Hippocampal neuron preparation procedures followed Kay and Wong (1987a) and Surmeier et al. (1992a), for the culturing of guinea pig pyramidal and neo-striatonigral neurons, respectively. C57BL/6J mice (6-12 weeks old) were anesthetized using halothane and decapitated. The brain was quickly removed and transversely sliced (500 µm). The slices were incubated in NaHCO₃ buffered Earle's balanced salt solution (pH 7.4; Sigma) saturated with 95% O₂ and 5% CO₂ at room temperature for 1-5 hours before mechanical dissociation. Single slices were kept in a low-Ca²⁺ buffer. Hippocampi were isolated and placed in a treatment chamber containing 1.5 mg/ml of protease XIV (Sigma) in HEPES-buffered balanced salt solution (Sigma) for 15 minutes at 35 ℃, pH 7.4. Tissue was then rinsed several times in the low Ca²⁺ buffer and triturated. The isolated cells were plated for five minutes on a glass coverslip coated with concanavalin A or poly-L-lysine (Kay and Wong, 1987a; Surmeier et al., 1992b).

Recordings from hippocampal neurons were performed immediately after isolation at room temperature (22-25 ℃). Patch pipettes were pulled from a thick wall borosilicate glass capillary tube on a P-97 micropipette puller (Sutter Instrument) and fire polished (final tip resistance, 2.5-5.0 MΩ). Voltage-clamp recording was obtained using an EPC-10 amplifier (HEKA) and visualized by an IX-51 inverted microscope (Olympus). Cells were held at -100 mV after achieving whole-cell voltage clamp. Voltage pulses were then delivered for 20 ms from -90 to 60 mV with 10 mV stepping at 0.5 Hz. 60-80% of the series resistance was compensated prior to recording. Voltage-clamp data were filtered at 10 kHz and digitally sampled at 200 kHz. Currents were recorded with pClamp 10 software (Axon Instruments). Data from cells with unstable or high holding currents were discarded. The whole-cell current clamp recording model was used for the recording of action potentials. A hyperpolarizating current was used to clamp the cell membrane potential at -90 mV and a 500 pA ramp depolarizing current was then input for 100 ms to generate a stimulated action potential. The amplitude and threshold values of the spike were recorded and analyzed. The external and internal electrode solutions followed Chen et al. (2005). Unless otherwise indicated, the external perfusion solution was (mM): 20 NaCl, 10 HEPES, 1 MgCl₂, 1 CdCl₂, 60 CsCl, 150 glucose, pH 7.4 (300-305 mOsm/L) with NaOH. The intracellular recording solution contained (mM): 189 N-methyl D-glucamine, 40 HEPES, 4 MgCl₂, 0.1 BAPTA, 1.0 NaCl, 25 phosphocreatine, 2 ATP-Na₂, 0.2 GTP, 0.1 leupeptin, pH 7.2 (270-275 mOsm/L).

To eliminate statistical error resulting from different cell membrane sizes, the current value was replaced by the current density. Current density was calculated as channel current divided by membrane capacitance. The current density-voltage curve was generated by plotting current density against step voltage.

Gene mutation was performed by means of homologous recombination as previously reported by Khasar et al. (1999). Homozygotes were obtained as follows (Chen et al., 2005). C57BL/6J mice were obtained from the Center of Laboratory Animal Science of Guangdong (Foshan, China). Wild-type female C57BL/6J mice (PKC-ε^+/+) were crossed with mutant male 129 SvJae mice (PKC-ε^+/-). The (C57BL/6J × 129 SvJae) F1 PKC-ε^+/- mice were then intercrossed. Mice were raised and propagated in separate cages in a sterile environment. The room temperature was maintained at 24 ± 2 ℃, humidity ~ 40-60% and light cycle period 12 : 12 (light : dark). Water and food were available ad libitum. During the experiments, all the mice were from two to six months old. Procedures involving animals and their care were in accordance with Guide for the Care and Use of Laboratory Animals (Institute of Laboratory Animal Resources, Commission on Life Sciences, National Research Council; NIH Pub. No.85-23, revised 1996), approved by Jinan Central Hospital affiliated to Shandong University (No. 2018-53-01).

Heated water-bath seizure induced mice were prepared for the FS mouse model as described previously (Zhao et al., 2016) and onset characteristics were taken as the evaluation standard. C57BL/6J mice were placed in water at 44.9 ℃ for five minutes and were assigned to the FS group once seizures occurred. εV1-2 treated mice were also exposed to 44.9 ℃ water for five minutes and are referred to as the FS + εV1-2 group. PKC-ε inhibitor εV12 (5 nM) (Feng et al., 2013) in FS + εV1-2 group was ICV injected into the right dorsal hippocampus 12 hours before the heated water test. Water immersion for control mice was at 37 ℃ for five minutes. If seizures occurred within this time, mice were immediately removed from the water. Under the FS model, mice were treated with heated water 10 times on each of two days. Hyperthermiainduced febrile seizure mice were characterized based on onset characters that included seizure latency, duration and frequency. Following a previous study (Racine, 1972), seizure frequency was classified into five stages: (1) bilateral forelimb clonus (BFC) with falling and rearing; (2) BFC and rearing; (3) unilateral forelimb clonus; (4) head nodding; and (5) facial movement. Seizure latency was measured as the time from being placed into the water to any sign of seizure onset.

PKC-ε inhibitor εV1-2 and its negative control were obtained from Biomol (Plymouth Meeting, PA). Recombinant IL-1β was obtained from Sigma.

Trace images and related data were processed for statistical analysis using Clampfit 10.0 (Axon Instruments). Numerical values are reported as mean ± SEM. In some cases, the means were tested for equality using a paired Student’s t test. Values of P < 0.05 were considered significant.

IL-1β is known to be involved in the modulation of voltagegated ion channels (Choi et al., 2007; Liu et al., 2006). To investigate the effects of IL-1β on Na⁺ channels, here, mouse hippocampal neurons were isolated and cultured with 10 ng/ml IL-1β for 12 hours. From the results of whole-cell current recording, IL-1β inhibited the Na⁺ currents in these cells (Fig. 1A). Current densityvoltage curves for untreated neurons and neurons treated with 10 ng/ml IL-1β were plotted. The current density was obtained from peak current divided by membrane capacitance determined by an EPC-10 amplifier in the lock-in simulative protocol. The peak current for control cells was obtained at a step voltage of -30 mV, while IL-1β depolarized the cell membrane to -20 mV (Fig. 1B). As shown in Fig. 1C, the average peak currents from 10 cells was significantly reduced by 45.3% after IL-1β treatment (Fig. 1C). To ascertain whether IL-1β reduced Na⁺ current via PKC-ε, 200 µM of εV1-2 a specific inhibitor of PKC-ε (Cang et al., 2009) was delivered to the culture medium for 30 minutes prior to IL-1β addition. As shown in Fig. 1D, εV1-2 treatment alone had no significant effect on hippocampal neurons. A cell viability assay also demonstrated that there were no alterations when using either IL1β or εV1-2 alone, Additionally, εV1-2 significantly alleviated the inhibitory effect of IL-1β on mouse hippocampal neurons (P < 0.01; n = 10, Fig. 1D). Taken together, this suggested that IL-1β inhibited Na⁺ currents in mouse hippocampal neurons via PKC-ε.

Figure 1.

The inhibitory effect of IL-1β on Na⁺ currents was dependent on PKC-ε in mouse hippocampal neurons. A, whole-cell current recording of the Na⁺ currents in mouse hippocampal neurons (n = 10). Neurons cultured without IL-1β treatment (Left) and with 10 ng/ml of IL-1β (Right) for 12 hours. B, Comparison of current density-voltage curves between no treatment and 10 ng/ml (12 h) of IL-1β (P < 0.05; n = 10; Student's t-test), as compared with no IL-1β (^*P < 0.05). C, Time course of the inhibitory effects of IL-1β on Na⁺ current (n = 10). Data normalized to current magnitude prior to IL-1β treatment. D, Inhibitory effect of IL-1β abolished by εV1-2, an inhibitor of PKC-ε (n = 10), as compared with the no IL-1β group (^**P < 0.01); as compared with IL-1β group (^##P < 0.01, Student's t-test).

PKC-ε null mutant mice (PKC-ε^-/-, Khasar et al., 1999) were used to examine and analyze the role of IL-1β modulation of Na⁺ currents in vivo. Mouse hippocampal neurons from PKC-ε^-/- mice and wild-type littermates (PKC-ε^+/+) were isolated and culhours. PKC-ε^-/- mice exhibited significantly less Na⁺ current reduction than PKC-ε^+/+ mice (P < 0.05). After 12 hours, as shown in Fig. 2, no significant reduction in Na⁺ current was observed in PKC-ε^-/- mice following IL-1β treatment as compared to the untreated mice, However, a significant reduction in Na⁺ current was observed in PKC-ε^+/+ mice following IL-1β treatment as compared to untreated mice (P < 0.05; n = 10). Thus, in PKC-ε^-/- mice, IL-1β showed no significant inhibitory effect on voltage-gated Na⁺ current and the presence of endogenous PKC-ε aggravated the effect of reducing voltage-gated Na⁺ current.

Figure 2.

The inhibitory effect of IL-1β on Na⁺ currents was dependent on PKC-ε in vivo. 10 ng/ml of IL-1β was added in the culture of the hippocampal neurons isolated from PKC-ε^-/- mice (n = 10) and wild-type littermates PKC-ε^+/+ (n = 10). Data were normalized to the current of neurons from PKC-ε^+/+ mice not treated with IL-1β. The mean minimal current of ε^+/+ mice was inhibited by 56.1 ± 5.1%, which was significantly different from PKC-ε^-/- mice (P < 0.05), as compared with 10 ng/ml of IL-1β, PKC-ε^-/- group (^*P < 0.05 and ^**P < 0.01).

IL-1β induces repetitive firing of action potentials. To further investigate the effects of IL-1β, the action potentials of hippocampal neurons were recorded using a ramp depolarizing current. IL1β and εV1-2 were administrated chronically to neuron cultures as described above. Results showed that neither IL-1β nor IL-1β + εV1-2 had any significant effect on the spike threshold (Fig. 3B), an important indicator of neuronal excitability. However, the peak depolarization of spikes was significantly reduced by IL-1β treatment (P < 0.01, n = 10), but this effect was abolished following administration of εV1-2 (n = 10, Fig. 3A). These results were consistent with the suppressing effect of current density mentioned above, which indicated that IL-1β reduced the amplitude of the action potential due to its inhibitory effect on Na⁺.

Figure 3.

Effect of IL-1β and PKC-ε inhibitor εV1-2 on spike amplitude and threshold examined by current-clamp recording (n = 10). as compared with no IL-1β group (^**P < 0.01); as compared with IL1β group (^##P < 0.01, Student's t-test).

To ascertain the role of PKC-ε in febrile seizures in vivo, a heated water-bath model of FS induction was employed as described previously (Zhao et al., 2016). Results demonstrated that the frequency and duration of FS in mice injected with εV1-2 were increased over the pretreatment times (Fig. 4A,C). Comparing the 10 th heat treatment to the 1st, the mean frequency was increased by 70.6% in the FS + εV1-2 group but was decreased by 11.1% in FS mice. Meanwhile, the mean FS duration was lengthened by 72.5% in the εV1-2 injection group and 14.4% in FS mice. Seizure latency of the FS group decreased slightly, and that of the FS + εV1-2 group demonstrated similar changes with no significant differences with the FS group (Fig. 4B). Compared with the 1st pretreatments in FS and FS + εV1-2 mice, the mean latency after the 10 th pretreatment was shortened by 19.0% and 18.3%, respectively. Moreover, a FS model using PKC-ε^-/- mice by means of heat-induction was also developed following the procedures employed to generate PKC-ε^+/+ mice. After 10 treatments, the outcomes were consistent with those of the FS + εV1-2 group. These results demonstrate that PKC-ε plays a highly significant role in mouse FS.

Figure 4.

The susceptibility and frequency of mice (n = 10) with FS after injection of εV1-2. (A) PKC-ε inhibitor significantly reduced seizure frequency (P < 0.01) and (C) duration (P < 0.01), (B) but had no significant impact on response latency (P > 0.05), as compared with the t-test). tured in 10 ng/ml of IL-1β for 12 FS group (^*P < 0.05 and ^**P < 0.01).

Inflammatory cytokines play functional roles in neuronal excitability and, specifically, receptors for IL-1β have been found in the hippocampus at high concentration (Dubé et al., 2007). Mechanisms of FS involve alterations to molecular and functional properties, such as their effect on neurotransmitter receptors (Sanchez et al., 2001; Zhang et al., 2004) or voltage-gated ion channels (Brewster et al., 2002; Chen et al., 2001). IL-1β is known to increase seizure susceptibility by augmenting nitric oxide formation and increases neuronal excitability by direct GABA(A) receptor inhibition, increasing NMDA receptor function or inhibiting K+ efflux (Meini et al., 2000; Miller et al., 1991; Viviani et al., 2003). IL-1β rapidly inhibitsing voltage-gated Ca²⁺ channel activity, an effect that occurs within a few seconds after the addition of IL-1β to acutely dissociated hippocampus CA1 neurons (Plata-Salamán and Ffrench-Mullen, 1992, 1994) and in cultured cortical neurons (Macmanus et al., 2000). Moreover, brief exposure (five minutes) to IL-1β led to a specific (receptor-mediated) reduction of sodium current in trigeminal nociceptive neurons, while chronic treatment with IL-1β enhanced this current (Liu et al., 2006), which may be one of the mechanisms accounting for the effect of IL-1β on the nerve conduction underlying pain perception. However, in the central nervous system, the interrelations of IL-1β and Na⁺ channels are unclear. In cortical neurons, IL-1β inhibits the Na⁺ current via its receptor in a time- and dose-dependent manner, but does not affect voltage-dependent activation and inactivation (Zhou et al., 2011). Previous real-time PCR data shows such inhibition is caused by reducing the protein expression of Na⁺ channels, which may be compensated for by other signaling pathways during long-term application of IL-1β (Zhou et al., 2006, 2011). It can be hypothesized that an effect specifically directed to the inhibition of voltage-gated Na⁺ channels by IL-1β may modulate neurotoxic or neuroprotective properties related to final seizure onset. It is worth noting that only using hippocampal neurons may have limitations in fully explaining the complex pathogenesis of febrile seizures.

Neuronal injury can be both neurotoxic and neuroprotective. The dosage for IL-1β neurotoxicity varies widely from 0.05 (Viviani et al., 2003) to 100 ng/ml (Li et al., 2000) or even 100 µg/ml (Strijbos and Rothwell, 1995). This suggests the presence of mul-tiple pathway mechanisms. Moreover, the dosage for neuroprotective effects varies from 5 to 100 ng/ml (Carlson et al., 1999; Macmanus et al., 2000; Prehn, 1996), while the level of IL-1β in cerebrospinal fluid has previously been reported at 4.2 ± 1.5 pg/ml (Lahat et al., 1997). In various pathophysiological conditions, the concentration of IL-1β observed in cerebrospinal fluid can be greater than 0.1 ng/ml (Plata-Salamán and Ffrench-Mullen, 1994). As estimated previously (Zhou et al., 2006), the concentration in sensitive tissues should be increased by 1-2 magnitudes, and the maximum reaches up to 10 ng/ml. Na⁺ current suppression at this dosage was found in preliminary experiments. Subsequently, mean percentages of Na⁺ currents were found to be reduced to 96% (IL-1β 0.1 ng/ml), 68% (IL-1β 1 ng/ml) and 62% (IL-1β 10 ng/ml). Therefore, a 10 ng/ml concentration of IL-1β was employed in the present experiment to probe its biological effect on voltage-gated Na⁺ channels. During the first four hours of the application of 10 ng/ml IL-1β, no significant changes of Na⁺ currents were obtained (Fig. 1C). This may be a consequence of a reduced number of Na⁺ channels in hippocampal neuron membranes, which was consistent with the modulation of their expression in both alveolar epithelial cells (Roux et al., 2005) and cortical neurons (Zhou et al., 2011)

εV1-2 has been used as an inhibitor of PKC-ε (Zhou et al., 2011). In Zhou's study, the inhibitor of PKC (chelerythrine; 1 µM) and p38 MAPK (SB 203580; 1 µM) were applied 30 minutes prior to IL-1β. Results indicated that the inhibitory effect of IL1β on Na⁺ current in cortical neurons was dependent on PKC and then p38 MAPK. Consequently, it can be hypothesized that Na⁺ channels are regulated through a mechanism similar to that existing in other neurons, the main reason being that the PKC-ε isoform is highly expressed in hippocampal neurons (C Tanaka and Nishizuka, 1994; Naik et al., 2000) and Na⁺ channels and PKC-ε are widely coexpressed (Chen et al., 2005). The present study explored the roles of PKC-ε via selective inhibition by a PKC-ε inhibitor εV1-2 and targeted gene deletion.

In mouse hippocampal neurons, the Na⁺ current was significantly reduced with extracellular treatment by IL-1β (P < 0.01), while the application of IL-1β + εV1-2 abrogated the inhibitory effect of IL-1β with no significant change in current magnitude when compared to control. Thus, PKC-ε contributes to the inhibitory effect of IL-1β. This was consistent with a previously reported effect of PKC on IL-1β (Zhou et al., 2011). Additionally, the endogenous PKC-ε existing in normal neurons does not inhibit Na⁺ currents in the absence of an increased concentration of IL1β. This indicates that the absence of PKC-ε does not impact Na⁺ currents even if IL-1β is increased and PKC-ε may be capable of co-opting this mechanism in pathological cells. Further an effect of PKC-ε on IL-1β was identified in PKC-ε^-/- mouse neurons, while the inhibition of Na⁺ current was similar to results obtained from experiments using PKC-ε inhibitor. As previously reported, PKC-ε is required for selective regulation of cardiac Na⁺ channels, which may work by interacting with Na⁺ channel proteins (Xiao et al., 2001). PKC-ε knock-out mice show no change in Na⁺ currents in response to PKC-ε activation in hippocampal neurons, suggesting PKC-ε as the PKC isozyme that serves as a regulator for voltage-gated Na⁺ channels (Barnett et al., 2007; Chen et al., 2005).

It has been reported that IL-1β shows proconvulsive effects before FS onset. IL-1β degrades electrographic activity and prolongs behavioral seizure duration in rodent epilepsy models (Vezzani et al., 1999, 2002). Seizure-like behavioral changes were significantly reduced following inhibition of IL-1β production by the treatment of an inhibitor of a IL-1β-converting enzyme (Donnelly et al., 2001). Moreover, administration of an extra-high dose (116 ng/2 µl) of IL-1β is sufficient to generate seizures in the absence of increased brain temperature (Dubé et al., 2005). Endogenous IL-1β contributes to hyperexcitability and seizures generated by fever and hyperthermia in mice, but this is improved by injecting IL-1β six minutes prior to seizure induction (Dubé et al., 2005).

Further, the chronic effect of IL-1β may be different to that in acute situations. Acute administration of IL-1β to trigeminal nociceptive neurons inhibits Na⁺ currents, whereas its chronic effect potentiates Na⁺ currents (Liu et al., 2006). Thus, the effect of PKC-ε on IL-1β may be determined by the type of application. To investigate the properties of PKC-ε during chronic IL-1β exposure, in the present study, the FS model mice were generated by heated-water treatment for two hours following each 12 hour application of εV1-2. The seizure frequency and duration following exposure to εV1-2 were exacerbated and prolonged significantly, possibly resulting from abrogation of the neuroprotective effect of IL-1β by εV1-2. Therefore, PKC-ε may be involved in an IL1β neuroprotective effect in FS. After ICV injection of εV1-2, the occurrence rate and characteristics of seizures were consistent with the FS group before the first heated water treatment (data not shown). That is to say, application of PKC-ε inhibitor cannot be the only factor that affects FS onset. Thus, endogenous IL-1β and PKC-εboth exert functions acting on FS via fever or hyperthermia.

Taken together, IL-1β inhibits the Na⁺ currents in mouse hippocampal neurons and the PKC-ε contributes the effect of IL-1β. Targeted gene deletion experiments identified the effects of IL1β and PKC-ε. IL-1β reduced the amplitude of action potentials resulting from its inhibitory effect on the Na⁺ channels, which may decrease the release of presynaptic transmitters of neuroexcitability which in turn is neuroprotective for excitotoxicity in hippocampal neurons. Moreover, PKC-ε reduces the susceptibility and frequency of FS onset via the neuroprotective effect of IL1β in hippocampal neurons. IL-1β's effect on the Na⁺ channel and the role of PKC-ε in FS has been demonstrated. However, effect of K⁺ currents or other calcium currents have not been ruled out in the present study.