Early Life Febrile Seizures Impair Hippocampal Synaptic Plasticity in Young Rats

Abstract

Febrile seizures (FSs) in early life are significant risk factors of neurological disorders and cognitive impairment in later life. However, existing data about the impact of FSs on the developing brain are conflicting. We aimed to investigate morphological and functional changes in the hippocampus of young rats exposed to hyperthermia-induced seizures at postnatal day 10. We found that FSs led to a slight morphological disturbance. The cell numbers decreased by 10% in the CA1 and hilus but did not reduce in the CA3 or dentate gyrus areas. In contrast, functional impairments were robust. Long-term potentiation (LTP) in CA3-CA1 synapses was strongly reduced, which we attribute to the insufficient activity of N-methyl-D-aspartate receptors (NMDARs). Using whole-cell recordings, we found higher desensitization of NMDAR currents in the FS group. Since the desensitization of NMDARs depends on subunit composition, we analyzed NMDAR current decays and gene expression of subunits, which revealed no differences between control and FS rats. We suggest that an increased desensitization is due to insufficient activation of the glycine site of NMDARs, as the application of D-serine, the glycine site agonist, allows the restoration of LTP to a control value. Our results reveal a new molecular mechanism of FS impact on the developing brain.

Keywords: NMDA receptor; febrile seizures; hippocampus; hyperthermia; long-term potentiation.

Figure 1

Nissl staining of the neurons in control (CTRL) and FS rats (FS) in the hippocampal areas CA1, CA3, hilus and dentate gyrus (DG). Group data of the counted Nissl stained neurons per 100 μm of the cellular layer. The rhombuses show individual values per brain. The bars indicate average values, and error bars show standard errors of the means. * p < 0.05, ** p < 0.01, t-test.

Figure 2

Stimulation–response relationships for fEPSP amplitudes (a) and presynaptic FV amplitudes (b) recorded from the hippocampal CA1 area in control (CTRL) and FS rats. Each point represents the mean ± SEM. (c) Representative examples of I/O relationships between the fEPSP and FV amplitudes in hippocampal slices of control and FS rats. (d) The maximal I/O slopes were smaller in FS rats than in control animals. Each rhombus represents a value obtained in individual brain slice. One to three slices were used per animal. * p < 0.05, ** p < 0.01—the difference between control and FS groups.

Figure 3

Short-term synaptic plasticity did not change in rat hippocampal slices after FS. (a) Representative examples of pair-pulse responses from the hippocampal CA1 area in control (CTRL) and post-FS rats (FS), interpulse intervals of 30 ms and 80 ms. (b) Diagram showing the paired-pulse facilitation in hippocampal slices across different inter-stimuli. Each point represents the mean ± SEM.

Figure 4

Long-term synaptic potentiation (LTP) was weakened in the CA1 of juvenile rats’ hippocampus after early life FSs. (a) Schema showing the positions of electrodes in the hippocampus. (b) Representative examples of fEPSP before induction (1) and 60 min after HFS or TBS (2). (c) Diagram showing changes in the value of the normalized slope of fEPSP in control (CTRL) and experimental (FS) animals after theta-burst stimulation (TBS) or after high-frequency stimulation (HFS) (stimulation was carried out at time 0). (d) Diagram illustrates the differences in LTP between control (CTRL) and experimental (FS) animals with different stimulation types. All data in this and the following figures are presented as a mean ± standard error of the mean. * p < 0.05—the significant difference with the control group (Tukey’s post hoc test).

Figure 5

LTP induction in the CA1 hippocampus of juvenile animals after early life FS was NMDAR-dependent. The normalized fEPSP slope in the control and experimental groups in the presence of the NMDAR blocker MK-801 (10 μM), before and after TBS (a) or HFS (b). (c,d) Diagrams illustrating the magnitude of plasticity in the control and experimental groups in the presence of MK-801, after TBS (c) or HFS (d). Note that no synaptic plasticity was detected in the presence of MK-801 in any group. Two-way ANOVA following Tukey post hoc tests were used. ** p < 0.01, *** p < 0.001.

Figure 6

GluN2B-containing NMDARs were involved in the induction of LTP in both control and post-FS animals. The relative fEPSP slope in the control and experimental groups in the presence of ifenprodil (Ifen, 3 μM), a selective GluN2B-containing NMDAR antagonist, before and after TBS (a) or HFS (b). Note that no synaptic plasticity was detected in the presence of ifenprodil in FS groups. (c,d) Diagrams illustrating the magnitude of plasticity in the control and post-FS groups in the presence of ifenprodil, after TBS (c) or HFS (d). Note that no synaptic plasticity was detected in the presence of ifenprodil in FS groups. Two-way ANOVA following Tukey post hoc tests were used. * p < 0.05.

Figure 7

The relative expression of the Grin1, Grin2a, Grin2b, Gria1, Gria2 genes, and Grin2a/Grin2b ratio in the dorsal hippocampal area after febrile seizures. CTRL—control group; FS—experimental group. One-way ANOVA followed by Tukey’s post hoc multiple comparison tests. Data are presented as a mean with a standard error of the mean. Each dot (circle, rhombus, square) represents a value obtained in individual animal.

Figure 8

The properties of NMDAR-mediated currents evoked by TBS were disturbed in FS rats. (a) The representative voltage-clamp recordings of NMDAR-mediated currents, induced by TBS in CA1 neurons in control (black) and FS rats (red). (b) The average amplitudes of the individual peaks of current within the responses, normalized to the amplitude of the first response. Mixed-model ANOVA revealed a significant interaction between the factors (FS × stimulus number). Asterisks indicate the significant difference between the values of control and FS rats for the same stimulus number according to Dunnett’s post hoc test. (c) The representative recordings of the 90–10% decays of NDMAR-mediated currents evoked during the TBS. The currents were fitted with the biexponential function with the time constants of ${\tau }_{fast}$ = 65 ms and ${\tau }_{slow}$ = 300 ms. (d) The relative contribution of the fast and slow components of the response during the TBS. Mixed-model ANOVA did not reveal a difference between the control and FS rats. Asterisks indicate the significant difference with the first response. * p < 0.05; ** p < 0.01; *** p < 0.001.

Figure 9

D-serine, a co-agonist of NMDARs, enhanced LTP in the FS but not control group. The normalized fEPSP slope in the control and FS groups in the presence of the NMDAR co-agonist D-serine (10 μM), before and after TBS (a) or HFS (b). (c,d) Diagrams illustrating the magnitude of plasticity in the control and experimental groups in the presence of D-serine after TBS (c) or HFS (d). Two-way ANOVA following Tukey post hoc tests: * p < 0.05, ** p < 0.01.

Figure 10

The kinetic scheme used for simulating the gating mechanisms of NMDARs. Rate constants above each arrow are given in s⁻¹, except for the glycine-binding and glutamate-binding rate constants, which are in µM⁻¹s⁻¹. Gly and Glu indicate the glycine and glutamate concentrations (in µM), respectively.

Figure 11

The simulations of the NMDAR-mediated response to TBS. (a) The sums of the open state probability (O1 + O1), which reflect the time course of the NMDAR-mediated synaptic currents induced by TBS. The upper traces (black) represent the 5 responses, simulated with 10 µM of ambient glycine. The lower traces (red) were simulated with 1 µM of ambient glycine. (b) The amplitudes of the individual peaks of open state probability within the responses normalized to the first peak amplitude of the first response.