Basal forebrain neurons suppress amygdala kindling via cortical but not hippocampal cholinergic projections in rats

Cholinergic Signaling, Neural Excitability, and Epilepsy

Epilepsy is a common brain disorder characterized by recurrent epileptic seizures with neuronal hyperexcitability. Apart from the classical imbalance between excitatory glutamatergic transmission and inhibitory γ-aminobutyric acidergic transmission, cumulative evidence suggest that cholinergic signaling is crucially involved in the modulation of neural excitability and epilepsy. In this review, we briefly describe the distribution of cholinergic neurons, muscarinic, and nicotinic receptors in the central nervous system and their relationship with neural excitability. Then, we summarize the findings from experimental and clinical research on the role of cholinergic signaling in epilepsy. Furthermore, we provide some perspectives on future investigation to reveal the precise role of the cholinergic system in epilepsy.

Direct Septum-Hippocampus Cholinergic Circuit Attenuates Seizure Through Driving Somatostatin Inhibition

BACKGROUND

Previous studies indicated the involvement of cholinergic neurons in seizure; however, the specific role of the medial septum (MS)-hippocampus cholinergic circuit in temporal lobe epilepsy (TLE) has not yet been completely elucidated.

METHODS

In the current study, we used magnetic resonance imaging and diffusion tensor imaging to characterize the pathological change of the MS-hippocampus circuit in 42 patients with TLE compared with 22 healthy volunteers. Using optogenetics and chemogenetics, combined with in vivo or in vitro electrophysiology and retrograde rabies virus tracing, we revealed a direct MS-hippocampus cholinergic circuit that potently attenuates seizure through driving somatostatin inhibition in animal TLE models.

RESULTS

We found that patients with TLE with hippocampal sclerosis showed a decrease of neuronal fiber connectivity of the MS-hippocampus compared with healthy people. In the mouse TLE model, MS cholinergic neurons ceased firing during hippocampal seizures. Optogenetic and chemogenetic activation of MS cholinergic neurons (but not glutamatergic or GABAergic [gamma-aminobutyric acidergic] neurons) significantly attenuated hippocampal seizures, while specific inhibition promoted hippocampal seizures. Electrophysiology combined with modified rabies virus tracing studies showed that direct (but not indirect) MS-hippocampal cholinergic projections mediated the antiseizure effect by preferentially targeting hippocampal GABAergic neurons. Furthermore, chemogenetic inhibition of hippocampal somatostatin-positive (rather than parvalbumin-positive) subtype of GABAergic neurons reversed the antiseizure effect of the MS-hippocampus cholinergic circuit, which was mimicked by activating somatostatin-positive neurons.

CONCLUSIONS

These findings underscore the notable antiseizure role of the direct cholinergic MS-hippocampus circuit in TLE through driving the downstream somatostatin effector. This may provide a better understanding of the changes of the seizure circuit and the precise spatiotemporal control of epilepsy.

Cholinergic denervation attenuates phencyclidine-induced c-fos responses in rat cortical neurons

The cortical cholinergic innervation, which is important for memory and cognition, has been implicated in schizophrenia. To experimentally analyze such a possible role of the cholinergic system, we have used the dissociative drug phencyclidine (PCP), known to produce schizophrenia-like psychosis in humans, to model aspects of schizophrenia in rats. We previously showed that induced cortical cholinergic hypofunction leads to enhanced PCP-induced locomotor activity and attenuated social interaction. After PCP, rats lacking cortical cholinergic innervation also show impaired declarative memory. To directly study the role of the basalo-cortical cholinergic projections for PCP-induced neural activation in different cortical areas, we have now monitored the rapid (30 and 60 min) effects of low doses of PCP (2 and 3mg/kg) on neural activation as reflected by transcriptional activation of c-fos in cortical areas, using quantitative in situ hybridization. We find an almost pan-cortical neural induction of c-fos mRNA with doses of PCP low enough not to alter levels of either BDNF or Nogo receptor mRNA levels. Specific unilateral lesioning of the uncrossed cholinergic projections to the cortical mantle by 192-IgG-saporin immunotoxin delivery to nc basalis (NBM) caused a striking ipsilateral decrease of the PCP-induced cortical c-fos mRNA induction, restricted to areas which had become effectively denervated. Because PCP at low doses is unlikely to directly influence cortical neurons, we suggest that it acts by activation of the cholinergic input, which in turn leads to cortical c-fos mRNA increases. Our results are compatible with a role for the cholinergic system in symptoms of schizophrenia, by showing that the basalo-cortical cholinergic projections are needed in order for PCP to have full activating effects on cortical neurons.

Transplantation of GABA-producing cells for seizure control in models of temporal lobe epilepsy

A high percentage of patients with temporal lobe epilepsy (TLE) are refractory to conventional pharmacotherapy. The progressive neurodegenerative processes associated with a lifetime of uncontrolled seizures mandate the development of alternative approaches to treat this disease. Transplantation of inhibitory cells has been suggested as a potential therapeutic strategy to achieve seizure suppression in humans with intractable TLE. Preclinical investigations over 20 years have demonstrated that multiple cell types from several sources can produce anticonvulsant, and antiepileptogenic, effects in animal models of TLE. Transplanting GABA-producing cells, in particular, has been shown to reduce seizures in several well-established models. This review addresses experimentation using different sources of transplantable GABAergic cells, highlighting progress with fetal tissue, neural cell lines, and stem cells. Regardless of the source of the GABAergic cells used in seizure studies, common challenges have emerged. Several variables influence the anticonvulsant potential of GABA-producing cells. For example, tissue availability, graft survival, immunogenicity, tumorigenicity, and varying levels of cell migration, differentiation, and integration into functional circuits and the microenvironment provided by sclerotic tissue all contribute to the efficacy of transplanted cells. The challenge of understanding how all of these variables work in concert, in a disease process that has no well-established etiology, suggests that there is still much basic research to be done before rational cell-based therapies can be developed for TLE.

Septal GABAergic neurons are selectively vulnerable to pilocarpine-induced status epilepticus and chronic spontaneous seizures

The septal region of the basal forebrain plays a critical role modulating hippocampal excitability and functional states. Septal circuits may also play a role in controlling abnormal hippocampal hyperexcitability in epilepsy. Both lateral and medial septal neurons are targets of hippocampal axons. Since the hippocampus is an important epileptogenic area in temporal lobe epilepsy, we hypothesize that excessive excitatory output will promote sustained neurodegeneration of septal region neurons. Pilocarpine-induced status epilepticus (SE) was chosen as a model to generate chronic epileptic animals. To determine whether septal neuronal populations are affected by hippocampal seizures, immunohistochemical assays were performed in brain sections obtained from age-matched control, latent period (7 days post-SE) and chronically epileptic (more than one month post-SE survival) rats. An anti-NeuN (neuronal nuclei) antibody was used to study total neuronal numbers. Anti-ChAT (choline acetyltransferase), anti-GAD (glutamic acid decarboxylase) isoenzymes (65 and 67), and anti-glutamate antibodies were used to reveal cholinergic, GABAergic and glutamatergic neurons, respectively. Our results revealed a significant atrophy of medial and lateral septal areas in all chronically epileptic rats. Overall neuronal density in the septum (medial and lateral septum), assessed by NeuN immunoreactivity, was significantly reduced by approximately 40% in chronically epileptic rats. The lessening of neuronal numbers in both regions was mainly due to the loss of GABAergic neurons (80-97% reduction in medial and lateral septum). In contrast, populations of cholinergic and glutamatergic neurons were spared. Overall, these data indicate that septal GABAergic neurons are selectively vulnerable to hippocampal hyperexcitability, and suggest that the processing of information in septohippocampal networks may be altered in chronic epilepsy.

Septo-hippocampal networks in chronically epileptic rats: potential antiepileptic effects of theta rhythm generation

A series of experiments was carried out testing the hypothesis that the septal region decreases the hippocampal susceptibility to hyperexcitability states through theta rhythm generation. Medial septal neurons were simultaneously recorded with hippocampal field potentials to investigate the septo-hippocampal function in the pilocarpine model of chronic epilepsy. The theta rhythm from chronically epileptic rats had lower amplitude (20% less) and higher frequency than controls (from 3.38 to 4.25 Hz), suggesting that both generator and pacemaker structures are affected during the epileptic process. At the cellular level, the group of rhythmically bursting firing medial septal neurons, in the epileptic animals, significantly and chronically increased their firing rates in relation to controls (from 13.86 to 29.14 spikes/s). Peristimulus histograms performed around hippocampal sharp waves showed that all high-frequency firing neurons, including rhythmically bursting neurons and most slow firing neurons, decreased firing rates immediately after hippocampal epileptic discharges. Thus inhibitory hippocampo-septal influences prevail during hippocampal epileptic discharges. The occurrence of epileptic discharges was reduced 86-97% of the number observed during spontaneous theta and theta induced by sensory (tail pinch) or chemical stimulation (carbachol), suggesting that the presence of the theta state regardless of how it was produced was responsible for the reduction in epileptic discharge frequency. The understanding of the theta rhythm "anti-epileptic" effect at the cellular and molecular levels may result in novel therapeutic approaches dedicated to protect the brain against abnormal excitability states.

Reorganization of CA3 area of the mouse hippocampus after pilocarpine induced temporal lobe epilepsy with special reference to the CA3-septum pathway

We showed that when CA3 pyramidal neurons in the caudal 80% of the dorsal hippocampus had almost disappeared completely, the efferent pathway of CA3 was rarely detectable. We used the mouse pilocarpine model of temporal lobe epilepsy (TLE), and injected iontophoretically the anterograde tracer phaseolus vulgaris leucoagglutinin (PHA-L) into gliotic CA3, medial septum and the nucleus of diagonal band of Broca, median raphe, and lateral supramammillary nuclei, or the retrograde tracer cholera toxin B subunit (CTB) into gliotic CA3 area of hippocampus. In the afferent pathway, the number of neurons projecting to CA3 from medial septum and the nucleus of diagonal band of Broca, median raphe, and lateral supramammillary nuclei increased significantly. In the hippocampus, where CA3 pyramidal neurons were partially lost, calbindin, calretinin, parvalbumin immunopositive back-projection neurons from CA1-CA3 area were observed. Sprouting of Schaffer collaterals with increased number of large boutons in both sides of CA1 area, particularly in the stratum pyramidale, was found. When CA3 pyramidal neurons in caudal 80% of the dorsal hippocampus have almost disappeared completely, surviving CA3 neurons in the rostral 20% of the dorsal hippocampus may play an important role in transmitting hyperactivity of granule cells to surviving CA1 neurons or to dorsal part of the lateral septum. We concluded that reorganization of CA3 area with its downstream or upstream nuclei may be involved in the occurrence of epilepsy.

Septal networks: relevance to theta rhythm, epilepsy and Alzheimer's disease

Information processing and storing by brain networks requires a highly coordinated operation of multiple neuronal groups. The function of septal neurons is to modulate the activity of archicortical (e.g. hippocampal) and neocortical circuits. This modulation is necessary for the development and normal occurrence of rhythmical cortical activities that control the processing of sensory information and memory functions. Damage or degeneration of septal neurons results in abnormal information processing in cortical circuits and consequent brain dysfunction. Septal neurons not only provide the optimal levels of excitatory background to cortical structures, but they may also inhibit the occurrence of abnormal excitability states.

Activation of nicotinic acetylcholine receptors increases the frequency of spontaneous GABAergic IPSCs in rat basolateral amygdala neurons

The basolateral amygdala (BLA) is a critical component of the amygdaloid circuit, which is thought to be involved in fear conditioned responses. Using whole cell patch-clamp recording, we found that activation of nicotinic acetylcholine receptors (nAChRs) leads to an action potential-dependent increase in the frequency of spontaneous GABAergic currents in principal neurons in the BLA. These spontaneous GABAergic currents were abolished by a low-Ca2+/high-Mg2+ bathing solution, suggesting that they are spontaneous inhibitory postsynaptic currents (sIPSCs). Blockade of ionotropic glutamate receptors did not prevent this increased frequency of sIPSCs nor did blockade of alpha7 nAChRs. Among the nAChR agonists tested, cystisine was more effective at increasing the frequency of the sIPSCs than nicotine or 1,1-dimethyl-4-phenyl piperazinium iodide, consistent with a major contribution of beta4 nAChR subunits. The nicotinic antagonist, dihydro-beta-erythroidine, was less effective than d-tubocurarine in blocking the increased sIPSC frequency induced by ACh, suggesting that alpha4-containing nAChR subunits do not play a major role in the ACh-induced increased sIPSC frequency. Although alpha2/3/4/7 and beta2/4 nAChR subunits were found in the BLA by RT-PCR, the agonist and antagonist profiles suggest that the ACh-induced increase in sIPSC frequency involves predominantly alpha3beta4-containing nAChR subunits. Consistent with this, alpha-conotoxin-AuIB, a nAChR antagonist selective for the alpha3beta4 subunit combination, inhibited the ACh-induced increase in the frequency of sIPSCs. The observations suggest that nicotinic activation increases the frequency of sIPSCs in the BLA by acting mainly on alpha3beta4-containing nicotinic receptors on GABAergic neurons and may play an important role in the modulation of synaptic transmission in the amygdala.

Forebrain acetylcholine regulates adult hippocampal neurogenesis and learning

Hippocampus-mediated learning enhances neurogenesis in the adult dentate gyrus (DG), and this process has been suggested to be involved in memory formation. The hippocampus receives abundant cholinergic innervation and acetylcholine (ACh) plays an important role in learning and Alzheimer's disease (AD) pathophysiology. Here, we show that a selective neurotoxic lesion of forebrain cholinergic input with 192 IgG-saporin reduces DG neurogenesis with a concurrent impairment in spatial memory. Conversely, systemic administration of the cholinergic agonist physostigmine increases DG neurogenesis. We find that changes of forebrain ACh levels primarily influence the proliferation and/or the short-term survival rather than the long-term survival or differentiation of the new neurons. We further demonstrate that these newly born cells express the muscarinic receptor subtypes M1 and M4. Our data provide evidence that forebrain ACh promotes neurogenesis, and suggest that the impaired cholinergic function in AD may in part contribute to deficits in learning and memory through reductions in the formation of new hippocampal neurons.

The epsilon theory: a novel synthesis of the underlying molecular and electrophysiological mechanisms of primary generalized epilepsy and the possible mechanism of action of valproate

Primary generalized epilepsy may be the result of maldevelopment of central nervous system and each seizure may be the consequence of a neuronal maladaptation to an unknown stimulus using the paleospinothalamical tract due to an overexpression of brain-derived neurotrophic factor and neurotrophin-3. The subsequent protein kinase C epsilon (PKC-epsilon) activation and intracellular Ca(2+) release causes a nociceptive hypersensitization and an increased cortical hyperexcitability because of increased frequency of synchronous Ca(2+) oscillations, cortical maldevelopment at the level of synapses and an attenuation of GABA(A) receptor mediated responses in reticular thalamic nucleus. Valproate may exert its antiepileptic effect as a PKC-epsilon inhibitor, and using with a PKC-epsilon activator that cannot pass blood brain barrier, its side effects may become avoidable.

Evaluation of cholinergic markers in Alzheimer's disease and in a model of cholinergic deficit

Cognitive deficits in neuropsychiatric disorders, such as Alzheimer's disease (AD), have been closely related to cholinergic deficits. We have compared different markers of cholinergic function to assess the best biomarker of cognitive deficits associated to cholinergic hypoactivity. In post-mortem frontal cortex from AD patients, acetylcholine (ACh) levels, cholinacetyltransferase (ChAT) and acetylcholinesterase (AChE) activity were all reduced compared to controls. Both ChAT and AChE activity showed a significant correlation with cognitive deficits. In the frontal cortex of rats with a selective cholinergic lesion, all cholinergic parameters measured (ACh levels, ChAT and AChE activities, "in vitro" and "in vivo" basal ACh release) were significantly reduced. AChE activity was associated to ChAT activity, and even more, to "in vivo" and "in vitro" basal ACh release. Quantification of AChE activity is performed by an easy and cheap method and therefore, these results suggest that determination of AChE activity may be used as an effective first step method to evaluate cholinergic deficits.

Kindling and status epilepticus models of epilepsy: rewiring the brain

This review focuses on the remodeling of brain circuitry associated with epilepsy, particularly in excitatory glutamate and inhibitory GABA systems, including alterations in synaptic efficacy, growth of new connections, and loss of existing connections. From recent studies on the kindling and status epilepticus models, which have been used most extensively to investigate temporal lobe epilepsy, it is now clear that the brain reorganizes itself in response to excess neural activation, such as seizure activity. The contributing factors to this reorganization include activation of glutamate receptors, second messengers, immediate early genes, transcription factors, neurotrophic factors, axon guidance molecules, protein synthesis, neurogenesis, and synaptogenesis. Some of the resulting changes may, in turn, contribute to the permanent alterations in seizure susceptibility. There is increasing evidence that neurogenesis and synaptogenesis can appear not only in the mossy fiber pathway in the hippocampus but also in other limbic structures. Neuronal loss, induced by prolonged seizure activity, may also contribute to circuit restructuring, particularly in the status epilepticus model. However, it is unlikely that any one structure, plastic system, neurotrophin, or downstream effector pathway is uniquely critical for epileptogenesis. The sensitivity of neural systems to the modulation of inhibition makes a disinhibition hypothesis compelling for both the triggering stage of the epileptic response and the long-term changes that promote the epileptic state. Loss of selective types of interneurons, alteration of GABA receptor configuration, and/or decrease in dendritic inhibition could contribute to the development of spontaneous seizures.

Loss of cortical acetylcholine enhances amphetamine-induced locomotor activity

Cholinergic disturbances have been implicated in schizophrenia. In a recent study we found that intracerebroventricular (i.c.v.) delivery of the immunotoxin 192 IgG-saporin, that effectively destroys cholinergic projections from the basal forebrain to hippocampus and cortex cerebri, leads to a marked facilitation of amphetamine-induced locomotor activity in adult rats. The aim of the present experiments was to evaluate the contribution of the septohippocampal versus the basalocortical cholinergic projections for the amphetamine hyper-response seen previously in i.c.v. 192 IgG-saporin injected rats. Since i.c.v. delivery of 192 IgG-saporin also destroys a population of Purkinje neurons in cerebellum, this cell loss needs to be taken into consideration as well. Cortex cerebri and hippocampus were selectively cholinergically denervated by intraparenchymal injections of 192 IgG-saporin into nucleus basalis magnocellularis and the medial septum/diagonal band of Broca, respectively. Selective loss of Purkinje cells in cerebellum was achieved by i.c.v. delivery of OX7 saporin. Possible effects of these three lesions on spontaneous and amphetamine-induced locomotor activity were assessed in locomotor activity cages. We find that selective cholinergic denervation of cortex cerebri, but not denervation of hippocampus or damage to cerebellum can elicit dopaminergic hyper-reactivity similar to that seen in previous i.c.v. 192 IgG-saporin experiments. Our data are compatible with the hypothesis that disturbances of cholinergic neurotransmission in cortex cerebri may be causally involved in forms of schizophrenia.

Septal cholinergic neurons suppress seizure development in hippocampal kindling in rats: comparison with noradrenergic neurons

Widespread lesions of forebrain cholinergic or noradrenergic projections by intraventricular administration of 192 IgG-saporin or 6-hydroxydopamine, respectively, accelerate kindling epileptogenesis. Here we demonstrate both quantitative and qualitative differences between the two lesions in their effects on hippocampal kindling in rats. Epileptogenesis was significantly faster after noradrenergic as compared to cholinergic denervation, and when both lesions were combined, kindling development resembled that in animals with 6-hydroxydopamine lesion alone. Furthermore, whereas the 192 IgG-saporin lesion promoted the development only of the early stages of kindling, administration of 6-hydroxydopamine or both neurotoxins accelerated the late stages also. To investigate the contribution of different subparts of the basal forebrain cholinergic system to its seizure-suppressant action in hippocampal kindling, 192 IgG-saporin was injected into medial septum/vertical limb of the diagonal band of Broca or nucleus basalis magnocellularis, leading to selective hippocampal or cortical cholinergic deafferentation, respectively. The denervation of the hippocampus facilitated kindling similar to the extensive lesion caused by intraventricular 192 IgG-saporin, whereas the cortical lesion had no effect. These results indicate that although both noradrenergic and cholinergic projections to the forebrain exert powerful inhibitory effects on hippocampal kindling epileptogenesis, the action of the cholinergic system is less pronounced and occurs specifically prior to seizure generalization. In contrast, noradrenergic neurons inhibit the development of both focal and generalized seizures. The septo-hippocampal neurons are responsible for the antiepileptogenic effect of the cholinergic system in hippocampal kindling, whereas the cortical projection is not significantly involved. Conversely, we have previously shown [Ferencz I. et al. (2000) Eur. J. Neurosci., 12, 2107-2116] that seizure-suppression in amygdala kindling is exerted through the cortical and not the hippocampal cholinergic projection. This shows that, depending on the location of the primary epileptic focus, i.e. the site of stimulation, basal forebrain cholinergic neurons operate through different subsystems to counteract seizure development in kindling.