Main

Hypoxic-ischemic cerebral damage is an important contributor to perinatal mortality and morbidity including long-term neurologic sequelae in term and preterm fetuses (1). Although over the last decade many therapeutic strategies have been developed in various animal models to reduce neuronal cell damage caused by ischemic insults, their clinical application has often been rejected due to major drug-related side-effects (2). Especially after application of calcium antagonists and inhibitors of NOS, alterations of fetal cardiovascular control have been observed (3, 4). However, the recently developed benzothiazole compound lubeluzole seems to justify further research in this field. The neuroprotective efficiency of this drug has been shown in a variety of in vitro as well as in vivo experiments (510). In clinical safety studies, severe drug-related side-effects could be excluded (11). Phase III trials have provided different results. Whereas Grotta (12) described a neuroprotective effect of lubeluzole in patients suffering from acute ischemic stroke, this finding could not be confirmed in other trials (13, 14).

The present knowledge on the safety profile and efficiency of lubeluzole seems to be promising enough to warrant further studies in immature animals. Recently, we were able to show that lubeluzole does not affect fetal circulatory centralization during acute asphyxia. This mechanism is of major importance, because it protects the fetal brain from neuronal injury by increasing blood flow to the central organs, i.e. brain, heart, adrenals, when oxygen is in short supply (15). The present study was performed to clarify whether lubeluzole protects the brain of fetal sheep near term from neuronal cell damage. Global cerebral ischemia was induced by occluding both carotid arteries for 30 min. To account for drug-related changes in cerebral blood flow that might affect neuronal cell damage, we measured cerebral blood flow using the microspheres method.

METHODS

Animal preparation.

Thirteen fetal sheep were prepared at a mean gestational age of 127 ± 1 d (term is at 147 d). All ewes were anesthetized by subarachnoid injection of 8 mL of 0.75% (wt/vol) bupivacaine at the lower spine, and were operated under sterile conditions. Polyvinyl catheters were placed in a maternal iliac artery and vein through tibial vessels. The ewe's abdominal wall was opened along the midline and the fetal hindlimbs were exposed through a small uterine incision. Using local anesthesia (prilocaine HCl 1.0%), polyvinyl catheters were inserted via the pedal vein of each hindlimb into the inferior vena cava. The uterine incision was closed and a second uterine incision was performed over the fetal snout. Head and neck of the fetus were exteriorized. To prevent the fetus from breathing, its head was covered by a water-filled rubber glove. Catheters were inserted into the fetal ascending aorta via both brachial arteries. Furthermore, both fetal common carotid arteries were prepared. Cerebral ischemia was induced by occluding the carotid arteries on both sides simultaneously below the thyroid and above the lingual artery for 30 min as described recently (16, 17). Thus, blood flow to the cerebrum via anastomoses between the carotid and vertebral arteries was arrested. After cerebral ischemia, a catheter was inserted into the amniotic cavity, and the second intrauterine incision was closed. All catheters were filled with heparin (1000 IU/mL), plugged, and passed subcutaneously to the ewe's flank, where they were exteriorized and protected by a pouch sewn to the skin. On the day of surgery and each day thereafter, the ewe received 2 million units of penicillin G (Grünenthal, Aachen, Germany) and 80 mg of gentamicin sulfate (Merck, Darmstadt, Germany), half intravenously and half into the amniotic cavity.

Experimental protocol.

After insertion of polyvinyl catheters into the inferior vena cava, six fetuses received three bolus injections of lubeluzole at 30 min intervals (3 × 0.11 mg/kg estimated fetal body weight), while an equal volume (3 × 1.3 mL) of the solvent was administered to the remaining fetuses. The drugs were donated by Janssen Pharmaceutica, Beerse, Belgium. The concentration of lubeluzole used in the present study was within the same range as that applied in previous investigations. In humans, the injected dose ranged from 2.5 to 25 mg/d, i.e. about 0.036 to 0.36 mg/kg/d (12, 13, 18), whereas in rats lubeluzole was applied at a concentration of between 0.62 and 0.8 mg/kg (9, 10, 19). In vitro studies have even shown that lubeluzole has neuroprotective effects down to the picomolar range (molecular weight: 43,352 g/mol) (7). In preliminary experiments, we applied lubeluzole to fetal sheep by a single i.v. injection at a concentration of 0.33 mg/kg. This resulted in alterations of FHR as well as arterial hypotension. We therefore considered the dosage applied in the present study to be at the upper limit of the therapeutic range in fetal sheep near term. A further increase in the administered concentration might have altered fetal cardiovascular control and would therefore have interfered with a possible neuroprotective effect of lubeluzole.

Cerebral ischemia was induced 30 min after the final injection of lubeluzole. To determine the time course of changes in fetal cerebral blood flow, six batches of differently isotope-labeled microspheres (113Sn, 103Ru, 46Sc, 114In, 141Ce, and 95Nb, 16 μm diameter, New England Nuclear, Boston, MA, U.S.A.) were used. The microspheres, suspended in 10% dextran containing 0.01% Tween 80, were sonicated and checked for size, shape, and aggregation. Depending on the specific activity, 1.2–1.8 million microspheres per batch were injected into the inferior vena cava before (−90 min), during (+3 min and +27 min), and after (+40 min, +3 h, and +72 h) 30 min of cerebral ischemia. The number of injected microspheres was large enough to ensure both an adequate number of microspheres per sample and valid blood flow measurements during cerebral ischemia (20, 21). Specific calculations revealed that about 400 microspheres were trapped in low-flow cerebral areas during ischemia. Thus, for theoretical considerations, the blood flow estimates in these areas are within 5% of the true values (20). During and shortly after injection of the microspheres no significant changes in FHR or arterial blood pressure were found. Reference blood samples were withdrawn from the brachial artery at a rate of 2.5 mL·min−1 for 90 s (16, 17). The volume of blood withdrawn was about 22.5 mL and was simultaneously replaced by maternal blood maintained at 39°C in a water bath.

During and shortly after cerebral ischemia FHR, ascending aortic and intrauterine pressure were recorded simultaneously. After having injected the fourth batch of microspheres (at +40 min) all catheters were closed and secured as described above, the abdominal wall was closed and the ewe was brought back to the metabolic cage, where the fifth injection (at +3 h) was performed. For technical reasons, there were no pressure measurements at this point in time. Before each injection, blood samples were obtained from the brachial artery to measure blood gases, So2, and acid-base balance. At the end of the experiment (at +72 h), the ewe was given a lethal dose of sodium pentobarbitone and saturated potassium chloride intravenously, and the fetus was perfused with 300 mL of formalin (15%, wt/vol, saline).

The experimental protocols were approved by the appropriate institutional review committee and met the guidelines of the governmental agency responsible.

Measurements.

Ascending aortic and intrauterine pressure, and FHR were recorded on a polygraph (Hellige, Freiburg, Germany). Blood gases and pH were measured in an automatic blood gas analyzer (278 Blood Gas System, Ciba Corning, Frankfurt, Germany), and BE was calculated. Hb concentration and oxygen saturation of Hb were measured photometrically (OSM 2 Hemoximeter, Radiometer, Copenhagen, Denmark) in duplicate.

Fetal cerebral blood flow and the extent of neuronal cell damage were assessed in identical brain specimens. To determine fetal cerebral blood flow using the microspheres method, the fetal brain was removed and fixed in formalin for at least 7 d. Afterward, the cerebrum was separated from the basal ganglia and divided into four frontal sections (rostral, pericentral, postcentral, occipital) with a thickness of about 1.5 cm. The right and left parts of these four sections were further subdivided into four equal segments (parasagittal 1 and 2, lateral 1 and 2) each weighing 1–2 g. In addition to these 32 cerebral specimens, caudate nucleus, thalamus, hippocampus, tegmentum-colliculi-pons, cerebellum, and medulla oblongata were separated. These brain structures were placed in vials, which were filled to the same height to reduce variations in geometry during gamma counting.

After cerebral blood flow analysis, the specimens of the fetal brain were removed from the counting vials and embedded in paraffin. Coronal subserial sections of 10 μm were obtained and stained with cresyl violet/fuchsin. Every 40th section was mounted to evaluate the extent of neuronal cell damage. Neuronal cell damage was assessed at a magnification of 250×. Neurons with ischemic cell damage were identified according to the criteria of Brown and Brierley (22). Neuronal cell damage in each microscopical visual field was quantified by the following score: 0–5% damage (score 1), 5–50% damage (score 2), 50–95% damage (score 3), 95–99% damage (score 4), and 100% damage (score 5). In the present investigation, a scoring system was preferred, inasmuch as evaluation of neuronal cell loss by neuron counts was unfeasible due to the considerable number of histologic sections studied. As assessed by the coefficient of variation, the intra- and interobserver reliability of this procedure was 4% and 10%, respectively (17).

Calculations.

Fetal cerebral blood flow was calculated from counts of the injected nuclide recovered in the fetal cerebrum and the appropriate reference sample, and from the withdrawal rate of the reference sample (16, 17). The histologic score of each cerebral specimen was calculated by averaging the scores of all visual fields analyzed from three sections of that specimen. The number of visual fields evaluated per specimen ranged between 400 and 500 each. The scores from corresponding specimens from the right and left hemisphere were averaged.

Statistics.

Values are given as means ± SD. The data were analyzed for intra- and intergroup differences by two-way multivariate ANOVA for repeated measures. The Games-Howell test was used as a posthoc testing procedure.

RESULTS

The physiologic variables of the control and study groups before, during, and after 30 min of global cerebral ischemia are shown in Table 1. Before injection of lubeluzole or carrier no significant differences in these variables were found between the control and study groups. At control, the plasma concentrations of both glucose and lactate were slightly above the normal range for chronically prepared fetal sheep (23), but had normalized by the end of the experiments.

Table 1 Physiological variables, acid-base balance, and blood gases before, during, and after transient global cerebral ischemia in fetal sheep near term.

Under control conditions, blood flow to the cerebrum averaged 100 ± 25 mL·100 g−1 min−1 and was reduced by more than 80% after carotid blood flow had been arrested. In the immediate recovery period, i.e. 10 min after ischemia, hyperperfusion occurred followed by a variable hypoperfusion later on (+3 h). Finally, at 72 h after ischemia, blood flow to the cerebrum was higher than at control (p ≤ 0.05). This pattern of dynamic changes in blood flow could be observed in each of the 32 cerebral specimens studied as well as in almost all deeper brain structures. Differences between the control and study groups could not be detected (Table 2). This was also true for each of the 32 cerebral specimens.

Table 2 Blood flow to various brain areas before, during, and after transient global cerebral ischemia in fetal sheep near term

Neuronal cell damage was most pronounced in the parasagittal regions, whereas in the more lateral part of the cortex only minor neuronal damage could be seen (Table 3). Furthermore, cell damage was observed in the cerebellum, hippocampus, and caudate nucleus (Table 4). There were almost no significant differences in neuronal cell damage in any parts of the cerebrum or deeper brain structures between treated and untreated fetuses (Tables 3 and 4).

Table 3 Neuronal cell damage in the cerebrum of fetal sheep near term 72 h after global cerebral ischemia
Table 4 Neuronal cell damage in various areas of the diencephalon, mesencephalon, and brain stem in fetal sheep near term 72 h after 30 min of global cerebral ischemia

DISCUSSION

Lubeluzole, the S-isomer of a novel 3,4-difluoro benzothiazole, has been shown to reduce ischemic neuronal cell damage in a variety of in vitro as well as in vivo studies. The neuroprotective property of lubeluzole may result from various effects on neuronal tissue. First, lubeluzole has been shown to block ischemia-induced increases in extracellular levels of glutamate and may therefore reduce excitotoxic cell injury (24). Second, lubeluzole inhibits glutamate-stimulated NO production (25). NO combines with superoxide anions to synthesize peroxynitrite, a compound that spontaneously decomposes to form hydroxyl radicals, nitrogen dioxide, and NO2+. All these radicals are able to destroy cell membranes and various intracellular structures. Third, lubeluzole has been shown to activate voltage-sensitive Ca2+ channels in isolated rat dorsal root ganglion cells, thus possibly reducing the tremendous intracellular influx of this ion during an ischemic insult (26). This so-called calcium overload leads to cell damage by activating proteases, lipases, and endonucleases (27). Fourth, lubeluzole decreased DNA fragmentation and annexin-V binding in primary hippocampal neurons (28, 29). Because these two phenomena are specific markers of apoptosis, lubeluzole might protect neurons from ischemic injury through its inhibitory effects on pathophysiological pathways that trigger the cellular suicide program. However, these last three lubeluzole-mediated effects have so far only been observed in neuronal tissue in vitro. Their in vivo significance has still to be confirmed.

Unfortunately, in the present study we were unable to show any neuroprotective effect of lubeluzole on neuronal cell damage in fetal sheep near term after global ischemia. This appears to be inconsistent with various investigations on adult animals (510). However, in almost all of these studies models of focal cerebral ischemia were used (59). This type of cerebral ischemia is characterized by an ischemic core and a peri-infarct region known as ischemic penumbra (30, 31). In this area, brain tissue is perfused at a level between the thresholds of functional impairment and of morphologic integrity. Unless cerebral blood flow is rapidly improved in this region within a few hours after the insult, neuronal cell damage is the inevitable result (30, 31). As shown in previous studies, glutamate is released in tremendous amounts from the infarct core into the extracellular space after focal ischemia. Increases of up to 80 times above baseline levels have been observed (32). The released glutamate activates the neuronal NOS via calcium influx through N-methyl-d-aspartate (NMDA)-regulated calcium channels (33, 34). In focal ischemia, this pathway may be of greater importance for the development of neuronal cell damage than in global ischemia, since in the latter type of ischemia only a moderate and short-lasting increase in glutamate release has been observed (35, 36). This may explain, in part, the differences in neuronal outcome between the present study and previous investigations after treatment of brain injury by lubeluzole. In addition, many other pathophysiological mechanisms such as inhibition of protein synthesis, generation of epileptogenic impulses, etc. are differently regulated in focal and global ischemia and may therefore alter the neuroprotective efficiency of lubeluzole (3741).

However, there is one study in adult rats indicating that lubeluzole protects the brain from neuronal cell loss after global ischemia (10). Posttreatment with lubeluzole significantly increased the number of viable neurons in the hippocampus. There could be a number of reasons for the difference in neuronal outcome between this study and our experiments. First, the postischemic pattern of pathophysiological changes in the immature brain may not be the same as in the adult. For example, glutamate release has been reported to be significantly lower in neonates than in adults (42). Although glutamate toxicity has been reported in fetal sheep after cerebral ischemia, this may also affect the induction of NOS via NMDA-receptor regulated calcium channels. Because the neuroprotective properties of lubeluzole are mediated, in part, by its effects on the NO system (25), this may explain the differences in neuronal outcome between the present study and that of Haseldonckx (10). Furthermore, inhibition of the NOS in the fetal sheep had no influence on neuronal cell damage, indicating that NO release may not play a major pathophysiological role in the development of brain injury under these conditions (43). Second, in our study, lubeluzole was injected before cerebral ischemia, whereas Haseldonckx and co-workers (10) applied the drug shortly after the insult. One might therefore speculate that bolus injection of lubeluzole before ischemia results in an insufficient drug supply during the postischemic period. However, as demonstrated in patients with acute ischemic stroke, the mean terminal half-life of this compound in the plasma is 27.7 h (18). Thus, application of lubeluzole 1 h before ischemia guarantees postischemic plasma concentrations that are within the therapeutic range. When designing the present study, a pretreatment protocol was chosen, inasmuch as we considered this to be more effective than a posttreatment strategy. Especially under clinical conditions, pretreatment of babies at risk of hypoxic-ischemic encephalopathy by maternal application of various neuroprotective drugs would seem conceivable. Whether an additional application of lubeluzole after cerebral ischemia provides neuroprotection in the present experimental model will have to be examined in further investigations. Third, the end point in our study and that of Haseldonckx and co-workers was different. Whereas neuronal cell damage was evaluated only 3 d after the insult in the present investigation, it was not assessed until 7 d after cerebral ischemia in the latter study. Unfortunately, the sheep model does not allow for longer postischemic observation periods, inasmuch as this would result in a tremendous increase in the abortion rate due to intrauterine infection and fetal distress. It is conceivable that even several days after the ischemic insult the continued activation of the inducible form of NOS in astrocytes and microglia might result in an increased production of NO (44). One might therefore object that delayed neuronal cell death due to an increased NO production cannot be evaluated using this model (4547). Because modulation of NO toxicity is a probable target for lubeluzole (25), the negative result of the present study might be due to this phenomenon. This objection cannot be fully excluded. However, as mentioned above, NO toxicity may not play any major part in the development of neuronal cell death in this model, inasmuch as inhibition of the NO system had no effect on brain injury (43). Furthermore, as shown in a recent study using the MRI technique, neuroprotective effects of lubeluzole can already be observed within a few hours after ischemia (9). If the main target of lubeluzole is NO toxicity, then a neuroprotective effect should have been observed as early as 3 d after ischemia as demonstrated in a recent study on neonatal rats using an inhibitor of the NOS (48). To exclude very delayed neuroprotective effects of lubeluzole on the neuropathologic outcome in immature animals, additional experiments will have to be performed using other ischemic models, e.g. unilateral carotid occlusion in the neonatal rat (49, 50). However, it should be taken into account that drug-related side effects of various neuroprotective substances, such as lubeluzole, on the cardiovascular system cannot be monitored in such small animals. We therefore feel that in a first approach, the present model was highly appropriate for investigation of the neuroprotective efficacy of lubeluzole.

As already shown for a variety of other drugs (for review, Ref. 40), the neuroprotective properties of lubeluzole might depend on the type of ischemia studied and the animal model used. It is therefore conceivable that lubeluzole might protect the fetal brain from ischemic insults in animal models of systemic asphyxia induced by occlusion of the umbilical cord or the uterine arteries. Neuronal loss has been observed after repetitive cord occlusion in the striatum, an area with high glutaminergic input (51). Inasmuch as lubeluzole is known to inhibit glutamate release during an ischemic insult (24), this drug may be neuroprotective in such an animal model. However, when brain injury is induced by systemic asphyxia, the extent and location of neuronal cell loss often vary widely (5254) and there is a considerable rise in the fetal abortion rate. Because hardly any neuroprotective effect can be demonstrated under such conditions, we preferred the fetal sheep model of global cerebral ischemia caused by occlusion of both carotid arteries. The fact that lubeluzole did not appear to have any neuroprotective effect in the present model cannot be attributed to a lack of glutamate toxicity. As already shown in a previous study the application of MK-801 after global cerebral ischemia caused a significant improvement in neuropathological outcome in fetal sheep (55).

From the present results, we conclude that pretreatment with lubeluzole fails to protect the brain of fetal sheep near term from injury after transient global cerebral ischemia. However, because the observation period lasted only 3 d, a possible effect of lubeluzole on pathophysiological mechanisms inducing delayed neuronal cell death cannot be fully excluded.