Endogenous nitric oxide in the maintenance of rat microvascular integrity against widespread plasma leakage following abdominal laparotomy

Abstract

1. The role of nitric oxide (NO) in the maintenance of microvascular integrity during minor surgical manipulation has been evaluated in the rat.

2. The NO synthase inhibitors, N^G-nitro-L-arginine methyl ester (L-NAME, 5 mg kg⁻¹, s.c.) and N^G-monomethyl-L-arginine (L-NMMA, 50 mg kg⁻¹, s.c.) had no effect on microvascular leakage of radiolabelled albumin over 1 h in the stomach, duodenum, jejunum, colon, lung and kidney in the un-operated conscious or pentobarbitone-anaesthetized rat.

3. In contrast, in anaesthetized rats with a midline abdominal laparotomy (5 cm), L-NAME (1–5 mg kg⁻¹, s.c.) or L-NMMA (12.5–50 mg kg⁻¹, s.c.) dose-dependently increased gastrointestinal, renal and pulmonary vascular leakage, effects reversed by L-arginine pretreatment (300 mg kg⁻¹, s.c., 15 min). These actions were not observed in anaesthetized rats that had only received a midline abdominal skin incision (5 cm).

4. Pretreatment with a rabbit anti-rat neutrophil serum (0.4 ml kg⁻¹, i.p.), 4 h before laparotomy, abolished the plasma leakage induced by L-NAME in all the organs investigated.

5. These results indicate that the following abdominal laparotomy, inhibition of constitutive NO synthase provokes vascular leakage in the general microcirculation, by a process that may involve neutrophils. Such effects could thus confound studies on the microvascular actions of NO synthase inhibitors using acute surgically prepared in vivo models. The findings thus suggest that constitutively-formed NO has a crucial role in the maintenance of acute microvascular integrity following abdominal surgical intervention.

Introduction

Nitric oxide (NO), synthesized in the vascular endothelium, by a calcium and calmodulin-dependent constitutive isoform of nitric oxide synthase (eNOS), plays a significant role in the regulation of vascular tone (Moncada & Higgs, 1995). Moreover, NO inhibits aggregation and adhesion of platelets to the vascular endothelium (Radomski et al., 1992; Moncada & Higgs, 1995) and may be involved in modulating the interaction of circulating leukocytes with the microvasculature (Kubes et al., 1991), thus further participating in the general homeostatic control of vascular beds.

NO, formed by eNOS has also been proposed to regulate resting microvascular permeability, since administration of the nitric oxide synthase inhibitor, N^G-nitro-L-arginine methyl ester (L-NAME) led to an increase in microvascular protein efflux in the cat mesentery, an effect reversed by the nitric oxide donor, nitroprusside (Kubes & Granger, 1992). Moreover, L-NAME elevated vascular protein extravasation in a number of organs of the rat (Filep & Földes-Filep, 1993; Filep et al., 1993). In contrast to these findings, in our previous studies we could not demonstrate that either L-NAME or the other nitric oxide synthase inhibitors including N^G-monomethyl-L-arginine (L-NMMA), aminoguanidine or N^G-iminoethyl-L-ornithine could cause increased plasma leakage in the rat intestine or other major organs (László & Whittle, 1997; László et al., 1994a,1994b, 1995a1995b1995c). Acute plasma leakage was, however, provoked by these NO synthase inhibitors, when administered at a time of low-dose endotoxin challenge, which itself did not increase microvascular leakage (László & Whittle, 1997; László et al., 1994a,1994b, 1995a1995b1995c). These latter findings imply that other factors, presumably acting on the microvasculature, in addition to eNOS inhibition, are required to provoke plasma leakage and accumulation in this model.

In the original studies where NO synthase inhibitors were shown to increase resting vascular permeability, the animals had been acutely surgically manipulated to prepare an exteriorized vascular bed (Kubes & Granger, 1992). In other studies where an increase in resting vascular permeability with L-NAME was observed, the rats had catheters implanted into the abdominal aorta and into the vena cava. 4 days before the study (Filep & Földes-Filep, 1993; Filep et al., 1993). Since it is known that wound healing takes some 10 days, in the latter studies a sterile wound-healing inflammatory process would likely to be in progress. In addition, vascular endothelial cells release a number of vasoactive mediators following irritation (Burnstock, 1990) a process which also could be activated during arterial or venous catheterization. However, in our studies where L-NAME failed to affect resting leakage of albumin, the animals had not undergone any surgical manipulation, although there were also differences between the actual measurements of permeability made in these studies.

In order to evaluate the possible effects of surgical intervention on plasma exudation using a standardized technique, in the present study, the acute effects of the NO synthase inhibitors, L-NAME and L-NMMA on microvascular leakage and accumulation of radiolabelled human serum albumin was evaluated in animals that had undergone various degrees of surgical manipulation. In addition, the involvement of neutrophils in any microvascular permeability changes following surgical intervention has also been investigated.

Methods

Surgical manipulation

Male Wistar rats (225–275 g) were fasted overnight, but allowed free access to water. The animals were separated into four groups. In the first group the animals were deemed to be conscious for the majority of the experimentation period, since the treatments were performed under transient halothane anaesthesia from which the animals had completely recovered within 2 min. Autopsy in this group, which was designated as the conscious surgically un-operated group, was performed under halothane anaesthesia within 1 min. In the second group, the animals received pentobarbitone (60 mg kg⁻¹, i.p.) to induce anaesthesia and were tracheotomized. In the third group, under pentobarbitone anaesthesia, a 5 cm long skin incision was performed on the abdominal region. The animals were tracheotomized and a gauze pad moistened with saline was placed on the skin incision. In the fourth group, in pentobarbitone-anaesthetized animals, a 5 cm long midline laparotomy in the abdominal wall was performed, without significant bleeding. Rats were tracheotomized and gauze moistened with saline was placed over the incision.

All groups were administered [¹²⁵I]human serum albumin ([¹²⁵I]HSA, 2 μCi kg⁻¹, i.v.) via a needle inserted into the tail vein for 3–4 s, and autopsy was performed 1 h later. The 1 h maximum time-point was chosen to exclude the involvement of inducible NO synthase (iNOS), that requires 2–3 h following challenge for expression, since this could modify vascular leakage and hence confound interpretation of the findings (Boughton-Smith et al., 1993). In one group, autopsy was performed 30 min after administration of radiolabelled albumin. In all of the anaesthetized groups, the body temperature was maintained on 36.5–37°C using a homeo-thermic control unit and underblanket (Harvard Instruments).

Plasma leakage

As a measure of vascular endothelial permeability, leakage of [¹²⁵I]HSA into tissue was determined in segments of the stomach, duodenum, jejunum, colon, lung and kidney. Blood was collected from the abdominal aorta into syringes containing trisodium citrate (final concentration 0.318%) and centrifuged (10,000×g, 10 min, 4°C). The [¹²⁵I]HSA content of the plasma and segments of tissues was determined in a gamma-spectrometer (Nuclear Enterprises NE 1600) and the albumin content in tissues were calculated.

The control value for albumin accumulation was taken as the mean of the data of a group of control animals which received albumin only, which reflects basal albumin movement into tissues. In each experiment and for each procedure, this basal control mean value was calculated and subtracted from the value from each of the animals in each treatment group. The data were expressed as changes in albumin accumulation (Δ plasma leakage,  μl plasma g⁻¹ tissue), corrected for intravascular volume (László et al., 1994a, 1995c).

Intravascular volume

Changes in intravascular volume in gastric, duodenal, jejunal, colonic, pulmonary and renal tissues were determined in additional groups of rats by administering [¹²⁵I]HSA (2 μCi kg⁻¹) intravenously via the tail vein 2 min before tissue removal, in all groups investigated. The tissue and plasma con-tent of radiolabel was determined and the intravascular volume was expressed as μl g⁻¹ tissue (László et al., 1994a, 1995c).

Effect of L-NAME and L-NMMA on intestinal plasma leakage

In a set of rats from each of the groups, L-NAME (1–5 mg kg⁻¹, s.c.) or L-NMMA (1–50 mg kg⁻¹, s.c.) was injected concurrently with [¹²⁵I]HSA. Plasma leakage in the jejunum and colon was evaluated after 1 h. In separate groups, rats were pretreated with L-arginine (300 mg kg⁻¹, s.c.) 15 min before L-NAME (5 mg kg⁻¹, s.c.) administration, and jejunal and colonic plasma leakage was determined 1 h after L-NAME.

In a separate study, L-NAME (5 mg kg⁻¹, s.c.) was administered concurrently with laparotomy and [¹²⁵I]HSA, and plasma leakage in the jejunum and colon was measured 30 min later.

Plasma leakage in other organs

To evaluate the action of NO synthase inhibitors on plasma leakage in the stomach, duodenum, lung and kidney in conscious and laparotomized rats L-NAME (5 mg kg⁻¹, s.c.) was administered concurrently with [¹²⁵I]HSA, and plasma leakage in these tissues was measured 1 h later.

Effect of phenylephrine infusion on intestinal plasma leakage and blood pressure

In laparotomized rats, phenylephrine infusion (10 μg kg⁻¹ min⁻¹, into the tail vein) was commenced immediately following administration of [¹²⁵I]HSA. Plasma leakage in the jejunum and colon was determined 1 h later.

In additional groups of anaesthetized rats, and rats with abdominal laparotomy, the effect of L-NAME (5 mg kg⁻¹, s.c.) or phenylephrine infusion (10 μg kg⁻¹ min⁻¹, i.v.) on systemic arterial blood pressure was measured from a cannula inserted into the right carotid artery using a pressure transducer (Elcomatic) connected to a Grass Polygraph. The blood pressure was monitored over a 1 h period. The dose of phenylephrine was taken from previous studies in this model (László et al., 1994b).

Effect of S-nitroso-glutathione on intestinal plasma leakage and blood pressure

In laparotomized rats, infusion of the NO donor S-nitroso-glutathione (SNOG, 1 μg kg⁻¹ min⁻¹) into the tail vein was commenced concurrently with the administration of L-NAME (5 mg kg⁻¹, s.c.). Plasma leakage in the jejunum and colon was measured 1 h later.

In another group of rats with abdominal laparotomy, the effect of infusion of SNOG (1 μg kg⁻¹ min⁻¹) into the tail vein, on the actions of L-NAME (5 mg kg⁻¹, s.c.) on arterial blood pressure was monitored over a 1 h period. The dose of SNOG was taken from previous studies in this model (László et al., 1995c).

Effect of anti-neutrophil serum on plasma leakage

Rabbit anti-rat neutrophil serum (0.4 ml kg⁻¹, i.p.) Was administered 4 h before laparotomy. L-NAME (5 mg kg⁻¹, s.c.) was injected immediately after abdominal laparotomy, and plasma leakage was determined 1 h later, being thus 5 h after anti-neutrophil serum administration.

Chemicals

[¹²⁵I]human serum albumin was obtained from Amersham International (U.K.) and IZINTA, Budapest (Hungary). Rabbit anti-rat neutrophil serum was purchased from Accurate Chemical and Scientific Corporation (Westbury, New York, U.S.A.). All the other compounds were from Sigma Chemical Company (Poole, Dorset, U.K.).

Statistics

The data are expressed as means±s.e.mean from (n) rats per experimental group. For statistical comparisons, analysis of variance with the Bonferroni test was utilized, where P<0.05 was taken as significant.

Results

Effect of L-NAME or L-NMMA on intestinal plasma leakage following various surgical manipulations

In conscious un-operated rats, administration of L-NAME (5 mg kg⁻¹, s.c.) or L-NMMA (50 mg kg⁻¹, s.c.) did not affect intestinal plasma leakage over 1 h (Table 1).

Table 1.

Jejunal and colonic plasma leakage (1  h) in conscious un-operated or barbiturate-anaesthetized rats or following incision of the abdominal skin in anaesthetized rats

In pentobarbitone-anaesthetized rats with or without incision of the abdominal skin, no significant change in plasma leakage was found in the jejunum (Δ19±14 and Δ18±8 μl g⁻¹ tissue, respectively, n=8–11) or colon (Δ8±3 and Δ0±3 μl g⁻¹ tissue, respectively, n=8–11). Furthermore, neither L-NAME (5 mg kg⁻¹, s.c.) nor L-NMMA (50 mg kg⁻¹, s.c.) affected jejunal or colonic plasma leakage in pentobarbitone-anaesthetized rats with or without skin incision (Table 1).

Abdominal laparotomy alone did not affect plasma leakage in the jejunum and colon, being Δ3±6 and Δ2±3 μl g⁻¹ tissue over 1 h, respectively (n=6), as shown in Figures 1 and 2. However, in laparotomized animals, administration of L-NAME (1–5 mg kg⁻¹, s.c.) or L-NMMA (12.5–50 mg kg⁻¹, s.c.) provoked significant dose-dependent jejunal and colonic plasma leakage over 1 h. This effect was inhibited by pretreatment with L-arginine (300 mg kg⁻¹, s.c., 15 min before L-NAME), as shown in Figures 1 and 2.

Figure 1.

The vascular leakage, determined using radiolabelled albumin, following administration of N^G-nitro-L-arginine methyl ester (L-NAME, 5 mg kg⁻¹, s.c.) in the jejunum and colon of the conscious un-operated or anaesthetized rats with abdominal laparotomy over 1 h. The reversal by L-arginine (L-Arg, 300 mg kg⁻¹, s.c., 15 min before L-NAME) of the jejunal and colonic plasma leakage induced by L-NAME in the laparotomized rat is also shown. The columns indicate the leakage of plasma in Δ μl g⁻¹ tissue. Data are given as the means±s.e.mean of five rats per group; statistical significance is shown as ^*P<0.05 compared to the control conscious untreated group.

Figure 2.

Induction of vascular leakage, determined using radio-labelled albumin, in the jejunum (upper panels) and colon (lower panels) by the administration of N^G-nitro-L-arginine methyl ester (L-NAME, 1–5 mg kg⁻¹, s.c.) or N^G-monomethyl-L-arginine (L-NMMA, 12.5–50 mg kg⁻¹, s.c.) in the anaesthetized rat with abdominal laparotomy over 1 h. Plasma leakage is expressed as Δ μl g⁻¹ tissue. Data are given as the means±s.e.mean of 5–8 rats per group; statistical significance is shown as ^*P<0.05 compared to the control (Cont.) untreated laparotomized group.

In additional studies, plasma leakage in the jejunum and colon was observed 30 min after the administration of L-NAME (5 mg kg⁻¹, s.c.) with laparotomy (Δ35±8 and Δ32±5 μl g⁻¹ tissue, respectively; n=6, P<0.05). This plasma leakage was significantly (P<0.05) smaller than that observed when L-NAME (5 mg kg⁻¹, s.c.) was administered concurrently with laparotomy and plasma leakage was investigated 1 h later (Δ88±12 μl g⁻¹ tissue in the jejunum and Δ55±9 μl g⁻¹ tissue in the colon, n=8).

Effect of L-NAME on plasma leakage in other organs

Administration of L-NAME (5 mg kg⁻¹, s.c.) had no action on resting plasma leakage in the stomach (Δ5±3 μl g⁻¹ tissue, n=12), duodenum (Δ0±11 μl g⁻¹ tissue, n=11), lung (Δ21±9 μl g⁻¹ tissue, n=5) and kidney (Δ3±3 μl g⁻¹ tissue, n=5) in conscious animals over 1 h (Figure 3).

Figure 3.

Vascular leakage, determined using radiolabelled albumin, induced by N^G-nitro-L-arginine methyl ester (L-NAME, 5 mg kg⁻¹, s.c.) in the stomach (a), duodenum (b), lung (c) and kidney (d) of the anaesthetized rat with abdominal laparotomy over 1 h. Plasma leakage is expressed as Δ μl g⁻¹ tissue. Data are shown as the means±s.e.mean of five rats per group; statistical significance is shown as ^*P<0.05 compared to the control untreated conscious group.

Abdominal laparotomy alone did not provoke significant plasma leakage in the stomach, duodenum, lung and kidney over 1 h, being Δ0±4, Δ0±8, Δ2±11 and Δ8±3 μl g⁻¹ tissue, respectively (n=5). However, administration of L-NAME (5 mg kg⁻¹, s.c.) following laparotomy caused significant plasma leakage in gastric, duodenal, pulmonary and renal tissues after 1 h (Figure 3).

Effect of phenylephrine on blood pressure and intestinal plasma leakage

In laparotomized rats, infusion of phenylephrine (10 μg kg⁻¹ min⁻¹, i.v.) elevated systemic arterial blood pressure over the 1 h period of administration (Figure 4). The increase in blood pressure by phenylephrine infusion was not significantly different from that caused by L-NAME (5 mg kg⁻¹, s.c.) as shown in Figure 4.

Figure 4.

Increase of systemic arterial blood pressure following the administration of N^G-nitro-L-arginine methyl ester (L-NAME, 5 mg kg⁻¹, s.c.) or phenylephrine (10 μg kg⁻¹ min⁻¹, i.v.) over 1 h in the anaesthetized rat with abdominal laparotomy (upper panel). Actions of L-NAME (5 mg kg⁻¹, s.c.) or phenylephrine (10 μg kg⁻¹ min⁻¹, i.v.) on plasma leakage in the jejunum after 1 h in laparotomized rats are shown in the lower panel. Blood pressure changes are given as Δ mmHg compared to the untreated laparotomized group and plasma leakage is expressed as Δ μl g⁻¹ tissue. Data are shown as the means±s.e.mean of four rats per group; statistical significance is shown as #P<0.05 for difference in blood pressure difference compared to the control abdominal laparotomized group, and as ^*P<0.05 for difference in plasma leakage compared to the control laparotomized group.

Infusion of phenylephrine (10 μg kg⁻¹ min⁻¹, i.v.) did not provoke plasma leakage in the jejunum and colon over 1 h in laparotomized rats, being Δ5±11 and Δ3±4 μl g⁻¹ tissue, respectively (n=4) as demonstrated in Figure 4.

Effect of S-nitroso-glutathione on intestinal plasma leakage and blood pressure

Intravenous infusion of the NO donor, S-nitroso-glutathione (SNOG, 1 μg kg⁻¹ min⁻¹ for 1 h) did not reduce blood pressure in laparotomized rats (Δ2±4, Δ0±3, Δ-3±5 and Δ4±5 mmHg 15, 30, 45 and 60 min later, respectively; n=4). Furthermore, SNOG did not affect the elevation and blood pressure induced by L-NAME (5 mg kg⁻¹, s.c.) over 1 h (Figure 5).

Figure 5.

Changes in systemic arterial blood pressure over 1 h in the anaesthetized rat with abdominal laparotomy by the concurrent infusion of S-nitroso-glutathione (SNOG, 1 μg kg⁻¹ min⁻¹, i.v.) with N^G-nitro-L-arginine methyl ester (L-NAME, 5 mg kg⁻¹, s.c.; upper panel). Actions of L-NAME (5 mg kg⁻¹, s.c.) with or without the concurrent intravenous infusion of SNOG (1 μg kg⁻¹ min⁻¹) on plasma leakage in the jejunum after 1 h in laparotomized rats are shown in the lower panel. Blood pressure changes are shown as Δ mmHg compared to the untreated laparotomized group. Plasma leakage is expressed as Δ μl g⁻¹ tissue. Data are shown as the means±s.e.mean of four rats per group; statistical significance is shown as #P<0.05 for difference in blood pressure compared to the control laparotomized group, and as ^*P<0.05 for difference in plasma leakage compared to the control laparotomized group.

Administration of this dose of SNOG abolished the increase in jejunal and colonic plasma leakage induced by L-NAME (5 mg kg⁻¹, s.c.) in laparotomized rats, as shown in Figure 5.

Effect of anti-neutrophil serum on plasma leakage

Pretreatment of rats with a rabbit anti-rat neutrophil serum (0.4 ml kg⁻¹, i.p.) 4 h before laparotomy, substantially reduced L-NAME (5 mg kg⁻¹, s.c.)-provoked plasma leakage in the stomach, duodenum, jejunum, colon and lung (Figure 6). This dose of anti-neutrophil serum reduced the circulating neutrophil count by 85±5%, as determined on blood smears.

Figure 6.

Inhibition of vascular leakage, determined using radio-labelled albumin, induced by N^G-nitro-L-arginine methyl ester (L-NAME, 5 mg kg⁻¹, s.c.) in the stomach (a), duodenum (b), jejunum (c), colon (d), lung (e) and kidney (f), of the anaesthetized rat with abdominal laparotomy over 1 h by the pretreatment with rabbit anti-rat neutrophil serum (ANS. 0.4 ml kg⁻¹, i.p., 4 h before laparotomy). Plasma leakage is expressed as Δ μl g⁻¹ tissue. Data are shown as the means±s.e.mean of four rats per group; statistical significance is shown as ^*P<0.05 compared to the control (Cont.) laparotomized group and #P<0.05 compared to the L-NAME-treated laparotomized group.

Intravascular volume

In control conscious un-operated rats, the intravascular volume was 71±8 μl g⁻¹ tissue in the jejunum and 39±4 μl g⁻¹ tissue on the colon (n=6). This intravascular volume did not change significantly in the jejunum and colon of anaesthetized rats, and in animals following skin incision or abdominal laparotomy. Neither L-NAME (5 mg kg⁻¹, s.c.) nor L-NMMA (50 mg kg⁻¹, s.c.) affected intravascular volume in any of the experimental groups and tissues investigated (n=3–4, data not shown). Infusion of phenylephrine (10 μg kg⁻¹ min⁻¹, i.v.) did, however, significantly reduce intravascular volume, as measured in the jejunum and colon (by 44±1 and 46±4%, respectively, n=4, P<0.01).

Discussion

In the current study, no significant difference in albumin leakage and accumulation has been observed in tissues from un-operated conscious or anaesthetized rats, or in anaesthetized rats with a skin-incision or abdominal laparotomy under resting conditions, showing that these minor surgical interventions alone do not provoke changes in microvascular permeability to albumin. In contrast, abdominal laparotomy produced a time-dependent significant elevation in microvascular leakage in the jejunum and colon over 1 h following administration of the NO synthase inhibitors, L-NAME or L-NMMA. These NO synthase inhibitors had no such effect in conscious, anaesthetized or skin-incised rats. The increase in vascular leakage provoked by abdominal laparotomy following treatment with L-NAME was reversed by L-arginine, indicating that these effects involved the NO-L-arginine pathway. Involvement of constitutive NO synthase is more likely than the inducible isoform, since in previous studies it has been demonstrated that a minimum of 2 h is needed for the expression of iNOS (Salter et al., 1991; Boughton-Smith et al., 1993).

Inhibition of eNOS elevates blood pressure (Moncada et al., 1991; Moncada & Higgs, 1995). It is therefore possible that the increased vascular leakage by L-NAME and L-NMMA in the laparotomized rat in our study could in part reflect the enhanced perfusion pressure leading to increased blood flow (Granger et al., 1989). Administration of phenylephrine, which caused a similar elevation in blood pressure as L-NAME, did not provoke vascular leakage in the laparotomized rat, although any local actions on the microcirculation exerted by phenylephrine but not L-NAME, could obscure interpretation of these findings. In previous studies, L-NAME has been reported to enhance transcapillary protein flux without increasing capillary hydrostatic pressure in the surgically manipulated cat (Kubes & Granger, 1992). It is also unlikely that the accumulation of albumin provoked by L-NAME following laparotomy in our model could reflect an impairment of lymphatic drainage, since in previous studies, administration of L-NAME following surgical manipulation was shown to lead to 5 fold increase in lymphatic flow (Kubes & Granger, 1992). Thus, since all of the experimental groups in the present study received the same low dose of radiolabelled albumin, and the tissues subsequently processed in an identical fashion, the most likely explanation for the current findings is that the enhanced tissue albumin accumulation is a consequence of endothelial dysfunction due to inhibition of eNOS by L-NAME or L-NMMA following laparotomy.

Infusion of the NO donor, SNOG in a dose that had no effect on the elevation in blood pressure caused by L-NAME substantially reduced the vascular leakage induced by L-NAME in the laparotomized rat. This observation is in agreement with previous findings by others that the NO donor nitroprusside reversed the elevation of vascular permeability provoked by L-NAME in a surgical preparation, an effect brought about without actions on capillary pressure (Kubes & Granger, 1992). NO donors also significantly attenuate intestinal microvascular injury following the administration of high doses of endotoxin (Boughton-Smith et al., 1990), infusion of platelet-activating factor (Boughton-Smith et al., 1992), or by the concurrent administration of L-NAME and low doses of endotoxin (László et al., 1995c).

NO donors decrease neutrophil function and adhesion both in in vivo and in vitro (Moilanen et al., 1993; Ma et al., 1993; Granger & Kubes, 1994). Administration of L-NAME enhances the adhesion of leukocytes to the vascular endothelium, assessed by in vivo microscopy in surgically prepared animals (Kubes et al., 1991; Arndt et al., 1993). Such events involve platelet-activating factor and leukotrienes (Arndt et al., 1993), as also found with the intestinal microvascular leakage provoked by L-NAME in endotoxin-challenged rats (László et al., 1994b; László & Whittle, 1995). Polymorphonuclear leukocytes are well-known to play a crucial role in the changes in microvascular permeability during inflammatory processes leading to tissue oedema (Wedmore & Williams, 1981). In the present study, the increase in plasma leakage by L-NAME in laparotomized rats was abolished by the pretreatment of a rabbit anti-rat neutrophil serum suggesting the involvement of neutrophils in these events. Thus, it is feasible that endogenous NO formed by eNOS could counteract the actions of neutrophil-derived mediators that are released subsequent to laparotomy.

On the basis of the present results, it is proposed that the relatively minor surgical intervention of opening the abdominal cavity during laparotomy causes wide-spread microvascular events throughout the body. These microvascular changes are modulated by NO, since following inhibition of constitutive NOS, laparotomy causes increased plasma loss from gastrointestinal organs, lung and kidney. The mechanisms underlying this apparent microvascular priming are unknown, but may reflect neuronal or humoral stimulation as a consequence of the surgical stress of laparotomy, as well as the activation or priming of the neutrophil. Although vascular endothelial dysfunction during and following surgical operations has been described earlier (Jarnum, 1961; Krakelund, 1971; Robarts, 1979; Akerström & Lisander, 1991), this was not observed with our acute abdominal laparotomy procedures, where no manipulation of the intestinal organs was performed. However, the current data suggests that should NO formation be compromized during surgical procedures, plasma and fluid loss would be augmented which could lead to hypovolaemia, as well as oedema formation and a decreased tissue oxygenization. It is possible therefore that administration of NO donors during major surgical interventions, in low doses which do not affect systemic arterial blood pressure, may be beneficial in preventing any vascular endothelial dysfunction.

The current studies thus may help to resolve the earlier apparent conflict that NO synthase inhibitors increase vascular permeability in studies with surgically prepared models (Kubes & Granger, 1992) whereas they have no effect in intact animals (László et al., 1994a,1994b, 1995a1995b1995c), although there are also differences between the experimental models utilized. However, such effects observed in the present study could therefore confound studies on the microvascular actions of NO synthase inhibitors using acute surgically manipulated ex vivo models. Moreover, since these vascular events appear to involve neutrophils, the action of NO synthase inhibitors on neutrophil activation and adhesion in surgically prepared vascular beds may require careful interpretation. The findings thus suggest that constitutively formed NO plays a more-crucial role in the maintenance of microvascular integrity during or following minor abdominal surgical intervention, than under more physiological conditions.

Abbreviations

eNOS

endothelial nitric oxide synthase

[¹²⁵I]HSA

[¹²⁵I]human serum albumin

iNOS

inducible nitrix oxide synthase

L-NAME

N^G-nitro-L-arginine methyl ester

L-NMMA

N^G-monomethyl-L-arginine

NO

nitric oxide

SNOG

S-nitroso-glutathione