- Email: [email protected]
Soybean trypsin inhibitor attenuates ischemic injury to the feline small intestine
GASTROENTEROLOGY
1985:89:6-12
Soybean Trypsin Inhibitor Attenuates Ischemic Injury to the Feline Small Intestine DALE A. PARKS, D. NEIL GRANGER, ARVIND K. SHAH
GREGORY
B. BULKLEY,
and
Department of Physiology, College of Medicine, University of Cincinnati, Cincinnati: Departments of Physiology and Mathematics and Statistics, University of South Alabama, Mobile, Alabama; and Department of Surgery, Johns Hopkins University, School of Medicine, Baltimore, Maryland
Recent evidence suggests that oxygen free radicals are largely responsible for the increased vascular permeability and early mucosal lesions associated with partial intestinal ischemia. It is postulated that oxygen radicals are produced by the reaction of the enzyme xanthine oxidase with hypoxanthine and molecular oxygen. In normal healthy cells, xanthine oxidase exists as a nicotinamide adenine dinucleotide-reducing dehydrogenase and not the oxygen radical-producing oxidase. In the intestine, dehydrogenase-to-oxidase conversion is nearly complete with Cl min of ischemia. Biochemical evidence from the intestine and liver indicate that ischemiainduced conversion of xanthine dehydrogenase to xanthine oxidase can be prevented by administration of protease inhibitors such as soybean trypsin inhibitor. In order to assess the role of proteases in oxygen radical-mediated ischemic injury to the small bowel, quantitative analyses of mucosal Jesion development and vascular permeability were performed in autoperfused segments of cat ileum subjected to 1 or 3 h of ischemia and pretreated with 15 mgkg (i.v.) soybean trypsin inhibitor. One hour of ischemia produced a significant increase in intestinal vascular permeability. The ischemia-induced increase in vascular permeability was significantly attenuated by soybean trypsin inhibitor pretreatment. Three hours of ischemia led to the developReceived May 4, 1984. Accepted December 20, 1984. Address requests for reprints to: Dale A. Parks, Ph.D., Department of Physiology, College of Medicine, University of Cincinnati, 231 Bethesda Avenue, Cincinnati, Ohio 45267-0576. This research was supported by a grant from the National Institutes of Health (AM 33594) and the American Heart Association (AM 31764). D. N. Granger is the recipient of a Research Career Development Award from the National Heart Lung and Blood Institute (HL 00816). 0 1985 by the American Gastroenterological Association 0016-5085/85/$3.30
ment of mucosal lesions in untreated animals. Pretreatment with soybean trypsin inhibitor largely prevented the development of the mucosal lesions. The findings of our study are consistent with biochemical evidence that, during ischemia, proteases trigger the conversion of xanthine dehydrogenase to xanthine oxidase and thereby lead to oxygen radical production and subsequent tissue injury. Major early consequences in the pathogenesis of intestinal &hernia include an increased vascular permeability and destruction of the mucosal membrane. Recent evidence from our laboratory suggests that oxygen-derived free radicals are largely responsible for the endothelial (1,~) and epithelial (3,4) damage associated with a partial arterial occlusion model of intestinal ischemia. Furthermore, the evidence indicates that oxygen radicals are produced by the reaction of the enzyme xanthine oxidase with hypoxanthine, an endogenous substrate produced by adenosine triphosphate catabolism, and molecular oxygen (l-3). Normal xanthine oxidase exists in the cell as a nicotinamide adenine dinucleotide (oxidized form)-reducing dehydrogenase (type D) and not the oxygen radical-producing oxidase (type 0). Conversion of xanthine dehydrogenase to xanthine oxidase can be induced by a number of conditions, including proteolysis and incubation under low oxygen tension in the presence of substrate (5). Addition of soybean trypsin inhibitor (STI) to rat liver homogenate prevents dehydrogenase-to-oxidase conversion proportional to the amount of ST1 added (6), whereas addition of any of several purified proteases induces dehydrogenase-to-oxidase conversion (7). McCord and Roy (7) have reported that, in the Abbreviation
used in this paper:
STI, soybean
trypsin
inhibitor.
INTESTINAL
July 1985
intestine, nearly complete dehydrogenase-to-oxidase conversion occurs with
of intestinal ischemia is the development of mucosal lesions. Therefore, morphologic alterations in the mucosal region of the small intestine were used as an index for assessment of tissue injury produced by longer (3 h) durations of ischemia. Materials
and Methods
Surgical Procedure The experimental preparation used in this study is similar, except for minor modifications, to that described in detail in previous studies (1,lO). Briefly, 42 cats previously fasted for 18-24 h were initially anesthetized with 50 mg/kg of ketamine-HCl. The right femoral artery and vein were cannulated and anesthesia was maintained by administration of sodium pentobarbital into the femoral vein. Systemic arterial pressure was measured with a Statham P23A transducer which was connected to a carotid artery cannula. A tracheotomy was performed to facilitate breathing and as a means of artificial ventilation if the cats failed to breathe spontaneously during the experiment. A midline abdominal incision was made and a 15-ZOcm segment of ileum was isolated, maintaining blood and lymph vessels intact. The remainder of the small and large intestine were extirpated. Immediately after isolation of the ileal segment and before cannulation of the lymphatic and blood vessels, heparin (5 mg/kg) was administered intravenously. Body temperature was maintained at 37°C with a thermistor-controlled infrared lamp. To minimize evaporation and tissue dehydration, all exposed tissues
ISCHEMIA
7
were moistened with saline-soaked gauze and placed in a plastic bag. In the vascular permeability studies, a large lymphatic vessel emerging from the mesenteric pedicle was cannulated and lymph flow was determined by observing the movement of lymph in a calibrated pipette (50~ZOO ~1, full scale). Lymph (C,) and plasma (Cp) total protein concentrations were measured with a refractometer [American Optical Corp., Buffalo, New York]. A large cannula was inserted into the superior mesenteric vein, and venous outflow drained into a reservoir mounted on a vertically positioned pulley system. Blood from the reservoir was returned to the animal via the cannula inserted into the femoral vein. Venous outflow pressure of the intestinal segment was set by adjusting the height of the reservoir and monitored from a T-connector in the venous circuit. In all experiments (vascular permeability and morphologic), an arterial circuit was established between the superior mesenteric and femoral arteries. Superior mesenteric arterial pressure was measured via a T-tube interposed within the arterial circuit using a Statham P23A transducer. All pressure cannulas and associated transducers were positioned at heart level. Heparinized blood from a donor animal was used to prime all the extracorporeal blood circuits. Systemic and superior mesenteric arterial pressures and venous pressure were continually recorded with a Grass physiologic recorder (Grass Instrument Co., Quincy, Mass.). At the end of each experiment the intestinal segment was weighed and lymph flow was normalized to milliliters per minute times 100 grams of used in the morphotissue. The experimental preparation logic studies was similar to that described above, except that the venous and lymphatic vessels were not cannulated.
Experimental
Protocols
In the vascular permeability studies, control values for intestinal lymph flow and lymph and plasma protein concentrations were obtained at a normal local arterial pressure and a venous pressure of 0 mmHg. Local arterial pressure was reduced to 25-35 mmHg using an adjustable clamp interposed within the local arterial circuit. Superior mesenteric arterial pressure was maintained at the reduced pressure for 60 min and then released. Forty-five minutes after reducing arterial pressure, additional lymph flow and lymph and plasma protein concentration measurements were made. After 60 min of local arterial hypotension, the arterial clamp was released and all measured parameters were allowed to reach a new steady state. Venous pressure was then elevated in lo-mmHg increments up to 40 mmHg. Intestinal venous pressure was maintained constant at each pressure level until all parameters were in a steady state, at which time lymph and plasma samples were acquired for protein concentration determination. In the morphologic studies, the animals (n = 14) were allowed to stabilize for 30 min after the operative procedure. Then, a “control” biopsy specimen was taken for histologic evaluation. Local arterial pressure was then reduced to 25-35 mmHg using an adjustable Gaskell
8
PARKS ET AL.
GASTROENTEROLOGY
clamp interposed within the arterial circuit. Superior mesenteric arterial pressure was maintained at the reduced level for 3 h and then released. The cats were then observed for 1 h after release of the arterial clamp, at which time an additional tissue biopsy specimen was taken for histologic examination. Tissue sections were always taken at the midpoint (r3 cm) of the intestinal loop. Soybean trypsin inhibitor (Sigma Chemical Co., St. Louis, MO.) was administered intravenously (15 mgikg) immediately before the onset of the 1 h (n = 7) or 3 h (n = 7) partial arterial occlusion. This dose of ST1 has been previously shown to prevent dehydrogenase-to-oxidase conversion in rat small intestine (7).
Osmotic
Reflection
The osmotic reflection coefficient (od) was estimated using the steady state relationship between the lymphto-plasma protein concentration ratio (C&p) and lymph flow in the postocclusion period. As lymph flow is increased, C&p rapidly decreases (filtration rate dependent) and then becomes relatively constant at a minimal value (filtration rate independent) when lymph flow is high. At a normal portal pressure, the exchange of macromolecules across the intestinal capillary wall occurs by both diffusion and convection. Elevation of venous pressure increases the convective movement of macromolecules across the capillary wall while at the same time the diffusive contribution to total exchange is reduced to a negligible level. Theoretical and experimental evidence suggests that ad = 1 - C&p, when C&p is filtration rate independent, i.e., when diffusive exchange is negligible (10). In the present study, fld for total plasma proteins was assessed using this approach. The theoretical basis for this approach is described in Reference 10.
Histologic
Technique
Tissue samples taken for histologic examination were -1 x 1.5 cm in size. After excision, each sample was rapidly mounted on a piece of Styrofoam with the mucosal surface exposed. The samples were immediately fixed in 10% neutral buffered formalin (Carson’s fixative). The tissues were dehydrated, cleared, embedded in paraffin, cut at about 6 pm, and stained with hemotoxylin and eosin.
Morphologic
and
Morphometric
affected; and 4+, epithelium of tips, midportions, and lower-portions of majority of villi affected. Crypt epithelial cell inflammation and necrosis were graded as follows: 1+, epithelium of occasional crypts affected; 2 +, scattered crypts affected; 3 +, many crypts affected; and 4 +, majority of crypts affected. Villus height, crypt depth, and mucosal thickness in each section were determined for all welloriented villi by measurement with an ocular micrometer. Each villus on each specimen was individually evaluated by the above criteria and the individual scores, as well as the mean score for each specimen, were recorded.
Statistical Analyses Vascular
Coefficients
Evaluation
Coded histopathoiogic sections were examined for ischemic injury, as evidenced by inflammation and necrosis of villus and crypt epithelial cells and destruction of the villus architecture of the mucosa, using a modification of a morphometric system we have previously described (3). This evaluation was made without knowledge of the treatment group from which the specimen came. Villus epithelial cell inflammation and necrosis were graded as follows: 0, no visible injury; 1 + ,epithelium of occasional villus tips affected: 2+, majority of tips affected; 3+, epithelium of majority of tips and midportion of some villi
Vol. 89, No. 1
Permeability
Studies
Two different statistical approaches were used. The first approach involved a one-way analysis of variance and Scheffe’s method of multiple comparison between the filtration-independent values of the lymph-to-plasma protein concentration ratio (CL/Cp) of the control group and of the groups subjected to 1 h local hypotension and local hypotension plus ST1 pretreatment. The relationship between C& and lymph flow for the various experimental protocols was compared by fitting a regression model for each condition and testing for coincidence of the regression curves. In terms of a linearized model, a relationship of the form In y = B, + B,x + B2xZ + Bgz + B,xz + BSxz + E,
where In y is the natural log of y and E is a random error term, is assumed between C&p (variable y) and lymph flow (variable x). A regression model was fit for each experimental protocol with the statistical analysis system package utilizing indicator (dummy) variables. After fitting the model, the null hypothesis (i.e., regression curves are coincident) was tested against the alternate hypothesis (i.e., regression curves are not coincident) and the observed significance level (p value) was calculated. Based on the p value, a decision was made as to whether the sample evidence was sufficient to refute the null hypothesis (p < 0.05). Morphologic
Studies
For these studies, each individual villus was considered to be the unit of observation and was scored using the criteria previously described. Each treatment group (i.e., control, ischemia, and STI) contained 89-213 units. A one-way analysis of variance and Scheffe’s method of multiple comparison was used to statistically compare the individual units of each experimental group. A significant difference between the groups was assumed when a level of significance (p value) was cO.05.
Results Figure 1 illustrates the relationship between the lymph-to-plasma protein concentration ratio (CL/ C,) and lymph flow for control conditions and after 1 h of local arterial hypotension. The curves repre-
July
INTESTINAL
1985
ISCHEMIA
9
sentative of the mean injury score obtained in the control group, the ischemia group, and the ischemia plus STI-pretreatment group. The morphologic and morphometric changes produced by ischemia are summarized for all groups in Table 1.The results in Table 1 clearly demonstrate that pretreatment with ST1 significantly diminished the severity of the mucosal lesions induced by 3 h of regional ischemia. The mean mucosal thickness and mean villus height were significantly greater in the ischemic preparations pretreated with ST1 compared with ischemia alone. In addition, ST1 significantly reduced the villus and crypt epithelial necrosis produced by ischemia. Both statistical analyses (regression model and one-way analysis of variance) predicted significant differences (p < 0.05) between the pretreatment group (STI) and ischemia alone for mucosal thickness, villus height, villus epithelial necrosis, and crypt epithelial necrosis. No injury to the muscularis propria or deeper bowel wall layers was present in any group.
l.OOf c
Discussion Lymph (ml/min
Figure
Flow x
1OOg)
1. Relationship between intestinal lymph-to-plasma total protein concentration ratio (C&p) and lymph flow under control conditions (open and closed circles) and after 1 h of ischemia (open and closed squares). The open symbols represent data from the present study and the closed symbols are data from previous studies (2,9). The capillary reflection coefficient (u(j) was derived assuming ad = 1 - C&e at lymph flows >0.40 mlimin X 100 g. Reflection coefficient values of 0.92 t 0.009 and 0.58 + 9.012 were obtained for the control and ischemia groups, respectively.
senting control conditions and ischemic conditions were acquired in previous studies (2,9) plus data from seven additional experiments. Using 1 - CL/Cp as an estimate of the osmotic reflection coefficient (ad) when CL/& is filtration rate independent, one would predict fld values of 0.92 2 0.009 for control conditions and 0.58 t 0.012 after ischemia. The ischemia group was significantly different (p < 0.01) from the control groups. The relationships between the lymph-to-plasma protein concentration ratio (CLICP) and lymph flow for control conditions (solid line), after local arterial hypotension (broken line), and after arterial hypotension with ST1 pretreatment [solid circles) are illustrated in Figure 2. One would predict a Od value of 0.84 +- 0.012 for ischemia plus STI. The STIpretreated group was significantly different (p < 0.01) from the ischemia alone and control groups. Figure 3 illustrates mucosal sections repre-
Oxygen-derived free radicals have recently been implicated in the vascular endothelial and mucosal epithelial damage associated with intestinal ischemia (l-3). Short durations of intestinal ischemia (i.e., 1 h) are associated with subtle morphologic changes such as mitochondrial swelling and moderate interstitial edema; the mucosal membrane remains intact (8). The permeability of intestinal
Lym~~h (mllmtn
Figure
Flow x 1009)
2. Effect of 15 mgikg soybean trypsin inhibitor on the relationship between the lymph-to-plasma protein concentration ratio and lymph flow after 1 h of ischemia. Solid line represents control curve, broken line represents ischemia alone. A reflection coefficient of 0.84 ? 9.012 was obtained for the soybean trypsin inhibitorpretreatment (solid circles) group.
10
PARKS
Figure
ET AL.
GASTROENTEROLOGY
1. Morphologic
and
Morphometric
Changes
Mean mucosal thickness Group
n
(Km)
Normal Ischemia Ischemia-soybean trypsin inhibitor
14 7 7
1000.9 ? 8.8 547.2 ? 13.2" 982.9 + 128.5"
Values
89, No. 1
3. Illustration of mucosal injury in the typical animal of the ischemic and soybean trypsin inhibitor groups. Panel A. Nonischemic bowel showed no histopathologic abnormalities. Panel B. Ischemic bowel showed marked reduction in villus height, denudation of villus epithelium to the level of the crypts, and focal necrosis of crypt epithelium. Panel C. Soybean trypsin inhibitor-treated ischemic segments showed less severe changes with better preservation of villus height, preservation of epithelium on the lower portions of the villi, and no necrosis in crypt epithelium (H & E, all x 100).
capillaries to macromolecules, however, is dramatically increased (9). This conclusion is based upon postischemic estimates of the capillary osmotic reflection coefficient (Ed), a parameter that relates the degree of macromolecular restriction by the capillary wall to osmotically induced water movement. Reperfusion of the intestine after 1 h of ischemia significantly reduces the osmotic reflection coefficient for total proteins from a normal value of 0.92 to 0.58. This reduction is consistent with a dramatic increase in vascular permeability. A characteristic feature of longer (3 h) durations of Table
Vol.
are mean
? SE. a p < 0.05 vs. normal
ischemia is the development of histologically identifiable mucosal lesions (3). The morphologic changes in the mucosa produced by 3 h of ischemia in the present study are similar to those changes described previously (3). Generally, the lesions were characterized by massive epithelial lifting extending down the sides of the villi, sometimes completely denuded villi, and most frequently, disintegration of the lamina propria, hemorrhage, and ulceration. Thus, vascular permeability can serve as index of tissue injury for brief durations (1 h) of ischemia and morphologic alterations in the mucosal region of the small intes-
in the Small Intestine Produced
Mean villus height
Mean crypt depth
(pm)
(pm)
632.6 2 7.4 242.6 2 9.3" 581.9 + 25.8",b
(one-way
analysis
371.2 + 4.5 304.3 + 6.9" 403.9 + 7.9O.b
of variance).
by 3 Hours of lschemia
Villus epithelial necrosis (o-4+)
Crypt epithelial necrosis (o-4+)
0 3.9 2 O.l"." 1.6 f 0.4"."
’ p < 0.05 vs. ischemia
0 2.7 2 10.4",b 0.9 t 0.4"."
alone.
Muscularis propria injury (O-4f) 0 0 0
July
INTESTINAL
1985
tine can be used as an index for assessment of tissue injury produced by longer (3 h) durations of ischemia. Previous studies by the authors indicate that intravenous infusion of superoxide dismutase, a specific scavenger of superoxide anion (O,-), largely prevents (Up = 0.86) the increased intestinal vascular permeability (1) and mucosal lesions (3) produced by 1 and 3 h of ischemia, respectively. The major source of 02- in the ischemic small intestine is the enzyme xanthine oxidase (l-4), the first documented biologic source of superoxide anion (11). This conclusion is based largely upon the observation that allopurinol, a specific competitive inhibitor of xanthine oxidase, is as effective as superoxide dismutase in preventing the increase in vascular permeability (2) and mucosal lesions (3) produced by 1 and 3 h of intestinal ischemia, respectively. These findings are consistent with the observation that the intestinal mucosa is one of the richest sources of xanthine oxidase (6) and that the enzyme is most highly concentrated in the villus tip region (12), the portion of the intestine most sensitive to ischemic injury (8). There is now considerable evidence that xanthine oxidase exists in normal healthy cells as a nicotinamide adenine dinucleotide (oxidized form)-reducing dehydrogenase (type D) which can be converted to the superoxide-producing oxidase (type 0) by a variety of conditions (5,6), including oxidation of thiol groups (sulfhydryl oxidation) and proteolysis. Sulfhydryl oxidation-induced conversion of xanthine dehydrogenase to xanthine oxidase is reversible, i.e., conversion is readily reversed by thiol reductants such as dithioerythritol. Proteolytic dehydrogenase-to-oxidase conversion is irreversible, i.e., thiol reductants do not reverse the conversion. McCord and Roy (7) have reported that during ischemia in the small intestine, dehydrogenase-to-oxidase conversion is extremely rapid and irreversible; suggesting that the conversion is proteolytically mediated. Furthermore, they observed that intraperitoneal administration of ST1 (serine protease inhibitor) in the rat effectively prevents ischemia-induced conversion of xanthine dehydrogenase to xanthine oxidase in the intestine. In our studies, pretreatment with ST1 significantly attenuates the increase in vascular permeability induced by 1 h of ischemia as well as the morphologic alterations produced by 3 h of ischemia in the intestine. These results support the hypothesis that xanthine dehydrogenase is converted proteolytically to the oxygen radical-producing xanthine oxidase during ischemia in the intestine. Bounous and coworkers (13-16) have postulated that a circulatory deficiency in the small bowel,
ISCHEMIA
11
regardless of etiology, results in increased vulnerability of the intestinal mucosa to the digestive action of pancreatic proteases, most notably, trypsin and chymotrypsin. During low-flow states, there is a reduction in oxygen availability and an inability to maintain the concentration of high-energy phosphates in the tissues. Because biosynthesis of mucus is an energy-requiring process, depletion of tissue energy stores would largely prevent the production of mucus. The loss of the protective mucus coat could allow the proteases contained in thyme to damage the epithelial lining of the intestine (14). Support for this hypothesis is provided by the observation that inhibition of pancreatic proteases by aprotinin (Trasylol; FBA Pharmaceuticals, Inc., New York) largely prevents the formation of mucosal lesions, whereas intraluminal instillation of trypsin aggravates the damage induced by ischemia (15). The digestive action of the pancreatic enzymes is apparently due to the enzymes already present along the intestinal wall, as removal of the pancreas just before the onset of ischemia has no effect on the severity of the mucosal lesions (16). Pancreatectomy or pancreatic duct ligation 5 days or more before the animal is subjected to ischemia, however, greatly attenuates the mucosal lesions produced by ischemia (14). Although previous reports of beneficial effects of protease inhibitors in ischemic injury to the small bowel have been interpreted relative to protection against proteolytic digestion of the mucosal epitheliurn, these results are also consistent with the oxygen radical theory of ischemia-induced injury to the intestine. Under normal conditions, there is a small transmucosal (lumen-to-interstitium) flux of proteins the size of pancreatic proteases (14). With ischemia, the mucosal barrier becomes even less restrictive to the proteases (171, and a greater amount could gain access to the interstitium. These proteases could trigger the conversion of xanthine dehydrogenase to xanthine oxidase, thereby allowing production of oxygen radicals and subsequent tissue injury. In summary, the results of our study indicate that pretreatment with ST1 attenuates the increment in vascular permeability and the mucosal lesions associated with ischemia in the small intestine. These findings are consistent with the hypothesis that xanthine dehydrogenase is converted to xanthine oxidase by a protease, thereby allowing the production of oxygen radicals and subsequent tissue injury. References 1.
Granger DN, Rutili G, McCord JM. Superoxide radicals in feline intestinal ischemia. Gastroenterology 1981;81:22-9. 2. Parks DA, Granger DN. Ischemia-induced microvascular
12 PARKS ET AL.
3.
4.
5.
6.
7.
8.
changes: role of xanthine oxidase and hydroxyl radicals. Am J Physiol 1983;245:G285-9. Parks DA, Bulkley GB, Granger DN, Hamilton SR, McCord JM. Ischemic injury to the cat small intestine: role of superoxide radicals. Gastroenterology 1982;82:9-15. Schoenberg MH, Younes M, Muhl E, Haglund U, Sellin D, Schildberg FW. Free radical involvement in ischemic damage of the small intestine. In: Greenwald RA, Cohen G, eds. Oxy radicals and their scavenger system. Vol. II. Cellular and medical aspects. New York: Elsevier Biomedical, 1983:154-8. Battelli MG, Lorenzoni E, Stirpe F. Milk xanthine oxidase type D (dehydrogenase) and type 0 (oxidase): purification, interconversion and some properties. Biochem J 1973;131: 191-8. Stirpe F, Della Corte E. The regulation of rat liver xanthine oxidase: conversion in vitro of the enzyme activity from dehydrogenase (type D) to oxidase (type 0). J Biol Chem 1969;244:3855-63. McCord JM, Roy RS. The pathophysiology of superoxide: roles in inflammation and ischemia. Can J Physiol Pharmacol 1982;60:1346-52. Chiu CJ, McArdle AH, Brown R, Scott HJ, Gurd FN. Intestinal mucosal lesions in low flow states. I. A morphologic, hemodynamic and metabolic reappraisal. Arch Surg 1970;101:47883.
GASTROENTEROLOGY
Vol. 89, No. 1
9. Granger DN, Sennett M, McElearney P, Taylor AE. Effect of local arterial hypotension on cat intestinal capillary permeability. Gastroenterology 1980;78:474-80. 10. Granger DN, Taylor AE. Permeability of intestinal capillaries to endogenous macromolecules. Am J Physiol 1980;238: H457-64. 11. McCord JM, Fridovich I. The reduction of cytochrome c by milk xanthine oxidase. J Biol Chem 1968;243:5753-60. 12. Auscher C, Amory N, Pasquier C, Delbarre F. Localization of xanthine oxidase activity in hepatic tissue. A new histochemical method. Adv Exp Med Biol 1977;76:605-9. 13. Bounous G. Role of intestinal contents in the pathophysiology of acute intestinal ischemia. Am J Surg 1967;114:386-95. 14. Bounous G. Acute necrosis of the intestinal mucosa. Gastroenterology 1982;82:1457-67. 15. Bounous G, McArdle AH, Hodges D, Gurd FN. Biosynthesis of intestinal mucin in shock. Ann Surg 1966;164:13-22. 16. Bounous G, Menard D, DeMedicis E. Role of pancreatic proteases in the pathogenesis of ischemic enteropathy. Gastroenterology 1977;73:102-8. 17. Gregaard B, Parks D, Granger DN, McCord JM, Forsberg JO. Effects of ischemia and oxygen radicals on mucosal albumin clearance in intestine. Am J Physiol 1982;242:G448-54.
1985:89:6-12
Soybean Trypsin Inhibitor Attenuates Ischemic Injury to the Feline Small Intestine DALE A. PARKS, D. NEIL GRANGER, ARVIND K. SHAH
GREGORY
B. BULKLEY,
and
Department of Physiology, College of Medicine, University of Cincinnati, Cincinnati: Departments of Physiology and Mathematics and Statistics, University of South Alabama, Mobile, Alabama; and Department of Surgery, Johns Hopkins University, School of Medicine, Baltimore, Maryland
Recent evidence suggests that oxygen free radicals are largely responsible for the increased vascular permeability and early mucosal lesions associated with partial intestinal ischemia. It is postulated that oxygen radicals are produced by the reaction of the enzyme xanthine oxidase with hypoxanthine and molecular oxygen. In normal healthy cells, xanthine oxidase exists as a nicotinamide adenine dinucleotide-reducing dehydrogenase and not the oxygen radical-producing oxidase. In the intestine, dehydrogenase-to-oxidase conversion is nearly complete with Cl min of ischemia. Biochemical evidence from the intestine and liver indicate that ischemiainduced conversion of xanthine dehydrogenase to xanthine oxidase can be prevented by administration of protease inhibitors such as soybean trypsin inhibitor. In order to assess the role of proteases in oxygen radical-mediated ischemic injury to the small bowel, quantitative analyses of mucosal Jesion development and vascular permeability were performed in autoperfused segments of cat ileum subjected to 1 or 3 h of ischemia and pretreated with 15 mgkg (i.v.) soybean trypsin inhibitor. One hour of ischemia produced a significant increase in intestinal vascular permeability. The ischemia-induced increase in vascular permeability was significantly attenuated by soybean trypsin inhibitor pretreatment. Three hours of ischemia led to the developReceived May 4, 1984. Accepted December 20, 1984. Address requests for reprints to: Dale A. Parks, Ph.D., Department of Physiology, College of Medicine, University of Cincinnati, 231 Bethesda Avenue, Cincinnati, Ohio 45267-0576. This research was supported by a grant from the National Institutes of Health (AM 33594) and the American Heart Association (AM 31764). D. N. Granger is the recipient of a Research Career Development Award from the National Heart Lung and Blood Institute (HL 00816). 0 1985 by the American Gastroenterological Association 0016-5085/85/$3.30
ment of mucosal lesions in untreated animals. Pretreatment with soybean trypsin inhibitor largely prevented the development of the mucosal lesions. The findings of our study are consistent with biochemical evidence that, during ischemia, proteases trigger the conversion of xanthine dehydrogenase to xanthine oxidase and thereby lead to oxygen radical production and subsequent tissue injury. Major early consequences in the pathogenesis of intestinal &hernia include an increased vascular permeability and destruction of the mucosal membrane. Recent evidence from our laboratory suggests that oxygen-derived free radicals are largely responsible for the endothelial (1,~) and epithelial (3,4) damage associated with a partial arterial occlusion model of intestinal ischemia. Furthermore, the evidence indicates that oxygen radicals are produced by the reaction of the enzyme xanthine oxidase with hypoxanthine, an endogenous substrate produced by adenosine triphosphate catabolism, and molecular oxygen (l-3). Normal xanthine oxidase exists in the cell as a nicotinamide adenine dinucleotide (oxidized form)-reducing dehydrogenase (type D) and not the oxygen radical-producing oxidase (type 0). Conversion of xanthine dehydrogenase to xanthine oxidase can be induced by a number of conditions, including proteolysis and incubation under low oxygen tension in the presence of substrate (5). Addition of soybean trypsin inhibitor (STI) to rat liver homogenate prevents dehydrogenase-to-oxidase conversion proportional to the amount of ST1 added (6), whereas addition of any of several purified proteases induces dehydrogenase-to-oxidase conversion (7). McCord and Roy (7) have reported that, in the Abbreviation
used in this paper:
STI, soybean
trypsin
inhibitor.
INTESTINAL
July 1985
intestine, nearly complete dehydrogenase-to-oxidase conversion occurs with
of intestinal ischemia is the development of mucosal lesions. Therefore, morphologic alterations in the mucosal region of the small intestine were used as an index for assessment of tissue injury produced by longer (3 h) durations of ischemia. Materials
and Methods
Surgical Procedure The experimental preparation used in this study is similar, except for minor modifications, to that described in detail in previous studies (1,lO). Briefly, 42 cats previously fasted for 18-24 h were initially anesthetized with 50 mg/kg of ketamine-HCl. The right femoral artery and vein were cannulated and anesthesia was maintained by administration of sodium pentobarbital into the femoral vein. Systemic arterial pressure was measured with a Statham P23A transducer which was connected to a carotid artery cannula. A tracheotomy was performed to facilitate breathing and as a means of artificial ventilation if the cats failed to breathe spontaneously during the experiment. A midline abdominal incision was made and a 15-ZOcm segment of ileum was isolated, maintaining blood and lymph vessels intact. The remainder of the small and large intestine were extirpated. Immediately after isolation of the ileal segment and before cannulation of the lymphatic and blood vessels, heparin (5 mg/kg) was administered intravenously. Body temperature was maintained at 37°C with a thermistor-controlled infrared lamp. To minimize evaporation and tissue dehydration, all exposed tissues
ISCHEMIA
7
were moistened with saline-soaked gauze and placed in a plastic bag. In the vascular permeability studies, a large lymphatic vessel emerging from the mesenteric pedicle was cannulated and lymph flow was determined by observing the movement of lymph in a calibrated pipette (50~ZOO ~1, full scale). Lymph (C,) and plasma (Cp) total protein concentrations were measured with a refractometer [American Optical Corp., Buffalo, New York]. A large cannula was inserted into the superior mesenteric vein, and venous outflow drained into a reservoir mounted on a vertically positioned pulley system. Blood from the reservoir was returned to the animal via the cannula inserted into the femoral vein. Venous outflow pressure of the intestinal segment was set by adjusting the height of the reservoir and monitored from a T-connector in the venous circuit. In all experiments (vascular permeability and morphologic), an arterial circuit was established between the superior mesenteric and femoral arteries. Superior mesenteric arterial pressure was measured via a T-tube interposed within the arterial circuit using a Statham P23A transducer. All pressure cannulas and associated transducers were positioned at heart level. Heparinized blood from a donor animal was used to prime all the extracorporeal blood circuits. Systemic and superior mesenteric arterial pressures and venous pressure were continually recorded with a Grass physiologic recorder (Grass Instrument Co., Quincy, Mass.). At the end of each experiment the intestinal segment was weighed and lymph flow was normalized to milliliters per minute times 100 grams of used in the morphotissue. The experimental preparation logic studies was similar to that described above, except that the venous and lymphatic vessels were not cannulated.
Experimental
Protocols
In the vascular permeability studies, control values for intestinal lymph flow and lymph and plasma protein concentrations were obtained at a normal local arterial pressure and a venous pressure of 0 mmHg. Local arterial pressure was reduced to 25-35 mmHg using an adjustable clamp interposed within the local arterial circuit. Superior mesenteric arterial pressure was maintained at the reduced pressure for 60 min and then released. Forty-five minutes after reducing arterial pressure, additional lymph flow and lymph and plasma protein concentration measurements were made. After 60 min of local arterial hypotension, the arterial clamp was released and all measured parameters were allowed to reach a new steady state. Venous pressure was then elevated in lo-mmHg increments up to 40 mmHg. Intestinal venous pressure was maintained constant at each pressure level until all parameters were in a steady state, at which time lymph and plasma samples were acquired for protein concentration determination. In the morphologic studies, the animals (n = 14) were allowed to stabilize for 30 min after the operative procedure. Then, a “control” biopsy specimen was taken for histologic evaluation. Local arterial pressure was then reduced to 25-35 mmHg using an adjustable Gaskell
8
PARKS ET AL.
GASTROENTEROLOGY
clamp interposed within the arterial circuit. Superior mesenteric arterial pressure was maintained at the reduced level for 3 h and then released. The cats were then observed for 1 h after release of the arterial clamp, at which time an additional tissue biopsy specimen was taken for histologic examination. Tissue sections were always taken at the midpoint (r3 cm) of the intestinal loop. Soybean trypsin inhibitor (Sigma Chemical Co., St. Louis, MO.) was administered intravenously (15 mgikg) immediately before the onset of the 1 h (n = 7) or 3 h (n = 7) partial arterial occlusion. This dose of ST1 has been previously shown to prevent dehydrogenase-to-oxidase conversion in rat small intestine (7).
Osmotic
Reflection
The osmotic reflection coefficient (od) was estimated using the steady state relationship between the lymphto-plasma protein concentration ratio (C&p) and lymph flow in the postocclusion period. As lymph flow is increased, C&p rapidly decreases (filtration rate dependent) and then becomes relatively constant at a minimal value (filtration rate independent) when lymph flow is high. At a normal portal pressure, the exchange of macromolecules across the intestinal capillary wall occurs by both diffusion and convection. Elevation of venous pressure increases the convective movement of macromolecules across the capillary wall while at the same time the diffusive contribution to total exchange is reduced to a negligible level. Theoretical and experimental evidence suggests that ad = 1 - C&p, when C&p is filtration rate independent, i.e., when diffusive exchange is negligible (10). In the present study, fld for total plasma proteins was assessed using this approach. The theoretical basis for this approach is described in Reference 10.
Histologic
Technique
Tissue samples taken for histologic examination were -1 x 1.5 cm in size. After excision, each sample was rapidly mounted on a piece of Styrofoam with the mucosal surface exposed. The samples were immediately fixed in 10% neutral buffered formalin (Carson’s fixative). The tissues were dehydrated, cleared, embedded in paraffin, cut at about 6 pm, and stained with hemotoxylin and eosin.
Morphologic
and
Morphometric
affected; and 4+, epithelium of tips, midportions, and lower-portions of majority of villi affected. Crypt epithelial cell inflammation and necrosis were graded as follows: 1+, epithelium of occasional crypts affected; 2 +, scattered crypts affected; 3 +, many crypts affected; and 4 +, majority of crypts affected. Villus height, crypt depth, and mucosal thickness in each section were determined for all welloriented villi by measurement with an ocular micrometer. Each villus on each specimen was individually evaluated by the above criteria and the individual scores, as well as the mean score for each specimen, were recorded.
Statistical Analyses Vascular
Coefficients
Evaluation
Coded histopathoiogic sections were examined for ischemic injury, as evidenced by inflammation and necrosis of villus and crypt epithelial cells and destruction of the villus architecture of the mucosa, using a modification of a morphometric system we have previously described (3). This evaluation was made without knowledge of the treatment group from which the specimen came. Villus epithelial cell inflammation and necrosis were graded as follows: 0, no visible injury; 1 + ,epithelium of occasional villus tips affected: 2+, majority of tips affected; 3+, epithelium of majority of tips and midportion of some villi
Vol. 89, No. 1
Permeability
Studies
Two different statistical approaches were used. The first approach involved a one-way analysis of variance and Scheffe’s method of multiple comparison between the filtration-independent values of the lymph-to-plasma protein concentration ratio (CL/Cp) of the control group and of the groups subjected to 1 h local hypotension and local hypotension plus ST1 pretreatment. The relationship between C& and lymph flow for the various experimental protocols was compared by fitting a regression model for each condition and testing for coincidence of the regression curves. In terms of a linearized model, a relationship of the form In y = B, + B,x + B2xZ + Bgz + B,xz + BSxz + E,
where In y is the natural log of y and E is a random error term, is assumed between C&p (variable y) and lymph flow (variable x). A regression model was fit for each experimental protocol with the statistical analysis system package utilizing indicator (dummy) variables. After fitting the model, the null hypothesis (i.e., regression curves are coincident) was tested against the alternate hypothesis (i.e., regression curves are not coincident) and the observed significance level (p value) was calculated. Based on the p value, a decision was made as to whether the sample evidence was sufficient to refute the null hypothesis (p < 0.05). Morphologic
Studies
For these studies, each individual villus was considered to be the unit of observation and was scored using the criteria previously described. Each treatment group (i.e., control, ischemia, and STI) contained 89-213 units. A one-way analysis of variance and Scheffe’s method of multiple comparison was used to statistically compare the individual units of each experimental group. A significant difference between the groups was assumed when a level of significance (p value) was cO.05.
Results Figure 1 illustrates the relationship between the lymph-to-plasma protein concentration ratio (CL/ C,) and lymph flow for control conditions and after 1 h of local arterial hypotension. The curves repre-
July
INTESTINAL
1985
ISCHEMIA
9
sentative of the mean injury score obtained in the control group, the ischemia group, and the ischemia plus STI-pretreatment group. The morphologic and morphometric changes produced by ischemia are summarized for all groups in Table 1.The results in Table 1 clearly demonstrate that pretreatment with ST1 significantly diminished the severity of the mucosal lesions induced by 3 h of regional ischemia. The mean mucosal thickness and mean villus height were significantly greater in the ischemic preparations pretreated with ST1 compared with ischemia alone. In addition, ST1 significantly reduced the villus and crypt epithelial necrosis produced by ischemia. Both statistical analyses (regression model and one-way analysis of variance) predicted significant differences (p < 0.05) between the pretreatment group (STI) and ischemia alone for mucosal thickness, villus height, villus epithelial necrosis, and crypt epithelial necrosis. No injury to the muscularis propria or deeper bowel wall layers was present in any group.
l.OOf c
Discussion Lymph (ml/min
Figure
Flow x
1OOg)
1. Relationship between intestinal lymph-to-plasma total protein concentration ratio (C&p) and lymph flow under control conditions (open and closed circles) and after 1 h of ischemia (open and closed squares). The open symbols represent data from the present study and the closed symbols are data from previous studies (2,9). The capillary reflection coefficient (u(j) was derived assuming ad = 1 - C&e at lymph flows >0.40 mlimin X 100 g. Reflection coefficient values of 0.92 t 0.009 and 0.58 + 9.012 were obtained for the control and ischemia groups, respectively.
senting control conditions and ischemic conditions were acquired in previous studies (2,9) plus data from seven additional experiments. Using 1 - CL/Cp as an estimate of the osmotic reflection coefficient (ad) when CL/& is filtration rate independent, one would predict fld values of 0.92 2 0.009 for control conditions and 0.58 t 0.012 after ischemia. The ischemia group was significantly different (p < 0.01) from the control groups. The relationships between the lymph-to-plasma protein concentration ratio (CLICP) and lymph flow for control conditions (solid line), after local arterial hypotension (broken line), and after arterial hypotension with ST1 pretreatment [solid circles) are illustrated in Figure 2. One would predict a Od value of 0.84 +- 0.012 for ischemia plus STI. The STIpretreated group was significantly different (p < 0.01) from the ischemia alone and control groups. Figure 3 illustrates mucosal sections repre-
Oxygen-derived free radicals have recently been implicated in the vascular endothelial and mucosal epithelial damage associated with intestinal ischemia (l-3). Short durations of intestinal ischemia (i.e., 1 h) are associated with subtle morphologic changes such as mitochondrial swelling and moderate interstitial edema; the mucosal membrane remains intact (8). The permeability of intestinal
Lym~~h (mllmtn
Figure
Flow x 1009)
2. Effect of 15 mgikg soybean trypsin inhibitor on the relationship between the lymph-to-plasma protein concentration ratio and lymph flow after 1 h of ischemia. Solid line represents control curve, broken line represents ischemia alone. A reflection coefficient of 0.84 ? 9.012 was obtained for the soybean trypsin inhibitorpretreatment (solid circles) group.
10
PARKS
Figure
ET AL.
GASTROENTEROLOGY
1. Morphologic
and
Morphometric
Changes
Mean mucosal thickness Group
n
(Km)
Normal Ischemia Ischemia-soybean trypsin inhibitor
14 7 7
1000.9 ? 8.8 547.2 ? 13.2" 982.9 + 128.5"
Values
89, No. 1
3. Illustration of mucosal injury in the typical animal of the ischemic and soybean trypsin inhibitor groups. Panel A. Nonischemic bowel showed no histopathologic abnormalities. Panel B. Ischemic bowel showed marked reduction in villus height, denudation of villus epithelium to the level of the crypts, and focal necrosis of crypt epithelium. Panel C. Soybean trypsin inhibitor-treated ischemic segments showed less severe changes with better preservation of villus height, preservation of epithelium on the lower portions of the villi, and no necrosis in crypt epithelium (H & E, all x 100).
capillaries to macromolecules, however, is dramatically increased (9). This conclusion is based upon postischemic estimates of the capillary osmotic reflection coefficient (Ed), a parameter that relates the degree of macromolecular restriction by the capillary wall to osmotically induced water movement. Reperfusion of the intestine after 1 h of ischemia significantly reduces the osmotic reflection coefficient for total proteins from a normal value of 0.92 to 0.58. This reduction is consistent with a dramatic increase in vascular permeability. A characteristic feature of longer (3 h) durations of Table
Vol.
are mean
? SE. a p < 0.05 vs. normal
ischemia is the development of histologically identifiable mucosal lesions (3). The morphologic changes in the mucosa produced by 3 h of ischemia in the present study are similar to those changes described previously (3). Generally, the lesions were characterized by massive epithelial lifting extending down the sides of the villi, sometimes completely denuded villi, and most frequently, disintegration of the lamina propria, hemorrhage, and ulceration. Thus, vascular permeability can serve as index of tissue injury for brief durations (1 h) of ischemia and morphologic alterations in the mucosal region of the small intes-
in the Small Intestine Produced
Mean villus height
Mean crypt depth
(pm)
(pm)
632.6 2 7.4 242.6 2 9.3" 581.9 + 25.8",b
(one-way
analysis
371.2 + 4.5 304.3 + 6.9" 403.9 + 7.9O.b
of variance).
by 3 Hours of lschemia
Villus epithelial necrosis (o-4+)
Crypt epithelial necrosis (o-4+)
0 3.9 2 O.l"." 1.6 f 0.4"."
’ p < 0.05 vs. ischemia
0 2.7 2 10.4",b 0.9 t 0.4"."
alone.
Muscularis propria injury (O-4f) 0 0 0
July
INTESTINAL
1985
tine can be used as an index for assessment of tissue injury produced by longer (3 h) durations of ischemia. Previous studies by the authors indicate that intravenous infusion of superoxide dismutase, a specific scavenger of superoxide anion (O,-), largely prevents (Up = 0.86) the increased intestinal vascular permeability (1) and mucosal lesions (3) produced by 1 and 3 h of ischemia, respectively. The major source of 02- in the ischemic small intestine is the enzyme xanthine oxidase (l-4), the first documented biologic source of superoxide anion (11). This conclusion is based largely upon the observation that allopurinol, a specific competitive inhibitor of xanthine oxidase, is as effective as superoxide dismutase in preventing the increase in vascular permeability (2) and mucosal lesions (3) produced by 1 and 3 h of intestinal ischemia, respectively. These findings are consistent with the observation that the intestinal mucosa is one of the richest sources of xanthine oxidase (6) and that the enzyme is most highly concentrated in the villus tip region (12), the portion of the intestine most sensitive to ischemic injury (8). There is now considerable evidence that xanthine oxidase exists in normal healthy cells as a nicotinamide adenine dinucleotide (oxidized form)-reducing dehydrogenase (type D) which can be converted to the superoxide-producing oxidase (type 0) by a variety of conditions (5,6), including oxidation of thiol groups (sulfhydryl oxidation) and proteolysis. Sulfhydryl oxidation-induced conversion of xanthine dehydrogenase to xanthine oxidase is reversible, i.e., conversion is readily reversed by thiol reductants such as dithioerythritol. Proteolytic dehydrogenase-to-oxidase conversion is irreversible, i.e., thiol reductants do not reverse the conversion. McCord and Roy (7) have reported that during ischemia in the small intestine, dehydrogenase-to-oxidase conversion is extremely rapid and irreversible; suggesting that the conversion is proteolytically mediated. Furthermore, they observed that intraperitoneal administration of ST1 (serine protease inhibitor) in the rat effectively prevents ischemia-induced conversion of xanthine dehydrogenase to xanthine oxidase in the intestine. In our studies, pretreatment with ST1 significantly attenuates the increase in vascular permeability induced by 1 h of ischemia as well as the morphologic alterations produced by 3 h of ischemia in the intestine. These results support the hypothesis that xanthine dehydrogenase is converted proteolytically to the oxygen radical-producing xanthine oxidase during ischemia in the intestine. Bounous and coworkers (13-16) have postulated that a circulatory deficiency in the small bowel,
ISCHEMIA
11
regardless of etiology, results in increased vulnerability of the intestinal mucosa to the digestive action of pancreatic proteases, most notably, trypsin and chymotrypsin. During low-flow states, there is a reduction in oxygen availability and an inability to maintain the concentration of high-energy phosphates in the tissues. Because biosynthesis of mucus is an energy-requiring process, depletion of tissue energy stores would largely prevent the production of mucus. The loss of the protective mucus coat could allow the proteases contained in thyme to damage the epithelial lining of the intestine (14). Support for this hypothesis is provided by the observation that inhibition of pancreatic proteases by aprotinin (Trasylol; FBA Pharmaceuticals, Inc., New York) largely prevents the formation of mucosal lesions, whereas intraluminal instillation of trypsin aggravates the damage induced by ischemia (15). The digestive action of the pancreatic enzymes is apparently due to the enzymes already present along the intestinal wall, as removal of the pancreas just before the onset of ischemia has no effect on the severity of the mucosal lesions (16). Pancreatectomy or pancreatic duct ligation 5 days or more before the animal is subjected to ischemia, however, greatly attenuates the mucosal lesions produced by ischemia (14). Although previous reports of beneficial effects of protease inhibitors in ischemic injury to the small bowel have been interpreted relative to protection against proteolytic digestion of the mucosal epitheliurn, these results are also consistent with the oxygen radical theory of ischemia-induced injury to the intestine. Under normal conditions, there is a small transmucosal (lumen-to-interstitium) flux of proteins the size of pancreatic proteases (14). With ischemia, the mucosal barrier becomes even less restrictive to the proteases (171, and a greater amount could gain access to the interstitium. These proteases could trigger the conversion of xanthine dehydrogenase to xanthine oxidase, thereby allowing production of oxygen radicals and subsequent tissue injury. In summary, the results of our study indicate that pretreatment with ST1 attenuates the increment in vascular permeability and the mucosal lesions associated with ischemia in the small intestine. These findings are consistent with the hypothesis that xanthine dehydrogenase is converted to xanthine oxidase by a protease, thereby allowing the production of oxygen radicals and subsequent tissue injury. References 1.
Granger DN, Rutili G, McCord JM. Superoxide radicals in feline intestinal ischemia. Gastroenterology 1981;81:22-9. 2. Parks DA, Granger DN. Ischemia-induced microvascular
12 PARKS ET AL.
3.
4.
5.
6.
7.
8.
changes: role of xanthine oxidase and hydroxyl radicals. Am J Physiol 1983;245:G285-9. Parks DA, Bulkley GB, Granger DN, Hamilton SR, McCord JM. Ischemic injury to the cat small intestine: role of superoxide radicals. Gastroenterology 1982;82:9-15. Schoenberg MH, Younes M, Muhl E, Haglund U, Sellin D, Schildberg FW. Free radical involvement in ischemic damage of the small intestine. In: Greenwald RA, Cohen G, eds. Oxy radicals and their scavenger system. Vol. II. Cellular and medical aspects. New York: Elsevier Biomedical, 1983:154-8. Battelli MG, Lorenzoni E, Stirpe F. Milk xanthine oxidase type D (dehydrogenase) and type 0 (oxidase): purification, interconversion and some properties. Biochem J 1973;131: 191-8. Stirpe F, Della Corte E. The regulation of rat liver xanthine oxidase: conversion in vitro of the enzyme activity from dehydrogenase (type D) to oxidase (type 0). J Biol Chem 1969;244:3855-63. McCord JM, Roy RS. The pathophysiology of superoxide: roles in inflammation and ischemia. Can J Physiol Pharmacol 1982;60:1346-52. Chiu CJ, McArdle AH, Brown R, Scott HJ, Gurd FN. Intestinal mucosal lesions in low flow states. I. A morphologic, hemodynamic and metabolic reappraisal. Arch Surg 1970;101:47883.
GASTROENTEROLOGY
Vol. 89, No. 1
9. Granger DN, Sennett M, McElearney P, Taylor AE. Effect of local arterial hypotension on cat intestinal capillary permeability. Gastroenterology 1980;78:474-80. 10. Granger DN, Taylor AE. Permeability of intestinal capillaries to endogenous macromolecules. Am J Physiol 1980;238: H457-64. 11. McCord JM, Fridovich I. The reduction of cytochrome c by milk xanthine oxidase. J Biol Chem 1968;243:5753-60. 12. Auscher C, Amory N, Pasquier C, Delbarre F. Localization of xanthine oxidase activity in hepatic tissue. A new histochemical method. Adv Exp Med Biol 1977;76:605-9. 13. Bounous G. Role of intestinal contents in the pathophysiology of acute intestinal ischemia. Am J Surg 1967;114:386-95. 14. Bounous G. Acute necrosis of the intestinal mucosa. Gastroenterology 1982;82:1457-67. 15. Bounous G, McArdle AH, Hodges D, Gurd FN. Biosynthesis of intestinal mucin in shock. Ann Surg 1966;164:13-22. 16. Bounous G, Menard D, DeMedicis E. Role of pancreatic proteases in the pathogenesis of ischemic enteropathy. Gastroenterology 1977;73:102-8. 17. Gregaard B, Parks D, Granger DN, McCord JM, Forsberg JO. Effects of ischemia and oxygen radicals on mucosal albumin clearance in intestine. Am J Physiol 1982;242:G448-54.