JOURNAL

OF SURGICAL

Limiting

RESEARCH

53,486-&g

(19%)

Oxygen Delivery Attenuates

Intestinal

Reperfusion

Injury’

ELIZABETH T. CLARK, M.D., AND BRUCE L. GEWERTZ, M.D. Department

of Surgery,

University

of Chicago, Submitted

5841 South

Maryland

Avenue,

for publication

March

14, 1991

Academic

Press,

Experimental model. The experimental preparation was a modification of that previously described by Anzueto et al. [3] and Kim and Gewertz [4] (see Fig. 1). Male Wistar rats weighing 400-600 g (retired breeders, Harlan Sprague-Dawley) were anesthetized by either inhalation augmented by intraperitoneal sodium pentothal (50 mg/ml at 0.1 ml/100 g). The trachea was intubated with PE 240 tubing (1.67-mm i.d., Clay-Adams) and supplemental oxygen was supplied at a rate of 4 liters/min to ensure hemoglobin saturation. Surgical re-

Inc.

Exposure of previously ischemic intestines to oxygen during reperfusion generates oxygen-derived free radisupported and NIH

by American Heart Association Grant Grant T32 HL07665-02 (Dr. Clark).

60637

METHODS

INTRODUCTION

1 Research (Dr. Gewertz)

Illinois

cals (OFRs) via the xanthine oxidase and neutrophilic NADPH oxidase system [l]. The importance of these oxidants is supported by a large body of experimental work which documents that pretreatment with OFR scavengers protects against ischemia-reperfusion injury (IR) [2]. These data also raise the possibility that simply limiting the availability of molecular oxygen during reperfusion of ischemic gut may be effective in reducing damage. Such a management strategy is especially attractive to clinicians since, in practice, treatment of mesenteric ischemia is not initiated until sometime after symptoms have developed and arterial insufficiency is established. Restriction of oxygen delivery may be accomplished by reperfusion with blood with decreased oxygen saturation (i.e., venous blood) or decreased hemoglobin concentration (i.e., hemodiluted blood). Both of these methods are clinically feasible, but must take into consideration that there are limits to the degree of oxygen reduction; severe decreases in reperfusate O2 content may actually prolong the hypoxic insult. The experiments described below tested the benefits of reduced 0, delivery in a well-characterized rat small intestinal preparation. Animals were subjected to 30 min of complete intestinal ischemia. This was followed by 1 hr of reperfusion with one of four solutions: arterial blood (A), venous blood (V), hemodiluted arterial blood at high flow (HHD), and hemodiluted arterial blood at normal flow (NHD). Flow rates and 0, contents were adjusted to produce either high (A or HHD - 12 ml O,/min/lOO g) or low (V or NHD -8 ml O,/min/lOO g) levels of 0, delivery.

Since free radical-mediated injury is dependent on the reintroduction of oxygen into ischemic tissues, restriction of oxygen content in the initial reperfusate has therapeutic potential. The degree to which oxygen must be restricted is crucial since hypoxic injury would continue if reperfusion 0, delivery remained below the ischemic threshold of the tissue. We examined this treatment strategy in 20 pump-perfused intestinal preparations subjected to 30 min of flow interruption. The oxygen content of the reperfusate was varied by utilizing arterial (A) or venous (V) blood; as a further modification, we also performed experiments in which hemodiluted arterial blood (HD) was the reperfusate at normal (NHD) and high (HHD) flow rates. The flow rates and Oz contents of the reperfusates were adjusted to produce either high (-12 ml O,/min/lOO g) or low (-8 ml O,/min/lOO g) levels of Oz delivery. Histologic sections, obtained after ischemia and after 1 hr of reperfusion, were blindly evaluated for mucosal injury (1 = normal to 5 = severe injury). Immediately after 30 min of &hernia, all groups had comparable histologic grades (A 2.0 + 0.3, V 1.8 f 0.3, NHD 1.6 -t 0.3, HHD 2.3 + 0.3). One hour after reperfusion, intestines reperfused with blood with high 0% content and hence high Oz delivery showed significantly more damage (P < 0.001) than those with exposed to low Oz delivery during reperfusion: A 3.9 + 0.5 and HHD 4.4 + 0.4 versus V 2.7 f 0.5 and NHD 2.9 + 0.3. While it is accepted that reintroduction of Oz during reperfusion is the crucial step in the genesis of reperfusion injury, it is remarkable that such modest reductions in total Oz delivery (approximately 30%) are effective in decreasing mucosal injury. In our model of intestinal ischemia, this degree of 0, restriction does not result in significant depression of intestinal 0, consumption or continued hypoxic injury and therefore has therapeutic potential. 0 1992

Box 129, Chicago,

88-786

485 All

Copyright 0 1992 rights of reproduction

0022-4804/92 $4.00 by Academic Press. Inc. in any form reserved.

486

JOURNAL

OF

SURGICAL

RESEARCH:

VOL.

53, NO.

5, NOVEMBER

1992

TRANSDUCERS

MESENTERIC

RESERVOIR

WATER

FIG. pumped

1. The through

BATH

model used in our laboratory is a denervated, isolated pump-perfused the intestine via the superior mesenteric artery at rates calculated

moval of the duodenum, pancreas, and large intestine was followed by bilateral ligation of the renal arteries and veins. Periadventitial stripping of the superior mesenteric artery (SMA) was performed to denervate the preparation. The small intestine was wrapped in salinesoaked gauze and cellophane to minimize evaporative losses. Temperature was maintained at 37°C with an intermittent, thermistor-controlled incandescent lamp. Direct cannulation of the right femoral artery allowed continuous monitoring of systemic arterial pressure. An isolated pump-perfused midgut segment was created. Following systemic heparinization (100 U sodium heparin/lOO g body wt), an extracorporeal circuit was constructed by catheterization of the left internal carotid artery (A, NHD, HHD) or left femoral vein (V). Hemodilution was performed using Hespan, a colloid solution that approximates the oncotic pressure of plasma. In hemodiluted preparations, WBC counts were assumed to parallel the decrease in hemoglobin seen with this perturbation (approximately one-half normal). Blood was routed through an infusion pump (Gilco minipuls II) and into the superior mesenteric artery at either 70-80 ml/min/lOO g (normal flow) or 110-120 ml/min/lOO g (high flow). Superior mesenteric venous effluent was collected via cannulation (PE 190 tubing, 1.19-mm i.d.,

rat small to provide

intestinal normal

preparation. Various reperfusates or high intestinal blood flow.

are

Clay-Adams) and drained into a reservoir. All extracorporeal blood was then pumped through a 37°C water bath and returned to the rat via the right femoral vein. Protocol. Preparations (n = 20; 5 animals per group) were allowed to stabilize for at least 30 min following manipulation. Intestines were rendered ischemic for 30 min by complete flow interruption (cessation of pump perfusion). Prior to reperfusion, a full-thickness biopsy of the intestine was obtained, weighed, and fixed in 10% neutral buffered formalin. Blood flow was then initiated with one of the four perfusion solutions. Intestinal blood flow was pump controlled at either normal (A, V, NHD) or high (HHD) flow. Each of the perfusion solutions was assayed for hemoglobin content, oxygen saturation, and partial pressure of oxygen determinations. 0, contents and total 0, deliveries were calculated using standard formulas (see Table 1). Following experiments, rats were sacrificed by exsanguination. Histologic evaluation of intestinal biopsies was performed using a modification of the Chiu system in a blinded manner with 1 = normal and 5 = severe injury [5]. Both villous height and morphology are considered. Using blinded evaluations, this method has been quite reproducible in our laboratory and represents a continuous function at half-grade intervals (see Fig. 2).

CLARK

AND

GEWERTZ:

LIMITING

TABLE Character

OXYGEN

1

of the Four

A V NHD HHD

11.4 12.7 7.6 7.4

(ml/min

Wdl) +f f f

0.8 1.2 0.3 0.2

76.3 77.4 79.9 114.5

a Hgb, hemoglobin content; IBF, intestinal blood flow. b Oxygen content and delivery were calculated using standard and HHD groups had higher oxygen delivery than V and NHD

Reperfusates 0, content* (ml O,/lOO ml)

IBF Hgb”

- 100 g)

15.5 10.5 10.4 10.2

k 4.9 f 7.7 f 11.1 f 8.1

formulas. groups.

487

DELIVERY

Due to variation

Data analysis. All values are reported as means + SD. Intestinal blood flow was normalized to 100-g gut weight stripped of mesentery. Statistical analysis was performed with Student’s t test utilizing the Bonferroni correction as appropriate. Significance was attained at P < 0.008.

+ ” + +

0, delivery (ml OJmin * 100 g)

1.2 1.8 0.6 0.3

11.8 8.1 8.3 11.7

in hemoglobin

saturation,

content,

k f +f

0.5 1.5 1.6 0.9

and blood

flow, A

DISCUSSION

Ischemic injury to an organ results from both hypoxia during flow interruption and deleterious effects of reperfusion. The initial hypoxic injury reflects the minimal 0, uptake required to maintain cellular integrity and the duration and degree of ischemia. In contrast, reperfusion injury may occur following even relatively brief peRESULTS riods of flow interruption. The susceptibility of tissue to reperfusion injury is quite variable and is thought to reflect, at least in part, the generation of oxygen-derived Immediately following ischemia and prior to reperfusion, there were no differences in histologic grades (see free radicals [21. In 1968, McCord and Fridovich documented that the Fig. 3; A 2.0 f 0.3, V 1.8 f 0.3, NHD 1.6 + 0.3; HHD 2.3 f 0.3). One hour after reperfusion, intestines provided biological enzyme system of xanthine oxidase generated active molecules with unpaired electrons in outer orwith “high” oxygen delivery (A and HHD) showed signifbitals (free radicals) [6]. Using histochemistry, high icantly more histologic damage than those with “low” oxygen delivery (V and NHD) (see Fig. 4; A 3.9 + 0.5 and concentrations of xanthine oxidase were localized in the rat small intestine [7]. Batelli et al. documented the exHHD4.4+0.4versusV2.7fOJiandNHD2.9?0.3,P< istence of two interconvertible forms of the enzyme [8]. 0.008).

Grade 1

Grade 2

Grade 3

Grade 4

Grade 5

FIG. 2. Histologic evaluation on biopsied sections was performed using a grading scale modified from that developed by Chiu and others. Grade 1 represents normal villous morphology. Grade 2 shows evidence of necrosis at the villous tip with progression to the midportion of the villous in grade 3 and the lower one-third of the villous in grade 4. Complete necrosis and sloughing of the entire villous are seen in grade 5. Increments of 0.5 represent intermediary damage between groups.

488

JOURNAL

A

V

OF

NHD

SURGICAL

RESEARCH:

HHD

30 MIN. ISCHEMIA FIG. 3. NHD,

HHD)

No differences were seen in experimental groups (A, V. following 30 min of ischemia (prior to reperfusion).

The predominant form, xanthine dehydrogenase, transfers electrons from various substrates specifically to nicotinamide adenine dinucleotide (NAD). Ischemia converts this enzyme to xanthine oxidase, a lessdiscriminant enzyme capable of transfer of unpaired electrons to oxygen. In the rat intestine nearly complete conversion of dehydrogenase to oxidase takes place within 1 min of induction of ischemia [2]. Free radicals have been documented to cause cell death by the related mechanisms of lipid peroxidation [9], membrane injury [lo], increased microvascular permeability and leukocyte adhesion [ 11, 121, and degradation of hyaluronic acid [ 111. Grogaard et al. placed the free radical-generating system of xanthine oxidase and hypoxanthine in the lumen of the small intestine and produced a histologic injury equivalent to that of 1 hr of ischemia followed by reperfusion [13]. In subsequent experiments, protection against reperfusion injury with free radical scavengers or xanthine oxidase inhibitors has been consistently documented regardless of the method utilized to demonstrate injury [14-171. It is also well known that leukocytes, the final mediators of most inflammatory responses, are attracted by and produce free radicals [18]. Hernandez et al. demonstrated decreased injury when ischemic intestine was reperfused with leukopenic blood [ 191. Further work in their laboratory utilized a monoclonal antibody against a membrane-associated glycoprotein which regulates the adherence of neutrophils (PMNs) to endothelium. Inhibition of PMN adherence was as effective as leukocyte depletion in preventing reperfusion injury. Although it is generally accepted that the bulk of mucosal injury occurs with reperfusion, the substrate most responsible for IR damage has been debated. Hypoxanthine increases tenfold during ischemia and is necessary for OFR production via the reaction hypoxanthine + 0,

VOL.

53, NO.

5, NOVEMBER

1992

+ OH * + H,O, [20]. However, Mousson et al. have shown conclusively that hypoxanthine is not the limiting factor in OFR formation [21]. Even the normal nonischemic level of the substrate (20-40 PM) exceeds the K,,, for hypoxanthine in the intestinal xanthine oxidase reaction (11 PM). There are more indications that oxygen is the limiting substrate. This evidence includes the timing of the histologic injury, the neglible levels of molecular oxygen during ischemia (much less than the required 20 mm Hg Km for 0,) and the dramatic increase in tissue oxygen tension with reperfusion (greater than 70 mm Hg). Korthuis et al. showed that hypoxic reperfusion of an isolated skeletal muscle preparation (using 95% N,/5% CO, to achieve P,O, 3-5 mm Hg) significantly reduced postischemic increases in both vascular resistance and vascular permeability [22]. In a longer-term hindlimb preparation, Walker et al. showed that gradual reintroduction of oxygen over a 1-hr period lessened the muscle necrosis observed after 5 hr of ischemia and 48 hr of reperfusion [23]. Similar observations have been made in a feline stomach preparation in which reperfusion with venous blood (P,O, 34 mm Hg) lessened injury [24]. Our results extend this previous work to the small intestine and show that even modest reductions in total oxygen delivery, to approximately 70% of that associated with arterial reperfusion, attenuate the early mucosal injury following 30 min of ischemia and 1 hr of reperfusion. Based on earlier experiments, reduction in 0, delivery to this level does not decrease oxygen consumption below 2 ml/min/lOO g or result in histologic injury [25]. Furthermore, although neutrophils are undoubtedly involved in IR, the beneficial effect of 0, re-

A

V

NHD

HHD

1 HR. REPERFUSION

FIG. 4. histologic reperfusion increased

At 1 hr of reperfusion there were significant differences in grades. Intestines with restricted oxygen delivery during (V, NHD) suffered less injury as compared to those with oxygen delivery (A, HHD).

CLARK

AND

GEWERTZ:

LIMITING

duction appeared to be independent of leukocyte concentration as WBC counts in hemodiluted preparations were roughly one-half normal. Our observations were consistent whether the reperfusate was venous blood with normal leukocyte concentration or hemodiluted blood with reduced leukocyte concentration. Due to the impracticality of direct measurements of OFR generation, the specific mechanism of this protective effect cannot be definitively determined. Nonetheless, these data would support the strategy of decreasing the initial oxygen delivery to previously ischemic intestine as a means to attenuate reperfusion phenomena. The use of hypoxic or hemodiluted blood in the initial stages of flow reestablishment is not overly complex and should be considered in future clinical efforts to lower the high morbidity and mortality of mesenteric ischemic syndromes.

OXYGEN

9. Kellog,

E. W., and Fridovich, I. Superoxide, hydrogen and singlet oxygen in lipid peroxidation by a xanthine system. J. Biol. Chem. 250: 8812, 1975.

10.

11.

assistance of Eileen of this manuscript.

1.

Parks, D. A., and Granger, D. N. Contributions reperfusion to mucosal lesion formation. Am. (Gustrointest Liver Physiol. 13): G749, 1986.

of ischemia J. Physiol.

and 250

Grogaard, B., Parks, D. A., Granger, D. N., McCord, J. N., and Folsberg, J. 0. Effects of ischemia and oxygen radicals on mucosal albumin clearance in intestine. Am. J. Physiol. 242(5): G448, 1982. Granger, D. N., Rutili, G., and McCord, J. N. Superoxide radicals in feline intestinal ischemia. Gastroenterology 81: 22, 1981. Parks, D. A., Bulkley, G. B., Granger, D. N., Hamilton, S. R., and McCord, J. N. Ischemic injury in the cat small intestine: Role of superoxide radicals. Gastroenterology 82: 9,1982.

14.

16.

D. A. Ischemia-reradicals. Actu Phys-

19.

3.

N. A rat model for Physiol. 246: G56,

20.

4. Kim, talis

E. S., and Gewertz, B. L. Chronic administration alters mesenteric vascular reactivity. J. Vast. 1987.

382, 5. Chiu, C., McArdle, Intestinal

mucosal

A. H., Brown, R., Scot, H. J., and Gurd, lesion in low-flow states. Arch. Surg.

478,197O. 6. McCord, J. N., and Fridovich, 7. 8.

of digiSurg. 5(2): F. N.

101:

I. The reduction of cytochrome c by milk xanthine oxidase. J. Biol. Chem. 243(21): 5753, 1968. Sackler, M. L. Xanthine oxidase from liver and duodenum of the rat: Histochemical localization and electrophoretic heterogeneity. J. Histochem. Biochem. 14(4): 326, 1966. Batelli, M. G., Lorenzoni, E., and Strippe, F. Milk xanthine oxidase type d (dehydrogenase) and type o (ox&se). Biochem. J. 131: 191,1973.

Parks, changes: Physiol. Perry, Granger, sions in

D. A., and Granger, D. N. Ischemia-induced vascular Role of xanthine oxidase and hydroxyl radicals. Am. J. 245(8): G285, 1983. M. A., Wadhwa, S., Parks, D. A., Pickard, W., and D. N. Role of oxygen radicals in ischemia-induced lethe cat stomach. Gastroenterology 90: 362, 1986.

18. McCord,

2. Granger,

D. N., Hollwarth, M. E., and Parks, perfusion injury: Role of oxygen-derived free iol. &and. Suppl. 548: 47, 1986. Anzueto, A., Benoit, J. N., and Granger, D. studying the intestinal circulation. Am. J. 1984.

Kellog, E. W., and Fridovich, I. Liposome oxidation and erythrocyte lysis by enzymatically generated superoxide and hydrogen peroxide. J. Biol. Chem. 252: 6721, 1977. DelMaestro, R., Thaw, H. H., Bjork, J., Planker, M., and Arfors, K. E. Free radicals as mediators of tissue injury. Acta Physiol. Stand. 492(Suppl.): 43,198O.

13.

17.

REFERENCES

peroxide oxidase

DelMaestro, R. F., Bjork, J., and Arfors, K. E. Increase in microvascular permeability induced by enzymatically generated free radicals. II. Role of superoxide anion radical, hydrogen peroxide and hydroxyl radical. Microvasc. Res. 22: 255, 1981.

15.

authors gratefully acknowledge the expert and Deborah Pieczynski in the preparation

489

12.

ACKNOWLEDGMENTS The Wayte

DELIVERY

J. N., and Wong, K. Phagocyte-Produced Free Radicals: Roles in Cyto-toxicity and Znflammation. New York: Excerpta Medica, 1979. Ciba Found. Series 65, pp. 343-360. Hernandez, L. A., Grisham, M. B., Twohig, B., Arfors, K. E., and Granger, D. N. Role of neutrophils in ischemia-reperfusion injury. Am. J. Physiol. 253: H699,1987. Parks, D. A., and Granger, D. N. Xanthine oxidase: Biochemistry, distribution and physiology. Acta Physiol. Stand. Suppl.

548:97,1986. 21. Mousson, B., Desjacques, 22.

P., and Baltasatt, P. Measurement of xanthine oxidase activity in some human tissues. Enzyme 29: 32, 1983. Korthuis, R. J., Smith, J. K., and Carden, D. L. Hypoxic reperfusion attenuates postischemic microvascular injury. Am. J. Physiol. 256 (Heart Circ. Physiol. 25): H315, 1989.

23. Walker, 24. 25.

P. M., Lindsay, T. F., Labbe, R., Mickle, D. A., and Romaschin, A. D. Salvage of skeletal muscle with free radical scavengers. J. Vast. Surg. 5: 68, 1987. Perry, M. A., and Wadhwa, S. S. Gradual reintroduction of oxygen reduces reperfusion injury in cat stomach. Am. J. Physiol. 254 (Gastrointest. Liver Physiol. 17): G366, 1988. Clark, E. T., and Gewertz, B. L.: Glucagon potentiates intestinal reperfusion injury. J. Vusc. Surg. 2: 270, 1990.