Experimental early‐life febrile seizures cause a sustained increase in excitatory neurotransmission in newborn dentate granule cells

Abstract Prolonged febrile seizures (FS) are a risk factor for the development of hippocampal‐associated temporal lobe epilepsy. The dentate gyrus is the major gateway to the hippocampal network and one of the sites in the brain where neurogenesis continues postnatally. Previously, we found that experimental FS increase the survival rate and structural integration of newborn dentate granule cells (DGCs). In addition, mature post‐FS born DGCs express an altered receptor panel. Here, we aimed to study if these molecular and structural changes are accompanied by an altered cellular functioning. Experimental FS were induced by hyperthermia in 10‐days‐old Sprague‐Dawley rats. Proliferating progenitor cells were labeled the next day by injecting green fluorescent protein expressing retroviral particles bilaterally in the dentate gyri. Eight weeks later, spontaneous excitatory and inhibitory postsynaptic events (sEPSCs and sIPSCs, respectively) were recorded from labeled DGCs using the whole‐cell patch‐clamp technique. Experimental FS resulted in a robust decrease of the inter event interval (p < .0001) and a small decrease of the amplitude of sEPSCs (p < .001). Collectively the spontaneous excitatory charge transfer increased (p < .01). Experimental FS also slightly increased the frequency of sIPSCs (p < .05), while the amplitude of these events decreased strongly (p < .0001). The net inhibitory charge transfer remained unchanged. Experimental, early‐life FS have a long‐term effect on post‐FS born DGCs, as they display an increased spontaneous excitatory input when matured. It remains to be established if this presents a mechanism for FS‐induced epileptogenesis.


INTRODUCTION
Febrile seizures (FS) are fever-induced seizures in children till the age of years and are considered a risk factor for contracting temporal lobe epilepsy (TLE) (Berg & Shinnar, 1996;Hauser, 1994). The suggestion that FS cause or signify hippocampal injury or aggravate a preexisting hippocampal lesion, thereby contributing to the development of TLE, comes from clinical retrospective cohort and case reports, and experimental studies (Abou-Khalil et al., 1993;French et al., 1993;Hamati-Haddad & Abou-Khalil, 1998). One hypothesis is that FS initiate the development of a hyperexcitable hippocampal circuitry.
Within the hippocampal formation, the dentate gyrus (DG) is a major source of input that functions as a gatekeeper. Here the excitable input is controlled by inhibitory postsynaptic currents (sIPSCs) in dentate granule cells (DGCs). Experimental FS, a known risk factor for TLE, have shown to increase the amplitude of sIPSCs.
However, also a decrease in frequency was found resulting in an unaltered net charge transfer (Swijsen, Avila, et al., 2012). On the short term, the glutamatergic transmission and LTP are not affected in the DG by hyperthermia treatment (van Campen et al., 2018). Another feature of the DG is that it is one of three regions where neurogenesis continues during adulthood (Altman & Das, 1965;Danzer, 2019;Eriksson et al., 1998). Neurogenic activity is especially high at the ages when FS occur. Previously we have shown that experimental FS increase the number of surviving post-FS born DGCs and that these matured cells show a higher degree of dendritic complexity (Raijmakers et al., 2016). This was also confirmed in other experimental FS models (Gibbs et al., 2011;van Campen et al., 2018). Also, animal models for TLE are characterized by a seizure-induced altered morphology of DGCs, increase in mitotic activity, accelerated maturation and integration (Jessberger et al., 2007;Parent et al., 1997;Singh et al., 2013). Though, five days after seizure induction DGCs already respond to glutamatergic perforant path stimulation (Overstreet-Wadiche et al., 2006), these cells receive increased GABAergic synaptic input once they are fully integrated (Jakubs et al., 2006).
Since GABA has an initial depolarizing effect on newborn cells (Ben-Ari, 2002), this may further enhance hyperexcitability and thereby promote maturation and synaptic integration of these cells (Ge et al., 2006).
Experimental FS can induce ectopic expression of DGCs, that display increased excitatory GABAergic signaling (Koyama et al., 2012). As most post-FS born DGCs migrate normotopically, we aimed to analyze whether these normotopic newborn cells also display an altered functioning. To this end, we modeled FS by exposing 10days-old rats to a stream of heated air. Proliferating progenitor cells were labelled the next day by an intrahippocampal injection with retroviral particles expressing enhanced green fluorescent protein (eGFP). Eight weeks later, spontaneous excitatory and inhibitory postsynaptic currents (sEPSCs and sIPSCs, respectively) were measured using whole-cell patch-clamp recordings of eGFP-positive newborn DGCs. FS were induced as described earlier (Baram et al., 1997;Lemmens et al., 2005;Swijsen, Nelissen et al., 2012). Briefly, on postnatal day (P)10, pups were placed in a perspex cylinder and exposed to a stream of regulated heated air (Baram et al., 1997;Lemmens et al., 2005;Raijmakers et al., 2016;Swijsen, Avila, et al., 2012;Swijsen, Nelissen, et al., 2012). During this treatment, body temperature was measured rectally every 2.5 min and maintained for 35 min between 39.5 and 42.5 • C. The occurrence of seizures, characterized by clonic contractions of fore-and hind limbs while lying on side or back, was monitored by two observers independently. Hyperthermia-induced seizures typically lasted 8-9 min (Jansen et al., 2008;Lemmens et al., 2005). Animals not displaying behavioral seizures were excluded from the study. After hyperthermia treatment (HT), pups were cooled down to pre-exposure body temperature by rubbing them with room temperature water-soaked paper wipes and then returned to their nest. Control pups underwent the same procedure, with the exception that the stream of heated air was adjusted to keep them normotherm (NT), that is, to the body temperature that was measured at the start of the experiment (∼35 • C). On P22, the dam was removed from the nest and from P36 onward animals were housed individually.

Statistical analysis
Statistical analyses were performed using GraphPad Prism 9 software

Experimental FS cause a sustained increase in excitatory input to post-FS born DGCs
Excitatory synaptic activity was examined by recording sEPSCs in the presence of the GABA A receptor blocker GABAzine to isolate excitatory glutamatergic currents. Figure 2a shows sEPSC recordings  (Figure 2h). Also, the estimated single channel or unitary current (about 1 pA) did not differ between groups (Figure 2i). Finally, analysis of the sEPSC kinetics showed that the 10-90% rise time in HT animals was comparable to that in NT controls (Figure 2j).

FS do not increase the net charge transfer of inhibitory currents of post-FS born DGCs
Next, we recorded sIPSCs to evaluate inhibitory synaptic transmission from post-FS born DGCs. Spontaneous IPSC recordings were performed in the presence of the glutamate receptor blockers DL-APV and CNQX in order to eliminate glutamatergic spontaneous events.
Action potentials in the target neuron were blocked by adding the sodium channel blocker QX-314 to the recording pipette. Figure 3a shows sIPSC recordings from 8-weeks-old DGCs that were born

DISCUSSION
Neurogenesis in the DG has been hypothesized to play a role in the development of TLE. Several studies have reported an aberrant development and/or network integration of hippocampal cells that are born after seizures (Gibbs et al., 2011;Raijmakers et al., 2016;van Campen et al., 2018). Though the long-term effect of FS on the behavior of the hippocampal network remains largely unknown, a significant proportion of animals develops limbic-onset epilepsy (Dube et al., 2006).
Previously, we observed that eight weeks after experimental FS the number of new DGCs was increased and that their dendrites developed faster and exhibited more mushroom spines and increased dendritic complexity (Figure 1; Lemmens et al., 2005;Raijmakers et al., 2016;. This suggests that experimental FS drive an increased production and accelerated synaptic integration of DGCs. In this study, we aimed to address the question if these morphological changes are accompanied by cell electrophysiological changes.
By whole-cell patch-clamp, we measured spontaneous excitatory and inhibitory currents in post-FS born, 8-weeks-old DGCs. Compared to 8-weeks-old DGCs from control animals, those from FS animals exhibited a similar inhibitory charge transfer and an increased excitatory input.
As these new DGCs are a minor part of the total number of DGCs it is not clear to what extend their increased excitability affects the dentate gyrus output and if this is a sign of epileptogenesis. By comparison, the pilocarpine model of TLE is also characterized by prolonged excitation of DGCs (Kobayashi & Buckmaster, 2003). In that model, it precedes the onset of spontaneous recurrent seizures by days to weeks suggesting that it may contribute, but is insufficient, to cause epilepsy.
A cellular mechanism of excitatory neuronal activity is that it controls spine density (Swann et al., 2000). For instance, blocking the NMDA receptor results in an increased dendritic spine density while activation leads to a decreased density. Accordingly, in different TLE models, the excessive glutamate release during seizures result in a decreased spine density (Isokawa, 1998;Santos et al., 2011;Singh et al., 2013;Tejada et al., 2014). This may serve as a protective mechanism to counteract high DG activity during seizures. The higher dendritic complexity and increased number of mushroom spines on dendrites of post-FS born, 8-weeks-old DGCs previously reported (Raijmakers et al., 2016), corroborates with the increased excitatory input reported here. Moreover, this increased excitatory input, reflected by an increase in sEPSC frequency, is accompanied by a decreased sEPSC amplitude resulting together in a higher net charge transfer by post-FS born DGCs. Similar to the prior mentioned pilocarpine model, it is therefore tempting to speculate that these cellular changes are part of an epileptogenesis process. Alternatively, or additionally, the increased excitatory input to post-FS born DGCs may also due to alterations of input circuits.
Interestingly, we previously detected increased fiber density in the hippocampus two months after FS (Jansen et al., 2008). Also, Bender et al. (2003) revealed abnormal mossy fiber innervation of the granule cell and molecular layer three months after FS . In several animal seizure models, mossy fiber sprouting correlates with abnormal recurrent excitation of DGCs (Lynch & Sutula, 2000;Wuarin & Dudek, 2001). Yet, it remains speculative if this phenomenon of mossy fiber sprouting contributes to the observed altered excitatory input to post-FS born DGCs.
In another FS model, Kwak et al. reported immunohistochemical findings of excitatory synaptogenesis in newly generated DGCs at seven weeks after FS. These morphological changes were accompanied by an elevated excitability ratio of DG paired-pulse responses and preceded the onset of recurrent seizures at ten weeks after FS (Kwak et al., 2008). Also here, the contribution that these new DGCs may have had to alter the hippocampal network physiology remained unanswered. Interestingly, as soon as one week after pilocarpine-induced F I G U R E 3 Electrophysiological properties of sIPSCs in DGCs. Ten-days-old rats were submitted to a 35 min normothermia (NT) or hyperthermia (HT) treatment. The next day, the animals received an intrahippocampal injection with eGFP-expressing viral particles to label newborn DG cells. After 8 weeks, sIPSCs of labelled DGCs were recorded by a whole-cell patch-clamp configuration with 5 mM   The sensitivity of new DGCs to excitatory input shown here may thus contribute to hyperexcitability during the latent period following FS.
In addition to an increased excitatory state, we also found that the inhibitory charge transfer in new DGCs of FS animals was similar to that in controls. By patch-clamp recordings one week after FS, we previously reported a decreased sIPSC frequency in DGCs (Swijsen, Avila, et al., 2012). Those recordings were made in unlabeled DGCs and therefore did not permit to distinguish pre-FS, existing DGCs from post-FS generated ones. The relatively small number of new DGCs however favors the assumption that these recordings mostly came from pre-FS, existing DGCs. Moreover, our patch-clamp recordings on 1-week-old post-FS born DGCs revealed no inhibitory or excitatory synaptic activity in those cells (non-published results). In contrast, the current observations were made in post-FS generated DGCs, which show a slightly increased sIPSC frequency. As new DGCs are more plastic, this may be a mechanism to compensate the previously observed loss of inhibition in pre-FS, existing DGCs. The iPSCs of post-FS born DGCs also exhibited a decreased amplitude, possibly indicating a decreased number and/or altered receptor kinetics. The latter suggestion is further supported by the slight change in 10-90% rise time.
Again, this characteristic may be specific to new DGCs as we previously reported an increased sIPSC amplitude in unlabeled DGCs (Swijsen, Avila, et al., 2012).
It is undisputed that brain plasticity, such as the degree of neurogenesis, decreases with age. For example, in adult rats, about 50% of newborn DGCs die within 4 weeks yet matured neurons are relatively stable. In contrast, in postnatal day 6 rats almost all newborn DGCs survive up to 8 weeks and about 20% of them die after reaching maturity at 5-6 months (Cahill et al., 2017). This suggests that a substantial number of DGCs that are born at infancy survive through young adulthood.
It is known that mature DGCs regulate the recruitment of DGCs that are born in adulthood (Alvarez et al., 2016). Altering early-life neurogenesis may therefore have a long-lasting impact on the DGC network and presents a risk for pathogenesis as well as a target for modification of hippocampal (dys)functioning (Danzer, 2019).
In conclusion, these data show that experimental FS result in an increased excitatory and unaltered inhibitory input to mature, post-FS born DGCs. The consequence of this cellular change to the overall hippocampal physiology and behavior remains to be established.
Yet, this finding supports the hypothesis that neurogenesis can be a mechanism that facilitates FS-induced epileptogenesis. Conclusive evidence for this suggestion may come from experiments in which post-FS neurogenesis is ablated and followed up by detection of spontaneous recurrent seizures.

ACKNOWLEDGMENTS
We would like to thank H. van Praag for generously providing the plasmids that we used for the production of viral particles. Furthermore, we greatly appreciate the help of Petra Bex and Rosette Beenaerts for their assistance with this production. This research was financially supported by a 'Bijzonder Onderzoeksfonds' grant from Hasselt University.

CONFLICT OF INTEREST
The authors declare no conflict of interest.

DATA AVAILABILITY STATEMENT
The data that support the findings of this study are available from the corresponding author upon reasonable request.

PEER REVIEW
The peer review history for this article is available at https://publons.