A selective loss of somatostatin in the hippocampus of patients with temporal lobe epilepsy

Up-regulation of P/Q-type voltage-gated Ca2+ channel immunoreactivity within parvalbumin positive neurons in the rat hippocampus following status epilepticus

To identify the roles of VGCC subtypes in damages/impairs of inhibitory transmission during epileptogenesis, we investigated temporal- and spatial-specific alterations in voltage-gated Ca(2+) channel (VGCC) immunoreactivities within parvalbumin (PV, a Ca(2+) binding protein) positive neurons in the rat hippocampus following status epilepticus (SE). Compared to controls, only P/Q-type (alpha1A) VGCC immunoreactivity was enhanced in PV positive neurons at the early point following SE. The alteration in P/Q-type (alpha1A) VGCC immunoreactivity showed an inverse proportionality to that in PV immunoreactivity in the dentate gyrus and the CA1 region. These findings suggest that SE may induce prolonged up-regulation in P/Q-type VGCC expression within PV positive neurons.

Selective loss of dentate hilar interneurons contributes to reduced synaptic inhibition of granule cells in an electrical stimulation-based animal model of temporal lobe epilepsy

Neuropeptide-containing hippocampal interneurons and dentate granule cell inhibition were investigated at different periods following electrical stimulation-induced, self-sustaining status epilepticus (SE) in rats. Immunohistochemistry for somatostatin (SOM), neuropeptide Y (NPY), parvalbumin (PV), cholecystokinin (CCK), and Fluoro-Jade B was performed on sections from hippocampus contralateral to the stimulated side and studied by confocal laser scanning microscopy. Compared to paired age-matched control animals, there were fewer SOM and NPY-immunoreactive (IR) interneurons in the hilus of the dentate gyrus in animals with epilepsy (40-60 days after SE), and 1, 3, and 7 days following SE. In the hilus of animals that had recently undergone SE, some SOM-IR and NPY-IR interneurons also stained for Fluoro-Jade B. Furthermore, there was electron microscopic evidence of the degeneration of SOM-IR interneurons following SE. In contrast, the number of CCK and PV-IR basket cells in epileptic animals was similar to that in controls, although it was transiently diminished following SE; there was no evidence of degeneration of CCK or PV-IR interneurons. Patch-clamp recordings revealed a diminished frequency of inhibitory postsynaptic currents in dentate granule cells (DGCs) recorded from epileptic animals and animals that had recently undergone SE compared with controls. These results confirm the selective vulnerability of a particular subset of dentate hilar interneurons after prolonged SE. This loss may contribute to the reduced GABAergic synaptic inhibition of granule cells in epileptic animals.

GABA excitation in mouse hilar neuropeptide Y neurons

Neuropeptide Y-containing interneurons in the dentate hilar area play an important role in inhibiting the activity of hippocampal circuitry. Hilar cells are often among the first lost in hippocampal epilepsy. As many types of neurons are found in the hilus, we used a new transgenic mouse expressing green fluorescent protein (GFP) in a subset of neurons that colocalized neuropeptide Y (NPY), somatostatin (SST), and GABA for whole-cell, perforated, and cell-attached recording in 240 neurons. As these neurons have not previously been identifiable in live slices, they have not been the focus of physiological analysis. Hilar NPY neurons showed modest spike frequency adaptation, a large 15.6 +/- 1.0 mV afterhyperpolarization, a mean input resistance of 335 +/- 26 M Omega, and were capable of fast-firing. Muscimol-mediated excitatory actions were found in a nominally Ca(2+)-free/high-Mg(2+) bath solution using cell-attached recording. GABA(A) receptor antagonists inhibited half the recorded neurons and blocked burst firing. Gramicidin perforated-patch recording revealed a GABA reversal potential positive to both the resting membrane potential and spike threshold. Together, these data suggest GABA is excitatory to many NPY cells. NPY and SST consistently hyperpolarized and reduced spike frequency in these neurons. No hyperpolarization of NPY on membrane potential was detected in the presence of tetrodotoxin, AP5, CNQX and bicuculline, supporting an indirect effect. Under similar conditions, SST hyperpolarized the cells, suggesting a direct postsynaptic action. Depolarizing actions of GABA and GABA-dependent burst-firing may synchronize a rapid release of GABA, NPY, and SST, leading to pre- and postsynaptic inhibition of excitatory hippocampal circuits.

Somatostatin in the dentate gyrus

The neuropeptide somatostatin (SST) is expressed in a discrete population of interneurons in the dentate gyrus. These interneurons have their soma in the hilus and project to the outer molecular layer onto dendrites of dentate granule cells, adjacent to perforant path input. SST-containing interneurons are very sensitive to excitotoxicty, and thus are vulnerable to a variety of neurological diseases and insults, including epilepsy, Alzheimer's disease, traumatic brain injury, and ischemia. The SST gene contains a prototypical cyclic AMP response element (CRE) site. Such a regulatory site confers activity-dependence to the gene, such that it is turned on when neuronal activity is high. Thus SST expression is increased by pathological conditions such as seizures and by natural stimulation such as environmental enrichment. SST may play an important role in cognition by modulating the response of neurons to synaptic input. In the dentate, SST and the related peptide cortistatin (CST) reduce the likelihood of generating long-term potentiation, a cellular process involved in learning and memory. Thus these neuropeptides would increase the threshold of input required for acquisition of new memories, increasing "signal to noise" to filter out irrelevant environmental cues. The major mechanism through which SST inhibits LTP is likely through inhibition of voltage-gated Ca(2+) channels on dentate granule cell dendrites. Transgenic overexpression of CST in the dentate leads to profound deficits in spatial learning and memory, validating its role in cognitive processing. A reduction of synaptic potentiation by SST and CST in dentate may also contribute to the well-characterized antiepileptic properties of these neuropeptides. Thus SST and CST are important neuromodulators in the dentate gyrus, and disruption of this signaling system may have major impact on hippocampal function.

Hippocampal Cell Loss in Posttraumatic Human Epilepsy

PURPOSE

We performed this study to determine whether significant head trauma in human adults can result in hippocampal cell loss, particularly in hilar (polymorph) and CA3 neurons, similar to that observed in animal models of traumatic brain injury. We examined the incidence of hippocampal pathology and its relation to temporal neocortical pathology, neuronal reorganization, and other variables.

METHODS

Twenty-one of 200 sequential temporal lobectomies had only trauma as a risk factor for epilepsy. Tissue specimens from temporal neocortex and hippocampus were stained with glial fibrillary acidic protein (GFAP) and hematoxylin and eosin (H&E). Eleven hippocampal specimens had additional analysis of neuronal distributions by using cresyl violet and immunolabeling of a neuron-specific nuclear protein.

RESULTS

The median age at onset of trauma was 19 years, the median time between trauma and onset of seizures was 2 years, and the median epilepsy duration was 16 years. The length of the latent period was inversely related to the age at the time of trauma (r=0.75; Spearman). The neocortex showed gliosis in all specimens, with hemosiderosis (n=8) or heterotopias (n=6) in some, a distribution differing from chance (p=0.02; Fisher). Hippocampal neuronal loss was found in 94% of specimens, and all of these had cell loss in the polymorph (hilar) region of the dentate gyrus. Hilar cell loss ranged from mild, when cell loss was confined to the hilus, to severe, when cell loss extended into CA3 and CA1. Some degree of mossy fiber sprouting was found in the dentate gyrus of all 10 specimens in which it was evaluated. Granule cell dispersion (n=4) was seen only in specimens with moderate to severe neuronal loss.

CONCLUSIONS

Neocortical pathology was universally present after trauma. Neuronal loss in the hilar region was the most consistent finding in the hippocampal formation, similar to that found in the fluid-percussion model of traumatic head injury. These findings support the idea that head trauma can induce hippocampal epilepsy in humans in the absence of other known risk factors.

Somatostatin receptor type 2 undergoes plastic changes in the human epileptic dentate gyrus

Temporal lobe epilepsy (TLE) is characterized by hippocampal sclerosis together with profound losses and phenotypic changes of different classes of interneurons, including those expressing somatostatin (SRIF). To understand the functional significance of the plasticity of SRIF transmission in TLE, unraveling the status of SRIF receptors is, however, a prerequisite. To address this issue, we characterized expression and distribution of the major SRIF receptor, the sst2 subtype, in hippocampal tissue resected in patients with TLE using complementary neuroanatomic approaches. In patients with hippocampal sclerosis, the number of cells expressing sst2 receptor mRNA as well as sst2 receptor-binding sites and immunoreactivity decreased significantly in the CA1-3, reflecting neuronal loss. By contrast, in the dentate gyrus, sst2 receptor mRNA expression was strongly increased in the granule cell layer, and sst2 receptor-binding sites and immunoreactivity was preserved in the inner but decreased significantly in the outer molecular layer. In this latter region, pronounced changes in SRIF terminal fields were observed. Decreased receptor density in the distal dendrites of granule cells is likely to reflect downregulation of sst2 receptors in response to physiopathologic release of SRIF. Because sst2 receptors have anticonvulsant and antiepileptogenic properties, this phenomenon may contribute to the etiology of TLE seizures.

Axonal sprouting of GABAergic interneurons in temporal lobe epilepsy

Temporal lobe epilepsy is one of the most common forms of epilepsy. Numerous contributing factors and compensatory mechanisms have been associated with temporal lobe epilepsy. One feature found in both humans and animal models is sprouting of hippocampal principal cell axons, which suggests that axonal sprouting may be a general phenomenon associated with temporal lobe epilepsy. This article highlights the evidence showing that hippocampal GABAergic interneurons also undergo axonal sprouting in temporal lobe epilepsy. The caveats and unanswered questions associated with the current data and the potential physiological consequences of reorganizations in GABAergic circuits are discussed.

Antiepileptic drug-resistant rats differ from drug-responsive rats in hippocampal neurodegeneration and GABA(A) receptor ligand binding in a model of temporal lobe epilepsy

The disabling seizures associated with mesial temporal lobe epilepsy (TLE) are often resistant to antiepileptic drugs (AEDs). The biological basis of this refractoriness is unknown but may include alterations in AED targets in the epileptogenic brain tissue, reduced AED penetration to the seizure focus, and neuropathological brain alterations such as hippocampal sclerosis typically found in patients with refractory TLE. In the present study, we used a rat model of TLE to examine whether AED responders differ from non-responders in their structural alterations and GABA(A) receptor characteristics in the hippocampal formation. In this model, spontaneous recurrent seizures develop after a status epilepticus induced by prolonged electrical stimulation of the basolateral amygdala. The frequency of these seizures was recorded by continuous video/EEG monitoring before, during, and after daily treatment with phenobarbital, which was given at maximum tolerated doses for 2 weeks. Based on their individual response to phenobarbital, rats were grouped into responders and non-responders. The severity or duration of the initial brain insult (the status epilepticus) did not differ between responders and non-responders, indicating that the difference between the two subgroups is genetically determined. Subsequent histological examination showed a significant loss of neurons in the CA1, CA3c/CA4, and dentate hilus of non-responders, whereas responders did not differ in this respect from non-epileptic controls. The morphological alterations in the non-responders were associated with striking alterations in autoradiographic imaging of diazepam-sensitive and diazepam-insensitive GABA(A) receptor binding in the dentate gyrus with a significant shift to enhanced diazepam-insensitive binding. The present data indicate that neurodegeneration and associated GABA(A) receptor changes in the dentate gyrus are critically involved in the mechanisms underlying refractoriness of seizures in TLE.

Reduced calcium binding protein immunoreactivity induced by electroconvulsive shock indicates neuronal hyperactivity, not neuronal death or deactivation

Calcium-binding proteins (CBPs), such as parvalbumin and calbindin D-28k, are useful markers of specific neuronal types in the CNS. In recent studies, expression of CBPs may be indicative of a deactivated neuronal state, particularly epilepsy. However, it is controversial whether altered expression of CBPs in the hippocampus practically indicate neuronal activity. Therefore, the present study was performed to investigate the extent of profiles of expression of CBPs in the rat hippocampus affected by several episodes induced by electroconvulsive shock. In the present study, following electroconvulsive shock expression of CBPs were reduced in the hippocampus in a stimulus-dependent manner, and recovered to the control level at 6 h after electroconvulsive shock. However, paired-pulse responses of the dentate gyrus were transiently impaired by electroconvulsive shock, and immediately normalized to baseline value. In addition, effects of electroconvulsive shock on expression of CBPs and paired-pulse responses were prevented by pretreatment of vigabatrin. These findings suggest that reduced expression of CBPs induced by seizure activity may be indicative of hyperactivity of CBP positive neurons, which is a practical consequence of the abnormal discharge, and that they may play an important role in regulating seizure activity.

Region- and age-specific deficits in γ-aminobutyric acidergic neuron development in the telencephalon of theuPAR-/- mouse

We have previously shown that in adult mice with a null mutation in the urokinase-type plasminogen activator receptor (uPAR) gene, maintained on a C57BL/6J/129Sv background, there is a selective loss of GABAergic interneurons in anterior cingulate and parietal cortex, with the parvalbumin-expressing subpopulation preferentially affected. Here, we performed a more detailed anatomical analysis of uPAR(-/-) mutation on the congenic C57BL/6J background. With glutamic acid decarboxylase-67 and gamma-aminobutyric acid (GABA) immunostaining, there is a similar region-selective loss of cortical interneurons in the congenic uPAR(-/-) mice from the earliest age examined (P21). In contrast, the loss of parvalbumin-immunoreactive cells is observed only in adult cortex, and the extent of this loss is less than in the mixed background. Moreover, earlier in development, although there are normal numbers of parvalbumin cells in the uPAR(-/-) cortex, fewer cells coexpress GABA, suggesting that the parvalbumin subpopulation migrates appropriately to the cortex, but does not differentiate normally. Among the other forebrain regions examined, only the adult hippocampus shows a loss of GABAergic interneurons, although the somatostatin, rather than the parvalbumin, subpopulation contributes to this loss. The data suggest that uPAR function is necessary for the normal development of a subpopulation of GABAergic neurons in the telencephalon. It is likely that the late-onset parvalbumin phenotype is due to the effects of an altered local environment on selectively vulnerable neurons and that the extent of this loss is strain dependent. Thus, an interplay between complex genetic factors and the environment may influence the phenotypic impact of the uPAR mutation both pre- and postnatally.

Plasticity of somatostatin and somatostatin sst2A receptors in the rat dentate gyrus during kindling epileptogenesis

Increasing evidence suggests that somatostatin may control neuronal excitability during epileptogenesis. In the hippocampus, sst2A receptors are likely to mediate somatostatin inhibitory actions but little is known about their status in kindled tissues. In the present study, sst2A receptor and somatostatin immunoreactivity were examined by confocal microscopy in the hippocampus during and after kindling acquisition. In control rats, somatostatin-positive axon terminals were mainly found in the stratum lacunosum moleculare of CA1 area and in the outer molecular layer of the dentate gyrus. sst2A receptor immunoreactivity was diffusely distributed in the strata radiatum and oriens of CA1 and in the stratum moleculare of the dentate gyrus. Immunogold electron microscopy revealed that sst2A receptors were predominantly localized postsynaptically, at the plasma membrane of dendritic shafts and spines of principal neurons. During kindling epileptogenesis, qualitative and semiquantitative analysis revealed a progressive decrease of sst2A immunoreactivity in the outer molecular layer, which was spatially associated with an increase in somatostatin immunoreactivity. No obvious changes in sst2A receptor immunoreactivity were observed in other hippocampal subfields. These results suggest that the decrease of sst2A receptor immunoreactivity in the outer molecular layer reflects receptor down-regulation in distal dendrites of granule cells in response to chronic somatostatin release. Because the sst2A receptor appears to mediate anticonvulsant and antiepileptogenic effects of somatostatin, this may represent a pivotal mechanism contributing to epileptogenesis.

Rapid and long-term alterations of hippocampal GABAB receptors in a mouse model of temporal lobe epilepsy

Alterations of gamma-aminobutyric acid (GABA)B receptor expression have been reported in human temporal lobe epilepsy (TLE). Here, changes in regional and cellular expression of the GABAB receptor subunits R1 (GBR1) and R2 (GBR2) were investigated in a mouse model that replicates major functional and histopathological features of TLE. Adult mice received a single, unilateral injection of kainic acid (KA) into the dorsal hippocampus, and GABAB receptor immunoreactivity was analysed between 1 day and 3 months thereafter. In control mice, GBR1 and GBR2 were distributed uniformly across the dendritic layers of CA1-CA3 and dentate gyrus. In addition, some interneurons were labelled selectively for GBR1. At 1 day post-KA, staining for both GBR1 and GBR2 was profoundly reduced in CA1, CA3c and the hilus, and no interneurons were visible anymore. At later stages, the loss of GABAB receptors persisted in CA1 and CA3, whereas staining increased gradually in dentate gyrus granule cells, which become dispersed in this model. Most strikingly, a subpopulation of strongly labelled interneurons reappeared, mainly in the hilus and CA3 starting at 1 week post-KA. In double-staining experiments, these cells were selectively labelled for neuropeptide Y. The number of GBR1-positive interneurons also increased contralaterally in the hilus. The rapid KA-induced loss of GABAB receptors might contribute to epileptogenesis because of a reduction in both presynaptic control of transmitter release and postsynaptic inhibition. In turn, the long-term increase in GABAB receptors in granule cells and specific subtypes of interneurons may represent a compensatory response to recurrent seizures.

Cellular compartmentalization of phosphorylated eIF2alpha and neuronal NOS in human temporal lobe epilepsy with hippocampal sclerosis

Hippocampal sclerosis (HS) is the most common neuropathologic finding in patients with medically refractory temporal lobe epilepsy (TLE). The mechanisms resulting in neuronal injury and cell loss in HS are incompletely understood, but inhibition of protein synthesis may play a pivotal role in these processes. This study examined the relationships between two molecules known to be involved in reduced protein synthesis in animals subjected to traumatic brain injury. Translational initiation of protein synthesis is inhibited when 2alpha (eIF2alpha) is phosphorylated. Recently, nitric oxide (NO) has been shown to reduce protein synthesis by inducing phosphorylation of eIF2alpha. We performed immunocytochemistry for eIF2alpha(P) and histochemistry (NADPH-D reaction) for nitric oxide synthase (NOS) to determine the distribution of these molecules in hippocampi removed from patients undergoing anterior temporal lobectomy (ATL) for medically intractable TLE due to HS. The greatest number of eIF2alpha(P) positive cells was in the CA1 sector of the hippocampus, followed by the hilus of the dentate gyrus. NADPH-D positive neurons were observed most often in the hilus. Labeling in both instances involved neuronal cell body cytoplasm and varicose processes. Combination of both staining procedures revealed close relationships between differentially labeled neurons within the hilus. The results suggest that NO participates in the phosphorylation of eIF2alpha since we demonstrated that nNOS processes are closely related to eIF2alpha(P) positive cells. This may occur through activation of kinases such as PERK, which was recently revealed. In human, TLE protein synthesis inhibition may occur at the translational level since the eIF2alpha (P) labeling is cytoplasmic. Protein synthesis inhibition may contribute to neuronal cell injury and death in HS.

Somatostatin depresses long-term potentiation and Ca2+ signaling in mouse dentate gyrus

The selective loss of somatostatin (SST)-containing interneurons from the hilus of the dentate gyrus is a hallmark of epileptic hippocampus. The functional consequence of this loss, including its contribution to postseizure hyperexcitability, remains unclear. We address this issue by characterizing the actions of SST in mouse dentate gyrus using electrophysiological techniques. Although the majority of dentate SST receptors are located in the outer molecular layer adjacent to lateral perforant path (LPP) synapses, we found no consistent action of SST on standard synaptic responses generated at these synapses. However, when SST was present during application of high-frequency trains that normally generate long-term potentiation (LTP), the induction of LTP was impaired. SST did not alter the maintenance of LTP when applied after its induction. To examine the mechanism by which SST inhibits LTP, we recorded from dentate granule cells and examined the actions of this neuropeptide on synaptic transmission and postsynaptic currents. Unlike findings in the CA1 hippocampus, we observed no postsynaptic actions on K(+) currents. Instead, SST inhibited Ca(2+)/Ba(2+) spikes evoked by depolarization. This inhibition was dependent on N-type Ca(2+)currents. Blocking these currents also blocked LTP, suggesting a mechanism through which SST may inhibit LTP. Our results indicate that SST reduction of dendritic Ca(2+) through N-type Ca(2+) channels may contribute to modulation of synaptic plasticity at LPP synapses. Therefore the loss of SST function postseizure could result in abnormal synaptic potentiation that contributes to epileptogenesis.

Preservation of perisomatic inhibitory input of granule cells in the epileptic human dentate gyrus

Temporal lobe epilepsy is known to be associated with hyperactivity that is likely to be generated or amplified in the hippocampal formation. The majority of granule cells, the principal cells of the dentate gyrus, are found to be resistant to damage in epilepsy, and may serve as generators of seizures if their inhibition is impaired. Therefore, the parvalbumin-containing subset of interneurons, known to provide the most powerful inhibitory input to granule cell somata and axon initial segments, were examined in human control and epileptic dentate gyrus. A strong reduction in the number of parvalbumin-containing cells was found in the epileptic samples especially in the hilar region, although in some patches of the granule cell layer parvalbumin-positive terminals that form vertical clusters characteristic of axo-axonic cells were more numerous than in controls. Analysis of the postsynaptic target elements of parvalbumin-positive axon terminals showed that they form symmetric synapses with somata, dendrites, axon initial segments and spines as in the control, but the ratio of axon initial segment synapses was increased in the epileptic tissue (control: 15.9%, epileptic: 31.3%). Furthermore, the synaptic coverage of granule cell axon initial segments increased more than three times (control: 0.52, epileptic: 2.10 microm synaptic length/100 microm axon initial segment membrane) in the epileptic samples, whereas the amount of somatic symmetric synapses did not change significantly. Although the number of parvalbumin-positive interneurons is decreased, the perisomatic inhibitory input of dentate granule cells is preserved in temporal lobe epilepsy. Basket and axo-axonic cell terminals - whether positive or negative for parvalbumin - are present, moreover, the axon collaterals targeting axon initial segments sprout in the epileptic dentate gyrus. We suggest that perisomatic inhibitory interneurons survive in epilepsy, but their somadendritic compartment and partly the axon loses parvalbumin or immunoreactivity for parvalbumin. The hyperinnervation of axon initial segments might be a compensatory change in the inhibitory network, but at the same time may lead to a more effective synchronization of granule cell firing that could contribute to the generation or amplification of epileptic seizures.

Heightened seizure severity in somatostatin knockout mice

Patients and experimental models of temporal lobe epilepsy display loss of somatostatinergic neurons in the dentate gyrus. To determine if loss of the peptide somatostatin contributes to epileptic seizures we examined kainate-evoked seizures and kindling in somatostatin knockout mice. Somatostatin knockout mice were not observed to experience spontaneous seizures. Timm staining, acetylcholinesterase histochemistry, and immunocytochemistry for NPY, calbindin, calretinin, and parvalbumin revealed no compensatory changes or developmental abnormalities in the dentate gyrus of somatostatin knockout mice. Optical fractionator counting of Nissl-stained hilar neurons showed similar numbers of neurons in wild type and somatostatin knockout mice. Mice were treated systemically with kainic acid to evoke limbic seizures. Somatostatin knockout mice tended to have a shorter average latency to stage 5 seizures, their average maximal behavioral seizure score was higher, and they tended to be more likely to die than controls. In response to kindling by daily electrical stimulation of the perforant path, to more specifically challenge the dentate gyrus, mean afterdischarge duration in somatostatin knockout mice was slightly longer, but the number of treatments to five stage 4-5 seizures was similar to controls. Although we cannot exclude the possibility of undetected compensatory mechanisms in somatostatin knockout mice, these findings suggest that somatostatin may be mildly anticonvulsant, but its loss alone is unlikely to account for seizures in temporal lobe epilepsy.

Expression of metabotropic glutamate receptor 1alpha in the hippocampus of rat pilocarpine model of status epilepticus

The expression of metabotropic glutamate receptor 1alpha was studied in the rat hippocampus after pilocarpine-induced status epilepticus by Western blot and immunocytochemistry at both light and electron microscopic levels. At 1 day after pilocarpine-induced status epilepticus, there was marked decrease in metabotropic glutamate receptor 1alpha immunoreactivity at the border between stratum oriens and alveus in CA1 and CA3, and in the hilus of dentate gyrus. Between 3 and 31 days after pilocarpine-induced status epilepticus, metabotropic glutamate receptor 1alpha-immunoreactive dendrites and cell bodies in the border between stratum oriens and alveus gradually reappeared. Upregulation of metabotropic glutamate receptor 1alpha, however, was observed in the stratum oriens of CA1 at day 1, but returned to baseline by day 7. By electron microscopy, the metabotropic glutamate receptor 1alpha-immunoreactive product was demonstrated only in the post-synaptic elements in the border between the stratum oriens and alveus of CA1 and the hilus of the dentate gyrus in both control and experimental rats. At 1 day after pilocarpine-induced status epilepticus, metabotropic glutamate receptor 1alpha-immunoreactive degenerating neurons were identified in the border between stratum oriens and alveus of CA1 and the hilus of the dentate gyrus. At 7 and 31 days, many degenerating axons were also found. Present results suggest that excitoneurotoxicity mediated through post-synaptic metabotropic glutamate receptor 1alpha may be involved in degeneration and death of interneurons in the hilus of dentate gyrus, and the border between stratum oriens and alveus of CA1 in the early stage after pilocarpine-induced status epilepticus.

Early loss of interneurons and delayed subunit-specific changes in GABA(A)-receptor expression in a mouse model of mesial temporal lobe epilepsy

Unilateral injection of kainic acid (KA) into the dorsal hippocampus of adult mice induces spontaneous recurrent partial seizures and replicates histopathological changes observed in human mesial temporal lobe epilepsy (MTLE) (Bouilleret V et al., Neuroscience 1999; 89:717-729). Alterations in pre- and postsynaptic components of GABAergic neurotransmission were investigated immunohistochemically at different time points (1-120 days) in this mouse model of MTLE. Markers of GABAergic interneurons (parvalbumin, calbindin-D28k, and calretinin), the type-1 GABA transporter (GAT1), and major GABA(A)-receptor subunits expressed in the hippocampal formation were analyzed. Acutely, KA injection produced a profound loss of hilar cells but only limited damage to CA1 and CA3 pyramidal cells. In addition, parvalbumin and calbindin-D28k staining of interneurons disappeared irreversibly in CA1 and dentate gyrus (DG), whereas calretinin staining was spared. The prominent GABA(A)-receptor alpha1 subunit staining of interneurons also disappeared after KA treatment, suggesting acute degeneration of these cells. Likewise, GAT1 immunoreactivity revealed degenerating terminals at 24 h post-KA in CA1 and DC and subsided almost completely thereafter. Loss of CA1 and, to a lesser extent, CA3 neurons became evident at 7-15 days post-KA. It was more accentuated after 1 month, accompanied by a corresponding reduction of GABA(A)-receptor staining. In contrast, DC granule cells were markedly enlarged and dispersed in the molecular layer and exhibited a prominent increase in GABA(A)-receptor subunit staining. After 4 months, the dorsal CA1 area was lost almost entirely, CA3 was reduced, and the DG represented most of the remaining dorsal hippocampal formation. No significant morphological alterations were detected contralaterally. These results suggest that loss of hilar cells and GABAergic neurons contributes to epileptogenesis in this model of MTLE. In contrast, long-term degeneration of pyramidal cells and granule cell dispersion may reflect distinct responses to recurrent seizures. Finally, GABA(A)-receptor upregulation in the DG may represent a compensatory response persisting for several months in epileptic mice.

Kainic acid-induced mossy fiber sprouting and synapse formation in the dentate gyrus of rats

In the kainic acid (KA) model of temporal lobe epilepsy, mossy fibers (MFs) are thought to establish recurrent excitatory synaptic contacts onto granule cells. This hypothesis was tested by intracellular labeling of granule cells with biocytin and identifying their synaptic contacts in the dentate molecular layer with electron microscopic (EM) techniques. Twenty-three granule cells from KA-treated animals and 14 granule cells from control rats were examined 2 to 4 months following KA at the light microscopic (LM) level; four cells showing MF sprouting were further characterized at the EM level. Timm staining revealed a time-dependent growth of aberrant MFs into the dentate inner molecular layer. The degree of sprouting was generally (but not invariably) correlated with the severity and frequency of seizures. LM examination of individual biocytin-labeled MF axon collaterals revealed enhanced collateralization and significantly increased numbers of synaptic MF boutons in the hilus compared to controls, as well as aberrant MF growth into the granule cell and molecular layers. EM examination of serially reconstructed, biocytin-labeled MF collaterals in the molecular layer revealed MF boutons that form asymmetrical synapses with dendritic shafts and spines of granule cells, including likely autaptic contacts on parent dendrites of the biocytin-labeled granule cell. These results constitute ultrastructural evidence for newly formed excitatory recurrent circuits, which might provide a structural basis for enhanced excitation and epileptogenesis in the hippocampus of KA-treated rats.

Cellular biology of somatostatin receptors

Somatostatin, and the recently discovered neuropeptide cortistatin, exert their physiological actions via a family of six G protein-coupled receptors (sst1, sst2A, sst2B, sst3, sst4, sst5). Following the cloning of somatostatin receptors significant advances have been made in our understanding of their molecular, pharmacological and signaling properties although much progress remains to be done to define their physiological role in vivo. In this review, the present knowledge regarding neuroanatomical localization, signal transduction pathways, desensitization and internalization properties of somatostatin receptors is summarized. Evidence that somatostatin receptors can form homo- and heterodimers and can physically interact with members of the SSTRIP/Shank/ProSAP1/CortBP1 family is also discussed.

Dendritic but not somatic GABAergic inhibition is decreased in experimental epilepsy

Impaired inhibition is thought to be important in temporal lobe epilepsy (TLE), the most common form of epilepsy in adult patients. We report that, in experimental TLE, spontaneous GABAergic inhibition was increased in the soma but reduced in the dendrites of pyramidal neurons. The former resulted from the hyperactivity of somatic projecting interneurons, whereas the latter was probably due to the degeneration of a subpopulation of dendritic projecting interneurons. A deficit in dendritic inhibition could reduce seizure threshold, whereas enhanced somatic inhibition would prevent the continuous occurrence of epileptiform activity.

The over-expression of somatostatin in the gerbil entorhinal cortex induced by seizure

In present study, we investigated the immunohistochemical distribution of somatostatin (SRIF) in the hippocampal complex of the Mongolian gerbil and its association with different sequelae of spontaneous seizures, in an effort to identify the roles of SRIF in the self-recovery mechanisms in these animals. In the dentate gyrus and subiculum, SRIF immunoreactive (SRIF(+)) cells were similar in both the seizure resistant and the pre-seizure group of seizure sensitive gerbils. Interestingly, SRIF immunoreactivity was markedly decreased until 12 h postictal. Twenty-four hours after the on-set of seizure, the distribution of SRIF immunoreactivity in these regions had slightly increased. In contrast, in the entorhinal cortex the population of SRIF(+) cells and their density were significantly elevated compared to pre-seizure group 30 min postictal. Twelve hours after the on-set of seizure, however, the population of SRIF(+) cells and their density declined, approximately 70-80% compared to the situation at 30 min postictal. These findings suggest that the enhancement of SRIF expression in gerbil entorhinal cortex may affect tissue excitability and have a role in modulating recurrent excitation following seizures.

Loss of hilar mossy cells in Ammon's horn sclerosis

PURPOSE

Hilar mossy cells represent an important excitatory subpopulation of the hippocampal formation. Several studies have identified this cell type as particularly vulnerable to seizure activity in rat models of limbic epilepsy. Here we have subjected hilar mossy cell loss in the hippocampus of patients with chronic temporal lobe epilepsy (TLE) to a systematic morphological and immunohistochemical analysis.

METHODS

Hippocampal specimens from 30 TLE patients were included; 21 patients presented with segmental neuronal cell loss [Ammon's horns clerosis (AHS)] and 8 with focal lesions (tumors, scars, malformations) not involving the hippocampus proper. In one additional TLE patient, no histopathological alteration could be observed. Surgical specimens from tumor patients without epilepsy (n = 2) and nonepileptic autopsy brains (n = 8) were used as controls. Hilar mossy cells in the human hippocampus were visualized using a novel polycloncal antiserum directed against the metabotropic glutamate receptor subtype mGluR7b or by intracellular Lucifer Yellow injection, confocal laser scanning microscopy, and three-dimensional morphological reconstruction.

RESULTS

Compared with controls, a significant loss of mGluR7 immunoreactive mossy cells was observed in patients with AHS (p < 0.05). In contrast, TLE patients with focal lesions but structurally intact hippocampus demonstrated only a discrete, nonsignificant reduction of this neuronal subpopulation. This observation was confirmed by analysis of 62 randomly injected hilar neurons from AHS patients, in which we were unable to detect neurons with a morphology like that of hilar mossy cells.

CONCLUSION

Our present data indicate significant hilar mossy cell loss in TLE patients with AHS. In contrast, hilar mossy cells appear to be less vulnerable in patients with lesion-associated TLE. Although the significance of mGluR7 immunoreactivity in mossy cells remains to be studied, loss of this cell population is compatible with alterations in hippocampal networks and regional hyperexcitability as pathogenic mechanism of AHS and TLE.

What is GABAergic inhibition? How is it modified in epilepsy?

A deficit of gamma-aminobutyric acid-ergic (GABAergic) inhibition is hypothesized to underlie most forms of epilepsy. Although apparently a straightforward and logical hypothesis to test, the search for a deficit of GABAergic inhibition in epileptic tissue has revealed itself to be as difficult as the quest for the Holy Grail. The investigator faces many obstacles, including the multiplicity of GABAergic inhibitory pathways and the multiplicity of variables that characterize the potency of inhibition within each inhibitory pathway. Perhaps more importantly, there seems to be no consensual definition of GABAergic inhibition. The first goal of this review is to try to clarify the notion of GABAergic inhibition. The second goal is to summarize our current knowledge of the various alterations that occur in the GABAergic pathways in temporal lobe epilepsy. Two important features will emerge: (a) according to the variable used to measure GABAergic inhibition, it may appear increased, decreased, or unchanged; and (b) these modifications are brain area- and inhibitory pathway-specific. The possible functional consequences of these alterations are discussed.

Selective alterations in GABAA receptor subtypes in human temporal lobe epilepsy

Temporal lobe epilepsy (TLE) is associated with impaired inhibitory neurotransmission. Studies in animal models suggest that GABA(A) receptor dysfunction contributes to epileptogenesis. To understand the mechanisms underlying TLE in humans, it is fundamental to determine whether and how GABA(A) receptor subtypes are altered. Furthermore, identifying novel receptor targets is a prerequisite for developing selective antiepileptic drugs. We have therefore analyzed subunit composition and distribution of the three major GABA(A) receptor subtypes immunohistochemically with subunit-specific antibodies (alpha1, alpha2, alpha3, beta2,3, and gamma2) in surgical specimens from TLE patients with hippocampal sclerosis (n = 16). Profound alterations in GABA(A) receptor subtype expression were observed when compared with control hippocampi (n = 10). Although decreased GABA(A) receptor subunit staining, reflecting cell loss, was observed in CA1, CA3, and hilus, the distinct neuron-specific expression pattern of the alpha-subunit variants observed in controls was markedly changed in surviving neurons. In granule cells, prominent upregulation mainly of the alpha2-subunit was seen on somata and apical dendrites with reduced labeling on basal dendrites. In CA2, differential rearrangement of all three alpha-subunits occurred. Moreover, there was layer-specific loss of alpha1-subunit-immunoreactive interneurons in hippocampus proper, whereas surviving interneurons exhibited extensive changes in dendritic morphology. Throughout, expression patterns of beta2,3- and gamma2-subunits largely followed those of alpha-subunit variants. These results demonstrate unique subtype-specific expression of GABA(A) receptors in human hippocampus. The significant reorganization of distinct receptor subtypes in surviving hippocampal neurons of TLE patients with hippocampal sclerosis underlines the potential for synaptic plasticity in the human GABA system.

Neurodegenerative and morphogenic changes in a mouse model of temporal lobe epilepsy do not depend on the expression of the calcium-binding proteins parvalbumin, calbindin, or calretinin

The functional role of the calcium-binding proteins parvalbumin, calretinin, and calbindin D-28k for epileptogenesis and long-term seizure-related alterations of the hippocampal formation was assessed in single- and double-knockout mice, using a kainate model of mesial temporal lobe epilepsy. The effects of a unilateral intrahippocampal injection of kainic acid were assessed at one day, 30 days, and four months post-injection, using various markers of GABAergic interneurons (GABA-transporter type 1, GABA(A)-receptor alpha1 subunit, calretinin, calbindin D-28k, somatostatin, and neuropeptide Y). Parvalbumin-deficient, parvalbumin/calbindin-deficient, and parvalbumin/calretinin-deficient mice exhibited no difference in cytoarchitecture of the hippocampal formation and in the number, distribution, or morphology of interneurons compared to wild-type mice. Likewise, mutant mice were not more vulnerable to acute kainate-induced excitotoxicity or to long-term effects of recurrent focal seizures, and exhibited the same pattern of neurochemical alterations (e.g., bilateral induction of neuropeptide Y in granule cells) and morphogenic changes (enlargement and dispersion of dentate gyrus granule cells) as wild-type animals. Quantification of interneurons revealed no significant difference in neuronal vulnerability among the genotypes.These results indicate that the calcium-binding proteins investigated here are not essential for determining the neurochemical phenotype of interneurons. Furthermore, they are not protective against kainate-induced excitotoxicity in this model, and do not appear to modulate the overall level of excitability of the hippocampus. Finally, seizure-induced changes in gene expression in granule cells, which normally express high levels of calcium-binding proteins, apparently were not affected by the gene deletions analysed.

Acute stress increases neuropeptide Y mRNA within the arcuate nucleus and hilus of the dentate gyrus

The effects acute restraint stress on neuropeptide Y (NPY) mRNA expression were determined within the dentate gyrus and arcuate nucleus, where the effects of adrenal steroid action were previously reported. Adult male rats were exposed to 1 h of restraint stress and then sacrificed immediately, 6 h, or 24 h later. Controls were undisturbed. Stress increased NPY mRNA levels in both the arcuate nucleus and in the hilar region of the hippocampus with different time courses. NPY mRNA increased in the arcuate at 24 h, but not earlier, as determined by film autoradiography. Single cell grain analysis was performed in the dentate gyrus hilus because the NPY mRNA was heterogeneously distributed and revealed that the number of cells expressing NPY mRNA increased 6 h after stress, returning to control levels within 24 h. These results fit with previously reported effects of adrenal steroids modulating arcuate nucleus NPY expression through the adrenal steroid Type II receptors. In the hilus where adrenal steroid Type I receptors have been reported to suppress NPY mRNA levels, the effect of stress is in the opposite direction to that of adrenal steroid action and a more complex regulation of NPY expression is indicated.

Distribution, targeting, and internalization of the sst4 somatostatin receptor in rat brain

Somatostatin mediates its diverse physiological effects through a family of five G-protein-coupled receptors (sst(1)-sst(5)); however, knowledge about the distribution of individual somatostatin receptor proteins in mammalian brain is incomplete. In the present study, we have examined the regional and subcellular distribution of the somatostatin receptor sst(4) in the rat CNS by raising anti-peptide antisera to the C-terminal tail of sst(4). The specificity of affinity-purified antibodies was demonstrated using immunofluorescent staining of HEK 293 cells stably transfected with an epitope-tagged sst(4) receptor. In Western blotting, the antiserum reacted specifically with a broad band in rat brain, which migrated at approximately 70 kDa before and approximately 50 kDa after enzymatic deglycosylation. sst(4)-Like immunoreactivity was most prominent in many forebrain regions, including the cerebral cortex, hippocampus, striatum, amygdala, and hypothalamus. Analysis at the electron microscopic level revealed that sst(4)-expressing neurons target this receptor preferentially to their somatodendritic domain. Like the sst(2A) receptor, sst(4)-immunoreactive dendrites were often closely apposed by somatostatin-14-containing fibers and terminals. However, unlike the sst(2A) receptor, sst(4) was not internalized in response to intracerebroventricular administration of somatostatin-14. After percussion trauma of the cortex, neuronal sst(4) receptors progressively declined at the sites of damage. This decline coincided with an induction of sst(4) expression in cells with a glial-like morphology. Together, this study provides the first description of the distribution of immunoreactive sst(4) receptor proteins in rat brain. We show that sst(4) is strictly somatodendritic and most likely functions in a postsynaptic manner. In addition, the sst(4) receptor may have a previously unappreciated function during the neuronal degeneration-regeneration process.

Cerebrospinal fluid somatostatin levels in febrile seizures and epilepsy in children

We analysed the level of cerebrospinal fluid (CSF) somatostatin in children with febrile seizures and epilepsy. In the febrile seizure group (n = 23), the somatostatin level was 83.9 +/- 11.2 pg/ml, which was significantly higher than that of age-matched controls. CSF samples obtained within 3 h of the last seizure had higher somatostatin levels (106.1 +/- 12.4 pg/ml;n = 14) than did the CSF obtained after 3 h (49.4 +/- 15.6 pg/ml;n = 9). The mean somatostatin level in the epilepsy group was 35.3 +/- 4.3 pg/ml (n = 34), and was distributed as follows: 27.6 +/- 3.6 pg/ml in the idiopathic generalized epilepsy group (n = 16), 44.0 +/- 9.4 pg/ml in the symptomatic generalized epilepsy group (n = 13), and 37.2 +/- 10.1 pg/ml in the partial epilepsy group (n = 5). The levels in each group were significantly higher than those in age-matched controls. Somatostatin is a hypothalamic tetradecapeptide with excitatory effects on neurons in children with febrile seizures and epilepsy. The finding that patients with convulsive disease had elevated levels of CSF somatostatin suggests that somatostatin release is somehow related to seizure activity. It remains to be determined whether this is due to increased release from over-active excitatory neurons or leakage from damaged or anoxic neurons, secondary to seizure activity.

Highly specific neuron loss preserves lateral inhibitory circuits in the dentate gyrus of kainate-induced epileptic rats

Patients with temporal lobe epilepsy display neuron loss in the hilus of the dentate gyrus. This has been proposed to be epileptogenic by a variety of different mechanisms. The present study examines the specificity and extent of neuron loss in the dentate gyrus of kainate-treated rats, a model of temporal lobe epilepsy. Kainate-treated rats lose an average of 52% of their GAD-negative hilar neurons (putative mossy cells) and 13% of their GAD-positive cells (GABAergic interneurons) in the dentate gyrus. Interneuron loss is remarkably specific; 83% of the missing GAD-positive neurons are somatostatin-immunoreactive. Of the total neuron loss in the hilus, 97% is attributed to two cell types-mossy cells and somatostatinergic interneurons. The retrograde tracer wheat germ agglutinin (WGA)-apoHRP-gold was used to identify neurons with appropriate axon projections for generating lateral inhibition. Previously, it was shown that lateral inhibition between regions separated by 1 mm persists in the dentate gyrus of kainate-treated rats with hilar neuron loss. Retrogradely labeled GABAergic interneurons are found consistently in sections extending 1 mm septotemporally from the tracer injection site in control and kainate-treated rats. Retrogradely labeled putative mossy cells are found up to 4 mm from the injection site, but kainate-treated rats have fewer than controls, and in several kainate-treated rats virtually all of these cells are missing. These findings support hypotheses of temporal lobe epileptogenesis that involve mossy cell and somatostatinergic neuron loss and suggest that lateral inhibition in the dentate gyrus does not require mossy cells but, instead, may be generated directly by GABAergic interneurons.

Brain somatostatin: a candidate inhibitory role in seizures and epileptogenesis

Recent evidence shows that neuropeptide expression in the CNS is markedly affected by seizure activity, particularly in the limbic system. Changes in neuropeptides in specific neuronal populations depend on the type and intensity of seizures and on their chronic sequelae (i.e. neurodegeneration and spontaneous convulsions). This paper reviews the effects of seizures on somatostatin-containing neurons, somatostatin mRNA and immunoreactivity, the release of this peptide and its receptor subtypes in the CNS. Differences between kindling and status epilepticus in rats are emphasized and discussed in the light of an inhibitory role of somatostatin on hippocampal excitability. Pharmacological studies show that somatostatin affects electrophysiological properties of neurons, modulates classical neurotransmission and has anticonvulsant properties in experimental models of seizures. This peptidergic system may be an interesting target for pharmacological attempts to control pathological hyperactivity in neurons, thus providing new directions for the development of novel anticonvulsant treatments.

Molecular neuropathology of human mesial temporal lobe epilepsy

With the recent progress in surgical treatment modalities, human brain tissue from patients with intractable focal epilepsies will increasingly become available for studies on the molecular pathology, electrophysiological changes and pathogenesis of human focal epilepsies. An inherent problem for studies on human temporal lobe epilepsy (TLE) is the lack of suitable controls. Strategies to alleviate this obstacle include the use of human post mortem samples, hippocampus from experimental animals and, in particular, the comparative analysis of surgical specimens from patients with Ammon's horn sclerosis (AHS) and with focal temporal lesions but anatomically preserved hippocampal structures. In this review we focus on selected aspects of the molecular neuropathology of TLE: (1) the potential impact of persisting calretinin-immunoreactive neurons with Cajal-Retzius cell morphology, (2) astrocytic tenascin-C induction and redistribution as potential regulator of aberrant axonal sprouting and (3) alterations of Ca2+ -mediated hippocampal signalling pathways. The diverse and complex changes described so far in human TLE specimens require a systematic interdisciplinary approach to distinguish primary, epileptogenic alterations and secondary, compensatory mechanisms in the pathogenesis of human temporal lobe epilepsies.

Heterogeneous distribution of polyamines in temporal lobe epilepsy

Polyamine contents were determined in human temporal lobe epilepsy. In the seven patients studied, stereoelectroencephalography (SEEG) located the epileptogenic focus in Ammon's horn and neuropathological findings were limited to hippocampal gliosis and sclerosis. Each polyamine exhibited a specific regional distribution. The most important variations were observed for spermidine and spermine while putrescine levels varied less. The regional variation was predominant in middle > posterior > anterior parts of the temporal lobe. Spermine contents and the spermidine/spermine (SPD/SPM) index varied especially in the middle and posterior parts of the hippocampus. Metabolic SPD/SPM index and spermidine levels were found to be drastically increased in almost all limbic parts when compared to neocortical regions. The opposite was observed for spermine. The heterogeneous distribution of polyamines was compared to abnormal electrical activities recorded by SEEG: SPD/SPM index and spermidine levels were sharply increased in seizure onset areas and high levels of spermine were detected in temporal cortex propagation areas. The presently reported heterogeneity of polyamine contents might contribute to modulate differentially the local control of excitability in human temporal epilepsy.

GABAergic neurons and GABA(A)-receptors in temporal lobe epilepsy

Mesial temporal lobe epilepsy (MTLE) is the most prevalent form of epilepsy, characterized by recurrent complex partial seizures and hippocampal sclerosis. The pathophysiology underlying this disorder remains unidentified. While a loss of benzodiazepine binding sites is a key diagnostic feature of MTLE, experimental studies have shown enhanced inhibitory transmission and increased expression of GABA(A)-receptors, suggesting that compensatory mechanisms are operative in epileptic hippocampus. In the present study, changes in the expression and cellular distribution of major GABA(A)-receptor subunits were investigated in the hippocampus of pilocarpine-treated rats during the phase of spontaneous recurrent seizures. A uniform decrease in GABA(A)-receptor subunit-immunoreactivity was observed in regions of extensive neuronal death (i.e. CA1, CA3, hilus). whereas a prominent increase occurred in the dentate gyrus (DG). Most strikingly, the increase was largest for the alpha3- and alpha5-subunits, which are expressed at very low levels in the DG of control rats, suggesting the formation of novel GABA(A)-receptor subtypes in epileptic tissue. Furthermore, an extensive loss of interneurons expressing the alpha1-subunit, representing presumptive basket cells, was seen in the DG. These changes were very similar to those reported in a novel mouse model of MTLE, based on the unilateral injection of kainic acid into the dorsal hippocampus (Bouilleret et al., 1999). This indicates that the regulation of GABA(A)-receptor expression is related to chronic recurrent seizures, and is not due to the extrahippocampal neuronal damage affecting pilocarpine-treated rats. These results allow causal relationships in the induction and maintenance of chronic recurrent seizures to be distinguished. The loss of a critical number of interneurons in the DG is a possible cause of seizure initiation, whereas the long-lasting upregulation of GABA(A)-receptors in granule cells represents a compensatory response to seizure activity.

Somatostatin acts in CA1 and CA3 to reduce hippocampal epileptiform activity

Although the peptide somatostatin (SST) has been speculated to function in temporal lobe epilepsy, its exact role is unclear, as in vivo studies have suggested both pro- and anticonvulsant properties. We have shown previously that SST has multiple inhibitory cellular actions in the CA1 region of the hippocampus, suggesting that in this region SST should have antiepileptic actions. To directly assess the effect of SST on epileptiform activity, we studied two in vitro models of epilepsy in the rat hippocampal slice preparation using extracellular and intracellular recording techniques. In one, GABA-mediated neurotransmission was inhibited by superfusion of the GABAA receptor antagonist bicuculline. In the second, we superfused Mg2+-free artificial cerebrospinal fluid to remove the Mg2+ block of the N-methyl-D-aspartate (NMDA) subtype of glutamate receptor. We show here that SST markedly reduces the intensity of evoked epileptiform afterdischarges and the frequency of spontaneous bursts in both CA1 and CA3. SST appears to act additively in the two regions to suppress the transmission of epileptiform events through the hippocampus. We further examined SST's actions in CA3 and found that SST dramatically reduced the frequency of paroxysmal depolarizing shifts (PDSs) recorded intracellularly in current clamp, as well as increasing the threshold for evoking "giant" excitatory postsynaptic currents (EPSCs), large polysynaptically mediated EPSCs that are the voltage-clamp correlate of PDSs. We also examined the actions of SST on pharmacologically isolated EPSCs generated at both mossy fiber (MF) and associational/commissural (A/C) synapses. SST appears to act specifically to reduce recurrent excitation between CA3 neurons because it depresses A/C- but not MF-evoked EPSCs. SST also increased paired-pulse facilitation of A/C EPSCs, suggesting a presynaptic site of action. Reciprocal activation of CA3 neurons through A/C fibers is critical for generation of epileptiform activity in hippocampus. Thus SST reduces feedforward excitation in rat hippocampus, acting to "brake" hyperexcitation. This is a function unique from that described for other hippocampal neuropeptides, which affect more standard neurotransmission. Our results suggest that SST receptors could be a unique, selective clinical target for treatment of limbic seizures.

Hippocampal sclerosis revisited

Studies dating back more than 150 years reported a relationship between hippocampal sclerosis and epilepsy. Retrospective studies of patients who underwent temporal lobectomy for intractable partial epilepsy found a relationship between a history of early childhood convulsions, hippocampal sclerosis, and the development of temporal lobe epilepsy. Many believe that febrile seizures lead to hippocampal damage and this in turn predisposes the patient to the development of temporal lobe epilepsy. Studies in adult rats have shown that seizures can lead to hippocampal damage and unprovoked recurrent seizures. However, many questions remain as to the relevance of early childhood seizures to hippocampal sclerosis and temporal lobe epilepsy. Human prospective epidemiologic studies have not shown a relationship between early childhood seizures and temporal lobe epilepsy. Recent MRI studies in humans suggest that a preexisting hippocampal lesion may predispose infants to experience febrile seizures, later on hippocampal sclerosis, and possibly temporal lobe epilepsy may occur. Unlike the studies in adult rats, normal immature rats with seizures have not been shown to develop hippocampal damage or unprovoked seizures in adulthood. Furthermore, animal studies reveal that preexisting brain abnormalities can predispose to hippocampal damage following seizures early in life. This paper reviews evidence for and against the view that early childhood convulsions, hippocampal sclerosis, and temporal lobe epilepsy are related, while also exploring clinical and animal studies on how seizures can lead to hippocampal damage, and how this can result in temporal lobe epilepsy. By better understanding the cause and effect relationship between early childhood seizures and hippocampal injury in normal and abnormal brains specific treatments can be developed that target the pathogenesis of epilepsy.

Selective loss of GABA neurons in area CA1 of the rat hippocampus after intraventricular kainate

The intraventricular injection of kainic acid (KA) in rats produces a loss of dentate hilar neurons and hippocampal CA3 pyramidal cells, and renders the dentate granule cells and the CA1 pyramidal cells hyperexcitable. We have used immunocytochemical detection of glutamic acid decarboxylase (GAD), a marker of gamma-aminobutyric acid (GABA) cells, as well as stereological cell counting techniques, to determine whether inhibitory cell loss was present 2 weeks after KA treatment. In area CA1, we found that the density of GAD-positive cells was reduced by KA, but only in stratum oriens and the alveus. Counts of Nissl-stained neurons were also significantly reduced in this layer. These results demonstrate a loss of GABA cells in the basal dendritic layer of the CA1 region, which may underlie the hyperexcitability of CA1 pyramidal cells following KA treatment.

Expression of GABA(A) receptor subunits in the hippocampus of the rat after kainic acid-induced seizures

The GABA(A) receptor is a ligand gated chloride channel consisting of five membrane spanning proteins for which 13 different genes have been identified in the mammalian brain. The present review summarizes recent work from our laboratory on the characterization of the immunocytochemical distribution of these GABA(A) receptor subunits in the rat brain and changes in immunoreactivity and mRNA expression after kainic acid-induced status epilepticus. A heterogeneous distribution of immunoreactive GABA(A) receptor subunits was observed. The most abundant ones were: alpha1, alpha2, alpha4, alpha5, beta2, beta3, gamma2, and delta. Alpha1, beta2, and gamma2 were about equally distributed in all subfields of the hippocampus; alpha4- and delta-subunits were preferentially found in the dentate molecular layer and in CA1; alpha2 was localized to the dentate molecular layer and CA3; alpha5 was found in the dendritic areas of CA1 to CA3; and beta1 was preferentially seen in CA2. Alpha1, beta2, gamma2 and delta were highly concentrated in interneurons. Kainic acid-induced seizures caused acute and chronic changes in the expression of mRNAs and immunoreactive proteins. Acute changes included decreases in alpha2, alpha5, beta1, beta3, gamma2 and delta mRNA levels (by about 25-50%), accompanied by increases (by about 50%) in alpha1, alpha4, and beta2 messages in granule cells (after 6-12 h). Chronic changes, characterized by losses in mRNA and immunoreactive proteins in CA1 and CA3, are undoubtedly due to seizure-related cell damage. However, compensatory expression of alpha2 and beta3 subunits, especially in CA3b/c, was observed. Furthermore, increases in mRNAs and immunoreactive proteins were seen for alpha1, alpha2 alpha4, beta1, beta2, beta3 and gamma2 in granule cells and in the molecular layer of the dentate gyrus at 7-30 days after kainic acid injection. The changes in the expression of GABA(A) receptor subunits, observed in practically all hippocampal subfields, may reflect altered GABA-ergic transmission during development of the epileptic syndrome. Increased expression of GABA(A) receptor subunits in the dendritic field of granule cells and CA3 suggest that GABA-ergic inhibition may be augmented at these levels. However, the lasting preservation of alpha1-, beta2-, and gamma2-subunits in interneurons could provide a basis for augmented inhibition of GABA-ergic interneurons, leading to net disinhibition.

Interneurones are not so dormant in temporal lobe epilepsy: a critical reappraisal of the dormant basket cell hypothesis

One axiom at the basis of epilepsy research is that there exists an imbalance between excitation and inhibition. This abnormality can be achieved by an increase of excitation on principal cells, a decreased inhibition (i.e. disinhibition) or both. This review focuses on dysfunction of inhibition, and in particular on the 'dormant basket cell hypothesis'. This hypothesis states that, (1) interneurones are functionally disconnected from excitatory afferents, resulting in hyperexcitability of principal neurones and loss of paired pulse inhibition, (2) when properly activated, interneurones can still perform their task, i.e. suppress epileptiform activity and restore paired pulse inhibition. The aim of this review is to discuss the evidence in support of the 'dormant basket cell hypothesis'. We will first discuss the rationale underlying the hypothesis and the criteria needed to validate the hypothesis. We will then show that, (1) the key experimental data offered in support of the hypothesis (Bekenstein and Lothman, 1993. Dormancy of inhibitory interneurones in a model of temporal lobe epilepsy. Science 259, 97-100; Sloviter, 1991. Permanently altered hippocampal structure, excitability, and inhibition after experimental status epilepticus in the rat: the 'dormant basket cell' hypothesis and its relevance to temporal lobe epilepsy. Hippocampus 1, 41-66) are difficult to interpret, and (2) recent recordings from interneurones in epileptic tissue argue against the hypothesis. The 'dormant basket cell hypothesis' is then discussed in the broader context of disinhibition.

Pathologic basis of the symptomatic epilepsies in childhood

The epilepsies in childhood are classified as primary (or idiopathic) and secondary (or symptomatic). The primary epilepsies account for two thirds of all childhood epilepsies and are presumed to be genetically determined. In the remaining one third of cases, a neuropathologic lesion can be identified. This paper summarizes the etiologies of the symptomatic epilepsies. They are classified according to the pathologic processes; malformative, metabolic, neoplastic and phakomatoses, hypoxic-ischemic, infectious, and of unknown pathologic process.

Apoptosis of hippocampal neurons after amygdala kindled seizures

Seizure-induced neuronal damage may involve both excitotoxic and apoptotic (programmed cell death) mechanisms. In the present study, we used an amygdala kindled seizure model to study whether apoptotic cell death occurs. To evaluate apoptosis, we counted the numbers of cells that had DNA fragments labeled at the 3' end with digoxigenin using terminal transferase (ApopTag, Oncor). Additionally, the expression of Bax and Bcl-2, two genes associated with apoptotic cell death, was also measured following kindled seizures. We found that the number of ApopTag-positive cells in the hippocampus increased 30.4% after one kindled seizure and 82.5% after 20 seizures compared to sham controls. The ApopTag-labeled cells could be mainly interneurons of the hippocampal formation, although additional studies are required. Preferential vulnerability of inhibitory interneurons is consistent with previous studies on seizure-induced cell loss. These results, coupled with our findings that the ratio of Bax/Bcl-2 expression is increased in the hippocampus by seizures, suggest that apoptosis of hippocampal interneurons may lead to dysinhibition in the hippocampus and increased seizure susceptibility.

GABA(A) receptor subunits in the rat hippocampus III: altered messenger RNA expression in kainic acid-induced epilepsy

Kainic acid-induced seizures in rats represent an established animal model for human temporal lobe epilepsy. The neuropathological sequelae include acute status epilepticus followed by neurodegeneration in the CA1 and CA3 sector of the Ammon's horn and of interneurons in the hilus of the dentate gyrus. After about three weeks spontaneous recurrent seizures become manifest. We investigated changes in messenger RNA expression of 13 GABA(A) receptor subunits in the hippocampus of rats in the initial phase (6 h, 12 h and 24 h) after acute kainic acid-induced status epilepticus and seizure-related neuronal cell damage during and after acquisition of spontaneous recurrent seizures (seven and 30 days after kainic acid injection). In the granule cell layer, initial (after 6 to 12 h) decreases in (alpha2, alpha3, alpha5, beta1, beta3, gamma2 and delta messenger RNAs (by about 25 to 50%) were accompanied by increases (by about 50%) in alpha1, alpha4, and beta2 messages. At later intervals (after seven to 30 days), expression of alpha2, alpha4, beta3 and gamma2 messenger RNAs recovered to control values, with alpha5 and delta messenger RNA still being reduced (by 15 and 40% below control levels, respectively). Concentrations of the transcripts encoding for alpha1, alpha3, beta1, beta2, became markedly enhanced (between 20 and 50% of controls). Within the pyramidal cell layers CA1 and CA3, decreases in alpha2, alpha4, alpha5, beta(1-3) and gamma2 messenger RNAs were detected after seven to 30 days, reflecting pronounced neurodegeneration in these areas. The alpha1 transcript was decreased in CA3 after 24 h and increased to control levels indicating compensatory up-regulation of this message after seven days. Messenger RNAs encoding for alpha3-, gamma1-, and gamma3-subunits were detected at rather low levels, alpha6 was not present in the hippocampus. Our data suggest a fast but transient change in the expression of messenger RNAs encoding for different subunits of the GABA(A) receptor in the granule cell layer of the dentate gyrus. This is followed by a lasting augmentation of messenger RNAs encoding different GABA(A) receptor subunits in the same cell layer indicating long-lasting GABAergic inhibition. Changes within the pyramidal cell layer are mostly determined by concomitant neurodegenerative processes.

Quantitative autoradiographic analysis of ionotropic glutamate receptor subtypes in human temporal lobe epilepsy: up-regulation in reorganized epileptogenic hippocampus

Medically intractable temporal lobe epilepsy is a common disease typically associated with hippocampal damage (sclerosis) and synaptic remodelling. These changes could include increased glutamate receptor expression, enhancing excitability and the potential for neuronal injury. We directly assessed this hypothesis using quantitative in vitro receptor autoradiography to determine the densities of glutamate-, NMDA-, quisqualate/alpha-amino-3-hydroxy-5-methyl-isoxazoleproprionic acid (AMPA)- and kainic acid-preferring binding sites in surgically removed hippocampi from patients with mesial temporal lobe epilepsy (sclerosis; MTLE) and patients with mass-associated temporal lobe epilepsy (no sclerosis; MaTLE), compared with autopsy material. Neuronal cell counts and in situ total protein densities were also obtained. In general, MaTLE and autopsy binding densities were indistinguishable. In contrast, some regions of MTLE hippocampi exhibited decreased receptor densities, with a corresponding loss of protein. In the hilus and CA1, however, ligand binding densities did not differ from the comparison groups in spite of markedly reduced protein content, consistent with increased glutamate receptor density. Kainate-preferring sites were distributed differently from the other glutamate subtypes and were uniformly decreased throughout the MTLE hippocampus, except for a unique expression within the outer dentate molecular layer. Along with increased NMDA and AMPA receptor densities in the hilus and CA1, this distinctive population of kainate receptors establishes that increased glutamate receptor expression is a feature of the remodelled MTLE hippocampus. These observations suggest that enhanced sensitivity to glutamate may be an important element in the pathophysiology of temporal lobe epilepsy.

Connections of the hippocampal formation in humans: I. The mossy fiber pathway

The hippocampal formation has been one of the most extensively studied cortical regions in rats, yet little is known about the anatomical connections of the hippocampus in primates, especially humans. With the use of an antibody against the calcium-binding protein, calbindin-D28K, in normal autopsy tissue and the neuronal tracers biocytin or biotinylated dextrans in in vitro slice preparations from tissue removed during surgery for intractable epilepsy, we examined the human hippocampal mossy fiber pathway. The injections of biocytin into the dentate granule cell layer labeled neurons in a Golgi-like manner, revealing the presence of basal dendrites on about 30% of the granule cells. The granule cell axons, the mossy fibers, initially formed a diffuse plexus of fibers in the polymorphic layer before organizing into fiber fascicles in the hilar pyramidal region. These fiber fascicles were much more prominent rostrally than caudally. Within the hilus and proximal portions of the extrahilar CA3 field, the mossy fibers ran through the pyramidal cell layer, and while near the transition to field CA2, the fibers turned superficially and crossed the pyramidal layer to run in the stratum lucidum. All of these features, seen following injections of tracer into hippocampal slices from the brains of epileptics, were confirmed by calbindin-staining of mossy fibers in normal brains. Biocytin-labeled mossy fiber axons revealed two characteristic types of enlargements: small varicosities and larger expansions. The expansions were found throughout the neuropil and were highly irregular, diaminobenzidine-dense profiles that had pleiomorphic modes of attachment to the parent axon. Electron microscopic images of these biocytin labeled expansions revealed that they were large synaptic boutons bearing asymmetric synapses. This study indicates that the human mossy fiber pathway shows some minor deviations from the rodent brain but little difference from monkeys. We argue that these changes mirror a phylogenetic growth of the CA3 pyramidal neurons (subfield CA3c) into the hilus rather than an evolutionary change of the mossy fiber pathway. This growth of subfield CA3c and the increase in mossy fibers running through the pyramidal layer (and a presumed accompanying increase in proximal basal dendritic contacts) may reflect a growing role of the projection from the dentate granule cells to subfield CA3c and from there to field CA1 in the primate hippocampus.

Comparison of NADPH diaphorase histochemistry, somatostatin immunohistochemistry, and silver impregnation in detecting structural and functional impairment in experimental status epilepticus

Nitric oxide has been postulated as a retrograde intercellular messenger for long-term potentiation, a form of synaptic plasticity that is associated with learning and memory processes. In the present study we investigated whether the loss or survival of nicotinamide adenine dinucleotide phosphate (NADPH) diaphorase-containing neurons, which are known to synthesize nitric oxide, would be an useful indicator for evaluating the structural and functional state of the rat hippocampus after status epilepticus that is induced by intraperitoneal injection of kainic acid. Besides NADPH diaphorase histochemistry, two other histological parameters were studied: the grade of cell damage evaluated from silver-impregnated sections, and the number of somatostatin-containing neurons in different hippocampal subfields. We found that the number of NADPH diaphorase-containing neurons in the hilus and granule cell layer correlated well with spatial learning and memory performance as assessed by the Morris water-maze test. The extent of cell damage in the CA1 subfield analysed in silver-impregnated sections and the number of hilar somatostatin-containing neurons also significantly correlated with latencies in the water-maze test. Furthermore, linear regression analysis revealed that the number of somatostatin-containing neurons in the hilus explains about 50% of the variation in water-maze learning. These findings emphasize that although general structural preservation is of crucial importance for the function of the hippocampus also interneurons, such as somatostatin- and NADPH diaphorase-containing neurons, may play an important role during the acquisition phase and processing of information in hippocampal circuitry. Therefore, in addition to evaluating general cell damage, analysis of the cell loss that occurs in the interneuron subpopulations will be beneficial in verifying structural and functional deficits of the hippocampus after status epilepticus.