The mTOR pathway genes MTOR, Rheb, Depdc5, Pten, and Tsc1 have convergent and divergent impacts on cortical neuron development and function

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 activator genes, Rheb or MTOR, or biallelic inactivation of the mTORC1 repressor genes, Depdc5, Tsc1, or Pten in the 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 children 1,2 .These disorders are characterized by abnormal brain cytoarchitecture with cortical overgrowth, mislamination, and cellular anomalies, and are often associated with developmental delay and intellectual disability 3 .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 HME [5][6][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 mosaicism 8,9 .
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 autophagy 10 .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 regulators 11 .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) proteins 12 .
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 FCDII 9 .Genetic targeting of these genes in mouse models consistently recapitulates the epilepsy phenotype, supporting an important role for these genes in seizure generation 13,14 .However, while all these genes impinge on the mTORC1 pathway, many of them also participate in mTORC1-independent functions through parallel signaling pathways 13,[15][16][17][18][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.

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 Rheb Y35L or mTOR S2215Y , 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 seizures [20][21][22][23][24][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 Depdc5 26 , Tsc1 27 , or Pten 28 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 Rheb Y35L , mTOR S2215Y , Dedpdc5 KO , Pten KO , and Tsc1 KO 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 Rheb Y35L and mTOR S2215Y led to the largest soma size changes, with a >3-fold increase from control.
Expression of Pten KO led to a similarly large increase of 2.7-fold, while expression of Dedpdc5 KO and Tsc1 KO led to a 1.7 and 2.1-fold increase, respectively.The increase in the Pten KO condition was significantly higher than both the Dedpdc5 KO and Tsc1 KO conditions, and the increase in the Tsc1 KO condition was significantly higher than the Dedpdc5 KO 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 Pten KO , resulted in neuron misplacement (Fig. 1f-h, Table 1).The Rheb Y35L and mTOR S2215Y conditions 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 Dedpdc5 KO and Tsc1 KO 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 Rheb Y35L , mTOR S2215Y , Dedpdc5 KO , Pten KO , and Tsc1 KO leads to varying magnitudes of neuronal enlargement and mispositioning in the cortex.Of note, Pten KO 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 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 Rheb Y35L and mTOR S2215Y conditions were compared to a control group expressing fluorescent proteins under the same CAG promoter.The Dedpdc5 KO , Pten KO , and Tsc1 KO 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 mTOR S2215Y variant had a larger capacitance increase than those expressing the Rheb Y35L variant.Pten KO and Tsc1 KO neurons had similar increases in capacitance that were both larger than that of the Depdc5 KO 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 Rheb Y35L , mTOR S2215Y , and Tsc1 KO expression led to depolarized resting membrane potential (RMP), Dedpdc5 KO and Pten KO 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 mTOR S2215Y and Rheb Y35L conditions leading to the largest rheobase increases (Fig. 2g, Table 1, Fig. S3d).Collectively, these findings indicate that Rheb Y35L , mTOR S2215Y , and Tsc1 KO 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 neurons 29 , this was observed in a smaller fraction of neurons in the experimental conditions [number of neurons displaying spike doublets: Rheb Y35L : 4/23 (17%), mTOR S2215Y : 7/28 (25%), Dedpdc5 KO : 11/27 (41%), Pten KO : 21/28 (75%) -notably, 6 of the 21 doublets are parts of triplets, Tsc1 KO : 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 mTOR S2215Y and Tsc1 KO conditions (Fig. 2i, Table 1, Fig. S3f), while the AP half-width was decreased in the Rheb Y35L , mTOR S2215Y, and Pten KO 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.

presence of HCN4 channels with variations in functional expression
We recently reported that neurons expressing Rheb S16H , an mTOR activating variant of Rheb, display abnormal expression of HCN4 channels 30,31 .These channels give rise to a hyperpolarization-activated cation current (Ih) that is normally absent in layer 2/3 pyramidal neurons 30,31 .The aberrant Ih, which has implications for neuronal excitability, preceded seizure onset and was detected by P8-12 in mice 30 .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, respectively 30,31 .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 antibodies 30 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 mTOR S2215Y electroporated neurons and absent in the non-electroporated contralateral cortex (Fig. S5a).These data indicate the presence of aberrant HCN4 channel expression following Rheb Y35L , mTOR S2215Y , Depdc5 KO , Pten KO , or Tsc1 KO 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 Rheb S16H neurons 30,31 , hyperpolarizing voltage pulses elicited significantly larger inward currents in all experimental conditions compared to their respective controls (Fig. 3c, d, Table 1).The mTOR S2215Y condition displayed larger inward currents than the Rheb Y35L condition, while the Pten KO condition displayed the largest inward currents compared to the Depdc5 KO and Tsc1 KO conditions (Fig. 3d).These data were proportional to the changes in neuron soma size (Fig. 1e,   2b).The inward currents in Rheb S16H neurons are thought to result from the activation of both inward-rectifier K+ (Kir) channels and HCN channels 30 .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 ΔIvoltage (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 mTOR S2215Y and Tsc1 KO conditions compared to controls (Fig. 3f, Table 1, Fig. S4b).Of note, although the Depdc5 KO and Pten KO conditions did not display a significant increase in Ih amplitudes at -90 mV, 1 out of 28 Depdc5 KO neurons and 4 out of 27 Pten KO neurons had Ih amplitudes that were 2-fold greater than the mean Ih amplitude of the Tsc1 KO condition.6 out of 24 Rheb Y35L 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 Rheb Y35L , Depdc5 KO , and Pten KO neurons, and most mTOR S2215Y and Tsc1 KO neurons.
Considering that the mTOR S2215Y condition led to the largest Ih, we examined the impact of the selective HCN channel blocker zatebradine on hyperpolarization-induced inward currents in mTOR S2215Y 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 potentials 33 (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 Tsc1 KO neurons (Fig. S6a-i).Consistent with the function of Ih in maintaining RMP at depolarized levels [34][35][36][37][38] , application of zatebradine hyperpolarized RMP in mTOR S2215Y neurons but did not affect the RMP of control neurons that exhibited no Ih (Fig. 3l, Table 1).Collectively, these findings suggest that Rheb Y35L , mTOR S2215Y , Dedpdc5 KO , Pten KO , and Tsc1 KO expression in layer 2/3 pyramidal neurons lead to the abnormal presence of HCN4 channels with variations in functional expression.
Expression of Rheb Y35L , mTOR S2215Y , Dedpdc5 KO , Pten KO , and Tsc1 KO leads 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 Tsc1 KO condition, in which the frequency was significantly increased (Fig. 4a, b, Table 1, Fig. S7a).Unlike the other experimental conditions, the Rheb Y35L condition displayed a slight decrease in sEPSC frequency, consistent with previous findings in Rheb S16H neurons; however, this did not reach statistical significance (Fig. 4b).The amplitude of sEPSCs was larger in the Rheb Y35L , mTOR S2215Y, and Pten KO conditions (Fig. 4a, c, Table 1, Fig. S7b).Although the amplitudes were slightly larger in the Depdc5 KO and Tsc1 KO conditions, these changes were not significant (Fig. 4c).Thus, Tsc1 KO neurons display increased sEPSC frequency but unchanged amplitude, while Rheb Y35L , mTOR S2215Y, and Pten KO 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 Rheb Y35L condition (Fig. 4d, Table 1, Fig. S7c).Collectively, these findings suggest all experimental conditions, except for Rheb Y35L , 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 patchclamp 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 reports 24,26,27,[39][40][41][42] .Additionally, all conditions, except for Pten KO , resulted in neuronal mispositioning in the mPFC, with Rheb Y35L and mTOR S2215Y conditions being the most severe.Interestingly, Pten KO 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 Pten KO neurons was surprising as it is thought that increased mTORC1 activity leads to neuronal misplacement, and aberrant migration of Pten KO neurons has been reported in the hippocampus 43 .Given that PTEN has a long half-life, with a reported range of 5 to 20 hours or more depending on the cell type [44][45][46][47][48][49][50][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 hours 52,53 , 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 conditions 39 .Therefore, these data suggest that PTEN has genespecific biological mechanisms that contribute to the lack of severe neuronal mispositioning in the developing cortex.mTOR S2215Y exhibited 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.Depdc5 KO 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 levels 54 .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 cortex 26,39,42,[55][56][57][58] and in the mPFC for the Rheb S16H condition 30,31 .
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 Rheb S16H mutation via co-expression with Kir2.1 channels (to hyperpolarize neurons) has been shown to prevent seizures, supporting a cellautonomous mechanism 30 .This discrepancy may be reconciled by the identification of abnormal HCN4 channel expression in the Rheb S16H 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 AMP 30 .The abnormal HCN4 channel expression has been verified in resected human FCDII and HME tissue (of unknown genetic etiology) 30 .Here, we found that Rheb Y35L , mTOR S2215Y , Depdc5 KO , Pten KO , and Tsc1 KO cortical neurons also display aberrant HCN4 channel expression, which is consistent with the rapamycin-and 4E-BP1-dependent mode of expression in Rheb S16H neurons 30,31 .The functional expression of these channels was variable, and the corresponding changes in Ih did not reach statistical significance for Rheb Y35L , Depdc5 KO , and Pten KO neurons.Nonetheless, consistent with the function of HCN4 channels in maintaining the RMP at depolarized levels 38 , Rheb Y35L , mTOR S2215Y , and Tsc1 KO neurons exhibited depolarized RMPs.Interestingly, the RMP was unchanged in Pten KO and Depdc5 KO 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 Tsc1 KO neurons where application of BaCl2 to block Kir channels reduced the overall inward current and depolarized RMPs.Since Pten KO neurons have a larger overall inward current but a smaller Ih compared to Tsc1 KO neurons, we postulate that Pten KO 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 mTORC1related epilepsy.
While membrane excitability was largely similar across all gene conditions, the excitatory synaptic properties were more variable.All conditions, except for Rheb Y35L , led to increased excitatory synaptic activity, with variable impact on sEPSC frequency and amplitude.Tsc1 KO 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 mice 58 .Interestingly, as previously reported for Rheb S16H neurons 59 , the sEPSC frequency in Rheb Y35L 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.Rheb Y35L , mTOR S2215Y , and Pten KO neurons exhibited increased sEPSC amplitudes.We found no changes in the sEPSC amplitude of Depdc5 KO neurons, which differs from a previous study reporting increased sEPSC amplitude in these neurons 26 .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 P21 60 .Despite the above differences in sEPSC frequencies and amplitudes, the total charge was increased in almost all conditions except for in Rheb Y35L , 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 Rheb Y35L 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 activity 13 .One limitation of this study is that we did not examine inhibitory synaptic properties and the impact on excitatoryinhibitory 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.

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.

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 misplaced 31,41,61 .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.
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 groups 62,63 ), 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).Left diagram depicts the approach for analysis: the total number of tdTomato+ neurons within layer 2/3 (white square) was counted and expressed as a % of total neurons in the imaged area.Right bar graphs show the quantification.(h) Quantification of tdTomato+ neuron distribution across cortical layers.Left diagram depicts the approach for analysis: the imaged area was divided into 10 equal-sized bins across the cortex, spanning the corpus callosum to the pial surface (white grids); the total number of tdTomato+ neurons within each bin was counted and expressed as a % of total neurons in the imaged area.Right bar graphs show the quantification.For graphs d, e: n = 4-8 mice per group, with 6-15 cells analyzed per animal.For graphs g, h: n = 3-7 mice per group, with 1 brain section analyzed per animal.Statistical comparisons were performed using (d, e) nested one-way ANOVA (fitted to a mixed-effects model) to account for correlated data within individual animals, (g) one-way ANOVA, or (h) two-way repeated measures ANOVA.Post-hoc analyses were performed using Holm-Šídák multiple comparison test.*p<0.05,**p<0.01,***p<0.001,****p<0.0001.All data are reported as the mean of all neurons or brain sections in each group ± SEM.For graphs d, e, f: n = 5-10 mice per group, with 24-47 cells analyzed per animal.For graphs h, i, k, l: n = 4-6 neurons (paired).Statistical comparisons were performed using (b, f) nested ANOVA (fitted to a mixedeffects model) to account for correlated data within individual animals, (d, e) mixed-effects ANOVA accounting for repeated measures, or (h, i, l) two-way repeated measures ANOVA.Post-hoc analyses were performed using Holm-Šídák multiple comparison test.*p<0.05,**p<0.01,****p<0.0001,for the IV curves in d, e, h, i: *p<0.05(vs.control), #p<0.05(vs.Rheb Y35L or Dedpdc5 KO ), Ɏp<0.05 (vs.Pten KO ).All data are reported as the mean of all neurons in each group ± SEM.SD -8.1 ± 2.1 -11.5 ± 2.5 -12.6 ± 3.0 19.20, 2, 15 <0.0001 -6.7 ± 1.3 -8.7 ± 1.9 -10.1 ± 4.1 -8.5 ± 1.8 4.280, 3, 23  a The nested one-way ANOVA fits a mixed-effects model wherein the main factor is treated as a fixed factor and the nested factor is treated as a random factor.b Post-hoc analyses were performed using Holm-Šídák multiple comparison test.Significant post-hoc results (p<0.05) are denoted with symbols (*, #, Ɏ) on the graphs, with the number of symbols 1-4 denoting the significant levels p<0.05, <0.01, <0.001, and <0.0001, respectively.For all two-way repeated measured and mixed-effects model ANOVA, all significant results (p<0.05) are denoted with one symbol regardless of the significant level for clearness on the graphs.a Working plasmid solutions were diluted in water and contained 0.03% Fast Green dye to visualize the injection.For each embryo, 1.5 µl of the plasmid mixture was injected into the right ventricle.b Control groups for Rheb Y35L and mTOR S2215Y contained pCAG tdTomato and pCAG-GFP+pCAG-tdTomato, respectively, to distinguish between control and experimental animals for each litter.The total plasmid concentration was kept equal between the two control groups.No differences were observed between the two groups, and therefore, they were combined into one control group.c Equal concentrations of pCAG-tagBFP and pCAG-GFP were added into the control and experimental (Depdc5 KO , Pten KO , and Tsc1 KO ) groups, respectively, to distinguish the control and experimental animals for each litter.KO 13 a The nested t-test and ANOVA fit a mixed-effects model wherein the main factor is treated as a fixed factor and the nested factor is treated as a random factor.b Post-hoc analyses were performed using Holm-Šídák multiple comparison test.Significant post-hoc results (p<0.05) are denoted with symbols (*) on the graphs, with the number of symbols 1-4 denoting the significant levels p<0.05, <0.01, <0.001, and <0.0001, respectively.

Figure 1 :
Figure 1: Expression of Rheb Y35L , mTOR S2215Y , Dedpdc5 KO , Pten KO , and Tsc1 KO leads to varying magnitudes of neuronal enlargement and mispositioning in the cortex.(a) Diagram of the PI3K-mTORC1 pathway.Activation of mTORC1 signaling is controlled by positive (blue) and negative (purple) regulators within the pathway.(b) Diagram of overall experimental timeline and approach.IUE was performed at E15.5.A cohort of animals was used for patch clamp recording at P26-51 and another cohort was used for histology at P28-43.(c) Representative images of tdTomato+ cells (red) and p-S6 staining (green) in mouse mPFC at P28-43.(d) Quantification of p-S6 staining intensity (normalized to the mean control) in tdTomato+ neurons.(e) Quantification of tdTomato+ neuron soma size.(f) Representative images of tdTomato+ neuron (red) placement and distribution in coronal mPFC sections.Red square on the diagram denotes the imaged area for all groups.CC, corpus callosum.(g) Quantification of tdTomato+ neuron placement in layer 2/3.Left diagram depicts the approach for analysis: the total number of tdTomato+ neurons within layer 2/3 (white square) was counted and expressed as a % of total neurons in the imaged area.Right bar graphs show the quantification.(h) Quantification of tdTomato+ neuron distribution across cortical layers.Left diagram depicts the approach for analysis: the imaged area was divided into 10 equal-sized bins across the cortex, spanning the corpus callosum to the pial surface (white grids); the total number of tdTomato+ neurons within each bin was counted and expressed as a % of total neurons in the imaged area.Right bar graphs show the quantification.For graphs d, e: n = 4-8 mice per group, with 6-15 cells analyzed per animal.For graphs g, h: n = 3-7 mice per group, with 1 brain section analyzed per animal.Statistical comparisons were performed using (d, e) nested one-way ANOVA (fitted to a mixed-effects model) to account for correlated data within individual animals, (g) one-way ANOVA, or (h) two-way repeated measures ANOVA.Post-hoc analyses were performed using Holm-Šídák multiple comparison test.*p<0.05,**p<0.01,***p<0.001,****p<0.0001.All data are reported as the mean of all neurons or brain sections in each group ± SEM.

Figure 2 :
Figure 2: Expression of Rheb Y35L , mTOR S2215Y , Dedpdc5 KO , Pten KO , and Tsc1 KO universally leads to decreased depolarization-induced excitability, but only Rheb Y35L , mTOR S2215Y , and Tsc1 KO expression leads to depolarized RMPs.(a) Diagram of experimental approach for whole-cell patch clamp recording.Recordings were performed in layer 2/3 pyramidal neurons in acute coronal slices from P26-51 mice, expressing either control, Rheb Y35L , mTOR S2215Y , Dedpdc5 KO , Pten KO , or Tsc1 KO plasmids.(b-d) Graphs of membrane capacitance, resting membrane conductance, and RMP.(e) Representative traces of the AP firing response to a 300 pA depolarizing current injection.(f) Input-output curves showing the mean number of APs fired in response to 500 ms-long depolarizing current steps from 0 to 400 pA.Arrows point to initial spike doublets.(g-j) Graphs of rheobase, AP threshold, AP peak amplitude, and AP half-width.For all graphs: n = 5-10 mice per group, with 22-50 cells analyzed per animal.Statistical comparisons were performed using (b-d, g-j) nested one-way ANOVA (fitted to a mixed-effects model) to account for correlated data within individual animals or (f) mixed-effects ANOVA accounting for repeated measures.Post-hoc analyses were performed using Holm-Šídák multiple comparison test.*p<0.05,**p<0.01,***p<0.001,****p<0.0001,for the input-output curves in f: *p<0.05(vs.control), #<0.05 (vs.Dedpdc5 KO ), Ɏp<0.05 (vs.Tsc1 KO ).All data are reported as the mean of all neurons in each group ± SEM.

Figure 3 :
Figure 3: Expression of Rheb Y35L , mTOR S2215Y , Dedpdc5 KO , Pten KO , and Tsc1 KO leads to the abnormal presence of HCN4 channels with variations in functional expression.(a) Representative images of tdTomato+ cells (red) and HCN4 staining (green) in mouse mPFC at P28-43.(b) Quantification of HCN4 intensity (normalized to the mean control) in tdTomato+ neurons.(c) Representative current traces in response to a series of 3 s-long hyperpolarizing voltage steps from -120 to -40 mV, with a holding potential of -70 mV.Current traces from the -40 and -50 mV steps were not included due to contamination from unclamped Na+ spikes.Orange lines denote the current traces at -90 mV.(d) IV curve obtained from I ss amplitudes.(e) ΔIV curve obtained from I h amplitudes (i.e., ΔI, where ΔI=I ss -I inst ).(f) Graphs of I h amplitudes at -90 mV.(g) Representative current traces in response to a series of 3-s long hyperpolarizing voltage steps from -110 mV to -50 mV in the mTOR S2215Y condition pre-and post-zatebradine application.Orange lines denote the current traces at -90 mV.(h) IV curve obtained from I ss amplitudes in the mTOR S2215Y condition pre-and post-zatebradine application.(i) ΔIV curve obtained from I h amplitudes (i.e., ΔI) in the mTOR S2215Y condition pre-and post-zatebradine application.Arrow points to the post-zatebradine I h amplitude at -90 mV.(j) Representative traces of the zatebradine-sensitive current obtained after subtraction of the post-from the pre-zatebradine current traces in response to -110 mV to -50 mV voltage steps.Orange lines denote the current traces at -90 mV.(k) IV curve of the zatebradine-sensitive current obtained after subtraction of the post-from the pre-zatebradine current traces.(l) Graph of RMP in the control and mTOR S2215Y conditions pre-and post-zatebradine application.Connecting lines denote paired values from the same neuron.For graph b: n = 4-8 mice per group, with 4-15 cells analyzed per animal.For graphs d, e, f: n = 5-10 mice per group, with 24-47 cells analyzed per animal.For graphs h, i, k, l: n = 4-6 neurons (paired).Statistical comparisons were performed using (b, f) nested ANOVA (fitted to a mixedeffects model) to account for correlated data within individual animals, (d, e) mixed-effects ANOVA accounting for repeated measures, or (h, i, l) two-way repeated measures ANOVA.Post-hoc analyses were performed using Holm-Šídák multiple comparison test.*p<0.05,**p<0.01,****p<0.0001,for the IV curves in d, e, h, i: *p<0.05(vs.control), #p<0.05(vs.Rheb Y35L or Dedpdc5 KO ), Ɏp<0.05 (vs.Pten KO ).All data are reported as the mean of all neurons in each group ± SEM.

Figure 4 :
Figure 4: Expression of Rheb Y35L , mTOR S2215Y , Dedpdc5 KO , Pten KO , and Tsc1 KO leads to different impacts on sEPSC properties.(a) Representative sEPSC traces recorded at a holding voltage of -70 mV.Top and bottom traces are from the same neuron.(b-d) Graphs of sEPSC frequency, amplitude, and total charge.For all graphs: n = 5-9 mice per group, with 17-34 cells analyzed per animal.Statistical comparisons were performed using a nested ANOVA (fitted to a mixed effects model) to account for correlated data within individual animals.Post-hoc analyses were performed using Holm-Šídák multiple comparison test.*p<0.05,**p<0.01,***p<0.001,****p<0.0001.All data are reported as the mean of all neurons in each group ± SEM.

Figure S1 :
Figure S1: Distribution of p-S6 staining intensity and neuron soma size among individual animals P28-43.(a) Distribution of p-S6 staining intensity among individual animals.(b) Distribution of neuron soma size among individual animals.

Figure S2 :
Figure S2: Neuron soma sizes at P7-9.(a) Representative images of tdTomato+ cells (red) in mouse mPFC at P7-9.(b) Quantification of tdTomato+ neuron soma size.(c) Graph showing the distribution of neuron soma size among individual animals in each group.n = 3-8 mice per group, with 30 cells analyzed per animal.Statistical comparisons were performed using a nested t-test (control vs. mTOR S2215Y ) or nested one-way ANOVA (control vs. Dedpdc5 KO vs. Pten KO , vs. Tsc1 KO ) (fitted to a mixed-effects model) to account for correlated data within individual animals.Post-hoc analyses were performed using Holm-Šídák multiple comparison test.*p<0.05,***p<0.001,****p<0.0001.Data are reported as the mean of all neurons in each group ± SEM.

Figure S3 :
Figure S3: Distribution of membrane and AP properties among individual animals P26-51.(a-c) Distribution of membrane capacitance, conductance, and RMP among individual animals.(d-g) Distribution of rheobase, AP threshold, AP peak amplitude, and AP half-width among individual animals.

Figure S4 :
Figure S4: Distribution of HCN4 staining intensity and I h amplitudes (at -90 mV) among individual animals at P28-43 and P26-51, respectively.(a) Distribution of HCN4 staining intensity among individual animals.(d-g) Distribution of I h amplitudes (at -90 mV) among individual animals.

Figure S5 :
Figure S5: Additional images of HCN4 staining in mPFC sections from P28-43 mice expressing mTOR S2215Y .(a) Representative images of tdTomato+ cells (red) and HCN4 staining (green).Top panels show overlay images.Bottom panels show HCN4 single-channel images.a1 and a2 show the ipsilateral and contralateral cortex, respectively.a3 and a4 are high-magnification images demonstrating HCN4 staining within the cell.(b) Representative images of immunostaining with (left) or without (right) HCN4 primary antibodies (control for secondary antibody specificity).

Figure S6 :
Figure S6: BaCl 2 application decreases overall inward currents without affecting I h in Tsc1 KO neurons.(a) Representative current traces in response to a series of 3-s long hyperpolarizing voltage steps from -130 mV to -40 mV in the Tsc1 KO condition pre-and post-BaCl 2 application.Orange lines denote the current traces at -90 mV.(b) IV curve obtained from I ss amplitudes in the Tsc1 KO condition pre-and post-BaCl 2 application.(c) IV curve obtained from I h amplitudes (i.e., ΔI, where ΔI=I ss -I inst ) in the Tsc1 KO condition pre-and post-BaCl 2 application.(d) Representative traces of the BaCl 2 -sensitive current obtained after subtraction of the post-from the pre-BaCl 2 current traces.Orange lines denote the current traces at -90 mV.(e) IV curve of the BaCl 2 -sensitive current obtained after subtraction of the post-from the pre-BaCl 2 current traces.The isolated BaCl 2 -sensitive current reversed near -80 mV.(f) Graph of conductances (at -500 pA) in the control and Tsc1 KO conditions pre-and post-BaCl 2 application.Connecting lines denote paired values from the same neuron.(g) Graph of % change in conductances pre-and post-BaCl 2 application in the control and Tsc1 KO conditions.(h) Graph of RMP in the control and Tsc1 KO conditions pre-and post-BaCl 2 application.Connecting lines denote paired values from the same neuron.(i) Graph of % change in RMP pre-and post-BaCl 2 application in the control and Tsc1 KO conditions.For all graphs: n = 10-13 neurons per group.Statistical comparisons were performed using (b, c, e) mixed-effects ANOVA, (f, h) two-way repeated measures ANOVA, or (g, i) unpaired t-test.Post-hoc analyses were performed using Holm-Šídák multiple comparison test.*p<0.05,**p<0.01,****p<0.0001.All data are reported as the mean of all neurons or brain sections in each group ± SEM.

Figure S7 :
Figure S7: Distribution of sEPSC properties among individual animals at P26-51.(a-c) Distribution of sEPSC properties among individual animals.