Over the past decade, both clinical and experimental studies have implicated gut injury with the development of an exaggerated systemic inflammatory state and distant organ dysfunction (1). The clinical relevance of gut origin sepsis and gut-induced distant organ failure remains controversial largely because of the failure to find translocating bacteria or endotoxin in the portal blood of severely injured trauma patients (2). However, based on recent work, it appears that proinflammatory and/or toxic factors are reaching the circulation via the mesenteric lymphatics rather than by the portal venous route (3). Not only does the mesenteric lymph of rats subjected to a combination of trauma (laparotomy) plus hemorrhagic shock (T/HS) contain humoral factors that prime neutrophils (4–6), potentiate endothelial cell injury and increase endothelial cell permeability (7,8), but ligation of the mesenteric lymph duct exiting the intestine prevents T/HS-induced lung injury (8,9) and neutrophil activation (10). Although these results provide information on the pathophysiology of T/HS-induced gut-mediated lung injury and the induction of a septic state, several important questions about the basic biology and physiology of this process remain to be answered. These include: 1) determining the identity of the biologically active factors in T/HS lymph; 2) the exact tissue and cellular sources of these factors; and 3) the signals and processes leading to their generation. Thus, the goal of this study was to test the hypothesis that pancreatic-derived serine proteases are involved in the generation of the toxic and biologically active factors contained in mesenteric lymph after T/HS and that inhibition of serine proteases activity will reduce T/HS-induced lung injury by reducing gut injury. To accomplish this goal, we used the broad-spectrum serine protease inhibitor nafamostat mesilate (6-amidino-2-naphthyl p-guanidinobenzoate dimethanesulfate) to test whether it would protect against T/HS-induced lung and/or intestinal injury as well as reduce the neutrophil activating properties of T/HS mesenteric lymph. Nafamostat is a novel synthetic broad-acting serine protease inhibitor that has been documented to inhibit the enzyme activities of trypsin, chymotrypsin, C1r, C1s, thrombin, kallikrein, and plasmin, as well as to protect a wide range of animals from lethal trypsin-induced shock (11).
MATERIALS AND METHODS
Specific pathogen-free male Sprague-Dawley rats (Taconic Farms, Germantown, NY) weighing 270–400 g were housed under barrier-sustained conditions and kept at 25°C with 12-h light/dark cycles. The rats had free access to water and chow (Teklad 22/5 Rodent Diet W-8640, Harlan Teklad, Madison, WI). All rats were maintained in accordance with the recommendations of the Guide for the Care and Use of Laboratory Animals. The New Jersey Medical School Animal Care Committee approved the experiments.
The model used in this study is a trauma-hemorrhagic shock (T/HS) model, where the rats are subjected to a laparotomy (trauma) and then either hemorrhagic shock (T/HS) or sham-shock (T/SS) as previously described (9). Rats subjected to T/HS were exsanguinated to a mean arterial blood pressure of 30 mmHg and maintained at this blood pressure for 90 min by withdrawing or reinfusing shed blood (kept at 37°C) as needed. T/SS rats had their vessels cannulated but no blood was withdrawn or infused. At the end of the shock period, the T/HS animals were resuscitated with Ringers lactate solution at three times the volume of shed blood.
The major goal of this study was to test the hypothesis that intraluminal serine proteases produced during T/HS contribute to T/HS-induced gut and lung injury. To accomplish this goal, the ability of the broad acting serine protease inhibitor, 6-amino-2-napthyl p-guanidino-benzoate dimethanesulfate (nafamostat; gift of Torrii Pharmaceuticals, Chiba Japan) to prevent gut and lung injury as well as reduce the neutrophil activating ability of T/HS mesenteric lymph was tested. In the first set of experiments, the intestinal lumen of non-fasted rats subjected to T/HS or T/SS was perfused at a rate of 0.5 mL/min with nafamostat or vehicle (Ringers lactate) through a duodenotomy. Intestinal perfusion was begun before the shock or sham shock period and continued for 10 min. After the 10-min perfusion period, the rats were subjected to shock or sham shock. Pilot studies were performed with two different doses of intraluminal nafamostat (5 mg or 10 mg per rat). At the 10 mg/rat but not the 5 g/rat dose, nafamostat caused intestinal mucosal injury in the T/SS rats (data not shown). Thus, only the 5 mg/rat dose of nafamostat was tested. Four groups of rats were studied; T/HS rats intestinally-perfused with nafamostat, T/HS rats intestinally perfused with vehicle, T/SS rats intestinally-perfused with nafamostat and T/SS rats intestinally perfused with vehicle.
In the second set of experiments, after catheter placement and the performance of a laparotomy, but before the induction of shock, the rats were given nafamostat dissolved in 1 mL of saline or the same volume of Ringer's lactate (vehicle) respectively via the internal jugular vein catheters at the rate of 0.2 mL/min. The rats were then subjected to hemorrhage or sham shock. In pilot studies, two intravenous doses of nafamostat were tested (5 mg/rat and 5 mg/kg). The rats receiving 5 mg/kg of nafamostat showed no adverse effects of the nafamostat, but all of the T/HS or T/SS rats receiving 5 mg/rat died of respiratory depression during the experimental period (data not shown). Thus, the lower nafamostat dose of 5mg/kg was used. As above, four groups of rats were studied. The initial doses of nafasmostat used in both sets of experiments were chosen based on previous studies by other investigators (11–13).
Three hours after the end of the T/HS or T/SS period, the rats were killed and lung permeability was measured. In addition, lung samples were obtained for quantitating pulmonary leukosequestration and a specimen of the ileum was obtained for histologic study.
The third group of studies was performed to compare the ability of mesenteric lymph from T/HS rats receiving intraluminal nafamostat versus vehicle to activate neutrophils. In this experiment mesenteric lymph was collected from T/HS and T/SS rats receiving either intraluminal nafamostat or vehicle. Mesenteric lymph was incubated in vitro with normal human PMNs. The effect of T/HS and T/SS mesenteric on neutrophil adhesion molecule expression (CD11b) as well as the ability of these lymph samples to prime for an augmented respiratory burst was measured. Human neutrophils were used, since our previous studies have documented that human and rat PMNs respond in a similar fashion to mesenteric lymph (14).
Lung permeability was measured both by the Evans blue dye (EBD) technique and by determining the bronchoalveolar fluid (BALF) to plasma protein ratio as previously described (8). The total protein content (g/dL) of both BALF and plasma collected from each animal was determined using a hand refractometer (Milton Roy, Rochester, NY). Using this data, the BALF to plasma protein ratio was determined (8).
Pulmonary neutrophil (PMN) sequestration
PMN sequestration was quantitated as myeloperoxidase (MPO) as previously described (8). MPO activity per gram of wet weight lung tissue was calculated as follows: MPO activity (units/g tissue) = (δA460 × 13.5)/ weight (g), where δA460 equals the rate of change in absorbance at 460 nm between 1 and 3 min. One unit of MPO activity is the amount of enzyme that will reduce 1 μmole peroxide per minute.
At sacrifice, a segment of the terminal ileum was excised and fixed in 10% buffered formalin. After processing, semithin (2–4 μm) sections were cut and stained with 1% toluidine blue. Five random fields with 100–250 villi from each animal were then analyzed in a blinded fashion using light microscopy at 100 magnification as previously described (15). Villus injury was deemed present if part of the villus tip had sloughed or if submucosal edema was present. The overall percentage of ileal villous damage was determined by dividing the number of injured villi by the total number of villi examined.
Mesenteric lymph-induced PMN adhesion molecule expression
Mesenteric lymph was collected as previously described (8). To assess the direct effects of the T/HS and T/SS lymph specimens on CD11b expression, lymph (5% final lymph concentration) was added to 100 μL aliquots of whole blood from normal volunteers, which had been diluted 1:1 with 100 μL of medium (14). After a 4-h incubation period, 10 μL of anti-human CD11b FITC-labeled monoclonal antibody (BD Pharmingen, San Diego CA) was added to the lymph-blood samples, after which the samples were gently and briefly vortexed, then placed on ice for 45 min in the dark. Subsequently, the red cells were lysed, the PMNs washed three times and then the level of CD11b expression was determined by flow cytometry as previously described (14). The data were expressed as mean fluorescence intensity. FITC-conjugated, isotype mouse IgG1 antibody was used as an isotype control for nonspecific antibody binding. Additionally, to determine whether mesenteric lymph primed the PMNs for increased CD11b expression, the PMNs were also tested in the presence of the PMN activator phorbol 12-myristate 13 acetate (PMA) at a concentration of 90 ng/mL. This dose of PMA was found in pilot studies to be a submaximal stimulatory dose of PMA.
Mesenteric lymph-induced PMN respiratory burst
Heparinized whole blood samples (100 μL) were incubated with an equal volume of medium, following which the red blood cells were lysed. The white blood cell pellets were washed twice and then resuspended in 400 μL of Hanks medium. The lymph samples (5% final concentration) were then added and the tubes placed into an incubator for 5 min after which dihydrorhodamine (DHR) (15 ng/mL) was added to the tubes. Five minutes after the DHR was added, the PMNs were stimulated with PMA (90 ng/mL). After 15 min at 37°C, the PMN respiratory burst was measured by flow cytometry as previously described (14).
All measurements are expressed as the mean ± SD and were analyzed by one-way analysis of variance with the Student-Newman-Keuls multiple comparisons test when comparing the differences between the means of four groups at the same time point. The relationship between two continuous variables was evaluated by linear regression analysis. Probabilities less than 0.05 were considered to be statistically significant.
As previously reported (8,9), T/HS caused lung injury as reflected in increased permeability to Evans blue dye, BALF protein levels and the BALF to plasma protein ratio (Table 1). The intraluminal administration of nafamostat significantly reduced but did not completely prevent lung injury after T/HS (Table 1). In contrast to what was observed with nafamostat perfusion of the intestine, intravenously administered nafamostat did not protect against T/HS-induced lung injury (Table 2). Similar results were observed when pulmonary neutrophil sequestration, as reflected by lung MPO levels, was measured (Table 3). Intraluminal nafamostat reduced but did not prevent T/HS-induced pulmonary leukosequestration, whereas intravenous nafamostat was not protective. Furthermore, intestinal luminal perfusion, even in the absence of nafamostat, appeared to reduce the extent of both T/HS-induced lung injury and pulmonary leukosequestration (Table 4) indicating that intestinal luminal perfusion itself has a modest protective effect on lung injury in rats subjected to T/HS.
The intraluminal administration of nafamostat largely prevented T/HS-induced intestinal ileal villous injury, since nafamostat reduced the incidence of villus injury from 17 ± 11% to 8 ± 3% (P < 0.05). The incidence of villus injury in the T/SS vehicle-treated (6 ± 1%) and the T/SS nafamostat-treated (9 ± 7%) was similar to that of the T/HS rats receiving intraluminal nafamostat. Furthermore, consistent with our previous studies indicating an association between gut and lung injury (16), linear regression analyses documented that there was a direct relationship between the extent of ileal villous injury and lung permeability (r2 = 0.67) as well as pulmonary MPO levels (r2 = 0.43;Fig. 1).
The mesenteric lymph from the T/HS rats receiving intraluminal vehicle (Ringers lactate) significantly increased PMN CD11b expression, both in the presence and absence of PMA (Table 5). These T/HS lymph samples from vehicle-treated rats also augmented the PMN respiratory burst. In contrast, the intraluminal administration of nafamostat reduced the neutrophil activating ability of T/HS mesenteric lymph to that observed in the T/SS lymph samples (Table 5). Thus, nafamostat reduced the biologic activity of T/HS lymph as well as reducing T/HS-induced gut injury.
The results of the current study shows that the intestinal, but not the intravenous, administration of the serine protease inhibitor, nafamostat, reduces T/HS-induced intestinal and pulmonary injury as well as reduces the ability of T/HS lymph to activate neutrophils. Our results, implicating serine proteases as being involved in T/HS-induced organ injury, are consistent with a recent study in a splanchnic arterial intestinal ischemia model indicating that pancreatic enzymes acting on the ischemic intestine are responsible for the production of factors leading to hypotension, neutrophil sequestration in organs, pulmonary edema and gut injury (13). Furthermore, Kistler working with Schmid-Schonbein and Hugli showed that the injection of fresh pancreatic, but not intestinal or other organ extracts into rats caused hypotension, activated leukocytes, led to microvascular injury and decreased survival (17). They also found that incubating substimulatory concentrations of pancreatic homogenates with other tissues resulted in these tissue extracts becoming toxic. Finally, they documented that incubating pancreas homogenates with nafamostat neutralized the toxicity of the pancreatic extracts (17). They interpreted their results as indicating that pancreas-derived proteolytic enzymes acting directly or indirectly, through their effect on the ischemic gut as well as other potential organs, produce activators that stimulate inflammatory cell functions in the microvasculature. Based on this work and other studies performed in their laboratory (12,18), they conclude that the ischemic pancreas is the source of these activating factors and stress that it is not the pancreatic proteases themselves that are the stimulatory factors, but that it is the proteolytic products of the pancreatic serine proteases (trypsin and chymotrypsin) that have activity. This notion that pancreatic serine proteases acting on the ischemic intestine contribute to gut injury and the production of toxic factors that lead to lung injury is consistent with our observations that intraluminal but not intravenous nafamostat reduced T/HS-induced gut and lung injury, since if the serine proteases were acting systemically, we would have expected intravenous nafamostat to have reduced lung injury in our T/HS model. In fact, our findings that intestinal perfusion with vehicle (Ringers lactate) displayed a modest but statistically significant protective effect on lung injury indicates that washing proteases as well as other potentially active factors out of the intestine limits T/HS-induced lung injury.
Even two decades before the recent work of Schmid-Schonbein (12,13,17,18), the pancreas had been proposed as playing an important pathologic role in the organ failure observed after hemorrhagic or traumatic shock (19–24). For example, the work of Lefer et al. implicated the pancreas as the source of myocardial depressant factor (22), whereas Bounous et al. (23) showed that the magnitude of ischemic gut injury was mitigated by pancreatic protease inhibition or pancreatic duct ligation. Although it is recognized that severe pancreatitis can lead to respiratory failure, shock and multiple organ dysfunction (1), studies also indicate that splanchnic ischemia can lead to the development of acute pancreatitis and that an episode of splanchnic ischemia can convert edematous to hemorrhagic pancreatitis (25,26). This recognized relationship between pancreatic protease activation and pancreatitis has resulted in clinical trials of pancreatic protease inhibition in patients with acute pancreatitis. A recent meta-analysis of the effectiveness of the serine protease inhibitor gabexate mesilate in five randomized clinical trials of patients with pancreatitis has been recently reported (27). Although the use of gabexate did not reduce the 90-day mortality rate, it did decrease the overall rate of complications and the incidence of complications requiring surgery. In light of these clinical trials, it is important to note that nafamostat has been shown experimentally to be more effective than gabexate in blocking the production of activating factors in the presence of pancreatic proteases (12).
Because interpretation of the actions of nafamostat must consider its biologic activity, it is important to stress that this broad-acting serine protease inhibitor also inhibits a number of non-pancreatic serine proteases besides the pancreatic serine proteases trypsin and chymotrypsin. These include certain complement proteins, thrombin, plasmin and kallikren (11) as well as lipase and phospholipase A2 (28). Nafamostat has also been documented to stabilize pancreatic acinar cell membranes (29).
The exact mechanism by which nafamostat limits T/HS-induced gut injury is not known. However, gut injury after T/HS or splanchnic ischemia appears to involve xanthine oxidase-generated oxidants (30) as well as the increased production of nitric oxide by the inducible form of nitric oxide synthase (14,31). Because proteases are involved in the conversion of xanthine dehydrogenase to xanthine oxidase, further study will be required to determine whether, and to what extent, nafamostat protects against ischemic gut injury by preventing the proteolytic conversion of xanthine dehydrogenase to xanthine oxidase. Likewise, little work has been conducted on the role of protease-induced intestinal nitric oxide production. Because pancreatic tissue obtained within minutes of death from patients dying with all types of shock showed a correlation between the subcellular appearance of the acinar cells and the type, severity and duration of shock (32) and nafamostat stabilizes acinar cell membranes, part of nafamostat's protective effect on the gut may be secondary to its ability to protect the pancreas. However, this possibility is less likely, since the intravenous administration of nafamostat was not effective in decreasing lung injury and the ability of nafamostat to prevent acinar cell injury would be expected to be better after the intravenous than the intraluminal administration of nafamostat.
Our results showing a direct relationship between the magnitude of T/HS-induced ileal mucosal injury and lung injury as well as pulmonary neutrophil sequestration add to the body of literature documenting a relationship between gut and lung injury (1). Because of the association between gut injury and distant organ dysfunction, several clinical trials have been performed testing various gut-protective strategies in high-risk trauma patients (33,34). These clinical trials indicate that gut-directed anti-oxidant therapy reduces organ failure. Thus, based on the results of the current experimental study and the work of others implicating pancreatic-derived proteases in the pathogenesis of gut injury as well as the dysfunction of a number of distant organs and hemostatic systems, it appears that therapy aimed at limiting the potentially deleterious effects of serine proteases in shock states might be worthwhile.
In summary, the results of the current study indicate that serine protease inhibition by the intestinal administration of nafamostat reduces both T/HS-induced gut and lung injury as well as limits the ability of T/HS mesenteric lymph to activate PMNs. Thus, the results of this study implicate serine proteases in the pathogenesis of shock-induced gut injury and the generation of biologically active mesenteric lymph.
1. Deitch EA: Multiple organ failure: pathophysiology and potential future therapy. Ann Surg 216:117–134, 1992.
2. Moore FA, Moore EE, Poggetti R: Gut bacterial translocation via the portal vein: a clinical perspective with major torso trauma. J Trauma 31:629–638, 1991.
3. Deitch EA: Role of the gut lymphatic system in multiple organ failure. Curr Opin Crit Care 7:92–98, 2001.
4. Adams CA, Hauser CJ, Adams JM, Fekete Z, Xu DZ, Sambol JT, Deitch EA: Trauma-hemorrhage-induced neutrophil priming is prevented by mesenteric lymph duct ligation. Shock 18:513–517, 2002.
5. Upperman JS, Deitch EA, Guo W, Lu Q, Xu DZ: Post-hemorrhagic shock mesenteric lymph is cytotoxic to endothelial cells and activates neutrophils. Shock 10:407–411, 1988.
6. Zallen G, Moore EE, Johnson JL, Tamura DY, Ciesla DJ, Silliman CC: Posthemorrhagic shock mesenteric lymph primes circulating neutrophils and provokes lung injury. J Surg Res 83:83–88, 1999.
7. Deitch EA, Adams CA, Lu Q, Xu DZ: Mesenteric lymph from rats subjected to trauma-hemorrhagic shock are injurious to rat pulmonary microvascular endothelial cells as well as human umbilical vein endothelial cells. Shock 16:290–293, 2001.
8. Magnotti LJ, Upperman JS, Xu DZ, Lu Q, Deitch EA: Gut-derived mesenteric lymph but not portal blood increases endothelial cell permeability and promotes lung injury after hemorrhagic shock. Ann Surg 228:518–527, 1998.
9. Deitch EA, Adams CA, Lu Q, Xu DZ: A time course of the protective effect of mesenteric lymph duct ligation on hemorrhagic shock-induced pulmonary injury and the toxic effects of lymph from shocked rats on endothelial cell monolayer permeability. Surgery 129:39–47, 2001.
10. Adams JM, Hauser CJ, Adamas CA, Xu DZ, Livingston DH, Deitch EA: Entry of gut lymph into the circulation primes rat neutrophil respiratory burst. Crit Care Med 29:2194–2198, 2001.
11. Aoyama T, Ino Y, Ozeki M, Oda M, Sato T, Koshiyama Y, Suzuki S, Fujita M: Pharmacological studies of FUT-175, nafamstat mesilate. I. Inhibition of protease activity in vitro and in vivo experiment. Japan J Pharmacol 35:203–227, 1984.
12. Kistler EB, Lefer AM, Hugli TE, Schmid-schonbein GW: Plasma activation during splanchnic arterial occlusion shock. Shock 14:30–34, 2000.
13. Mitsuoka H, Kistler EB, Schmid-schonbein GW: Generation of in vivo activating factors in the ischemic intestine by pancreatic enzymes. Proc Natl Acad Sci USA 97:1772–1777, 2000.
14. Deitch EA, Shi HP, Feketeova E, Hauser CJ, Xu DZ: Hypertonic saline resuscitation limits neutrophil activation after trauma-hemorrhagic shock. Shock 17:496–501, 2002.
15. Suzuki Y, Deitch EA, Mishima S, Lu Q, Xu DZ: Inducible nitric oxide synthase gene knockout mice have increased resistance to gut injury and bacterial translocation after an intestinal ischemia-reperfusion injury. Crit Care Med 28:3692–3696, 2000.
16. Adams Jr, CA Magnotti LJ, Xu DZ, Lu Q, Deitch EA: Acute lung injury after hemorrhagic shock is dependent on gut injury and sex. Am Surg 66:905–913, 2000.
17. Kistler EB, Hugli TE, Schmid-Schonbein GW: The pancreas as a source of cardiovascular cell activating factors. Microcirculation 7:183–192, 2000.
18. Schmid-Schonbein GW, Kistler EB, Hugli TE: Mechanisms for cell activation and its consequences for biorheology and microcirculation: multi-organ failure in shock. Biorheology 38:185–201, 2001.
19. Lefer AM, Glenn TM: Role of the pancreas in the pathogenesis of circulatory shock. Adv Exp Med Biol. 23:311–335, 1971.
20. Glenn TM, Lefer AM: Significance of splanchnic proteases in the production of a toxic factor in hemorrhagic shock. Circ Res 29:338–349, 1971.
21. Glenn TM, Herlihy BL, Ferguson WW, Lefer AM: Protective effect of pancreatic duct ligation in splanchnic ischemia shock. Am J Physiol. 222:1278–1284, 1972.
22. Lefer AM, Spath Jr: JA Pancreatic hypoperfusion and the production of a myocardial depressant factor in hemorrhagic shock. Ann Surg. 179:868–876, 1974.
23. Soma LR, Neufeld GR, Dodd DC, Marshall BE: Pulmonary function in hemorrhagic shock: the effect of pancreatic ligation and blood filtration. Ann Surg. 179:395–402, 1974.
24. Bounous G, Brown RA, Mulder DS, Hampson LG, Gurd FN: Abolition of “tryptic enteritis” in the shocked dog. Arch Surg 91:371–375, 1965.
25. Sakorafas GH, Tsiotos GG, Sarr MG: Ischemia/reperfusion-induced pancreatitis. Dig Surg 17:3–14, 2000.
26. Kyogoku T, Manabe T, Tobe T: Role of ischemia in acute pancreatitis. Hemorrhagic shock converts edematous pancreatitis to hemorrhagic pancreatitis in rats. Dig Dis Sci 37:1409–1417, 1992.
27. Messori A, Rampazzo R, Scroccaro G, Olivato R, Bossi C, Falconi M, Pederzoli P, Martini N: Effectiveness of gabexate mesilate in acute pancreatitis: a metaanalysis. Dig Dis Sci 40:734–738, 1995.
28. Fujii S, Hitomi Y: New synthetic inhibitors of C1r, C1 esterase, thrombin, plasmin, kallikrein and trypsin. Biochim Biophys Acta 661:342–345, 1981.
29. Manabe T, Hirano T, Imanishi K, Ando K, Yotsumoto F, Tobe T: Protective effect of nafamostat mesilate on cellular and lysosomal fragility of acinar cells in rat cerulein pancreatitis. In J Pancreatol 12:167–172, 1992.
30. Parks DA, Granger DN: Ischemia-induced vascular changes: role of xanthine oxidase and hydroxyl radicals. Am J Physiol 245:G285–G289, 1983.
31. Salzman AL: Nitric oxide and the gut. New Horiz 3:33–45, 1995.
32. Jones RT, Garcia JH, Mergner WJ, Pendergrass RE, Valigorsky JM, Trump BF: Effects of shock on the pancreatic acinar cell. Cellular and subcellular effects in humans. Arch Pathol 99:634–644, 1975.
33. Barquist E, Kirton O, Windsor J, Hudson-Civetta J, Lynn M, Herman M, Civetta J: The impact of antioxidant and splanchnic-directed therapy on persistent uncorrected gastric mucosal pH in the critically injured trauma patient. J Trauma 44:355–360, 1998.
34. Porter JM, Ivatoury RR, Azimuddin K, Swami R: Antioxidant therapy in the prevention of organ dysfunction syndrome and infectious complications after trauma: early results of a prospective randomized trial. Am Surg 65:478–483, 1999.