Main

OFR have been implicated in postischemic injury in several organs(1–3) including the brain(4–6). Attempts to measure radical formation directly include a report from Zini et al.(7) who demonstrated the formation of a radical adduct in the microdialysate fluid from the brain of adult rats during ischemia and a considerable increase during reperfusion. Armstead et al.(8) have shown that superoxide radical is produced during postischemic reperfusion in the newborn pig, and recently Hasegawa et al.(9) demonstrated an increased free radical density in the neonatal mouse brain during posthypoxic recovery. The situation during oxygen deficiency/reoxygenation in the fetal brain may be grossly different from the adult brain for two reasons: the fetal and immature brain has considerably greater tolerance to oxygen depletion than the adult, mature brain(10, 11), and the reoxygenation occurs at much lower oxygen tensions (arterial Po2 being at most 4 kPa in the fetus compared with 13 kPa in the adult).

Sources for the formation of OFR during reperfusion include the oxidation of hypoxanthine in the presence of xanthine oxidase and of arachidonic acid in the presence of lipoxygenase and cyclooxygenase. Indeed, studying the fetal sheep we have shown that hypoxanthine was produced by the brain at the degree of oxygen deficiency that caused the somatosensory evoked potentials to disappear(12) and that interstitial concentrations of hypoxanthine were quadrupled during severe fetal asphyxia(13). But is the low arterial oxygen tension high enough to support the formation of OFR?

Evidence for formation of free radicals in the prenatal or neonatal circulation is indirect(14). It is related to the reduction of the antioxidant capacity of the tissue during oxygen deficiency or the accumulation of metabolites that are known to produce OFR when oxidized(13, 15). It also may be related to the fact that postischemic injury can be ameliorated by treatment aimed at preventing the formation of OFR(16) by supplying extra antioxidants or radical scavengers(17–20).

The direct demonstration and quantification of OFR is rendered impossible by the extreme reactivity and resulting short biologic half-life of these species. However, short lived radicals can be made detectable by the use of chemical substances that form secondary stable radicals upon reaction with e.g. oxygen-derived free radicals. When such a spin trap is applied, the possibility of quantifying the amount of radicals produced by the use of EPR spectrometry is at hand. In this study we applied a modification of the spin trap methodology(21) and EPR to investigate the possible formation of free radicals in the brain of the anesthetized fetal sheep during ischemia/reperfusion.

METHODS

Eleven pregnant ewes of mixed breed with gestation between 135 and 138 d were anesthetized with sodium pentothal 500 mg and chloralose 35 mg/kg i.v. and maintained with repeated doses of chloralose. The ewes were ventilated with extra oxygen to keep arterial blood gases between Po2 12-15 kPa and Pco2 4-5 kPa.

Fetuses were exteriorized to a thermostat-controlled, heated table, keeping body temperature at 38-39°C. The uterus was marsupialized to the abdominal wall, and only an ample opening for the umbilical cord was left unsutured. The cord was wrapped in saline-soaked gauze covered with a thin plastic foil and kept at 39°C. The right brachial artery and vein were cannulated to allow i.v. injections and the recording of blood pressure and sampling of preductal arterial blood for blood gas analysis.

The neck of the fetus was opened by a midline incision, and the carotid arteries were dissected free for a length of at least 3 cm, including the portion close to the base of the skull where the communicant from the vertebral artery enters the carotid artery(22). Vascular connections with the vertebral arteries were tied as suggested by Williams et al.(23), thus converting the circulation of the brain into a two-vessel system. The left lingual artery was cannulated with a catheter (PE50) in the retrograde direction so the tip of the catheter was left in the mainstream of the carotid artery. This catheter was used for the intraarterial infusion of spin trap. One flowmeter probe(Transonic, 6 mm) was placed around each carotid artery together with a loop of thin rubber tubing to enable reversible and complete occlusion of both carotids. An indwelling butterfly needle (G23) was placed in the sagittal sinus and glued in place to the bone of the skull. This needle with its catheter was connected to a three-way stopcock to allow intermittent sampling of venous blood from the brain and recording of sinus sagittal pressure. Venous and arterial blood from the nonheparinized animals were sampled in heparinized syringes.

Procedure. When the preparation was finished, a 30-min period of stabilization was allowed. Thereafter, baseline values for radical production were collected, and the control situation was considered stable when three to four consecutive samples, 5 min apart, were equal. Thus, the postsurgery recovery period always extended at least 60 min and during that time four to six samples were analyzed. Thereafter a 30-min period of complete cerebral ischemia was instituted in six animals by tying the rubber loops around the carotids. Five animals served as control fetuses. These animals underwent the same preparation and surgical procedures as the other six animals, but they were never subjected to carotid artery occlusion. After exactly 30 min ischemia was discontinued, and new blood samples from the sagittal sinus were collected at 2, 4, 6, 10, 20, 30, 40, 60, and 90 min. Samples for blood gases were obtained from the brachial artery at the same time intervals and analyzed for oxygen saturation, Hb concentration, Po2, Pco2, and pH using a Radiometer ABL 550 blood gas analyzer.

Spin trap synthesis and EPR analysis. OXANO˙ was synthesized according to previously described methods(24). The corresponding hydroxylamine spin trap OXANOH was prepared immediately before each experiment by gassing a 10 mM solution of OXANO˙ with hydrogen(25) and kept on ice until used.

The principle of the method is to infuse the hydroxylamine OXANOH into the arterial circulation of the brain at a constant rate and measure how much has been converted into the stable nitroxide radical OXANO˙, by analyzing the venous blood. In these experiments an intraarterial infusion rate of 1.65 mL/min was chosen, and intermittent infusions were given 2 min before sampling. Thereby a stable spin trap concentration was obtained in both the perfusing blood and the tissue(26). The OXANO˙ content of the blood from the sagittal sinus was measured within 30 s after blood sampling with a Bruker ECS 106 EPR spectrometer. Spectrometer settings were as follows: field center, 3478.5 Gauss; modulation amplitude, 1.0 Gauss; microwave power, 10 mW; microwave frequency, 9.74 GHz; scan range, 5 Gauss; scan rate, 60 Gauss/min; time constant, 0.02 s.

The area under the absorption peaks in the EPR spectra was compared with that of a known standard, giving the concentration of OXANO˙ in the venous blood. This concentration depends on the intraarterial infusion rate of OXANOH and the carotid blood flow rate at the time when the sample is taken. Therefore, the concentration of OXANO˙ in the sagittal sinus blood was multiplied by the combined carotid blood flows per 100 g of brain to obtain the release rate of OXANO˙ from the tissue (mmol × min-1× 100 g-1). This parameter is thus directly proportional to the formation of radicals in the brain and is independent of cerebral blood flow(26).

After the experiment the fetus was weighed, and the crownrump length was measured. The brain was dissected out, including the medulla oblongata, and weighed.

Statistical analysis. Mean percentage increase in reperfusion values of OXANO˙ release compared with baseline values were analyzed using a Mann-Whitney U test with Bonferroni adjustment. Significance was set at p < 0.05. Differences in OXANO˙ release at specific time intervals between animals subjected to cerebral ischemia and sham-operated control animals were analyzed using the Mann-Whitney U test. Commercial software (Statistica/w, StatSoft, Tulsa, OK) was used for all calculations.

RESULTS

The 11 fetuses studied weighed 3.9 ± 0.4 kg (mean ± SEM) and had a mean brain weight of 43.3 ± 2.8 g. Mean arterial blood pressure and heart rate in animals subjected to 30 min of cerebral ischemia are displayed in Figure 1, demonstrating a sustained increase of both variables during carotid occlusion and return to preocclusion values immediately on release of the occlusion. During the first 90 min of reperfusion a slight increase of the heart rate and small changes of mean arterial blood pressure are seen. Oxygen tension and pH values in the brachial artery and in the sagittal sinus during the experiment are demonstrated in Table 1.

Figure 1
figure 1

Heart rate (HR) and mean arterial pressure(MAP) during baseline, cerebral ischemia, and reperfusion period in animals subjected to 30 min of cerebral ischemia (n = 6) and in sham-operated control animals (n = 5). Values are mean ± SEM.

Full size image
Table 1 pH and Po2 levels during baseline,“ischemia,” and reperfusion periods in the brachial artery and the sagittal sinus in sham-operated control animals (n = 5) and animals subjected to 30 min of cerebral ischemia (n = 6)
Full size table

The release of OXANO˙ from the brain during basal conditions was 0.26± 0.06 mmol × min-1 × 100 g-1. The rate of release varied greatly between experiments but was constant within 25% when repeated measurements in the same animal were performed.

A more than 2-fold increase of OXANO˙ took place, reaching a maximum at 10 min of reperfusion in the animals subjected to 30 min of cerebral ischemia (Fig. 2). A shortlived, moderate increase of OXANO˙ was also seen during the initial phase of “reperfusing” in the control animals (Fig. 3). Figure 2 also gives the blood flow in the carotid arteries and demonstrates a short lasting reactive hyperemia that has already waned at 6 min of reperfusion. The increase of OXANO˙ release was significantly higher in animals subjected to cerebral ischemia compared with sham-operated control animals (Fig. 3).

Figure 2
figure 2

OXANO˙ release from the brain is shown together with the combined blood flow in the two carotid arteries during the reperfusion period. Release of OXANO˙ from the brain is normalized to 100% before start of carotid occlusion (mean of three consecutive measurements). Values are mean± SEM; n = 6; * = p < 0.05. The postischemic OXANO˙ values are compared with the baseline value.

Full size image
Figure 3
figure 3

OXANO˙ release from the brain in animals subjected to 30 min of cerebral ischemia (n = 6) and in sham-operated control animals (n = 5). Release of OXANO˙ from the brain is normalized to 100% before start of carotid occlusion (mean of three consecutive measurements). Values are mean ± SEM; the asterisk (*) indicates a p value <0.05 when animals subjected to cerebral ischemia are compared with control animals. The dagger (†) indicates a p value <0.05 when “reperfusion” values in control animals are compared with baseline values in the same animals.

Full size image

pH and Po2 levels during baseline, cerebral ischemia, and reperfusion periods in the brachial artery and the sagittal sinus in animals subjected to 30 min of cerebral ischemia and in sham-operated controls are displayed in Table 1.

DISCUSSION

The method used for spin trapping takes advantage of the fact that the hydroxylamine OXANOH traps the spin of the primary radical via one-electron oxidation to the stable nitroxide radical OXANO˙. This compound accumulates to concentrations that allow the detection and quantification with EPR. The method has been validated both in vitro(25, 27, 28) and in vivo(26, 29, 30).

The method uses the constant intraarterial infusion of OXANOH into the arterial circulation of the brain. Thus, measurements during total ischemia were precluded, and instead they began immediately after release of the carotid occlusion. From 4 to 30 min of reperfusion, a significant elevation of OXANO˙ release was measured, peaking at 10 min. This cannot be ascribed to a washout effect of radicals accumulated in the brain during ischemia, both because of the short half-lives of radicals and because the hyperemia was only transient and no time relation between the two phenomena was observed. Thus, at 2 min of reperfusion no change of radical production was observed, although the peak of reactive hyperemia had just passed (Fig. 2). Instead, the results demonstrate a true increase of radical production during the reperfusion period. A moderate but significant increase of OXANO˙ release was also observed during the first 10 min of “reperfusing” in the control animals (Fig. 3). This demonstrates that repeated infusions in short intervals of OXANOH create a transient increase of OXANO˙ concentration, presumably because of autooxidation. Because the spin trap was infused into the blood stream and recovered from the venous effluent of the brain, the exact origin of the radicals cannot be specified. OXANOH passes rapidly over the blood-brain barrier and also over cell membranes(21). Therefore, it may well trap radicals in the brain tissue. However, the possibility that the radicals are formed in the cerebral microcirculation, at the site of the endothelium, cannot be excluded. Zini et al.(7) were the first to use spin trap methodology in vivo to examine the brain during ischemia. They used the microdialysis technique and recovered a radical adduct from the striatum of the adult rat. The increased production rate of radicals during reperfusion must thus have originated in the brain parenchyma in their experiments. Armstead et al.(8) used the superoxide dismutase-inhibitable nitro blue tetrazolium with a closed cranial window technique and showed that the superoxide radical is produced during postischemic reperfusion in the newborn pig. Hasegawa et al.(9) studied EPR spectra in tissue sections of the brains from 1-d-old mice during anoxia and recovery in room air and demonstrated a transient, moderate increase of radical density during 10-20 min of recovery. In both these studies recovery after ischemia or anoxia took place at extrauterine levels of oxygenation. In the fetal lambs the Po2 of the arterial blood perfusing the brain was at a typical fetal level,i.e. 2.5-3.5 kPa (Table 1). Apparently, this low oxygen tension is enough to allow the production of free radicals, whether derived from oxygen or not.

Postischemic activation of circulating neutrophils with subsequent oxygen radical production may also mediate brain injury. Matsuo et al.(31) demonstrated that anti-neutrophil MAb inhibited the increase in radical production in adult rats subjected to 1 h of middle cerebral artery occlusion.

During and immediately after cerebral ischemia some characteristic hemodynamic changes took place (Fig. 1). The elevation of both mean arterial blood pressure and heart rate during ischemia should be ascribed to the fact that both carotid arteries were occluded, suddenly taking the load off the carotid baroreceptors. The transient augmentation of the blood flow in the carotid arteries during reperfusion is ascribed to the fact that the ischemia was complete and suggests that metabolism in the brain ceased very rapidly after cessation of blood supply with only a small accumulation of lactate. This is in line with findings in the rat brain during complete and incomplete ischemia(32).

These experiments thus demonstrate an increased release of free radicals from the fetal brain during the phase of reperfusion after ischemia in spite of the fact that reoxygenation occurs at fetal levels of oxygen tension.