Main text

An acute impairment in the biological response to insulin in the brain, known as central insulin resistance, was identified during stroke [1] and traumatic brain injury [2], for which excitotoxicity, which is caused by excessive glutamate release, is a key pathogenic factor [3].

The exact molecular link between insulin resistance and glutamate excitotoxicity remains unclear. Based on published data, an abnormal rise in intracellular free Ca2+ concentration ([Ca2+]i) and decreased mitochondrial inner membrane potential (ΔΨm) are factors associated with the excitotoxic glutamate [4,5,6], which could potentially affect insulin signalling. The presence of Ca2+ (1 mM) has been shown to reduce the insulin-induced tyrosine phosphorylation of the insulin receptor (IR) in hippocampal synaptic preparations [7]. Glutamate has been shown to reduce the tyrosine phosphorylation of the IR when added after the prolonged insulin-mediated stimulation of hippocampal neuronal cultures [8]. Protonophore-induced decreases in ΔΨm have been shown to evoke concomitant decreases in the IR tyrosine phosphorylation in response to insulin, indicating that mitochondrial depolarisation is an independent causative factor for neuronal insulin resistance [9]. However, the effects of glutamate-evoked changes in [Ca2+]i and ΔΨm on the insulin-induced activation of the IR/Akt/mammalian target of rapamycin (mTOR) and glycogen synthase kinase (GSK)3β pathways have never been studied. Here, we investigated whether excitotoxic glutamate affects the insulin-induced phosphorylation of IR/Akt/mTOR pathway components during significant mitochondrial depolarisations caused by the massive influx of Ca2+ into the cells.

To determine the times during which glutamate induces significant mitochondrial depolarisation, rat cortical neurons were exposed to 100 μM glutamate and changes in [Ca2+]i, and ΔΨm were monitored for 30 min. Detailed methods for the preparation of primary cortical neuronal cultures and measuring [Ca2+]i and ΔΨm are described in Additional file 1. As expected, glutamate evoked a rapid increase in [Ca2+]i, followed by a decreased in ΔΨm in the cells (Fig. 1a and b, Additional file 2: Tables S3 and S4). After 30 min of glutamate exposure, the mean [Ca2+]i level increased by 2.4-fold above baseline (F148,8732 = 44.8, P < 0.0001), and the mean ΔΨm value significantly decreased by 1.6-fold below baseline (F148,8732 = 182.8, P < 0.0001) (Fig. 1a–c, one-way analysis of variance [ANOVA] with repeated measures, followed by Tukey’s post hoc test). Therefore, the 30-min interval for glutamate exposure was selected for subsequent experiments.

Fig. 1
figure 1

Excitotoxic glutamate inhibits IR/Akt/mTOR pathway

a Dynamics of [Ca2+]i and b ΔΨm in single rat cortical neurons, loaded simultaneously with Fura-FF and Rh123 dyes and exposed to 100 μM glutamate. Grey lines represent sixty single neurons. Blue and red lines represent the respective means of [Ca2+]i and ΔΨm, averaged across sixty individual neurons at every time point. c Fura-FF and Rh123 fluorescence 30 min after the onset of glutamate exposure, expressed as the fold increase over baseline (at − 5 min). Data are the mean ± SEM from sixty neurons. ****P < 0.0001 compared to respective baselines (one-way ANOVA with repeated measures, followed by Tukey’s post hoc test). d-g Levels of d IRβ pY1150/1151, e Akt pS473, f mTOR pS2448, and g GSK3β pS9 in rat cortical neurons exposed to 0 nM (C) or 100 nM insulin for 15 min (I), 100 μM glutamate for 30 min g, or sequentially to 100 μM glutamate for 30 min and 100 nM insulin for 15 min (G + I). Bar graphs represent the levels of the phosphoproteins, normalised against respective total protein levels, in cell lysates and expressed as a percentage of levels in insulin-treated cells (group I). Each value represents the mean ± SEM from six independent cultures (cell populations obtained from twelve separate rats, two per culture). ###P < 0.001, ####P < 0.0001 compared with untreated control c; **P < 0.01, ***P < 0.001, ****P < 0.0001 compared with insulin i (one-way ANOVA, followed by Tukey’s post hoc test for multiple comparisons). h Scheme illustrating the inhibitory effects of glutamate on the insulin-induced activation of the IR/Akt/mTOR pathway.

Full size image

Next, we investigated the effects of glutamate on the insulin-induced activation of signalling molecules in the IR/Akt/mTOR and GSK3β pathways. Rat cortical neuronal cultures were sequentially exposed to either 100 μM excitotoxic glutamate concentration or no glutamate for 30 min, then stimulated with 100 nM insulin for 15 min and lysed. The levels of the tyrosine-phosphorylated IR β-subunit (IRβ pY1150/1151), the serine-phosphorylated Ser/Thr kinase Akt (Akt pS473), the serine-phosphorylated mammalian target of rapamycin (mTOR pS2448), and the serine-phosphorylated glycogen synthase kinase 3β (GSK3β pS9) were measured in the cell lysates and normalised against the total levels of these proteins, as outlined in Additional file 1. A one-way ANOVA analysis revealed a significant difference between groups for IRβ pY1150/1151 (Fig. 1d, F3,20 = 23.62, P < 0.0001), Akt pS473 (Fig. 1e, F3,20 = 388.6, P < 0.0001), mTOR pS2448 (Fig. 1f, F3,20 = 19.69, P < 0.0001), and GSK3β pS9 (Fig. 1g, F3,20 = 4.48, P = 0.0146). Tukey’s multiple comparisons test showed that insulin stimulation resulted in a significant increase in IRβ pY1150/1151 (Fig. 1d, P < 0.0001), Akt pS473 (Fig. 1e, P < 0.0001), and mTOR pS2448 (Fig. 1f, P < 0.001) levels compared with their respective controls (Additional file 1: Tables S1 and S2). In the glutamate pre-treated neurons, insulin induced 48% less IR phosphorylation at pY1150/1151 (Fig. 1d, P < 0.001), 72% less Akt phosphorylation at pS473 (Fig. 1e, P < 0.0001), 44% less mTOR phosphorylation at pS2448 (Fig. 1f, P < 0.0001), and 38% less GSK3β phosphorylation at pS9 (Fig. 1g, P < 0.01) compared with non-glutamate treated controls. Neuronal IR becomes fully active following autophosphorylation at Y1150/1151 [10], which triggers downstream signalling via the Akt/mTOR and GSK3β pathways. The serine-phosphorylation of Akt at S473 [11] and mTOR at S2448 [12] are crucial for their activation, whereas the serine-phosphorylation of GSK3β at S9 [13] leads to its inhibition. Therefore, our results suggested that excitotoxic glutamate inhibits insulin-induced IR activation and the downstream IR/Akt/mTOR signalling pathway (Fig. 1h).

The primary finding of the present study was that excitotoxic glutamate inhibits the IR/Akt/mTOR pathway, resulting in the development of acute neuronal insulin resistance during periods of significant mitochondrial depolarisation caused by glutamate-evoked massive influxes of Ca2+. This rapid loss of neuronal insulin sensitivity appears to be one of the earliest pathological events associated with glutamate excitotoxicity. These results are in complete agreement with our previous findings that mitochondria control IR autophosphorylation in neurons and that mitochondrial depolarisation causes the loss of insulin sensitivity during the IR autophosphorylation stage [9, 14]. Recently we showed that pre-treatment with insulin prevents the glutamate-evoked increases in [Ca2+]i and decreases in ΔΨm, protecting rat cortical neurons against excitotoxicity [15]. The glutamate effect and the protective effects of insulin were both completely abrogated by MK 801, an inhibitor of Ca2+ influx, via the N-methyl-D-aspartate (NMDA) receptor and the plasmalemmal Na+/Ca2+ exchanger operating in reverse mode [15]. Collectively, these findings suggested that the modulation of intracellular Ca2+ levels plays a critical role in negative cross-talk between insulin and glutamate signalling during excitotoxicity. Glutamate induces an increase in [Ca2+]i and a decrease in ΔΨm, which inhibit IR activation. In turn, insulin prevents the glutamate-evoked rise in [Ca2+]i and mitochondrial depolarisation, protecting against excitotoxicity.

In conclusion, this study showed that glutamate excitotoxicity is causative for central insulin resistance and may induce the acute loss of insulin signalling within minutes under mitochondrial depolarisation conditions. Therefore, the use of agents designed to prevent mitochondrial depolarisation may be a reasonable approach to the treatment of acute neuronal insulin resistance during excitotoxicity.