This study evaluated the effects of retinal ischemia-reperfusion (IR) injury and pre-treatment with the potent and specific aldose reductase inhibitor fidarestat on apoptosis, aldose reductase and sorbitol dehydrogenase expression, sorbitol pathway intermediate concentrations, and oxidative-nitrosative stress. Female Wistar rats were pre-treated with either vehicle (N-methyl-D-glucamine) or fidarestat, 32 mgkg-1d-1 for both, in the right jugular vein, for 3 consecutive days. A group of vehicle- and fidarestat-treated rats were subjected to 45-min retinal ischemia followed by 24-h reperfusion. Ischemia was induced 30 min after the last vehicle or fidarestat administration. Retinal IR resulted in a remarkable increase in retinal cell death. The number of TUNEL-positive nuclei increased 48-fold in the IR group compared with non-ischemic controls (p<0.01), and this increase was partially prevented by fidarestat. AR expression (Western blot analysis) increased by 19% in the IR group (p<0.05), and this increase was prevented by fidarestat. Sorbitol dehydrogenase and nitrated protein expressions were similar among all experimental groups. Retinal sorbitol concentrations tended to increase in the IR group but the difference with non-ischemic controls did not achieve statistical significance (p=0.08). Retinal fructose concentrations were 2.2-fold greater in the IR group than in the non-ischemic controls (p<0.05). Fidarestat pre-treatment of rats subjected to IR reduced retinal sorbitol concentration to the levels in non-ischemic controls. Retinal fructose concentrations were reduced by 41% in fidarestat-pre-treated IR group vs. untreated ischemic controls (p=0.0517), but remained 30% higher than in the non-ischemic control group. In conclusion, IR injury to rat retina is associated with a dramatic increase in cell death, elevated AR expression and sorbitol pathway intermediate accumulation. These changes were prevented or alleviated by the AR inhibitor fidarestat. The results identify AR as an important therapeutic target for diseases involving IR injury, and provide the rationale for development of fidarestat and other AR inhibitors.

Hyperglycemia is a recognized risk factor for cardiovascular disease in diabetes. Recently we reported that high glucose activates the transcription factor NFAT in arteries ex vivo. Here, we investigate if hyperglycemia activates NFAT in vivo and whether this leads to vascular complications. Intra peritoneal glucose tolerance test in mice increased NFATc3 nuclear accumulation in vascular smooth muscle. Streptozotocin‐induced diabetes augmented NFATc3 transcriptional activity in arteries of NFAT‐luciferase transgenic mice. Two NFAT responsive sequences were identified in the promoter of the proinflammatory cytokine osteopontin (OPN), known to exacerbate atherosclerosis and restenosis. Activation of NFAT resulted in higher OPN mRNA and protein in native arteries. Glucose‐induced OPN expression was prevented by the calcineurin inhibitor cyclosporin A, the NFAT blocker A‐285222 or the ectonucleotidase apyrase, suggesting a mechanism involving the release of extracellular nucleotides. Further, diabetes resulted in higher OPN expression, which was significantly decreased by in vivo treatment with A‐285222 for 4 weeks or prevented in arteries from NFATc3−/− mice. These results identify a novel molecular mechanism underlying vascular complications of diabetes. Funding: Swedish Heart & Lung Foundations

Recent studies suggest that increased aldose reductase (AR) activity plays an important role in ischemia-reperfusion injury in the retina. The mechanisms are not completely understood, but may be linked to inflammation. In the present study, we investigated whether the AR inhibitor fidarestat suppressed the retinal inflammatory response induced by ischemia-reperfusion in a rat model. The inflammatory response was manifested by increased gene expression of tumor necrosis factor-α and intercellular adhesion molecule-1 (ICAM-1) as well as elevated protein levels of soluble ICAM-1. This response was partially suppressed by the AR inhibitor fidarestat. The findings may reveal beneficial effects of AR inhibition on retinal inflammation associated with ischemia-reperfusion and are in agreement with recent developments in pharmacological research suggesting that pathological conditions other than diabetes may benefit from AR inhibitors.

Objective-Vascular inflammation is a key feature of both micro- and macrovascular complications in diabetes. Several lines of evidence have implicated the cytokine tumor necrosis factor (TNF)alpha as an important mediator of inflammation in diabetes. In the present study we evaluated the role of TNF alpha in streptozotocin (STZ)-induced diabetes on vascular inflammation in C57BL/6 wild-type and apoE(-/-) mice.Methods and Results-Diabetes increased the expression of vascular cell adhesion molecule (VCAM)-1 in cerebral arteries > 150 mu m in diameter as well as the macrophage accumulation in aortic root atherosclerotic plaques in apoE(-/-) mice. A more pronounced vascular inflammatory response was observed in diabetic TNF alpha-deficient apoE(-/-) mice. These mice were also characterized by increased accumulation of IgG and IgM autoantibodies in atherosclerotic lesions. Diabetes also increased VCAM-1 expression and plaque formation in apoE-competent TNF alpha(-/-) mice, whereas no such effects were observed in C57BL/6 wild-type mice.Conclusions-The present findings suggest that TNF alpha does not mediate diabetic-induced vascular inflammation in mice and reveal an unexpected protective role for TNF alpha. These effects are partly attributable to a direct antiinflammatory role of TNF alpha, but may also reflect a defective development of the immune system in these mice. (Arterioscler Thromb Vasc Biol. 2009; 29: 1465-1470.)

Heat shock protein 70 (Hsp70) is induced in cells by a variety of stress conditions, is known to be cytoprotective, and has been proposed to be neuroprotective during brain ischemia. A recently developed mouse model of 12-min global cerebral ischemia by bilateral common carotid artery occlusion with artificial ventilation and bilateral monitoring of regional cerebral blood flow by laser Doppler was applied. We examined the expression and possible neuroprotective role of the inducible form of Hsp70 in the mouse brain following global cerebral ischemia. Ischemia induced a marked expression of Hsp70 in the ischemia vulnerable CA1-CA3 region of the hippocampus. Intraischemic hypothermia (33degreesC) prevented cell damage without noticeable expression of Hsp70. A transgenic mouse overexpressing Hsp70 was subjected to 12 min of global cerebral ischemia, and the brain damage was evaluated after 4 days. No neuroprotection of ischemia-induced brain damage in hippocampus, striatum, cortex or thalamus was found in Hsp70 transgenic animals compared with wild-type littermate mice. We suggest that overexpression of Hsp70 following cerebral ischemia is an indicator of cell stress. Also, constitutively overexpression of Hsp70 is insufficient to effectively influence cell death after global cerebral ischemia in the mouse.

Extracellular adenosine is dramatically increased during cerebral ischaemia and is considered to be neuroprotective due to its inhibitory effect on synaptic transmission mediated by the adenosine A1 receptor (A1R). We investigated the importance of the A1R in a mouse model of global ischaemia and in a murine hippocampal slice culture model of in vitro ischaemia, using mice with the A1R gene deleted. In brains from mice lacking the A1R, damage induced by global ischaemia was similar to that in wild-type animals. In contrast, treatment with a selective A1R antagonist [8-cyclo-pentyl theophylline (8-CPT)], administered before the ischaemic insult in naive wild-type mice, exacerbated the neuronal damage following global ischaemia. Although the inhibitory action of adenosine on excitatory neurotransmission in hippocampal slices was lost in A1R knockout mice, there was no difference in damage between slices from wild-type and knockout mice after in vitro ischaemia. The results suggest that some effects of the A1R are compensated for in knockout animals.

Cystatin C is distributed in all human tissues and fluids with a particular abundance in the cerebrospinal fluid. Cystatin C is a strong endogenous inhibitor of lysosomal cysteine proteases, such as cathepsin B, L, H and S, that are involved in various biological processes such as degradation of cellular proteins and regulation of enzymes, as well as in pathological processes. Pharmacological inhibition of cathepsins has been shown to reduce neuronal damage after brain ischemia, suggesting that cystatin C is an endogenous neuroprotectant. Cystatin C has also amyloidogenic properties and is co-localized with beta-amyloid in degenerated neurons in Alzheimer's disease, suggesting a role in neuronal degeneration. To test the hypothesis that endogenous cystatin C is neuroprotective during brain ischemia, global and focal brain ischemia was induced in mice with the cystatin C gene knocked out. Following focal ischemia, larger brain infarcts were found in cystatin C knockout mice, probably due to a reduced inhibition of the cathepsins during ischemia. In contrast, brain damage after global ischemia was diminished in cystatin C knockout mice, suggesting that cystatin C has an aggravating effect on selective neuronal damage after global ischemia. (C) 2004 IBRO. Published by Elsevier Ltd. All rights reserved.

The importance of particular genes in neuronal death following global cerebral ischemia can readily be studied in genetically modified mice provided a reliable model of ischemia is available. For that purpose, we developed a mouse model of global cerebral ischemia that induces consistent damage to different regions of the brain and with a low mortality rate. Twelve minutes of ischemia was induced in C57BL/6 mice by bilateral common carotid artery occlusion under halothane anesthesia and artificial ventilation. Body and brain temperature were monitored and cortical cerebral blood flow in each hemisphere was measured by laser Doppler flowmeter before, during, and for 5 min after ischemia. Extensive damage was found in the striatum and marked cell damage was observed in the CA1 and CA2 regions of hippocampus and in thalamus. Mild damage was seen in the CA3 region, dentate gyrus and cortex. Hippocampal damage in the CA1 region is delayed and developed over 48 h. Intraischemic hypothermia of 33°C provided a robust neuroprotection. The non-competitive N-methyl-d-aspartate receptor blocker, MK-801, did not provide protection in the hippocampus, cortex, striatum or thalamus when administered 30 min prior to ischemia or 2 h after the end of ischemia, but selectively mitigated damage in the hippocampus, when administered immediately following ischemia. This model of global cerebral ischemia may be useful in pharmacological and genomic studies of ischemic brain damage.

In evaluating the damage impact on the brain from an ischemic insult, it is of importance to take into account several different parameters, which each may strongly, influence the extent of tissue injury. The duration of the insult and the level of blood flow reduction will naturally be of major importance im determining the extent of damage, but also the temperature of the brain during as well as after ischemia, and the pre- and intraischemic blood glucose levels are significant determinants of outcome. In addition, the cerebral structure affected and, if applicable, the time of recirculation after ischemia, further complicates the picture.

Status epilepticus (SE), i.e. ongoing seizures of more than 30 min duration, gives rise to bilateral pan-necrotic lesions of the substantia nigra, pars reticulata (SNPR). These are known to be preceded by an initial increase, followed by a depression of metabolic rate, and by failure of the bioenergetic state, suggesting mitochondrial dysfunction. We have previously shown that the spin trap α-phenyl-N-tert-butyl nitrone (PBN) prevents the lesions caused by 45 min of SE from occurring, in spite of ongoing seizure activity. In this article, we demonstrate that PBN, given 30 min before seizure induction, reduces or prevents the decrease in ATP concentration and adenylate energy charge, without significantly reducing the amount of lactate accumulated, or the decrease in intracellular pH (pHi). The results suggest that the spin trap nitrone preserves the structural and functional integrity of SNPR neurons by protecting the mitochondria against oxidative damage.

To study the effects of pronounced hypoglycemia on brain osmolality and brain edema formation, fasted rats were rendered hypoglycemic by injection of insulin, and subjected to 30 min of hypoglycemic coma. Recovery was accomplished by glucose administration. The change in water content in different brain regions was measured as a change in specific gravity after 30 min of hypoglycemic coma, or 30, 60, and 180 min after glucose administration. Plasma and brain tissue osmolality were measured in separate animals. The results show a significant decrease in specific gravity (increase in water content) in all structures measured (caudoputamen, neocortex, hippocampus, and cerebellum) at the end of the period of coma, as well as after 30 min and 60 min of recovery. At 180 min of recovery, brain water was normalized. The edema affected all structures to the same degree regardless of their vulnerability to hypoglycemic damage. Brain tissue osmolality showed a tendency to decrease with decreasing tissue glucose content. The decrease was significant (P<0.01) at 30 min of isoelectric coma. In the recovery phase, normal brain osmolality was restored within 30 min. Measurements of blood-brain barrier (BBB) permeability after 30 min of hypoglycemic coma showed no extravasation of Evan’s blue, though a small but significant increase in the permeability for aminoisobutyric acid (AIB) in caudoputamen and in cerebellum was found. To analyze the importance of tissue acidosis for formation of edema, hypoglycemic animals were made acidotic by increasing the CO2 concentration in inspired air to produce an arterial plasma pH of 6.8–6.9. In these animals the edema was of a similar degree to the normocapnic animals, and the permeability for AIB was normal. We conclude that osmolytic mechanisms are not the primary cause of the selective neuronal vulnerability in hypoglycemic coma. Furthermore, the BBB is largely intact during a hypoglycemic insult.

The immunosuppressant drug cyclosporin A (CsA) is considered to be inherently protective in conditions of ischemia, e.g. in hepatic and cardiac tissue. However, investigations of effects of CsA on neuronal tissue have been contradictory, probably because the blood–brain barrier (BBB) is virtually impermeable to CsA. In the present study, we exploited the finding that the insertion of a syringe needle into brain parenchyma obviously disrupts the BBB and allows influx of CsA, and explored whether CsA, given as intraperitoneal injections daily for 1 week before and 1 week after forebrain ischemia of 7 or 10 min duration, ameliorates the damage incurred to the hippocampal CA 1 sector. In other experiments, the needle insertion and the first i.p. injection of CsA were made 30 min after the start of recirculation, with continued daily administration of CsA during the postinsult week. In animals which were injected with CsA in daily doses of 10 mg kg−1, but in which no needle was inserted, the drug failed to ameliorate CA1 damage, whether the ischemia had a duration of 7 or 10 min. Likewise, needle insertion had no effect on CA1 damage if CsA was not administered. In contrast, when CsA was given to animals with a needle insertion, CA1 damage was dramatically ameliorated, whether treatment was initiated 1 week before ischemia, or 30 min after the start of recirculation. The effect of CsA seemed larger than that of any other drug proposed to have an anti-ischemic effect in forebrain/global ischemia. Injection of tritiated CsA in one animal with BBB disruption lead to detectable radioactivity throughout the ventricular system, suggesting a generalised increase of the entry of CsA across the BBB. The results demonstrate that immunosuppressants of the type represented by CsA markedly ameliorate delayed neuronal damage after transient forebrain ischemia, provided that they can pass the BBB. It is discussed whether the effect of the drug is one involving calcineurin, a protein phosphatase, or if CsA counteracts a permeability transition of the inner mitochondrial membrane, assumed to occur in response to adverse conditions, e.g. gradual accumulation of Ca2+ in the mitochondria in the postischemic period.

A recent study from this laboratory has shown that brief transient ischemia (2 min 30 s) in normo- and hyperglycemic rats leads to moderate neuronal necrosis in CA1 cells of the hippocampus, of equal density in the two groups. However, hyperglycemic animals failed to depolarize during the ischemia, nor did they show a decrease in extracellular calcium concentration. The present study was undertaken to study the metabolic correlates to these unexpected findings. Normoglycemic (plasma glucose ∼6 mM) and hyperglycemic (∼20 mM) rats were subjected to ischemic periods of 1 min and 2 min 15 s (2 min 30 s with freezing delay considered), and their brains were frozen in situ. Samples of dorsal hippocampus were dissected at –22°C and extracted for the measurement of phosphocreatine (PCr), creatine, ATP, ADP, AMP, glucose, glycogen, and lactate. Normoglycemic animals showed rapid depletion of PCr, ATP, glucose, and glycogen, and a rise in lactate content to 10–12 mM·kg –1 during the ischemia. Hyperglycemic animals displayed a more moderate rate of fall of PCr and ATP, with ATP values exceeding 50% of control after 2 min 30 s. Glycogen stores were largely maintained, but degradation of glucose somewhat enhanced the lactic acidosis. The results demonstrate that hyperglycemic rats maintained ATP at levels sufficient to prevent cell depolarization and calcium influx during the ischemic period. However, the metabolic perturbation observed must have been responsible for the delayed neuronal damage. We speculate that lowered ATP, increased inorganic P, and oxidative stress triggered a delayed mitochondrial permeability transition (MPT), which led to delayed neuronal necrosis. This assumption was supported by a second series of experiments in which CA1 damage in hyperglycemic rats was prevented by cyclosporin A, a virtually specific inhibitor of the MPT.

In this study we explored if the secondary bioenergetic failure, which occurs a few hours after recirculation, following transient middle cerebral artery occlusion (MCAO) in rats, is caused by a compromised reflow. We induced 2 hours of MCAO and measured CBF at the end of the ischemia, as well as 15 minutes, 1, 2, and 4 hours after the start of recirculation, using autoradiographic or tissue sampling 14C-iodoantipyrine techniques. After 2 hours of MCAO, the autoradiographically measured CBF in the ischemic core areas was reduced to 3 to 5% of contralateral values. The reduction in CBF was less in neighboring, penumbral areas. After recirculation, flow already normalized in core tissues after 15 minutes, and remained close to normal for the 4 hours recirculation period studied. However, in penumbral tissues, recovery CBF values were usually below normal. The results show that tissues that are heavily compromised by the 2-hour period of ischemia and are destined to incur infarction, show a "relative hyperemia" during recirculation. In fact, some areas of the previously densely ischemic tissue showed overt hyperperfusion. This finding raises the question whether the relative or absolute hyperemia reflects events that are pathogenetically important. Because drugs that clearly ameliorate the final damage incurred fail to alter the relative hyperperfusion of previously ischemic tissues, it is concluded that vascular events in the reperfusion period do not play a major role in causing the final damage.

This chapter discusses acidosis as a factor for aggravating ischemic brain damage. Acidosis is always observed during dense ischemia, because bioenergetic failure is accompanied by enhanced production of lactate − plus H + and because additional H + is formed in other reactions triggered by anaerobiosis. This acidosis is probably responsible for some of the aspects of the ischemic damage, for example, its pannecrotic character. In support, infarction is not observed in hypoglycemic coma, a condition in which bioenergetic failure and loss of ion homeostasis occurs in the absence of acidosis. There is no direct evidence that acidosis leads to DNA fragmentation, but two series of observations provide a hint. Thus, ultrastructural results obtained in animals with an exaggerated lactic acidosis during a 30-min period of ischemia in rats showed condensation of nuclear chromatin, suggesting fragmentation of DNA. Furthermore, results have been described in which acidosis activates an endonuclease (DNAase II) leading to DNA fragmentation. Therefore, the possibility exists that acidosis causes release of relatively loosely bound iron, thereby triggering iron-catalyzed free radical damage to DNA.

The objective of the present study was to explore whether a diffusible free radical scavenger can ameliorate the pan-neurotic lesions of the substantia nigra, pars reticulate (SNPR), which are incurred in rats subjected to status epilepticus of more than 30 min duration. Vehicle-injected animals had flurothyl seizures induced for 45 min. The seizures were then terminated and the animals were recovered for 7 d to allow histopathological evaluation of the SNPR lesions. Drug-treated animals, which were otherwise treated identically, were given either 100–800 mg/kg of dimethylthiourea (DMTU), a diffusible hydroxyl ion scavenger, or the diffusible spin trap α-phenyl N-tert-butyl nitrone (PBN) in a dose of 100 mg/kg IP. All animals given DMTU died 2 to 8 h after status epilepticus, but PBN was tolerated well by the animals. The amount of flurothyl required to sustain the electrographic seizures was identical in the vehicle- and drug-injected groups, demonstrating that PBN did not suppress seizure activity. Vehicle-injected animals had large bilateral infarcts localized to the SNPR. Of the six animals treated with PBN, one had a small, unilateral lesions, and in all other animals the SNPR had a normal histological appearance. The results strongly suggest that the pan-necrotic lesions of the SNPR incurred during ongoing seizure activity represent a free radical-mediated lesion. Copyright © 1997 Elsevier Science Inc.

Prolonged ischemia is known to damage the blood-brain barrier, causing an increase in vascular permeability to proteins. We studied the time course of extravasation of endogenous albumin in rats after 1 and 2 h of middle cerebral artery (MCA) occlusion followed by 6, 12, and 24 h of recirculation. In a separate group of rats that had undergone 1 h of MCA occlusion and 6 h of recirculation, influx of [14C]aminoisobutyric acid (AIB) from blood to brain was also measured. After 1 h of occlusion followed by 6 h of recirculation, neuronal damage was evident in caudoputamen, but there were no signs of blood-brain barrier leakage to either AIB or albumin. At 12 h, the caudoputamen contained extravasated albumin, and at 24 h extravasation was extended to the somatosensory cortex. Animals subjected to 2 h of MCA occlusion showed albumin extravasation in caudoputamen already at 6 h of recirculation, and at 12 and 24 h albumin was abundant in the major part of the right hemisphere. This study suggests that damage to neurons precedes leakage of the blood-brain barrier. Even a relatively short period of ischemia such as 1 h will result in markedly increased vascular permeability. However, a longer transient ischemic insult disrupts the blood-brain barrier earlier than a shorter one.

Levels of BDNF mRNA and protein were measured in the rat brain using in situ hybridization and a two-site enzyme immunoassay. Under basal conditions, the highest BDNF concentration was found in the dentate gyrus (88 ng/g), while the levels in CA3 (50 ng/g), CA1 (18 ng/g) and parietal cortex (8 ng/g) were markedly lower. Following 10 min of forebrain ischemia, BDNF protein increased transiently in the dentate gyrus (to 124% of control at 6 h after the insult) and CA3 region (to 131% of control, at 1 week after the insult). In CA1 and parietal cortex, BDNF protein decreased to 73-75% of control at 24 h. In contrast, BDNF mRNA expression in dentate granule cells and CA3 pyramidal layer was transiently elevated to 287 and 293% of control, respectively, at 2 h, whereas no change was detected in CA1 or neocortex. The regional BDNF protein levels shown here correlate at least partly with regional differences in cellular resistance to ischemic damage, which is consistent with the hypothesis of a neuroprotective role of BDNF.

Normoglycemic animals subjected to 10-20 min of transient ischemia survive without major neurological symptoms, but incur delayed neuronal damage selectively affecting vulnerable neuronal populations. If the animals are hyperglycemic before ischemia is induced, cell damage develops more rapidly, and postischemic seizures appear after a delay of 18-24 h. This study was designed to assess whether the primary insult, i.e., transient ischemia in hyperglycemic animals, triggers early epileptogenic activity which 'matures' into clinical seizures, or if the seizures arise as a result of secondary events occurring after many hours of recirculation. EEG activity during 20-24 h of postischemic recirculation was recorded from electrodes implanted in the neocortex and hippocampus of freely moving rats which had been subjected to 10 min of ischemia under normoglycemic or hyperglycemic conditions. Normoglycemic animals showed a transient postischemic reduction of EEG amplitude and frequency, and sparse and temporary epileptiform activity. In contrast, hyperglycemic animals showed a more pronounced reduction of EEG amplitude and frequency, and early appearing epileptiform activity which was sustained, and ultimately transformed into overt electrographic seizures. The EEG changes were more pronounced in the neocortex than in the hippocampus. The results thus demonstrate that the initial ischemic insult, and not the secondary damage appearing many hours after the initiation of recirculation, triggers epileptiform activity that 'matures' into status epilepticus.

Levels of mRNA for c-fos, nerve growth factor (NGF), brain-derived neurotrophic factor (BDNF), neurotrophin-3 (NT-3), TrkB, and TrkC were studied using in situ hybridization in the rat brain at different reperfusion times after unilateral middle cerebral artery occlusion (MCAO). Short-term (15 min) MCAO, which does not cause neuronal death, induced elevated BDNF mRNA expression confined to ipsilateral frontal and cingulate cortices outside the ischemic area. With a longer duration of MCAO (2 h), which leads to cortical infarction, the increase was more marked and elevated BDNF mRNA levels were also detected bilaterally in dentate granule cells and CA1 and CA3 pyramidal neurons. Maximum expression was found after 2 h of reperfusion. At 24 h BDNF mRNA expression had returned to control values. In the ischemic core of the parietal cortex only scattered neurons were expressing high levels of BDNF mRNA after 15 min and 2 h of MCAO. Analysis of different BDNF transcripts showed that MCAO induced a marked increase of exon III mRNA but only small increases of exon I and II mRNAs in cortex and hippocampus. In contrast to BDNF mRNA, elevated expression of c-fos mRNA was observed in the entire ipsilateral cerebral cortex, including the ischemic core, after both 15 min and 2 h of MCAO. Two hours of MCAO also induced transient, bilateral increases of NGF and TrkB mRNA levels and a decrease of NT-3 mRNA expression, confined to dentate granule cells. The upregulation of BDNF mRNA expression in cortical neurons after MCAO is probably triggered by glutamate through a spreading depression-like mechanism. The lack of response of the BDNF gene in the ischemic core may be due to suppression of signal transduction or transcription factor synthesis caused by the ischemia. The observed pattern of gene expression after MCAO agrees well with a neuroprotective role of BDNF in cortical neurons. However, elevated levels of NGF and BDNF protein could also increase synaptic efficacy in the postischemic phase, which may promote epileptogenesis.