Correspondence: Ryuta Koyama, PhD, Laboratory of Chemical Pharmacology, Graduate School of Pharmaceutical Sciences, The University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo 113-0033, Japan. Email: firstname.lastname@example.org
Epilepsy is a nervous system disorder characterized by recurrent seizures. Among several types of epilepsy, which accounts for a significant portion of the disease worldwide, temporal lobe epilepsy (TLE) is one of the most common types of intractable epilepsy in adulthood. It has been suggested that complex febrile seizures in early life are associated with the development of TLE later in life; however, cellular and molecular links between febrile seizures and TLE remain unclear because of the lack of an appropriate in vitro system. Using rat hippocampal slice cultures, in which many features of native organotypic organization are retained, we found that the dentate granule cells exhibit aberrant migration in the dentate hilus via enhanced excitatory GABAA receptor (GABAA-R) signaling, which results in granule cell ectopia that persists into adulthood. We further found that the granule cell ectopia is associated with spontaneous limbic seizures in adulthood. Importantly, both of these phenomena were prevented by inhibiting Na+K+2Cl− co-transporter (NKCC1) which mediates the excitatory action of GABA.
The hippocampi of individuals with mesial temporal lobe epilepsy (TLE) and corresponding animal models are accompanied by several pathological changes, such as the dispersion of dentate granule cells,[1-3] the emergence of ectopic granule cells,[4-7] the sprouting of hippocampal mossy fibers,[8-10] and hippocampal sclerosis, including selective neuronal loss and reactive gliosis in Ammon's horn. Each of these features has been suggested to play a role in the initiation and propagation of epileptic activity in the hippocampus. These pathological changes may be triggered by early-life seizures considering that retrospective studies have suggested a correlation between a history of early-life seizures and hippocampal sclerosis;[12-16] however, direct evidence is lacking.
Febrile seizures, which are associated with fevers (typically greater than 38.5°C), are the most common convulsive events in infancy and childhood between 6 months and 5 years of age with a prevalence of 2–14% of the population. Although febrile seizures are benign in most instances, 30–40% of them are “complex”,[18, 19] with a prolonged seizure duration of >15 min, and are subsequently associated with 30–70% of the cases of adult TLE.[12, 20] Importantly, evidence of hippocampal injury has been reported after febrile status epilepticus, particularly when febrile seizures were longer than 90 min. However, cellular and molecular approaches are necessary to directly investigate epileptogenic changes in neural circuits; these approaches cannot be adequately applied to resected and often fixed human tissues. For this purpose, an organotypic slice culture system that retains the characteristic anatomic organization of the tissue of origin suits well to these requirements. Further, in the slice cultures derived from neonatal brain tissues, several developmental changes of neural circuits take place, including neuronal migration, axonal and dendritic growth, and synaptogenesis.
In a recent study, we utilized organotypic slice cultures that were prepared from rat pups which experienced experimental febrile seizures, to investigate the mechanisms underlying the emergence of ectopic granule cells, because the ectopic granule cells have been suggested to be abnormally incorporated into excitatory hippocampal networks and may be epileptogenic (the morphological and functional properties of ectopic granule cells were excellently reviewed in Scharfman et al., Pierce et al. and Scharfman and Pierce).[22-24] The slice culture system allowed us to perform time-lapse imaging of the migrating granule cells, revealing that neonatally generated granule cells exhibit aberrant migration after febrile seizures, which results in granule cell ectopia. We further determined that the aberrant migration is mediated by the excitatory action of GABA. In this article, I will introduce our study mainly focusing on the use of hippocampal slice cultures.
Experimental Febrile Seizures Induce Ectopic Granule Cells in Vivo
First, we examined whether complex febrile seizures affect the localization of neonatally generated granule cells using a rat model of febrile seizures. Experimental febrile seizures were induced by placing rats at post natal day 11 (P11) under hyperthermic conditions. To examine the localization of neonatally generated granule cells, P5 rats were injected with the S-phase marker 5-bromo-2′-deoxyuridine (BrdU), and the localization of BrdU-labelled granule cells were examined at P60. Immunohistochemical analysis revealed that the number of BrdU-labelled ectopic granule cells which failed to migrate into the granule cell layer and remained in the dentate hilus was significantly higher in the rats that experienced febrile seizures compared to control rats. In the same experimental paradigm, except that a retrovirus that encodes membrane-targeted green fluorescent protein (GFP) instead of BrdU was injected into P5 rats, we found ectopic granule cells which had bipolar dendrites that extended into the hilus and axons that projected to the granule cell layer, as well as into the CA3 region in seizure animals at P60. These results suggested that febrile seizures attenuated the proper migration of neonatally generated granule cells, inducing granule cell ectopia that persists into adulthood.
GABAA-R Signaling Regulates Granule Cell Migration and Localization
The neurotransmitter GABA is crucial in neuronal migration during cortical development.[26, 27] To examine whether GABAA receptor (GABAA-R) signaling is involved in granule cell ectopia, we treated rat pups with either the GABAA-R antagonist picrotoxin or the positive modulator of GABAA-R phenobarbital, finding that picrotoxin inhibited febrile seizure-induced granule cell ectopia, whereas phenobarbital accelerated the cell ectopia. These results suggested that GABAA-R signaling regulates granule cell migration in vivo.
To determine the specificity of GABAA-R signaling in regulating granule cell migration, we took advantage of the slice culture system in which pharmacological experiments can be easily performed. Hippocampal slices were obtained from P6 rats that received a BrdU injection at P5 to label neonatally generated granule cells. By chronically applying several agonists or antagonists for the receptors of neurotransmitters for 5 days in vitro, we found that the GABAA-R agonist muscimol retarded, and the GABAA-R antagonist bicuculline facilitated, granule cell migration, whereas glutamatergic receptor signaling was probably not involved.
Another advantage of the slice culture system is that time-lapse imaging of the neuronal maturation is available under a proper environment in which CO2 concentration and temperature are well-regulated. Direct time-lapse imaging for radially migrating granule cells was lacking, even though it was reported that granule cell progenitors are associated with radial glia in the dentate gyrus.[28, 29] To visualize granule cell migration and further determine the effects of neurotransmitters on the migrating granule cells, we developed a slice coculture system in which we replaced the hilar region of the hippocampal slice from wild-type rats with the hilar graft slices prepared from transgenic rats expressing GFP (GFP+ transgenic rats) (Fig. 1A). A 24-h time-lapse analysis revealed that GFP+ granule cells migrated radially to the granule cell layer (Fig. 1B). Using this slice coculture system, we could also examine the functional properties of migrating granule cells by directly recording electrophysiological properties from GFP+ migrating granule cells, finding that granule cells receive excitatory GABAergic but not glutamatergic inputs during migration.
Aberrant Migration of Granule Cells after Febrile Seizures: Findings from a Novel Coculture System
The above results indicated the possibility that enhanced GABAA-R signaling induced aberrant migration of granule cells after febrile seizures. This hypothesis led us to examine mainly two possible mechanisms that take place after experiencing febrile seizures: (i) the increased GABA amount in the environment (the hilus) where neonatally generated granule cells migrate; and (ii) the increased GABAA-R response of migrating granule cells to GABA. We examined the first possibility by immunohistochemistry, finding that febrile seizures did not significantly affect the expression of glutamate decarboxylase (GAD)-67 or GABA in the dentate gyrus. To examine the second possibility, we next designed a slice coculture system in which we prepared GFP+ hilar grafts and wild-type host slices from P12 rats in the normothermic control groups and the hyperthermic groups that experienced febrile seizures at P11 (Fig. 1C). A time-lapse analysis revealed that approximately half of the GFP+ cells (46%) from the hyperthermic grafts migrated in the opposite direction in host normothermic slices (hyperthermic to normothermic cocultures, Fig. 1Ciii), a phenomenon prevented by administering the GABAA-R blocker bicuculline. In contrast, most of the cells from the normothermic to normothermic (93%) (Fig. 1Ci) and normothermic to hyperthermic (92%) cocultures (Fig. 1Cii) migrated correctly to the granule cell layer. These results indicated that functional changes that mediate enhanced GABAA-R signaling were induced by febrile seizures in migrating granule cells.
To determine the febrile seizure-induced changes in a single granule cell level, we isolated the hilar explants from P12 rats in either the normothermic or hyperthermic group (24 h after the induction of febrile seizures). In the explant culture of the hilus, we found a large number of granule cells with a polarized morphology typical of migrating neurons around the explants. Immunocytochemical and immunoblot analyses in the explant culture system revealed that the surface expression of GABAA-R β subunits was upregulated in migrating granule cells from the hyperthermic group (Fig. 1D). Using this explant culture system, we found that pharmacological activation of GABAA-R caused a reversal in the direction of the migration of the migrating hyperthermic cells but not the migrating normothermic cells, suggesting an increased sensitivity of hyperthermic granule cells to GABA. The excitatory action of GABA on immature neurons is mediated by the accumulation of Cl− through the Na+K+2Cl− co-transporter (NKCC1). In agreement with this, GABA-mediated attenuation of the granule cell migration in the explant cultures was prevented by either applying the NKCC1 blocker bumetanide, a widely used loop diuretic, or short hairpin RNA (shRNA)-mediated knock down of NKCC1 in migrating granule cells.
NKCC1 Inhibition Prevents Granule Cell Ectopia and Development of Epilepsy
Finally, we investigated the link between ectopic granule cells and the future development of epilepsy. We found an increased susceptibility to pilocarpine-induced limbic seizures in adult rats that had experienced febrile seizures at P11. More importantly, 8/16 adult rats that experienced febrile seizures exhibited spontaneous limbic seizures with their frequencies positively correlated with the number of ectopic granule cells.
Because a series of in vitro experiments in our study suggested that the function of NKCC1 underlies the excitatory GABAA-R signaling-mediated granule cell ectopia, we injected bumetanide daily for a week after inducing experimental febrile seizures at P11, finding that granule cell ectopia, susceptibility to limbic seizures and the development of epilepsy in adulthood are all prevented.
The impact of complex febrile seizures in early life on neural circuit formation and its relationship to future development of epilepsy remains unclear. In our recent work, we cocultured the hippocampal slices from control and seizure animals to visualize what is going on in the brain during epileptogenesis. Even though it must be noted that the brain slice culture system includes reorganization of some neural circuits which are not observed in vivo, it still offers the investigator the opportunity to examine the cellular and molecular mechanisms underlying epileptogenesis-related changes in the neural circuits. With these properties of the slice culture system, in addition to a relatively simpler experimental manipulation compared to that with in vivo, the use of organotypic slice cultures will thus contribute to the discovery of novel therapeutic targets and strategies for preventing the emergence of epileptogenic foci.
I thank Dr. Norio Matsuki, Dr. Yuji Ikegaya, and the members of Laboratory of Chemical Pharmacology (Yaku-Saku Lab) for supporting the projects on experimental febrile seizures. This work was supported by a Grant-in-Aid for Science Research on Young Scientists (B) (No. 19790048) and the Research Foundation for Pharmaceutical Sciences.