Neuroprotective effects of Ginkgo biloba extract.

Ginkgo biloba extract has been therapeutically used for several decades to increase peripheral and cerebral blood flow as well as for the treatment of dementia. The extract contains multiple compounds such as flavonoids and terpenoids that are thought to contribute to its neuroprotective and vasotropic effects. In this review, we summarize the experimental results on the mechanism of neuroprotection induced by standardized extract of Ginkgo biloba leaves (EGb 761) and its constituents. The effects described mostly in animals include those on cerebral blood flow, neurotransmitter systems, cellular redox state and nitric oxide level. Furthermore, we discuss the current status of clinical trials as well as undesired side effects of EGb 761.

Pharmacological studies supporting the therapeutic use of Ginkgo biloba extract for Alzheimer's disease.

The standardized Ginkgo biloba extract EGb 761(definition see editorial) has been shown to produce neuroprotective effects in different in vivo and in vitro models. Since EGb 761 is a complex mixture containing flavonoid glycosides, terpene lactones (non-flavone fraction) and various other constituents, the question arises as to which of these compounds mediates the protective activity of EGb 761. Previous studies have demonstrated that the non-flavone fraction was responsible for the antihypoxic activity of EGb 761. Thus, we examined the neuroprotective and anti-apoptotic ability of the main constituents of the non-flavone fraction, the ginkgolides A, B, C, J and bilobalide. In focal cerebral ischemia models, the administration of bilobalide (5-20 mg/kg, s. c.) 60 min before ischemia dose-dependently reduced the infarct area in mouse brain and the infarct volume in rat brain 2 days after the onset of the injury. 30 minutes of pretreatment with ginkgolide A (50 mg/kg, s. c.) and ginkgolide B (100 mg/kg, s. c.) reduced the infarct area in the mouse model of focal ischemia. In primary cultures of hippocampal neurons and astrocytes from neonatal rats, ginkgolide B (1 microM) and bilobalide (10 microM) protected the neurons against damage caused by glutamate (1 mM, 1 h) as evaluated by trypan blue staining. In addition, bilobalide (0.1 microM) was able to increase the viability of cultured neurons from chick embryo telencepalon when exposed to cyanide (1 mM, 1h). Furthermore, we attempted to find out whether ginkgolides A, B, and J and bilobalide were also able to inhibit neuronal apoptosis (determined by nuclear staining with Hoechst 33 258 and TUNEL-staining). Ginkgolide B (10 microM), ginkgolide J (100 microM) and bilobalide (1 microM) reduced the apoptotic damage induced by serum deprivation (24h) or treatment with staurosporine (200 nM, 24h) in cultured chick embryonic neurons. Bilobalide (100 microM) rescued cultured rat hippocampal neurons from apoptosis caused by serum deprivation (24h), whereas ginkgolide B (100 microM) and bilobalide (100 microM) reduced apoptotic damage induced by staurosporine (300 nM, 24h). Ginkgolide A failed to affect apoptotic damage neither in serum-deprived nor in staurosporine-treated neurons. The results suggest that some of the constituents of the non-flavone fraction of EGb 761 possess neuroprotective and anti-apoptotic capacity, and that bilobalide is the most potent one. In contrast, ginkgolic acids (100-500 microM) induced neuronal death, which showed features of apoptosis as well as of necrosis, but these constituents were removed from EGb 761 below an amount of 0.0005 %. Taking together, there is experimental evidence for a neuroprotective effect of EGb 761 that agrees with clinical studies showing the efficacy of an oral treatment in patients with mild and moderate dementia.

Ginkgolic acids induce neuronal death and activate protein phosphatase type-2C.

The standardized extract from Ginkgo biloba (EGb 761) is used for the treatment of dementia. Because of allergenic and genotoxic effects, ginkgolic acids are restricted in EGb 761 to 5 ppm. The question arises whether ginkgolic acids also have neurotoxic effects. In the present study, ginkgolic acids caused death of cultured chick embryonic neurons in a concentration-dependent manner, in the presence and in the absence of serum. Ginkgolic acids-induced death showed features of apoptosis as we observed chromatin condensation, shrinkage of the nucleus and reduction of the damage by the protein synthesis inhibitor cycloheximide, demonstrating an active type of cell death. However, DNA fragmentation detected by the terminal-transferase-mediated ddUTP-digoxigenin nick-end labeling (TUNEL) assay and caspase-3 activation, which are also considered as hallmarks of apoptosis, were not seen after treatment with 150 microM ginkgolic acids in serum-free medium, a dose which increased the percentage of neurons with chromatin condensation and shrunken nuclei to 88% compared with 25% in serum-deprived, vehicle-treated controls. This suggests that ginkgolic acid-induced death showed signs of apoptosis as well as of necrosis. Ginkgolic acids specifically increased the activity of protein phosphatase type-2C, whereas other protein phosphatases such as protein phosphatases 1A, 2A and 2B, tyrosine phosphatase, and unspecific acid- and alkaline phosphatases were inhibited or remained unchanged, suggesting protein phosphatase 2C to play a role in the neurotoxic effect mediated by ginkgolic acids.

Neuroprotective effects of NV-31, a bilobalide-derived compound: evidence for an antioxidative mechanism.

In previous studies we have already shown that the extract of Ginkgo biloba, and some of its constituents, such as ginkgolide B and bilobalide, protected cultured neurons against apoptotic and excitotoxic damage and reduced the infarct volume after focal cerebral ischemia in mice and rats. In this work, we determined the neuroprotective and antioxidative effects of 4-hydroxy-4-tert-butyl-2,3,5,6-tetrahydrothiopyran-1-oxide (NV-31), a stable compound which was synthesized to mimic the pharmacological activity profile of bilobalide. In pure neuronal cultures from chick embryo telencephalon, damage was induced by serum deprivation (24 h) and exposure to staurosporine (200 nM, 24 h) which caused an increase in the percentage of apoptotic neurons from 14 (controls) to 30 and 55%, respectively. NV-31 (1-100 nM) protected dose-dependently chick neurons against both serum deprivation- and staurosporine-induced apoptosis. Similarly, NV-31 (100 nM) reduced staurosporine (300 nM, 24 h)-induced neuronal damage in mixed cultures of neurons and astrocytes from neonatal rat hippocampus. The cellular ROS content increased 6-fold 4 h after serum deprivation as well as 4 h after the exposure to staurosporine and this increase was reduced by 50% in the presence of 10 and 100 nM NV-31, respectively. In mice, a treatment with 10 and 20 mg/kg NV-31 60 min before and immediately after focal cerebral ischemia, respectively, significantly reduced the infarct area compared with vehicle-treated animals. In the present study, we show that NV-31 promotes neuronal survival and we suggest that its antioxidative property contributes to the mechanism of neuroprotection.

Inhibition of serum deprivation- and staurosporine-induced neuronal apoptosis by Ginkgo biloba extract and some of its constituents.

Previous studies have already demonstrated that some constituents of an extract of Ginkgo biloba (EGb), such as ginkgolide B and bilobalide, protect cultured neurons from hypoxia- and glutamate-induced damage. This prompted us to investigate whether they were also able to inhibit neuronal apoptosis. We induced apoptosis in cultured chick embryonic neurons as well as in mixed cultures of neurons and astrocytes from neonatal rat hippocampus by serum deprivation and staurosporine. The increase in the percentage of apoptotic chick neurons from 12% in controls to 30% after 24 h of serum deprivation was reduced to control level by EGb (10 mg/l), ginkgolide B (10 microM), ginkgolide J (100 microM) and bilobalide (1 microM). After treatment with staurosporine (200 nM) for 24 h we observed 74% apoptotic chick neurons. This percentage of apoptotic neurons was reduced to 24%, 62% and 31% in the presence of EGb (100 mg/l), ginkgolide J (100 microM) and ginkgolide B (10 microM), respectively. Bilobalide (10 microM) decreased apoptotic damage induced by staurosporine treatment for 12 h nearly to the control level. In mixed neuronal/glial cultures, the extract of EGb (100 mg/l) and bilobalide (100 microM) rescued rat neurons from apoptosis caused by serum deprivation, whereas, bilobalide (100 microM) and ginkgolide B (100 microM) reduced staurosporine-induced apoptotic damage. Ginkgolide A revealed no anti-apoptotic effect in either serum-deprived or staurosporine-treated neurons. Our results suggest that EGb and some of its constituents possess anti-apoptotic capacity and that bilobalide is the most potent constituent.

Ca2+ and reactive oxygen species in staurosporine-induced neuronal apoptosis.

Staurosporine (0.03-0.5 microM) induced a dose-dependent, apoptotic degeneration in cultured rat hippocampal neurons that was sensitive to 24-h pretreatments with the protein synthesis inhibitor cycloheximide (1 microM) or the cell cycle inhibitor mimosine (100 microM). To investigate the role of Ca2+ and reactive oxygen species in staurosporine-induced neuronal apoptosis, we overexpressed calbindin D28K, a Ca2+ binding protein, and Cu/ Zn superoxide dismutase, an antioxidative enzyme, in the hippocampal neurons using adenovirus-mediated gene transfer. Infection of the cultures with the recombinant adenoviruses (100 multiplicity of infection) resulted in a stable expression of the respective proteins assessed 48 h later. Overexpression of both calbindin D28K and Cu/Zn superoxide dismutase significantly reduced staurosporine neurotoxicity compared with control cultures infected with a beta-galactosidase overexpressing adenovirus. Staurosporine-induced neuronal apoptosis was also significantly reduced when the culture medium was supplemented with 10 or 30 mM K+, suggesting that Ca2+ influx via voltage-sensitive Ca2+ channels reduces this apoptotic cell death. In contrast, neither the glutamate receptor agonist NMDA (1-10 microM) nor the NMDA receptor antagonist dizocilpine (MK-801; 1 microM) was able to reduce staurosporine neurotoxicity. Cultures treated with the antioxidants U-74500A (1-10 microM) and N-acetylcysteine (100 microM) also demonstrated reduced staurosporine neurotoxicity. These results suggest a fundamental role for both Ca2+ and reactive oxygen species in staurosprine-induced neuronal apoptosis.

Primary cultures of neurons for testing neuroprotective drug effects.

Primary cultures of neurons are widely used for the investigation of pathomechanisms of neuronal damage and for the evaluation of neuroprotective drug effects. The present paper gives a short survey of frequently used primary neuronal culture systems and of experimental measures for the induction of defined neuronal damage with particular respect to the pathomechanisms of cerebral ischemia. Neuroprotective drug effects as achieved under these conditions are reviewed, and the neuroprotective effects of glutamate antagonists, radical scavengers, and neural growth factors are discussed in some more detail.

Isoform-specific effects of transforming growth factors-beta on degeneration of primary neuronal cultures induced by cytotoxic hypoxia or glutamate.

The transforming growth factors-beta (TGFs-beta) are multifunctional peptide growth factors that have been localized in neuronal and glial cells of the CNS of mice, rats, and chick embryos. We tested the TGF-beta isoforms 1, 2, and 3 for their protective effects against neuronal degeneration caused by cytotoxic hypoxia or by the excitatory amino acid L-glutamate. A cytotoxic hypoxia was induced in cultured chick embryo telencephalic neurons by adding 1 mM sodium cyanide to the culture medium for a period of 30 min. Treatment with TGF-beta 1 (1-30 ng/ml) led to a statistically significant increase in cell viability, neuronal ATP levels, and protein content of the cultures assessed 72 h after the toxic insult. TGF-beta 3 was able to reduce the cyanide-induced neuronal damage at concentrations of 0.3 and 1 ng/ml, whereas TGF-beta 2 only showed neuroprotective activity at concentrations of 30 and 50 ng/ml. Both pre- and post-treatment with TGF-beta 1 also prevented the degeneration of cultured chick embryo telencephalic neurons that had been exposed to 1 mM L-glutamate in a buffered salt solution for a period of 60 min. Furthermore, TGF-beta 1 (0.3-3 ng/ml), and to a lesser extent TGF-beta 3 (0.1-1 ng/ml), significantly reduced excitotoxic injury of cultured neurons from rat cerebral cortex that had been exposed to serum-free culture medium supplemented with 1 mM L-glutamate. These results demonstrate that the TGFs-beta are able to prevent the degeneration of primary neuronal cultures, which was caused by energy depletion and activation of glutamate receptors, in an isoform-specific manner.

Extract of kava (Piper methysticum) and its methysticin constituents protect brain tissue against ischemic damage in rodents.

The purpose of the present study was to test whether kava extract and its constituents kawain, dihydrokawain, methysticin, dihydromethysticin and yangonin provide protection against ischemic brain damage. To this end, we used a model of focal cerebral ischemia in mice and rats. Ischemia was induced by microbipolar coagulation of the left middle cerebral artery (MCA). To quantify the size of the lesion in mice, the area of the infarct on the brain surface was assessed planimetrically 48 h after MCA occlusion by transcardial perfusion of carbon black. In the rat model infarct volume was determined 48 h after MCA occlusion by planimetric analysis and subsequent integration of the infarct areas on serial coronal slices. Compounds were administered i.p., except the kava extract, which was administered orally. The effects of the kava extract and its constituents were compared with those produced by the typical anticonvulsant, memantine. The kava extract, methysticin and dihydromethysticin produced effects similar to those of the reference substance memantine. The kava extract (150 mg/kg, 1 h before ischemia) diminished the infarct area (P less than 0.05) in mouse brains and the infarct volume (P less than 0.05) in rat brains. Methysticin, dihydromethysticin (both 10 and 30 mg/kg, 15 min before ischemia) and memantine (20 mg/kg, 30 min before ischemia) significantly reduced the infarct area in mouse brains. All other compounds failed to produce a beneficial effect on the infarct area in mouse brains. In conclusion, the kava extract exhibited neuroprotective activity, which was probably mediated by its constituents methysticin and dihydromethysticin.

A mouse model of focal cerebral ischemia for screening neuroprotective drug effects.

The aim of the present study was to investigate if the infarct area on the brain surface after middle cerebral artery (MCA) occlusion in the mouse is representative for the infarct volume and if this determination of brain injury can be used for screening neuroprotective drug effects. Cerebral infarction was induced by coagulating electrically the stem of the left MCA. After 48 hr, the brains were perfused with carbon black and the unstained infarct area was determined by means of an image analyzing system. The infarct volume was determined by calculating the infarct area on coronal slices and the distance between succeeding slices. The correlation between the area and the volume of infarction was significant (r = 0.81; p less than 0.001). N-methyl-D-aspartate (NMDA) antagonists, calcium antagonists, 5-hydroxytryptamine-1A (5-HT-1A) agonists, radical scavengers, and various drugs were investigated in the mouse model of MCA occlusion. Drugs were usually applicated before ischemia. The drugs that were found to be neuroprotective in the mouse model revealed similar effects in rat models of focal or global cerebral ischemia. These findings show that the presented mouse model with its simple technique of measuring the infarct size is suitable for screening purposes.

Neuroprotective effect of nimodipine is not mediated by increased cerebral blood flow after transient forebrain ischemia in rats.

In the present study, we investigated the protective effect of nimodipine against postischemic neuronal damage in the rat and considered the question of whether this histologic finding coincides with an improvement of cerebral circulation. Male Wistar rats were subjected to 10 minutes of forebrain ischemia by clamping both common carotid arteries and lowering blood pressure to 40 mm Hg. Histologic evaluation was performed 7 days after ischemia. Local cerebral blood flow was determined with the 14C-iodoantipyrine technique in anatomically defined areas of the brain, including hippocampus. Preischemic application of nimodipine (3.0 mg/kg p.o.) significantly reduced the percentage of damaged neurons in hippocampal CA1 subfield from 78 +/- 4% in controls to 59 +/- 6% in treated rats (mean +/- SEM; p less than 0.05, Mann-Whitney U test). After 10, 60, and 180 minutes of recirculation no differences in local cerebral blood flow between control and drug-treated animals were observed. Our results demonstrate that nimodipine reduces ischemia-induced neuronal damage in rat hippocampus. We did not consider increased cerebral blood flow in the hypoperfusion state in the applied experimental design since improvement of cerebral blood flow seems to bear little relation to the neuroprotective activity of nimodipine.

Cultured neurons for testing antihypoxic drug effects.

Cultured neurons of chick embryo cerebral hemispheres were used as an in vitro system for investigating the influence of several drugs on neuronal cell viability and metabolic activity under hypoxic conditions. Hypoxia was induced by addition of sodium cyanide to the nutrient medium, which led to a rapid depletion of energy stores. The ATP level of the cells and the protein content of the cultures were used to characterize the degree of neuronal damage after cytotoxic hypoxia and recovery, respectively, recovery lasting 15 min or 3 days. Various calcium antagonists, NMDA-antagonists, central depressants, central stimulants, nootropics, and miscellaneous drugs were tested. NMDA-antagonists and central depressants consistently protected the neurons against alterations caused by hypoxia. However, only one (flunarizine) out of five calcium antagonists, two (naftidrofuryl, pyritinol) out of 13 nootropics, the kappa-agonist ketazocine, and the ATPase inhibitor ouabaine exerted neuroprotection. The in vitro model seems to be suitable for testing neuroprotective drug effects and to be a valuable supplement for in vivo experiments, especially when the cellular mechanism of drug action has to be clarified.

PAF antagonist ginkgolide B reduces postischemic neuronal damage in rat brain hippocampus.

We investigated the effect of the known antagonist of platelet-activating factor (PAF), ginkgolide B, on postischemic neuronal damage in the rat. Neuronal necroses were evaluated in the hippocampus 7 days after a 10-min forebrain ischemia. Preischemic application of ginkgolide B (50 mg/kg p.o.) significantly reduced neuronal damage. It is suggested that the antagonism of PAF is responsible for this beneficial effect of ginkgolide B.

Effects of Ginkgo biloba constituents related to protection against brain damage caused by hypoxia.

The purpose of the present study was to approach the compound(s) responsible for the beneficial effects of an extract of Ginkgo biloba leaves (EGB) on animals subjected to hypoxia. In this first approach we compared the effects of the flavone and the non-flavone fraction of EGB with those of the whole extract on mice in lethal hypoxia (3.5% O2), on brain energy metabolism of artificially ventilated rats inspiring 7% O2, and on local cerebral blood flow (LCBF) of normoxic rats. The latter two experimental settings should also extend the knowledge about the underlying mechanisms of the antihypoxidotic actions. EGB as well as its non-flavone fraction considerably prolonged the survival time of mice under lethal hypoxia. EGB retarded the breakdown of brain energy metabolism in the hypoxic artificially ventilated rat. A corresponding effect was exerted by the non-flavone fraction while the flavone fraction even worsened the metabolic state. The non-flavone fraction increased LCBF in the majority of 35 examined brain regions; a similar effect could be seen after EGB-treatment, while the flavone fraction caused only minor alterations of LCBF. We conclude that the non-flavone fraction of EGB carries the antihypoxidotic activity. Metabolic effects are suggested to cause this activity. Further studies are necessary to elucidate the effective compound within this fraction.

Influence of an extract of Ginkgo biloba on cerebral blood flow and metabolism.

The purpose of the present investigation was to examine the effects of an extract of Ginkgo biloba (EGB) on blood glucose levels, on local cerebral blood flow as well as on cerebral glucose concentration and consumption. The local cerebral blood flow (LCBF) was measured in conscious rats by means of the 14C-iodoantipyrine technique and local cerebral glucose utilization (LCGU) by 14C-2-deoxy-glucose autoradiography. EGB increased the LCBF in 39 analyzed, anatomically defined brain structures by 50 to 100 per cent. No influence of EGB on LCGU was demonstrable. However, EGB enhanced the blood glucose level dose-dependently. Substrates and metabolites of energy metabolism were measured in the cortex of the isolated rat brain perfused at constant rate and with 7 mmol/l glucose added to the perfusion medium. In these experiments EGB decreased the cortical glucose concentration without other substrate levels being changed. These results suggest that glucose uptake may be inhibited by EGB. It is argued that the effects of EGB on brain glucose concentration and blood flow may contribute to its protection of brain tissue against ischemic or hypoxic damage.

[The usefulness of a model of acute hypoxia for the testing of cerebroprotective drugs]. / Uber die Brauchbarkeit eines Modells der akuten Hypoxie zur Testung zerebroprotektiver Pharmaka.

The aim of this study was to find out whether the mouse in acute normobaric hypoxia (3.5% O2) may be appropriate as a screening model for testing cerebroprotective drugs. The survival time of the mice in hypoxia was used to determine drug effects. 39 agents with various pharmacological and toxicological properties were investigated. Cerebroprotective, centrally depressant and stimulating as well as cardiovascular drugs were included. Only 6 out of 16 cerebroprotective drugs used therapeutically prolonged the survival time of mice in acute hypoxia. However, a positive effect was demonstrable especially for those cerebroprotective drugs (extractum Ginkgo bilobae, meclofenoxate, naftidrofuryl, pentoxifylline and pyritinol), which are generally accepted to be active. On the other hand, 12 out of 23 drugs which are therapeutically used as cardiovascular drugs, central stimulants or depressants or have known toxicological effects also prolonged the survival time of mice in hypoxia. Thus, it became clear that the result of this hypoxia test may be influenced by various pharmacological actions. Especially drugs stimulating the cardiovascular system or depressing CNS activity could prolong the survival time of mice in hypoxia, whereas drugs with vasodilating effects could even shorten it. Opposite effects could superpose or neutralize each other. Therefore, the mouse in acute hypoxia seems to be of limited value as a screening model for testing cerebroprotective drugs.

Effects of cerebro-protective agents on enzyme activities of rat primary glial cultures and rat cerebral cortex.

The effects of different cerebro-protective agents on selected key enzymes of the energy metabolism of rat primary glial cultures and rat cerebral cortex were studied. As indicators for the capacity of the most important pathways of energy metabolism the following enzyme activities were determined: hexokinase (HK), phosphofructokinase (PFK), pyruvate kinase (PK), lactate dehydrogenase (LDH), glucose-6-phosphate dehydrogenase (G-6-P-DH), malate dehydrogenase (MDH), glutamate dehydrogenase (GDH), and cytochrome-c-reductase (CCR). After a one week growth period, rat glial cultures were incubated for 3 or 4 weeks with the substances to be tested. Bencyclane (5 X 10(-5) mol/l) increased the activities of HK, G-6-P-DH, and LDH, whereas PFK and CCR were reduced. Pyritinol (10(-4) mol/l) led to a higher G-6-P-DH activity, simultaneously lowering the values for PFK, CCR, PK, LDH, and MDH. Under the influence of an extract of the leaves of Ginkgo bilobae (EGB; 100 mg/l) PFK, LDH, and MDH activities were reduced. All these alterations in enzyme activities went along with simultaneous reductions in protein content, therefore not allowing to exclude toxic effects with regard to the doses used. Moreover, direct interference with the analytical procedure was demonstrable for bencyclane and EGB. Piracetam (10(-3) mol/l), flunarizine (10(-6) mol/l), dihydroergocristine (5 X 10(-6) mol/l), and nicergoline (5 X 10(-6) mol/l) failed to induce any alteration in the employed doses. The most striking effects were obtained with meclofenoxate which was tested at 10(-3) and 10(-4) mol/l. The higher dose caused an elevation of HK, PFK, CCR, G-6-P-DH, GDH and MDH activities, while slightly reducing PK. With the lower dose of meclofenoxate CCR and G-6-P-DH activities were increased. Short-term incubation of the cultures with 10(-3) mol/l meclofenoxate for 24 hr led to an increase in LDH, G-6-P-DH, and GDH activities. Chronic incubation with meclofenoxate (10(-3) mol/l) followed by 48 hr deprivation of the drug resulted in elevated HK, PFK, CCR, G-6-P-DH, GDH, and MDH activities. These changes were accompanied by alterations in related metabolite levels. These include elevations in the concentration of creatine phosphate and fructose-1,6-bisphosphate, whereas glucose-6-phosphate levels were reduced. After one week of meclofenoxate deprivation the activities of CCR and G-6-P-DH were still elevated. The metabolites of meclofenoxate dimethylaminoethanol (DMAE; 10(-3) mol/l) and p-chlorophenoxyacetic acid (10(-3) mol/l) were also investigated.(ABSTRACT TRUNCATED AT 400 WORDS)

Effect of an extract of Ginkgo biloba on rat brain energy metabolism in hypoxia.

The purpose of the present investigation was to determine brain energy metabolism under hypoxic conditions as influenced by an extract of Ginkgo biloba (EGB). Male Sprague-Dawley rats treated with EGB were exposed to hypobaric or hypoxic hypoxia, and at various time points during or after hypoxia the levels of high-energy phosphates and some substrates of glycolysis were measured in brain cortical tissue. Rats treated with EGB (100 mg/kg, intraperitoneally) survived hypobaric hypoxia for a much longer period than controls (e.g. controls: 3.9 +/- 1.8 min, EGB-treated: 23.6 +/- 10.5 min). The brain glucose level was elevated by EGB in most experimental series, and the lactate concentration was slightly lower than in control brains. The lowering of lactate/pyruvate ratio was due to the decreased level of lactate and to the enhanced concentration of pyruvate as well. When hypoxia was sufficiently severe the breakdown of high-energy phosphates was less pronounced in EGB-treated animals. After oral application of EGB for 14 days the rats survived hypobaric hypoxia for 25.7 +/- 2.5 min whereas controls survived for 11.5 +/- 5.1 min. However, brain energy metabolism was not significantly influenced by this oral treatment. It is suggested that changes in brain energy metabolism and blood flow may contribute to the protective effect of EGB against hypoxia.