Introduction

The primary physiological function of glucocorticoids (cortisol in humans and corticosterone in rodents) (Grinevich et al. 2012) is to mediate the specificity and magnitude of neural, behavioral and hormonal responses to stress (Lucassen et al. 2014). The impaired feedback regulation of hypothalamic-pituitary- adrenal (HPA) axis results in the increased level of corticosterone (CORT) (Buhl et al. 2010; Lee et al. 2015). Short-term and long-term of excessive CORT have the profound influence on physiological and neurobehavioral functions during adolescent development (Kinlein et al. 2017).

Insulin signaling has been studied extensively in peripheral organ systems and central nervous system (CNS) (Chiu and Cline 2010; Dai et al. 2014). Particularly, it is reported that insulin signaling plays various roles in CNS, including the regulation of neuronal survival, synaptic plasticity, cognition, and memory (Chiu et al. 2008). Insulin resistance can be induced by CORT treatment, including the reduction of associated intracellular pathways and insulin receptor (IR) activation (Buhl et al. 2010; Dobarro et al. 2013; Solas et al. 2013). Chronic CORT treatment can alter gene expression of insulin-signaling pathway, resulting in the impairment of learning and memory function (Osmanovic et al. 2010). In addition to responding through gene-mediated signaling pathways, CORT can also rapidly and reversibly change nongenomic hippocampal signaling within minutes after stress-induced elevation of corticosteroid level (Karst et al. 2005). Short-term CORT administration and peripheral insulin resistance can damage insulin signaling in rat hippocampus (Piroli et al. 2007).

Ion channels are vital to the activation of action potential (AP) and the maintaining of resting membrane potential. AP is a fundamental property of hippocampal neurons, which mainly reflects the excitability of cell membrane. It has been demonstrated that toxic impairments could lead to neuronal excitotoxicity by virtue of modifying AP (Chen et al. 2013). Voltage-gated potassium ion (K+) channels have been demonstrated similar characteristics in pancreatic β-cells, smooth muscle cells and central neurons (Teramoto 2006). K+ channels affect resting membrane potential, influence repolarization and depolarization, and alter waveform and frequency of AP.

Depression is a common debilitating mood disorder and becomes the second disease contributing to the disability worldwide (Cai et al. 2015). Some studies showed that CORT was associated with cognitive impairments and behavior changes, such as depression (Aurand et al. 2016; Grinevich et al. 2012; Lee et al. 2015). CORT suggested a possibility of exploring the mechanism of antidepressants (Aurand et al. 2016). Several effective antidepressant drugs and therapies that regulate cortisol secretion have shown promise in clinical trials (Johnson et al. 2006). However, the definite changes of excessive CORT in voltage-gated ion channels in CNS and its upstream signaling effectors remain unclear. In this study, we constructed an in vitro model of excessive CORT (Karst et al. 2005) to explore its impact on voltage-gated ion channels and insulin-signaling pathway in hippocampal CA1 pyramidal neurons. We reported that the stimulation of excessive CORT induced excitotoxicity of hippocampal neurons and sensitivity of potassium channels via phosphatidylinositol 3 kinase (PI3K)-mediated insulin-signaling pathway, which played an important role in cognitive diseases.

Material and methods

Slice preparation

The experimental protocol was approved by the local Ethical Committee of Nankai University. Male Wistar rats (10-14d on postnatal days) were used in the experiment (Yin et al. 2017). They were purchased from the Laboratory Animal Center, Academy of Military Medical Science of People’s Liberation Army, and temporarily fed in the animal house of Medicine School, Nankai University. After the rats were decapitated, their brains were quickly removed and immersed in an ice-cold and oxygenated (95% O2 and 5% CO2) dissection buffer, containing the following (in mM): 220 sucrose, 2.5 KCl, 6 MgCl2, 1 CaCl2, 1.23 NaH2PO4, 26 NaHCO3 and 10 Glucose, and pH 7.4. Subsequently, hippocampal slices (400 μm thick) were cut using a vibratome (VT1000S, Leica, Germany). They were placed in a water bath kettle and incubated with artificial cerebrospinal fluid (ACSF) containing the following (in mM):125 NaCl, 25 NaHCO3, 1.25 NaH2PO4, 1.25 KCl, 1.5 MgCl2, 2.0 CaCl2 and 10 Glucose, and pH 7.4 (saturated with 95% O2–5% CO2). Then, slices were used for whole-cell patch-clamp recording and Western blot assay.

Drug application

The final concentrations of CORT and insulin were 0.1 μM and 0.5 μM, respectively. CORT was dissolved in ethanol. The concentration of ethanol was less than 0.009% (Karst et al. 2005). Transient outward potassium current (IA) and delay rectifier potassium current (IK) were two types of voltage-dependent potassium channels. IA and IK were separated by 25 mM tetraethylammonium chloride (TEA-Cl) and 3 mM 4-Aminopyrodine (4-AP) (Chen et al. 2014; Liu et al. 2014). Tetrodotoxin (TTX, 1 μM) and CdCl2 (0.2 mM) blocked sodium and calcium channels, respectively. TTX was purchased from the Research Institute of the Aquatic Products of Hebei (China). TEA-Cl, 4-AP, CdCl2, EGTA, HEPES and ATP-Na2 were obtained from Sigma (USA), and other reagents were of A.R. grade.

Whole-cell patch-clamp recording

Whole-cell patch-clamp recordings obtained from the CA1 hippocampal neurons using a patch-clamp amplifier (EPC-10, HEKA, Germany). It was viewed with an upright microscope (BX51WI, Olympus, Japan) and visualized on a television monitor connected to a low-light-sensitive CCD camera (710 M, DVC, USA). After transferred into a glass-bottomed recording chamber, slices were submerged in 1 ml oxygenated ACSF solution. The patch electrodes were made into opening tips of 1-2 μm by a vertical puller (PIP5, HEKA, Germany). Then, electrodes with 4–8 MΩ electrical resistance were filled with standard solution, containing the following (in mM): 140 KCl, 10 HEPES, 10 EGTA, 2 MgCl2·6H2O and 2 ATP-Na2, and pH 7.4. All cells were held at −70 mV when slow and fast capacitance compensation was automatically performed.

Western blot assay

Hippocampal slices were used for Western blot assay. Slices were incubated in ACSF for 1 h before treatment with CORT and insulin. Preparation of tissue lysates has been described in our previous studies (Xu et al. 2012). Equal protein loading was separated by SDS-PAGE, and it was transferred to polyvinylidene-difluoride (PVDF) membranes. Non-specific binding sites on membranes were incubated with blocking buffer, which was 5% fat-free milk powder in Tris-buffered saline including 5% Tween 20 (TBST) for 1 h at room temperature. Membranes were incubated in the following antibodies: primary antibodies (anti-PI3K, 1:2000 dilution, CST; anti-Akt, 1:2000 dilution, CST; anti-p-Akt, 1:2000 dilution, CST; anti-β-actin, 1:2000 dilution, Abcam) overnight at 4 °C and horseradish peroxidase-conjugated secondary antibodies (anti-rabbit IgG, 1:2500 dilution, Invitrogen) for 1 h at room temperature. Blots was detected by a chemiluminescent imaging system (Tanon 5500, Tanon Science & Technology Co., Ltd., China). Ser473 of p-Akt was tested. β-actin was served as an internal control on the same Western blots. The experiments repeated at least three times.

Data acquisition and analysis

Whole-cell patch-clamp data were acquired with a patch-clamp amplifier (EPC-10, HEKA, Germany) and analyzed by Clampfit 10.5, Origin 9 and SPSS 22. Western blot assay data were detected by a chemiluminescent imaging system (Tanon 5500, Tanon Science & Technology Co., Ltd., China) and analyzed by Photoshop CS6, Origin 9 and SPSS 22. Paired sample t test was applied in vehicle and vehicle-insulin groups, CORT and CORT-insulin groups. The activated currents were converted to conductance (G) using the formula: G = I/(Vm - Vr), where Vr was reversal potential. The current-voltage curves, activation curves and steady-state inactivation curves were respectively fitted with Boltzmann function: I = 1/{1 + exp[(Vm – Vh1)/k]}, G/Gmax = 1/{1 + exp[(Vm – Vh2)/k]} and I/Imax = 1/{1 + exp[(Vm – Vh3)/k]}. Vh1 was the voltage of half-maximal activation. Vh2 was the voltage of half-maximal conductance. Vh3 was the voltage of half-maximal inactivation. k was a slope factor. The recovery time of curves was fitted with a mono-exponential function: I/I max = A{1 - exp[Δt/τ]}, where I was current amplitude at time of Δt, A was the coefficient of amplitude and τ was the time constant.

The values were expressed as means ± SEM. p < 0.05 was recognized as significant difference.

Results

As shown in Fig. 1, there were four groups in our experiment. Before recording, the vehicle group received no treatment and the CORT group was pretreated with CORT for 5 min. The recordings of vehicle and CORT groups were performed 5 min after the establishment of recording pattern. The recordings of vehicle-insulin and CORT -insulin groups were performed 5 min after insulin treatment.

Fig. 1
figure 1

The schematic of four groups with different treatments. -5 min represented the time before the establishment of recording pattern, defined as 0 min. Drugs in the ovals added at the appointed time. Rectangles represented group and recording of electrophysiological indexes at a particular time. The concentration of ethanol was less than 0.009%. cort = corticosterone; ins = insulin

Full size image

Effects of corticosterone on the excitability of hippocampal neurons

AP is used for describing the property of excitability of hippocampal pyramidal neurons. The frequency of AP was evoked by a 50 pA depolarizing current injection, 500-ms maintained (Fig. 2a). After insulin application, the frequency of AP was significantly increased by 15.50 ± 5.80% (Fig. 2c, p < 0.05, t = −2.67, paired sample t test; n = 6; % of vehicle). However, after pretreatment with CORT, the frequency of AP was not influenced by insulin (Fig. 2c, p > 0.05, t = −0.70, paired sample t test; n = 6; % of CORT). There was a marked increase in the frequency of AP (Fig. 2e, p < 0.01,t = −3.20, independence sample t test; n = 6) induced by CORT itself. The results indicated that insulin or CORT could increase the excitability of neurons. However, insulin had no effect on the AP frequency increased by CORT.

Fig. 2
figure 2

Comparison of firing frequency and half-width of AP before and after application of drugs in four groups. (a, b) A recording example before and after application of drugs in four groups. The evoking firings, depolarizing current (500 ms, 50 pA), stimulated the neurons to acquire the distinction of insulin sensitivity in four groups. AP was evoked by 5 ms brief depolarizing current pulses in four groups. (c, d) Comparison of firing frequency and half-width in four groups. (e) The contrast of firing frequency and half-width between vehicle and cort groups. Data were presented as means ± SEM. * p < 0.05, ** p < 0.01. cort = corticosterone

Full size image

The half-width of AP was elicited by a 5-ms depolarizing current pulse of 100 pA (Fig. 2b). After treatment with insulin, the half-width of AP increased by 10.06 ± 3.65% (Fig. 2d, p < 0.05, t = −2.76, paired sample t test; n = 6; % of vehicle). After pretreatment with CORT, insulin increased the half-width by 14.37 ± 2.8% (Fig. 2d, p < 0.01, t = −5.12, paired sample t test; n = 6; % of CORT). CORT itself showed no effect on the half-width of AP (Fig. 2e, p > 0.05, t = −1.21, independence sample t test; n = 6). The results implied that insulin had an evident influence on the half-width of AP. CORT showed no effect on the half-width of AP and no mutual interference with insulin.

Effects of corticosterone on the I-V changes of IA and IK

To record IA, the holding potential was −70 mV, and the current traces of IA were evoked by using an 80 ms constant depolarizing pulse from −50 to +90 mV in increments of 10 mV (Fig. 3a). IK was obtained using a 300 ms constant depolarizing pulse by a similar pulse protocol (Fig. 3b). The current-voltage curves of IA and IK in four groups were fitted well with the Boltzmann equation: I = 1/{1 + exp[(Vm–Vh)/k]}. The detailed parameter list was showed in Table 1. After insulin application, IA currents were decreased significantly (Fig. 3c, p < 0.05, t = −3.10, paired sample t test; n = 6; % of vehicle) while IK currents had no significant change (Fig. 3d, p > 0.05, t = −0.14, paired t test; n = 8; % of vehicle) at different command potentials. After pretreatment with CORT, the effect of insulin on IA currents was abolished (Fig. 3c, p > 0.05, t = −0.10, paired t test; n = 6; % of CORT), and the effect of insulin on IK currents had no significant change (Fig. 3d, p > 0.05, t = −2.14, paired sample t test; n = 6; % of CORT). CORT alone had no effect on IA and I K currents (Fig.S1a, b). It showed that insulin could effectively reduce the amplitudes of IA currents. After pretreatment with CORT, however, insulin had no effect on the amplitudes of IA currents.

Fig. 3
figure 3

Effects of corticosterone on the I-V relationship of IA and IK. (a, b) A recording example before and after application of drugs in four groups. IA and IK were obtained by 80 and 300 ms depolarizing pulses, respectively, from a command potential of −50 to +90 mV in increments of 10 mV, and the holding potential was −70 and − 50 mV, respectively. Comparison of the effects of corticosterone on current-voltage activation kinetics curves of IA (c) and IK (d) in four groups. The steady-state activation kinetics curves of IA (e) and IK (f). The current-voltage curves were fitted with a Boltzmann equation. Utilized the eq. G = I/(Vm - Vr), amplitudes of IA and IK currents were converted into conductance and fitted with a Boltzmann function. Data were presented as means ± SEM. * p < 0.05, ** p < 0.01. cort = corticosterone

Full size image
Table 1 Effect of corticosterone on I-V curve of IA and IK
Full size table

Effects of corticosterone on the activation kinetics of IA and IK

The activation kinetics curves of IA and IK in four groups were visible (Fig. 3e, Fig. 3f), which were fitted well with the Boltzmann equation: G/Gmax = 1/{1 + exp[(Vm–Vh)/k]}. It showed that after the addition of insulin for 5 min, the activation curves of IA and IK showed no visible alteration. Moreover, after pretreatment with CORT, there were no significant change in the activation curves of IA and IK with insulin application. The detailed parameter list was shown in Table 2. The CORT alone had no influence on the activation curves of IA and IK (Fig.S1c, Fig.S1d).

Table 2 Effect of corticosterone on the activation kinetics of IA and IK
Full size table

Effects of corticosterone on the inactivation kinetics of IA

To obtain the steady-state inactivation kinetics, neurons were under the parameter: being at the holding potential of −70 mV, elicited with an 80 ms test pulse to +50 mV, proceeded by 80 ms prepulse to potentials between −110 and + 10 mV of currents (Fig.4a). The amplitudes for IA currents were normalized as I/I max, and the curves were fitted up with the Boltzmann equation: I/Imax = 1/{1 + exp [(Vm – Vh)/k]. The parameter of inactivation curves was displayed in Table 3. It showed that insulin significantly reduced the Vh and increased the slope factor of IA inactivation curves (Fig.4c). With CORT pretreatment, there was no further change following insulin administration. CORT alone showed no effect on the steady-state inactivation (Fig.S1e).

Fig. 4
figure 4

The steady-state inactivation kinetics curves of IA and recovery from inactivation curves kinetics of IA. (a) Normalized steady-state inactivation of IA before and after application of drugs. Currents were elicited with a 80 ms test pulse to +50 mV proceeded by 80 ms prepulse to potentials between −110 and + 10 mV. Steady-state inactivation of IA was normalized as I/Imax and fitted with a Boltzmann function. (c) Effects of corticosterone on IA steady-state inactivation kinetics curves. (b) The recovery from inactivation of IA was well fitted with a mono-exponential function. (d) Effects of corticosterone on the recovery from inactivation of IA. Data were presented as means ± SEM. cort = corticosterone

Full size image
Table 3 Effect of corticosterone on the inactivation parameters of IA
Full size table

Effects of corticosterone on the recovery from inactivation of IA

To explore the kinetics of recovery from activation, neurons were under the holding potential of −70 mV and applied with an 80 ms conditioning depolarizing pulse of +50 mV that completely inactivated the transient outward potassium channels. Then neurons were treated with an 80 ms test pulse of +50 mV that was following a series of 80 mV intervals varying from 10 to 265 ms (Fig. 4b). The I2/I1 ratio signified the recovery from inactivation, in which I1 was the peak value of IA evoked by the conditioning pulse, and that of I2 was evoked by the test pulse. The recovery from inactivation of I2/I1 vs. the interval of the 80 mV was well fitted up with a mono-exponential function: I/Imax = A + B exp(−t/τ). It showed that insulin did not influence the recovery from inactivation of IA (Fig.4d). With CORT pretreatment, there were no further change following insulin administration. CORT alone did not make significant change in the recovery from inactivation (Fig.S1f). τ, the parameter of recovery from inactivation of IA, listed in Table 4.

Table 4 Effect of corticosterone on the recovery from inactivation of IA
Full size table

Corticosterone alleviated the activation of insulin-signaling pathway induced by insulin

As we know, the activation of insulin-signaling pathway can upregulate the protein expression of Kv4.2 (Yao et al. 2012), which is the main α-subunit of the transient potassium channels. Insulin activates the IR tyrosine kinase, which phosphorylates and recruits different substrate adaptors. PI3K plays a major role in insulin function and mainly phosphorylates the downstream target protein, Akt. PI3K and Akt were determined by Western blot assay. The expression of PI3K was significantly increased in the vehicle and vehicle-insulin groups (Fig.5b, p < 0.05, t = −5.05, paired sample t test; n = 3; % of vehicle) while no significant difference exists between CORT and CORT-insulin groups (Fig.5c, p > 0.05, t = −0.61, paired sample t test; n = 3; % of CORT). We also examined the possible downstream proteins, Akt. It showed that insulin treatment significantly increased the phosphorylation of Akt compared to that of vehicle group (Fig.5d, p < 0.05, t = 4.53, paired sample t test; n = 3), and the phosphorylation of Akt was increased in the CORT and CORT-insulin groups (Fig.5e, p < 0.05, t = −4.44, paired sample t test; n = 3). CORT alone did not change the expression of PI3K (Fig.5f, p > 0.05, t = −0.36, independence sample t test; n = 3) and the phosphorylation of Akt (Fig.5g, p > 0.05, t = −0.66, independence sample t test; n = 3). These results demonstrated that insulin possibly inhibited transient potassium currents via the activation of PI3K signaling pathway, and CORT alleviated insulin-induced inhibition of transient potassium currents by affecting the activation of PI3K/Akt signaling pathway.

Fig. 5
figure 5

Effects of corticosterone on PI3K/Akt of CA1 neurons. (a) The representative immunoreactive bands of PI3K (80 kDa), p-Akt (60 kDa) and β-actin (43 kDa) of four groups. (b, c) Quantitative analysis of PI3K/β-actin in four groups. (d, e) Quantitative analysis of Akt/p-Akt in four groups. (f, g) Quantitative analysis of PI3K/β-actin and Akt/p-Akt between vehicle and cort groups.* p < 0.05, ** p < 0.01. (n = 3)

Full size image

Discussion

Stress can induce activation of HPA axis and lead to excessive corticosteroids released into the systemic bloodstream (Russell et al. 2018). Excessive CORT, resulting from the disorder of HPA axis, is vital to the progression of depression (Grinevich et al. 2012; Johnson et al. 2006; Lee et al. 2015; Wieczorek et al. 2015). Chronic CORT treatment induces cellular changes of hippocampus, which can be prevented by antidepressants (Buhl et al. 2010; Murray et al. 2008). Neuroscientists have put forward a principal hypothesis that depression could result from excitotoxic damage to neurons in hippocampus (Kudryashova 2015), which causes cognitive impairments and behavior changes. Exposure to CORT results in cognitive deficits in neurogenesis and synaptic plasticity (Dobarro et al. 2013; Stranahan et al. 2008) or insulin resistant with decreased IR phosphorylation (Solas et al. 2013). In this study, we reported that the stimulation of excessive CORT induced excitotoxicity of hippocampal neurons and sensitivity of potassium channels by PI3K-mediated insulin-signaling pathway,which played an important role in neuronal development and maturation.

AP is the basic characteristic reflecting neuronal excitability on CNS, which is regulated by ion channels in membrane. Toxic impairment or activation of signaling pathways can modify the excitability of neurons (Joseph and Turrigiano 2017; Larimore et al. 2017), which ultimately leads to neuronal excitotoxicity (Chen et al. 2013). We observed that insulin and CORT could increase the excitability of CA1 pyramidal neurons. The increased frequency, induced by CORT, was not increased further by insulin treatment (Fig.2). These results suggest that CORT has mutual effect against insulin-induced excitability.

Many toxins and drugs aim at ion channels of cell membrane. K+ channels are important to repolarization and hyperpolarization, which determine the excitability of neurons. Transient potassium channel (IA) and delay rectifier potassium channel (IK) are two main types of neuronal K+ currents. They can be activated by depolarizing voltage gradually (Schroder et al. 2000). IA plays an important role in repetitive firing of AP while IK is associated with the process of repolarization (Zhang and McBain 1995). In the present study, insulin effectively restrained IA while insulin had no effect on IA with CORT pretreatment (Fig.3). It suggests that CORT alters IA to produce excitotoxicity on neurons and induces insulin resistant. The block of potassium currents, IA in our study, could increase the neuronal excitability. Short-term modulation of IA can arise from a rapid mechanism due to changes in voltage-gating properties or intracellular trafficking of the channel proteins (Yao et al. 2013). We observed that CORT applied acutely to the bath solution significantly increased neuronal excitability without modification of IA activation and inactivation properties, indicating that the mechanism of action of CORT involved long-term effects. Therefore, we explored its upstream signaling mechanism to explain this phenomenon.

The dysregulation of PI3K/Akt pathway is implicated in a number of human diseases including cancer, diabetes, cardiovascular disease and neurological diseases. Insulin enhanced IR phosphorylation and activated potassium channels (Yasui et al. 2008). Chronic CORT inactivated the IR-mediated pathway, including the Akt/GSK-3β pathway, and resulted in the cognitive deficits (Solas et al. 2013). In the present study, we observed that insulin inhibited transient potassium currents via the activation of PI3K-mediated signaling pathway, and CORT possibly alleviated insulin-induced inhibition of transient potassium currents by affecting the activation of PI3K/Akt signaling pathway. However, CORT could not inhibit the activation of downstream signaling effectors of PI3K signaling in hippocampal neurons (Fig. 5). It confirmed that Akt was the important intermediate medium of multiple signaling pathways. A pathway-like picture was to describe the results (Fig. 6).

Fig. 6
figure 6

A pathway-like figure of conclusion description. Insulin could cause changes of AP, K-currents and protein levels. However, all these effects were abolished with CORT pretreatment. The mechanism would rely on the PI3K-mediated insulin-signaling pathway. Cort = corticosterone; ins = insulin; IR = insulin receptor

Full size image

The connection between stress and depression was drawn from observations of elevated cortisol levels in depressed patients (Gold et al. 2015; Ulloa et al. 2010), which was one of the most robust findings in biological psychiatry (Scott and Dinan 1998). Cognitive impairments in hippocampal function was related to cortisol levels in depressed patients (Brown et al. 2004). The acute increase of glucocorticoid could effectively induce depression (Xiao et al. 2017). In this study, we found that CORT regulated the excitability of hippocampal neurons and the sensitivity of potassium channels through insulin-signaling pathway. It implied a possible mechanism of depression induced by elevated CORT. However, the deeper mechanism of potassium channel activity in depression process remained unknown. Other ion channels may have interferences in potassium channels, such as sodium channels. Due to all these possibilities, further studies need to explore the deeper mechanism of depression, ion channels and insulin-signaling pathway.

Conclusions

Corticosterone induces the neuronal excitotoxicity and possibly alleviates insulin-induced inhibition of transient potassium currents by affecting the activation of PI3K/Akt signaling pathway. Insulin activates insulin-signaling pathway, and the downstream effectors involves potassium channels. Therefore, insulin-signaling pathway is a possible pathway that CORT leads to insulin resistant and depression. Thus, the CORT → insulin resistant → PI3K/Akt → IA axis may play an important role in neuronal development and depression treatment.