Expression of 4E-BP1 in juvenile mice alleviates mTOR-induced neuronal dysfunction and epilepsy

Abstract

Hyperactivation of the mTOR pathway during foetal neurodevelopment alters neuron structure and function, leading to focal malformation of cortical development and intractable epilepsy. Recent evidence suggests a role for dysregulated cap-dependent translation downstream of mTOR signalling in the formation of focal malformation of cortical development and seizures. However, it is unknown whether modifying translation once the developmental pathologies are established can reverse neuronal abnormalities and seizures. Addressing these issues is crucial with regards to therapeutics because these neurodevelopmental disorders are predominantly diagnosed during childhood, when patients present with symptoms. Here, we report increased phosphorylation of the mTOR effector and translational repressor, 4E-BP1, in patient focal malformation of cortical development tissue and in a mouse model of focal malformation of cortical development. Using temporally regulated conditional gene expression systems, we found that expression of a constitutively active form of 4E-BP1 that resists phosphorylation by focal malformation of cortical development in juvenile mice reduced neuronal cytomegaly and corrected several neuronal electrophysiological alterations, including depolarized resting membrane potential, irregular firing pattern and aberrant expression of HCN4 ion channels. Further, 4E-BP1 expression in juvenile focal malformation of cortical development mice after epilepsy onset resulted in improved cortical spectral activity and decreased spontaneous seizure frequency in adults. Overall, our study uncovered a remarkable plasticity of the juvenile brain that facilitates novel therapeutic opportunities to treat focal malformation of cortical development-related epilepsy during childhood with potentially long-lasting effects in adults.

Keywords: in utero electroporation; cap-dependent translation; hyperpolarization-activated cyclic nucleotide-gated (HCN) channels; malformation of cortical development; seizures.

Figure 1

4E-BP1 is hyperphosphorylated in human FMCD tissue and in the Rheb^CA mouse model of FMCD. (A) Diagram of the PI3K-mTORC1 pathway and downstream regulation of cap-dependent translation via 4E-BP1/2. Activated mTORC1 phosphorylates and inhibits 4E-BP1/2, which disinhibits the eIF4F complex (via release of eIF4E) and promotes cap-dependent translation. Conversely, inactivated mTORC1 disinhibits 4E-BP1/2, which inhibits 4E-BP1/2 and blocks cap-dependent translation. (B) Representative images of DAPI (blue), SMI-311 (green) and p-4E-BP1 (red) staining in resected brain tissue samples from three individuals with TSC and one individual with FCDII who underwent epilepsy surgery. For Individual TSC #1, low-magnification images are shown on the left and high-magnification images are shown on the right. White squares denote the magnified areas. For all images, filled arrows point to SMI-311+ dysmorphic neurons with high p-4E-BP1 immunoreactivity. Unfilled arrows point to surrounding SMI-311− cells with low or no p-4E-BP1 immunoreactivity. Note that DAPI-stained nuclei in SMI-311+ cells appear fainter and are often enlarged compared to SMI-311− cells. Scale bars = 25 μm. (C) Diagram of IUE and plasmids used to generate Rheb^CA mice with FMCD. Mouse embryos were electroporated with a BFP (control) or Rheb^CA plasmid at E15.5, targeting radial glia generating pyramidal neurons destined to L2/3 in the mPFC. tdTomato was co-electroporated in both conditions to label the targeted neurons. (D) Image of a P0 pup head showing tdTomato fluorescence in the targeted region following electroporation at E15.5. (E) Diagram showing the targeted region (red) in an adult mouse brain and the corresponding area in coronal view. (F) Representative images of tdTomato+ cells (red) and p-4E-BP1 staining (green, pseudocoloured) in coronal cortical sections from P30 BFP control and Rheb^CA mice. Low-magnification tile scan images of tdTomato+ cells are shown on the left. tdTomato+ cells are strictly found in L2/3 in BFP control mice and misplaced across the cortical layers in Rheb^CA mice. High-magnification images of tdTomato+ cells and p-4E-BP1 staining are shown on the right. tdTomato+ cells display basal level p-4E-BP1 immunoreactivity in BFP control mice and high p-4E-BP1 immunoreactivity in Rheb^CA mice. White squares denote the magnified areas. Scale bars = 1000 μm (left), 25 μm (right). (G) Quantification of tdTomato+ cell soma size. n = 5 BFP, n = 6 Rheb^CA mice; each data-point represents averaged values from 22 to 30 cells per animal. Data were analysed by unpaired t-test; ****P < 0.0001. (H) Quantification of p-4E-BP1 intensity in tdTomato+ cells. n = 5 BFP, n = 6 Rheb^CA mice; each data-point represents averaged values from 22 to 30 cells per animal. Data were normalized to the mean control (BFP) and analysed by unpaired t-test; *P = 0.0186. Error bars are ± SEM.

Figure 2

Juvenile expression of c4E-BP1^CA restores the excitability of Rheb^CA neurons. (A) Diagram of plasmids used to generate BFP+cGFP (control), Rheb^CA+cGFP and Rheb^CA+c4E-BP1^CA mice and experimental strategy. Mouse embryos were electroporated at E15.5. For each litter, 1/3 of the embryos received BFP+cGFP, 1/3 received Rheb^CA+cGFP and 1/3 received Rheb^CA+c4E-BP1^CA plasmids. A tamoxifen-inducible Cre (ER^T2-Cre-ER^T2) plasmid and a tdTomato reporter plasmid were co-electroporated in all three conditions. Tamoxifen was administered to all mice from P12 to P16 to induce GFP or 4E-BP1^CA expression. Whole-cell patch clamp recordings were performed in L2/3 pyramidal neurons in acute coronal slices from P23–32 mice. (B) Representative images of tdTomato+ (electroporated) neurons from brain slices used in patch clamp recording. Images are maximal intensity projections of 25-µm thick z-stack sections. Scale bars = 20 μm. (C–E) Bar graphs of (C) membrane capacitance, (D) resting membrane conductance and (E) RMP. (F) Representative current traces in response to a 1-s long conditioning step to −40 mV from a holding potential of −70 mV, followed by a series of 3-s long hyperpolarizing voltage steps from −130 mV to −40 mV in 10-mV increments. Traces at the −40 mV conditioning step are not shown due to overlapping traces from unclamped sodium spikes. (G) IV curve obtained from I_ss amplitudes. (H) IV curve obtained from I_h amplitudes or ΔI, where ΔI = I_ss − I_inst. (I) Quantification of I_h amplitude at −90 mV. (J) Scatterplot of I_h amplitudes at −90 mV versus RMP. n = 47 XY pairs. P = 0.0021 by Pearson product–moment correlation. (K) Representative voltage traces in response to 1-s long hyperpolarizing current steps from −500 pA to 0 pA in 100-pA increments from RMP. Arrow points I_h-associated voltage sags induced by hyperpolarizing currents. (L) Representative voltage traces in response to −500 pA current steps. Traces were rescaled and superimposed post-recording to visualize the differences in voltage sag size between groups. (M) Quantification of voltage sag ratio at −500 pA, where sag ratio = (V_peak − V_ss)/V_peak × 100. (N) Representative traces of AP firing response to depolarizing current injections. (O) Input–output curve showing the mean number of APs fired in response to 500-ms long depolarizing current steps from 0 to 700 pA in 50-pA increments. For all graphs, n = 10–14 BFP+cGFP, n = 13–16 Rheb^CA+cGFP, n = 17–18 Rheb^CA+c4E-BP1^CA neurons. Data were analysed using (C–E, I and M) one-way ANOVA with Tukey’s post hoc test; *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001 or (G, H and O) mixed-effects ANOVA with Tukey’s post hoc test; *P < 0.05, **P < 0.01, ***P < 0.001 (versus BFP+cGFP), ^#P < 0.05 (versus Rheb^CA+cGFP). Error bars are ± SEM.

Figure 3

c4E-BP1^CA expression in juvenile Rheb^CA mice reduces neuronal cytomegaly and aberrant HCN4 channel expression. (A) Diagram of plasmids used to generate Rheb^CA+cGFP and Rheb^CA+c4E-BP1^CA mice and experimental strategy. Mouse embryos were electroporated at E15.5. For each litter, half of the embryos received Rheb^CA+cGFP and the other half received Rheb^CA+c4E-BP1^CA. A tamoxifen-inducible Cre (ER^T2-Cre-ER^T2) plasmid and a tdTomato reporter plasmid were co-electroporated in both conditions. Tamoxifen was administered from P28 to P32 to induce GFP or 4E-BP1^CA expression. (B) Representative images of tdTomato+ cells in coronal sections from P84–106 Rheb^CA+cGFP and Rheb^CA+c4E-BP1^CA mice. Low-magnification tile scan images showing tdTomato+ cell distribution across the cortical layers are on the left. High-magnification images showing differences in tdTomato+ cell size are on the right. Scale bars = 500 μm (left), 50 μm (right). (C) Quantification of tdTomato+ cell placement in L2/3. n = 11 Rheb^CA+cGFP, 18 Rheb^CA+c4E-BP1^CA mice; each data-point represents averaged values from three brain sections per animal. Data were analysed by unpaired t-test. (D) Quantification of tdTomato+ cell soma size. n = 11 Rheb^CA+cGFP, n = 18 Rheb^CA+c4E-BP1^CA mice; each data-point represents averaged values from 50 cells per animal. Data were analysed by unpaired t-test; ****P < 0.0001. For graphs C and D, previously reported levels for age-matched control mice electroporated with GPF only (Nguyen et al.) are shown for comparison. (E) Representative images of tdTomato+ cells (red) and HCN4 staining (green, pseudocoloured) in coronal sections from P84–106 Rheb^CA+cGFP and Rheb^CA+c4E-BP1^CA mice. Low-magnification tile scan images showing HCN4 staining in whole-brain sections are on the left. High-magnification images showing somatic HCN4 staining in Rheb^CA+cGFP neurons and lack of HCN4 staining in Rheb^CA+c4E-BP1^CA neurons are on the right. White squares denote the areas targeted by IUE. Arrows point to HCN4 staining in the medial septum. Scale bars = 500 μm (left), 25 μm (right). (F) Quantification of HCN4 staining intensity. n = 11 Rheb^CA+cGFP, n = 17 Rheb^CA+c4E-BP1^CA mice; each data-point represents averaged values from two brain sections per animal. Data were normalized to the mean control (Rheb^CA+cGFP contralateral cortex) and analysed using two-way repeated measures ANOVA with Bonferroni’s post hoc test; **P < 0.01, ****P < 0.0001. Error bars are ± SEM. Ipsi ctx = ipsilateral cortex; contra ctx = contralateral cortex.

Figure 4

c4E-BP1^CA expression after epilepsy onset in juvenile Rheb^CA mice normalizes cortical spectral activity. (A) Representative spectrograms of background EEG during the light (12–2 p.m., top) and dark (12–2 a.m., bottom) cycles. (B and C) Relative EEG power spectra of background EEG during the (B) light and (C) dark cycles. Insets show enlarged graphs from the areas denoted by the red boxes. (D) Sample cortical EEG traces showing the delta, theta, alpha, beta and gamma frequency bands. The bands were decomposed from the composite EEG signal shown at the bottom. (E and F) Bar graphs of the delta, theta, alpha, beta and gamma relative bandpower during the (E) light and (F) dark cycles. For all graphs, n = 6 GFP, n = 11 Rheb^CA+cGFP, n = 18 Rheb^CA+c4E-BP1^CA mice; each data-point represents averaged values from 3 to 10 cycles (epochs) per animal. Data were analysed using (B and C) two-way repeated measures ANOVA with Tukey’s post hoc test; *P < 0.05 (versus GFP), ^#P < 0.05 (versus Rheb^CA+cGFP) or (E and F) one-way ANOVA with Tukey’s post hoc test; *P < 0.05. Error bars are ± SEM.

Figure 5

c4E-BP1^CA expression after epilepsy onset in juvenile Rheb^CA mice alleviates seizures. (A) Sample cortical EEG trace showing a typical seizure in Rheb^CA mice. Expanded EEG traces from representative preictal, ictal and postictal periods are shown at the bottom. (B) Heat map showing the daily number of seizures of individual animals over seven consecutive days of vEEG recording. (C) Bar graph showing the proportion of animals in each group with no seizures, <1 seizure/day and >1 seizure/day. n = 11 Rheb^CA+cGFP, n = 18 Rheb^CA+c4E-BP1^CA mice. (D) Quantification of seizure frequency. n = 11 Rheb^CA+cGFP, n = 18 Rheb^CA+c4E-BP1^CA mice; each data-point represents mean seizures/day over a 7-day recording period per animal. Data were analysed using Mann–Whitney U-test; *P = 0.0206. (E–G) Scatterplots of (E) seizure frequency versus cell placement in L2/3, (F) seizure frequency versus cell size and (G) seizure frequency versus HCN4 staining intensity. For plots (F and G), the areas denoted with the grey boxes are enlarged below to reveal overlapping values. n = 29 (E and F), n = 28 (G) XY pairs. Data were analysed with Spearman’s rank-order correlation. Error bars are ± SEM.