Effect on biochemical markers of brain injury of therapy with deferoxamine or superoxide dismutase following cardiac arrest

Effect on Biochemical Markers of Brain Injury or Superoxide of Therapy With Deferoxamine Dismutase Following Cardiac Arrest BLAINE C. WHITE, MD,* NARSIMHA R. NAYINI, PtiD,t GARY S. KRAUSE, MD, MS,* STEVEN D. AUST, PtiD,t GARY G. MARCH, DO,* JAMES S. BICKNELL IV, MD,+ MARK GOOSMANN, MD5 iron-mediated lipid pemxidation by oxygen radical mechanisms is thought to be a contributingfactor to neurological injury during reperfusion following resuscitation fmm cardiac arrest. This studywas designed to examine and compare the effects of an imn chelator (deferoxamine) and superoxide dismutase (SOD) on brain lipid pemxldation and tissue ions after eight hours of reperfusion following a 15minute cardiac arrest. This sampling time was chosen because other work with this model has shown severe ionic and uitrastmcturai derangement at this point. Twentythree dogs were anesthetized with ketamine and haiothane and divided into four groups. Six dogs were nonischemic controls (group I). in the remaining dogs, a 15-minute cardiac arrest was induced with KCI. Resuscitation was begun with internal cardiac massage and artificial ventilation. After five minutes of artificial perfusion, internal defibrillation was performed to restart the heart. Ail dogs were resuscitated and supported by a standard intensive care (SIC) pmtacoi for eight hours. Six resuscitated dogs served as SIC controls (group ii). Six were treated with defemxamine, 200 me/kg loading dose and 100 mglkglh maintenance drip (group iii), and five were treated with SOD, l,OOO,OOOunits boius and 500,000 units/h drip (gmup IV). Ail drugs were administered intravenously immediately postresuscitation. At eight hours postresuscitation, a 3-g portion of parietai cerebral cortex was obtained through a trephine hole. The sample was assayed for tissue maiondiaidehyde (MDA) by the thiobarbituric acid test, the double bond content of the tissue lipids (lipid

From the ‘Section of Emergency Medicine, Wayne State University, Detroit; the tDepartment of Biochemistry, Michigan State University, East Lansing; the *Department of Emergency Medicine, Mt Carmel Mercy Hospital, Detroit, and the SDepartment of Emergency Medicine, Butterworth Hospital, Grand Rapids, Michigan. Supported in part by the US Army Medical Research and Development Command, Contract DAMD-84-17-4200. Manuscript received June 251987; revision accepted December 24, 1987. Address reprint requests to Dr. White: Section of Emergency Medicine, Wayne State University, Detroit Receiving Hospital, 4201 St Antoine, Detroit, Ml 48201. Key

Words:

Cardiac arrest, deferoxamine,

tase. 0 1988 by W.B. Saunders Company. 0735-6757/88/0605-0005$5.00/0

superoxide

dismu-

unsaturation index, LUSI), and total tissue content of K and Na. Defemxamine treatment inhibited pemxidation of brain lipids by both the MDA and LUSI assays, and protected tissue WNa ratio. SOD treatment inhibited only MDA. iron chelation therapy appears to be promising for the membrane component of reperfusion injury.(Am J Emerg Med 1988;8:589-575. 0 1988 by W.8. Saunders Company.)

We have suggested that the critical phase of brain injury as a consequence of complete ischemia may occur during reperfusion. 1,2In particular, iron-mediated lipid peroxidation by oxygen radical mechanisms is thought to play an important role. The initiation reaction, the rate-limiting step in lipid peroxidation, requires the presence of a transitional metal, such as iron. 3,4 Reduction of ferric iron (Fe+3) stored intracellularly in ferritin to the ferrous form (Fe+2) by superoxide (O,-) delocalizes the iron from ferritin.5 This pathway points to two approaches to preventing ironinitiated lipid peroxidation. The first is removing the superoxide stimulus for iron-initiated lipid peroxidation. The first is removing the superoxide stimulus for iron release from ferritin. The other is chelating the iron, thereby rendering it inactive. This study explores the effects of the enzyme superoxide dismutase (SOD), which converts superoxide to hydrogen peroxide and water,6 and the iron chelator deferoxamine (DEF)7 on lipid peroxidation and the membrane function of ion partition in the brain following resuscitation from a 15-minute cardiac arrest. METHODS Animal Model Twenty-three large mongrel dogs weighing 20 to 30 kg were screened for heartworm by microfllarial and immunofluorescence assays. They were anesthetized with ketamine, 7 mg/kg intravenous (IV), intubated, and placed on a volume ventilator using room air. Surgical anesthesia was maintained with halothane 1% to 569

AMERICAN

JOURNAL

OF EMERGENCY

MEDICINE

n Volume 6. Number 6 n November

2% inhalation. A pulmonary arterial catheter was placed by right jugular venous cutdown. Placement was guided by observing the pulse pressure waveform. An arterial catheter was placed by femoral artery cutdown. Both lines were connected by pressure transducers to a physiograph recorder for continuous pressure display. The ECG was monitored on a Lifepak 2 (PhysioControl Cot-p, Richmond, WA) equipped with internal defibrillation paddles. A urinary catheter was connected to gravity drainage. Supplemental oxygen (1 L/min) was used to maintain prearrest PaO, at 80 to 100 torr. The dogs were divided into four groups: (1) nonischemic controls (n = 6) in which the tissue samples were taken without an ischemic insult; (2) 15minute cardiac arrest, resuscitation, and eight hours of standard intensive care (SIC) without specific therapy directed at cerebral protection (n = 6); (3) U-minute cardiac arrest, resuscitation, deferoxamine treatment (200 mg/kg rapid IV infusion followed by 100 mg/kg/h IV drip*), and eight hours of SIC (n = 6); and (4) 15minute cardiac arrest, resuscitation, treatment with bovine SOD (EC. 1.15.1.1, DDI Pharmaceuticals, Mountain View, CA) (one million units IV bolus followed by 500,000 units/h IV drip) and 8 hours of SIC (n = 5). The commercially prepared SOD was pooled and assayed for activity (1,458,OOO units/ml) by the method of McCord and Fridovich.8 The pool was then divided into aliquots, which were kept frozen until immediately before administration. Activity of each aliquot was again verified by assay after it was used. All drugs were administered IV immediately postresuscitation. In those dogs subjected to a cardiac arrest, it was induced by injection of a bolus of 0.75 mEq/kg KC1 into the pulmonary artery catheter. The onset of cardiac arrest was confirmed by the ECG with arterial pressure monitors, and the arrest period was timed by stopwatch. Ventilation and anesthesia ceased with the onset of cardiac arrest. After 13 minutes of arrest, the chest was entered through the left lateral thoracotomy The dose of deferoxamine was chosen based on our previous studies of iron delocalization in the brain. After two hours of reperfusion we have found a total of 40 nmol/lOO mg tissue (wet weight) in the low molecular weight pool of iron in the brain. This translates to a concentration of 400 pmollkg. The molecular weight of deferoxamine is about 600. The dose of deferoxamine may then be calculated against the available LMWS pool as: 600 pglkg = 1 crmol/kg; therefore, 240 mg/kg = 400 kmol/kg, and estimating total body water at 75% a dose of 200 mg/kg and assuming initial general distribution through the body, we concluded that a dose of 200 mg/kg should be adequate to control the LMWS iron pool. The 100 mglkglh maintenance drip is based on a 1 hr half-life of the polypeptide in body fluids according to the manufacturer (Ciba Pharmaceutical). We did not notice hypotension in the treated group with our controlled infusions of this dose of the drug. l

570

1988

in the sixth interspace. Rib retractors were placed and opened sufficiently to perform internal massage. The pericardium was not opened. At the end of the 15-minute arrest period, internal cardiac massage (IC) was begun and mechanical ventilation resumed with 100% oxygen. Simultaneously, sodium bicarbonate (8 mEq/kg) and epinephrine (20 kg/kg) were given by bolus injection through the pulmonary arterial line, as we have previously discussed.30 This was followed by an epinephrine drip (5 Fg/kg/min) using a Harvard pump. After five minutes of IC, internal defibrillation was attempted at 1 J/kg and repeated if necessary. All animals resumed spontaneous perfusion, after which the epinephrine drip was discontinued. The ribs were approximated with umbilical tape and the chest was closed surgically in two layers with the incorporation of a thoracostomy tube with water-sealed vacuum drainage to ensure full expansion of the lungs. Standard intensive care postarrest consisted of maintenance of ventilation, hemodynamics, and urinary output. Arterial diastolic pressures were maintained above 90 mmHg with low-dose dopamine ( PCO, > 35 and 200 > POZ > 150 during the postresuscitation period. Pancuronium was administered immediately postresuscitation to block spontaneous respiratory effort during closure of the thoracotomy, and oxymorphone was administered thereafter as needed. Brain Tissue Sampling

Before obtaining the brain sample, anesthesia was reinstituted with 1% to 2% halothane. The scalp was surgically reflected and the brain exposed by removing a 15 mm diameter piece of the parietal skull with a surgical trephine. At the appropriate time, a specimen of the parietal cortex weighing about 3 g was obtained. This was immediately placed in Ringers lactate solution which had been deoxygenated by bubbling with argon gas and cooled to 0°C. The animals were then killed.

Chemical Analysis All reagents were analytical grade and were used without further purification. All solutions were made with distilled water which had been passed over a Chelex-100 column (Chelex water, Bio-Rad, Richmond, CA) to remove any trace iron. Analytical stock solutions were prepared using chemicals from the fol-

WHITE ET AL n DEFEROXAMINE OR SUPEROXIDE DISMUTASE THERAPY

lowing sources: EDTA (Mallinckrodt Inc, St. Louis), thiobarbituric acid (TBA; Sigma Chemical Company, St. Louis), butylated hydroxytoluene (BHT); Sigma), 0-phenanthroline (Sigma), trichloroacetic acid (TCA; Sigma), ascorbic acid solution (Sigma), and ammonium acetate (J. T. Baker Chemicals, Phillipsburg, NJ). Tissue KINa Ratio One gram of the tissue sample (weighed out using a Metler balance) was used for the K/Na analysis. The specimen was placed in a 50 mL volumetric flask, and 10 mL absolute nitric acid was added. This was gently heated until the tissue was fully dissolved and the solution was clear. The flask was then capped and allowed to cool overnight. The solution was diluted with triple distilled water to approximately 25 mL, and 500 p.L of yttrium standard for the atomic emission spectrometer was added. The solution was then diluted to exactly 50 mL. The content of Na and K were determined in triplicate using a Jarrel-Ashe 955 inductively coupled plasma atomic emission spectrometer with a detection limit of lop9 ppm. The value of the averaged analytical determinations was reported as pEq/g tissue wet weight. The K/Na ratios were calculated for each dog from this data. Thiobarbituric Acid-Reactive

Substances

Thiobarbituric acid-reactive substances (TBARS) assays were conducted by the method of Buege and Aust ** The TCA-TBA-HCl reagent was prepared in 0.25 N HCl with 15% w/v TCA and 0.375% w/v TBA. This solution was warmed to assist in dissolving the TBA. The tissue specimen obtained from the animal model described above was immediately rinsed in 4°C deoxygenated Chelex water and blotted. One gram of tissue was weighed out and homogenized in 10 mL of 1 mmol EDTA using a Potter-Elvehjam homogenizer. To 1 mL of the homogenate, 2 mL of the TCA-TBA-HCl stock solution and 60 ~.LLof 2% BHT in ethanol were added to prevent further lipid peroxidation. This suspension was thoroughly mixed and heated for 15 minutes in boiling water. After cooling, the flocculant precipitate was removed by centrifugation at 1,000 G for ten minutes. The absorbance of the supernatent was read at 535 nm against a blank containing all reagents minus the tissue homogenate. The TBARS concentration was expressed as nmol of malondialdehyde (MDA)/ 100 mg tissue using a standard curve. Lipid Unsaturation Index Brain lipids were extracted from the tissue sample by the method of Folch et al” and stored at - 20°C in a 2/l mixture of CHCl,/CH,OH. All solvents used in lipid isolation were purged of oxygen and stored with argon, and the solvents were kept at 4°C to prevent

autooxidation of lipids. The fatty acid composition of the brain phospholipids was determined by gas-liquid chromatographic analysis of the methyl esthers. Methyl esthers were prepared by transesterification of lipid with 10% H,SO, in methanol at 60°C for 24 hours. Chromatography was performed on a Varian model 3700 gas chromatograph equipped with a flame ionization detector. The glass column (6 ft x 112in outer diameter) was packed with 10% DEGS-PS on 80/100 Suplecoport. The instrument conditions were: (1) carrier gas flow rate, 30 mL/min; (2) temperature program of initial temperature at 150°C and final temperature of 200°C at 3”C/min, with the final temperature of 200°C at 3”C/min, with the final temperature held for 22 minutes. Areas under the fatty acid peaks were integrated by a 3390A Hewlett Packard (Hewlett Packard, Palo Alto, CA) integrator connected in line with the gas chromatograph. The fatty acid composition was standardized against palmitic acid in each sample by the method of Tien and Aust.12 This is done because palmitate is a saturated fatty acid and does not readily undergo degradation reactions such as lipid peroxidation. After the polyunsaturated fatty acids (PUFA) are standardized against the amount of palmitate in the sample, the quantity of each PUFA is multiplied by its number of double bonds. Addition of this product for the naturally occurring PUFAs , arachidonic , linolenic , linoleic, and docosahexenoic, is expressed as the lipid unsaturation index (LUSI). Thus, the units of LUSI are total PUFA double bonds standardized against the concentration of palmitate in the tissue lipid fraction.

Statistical Method The data were examined by the Wilks analytical method of multivariance (MANOVA) using the SPSS program on an IBM-AT computer (IBM, Boca Raton, FL). Each dependent variable data pool then was tested individually for significance by the univariant analysis of variance (ANOVA). When significance was found by these tests, the data was examined for significant patterns of difference between treatment groups by the Scheffe posthoc comparison method.

RESULTS Normal acid-base status and arterial PO2 and PC02 within the selected parameters were maintained in all dogs throughout the postresuscitation period. There were no significant differences in prearrest arterial blood gasses among all treatment groups, nor in postresuscitation arterial blood gasses among the groups of resuscitated animals. Further NaHCO, was not required in any dog. The biochemical data are displayed as mean values + one SD in Table 1. MANOVA analysis yielded F = 571

AMERICAN JOURNAL OF EMERGENCY MEDICINE W Volume 6, Number 6 W November 1988

TABLE 1.

Results

Group Nonischemic controls 8 hours standard care 8 hours with deferoxamine 8 hours with SOD

$0 TEARS’

LUSIT

WNa

7.15 * 1.48

417.0 t 64.5

1.35 -t 0.10

40.10 rt 2.54

286.7 + 55.7

0.47 ? 0.21

20.37 -c 8.07

436.2 ? 40.8

1.22 i 0.27

20.68 2 11.26

268.8 k 35.8

0.93 + 0.11

T

Data is shown as mean + SD. nm/lOO mg. t Lipid double bondsfpalmitate. l

11.60 and P < .OOl. Univariate parametric ANOVA yielded for MDA F = 19.5 and P < .OOOl;for LUSI, F = 18.9 and P < .OOOl;and for K/Na, F = 18.5 and P < .OOOl. results of Scheffe analysis of the

NIC

STD

DEF

SOD

GROUPS

1. Tissue malondialdehyde turic acid assay.

concentrations

by thiobarbi-

NS). There was no protection of LUSI with SOD (P< .Ol compared with NIC). The K/Na ratio with SOD therapy is intermediate between the SIC (P = NS)and DEF groups (P = NS),but is clearly reduced in comparison to nonischemic controls (P< .05). Deferoxamine treatment significantly reduced brain tissue TBARS as compared to the untreated SIC group (P< .Ol), although TBARS remained elevated above NIC (P < .Ol). DEF preserved LUSI (P < .Ol) and K/Na (P< .Ol) at preischemic levels. This data shows that membrane damage by an iron mediated process was responsible for the loss of tissue PUFAs and deterioration of the K/Na ratio. SOD significantly reduced TBARS as compared with the SIC group (P< .Ol), but TBARS remained above the NIC group (P< .Ol). DEF and SOD provided similar degrees of reduction in TBARS (P = TABLE 2.

Minimum

Significance

< .Ol < .Ol < .05 < .Ol NS < .a1

2

3

These experiments continue our work to characterize the sequence and mechanisms of brain injury following cardiac arrest and resuscitation. Previous experiments in our laboratory demonstrated that brain tissue metals (Ca, Na, and K) rapidly return to normal during the early reperfusion phase following a 15minute cardiac arrest.13 They remain normal after two and four hours of reperfusion. However, by eight hours of reperfusion, prominent shifts in the tissue concentration of Na, K, and Ca in the direction of

< .05 < .Ol

NS

-

< .Ol NS

C.01

-

450

-’

400

-.

560

-.

_

500

v) =

260

-1

200 I

-

160 100

< .Ol NS < .05

500

4

< .Ol NS

60

NS

-

P values are shown, comparing the various groups against one another within individual variables. ABBREVIATION: NS, not significant. 512

Cations

Levels Among Groups 1

TBARS (1) Nonischemic controls (2) 8 hours standard care (3) 8 hours deferoxamine (4) 8 hours SOD LUSI (1) Nonischemic controls (2) 8 hours standard care (3) 8 hours deferoxamine (4) 8 hours SOD WNa (1) Nonischemic controls (2) 8 hours standard care (3) 8 hours deferoxamine (4) 8 hours SOD

DISCUSSION

0

L MC

SrD

CARL

OEF

GROUPS FIGURE 2. mitatej.

Lipid unsaturation

index (lipid double bonds per pal-

WHITE ET AL n DEFEROXAMINE OR SUPEROXIDE DISMUTASE THERAPY

1.6

I

t

1.4 0 i=

1.2

d m

10 -

T Y

.a .6 .4 .2 0 NIC

STD

CARE

DLF

SOD

GROUPS

FIGURE 3.

Brain tissue K/Na ratio.

equilibration with the interstitial fluid have occurred.‘3 This phenomenon is not effected by the calcium antagonist flunarizine,‘4 and seems to involve a general loss of membrane integrity in terms of maintenance of normal cellular ionic gradients. The previous demonstration of these delayed cations shifts is reconfirmed by the K/Na ratios in the current experiments. Any proposed therapy for reperfusion brain injury must be able to prevent these cation shifts. DEF provided complete protection of the K/Na ratio. There was a trend toward protection of K/Na ratio with SOD treatment, but this did not reach statistical significance.

Lipid Peroxidation Oxygen radical species that attack biological molecules require a transition metal such as iron for their formation. 3*4During the first two hours of reperfusion, a large amount of brain tissue iron is delocalized into species weighing <30,000 daltons (low molecular weight species iron; LMWS iron). l5 Increased LMWS iron is not seen during ischemia. We believe iron delocalization is caused by the reduction of iron in ferritin by superoxide (O,-)’ formed in excess during reperfusion. There are several potential sources of O,- in brain tissue during postischemic reperfusion: (1) Mitochondria produce O,- during the first single electron reduction of oxygen in the oxidative phosphorylation pathway. (2) Superoxide is also produced by metabolizing arachidonate accumulated during ischemia16 via the cyclooxygenase pathway. 17,18Reperfusion results in a burst of cyclooxygenase activity, as measured by the formation of prostaglandin products. l9 (3) A further source of 02- during reperfusion is the reaction between xanthine oxidase and hypoxanthine.

Xanthine oxidase, present in brain capillary endothelium,” is produced by Ca2-mediated proteolysis of xanthine dehydrogenase during ischemia.21 Hypoxanthine accumulates during ischemia as a result of adenine nucleotide degradation.22 Two of these sources of 02- are intracellular in brain cells proper, and one is in the capillaries. While the superoxide anion appears to cross brain cellular membranes rather readily through an anion channel,” SOD (Mr 32,600) crosses membranes poorly23 and is virtually not cleared from plasma into the brain at a11.24Thus, it is likely that the limited protective effect of SOD seen in this study only in the TBARS parameter represents the effect of scavenging O,- only in the brain microvasculature. The loss of LUSI and K/Na gradient were not inhibited by SOD suggests that the more important sources of 02- during reperfusion are from the reactions occurring within the brain cells themselves. There are at least eight methods for measuring products of lipid peroxidation in in vivo experiments.25 Some of these, such as serum concentrations of lipid hydroperoxides and exhaled concentrations of hydrocarbons (ie, pentane), are not suitable for organspecific work. Others, such as measurement of Schiff base fluorescence, are limited by the same theoretical problems as determination of MDA by the thiobarbituric acid test. Three methods are widely used for demonstration of lipid peroxidation in tissue: (1) determination of the lipid peroxidation product MDA by the thiobarbituric acid method (TBARS); (2) determination of conjugated diene formation in the lipids; and (3) demonstration of a selective loss of PUFAs. We have used two of these three methods in this study. The thiobarbituric acid assay is useful, and is the most widely used of the assays for lipid peroxidation, although it is less specific than the demonstration of a selective loss of PUFAs. TBARS actually measures predominantly lipid peroxides in the sample that are converted to MDA during the acid heating portion of the assay.25 Other compounds that can produce a similar calorimetric product in the reaction with thiobarbituric acid include ethanal, biliverdin, glyoxal, and furfuraldehyde; however, none of these is likely to be significant in the in vivo system we have studied. It is of increasing interest to us that reactions between hydroxyl radical (OH*) and DNA also produce peroxidation products that form MDA during the acid heating portion of the thiobarbituric acid assay.26727 The thiobarbituric acid assay does not appear to detect products of normal enzymatic lipid metabolism.28 In summary, the TBARS assay measures peroxides derived from oxygen radical reactions with lipids and DNA. Thus, the assignment of TBARS to iron dependent lipid peroxidation rests on the simultaneous loss of PUFAs (LUSI) and the inhibition of both TBARS 573

AMERICAN

JOURNAL

OF EMERGENCY

MEDICINE

W Volume

6, Number

and LUSI changes by iron chelators such as deferoxamine. At two and four hours of reperfusion, the accumulation of lipid peroxidation products is inhibited by DEF.29*30 It was reasonable to propose that the tissue ionic gradient failure seen between four and eight hours of reperfusion13 reflected the consequences of membrane destruction by continuing progression of the lipid peroxidation reactions. In this study intervention with either DEF or SOD demonstrated significant protective effects during reperfusion on the accumulation of TBARS. This result argues that the accumulation of TBARS after eight hours of reperfusion results from reactions in which both O,- and iron are involved. Intervention with only DEF protected all three study parameters, and values for LUSI and K/Na were maintained at preischemic levels. These results represent strong evidence implicating O,- and iron dependent peroxidation of PUFAs in the pathochemistry of brain tissue injury in reperfusion following complete global ischemia.

Additional f&oxygenation Injury Mechanisms Despite the ability of DEF to preserve ionic homeostasis and prevent lipid peroxidation, it does not provide complete protection against reperfusion injury. Three aspects of our several studies on irondependent, oxygen-free radical reperfusion injury of the brain are disturbing. First, intervention with deferoxamine does not improve neurological outcome after 40 hours with this mode1.32 Second, there is a persistent elevation of TBARS observed in this study in spite of evidence of virtually complete protection of the polyunsaturated fatty acid pool. Finally, DEF did not prevent the accumulation of pyknotic neurons in the selectively vulnerable areas and the massive disorganization of nuclear chromatin during reperfusion.31 DEF did reduce the incidence of microhemorrhages to almost zero after eight hours of reperfusion. 31 It is interesting to note that Negovskii observed these microhemorrhages 30 years ago anddeveloped preliminary evidence that they were related to tissue reactions with oxygen.33 These observations may point to DNA damage by oxygen radicals. As we discussed above, MDA is formed not only by degradation of peroxidized lipids but also by degradation of DNA damaged by oxygen radical reactions. These two reactions appear to be the major sources of TBARS in tissue. Unlike lipid peroxidation reactions which progress relatively slowly in the membrane, the initial reactions between hydroxyl radical (OH*) and DNA may go to completion within a few minutes. 34 DNA damage by oxyg en radical mechanisms may be uncontrolled by DEF because of the relatively slow penetration of the blood brain barrier 574

6 n November

1988

by DEF, even after ischemia (Bo Hedlund, University of Minnesota, personal communication). In this context we have recently substantially expanded our hypothesis of the mechanisms of brain injury by ischemia and reperfusion to include potential DNA damage. 35 Briefly, mechanisms initiating injury during ischemia appear to include cytosolic Ca*+ accumulation,36 with consequent activation of cellular phospholipases, ‘6*37-40and conversion of xanthine dehydrogenase to xanthine oxidase,*’ which will produce superoxide (02-). It also now appears that Ca*+ may be involved in nuclease activation,41 with consequent generation of many single-stranded zones in nuclear DNA during ischemia.42 Nuclease reactions may be the biochemical correlate of the margination and clumping of nuclear chromatin that occurs in the brain during ischemia.43 O,- formed upon initiation of reperfusion will react with ferritin to reduce stored Fe3 + to soluble Fe*+ ,5 resulting in release of large quantities of brain tissue iron into low molecular weight forms. l5 Fe*+ readily forms complexes with polar organic species such as ADP, AMP, and DNA.3*4*44These complexes, in the presence of 02or H,O, (which is the dismutation product of 02-) form extremely avid oxidizing agents such as pet-ferry1 type species or OH*,3,4 which can not only degrade membrane lipids by lipid peroxidation but also produce DNA single-strand breaks by attack on the sugarPO4 backbone.45-50 This would convert single-strand DNA zones accumulated during ischemia to doublestrand breaks upon reperfusion. Double-strand breaks are not repaired” and are thought to be letha15* Major tissue detoxification mechanisms against oxygen radicals include superoxide dismutase, catalase, and glutathione peroxidase.3 Evidence suggests that all three of these defensive systems are limited in the nuclei of the selectively vulnerable zones of the brain during postischemic reperfusion.53-55 Both the extensive ultrastructural damage observed in chromatin43 and the failure of protein synthesis in the selectively vulnerable zones during brain reperfusions6 fit this biochemical hypothesis for DNA injury. The persistent TBARS elevations in the DEF treated group, observed in spite of complete protection of LUSI and the K/Na ratio, may arise from peroxidized DNA products formed before adequate tissue levels of DEF were achieved.

CONCLUSION Lipid peroxidation occurs in the brain during the first eight hours following a 15minute cardiac arrest and resuscitation. MDA levels are five times normal, and a 3 1% loss of total PUFAs has occurred after eight hours of reperfusion. This is accomplished by a shift of the tissue K/Na ratio toward equilibration with that of the interstitial fluid. Intervention with DEF protects

WHITE ET AL n DEFEROXAMINE OR SUPEROXIDE DISMUTASE THERAPY

PUFA content and the K/Na ratio at levels indistinguishable from those seen in nonischemic controls, but TBABS accumulation is only 33% inhibited. We conclude that iron-mediated lipid peroxidation is an important mechanism of reperfusion injury in the brain; however, DEF therapy following a 15minute cardiac arrest and resuscitation does not provide complete protection against reperfusion injury.

17. 18.

19.

The authors thank Dr Raywin Huang for statistical analysis of the biochemical data. We would also like to thank DDI Biochemicals, Mountain View, California for their generous donation of superoxide dismutase.

20.

REFERENCES

22.

1. White BC, Aust SD, Arfors KE, et al: Brain injury by ischemic anoxia: Hypothesis extension-A tale of two ions? Ann Emerg Med 1984;13:862-867 2. Krause GS, White BC, Aust SD, et al: Ischemia, resuscitation, and reperfusion: Mechanisms of tissue injury and prospects for protection. Am Heart J 1986;111:768-780 3. Halliwell B, Gutteridge J: Oxygen toxicity, oxygen radicals, transition metals, and disease. Biochem J 1984;219:1-4 4. Aust SD, Morehouse SA, Thomas CE: Role of metals in oxygen radical reactions. J Free Radicals Biol Med 1985;1:325 5. Thomas CE, Morehouse LA, Aust SD: Ferritin and superoxide dependent lipid peroxidation. J Biol Chem 1985; 260:3275-3280 6. Brawn K, Fridovich I: Superoxide radical and superoxide dismutases: Threat and defense. Acta Physiol Stand 1980;492:9-18 (suppl) 7. Keberle H: The biochemistry of desferioxamine and its relation to iron metabolism. Ann NY Acad Sci 1964;119:758768 8. McCord JM, Fridovich I: Superoxide dismutase. J Biol Chem 1969;244:6049-6055 9. Jelenko C, Solenberger RI, Wheeler ML, et al: Shock and resuscitation. Ill: Accurate refractrometric COP determination in hypovolemia treated with HALFD. J Am Coll Emerg Phys 1979;8:253-256 10. Beuge JA, Aust SD:/n Packer L, Fleischer S (eds): Methods in Enzymology and Biomembranes. San Diego, Academic, 1980, pp 302-310 11. Folch J, Lees M, Sloane-Stanley GH: A simple method for the isolation and purification of total lipids from animal tissue. J Biol Chem 1957;226:497-509 12. Tien M, Aust SD: Rabbit liver microsomal lipid peroxidation: The effects of lipid on the rate of peroxidation. Biochem Biophys Acta 1982;712:1-9 13. Hoehner TJ, Garritano AM, DiLorenzo RA, et al: Brain cortex calcium, magnesium, iron, sodium, and potassium following resuscitation from a 15 minute cardiac arrest in dogs. Am J Emerg Med 1987;5:19-23 14. White BC, Kumar K, Nayini NR, et al: Failure of flunarizine to protect the brain during reperfusion following a 15-min cardiac arrest: Evaluation by markers of membrane injury and quantitative morphometry. Abstract, Ann Emerg Med 1987;16:494 15. Krause GS, Joyce KM, Nayini NR, et al: Cardiac arrest and resuscitation: Brain iron delocalization during reperfusion. Ann Emerg Med 1985;14:1037-1043 16. Nemoto EM, Shin GK, Nemmer JP, et al: Free fatty acid ac-

21.

23.

24.

25.

26.

27.

28.

29.

30.

31.

32.

33. 34. 35.

36.

37.

38.

cumulation in the pathogenesis of cerebral ischemicanoxic injury. Am J Emerg Med 1983;1:175-179 Bakhle YS: Synthesis and catabolism of cycle-oxygenase products. Br Med Bull 1983;39:214-218 Kontos HA, Wei EP, Ellis EF, et al: Appearance of superoxide anion radical in cerebral extracellular space during increased prostaglandin synthesis in cats. Circ Res 1985; 57:142-151 Gaudet RJ, Levine L: Transient cerebral ischemia and brain prostaglandins. Biochem Biophys Res Comm 1979; 86:893-901 Betz AL: Identification of hypoxanthine transport and xanthine oxidase in brain capillaries. J Neurochem 1985; 44574-579 McCord JM: Oxygen-derived free radicals in post-ischemic tissue injury. N Engl J Med 1985;312:159-163 Saugstad OD, Schrader H: The determination of inosine and hypoxanthine in rat brain during normothermic and hypothermic anoxia. Acta Neurol Stand 1978;57:281-288 Michelson AM, Puget K: Cell penetration by exogenous superoxide dismutase. Acta Physiol Stand 1980;492:67-80 (SUPPI) Petkau A, Chelack WS, Kelly K, et al: Tissue distribution of bovine ‘251-superoxide dismutase in mice. Res Comm Chem Pathol Pharmacol 1976;15:641-654 Halliwell B, Gutteridge JMC: Lipid peroxidation: A radical chain reaction. In Oxygen Radicals in Biology and Medicine. Oxford, Oxford Scientific, 1985; pp 139-189 Halliwell B, Gutteridge JMC: Formation of a thiobarbituric acid reactive substance from deoxyribose in the presence of iron salts. FEBS LETT 1981;128:347-352 Gutteridge JMC, Toeg D: Iron-dependent free radical damage to DNA and deoxyribose. Separation of TBA-reactive intermediates. Int J Biochem 1982;14:891-893 Morisaki N, Lindsey JA, Stitts JM, et al: Fatty acid metabolism and cell proliferation. V. Evaluation of pathways for the generation of lipid peroxides. Lipids 1984;19:381-394 Babbs CF, Ralston SH, Badylak SF, et al: Evidence of preventable lipid peroxidation in brain and heart four hours after cardiac arrest and reper-fusion in rats. Am J Emerg Med 1986;4:414 Komara JS, Nayini NR, Bialick HA, et al: Brain iron delocalization and lipid peroxidation following cardiac arrest. Ann Emerg Med 1986;15:384-389 Kumar K, White BC, Krause GS, et al: A quantitative morphologic assessment of the effect of lidoflazine and deferoxamine therapy for global brain ischemia. Neurol Res, 1988;16:714-727 Krause GS, White BC, Evans AT, et al: Neurological outcome of histological evaluation of lidoflazine and deferoxamine therapy for post-arrest encephalopathy. Ann Emerg Med 1986;15:15 (abstr) Negovskii VA: Resuscitation and Artificial Hypothermia. New York, Consultants Bureau, 1962, pp 95-97 Ward JF: Molecular mechanisms of radiation induced damage to nucleic acids. Adv Radiat Biol 1977;5:181-239 Krause GS, White BC, Aust SD, et al: Brain cell death following ischemia and reperfusion: A proposed biochemical sequence. Crit Care Med, 1987 (in press) Harris RJ, Symon L, Branston NM, et al: Changes in extracellular calcium activity in cerebral ischemia. J Cereb Blood Flow Metab 1981 ;1:203-209 Drenth J, Enzing CM, Kalk KH, et al: Structure of porcine pancreatic phospholipase A2. Nature (London) 1976; 264~373-377 Van Den Bosch H: Intracellular phospholipases A. Biochem Biophys Acta 1980;604:191-246 575

AMERICAN JOURNAL OF EMERGENCY MEDICINE n Volume 6. Number 6 n November 1988

39. Edgar AD, Stroszanjder J, Horrocks LA: Activation of ethanolamine phospholipase A2 in brain during ischemia. J Neurochem 1982;39:111 l-l 116 40. Rehncrona S, Westerberg E, Akesson 6, et al: Brain cortical fatty acids and phospholipids during and following complete and severe incomplete ischemia. J Neurochem 1982;38:84-93 41. Tullis RH, Rubin H: Calcium protects DNAsel from proteinase K: A new method for the removal of spin-trapped radicals in aqueous solutions of pyrimidine nucleosides and nucleotides. Reactions of the hydroxyl radical. Int J Radiat Biol 1982;41:241-259 42. Warnick CT, Dierenfeldt SM, Lazarus HM: A description of the damaged DNA produced during tissue ischemia. Exp Mol Pathol 1984;41:397-408 43. Jenkins LW, Povlishock JT, Lewelt W, et al: The role of postischemic recirculation in the development of ischemic neuronal injury following complete cerebral ischemia. Acta Neuropathol (Berl) 1981;55:205-220 44. Floyd RA: DNA-ferrous iron catalyzed hydroxyl radical formation from H,O,. Biochem Biophys Res Comm 1981;99:1209-1215 45. Brawn K, Fridovich I: DNA strand scission by enzymatically generated oxygen radicals. Acta Biochem Biophys 1981;204:414-419 46. Leski SA, Lorentzen RJ, Tso PO: Role of O,- in DNA strand scission. Biochemistry 1980;19:3023-3028 47. Repine JE, Pfenninger OW, Talmage DW, et al: Dimethyl sulfoxide prevents DNA nicking mediated by ionizing radiation or iron/hydrogen peroxide-generated hydroxyl radical. Proc Natl Acad Sci 1981;78:1001-1003

576

48. Filho ACM, Meneghini R: In-vivo formation of single strand breaks in DNA by H,OP is mediated by the Haber-Weiss reaction. Biochem Biophys Acta 1984;781:56-63 49. Bradlen MO, Erikson LC: Comparison of the effects of H,OP and x-ray on toxicity, mutation, and DNA damage/repair in mammalian cells. Biochem Biophys Acta 1981;654:135141 50. Dizdaroglu M, Von Sonntag C, Schulte-Frohlinde D: DNA strand breaks and sugar release by gamma irradiation of DNA in aqueous solution. J Am Chem Sot 1975;97:22772278 51. Hutchinson KF: The repair of double strand breaks. In Okada S, lmanura M, Terasima T, et al (eds): Proceedings of the Sixth Congress of Radiation Research. Tokyo, Tappan, 1979, pp 4446-4453 52. Painter RB: The role of DNA damage and repair in cell killing induced by ionizing radiation. In Meyuen RE, Withers HR (eds): Radiation Biology in Cancer Research. New York, Raven, 1980, pp 59-68 53. White BC, Hildebrandt JF, Johns D, et al: Prolonged cardiac arrest and resuscitation in dogs: Brain mitochondrial function with different artificial perfusion methods. Ann Emerg Med 1985;14:383-389 54. Ushijima K, Miyazaki H, Morioka T: lmmunohistochemical localization of glutathione peroxidase in the brain of the rat. Resuscitation 1986;13:97-105 55. Hartz JW, Funkoshi S, Deutsch HF: The levels of SOD and catalase in human tissues as determined immunochemitally. Clin Chim Acta 1973;46:125-132 56. Hossman KA, Kleihues P: Reversibility of ischemic brain damage. Arch Neurol 1973;29:375-384