Extra-Synaptic Functions of NRG1 Type I Signaling in Hippocampal γ-Oscillations.

Extra-Synaptic Functions of NRG1 Type I Signaling in Hippocampal γ-Oscillations

Abstract

Background and Hypothesis

The neurodevelopmental hypothesis of schizophrenia posits that early-acting genetic risk factors precipitate pathological neurodevelopmental processes that result in neural network dysfunctions during adolescence. Human gene variants of the epidermal growth factor-like ligand Neuregulin 1 (NRG1) and its tyrosine kinase receptor ErbB4 are putative risk factors in schizophrenia, and aberrant NRG1 signaling has been implicated in schizophrenia-relevant endophenotypes. However, insight into associated pathomechanisms remains limited.

Study Design

We modeled embryonic glutamatergic neuron-specific NRG1 type I overexpression in a conditional transgenic mouse line to investigate the impact of early hyperactive NRG1 type I signaling on neuronal network functions in the hippocampus in vivo.

Study Results

NRG1 type I transgenic mice display abnormal gamma-frequency network oscillations in the adolescent hippocampus, in the absence of neuroinflammation or changes in hippocampal morphology. Instead, we found extra-synaptic NRG1 type I accumulation, which was associated with ErbB4 hyperphosphorylation, an increased size and number of Kv2.1 potassium channel–marked endoplasmic reticulum (ER)-plasma membrane (PM) junctions, and a reduction of A-type potassium channel currents in pyramidal cells.

Conclusions

Together, these findings suggest a model in which early-onset extra-synaptic NRG1 type I accumulation at ER-PM junctions causes sustained changes in neuronal excitability, ultimately contributing to deficits in neuronal network oscillations as observed in various neuropsychiatric disorders.

Introduction

Schizophrenia is a highly heritable disorder, much of which is attributable to a large number of common genetic variants with small effect size.1 According to the neurodevelopmental hypothesis of schizophrenia, early genetic insults induce pathological processes in the developing brain. During adolescence this may result in disturbed neuronal networks that are more prone to produce psychotic symptoms, particularly when triggered by a second “hit”, eg, heightened psychological stress.2,3 This concept is supported by genome-wide association studies (GWAS) with human samples in which genetic risk was enriched in biological processes of neuronal development and synapse formation,4 as well as in regulatory DNA regions linked to transcriptional dysregulation of genes with neurodevelopmental functions.5,6 However, early-acting genetic risk factor–mediated pathomechanisms are largely unexplored, mostly due to the limited availability of suitable mouse models.

Human NRG1 and ERBB4 gene variants are putative risk factors in schizophrenia,7–9 and ErbB receptor signaling was identified as a schizophrenia–relevant signaling pathway in a protein network-based analysis of GWAS–derived genetic risk factors.10 While recent GWAS studies have not consistently replicated earlier findings, genetic, functional, and preclinical studies continue to support an important role for NRG1-ErbB4 signaling in psychosis-relevant endophenotypes, including neuroinflammation, synaptic and network dysfunctions, aberrant neuroimaging, and behavioral deficits.11–14

Altered oscillatory neuronal network activity has been repeatedly observed in neuropsychiatric disorders, including schizophrenia.15 NRG1 signaling to ErbB4, its main tyrosine kinase receptor in the central nervous system, is required for the development of glutamatergic, GABAergic, and dopaminergic neuronal networks,16 and altered γ-oscillation power in ErbB4 mouse mutants17,18 indicates a role of NRG1/ErbB4 signaling in regulating oscillatory neuronal network activity. Together with postmortem studies of increased expression of the NRG1 type I isoform and ErbB4 hyperphosphorylation in schizophrenia patients,19,20 these findings suggest that hyperactive NRG1/ErbB4 signaling could contribute to abnormal oscillatory activity in neuropsychiatric disorders. Notably, cortical expression of NRG1 type I peaks already during embryonic development,21 and embryonic, but not postnatal loss of ErbB4 in parvalbumin-expressing (PV+) interneurons was shown to disrupt patterned cortical activity in the adult brain.22 Together, these data indicate a pathomechanistic impact of early disturbances in NRG1 type I signaling on neuronal network functions, but this has not been studied in vivo.

In this study we show in a conditional transgenic mouse model that an embryonic onset of moderate NRG1 type I overexpression in glutamatergic neurons results in extra-synaptic NRG1 type I accumulation, increases the size and number of Kv2.1-marked ER-PM junctions, reduces A-type potassium channel currents and diminishes the power of hippocampal γ-oscillations at young adult stages.

Materials and Methods

Transgenic Mice

STOP-Nrg1 mice were generated as previously described.23 Mice were maintained on a C57Bl/6N background in compliance with animal policies of the Max-Planck-Institute of Experimental Medicine and approved by the German Federal State of Lower Saxony.

Fluorescence of Stop cassette-encoded EGFP was used for routine phenotyping of Stop-Nrg1 mice. Tail biopsies were examined under fluorescent light (488 nm) with a Leica MZ16F fluorescent stereomicroscope. For PCR-based genotyping, genomic DNA was isolated from tail biopsies according to the manufacturer’s manual (1-Step Tissue & Cells genomic DNA isolation kit; nexttec™) and amplified using primer 1 (5′-GGTGGCTATAAAGAGGTCATCAG-3′) and primer 2 (5′-CTCCTGGCTTTTCATCTCTTTCAA-3′) (Suppl. Figure 1A). Genotyping of NEX-Cre mice was performed as described.24

Biochemistry

Sucrose homogenization buffer (320 mM Sucrose, 10 mM Tris (pH 7.4), 1 mM NaHCO3 and 1 mM MgCl2) with protease and phosphatase inhibitors (Roche) was used for protein extraction as described.25 Proteins were separated on 8% SDS-polyacrylamide gels and blotted onto PVDF (Hybond-P, Invitrogen). Membranes were incubated with primary antibodies (Actin ms 1:1000 (Millipore), ErbB4 rb 1:1000 (Abcam), p-HER4/ErbB4 (Tyr1284) rb 1:1000 (Cell Signaling), AKT rb 1:1000 (Cell Signaling), p-AKT rb 1:1000 (Ser473) (Cell Signaling), p44/42 MAPK (Erk1/2) rb 1:1000 (Cell Signaling), p-p44/42 MAPK (Erk1/2) (Thr202/Tyr204) rb 1:1000 (Cell Signaling), HA ms 1:250 (Covance), Neuregulin-1a/b1/2 rb 1:500 (Santa Cruz)), overnight at 4°C. Membranes were washed 3× in 1× TBS-Tween and incubated in secondary antibodies (goat anti-ms-HRP, Dianova, 1:5000; goat anti-rb-HRP, Dianova, 1:5000) for 1 hour at RT, developed with Western Lightning Plus-ECL Oxidizing reagent plus and Enhanced luminol reagent plus solutions (Perkin Elmer Life Science, Inc.), and scanned using an Intas ChemoCam Imager. Densitometric analysis of band intensities was carried out using ImageJ software.

Cell Culture, RNA Extraction, cDNA Synthesis, and qPCR

Hippocampal and cortical cultures were prepared from C57BL6/J mice at P1-2 according to an optimized protocol.26,27 In brief, tissue was isolated and dissociated neurons were plated onto sterilized 18 mm coverslips coated with 250 μg/ml PLL and 50 μg/ml PDL. After 1 hour incubation at 37°C, Neurobasal-A medium (Gibco) containing 2% (v/v) B27 supplement, 1% (v/v) GlutaMAX® and 0.2% (v/v) penicillin–streptomycin was added. Cultures were maintained at 37°C and 5% CO2. One-third of the medium was changed every second day. On DIV12, cultures were treated with 50 mM KCl or ddH2O (control) for 6 hours. Subsequently, mRNA was purified using the RNeasy® Plus Mini Kit (Qiagen) and transcribed into cDNA using the SuperScriptIII First-Strand Synthesis System (reverse transcription) (Life Technologies) according to the manufacturer’s protocol.

qPCR was performed using the standard protocol for relative quantification of the 7500 Fast Real-Time PCR (Applied Biosystems). 10 μl reactions were set up with 4 μl DNA (5 ng/μl), 0.1 μl sense primer (10 pM), 0.1 μl antisense primer (10 pM), 5 μl GoTaq qPCR Master Mix (Promega), and 0.8 μl ddH2O. The following primers were used: F5′-TTCTCACCACTGCTCCACAG-3′, R5′-CTCCACGTGTGGC-TCTCAT-3′ (NRG2); F5′-ACTCAGCCACAAACAACAGA-AAC-3′, R5′- GAAGCACTCGCCTCCATT-3′ (NRG1 type III); F5′-GGAAGGGCAAGAAGAAGGAC -3′, R5′-CCTGGCTTT-TCATCTCTTTCA -3′ (NRG1 type I); F5′-CTGCTCTTTCCCA-GACGAGG-3′, R5′-AAGGCCACTTATCACCAGCC-3′ (β-actin). To assess the expression of NRG1 type II, 5 primer pairs were used F5′-CTGACAACTGGACCCTACTC-3′, R5′-TTTCACAGCCCACACCTG-3′ (Makinodan et al 2012), F5′- CCTCGTTTCCTCCGCTG-3′, R5′- GACTCCTGGCTTTTCA-TCTCTT-3′, F5′-GGGGGCTTGAAGAAGGACTC-3′, R5′-ACCTCCTTCTTGAGGTTGCG-3′, F5′- GAAGGACTCGCTA-CTCACCG-3′, R5′-CGGCTGACCTCCTTCTTGAG-3′, F5′-CGCCCTATCTGGTGAAGGTG-3′, R5′-CGGTGAGTAGCG-AGTCCTTC-3′. Ct values from quadruplicates were analyzed, and outliers were eliminated when the standard deviation was higher than to 0.5. Comparative quantification analysis was performed by the ΔΔCt method. The fold difference in expression (2-ΔΔCt) was determined. T-students were performed to compare the expression of each gene in treated cells and controls.

Immunocytochemistry

Hippocampal primary neurons cultured on PLL-coated coverslips, including KCl-treated and untreated control conditions, were fixed with 4% PFA for 20 minutes at RT, then washed three times with 1× PBS. Cells were permeabilized with 100% methanol for 5 minutes and incubated in blocking buffer (2% bovine-serum albumin, 2% horse-serum, 0.1% gelatin) 1 hour at RT. Coverslips were incubated overnight at 4°C with primary antibodies (Neurogranin, mouse, 1:500, Novusbio; Parvalbumin, mouse, 1:500, Swant) diluted in 10% blocking solution, followed by three 1× PBS washes. Afterwards, coverslips were incubated with secondary antibodies diluted in 5% blocking solution (αmouse-Cy3, Dianova; αrabbit-Alexa488, 1:750, Invitrogen) together with DAPI (300 nm) for 2 hours at RT. After final washes, coverslips were mounted with Aqua-Poly/Mount (Polysciences, Inc.) and stored at 4°C. Images were acquired using a Leica TCS SP2 (Leica Microsystems GmbH, Wetzlar, Germany).

Histology and Immunohistochemistry

Paraplast embedding of brain tissue and fluorescent immunostaining was performed as described.28 Cryo- and vibratome sections were blocked and permeabilized with 5 % horse-serum, 0.1 % Triton X-100 in 1× PBS for 30 minutes at RT. Primary antibodies (αCre, rabbit, 1:1000, Covance; αGFAP, rabbit, 1:500, Dako; αGFP, goat, 1:500, Rockland; αHA, mouse, 1:100, Biolegend; αHA, chicken, 1:100, Sigma Aldrich; αHA, rabbit, Cell Signaling, 1:400; αNeuN, mouse, 1:100, αChemicon; NRG1, rabbit, 1:100, Santa Cruz; αOlig2, rabbit, 1:200, John Alberta, Harvard; αKv2.1, mouse, 1:500, Abcam; IBA, rabbit, 1:700, Wako; VGlut, guinea pig, 1:100, Synaptic Systems; PSD95 mouse, 1:500, Neuromab; Gephyrin, mouse, 1:500, Synaptic Systems; Synaptophysin, mouse, 1:1000; Synaptic Systems; SERT, goat, 1:100, Santa Cruz; GM130, ms, 1:200, BD Biosciences; ChAT, chicken, 1:500, Aves Labs; Parvalbumin, rabbit, 1:200, Swant) were diluted in blocking solution and incubated overnight at 4°C. After three washing steps in 1× PBS at RT secondary antibodies diluted in blocking solution (αchicken-cy3, 1:1000; αchicken-Alexa488, 1:1000; αgoat-Alexa488, 1:500; αmouse-Alexa555, 1:500; αmouse-DL649, 1:250; αmouse-Cy2, 1:100; αmouse-Cy3, 1:1000; αmouse-Alexa 488, 1:1000; αrabbit-Alexa488, 1:100; αrabbit-Alexa555, 1:500; αrabbit-Cy2, 1:100; αrabbit-Cy3, 1:1000; αrabbit-Cy5, 1:250; αrabbit-DL649, 1:250; α-guinea pig-cy3 1:1000, α-guinea pig-cy5 1:250 from Dianova and Molecular Probes) and DAPI (300 nM) were incubated for 2 hours at RT. Slides were washed three times in 1× PBS (protected from light), mounted in Aqua-Poly/Mount (Polysciences, Inc.) and stored at 4°C.

Chromogenic immunostaining on paraplast embedded brain tissue was performed as for fluorescent immunostaining, except for inactivation of endogenous peroxidases by 5 minutes incubation with 3 % hydrogen peroxide prior to blocking. Sections were incubated overnight at 4°C with primary antibodies (GFAP rb 1:500 (Dako), IBA1 rb 1:700 (Wako), NeuN ms 1:100 (Chemicon), Parvalbumin 28 rb 1:200 (Swant)), followed by secondary antibody (biotinylated, Dako LSAB2) and streptavidin conjugated to Horseradish Peroxidase (Dako LSAB2). Stainings were developed by addition of DAB substrate solution, counterstained with Hematoxylin, dehydrated in an ascending alcohol series, and mounted in Eukitt® (O. Kindler). Images were acquired using a Leica TCS SP2 (Leica Microsystems GmbH, Wetzlar, Germany), inverted confocal-laser scanning microscope, Zeiss LSM 880 Airyscan confocal microscope, and a Zeiss AxioZ brightfield microscope (Carl Zeiss Jena, Germany). Images were processed using ImageJ and Adobe Photoshop software. 3D images of Kv2.1+ and NRG1+ signals were acquired using IMARIS ×64 v.10.0.1 software (Oxford Instruments).

Cell Counting of Chromogenic Stainings

Immunostainings on 5 μm thick coronal paraffin sections (bregma –1.3 or –1.7) were imaged as tile overviews at 10× magnification using a Zeiss Imager.Z1 microscope. Cell numbers were counted using the Cell Counter plugin for ImageJ software and expressed as cell densities (cells per area counted, in mm2) or number of positive cells per region. Dentate gyrus areas were measured using ImageJ software. Cell numbers or areas of both hemispheres were averaged.

Quantification of GFAP+, IBA1+, and Kv2.1+ Areas

GFAP+ and IBA1+ areas were quantified on chromogenic immunostainings of coronal brain sections with a semi-automatic analysis using ImageJ software as described.29 Overview pictures (10×) were acquired with the Zeiss Imager.Z1 microscope, and ratio of DAB-positive versus total area was quantified in regions of interest (ROI). Area fractions from both hemispheres were averaged. Kv2.1+ area fraction and number of Kv2.1+ puncta were quantified using fluorescent immunostainings of frozen coronal brain sections. Images were acquired with a 40×/1.0 or 63× DIC objective of a Zeiss LSM 880 Airyscan confocal microscope and examined by semi-automatic analysis using a custom-made macro in FIJI software. The number of Kv2.1+ puncta was normalized to the number of DAPI+ cells in designated ROI.

Electrophysiology

Mice (P30) were anesthetized and sacrificed using isoflurane. Brains were removed and 300 μm thick transverse brain sections were prepared from both hemispheres using a Leica VT1200S vibratome in ice-cold sucrose-based Slicing Solution (230 mM Sucrose, 26 mM NaHCO3, 2 mM KCl, 1 mM KH2PO4, 1 mM MgCl2, 10 mM Glucose, 0.5 mM CaCl2) saturated with carbogen gas (95% oxygen, 5% carbon dioxide). Brain slices were transferred into a holding chamber filled with Artificial Cerebrospinal Fluid (ACSF; 120 mM NaCl, 26 mM NaHCO3, 1 mM KH2PO4, 2 mM KCl, 2 mM MgCl2, 10 mM Glucose, 2 mM CaCl2 saturated with carbogen gas) and incubated for at least 20 minutes before recording. An interface recording chamber (BSCBU Base Unit with the BSC-HT Haas Top, Harvard Apparatus) with a constant flow of ACSF at 33°C was used to record γ-oscillations in the CA3 region of the hippocampus. Extracellular recording electrodes (with a resistance between 2.0-3.0 MΩ) filled with ACSF solution were placed on the CA3 pyramidal cell layer of slices perfused constantly with the same solution. Baseline recordings were performed before kainate application for 10 minutes. Oscillatory recordings were performed for 20 minutes upon perfusion with 100 nM kainate (in ACSFsolution). In order to find the maximum oscillations, electrodes were then re-positioned and slices were recorded for additional 20 minutes in the presence of kainate. To assess the effect of PD158780 (ErBb receptor inhibitor, MedChemExpress, HY-18609) on γ-oscillations, baseline recordings were performed for 20 minutes under ACSF perfusion. Oscillations were induced by 100 nM kainate and recorded for 1 hour to achieve stable maximum γ-oscillations. Subsequently, 100 uM of PD158780 was applied in ACSF for 30 minutes, followed by a 20 minutes washout period. Signals were acquired using a Multiclamp 700B amplifier and Digidata 1440A data acquisition system (Axon Instruments, Molecular Devices), with 3 kHz Bessel filter. Data were analyzed using Axograph X software (Axon Instruments). The power spectra were calculated for 10 minutes recording epochs. Baseline power spectra were subtracted from kainate-induced oscillatory spectra, and peak frequency and average power were determined within the 25-45 Hz frequency band.

Voltage–dependent potassium currents were measured in voltage-clamp condition at –70 mV using the HEKA EPC-10 amplifier and Patch-Master (Ver. 2X73.5) software (HEKA electronics). The internal patch pipette solution contained 138 mM K-gluconate, 16.8 mM HEPES, 10 mM NaCl, 1 mM MgCl2·6 H2O, 0.25 mM EGTA, 4 mM ATP-Mg2+, 0.3 mM GTP–Na+ (adjusted to pH 7.4, ∼320 mOsmol); the external solution adds TTX (0.3 μM) to ACSF solution. Resistance of recording pipette filled with internal solution was 3–4 MΩ. Membrane current recordings were collected after adjustment of capacitance and series resistance compensation. Serial resistance was compensated by ~50% and only cells with serial resistance below 10 MΩ were analyzed. IK currents were evoked by 500 ms depolarizing test pulsed from –80 mV to +70 mV in 10 mV increment after –70 mV. IDR was evoked by same voltage steps with a 1000 ms preconditioning pulse of –40 mV to inactivate IA. IA were isolated by digital subtraction of currents between two specific voltage-clamp protocols (IK-IDR). Data were analyzed using Axograph (Ver.1.8.0) software (Axograph Scientific). Experiments using Guangxitonxin-1E (GxTX; Alomone Laboratories) were performed to isolate Kv2.1 currents in CA3 pyramidal neurons. IK currents were evoked by 500 ms depolarizing test pulses from –80 mV to +70 mV in 10 mV increment from a holding potential of –70 mV in ACSF containing 0.3 μM TTX in presence or absence of 1 μM GxTX. Kv2.1-mediated current (IKv2.1) were isolated by subtracting the currents recorded in the presence of GxTX from total IK currents.

Potassium current density (pA/pF) was normalized to cell membrane capacitance (pF) from peak outward current amplitude (pA). I-V curves were generated by plotting current density values against test potential. Chord conductance of potassium currents was calculated from “G = I/(Vm-Vrev)”, where I is current amplitude, Vm is test potential, and Vrev is virtual reversal potential. Chord conductance values for each test potential were normalized with maximal chord conductance values (Gmax) and then average normalized conductance values (G/Gmax) were plotted against test potentials fitted with Boltzmann equation built in Prism program.

Spontaneous miniature excitatory/inhibitory postsynaptic currents (mEPSC/mIPSC) were recorded from CA3 pyramidal neurons in mouse hippocampal brain slices during continuous perfusion with ACSF containing 0.3 μM TTX to block AP firing. In whole cell patch clamp configuration, spontaneous mEPSC and mIPSC signals were recorded at a holding potential of –70 mV without any stimulation for 3-5 minutes. To the same slice to examine GxTX effects on spontaneous miniature responses, 1 μM GxTX was additionally applied to the external solution for 10-15 minutes in the same slice. Spontaneous mEPSC or mIPSC events were detected using template-matching algorithm in AxoGraph software (Axograph Scientific).

Results

Neuronal activity has been implicated in the regulation of NRG1 expression and signaling functions.21,30,31 In line with a previous study,21 we detected NRG1 isoform-specific transcriptional responses upon neuronal stimulation of primary hippocampal and neocortical neurons by KCl treatment. Specifically, we confirmed by qRT-PCR elevated expression of NRG1 type I transcripts, while NRG1 type III expression was not affected, and expression of NRG2, a highly related neuregulin, was reduced under these conditions (Figure 1A; Suppl. Figure 2A, B). Fluorescent in situ hybridization (FISH) demonstrated increased NRG1 expression in a majority of KCl-treated cultured neurons (Suppl. Methods, Suppl. Figure 2C, D), most of which expressed the glutamatergic marker neurogranin (Suppl. Figure 2E). This finding is consistent with FISH analysis of adult brain sections that detected NRG1 type I and type III transcripts in cortical layer VI and the pyramidal cell layer in CA3 (Suppl. Methods, Suppl. Figure 2G, H). These data prompted us to model stage- and glutamatergic neuron-specific NRG1 type I overexpression in vivo, using a conditional transgenic mouse line (Stop-Nrg1) for Cre recombinase-mediated expression of N-terminally haemagglutinin (HA) epitope-tagged NRG1 type I23 (Figure 1B, C). Based on green fluorescent protein (GFP) expression from a loxP sites-flanked (“floxed”) “Stop” element and immunostaining for neural cell-specific markers, we verified the suitability of this mouse line for Cre-mediated expression in various neural cell types, including glutamatergic neurons (Suppl. Figure 1A, B; Suppl. Figure 3). NRG1 type I expression starts in the embryonic neocortex,21,32 consistent with signaling functions in early brain development. Thus, to model elevated NRG1 type I signaling with an onset during embryonic cortical development, we bred Stop-Nrg1 mice to NEX-Cre driver mice, in which Cre-dependent recombination is initiated around embryonic day (E) 11.5 exclusively in postmitotic glutamatergic projection neurons24 (Figure 1C). Double transgenic NEX-Cre*Stop-Nrg1 mice (referred to as NC-Nrg1 mice) were born at expected mendelian frequency and displayed normal lifespan, cage behavior, litter size, and body weight (Suppl. Figure 4A). Immunostaining demonstrated transgenic NRG1 type I expression (and a concomitant loss of GFP) in cortical plate neurons from NC-Nrg1 mice at E16 (Figure 1D, E). NRG1 type I expression faithfully mimicked NEXCre recombination patterns in the hippocampus24 (Figure 1F), and co-immunostaining confirmed the absence of NRG1 type I expression from parvalbumin (PV)+ GABAergic interneurons in the adult brain (Figure 1G).

Biochemical analysis verified NEXCre-dependent expression of full-length NRG1 type I in the neocortex of NC-Nrg1 mice (Figure 2A), and co-immunostaining for N- and C-terminal domains of NRG1 type I (Figure 2B) was consistent with proteolytic processing in the juxtamembrane stalk region.33 Next, we compared NGR1 type I expression, ErbB4 receptor activation, and downstream signaling in NC-Nrg1 mice to a “constitutive” transgenic mouse line with prominent NRG1 type I overexpression in hippocampus and neocortex under control of the neuronal Thy1.2 promoter (T-Nrg134; Suppl. Figure 4B). Biochemical analysis of adult brain revealed a ~ 6-fold higher NRG1 type I overexpression in adult T-Nrg1 mice compared to NC-Nrg1 mice (Suppl. Figure 4C). To investigate ErbB4 stimulation, we employed an antibody directed against phosphorylated Tyr1284, a well-characterized marker of NRG1–mediated ErbB4 activation by auto-phosphorylation and downstream signaling via adaptor protein Shc,35 upstream from both MAP and AKT kinase pathways.36 Consistent with higher NRG1 type I overexpression in T-Nrg1 mice, ErbB4 phosphorylation at Tyr1284 was ~4-fold higher in T-Nrg1 mice compared to NC-Nrg1 mice (Figure 2C, D). In line with the embryonic onset of Cre-mediated recombination, NRG1 type I overexpression was detectable in NC-Nrg1 mice at P7 (Suppl. Figure 4D) and associated with ~1.5-fold ErbB4 hyperphosphorylation (Figure 2E). NRG1 type I overexpression and ErbB4 hyperphosphorylation (~3-fold) were maintained in adult NC-Nrg1 mice (Suppl. Figure 4E; Figure 2F).

Sustained ErbB4 hyperstimulation was linked to increased MAP kinase phosphorylation in T-Nrg1 mice but not in NC-Nrg1 mice (Figure 2G, H). In contrast, AKT phosphorylation was significantly increased both in NC-Nrg1 and T-Nrg1 mice, with a greater increase observed in T-Nrg1 mice (Figure 2I, J). These findings imply sustained hyperstimulation of AKT signaling as a possible component of molecular pathophysiology in response to moderate NRG1 type I overexpression.

T-Nrg1 mice display an elevated hippocampal expression of inflammation-associated markers, including GFAP, and an enlarged dentate gyrus.37 In contrast, we found no evidence for astrogliosis or activated microglia in the hippocampus and fimbria of NC-Nrg1 mice at 16 weeks and even 1 year of age (Figure 2K, L; Suppl. Figure 4F-K). In addition, the number of NeuN+ cells in CA1-3 and neocortex (Suppl. Figure 4L, M, O-R), and both infra- (Figure 2M, N) and supra-pyramidal dentate gyrus areas (Suppl. Figure 4N) were not altered in NC-Nrg1 mice. We conclude that NC-Nrg1 mice are a physiologically relevant model to examine the effects of moderately increased NRG1 type I signaling activity during embryonic development on juvenile brain functions.

Treatment of hippocampal slices with the recombinant EGF domain of NRG1 increases the power of kainate-induced γ-oscillations, proposedly acting via ErbB4 receptors in PV+ interneurons.17 Adhering to the protocol used for testing the EGF domain,17 we induced γ-oscillations in hippocampal slices at P30 using 100 nM kainate and performed extracellular field recordings in the stratum pyramidale of CA3 (Figure 3A). In contrast to acute EGF domain treatment, sustained NRG1 type I overexpression in NC-Nrg1 mice reduced the power of kainate-induced γ-oscillations, whereas γ-oscillation frequency was not altered (Figure 3B-E; Suppl. Table 1). To rule out a possible effect of NEX (Neurod6) haploinsufficiency in “knock in” NEXCre driver mice, we compared NEXCre mice to wildtype, but found no difference in γ-oscillation power or frequency (Figure 3F-H; Suppl. Figure 5A-B). Considering the potency of EGF domain treatment in ErbB4 receptor activation in vivo,38 we additionally tested NC-Nrg1 mice homozygous for the Stop-Nrg1 cassette (NC-Nrg1hom;  Suppl. Figure 5C). However, γ-power was also reduced in NC-Nrg1hom mice (Suppl. Figure 5D-G), suggesting that increasing NRG1 type I signaling levels in vivo is unlikely to switch NRG1 type I functions in hippocampal γ-oscillations from inhibitory to stimulatory.

To test for a role of ErbB signaling in diminished γ-oscillation power in NC-Nrg1 mice, we repeated recordings in the presence of PD158780, a pan-specific antagonist of ErbB receptor tyrosine kinase activity (Figure 3I). PD158780 treatment had no effect on γ-frequency, but increased γ-power in controls and restored γ-power in NC-Nrg1 mice to PD158780-treated control levels (Figure 3J, K). As oscillatory activity in the γ-band range critically depends on ErbB4-expressing PV+ interneuron functions, we next addressed possible changes in PV+ interneurons in NC-Nrg1 mice. Chromogenic immunostainings revealed no differences in the number or localization of these cells in the hippocampus at P30 (Figure 3L-N), as well as hippocampus and neocortex of NC-Nrg1 mice at 16 weeks and 52 weeks of age (Suppl. Figure 6). Moreover, in contrast to glutamatergic pyramidal cells (see below), we observed no changes in potassium currents, rheobase or AP firing properties in PV+ interneurons from NC-Nrg1 mice compared to controls (Figure 3O-R; Suppl. Figure 5H-L), suggesting unaltered PV+ interneuron excitability.

In search for possible glutamatergic neuron-intrinsic deficits that could contribute to diminished γ-oscillations under chronic conditions in NC-Nrg1 mice, we performed immunostainings for NRG1 and a battery of synaptic markers. These studies revealed distinct clusters of NRG1 type I protein in the somatodendritic compartment of glutamatergic neurons from NC-Nrg1 and T-Nrg1 mice (Figure 4A, B; Suppl. Figure 7A). Notably, we observed no overlap of NRG1 type I with pre- and postsynaptic compartments of glutamatergic, GABAergic, and serotonergic synapses (Figure 4B; Suppl. Figure 7B-H). We recently showed that NRG1 type I accumulates in the PM of spinal α-motoneurons atop of ER–derived subsurface cisterns.39 Thus, we performed co-immunostainings for NRG1 and Kv2.1, a potassium channel that forms micron-sized clusters at homologous ER-PM junctions of cortical neurons.40 This revealed a partial overlap of NRG1 type I with Kv2.1-marked ER-PM junctions in NC-Nrg1 and T-Nrg1 mice (Figure 4C, D), consistent with previous findings in cultured hippocampal neurons.41 To distinguish NRG1 type I accumulation in ER-PM junctions from overexpression-induced “clogging” in the Golgi complex, we carried out co-immunostaining for Golgi marker GM130, but found no evidence for Golgi–associated NRG1 type I expression in T-Nrg1 mice (Suppl. Figure 7I).

Next, we investigated if increased extra-synaptic NRG1 type I accumulation in transgenic mice has an impact on ER-PM junctions. Immunostaining revealed that CA1 and CA3 pyramidal neurons in T-Nrg1 mice displayed an increase in the number and size of Kv2.1 clusters compared to wildtype (Figure 4E-G; Suppl. Figure 7J, K), without affecting pyramidal cell density in CA3 (Suppl. Figure 7L). A corresponding analysis confirmed that also a lower level of NRG1 type I overexpression in NC-Nrg1 mice increases Kv2.1 cluster number and size in CA3 pyramidal neurons (Figure 4H-J), without affecting CA3 pyramidal cell density (Suppl. Figure 7M). Examination of individual CA3 pyramidal cells in NC-Nrg1 mice confirmed an increased average size of Kv2.1 clusters that overlap with NRG1 type I compared to Kv2.1 clusters that are non-overlapping with NRG1 type I (Figure 4K-M). We also noted a large variability in size for NRG1 type I-overlapping Kv2.1 clusters (Figure 4M), consistent with the notion that NRG1 type I may serve “seeding” functions to promote dynamic changes in Kv2.1-marked ER-PM junctions.

The voltage–gated potassium channel Kv2.1 produces delayed rectifying potassium currents, but also serves non-conducting functions in ER-PM junctions40 that may impact on hippocampal network–relevant pyramidal cell properties. Electrophysiological recordings in CA3 pyramidal neurons revealed that total potassium current (IK) was reduced in NC-Nrg1 mice at depolarized potentials, whereas delayed rectifying current (IDR) was unaffected (Figure 5C). In contrast, A-type current (IA), isolated by voltage-step subtraction, was significantly reduced at depolarizing voltages (Figure 5A-D). Notably, the conductance–voltage (G/Gmax) curves were superimposable between NC-Nrg1 mice and controls (Suppl. Figure 8B), indicating preserved voltage dependence of channel activation and inactivation, and that IA reduction likely reflects altered channel expression or membrane trafficking rather than changes in gating properties. Pyramidal cell capacitance was equivalent between control and NC-Nrg1 mice (Suppl. Figure 8C), ruling out differences in somatic size as a contributor to reduced current density. To test a role of Kv2.1-mediated currents in diminished IA, we applied the Kv2–specific blocker Guangxitoxin-1E (GxTX) to hippocampal slices. Baseline recordings in response to depolarizing voltage-step pulses before GxTX treatment confirmed a reduction of total IK in CA3 pyramidal neurons from NC-Nrg1 mice. Following GxTX application, recordings from the same pyramidal cells showed a similar reduction of IK + GxTX currents in controls and NC-Nrg1 mice, ie, GxTX treatment had no ameliorating effect on IA in NC-Nrg1 mice (Suppl. Figure 8D). These findings indicate that Kv2.1-mediated delayed rectifying currents have no major role in diminishing IA in pyramidal cells of NC-Nrg1 mice.

Because IA sets action potential (AP) threshold and limits repetitive firing, we next examined whether its reduction translates into changes in pyramidal neuron excitability. Current-clamp recordings from CA3 pyramidal neurons revealed a lower rheobase, an increased number of evoked APs, a greater AP number at 150 pA current injection, and a shorter latency to the first AP (Figure 5E-H). These findings were further supported by representative recordings from CA3 pyramidal neurons, which showed more robust AP discharge at rheobase and suprathreshold stimuli in NC-Nrg1 mice (Suppl. Figure 8E). A corresponding analysis in CA1 pyramidal cells demonstrated a reduction in total, A-type, and even delayed rectifier potassium currents (Figure 5I-L), demonstrating that suppression of potassium currents, particularly IA, is a common feature of hippocampal pyramidal neurons, and consistent with a general increase in pyramidal cell excitability and altered spike timing coherence in the hippocampus of NC-Nrg1 mice.

In summary, we found no evidence for NRG1 type I expression in presynaptic glutamatergic terminals, even under conditions of prominent overexpression in transgenic mice. Instead, we suggest a working model (Figure 5M), according to which NRG1 type I accumulates in the somatodendritic compartment of glutamatergic neurons outside postsynaptic sites, with a substantial fraction of NRG1 type I located to ER-PM junctions. Chronically elevated NRG1 type I expression in glutamatergic neurons impairs hippocampal γ-oscillations, as a consequence of ErbB4 hyperstimulation in PV+ interneurons and maladaptive remodeling in glutamatergic networks, including altered ER-PM junctions, decreased A-type potassium currents, hyperexcitability, and reduced spike timing coherence (Figure 5M).

Discussion

Insight into pathological processes linked to individual genetic risk factors of schizophrenia is limited and requires the interrogation of adequate mouse models. We provide evidence that moderate NRG1 type I overexpression in NC-Nrg1 mice causes persistent ErbB4 receptor hyperstimulation in the brain without affecting hippocampal morphology or inducing neuroinflammation, in stark contrast to “conventional” T-Nrg1 mice.37 We conclude that NC-Nrg1 mice are a physiologically more relevant model to examine how developmental hyperactivity of a “genetic risk pathway” translates into neuronal network dysfunctions with relevance for neuropsychiatric disorders.

Our electrophysiological examination of NC-Nrg1 mice revealed a reduction in the power of kainate-induced γ-oscillations in the hippocampal CA3 region. This suggests abnormal oscillatory network activity in adolescent NC-Nrg1 mice as a consequence of developmental hyperactivity of NRG1 type I signaling and matches increased γ-power in mouse mutants with ErbB4 disruption in PV+ interneurons.18 However, it contrasts previous studies, in which acute treatment with the soluble EGF domain of NRG1 (NRG1β increased kainate-induced γ-power in CA3).17 These discrepancies suggest differences in mechanism of action between acute exposure to extremely potent signaling activity of recombinant NRG1β versus chronic signaling of full-length NRG1 type I to ErbB4 receptors, which may induce long-term homeostatic and maladaptive changes in hippocampal networks. Of note, T-Nrg1 mice display a reduced frequency, but normal power of carbachol-induced γ-oscillations in CA3,42 pointing to distinct network dysfunctions in response to significantly higher NRG1 type I expression in T-Nrg1 mice.

What are potential pathomechanisms for reduced γ-oscillation power in NC-Nrg1 mice? NRG1-mediated ErbB4 signaling in PV+ interneurons regulates the development of GABAergic circuitry43 and cortically expressed NRG1 type I serves as an attractant for cortically migrating PV/ErbB4+ interneurons,32 suggesting that embryonic NRG1 type I/ErbB4 hyperstimulation could impair PV+ interneuron development in NC-Nrg1 mice. We observed unaltered numbers and location of PV+ interneurons in the hippocampus and neocortex of NC-Nrg1 mice. However, glutamatergic neuron-specific overexpression of cysteine-rich domain-containing NRG1 diminishes inhibitory functions44 and PV+ interneurons from T-Nrg1 mice exhibit reduced NMDA receptor-mediated currents.45 Thus, more subtle developmental changes in PV+ interneuron functions may add to impaired γ-oscillations in NC-Nrg1 mice. Importantly, we show that acute pharmacological blockade with a panErbB inhibitor fully recovered γ-power in hippocampal slices from NC-Nrg1 mice, in line with enhanced oscillatory activity after ErbB4 deletion in PV+ interneurons.18 This indicates that NRG1 type I/ErbB4 signaling mediates “constraining” functions that counteract γ-oscillatory activity, eg, via regulation of PV+ interneuron excitability through potassium channel Kv1.1.46 However, as we found no evidence for altered PV+ interneuron excitability or inhibitory drive on pyramidal cells in NC-Nrg1 mice, PV+ interneuron-intrinsic (patho)mechanisms downstream from NRG1 type I/ErbB4 hyperstimulation require further investigation.

Optogenetic interference with pyramidal cell excitability reduces hippocampal γ-power,47 in line with pyramidal cell-intrinsic changes in excitability that could contribute to altered γ-oscillations under chronic conditions in NC-Nrg1 mice. This notion is supported by a glutamatergic hypofunction in mice with CamKII/tTA-dependent postnatal NRG1 type I overexpression, which was reported to result from pyramidal cell-intrinsic, ErbB4-independent processes.48 Accordingly, we found no evidence for NRG1 type I expression in presynaptic glutamatergic terminals, calling into question “point-to-point” presynaptic signaling to postsynaptic ErbB4 receptors in PV+ interneurons as a major driver of abnormal γ-oscillations under chronic conditions. Instead, we demonstrate extra-synaptic NRG1 type I accumulation at Kv2.1-marked ER-PM junctions, which is congruent with findings in NRG1 type I-transfected hippocampal neurons. 41 Similar to effects of NRG2 overexpression in cultured ErbB4+ interneurons,49,50 we found that elevated NRG1 type I expression in NC-Nrg1 mice was associated with an increase in number, size, and possibly structural dynamics of somatic Kv2.1 puncta. In line with biochemical studies on NRG2,50 we speculate that NRG1 type I and Kv2.1 share the same tethering mechanisms at ER-PM junctions via molecular interactions with the ER scaffolding protein VAP (VAMP-associated protein), thereby tuning their activity in response to changes in excitatory activity. Importantly, aggregated KV2.1 channels most likely serve non-conducting functions.51 Thus, an increased number or size of Kv2.1 puncta in NC-Nrg1 mice is not necessarily linked to elevated Kv2.1-mediated currents. Rather, we suggest that aggregated NRG1 type I and Kv2.1 channels organize ER-PM junctions, including clustering and functional coupling of voltage-gated L-type Ca2+ channels in the PM with ER-located ryanodine receptors, thereby linking neuronal activity with local Ca2+ signaling.52–54 Moreover, ryanodine receptor activation has been associated with PKA-mediated downregulation of A-type potassium channel currents in hippocampal neurons,55,56 providing a possible mechanism for reduced A-type potassium currents in pyramidal neurons from NC-Nrg1 mice. Ultimately, diminished A-type potassium currents may result in pyramidal cell hyperexcitability and loss of coherent spike timing, with relevance for the expression of γ-oscillatory activity.

We recently reported an altered structure of specialized ER-PM junctions in spinal cord motor neurons and their associated cholinergic C-boutons in NRG1 type I and type III transgenic mice,39,57 consistent with a role for NRG1/Kv2.1 interactions at ER-PM junctions in the regulation of α-motorneuron excitability. Finally, we note that NRG1 type I overexpression using NEX-Cre driver mice occurs also outside the hippocampus, mainly in glutamatergic neurons of the neocortex, and in a smaller number of mid- and hindbrain neurons.24 Based on our observations of extra-synaptic NRG1 type I accumulation in neocortical glutamatergic neurons, we speculate that altered excitability and abnormal γ-power may also a feature in the neocortex of NC-Nrg1mice.

Taken together, we propose a working model in which NRG1 type I accumulation in pyramidal cells causes sustained ER-PM junction remodeling, including an increased (indirect) Kv2.1 recruitment, which in turn diminishes A-type potassium channel functions. NRG1 type I accumulation may also affect other ER-PM junction-associated ion channels with important functions in neuronal excitability and γ-oscillations, eg, calcium-gated (SK) potassium channels.58 Ultimately, this could interfere with optimal reciprocal coupling between pyramidal cells and PV+ basket cells, adding to impaired γ-oscillation power. Exactly how NRG1 type I accumulation alters ER-PM junction remodeling is currently unclear. Formally, ErbB4 hyperphosphorylation in NC-Nrg1 mice implies paracrine NRG1 type I signaling to PV+ interneuron ErbB4 receptors.41 However, a recent study reported a reduction of Nav1.1 sodium channel currents in transgenic mice with NRG1 type I overexpression in GABAergic interneurons in the absence of elevated ErbB4 stimulation.59 Thus, extra-synaptic NRG1 type I accumulation at ER-PM junctions may primarily act by providing a platform for (abnormal) protein/protein interactions and/or nuclear back signaling via its cytoplasmic domain.60

A remarkable convergence of abnormal NRG1/ErbB4 expression, PV+ interneuron deficits, and disordered oscillatory network activity has been noted in several neurological disorders, including schizophrenia, multiple sclerosis, epilepsy, and Alzheimer’s disease.61–64 We propose the concept of a NRG1-mediated “signalopathy” as a common pathophysiological outcome, in which (NRG1-dependent and -independent) PV+ interneuron deficits drive neuronal hyperexcitability. This could result in a vicious cycle of NRG1 accumulation at ER-PM junctions, possibly amplified by human NRG1 gene variants prone to increased Nrg1 transcription,65 and, ultimately, abnormal γ-oscillations. Aberrant oscillatory network activity may serve as a sensitive biomarker during prodromal neuropsychiatric disease stages that could be read out by non-invasive EEG recordings.66 Thus, pathomechanistic insight into NRG1-mediated neuronal network dysfunctions will contribute to better diagnosis and treatment strategies in neuropsychiatric disorders.

Supplementary Material

References

1.Owen  MJ , Legge  SE , Rees  E , Walters  JTR , O’Donovan  MC . Genomic findings in schizophrenia and their implications . Mol Psychiatry . 2023 ;39 :3638 –3647 . 10.1038/s41380-023-02293-8 PMC10730422 37853064

2.Birnbaum  R , Weinberger  DR . Genetic insights into the neurodevelopmental origins of schizophrenia . Nat Rev Neurosci . 2017 ;18 :727 –740 . 10.1038/nrn.2017.125 29070826

3.Schmitt  A , Falkai  P , Papiol  S . Neurodevelopmental disturbances in schizophrenia: evidence from genetic and environmental factors . J Neural Transm . 2023 ;130 :195 –205 . 10.1007/s00702-022-02567-5 PMC9660136 36370183

4.Trubetskoy  V , Pardiñas  AF , Qi  T , et al.  Mapping genomic loci implicates genes and synaptic biology in schizophrenia . Nature . 2022 ;604 :502 –508 . 10.1038/s41586-022-04434-5 PMC9392466 35396580

5.Walker  RL , Ramaswami  G , Hartl  C , et al.  Genetic control of expression and splicing in developing human brain informs disease mechanisms . Cell . 2019 ;179 :750 –771.e22 . 10.1016/j.cell.2019.09.021 PMC8963725 31626773

6.Schork  AJ , Won  H , Appadurai  V , et al.  A genome-wide association study of shared risk across psychiatric disorders implicates gene regulation during fetal neurodevelopment . Nat Neurosci . 2019 ;22 :353 –361 . 10.1038/s41593-018-0320-0 PMC6497521 30692689

7.Agim  ZS , Esendal  M , Briollais  L , et al.  Discovery, validation and characterization of Erbb4 and Nrg1 haplotypes using data from three genome-wide association studies of schizophrenia . PLoS One . 2013 ;8 :e53042 . 10.1371/journal.pone.0053042 PMC3536812 23301017

8.Pardiñas  AF , Holmans  P , Pocklington  AJ , et al.  Common schizophrenia alleles are enriched in mutation-intolerant genes and in regions under strong background selection . Nat Genet . 2018 ;50 :381 –389 . 10.1038/s41588-018-0059-2 PMC5918692 29483656

9.Mostaid  MS , Mancuso  SG , Liu  C , et al.  Meta-analysis reveals associations between genetic variation in the 5’ and 3’ regions of Neuregulin-1 and schizophrenia . Transl Psychiatry . 2017 ;7 . 10.1038/tp.2016.279 PMC5545738 28094814

10.Chang  X , de A  LL , Liu  Y , et al.  Common and rare genetic risk factors converge in protein interaction networks underlying schizophrenia . Front Genet . 2018 ;9 :1 –10 . 10.3389/fgene.2018.00434 PMC6172705 30323833

11.Shen  J , Xiao  C . Spatiotemporal dynamics of RERE in schizophrenia pathogenesis: insights from multi-omics and single-cell sequencing . Schizophrenia (Heidelb) . 2025 ;11 :155 . 10.1038/S41537-025-00705-Y PMC12743055 41298505

12.Dong  J , Chen  W , Liu  N , Chang  S , Zhu  W , Kang  J . NRG1 knockdown rescues PV interneuron GABAergic maturation deficits and schizophrenia behaviors in fetal growth restriction mice . Cell Death Discov . 2022 ;8 :476 . 10.1038/S41420-022-01271-3 PMC9718849 36460658

13.Chung  Y , Bazmi  HH , Lewis  DA , Chung  DW . Disrupted ErbB4 splicing with region-specific severity across the cortical visuospatial working memory network in schizophrenia . Cereb Cortex . 2026 ;36 :bhag045 . 10.1093/cercor/bhag045 42030198

14.Kiemes  A , Navacerrada  MES , Kim  E , et al.  Erbb4 deletion from inhibitory interneurons causes psychosis-relevant neuroimaging phenotypes . Schizophr Bull . 2023 ;49 :569 –580 . 10.1093/schbul/sbac192 PMC10154722 36573631

15.Uhlhaas  PJ , Singer  W . Abnormal neural oscillations and synchrony in schizophrenia . Nat Rev Neurosci . 2010 ;11 :100 –113 . 10.1038/nrn2774 20087360

16.Mei  L , Nave  KA . Neuregulin-ERBB signaling in the nervous system and neuropsychiatric diseases . Neuron . 2014 ;83 :27 –49 . 10.1016/j.neuron.2014.06.007 PMC4189115 24991953

17.Fisahn  A , Neddens  J , Yan  L , Buonanno  A . Neuregulin-1 modulates hippocampal gamma oscillations: implications for schizophrenia . Cereb Cortex . 2008 ;19 :612 –618 . 10.1093/cercor/bhn107 PMC2638818 18632742

18.del Pino  I , García-Frigola  C , Dehorter  N , et al.  Erbb4 deletion from fast-spiking interneurons causes schizophrenia-like phenotypes . Neuron . 2013 ;79 :1152 –1168 . 10.1016/j.neuron.2013.07.010 24050403

19.Hahn  CG , Wang  HY , Cho  DS , et al.  Altered neuregulin 1-erbB4 signaling contributes to NMDA receptor hypofunction in schizophrenia . Nat Med . 2006 ;12 :824 –828 . 10.1038/nm1418 16767099

20.Law  AJ , Lipska  BK , Weickert  CS , et al.  Neuregulin 1 transcripts are differentially expressed in schizophrenia and regulated by 5’ SNPs associated with the disease . Proc Natl Acad Sci USA . 2006 ;103 :6747 –6752 . 10.1073/pnas.0602002103 PMC1458952 16618933

21.Liu  X , Bates  R , Yin  DM , et al.  Specific regulation of NRG1 isoform expression by neuronal activity . J Neurosci . 2011 ;31 :8491 –8501 . 10.1523/jneurosci.5317-10.2011 PMC3154699 21653853

22.Batista-Brito  R , Majumdar  A , Nuño  A , et al.  Developmental loss of ErbB4 in PV interneurons disrupts state-dependent cortical circuit dynamics . Mol Psychiatry . 2023 ;28 :3133 –3143 . 10.1038/s41380-023-02066-3 PMC10618960 37069344

23.Fledrich  R , Akkermann  D , Schütza  V , et al.  NRG1 type I dependent autoparacrine stimulation of Schwann cells in onion bulbs of peripheral neuropathies . Nat Commun . 2019 ;10 :1467 . 10.1038/s41467-019-09385-6 PMC6443727 30931926

24.Goebbels  S , Bormuth  I , Bode  U , Hermanson  O , Schwab  MH , Nave  KA . Genetic targeting of principal neurons in neocortex and hippocampus of NEX-Cre mice . Genesis  Published online. 2006 ;44 :611 –621 . 10.1002/dvg.20256 17146780

25.Agarwal  A , Zhang  M , Trembak-Duff  I , et al.  Dysregulated expression of neuregulin-1 by cortical pyramidal neurons disrupts synaptic plasticity . Cell Rep . 2014 ;8 :1130 –1145 . 10.1016/j.celrep.2014.07.026 25131210

26.Schill  Y , Bijata  M , Kopach  O , et al.  Serotonin 5-HT4 receptor boosts functional maturation of dendritic spines via RhoA-dependent control of F-actin . Commun Biol . 2020 ;3 :1 –16 . 10.1038/s42003-020-0791-x PMC7021812 32060357

27.Speranza  L , Chambery  A , Di Domenico  M , et al.  The serotonin receptor 7 promotes neurite outgrowth via ERK and Cdk5 signaling pathways . Neuropharmacology . 2013 ;67 :155 –167 . 10.1016/j.neuropharm.2012.10.026 23164613

28.Velanac  V , Unterbarnscheidt  T , Hinrichs  W , et al.  Bace1 processing of NRG1 type III produces a myelin-inducing signal but is not essential for the stimulation of myelination . Glia . 2012 ;60 :203 –217 . 10.1002/glia.21255 PMC3267053 22052506

29.Wieser  GL , Gerwig  UC , Adamcio  B , et al.  Neuroinflammation in white matter tracts of Cnp1 mutant mice amplified by a minor brain injury . Glia . 2013 ;61 :869 –880 . 10.1002/glia.22480 23483656

30.Tan  GH , Liu  YY , Hu  XL , Yin  DM , Mei  L , Xiong  ZQ . Neuregulin 1 represses limbic epileptogenesis through ErbB4 in parvalbumin-expressing interneurons . Nat Neurosci . 2012 ;15 :258 –266 . 10.1038/nn.3005 22158510

31.Tamura  H , Shiosaka  S , Morikawa  S . Trophic modulation of gamma oscillations: the key role of processing protease for Neuregulin-1 and BDNF precursors . Neurochem Int . 2018 ;119 :2 –10 . 10.1016/j.neuint.2017.12.002 29233654

32.Flames  N , Long  JE , Garratt  AN , et al.  Short- and long-range attraction of cortical GABAergic interneurons by neuregulin-1 . Neuron . 2004 ;44 :251 –261 . 10.1016/j.neuron.2004.09.028 15473965

33.Willem  M . Proteolytic processing of Neuregulin-1 . Brain Res Bull . 2016 ;126 :178 –182 . 10.1016/j.brainresbull.2016.07.003 27393467

34.Michailov  GV , Sereda  MW , Brinkmann  BG , et al.  Axonal Neuregulin-1 regulates myelin sheath thickness . Science . 2004 ;304 :700 –703 . 10.1126/science.1095862 15044753

35.Schulze  WX , Deng  L , Mann  M . Phosphotyrosine interactome of the ErbB-receptor kinase family . Mol Syst Biol . 2005 ;2005 :0008 . 10.1038/msb4100012 PMC1681463 16729043

36.Ahmed  SBM , Prigent  SA . Insights into the Shc family of adaptor proteins . J Mol Signal . 2017 ;12 :1 –17 . 10.5334/1750-2187-12-2 PMC5624076 30210578

37.Deakin  IH , Godlewska  BR , Walker  MA , et al.  Altered hippocampal gene expression and structure in transgenic mice overexpressing neuregulin 1 (Nrg1) type I . Transl Psychiatry . 2018 ;8 :229 . 10.1038/s41398-018-0288-2 PMC6197224 30348978

38.Abe  Y , Namba  H , Kato  T , Iwakura  Y , Nawa  H . Neuregulin-1 signals from the periphery regulate AMPA receptor sensitivity and expression in GABAergic interneurons in developing neocortex . J Neurosci . 2011 ;31 :5699 –5709 . 10.1523/JNEUROSCI.3477-10.2011 PMC6622838 21490211

39.Salvany  S , Casanovas  A , Tarabal  O , et al.  Localization and dynamic changes of neuregulin-1 at C-type synaptic boutons in association with motor neuron injury and repair . FASEB J . 2019 ;33 :7833 –7851 . 10.1096/fj.201802329R 30912977

40.Johnson  B , Leek  AN , Tamkun  MM . Kv2 channels create endoplasmic reticulum / plasma membrane junctions: A brief history of Kv2 channel subcellular localization . Channels . 2019 ;13 : 88 –101 . 10.1080/19336950.2019.1568824 PMC6380216 30712450

41.Vullhorst  D , Ahmad  T , Karavanova  I , Keating  C , Buonanno  A . Structural similarities between neuregulin 1–3 isoforms determine their subcellular distribution and signaling mode in central neurons . J Neurosci . 2017 ;37 :5232 –5249 . 10.1523/JNEUROSCI.2630-16.2017 PMC5456106 28432142

42.Deakin  IH , Nissen  W , Law  AJ , et al.  Transgenic overexpression of the type i isoform of neuregulin 1 affects working memory and hippocampal oscillations but not long-term potentiation . Cereb Cortex . 2012 ;22 :1520 –1529 . 10.1093/cercor/bhr223 PMC3377963 21878485

43.Rico  B , Marín  O . Neuregulin signaling, cortical circuitry development and schizophrenia . Curr Opin Genet Dev . 2011 ;21 :262 –270 . 10.1016/j.gde.2010.12.010 21295966

44.Götze  T , Soto-Bernardini  MC , Zhang  M , et al.  Hyperactivity is a core endophenotype of elevated neuregulin-1 signaling in embryonic glutamatergic networks . Schizophr Bull . 2021 ;47 :1409 –1420 . 10.1093/schbul/sbab027 PMC8379540 33871014

45.Kotzadimitriou  D , Nissen  W , Paizs  M , et al.  Neuregulin 1 type i overexpression is associated with reduced NMDA receptor–mediated synaptic signaling in hippocampal interneurons expressing PV or CCK . eNeuro . 2018 ;5 :eNEURO.0418-17.2018 . 10.1523/ENEURO.0418-17.2018 PMC5938717 29740596

46.Li  KX , Lu  YM , Xu  ZH , et al.  Neuregulin 1 regulates excitability of fast-spiking neurons through Kv1.1 and acts in epilepsy . Nat Neurosci . 2011 ;15 :267 –273 . 10.1038/nn.3006 22158511

47.Craig  MT , McBain  CJ . Fast gamma oscillations are generated intrinsically in CA1 without the involvement of fast-spiking basket cells . J Neurosci . 2015 ;35 :3616 –3624 . 10.1523/JNEUROSCI.4166-14.2015 PMC4339363 25716860

48.Yin  DM , Chen  YJ , Lu  YS , et al.  Reversal of behavioral deficits and synaptic dysfunction in mice overexpressing neuregulin 1 . Neuron . 2013 ;78 :644 –657 . 10.1016/j.neuron.2013.03.028 PMC4176764 23719163

49.Vullhorst  D , Mitchell  RM , Keating  C , et al.  A negative feedback loop controls NMDA receptor function in cortical interneurons via neuregulin 2/ErbB4 signalling . Nat Commun . 2015 ;6 :7222 . 10.1038/ncomms8222 PMC4451617 26027736

50.Vullhorst  D , Bloom  MS , Akella  N , Buonanno  A . ER-PM junctions on GABAergic interneurons are organized by Neuregulin 2/VAP interactions and regulated by NMDA receptors . Int J Mol Sci . 2023 ;24 :2908 . 10.3390/IJMS24032908 PMC9917868 36769244

51.Deardorff  AS , Romer  SH , Fyffe  REW . Location, location, location: the organization and roles of potassium channels in mammalian motoneurons . J Physiol . 2021 ;599 :1391 –1420 . 10.1113/JP278675 PMC12993820 33449375

52.Kirmiz  M , Palacio  S , Thapa  P , King  AN , Sack  JT , Trimmer  JS . Remodeling neuronal ER-PM junctions is a conserved nonconducting function of Kv2 plasma membrane ion channels . Mol Biol Cell . 2018 ;29 :2410 –2432 . 10.1091/mbc.E18-05-0337 PMC6233057 30091655

53.Vierra  NC , Kirmiz  M , van der List  D , Santana  LF , Trimmer  JS . KV2.1 mediates spatial and functional coupling of l-type calcium channels and ryanodine receptors in mammalian neurons . elife . 2019 ;8 :1 –42 . 10.7554/eLife.49953 PMC6839919 31663850

54.Vierra  NC , O’Dwyer  SC , Matsumoto  C , Fernando Santana  L , Trimmer  JS . Regulation of neuronal excitation-transcription coupling by Kv2.1-induced clustering of somatic L-type Ca2+ channels at ER-PM junctions . Proc Natl Acad Sci USA . 2021 ;118 :1 –11 . 10.1073/pnas.2110094118 PMC8609631 34750263

55.Yang  YS , Jeon  SC , Kang  MS , et al.  Activation of ryanodine receptors is required for PKA-mediated downregulation of A-type K+ channels in rat hippocampal neurons . J Neurosci Res . 2017 ;95 :2469 –2482 . 10.1002/jnr.24076 28481020

56.Hiess  F , Yao  J , Song  Z , et al.  Subcellular localization of hippocampal ryanodine receptor 2 and its role in neuronal excitability and memory . Commun Biol . 2022 ;5 :1 –16 . 10.1038/s42003-022-03124-2 PMC8888588 35233070

57.Hernández  S , Salvany  S , Casanovas  A , et al.  Persistent NRG1 type III overexpression in spinal motor neurons has No therapeutic effect on ALS - related pathology in SOD1 G93A mice . Neurotherapeutics . 2023 ;20 :1820 –1834 . 10.1007/s13311-023-01424-x PMC10684470 37733208

58.Klemz  A , Wildner  F , Tütüncü  E , Gerevich  Z . Regulation of hippocampal gamma oscillations by modulation of intrinsic neuronal excitability . Front Neural Circuits . 2022 ;15 :778022 . 10.3389/fncir.2021.778022 PMC8845518 35177966

59.Wang  YY , Zhao  B , Wu  MM , Zheng  XL , Lin  L , Yin  DM . Overexpression of neuregulin 1 in GABAergic interneurons results in reversible cortical disinhibition . Nat Commun . 2021 ;12 :1 –13 . 10.1038/s41467-020-20552-y PMC7804852 33436636

60.Bao  J , Wolpowitz  D , Role  LW , Talmage  DA . Back signaling by the Nrg-1 intracellular domain . J Cell Biol . 2003 ;161 :1133 –1141 . 10.1083/jcb.200212085 PMC2172983 12821646

61.Zhu  JM , Li  KX , Cao  SX , et al.  Increased NRG1-ErbB4 signaling in human symptomatic epilepsy . Sci Rep . 2017 ;7 :1 –7 . 10.1038/s41598-017-00207-7 PMC5428003 28273943

62.Colmers  PLW , Maguire  J . Network dysfunction in comorbid psychiatric illnesses and epilepsy . Epilepsy Curr . 2020 ;20 :205 –210 . 10.1177/1535759720934787 PMC7427163 32628514

63.Zoupi  L , Booker  SA , Eigel  D , et al.  Selective vulnerability of inhibitory networks in multiple sclerosis . Acta Neuropathol . 2021 ;141 :415 –429 . 10.1007/s00401-020-02258-z PMC7882577 33449171

64.Vincent  B , Maitra  S . BACE1-dependent metabolism of neuregulin 1: bridging the gap in explaining the occurrence of schizophrenia-like symptoms in Alzheimer’s disease with psychosis?  Ageing Res Rev . 2023 ;89 :101988 . 10.1016/j.arr.2023.101988 .37331479

65.Tan  W , Wang  Y , Gold  B , et al.  Molecular cloning of a brain-specific, developmentally regulated neuregulin 1 (NRG1) isoform and identification of a functional promoter variant associated with schizophrenia . J Biol Chem . 2007 ;282 :24343 –24351 . 10.1074/jbc.M702953200 17565985

66.Günther  A , Hanganu-Opatz  IL . Neuronal oscillations: early biomarkers of psychiatric disease?  Front Behav Neurosci . 2022 ;16 :1 –9 . 10.3389/fnbeh.2022.1038981 PMC9806131 36600993