| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1Sunnybrook and Women's College Health Sciences Centre, University of Toronto, Toronto M4N 3M5; and 2Lawson Health Research Institute and Department of Medical Biophysics, University of Western Ontario, London, Ontario, Canada N6A 5B8
Submitted 18 March 2003 ; accepted in final form 8 September 2003
 |
ABSTRACT |
|---|
 TOP ABSTRACT METHODSRESULTSDISCUSSIONREFERENCES |
|---|
] levels were doubled by 21 h (P < 0.05). To assess the effect of NO overproduction on microvascular perfusion, N6-(1-iminoethyl)-L-lysine (L-NIL) was administered to maintain 
levels at baseline. Pneumonia did cause a decrease in RBC velocity of 23% by 6 h, but by 24 h RBC velocity and supply rate had increased relative to sham by 22 and 38%, respectively (P < 0.05). L-NIL treatment demonstrated that this increase was due to NO overproduction. With pneumonia, there was no change in functional capillary density and only modest increases in cellular damage. We conclude that, in this normotensive pneumonia model of sepsis, NO overproduction was protective of microvascular perfusion in the pancreas.
sepsis syndrome; bacterial pneumonia; microcirculation
However, recent clinical studies of mechanically ventilated intensive care patients with sepsis and septic shock have presented a very different picture (25, 26). The septic shock patients demonstrated severe exocrine pancreatic dysfunction without histological evidence of necrosis or biochemical evidence of tissue injury as indicated by elevated levels of amylase or lipase (25). Although no blood flow measurements were made, it is unlikely that these patients experienced pancreatic ischemia as predicted by the experimental models. The difference between the experimental studies and this clinical report may be the experimental models employed in these studies. Endotoxin models have been criticized as not adequately mimicking the clinical situation (9). Peritonitis models, which more closely model the progression of sepsis in patients, may cause direct inflammatory injury to the pancreas and its microvasculature resulting in pancreatitis. However, in the clinical study, only a small number of patients had an abdominal focus of inflammation. Sepsis arising from a focus of injury outside of the abdominal cavity and causing a remote injury to the pancreas may have very different characteristics.
Studies of remote injury models of sepsis have shown microvascular dysfunction without extensive areas of ischemia (3, 4, 10). A focus of infection in the gut caused a significant increase in stopped flow capillaries in rat skeletal muscle (from <10% in sham to >25% in sepsis), but the loss of perfused capillaries was very heterogeneous with both stopped flow and perfused capillaries observed in the same microscopic field of view (3, 10). Inhibition of NO overproduction with aminoguanidine prevented this loss of perfused capillaries. The goals of the present study were to investigate whether remote inflammatory injury to the pancreas would cause a similar pattern of microvascular dysfunction and whether inhibition of NO overproduction would prevent the remote injury. We chose to study remote injury in the pancreas with intravital video microscopy using a nonlethal Pseudomonas pneumonia model of sepsis in the rat.
|
METHODS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Male Sprague-Dawley rats were anesthetized with halothane (1–2%) and oxygen (100%). Animal temperature (36.5–37.5°C) was maintained by heat lamp and monitored by a rectal probe (Simpson 383). The right carotid artery was cannulated [polyethylene (PE)-10/PE-50, Clay Adams] to enable monitoring of mean arterial pressure (MAP; Digi-Med pressure analyzer and Baxter pressure transducers). The left jugular vein was cannulated (Bio-Sil tubing, Silmed) for intravenous fluid resuscitation. In pneumonia rats, a 2-mm transverse incision was made in the anterior trachea, through which a 20-gauge Angiocath catheter (Becton Dickinson) was inserted; 40 µl/100 g body wt of broth of 2 x 108 colony-forming units/ml Pseudomonas aeruginosa were instilled through the catheter (prepared by the Department of Microbiology, Victoria Campus, London Health Sciences Centre). The tracheotomy was covered with a 5 x 5-mm piece of Gelfoam, and the wound was closed. The control procedure consisted of cutting the trachea and placing the Gelfoam. Dextrose (5%) and sodium chloride (0.9%), 1 ml·100 g body wt–1·h–1, was administered during surgery and in the postoperative period to prevent hypotension and support the emergence of the hyperdynamic circulatory state (29). Fentanyl was added to the infusion at a rate of 0.5 mg·100 g body wt–1·h–1 for analgesia, and 0.5 U·100 g body wt–1·h–1 heparin was added to maintain catheter patency. Catheters were tunnelled subcutaneously to the interscapular region where they exited and were connected to a swivel-harness device. Rats were maintained in cages with free access to water and standard rat chow.
Pancreas preparation. At 6 and 24 h after P. aeruginosa instillation, rats were reanesthetized by an intravenous injection of pentobarbital (with temperature monitored and maintained as before). The pancreas was exteriorized through a left lateral abdominal incision by pulling the spleen. After the animal was transferred onto the microscopic stage, the pancreas was placed at the center of the microscopic stage, kept moist with physiological saline solution, covered with a thin, transparent plastic wrap (to prevent tissue drying), and slightly compressed by a coverslip. The microvascular preparation was allowed to stabilize for 30 min. To avoid any bias introduced by the preparation technique, pancreas preparation was done by a single person throughout the study. Animals in all groups were exposed to 30% oxygen via a mask.
Intravital video microscopy of the pancreas. The pancreas was transilluminated by a 100-W xenon lamp, and its surface microcirculation was observed using a Nikon Diaphot 300 microscope at x20/0.4, with an effective magnification on the monitor of x1,740. High-contrast images were achieved with a 440-nm band-pass filter (Nikon DF440-DF30). Video images of the microcirculation were obtained with a charge-coupled device camera (Dage-MTI-CCDC72), viewed on a monitor (Panasonic WV-BM1400) via a closed-circuit video system, and recorded (Panasonic AG 7350 VCR) on S-VHS videotape for off-line analysis. Time code (Telekom Research) was placed on the audio track. After the stabilization period, 3-min video recordings were made of 10 random fields/experiment.
Analysis of capillary hemodynamics. A Silicon Graphics O2 Image Analysis workstation was used to capture and store 45 s of video recording (1,350 frames) for each field of view to be analyzed for capillary hemodynamics. Each video sequence was processed to show the location of perfused capillaries within the field of view. This "functional" image was used for locating each in-focus capillary segment to be analyzed for hemodynamic data. From the functional image, the user selected each capillary segment and the analysis software automatically processed the light intensity data for each capillary from the stored frame-by-frame video data. Red blood cell (RBC) velocity (in mm/s) was calculated using a spatial correlation technique that determines how far RBCs move from frame to frame. The number of RBCs in the capillary segment in each frame was determined from the RBC optical density [OD = log (Io/I), where Io is the plasma light intensity and I is the RBC light intensity]. The total OD in the segment in each frame was divided by the OD for a single RBC to yield the number of RBCs in that length of capillary and hence the number of RBCs/mm, i.e., RBC lineal density. The RBC supply rate (in RBCs/s) was calculated from the product of RBC velocity and RBC lineal density. Each field of view yielded from 1 to 10 capillary segments with the hemodynamic data for each expressed as mean ± SD of the 45-s sample.
Functional capillary density analysis. Functional capillary density (FCD; in cm–1), defined as the length of all erythrocyte-perfused capillaries per observation area, was assessed in each of 10 randomly selected areas in accordance to the method of Schmid-Schoenbein et al. (23). A grid system with a grid width representing 50 µm was superimposed on the video monitor. By counting the number of intersections between the grid and the capillaries (Nc), the FCD for each observation area was calculated in accordance with the following equation: FCD = 
/2 x Nc/L (cm–1), where L represents the total length of the grid system.
Fluorescence microscopy of lethal cell injury. Severely injured cells were labeled in vivo with an intra-arterial bolus of the fluorescent vital dye propidium iodine (PI; 0.05 mg/100 g body wt, Sigma-Adrich) just before the end of each experiment. Nuclei labeled with PI (excitation wavelengths, 500–545 nm; emission barrier filter, >600 nm, Omega Opticals XF 103–2) were used to identify irreversibly damaged cells (15–17) and are expressed as the number of PI-labeled nuclei per field.
Drug dose and delivery. N6-(1-iminoethyl)-L-lysine (L-NIL; Sigma-Aldrich; St. Louis, MO) was administered intravenously at 0.3 mg·kg–1·h–1 starting at 10 h to maintain plasma nitrite/nitrate () levels at ±20% baseline throughout the experiment.
Measurement of plasma levels of NO metabolites. Plasma measurements were made using an optimized NO chemiluminescent system (2). In brief, whole blood was collected isovolemically from the arterial catheter at 1, 9, 15, and 21 h and at the end of the experiment and immediately centrifuged (3,000 g for 5 min at 4°C). Plasma was stored at –20°C before analysis. species (
and 
) were reduced to NO [0.05 M V(III) in 1 M HCl at 90°C] and quantified by chemiluminescent reaction with ozone using a Sievers 270B NO analyzer. The system was calibrated against known concentrations of NO–.
Statistical analysis. All values are expressed as means ± SD unless otherwise stated. For all tests of significance, P values of < 0.05 were considered statistically significant. A 2 x 2 ANOVA model was used for statistical analysis of physiological and microcirculation data. Accordingly, the injury effect, the time effect, and the interaction between injury and time were analyzed as described previously (3). A one-way ANOVA was used to assess differences in physiological and microcirculation parameters between 24-h sham and 24-h pneumonia animals with and without L-NIL treatment. Tukey's multiple-comparison tests for all pairwise differences were performed to identify specific pairwise group mean differences (post hoc analysis). Results of the Tukey's test applied on the results of the two-way ANOVA are reported in RESULTS and in Tables 1 and 3. Results of the Tukey's test applied on the results of the one-way ANOVA are indicated in Tables 2 and 4. Statistical analysis was done using the software SigmaStat for Windows version 2.0 (Jandel Scientific; San Rafael, CA).
|
|
|
|
|
RESULTS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Physiological data. Physiological data are presented in Tables 1 and 2. MAP was well controlled in all groups, ranging from a low of 98 mmHg in the 24-h sham group to a high of 109 mmHg in the L-NIL group. Analysis of the MAP data indicated that there was a significant interaction of time and injury (2-way ANOVA; Table 1) and significant differences between the 24-h groups (1-way ANOVA; Table 2), which reflected only small differences between groups. The white blood cell (WBC) data in Table 1 demonstrated a significant injury effect with the pneumonia animals having a lower WBC count. L-NIL treatment had a significant effect on WBCs compared with the untreated pneumonia group (Table 2). Erythrocyte count increased slightly over time in the sham group and decreased slightly over time in the pneumonia group, resulting in a significant interaction (Table 1). There were no significant differences with erythrocyte count in the 24-h group. Amylase increased from 6 to 24 h in both sham and pneumonia groups (significant time effect), but there was no injury effect. Platelet, lipase, and lactate did not show any significant changes in any of the groups.
Pancreas FCD and capillary hemodynamic parameters. Microvascular data are shown in Tables 3 and 4. FCD was not significantly different between groups at either time point and was not altered by the application of L-NIL.
RBC velocity was significantly decreased in pneumonia versus sham animals at 6 h (Tukey's test, P < 0.05), but at 24 h RBC velocity was significantly increased (Tukey's test, P < 0.05) in the pneumonia versus sham group. L-NIL treatment resulted in a significant decrease in RBC velocity compared with the untreated 24-h pneumonia group. Although there was no time or injury effect for RBC lineal density (Table 3), lineal density was significantly lower in the L-NIL-treated animals compared with both sham and untreated pneumonia animals (Table 4). The significant time, injury, and interaction effects for RBC supply rate (Table 3) were due to the significant increase in supply rate in the 24-h pneumonia animals compared with the 6-h pneumonia animals and with the sham animals at both time points (Tukey's test, P < 0.05). The decrease in velocity and lineal density with L-NIL treatment resulted in a significant 62% decrease (P < 0.01) in the RBC supply rate (Table 4).
NO inhibition. Figure 1 shows the time course of serum levels in three animals after the induction of pneumonia. All the animals showed a drop in levels at 9 h. Subsequent studies in our laboratory have shown that levels increase briefly immediately after surgery and hence the values at 9 h are likely closer to the presurgery baseline. Over the next 12 h (21-h time point), levels increased significantly (P < 0.05) by at least twofold in all three animals. The L-NIL dosing protocol (0.3 mg·kg–1·h–1 beginning at 10 h) inhibited the overproduction of (no significant change in the levels at the end of the experiment compared with baseline).
|
PI staining. Fluorescence microscopy after an intra-arterial bolus of PI did show a time and injury effect on the number of staining cells. Septic animals at 24 h after the induction of pneumonia (Fig. 2) had an increased number of staining cells. L-NIL treatment did not effect the number of PI-stained cells compared with 24-h pneumonia animals without L-NIL.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Pseudomonas pneumonia model of sepsis. A nonlethal Pseudomonas pneumonia model of normotensive sepsis in the rat was chosen as the focus of inflammation for the remote injury to the pancreas. This model has been demonstrated to be hyperdynamic, to be associated with gut mucosal and microvascular injury, and to lead to the development of gut barrier dysfunction with bacterial translocation as early as 24 h (29). We observed that rats with pneumonia developed severe dyspnoea, tachypnoea, piloerection, and a significant drop in WBCs by 24 h. There was a significant increase in plasma levels to twice baseline, which is another indicator of systemic inflammation (3). Although there was a statistically significant difference in MAP between the 24-h sham and 24-h pneumonia animals, it was not physiologically relevant, indicating that our fluid resuscitation protocol maintained MAP. All of these signs demonstrate that our pneumonia animals developed normotensive sepsis by 24 h.
Choice of L-NIL. L-NIL was chosen as a selective inhibitor of inducible NO synthase (iNOS), which is responsible for the increased production of NO in sepsis (1). Aminoguanidine, another selective iNOS inhibitor, was not chosen because it has antioxidant properties that prevent one from drawing mechanistic conclusions concerning the role of NO overproduction in the pathogenesis of sepsis (28). Nonselective inhibitors, such as N-monomethyl-L-arginine, used by Booke et al. (5), have adverse effects on organ perfusion because they also inhibit ecNOS, the constitutive form of NOS, which plays a key role in regulating microvascular blood flow.
Time course of NO overproduction. As shown in Fig. 1, levels were not significantly increased until 21 h after the induction of pneumomia, which is rather late compared with several lipopolysaccharide (LPS) and cecal ligation and perforation models with a more rapid increase in (19, 22, 24). Because levels began to increase between 9 and 15 h after the induction of pneumonia, L-NIL infusion was not started until 10 h.
Alterations in pancreatic perfusion with pneumonia. The results of the intravital video microscopy study of microvascular perfusion of the pancreas after the induction of pneumonia were unexpected. Although velocities in capillaries at 6 h after the induction of pneumonia were significantly decreased by 23% relative to sham animals, this occurred without a loss of FCD or a decrease in the RBC supply rate. By 24 h, when sepsis was clearly established, the expected progressive deterioration of microvascular perfusion of the pancreas had not occurred. Surprisingly, there was no change in FCD, but there was a 22% increase in RBC velocity and a 38% increase in the RBC supply rate above the 24-h sham values. Relative to the 6-h pneumonia group, the increases in RBC velocity and RBC supply rate were 74% and 62%, respectively, indicating a substantial rebound and increase in microvascular blood supply in these capillaries. This substantial increase in capillary blood flow with the onset of normotensive sepsis has not been previously reported in the pancreas.
Mechanisms responsible for changes in pancreatic perfusion. The decrease in velocity at 6 h cannot be attributed to NO overproduction because plasma levels did not begin to increase until sometime after 9 h. A possible cause for the decrease in velocity could be increased levels of endothelin. Krejci et al. (14) reported that an endothelin antagonist, bosentan, restored blood flow in the pancreas in a peritonitis septic shock model in pigs, thus indicating that endothelin was responsible for the decreased pancreatic blood flow in their model. If the decrease in velocity we observed was due to endothelin, then endothelin levels have increased in the pancreas before the upregulation of iNOS and overproduction of NO.
Although we had not expected the increase in velocity and supply rate at 24 h, the L-NIL data does allow us to address the mechanism responsible for this increase. When L-NIL was administered to maintain plasma levels at baseline, mean RBC velocity fell relative to the 24-h pneumonia group (significantly different, Table 4) to the level of the 24-h sham group (not significantly different). Because L-NIL is a highly selective inhibitor of iNOS and only sufficient L-NIL was administered to inhibit the overproduction of NO [ levels remained at baseline (Fig. 1) with no significant change in MAP (Table 2)], we can state that the mechanism responsible for the increase in RBC velocity with pneumonia at 24 h was NO overproduction in the pancreas.
L-NIL treatment also resulted in a significant decrease in mean RBC lineal density (relative to both the 24-h sham and 24-h pneumonia groups, Table 4). RBC lineal density is a measure of capillary hematocrit. Because the decrease in lineal density occurred without a decrease in systemic RBC count (Table 2), there must have been a redistribution of RBC flow away from the capillaries being observed at the surface to other regions deeper within the pancreas (capillaries in these regions would have higher lineal densities). These data demonstrate that without the overproduction of NO in sepsis, there would have been a maldistribution of RBC flow within the pancreas.
The decrease in RBC velocity and lineal density with L-NIL treatment resulted in a substantial decrease in RBC supply rate, from 10.9 ± 8.8 RBCs/s in the untreated pneumonia animals to 4.2 ± 4.2 RBCs/s (Table 4). Without NO overproduction, the RBC supply would have been seriously impaired, a situation that would be consistent with the decrease in flow observed in LPS and peritonitis sepsis models. These results suggest that the difference between our study and these models likely reflects whether vasodilation (NO) or vasoconstriction (endothelin) dominated the inflammatory response (4).
Pancreatic injury in this pneumonia model. Unlike the endotoxin models of sepsis, which did demonstrate evidence of pancreatitis and pancreatic injury (21, 27), there was no significant increase in amylase or lipase levels indicative of substantial pancreatic damage in this model of sepsis. This is consistent with the microcirculatory data, which showed no evidence of pancreatic ischemia. Amylase did increase significantly from 6 to 24 h, but this occurred in both sham and pneumonia groups. Although PI staining did demonstrate a statistically significant fourfold increase in lethal cell injury in the 24-h pneumonia group (Fig. 2), this injury was not sufficient to be reflected in increased amylase or lipase levels. We speculate that the PI staining data reflect an underlying impairment of pancreatic function as found in the clinical study of septic shock patients, but clearly, additional studies are required.
The mechanism responsible for the increased lethal cell injury is not apparent from the present study. L-NIL treatment had no effect on amylase and lipase results or on PI staining data, indicating that elevated levels of NO were not responsible for the cell injury. It is unlikely that WBC accumulation in the pancreas was responsible because L-NIL treatment prevented the fall in WBC counts at 24 h (Table 2), which is consistent with several other animal models of inflammation where iNOS inhibitors suppressed the infiltration of inflammatory cells (7, 8, 18).
In conclusion, pneumonia-induced sepsis caused a reduction in pancreatic capillary perfusion at 6 h and an unexpected rebound and increase in RBC velocity and supply rate at 24 h, without a loss in perfused capillary density. NO overproduction was beneficial to pancreatic capillary perfusion and not responsible for cell injury. Further studies are needed to establish whether this normotensive pneumonia model of sepsis induces pancreatic cellular dysfunction similar to what was demonstrated in the previous clinical study with septic shock patients (25). Clearly, the time course of upregulation of mediators involved in the sepsis-induced changes of microvascular perfusion and of cell injury and pancreatic dysfunction deserve further investigation.
|
ACKNOWLEDGMENTS |
|---|
GRANTS
This research was supported by Heart and Stroke Foundation of Ontario Grant T4476 (to W. J. Sibbald) and Canadian Institutes of Health Research Grant MOP-49416 (to C. G. Ellis). B. Tribl was supported by a Charlotte-Bühler stipend from the Fonds zur Förderung der Wissenschaftlichen Forschung (Austria).
|
FOOTNOTES |
|---|
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
|
REFERENCES |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  J. H. Crawford, B. K. Chacko, H. M. Pruitt, B. Piknova, N. Hogg, and R. P. Patel Transduction of NO-bioactivity by the red blood cell in sepsis: novel mechanisms of vasodilation during acute inflammatory disease Blood, September 1, 2004; 104(5): 1375 - 1382. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Visit Other APS Journals Online |
