Abstract
Brain somatic mutations in various components of the mTOR complex 1 (mTORC1) pathway have emerged as major causes of focal malformations of cortical development and intractable epilepsy. While these distinct gene mutations converge on excessive mTORC1 signaling and lead to common clinical manifestations, it remains unclear whether they cause similar cellular and synaptic disruptions underlying cortical network hyperexcitability. Here, we show that in utero activation of the mTORC1 activators, Rheb or mTOR, or biallelic inactivation of the mTORC1 repressors, Depdc5, Tsc1, or Pten in mouse medial prefrontal cortex leads to shared alterations in pyramidal neuron morphology, positioning, and membrane excitability but different changes in excitatory synaptic transmission. Our findings suggest that, despite converging on mTORC1 signaling, mutations in different mTORC1 pathway genes differentially impact cortical excitatory synaptic activity, which may confer gene-specific mechanisms of hyperexcitability and responses to therapeutic intervention.
Introduction
Focal malformations of cortical development (FMCDs), including focal cortical dysplasia type II (FCDII) and hemimegalencephaly (HME), are the most common causes of intractable epilepsy in children1,2. These disorders are characterized by abnormal brain cytoarchitecture with cortical overgrowth, mislamination, and cellular anomalies, and are often associated with developmental delay and intellectual disability3. Early immunohistochemical studies identified hyperactivation of the mechanistic target of rapamycin complex 1 (mTORC1) signaling pathway in resected brain tissue from individuals with FCDII and HME, leading to the classification of these disorders as “mTORopathies”4. More recently, somatic mutations in numerous regulators of the mTORC1 pathway were identified as the genetic causes of FCDII and HME5–7. Accumulating evidence shows that these mutations occur in a subset of dorsal telencephalic progenitor cells that give rise to excitatory neurons during fetal development, resulting in brain somatic mosaicism8,9. These genetic findings provide opportunities for gene therapy approaches targeting the mTORC1 pathway for FCDII and HME.
mTORC1 is an evolutionarily conserved protein complex that promotes cell growth and differentiation through the regulation of protein synthesis, metabolism, and autophagy10. mTORC1 is composed of mTOR, a serine/threonine kinase that exerts the complex’s catalytic function, and several companion proteins. Activation of mTORC1 is controlled by a well-described upstream cascade involving multiple protein regulators11. Stimulation by growth factors activates phosphoinositide 3-kinase (PI3K), which leads to the activation of phosphoinositide-dependent kinase 1 (PDK1) and subsequent phosphorylation and activation of AKT. Activated AKT phosphorylates and inactivates tuberous sclerosis complex 1/2 (TSC1/2 complex; consisting of TSC1 and TSC2 proteins), which releases the brake on Ras homolog enriched in brain (RHEB), a GTP-binding protein that directly activates mTORC1. This mTORC1-activating cascade is negatively controlled by the phosphatase and tensin homolog (PTEN) protein, which inhibits PI3K activation of PDK1. Additionally, mTORC1 signaling is regulated by a separate nutrient-sensing GAP Activity Towards Rags 1 (GATOR1) complex, which consists of DEP domain containing 5 (DEPDC5) and the nitrogen permease regulator 2-like (NPRL2) and 3-like (NPRL3) proteins12. The GATOR1 complex serves as a negative regulator of mTORC1 that inhibits mTORC1 at low amino acid levels.
Pathogenic mutations in numerous regulators that activate mTORC1 signaling, including PIK3CA, PTEN, AKT3, TSC1, TSC2, MTOR, RHEB, DEPDC5, NPRL2, and NPRL3, have been identified in HME and FCDII9. Genetic targeting of these genes in mouse models consistently recapitulates the epilepsy phenotype, supporting an important role for these genes in seizure generation13,14. However, while all these genes impinge on the mTORC1 pathway, many of them also participate in mTORC1-independent functions through parallel signaling pathways13,15–19, and it is unclear whether mutations affecting different mTORC1 pathway genes lead to cortical hyperexcitability through common neural mechanisms. In this study, we examined how disruption of five distinct mTORC1 pathway genes, Rheb, mTOR, Depdc5, Pten, and Tsc1, individually impact pyramidal neuron development and electrophysiological function in the mouse medial prefrontal cortex (mPFC). Collectively, we found that activation of the mTORC1 activators, Rheb and mTOR, or inactivation of the mTORC1 repressors, Depdc5, Tsc1, and Pten, largely leads to similar alterations in neuron morphology and membrane excitability but differentially impacts excitatory synaptic activity. The latter has implications for cortical network function and seizure vulnerability and may affect how individuals with different genotypes respond to targeted therapeutics.
Results
Expression of RhebY35L, mTORS2215Y, Dedpdc5KO, PtenKO, and Tsc1KOleads to varying magnitudes of neuronal enlargement and mispositioning in the cortex
To model somatic gain-of-function mutations in the mTORC1 activators, we individually expressed plasmids encoding RhebY35L or mTORS2215Y, respectively, in select mouse neural progenitor cells during late corticogenesis via in utero electroporation (IUE) (Fig. 1a, b). These pathogenic variants were previously detected in brain tissue from children with FCDII and HME associated with intractable seizures20–25. We specifically targeted a subset of late-born progenitor cells that generate excitatory pyramidal neurons destined to layer 2/3 in the medial prefrontal cortex (mPFC) to mimic frontal lobe somatic mutations and the genetic mosaicism characteristic of these lesions. To model the somatic loss-of-function mutations in mTORC1 repressors, we expressed plasmids encoding previously validated CRISPR/Cas9 guide RNAs against Depdc526, Tsc127, or Pten28 to individually knockout (KO) the respective genes using the same IUE approach (Fig. 1a, b). As a control for the activating mutations, we expressed plasmids encoding fluorescent proteins under the same CAG promoter. As a control for the CRISPR/Cas9-mediated knockouts, we used an empty CRISPR/Cas9 vector containing no guide RNA sequences. To verify that the expression of these plasmids leads to mTORC1 hyperactivation, we assessed the phosphorylated levels of S6 (i.e., p-S6), a downstream substrate of mTORC1, using immunostaining in brain sections from postnatal day (P) 28-43 mice. As expected, we found that expression of the RhebY35L, mTORS2215Y, Dedpdc5KO, PtenKO, and Tsc1KO plasmids all led to significantly increased p-S6 staining intensity, supporting that the expression of each of these plasmids leads to increased mTORC1 signaling (Fig. 1c, d, Table 1, Fig. S1a).
Considering the cytoarchitectural abnormalities associated with mTORC1 hyperactivation, we compared neuron soma size and positioning in the cortex following in utero activation of Rheb and mTOR and inactivation of Depdc5, Pten, and Tsc1 in P28-43 mice. While all experimental conditions led to increased neuron soma size, the magnitude of the enlargement was dependent on the specific gene that was targeted (Fig. 1c, e, Table 1. Fig. S1b). In particular, expression of RhebY35L and mTORS2215Y led to the largest soma size changes, with a >3-fold increase from control. Expression of PtenKO led to a similarly large increase of 2.7-fold, while expression of Dedpdc5KOand Tsc1KO led to a 1.7 and 2.1-fold increase, respectively. The increase in the PtenKO condition was significantly higher than both the Dedpdc5KO and Tsc1KO conditions, and the increase in the Tsc1KO condition was significantly higher than the Dedpdc5KO condition. Although the above analysis was performed at P28-43, the enlargement of neuron soma sizes was already detected by P7-9 (Fig. S2a-c, Table S1). In terms of neuronal positioning, all experimental conditions, except for PtenKO, resulted in neuron misplacement (Fig. 1f-h, Table 1). The RhebY35L and mTORS2215Yconditions led to the most severe phenotype: ∼70-80% of the neurons were misplaced outside of layer 2/3 (Fig. 1g), with the mispositioned neurons evenly scattered across the deeper layers (Fig. 1h). For the Dedpdc5KO and Tsc1KO conditions, ∼45-60% of the neurons were misplaced outside of layer 2/3 (Fig. 1g), with the mispositioned neurons found closer to layer 2/3 (Fig. 1h). Taken together, these studies show that the expression of RhebY35L, mTORS2215Y, Dedpdc5KO, PtenKO, and Tsc1KO leads to varying magnitudes of neuronal enlargement and mispositioning in the cortex. Of note, PtenKO neurons, despite exhibiting a 2.7-fold increase in neuron soma size, were mostly correctly positioned in layer 2/3. These findings suggest that while all experimental conditions lead to increased soma size, not all lead to neuron mispositioning, suggesting defective migration and the subsequent impact on neuron positioning occur independently of cell size.
Expression of RhebY35L, mTORS2215Y, Dedpdc5KO, PtenKO, and Tsc1KOuniversally leads to decreased depolarization-induced excitability, but only RhebY35L, mTORS2215Y, and Tsc1KOexpression leads to depolarized resting membrane potential
To elucidate the contribution of each experimental condition to the function of cortical neurons, we obtained whole-cell patch clamp recordings of layer 2/3 pyramidal neurons at P26-P51 (Fig. 2a). The RhebY35Land mTORS2215Yconditions were compared to a control group expressing fluorescent proteins under the same CAG promoter. The Dedpdc5KO, PtenKO, and Tsc1KO conditions were compared to a CRISPR/Cas9 empty vector control. Recordings of control and experimental conditions were alternated to match the animal ages.
Consistent with the changes in soma size (Fig. 1c, e), recorded neurons displayed increased membrane capacitance in all experimental conditions (Fig. 2b, Table 1, Fig. S3a). Neurons expressing the mTORS2215Yvariant had a larger capacitance increase than those expressing the RhebY35L variant. PtenKO and Tsc1KO neurons had similar increases in capacitance that were both larger than that of the Depdc5KO neurons. All neurons across the experimental conditions also had increased resting membrane conductance in a pattern that followed that of the capacitance (Fig. 2c, Table 1, Fig. S3b). However, while RhebY35L, mTORS2215Y, and Tsc1KOexpression led to depolarized resting membrane potential (RMP), Dedpdc5KOand PtenKO expression did not significantly affect the RMP (Fig. 2d, Table 1, Fig. S3c). To assess whether these changes impacted neuron intrinsic excitability, we examined the action potential (AP) firing response to depolarizing current injections. For all experimental conditions, neurons fired fewer APs for current injections above 100 pA compared to the respective control neurons (Fig. 2e, f, Table 1). This decrease in intrinsic excitability is reflected in the increased rheobase (i.e., the minimum current required to induce an AP) in all experimental conditions, with the mTORS2215Y and RhebY35Lconditions leading to the largest rheobase increases (Fig. 2g, Table 1, Fig. S3d). Collectively, these findings indicate that RhebY35L, mTORS2215Y, and Tsc1KO neurons display a decreased ability to generate APs upon depolarization despite having depolarized RMPs. In terms of firing pattern, neurons in all groups displayed a regular-spiking pattern with spike-frequency adaptation (Fig. 2e). However, while an initial spike doublet (with an interspike interval of 10-15 ms) was observed in the majority of control neurons, consistent with the expected firing pattern for layer 2/3 mPFC pyramidal neurons29, this was observed in a smaller fraction of neurons in the experimental conditions [number of neurons displaying spike doublets: RhebY35L: 4/23 (17%), mTORS2215Y: 7/28 (25%), Dedpdc5KO: 11/27 (41%), PtenKO: 21/28 (75%) - notably, 6 of the 21 doublets are parts of triplets, Tsc1KO: 7/28 (25%), control: 24/26 (92%)]. There were no differences in the AP threshold across the conditions (Fig. 2h, Table 1, Fig. S3e). The AP peak amplitude was decreased in the mTORS2215Yand Tsc1KO conditions (Fig. 2i, Table 1, Fig. S3f), while the AP half-width was decreased in the RhebY35L, mTORS2215Y, and PtenKO conditions (Fig. 2j, Table 1, Fig. S3g). Taken together, these findings show that various genetic conditions that activate the mTORC1 pathway universally lead to decreased depolarization-induced excitability in layer 2/3 pyramidal neurons, with gene-dependent changes in RMP and several AP properties.
Expression of RhebY35L, mTORS2215Y, Dedpdc5KO, PtenKO, and Tsc1KOleads to the abnormal presence of HCN4 channels with variations in functional expression
We recently reported that neurons expressing RhebS16H, an mTOR activating variant of Rheb, display abnormal expression of HCN4 channels30,31. These channels give rise to a hyperpolarization-activated cation current (Ih) that is normally absent in layer 2/3 pyramidal neurons30,31. The aberrant Ih, which has implications for neuronal excitability, preceded seizure onset and was detected by P8-12 in mice30. Rapamycin treatment starting at P1 and expression of constitutive active 4E-BP1, a translational repressor downstream of mTORC1, prevented and alleviated the aberrant HCN4 channel expression, respectively30,31. These findings suggest that the abnormal HCN4 channel expression is mTORC1-dependent. Given that all the experimental conditions in the present study led to increased mTORC1 activity, we investigated whether they also resulted in abnormal HCN4 channel expression. Immunostaining for HCN4 channels using previously validated antibodies30 in brain sections from P28-43 mice revealed significantly increased HCN4 staining intensity in the electroporated neurons in all experimental conditions compared to the respective controls, which exhibited no HCN4 staining (Fig 3a, b, Table 1, Fig. S4a). The increased staining was evident in the ipsilateral cortex containing mTORS2215Yelectroporated neurons and absent in the non-electroporated contralateral cortex (Fig. S5a). These data indicate the presence of aberrant HCN4 channel expression following RhebY35L, mTORS2215Y, Depdc5KO, PtenKO, or Tsc1KO expression in layer 2/3 pyramidal neurons.
To examine the functional impacts of the aberrant HCN4 channel expression, we examined Ih amplitudes in the various experimental conditions. To evoke Ih, we applied a series of 3 s-long hyperpolarizing voltage steps from -120 mV to -40 mV. Consistent with previous findings in RhebS16H neurons30,31, hyperpolarizing voltage pulses elicited significantly larger inward currents in all experimental conditions compared to their respective controls (Fig. 3c, d, Table 1). The mTORS2215Y condition displayed larger inward currents than the RhebY35L condition, while the PtenKO condition displayed the largest inward currents compared to the Depdc5KO and Tsc1KO conditions (Fig. 3d). These data were proportional to the changes in neuron soma size (Fig. 1e, 2b). The inward currents in RhebS16H neurons are thought to result from the activation of both inward-rectifier K+ (Kir) channels and HCN channels30. Kir-mediated currents activate fast whereas HCN-mediated currents, i.e., Ih, activate slowly during hyperpolarizing steps; therefore, to assess Ih amplitudes, we measured the difference between the instantaneous and steady-state currents at the beginning and end of the voltage pulses, respectively (i.e., ΔI)32. The resulting ΔI-voltage (V) curve revealed significantly larger Ih amplitudes in all experimental conditions (Fig. 3e, Table 1). To further isolate the Ih from Kir-mediated currents, we measured Ih amplitudes at -90 mV, near the reversal potential of Kir channels to eliminate Kir-mediated current contamination. At -90 mV, Ih amplitudes were significantly higher in the mTORS2215Y and Tsc1KO conditions compared to controls (Fig. 3f, Table 1, Fig. S4b). Of note, although the Depdc5KO and PtenKO conditions did not display a significant increase in Ih amplitudes at -90 mV, 1 out of 28 Depdc5KO neurons and 4 out of 27 PtenKO neurons had Ih amplitudes that were 2-fold greater than the mean Ih amplitude of the Tsc1KO condition. 6 out of 24 RhebY35L neurons also had Ih amplitudes 2-fold greater than this value (Fig. 3f). These data suggest that Ih currents are present in a subset of RhebY35L, Depdc5KO, and PtenKO neurons, and most mTORS2215Y and Tsc1KO neurons.
Considering that the mTORS2215Y condition led to the largest Ih, we examined the impact of the selective HCN channel blocker zatebradine on hyperpolarization-induced inward currents in mTORS2215Y neurons to further confirm the identity of ΔI as Ih. Application of 40 µM zatebradine reduced the overall inward currents (Fig. 3g, h, Table 1) and ΔI (Fig. 3i, Table 1). At -90 mV, ΔI was significantly decreased from -167.7 ± 54.2 pA to 0.75 ± 8.2 pA (± SD) (Fig. 3i, arrow). Subtraction of the post-from the pre-zatebradine current traces isolated zatebradine-sensitive inward currents which reversed near -50 mV, as previously reported for HCN4 channel reversal potentials33 (Fig. 3j, k, Table 1). These experiments verified the identity of ΔI as Ih. In comparison, application of the Kir channel blocker barium chloride (BaCl2) substantially reduced the overall inward currents but had no effects on ΔI (i.e., Ih) in Tsc1KOneurons (Fig. S6a-i). Consistent with the function of Ih in maintaining RMP at depolarized levels34–38, application of zatebradine hyperpolarized RMP in mTORS2215Y neurons but did not affect the RMP of control neurons that exhibited no Ih (Fig. 3l, Table 1). Collectively, these findings suggest that RhebY35L, mTORS2215Y, Dedpdc5KO, PtenKO, and Tsc1KO expression in layer 2/3 pyramidal neurons lead to the abnormal presence of HCN4 channels with variations in functional expression.
Expression of RhebY35L, mTORS2215Y, Dedpdc5KO, PtenKO, and Tsc1KOleads to different impacts on excitatory synaptic activity
As part of examining neuron excitability, we recorded spontaneous excitatory postsynaptic currents (sEPSCs) in all the gene conditions. The frequency of sEPSCs was unchanged in all experimental conditions except for the Tsc1KO condition, in which the frequency was significantly increased (Fig. 4a, b, Table 1, Fig. S7a). Unlike the other experimental conditions, the RhebY35Lcondition displayed a slight decrease in sEPSC frequency, consistent with previous findings in RhebS16H neurons; however, this did not reach statistical significance (Fig. 4b). The amplitude of sEPSCs was larger in the RhebY35L, mTORS2215Y, and PtenKO conditions (Fig. 4a, c, Table 1, Fig. S7b). Although the amplitudes were slightly larger in the Depdc5KO and Tsc1KO conditions, these changes were not significant (Fig. 4c). Thus, Tsc1KO neurons display increased sEPSC frequency but unchanged amplitude, while RhebY35L, mTORS2215Y,and PtenKO neurons display increased sEPSC amplitude but unchanged frequency. Finally, there was an increase in the sEPSC total charge in all experimental conditions except for the RhebY35L condition (Fig. 4d, Table 1, Fig. S7c). Collectively, these findings suggest all experimental conditions, except for RhebY35L, lead to increased synaptic excitability, with variable impact on sEPSC frequency and amplitude.
Discussion
In this unprecedented comparison study, we examined the impacts of several distinct epilepsy-associated mTORC1 pathway gene mutations on cortical pyramidal neuron development and their electrophysiological properties and synaptic integration in the mouse mPFC. Through a combination of IUE to model the genetic mosaicism of FMCDs, histological analyses, and patch-clamp electrophysiological recordings, we found that activation of either Rheb or mTOR or biallelic inactivation Depdc5, Tsc1, or Pten, all of which increase mTORC1 activity, largely leads to similar alterations in neuron morphology and membrane excitability but differentially impacts excitatory synaptic transmission. These findings have significant implications for understanding the mechanisms leading to cortical hyperexcitability and seizures in mTORC1-driven FMCDs and highlight the utility of personalized medicine dictated by patient gene variants.
Several histological phenotypes associated with mTORC1 hyperactivation were anticipated and confirmed in our studies. We found increased neuron soma size across all gene conditions, consistent with previous reports24,26,27,39–42. Additionally, all conditions, except for PtenKO, resulted in neuronal mispositioning in the mPFC, with RhebY35L and mTORS2215Y conditions being the most severe. Interestingly, PtenKO neurons were correctly placed in layer 2/3, despite having one of the largest soma size increases, suggesting that cell positioning is independent of cell size. The lack of mispositioning of PtenKO neurons was surprising as it is thought that increased mTORC1 activity leads to neuronal misplacement, and aberrant migration of PtenKO neurons has been reported in the hippocampus43. Given that PTEN has a long half-life, with a reported range of 5 to 20 hours or more depending on the cell type44–51, it is possible that following knockout at E15, the decreases in existing protein levels lagged behind the time window to affect neuronal migration. However, TSC1 also has a long half-life of ∼22 hours52,53, and knockout of Tsc1 at the same time point was sufficient to affect neuronal positioning. Mispositioning of neurons following Pten knockout in E15 rats was reported by one study, but this was a very mild phenotype compared to findings in the other gene conditions39. Therefore, these data suggest that PTEN has gene-specific biological mechanisms that contribute to the lack of severe neuronal mispositioning in the developing cortex. mTORS2215Yexhibited the strongest phenotypes in terms of neuron size and positioning, which is perhaps not surprising as the mTOR protein itself forms the catalytic subunit of mTORC1. Depdc5KO exhibited the mildest changes in these parameters. Unlike PTEN, TSC1, RHEB, and mTOR, which regulate mTORC1 via the canonical PI3K-mTORC1 pathway in response to growth factor stimulation, DEPDC5 regulates mTORC1 via the GATOR1 complex in response to changes in cellular amino acid levels54. The different modes of mTORC1 regulation by DEPDC5 may contribute to the differences in the severity of the phenotypes.
To assess the impact of RHEB, mTOR, DEPDC5, PTEN, and TSC1 disruption on cortical neuron excitability, we examined the intrinsic biophysical and synaptic properties of layer 2/3 pyramidal neurons in the mPFC. We found that all gene manipulations led to increased membrane conductance but decreased AP firing upon membrane depolarization, consistent with previous reports in the somatosensory cortex26,39,42,55–58 and in the mPFC for the RhebS16Hcondition30,31. Neurons in these conditions required a larger depolarization to generate an AP (increased rheobase), likely due to their enlarged cell size. These findings seem to contradict the theory that seizure initiation originates from these neurons. However, inhibiting firing in cortical pyramidal neurons expressing the mTORC1-activating RhebS16H mutation via co-expression with Kir2.1 channels (to hyperpolarize neurons) has been shown to prevent seizures, supporting a cell-autonomous mechanism30. This discrepancy may be reconciled by the identification of abnormal HCN4 channel expression in the RhebS16H neurons, which was shown to contribute to increased excitability by depolarizing RMPs, bringing neurons closer to the AP threshold, and conferring firing sensitivity to cyclic AMP30. The abnormal HCN4 channel expression has been verified in resected human FCDII and HME tissue (of unknown genetic etiology)30. Here, we found that RhebY35L, mTORS2215Y, Depdc5KO, PtenKO, and Tsc1KO cortical neurons also display aberrant HCN4 channel expression, which is consistent with the rapamycin- and 4E-BP1-dependent mode of expression in RhebS16H neurons30,31. The functional expression of these channels was variable, and the corresponding changes in Ih did not reach statistical significance for RhebY35L, Depdc5KO, and PtenKO neurons. Nonetheless, consistent with the function of HCN4 channels in maintaining the RMP at depolarized levels38, RhebY35L, mTORS2215Y, and Tsc1KO neurons exhibited depolarized RMPs. Interestingly, the RMP was unchanged in PtenKO and Depdc5KO neurons. The lack of RMP changes may be explained by the milder HCN4 phenotype or the presence of different ion channel complements in these neurons. In addition to HCN4 channels, it is thought that increased Kir channels, which are well-known to hyperpolarize RMPs, contribute to the enlarged inward currents in all the conditions we investigated. This was confirmed in the Tsc1KO neurons where application of BaCl2 to block Kir channels reduced the overall inward current and depolarized RMPs. Since PtenKO neurons have a larger overall inward current but a smaller Ih compared to Tsc1KO neurons, we postulate that PtenKO neurons have a higher expression of Kir channels neurons that could counteract the HCN4 channels’ depolarizing effect on RMP. The mechanisms and functional significance of the Kir channel increases in these neurons are yet to be validated. Nevertheless, our data suggest that abnormal HCN4 channel expression is a conserved mechanism across these gene conditions and warrant further investigation of HCN-mediated excitability in mTORC1-related epilepsy.
While membrane excitability was largely similar across all gene conditions, the excitatory synaptic properties were more variable. All conditions, except for RhebY35L, led to increased excitatory synaptic activity, with variable impact on sEPSC frequency and amplitude. Tsc1KO neurons were the only investigated condition that displayed increased sEPSC frequency. This finding corroborates a previous study showing increased sEPSC frequency with no sEPSC amplitude changes in layer 2/3 cortical pyramidal neurons from 4-week-old Tsc1 conditional KO mice58. Interestingly, as previously reported for RhebS16H neurons59, the sEPSC frequency in RhebY35L neurons was reduced by 25% but this did not reach statistical significance, possibly due to a generally weaker effect of the Y35L mutation compared to the S16H mutation. RhebY35L, mTORS2215Y, and PtenKO neurons exhibited increased sEPSC amplitudes. We found no changes in the sEPSC amplitude of Depdc5KOneurons, which differs from a previous study reporting increased sEPSC amplitude in these neurons26. This discrepancy may likely be explained by the differences in the age of the animals examined in their study (P20-24) versus ours (P26-43), given that dynamic changes in synaptic properties are still ongoing past P2160. Despite the above differences in sEPSC frequencies and amplitudes, the total charge was increased in almost all conditions except for in RhebY35L, suggesting enhanced synaptic drive and connectivity. These excitatory synaptic changes may counteract the increases in rheobase (i.e., decreased membrane excitability) and thereby impact circuit hyperexcitability and seizure vulnerability. Overall, the changes in synaptic activity for the RhebY35L condition are in stark contrast to the other gene conditions. These observations suggest a Rheb-specific impact on synaptic activity that differs from the other mTORC1 pathway genes. Thus, despite impinging on mTORC1 signaling, different mTORC1 pathway gene mutations can affect synaptic activity, and thereby, excitability differently. The mechanisms accounting for the observed differences are not clear, but it is increasingly acknowledged that additional pathways beyond mTORC1 are activated by each gene condition that could contribute to the differential impacts on synaptic activity13. One limitation of this study is that we did not examine inhibitory synaptic properties and the impact on excitatory-inhibitory balance. These changes could potentially contribute to gene-specific mechanisms of cortical hyperexcitability and should be the subject of another study.
In summary, we have shown that mutations affecting different mTORC1 pathway genes have similar and dissimilar consequences on cortical pyramidal neuron development and function, which may affect how neurons behave in cortical circuits. Our findings suggest that cortical neurons harboring different mTORC1 pathway gene mutations may differentially affect how neurons receive and process cortical inputs, which have implications for the mechanisms of cortical hyperexcitability and seizures in FMCDs, and potentially affect how neurons and their networks respond to therapeutic intervention.
Author contribution
LN: conceptualization, investigation, analysis, visualization, writing-original draft, review & editing, YX: investigation, analysis, MN: analysis, AB: conceptualization, writing-original draft, review & editing.
Data availability
The datasets generated during and/or analyzed during the current study are available from the corresponding authors upon reasonable request.
Competing interest
AB is an inventor on a patent application “Methods of treating and diagnosing epilepsy” (Pub. no. US2022/0143219 A1).
Grant funding
National Institutes of Health (NIH) F32 HD095567 (LHN), R01 NS111980 (AB)
Acknowledgements
We thank Drs. Stéphanie Baulac (Paris Brain Institute, France) for providing the pX330-Depdc5 plasmid and Dr. Joseph LoTurco (University of Connecticut) for insightful discussions and advice on the pX330-Tsc1 plasmid.
Methods
Animals
All animal procedures were performed in accordance with Yale University Institutional Animal Care and Use Committee’s regulations. All experiments were performed on male and female CD-1 mice (Charles River Laboratories).
Plasmid DNA
Information on the plasmids used in this study is listed in Table 2. Concentrations of plasmids used for each control and experimental condition are listed in Table 3.
In utero electroporation (IUE)
Timed-pregnant embryonic day (E) 15.5 mice were anesthetized with isoflurane, and a midline laparotomy was performed to expose the uterine horns. A DNA plasmid solution was injected into the right lateral ventricle of each embryo using a glass pipette. For each litter, half of the embryos received plasmids for the experimental condition and the other half received plasmids for the respective control condition. A 5 mm tweezer electrode was positioned on the embryo head and 6 x 42V, 50 ms pulses at 950 ms intervals were applied using a pulse generator (ECM830, BTX) to electroporate the plasmids into neural progenitor cells. Electrodes were positioned to target expression in the mPFC. The embryos were returned to the abdominal cavity and allowed to continue with development. At P0, mice were screened under a fluorescence stereomicroscope to ensure electroporation success, as defined by fluorescence in the targeted brain region, before proceeding with downstream experiments
Brain fixation and immunofluorescent staining
P7-9 (neonates) and P28-43 (young adults) mice were deeply anesthetized with isoflurane and sodium pentobarbital (85 mg/kg i.p. injection) and perfused with ice-cold phosphate buffered saline (PBS) followed by ice-cold 4% PFA. Whole brains were dissected and post-fixed in 4% PFA for 2 hours and then cryoprotected in 30% sucrose for 24-48 hours at 4°C until they sank to the bottom of the tubes. Brains were serially cut into 50 μm-thick coronal sections using a freezing microtome and stored in PBS + 0.05% sodium azide at 4°C until use.
For immunofluorescence staining, free-floating brain sections were washed in PBS + 0.1% triton X-100 (PBS-T) for 2×10 min and permeabilized in PBS + 0.3% triton X-100 for 20-30 min. Sections were then incubated in blocking buffer (10% normal goat serum + 0.3% BSA + 0.3% triton X-100 in PBS) for 1 hour at room temperature and in primary antibodies [anti-p-S6 S240/244 (Cell Signaling Technology #5364, 1:1000) or anti-HCN4 (Alomone Labs APC-052, 1:500), diluted in 5% normal goat serum + 0.3% BSA + 0.1% triton X-100 in PBS] for 2 days at 4°C. Sections were then washed in PBS-T for 3×10 min, incubated in secondary antibodies [goat anti-rabbit IgG Alexa Fluor Plus 647 (Invitrogen #A32733, 1:500)] for 2 hours at room temperature, and again washed in PBS-T for 3×10 min. All sections were mounted onto microscope glass slides and coverslipped with ProLong Diamond Antifade Mountant (Invitrogen) for imaging. The specificity of the HCN4 antibodies was previously validated in our lab30. Additionally, a no primary antibody control was included to confirm the specificity of the secondary antibodies (Fig. S5b).
Confocal microscopy and image analysis
Images were acquired using a Zeiss LSM 880 confocal microscope. All image analyses were done using ImageJ software (NIH) and were performed blinded to experimental groups. Data were quantified using grayscale images of single optical sections. Representative images were prepared using Adobe Photoshop CC. All quantified images meant for direct comparison were taken with the same settings and uniformly processed.
P28-43 neuron soma size and p-S6 staining intensity were quantified from 20x magnification images of p-S6 stained brain sections by tracing the soma of randomly selected tdTomato+ cells and measuring the area and p-S6 intensity (mean gray value) within the same cell, respectively. HCN4 staining intensity was quantified from 20x magnification images by tracing the soma of randomly selected tdTomato+ cells and measuring HCN4 intensity (mean gray value). Staining intensities were normalized to the mean control. 15 cells from 2 brain sections per animal were analyzed for each of the parameters. Neuron positioning (% cells in layer 2/3) was quantified by counting the number of tdTomato+ cells within an 800 μm x 800 μm region of interest (ROI) on the electroporated cortex. Cells within 300 μm from the pial surface were considered correctly located in layer 2/3 whereas cells outside that boundary were considered misplaced31,41,61. The distribution of neurons in the cortex was further quantified by dividing the 800 μm x 800 μm ROI into 10 evenly spaced bins (bin width = 80 μm) parallel to the pial surface and counting the number of tdTomato+ cells in each bin. Only cells within the gray matter of the cortex were quantified. Data are shown as % of total tdTomato+ cell count. One brain section per animal was analyzed. P7-9 neuron soma size (supplemental data) was quantified from 10x magnification images of unstained brain sections by tracing the soma of randomly selected tdTomato+ cells and measuring the area. 30 cells from 2 sections per animal were analyzed.
Acute brain slice preparation, patch clamp recording, and analysis
P26-P51 mice were deeply anesthetized by carbon dioxide inhalation and sacrificed by decapitation. Whole brains were rapidly removed and immersed in ice-cold oxygenated (95% O2/5%CO2) high-sucrose cutting solution (in mM: 213 sucrose, 2.6 KCl, 1.25 NaH2PO4, 3 MgSO4, 26 NaHCO3, 10 Dextrose, 1 CaCl2, 0.4 ascorbate, 4 Na-Lactate, 2 Na-Pyruvate, pH 7.4 with NaOH). 300 μm-thick coronal brain slices containing the mPFC were cut using a vibratome (Leica VT1000) and allowed to recover in a holding chamber with oxygenated artificial cerebrospinal fluid (aCSF, in mM: 118 NaCl, 3 KCl, 1.25 NaH2PO4, 1 MgSO4, 26 NaHCO3, 10 Dextrose, 2 CaCl2, 0.4 ascorbate, 4 Na-Lactate, 2 Na-Pyruvate, 300 mOsm/kg, pH 7.4 with NaOH) for 30 min at 32°C before returning to room temperature (22°C) where they were kept for 6-8 hours during the experiment.
Whole-cell current- and voltage-clamp recordings were performed in a recording chamber at room temperature using pulled borosilicate glass pipettes (4-7 MΩ resistance, Sutter Instrument) filled with internal solution (in mM: 125 K-gluconate, 4 KCl, 10 HEPES, 1 EGTA, 0.2 CaCl2, 10 di-tris-phosphocreatine, 4 Mg-ATP, 0.3 Na-GTP, 280 mOsm/kg, pH 7.3 with KOH). Fluorescent (i.e., electroporated) neurons in the mPFC were visualized using epifluorescence on an Olympus BX51WI microscope with a 40X water immersion objective. Recordings were performed on neurons in layer 2/3. Data were acquired using Axopatch 200B amplifier and pClamp 10 software (Molecular Devices) and filtered (at 5 kHz) and digitized using Digidata 1320 (Molecular Devices). All data analysis was performed offline using pClamp 10 software (Clampfit) and exported to GraphPad Prism 9 software for graphing and statistical analysis.
The RMP was recorded within the first 10 s after achieving whole-cell configuration at 0 pA in current-clamp mode. The membrane capacitance was measured in voltage-clamp mode and calculated by dividing the average membrane time constant by the average input resistance obtained from the current response to a 500-ms long (±5 mV) voltage step from -70 mV holding potential. The membrane time constant was determined from the biexponential curve best fitting the decay phase of the current; the slower component of the biexponential fit was used for the membrane time constant. The resting membrane conductance was measured in current-clamp mode and calculated using the membrane potential change induced by -500 pA hyperpolarizing current injections from rest. The AP input-output curve was generated by injecting 500 ms-long depolarizing currents steps from 0 to 400 pA in 20 pA increments from the RMP of each condition in current-clamp mode. The number of elicited APs was counted using the threshold search algorithm in Clampfit (pClamp). To determine the minimum amount of current needed to induce the first AP, i.e, rheobase, 5 ms-long depolarizing current steps in increments of 20 pA were applied every 3 s until an AP was elicited. The AP threshold, peak amplitude, and half-widths were analyzed from averaged traces of 5-10 consecutive APs induced by the rheobase +10 pA. The AP threshold was defined as the membrane potential at which the first derivative of an evoked AP achieved 10% of its peak velocity (dV/dt). The AP peak amplitude was defined as the difference between the peak and baseline. The AP half-width was defined as the duration of the AP at the voltage halfway between the peak and baseline. Ih was evoked by a series of 3 s-long hyperpolarizing voltage steps from -120 mV to -40 mV in 10 mV increments., with a holding potential of -70 mV. The Ih amplitudes (ΔI) were measured as the difference between the instantaneous current immediately following each test potential (Iinst) and the steady-state current at the end of each test potential (Iss)32.
Zatebradine (40 µM, Toris Bioscience) and BaCl2 (200 µM, Sigma-Aldrich) were applied locally to the recorded neurons via a large-tip (350 µM diameter) flow pipe. When no drugs were applied, a continuous flow of aCSF was supplied from the flow pipe. The IV curve, ΔIV curve, and RMP were measured before and after drug application as described above. The zatebradine-sensitive and BaCl2-sensitive currents were obtained by subtracting the current traces obtained after from before drug application. The IV curve for the zatebradine-sensitive current was obtained by measuring the steady-state of the resulting current, and the IV curve for the BaCl2-sensitive current was obtained by measuring the peak of the resulting current.
sEPSCs were recorded at a holding potential of -70 mV. Synaptic currents were recorded for a period of 2–5 min and analyzed by using the template search algorithm in Clampfit. The template was constructed by averaging 5-10 synaptic events, and the template match threshold parameter was adjusted to minimize false positives. All synaptic events identified by the program were manually inspected and non-synaptic currents (based on the fast-rising time) were discarded. The total charge (pA/ms) was calculated as area of sEPSC events (pA/ms) x frequency (Hz)/1000 for each neuron.
Statistical analysis
All statistical analyses were performed using GraphPad Prism 9 software with the significance level set at p< 0.05. Data were analyzed using nested t-test or nested one-way ANOVA (obtained by fitting a mixed-effects model wherein the main factor is treated as a fixed factor and the nested factor is treated as a random factor; to account for correlated data within individual animals within groups62,63), one-way ANOVA, two-way repeated measure ANOVA, or mixed-effects ANOVA (fitted to a mixed-effects model for when values were missing values in repeated measures analyses), as appropriate. For all nested statistics, the distribution of data among individual animals is shown in Supplemental Figs. S1, S3, S4, and S7. All post-hoc analyses were performed using Holm-Šídák’s multiple comparison test. The specific tests applied for each dataset, test results, and sample size (n, number of animals or neurons) are summarized in Tables 1 and S1 and described in the figure legends. All data are reported as the mean of all neurons or brain sections in each group ± standard error of the mean (SEM).
Supplemental figures and tables
References
- 1Defining the spectrum of international practice in pediatric epilepsy surgery patientsEpilepsia 49:146–155
- 2Histopathological Findings in Brain Tissue Obtained during Epilepsy SurgeryN Engl J Med 377:1648–1656
- 3Malformations of cortical development: clinical features and genetic causesLancet Neurol 13:710–726
- 4Focal brain malformations: seizures, signaling, sequencingEpilepsia 50:3–8
- 5Review: Mechanistic target of rapamycin (mTOR) pathway, focal cortical dysplasia and epilepsyNeuropathol Appl Neurobiol 44:6–17
- 6New insights into a spectrum of developmental malformations related to mTOR dysregulations: challenges and perspectivesJ Anat 235:521–542
- 7mTOR pathway: Insights into an established pathway for brain mosaicism in epilepsyNeurobiol Dis 182
- 8Somatic Mutations Activating the mTOR Pathway in Dorsal Telencephalic Progenitors Cause a Continuum of Cortical DysplasiasCell Rep 21:3754–3766
- 9Comprehensive multi-omic profiling of somatic mutations in malformations of cortical developmentNat Genet 55:209–220
- 10mTOR signaling in growth control and diseaseCell 149:274–293
- 11Regulation of mTORC1 by PI3K signalingTrends Cell Biol 25:545–555
- 12Regulation of mTORC1 by amino acidsTrends Cell Biol 24:400–406
- 13Convergent and Divergent Mechanisms of Epileptogenesis in mTORopathiesFront Neuroanat 15
- 14Current Review in Basic Science: Animal Models of Focal Cortical Dysplasia and EpilepsyEpilepsy Curr 22:234–240
- 15Non-canonical functions of the tuberous sclerosis complex-Rheb signalling axisEMBO Mol Med 3:189–200
- 16PTEN regulation of ERK1/2 signaling in cancerJ Recept Signal Transduct Res 32:190–195
- 17PI3K signaling in cancer: beyond AKTCurr Opin Cell Biol 45:62–71
- 18AKT/PKB Signaling: Navigating the NetworkCell 169:381–405
- 19Non-canonical mTOR-Independent Role of DEPDC5 in Regulating GABAergic Network DevelopmentCurr Biol 28:1924–1937
- 20Somatic Mutations in the MTOR gene cause focal cortical dysplasia type IIbAnn Neurol 78:375–386
- 21Association of MTOR Mutations With Developmental Brain Disorders, Including Megalencephaly, Focal Cortical Dysplasia, and Pigmentary MosaicismJAMA Neurol 73:836–845
- 22Germline and somatic mutations in the MTOR gene in focal cortical dysplasia and epilepsyNeurol Genet 2
- 23Dissecting the genetic basis of focal cortical dysplasia: a large cohort studyActa Neuropathol 138:885–900
- 24A brain somatic RHEB doublet mutation causes focal cortical dysplasia type IIExp Mol Med 51:1–11
- 25Gradient of brain mosaic RHEB variants causes a continuum of cortical dysplasiaAnn Clin Transl Neurol 8:485–490
- 26Second-hit mosaic mutation in mTORC1 repressor DEPDC5 causes focal cortical dysplasia-associated epilepsyJ Clin Invest 128:2452–2458
- 27Somatic Mutations in TSC1 and TSC2 Cause Focal Cortical DysplasiaAm J Hum Genet 100:454–472
- 28CRISPR-mediated direct mutation of cancer genes in the mouse liverNature 514:380–384
- 29Early postnatal development of pyramidal neurons across layers of the mouse medial prefrontal cortexSci Rep 9
- 30Ectopic HCN4 expression drives mTOR-dependent epilepsy in miceSci Transl Med 12
- 31Expression of 4E-BP1 in juvenile mice alleviates mTOR-induced neuronal dysfunction and epilepsyBrain 145:1310–1325
- 32The role of the hyperpolarization-activated current in modulating rhythmic activity in the isolated respiratory network of miceJ Neurosci 20:2994–3005
- 33Gabapentin Modulates HCN4 Channel Voltage-DependenceFront Pharmacol 8
- 34A hyperpolarization-activated cation current (Ih) contributes to resting membrane potential in rat superior cervical sympathetic neuronesPflugers Arch 436:429–435
- 35Contribution of the hyperpolarization-activated current to the resting membrane potential of rat nodose sensory neuronsJ Physiol 514:125–138
- 36Contribution of the hyperpolarization-activated current (I(h)) to membrane potential and GABA release in hippocampal interneuronsJ Neurophysiol 86:261–268
- 37Role of the hyperpolarization-activated cation current (Ih) in pacemaker activity in area postrema neurons of rat brain slicesJ Physiol 552:135–148
- 38The Role of HCN Channels on Membrane Excitability in the Nervous SystemJ Signal Transduct 619747
- 39Tracking and transforming neocortical progenitors by CRISPR/Cas9 gene targeting and piggyBac transposase lineage labelingDevelopment 142:3601–3611
- 40Brain somatic mutations in MTOR cause focal cortical dysplasia type II leading to intractable epilepsyNat Med 21:395–400
- 41Convulsive seizures from experimental focal cortical dysplasia occur independently of cell misplacementNat Commun 7
- 42RHEB/mTOR hyperactivity causes cortical malformations and epileptic seizures through increased axonal connectivityPLoS Biol 19
- 43Rapamycin prevents, but does not reverse, aberrant migration in Pten knockout neuronsNeurobiol Dis 93:12–20
- 44Ubiquitination regulates PTEN nuclear import and tumor suppressionCell 128:141–156
- 45Regulation of PTEN translation by PI3K signaling maintains pathway homeostasisMol Cell 81:708–723
- 46WWP2 is a physiological ubiquitin ligase for phosphatase and tensin homolog (PTEN) in miceJ Biol Chem 293:8886–8899
- 47Identification of a PTEN mutation with reduced protein stability, phosphatase activity, and nuclear localization in Hong Kong patients with autistic features, neurodevelopmental delays, and macrocephalyAutism Res 11:1098–1109
- 48Post-transcriptional and post-translational regulation of PTEN by transforming growth factor-beta1J Cell Biochem 106:1102–1112
- 49Rak functions as a tumor suppressor by regulating PTEN protein stability and functionCancer Cell 15:304–314
- 50Phosphorylation of the PTEN tail regulates protein stability and functionMol Cell Biol 20:5010–5018
- 51The deubiquitylase Ataxin-3 restricts PTEN transcription in lung cancer cellsOncogene 33:4265–4272
- 52IKKbeta suppression of TSC1 function links the mTOR pathway with insulin resistanceInt J Mol Med 22:633–638
- 53TSC1 loss increases risk for tauopathy by inducing tau acetylation and preventing tau clearance via chaperone-mediated autophagySci Adv 7
- 54GATORopathies: The role of amino acid regulatory gene mutations in epilepsy and cortical malformationsEpilepsia 60:2163–2173
- 55Somatic Depdc5 deletion recapitulates electroclinical features of human focal cortical dysplasia type IIAAnn Neurol 84:140–146
- 56BRAFV600E expression in neural progenitors results in a hyperexcitable phenotype in neocortical pyramidal neuronsJ Neurophysiol 123:2449–2464
- 57Non-Cell Autonomous Epileptogenesis in Focal Cortical DysplasiaAnn Neurol 90:285–299
- 58Synaptic hyperexcitability of cytomegalic pyramidal neurons contributes to epileptogenesis in tuberous sclerosis complexCell Rep 40
- 59Normalizing translation through 4E-BP prevents mTOR-driven cortical mislamination and ameliorates aberrant neuron integrationP Natl Acad Sci USA 113:11330–11335
- 60The developmental stages of synaptic plasticityJ Physiol 592:13–31
- 61mTOR Hyperactivity Levels Influence the Severity of Epilepsy and Associated Neuropathology in an Experimental Model of Tuberous Sclerosis Complex and Focal Cortical DysplasiaJ Neurosci 39:2762–2773
- 62Analyzing Clustered Data: Why and How to Account for Multiple Observations Nested within a Study Participant?PLoS One 11
- 63Beyond t test and ANOVA: applications of mixed-effects models for more rigorous statistical analysis in neuroscience researchNeuron 110:21–35
- 64Rapid genetic targeting of pial surface neural progenitors and immature neurons by neonatal electroporationNeural Dev 7
Article and author information
Author information
Version history
- Sent for peer review:
- Preprint posted:
- Reviewed Preprint version 1:
- Reviewed Preprint version 2:
- Version of Record published:
Copyright
© 2023, Nguyen et al.
This article is distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use and redistribution provided that the original author and source are credited.
Metrics
- views
- 1,530
- downloads
- 114
- citations
- 5
Views, downloads and citations are aggregated across all versions of this paper published by eLife.