Integrin-KCNB1 potassium channel complexes regulate neocortical neuronal development and are implicated in epilepsy

Potassium (K+) channels are robustly expressed during prenatal brain development, including in progenitor cells and migrating neurons, but their function is poorly understood. Here, we investigate the role of voltage-gated K+ channel KCNB1 (Kv2.1) in neocortical development. Neuronal migration of glutamatergic neurons was impaired in the neocortices of KCNB1 null mice. Migratory defects persisted into the adult brains, along with disrupted morphology and synaptic connectivity. Mice developed seizure phenotype, anxiety, and compulsive behavior. To determine whether defective KCNB1 can give rise to developmental channelopathy, we constructed Knock In (KI) mice, harboring the gene variant Kcnb1R312H (R312H mice) found in children with developmental and epileptic encephalopathies (DEEs). The R312H mice exhibited a similar phenotype to the null mice. Wild type (WT) and R312H KCNB1 channels made complexes with integrins α5β5 (Integrin_K+ channel_Complexes, IKCs), whose biochemical signaling was impaired in R312H brains. Treatment with Angiotensin II in vitro, an agonist of Focal Adhesion kinase, a key component of IKC signaling machinery, corrected the neuronal abnormalities. Thus, a genetic mutation in a K+ channel induces severe neuromorphological abnormalities through non-conducting mechanisms, that can be rescued by pharmacological intervention. This underscores a previously unknown role of IKCs as key players in neuronal development, and implicate developmental channelopathies in the etiology of DEEs.


INTRODUCTION
The voltage-gated K + channel KCNB1, forms macromolecular complexes with integrins, named Integrin_K + channel_Complexes (IKCs) [1][2][3]. Evidence suggests that IKCs regulate fundamental cellular functions, such as migration, proliferation, survival and death through non-conducting (non-ionic) mechanisms [1]. The crucial role of IKCs is further underscored by the fact that mutations in the KCNB1 gene are found in children affected by developmental and epileptic encephalopathies (DEEs), neurological conditions characterized by severe developmental delays, that often co-exist with seizures and abundant epileptiform abnormalities [4,5]. While the mechanisms underlying DEEs are currently under investigation, accumulating evidence points out to abnormal neuronal development and resulting synaptic connectivity disturbances, as one of the underlying causes [6,7]. It is not coincidental that in vitro IKCs formed with certain KCNB1 DEE-variants, impair cellular functions such as migration and neuritogenesis, through non-ionic mechanisms [2]. Given the established importance of integrins in determining neocortical developmental processes, these findings support an argument that IKCs may be implicated in prenatal brain development and consequently, that developmental channelopathies may contribute to the etiology of DEEs [2,[8][9][10][11][12].
To elucidate the role of IKCs in the mechanisms governing the development of the brain and more broadly, their impact on neurological disease, we generated a knock in (KI) mouse model of DEEs harboring the Kcnb1 R312H gene variant, and a null (NULL) mouse bearing a null allele encoding a 337 amino acids truncated KCNB1 protein. The NM_004975.2:c.935G>A mutation, which causes an arginine to histidine replacement in the voltage sensor of the channel (R312H), was initially identified in two children affected by severe developmental delays and severe epilepsy that was highly refractory to anti-epileptic drugs [7].
Here, we report that IKCs are essential for the migration of glutamatergic neurons during prenatal brain development that they control through non-ionic functions. Accordingly, IKC formed with R312H KCNB1 subunits impair neuronal migration causing significant non-structural abnormalities in the adult brain, and are associated with epilepsy and behavioral deficit.

In vivo animal studies
Mice of either sexes at developmental stages: E13, P0, P7, and 3 month old was used. Littermates of either sexes were randomly assigned (by flipping of a coin) to experimental groups. Animals were housed in an AAALAC approved vivarium under the care of a veterinarian. The animals were housed in large cages, with a 12-h light-dark cycle, and fed ad libitum. We adhere and thus followed the guiding principles of animal care as approved by the American Physiological Society, the Guide for the Care of Laboratory Animals and our institution's Animal Care and Use Committee (IACUC). All experiments with animals performed in this study were IACUC approved.
EGG recording and behavioral protocols EGG recordings and behavioral tests were performed with prior knowledge of the genotype.
Open field. We followed a standard protocol as described by Seibenhener and Wooten [13]. Briefly, testing was performed in dim light, in a room equipped with a plexiglass arena (50 cm × 50 cm × 40 cm) whose walls were covered by tissue paper to prevent the mouse from seeing outside, and a video-tracking system (EthoVision XT; Noldus Information Technology, Leesburg, VA). The mouse was allowed to acclimate to the procedure room for 30 min. Then the mouse was placed in the middle of the arena and the test administrator leaved the room. The mouse was kept in the arena for a single 10 min period during which movement was recorded.
Self grooming. Mice were recorded for 6 h with time-locked Lorex cameras and DVR. The overall time spent grooming was divided by the duration of the recording and expressed as the fraction of time spent grooming per minute.
EEG recordings and seizure analysis. Eight-nine week old animals heterozygous and homozygous in Kcnb1 R312H and NULL mice were equipped with a bipolar electrode surgically implanted into the hippocampus (AP: −1.94; ML:−1.75, DV: −1.6, relative to bregma), a surface cortical monopolar screw electrode and cerebellum reference screw electrode while under 1.5% isoflurane anesthesia. The mice were allowed to recover for 1 week before the EEG recording. Mice were tethered for the EEG recording for a period of 24 h. Electrical brain activity was amplified (Dual Bio Amp amplifier) and digitized by Powerlab 16/35 data acquisition device using Labchart application interface (ADInstuments, Dunedin, New Zealand). For the seizure analysis, the EEG recordings were systematically reviewed, and the seizures were manually scored and defined as sustained rhythmic synchronous discharges in the EEG, clearly distinguished from background EEG and interictal activity as described previously [14]. Brief bursts of spikes and periodic spikes were not considered part of a seizure.

Biochemistry
Western blotting was performed with prior knowledge of the genotype. Immunofluorescence and Golgi staining and Sholl analysis, were performed without prior knowledge of the genotype.
Crude brain lysates. Half sagittal brains were homogenized with a plastic tissue grinder in lysis buffer (0.32 M sucrose, 5 mM Tris-Cl pH 6.8, 0.5 mM EDTA, and protease inhibitor cocktail (Sigma-Aldrich). The brains samples were further homogenized by being passed 5-6 times with a syringe and/ or sonicated for 1-3 min and then centrifuged for 10 min at 14000 rpm at 4°C. The supernatant was collected and stored at −80°C for further analyses.
Immunoprecipitations. To minimize unspecific binding, lysates were preincubated with Protein A/G PLUS-Agarose beads (Santa Cruz Biotechnology) for 20 min at room temperature. The beads were then collected by centrifugation 5000 RPM for 3 min and discharged.

Immunofluorescence
Immunofluorescence experiments were performed without prior knowledge of the genotype.
Neonatal brains. The detailed biochemical procedures were previously described [15,16]. Briefly 5-chloro-2-deoxyuridine (CldU) solution at a concentration of 5 ml/kg of mouse weight in sterile PBS, from a 10 mg/mL stock, was prepared freshly and was intraperitoneally injected into pregnant dams at E17. The P7 anesthetized pups (ketamine/xylazine) were washed with 0.9% filtered NaCl and then perfused with 4% paraformaldehyde solutions. Brains were removed, post-fixed in 4% paraformaldehyde overnight at 4°C, and cryoprotected in 30% sucrose. Brains were embedded in agarose, and slices were prepared throughout the cortex and the hippocampus in a 1:6 series (70 μm thick) using a vibratome (Leica VT1000S), so that the same set of tissue samples could be used for expression of different makers. Samples were incubated in 1.0 M HCl for 30 min under gentle shaking followed by incubation in 2.0 M HCl for 10 min under gentle shaking and washed 4 times with PBS for 5 min each time under vigorous shaking (antigen retrieval). Slices were incubated in blocking/permeabilization buffer (donkey blocking serum pH = 7.43 + Triton-X 0.4%). Then, free-floating slices were processed for labeling with primary antibodies (according to manufacture recommendations). After 24-48 h incubation at 4°C in primary antibodies, sections were washed three times with PBS for 10 min each time, followed by application of the appropriate secondary conjugated antibodies. After incubation for 1 h at room temperature, sections were washed three times with PBS.
Adult brains. The detailed biochemical procedures were previously described [17]. Briefly anesthetized mice (ketamine/xylazine) were washed with 0.9% filtered NaCl and then perfused with 4% paraformaldehyde solutions. Brains were removed, post-fixed in 4% paraformaldehyde overnight at 4°C, and cryoprotected in 30% sucrose. Brain slices were cut 20 µm thick throughout the cortex and the hippocampus in a 1:20 series so that the same set of tissue samples could be used for expression of different makers. After permeabilization procedures (7 min in 0.1% Triton-X in PBS) when required, mounted sections were processed for double labeling with primary antibodies. After 24-48 h incubation at 4°C in A. Bortolami et al.
primary antibodies, sections were washed three times with PBS for 10 min each time and incubated with the appropriate secondary conjugated antibody. After 1 h at room temperature, the sections were washed three times with PBS.
Primary neurons. Cells grown in six wells plates were fixed using a 50% methanol, 50% acetone solution for 20 min on dry ice. Cells were washed 3 times with PBS for 10 min each time, and blocked for 1 h at room temperature with 3% BSA in PBS. After 24-48 h incubation at 4°C in primary antibodies, the cells were washed three times with PBS, followed by application of the appropriate secondary conjugated antibodies. After 1 h at room temperature, the cells were washed three times with PBS. All slides were mounted in VECTASHIELD Antifade Mounting Medium with DAPI mounting buffer (Vector Laboratories, Burlingame, CA) and stored at 4°C. Staining was visualized with a Zeiss Axiophot microscope or with an Olympus FV1000MPE (Orangeburg, NY) multi-photon microscope or Nikon Eclipse Ti2 series (Tokyo, Japan) confocal microscope, all equipped with dedicated software.
Golgi staining. Golgi staining was performed using the FD rapid GolgiStain kit (FD Neurotechnologies, Columbia, MD) according to manufacturer's instructions. Brains were embedded in Tissue Freezing Medium, sectioned at a 150 µm thickness, mounted on gelatin precoated slides and photographed with a Zeiss Axio Imager M1 at 100×. Sholl analysis of cropped neurons was performed using ImageJ/Fiji software.

Cell cultures
Chinese hamster ovary cells. Chinese hamster ovary (CHO-K1, simply CHO) cells were maintained at 37°C in a 95% air/5% CO 2 humidified incubator. They were typically seeded at a density of 0.5 million cells/mL using a hemocytometer, and grown in Dulbecco Modified Eagle's medium (DMEM) as described [2]. The cells were transfected using Lipofectamine 2000 according to manufacturer's instructions. Twenty μl of reagent and 4 μg plasmid DNA were used for each transfection.
Mouse neuroblastoma cells. Undifferentiated mouse neuroblastoma N2a cells cells were cultured in DMEM supplemented with 10% fetal bovine serum and 1% sodium pyruvate at 37°C in a 95% air/5% CO 2 humidified incubator. They were typically seeded at a density of 0.5 million cells/mL. After reaching ∼60% confluence the cells were transfected with Lipofectamine, incubated in serum-free DMEM for 48 h and then analyzed.
Primary cortical neurons. The detailed procedure was previously described. Experiments involving primary neurons were performed without prior knowledge of the genotype [17]. Briefly, cortices were obtained from time-mated embryonic day 13 (E13) embryos (pure cultures) or from P0 pups (co-cultures). Cortical tissue from individual embryos/pups was mechanically triturated and neuron isolated and plated in 12 mm poly-dlysine-precoated petri dishes (GG-12-1.5-PDL, Neuvitro, Camas, WA) at 250,000 cells/dish (1.5 ml medium/dish). Cultures were maintained in Plating medium at 37°C in a 95% air/5% CO2 humidified incubator for 24 h. After 24 h the medium was removed and replaced with Neurobasal medium at 37°C in a 95% air/5% CO2 humidified incubator. Ear samples from individual embryos or pups were processed for genotyping.

In vitro assays
In vitro experiments were performed without prior knowledge of the genotype.
Treatment with Ang II and PAF C-16 agonists. Ang II and PAF C-1 were freshly diluted from ethanol stocks (Ang II: 4.7 mM and PAF C-16: 20 mM) and added to the medium at the indicated concentrations. In the wound healing assay, Ang II or PAF were added immediately after the scratch and maintained until the end of the experiment. In the N2a assay, PAF or Ang II were added immediately after transfection and maintained until the end of the experiment. In primary neuron cultures, Ang II was added to the medium on DIV2 and maintained until the end of the experiment.
Proliferation. CHO cells were seeded into 96-well tissue culture plates in equal numbers. Twenty-four hours after seeding, the cells (∼250,000 cells/ well), were transfected with various plasmids using Lipofectamine. One and two days post-transfection (dpt1 and dpt2), the numbers of viable cells were assessed using the microtiter-plate colorimetric MTT assay.
Briefly, cells were incubated in serum-free media and MTT reagent for 3 h at 37°C. Then, MTT solvent was added, the cells were incubated for additional 15 min and then measurements taken using a Tecan Infinite M200pro (Männedorf, Switzerland) microplate reader (590 nm absorbance).
Wound healing assay. CHO cells were seeded into 6-well tissue culture plates in equal numbers. Twenty-four hours after seeding, the CHO cells (∼75% confluency), were transfected with Lipofectamine as already described. The monolayer was perpendicularly scratched twice with a 200 μl (yellow) pipette tip. The well was washed with DMEM medium to remove the detached cells and the scratches were photographed with a Zeiss Axiovert 200 M (Oberkochen, Germany) microscope at ×10 magnification. After 24 h the scratches were re-photographed. The gap covered in 24 h was measured using ImageJ 1.52a software (National Institutes of Health, Bethesda, MA). Gap distance was calculated as: where d 0 and d 24  . Pipettes (∼5 MΩ) were obtained by pulling borosilicate glass with a Sutter P-97 puller (Sutter Instruments, Novato, CA). Whole-cell currents were evoked by 1 s voltage sweeps from an holding potential of −80 mV to +80 mV in 20 mV increments and leak subtraction was performed digitally using Clampfit software (Molecular Devices). Macroscopic conductance (G) curves were calculated as: where I is the macroscopic current at steady-state (at the end of the voltage pulse) and V rev. is the reversal potential, obtained by linear fitting the current-voltage relationships around the I = 0 region. Offset potentials, including series resistance were estimated to be ≤5 mV and were not compensated for when generating current-voltage relationships. G/G Max curves were fitted to the Boltzmann function: where V is the membrane voltage, V 1/2 is the value of the voltage at which Eq. 3 is equal to 0.5 and V S is the slope coefficient (in mV).

Statistical analysis
Quantitative data are presented as mean ± standard error of the mean (SEM). The estimated sample size, N, needed to detect a meaningful difference between groups was calculated using power analysis with alpha = 0.05 and power = 0.8 [18]. Normality and log-normality tests (D′ Agostino and Pierson) were calculated for normal distribution. The level of significance, assumed at the 95% confidence limit or greater (P < 0.05), was calculated using one-way ANOVA with a Tukey's post hoc test or two-way ANOVA, with a Dunnett's post hoc test or a two-sample Kolmogorov-Smirnov, that were computed by Prism software.

R312H mice exhibit compulsive behaviors
All KI animals were grossly normal in appearance. However, R312H homozygotes (HOMO) did not mate, and were obtained by breeding of heterozygous (HETERO) pairs. The analysis of time-locked video recordings revealed excessive self grooming in the KI population, which often caused skin excoriations (Fig. 1A), and may reflect defects at the level of the neocortex [19]. Compulsive behaviors are associated with anxiety in mice. When tested in the open field maze, R312H homozygotes and NULL spent ∼10-fold more time on the sides than in the center of the arena compared to WT (Fig. 1B) [13]. In contrast, the heterozygotes behaved similarly to control animals, even though the former showed a trend toward more anxiety.
R312H KI mice are affected by frequent seizures KI animals exhibited spontaneous convulsive and non-convulsive seizures. Behavioral seizures were both tonic and myoclonic (Supplementary V1). To investigate the epileptic phenotype of the KI animals in more detail, we analyzed EEG recordings in 2-3month-old mice of either sexes. Both heterozygotes and homozygotes displayed spontaneous electrographic seizures (Fig. 1C).
There was no significant difference in seizure frequency or duration across KI mice, so that the seizure burden was comparable in all genotypes (Fig. 1D-F). Thus, a single Kcnb1 R312H allele is sufficient to induce the full seizure phenotype.

R312H protein is downregulated in the KI brains
To assess the impact of Kcnb1 R312H on the mouse brain, we characterized the expression of the KCNB1 protein. The neurons reactive to KCNB1 antibody (KCNB1+), were decreased in brain slices of heterozygotes and to a larger extent R312H homozygotes compared to control, and were not detected in NULL mice, as expected ( Fig. 2A, B. Bright-field images, Fig. S1A). Accordingly, KCNB1 protein was decreased in Western blots of crude KI brain lysates compared to control, and was absent in NULL mice (Fig. 2C. Uncropped gels in Supplementary Information). The downregulation of R312H protein could be due to defective trafficking to the plasma membrane. However, eye inspection revealed normal expression/distribution of the channel on the membrane of immunoreactive neurons of Kcnb1 R312H genotype (Fig. 2B, arrows). This impression was corroborated by biochemical assessment of surface expression (Fig. 2D), ruling out major trafficking defects (the ratio of R312H/WT protein, total, and at the surface, is ∼0.4 for both).
R312H channels assemble with integrins WT KCNB1 channels form macromolecular complexes with integrins in the mouse and human brain [20]. Therefore, we determined whether R312H mutant channels retained the ability to form IKCs. In three experiments, R312H co-immunoprecipitated with integrin-α5 and integrin-β5 (Fig. 2E). We refer to complexes containing R312H subunits as IKC R312H . Complexes formed with WT channels are referred to as IKC WT or simply IKCs.

R312H neuronal development is disrupted in vitro
The low levels of R312H protein may be due to decreased cell viability. Accordingly, immunoreactivity to neuronal marker microtubule-associated protein 2 (Map2), was reduced in the cortices of KI animals, suggesting disrupted neuronal maturation, density or organization-with large areas possibly depleted of pyramidal cells (Fig. S2) [21]. The sick cells may be eliminated during prenatal development, when neocortical progenitors differentiate into Map2+ glutamatergic neurons (KCNB1 is embryonically expressed, Fig. S3, [22]), or they may slowly degenerate and die, or both. To distinguish between these possibilities, we examined pure cultures harvested from the neocortices of embryonic day 13 (E13) embryos (when neocortical wall contains neuronal progenitors), and cocultures of neurons and glia cells, when all neocortical neurons are already born [23][24][25]. At day in vitro 3 (DIV3), the number of cells co-stained for KCNB1 and Map2 was comparable in all pure cultures, irrespective of the genotype, as expected (Fig. S4A, C) [26]. In contrast, when compared to WT in co-cultures, the number of cells co-stained for KCNB1 and Map2 was decreased by roughly 20% and 40% in R312H heterozygote and homozygote cells, respectively, (Fig. S4B, D). Notably, the decrease of KCNB1+ cells in co-cultures, matched the decrease of total and surface KCNB1 protein. Neuronal mortality from DIV3 to DIV14 was ∼10-15%, irrespective of the genotype or culture type. At DIV14, in all cultures, KCNB1+ neurons had assumed typical pyramidal cell shape, and likewise in vivo neurons, the R312H variant was normally present and formed clusters, at the surface (Fig. S4E, F; 3-D reconstructions of individual neurons, Supplementary V2). Taken together, these data reveal disrupted neuronal development or loss of neurons expressing KCNB1 in KI animals taking place between E13 and P0, a stage in which the neocortex is formed-suggesting that corticogenesis might be impaired in the KI brains.
Neuronal migration is hindered in the neocortices of R312H mice To begin revealing the role of IKCs in neurodevelopment, we injected thymidine analog CldU at E17, to label progenitors of glutamatergic intracortically projecting neurons and their progenies destined for upper layers (UL). Remarkably, at P7 we found significant numbers of E17 CldU+ neuronal progenies in the deep layers (DL) instead of UL, in heterozygotes and to a larger extent R312H homozygotes (Fig. 3. Bright-field images, Fig. S1B). This migratory defect was further exacerbated in NULL animals, where the majority of E17 born neurons remained in the DL. These data suggest that IKCs are required for neuronal migration to UL in developing neocortices.

R312H mice exhibit neuromorphological malformations
Deficits in neuronal migration are a primary cause of brain malformations, and can be associated with seizures [27]. Therefore, we sought to determine the potential impact of the migratory defects on different subpopulations of neocortical glutamatergic neurons, long-term. Coronal sections of 3-monthold mice were immunostained with Satb2, a transcription factor expressed in intracortically projecting neurons, including UL, regulating their development, and with Ctip2, a transcription factor expressed in subcortically projecting neurons in DL, regulating their development [28,29]. The cortices of WT mice, consistently exhibited immunoreactivity to Satb2 glutamatergic neurons between layers 2 and 5 (Fig. 4A, B). Ctip2+ neurons were detected only in DL, where they overlapped with Satb2+ cells, as expected (Fig. 4A, B). Conversely, the number of Satb2+ neurons in homozygous R312H, and NULL cortices, was significantly decreased in the ULs (corresponding to bins 1-3 in Fig. 4B) and increased in DLs (bins 4-5), while Ctip2+ neurons were also present in the mid layers. Western blot analysis revealed significant decreases in the levels of Satb2 protein in the brains of homozygotes, whereas the levels of Ctip2 did not significantly vary across genotypes (Fig. 4C).

Synaptic functionality is impaired in homozygous R312H KI mice
The emergence of functional neocortical connectivity depends on neuronal development driving synaptic formation, and may be affected by neuron viability [30]. Accordingly, immunoreactivity to Synapsin-1 (SYN-1), a pre-synaptic marker was comparable to control in heterozygous sections, but was significantly decreased in the cortices and the hippocampi of homozygous and NULL animals (Fig. S5. See also the immunoblots in Fig. 5C) [31]. Representative examples of cortical sections co-stained with SYN-1 and PSD-95 (a post-synaptic marker), are illustrated in Fig. 5A [32]. We found lower numbers of SYN-1/PSD-95 costainings in the KI cortices compared to WT, suggesting a decrease in establishing functional synapses (Fig. 5B). Furthermore, the amounts of SYN-1 protein, total and phosphorylated at Ser605 (pSYN-1, a proxy for functional SYN-1 in vivo [33]), were significantly lower in KI and NULL brains compared to control (Fig. 5C, D). PSD-95 showed a trend toward decreased protein levels in homozygotes and NULL animals (Fig. 5E).
Overall, these findings implicate IKCs in neocortical synapses and circuits formation.

Golgi staining underscores morphology abnormalities in R312H neurons
Disrupted neuronal migration and synaptogenesis are likely to be associated with altered morphology of neocortical neurons.
To test this, we performed silver Golgi staining of neocortices (Fig. S6). We observed less neurons Golgi impregnated in UL, along with disrupted columnar organization of apical dendrites in homozygotes, NULL and to a lesser extent, heterozygotes. There was accumulation of Golgi impregnated cells in Layer I. Finally, significant numbers of neurons in DL (e.g., layer 5) had apical dendrite disoriented or reversed. Individual KI neurons exhibited shorter apical dendrite and excessive arborizations around the soma (Figs. 6A, S6). To characterize the morphology of KI pyramidal cells, we performed Sholl analysis [34]. The Total Dendrite Length (TDL) was decreased in all KI genotypes (Fig. 6B). The Sholl intersection profiles (SIPs) exhibited typical shape corresponding to the pyramidal cell-with the number of interactions peaking at a shorter distance from the soma (Fig. 6C). However, the SIPs of the KI neurons were broader compared to control, reflecting increased dendritic complexity and abnormal arborization arising from the soma [35]. Furthermore, the dendrites of homozygous neurons had significantly less dendritic spines than WT (Fig. 6D, E), consistent with the lower number of mature synapses evidenced by the SYN-1/PSD-95 co-immunostainings (Fig. 5). Homozygote dendritic spines had immature morphological characteristics-they were thinner compared to control spines, and did not exhibit the typical mushroom-like shape (Fig. 6D). The total number of spines was comparable between heterozygotes and WT, but the former were more immature (Fig. 6E). Together, these data underscore a correlation between the severity of neuroabnormalities in the KI brains and the extent of impaired neuronal migration and abnormal development during corticogenesis. The IKC signaling machinery is conserved in IKC R312H IKCs signal through integrins, which have an established importance in determining neocortical developmental processes [12]. Therefore, we checked major components of the IKC signaling machinery (Fig. S7), by Western blot or ELISA (Fig. 7, Table 1) [36]. Notably, the amounts of phosphorylated (a proxy for activated) protein, were comparable between WT and R312H heterozygous brains, but were significantly decreased in R312H homozygous and NULL (Fig. 7A). Representative coimmunoprecipitations of KCNB1 with adhesome proteins, Paxillin, Vinculin and Talin-1, that link integrins to the actin cytoskeleton, are illustrated in Fig. 7B, C [37]. Phosphorylation of Paxillin by FAK, allows the recruitment of Talin-1 and Vinculin, whose phosphorylation, in turn, assists the assembly process [38]. R312H subunits Fig. 4 Adult R312H brains exhibit severe neuroabnormalities. A Representative images of Satb2 (green), Ctip2 (red) and DAPI (blue) staining and of Satb2 and Ctip2 and DAPI overlaps in central cortex of 3-month-old mice of the indicated genotypes. Scale bar 500 μm. B Relative distribution of Satb2, or Ctip2 positive cells per bin. N = 6 brains/genotype. Two technical replicates/brain. C Representative western blot of Satb2 protein and of Ctip2 in the indicated 3 brains/genotype, and densitometric quantification. Loading controls: Bradford assay. *P < 0.05, ***P < 0.001, and ****P < 0.0001 (two-way ANOVA, Dunnett's post hoc).
retained the ability to interact with the adhesome proteins; however significantly lower amounts of phosphorylated protein co-immunoprecipitated with R312H in homozygous brains, compared to control (Fig. 7D). We conclude that IKC signaling is grossly impaired in homozygous and NULL brains.

IKCs enhance cell proliferation
Overall, our findings suggested that the impaired signaling of IKC R312H , was the culprit for the neuromorphological abnormalities of the KI brains. To test this idea, we assessed cellular processes regulated by integrins, namely proliferation, motility and neuritogenesis, in heterologous expression systems, that provide wellestablished assays [2,39,40]. Proliferation rates in Chinese hamster ovary (CHO) cells, transfected with WT or R312H cDNA or empty pCi-neo vector (mock), up to 2 days post transfection (dpt), are illustrated in Fig. S8A. The WT channel significantly increased the rate of proliferation, compared to mock, in agreement with the work of others (20000 ± 980 cells/day vs. 7400 ± 136 cells/day, P = 0.017) [41,42]. In contrast, R312H was less effective in enhancing proliferation (rate = 12500 ± 1440 cells/ day, P = 0.049 vs. WT). Integrins regulate proliferative responses primarily via the PI3K/Akt and the Ras-MAPK pathways, which were down-activated in homozygous, and NULL brains (Fig. 7A, Table 1) [43,44]. Therefore, we determined whether overexpressing components of these pathways (FAK, Src, RAF, Akt), could rescue proliferation. Co-transfection with R312H or mock cDNA strongly enhanced the proliferation of the cells (Fig. S8B). In contrast, co-transfection with the WT channel only moderately increased the rate of proliferation. Two enzymatically inactive mutants, K295R Src and K179M Akt, reduced the proliferation rates in all experimental conditions, while a constitutively active RAF mutant, S259A, was more effective than WT RAF in enhancing proliferation of the cells [45][46][47].

IKCs affect cell motility
The motility of CHO cells transfected with WT--assessed in the wound healing assay--was increased compared to mock and R312H, in agreement with previous reports (Fig. S8C) [2,39,41]. However, pharmacological enhancement of the enzymatic activity of MEK, by agonist Platelet Activating Factor C-16 (PAF C-16, Fig. S8C) or FAK, by agonist Angiotensin II (Ang II, Fig. S8D), showed a trend toward accelerating wound closure in cells transfected with R312H, while had negligible, or even negative, effects on the motility of cells expressing WT or mock [48][49][50][51]. To rule out that the pharmacological agents did not alter KCNB1 current, we recorded whole-cell currents in cells incubated in the absence/presence of Ang II, following the same protocols used for cell migration. In agreement with a previous study, R312H macroscopic currents were smaller, and activated at more depolarizing voltages, compared to WT (Fig. S8E, F). We cannot rule out that the shift in the voltage-dependence was due to the presence of integrins in the complex, but this possibility seems  unlikely, since the R to H replacement occurs in the voltage sensor domain) [2]. Most importantly, the recordings did not reveal any effect of Ang II on KCNB1 current.

IKCs enhance neuritogenesis
Neurites were longer in differentiated mouse neuroblastoma N2a cells expressing WT compared to mock, in agreement with previous reports (Fig. S9A) [2,52]. Soma hypertrophy was also noticeable in cells expressing the WT channel (Fig. S9B). Most importantly, R312H only moderately enhanced neurite outgrowth and soma hypertrophy [2]. Treatment with Ang II (Fig. S9A, B), or PAF C-16 (Fig. S9C, D), significantly boosted neurite outgrowth and soma hypertrophy in cells expressing R312H, to WT levels, but were fairly ineffective, or even negative, in the cells expressing WT or mock.
Angiotensin II corrects morphological defects of R312H primary neurons We next sought to determine whether some of the defects of native cells were recapitulated by cultured neurons, and whether impaired IKC signaling was an underlying cause. To better reproduce physiological conditions, we employed co-cultures of cortical neurons and glia, that we analyzed using Sholl analysis [53]. Thus, the TDL of heterozygous, and homozygous KCNB1+ neurons, were significantly shorter, roughly halved in the latter, compared to control (Fig. 8A). Incubation with 0.7 μM Ang II significantly increased the TDL of the heterozygotes by ∼25%, and the homozygotes by ∼50%, while had a slightly negative effect on WT neurons. Likewise for native neurons, also the SIPs of the primary neurons were broader than control SIP, reflecting dysregulated branching. Ang II treatment reversed the shape of the SIP to control (Fig. 8B).
Angiotensin II increases the synaptic connectivity of R312H primary neurons At DIV14, both heterozygous and homozygous neurons had less dendritic spines than WT neurons. Most importantly, Ang II treatment significantly increased the number of spines in both genotypes, and had slightly negative effect on control cells (representative images in Fig. S10 and quantifications in Fig. 9C). The number of functional synapses (SYN-1/PSD-95 co-staining) was decreased in KI neurons, reflecting the fact that their dendrites had less spines than control (Fig. 9A, B, D). Following a script well established, Ang II treatment increased the number of functional synapses in heterozygous and homozygous co-cultures -in the case of the former, to WT levels-consistent with the enhancing effect of Ang II on spine maturation, and had no effect on WT co-cultures.  [3,[54][55][56]. The inhibitors caused a significant decrease in the total number of mature spines Fig. 7 IKCs signaling is impaired in the R312H brain. A Representative Western blots of major components of IKC signaling (total and phosphorylated, or in the case of Ras, GTP-bound) from the brains of the indicated genotypes. Loading control: Bradford assay. B Representative co-IPs of KCNB1 (IP) with the indicated adhesome proteins. Samples were loaded to have approximately the same amount of KCNB1 protein per genotype. The same membrane was used to visualize Vinculin and Talin-1 (membrane stripping). C Representative co-IPs of KCNB1 (IP) with the indicated phosphorylated adhesome proteins: Paxillin phosphorylated at Tyr118 (pPax), Talin-1 phosphorylated at Ser425 (pTal), and Vinculin phosphorylated at Tyr100 (pVin). Samples were loaded to have approximately the same amount of KCNB1 protein per genotype. The same membrane was used to visualize pVinculin and pTalin-1 (membrane stripping). Controls: mouse IgG and empty beads. D Densitometric quantifications of 3-4 experiments as those shown in (C). *P < 0.05, **P < 0.01, and ***P < 0.001 (one-way ANOVA, Tukey's post hoc).
( Fig. S11A, B), and synaptic connections (Fig. S11C, D), compared to control. Treated cells presented typical dysregulated branching around the soma, reflected in broader SIPs (Fig. S11E, F). In contrast, the TDL did not change upon treatment with the inhibitors (not shown), a result that was not investigated further. In summary, Kcnb1 R312H primary neurons exhibited morphological defects that recapitulated those of native neurons. These abnormalities could be suppressed, or induced, by impinging on the IKC signaling machinery. Hence, these data implicate the nonionic functions of IKCs into the mechanisms underlying the morphology of glutamatergic cortical neurons.

DISCUSSION
To answer fundamental questions about the role of IKCs in prenatal brain development and their associated channelopathies, we constructed a Kcnb1 null mouse and a KI mouse harboring the Kcnb1 R312H gene variant, originally mapped in children with developmental and epileptic encephalopathies [57]. We report that NULL and R312H mice exhibited gross disorganization of the cortical layers, along with disrupted synaptic connectivity, frequent, spontaneous seizures, anxiety and compulsive behavior. R312H KCNB1 subunits formed complexes with α5β5 integrins, but in homozygotes the signaling of those macromolecules was grossly compromised. Notably, rescuing IKC signaling in vitro, through overexpression, or pharmacological activation of its components, suppressed the IKC R312H phenotype. Overall, these results underscore a key role of IKCs in prenatal neuronal migration, that they appear to control through non-ionic functions.
The R312H KI mouse may contribute to the debate in epilepsy research, concerning the potential epileptogenic role of brain defects [58]. The ionic current appears to be the primary culprit for the severe epileptiform activity of the mice, as heterozygotes did not exhibit appreciable neuroabnormalities. In contrast, the malformations observed in homozygotes appear to be originated  P < 0.001 for WT vs. heterozygous R312H; P < 0.001 for WT vs. homozygous R312H in control and not statistically significant (NSS) in the presence of Ang II (Kolmogorov-Smirnov test). *P < 0.05, ***P < 0.001, and ****P < 0.0001 (one-way ANOVA, Tukey's post hoc).
by impaired IKC signaling during prenatal development. Thus, these studies may provide mechanistic insight into the underlying causes of brain defects, often detected in DEE patients harboring KCNB1 gene mutations--that were poorly understood [7,57]. Nonetheless, the interplay between ionic and non-ionic functions is synergetic and intrinsically linked in IKCs [1,2]. Therefore, it is likely that also IKC current plays a role in the context of neurodevelopment. In fact, many of the mechanisms underlying neurodevelopment are both hard-wired in the neurons and activity-dependent [59][60][61]. For example, spontaneous electrical activity influences neural circuit development and is crucial for laying out early connectivity maps in many areas of the brain [62][63][64]. Thus, it is possible that during the formation of the neocortex, IKCs translate the spontaneous electrical activity of emergent circuits into biochemical signals that help guiding their formation. Future investigations will elucidate the potential role of IKC ionic functions in those mechanisms. In conclusion, this study establishes integrin-K + channel complexes as major players in prenatal brain development, and supports an argument that developmental channelopathies contribute to the etiology of DEEs.

DATA AVAILABILITY
All the data used for this study that are not presented in the figures, and any additional information required to reanalyze the data reported in this paper are available on request from the corresponding author. Mouse lines generated in this study will be deposited to the Mutant Mouse Resource & Research Centers of the NIH, upon publication of this study. This study did not generate new unique chemical reagents. Two technical replicates/culture. Data were analyzed with ImageJ/Fiji software and normalized to WT in control conditions. *P < 0.05; ***P < 0.001, and ****P < 0.0001 (one-way ANOVA, Tukey's post hoc).