Pathophysiologic Role of Oxygen Free Radicals in Acute Pancreatitis

Abstract

Background and Objective

Oxidative stress is an important factor in the pathogenesis of acute pancreatitis, as shown in vivo by the beneficial effects of scavenger treatment and in vitro by the potential of free radicals to induce acinar cell damage. However, it is still unclear whether oxygen free radicals (OFR) act only as mediators of tissue damage or represent the initiating event in acute pancreatitis in vivo as well. In the present study the authors aimed to address this issue in an experimental set-up.

Materials and Methods

Two hundred male Wistar rats were randomly assigned to one of the following experimental groups. In two groups, acute necrotizing pancreatitis was induced by retrograde intraductal infusion of 3% sodium taurocholate. Through the abdominal aorta, a catheter was advanced to the origin of the celiac artery for continuous regional arterial (CRA) pretreatment with isotonic saline (NP-S group) or superoxide dismutase/catalase (NP-SOD/CAT group). In another group, oxidative stress was generated by CRA administration of xanthine oxidase and intravenous administration of hypoxanthine (HX/XOD group). Sham-operated rats received isotonic saline both arterially and intraductally. After observation periods of 5 and 30 minutes and 3 and 6 hours, the pancreas was removed for light microscopy and determination of reduced glutathione (GSH), oxidized glutathione (GSSG), conjugated dienes (CD), and malondialdehyde as a marker for OFR-induced lipid peroxidation as well as myeloperoxidase as a parameter for polymorphonuclear leukocyte accumulation.

Results

A significant decrease of GSH was paralleled by an increased ratio of GSSG per total glutathione and elevated CD levels after 5 minutes in the NP-S group versus the sham-operated group. Thereafter, the percentage of GSSG and GSH returned to normal levels until the 6-hour time point. After a temporary decrease after 30 minutes, CD levels increased again at 3 hours and were significantly higher at 6 hours in contrast to sham-operated rats. Myeloperoxidase levels were significantly elevated at 3 and 6 hours after pancreatitis induction. In contrast to NP-S rats, treatment with SOD/CAT significantly attenuated the changes in glutathione metabolism within the first 30 minutes and the increase of CDs after 6 hours. HX/XOD administration lead to changes in levels of GSH, GSSG, and CDs at 5 minutes as well as to increased myeloperoxidase levels at 3 hours; these changes were similar to those observed in NP-S rats. Acinar cell damage including necrosis was present after 5 minutes in both NP groups, but did not develop in HX/XOD rats. In addition, serum amylase and lipase levels did not increase in the latter group. SOD/CAT treatment significantly attenuated acinar cell damage and inflammatory infiltrate compared with NP-S animals during the later time intervals.

Conclusion

OFRs are important mediators of tissue damage. However, extracellular OFR generation alone does not induce the typical enzymatic and morphologic changes of acute pancreatitis. Factors other than OFRs must be involved for triggering acute pancreatitis in vivo.

Acute pancreatitis leads to various degrees of interstitial edema, acinar cell damage, hemorrhage, and the recruitment of leukocytes into the damaged gland. ^1,2 Oxygen free radicals (OFRs) have been implicated as an important factor in the pathogenesis and progress of this disease. As highly reactive biochemical species, OFRs exert their pathophysiologic effects by directly attacking lipids ³ and proteins ⁴ in the biologic membranes at the local site of generation and cause their dysfunction. ^5,6 Indirectly, they act on the arachidonic acid cascade by two mechanisms. First, they increase the production of thromboxane, which lowers tissue circulation by its potent platelet-aggregating and vasoconstricting effects. ⁷ Second, they enhance the production of leukotriene B4, which promotes the activation of leukocytes ⁸ and the discharge of lysosomal enzymes. ⁹ These changes contribute to further cell damage.

Since the first study by Sanfey et al, ¹⁰ many publications have addressed the role of OFRs in acute pancreatitis in both experimental ^11,12 and clinical respects. ¹³ In most of these reports, an indirect approach was chosen to investigate their pathophysiologic role by applying different types of radical scavengers. ^11,13 Most of these studies showed that free radical scavengers significantly ameliorated, but did not inhibit the characteristic changes observed with acute pancreatitis; OFRs are thus considered to be important mediators within this disease.

From in vitro studies, there is clear evidence that pancreatic acinar cells are susceptible to oxidative stress. Different oxygen radical species induce severe acinar cell damage in a dose- and time-dependent manner, which led to the conclusion that OFRs may be the crucial event for initiating the pathophysiologic changes of acute pancreatitis. ¹² Surprisingly, only a few studies have been performed to investigate the potential of OFRs to induce the typical enzymatic and morphologic alterations of acute pancreatitis in vivo, and their results are controversial. ^14,15 As a consequence, the question of whether OFRs act as mediator or as initiating event in acute pancreatitis under in vivo conditions remains unanswered. In the present study we aimed to address this issue in an in vivo experimental set-up.

MATERIALS AND METHODS

Experimental Procedures

Two hundred inbred male Wistar rats (285 ± 54 g) were acclimated for at least 1 week before use. They were maintained on a 12/12 hour light/dark cycle with free access to standard rodent chow and water. All studies were performed in accordance with the national guidelines for the use and care of laboratory animals and approved by the local animal care and use committee.

Rats were anesthetized with halothane (Fluothane; Zeneca, Plankstadt, Germany) after receiving 0.15 mg/kg buprenorphin subcutaneously (Temgesic; Grünethal, Aachen, Germany) and were placed in a supine position. A 26-gauge polyethylene catheter (Abbocath-T; Abbott, Wiesbaden, Germany) was inserted in the abdominal aorta below the renal arteries and advanced to the origin of the celiac artery for continuous regional arterial (CRA) perfusion of the pancreas. This technique was established and optimized in a separate series of 10 rats before the experiments and combined minimal surgical trauma to the pancreas with adequate perfusion of the gland. In these rats, CRA perfusion was visualized by the administration of 2 mL 1:25 diluted 10% fluorescein solution (Alcon-Thilo, Freiburg, Germany) under UVA illumination during several minutes. This provided a fluorescein distribution within the entire pancreas (Fig. 1). For CRA perfusion periods longer than 20 minutes, the abdomen was partially closed by few sutures with the catheter in situ to minimize temperature loss. During the entire CRA perfusion, the animals were kept on halothane inhalation anesthesia. After completion of the CRA perfusion, the aorta was clamped proximal to the insertion site, the catheter was removed, and the aortal puncture was oversewn with 8-0 nylon. For intravenous administration of hypoxanthine, an additional 24-gauge polyethylene catheter (Insyte; Beckton Dickinson GmbH, Heidelberg, Germany) was inserted into the inferior vena cava. Depending on the observation period, rats killed at 3 hours or later were allowed to recover from anesthesia and returned to their cages with free access to food and water.

Figure 1. Continuous regional arterial perfusion of the pancreas. Through a catheter placed at the origin of the celiac artery, 1:25 diluted 10% fluorescein solution was administered and visualized under UVA illumination.

Acute necrotizing pancreatitis was induced by a standardized retrograde infusion of 0.1 mL/100 g body weight of a freshly prepared 3% sodium taurocholate solution (Sigma-Aldrich, Steinheim, Germany) into the biliopancreatic duct, as previously described. ¹⁶

Experimental Groups

Rats were randomly assigned to one of the following experimental groups. After observation periods of 5 or 30 minutes and 3 or 6 hours after induction of pancreatitis, rats were exsanguinated under anesthesia by aortal puncture. Serum was taken for amylase and lipase determination. The duodenal and the splenic parts of the pancreas were rapidly excised and separately processed for light microscopic examination and determination of reduced glutathione (GSH), oxidized glutathione (GSSG), conjugated dienes (CDs), malondialdehyde (MDA), and myeloperoxidase (MPO).

Pancreatitis Group (n = 40)

Fifteen minutes before the induction of acute necrotizing pancreatitis, rats received CRA perfusion of isotonic saline at a flow rate of 1 mL/hr per 100 g body weight. This was continued for a maximum of 60 minutes after pancreatitis induction. Ten rats were killed at each of the time intervals.

Pancreatitis Scavenger Group (n = 40)

Fifteen minutes before the induction of acute necrotizing pancreatitis, rats received CRA perfusion of superoxide dismutase (SOD) (bovine SOD; Boehringer Mannheim, Mannheim, Germany) at an hourly dosage of 100,000 units/kg and a flow rate of 1 mL/hr per 100 g body weight. This was again continued for a maximum of 60 minutes after pancreatitis induction. Animals observed for 3 and 6 hours received a CRA bolus injection of 200,000 units catalase (CAT) (bovine CAT; Boehringer Mannheim) both at the beginning and the end of the CRA perfusion period. Rats killed at 5 and 30 minutes received only one bolus injection of 200,000 units at the beginning of the CRA perfusion. Ten rats were killed at each time interval.

Oxidative Stress Group (n = 40)

Xanthine oxidase (XOD) (bovine XOD; Boehringer Mannheim) was administered at a dosage of 4 units per hour and a flow rate of 0.5 mL/hr per 100 g body weight using the CRA catheter. After 5 minutes of XOD perfusion, continuous intravenous infusion of 10 mmol/L hypoxanthine (HX) (Sigma-Aldrich) was started at the same flow rate. Ten rats were killed at each time interval after the onset of the combined HX/XOD administration, which was continued for a maximum of 60 minutes.

Sham-Operated Group (n = 40)

Animals in this group received retrograde infusion of isotonic saline instead of 3% sodium taurocholate into the biliopancreatic duct after 15 minutes of CRA perfusion with isotonic saline. As with the other groups, CRA perfusion was performed for a maximum of 60 more minutes, with 10 rats killed at each time interval.

Measurements

α-amylase and lipase activity was determined in serum by a standard clinical method for automated analysis (both DADE Behring, Liederbach, Germany).

To determinate tissue concentrations of GSH, GSSG, CDs, MDA, MPO, and protein, portions of the duodenal part of the pancreas were snap-frozen in liquid nitrogen immediately after excision and stored at −80°C until later assay. GSH and GSSG concentrations were determined using a kinetic method described by Brehe and Burch, ¹⁷ modified by Griffith. ¹⁸ CDs were measured by the methods of Buege and Aust, ¹⁹ modified by Corongiu and Milia. ²⁰ Malondialdehyde levels were determined according to the method of Ohkawa et al. ²¹ Myeloperoxidase levels were measured as described by Bradley et al. ²² Tissue concentrations of the respective parameters were related to the protein content in the pancreatic samples. The protein concentration was determined by the method of Lowry et al. ²³

Light Microscopy

Tissue samples representing about half of both the duodenal and splenic parts of the pancreas were fixed in 4% phosphate-buffered formalin for 24 hours and embedded in paraffin. Five-μm-thick sections were stained with hematoxylin and eosin and examined and graded in a masked fashion as described below.

Assessment of Cell Damage, Extracellular Edema, and Inflammatory Infiltrate

Edema was assessed planimetrically at a magnification of ×40. The total areas occupied by the interlobular space and pancreatic parenchyma were determined in each histologic section. Areas containing vessels or adipose tissue were excluded. Edema was calculated as the percentage of the area of the interlobular space per the area of pancreatic parenchyma assessed in each section.

Cell damage was defined as zymogen degranulation, cytoplasmic vacuolization or shrinkage, loss of the basal basophilic/apical acidophilic staining of the cytoplasm, and pyknosis of acinar cell nuclei. At least 1,000 random acinar cells per histologic section were assessed at a magnification of ×160. The frequency of cell damage is given as a percentage.

Inflammatory cells were quantified by analyzing at least 10 random fields containing inter- or intralobular blood vessels at a magnification of ×400. The number of inflammatory cells (polymorphonuclear leukocytes, lymphocytes, and monocytes) within the vascular lumen, marginated or adherent to the endothelium, and in the perivascular pancreatic tissue were counted in each high-power field. The inflammatory infiltrate is presented as the mean number of inflammatory cells per high-power field, counted separately for each of the three compartments.

Statistics

All values are presented as median+upper quartile. Variables were tested for group differences with the Wilcoxon rank sum test or the Student t test, if the variables were normally distributed. In all instances, P values of less than .05 at α < .05 were considered significant. Statistical calculations were done with the MedCalc software package ²⁴ (MedCalc Software, Mariakerke, Belgium).

RESULTS

Untreated Pancreatitis Group

All untreated rats survived the 6-hour observation period. Ascites developed in all rats 3 hours after the induction of pancreatitis (P < .0001) and further increased until 6 hours (P < .0001) versus sham-operated animals, in whom no relevant ascites formation was found (Table 1). Five minutes after pancreatitis induction, serum amylase (P < .05) and lipase (P < .001) levels were markedly elevated versus the sham-operated group and consecutively increased up to the 6-hour interval. At all time points, both enzymes reached significantly higher levels than observed in sham-operated rats.

Table 1. ASCITES, α-AMYLASE, AND LIPASE ACTIVITY IN THE DIFFERENT EXPERIMENTAL GROUPS

HX/XOD, hypoxanthine/xanthine oxidase; NP, necrotizing pancreatitis; SOD/CAT, superoxide dismutase/catalase.

Data are expressed as medians + upper quartile.

* Significant differences between NP-S vs. sham: ascites, P < .0001 (3 hr and 6 hr); amylase, P < .05 (5 min and 3 hr), P < .001 (30 min and 6 hr); lipase, P < .001 (5 min, 30 min, 6 hr), P < .01 (3 hr)

° Significant differences between NP-SOD/CAT vs. NP-S: ascites, P < .0004 (3 hr), P < .007 (6 hr); amylase, P = n.s.; lipase, P = n.s.

# Significant differences between HX/XOD vs. sham: amylase, P < .02 (30 min), P < .001 (3 hr and 6 hr); lipase, P < .001 (30 min and 3 hr), P < .01 (6 hr)

Significant edema formed at 3 hours after the induction of pancreatitis and further increased until the 6-hour time interval (Table 2). In contrast to sham-operated rats, histologic examination revealed the first evidence of acinar cell damage as early as 5 minutes after pancreatitis induction. The cell damage became more severe until the end of the observation period at 6 hours. Inflammatory cells were present within the vessels at all time points. More extensive intravascular margination, adherence, and diapedesis were observed 3 hours after the induction of pancreatitis compared with sham-operated rats (P < .001). In the further course, inflammatory cells were continuously recruited to the extracellular compartment and invaded the whole pancreas, especially the areas of cellular damage, at 6 hours. In the sham-operated group, inflammatory infiltrate was almost absent (P < .001) at that time (Table 2).

Table 2. INTRAPANCREATIC MORPHOLOGIC ALTERATIONS IN THE EXPERIMENTAL GROUPS

HX/XOD, hypoxanthine/xanthine oxidase; NP, necrotizing pancreatitis; SOD/CAT, superoxide dismutase/catalase.

Data are expressed as medians + upper quartile.

* Significant differences between NP-S vs. sham: edema, P < .0001 (3 hr and 6 hr); cell damage, P < .001 (5 min), P < .0002 (30 min to 6 hr); inflammatory cells intravascular, P < .05 (3 hr); adherent, P < .03 (30 min), P < .001 (3 hr), P < .02 (6 hr); extravascular, P < .001 (6 hr).

° Significant differences between NP-SOD/CAT vs. NP-S: edema, P = n.s.; cell damage, P < .05 (30 min), P < .001 (3 hr), P < .05 (6 hr); inflammatory cells intravascular, P < .05 (6 hr); adherent, P < .04 (5 min); extravascular, P < .001 (6 hr).

# Significant differences between HX/XOD vs. sham: inflammatory cells intravascular, P = n.s.; adherent, P < .001 (3 hr); extravascular, P = n.s.

Parallel to the first histologic alterations and the pancreatic enzyme elevation, decreased tissue concentrations of GSH were present at 5 minutes (P < .04) and 30 minutes (P < .01) compared with sham-operated animals. At 3 and 6 hours, GSH levels had increased again and no longer differed from those of the sham-operated group (Table 3). Along with the decreased GSH levels, GSSG levels peaked as early as 5 minutes after pancreatitis initiation compared with sham-operated rats (P < .01). Thereafter, the percentage of GSSG returned in stepwise fashion to the same levels as found in the sham-operated group.

Table 3. REDUCED GLUTATHIONE, OXIDIZED GLUTATHIONE, AND PERCENTAGE OF OXIDIZED GLUTATHIONE PER TOTAL GLUTATHIONE IN THE EXPERIMENTAL GROUPS

* Significant differences between NP-S vs. sham: GSH, P < .04 (5 min), P < .01 (30 min); GSSG, P < .01 (5 min); % GSSG, P < .01 (5 min), P < .05 (30 min and 3 hr).

° Significant differences between NP-SOD/CAT vs. NP-S: GSH, P < .05 (5 min and 30 min); GSSG, P < .01 (3 hr); % GSSG, P < .05 (3 hr).

# Significant differences between HX/XOD vs. sham: GSH, P = n.s.; GSSG, P < .01 (5 min); % GSSG, P < .05 (5 min)

Data are expressed as medians + upper quartile.

GSH, reduced glutathione; GSSG, oxidized glutathione; HX/XOD, hypoxanthine/xanthine oxidase; NP, necrotizing pancreatitis; SOD/CAT, superoxide dismutase/catalase.

Enhanced lipid peroxidation in terms of elevated MDA and CD concentrations was present in all rats of the pancreatitis group. Malondialdehyde levels peaked 5 minutes after the induction of pancreatitis and quickly returned to low levels in the further course of the disease (Table 4). CD concentrations showed a biphasic course with an early peak at 5 minutes after pancreatitis induction, a subsequent decrease, and another peak at the 6-hour interval. Whereas differences in malondialdehyde levels were not significant, CD levels were higher in rats with pancreatitis after 5 minutes (P < .01), 3 hours (P < .05), and 6 hours (P < .01) compared with those in the sham-operated group.

Table 4. MALONDIALDEHYDE, CONJUGATED DIENES, AND MYELOPEROXIDASE IN THE EXPERIMENTAL GROUPS

Data are expressed as medians + upper quartile.

CD, conjugated dienes; HX/XOD, hypoxanthine/xanthine oxidase; MDA, malondialdehyde; MPO, myeloperoxidase; NP, necrotizing pancreatitis; SOD/CAT, superoxide dismutase/catalase.

* Significant differences between NP-S vs. sham: MDA, P = n.s.; CDs, P < .01 (5 min and 6 hr), P < .05 (3 hr); MPO, P < .01 (3 hr and 6 hr)

°Significant differences between NP-SOD/CAT vs. NP-S: MDA, P = n.s.; CDs, P < .05 (6 hr); MPO, P = n.s.

# Significant differences between HX/XOD vs. sham: MDA, P = n.s.; CDs, P < .02 (5 min); MPO, P < .01 (3 hr)

Consistently with the inflammatory infiltrate present in the histologic sections, MPO levels did not show an increase within 5 and 30 minutes. However, concentrations were markedly elevated at 3 hours (P < .01) and 6 hours (P < .01) after pancreatitis induction. This was not observed in the sham-operated animals.

Pancreatitis Scavenger Group

In SOD/CAT-treated animals, ascites formation was inhibited at 3 hours (P < .0004) and 6 hours (P < .007) compared with the nontreated pancreatitis group. Serum amylase and lipase levels increased despite the pretreatment and were not different from those in nontreated rats with pancreatitis until the 6-hour observation period (Table 1).

SOD/CAT pretreatment had no effect on the formation of edema (Table 2). Although acinar cell damage per se was not prevented, the extent of the damage was significantly ameliorated by scavenger treatment at 30 minutes (P < .05), 3 hours (P < .001), and 6 hours (P < .05) after pancreatitis induction. The number of inflammatory cells adherent to the vascular endothelium was lower at 5 minutes (P < .05), and their numbers within the vessels (P < .05) and infiltrating the extravascular compartment/pancreatic tissue (P < .001) were significantly reduced by the administration of SOD/CAT at 6 hours (Table 2).

Compared with nontreated pancreatitis rats, GSH levels remained high at 5 minutes (P < .05) and 30 minutes (P < .05) in the SOD/CAT group. In addition, the percentage of GSSG per total glutathione was lower at 3 hours (P < .05) in this group (Table 3). Whereas MDA levels were not influenced by scavenger treatment, the increase of CDs was prevented at the 6-hour observation interval (P < .05) (Table 4).

Concentrations of MPO were statistically not different from those obtained in nontreated pancreatitis rats.

Oxidative Stress Group

In the oxidative stress group, no ascites formation was observed and pancreatic enzymes showed no changes throughout the whole observation period (Table 1). Moreover, neither edema formation nor acinar cell damage, according to the definition, developed at any time after HX/XOD treatment (Table 2). Three hours after HX/XOD administration, the number of cells adherent to the vascular endothelium was higher than in the sham-operated group (P < .001) but was not different from nontreated rats with pancreatitis (Table 2).

Five minutes after the administration of HX/XOD, GSH levels decreased but were not statistically different from those found in either the nontreated pancreatitis group or the sham-operated animals. At 30 minutes, GSH levels had already returned to normal concentrations and remained at this level until the 6-hour observation period (Table 3). Concomitantly, the percentage of GSSG per total glutathione was markedly increased compared with the sham-operated group (P < .05) after 5 minutes. As was observed for GSH, the percentage of GSSG immediately dropped to normal levels in the further course of the experiment.

At all time intervals, MDA concentrations did not differ from the other three groups. However, CD levels peaked at 5 minutes and were higher than in the sham-operated group (P < .02). This was again followed by a dramatic decrease at 30 minutes, with no more changes until the end of the experiment at 6 hours (Table 4).

A marked increase in MPO levels was found 3 hours after HX/XOD administration compared with sham-operated rats (P < .01). At all other time intervals, MPO levels remained below the assay detection limits (Table 4).

DISCUSSION

Several lines of evidence suggest that OFRs play an important role in acute pancreatitis. 2,10,11,13 Studies in different experimental pancreatitis models have shown that OFR-induced lipid peroxidation and changes in glutathione metabolism occur at an early stage in the course of the disease. ^16,25,26 Our results demonstrate that these changes take place at an even earlier time point than shown by previous studies. The most dramatic changes in GSH, GSSG, and lipid peroxidation products were observed within the first minutes after pancreatitis induction and paralleled the first evidence of acinar cell damage. A second increase in lipid peroxidation products, especially CDs, paralleled the later time points in which inflammatory cells began to infiltrate the damaged gland. SOD/CAT pretreatment was effective in preventing the decrease of GSH within the first 30 minutes as well as the increase in GSSG and CDs during the later time intervals. The extent of acinar cell damage and of the inflammatory infiltrate remained at a lower level in contrast to nontreated animals.

These findings imply that there are two critical stages with possibly different sources of OFR generation in acute pancreatitis: first during the initial stages and second during the later time intervals. Within the early stages, it has been proposed that OFRs trigger acute pancreatitis. In contrast to several in vitro studies, ^27–29 our model of OFR generation failed to induce acute pancreatitis, although significant intrapancreatic oxidative stress in terms of increased GSSG and CD levels was present. Currently, only one study has reported that hemorrhagic necrotizing pancreatitis could be induced by CRA infusion of HX/XOD through the celiac artery for a maximum of 5 hours. ¹⁵ Possible reasons for the different findings may be the duration of HX/XOD administration, which was confined to only 1 hour in our protocol. This might have been insufficient to deplete all intra- and extracellular antioxidant stores in the animals and in consequence to induce pancreatitis. However, at least on the acinar cell level it has been shown that short-term peak production of OFRs is more injurious to the cells than radical generation at a lower level over a more extended period. ¹² Therefore, the threefold-higher HX/XOD concentration given over a shorter period in our study should be a more efficient regimen than that used in the experiments by Tamura et al. ¹⁵ A closer review of this study reveals that their experimental procedure included ligation of the celiac artery. This is a technique used to induce ischemia, an established pathogenetic factor in experimental ^30,31 and human ³² acute pancreatitis. Because ischemia causes the release of various radical species, including nitric oxide derivates and vasoactive mediators, ³² Tamura’s study failed the basic intention of assessing the isolated toxic effects of OFRs in vivo.

With respect to other in vivo experiments, our findings are in accord with the observations of Fu et al. ¹⁴ In that report, oxidative stress was induced by diethylmaleate, which depletes the total glutathione content and failed to induce acute pancreatitis as well.

Besides OFRs and leukocytes, other radical species, mediators, and cell systems are involved in the pathophysiologic process of inflammation in animals and humans. If OFRs were the sole triggering mechanism for inducing acute pancreatitis, scavenger treatment should theoretically prevent the alterations. This, however, has never been shown by any study.

Although there is still no valid proof that OFRs alone can initiate acute pancreatitis under in vivo conditions, our results strongly support the hypothesis that they represent potent mediators in the pathophysiologic process of this disease. It is well established that OFRs induce the attraction and adhesion of leukocytes, preferentially the polymorphonuclear subset. ^8,11,13 As a secondary effect, polymorphonuclear leukocytes are responsible for the so-called respiratory burst that leads to an enhanced production of radical species and activated enzymes, with the consequence of further cell damage. ^2,10 Both the SOD/CAT and the HX/XOD groups support this mediator hypothesis: scavenger treatment ameliorated the extent of acinar cell damage and the inflammatory infiltrate, and free radical generation led to a significant accumulation and adhesion of inflammatory cells as well as to an activation of polymorphonuclear leukocytes, as shown by increased MPO levels. Moreover, SOD and CAT have short half-lives and provide effective radical scavenging only as long as continuously administered. ^16,26 The fact that even the relatively short treatment period in our protocol was effective in ameliorating the morphologic changes at later time points underscores the important mediator function of OFRs during the early stages of acute pancreatitis.

Although the exact source of OFR generation remains unclear, our data suggest that inflammatory cells themselves contribute to enhanced OFR generation, because both the inflammatory infiltrate and the CD levels were reduced at the same time during the later stages of the disease. In this context, however, it is difficult to speculate about the cause-and-effect relation between acinar cell damage and infiltrating leukocytes. The extent of acinar cell damage was reduced hours before inflammatory cells began to invade the damaged gland. Therefore, the role of the damaged acinar cell in attracting inflammatory cells must be taken in account. In addition, infiltrating leukocytes may be deleterious effectors of cellular injury by mechanisms other than OFR release alone. Recent studies evaluating different approaches to block leukocyte function ³³ or adhesion ^34,35 have uniformly achieved a significant decrease in acinar cell damage.

Previous studies by our group were performed with the cerulein ²⁶ and the taurocholate model using a 5% taurocholate solution. ¹⁶ In the present experiment, the taurocholate concentration was decreased to 3% and the duration of scavenger treatment to 1 hour. These are two experimental factors that are well known for their impact on study results, especially in the field of acute pancreatitis. ^11,13 Together with modifications of the observation periods after pancreatitis induction, they are most likely the reasons why we found no beneficial effect on edema formation and pancreatic enzyme elevation in the SOD/CAT group compared with our previous investigations. However, independent from the type and the severity of the experimental model, our findings are the same with respect to the enhanced OFR generation and an amelioration of the morphologic changes by scavenger treatment.

In summary, our results demonstrate that OFRs play an important mediator function in the early and later course of acute pancreatitis. During the later stages of the disease, their major pathophysiologic role seems to be the attraction and activation of leukocytes, which in turn contribute to enhanced radical generation and pancreatic acinar cell damage. With respect to the initial stages of acute pancreatitis, the exact site and source of OFR generation remains unclear. However, extracellularly generated OFRs cannot induce the typical enzymatic and morphologic changes of acute pancreatitis and are thus not likely to be the triggering mechanism of this disease.

Acknowledgements

The authors thank Mrs. Gabriele Wiest, Mr. Michael Marzinzig, Mrs. Elke Marzinzig, Mrs. Erika Schmidt, and Ms. Sandra Esber for their support and excellent technical assistance.