Inducible nitric oxide synthase inhibition improves intestinal microcirculatory oxygenation and CO2 balance during endotoxemia in pigs
- First Online:
- Cite this article as:
- Siegemund, M., van Bommel, J., Schwarte, L.A. et al. Intensive Care Med (2005) 31: 985. doi:10.1007/s00134-005-2664-7
- 88 Views
We examined whether selective inhibition of inducible nitric oxide synthase (iNOS) promotes intestinal microvascular oxygenation (µPO2) and CO2 off-load after endotoxic shock.
Design and setting
Prospective, controlled experimental study in a university animal research laboratory.
13 domestic pigs.
After baseline measurements shock was induced by 1 µg kg−1 h−1 endotoxin until mean arterial pressure fell below 60 mmHg. After 30 min in shock the animals were resuscitated with either fluid alone (control, n=6) or fluid and the iNOS inhibitor N-[3-(aminomethyl)benzyl]acetamidine hydrochloride (1400W, n=7). As final experimental intervention all animals received the nonselective NOS inhibitor L-NAME.
Measurements and results
Systemic and regional hemodynamic and oxygenation parameters were measured at baseline, during endotoxemia and shock, hourly for 3 h of 1400W therapy, and 30 min after the final L-NAME administration. µPO2 was assessed by the Pd-porphyrin phosphorescence technique, and the arterial to intestinal PCO2 gap was determined by air tonometry. Endotoxemia and shock resulted in a decrease in ileal mucosal and serosal µPO2 and a rise in PCO2 gap. The combination of 1400W and fluid resuscitation, but not fluid alone, normalized both the serosal µPO2 and the intestinal PCO2 gap. Administration of L-NAME decreased cardiac output and oxygen delivery and intestinal µPO2 and blood flow in both groups.
Partial blockade of NO production by 1400W increased serosal microvascular oxygenation and decreased the intestinal CO2 gap. This findings are consistent with the idea that 1400W corrects pathological flow distribution and regional dysoxia within the intestinal wall following endotoxic shock.
KeywordsShock, septicEndotoxemia, physiopathologyRegional blood flow, drug effectsMicrocirculation, physiopathologyOxygen consumption, drug effectsNitric-oxide synthase
Endotoxemia and septic shock are characterized by components of hypovolemic, obstructive, cardiogenic, distributive, and cytopathic shock . Nitric oxide generated by the inducible isoform of nitric oxide synthase (iNOS) is believed to be responsible at least in part for the hypovolemic , cardiogenic , distributive [3, 4], and cytopathic [5, 6] alterations during septic shock. After adequate fluid resuscitation the presence and severity of hypotension are directly dependent on impairment of cardiac contractility and the degree of reduction in systemic vascular resistance. One of the main mechanisms causing this condition is thought to be due to the action of excessive nitric oxide amounts being synthesized by the sepsis-induced expression of iNOS. Under these circumstances the administration of an NOS inhibitor should restore blood pressure, increase vascular resistance, and normalize maldistribution of microvascular organ, blood, and tissue oxygenation . However, a recently published study using a nonselective NOS inhibitor showed a rise in mean arterial blood pressure and systemic vascular resistance but had to be terminated because of increased mortality in patients treated by NOS inhibition . A possible reason for the detrimental effect of unselective NOS inhibition is the simultaneous blockade of constitutive endothelial and neural NOS isoforms and their physiological effects, such as pulmonary vasodilatation. Such results have led to the idea that more selective inhibition of iNOS would have beneficial effects during endotoxemia and sepsis. In animal models selective iNOS inhibition prevented endotoxin induced organ dysfunction [8, 9, 10], normalized intestinal mucosal to arterial PCO2 gap [11, 12, 13, 14], and reversed respiratory inhibition .
The microcirculation in sepsis is severely altered with an increase in stopped flow and fast-flow capillaries, leading to a decreased capillary density and increased perfusion heterogeneity [16, 17, 18, 19]. The intestinal wall is especially vulnerable to endotoxin related heterogeneity in blood flow [20, 21]. In fluid-resuscitated animals the mucosa was perfused at the expense of submucosal and muscularis layers [22, 23]. Improvement in intestinal microcirculatory blood flow distribution, altered by iNOS generated NO, should normalize oxygen supply, support CO2 transport from tissues and even improve impaired oxygen consumption [3, 23]. This would be especially relevant in the light of a heterogeneous iNOS expression in the intestinal tract with a maximum in the villus tip and a smaller amount of expression in other locations, such as crypt cells . Inhibition of iNOS would also be expected to improve microvascular autoregulation and mitochondrial ATP production which are disturbed during excessive NO production such as in endotoxemia [21, 25].
In this study we hypothesized that treatment of endotoxic shock by selective iNOS blockade using 1400W  in combination with adequate fluid resuscitation would improve regional, microvascular oxygenation, and CO2 removal . To test this hypothesis we simultaneously measured global and regional oxygen consumption parameters and organ blood flow. Microcirculatory oxygen pressures (µPO2) in serosa and mucosa of the ileum and intestinal PCO2 (PiCO2) were measured in a normodynamic, low-dose endotoxic pig model.
Material and methods
This study was performed in accordance with the Dutch national guidelines for care of laboratory animals and approved by the Committee of ethics for Animal Research of the Academic Medical Center at the University of Amsterdam. Thirteen male crossbred Landrace Yorkshire pigs with a median body weight of 28.5 kg (range 25–31) were anesthetized by 5 mg/kg thiopental sodium (Nesdonal, Rhône-Poulenc Rorer, Amstelveen, The Netherlands) and maintained by continuous infusion of 0.2 mg kg−1 h−1 midazolam (Dormicum, Roche, Mijdrecht, The Netherlands), 0.01 mg kg−1 h−1 fentanyl (Fentanyl, Janssen Pharmaceutica, Tilburg, The Netherlands), and 0.1 mg kg−1 h−1 pancuronium (Pavulon, Organon Technik, Boxtel, The Netherlands). The animals were intubated and mechanically ventilated (tidal volume 15 ml/kg; AV-1, Dräger-Werke, Lübeck, Germany) using an oxygen in air gas mixture (FIO2 0.35). Lactated Ringer’s solution (15 ml kg−1 h−1) was administered as continuous infusion together with anesthetic drugs via the right brachial vein. The core temperature was maintained at 37.2±0.5°C using an electrical heating pad.
A midline laparatomy was performed and a catheter was inserted into the bladder for the measurement of urine production. A flow probe (4.0 mm, Transonic Systems, Ithaca, N.Y., USA) was placed around the superior mesenteric artery to continuously measure intestinal blood flow. A catheter (6 F) was inserted in an ileal branch of the superior mesenteric vein to collect blood for intestinal gas analysis during the experiment and for endotoxin infusion. A 10- to 15-cm segment of the terminal ileum was mobilized, placed outside the abdomen, and opened 3–5 cm on its antimesenteric border. The edges were fixed on the skin with the mucosa facing upward to measure serosal and mucosal µPO2. In a second ileal segment a TRIP sigmoid tonometer (Tonometrics Divison, Instrumentarium, Helsinki, Finland) was placed through a minimal antimesenteric wall incision, secured with a purse-string suture, and attached to a Tonocap (Tonometrics) to measure the CO2 partial pressure of the intestinal wall. After all preparations were completed the abdomen was loosely closed with sutures. The exteriorized surfaces were regularly irrigated with warmed Ringer’s solution and inspected for tissue viability.
Measurements and calculations
Mean arterial pressure (MAP) and arterial blood gases were measured by a catheter in the right brachial artery. A pulmonary artery thermodilution catheter (Edwards 7 F, Baxter Healthcare, Irvine, Calif., USA) was placed in the pulmonary artery to measure cardiac output, temperature, central venous pressure, pulmonary artery pressures, pulmonary artery occlusion pressure (PAOP) and mixed venous blood gases. Cardiac output was determined by triplicate measurements with a computer (Vigilance; Baxter Healthcare, Round Lake, Ill., USA) and normalized according to body weight (ml kg−1 min−1). At each time point hemoglobin the oxygen saturation and blood gas values were taken from the brachial artery, pulmonary artery, and superior mesenteric vein. They were analyzed by a hemoximeter and a blood gas analyzer (OSM 3 and ABL 505, Radiometer, Copenhagen, Denmark). Lactate measurements were determined from whole blood by an enzymatic procedure.
Systemic and intestinal O2 delivery (DO2) and consumption (VO2) were calculated using standard formulas. Intestinal DO2 and VO2 were calculated from the same formulas using mesenteric artery flow instead of cardiac output. Oxygen extraction rate was calculated as the venous-arterial O2 content difference divided by the arterial O2 content. The PiCO2 gap was calculated as difference between PiCO2 and arterial PCO2 to take systemic confounders into account .
Determination of the microvascular oxygen pressures
Microvascular oxygen partial pressure of ileal serosa and mucosa was measured by oxygen-dependent quenching of Pd-porphyrin phosphorescence. Excitation of Pd-porphyrin by a pulse of light causes emission of phosphorescence. The decay of phosphorescence is quantitatively related to the oxygen concentration, as described by the Stern-Volmer relationship [29, 30]. Since less oxygen results in longer decay times, this technique is especially sensitive for the detection of hypoxia. Pd-meso-tetra(4carboxy-phenyl)porphine (Porphyrin, Logan, Utah, USA) dissolved in 1.5 ml DMSO was bound to human albumin in saline (40 g/l). Injected intravenously this large molecular complex is confined mainly to the microvascular compound . Of the Pd-porphyrin solution 12 mg/kg was infused in experimental animals, and two optical fibers of a multifiber phosphorimeter described previously were placed near the intestinal mucosa and serosa without direct tissue contact [4, 29]. The measurement incorporates the µPO2 of blood vessels under the optic fiber over an area of approx. 1 cm2 to a penetration depth of about 0.5 mm . As the calibration constants in the calculation of the PO2 from the phosphorescence decay time are temperature dependent, intestinal surface temperature measurements were used for correction of these constants. An average value for µPO2 was calculated from 15 measurements taken during a 5-min period. In previous studies combining microscopic and fiberoptic phosphorimetry simultaneously, we demonstrated that the fibers selectively measure the PO2 in capillaries and first-order venules, which led us to term the measurement as microvascular PO2 (µPO2) .
After a stabilization period of 60 min following surgical preparation two baseline measurements were made 30 min apart. Because the consecutive measurements were not significantly different in any parameter, they were averaged and presented as a single data point (baseline). After the second baseline measurement endotoxemia was induced by infusion of 1 µg kg−1 h−1 Escherichia coli lipopolysaccharide (serotype O127:B8; Sigma Chemical, St. Louis, Mo., USA) via the mesenteric vein. The endotoxin infusion was stopped when the MAP fell below 60 mmHg. During the following 30 min the MAP was stabilized around 60 mmHg by the infusion of additional Ringer’s lactate, and two series of measurements were taken during shock (shock).
We randomly assigned 13 pigs to two groups using fluid resuscitation with (1400W group, n=7) or without 1400W (control group, n=6). A bolus of 500 ml hetastarch (HAES 6%) and a continuous infusion of 30 ml kg−1 h−1 Ringer’s lactate were administered after shock. Additional Ringer’s lactate and HAES 6% were infused if PAOP remained below 12 mmHg. The control group received fluid resuscitation only. In the treatment group 0.5 mg kg−1 h−1 1400W [N-(3-aminomethyl)benzyl]acetamidine 2 HCl, Alexis Biochemicals, Lausen, Switzerland) was administered via the right brachial vein in addition to the fluid resuscitation until the end of the experiment . All hemodynamic and oxygenation measurements were made every hour during resuscitation (t1, t2, t3). To study the effect of a blockade of all NOS isoforms all 13 animals received 5 mg/kg of the nonselective NOS inhibitor Nω -nitro-l-arginine methylester HCl (L-NAME; Sigma-Aldrich Chemicals, Steinheim, Germany) towards the end of the experiment. The last set of measurements was taken 30 min after administration of L-NAME. Euthanasia was performed by administering 30 mmol potassium chloride.
Data are reported as median and range unless otherwise stated. Differences compared to baseline were analyzed using Friedman’s repeated-measures analysis of variance on ranks within each group, and a subsequent Dunn’s test for multiple comparisons using GraphPad Prism version 4.0 for Windows (GraphPad Software, San Diego Calif., USA, http://www.graphpad.com). Differences between 1400-W and control animals during shock and resuscitation were analyzed using the Mann-Whitney rank sum test. Differences with p values less than 0.05 were considered statistically significant.
All hemodynamic and oxygenation variables were stable for 30 min before the first baseline measurement. Animals in the two groups received similar amounts of continuously administered endotoxin [control 49.0 µg (range 27–54) vs. 1400W 51.1 µg (39–65)] until shock was established. During the study period we administered comparable amounts in the two groups of Ringer’s lactate [control 6500 ml (6000–9300) vs. 1400W 6600 ml (5500–8000)] and HAES 6% [(control 1000 ml (500–1500) vs. 1400W 1300 ml (1000–1500)]. The mean urine production did not differ significantly between the groups [(control 820 ml (400–1200) vs. 1400W 670 ml (410–1300)].
Systemic hemodynamic, oxygen consumption and acid base parameters in the 1400W (n=7) control (n=6) groups: median (parentheses range). (t1, t2, t3 = 1, 2, 3 h of resuscitation, L-NAME 30 min after administration of Nω-nitro-l-arginine methylester, MAP mean arterial pressure, PAP mean pulmonary arterial pressure, PAOP pulmonary artery occlusion pressure, CI cardiac output normalized to body weight, SVR systemic vascular resistance, PVR pulmonary vascular resistance, DO2 whole-body oxygen delivery, VO2 whole-body oxygen consumption, O2ER whole-body oxygen extraction, pHa arterial pH, PaO2 arterial PO2, PaCO2 arterial PCO2)
Heat rate (min-1)
CI (ml kg−1 min−1)
SVR (dyne s−1 cm−5)
PVR (dyne s−1 cm−5)
DO2 (ml kg−1 min−1)
VO2 (ml kg−1 min−1)
Intestinal hemodynamic, oxygen consumption, and acid base parameters in the 1400W (n=7) and control (n=6) groups: median (parentheses range). (t1, t2, t3 = 1, 2, 3 h of resuscitation, L-NAME indicates measurements 30 min after administration of Nω-nitro-l-arginine methylester, Qsma superior mesenteric artery flow, DiO2 intestinal oxygen delivery, ViO2 intestinal oxygen consumption, O2ERi intestinal oxygen extraction, pHmv mesenteric venous pH, PaO2 arterial PO2, PmvO2 mesenteric venous PO2, PmvCO2 mesenteric venous PCO2, lactatemv mesenteric venous lactate)
Qsma (ml min−1 kg−1)
DiO2 (ml kg−1 min−1)
ViO2 (ml kg−1 min−1)
Administration of L-NAME towards the end of the experiment increased vascular resistances and decreased cardiac index and oxygen delivery in both groups significantly (Table 1). Intestinal microvascular oxygenation and oxygen delivery parameters decreased to similar values as in shock (Fig. 1). The negative effect of L-NAME on cardiac index and intestinal oxygenation was more pronounced in the control than the 1400W treatment group (Table 1).
This study tested the hypothesis that 1400W improves regional, microvascular oxygen availability and microcirculatory flow distribution and thereby oxygen utilization during fluid resuscitated endotoxic shock. The major finding was that 1400W, a highly selective iNOS inhibitor , increases mucosal and serosal µPO2 and normalizes the arterial-intestinal PCO2 gap after endotoxic shock. Fluid resuscitation alone successfully restored mucosal µPO2 but neither the serosal µPO2 nor the intestinal PCO2 gap. We believe that 1400W corrects pathological flow distribution and regional dysoxia within the intestinal wall following endotoxic shock. To investigate the effects of a blockade of all NOS isoforms the nonselective inhibitor L-NAME was applied towards the end of the experiment. L-NAME increased MAP but reduced blood flow and deteriorated microvascular oxygenation.
In our study cardiac output did not decrease after the rise in MAP and systemic vascular resistance after 1400W. In contrast, the administration of the nonselective NOS inhibitor L-NAME decreased cardiac output and organ blood flow to critical levels in both groups, indicated by a significant drop in mixed venous saturation. The stable cardiac output despite a significant increased afterload in the presence of a comparable pulmonary artery occlusion pressure suggests improved cardiac contractility due to 1400W. These findings are in accordance with other studies which have reported that selective iNOS inhibition has a positive inotropic effect [13, 14, 34]. Cohen et al.  found that iNOS is constitutively expressed in pig hearts, and that endotoxin downregulates iNOS mRNA and protein expression. The application of a selective iNOS inhibitor increased mean arterial pressure and prevented an increase in myocardial blood flow seen with fluid resuscitation alone. Whether a decreased amount of NO improved cardiac contractility due to lesser enzyme dysfunction (e.g., L-type calcium channel), a blunted inflammatory reaction and granulocyte accumulation or a prevented mitochondrial inhibition remains unclear . The unaffected CO2 gap after L-NAME despite a decreased blood flow may be due to the relative short accumulation and equilibration time for intestinal CO2 after L-NAME administration. The negative effect of L-NAME on cardiac function has been reported in earlier studies [14, 36, 37]. Remarkable was the more pronounced effect on cardiac index and blood flow in the control group compared to the 1400W treated animals. Improved contractility after 1400W possibly supports ventricular function during the rapid afterload increase by L-NAME due to a conserved coronary flow reserve .
That mucosa is preferentially perfused on the expense of muscularis and serosa during endotoxin induced shock and resuscitation has previously been reported in studies using microspheres [20, 21, 22]. A low-pressure left-to-right shunt in the mucosa induces heterogeneity of flow in other intestinal wall layers. An increased number of unperfused capillaries and a decreased ratio of normally perfused to overperfused capillaries have been found [16, 17, 38]. A possible underlying cause of this pathological flow distribution in the intestinal microcirculation is a heterogeneous distribution of iNOS along the crypt-villus axis . To date several reports have shown a maximal increase in intestinal iNOS 3–6 h after endotoxin either by messenger RNA expression, immunostaining or enzymatic activity in rodents, cats, and pigs [24, 35, 39, 40, 41]. Cohen et al.  reported a 7.6-fold increase in iNOS protein in the liver of endotoxic pigs. In their study the administration of the selective iNOS inhibitor S-methylisothiourea resulted in a normal hepatic blood flow . In our study the marked effect of 1400W on the µPO2 of the serosa and subserosa (e.g., muscularis) together with the improvement in the intestinal PCO2 gap strongly suggests redistribution of microcirculatory flow. We believe that selective blockade of NO production primarily decreases the blood flow in the mucosal villus tips and thereby redirects blood flow to deeper intestinal wall layers.
In three recent studies Radermacher and colleagues [11, 12, 13] found improvement in the intestinal to arterial PCO2 gap after the application of 1400W, whereas fluid resuscitation did not improve the PCO2 gap in their porcine long-term endotoxin model. In a short-term endotoxic shock model in pigs the administration of the iNOS inhibitor S-methylisothiourea restored the CO2 gap to baseline levels as well . Microsphere studies showed that hypoperfusion of the muscularis, and not the mucosa, was related to intestinal PCO2 [20, 21]. Van der Meer et al.  first reported that resuscitation of endotoxic shock by means of colloids improved the mucosal oxygenation to values above baseline while intestinal CO2 remained high. They concluded that dysfunction of oxygen utilization (e.g., cytopathic hypoxia) in mitochondria is a primary dysfunction in endotoxic shock and sepsis. Our study confirmed that reduced mucosal µPO2 after endotoxic shock was easily resuscitated in both groups, and that fluid alone did not decrease the intestinal CO2 gap. Furthermore we found a normal CO2 gap and an improved serosal microvascular oxygenation after the application of 1400W. Matejovic et al.  also found a partially restored microcirculatory flow and a normal CO2 gap after endotoxemia and the administration of the even more selective iNOS inhibitor l-N6-(1-iminoethyl)-lysine. These findings are consistent with the notion that the increased intestinal to arterial CO2 gap in early endotoxic and septic shock is caused by a hypoperfusion of other intestinal wall layers than the mucosa . The restored intestinal CO2 gap together with the increased serosal µPO2 may indicate an improved perfusion of submucosal, muscularis, and serosal wall layers by 1400W.
Cytopathic hypoxia, the impaired oxygen utilization by mitochondrial respiration in the presence of adequate amounts of oxygen may have been present in our experiment . Irreversible inhibition of mitochondrial respiration and damage to a variety of mitochondrial components is caused by peroxynitrite via oxidizing reactions and through activation of poly(ADP-ribose) polymerase . Higher concentrations of peroxynitrite in tissues resulted from increased amounts of NO during endotoxemia [39, 44]. In physiological concentrations NO specifically and reversibly inhibits cytochrome oxidase in competition with oxygen . Correlations between excessive NO production due to iNOS and dysfunction of the intestinal mucosa as a result of impaired oxygen metabolism have been reported [40, 45]. Changed mitochondrial ultrastructure  and decreased mucosal ATP concentration  have been found. Borutaite and colleagues , for example, showed a 90% inhibition of cellular oxygen consumption in vascular cells after endotoxin, an effect that was reversed by 1400W. In our experiments we measured only oxygen consumption, extraction, and lactate, all indirect parameters of oxygen utilization and mitochondrial respiration. Lactate values were not restored to baseline by1400W. Possible reasons for this could be persistent microcirculatory flow heterogeneity and respiratory inhibition in other organs, such as the kidney , or an impaired and prolonged lactate metabolism in the liver. Global and regional oxygen consumption did not change during endotoxic shock and resuscitation. Both total body and intestinal oxygen extraction showed a clear but not significant increase with 1400W compared to control. The increased O2ER in the control group contradicts an irreversible inhibition of mitochondria due to endotoxin. The administration of 1400W possibly prevented a reversible mitochondrial respiratory inhibition, either by decreased NO production or by improved microvascular oxygenation of the serosal and subserosal intestinal wall layers. Both mechanisms suspend a reversible inhibition of the cytochrome oxidase c by NO due to an increased relative amount of oxygen in mitochondria. Indeed our measurements could not definitely exclude the presence of cytopathic hypoxia.
The administration of 1400W together with a doubling of the infusion rate from 15 to 30 ml kg–1 h–1 significantly increased the mean arterial pressure and systemic vascular resistance compared to the control group. This is in accordance with other short- and long-term endotoxin models in pigs [11, 13, 47] showing a slight to moderate increase. The considerable effect on pressure and resistance in our experiment is possibly due to the larger quantity of infusion. Another possible explanation for this finding is a partial effect of 1400W on constitutively expressed endothelial NOS in the absence of iNOS. Previous experiments have been contradictory to the presence [35, 44, 48] or absence [47, 49] of intestinal iNOS after short-term endotoxemia in pigs. Our animals received 1400W 2–3 h after the start of endotoxin infusion via the superior mesenteric vein. Because this is approximately the same, albeit short, timeframe as in the above study by Cohen et al. , we presume that iNOS is present in the hepatosplanchnic circulation of the animals. Furthermore we found a microvascular oxygenation consistent with the normal intestinal blood flow distribution at baseline after the highly selective iNOS inhibitor 1400W was given . The doubled systemic and pulmonary vascular resistance in both groups after the administration of the nonselective NOS blocker L-NAME suggests the presence of comparable amounts of constitutive endothelial NOS. Whether 1400W improves intestinal perfusion by a blockade of endotoxin-induced iNOS or constitutively expressed endothelial NOS or iNOS is not definitively clear from our results. On the basis of the limitations discussed above our results may not be directly assigned to pathophysiological alterations in human bacterial sepsis.
In summary, our results demonstrate that 1400W increases intestinal microvascular oxygenation and PCO2 gap together with an increased oxygen extraction in a short-term, normodynamic, low-dose endotoxic shock model. Together with early goal-directed therapy  partial iNOS inhibition may be a future therapeutic alternative for the distributive and cardiogenic components of endotoxemia and septic shock.