Glutamatergic Fate of Neural Progenitor Cells of Rats with Inherited Audiogenic Epilepsy

Epilepsy is associated with aberrant neurogenesis in the hippocampus and may underlie the development of hereditary epilepsy. In the present study, we analyzed the differentiation fate of neural progenitor cells (NPC), which were isolated from the hippocampus of embryos of Krushinsky-Molodkina (KM) rats genetically prone to audiogenic epilepsy. NPCs from embryos of Wistar rats were used as the control. We found principal differences between Wistar and KM NPC in unstimulated controls: Wistar NPC culture contained both gamma-aminobutyric acid (GABA) and glutamatergic neurons; KM NPC culture was mainly represented by glutamatergic cells. The stimulation of glutamatergic differentiation of Wistar NPC resulted in a significant increase in glutamatergic cell number that was accompanied by the activation of protein kinase A. The stimulation of KM NPC led to a decrease in immature glutamatergic cell number and was associated with the activation of extracellular signal-regulated kinases 1 and 2 (ERK1/2) and protein kinase B/ glycogen synthase kinase 3 beta (Akt/GSK3β), which indicates the activation of glutamatergic cell maturation. These results suggest genetically programmed abnormalities in KM rats that determine the glutamatergic fate of NPC and contribute to the development of audiogenic epilepsy.


Introduction
Audiogenic epilepsy is a form of reflex epilepsy and its etiology is mainly hereditary, both in human and animal [1][2][3]. The development of audiogenic epilepsy, along with other epilepsy types, involves many factors, including aberrant neuronal excitability and a disturbed balance between inhibitory and excitatory synaptic transmission in the brain [4]. At the same time, it is known that epileptogenesis is associated with aberrant neurogenesis in the hippocampus [5,6].
Observations of patients with epilepsy and the results of experimental modeling of seizures indicate that epileptiform activity stimulates proliferation of neural progenitors in the subgranular layer of the dentate gyrus [7][8][9][10]. Part of these newborn cells migrates to the hilus and differentiates into ectopic excitatory glutamatergic neurons that exhibit stable hyperactivity and tend to synchronize with Cornu Ammonis area 3 (CA3) pyramid cells of the hippocampus [10,11]. On the other hand, the inhibition of aberrant neurogenesis caused by status epilepticus significantly reduced the frequency and severity of seizures [12,13].
Nowadays, genetic abnormalities in epilepsy are being actively investigated. Mutations in a variety of genes encoding voltage-dependent ion channels, receptors, synaptic proteins were identified

Evaluation of Cell Cultures and Statistical Analysis
Cell cultures were analyzed by counting the numbers of cells immunopositive for specific markers GAD65/67, VGLUT1/2, and DCX in each group. The total cell numbers in the groups were determined by counting DAPI-stained cell nuclei. Digital pictures were obtained with use of Leica DMI 6000B fluorescent microscope (Leica Microsystems GmbH, Wetzlar, Germany) and then processed by ImageJ software for the cell counting.
All data were processed statistically by the Mann-Whitney U test with use of GraphPad Prism 8 software. The results are presented as mean ± SD. Differences were regarded as significant at p < 0.05.

Analysis of Differentiation Fate
To investigate the effectiveness of glutamatergic differentiation protocol, we carried out immunofluorescent detection of vesicular glutamate transporters 1 and 2 (VGLUT1/2) and glutamate decarboxylases 65 and 67 (GAD65/67), which are markers of glutamatergic and GABAergic neurons, respectively. Percentages of VGLUT1/2-and GAD65/67-positive cells were evaluated in control (unstimulated) and neurotrophin-stimulated cultures of NPC isolated from Wistar and KM rat embryos. The obtained data revealed significant difference between control groups of KM and Wistar NPC. We showed that in culture of KM NPC, the number of VGLUT1/2-positive cells was significantly higher (Figure 1a    We also analyzed the expression of doublecortin (DCX), a marker of immature migrating neurons in the developing and adult brain [34]. Our data demonstrated that, in control KM NPC culture, the percentage of progenitor DCX-positive cells was increased in comparison with Wistar NPC culture (Figure 1a  We also analyzed the expression of doublecortin (DCX), a marker of immature migrating neurons in the developing and adult brain [34]. Our data demonstrated that, in control KM NPC culture, the percentage of progenitor DCX-positive cells was increased in comparison with Wistar NPC culture (Figure 1a (Figure 3a, Figure 4a), and about 25% of VGLUT1/2-positive cells expressed DCX (Figure 3c, Figure 4c). Simultaneously, the percentage of GAD65/67-positive cells was reduced (Figure 2b,d, Figure 4b,d). These data suggested that our differentiation protocol was successful to promote glutamatergic differentiation in wild-type NPC. However, the same stimulation of KM NPC towards glutamatergic differentiation changed neither glutamatergic (Figure 1a,c, Figure 4e) nor GABAergic cell percentages in the culture (Figure 1b,d,  Figure 4f,i), but the percentage of DCX/VGLUT1/2 double-positive cells was significantly decreased, which suggested the activation of glutamatergic maturation (Figure 1a,c, Figure 4g). percentage of GAD65/67-positive cells was reduced (Figure 2b,d, Figure 4b,d). These data suggested that our differentiation protocol was successful to promote glutamatergic differentiation in wild-type NPC. However, the same stimulation of KM NPC towards glutamatergic differentiation changed neither glutamatergic (Figure 1a,c, Figure 4e) nor GABAergic cell percentages in the culture ( Figure  1b,d, Figure 4f,i), but the percentage of DCX/VGLUT1/2 double-positive cells was significantly decreased, which suggested the activation of glutamatergic maturation (Figure 1a,c, Figure 4g).

Analysis of Cell Signaling
Then we analyzed activity of protein kinases ERK1/2, Akt, PKA, and GSK3β that participate in the regulation of neuronal differentiation.
The data showed that stimulation of glutamatergic differentiation in Wistar NPC did not change the activity of ERK1/2 ( Figure 5a) and Akt (Figure 6a) as compared with unstimulated Wistar NPC. At the same time, the activity of PKA was significantly enhanced (Figure 7a) that was accompanied with an increase in GSK3β phosphorylation at Ser9 (Figure 6c). VGLUT1/2/DCX double positive cells in KM NPC culture was decreased after stimulation (g). Axis x: experimental groups. Axis y: number of immunopositive or double immunopositive cells as % of the whole cell number calculated with use of DAPI-stained cell nuclei. The data are calculated from three independent experiments. Results are shown as mean ± SD. Significant differences from control: * p < 0.05.

Analysis of Cell Signaling
Then we analyzed activity of protein kinases ERK1/2, Akt, PKA, and GSK3β that participate in the regulation of neuronal differentiation.
The data showed that stimulation of glutamatergic differentiation in Wistar NPC did not change the activity of ERK1/2 ( Figure 5a) and Akt (Figure 6a) as compared with unstimulated Wistar NPC. At the same time, the activity of PKA was significantly enhanced (Figure 7a) that was accompanied with an increase in GSK3β phosphorylation at Ser9 (Figure 6c).  However, in stimulated KM NPC we observed significantly increased activity of ERK1/2 ( Figure 5b) and Akt (Figure 6b). Phosphorylation of GSK3β was also increased (Figure 6d), while the activity of PKA was diminished in comparison with corresponding KM control (Figure 7b). These data suggest that in Wistar, the NPC activity of GSK3β can be regulated by PKA, while in KM NPC it is mainly dependent on Akt. However, in stimulated KM NPC we observed significantly increased activity of ERK1/2 ( Figure  5b) and Akt (Figure 6b). Phosphorylation of GSK3β was also increased (Figure 6d), while the activity of PKA was diminished in comparison with corresponding KM control (Figure 7b). These data suggest that in Wistar, the NPC activity of GSK3β can be regulated by PKA, while in KM NPC it is mainly dependent on Akt.

Discussion
An imbalance between excitatory and inhibitory signals in the brain with excessive activity of excitatory glutamatergic system is one of the basic mechanisms of seizure susceptibility [35,36]. Our recent studies demonstrated increased glutamatergic transmission in the hippocampus of KM rats during the first month of postnatal development when the readiness to audiogenic seizures has not manifested yet [37]. These results indicate that the hyperactivity of glutamatergic system may be a result of aberrant glutamatergic differentiation of new-born cells in the hippocampus of KM rats. Indeed, here we revealed predominant glutamatergic differentiation of KM NPC. Comparison of NPC in control groups, where the cells were cultured in neurotrophin-free conditions, showed dramatic differences between Wistar and KM NPC behavior. While Wistar NPC culture contained mixed population of GABA and glutamatergic cells, KM NPC culture was represented mostly by glutamatergic cells. Thus, we revealed the susceptibility of KM NPC to differentiation into excitatory glutamatergic neurons along with the impaired formation of inhibitory GABAergic neurons. These results confirm that recent data were obtained with cultures of NPC isolated from hippocampus of KM rats at the early stage of postnatal development [38]. In particular, it was shown that stimulation by the retinoic acid of KM NPC, but not Wistar NPC, led to predominant differentiation into excitatory glutamate-and dopaminergic neurons [38].
In addition, we observed that control culture of KM NPC contained a large population of immature DCX-positive neurons, which mostly were glutamatergic. In previous studies, we

Discussion
An imbalance between excitatory and inhibitory signals in the brain with excessive activity of excitatory glutamatergic system is one of the basic mechanisms of seizure susceptibility [35,36]. Our recent studies demonstrated increased glutamatergic transmission in the hippocampus of KM rats during the first month of postnatal development when the readiness to audiogenic seizures has not manifested yet [37]. These results indicate that the hyperactivity of glutamatergic system may be a result of aberrant glutamatergic differentiation of new-born cells in the hippocampus of KM rats. Indeed, here we revealed predominant glutamatergic differentiation of KM NPC. Comparison of NPC in control groups, where the cells were cultured in neurotrophin-free conditions, showed dramatic differences between Wistar and KM NPC behavior. While Wistar NPC culture contained mixed population of GABA and glutamatergic cells, KM NPC culture was represented mostly by glutamatergic cells. Thus, we revealed the susceptibility of KM NPC to differentiation into excitatory glutamatergic neurons along with the impaired formation of inhibitory GABAergic neurons. These results confirm that recent data were obtained with cultures of NPC isolated from hippocampus of KM rats at the early stage of postnatal development [38]. In particular, it was shown that stimulation by the retinoic acid of KM NPC, but not Wistar NPC, led to predominant differentiation into excitatory glutamate-and dopaminergic neurons [38].
In addition, we observed that control culture of KM NPC contained a large population of immature DCX-positive neurons, which mostly were glutamatergic. In previous studies, we demonstrated an increase in VGLUT2 expression in the dentate gyrus of two-week old KM rats. However, in the dentate gyrus of one-month old KM rats expression of VGLUT2 was the same with Wistar rats [37]. Development of the dentate gyrus in rodents continues during two first weeks after birth and at the beginning is characterized by excessive formation of cells, which are then eliminated [39]. The increase in VGLUT2 expression in the dentate gyrus of two-week old KM demonstrated aberrantly activate glutamatergic transmission [37] that revealed the predisposition of newly born cells to differentiate into glutamatergic neurons. Here, we confirmed our in vivo data and demonstrated genetically determined glutamatergic differentiation fate of embryonic KM NPC in vitro. Previously, we have shown an increase in the expression of the NR2B subunit of N-methyl D-aspartate (NMDA) receptors in the hippocampus of two-week old KM pups [37]. It is known that not only neurons, but also astrocytes, both express glutamate receptors including NR2B [40] by which astrocytes synchronize Ca2+ signaling with neurons after the onset of epileptiform activity [41]. On the other hand, the astrocytes, like neurons, release glutamate and thus regulate the activity of neuronal NMDA receptors [42]. However, the inhibition of NMDA receptors [42][43][44] or disruption of astrocyte synchronization [41] reduces the severity of seizures. However, there are no data of the role of tripartite glutaminergic synapses in genetic epilepsy and this question should be address to further studies.
Then, we analyzed how the specific stimulation of glutamatergic differentiation can change NPC behavior. The stimulation of Wistar NPC led to a significant increase in glutamatergic cell number and a decrease in NPC differentiation into GABAergic cells, proving the specificity of the experimental protocol. Surprisingly, the same protocol failed to induce any significant changes in KM NPC culture, where the number of glutamatergic cells remained abnormally high.
To study the mechanisms underlying glutamatergic differentiation, we analyzed the activity of several key kinases, which regulate neuronal differentiation, such as ERK1/2, GSK3β, Akt and PKA. The first one was ERK1/2 kinase, the crucial participant of the ERK1/2 signaling cascade involved in neuronal differentiation [45] and the regulation of NMDA receptor expression [46]. The activation of ERK1/2 in the hippocampus was also demonstrated after seizure expression, stimulated by a different typed of convulsants [47,48], while the inhibition of ERK1/2 successfully prevented audiogenic seizure expression in KM rats [49]. Moreover, the hyperactivity of ERK1/2-dependent signaling cascade contributes to seizure expression [44,50]. It was shown that the activation of ERK1/2 in the hippocampus provoked spontaneous seizures in mice [44,50]. In human, high and persistence activity of ERK1/2 was revealed in the 'epileptic' cortex of patients with neocortical epilepsy [31]. The participation of Akt and PKA was also revealed in the patients with pharmacoresistant epilepsy. Recently, Valmiki and co-authors demonstrated sustained active Akt in the hippocampus of patients with temporal lobe epilepsy [29]. A role of PKA signaling was also confirmed in the development of pharmacoresistant temporal lobe epilepsy [30]. We determined that the activity of ERK1/2 was not changed after the differentiation of Wistar NPC into glutamatergic cells. At the same time, glutamatergic differentiation of KM NPC was accompanied with s significant elevation in ERK1/2 activity. Interestingly, during the early postnatal development of KM rats, the delayed development of the hippocampus was associated not only with the high activity of ERK1/2, but also with increased NR2B expression [32]. The present in vitro data showed an unchanged number of glutamatergic cells after stimulation that was accompanied with a significant decrease in immature glutamatergic cells and activated ERK1/2. Together, these data indicate the stimulation of glutamatergic maturation.
Protein kinase GSK3β also plays an important role in the regulation of neurogenesis during normal development and under pathological conditions [51,52]. It is well-known that GSK3β phosphorylates and inhibits several transcription factors, such as c-Myc, c-jun, β-catenine, which stimulate the proliferation of neural stem cells, therefore GSK3β activation induces neuronal differentiation [51]. The basic regulatory mechanisms that switch GSK3β activity in the brain include dynamical phosphorylation/dephosphorylation at Ser9 of GSK3β N-terminal domain. Thus, a number of neurotrophins stimulate Akt that, in turn, negatively regulates GSK3β by phosphorylation at Ser9 supporting stem cell survival or keeping the proliferative status of progenitor cells [53]. The same regulatory site of GSK3β is the target for PKA phosphorylation [24]. PKA-induced GSK3β inhibition is also supposed to be associated with cell survival and prevention of apoptosis during neurogenesis [54]. At the same time, it was shown that PKA activation stimulates neuronal differentiation in cultured cells [55,56]. Our data demonstrated that the stimulation of glutamatergic differentiation resulted in the increase in GSK3β phosphorylation at Ser9 in both Wistar and KM NPC. But in KM NPC we observed the activation of Akt, while in Wistar NPC, the activation of PKA was revealed. These data suggest that, in the case of Wistar NPC stimulation with neurotrophins, the activation of PKA is induced, which, in turn, phosphorylates multiple intracellular substrates including GSK3β and results in the expected glutamatergic differentiation. In KM NPC, the same stimulation resulted in the activation of ERK1/2 and Akt/GSK3β, signaling maintained aberrant glutamatergic differentiation.
Thus, our data suggest the genetically determined differentiation of NPC into glutamatergic neurons in the hippocampus of KM rats. We suppose that these defects contribute to the abnormal formation of the hippocampal glutamatergic system and, therefore, are responsible for the development of audiogenic seizure susceptibility in KM rats. Several studies have demonstrated that the PKA, Akt and ERK1/2 signaling pathways contribute to the development of human epilepsy too [29][30][31], which suggests that our findings may make some contribution to understanding the mechanisms of human epilepsy. However, it is important to note that our findings are based on 2D culture of NPC. The primary cultures of hippocampal cells, which contain both neurons and glia cells, or 3D perfused cellular models, can provide a more physiologically relevant insight into the mechanisms underlying signal transduction in hereditary reflex epilepsy.