Many studies have indicated that nitric oxide (NO) has a complex role when it is produced by inducible NO synthase (iNOS) in sepsis; namely, it has a beneficial effect in defending against infectious pathogens on the one hand and a detrimental effect on vascular responsiveness and disruption of endothelial permeability and integrity on the other (1,2).
In studies in an isolated perfused rat lung model, injection of bradykinin (BK) into the pulmonary artery caused concentration-dependent vasoconstriction in lipopolysaccharide (LPS)-treated rats but not in control animals (3,4). The selective iNOS inhibitor l-N 6-(1-iminoethyl)-lysine (l-NIL) attenuated BK-induced vasoconstriction, suggesting that selective iNOS inhibition has a protective role in the pulmonary circulation during sepsis due to LPS (3).
One of the most established sepsis models with clinical relevance is the cecal ligation and perforation procedure (CLP) followed by fluid resuscitation and analgesia. With this model, the time course of sepsis and the diversity of the bacterial flora mimic the clinical situation, in contrast to simple LPS infusion, which is only one of several triggers capable of initiating pathophysiological changes (5). Wichtermann et al. (5) cautioned that differences in variables between sepsis and endotoxemia (due to LPS) make it difficult to correlate between an endotoxic study and a septic model. The early hyperdynamic state of septic human patients corresponds well with the hemodynamic situation seen in rats after initiation of CLP (6,7). In contrast, LPS models often generate a situation with a profound hypodynamic state (8).
Because the LPS and the CLP septic models are characterized by differences in both the mediator and hemodynamic responses to the septic insult, we tested the hypothesis that they express different effects with respect to the induction of pulmonary endothelial dysfunction. We examined BK-induced vasoconstriction after the onset of sepsis induced by LPS or CLP at different time points. l-NIL, a selective iNOS inhibitor, was given concomitantly with either LPS or CLP to elucidate the influence of iNOS on sepsis-induced pulmonary injury. Additionally, lung homogenates were evaluated for iNOS protein expression.
This study was approved by our animal research committee. Rats were allowed food and water ad libitum. Rats made septic by intraperitoneal (ip) application of LPS were compared with rats treated with the CLP procedure, including appropriate control/sham groups. Pulmonary vascular reactivity was tested in CLP lungs at two times that represent an early and late septic state in this model. To study the role of iNOS in sepsis-induced pulmonary endothelial dysfunction, the selective iNOS inhibitor l-NIL was given concomitantly with the induction of sepsis in both models. Again, appropriate control and sham groups were also studied.
In detail, male Sprague-Dawley rats (250–300 g) were randomly divided into 9 groups (n = 6–8 per group). Groups 1–4 were given 1) saline (control), 2) l-NIL (3 mg/kg), 3) LPS (Salmonella typhimurium; 15 mg/kg), or 4) LPS plus l-NIL (3 mg/kg). All of these drugs were administered ip 6 h before lung isolation. Groups 5–9 received the following: 5) sham laparotomy at 6 h (only laparotomy; no CLP), 6) sham 24 h before lung isolation, 7) CLP at 6 h, 8) CLP at 24 h, and 9) CLP at 6 h plus l-NIL (3 mg/kg) before lung isolation. Because the 24-h mortality after ip application of 15 mg/kg LPS was extremely high in pilot studies, preventing any meaningful comparison with CLP-treated rats, we did not investigate a 24-h LPS group in this study (data not shown). Concentrations of drugs used in this study are described in a previously published article (3).
Rats were used after a 1-week acclimatization period in our laboratory. Before the start of the experiments, all animals were randomly assigned to the different experimental groups. Groups 1–4 received saline or LPS with or without treatment (see above) 6 h before lung isolation. Sham and CLP rats were instrumented under isoflurane anesthesia with catheters advanced into the superior vena cava via the right external jugular vein and the left carotid artery. Sham laparotomy or laparotomy and CLP was performed according to a previously standardized technique described by Farquhar et al. (9) to create sepsis. A mid-line laparotomy incision was made and the cecum was ligated just distal to the ileocecal valve so that bowel continuity was preserved. The cecum was then punctured twice with an 18-gauge needle and the bowel was returned to the peritoneal cavity. The incision was closed in two layers. The catheters were tunneled subcutaneously, exteriorized at the interscapular region, and guided into a swivel device to allow free movement. After surgery, rats were allowed to recover for 24 h and were provided with rodent food and water ad libitum. Fluid resuscitation with 0.9% saline (2 mL/100 g IV) was started after surgery. The carotid line was continuously flushed with heparin solution (45 IU/h) to maintain patency, and fentanyl (2 μg/100 g IV) was provided to ensure adequate analgesia. Animals were excluded if technical failure (e.g., damage or blocking of arterial or venous catheters) occurred before the completion of sepsis induction.
After 24 h of recovery, rat lungs were prepared by a second investigator. Anesthesia was applied with α-chloralose (50 mg/kg) and urethane (650 mg/kg) ip. A 17-gauge cannula was inserted into the trachea via a tracheostomy, and ventilation was started with warmed (35°C) and humidified 21% oxygen and 5% CO2 (balanced nitrogen) by using a rodent ventilator (Harvard Apparatus GmbH, Germany; tidal volume, 1 mL/100 g; frequency, 60 breaths/min). End-expiratory pressure was set at 1 mm Hg. Sternotomy was followed by excision of sections of the right and left anterior chest wall to expose the heart and lungs. After heparinization (100 U), rats were partially exsanguinated by needle aspiration (5–6 mL). A steel cannula (13-gauge) connected to the perfusion system was inserted through the pulmonic valve into the main pulmonary artery via an incision in the right ventricle. A suture tied around the pulmonary artery and aorta secured the cannula. The latter prevented systemic blood flow. The circle was closed by a cannula (3.5-mm outside diameter) that was inserted through the apex of the left ventricle and secured with umbilical tape.
Perfusate was a mixture out of the rat’s own blood, which was added to the perfusate after lung isolation and physiologic salt solution (hematocrit of 9%–12%). Indomethacin (30 μg/mL perfusate) was added to block prostaglandin synthesis. Perfusate was drained from the left ventricle to a glass reservoir and was heated to 38°C by a circumferential water jacket. A peristaltic pump (Harvard Apparatus GmbH) returned the perfusate to the pulmonary artery at constant flow (16 mL/min). A warmed and humidified chamber was placed over the thoracic cavity with the lung preparation inside to maintain thoracic temperature at 37°C; temperature was controlled by a thermometer and a heating lamp above the chamber. Reservoir pH was continuously monitored (Hanna-Instruments, Germany) and maintained at 7.35–7.45 by the addition of HCl or NaOH as required.
A pressure transducer (Becton Dickinson, Singapore) continuously monitored the pulmonary artery pressure (Pa). The mean pulmonary venous pressure was set at 2 mm Hg by adjusting the height of the reservoir and was held constant. The inspired oxygen concentration was monitored (Oxydig; Dräger, Germany) near the tracheal tube.
After tracheostomy and thoracotomy, blood samples were drawn from the right ventricle to assess hemoglobin, lactate (ABL 620; Radiometer, Copenhagen, Denmark) and leukocytes (Coulter MD II-8; Coulter-Beckmann, Krefeld, Germany). Next, lungs were isolated, and the perfusion rate was stepped up to 16 mL/ min and then remained constant. After stabilization, the potent vasoconstrictor angiotensin II (AngII) (0.1μg) was injected into the inflow tract. Ten minutes after Pa returned to baseline, pulmonary vascular responses were evaluated with increasing concentrations of BK (1, 3, and 6 μg), also injected into the inflow circuit. Each dose was administered 5 min after the perfusion pressure returned to baseline levels. Figure 1 summarizes the protocol for both sepsis models.
After experiments, lungs were immediately frozen and stored at −70°C until use. Briefly, equal amounts of tissue homogenate from all experimental groups were loaded in parallel onto a 7.5% sodium dodecyl sulfate trisglycine polyacrylamide gel and separated by electrophoresis (1.5 h; 120 V; MiniGell 2; Bio-Rad, Munich, Germany). Ten nanograms of authentic mouse macrophage iNOS (stimulated with interferon-γ and LPS; BD Transduction Laboratories, Germany) was always applied to one lane as a positive control. Protein was transferred to a nitrocellulose membrane overnight (300 mA; 0°C–5°C). Membranes were blocked with nonfat dry skim milk (5% in phosphate-buffered solution) blocking solution and incubated with iNOS antibodies (BD Transduction Laboratories; mouse primary monoclonal antibodies; 1:500). Membranes were subsequently incubated with a horseradish peroxidase-conjugated anti-mouse secondary antibody (1:3000; 1 h; 37°C; Amersham Life-science, Freiburg, Germany). Blots were developed by enhanced chemiluminescence (coumaric acid [0.4 nM; pH 8.5] and luminol 2.5 mM in tris-(hydroxymethyl) aminomethane [100 mM; pH 8.5]; 1-min exposure to the blot) and exposure of the blot to radiograph film (Hyperfilm ECL; Amersham Lifescience). The images were scanned and then analyzed with densitometry software (NIH ImageJ V1.61.1 image analysis).
AngII, BK, and LPS from S. typhimurium were freshly prepared for each experiment by dilution in NaCl 0.9% (Sigma Chemicals, Germany). l-NIL was prepared by dilution in NaCl 0.9% and stored at − 20°C (Cayman Chemicals, MI). Pilot studies showed that none of these vehicles had an effect on Pa.
Vasoconstriction from BK was expressed as the peak Pa − baseline Pa. Data are presented as mean ± sem. SigmaStat 2.0 was used for statistical analysis. The various treatments were compared by using one-way analysis of variance followed by the Student-Newman-Keuls method for pairwise multiple comparison procedures. When the data were not normally distributed, they were compared by using Kruskal-Wallis analysis of variance. P < 0.05 was considered significant.
All rats survived 6 and 24 h of sepsis due to LPS and CLP and underwent successful lung isolation. Independent of treatment with l-NIL, all animals treated with LPS and CLP demonstrated reduced activity, piloerection, and exudation around the eyes and nose, all of which appeared to be increased at CLP 24 h.
Baseline Pa after lung isolation was not significantly different among the control group (13.9 ± 0.7 mm Hg), LPS (14.1 ± 0.6 mm Hg), sham 6 h (13.8 ± 0.3 mm Hg), sham 24 h (15.4 ± 1.1 mm Hg), CLP 6 h (14.5 ± 0.9 mm Hg), and CLP 24 h (14.9 ± 1.4 mm Hg). Treatment with l-NIL did not alter Pa in the control group (14.6 ± 0.4 mm Hg), the LPS group (17.7 ± 1.2 mm Hg), or the CLP 6 h group (17.8 ± 1.4 mm Hg).
After 24 h of CLP sepsis, arterial hemoglobin concentration was decreased compared with the control and sham 24 h groups (Table 1). Only LPS, LPS plus l-NIL, and CLP 24 h (but not CLP 6 h) resulted in a significant increase in lactate concentration and decrease in leukocyte count compared with the control and sham 24 h groups (Table 1). Treatment with l-NIL in control animals did not alter these variables compared with controls (Table 1).
Baseline and preoperative (6 and 24 h after laparotomy, respectively) mean arterial blood pressures (MAP) of CLP and sham animals are shown in Figure 2. Baseline MAP was not different among groups. However, preoperative MAP in the CLP 24 h group decreased significantly compared with baseline. Pre-operative MAP was significantly increased in the CLP 6 h group compared with preoperative sham 6 h and CLP 24 h.
Vasoconstriction due to angiotensin II was not different between the control/sham groups and septic groups (Fig. 3). There was also no time-dependent difference in vasoconstriction in either the sham or the CLP groups (Fig. 3).
BK induced vasoconstriction in the pulmonary circulation of LPS-treated rats in a concentration-dependent manner (Fig. 4). Concurrent administration of l-NIL attenuated the vasoconstriction to control levels (Fig. 4).
In the CLP 6 h group, BK also caused pulmonary vasoconstriction (Fig. 4). Similar to the effects seen in LPS-treated rats, l-NIL reduced this vasoconstriction (Fig. 4). In contrast, the pulmonary circulation of CLP 24 h animals did not respond to BK and was not different from that of control animals (Fig. 4). Control plus l-NIL, sham 6 h, and sham 24 h animals demonstrated pulmonary vasoconstriction induced by BK similar to that of control animals (Fig. 4).
Injection of LPS and CLP 6 h elicited a significant increase of lung iNOS protein expression compared with control and sham rats, as demonstrated densitometrically and in a representative Western blot (Fig. 5). In contrast, iNOS protein was significantly attenuated 24 h after CLP (Fig. 5). The administration of l-NIL did not influence expression of iNOS protein in septic animals (data not shown).
The main findings of this study are as follows. First, sepsis induced by the CLP technique showed a time-dependent pulmonary vasoconstriction induced by BK in an isolated perfused rat lung model. The concentration-dependent responses to BK in LPS rats were comparable only to those of CLP after 6 hours but not after 24 hours. Second, this vasoconstriction was attenuated by concurrent treatment with the selective inhibitor of iNOS, l-NIL, in both the LPS and CLP six hour models. Third, the functional integrity of pulmonary smooth muscles was not different in LPS-and CLP-treated rats compared with control and sham animals, as was demonstrated by pulmonary vasoconstriction due to AngII. Fourth, expression of iNOS protein was confirmed in the LPS and CLP 6 hour groups but not in the CLP 24 hour group.
A variety of experimental models have been used to study the mechanisms involved in the pathophysiology of sepsis. A commonly used model is endotoxemia induced by the administration of LPS. LPS models, however, are often criticized for their lack of a complete simulation of the clinical situation (10). Another well established technique to create sepsis is the CLP model, in which fecal peritonitis is used to produce sepsislike conditions. This model is believed to more closely mimic the clinical situation (10). However, there are also interpretation problems associated with CLP-induced sepsis. For example, variations in the extent of fecal contamination of the abdominal cavity or the time of assessing end-points after the induction of CLP have led to disparate observations (11). Variations in the extent of fecal contamination in CLP models demonstrate, in our opinion, a strength and not a weakness of this model, because in the clinical setting, sepsis is also characterized by marked differences both in the severity and the time course of the disease.
After the CLP procedure, rats became hypotensive after 24 hours, and lactate concentrations increased over time from 6 to 24 hours. In addition, CLP animals were leukopenic at 24 hours compared with sham-operated animals. Interestingly, LPS animals were comparable to the CLP 24 hour group in this regard, whereas the CLP 6 hour group was not different from sham animals. These hemodynamic and biochemical changes are consistent with previous investigations using the CLP model (12,13) and are comparable to a profile characteristic of clinical sepsis (14). In contrast to our results, Scott et al. (15) reported a smaller count of leukocytes six hours after CLP with the same model and animal breed. However, the leukocyte count and lactate concentration of their study were comparable to ours 24 hours after the induction of CLP (15).
Results from the established and potent receptor-mediated vasoconstrictor AngII demonstrated that vasoreactivity of pulmonary smooth muscles was not altered during CLP- and LPS-mediated sepsis, thus confirming the functional integrity of pulmonary smooth muscles during septic injury. In accordance with our data, Schneidkraut and Carlson (16) showed in isolated perfused lungs that vasoconstriction by AngII was not decreased in rats four hours after CLP. Li et al. (17), however, demonstrated that 16 hours after CLP, pulmonary vasoconstriction due to KCl and angiotensin was depressed in lungs compared with sham groups. One explanation for these divergent reactivities seen in septic rats besides the different sepsis models could be the different times of measurement after initiation of sepsis. Pulmonary smooth muscles might be influenced to a different degree over the observed time periods. In addition, the type of vasoconstrictor (e.g., AngII, KCl, or phenylephrine) applied to the pulmonary circulation seems to be important. McCormack et al. (18) demonstrated attenuated vascular contractility to phenylephrine in rat pulmonary arterioles 24 hours after CLP. However, both AngII and phenylephrine are receptor-mediated vasoconstrictors; therefore, the reason for the observed discrepancy remains unclear.
Concentration-dependent vasoconstriction due to BK and attenuation by concomitant treatment with l-NIL in LPS animals was in agreement with previously published results (3,4). Lee et al. (19) already demonstrated that NO production can exert toxic effects on the pulmonary endothelial layer during sepsis. In addition, many investigators have pointed out that activation of iNOS in the lung is the major cause of microvascular injury associated with sepsis (20–22).
In our study, we focused on the vasoconstrictive properties of BK, although BK receptor subtypes (BK1 or BK2) are known to be responsible for both vasodilation and vasoconstriction. Vasoconstriction of pulmonary vessels induced by BK might be caused by direct receptor stimulation at the level of the vascular smooth muscle, suggesting endothelial dysfunction. We already demonstrated that BK2 receptors are involved in mediating BK-induced vasoconstrictions in LPS rats (4).
In this study, the important role of iNOS for pulmonary endothelial dysfunction was confirmed for both the LPS and the CLP sepsis models, because highly increased iNOS protein expression was detectable in both LPS six hour and CLP six hour rats (Fig. 5). This endothelial dysfunction was diminished by concomitant administration of l-NIL, a selective iNOS inhibitor. The fact that iNOS protein expression was significantly reduced 24 hours after CLP underlines the important role of iNOS in this respect (Fig. 5).
Determination of iNOS protein by the Western Blot technique referred to alterations in protein expression of the whole lung in this study. With our experimental setup, we could not determine the exact location of iNOS production. However, we think that iNOS must be induced in close proximity to the pulmonary endothelial layer, because the produced NO is rapidly inactivated by the hemoglobin in the perfusate. This is supported by the studies of Carraway et al. (20) and Kobzik et al. (21), who showed strong iNOS expression after CLP and LPS, respectively, in alveolar and bronchial epithelium, capillary endothelium, macrophages, and Clara cells. In this study, the expression of iNOS protein was confirmed in CLP rats at 6 hours, but not at 24 hours. In accordance with our results, Scott et al. (15) demonstrated significantly increased iNOS expression six hours after the induction of sepsis by the CLP procedure in lungs of Sprague-Dawley rats. The iNOS expression was markedly decreased at CLP 24 hours compared with CLP 6 hours (15). In contrast to our results, Okamoto et al. (23) observed iNOS bands from homogenates of CLP rat lungs from 6 to 42 hours. The density of bands tended not to decrease before 36 and 42 hours, respectively (23). They used a rabbit immunoglobulin G antibody against iNOS purified from RAW 264.7 cells (mouse leukemic monocyte cell line), whereas our monoclonal antibody was of a immunoglobulin G1 isotype and generated from mouse iNOS. In addition, they used a different strain of rats, namely, male Wistar King rats. These differences could explain the contradictory Western blotting results.
Concomitant administration of l-NIL did not change iNOS expression in our septic animals (data not shown). This is not surprising, because l-NIL is thought to be competitive versus the l-arginine substrate, with consecutively less production of NO (24). In contrast to dexamethasone (25), no inhibition of iNOS expression by l-NIL is known (26).
In summary, rats subjected to CLP sepsis clearly expressed BK-induced vasoconstriction in an isolated rat lung model. This endothelial injury was most pronounced early after the onset of sepsis. The iNOS appeared to play an important role in the pathogenesis of pulmonary endothelial dysfunction in both sepsis models; this was demonstrated by our approach to selectively inhibit iNOS and by the differences in pulmonary iNOS protein expression.
1. Moncada S, Palmer RM, Higgs EA. Nitric oxide: physiology, pathophysiology, and pharmacology. Pharmacol Rev 1991;43:109–42.
2. Bone RC. The pathogenesis of sepsis. Ann Intern Med 1991;115:457–69.
3. Fischer LG, Horstman DJ, Hahnenkamp K, et al. Selective iNOS-inhibition attenuates acetylcholine and bradykinin-induced vasoconstriction in lipopolysaccharide-exposed rat lungs. Anesthesiology 1999;91:1724–32.
4. Fischer LG, Horstman DJ, Hollmann MW, et al. Cyclooxygenase inhibitors attenuate bradykinin-induced vasoconstriction in septic isolated rat lungs. Anesth Analg 2000;90:625–31.
5. Wichtermann KA, Baue AE, Chaundry IH. Sepsis and septic shock: a review of laboratory models and a proposal. J Surg Res 1980;29:189–201.
6. Wang P, Yoo P, Zhou M, et al. Reduction in vascular responsiveness to adrenomedullin during sepsis. J Surg Res 1999;85:59–65.
7. Sielenkamper AW, Chin-Yee IH, Martin CM, et al. Diaspirin crosslinked hemoglobin improves systemic oxygen uptake in oxygen supply-dependent septic rats. Am J Respir Crit Care Med 1997;156:1066–72.
8. Fink MP, Heard SO. Laboratory models of sepsis and septic shock. J Surg Res 1990;49:186–96.
9. Farquhar I, Martin CM, Lam C, et al. Decreased capillary density in vivo in bowel mucosa of rats with normotensive sepsis. J Surg Res 1996;61:190–6.
10. Deitch EA. Animal models of sepsis and shock: a review and lessons learned. Shock 1998;9:1–11.
11. Lush CW, Kvietys PR. Microvascular dysfunction in sepsis. Microcirculation 2000;7:83–101.
12. Arkovitz MS, Wispe JR, Garcia VF, et al. Selective inhibition of the inducible isoform of nitric oxide synthase prevents pulmonary transvascular flux during acute endotoxemia. J Pediatr Surg 1996;31:1009–15.
13. Kengatharan KM, De Kimpe SJ, Thiemermann C. Role of nitric oxide in the circulatory failure and organ injury in a rodent model of gram-positive shock. Br J Pharmacol 1996;119:1411–21.
14. Bone RC, Balk RA, Cerra FB, et al. Definitions for sepsis and organ failure and guidelines for the use of innovative therapies in sepsis: the ACCP/SCCM Consensus Conference Committee—American College of Chest Physicians/Society of Critical Care Medicine. Chest 1992;101:1644–55.
15. Scott JA, Mehta S, Duggan M, et al. Functional inhibition of constitutive nitric oxide synthase in a rat model of sepsis. Am J Respir Crit Care Med 2002;165:1426–32.
16. Schneidkraut MJ, Carlson RW. Bacterial sepsis-induced decrease in lung vascular reactivity to 9,11-dideoxy-11a9a-epoxymethano-prostaglandin F2-α (U46619) in the rat. J Pharmacol Exp Ther 1990;253:1171–6.
17. Li S, Fan SX, McKenna TM. Role of nitric oxide in sepsis-induced hyporeactivity in isolated rat lungs. Shock 1996;5:122–9.
18. McCormack DG, Mehta S, Tyml K, et al. Pulmonary microvascular changes during sepsis: evaluation using intravital video-microscopy. Microvasc Res 2000;60:131–40.
19. Lee RP, Wang D, Kao SJ, et al. The lung is the major site that produces nitric oxide to induce acute pulmonary oedema in endotoxin shock. Clin Exp Pharmacol Physiol 2001;28:315–20.
20. Carraway MS, Piantadosi CA, Jenkinson CP, et al. Differential expression of arginase and iNOS in the lung in sepsis. Exp Lung Res 1998;24:253–68.
21. Kobzik L, Bredt DS, Lowenstein CJ, et al. Nitric oxide synthase in human and rat lung: immunocytochemical and histochemical localization. Am J Respir Cell Mol Biol 1993;9:371–7.
22. Fujii Y, Goldberg P, Hussain SN. Contribution of macrophages to pulmonary nitric oxide production in septic shock. Am J Respir Crit Care Med 1998;157:1645–51.
23. Okamoto I, Abe M, Shibata K, et al. Evaluating the role of inducible nitric oxide synthase using a novel and selective inducible nitric oxide synthase inhibitor in septic lung injury produced by cecal ligation and puncture. Am J Respir Crit Care Med 2000;162:716–22.
24. Bryk R, Wolff DJ. Mechanism of inducible nitric oxide synthase inactivation by aminoguanidine and L-N6
-(1-iminoethyl)lysine. Biochemistry 1998;37:4844–52.
25. Bryant CE, Perretti M, Flower RJ. Suppression by dexamethasone of inducible nitric oxide synthase protein expression in vivo: a possible role for lipocortin 1. Biochem Pharmacol 1998;55:279–85.
26. Zhang C, Walker LM, Hinson JA, et al. Oxidant stress in rat liver after lipopolysaccharide administration: effect of inducible nitric-oxide synthase inhibition. J Pharmacol Exp Ther 2000;293:968–72.