnometer; Hewlett-Packard, Waltham, MA). The gas analyzers were calibrated before the experiment. Expired minute volume (VE ) was measured with a spirometer (Haloscale Wright Respirometer; Edmonton, London, UK). Arterial and mixed venous blood samples were simultaneously withdrawn for immediate determination of blood gases and lactate concentration (ABL 500; Radiometer, Copenhagen, Denmark; lactate/glucose analyzer 2300 Stat Plus, Yellow Springs Instruments, Yellow Springs, OH). In each sample, hemoglobin and O2 saturation were measured (OSM 3 Hemoximeter, calibrated for dog blood; Radiometer). ˙ O2 was calculated as the product of arterial O2 content D ˙ O2 was determined by the formula and CI. V

˙ O2 (ml · kg21 · min21) 5 VE/wt (kg) V 3 5[(1 2 FEO2 2 FECO2)/(1 2 FIO2)] 3 FIO2 2 FEO26 where FECO2 and FIO2 represent expired CO2 fraction and inspired O2 fraction, respectively. ERO2 was derived from the ˙ O2/D˙ O2. ratio of V TNF-a levels were measured by using murine immunoglobulin G1 antibody to TNF-a (W. Buurman). This antibody to human TNF-a sees canine TNF-a in Western blot (unpublished data) and inhibits the biological activity of canine TNF-a to a lesser degree than it inhibits human TNF-a. NO production was determined spectrophotometrically by measuring the accumulation of both nitrite and nitrate (the latter is reduced to nitrite) in plasma and was reported by the percentage of changes in nitrite from the baseline. Nitrate was stoichometrically reduced to nitrite by incubation of sample (100 µl plasma) for 2 h at 37°C, in the presence of 0.1 U/ml nitrate reductase [NAD(P)H, nitrate oxidoreductase (EC 1.6.6.2; Aspergillus species) Sigma Chemical, St. Louis, MO], 120 µM NADPH, and 5 µM flavinadenine dinucleotide (Sigma Chemical) in a final volume of 103 µl. After nitrate had been reduced to nitrite, NADPH that interfered with the subsequent nitrite determination was oxidized with 10 U/ml L-lactic dehydrogenase (EC 1.1.1.27, type XI, from rabbit muscle; Sigma Chemical) and 10 mM sodium pyruvate for 30 min at 37°C in a final volume of 114 µl. Sodium nitrate was used as a standard. Nitrite concentration in plasma was assayed by a standard Griess reaction. Briefly, 100 µl plasma were incubated with an equal volume of Griess reagent (1% sulpanilamide-0.1% napthylenediamine dihydrochloride2.5% H3PO4 ) at room temperature for 10 min. The absorbance of the chromophore formed was determined at 540 nm by using a microtiter plate reader (Molecular Devices, Menlo Park, CA). Sodium nitrite was used as a standard, with control baseline plasma as a blank control reference. Experimental protocol. After surgical preparation, the dog was placed in supine position and allowed to stabilize for

1166

NITRIC OXIDE ON O2 EXTRACTION IN ENDOTOXEMIA

30 min before control measurements [baseline 1 (B1)]. Each dog received Escherichia coli endotoxin (E. coli 055:B5 lipopolysaccharide, no. 3120–10–7; Difco, Detroit, MI) as a slow intravenous bolus of 2 mg/kg over 2 min. Thirty minutes later, a second set of measurements [baseline 2 (B2)] was obtained. The dog then received a generous saline infusion to restore and maintain pulmonary artery occlusion pressure at baseline levels. A third set of measurements [baseline 3 (B3)] was obtained 30 min thereafter. Dogs were then randomly divided into three groups, receiving saline infusion at 20 ml · kg21 · h21 either alone (endotoxin, n 5 6 dogs), or in combination with L-NMMA (endotoxin 1 L-NMMA; n 5 6 dogs), or in combination with SIN-1 (endotoxin 1 SIN-1; n 5 6), respectively. In the endotoxin 1 L-NMMA group, L-NMMA (L-NMMA acetate salt; Calbiochem, La Jolla, CA) was continuously infused at 10 mg · kg21 · h21 (15 mg/ml solution). In the endotoxin 1 SIN-1 group, SIN-1 (Corvaton; Therabel Research, Brussels, Belgium) was continuously infused at 2 µg · kg21 · min21 (0.25 mg/ml solution). The SIN-1 solution was protected from light during infusion. Thirty minutes after the initial infusion of either L-NMMA or SIN-1, a fourth set of measurements [baseline 4 (B4)] was made. Cardiac tamponade was then induced by repeated bolus injections of warm normal saline heated to 37°C into the pericardial sac. The amount of saline injected was 30 ml ˙ O2 started to for the first two injections, then 10 ml until V fall from baseline levels, and finally 2–5 ml until the mean arterial pressure (MAP) fell by 80% from baseline. The experiment was then ended. After each injection, a time interval of 20 min was permitted to reach a steady state, characterized by a stable FEO2, PETCO2, arterial pressure, and heart rate before the next measurements were obtained. During the study, core temperature was kept constant at its initial level with warming lamps and a heating blanket. After the experiment was completed, dogs were killed with potassium chloride, and biopsies were immediately taken from the liver (left lobe), small intestine (30 cm from flexura duodenojejunalis), and left kidney. Tissue blocks were fixed overnight in 4% Formalin. Sections were stained with hematoxilin and eosin. Pathological examination was performed (M.-O. Peny). ˙ O2 crit was determined in each animal by Statistics. The D ˙ O2 vs. D˙ O2. For each plot, dual-line regression from a plot of V linear regression by best fit was used to calculate straight

Fig. 1. Changes in mean arterial pressure (MAP) and mean pulmonary arterial pressure (MPAP) in relation to incremental changes in intrapericardial pressure (IPP) in 3 groups of animals. B1, baseline; B2, 30 min after endotoxin; B3, 30 min after fluid resuscitation; B4, 30 min after administration of either NGmonomethyl-L-arginine (L-NMMA) or 3-morpholinosydnonimine (SIN-1) in treated groups. s, Endotoxin; n, endotoxin 1 L-NMMA; k, endotoxin 1 SIN-1. * P , 0.05 vs. endotoxin; † P , 0.05 vs. endotoxin 1 SIN-1.

lines for the O2 supply dependency and independency. The point of intersection of their regression lines defined the ˙ O2 crit and the corresponding critical V˙ O2 (29). ERO2 crit was D ˙ O2 to D˙ O2 at D˙ O2 crit. A two-way calculated as the ratio of V analysis of variance [for intrapericardial pressure (IPP) and group] followed by a Dunnett’s test was used for statistical ˙ O2/D˙ O2 was tested by the analysis. The difference in slopes of V analysis of covariance. A P value ,0.05 was considered statistically significant. All values are expressed as means 6 SD unless otherwise indicated. RESULTS

Effects of endotoxin. Endotoxin administration resulted in sharp decreases in MAP, cardiac filling pressures, CI, and regional blood flow, stroke index (SI), and left ventricular stroke work index (LVSWI) and in increases in pulmonary vascular resistance. Blood lactate levels also increased (Figs. 1-4). After initial fluid resuscitation, MAP remained low, but CI, regional blood flow, and SI increased above baseline, reflecting a hyperdynamic state associated with a low systemic vascular resistance (SVR; Figs. 1–4). ˙ O2/D˙ O2 relationship in indiFigures 5-7 show the V vidual dogs during progressive cardiac tamponade in each of the three groups. Effects of endotoxin 1 L-NMMA. L-NMMA administration after endotoxin significantly increased MAP compared with the control group. When blood flow was reduced by cardiac tamponade, L-NMMA-treated animals had a higher pulmonary arterial pressure (Ppa), but a lower CI, SI, and LVSWI than the SIN-1 group (Figs. 1 and 2). Both SVR and PVR were greater in the L-NMMA-treated group than in the other groups (Fig. 3). L-NMMA did not significantly influence hepatic artery blood flow but decreased blood flow to portal and mesenteric vasculature compared with the other groups (Fig. 4). L-NMMA had no significant effect on either ˙ O2 crit or ERO2 crit (Figs. 5–7). D Effects of endotoxin 1 SIN-1. SIN-1 administration had no significant influence on pressures but significantly increased CI and LVSWI compared with LNMMA-treated group (Figs. 1 and 2). SIN-1 increased

NITRIC OXIDE ON O2 EXTRACTION IN ENDOTOXEMIA

Fig. 2. Changes in cardiac index (CI), stroke index (SI), and left ventricular stroke work index (LVSWI) in relation to incremental changes in IPP in 3 groups of animals. s, Endotoxin; n, endotoxin 1 L-NMMA; k, endotoxin 1 SIN-1. § P , 0.05 vs. endotoxin 1 L-NMMA.

1167

blood flow to the hepatosplanchnic vasculature but decreased blood flow to the renal vasculature (Fig. 4). During tamponade, SIN-1 delayed the occurrence of ˙ O2 crit was significantly supply dependency, so that the D lower in the SIN-1-treated group (9.1 6 1.8 ml · kg21 · min21 ) than in the other groups (13.3 6 1.6 ml · kg21 · min21 in the L-NMMA-treated group and 12.8 6 3.3 ml · kg21 · min21 in the control group, both P , 0.05), and the ERO2 crit was higher in the SIN-1treated group (58.7 6 10.6%) than in the other groups [43.0 6 15.5% in the L-NMMA, P 5 not significant (NS); 42.2 6 7.6% in the control group, P , 0.05]. The slopes were identical in all three groups during the supplyindependent phase, whereas the slope was significantly steeper in the SIN-1-treated group than in the L-NMMAtreated group during the supply-dependent phase (P , 0.05). There were no significant differences in heart rate, cardiac filling pressures, and hematocrit level between any paired groups. Plasma lactate concentrations were identical in all three groups (data not shown). The amount of normal saline infusion was identical in the three groups of animals (4.2 6 0.8 vs. 3.5 6 0.5 liters in endotoxin alone vs. endotoxin 1 L-NMMA and vs. 3.7 6 0.5 liters in endotoxin 1 SIN-1 group; P 5 NS for both comparisons). In the three groups, hematocrit decreased from 39–40 to 29–33%. Plasma nitrite and nitrate levels were somewhat lower in the L-NMMA-treated than in the control animals at 240 min. Nitrite and nitrate levels were higher in the SIN-1-treated animals than in the other groups at 180 min and higher than the L-NMMA group at 240 min. The L-NMMA-treated group had higher TNF-a levels than the other groups at 150, 180, and 240 min (Fig. 8). Pathological examination of the liver biopsies showed a decreased number of polymorphonuclear cells, located either in the portal spaces or in the sinusoids, in both L-NMMA-treated and SIN-1-treated animals compared with the control group. The sinusoidal or centrolobular stasis and steatosis were less expressed in the SIN-1-treated group than the other groups. These

Fig. 3. Changes in systemic vascular resistance (SVR) and pulmonary vascular resistance (PVR) in relation to incremental changes in IPP in 3 groups of animals. s, endotoxin; n, endotoxin 1 L-NMMA; k, endotoxin 1 SIN-1.

1168

NITRIC OXIDE ON O2 EXTRACTION IN ENDOTOXEMIA

Fig. 4. Changes in regional blood flow during incremental changes in IPP in 3 groups of animals; values are means 6 SE. ˙ h, hepatic artery blood flow; B: Q˙ p, A: Q ˙ m, mesenportal venous blood flow; C: Q ˙ r, renal artery teric artery blood flow; D: Q blood flow. B1-B4, as in Fig. 1. s, Endotoxin; n, endotoxin 1 L-NMMA; k, endotoxin 1 SIN-1. * P , 0.05 vs. endotoxin; § P , 0.05 vs. endotoxin 1 L-NMMA; † P , 0.05 vs. endotoxin 1 SIN-1.

observations may suggest less inflammatory response in the SIN-1-treated animals than in the other groups of animals. Pathological examination of the small intestine and the kidney revealed only minor alterations, characterized by a vascular congestion, but there were no significant differences between the control and either L-NMMA-treated or SIN-1-treated animals.

˙ O2) and O2 delivFig. 5. Relation between O2 uptake (V ˙ O2 ) in control animals. Critical D˙ O2 (D˙ O2 crit ) and ery (D critical O2 extraction ratio (ERO2 crit ) are included in each plot. A-F: data for dogs 1–6 in group. Lines are dual-regression lines.

DISCUSSION

Although an increased production of NO in sepsis has been well demonstrated both in animals and humans (8, 28, 38), the beneficial effects of L-arginine analogs such as L-NMMA in septic shock have been much debated. These NOS inhibitors consistently increase

NITRIC OXIDE ON O2 EXTRACTION IN ENDOTOXEMIA

1169

˙ O2 and D˙ O2 in L-NMMAFig. 6. Relation between V ˙ O2 crit and ERO2 crit are included in treated animals. D each plot. A-F: data for dogs 1–6 in group. Lines are dual regression lines.

arterial pressure and SVR, but may reduce tissue blood flow and alter organ function (20). The present study investigated the effects of a NOS inhibitor and a NO donor on regional blood flow and O2 extraction capabilities in experimental endotoxic shock. We used a large dog model to have easy access to global and regional blood flow measurements. Also, the administration of endotoxin followed by generous fluid infusion in this model results in an initial hyperdynamic form of endotoxic shock. We applied cardiac tamponade in the second phase of the experiments to study the O2extraction capabilities of the animals. The timing of interventions may be of great importance. The rapid (5–30 min) development of hypotension in response to endotoxin in vitro and in vivo may be mediated by an enhanced release of mediators such as kinins, prostacyclin, or atrial natriuretic factor (13, 28), whereas the iNOS release contributing to the hypotension may take several hours to occur (27). It is, therefore, reasonable to hypothesize that early inhibition of NOS may produce greater vasoconstriction than needed, resulting in tissue hypoperfusion. On the other hand, the activity of the cNOS may be depressed during the early phase of endotoxemia, resulting in impaired endothelium-dependent vasodilation (24), suggesting that early administration of a NO donor may be beneficial to preserve tissue perfusion. We started L-NMMA or SIN-1 infusion 1 h after endotoxin challenge to study the effects of L-NMMA preventing excessive iNOS and

the effects of SIN-1 maintaining cNOS during early endotoxic shock. The increase in plasma nitrite/nitrate levels 180 min after endotoxin administration suggests that iNOS was significantly expressed. We recently observed in the same model that tissue iNOS was expressed in the heart, lung, liver, and small intestine 3 h after endotoxemia (unpublished data). As a NOS inhibitor, we used L-NMMA, which has been widely used in experimental studies and has been tested in humans (12). Because only one dose of the compound could be administered, we selected a moderate dose of 10 mg · kg21 · h21 to avoid excessive vasoconstricting effects. Higher doses, especially those exceeding 30 mg/kg may have harmful hemodynamic effects and may even increase mortality rates (38). The dose of 10 mg/kg has been shown to be effective and well tolerated in humans with hyperdynamic septic shock (12). L-NMMA increased arterial pressure and SVR but reduced CI and SI. Similar observations have been reported previously (6, 20, 22), although some of them involved hypodynamic models in which the fluid status might have been inadequate for endotoxic shock. In the second phase of the study, when tamponade was induced, the effects of L-NMMA on arterial pressure were not sustained. This observation was at least in part related to a more profound myocardial depression, as shown by a lower LVSWI for a similar degree of ventricular filling as in the other animals. Although an

1170

NITRIC OXIDE ON O2 EXTRACTION IN ENDOTOXEMIA

˙ O2 and D˙ O2in SIN-1-treated Fig. 7. Relation between V ˙ O2 crit and ERO2 crit are included in each plot. animals. D A-F: data for dogs 1–6 in group. Lines are dual regression lines.

excessive release of NO has been implicated in the development of sepsis-related myocardial depression (5), the effects of NOS inhibition on the endotoxininduced myocardial depression are still controversial. The studies reporting that NOS inhibitors may improve myocardial contractility usually have involved isolated myocardiums (5), but other studies involving entire organisms failed to observe such effects (16, 35). NG-nitro-L-arginine could even alter myocardial function in endotoxic shock in rats (16). Treatment with aminoguanidine, a selective iNOS inhibitor, reduced plasma nitrite and nitrate but did not affect cardiac depression (14). Results of a preliminary clinical trial showed that L-NMMA treatment failed to increase ventricular stroke work despite the frequent addition of dobutamine to L-NMMA administration (35). Taken together, these observations bring into question any beneficial effect of NOS inhibition on myocardial function in septic shock. The vasoconstricting effects of L-NMMA on the pulmonary circulation were more dramatic than on the systemic circulation. In the present model, with only moderate pulmonary hypertension, L-NMMA dramatically increased Ppa and PVR. Other studies have also emphasized that NOS inhibitors can potentiate endotoxin-induced pulmonary hypertension in endotoxic shock (20). The increase in Ppa induced by NOS inhibition may represent a major limitation to clinical use of NOS inhibitors.

The administration of L-NMMA decreased blood flow to the hepatosplanchnic circulation. Our observations are consistent with a number of recent studies underscoring that NOS inhibitors can significantly exacerbate regional vasoconstriction and ischemia (1, 22, 23, 25, 38). In endotoxic rats, Mulder et al. (22) demonstrated that NOS inhibition increased organ vascular resistance in the splanchnic vasculature during the first hour of endotoxic shock. In endotoxic pigs, Ayuse et al. (1) showed that NOS inhibition could alter local control of liver blood flow and markedly increase resistance to venous return across the liver. In endotoxic rabbits, Pastor and Payen (26) reported that Nv-nitro-Larginine significantly reduced portal vein and hepatic artery blood flows. Walker et al. (36) showed that L-NAME reduced gut blood flow by increasing gut vascular resistance in endotoxic dogs. In most of these studies, NOS inhibitors were administered during early endotoxemia or as a pretreatment, suggesting that NO plays an important role in regional hemodynamic effects and that NOS inhibition may be deleterious in acute endotoxemia. Other studies (3, 22) showed that NOS inhibition enhanced both macroscopic and histological intestinal and liver damage, but these effects were not always attributed to a reduced blood flow in this region. The effects of NOS inhibition on renal blood flow have been less well studied. Our observation of a reduced renal blood flow is in keeping with those of Spain et al.

NITRIC OXIDE ON O2 EXTRACTION IN ENDOTOXEMIA

Fig. 8. Changes in TNF-a and nitrate/nitrite concentrations in 3 groups over time. * P , 0.05 vs. endotoxin; § P , 0.05 vs. endotoxin 1 L-NMMA; † P , 0.05 vs. endotoxin 1 SIN-1.

(32), indicating that NOS inhibition can further constrict renal vasculature and decrease interlobular artery flow during sepsis. However, another recent study (2) reported that L-NMMA did not reduce renal blood flow in a sheep model of hyperdynamic sepsis. We chose SIN-1, the vasoactive metabolite of molsidomide, as a NO donor. SIN-1 spontaneously decomposes into NO and the stable metabolite N-3-morpholinoiminoacetonitrile (SIN-1C). Although SIN-1 at high doses may simultaneously generate NO and superoxide anion (17), it is an effective agent in the treatment of coronary artery disease and has effects at least as powerful as nitroglycerin. Also, Pastor et al. (25) recently demonstrated that SIN-1 administration in the early phase of endotoxic shock in rabbits can preserve systemic and hepatic perfusion while preventing lactic acidosis. We chose a dose of 2 µg · kg21 · min · 21 of SIN-1, because the improvement in myocardial function and splanchnic blood flow were most evident with this dose in previous experiments on a similar canine model of fluid-resuscitated endotoxic shock (41). Although the model used was characterized by a low vascular resistance, SIN-1 had no deleterious effect on arterial pressure. The adequate fluid loading of the animal was probably a prerequisite of such good cardiovascular tolerance, as SIN-1 may otherwise reduce venous return to the heart by its effect of venous dilation. The lack of hypotension was also related to a

1171

significant increase in CI associated with an improvement in cardiac function, as reflected by a greater LVSWI despite identical cardiac filling pressures. Such an improvement in cardiac performance after SIN-1 administration in endotoxic shock has been previously observed (41), but the exact mechanism is unclear. It may be related to a protective effect of SIN-1 on endothelial cell function or to NO-independent mechanisms related to preserved cytosolic Ca21 levels. A reduced Ppa could also be involved. SIN-1 administration selectively increased hepatic artery blood flow, compared with other groups, and maintained portal and mesenteric blood flow. Mulder et al. (22) showed that splanchnic blood flow is critically dependent on NO during the first hour of endotoxemia in rats. Boughton-Smith et al. (4) reported that exogenous supplementation of NO by S-nitroso-N-acetylpenicillamine administration could preserve gut blood flow and attenuate endotoxin-induced macroscopic jejunal damage in the rat. In a rabbit endotoxic shock model, Pastor et al. (25) recently demonstrated that pretreatment with SIN-1 could maintain portal vein blood flow at control level and significantly increase hepatic artery blood flow without any further effect on MAP. These studies suggest that NO release is essential to maintain splanchnic blood flow in the early phase of endotoxic shock. Whether these beneficial effects of SIN-1 are also present late in the course of endotoxic shock requires further study. On the contrary, SIN-1 administration decreased blood flow to the renal bed at moderate IPP compared with controls, an effect which was unexpected. When O2 supply becomes limited, the release of NO may contribute to the recruitment of unperfused capillaries and the increase in capillary density (7). However, previous studies, including those of healthy animals, demonstrated that NO inhibitors did not influence O2 extraction capabilities when blood flow was acutely reduced (15, 33, 37). During hypoxic hypoxia or ischemia in dogs, L-NAME did not affect ERO2 (31, 33). ˙ O2 crit nor maxiNG-nitro-L-arginine influenced neither D ˙ O2 mum ERO2 of the diaphragm during reductions in D by hemorrhage in dogs (37). Recently, we observed (39) that the infusion of the NO donor sodium nitroprusside ˙ O2 crit and ERO2 crit when blood flow did not influence D was acutely reduced by cardiac tamponade in dogs. Thus it appears that, in basal conditions, neither increasing exogenous NO nor blocking its release can significantly influence the O2 extraction capabilities. We investigated the effects of NO on O2 extraction capabilities during endotoxic shock. L-NMMA at a dose of 10 mg · kg21 · h · 21 did not influence tissue O2 extrac˙ O2 crit nor ERO2 crit tion capabilities, because neither D was significantly altered. These observations are consistent with those reported by Schumacker et al. (31), who demonstrated that L-NAME administration did not reverse the O2 extraction impairment seen during endotoxemia, either in whole body or in isolated intestine. Signals other than NO, such as hydrogen ion, calcium ion, and ATP products are largely related to capillary recruitment, vascular tone, and endothelial

1172

NITRIC OXIDE ON O2 EXTRACTION IN ENDOTOXEMIA

function. Although NOS inhibitors may restore tissue perfusion pressure, they may also decrease capillary density, induce capillary leak (24), activate platelet aggregation, and promote leukocyte adhesion in postcapillary venules (18). On the other hand, the NO donor SIN-1 increased the tissue O2 extraction capabilities during endotoxic shock, ˙ O2 crit and an increase in as reflected by a decrease in D ERO2 crit. NO donors may promote the metabolic vasodilation of terminal arterioles governing flow through nutritive capillaries. However, if this were the only mechanism, L-NMMA might be expected to alter O2 extraction, and this was not the case. SIN-1, like NO, can exert important antiinflammatory effects mediated by the inhibition of TNF-a (17), oxygen free radicals (21), platelet-activating factor (3), or thromboxane A2 (3), and these effects may have played a role in the improved O2 extraction capabilities. To study the interaction between NO and TNF-a, we determined TNF-a levels and found they increased after L-NMMA and SIN-1 administration. Other studies have shown that NOS inhibitors can enhance endotoxin- or bacteria-induced TNF-a production in experimental sepsis, both in vitro and in vivo (10), suggesting a regulatory feedback mechanism by which NO inhibits its own TNF-a-mediated production. Another recent study, however, showed that NO can increase TNF-a production from human neutrophils independently of guanosine 38,58-cyclic monophosphate stimulated by endotoxin in vitro (34). Kumins et al. (17) more recently showed that the NO donor molsidomine, which acts only after it is converted by the liver to SIN-1, attenuated TNF-a, interleukin-1, and interleukin-6 production in endotoxic mice. The higher TNF-a release in the L-NMMA group may contribute to the detrimental global and regional hemodynamics. A sustained release of TNF-a caused by NOS inhibition was associated with an increased toxicity of bacteria in mice (10). Although SIN-1-treated animals had higher nitrite and nitrate concentrations than the other groups, they had a pattern of TNF-a levels similar to the endotoxin-alone group, showing that TNF-a can be released by both NO-dependent and NO-independent mechanisms (9). In conclusion, the present study shows that although the NOS inhibitor L-NMMA could transiently reverse endotoxin-induced hypotension, it decreased hepatosplanchnic blood flow. These observations, made when L-NMMA was given before iNOS was expressed, stress that L-NMMA should not be administered before iNOS is induced. L-NMMA did not influence O2 extraction capabilities, suggesting that the loss in O2 extraction capabilities observed early after endotoxemia is not caused by a NO-induced loss of vascular reactivity. On the other hand, the administration of the NO donor SIN-1 could significantly increase global and hepatosplanchnic blood flow without deleterious effects on arterial pressure. SIN-1 could also improve myocardial function and increase tissue O2 extraction capabilities early during endotoxic shock.

We thank Dr. Wim A. Buurman (Department of Surgery, Immunology Laboratory, Rijksuniversiteit Limburg, Maastricht, The Netherlands) for the TNF-a determinations, Dr. Bernard Vray (Department of Immunology, Faculty of Medicine, Free University of Brussels, Belgium) for the nitrite/nitrate measurement, and Therabel Research for supplying SIN-1. This study was supported in part by Grant GD294.1.CH0380 from Fonds de la Recherche Scientifique Medicale, Belgium. Address for reprint requests: J.-L. Vincent, Dept. of Intensive Care, Erasme Univ. Hospital, Route de Lennik 808, B-1070 Brussels, Belgium (E-mail:[email protected]). Received 14 January 1997; accepted in final form 2 June 1997. REFERENCES 1. Ayuse, T., N. Brienza, J. P. Revelly, J. K. Boitnott, and J. L. Robotham. Role of nitric oxide in porcine liver circulation under normal and endotoxemic conditions. J. Appl. Physiol. 78: 1319– 1329, 1995. 2. Booke, M., F. Hinder, R. McGuire, L. D. Traber, and D. L. Traber. Nitric oxide synthase inhibition versus norepinephrine for the treatment of hyperdynamic sepsis in sheep. Crit. Care Med. 24: 835–844, 1996. 3. Bordet, J. C., and M. Lagard. Modulation of prostacyclin/ thromboxane formation by molsidomine during platelet-endothelial cell interactions. Biochem. Pharmacol. 37: 3911–3914, 1988. 4. Boughton-Smith, N. K., I. R. Hucheson, and A. M. Deakin. Protective effect of S-nitroso-N-acetyl-penicillamine in endotoxininduced acute intestinal damage in the rat. Eur. J. Pharmacol. 191: 485–488, 1994. 5. Brady, A. J., P. A. Poole-Wilson, S. E. Harding, and J. B. Warren. Nitric oxide production within cardiac myocytes reduces their contractility in endotoxemia. Am. J. Physiol. 263 (Heart Circ. Physiol. 32): H1963–H1966, 1992. 6. Cobb, J. P., C. Natanson, W. D. Hoffman, R. F. Lodato, S. Banks, C. A. Koev, M. A. Solomon, R. J. Elin, J. M. Hosseini, and R. L. Danner. Nw-amino-L-arginine, an inhibitor of nitric oxide synthase, raises vascular resistance but increases mortality rates in awake canines challenged with endotoxin. J. Exp. Med. 176: 1175–1182, 1992. 7. Drazenovic, R., R. W. Samsel, M. E. Wylam, C. M. Doerschuk, and P. T. Schumacker. Regulation of perfused capillary density in canine intestinal mucosa during endotoxemia. J. Appl. Physiol. 72: 259–265, 1992. 8. Evans, T., A. Carpenter, H. Kinderman, and J. Cohen. Evidence of increased nitric oxide production in patients with the sepsis syndrome. Circ. Shock 41: 77–81, 1993. 9. Fahey, T. J., III, T. Yoshioka, G. T. Shires, and G. A. Fantini. The role of tumor necrosis factor and nitric oxide in the acute cardiovascular response to endotoxin. Ann. Surg. 223: 63–69, 1996. 10. Florquin, S., Z. Amraoui, C. Dubois, J. Decuyper, and M. Goldman. The protective role of endogenously synthesized nitric oxide in staphylococcal enterotoxin B-induced shock in mice. J. Exp. Med. 180: 1153–1158, 1994. 11. Fong, Y., M. A. Marano, and L. L. Moldawer. The acute splanchnic and peripheral tissue metabolic response to endotoxin in humans. J. Clin. Invest. 85: 1896–1904, 1990. 12. Grover, R., D. Zaccardelli, G. Colice, K. Guntupalli, D. Waston, and J.-L. Vincent. The cardiovascular effects of 546C88 in human septic shock (Abstract). Intensive Care Med. 21, Suppl.: S20, 1995. 13. Hinder, F., M. Booke, L. D. Traber, N. Matsumoto, K. Nishida, S. Rogers, and D. L. Traber. Nitric oxide synthase inhibition during experimental sepsis improves renal excretory function in the presence of chronically increased atrial natriuretic peptide. Crit. Care Med. 24: 131–136, 1996. 14. Hock, C. E., K. Yin, G. Yue, and P. Y. K. Wong. Effects of inhibition of nitric oxide synthase by aminoguanidine in acute endotoxemia. Am. J. Physiol. 272 (Heart Circ. Physiol. 41): H843–H850, 1997. 15. King, C. E., M. J. Melinyshyn, J. D. Mewburn, S. E. Curtis, M. J. Winn, S. M. Cain, and C. K. Chapler. Canine hindlimb blood flow and O2 uptake after inhibition of EDRF/NO synthesis. J. Appl. Physiol. 76: 1166–1171, 1994.

NITRIC OXIDE ON O2 EXTRACTION IN ENDOTOXEMIA 16. Klabunde, R. E., and A. F. Coston. Nitric oxide synthase inhibition does not prevent cardiac depression in endotoxic shock. Shock 3: 73–78, 1995. 17. Kumins, N. H., J. Hunt, R. L. Gamelli, and J. P. Filkins. Molsidomine increases endotoxic survival and decreases cytokine production. Shock 7: 200–205, 1997. 18. Kurose, I., R. Wolf, M. B. Grisham, and D. N. Granger. Effects of an endogenous inhibitor of nitric oxide synthesis on postcapillary venules. Am. J. Physiol. 268 (Heart Circ. Physiol. 37): H2224–H2231, 1995. 19. Meyer, J., F. Hinder, J. Stothert, Jr., L. D. Traber, D. N. Herndon, J. T. Flynn, and D. L. Traber. Increased organ blood flow in chronic endotoxemia is reversed by nitric oxide synthase inhibition. J. Appl. Physiol. 76: 2785–2793, 1994. 20. Mitaka, C., Y. Hirata, K. Ichikawa, T. Uchida, K. Yokoyama, T. Nagura, Y. Tsunoda, and K. Amaha. Effects of nitric oxide synthase inhibitor on hemodynamic change and O2 delivery in septic dogs. Am. J. Physiol. 268 (Heart Circ. Physiol. 37): H2017–H2023, 1995. 21. Moilanen, E., P. Vuorinen, H. Kankaanranta, T. MetsaKetela, and H. Vapaatalo. Inhibition by nitric oxide-donors of human polymorphonuclear leucocyte functions. Br. J. Pharmacol. 109: 852–858, 1993. 22. Mulder, M. F., A. A. Van Lambalgen, E. Huisman, J. J. Visser, G. C. Van Den Bos, and L. G. Thijs. Protective role of NO in the regional hemodynamic changes during acute endotoxemia in rats. Am. J. Physiol. 266 (Heart Circ. Physiol. 35): H1558–H1564, 1994. 23. Nishida, J., R. S. McCuskey, D. McDonnell, and E. S. Fox. Protective role of NO in hepatic microcirculatory dysfunction during endotoxemia. Am. J. Physiol. 267 (Gastrointest. Liver Physiol. 30): G1135–G1141, 1994. 24. Parker, J. L., and H. R. Adams. Selective inhibition of endothelium-dependent vasodilator capacity by Escherichia coli endotoxemia. Circ. Res. 72: 539–551, 1993. 25. Pastor, C. M., M.-R. Losser, and D. Payen. Nitric oxide donor prevents hepatic and systemic perfusion decrease induced by endotoxin in anesthetized rabbits. Hepatology 22: 1547–1553, 1995. 26. Pastor, C. M., and D. M. Payen. Effect of modifying nitric oxide pathway on liver circulation in a rabbit endotoxin shock model. Shock 2: 196–202, 1994. 27. Salter, M., R. G. Knowles, and S. Moncada. Widespread tissue distribution, species distribution and changes in activity of Ca21-dependent and Ca21-independent nitric oxide synthase. FEBS Lett. 291: 145–149, 1991. 28. Salvemini, D., R. Korbut, E. Anggard, and J. R. Vane. Immediate release of nitric oxide-like factor from bovine aortic

29. 30. 31.

32. 33. 34.

35.

36.

37. 38. 39.

40.

41.

1173

endothelial cells by Escherichia coli lipopolysaccharide. Proc. Natl. Acad. Sci. USA 87: 2593–2597, 1990. Samsel, R. W., and P. T. Schumacker. Determination of the critical O2 delivery from experimental data: sensitivity to error. J. Appl. Physiol. 64: 2074–2082, 1988. Schlichtig, R., H. A. Klions, D. J. Kramer, and E. M. Nemoto. Hepatic dysoxia commences during O2 supply dependence. J. Appl. Physiol. 72: 1499–1505, 1992. Schumacker, P. T., J. Kazaglis, H. V. Connolly, R. W. Samsel, M. F. O’Connor, and J. G. Umans. Systemic and gut O2 extraction during endotoxemia: role of nitric oxide synthesis. Am. J. Respir. Crit. Care Med. 151: 107–115, 1995. Spain, D. A., M. A. Wilson, I. T. M. Bloom, and R. N. Garrison. Renal microvascular responses to sepsis are dependent on nitric oxide. J. Surg. Res. 56: 524–529, 1994. Vallet, B., S. E. Curtis, M. J. Winn, C. E. King, C. K. Chapler, and S. M. Cain. Hypoxic vasodilation does not require nitric oxide (EDRF/NO) synthesis. J. Appl. Physiol. 76: 1256–1261, 1994. Van Dervort, A. L., L. Yan, P. J. Madara, J. P. Cobb, R. A. Wesley, C. C. Corriveau, M. M. Tropea, and R. L. Danner. Nitric oxide regulates endotoxin-induced TNF-a production by human neutrophils. J. Immunol. 152: 4102–4109, 1994. Vincent, J.-L., G. Colice, R. Grover, D. Zaccardelli, K. Guntupalli, and D. Watson. The effects of 546C88 on left ventricular performance in patients with septic shock (Abstract). Intensive Care Med. 21, Suppl.: S20, 1995. Walker, T. A., S. E. Curtis, C. E. King-Van Vlack, C. K. B. Vallet, and S. M. Cain. Effects of nitric oxide synthase inhibition on regional hemodynamics and oxygen transport in endotoxic dogs. Shock 4: 415–420, 1995. Ward, M. E., and S. N. A. Hussain. Effect of inhibition of nitric oxide release on the diaphragmatic oxygen delivery-consumption relationship. J. Crit. Care 9: 90–99, 1994. Wright, C. E., D. D. Rees, and S. Moncada. Protective and pathological roles of nitric oxide in endotoxin shock. Cardiovasc. Res. 26: 48–57, 1992. Zhang, H., D. N. Nguyen, H. Spapen, M. Moock, F. Maciel, and J.-L. Vincent. Sodium nitroprusside does not influence tissue oxygen extraction capabilities during a critical reduction in oxygen delivery. Cardiovasc. Res. 30: 240–245, 1995. Zhang, H., H. Spapen, D. N. Nguyen, M. Benlabed, W. A. Buurman, and J.-L. Vincent. Protective effects of N-acetyl-Lcysteine in endotoxemia. Am. J. Physiol. 266 (Heart Circ. Physiol. 35): H1746–H1754, 1994. Zhang, H., P. Rogiers, H. Spapen, G. Friedman, J.-C. Preiser, and J.-L. Vincent. Effects of nitric oxide donor SIN-1 on oxygen availability and regional blood flow during endotoxic shock. Arch. Surg. 131: 767–774, 1996.