Febrile seizures in the developing brain result in persistent modification of neuronal excitability in limbic circuits

Abstract

Febrile (fever-induced) seizures affect 3-5% of infants and young children. Despite the high incidence of febrile seizures, their contribution to the development of epilepsy later in life has remained controversial. Combining a new rat model of complex febrile seizures and patch clamp techniques, we determined that hyperthermia-induced seizures in the immature rat cause a selective presynaptic increase in inhibitory synaptic transmission in the hippocampus that lasts into adulthood. The long-lasting nature of these potent alterations in synaptic communication after febrile seizures does not support the prevalent view of the 'benign' nature of early-life febrile convulsions.

Fig. 1

Hyperthermia-induced seizures result in a GABA_A receptor-dependent, long-term depression of the discharges of CA1 pyramidal cells. a, Upper row, orthodromic responses recorded in the CA1 pyramidal cell layer of control rats and of rats with hyperthermia-induced seizures (HT), 1 week after HT, in control ACSF. Lower row, responses from the same brain slices in the presence of the GABA_A receptor blocker bicuculline. The population spike is considerably smaller in the HT rat than in the control rat in bicuculline-free, but not in bicuculline-containing ACSF. Thus, the difference between the population spikes in the upper row was mostly abolished by bicuculline. b, The depression of the population spikes in CA1 was significant (*, P < 0.05) at all stimulation intensities in control ACSF. Control rats (○), n = 14 brain slices; HT rats (●), n = 16 brain slices. Inset, The depression of the population spikes was also seen in the granule cell layer of the dentate gyrus after stimulation of the preporant path. Control rats, n = 7 brain slices; HT rats, n = 8 brain slices. c, The population spike amplitude in the CA1 pyramidal cell layer from the HT rats in control ACSF is significantly smaller than that of control rats (*, P < 0.05), whereas there is no difference in the population spike amplitudes of HT and the control rats in the presence of the bicuculline (stimulus intensity, 4 mA). Data are expressed as a percent of population spike amplitudes measured in control brain slices (■, control; □, HT). These data indicate that there is a long-lasting depression of population responses of CA1 pyramidal cells of the experimental rats, and that this depression is mainly due to the enhanced GABA_A receptor-mediated ‘feed-forward’ inhibitory input to CA1 pyramidal cells.

Fig. 2

The GABA_A receptor-mediated inhibitory postsynaptic currents in CA1 pyramidal cells is enhanced in rats that experienced seizures. a, Evoked IPSCs 1 week after hyperthermia-induced seizures from an age-matched control rat (Control), from a rat that experienced hyperthermic seizures (HT) and from a hyperthermic control rat exposed to hyperthermia but whose seizures were blocked with a barbiturate (HT with seizures blocked). b, Summary plot of data obtained from recordings similar to those in a. The evoked IPSCs of brain slices from HT rats were invariably larger in amplitude than those of either normothermic or hyperthermic controls, at all stimulation intensities (including the smallest stimulation intensity, at which the responses seem to overlap because of the relatively large scale on the vertical axis). Control, (○) n = 6 cells; HT, (●) n = 6 cells; hyperthermic control, (▼) n = 9 cells. *, P < 0.05. These results were replicated in brain slices from another set of control and HT rats used to test the effectiveness of various protein kinase blockers (Fig. 4). Because the hyperthermic controls^, were identical to normothermic controls, it is the hyperthermic seizures, and not the hyperthermia itself, that result in long-lasting potentiation of the evoked IPSCs. Inset, pentobarbital injection (at the same dose as used for the hyperthermic controls) did not have any effect on the amplitude of the evoked IPSCs. Control, (○) n = 6 cells; controls with pentobarbital, (▽) n = 6 cells. c, The potentiation of the IPSCs was also found even 10 weeks after hyperthermia-induced seizures, indicating that the alterations in inhibitory neurotransmission are long-lasting; n = 6 cells for both HT (●) and control (○). *, P < 0.05. Inset, The degree of potentiation (relative to age-matched controls) was essentially unchanged at various times after hyperthermia-induced seizures; stimulation intensity, 2 mA.

Fig. 3

The frequency, but not the amplitude, of the miniature IPSCs (mIPSCs) is increased in rats that experienced seizures. a, mIPSCs from CA1 pyramidal cells from a control rat and from a rat that experienced hyperthermia-induced seizures 1 week before recording (HT). The frequency of the miniature IPSCs in CA1 pyramidal cells was considerably increased in HT rats. b and c, Summary data of the mIPSC inter-event interval (b) and amplitude (c) from cells similar to those in a as well as from cells from hyperthermic control rats (n = 14 cells in all three groups). The near-doubling of the frequency of the mIPSCs in the HT group, corresponding to a large decrease in the inter-event interval in b: median inter-event interval in normothermic control, 113.1 ms (solid lines); in HT, 64.6 ms (dashed lines); in hyperthermic control, 119.9 ms (dotted lines), without any change in the amplitude (or kinetics), indicates that the potentiation of the IPSCs has a presynaptic locus. Insets, 1 week after kainic acid-induced seizures at postnatal day 10, no equivalent alterations in the frequency of the mIPSCs could be seen (in addition, in contrast to the unchanged mIPSC amplitude after hyperthermia-induced seizures, there was a small, but statistically significant increase in the amplitude of the mIPSCs after kainate injection, compared with that of littermate controls); n = 12 cells in both the control (solid lines) and kainate-injected (dotted lines) groups.

Fig. 4

The effects of protein kinase blockers on the enhanced amplitude of the evoked IPSCs in CA1 pyramidal cells after hyperthermia-induced seizures. a, Staurosporin (0.5 μM; 2 h of incubation) abolished the difference between the evoked IPSC amplitudes in CA1 cells from control and experimental (HT) rats (1 week after seizures). The amplitude of the evoked IPSCs without staurosporin from the same HT rats was increased (as expected from Fig. 2). Control with staurosporin (○), n = 7 cells; control without staurosporin (●), n = 6 cells; HT with staurosporin (△), n = 8 cells; HT without staurosporin (▲), n = 7 cells. *, P < 0.05. The ability of staurosporin to abolish the difference between control and HT rats indicates the involvement of a protein kinase in the presynaptic potentiation of the GABA_A responses in CA1 cells after hyperthermia-induced seizures. b, The PKA-specific inhibitor Rp-cAMPS (100 μM) also abolished the enhanced evoked IPSC amplitude recorded from CA1 pyramidal cells of HT rats 1 week after hyperthermia-induced seizures. Control with Rp-cAMPS (○), n = 7 cells; control without Rp-cAMPS (●), n = 7 cells; HT with Rp-cAMPS (△), n = 7 cells; HT without Rp-cAMPS (▲), n = 8 cells; the brain slices were incubated in Rp-cAMPS for 2 hours. *, P < 0.05. c, Incubation of the brain slices with the specific PKC inhibitor calphostin C (1 μM, 2 h of incubation) did not have any effect on the increased amplitude of the monosynaptically evoked IPSCs in CA1 pyramidal cells from HT or control rats 1 week after hyperthermia-induced seizures. Control with calphostin-C (○), n = 6 cells; control without calphostin-C (●), n = 6 cells; HT with calphostin-C (△), n = 9 cells; HT without calphostin-C (▲), n = 6 cells. In contrast, the same concentration of calphostin was able to block the PDBU-induced increase in the evoked IPSC amplitude, indicating that the PKC antagonist was effectively blocking PKC activity in these experiments. d, Forskolin (10 μM) enhanced the amplitude of the evoked IPSCs in CA1 cells from HT rats to a significantly greater degree than in control rats. Evoked IPSC after forskolin with respect to the amplitude of the evoked IPSC before forskolin: control, 155.0 ± 20%; HT, 381.0 ± 81%; n = 5 for both groups. *, P < 0.05. Thus PKA, but not PKC, is involved in the presynaptic enhancement of the inhibitory neurotransmission after HT-induced seizures.