Oxygen radicals
in cerebral ischemia
CHARLES W. NELSON, ENOCH P. WEI, JOHN T. POVLISHOCK, HERMES A. KONTOS, AND MICHAEL A. MOSKOWITZ Departments of Medicine and Anatomy, Medical College of Virginia, Virginia Commonwealth University, Richmond, Virginia 23298; and Neurosurgery and Neurology Services, Massachusetts General Hospital, Harvard Medical School, Boston, Massachusetts 02114 Nelson, Charles W., Enoch P. Wei, John T. Povlishock, Hermes A. Kontos, and Michael A. Moskowitz. Oxygen radicals in cerebral ischemia. Am. J. Physiol. 263 (Heart Circ. Physiol. 32): H1356-H1362, 1992.-Superoxide production was measured as the superoxide dismutase (SOD)-inhibitable portion of nitro blue tetrazolium (NBT) reduction after cerebral ischemia-reperfusion in anesthetized cats equipped with cranial windows. Significant superoxide production was found in the early reperfusion period and continued for more than 1 h after ischemia. Superoxide was not detected in control animals not subjected to ischemia, during ischemia, and at 120 min of reperfusion. After ischemia, the vasoconstrictor response to arterial hypocapnia was reduced. This effect was prevented by pretreatment with SOD plus catalase or by deferoxamine. The response to topical acetylcholine was converted to vasoconstriction after ischemia. The normal vasodilator response reappeared spontaneously at 120 min of reperfusion. The vasodilator response to acetylcholine was preserved in animals pretreated with SOD plus catalase. Blood-brain barrier permeability to labeled albumin and horseradish peroxidase was increased after ischemia. These effects were minimized by pretreatment with SOD and catalase. We conclude that superoxide generation occurs during reperfusion after cerebral ischemia for a fairly long period and that superoxide and its derivatives are responsible at least in part for the vasodilation and the abnormal reactivity as well as for the increase in blood-brain barrier permeability to macromolecules seen after ischemia. Furthermore, the findings suggest that the agent responsible for the vascular abnormalities is hydroxyl radical generated via the iron-catalyzed HaberWeiss reaction. cerebral microcirculation; hydroxyl radical; blood-brain barrier; endothelium-dependent vasodilation OF univalent reduction of oxygen, superoxide, hydrogen peroxide, and hydroxyl radical are very reactive and capable of inducing tissue injury (6). They have been suggested as possible mediators of the tissue injury in ischemia-reperfusion in many tissues, including the brain. The generation of reactive oxygen species and their role in tissue injury in ischemia-reperfusion of the brain have received experimental support from a variety of sources. These include direct demonstration of superoxide production by showing the presence of superoxide dismutase (SOD) -inhibitable reduction of nitro blue tetrazolium (NBT) in the early part of reperfusion after ischemia in newborn pigs (1) and the finding of increased hydrogen peroxide production in gerbils subjected to brain ischemia (17). The latter authors also showed that after inactivation of xanthine oxidase the increased hydrogen peroxide production was eliminated and mortality and brain edema were reduced, suggesting that xanthine oxidase was the source of hydrogen peroxide. A number of attempts to inhibit xanthine oxidase in the brain have met with mixed results. Some investigators found that inhibition of xanthine oxidase reduced infarct size and mortality, whereas others found
THE PRODUCTS
H1356
0363-6135/92
$2.00
Copyright
no effect (2). Liu et al. (13) found that pretreatment with polyethylene glycol-conjugated SOD and polyethylene glycol-conjugated catalase reduced infarct size in rats subjected to middle cerebral artery occlusion. Attempts to demonstrate products of interaction of reactive oxygen species with tissue components have met with irregular and limited success. For example, products of lipid peroxidation have not been found with uniformity after ischemia-reperfusion of the brain (21). In the present research, we investigated the time course of superoxide production after ischemia-reperfusion in cats, and we examined the role of superoxide and its derivatives in the cerebral vascular changes that occur after ischemia of the brain. METHODS Animal preparation. Experiments were conducted on cats anesthetized with pentobarbital sodium (30 mg/kg iv). After tracheostomy, each animal was ventilated with a positive pressure respirator. End-tidal CO, was measured continuously with an infrared CO, analyzer and was maintained at a constant level of -30 mmHg. After operative procedures were completed, the animals received 5 mg/kg gallamine triethiodide intravenously for skeletal muscle paralysis. The left subclavian artery was ligated, and an occluder was placed around the brachiocephalic artery. Large-bore (PE-240) tubing was placed in the abdominal aorta for subsequent phlebotomy. Arterial blood pressure was measured with a Statham transducer connected to a cannula introduced into the aorta via the femoral artery. Arterial blood samples were taken periodically for arterial blood gas and pH measurements using a Corning blood gas analyzer. Hematocrit was measured with a micromethod. Single or double cranial windows were implanted over the parietal cortex as described in detail previously (12). When double windows were used, rubber 0 rings were placed between the dura and skull to prevent mixing of the contents between window chambers. The space under the window was filled with artificial cerebrospinal fluid (CSF) identical in composition to endogenous CSF of cats. Each window had ports for superfusion of reagents and for monitoring intracranial pressure. In experiments in which vessel caliber was measured, we observed several small and large arterioles (smaller or larger than 100 pm, respectively). Arteriolar diameter was measured with a Vickers image-splitting device attached to a Wild dissecting microscope. Superoxide measurements. Superoxide production was measured as the SOD-inhibitable rate of NBT reduction as described in detail previously (8). During the period of interest, a 2.4 mM solution of NBT was placed under both cranial windows. SOD (60 U/ml; 3,000 U/mg protein from bovine blood) was placed under one window during the same period. At the end of the assay period, the reagents were flushed from the window with fresh CSF. The brain then was perfused via both carotid arteries, first with 500 ml of heparinized 0.9% sodium chloride solution and then by 500 ml of a freshly prepared mixture of 2.5% glutaraldehyde and 2% paraformaldehyde in 0.1 M phosphate buffer. The amounts of reduced NBT were determined after pyridine
0 1992 the American
Physiological
Society
Downloaded from www.physiology.org/journal/ajpheart by ${individualUser.givenNames} ${individualUser.surname} (129.186.138.035) on January 5, 2019.
OXYGEN
RADICALS
IN CEREBRAL
extraction of brain homogenates or by surface spectrophotometry using a two-wavelength technique described in detail previously (8). The surface spectrophotometric measurements were calibrated by comparing the results with direct chemical determination of the reduced NBT after extraction with pyridine. The two methods yielded identical results. Pyridine extraction involved first removing a cylindrical core of tissue from under each cranial window and from under another part of each fixed brain. A 2-mm-thick surface section from each sample then was homogenized in 0.1% sodium hydroxide and 1% sodium dodecyl sulfate (SDS). The homogenate then was centrifuged for 20 min at 20,000 g, and the pellet was removed and resuspended in 4 ml of pyridine. After 1 h of heating at 85”C, the samples were recentrifuged for 10 min at 10,000 g. The absorbance at 515 nm then was measured with a Beckman spectrophotometer to analyze samples from under each window and control areas. We used an extinction coefficient of 17.9 x 1O’j M-l cm-’ to calculate the amount of reduced NBT. By subtracting the amount of reduced NBT found in the presence of SOD from the amount reduced in the absence of SOD, the rate of SOD-inhibitable NBT reduction was calculated. The results are expressed in nanomoles per minute per liter assuming a constant rate of NBT deposition throughout the assay period. Comparisons between groups were made using analysis of variance followed by t tests modified for multiple comparisons. Ischemia-reperfusion protocol. Before the induction of complete global cerebral ischemia, the animals were given a constant intravenous infusion of 1 mM ATP solution titrated to lower mean arterial blood pressure to 70 mmHg. The cannula in the abdominal aorta was connected to a reservoir that could be raised or lowered after ischemia induction to keep arterial blood pressure at 70 mmHg while the ATP infusion was continued as above until controlled bleeding was sufficient by itself to control arterial blood pressure. When the ischemic period was completed, the animal’s blood was reinfused at rates regulated to keep mean arterial blood pressure below 130 mmHg. Ischemia was induced by completely occluding the brachiocephalic artery. This maneuver typically resulted in reflex hypertension that was attenuated with controlled bleeding and ATP infusion. The completeness of ischemia was verified visually by noting that blood flow in pial vessels ceased completely or that vessels emptied completely of blood and the brain surface became pale without visible blood vessels. Ischemic time was measured after complete ischemia was established. This occurred usually -5 min after arterial occlusion. In four cats we measured blood flow with radioactive microspheres and verified that blood flow in the brain under the cranial window was zero during the period of ischemia. Experimental design. We measured superoxide production in the following groups of cats: 1) during complete ischemia; 2) during reperfusion at O-15 min after complete ischemia; 3) during reperfusion at 60-75 min after complete ischemia; 4) during reperfusion at 120-135 min after complete ischemia; 5) during reperfusion at O-15 min after incomplete ischemia. This was accomplished using the same protocol as for complete ischemia, except that no attempt was made to control the blood pressure. Under these conditions, there was intermittent or continuous sluggish flow through the vessels under the cranial window. Superoxide production also was measured in control animals subjected to the same operative procedures but without the induction of ischemia. We conducted the following experiments to determine the effects of oxygen radicals on postischemic arteriolar reactivity. We studied the time course of endothelium-dependent vasodilation from acetylcholine after ischemia-reperfusion in eight cats equipped with single cranial windows. After control responses to topical application of acetylcholine (low7 M) were
ISCHEMIA
H1357
determined, 15 min of global complete ischemia was induced. Vessel responses to acetylcholine were measured again 30, 45, 60, 75, 90, and 120 min after the onset of reperfusion. The effect of oxygen radical scavengers on arteriolar responses to acetylcholine after ischemia-reperfusion was studied in five cats equipped with double cranial windows. In seven cats, one window was pretreated with SOD (60 U/ml) plus catalase (40 U/ml) while the other window was left untreated. Responses to topical acetylcholine ( 10e7 M) were determined before and 60 min after the onset of reperfusion after a 15-min period of complete global ischemia. The effect of oxygen radical scavenging agents on arteriolar responses to hypocapnia after ischemia-reperfusion was studied in 12 cats equipped with double cranial windows. One window was pretreated with SOD (60 U/ml) plus catalase (40 U/ml). Responses to hypocapnia were measured before and at 60 min of reperfusion after complete global ischemia. Hypocapnia was induced by hyperventilation via increasing the volume and rate of the respirator. In five cats we studied the same responses to hypocapnia except that, instead of pretreatment with SOD and catalase, we pretreated one window with deferoxamine (1 mM) to scavenge iron and thereby inhibit the generation of hydroxyl radical via the iron-catalyzed Haber-Weiss reaction. We also studied the effect of ischemia-reperfusion on the permeability of the blood-brain barrier to proteins by using two methods. The first method determined the permeability of the blood-brain barrier to human plasma albumin (mol wt 70,000) labeled with lZr,I . Ten minutes before testing, the animal received 50 &i of labeled albumin. At the end of the experiment, the blood containing the radioactive albumin was removed from the cerebral vessels by perfusion of the brain transcardially, first with 0.9% sodium chloride solution and then with fixative consisting of 2.5% glutaraldehyde, 2% paraformaldehyde, and 0.1 M phosphate buffer as described above. This technique reliably eliminated all blood from the blood vessels. Fixation permitted cutting the brain samples in precise fashion to measure radioactivity in the desired thickness of the cerebral cortex. Appropriate samples of the brain then were cut and weighed, and their radioactivity was determined in an LKB 1282 gamma counter. Blood samples were drawn immediately before the onset of reperfusion and 1 h later. A permeability index was calculated by expressing the concentration of labeled albumin per unit weight of brain as a percent of the mean concentration of radioactive albumin in blood. The effectiveness of perfusion in eliminating blood from the cerebral vessels was verified by examination of the brain surface. In a properly perfused preparation, the vessels were no longer visible because of the absence of red blood cells. Incomplete perfusion resulted in retention of red blood cells, which rendered the vessels visible by virtue of the presence of hemoglobin. The concentration of labeled albumin was measured in a 2-mm surface layer of brain in each of two cranial windows. One window was pretreated with SOD (60 U/ml) plus catalase (40 U/ml) before the induction of a 15-min period of complete global ischemia while the other window was left untreated. We also measured the extravasation of labeled albumin in animals that underwent the same operative procedures but were not subjected to ischemia and therefore served as controls. A second technique was used to measure the permeability of the blood-brain barrier to horseradish peroxidase (mol wt 40,000). The animals received 50 mg/kg of horseradish peroxidase intravenously (Sigma type VI) 5 min before the experimental interventions were begun. This type of peroxidase causes no significant change in the animal’s arterial blood pressure and no alteration in the caliber of pial vessels. After the experimental interventions were performed, the animal’s head was perfused with fixative as described above to eliminate the blood from the cerebral vessels and to fix the brain. The brain was removed and the cortex was serially sectioned at a thickness
Downloaded from www.physiology.org/journal/ajpheart by ${individualUser.givenNames} ${individualUser.surname} (129.186.138.035) on January 5, 2019.
H1358
OXYGEN
RADICALS
IN CEREBRAL
ISCHEMIA
of 40 pm on a Vibratome. The sections were reacted for light microscopic visualization of the protein reaction product through the use of the cobalt glucose oxidase technique (8). Next the sections were mounted on glass slides, dehydrated, and cleared for microscopic analysis. Corresponding anatomic areas from oxygen radical scavenger-treated and untreated windows were compared by a blinded neuroanatomist to assess qualitatively the intensity and localization of the peroxidase reaction product. RESULTS
Superoxide production. Figure 1 shows that superoxide generation was not seen in control animals not subjected to ischemia, during the period of complete ischemia, and 120-135 min of reperfusion after complete ischemia. There was significant superoxide production at O-15 and 60-75 min of reperfusion after complete ischemia. Superoxide production was much greater in the early part of reperfusion than later on. Thus superoxide production after ischemia was time dependent. NBT reduction rates in nanomoles per minute per liter during O-15 min of reperfusion after incomplete ischemia were as follows: NBT alone, 5.42 t 0.86; NBT plus SOD, 3.80 t 0.61; giving a rate of SOD-inhibitable NBT reduction of 1.62 t 0.75. The latter was significantly lower than that seen during the corresponding period after complete ischemia. Vascular caliber and reactivity. Figure 2 shows the changes in arteriolar caliber observed during the period of reperfusion after complete ischemia. Marked vasodilation, seen in the early period of reperfusion, became progressively less pronounced. However, at the end of the 2-h observation, the cerebral arterioles were still significantly dilated when compared with their caliber before ischemia. Figure 3 shows that the dilation of cerebral arterioles at 60 min of reperfusion after complete ischemia was significantly reduced in the windows pretreated with SOD plus catalase or with deferoxamine when compared with the untreated side. 10 T
.-$8 -1 b cl
.\2 : 54
I
NET
EE3 eZa
NBT + SOD DIFFERENCE
I
I \c
CONTROL
ISCHEMIA
REPERFUSION o-15
60-75
(MN) 120-135
Fig. 1. Superoxide production, measured as superoxide dismutase (SOD)-inhibitable reduction of nitro blue tetrazolium (NBT), in control animals not subjected to complete ischemia (n. = 4), during ischemia (n = 4), and at O-15 (n = 6), 60-75 (n = 7), and 120-135 (n = 6) min after complete ischemia. Columns, means k SE of NBT reduction with and without SOD and difference between them. Difference, SOD-inhibitable NBT reduction and a measure of superoxide production. No superoxide was detected under control conditions, during ischemia, and late in reperfusion period (120-135 min). Significant superoxide production (P < 0.05) was found at O-15 and 60-75 min during reperfusion after ischemia.
o0
30 TIME AFTER
Fig. 2. Arteriolar diameter after complete ischemia. chemic diameters (means significantly dilated (P < line diameters from which
60 REPERFUSION
120
90 (min)
changes at various times during reperfusion Values are %change compared with preis75 vessels). Vessels are t SE, n = 11 animals, 0.001) throughout reperfusion period. Base%changes were calculated are 98.2 k 4.8 pm.
a m
SOD + CP.rALASE DEFEROXAMINE
b
$ 20 .-6 ; 10 .a 0 UNTREATED
TREATED
Fig. 3. Effect of treatment with SOD + catalase or with deferoxamine on change in arteriolar diameter after complete ischemia. Results are %change measured 60 min after ischemia compared with preischemic values (means + SE; SOD + catalase: n = 7 animals, 42 vessels; deferoxamine: n = 5 animals, 36 vessels). Treatment with SOD + catalase or with deferoxamine resulted in significantly reduced vasodilation (P < 0.001) after ischemia compared with untreated vessels. Baseline diameters from which %changes were calculated are displayed above columns (means k SE).
Figure 4 shows the response to topical acetylcholine before and during the period of reperfusion after complete ischemia. The normal vasodilator response to acetylcholine seen before ischemia was converted to a vasoconstriction during the early phase of reperfusion. This abnormal vasoconstrictor response persisted until 60 min of reperfusion when the response to acetylcholine reverted back to vasodilation. At the end of the observation period of 2 h the response was vasodilator, but it was significantly less than that seen before ischemia. Figure 5 shows that at 60 min of reperfusion after complete ischemia, the response to acetylcholine was vasoconstrictor in the untreated window, whereas pretreatment with SOD and catalase resulted in a retained vasodilator response to acetylcholine. Figure 6 shows that the vasoconstrictor responses to hypocapnia were significantly reduced at 60 min of reperfusion after complete ischemia and that this effect was inhibited by pretreatment with SOD plus catalase or by deferoxamine. Figure 7 shows that the permeability index to albumin increased considerably in animals subjected to ischemiareperfusion without pretreatment. Pretreatment with
Downloaded from www.physiology.org/journal/ajpheart by ${individualUser.givenNames} ${individualUser.surname} (129.186.138.035) on January 5, 2019.
OXYGEN
RADICALS
IN CEREBRAL
H1359
ISCHEMIA BEFORE
ISCHEMIA
AFTER
ISCHEMIA
15, G
10
.-L -G
5
:: n
EL
25 b
55 . .
-I
1
O--
y -5i *: -lO.-c a -15-
E
.-0 u .-c
a
I
--15!
1
I/‘//I
pre-iachemic control
I
0
1I I 1 I 1 1 30 60 90 TIME AFTER REPERFUSION (MIN)
1 120
Fig. 4. Change in arteriolar diameter caused by topical lop7 M acetylcholine application before ischemia and at various times during reperfusion after complete ischemia. Results are %change compared with diameters immediately before acetylcholine application (means t SE; n = 8 animals, 45 vessels). Acetylcholine-induced vasoconstriction occurred during first 60 min of reperfusion. Significant vasodilation occurred with acetylcholine after 120 min of reperfusion but was reduced when compared with response seen before ischemia (P c 0.005). Baseline diameters from which %changes were calculated are displayed above or beside columns (means t SE). BEFORE
25 7.5 :: n
AFTER
deferoxamine
sod + catalase
catalase
deferoxamine
Fig. 6. Effect of treatment with SOD + catalase or with deferoxamine on change in arteriolar diameter in response to hyperventilation before and after ischemia. Values are %change from diameters immediately before hyperventilation (means t SE; SOD + catalase: n = 7 animals, 42 vessels; deferoxamine: n = 5 animals, 36 vessels). Treatment has no effect on vasoconstrictor response to hyperventilation before ischemia. After ischemia, treated vessels responded normally to hyperventilation, whereas untreated vessels had a significantly (P < 0.001) attenuated response. Baseline diameters from which %changes were calculated are displayed above columns (means t SE). 1.00
ISCHEMIA
20
z
.90 T
=
.00-
p ‘5
.70AO-
T
15
z 28
lo
&I .-i -u c .a
ISCHEMIA
syd
I
untreate
5 CONTROL
-0 -5 -10 untreated
sod + catalase
untreated
sod + catalase
Fig. 5. Effect of pretreatment with SOD + catalase on change in arteriolar diameter in response to topical 10ey M acetylcholine application at 60 min of reperfusion after complete ischemia. Results are %change compared with diameter immediately before acetylcholine application (means k SE; n = 5 animals, 28 vessels). SOD + catalase had no effect on acetylcholine responses before ischemia. Untreated vessels constricted in response to acetylcholine application after ischemia. Treated vessels dilated in response to acetylcholine after ischemia, but response was attenuated when compared with preischemic responses (see also Fig. 4). Baseline diameters from which %changes were calculated are displayed above columns (means 2 SE).
SOD and catalase reduced this effect but did not eliminate it. Extravasation of horseradish peroxidase was observed consistently within the brain parenchyma underlying the cranial windows in all animals subjected to ischemiareperfusion. The altered cerebral vascular permeability was seen in relation to the crest of the gyri as well as in the deeper brain structures. Oxygen radical scavengers did not completely prevent the permeability change induced by ischemia-reperfusion but reduced it significantly (Fig. 8). The brain parenchyma exposed to topically applied SOD plus catalase showed easily recognizable reduction in permeability when compared with the
ISCHEMIA + SOD + CATALASE
ISCHEMIA UNTREATED
Fig. 7. Effect of treatment with SOD + catalase on changes in permeability to albumin after ischemia (n = 5 per each group). Values are means +: SE of permeability index, which is ratio of radiolabeled albumin content in fixed brain to that in whole blood (x100) measured 60 min after ischemia or in control period. Treatment with SOD + catalase reduced but did not completely prevent increase in permeability index seen after ischemia.
untreated side. SOD alone (60 U/ml) had a consistent but small effect on the increased permeability induced by ischemia-reperfusion. SOD in a concentration of 120 U/ml had a more pronounced effect. Catalase (80 U/ml) alone had no detectable effect on the increased permeability from ischemia-reperfusion. DISCUSSION
The important findings of these experiments are as follows. 1) Superoxide generation occurs during the period of reperfusion after either complete or incomplete cerebral ischemia. There is no superoxide production under resting conditions in the absence of ischemia or during the period of complete cerebral ischemia. Superoxide generation is much greater after complete rather than incomplete ischemia, and superoxide generation in the former condition lasts for >l h. 2) The cerebral arteriolar changes induced by ischemia-reperfusion are characterized by sustained dilation, reduced responsiveness to the vasoconstrictor effects of arterial hypocapnia, reversal of
Downloaded from www.physiology.org/journal/ajpheart by ${individualUser.givenNames} ${individualUser.surname} (129.186.138.035) on January 5, 2019.
Fig. 8. These light micrographs demonstrate altered cerebral vascular permeability to horseradish peroxidase occurring after ischemia with reperfusion in presence of various concentrations of SOD and/or catalase. In A, note that combined use of 120 U/ml SOD + 80 U/ml catalase reduced overall cortical extravasation of peroxidase (*) compared with untreated controls (B, arrows). Note that in both control and treated hemispheres, permeability change deep within brain parenchyma (arrowheads) is unaffected by SOD-catalase treatment. This most likely reflects inability of employed agents to diffuse rapidly to deep cortical sites. With use of 120 U/ml SOD, a marked reduction in altered cerebrovascular permeability to horseradish peroxidase (C, *) is seen compared with control hemisphere (D, arrows). Note again that use of this superoxide anion scavenger did not affect permeability change deep within cortex. Downloaded from www.physiology.org/journal/ajpheart by ${individualUser.givenNames} ${individualUser.surname} (129.186.138.035) on January 5, 2019.
OXYGEN
RADICALS
IN CEREBRAL
the endothelium-dependent vasodilator response to acetylcholine, and disruption of the blood-brain barrier to proteins. These changes are induced by oxygen radicals because they are inhibited by pretreatment with specific scavengers. In the case of the sustained vasodilation and the abnormal response to hypocapnia after ischemia-reperfusion, we identified hydroxyl radical as the agent responsible for the abnormalities. Unlike what occurs in hearts in which the production of oxygen radicals during reperfusion after ischemia is brief (lo), in the in vivo blood-perfused brain superoxide production is prolonged, lasting for >l h. This is similar to what is seen in other experimental pathophysiological conditions such as acute hypertension and fluid-percussion brain injury, in which superoxide production lasts for at least 1 h after the insult (7, 22). The prolonged generation of superoxide in ischemia-reperfusion may have practical significance because it affords a better opportunity for therapeutic intervention. The sites of generation of superoxide and its enzymatic sources were not investigated in this study. In another investigation, we used a histochemical technique to localize superoxide in the same model used in the present experiments (5). We found that the production of superoxide occurred in endothelial and smooth muscle cells of the cerebral arterioles, and we also localized superoxide in the extracellular space in proximity to the blood vessels. We showed earlier that superoxide escapes into the extracellular space from the ceils of the vessel wall via the anion channel (8). In this histochemical investigation (5) we found no superoxide production in the brain parenchyma, but its absence there may be due to incomplete penetration of the reagents that were applied topically on the brain surface. The enzymatic sources of superoxide during reperfusion after brain ischemia have not been identified. As noted at the outset of this article, the available evidence with respect to xanthine oxidase is conflicting. Armstead et al. (1) found that indomethacin inhibited superoxide production during the early part of reperfusion after ischemia in newborn pigs. Therefore they concluded that cyclooxygenase was the source of superoxide. We have shown previously that cyclooxygenase generates superoxide in vitro (8, 11). It should be noted, however, that the rate of SOD-inhibitable reduction of NBT reported by Armstead et al. (1) is -100 times greater than what we found. The microvascular abnormalities seen during reperfusion after complete ischemia, including sustained arteriolar dilation, abnormal reactivity to hypocapnia, reversal of the endothelium-dependent vasodilation response to acetylcholine and increased permeability of the bloodbrain barrier to proteins, are clearly due to the generation of oxygen radicals because they were inhibited strongly by pretreatment with oxygen radical scavengers. We did not investigate the role of radicals in the more pronounced reactive dilation that occurs in the earlier phases of reperfusion. Although the time course of the dilation is similar to the time course of generation of superoxide, suggesting a causal relationship, it is likely that this earlier dilation is multifactorial. We found earlier that release of
ISCHEMIA
H1361
polypeptides from sensory fibers is involved (19). It is also likely that adenosine may contribute because its concentration during ischemia rises markedly (23). The time course of the abnormal response to acetylcholine correlates with that of the superoxide production, suggesting that the two are causally related. This is further supported by the inhibition of the abnormality after pretreatment with SOD and catalase. It is well known that oxygen radicals attack directly and inactivate the endothelium-derived relaxing factor from acetylcholine (3, 9, 18). We believe this is the mechanism involved in view of the spontaneous return of the vasodilator response to acetylcholine 2 h after reperfusion. Had the mechanism been irreversible damage of the endothelium, we would have expected that this would not have occurred. From the practical standpoint, the response to acetylcholine may be used to gauge the time course of superoxide production during reperfusion after ischemia. It is likely that the increase in baseline vascular caliber caused by oxygen radicals also may have contributed to the reduction in the vasodilator responses to acetylcholine by nonspecific mechanisms. However, this factor cannot fully explain the findings because it cannot explain the reversal of the response from vasodilator to vasoconstrictor. Other investigators have noted inconsistent responses to acetylcholine application after cerebral ischemia-reperfusion in cats (15). They started measuring responses after 60 min of reperfusion and saw both dilation and constriction. Our data show that 60 min is approximately the time that responses return toward normal; therefore heterogeneous findings would not be unexpected. Their period of ischemia was shorter than in our studies, which would affect the time course to recovery, because the amount of reperfusion injury is related to the intensity of initial ischemia. The response to hypocapnia in large cerebral vessels is not endothelium dependent (20). It would appear, therefore, that the inhibition of the vasoconstrictor response to hypocapnia observed in the present experiments after ischemia is due to effects of oxygen radicals directly on vascular smooth muscle. Because this effect was inhibited by both SOD plus catalase and by deferoxamine, we attributed it to the direct effects of hydroxyl radical. This radical is very reactive and short-lived. Consequently, it does not survive more than a few molecular diameters from its site of formation. Because it is accessible to deferoxamine, SOD and catalase, which are unlikely to enter into the interior of cells easily, it is likely that its action is on the cell membrane. It has been reported by others that oxygen radicals can damage membrane enzyme systems and affect ion fluxes leading to permeability changes after different types of cerebral injury (14, 16). Our findings show that the increased permeability of the blood-brain barrier to proteins during reperfusion after ischemia is mediated by oxygen radicals. The precise mechanism by which this occurs was not identified. It appears from the use of different scavengers that superoxide is the agent immediately responsible for the increased permeability because SOD, but not catalase alone, was effective in minimizing
Downloaded from www.physiology.org/journal/ajpheart by ${individualUser.givenNames} ${individualUser.surname} (129.186.138.035) on January 5, 2019.
H1362
OXYGEN
RADICALS
IN CEREBRAL
the response. The increased permeability to proteins coupled with the sustained arteriolar dilation, which should lead to increased capillary pressure, would be expected to contribute to the induction of edema with secondary adverse effects on brain parenchymal function. This research was supported by National Institutes of Health Grants HL-21851, NS-19316, NS-26361, and HL-07537. Address for reprint requests: H. A. Kontos, Box 662, MCV Station, Medical College of Virginia, Richmond, VA 23298. Received
31 March
1992; accepted
in final
form
16 June
1992.
REFERENCES 1. Armstead, W. M., R. Mirro, D. W. Busija, and C. W. Leffler. Postischemic generation of superoxide anion by newborn pig brain. Am. J. Physiol. 255 (Heart Circ. Physiol. 24): H401-H403, 1988. 2. Betz, A. L., J. Randall, and D. Martz. Xanthine oxidase is not a major source of free radicals in focal cerebral ischemia. Am. J. Physiol. 260 (Heart Circ. Physiol. 29): H563-H568, 1991. 3. Gryglewski, R. J., R. M. J. Palmer, and S. Moncada. Superoxide anion is involved in the breakdown of endothelium-derived vascular relaxing factor. Nature Lond. 320: 454-456, 1986. 4. Itoh, K., A. Konishi, S. Nomura, N. Mizuno, Y. Nakamura, and T. Sugimoto. Application of coupled oxidation reaction to electron microscopic demonstration of horseradish peroxidase: cobalt-glucose oxidase method. Brain Res. 175: 341-346, 1979. 5. Kontos, C. D., E. P. Wei, J. I. Williams, H. A. Kontos, and J. T. Povlishock. Cytochemical detection of superoxide in cerebral inflammation and ischemia. Am. J. Physiol. 263 (Heart Circ. Physiol. 32): H1234-H1242, 1992. 6. Kontos, H. A. Oxygen radicals in CNS damage. Chem. Biol. Interact. 72: 229-255, 1989. 7. Kontos, H. A., and E. P. Wei. Superoxide production in experimental brain injury. J. Neurosurg. 64: 803-807, 1986. 8. Kontos, H. A., E. P. Wei, E. F. Ellis, L. W. Jenkins, J. T. Povlishock, G. T. Rowe, and M. L. Hess. Appearance of superoxide anion radical in cerebral extracellular space during increased prostaglandin synthesis in cats. Circ. Res. 57: 142-151, 1985. 9. Kontos, H. A., E. P. Wei, J. T. Povlishock, R. C. Kukreja, and M. L. Hess. Inhibition by arachidonate of cerebral arteriolar dilation from acetylcholine. Am. J. Physiol. 256 (Heart Circ. Physiol. 25): H665-H671, 1989. 10. Kramer, J. H., C. M. Arroyo, B. F. Dickens, and W. B. Weglicki. Spin-trapping evidence that graded myocardial ischemia alters post-ischemic superoxide production. Free Radical Biol. Med. 3: 153-159, 1987.
ISCHEMIA
11. Kukreja, R., H. A. Kontos, M. L. Hess, and E. F. Ellis. PGH synthase and lipoxygenase generate superoxide in the presence of NADH or NADPH. Circ. Res. 59: 612-619, 1986. 112. Levasseur, J. E., E. P. Wei, A. J. Raper, H. A. Kontos, and J. L. Patterson, Jr. Detailed description of a cranial window technique for acute and chronic experiments. Stroke 6: 308-317, 1975. 13. Liu, T. A., J. S. Beckman, B. A. Freeman, E. L. Hogan, and C. Y. Hsu. Polyethylene glycol-conjugated superoxide dismutase and catalase reduce ischemic brain injury. Am. J. Physiol. 256 (Heart Circ. Physiol. 25): H589-H593, 1989. 14. Lo, W. D., and A. L. Betz. Oxygen free-radical reduction of brain capillary rubidium uptake. J. Neurochem. 46: 394-397, 1986. W. G., S. M. Amundsen, F. M. Faraci, and D. D. 15. Mayhan, Heistad. Responses of cerebral arteries after ischemia and reperfusion in cats. Am. J. Physiol. 255 (Heart Circ. Physiol. 24): H879H884, 1988. 16. Palmer, G. C. Free radicals generated by xanthine oxidase-hypoxanthine damage adenylate cyclase and ATPase in gerbil cerebral cortex. Metab. Brain Dis. 2: 243-257, 1987. 17. Patt, A., A. H. Harken, L. K. Burton, T. C. Rodell, D. Piermattei, W. J. Schorr, N. B. Parker, E. M. Berger, I. R. Horesh, L. S. Terada, S. L. Linas, J. C. Cheronis, and J. E. Repine. Xanthine oxidase-derived hydrogen peroxide contributes to ischemia reperfusion-induced edema in gerbil brains. J. Clin. Invest. 81: 1556-1562, 1988. G. M., and P. M. Vanhoutte. Superoxide anions and 18. Rubanyi, hyperoxia inactivate endothelium-derived relaxing factor. Am. J. Physiol. 250 (Heart Circ. Physiol. 19): H822-H827, 1986. 19. Sakas, D. E., M. A. Moskowitz, E. P. Wei, H. A. Kontos, M. Kano, and C. S. Ogilvy. Trigeminovascular fibers increase blood flow in cortical gray matter by axon reflex-like mechanisms during acute severe hypertension or seizures. Neurobiology CPH. 86: 1401-1405, 1989. 20. Toda, N., Y. Hatano, and K. Mori. Mechanisms underlying response to hypercapnia and bicarbonate of isolated dog cerebral arteries. Am. J. Physiol. 257 (Heart Circ. Physiol. 26): H141-H146, 1989. 21. Watson, B. D., R. Busto, W. J. Goldberg, M. Santiso, S. Yoshida, and M. D. Ginsberg. Lipid peroxidation in vivo induced by reversible global ischemia in rat brain. J. Neurochem. 42: 268-274, 1984. 22. Wei, E. P., H. A. Kontos, C. W. Christman, D. S. Dewitt, and J. T. Povlishock. Superoxide generation and reversal of acetylcholine-induced cerebral arteriolar dilation after acute hypertension. Circ. Res. 57: 781-787, 1985. 23. Winn, H. R., R. Rubio, and R. M. Berne. Brain adenosine production in the rat during 60 seconds of ischemia. Circ. Res. 45: 486-492, 1979.
Downloaded from www.physiology.org/journal/ajpheart by ${individualUser.givenNames} ${individualUser.surname} (129.186.138.035) on January 5, 2019.
in cerebral ischemia
CHARLES W. NELSON, ENOCH P. WEI, JOHN T. POVLISHOCK, HERMES A. KONTOS, AND MICHAEL A. MOSKOWITZ Departments of Medicine and Anatomy, Medical College of Virginia, Virginia Commonwealth University, Richmond, Virginia 23298; and Neurosurgery and Neurology Services, Massachusetts General Hospital, Harvard Medical School, Boston, Massachusetts 02114 Nelson, Charles W., Enoch P. Wei, John T. Povlishock, Hermes A. Kontos, and Michael A. Moskowitz. Oxygen radicals in cerebral ischemia. Am. J. Physiol. 263 (Heart Circ. Physiol. 32): H1356-H1362, 1992.-Superoxide production was measured as the superoxide dismutase (SOD)-inhibitable portion of nitro blue tetrazolium (NBT) reduction after cerebral ischemia-reperfusion in anesthetized cats equipped with cranial windows. Significant superoxide production was found in the early reperfusion period and continued for more than 1 h after ischemia. Superoxide was not detected in control animals not subjected to ischemia, during ischemia, and at 120 min of reperfusion. After ischemia, the vasoconstrictor response to arterial hypocapnia was reduced. This effect was prevented by pretreatment with SOD plus catalase or by deferoxamine. The response to topical acetylcholine was converted to vasoconstriction after ischemia. The normal vasodilator response reappeared spontaneously at 120 min of reperfusion. The vasodilator response to acetylcholine was preserved in animals pretreated with SOD plus catalase. Blood-brain barrier permeability to labeled albumin and horseradish peroxidase was increased after ischemia. These effects were minimized by pretreatment with SOD and catalase. We conclude that superoxide generation occurs during reperfusion after cerebral ischemia for a fairly long period and that superoxide and its derivatives are responsible at least in part for the vasodilation and the abnormal reactivity as well as for the increase in blood-brain barrier permeability to macromolecules seen after ischemia. Furthermore, the findings suggest that the agent responsible for the vascular abnormalities is hydroxyl radical generated via the iron-catalyzed HaberWeiss reaction. cerebral microcirculation; hydroxyl radical; blood-brain barrier; endothelium-dependent vasodilation OF univalent reduction of oxygen, superoxide, hydrogen peroxide, and hydroxyl radical are very reactive and capable of inducing tissue injury (6). They have been suggested as possible mediators of the tissue injury in ischemia-reperfusion in many tissues, including the brain. The generation of reactive oxygen species and their role in tissue injury in ischemia-reperfusion of the brain have received experimental support from a variety of sources. These include direct demonstration of superoxide production by showing the presence of superoxide dismutase (SOD) -inhibitable reduction of nitro blue tetrazolium (NBT) in the early part of reperfusion after ischemia in newborn pigs (1) and the finding of increased hydrogen peroxide production in gerbils subjected to brain ischemia (17). The latter authors also showed that after inactivation of xanthine oxidase the increased hydrogen peroxide production was eliminated and mortality and brain edema were reduced, suggesting that xanthine oxidase was the source of hydrogen peroxide. A number of attempts to inhibit xanthine oxidase in the brain have met with mixed results. Some investigators found that inhibition of xanthine oxidase reduced infarct size and mortality, whereas others found
THE PRODUCTS
H1356
0363-6135/92
$2.00
Copyright
no effect (2). Liu et al. (13) found that pretreatment with polyethylene glycol-conjugated SOD and polyethylene glycol-conjugated catalase reduced infarct size in rats subjected to middle cerebral artery occlusion. Attempts to demonstrate products of interaction of reactive oxygen species with tissue components have met with irregular and limited success. For example, products of lipid peroxidation have not been found with uniformity after ischemia-reperfusion of the brain (21). In the present research, we investigated the time course of superoxide production after ischemia-reperfusion in cats, and we examined the role of superoxide and its derivatives in the cerebral vascular changes that occur after ischemia of the brain. METHODS Animal preparation. Experiments were conducted on cats anesthetized with pentobarbital sodium (30 mg/kg iv). After tracheostomy, each animal was ventilated with a positive pressure respirator. End-tidal CO, was measured continuously with an infrared CO, analyzer and was maintained at a constant level of -30 mmHg. After operative procedures were completed, the animals received 5 mg/kg gallamine triethiodide intravenously for skeletal muscle paralysis. The left subclavian artery was ligated, and an occluder was placed around the brachiocephalic artery. Large-bore (PE-240) tubing was placed in the abdominal aorta for subsequent phlebotomy. Arterial blood pressure was measured with a Statham transducer connected to a cannula introduced into the aorta via the femoral artery. Arterial blood samples were taken periodically for arterial blood gas and pH measurements using a Corning blood gas analyzer. Hematocrit was measured with a micromethod. Single or double cranial windows were implanted over the parietal cortex as described in detail previously (12). When double windows were used, rubber 0 rings were placed between the dura and skull to prevent mixing of the contents between window chambers. The space under the window was filled with artificial cerebrospinal fluid (CSF) identical in composition to endogenous CSF of cats. Each window had ports for superfusion of reagents and for monitoring intracranial pressure. In experiments in which vessel caliber was measured, we observed several small and large arterioles (smaller or larger than 100 pm, respectively). Arteriolar diameter was measured with a Vickers image-splitting device attached to a Wild dissecting microscope. Superoxide measurements. Superoxide production was measured as the SOD-inhibitable rate of NBT reduction as described in detail previously (8). During the period of interest, a 2.4 mM solution of NBT was placed under both cranial windows. SOD (60 U/ml; 3,000 U/mg protein from bovine blood) was placed under one window during the same period. At the end of the assay period, the reagents were flushed from the window with fresh CSF. The brain then was perfused via both carotid arteries, first with 500 ml of heparinized 0.9% sodium chloride solution and then by 500 ml of a freshly prepared mixture of 2.5% glutaraldehyde and 2% paraformaldehyde in 0.1 M phosphate buffer. The amounts of reduced NBT were determined after pyridine
0 1992 the American
Physiological
Society
Downloaded from www.physiology.org/journal/ajpheart by ${individualUser.givenNames} ${individualUser.surname} (129.186.138.035) on January 5, 2019.
OXYGEN
RADICALS
IN CEREBRAL
extraction of brain homogenates or by surface spectrophotometry using a two-wavelength technique described in detail previously (8). The surface spectrophotometric measurements were calibrated by comparing the results with direct chemical determination of the reduced NBT after extraction with pyridine. The two methods yielded identical results. Pyridine extraction involved first removing a cylindrical core of tissue from under each cranial window and from under another part of each fixed brain. A 2-mm-thick surface section from each sample then was homogenized in 0.1% sodium hydroxide and 1% sodium dodecyl sulfate (SDS). The homogenate then was centrifuged for 20 min at 20,000 g, and the pellet was removed and resuspended in 4 ml of pyridine. After 1 h of heating at 85”C, the samples were recentrifuged for 10 min at 10,000 g. The absorbance at 515 nm then was measured with a Beckman spectrophotometer to analyze samples from under each window and control areas. We used an extinction coefficient of 17.9 x 1O’j M-l cm-’ to calculate the amount of reduced NBT. By subtracting the amount of reduced NBT found in the presence of SOD from the amount reduced in the absence of SOD, the rate of SOD-inhibitable NBT reduction was calculated. The results are expressed in nanomoles per minute per liter assuming a constant rate of NBT deposition throughout the assay period. Comparisons between groups were made using analysis of variance followed by t tests modified for multiple comparisons. Ischemia-reperfusion protocol. Before the induction of complete global cerebral ischemia, the animals were given a constant intravenous infusion of 1 mM ATP solution titrated to lower mean arterial blood pressure to 70 mmHg. The cannula in the abdominal aorta was connected to a reservoir that could be raised or lowered after ischemia induction to keep arterial blood pressure at 70 mmHg while the ATP infusion was continued as above until controlled bleeding was sufficient by itself to control arterial blood pressure. When the ischemic period was completed, the animal’s blood was reinfused at rates regulated to keep mean arterial blood pressure below 130 mmHg. Ischemia was induced by completely occluding the brachiocephalic artery. This maneuver typically resulted in reflex hypertension that was attenuated with controlled bleeding and ATP infusion. The completeness of ischemia was verified visually by noting that blood flow in pial vessels ceased completely or that vessels emptied completely of blood and the brain surface became pale without visible blood vessels. Ischemic time was measured after complete ischemia was established. This occurred usually -5 min after arterial occlusion. In four cats we measured blood flow with radioactive microspheres and verified that blood flow in the brain under the cranial window was zero during the period of ischemia. Experimental design. We measured superoxide production in the following groups of cats: 1) during complete ischemia; 2) during reperfusion at O-15 min after complete ischemia; 3) during reperfusion at 60-75 min after complete ischemia; 4) during reperfusion at 120-135 min after complete ischemia; 5) during reperfusion at O-15 min after incomplete ischemia. This was accomplished using the same protocol as for complete ischemia, except that no attempt was made to control the blood pressure. Under these conditions, there was intermittent or continuous sluggish flow through the vessels under the cranial window. Superoxide production also was measured in control animals subjected to the same operative procedures but without the induction of ischemia. We conducted the following experiments to determine the effects of oxygen radicals on postischemic arteriolar reactivity. We studied the time course of endothelium-dependent vasodilation from acetylcholine after ischemia-reperfusion in eight cats equipped with single cranial windows. After control responses to topical application of acetylcholine (low7 M) were
ISCHEMIA
H1357
determined, 15 min of global complete ischemia was induced. Vessel responses to acetylcholine were measured again 30, 45, 60, 75, 90, and 120 min after the onset of reperfusion. The effect of oxygen radical scavengers on arteriolar responses to acetylcholine after ischemia-reperfusion was studied in five cats equipped with double cranial windows. In seven cats, one window was pretreated with SOD (60 U/ml) plus catalase (40 U/ml) while the other window was left untreated. Responses to topical acetylcholine ( 10e7 M) were determined before and 60 min after the onset of reperfusion after a 15-min period of complete global ischemia. The effect of oxygen radical scavenging agents on arteriolar responses to hypocapnia after ischemia-reperfusion was studied in 12 cats equipped with double cranial windows. One window was pretreated with SOD (60 U/ml) plus catalase (40 U/ml). Responses to hypocapnia were measured before and at 60 min of reperfusion after complete global ischemia. Hypocapnia was induced by hyperventilation via increasing the volume and rate of the respirator. In five cats we studied the same responses to hypocapnia except that, instead of pretreatment with SOD and catalase, we pretreated one window with deferoxamine (1 mM) to scavenge iron and thereby inhibit the generation of hydroxyl radical via the iron-catalyzed Haber-Weiss reaction. We also studied the effect of ischemia-reperfusion on the permeability of the blood-brain barrier to proteins by using two methods. The first method determined the permeability of the blood-brain barrier to human plasma albumin (mol wt 70,000) labeled with lZr,I . Ten minutes before testing, the animal received 50 &i of labeled albumin. At the end of the experiment, the blood containing the radioactive albumin was removed from the cerebral vessels by perfusion of the brain transcardially, first with 0.9% sodium chloride solution and then with fixative consisting of 2.5% glutaraldehyde, 2% paraformaldehyde, and 0.1 M phosphate buffer as described above. This technique reliably eliminated all blood from the blood vessels. Fixation permitted cutting the brain samples in precise fashion to measure radioactivity in the desired thickness of the cerebral cortex. Appropriate samples of the brain then were cut and weighed, and their radioactivity was determined in an LKB 1282 gamma counter. Blood samples were drawn immediately before the onset of reperfusion and 1 h later. A permeability index was calculated by expressing the concentration of labeled albumin per unit weight of brain as a percent of the mean concentration of radioactive albumin in blood. The effectiveness of perfusion in eliminating blood from the cerebral vessels was verified by examination of the brain surface. In a properly perfused preparation, the vessels were no longer visible because of the absence of red blood cells. Incomplete perfusion resulted in retention of red blood cells, which rendered the vessels visible by virtue of the presence of hemoglobin. The concentration of labeled albumin was measured in a 2-mm surface layer of brain in each of two cranial windows. One window was pretreated with SOD (60 U/ml) plus catalase (40 U/ml) before the induction of a 15-min period of complete global ischemia while the other window was left untreated. We also measured the extravasation of labeled albumin in animals that underwent the same operative procedures but were not subjected to ischemia and therefore served as controls. A second technique was used to measure the permeability of the blood-brain barrier to horseradish peroxidase (mol wt 40,000). The animals received 50 mg/kg of horseradish peroxidase intravenously (Sigma type VI) 5 min before the experimental interventions were begun. This type of peroxidase causes no significant change in the animal’s arterial blood pressure and no alteration in the caliber of pial vessels. After the experimental interventions were performed, the animal’s head was perfused with fixative as described above to eliminate the blood from the cerebral vessels and to fix the brain. The brain was removed and the cortex was serially sectioned at a thickness
Downloaded from www.physiology.org/journal/ajpheart by ${individualUser.givenNames} ${individualUser.surname} (129.186.138.035) on January 5, 2019.
H1358
OXYGEN
RADICALS
IN CEREBRAL
ISCHEMIA
of 40 pm on a Vibratome. The sections were reacted for light microscopic visualization of the protein reaction product through the use of the cobalt glucose oxidase technique (8). Next the sections were mounted on glass slides, dehydrated, and cleared for microscopic analysis. Corresponding anatomic areas from oxygen radical scavenger-treated and untreated windows were compared by a blinded neuroanatomist to assess qualitatively the intensity and localization of the peroxidase reaction product. RESULTS
Superoxide production. Figure 1 shows that superoxide generation was not seen in control animals not subjected to ischemia, during the period of complete ischemia, and 120-135 min of reperfusion after complete ischemia. There was significant superoxide production at O-15 and 60-75 min of reperfusion after complete ischemia. Superoxide production was much greater in the early part of reperfusion than later on. Thus superoxide production after ischemia was time dependent. NBT reduction rates in nanomoles per minute per liter during O-15 min of reperfusion after incomplete ischemia were as follows: NBT alone, 5.42 t 0.86; NBT plus SOD, 3.80 t 0.61; giving a rate of SOD-inhibitable NBT reduction of 1.62 t 0.75. The latter was significantly lower than that seen during the corresponding period after complete ischemia. Vascular caliber and reactivity. Figure 2 shows the changes in arteriolar caliber observed during the period of reperfusion after complete ischemia. Marked vasodilation, seen in the early period of reperfusion, became progressively less pronounced. However, at the end of the 2-h observation, the cerebral arterioles were still significantly dilated when compared with their caliber before ischemia. Figure 3 shows that the dilation of cerebral arterioles at 60 min of reperfusion after complete ischemia was significantly reduced in the windows pretreated with SOD plus catalase or with deferoxamine when compared with the untreated side. 10 T
.-$8 -1 b cl
.\2 : 54
I
NET
EE3 eZa
NBT + SOD DIFFERENCE
I
I \c
CONTROL
ISCHEMIA
REPERFUSION o-15
60-75
(MN) 120-135
Fig. 1. Superoxide production, measured as superoxide dismutase (SOD)-inhibitable reduction of nitro blue tetrazolium (NBT), in control animals not subjected to complete ischemia (n. = 4), during ischemia (n = 4), and at O-15 (n = 6), 60-75 (n = 7), and 120-135 (n = 6) min after complete ischemia. Columns, means k SE of NBT reduction with and without SOD and difference between them. Difference, SOD-inhibitable NBT reduction and a measure of superoxide production. No superoxide was detected under control conditions, during ischemia, and late in reperfusion period (120-135 min). Significant superoxide production (P < 0.05) was found at O-15 and 60-75 min during reperfusion after ischemia.
o0
30 TIME AFTER
Fig. 2. Arteriolar diameter after complete ischemia. chemic diameters (means significantly dilated (P < line diameters from which
60 REPERFUSION
120
90 (min)
changes at various times during reperfusion Values are %change compared with preis75 vessels). Vessels are t SE, n = 11 animals, 0.001) throughout reperfusion period. Base%changes were calculated are 98.2 k 4.8 pm.
a m
SOD + CP.rALASE DEFEROXAMINE
b
$ 20 .-6 ; 10 .a 0 UNTREATED
TREATED
Fig. 3. Effect of treatment with SOD + catalase or with deferoxamine on change in arteriolar diameter after complete ischemia. Results are %change measured 60 min after ischemia compared with preischemic values (means + SE; SOD + catalase: n = 7 animals, 42 vessels; deferoxamine: n = 5 animals, 36 vessels). Treatment with SOD + catalase or with deferoxamine resulted in significantly reduced vasodilation (P < 0.001) after ischemia compared with untreated vessels. Baseline diameters from which %changes were calculated are displayed above columns (means k SE).
Figure 4 shows the response to topical acetylcholine before and during the period of reperfusion after complete ischemia. The normal vasodilator response to acetylcholine seen before ischemia was converted to a vasoconstriction during the early phase of reperfusion. This abnormal vasoconstrictor response persisted until 60 min of reperfusion when the response to acetylcholine reverted back to vasodilation. At the end of the observation period of 2 h the response was vasodilator, but it was significantly less than that seen before ischemia. Figure 5 shows that at 60 min of reperfusion after complete ischemia, the response to acetylcholine was vasoconstrictor in the untreated window, whereas pretreatment with SOD and catalase resulted in a retained vasodilator response to acetylcholine. Figure 6 shows that the vasoconstrictor responses to hypocapnia were significantly reduced at 60 min of reperfusion after complete ischemia and that this effect was inhibited by pretreatment with SOD plus catalase or by deferoxamine. Figure 7 shows that the permeability index to albumin increased considerably in animals subjected to ischemiareperfusion without pretreatment. Pretreatment with
Downloaded from www.physiology.org/journal/ajpheart by ${individualUser.givenNames} ${individualUser.surname} (129.186.138.035) on January 5, 2019.
OXYGEN
RADICALS
IN CEREBRAL
H1359
ISCHEMIA BEFORE
ISCHEMIA
AFTER
ISCHEMIA
15, G
10
.-L -G
5
:: n
EL
25 b
55 . .
-I
1
O--
y -5i *: -lO.-c a -15-
E
.-0 u .-c
a
I
--15!
1
I/‘//I
pre-iachemic control
I
0
1I I 1 I 1 1 30 60 90 TIME AFTER REPERFUSION (MIN)
1 120
Fig. 4. Change in arteriolar diameter caused by topical lop7 M acetylcholine application before ischemia and at various times during reperfusion after complete ischemia. Results are %change compared with diameters immediately before acetylcholine application (means t SE; n = 8 animals, 45 vessels). Acetylcholine-induced vasoconstriction occurred during first 60 min of reperfusion. Significant vasodilation occurred with acetylcholine after 120 min of reperfusion but was reduced when compared with response seen before ischemia (P c 0.005). Baseline diameters from which %changes were calculated are displayed above or beside columns (means t SE). BEFORE
25 7.5 :: n
AFTER
deferoxamine
sod + catalase
catalase
deferoxamine
Fig. 6. Effect of treatment with SOD + catalase or with deferoxamine on change in arteriolar diameter in response to hyperventilation before and after ischemia. Values are %change from diameters immediately before hyperventilation (means t SE; SOD + catalase: n = 7 animals, 42 vessels; deferoxamine: n = 5 animals, 36 vessels). Treatment has no effect on vasoconstrictor response to hyperventilation before ischemia. After ischemia, treated vessels responded normally to hyperventilation, whereas untreated vessels had a significantly (P < 0.001) attenuated response. Baseline diameters from which %changes were calculated are displayed above columns (means t SE). 1.00
ISCHEMIA
20
z
.90 T
=
.00-
p ‘5
.70AO-
T
15
z 28
lo
&I .-i -u c .a
ISCHEMIA
syd
I
untreate
5 CONTROL
-0 -5 -10 untreated
sod + catalase
untreated
sod + catalase
Fig. 5. Effect of pretreatment with SOD + catalase on change in arteriolar diameter in response to topical 10ey M acetylcholine application at 60 min of reperfusion after complete ischemia. Results are %change compared with diameter immediately before acetylcholine application (means k SE; n = 5 animals, 28 vessels). SOD + catalase had no effect on acetylcholine responses before ischemia. Untreated vessels constricted in response to acetylcholine application after ischemia. Treated vessels dilated in response to acetylcholine after ischemia, but response was attenuated when compared with preischemic responses (see also Fig. 4). Baseline diameters from which %changes were calculated are displayed above columns (means 2 SE).
SOD and catalase reduced this effect but did not eliminate it. Extravasation of horseradish peroxidase was observed consistently within the brain parenchyma underlying the cranial windows in all animals subjected to ischemiareperfusion. The altered cerebral vascular permeability was seen in relation to the crest of the gyri as well as in the deeper brain structures. Oxygen radical scavengers did not completely prevent the permeability change induced by ischemia-reperfusion but reduced it significantly (Fig. 8). The brain parenchyma exposed to topically applied SOD plus catalase showed easily recognizable reduction in permeability when compared with the
ISCHEMIA + SOD + CATALASE
ISCHEMIA UNTREATED
Fig. 7. Effect of treatment with SOD + catalase on changes in permeability to albumin after ischemia (n = 5 per each group). Values are means +: SE of permeability index, which is ratio of radiolabeled albumin content in fixed brain to that in whole blood (x100) measured 60 min after ischemia or in control period. Treatment with SOD + catalase reduced but did not completely prevent increase in permeability index seen after ischemia.
untreated side. SOD alone (60 U/ml) had a consistent but small effect on the increased permeability induced by ischemia-reperfusion. SOD in a concentration of 120 U/ml had a more pronounced effect. Catalase (80 U/ml) alone had no detectable effect on the increased permeability from ischemia-reperfusion. DISCUSSION
The important findings of these experiments are as follows. 1) Superoxide generation occurs during the period of reperfusion after either complete or incomplete cerebral ischemia. There is no superoxide production under resting conditions in the absence of ischemia or during the period of complete cerebral ischemia. Superoxide generation is much greater after complete rather than incomplete ischemia, and superoxide generation in the former condition lasts for >l h. 2) The cerebral arteriolar changes induced by ischemia-reperfusion are characterized by sustained dilation, reduced responsiveness to the vasoconstrictor effects of arterial hypocapnia, reversal of
Downloaded from www.physiology.org/journal/ajpheart by ${individualUser.givenNames} ${individualUser.surname} (129.186.138.035) on January 5, 2019.
Fig. 8. These light micrographs demonstrate altered cerebral vascular permeability to horseradish peroxidase occurring after ischemia with reperfusion in presence of various concentrations of SOD and/or catalase. In A, note that combined use of 120 U/ml SOD + 80 U/ml catalase reduced overall cortical extravasation of peroxidase (*) compared with untreated controls (B, arrows). Note that in both control and treated hemispheres, permeability change deep within brain parenchyma (arrowheads) is unaffected by SOD-catalase treatment. This most likely reflects inability of employed agents to diffuse rapidly to deep cortical sites. With use of 120 U/ml SOD, a marked reduction in altered cerebrovascular permeability to horseradish peroxidase (C, *) is seen compared with control hemisphere (D, arrows). Note again that use of this superoxide anion scavenger did not affect permeability change deep within cortex. Downloaded from www.physiology.org/journal/ajpheart by ${individualUser.givenNames} ${individualUser.surname} (129.186.138.035) on January 5, 2019.
OXYGEN
RADICALS
IN CEREBRAL
the endothelium-dependent vasodilator response to acetylcholine, and disruption of the blood-brain barrier to proteins. These changes are induced by oxygen radicals because they are inhibited by pretreatment with specific scavengers. In the case of the sustained vasodilation and the abnormal response to hypocapnia after ischemia-reperfusion, we identified hydroxyl radical as the agent responsible for the abnormalities. Unlike what occurs in hearts in which the production of oxygen radicals during reperfusion after ischemia is brief (lo), in the in vivo blood-perfused brain superoxide production is prolonged, lasting for >l h. This is similar to what is seen in other experimental pathophysiological conditions such as acute hypertension and fluid-percussion brain injury, in which superoxide production lasts for at least 1 h after the insult (7, 22). The prolonged generation of superoxide in ischemia-reperfusion may have practical significance because it affords a better opportunity for therapeutic intervention. The sites of generation of superoxide and its enzymatic sources were not investigated in this study. In another investigation, we used a histochemical technique to localize superoxide in the same model used in the present experiments (5). We found that the production of superoxide occurred in endothelial and smooth muscle cells of the cerebral arterioles, and we also localized superoxide in the extracellular space in proximity to the blood vessels. We showed earlier that superoxide escapes into the extracellular space from the ceils of the vessel wall via the anion channel (8). In this histochemical investigation (5) we found no superoxide production in the brain parenchyma, but its absence there may be due to incomplete penetration of the reagents that were applied topically on the brain surface. The enzymatic sources of superoxide during reperfusion after brain ischemia have not been identified. As noted at the outset of this article, the available evidence with respect to xanthine oxidase is conflicting. Armstead et al. (1) found that indomethacin inhibited superoxide production during the early part of reperfusion after ischemia in newborn pigs. Therefore they concluded that cyclooxygenase was the source of superoxide. We have shown previously that cyclooxygenase generates superoxide in vitro (8, 11). It should be noted, however, that the rate of SOD-inhibitable reduction of NBT reported by Armstead et al. (1) is -100 times greater than what we found. The microvascular abnormalities seen during reperfusion after complete ischemia, including sustained arteriolar dilation, abnormal reactivity to hypocapnia, reversal of the endothelium-dependent vasodilation response to acetylcholine and increased permeability of the bloodbrain barrier to proteins, are clearly due to the generation of oxygen radicals because they were inhibited strongly by pretreatment with oxygen radical scavengers. We did not investigate the role of radicals in the more pronounced reactive dilation that occurs in the earlier phases of reperfusion. Although the time course of the dilation is similar to the time course of generation of superoxide, suggesting a causal relationship, it is likely that this earlier dilation is multifactorial. We found earlier that release of
ISCHEMIA
H1361
polypeptides from sensory fibers is involved (19). It is also likely that adenosine may contribute because its concentration during ischemia rises markedly (23). The time course of the abnormal response to acetylcholine correlates with that of the superoxide production, suggesting that the two are causally related. This is further supported by the inhibition of the abnormality after pretreatment with SOD and catalase. It is well known that oxygen radicals attack directly and inactivate the endothelium-derived relaxing factor from acetylcholine (3, 9, 18). We believe this is the mechanism involved in view of the spontaneous return of the vasodilator response to acetylcholine 2 h after reperfusion. Had the mechanism been irreversible damage of the endothelium, we would have expected that this would not have occurred. From the practical standpoint, the response to acetylcholine may be used to gauge the time course of superoxide production during reperfusion after ischemia. It is likely that the increase in baseline vascular caliber caused by oxygen radicals also may have contributed to the reduction in the vasodilator responses to acetylcholine by nonspecific mechanisms. However, this factor cannot fully explain the findings because it cannot explain the reversal of the response from vasodilator to vasoconstrictor. Other investigators have noted inconsistent responses to acetylcholine application after cerebral ischemia-reperfusion in cats (15). They started measuring responses after 60 min of reperfusion and saw both dilation and constriction. Our data show that 60 min is approximately the time that responses return toward normal; therefore heterogeneous findings would not be unexpected. Their period of ischemia was shorter than in our studies, which would affect the time course to recovery, because the amount of reperfusion injury is related to the intensity of initial ischemia. The response to hypocapnia in large cerebral vessels is not endothelium dependent (20). It would appear, therefore, that the inhibition of the vasoconstrictor response to hypocapnia observed in the present experiments after ischemia is due to effects of oxygen radicals directly on vascular smooth muscle. Because this effect was inhibited by both SOD plus catalase and by deferoxamine, we attributed it to the direct effects of hydroxyl radical. This radical is very reactive and short-lived. Consequently, it does not survive more than a few molecular diameters from its site of formation. Because it is accessible to deferoxamine, SOD and catalase, which are unlikely to enter into the interior of cells easily, it is likely that its action is on the cell membrane. It has been reported by others that oxygen radicals can damage membrane enzyme systems and affect ion fluxes leading to permeability changes after different types of cerebral injury (14, 16). Our findings show that the increased permeability of the blood-brain barrier to proteins during reperfusion after ischemia is mediated by oxygen radicals. The precise mechanism by which this occurs was not identified. It appears from the use of different scavengers that superoxide is the agent immediately responsible for the increased permeability because SOD, but not catalase alone, was effective in minimizing
Downloaded from www.physiology.org/journal/ajpheart by ${individualUser.givenNames} ${individualUser.surname} (129.186.138.035) on January 5, 2019.
H1362
OXYGEN
RADICALS
IN CEREBRAL
the response. The increased permeability to proteins coupled with the sustained arteriolar dilation, which should lead to increased capillary pressure, would be expected to contribute to the induction of edema with secondary adverse effects on brain parenchymal function. This research was supported by National Institutes of Health Grants HL-21851, NS-19316, NS-26361, and HL-07537. Address for reprint requests: H. A. Kontos, Box 662, MCV Station, Medical College of Virginia, Richmond, VA 23298. Received
31 March
1992; accepted
in final
form
16 June
1992.
REFERENCES 1. Armstead, W. M., R. Mirro, D. W. Busija, and C. W. Leffler. Postischemic generation of superoxide anion by newborn pig brain. Am. J. Physiol. 255 (Heart Circ. Physiol. 24): H401-H403, 1988. 2. Betz, A. L., J. Randall, and D. Martz. Xanthine oxidase is not a major source of free radicals in focal cerebral ischemia. Am. J. Physiol. 260 (Heart Circ. Physiol. 29): H563-H568, 1991. 3. Gryglewski, R. J., R. M. J. Palmer, and S. Moncada. Superoxide anion is involved in the breakdown of endothelium-derived vascular relaxing factor. Nature Lond. 320: 454-456, 1986. 4. Itoh, K., A. Konishi, S. Nomura, N. Mizuno, Y. Nakamura, and T. Sugimoto. Application of coupled oxidation reaction to electron microscopic demonstration of horseradish peroxidase: cobalt-glucose oxidase method. Brain Res. 175: 341-346, 1979. 5. Kontos, C. D., E. P. Wei, J. I. Williams, H. A. Kontos, and J. T. Povlishock. Cytochemical detection of superoxide in cerebral inflammation and ischemia. Am. J. Physiol. 263 (Heart Circ. Physiol. 32): H1234-H1242, 1992. 6. Kontos, H. A. Oxygen radicals in CNS damage. Chem. Biol. Interact. 72: 229-255, 1989. 7. Kontos, H. A., and E. P. Wei. Superoxide production in experimental brain injury. J. Neurosurg. 64: 803-807, 1986. 8. Kontos, H. A., E. P. Wei, E. F. Ellis, L. W. Jenkins, J. T. Povlishock, G. T. Rowe, and M. L. Hess. Appearance of superoxide anion radical in cerebral extracellular space during increased prostaglandin synthesis in cats. Circ. Res. 57: 142-151, 1985. 9. Kontos, H. A., E. P. Wei, J. T. Povlishock, R. C. Kukreja, and M. L. Hess. Inhibition by arachidonate of cerebral arteriolar dilation from acetylcholine. Am. J. Physiol. 256 (Heart Circ. Physiol. 25): H665-H671, 1989. 10. Kramer, J. H., C. M. Arroyo, B. F. Dickens, and W. B. Weglicki. Spin-trapping evidence that graded myocardial ischemia alters post-ischemic superoxide production. Free Radical Biol. Med. 3: 153-159, 1987.
ISCHEMIA
11. Kukreja, R., H. A. Kontos, M. L. Hess, and E. F. Ellis. PGH synthase and lipoxygenase generate superoxide in the presence of NADH or NADPH. Circ. Res. 59: 612-619, 1986. 112. Levasseur, J. E., E. P. Wei, A. J. Raper, H. A. Kontos, and J. L. Patterson, Jr. Detailed description of a cranial window technique for acute and chronic experiments. Stroke 6: 308-317, 1975. 13. Liu, T. A., J. S. Beckman, B. A. Freeman, E. L. Hogan, and C. Y. Hsu. Polyethylene glycol-conjugated superoxide dismutase and catalase reduce ischemic brain injury. Am. J. Physiol. 256 (Heart Circ. Physiol. 25): H589-H593, 1989. 14. Lo, W. D., and A. L. Betz. Oxygen free-radical reduction of brain capillary rubidium uptake. J. Neurochem. 46: 394-397, 1986. W. G., S. M. Amundsen, F. M. Faraci, and D. D. 15. Mayhan, Heistad. Responses of cerebral arteries after ischemia and reperfusion in cats. Am. J. Physiol. 255 (Heart Circ. Physiol. 24): H879H884, 1988. 16. Palmer, G. C. Free radicals generated by xanthine oxidase-hypoxanthine damage adenylate cyclase and ATPase in gerbil cerebral cortex. Metab. Brain Dis. 2: 243-257, 1987. 17. Patt, A., A. H. Harken, L. K. Burton, T. C. Rodell, D. Piermattei, W. J. Schorr, N. B. Parker, E. M. Berger, I. R. Horesh, L. S. Terada, S. L. Linas, J. C. Cheronis, and J. E. Repine. Xanthine oxidase-derived hydrogen peroxide contributes to ischemia reperfusion-induced edema in gerbil brains. J. Clin. Invest. 81: 1556-1562, 1988. G. M., and P. M. Vanhoutte. Superoxide anions and 18. Rubanyi, hyperoxia inactivate endothelium-derived relaxing factor. Am. J. Physiol. 250 (Heart Circ. Physiol. 19): H822-H827, 1986. 19. Sakas, D. E., M. A. Moskowitz, E. P. Wei, H. A. Kontos, M. Kano, and C. S. Ogilvy. Trigeminovascular fibers increase blood flow in cortical gray matter by axon reflex-like mechanisms during acute severe hypertension or seizures. Neurobiology CPH. 86: 1401-1405, 1989. 20. Toda, N., Y. Hatano, and K. Mori. Mechanisms underlying response to hypercapnia and bicarbonate of isolated dog cerebral arteries. Am. J. Physiol. 257 (Heart Circ. Physiol. 26): H141-H146, 1989. 21. Watson, B. D., R. Busto, W. J. Goldberg, M. Santiso, S. Yoshida, and M. D. Ginsberg. Lipid peroxidation in vivo induced by reversible global ischemia in rat brain. J. Neurochem. 42: 268-274, 1984. 22. Wei, E. P., H. A. Kontos, C. W. Christman, D. S. Dewitt, and J. T. Povlishock. Superoxide generation and reversal of acetylcholine-induced cerebral arteriolar dilation after acute hypertension. Circ. Res. 57: 781-787, 1985. 23. Winn, H. R., R. Rubio, and R. M. Berne. Brain adenosine production in the rat during 60 seconds of ischemia. Circ. Res. 45: 486-492, 1979.
Downloaded from www.physiology.org/journal/ajpheart by ${individualUser.givenNames} ${individualUser.surname} (129.186.138.035) on January 5, 2019.
