Astrocytic NHERF‐1 Increases Seizure Susceptibility by Inhibiting Surface Expression of TREK‐1

ABSTRACT

Mature hippocampal astrocytes exhibit a linear current‐to‐voltage (I‐V) K⁺ membrane conductance called passive conductance. It is estimated to enable astrocytes to keep potassium homeostasis in the brain. We previously reported that the TWIK‐1/TREK‐1 heterodimeric channels are crucial for astrocytic passive conductance. However, the regulatory mechanism of these channels by other binding proteins remains elusive. Here, we identified Na+/H+ exchange regulator‐1 (NHERF‐1), a protein highly expressed in astrocytes, as a novel interaction partner for these channels. NHERF‐1 endogenously bound to TWIK‐1/TREK‐1 in hippocampal cultured astrocytes. When NHERF‐1 is overexpressed or silenced, surface expression and activity of TWIK‐1/TREK‐1 heterodimeric channels are inhibited or enhanced, respectively. Furthermore, we confirmed that reduced astrocytic passive conductance by NHERF‐1 overexpressing in the hippocampus increases kainic acid (KA)‐induced seizure sensitivity. Taken together, these results suggest that NHERF‐1 is a key regulator of TWIK‐1/TREK‐1 heterodimeric channels in astrocytes and suppression of TREK‐1 surface expression by NHERF‐1 increases KA‐induced seizure susceptibility via reduction of astrocytic passive conductance.

Main Points

• NHERF‐1 is identified as an interaction partner for TREK‐1 channel.

• NHERF‐1 oppositely regulates the surface expression and activity of TREK‐1 channel.

• Astrocytic overexpression of NHERF‐1 is increased KA‐induced seizure susceptibility.

1. Introduction

Astrocytes have been attributed to several functions in the CNS, including neurotransmitter clearance from synapses, direction during neuronal migration, and active control of neuronal synaptic transmission. Additionally, astrocytes help to maintain the extracellular ionic environment required for communication between neurons (Purushotham and Buskila 2023). In particular, the electrophysiological properties of astrocytes change during maturation processes. One of the prominent features in mature astrocytes are the more negative membrane potential and a linear I‐V relationship. These features enable astrocytic ion homeostasis in the brain, such as extracellular K⁺ buffering, called astrocytic passive conductance (Kofuji and Newman 2004; Ryoo and Park 2016; Bellot‐Saez et al. 2017). This may be suggested as a protective mechanism against the toxic accumulation of K⁺ in the extracellular space and for modulating neuronal synchronization and network (Wang et al. 2012; Bellot‐Saez et al. 2017). Previous studies have shown that TWIK‐1 (K2P 1.1, KCNK1) and TREK‐1 (K2P 2.1, KCNK2) channels, which belong to the two‐pore domain K⁺ channels (K2P) isoforms, form heterodimeric channels in a disulfide bond‐dependent manner, and these channels predominantly mediate astrocytic passive conductance in the hippocampus (Zhou et al. 2009; Hwang et al. 2014). Although our group has elucidated the molecular identity of adult astrocyte passive conductance (Hwang et al. 2014; Bae et al. 2020), its physiological or pathophysiological role is still poorly studied. To address these questions, we examined novel proteins that bind to the TWIK‐1/TREK‐1 heterodimer channel, targeting proteins abundantly expressed in adult astrocytes.

NHERF‐1 is a member of the PDZ (postsynaptic density protein 95/Drosophila disc large tumor suppressor/zonula occulens‐1 (PSD‐95/DIg/ZO‐1)) family and scaffolding protein (Dunn and Ferguson 2015; Bushau‐Sprinkle and Lederer 2020). Scaffold proteins was functioned as crucial regulators of many signaling pathways through organize functional complexes organization and kinase activity modulation. They were characterized by the presence of protein–protein interaction modules including PDZ domains (Vaquero et al. 2017). NHERF‐1 contains two PDZ domains and one ezrin/radixin/moesin/merlin‐binding domain (EBD) that attach to the cytoskeleton (Treat et al. 2016). PDZ domains play an important role in regulating receptor and channel protein localization in several tissues and tight junctions and function to scaffold intracellular signaling protein complexes (Dunn and Ferguson 2015). NHERF‐1 mRNA is present in several regions of the mammalian brain (Weinman et al. 1995; Ediger et al. 1999). In particular, NHERF‐1 protein expression is higher in astrocytes than in neurons and interacts with TRPC4 channels and GLAST (Song et al. 2005; Lee et al. 2007). The lack of NHERF‐1 in cultured astrocytes alters subcellular localization and the distribution of Group II metabotropic glutamate receptor 2 and 3 (mGluR2/3) through interactions (Ritter‐Makinson et al. 2017). Additionally, mutations and overexpression of NHERF‐1 have been reported in several cancers, including glioblastoma (Kislin et al. 2009; Kumada et al. 2010; Bhattacharya et al. 2019).

In the present study, we found that the PDZ domains of NHERF‐1 bind to the N‐terminus of the TREK‐1 channel. Specifically, NHERF‐1 knockdown or overexpression in cultured astrocytes regulates the surface expression of TREK‐1 and TREK‐1‐mediated channel activity. The heterogeneous NHERF‐1 knockout (KO) mice had increased astrocyte passive conductance, whereas overexpression of NHERF‐1 resulted in decreased astrocyte passive conductance and increased kainic acid (KA)‐induced seizure susceptibility. Additionally, the reduced astrocyte passive conductance under NHERF‐1 overexpression was completely rescued by overexpression of TREK‐1 N4, the binding domain (BD) of TREK‐1 for NHERF‐1. Therefore, these data demonstrate that NHERF‐1 plays an important role in the regulation of its surface expression, channel activity, and seizure susceptibility through interaction with TREK‐1 in astrocytes.

2. Materials and Methods

2.1. Chemicals

Spadin was purchased from Tocris (Tocris Bioscience, Bristol, UK). Barium, Triethylamine (TEA), and Apamin were purchased from Sigma Aldrich (Sigma Aldrich, St. Louis, MO, USA). KA was purchased from Sigma Aldrich (Sigma Aldrich, St. Louis, MO, USA). All substances were stored as stock solutions at −20°C and diluted to the required concentrations in standard bath solution immediately prior to experimentation.

2.2. Animals

Transgenic NHERF‐1 mouse was purchased in Jackson Laboratory (stock No: 012862). Littermates including NHERF‐1 wild‐type (WT) or heterozygote knockout (Het) mice were used for this study. Mice were housed and maintained in standard laboratory conditions of 12:12 h light: dark cycle. Regular chow and water were provided ad libitum. Male C57BL/6N mice 7–12 weeks‐old were used for all experiments. Animal care and handling were performed according to the institutional guidelines of Institutional Animal Care and Use Committee (IACUC) at Korea University (Seoul, Korea) and at the Korea Institute of Science and Technology (Seoul, Korea). The animals were genotyped by PCR of tail DNA samples. The following primers were used: NHERF‐1 common forward primer—5′‐GAGAAGGGTCCAAATGGCTA‐3′; NHERF‐1 WT reverse primer—5′‐TTCGGCCTCATTCTGGTC‐3′, and NHERF‐1 KO reverse primer—5′‐CGCCTTCTTGACGAGTTCTT‐3′.

2.3. HEK293T Cells Culture and Transfection

HEK293T cells were purchased from the Korean Cell Line Bank (Seoul National University, Seoul, Korea) and cultured in DMEM (Thermo Fisher, Waltham, MA, USA) supplemented with 10% fetal bovine serum and penicillin–streptomycin at 37°C in a 5% CO₂ incubator. Transfection of expression vectors was perfomed with Lipofectamine 2000 (Thermo Fisher, Waltham, MA, USA), according to the manufacturer's protocol.

2.4. Primary Astrocyte Culture and Transfection

Primary astrocytes were maintained in DMEM supplemented with 10% fetal bovine serum, 10% horse serum, and penicillin–streptomycin. Cultures were maintained at 37°C in a 5% CO₂ incubator. After 4 days, cells were washed by repeated pipetting, and the media was replaced to remove debris and other cells. For gene silencing or overexpressing, cultured astrocytes were transfected with Nherf‐1 shRNAs or GFP‐NHERF‐1 with an optimized voltage protocol (1300 V, 20 pulse width ms, 2 pulses number) using the Neon Electroporation instrument (Thermo Fisher, Waltham, MA, USA).

2.5. Construction of Expression Vectors

cDNAs encoding for full‐length mouse TREK‐1 (GenBank accession no. NM_010607), mouse TWIK‐1 (NM_008430), and mouse NHERF‐1 (NM_012030) were obtained by using an RT‐PCR‐based gateway cloning method (Thermo Fisher, Waltham, MA, USA). Deletion mutants of TREK‐1 and NHERF‐1 were also generated using full‐length cDNAs as templates via EZchange site‐directed mutagenesis kit (Enzynomics, Daejeon, Korea). All constructs were cloned into various vectors by gateway cloning.

2.6. Construction of shRNAs

For gene silencing in cultured astrocyte, Trek‐1 shRNA were constructed and validated as described previously. The target region of Nherf‐1 shRNA is 5'‐AGCGATACCAGTGAGGAGCTAAAT‐3'.

2.7. Construction of Viral Vectors

For Nherf‐1 gene or TREK‐1 N4 overexpression in vivo, the mouse NHERF‐1 or 2A‐mCherry, TREK‐1 N4 or 2A‐GFP was re‐cloned into pDONR207 P1P5R or pDONR207 P5P2 vectors respectively via two independent BP‐reactions (ThermoFisher, Waltham, MA, USA). These vectors were cloned into the pAAV‐GFAP promoter vector using Multisite Gateway LR recombination reaction (ThermoFisher, Waltham, MA, USA). AAV viral vectors were produced by KIST virus facility (Korea).

2.8. Co‐Immunoprecipitation (Co‐IP) and Immunoblotting

For co‐immunoprecipitation (Co‐IP) in overexpressed HEK293T cells or endogenous astrocytes, whole‐cell lysates were mixed overnight at 4°C with 1 mg/mL anti‐HA (Santa Cruz Biotechnology, Dallas, TX, USA, cat# sc‐7392), anti‐GFP (Santa Cruz Biotechnology, Dallas, TX, USA cat# sc‐9996) or anti‐NHERF‐1 (H‐100; Santa Cruz Biotechnology) antibodies in lysis buffer (50 mM Tris–HCl (pH 7.4), 150 mM NaCl, 2 mM EDTA, 1 mM PMSF, 1% Triton X‐100 and 1% NP‐40) containing a protease‐inhibitor cocktail (Roche). Immune complexes were incubated by binding to mixed protein A/G PLUS Agarose (Santa Cruz Biotechnology) for 1 h and then washed four times with lysis buffer. For immunoblotting, protein samples were separated by SDS–PAGE using 10% gels. The separated proteins were transferred to polyvinylidene fluoride (PVDF) membranes (Bio‐rad). The blots were incubated overnight at 4°C with anti‐HA antibody (Roche Applied Science, Penzberg, Germany cat# 11867431001, 1:1000), anti‐GFP (Santa Cruz Biotechnology, Dallas, TX, USA cat# sc‐9996, 1:1000), anti‐TREK‐1 (Alomone Labs, Jerusalem, Israel, cat#APC‐047, 1:500), or anti‐NHERF‐1 (Thermo Fisher, Waltham, MA, USA, cat# PA1‐090, 1:1000). Blots were then washed and incubated with horseradish peroxidase‐conjugated goat anti‐mouse (Jackson ImmunoResearch, West Grove, PA, USA, 1:3000), anti‐rat (Jackson ImmunoResearch, West Grove, PA, USA, 1:3000), or anti‐rabbit (Jackson ImmunoResearch, West Grove, PA, USA, 1:3000) followed by washing and detection of immunoreactivity with enhanced chemiluminescence (Bio‐rad, Hercules, CA, USA).

2.9. Cell Surface Biotinylation

Biotinylation assays were performed using Sulfo‐NHS‐SS‐Biotin (Thermo Fisher, Waltham, MA, USA) according to the manufacturer's instructions. Nherf‐1 shRNA, Scramble shRNA, GFP control, or GFP‐NHERF‐1‐transfected astrocytes were incubated at 4°C and washed three times with PBS. Then, cells were incubated with Sulfo‐NHS‐SS‐Biotin in PBS for 30 min at 4°C. After washing cells with ice‐cold PBS containing glycine, non‐reacted biotinylation reagent was removed by lysing cells in ice‐cold lysis buffer. Cell lysate proteins were then incubated with high capacity NeutrAvidin‐agarose resin (Thermo Fisher, Waltham, MA, USA), and after washing in lysis buffer, bound proteins were eluted with SDS sample buffer and used in western blot analyses as described above.

2.10. Western Blot Analysis

To validate NHERF‐1 knock‐down or overexpressing conditions in cultured astrocytes, protein of cell transfected with Nherf‐1 shRNAs or GFP‐NHERF‐1 were separated by SDS–PAGE using 10% gels. The separated proteins were transferred to PVDF membranes (Bio‐rad, Hercules, CA, USA). The blots were incubated overnight at 4°C with anti‐NHERF‐1 (Thermo Fisher, Waltham, MA, USA, cat# PA1‐090, 1:1000) or anti‐actin antibody (Sigma Aldrich, St. Louis, MO, USA, cat# A2066, 1:1000). Blots were then washed and incubated with horseradish peroxidase‐conjugated goat anti‐rabbit (Jackson ImmunoResearch, West Grove, PA, USA, 1:3000) followed by washing and detection of immunoreactivity with enhanced chemiluminescence (Bio‐rad, Hercules, CA, USA).

2.11. Reverse Transcription (RT)‐PCR and Real‐Time PCR

For NHERF‐1 KO validation, total RNA was isolated from mouse brain using an RNA purification Kit (GeneAll, Seoul, Korea) according to the manufacturer's instructions. RT was performed with 1 μg total RNA using a cDNA Synthesis Kit (Takara Bio Inc. Kusatsu, Japan cat# 639543), according to the manufacturer's instructions. PCR was performed using rTaq 5× PCR premix (ELPIS‐BIOTECH, Daejeon, Korea, cat# EBT‐1013) under the following cycle conditions: denaturation at 95°C for 20 s, annealing at 55°C for 20 s, and extension at 72°C for 20 s. This cycle was repeated a total of 35 times. The PCR products were separated by electrophoresis in a 2% agarose gel, and images were captured on a gel imaging system. qPCR was also performed with SYBR Green mix (Enzo Life Sciences, Farmingdale, NY, USA, cat# ENZ‐NUC104‐1000). Primers were used the following sense and antisense primers: for NHERF‐1, forward (5′‐AGATCTGCCTCCAGCGATAC‐3′) and reverse (5′‐TTCATTTTTCTTGCTCCAGTCC‐3′), and for GAPDH, forward (5′‐GTCTTCACCACCATGGAGAA‐3′) and reverse (5′‐GCATGGACTGTGGTCATGAG‐3′). GAPDH was used as a reference gene. The 2‐△△CT method was used to calculate fold changes in gene expression.

2.12. Yeast Two‐Hybrid Assay

The TREK‐1 N or C‐terminus was cloned into the GAL4 DNA BD and the full length of NHERF‐1 or NHERF‐1 mutants (NHERF‐1 PDZ1, PDZ2, C tail, NHERF‐1 ΔPDZ1, ΔPDZ2, and ΔC tail) was cloned into the activation domain (AD). Direct interactions of the two proteins were investigated by co‐transforming yeast AH190 cells with BD/TREK‐1 N or C‐terminus and AD/NHERF‐1 WT or mutants. Transformed cells were then plated on synthetic dropout medium lacking Trp and Leu (TL‐) at 30°C for 3 days and were subsequently transferred to the medium lacking Trp, Leu, and His (TLH‐) for growth selection.

2.13. Bimolecular Fluorescence Complementation (BiFC) Assay

For BiFC, TREK‐1, TWIK‐1, TRAAK, TRPC4, and NHERF‐1 were cloned into bimolecular fluorescence complement (pBiFC)‐VN173 and (pBiFC)‐VC155 vectors. To confirm the expression of each BiFC vector, additional Flag and HA tags were inserted in the C‐terminal region of both BiFC vectors. HEK293T cells were co‐transfected with cloned BiFC vectors in all possible pairwise combinations. The next day, these cells were fixed with 4% paraformaldehyde for 20 min at room temperature and permeabilized with 0.3% Triton X‐100 for 5 min. After blocking for 1 h in 5% BSA solution, the cells were incubated with mouse anti‐HA (Cell Signaling, Danvers, MA, USA cat# 2367S, 1: 250) and rabbit anti‐Flag (Cell Signaling, Danvers, MA, USA cat# 14793S, 1:250) antibodies at 4°C overnight. Then, the cells were incubated with Alexa Fluor 594‐ or 647‐conjugated secondary antibodies (Jackson ImmunoResearch, West Grove, PA, USA, 1:400) for 1 h and then stained with DAPI to visualize nuclei. All images were acquired by confocal microscopy on a Nikon A1 confocal microscope.

2.14. Bioluminescence Resonance Energy Transfer (BRET) Assay

For BRET, TREK‐1, TWIK‐1, TRAAK and TRPC4, and NHERF‐1 were cloned into mCit‐PA‐(pBRET) and NL‐myc‐(pBRET) vectors. HEK293T cells were cultured in 96‐well plate (SPL, Pocheon‐si, Gyeonggi‐do, Korea, cat# 33396) at a density of 1.5~2 × 10^5 cells per well. 24 h after incubation, plasmids encoding donor and acceptor proteins are transfected at a ratio 1:10 (total 500 ng of DNA) with lipofectamine 2000 (Thermo Fisher, Waltham, MA, USA). Coelenterazine‐h (Sigma Aldrich, St. Louis, MO, USA, cat# c3230‐50ug) was added to a final concentration of 5 μM per well (the total volume of 50 μL per well), and incubated for an additional 15 min. An infinite M200 Pro microplate reader (TECAN, Männedorf, Switzerland) was used with an integration time of 5 s to measure the short and long wavelengths with BLUE1 (370–480 nm) and GREEN1 (520–570 nm) filters for 1000 ms. BRET ratio was calculated as: (long‐wavelength emission/short‐wavelength emission)—(long‐wavelength emission for donor (NL) only transfected cells/short‐wavelength emission for donor (NL) only transfected cells).

2.15. Duolink Proximity Ligation Assay (PLA)

Interactions between endogenous proteins were detected using a Duolink PLA kit (Sigma Aldrich, St. Louis, MO, USA, cat# DUO92014), according to the manufacturer's instructions. The primary antibodies used for this assay were anti‐TREK‐1 (Santa Cruz Biotechnology, Dallas, TX, USA cat# sc‐398,449, 1:50) and anti‐NHERF‐1 (Bethyl Laboratories, Montgomery, TX, USA, cat# A302‐974A, 1:100) antibodies. The PLA probe anti‐rabbit minus binds to the anti‐NHERF‐1 antibody, whereas the PLA probe anti‐mouse plus binds to the anti‐TREK‐1. Astrocytes growing on PDL‐coated coverslips were observed using a Nikon Ti2 confocal microscope (Nikon Instruments Inc. Melville, NY, USA).

2.16. Immunohistochemistry

Virus infected mice were anesthetized using avertin and subjected to intracardiac perfusion with saline, followed by 4% PFA solution in PBS. Brains were fixed in 4% PFA overnight at 4°C, and then 40 μm‐thick sections were obtained using vibratome (Leica, Wetzlar, Germany, VT1200). The slices were permeabilized with 0.5% Triton X‐100 in PBS for 30 min at room temperature and subsequently incubated in blocking buffer (5% normal donkey serum, 3% BSA, and 0.2% Triton X‐100 in PBS) for 2 h at room temperature. The slices were then incubated with primary antibodies, Chicken anti‐GFAP antibody (Thermo Fisher, Waltham, MA, USA, cat# PA1‐10004, 1:500), Chicken anti‐MAP2 antibody (Thermo Fisher, Waltham, MA, USA, cat# PA1‐10005, 1:500), Rabbit anti‐NeuN antibody (Thermo Fisher, Waltham, MA, USA, cat# 711054, 1:500), Rabbit anti‐NHERF‐1 (Thermo Fisher, Waltham, MA, USA, cat# PA1‐090, 1:300) overnight at 4°C. The sections were washed thrice in PBS and incubated with suitable fluorescence Alexa Fluor‐conjugated secondary antibodies (Jackson ImmunoResearch, West Grove, PA, USA; 1:400). The tissue sections were counter stained with DAPI and were mounted on glass slides for microscopy. All images were acquired using a Nikon A1 confocal microscope (Nikon Instruments Inc. Melville, NY, USA).

2.17. Fluoro‐Jade B Staining

Fluoro‐Jade B staining was conducted according to the previous method (Schmued and Hopkins 2000). Virus‐infected and KA‐injected brain slices were mounted on coated slides and dried at 60°C on a slide warmer for 1 h. The slides were initially immersed in a basic alcohol solution for 5 min. Then, slides were rehydrated with 70% ethanol for 2 min and washed with D.W. for 2 min. After rehydration, the slides were transferred to 0.06% potassium permanganate for 5 min and rinsed with D.W. three times for 1 min each. The slides were incubated in 0.004% Fluoro‐Jade B (Sigma Aldrich, St. Louis, MO, USA, AG310) solution containing 0.1% acetic acid and DAPI solution for 10 min at RT. After incubation, the slides were rinsed with D.W for 1 min for three times and then slides were cleared by xylene for 5 min and dried in an amber condition. Dried slides were mounted with DPX (DPX mountant for histology, Sigma Aldrich, St. Louis, MO, USA, 06522‐100ML) and left to dry for 24 h under the hood.

2.18. Electrophysiological Recording in Cultured Astrocytes

Cultured astrocytes were plated onto coverslips for electrophysiological experiments. The standard solution for the pipette contained 150 mM KCl, 1 mM CaCl₂, 1 mM MgCl₂, 5 mM EGTA, and 10 mM HEPES (pH 7.2, adjusted with KOH). Standard bath solution contained in mM: 150 NaCl, 3 KCl, 2 CaCl₂, 1 MgCl₂, 10 HEPES, 5.5 D‐glucose, and 20 sucrose (pH 7.4, adjusted with NaOH). Patch pipettes were made from borosilicate glass capillaries (Warner Instruments, Washington, DC, USA). The pipette resistance was 5–6 MΩ. Whole‐cell currents were recorded using a patch clamp amplifier (Axopatch 700B, Axon Instruments, Union City, CA, USA). Current–voltage relations were measured by applying ramped pulses (from −150 mV to +50 mV over 1000‐ms) from a holding potential of −60 mV. A Digidata 1550. A interface (Axon Instruments, Union City, CA, USA) was used to convert digital–analog signals between the amplifier and computer. Data were sampled at 5 kHz and filtered at 1 kHz. Currents were analyzed with Clampfit software (Axon Instruments, Union City, CA, USA). All experiments were conducted at room temperature.

2.19. Electrophysiological Recording in Hippocampal Slices

AAV‐GFAPp‐mCherry, AAV‐GFAPp‐GFP, AAV‐GFAPp‐Nherf‐1‐2A‐mCherry, or AAV‐GFAPp‐TREK‐1 N4‐2A‐GFP injected brain slices (300 μm) containing the hippocampus were prepared using a vibrating blade microtome (Leica, Wetzlar, Germany, VT1200) in ice‐cold oxygenated artificial cerebrospinal fluid (ACSF) containing (in mM) 130 NaCl, 2.5 KCl, 1.25 KH2PO4, 3 MgCl₂, 1 CaCl₂, 26 NaHCO3, and 10 D‐glucose. Individual hippocampal slices were transferred to a recording chamber, which was constantly perfused with ACSF recording solution containing (in mM) 130 NaCl, 24 NaHCO3, 3.5 KCl, 1.25 NaH2PO4, 1.5 CaCl₂, 1.5 MgCl₂, and 10 glucose saturated with 95% O₂–5% CO₂ at pH 7.4. The slices were recovered at room temperature for at least 1 h and electrophysiological recording. Patch pipettes had a resistance of 3–5 MΩ when filled with pipette solution containing (in mM) 140 KCl, 10 HEPES, 5 EGTA, 2 Mg‐ATP, and 0.2 Na‐GTP, adjusted to pH 7.4 with KOH. Whole‐cell patch recordings were performed on hippocampal astrocytes with a voltage‐clamp configuration using an Axopatch 700B (Axon Instruments, San Jose, CA, USA).

2.20. Stereotaxic Injection

C57BL/6N mice (7–8 weeks old) were anesthetized with avertin (2,2,2‐tribromethanol in 2‐methyl 2‐butanol) and placed in a stereotaxic frame (Kopf Instruments, Tujunga, CA, USA). Briefly, the scale was opened and two holes were drilled in the skull (−1.7 mm AP, ± 1.4 mm ML from bregma). AAV‐GFAPp‐mCherry, AAV‐GFAPp‐Nherf‐1‐2A‐mCherry, and AAV‐GFAPp‐TREK‐1 N4‐2A‐GFP (2.21 × 10^14 GC/ml) were packaged in the serotype 5 at KIST Virus Facility. These viruses were bilaterally injected (250 nL per side) into the hippocampal CA1 stratum radiatum (SR) area (1.55–1.6 mm DV from the dura) through a Hamilton Syringe with a syringe pump (KD Scientific, Holliston, MA, USA) that infused the virus at a speed of 0.1 μL/min. The Hamilton Syringe was left undisturbed at the injected points for 10 min.

2.21. KA‐Induced Seizure Behaviors

KA (Sigma Aldrich, St. Louis, MO, USA.) dissolved in saline was administered intraperitoneal (ip) injection at a dose of 35 mg/kg. Animals were monitored for 90 min after the injection. Seizure scores were monitored every 5 min using a Racine scale; 0, normal behavior; 1, immobilization; 2, head nodding; 3, whole body myoclonus; 4, continuous rearing and falling; 5, clonic–tonic seizure; 6, death (Racine 1975). Nonresponsive animals with a seizure score of 0 over 90 min were excluded from analysis.

2.22. Statistical Analysis

All data are presented as means ± standard error of the mean (SEM). The significance of data for comparison was assessed by Student's t‐test (paired t‐test or unpaired t‐test) or one‐way ANOVA followed by Turkey's post hoc test and significance levels are given as: ns: not significant, * p < 0.05, ** p < 0.01, *** p < 0.001, and **** p < 0.0001. Prism9.0 software (GraphPad Software, San Diego, CA, USA) was used for carrying out the statistical analysis.

3. Results

3.1. NHERF‐1 Inversely Regulates K⁺ Currents in Cultured Astrocytes

The PDZ domains play an important role in regulating receptor and channel protein localization in several tissues and tight junctions and function to scaffold intracellular signaling protein complexes (Dunn and Ferguson 2015). We checked the expression level of Nherf‐1 in the astrocytes and neurons using Brain‐RNA seq database (https://www.brainrnaseq.org, Figure S1A). According to this database, the Nherf‐1 gene was more than expressed in the astrocytes compared to neurons. We also confirmed the protein expression level of NHERF‐1 in cultured astrocytes or neurons using a western blot. The NHERF‐1 protein was more expressed in cultured astrocytes than cultured neurons (Figure S1B) Therefore, NHERF‐1 was highly expressed in the astrocytes of the brain.

To assess the effect of NHERF‐1 on K⁺ channels in cultured astrocytes, we first recorded K⁺ currents under the NHERF‐1 overexpressed conditions. In previous studies, NHERF‐1 regulated the activity or surface expression of several channels in various cells or tissues (Storch et al. 2017; Martin et al. 2020; Saha et al. 2021). First, the NHERF‐1 overexpressed condition showed the increased expression of NHERF‐1 in cultured astrocytes (Figure 1A). K⁺ currents in NHERF‐1 overexpressed cultured astrocytes significantly decreased linearly (Figure 1B,C), and the reversal potential was depolarized from −82.69 ± 1.58 mV to −69.88 ± 3.38 (Figure 1D). To test the opposite effect, we produced NHERF‐1‐specific short hairpin‐forming interference RNA (shRNA). The expression level of NHERF‐1 was effectively reduced by Nherf‐1 shRNA in cultured astrocytes (Figure 1E). Next, we found that K⁺ currents were significantly increased in cultured astrocytes by Nherf‐1 shRNA (Figure 1F,G). The reversal potential was not changed (Figure 1H). These results suggest that NHERF‐1 reversely regulates K⁺ currents in cultured astrocytes.

FIGURE 1.

NHERF‐1 inversely regulates potassium currents in cultured astrocytes. (A) Top; western blotting was performed to verify NHERF‐1 overexpressed condition in astrocytes. Bottom; a bar graph showing the relative protein expression level of NHERF‐1 in the cultured astrocytes (n = 4). (B) Current densities showing to GFP‐control (GFP) or GFP‐NHERF‐1 transfected astrocytes. (C) Summary bar graph of GFP or GFP‐NHERF‐1 currents at +50 and −150 mV as in B. (D) Reversal potential (Vrev) in GFP‐control (GFP) or GFP‐NHERF‐1 transfected astrocytes. (E) Top; western blotting was performed to verify Nherf‐1 shRNA in astrocytes. Bottom; a bar graph showing the relative protein expression level of NHERF‐1 in the cultured astrocytes (n = 3). (F) Current densities showing to scrambled shRNA (Sc shRNA) or Nherf‐1 shRNA transfected astrocytes. (G) Summary bar graph of Sc shRNA or Nherf‐1 shRNA currents at +50 mV and −150 mV as in F. (H) Reversal potential (Vrev) in Sc shRNA or Nherf‐1 shRNA transfected astrocytes. The number on each bar indicates n for each condition. All values are mean ± SEM. p‐values were obtained using Student's t‐test. ns: not significant, *p < 0.05, **p < 0.01, and ***p < 0.001.

3.2. NHERF‐1 Mainly Contributes to TWIK‐1/TREK‐1‐Mediated K⁺ Currents in Cultured Astrocytes

To identify the channels contributing to the K⁺ current regulated by NHERF‐1, we applied various K⁺ channel inhibitors. K⁺ channels are a superfamily of diverse members involved in K⁺ currents in a variety of cell types, including astrocytes. Based on protein structure, K⁺ channels are classified into the voltage‐gated K⁺ channels (Kv), inwardly rectifying K⁺ channels (Kir), two‐pore domain K⁺ channels (K2P), and Ca²⁺–activated K⁺ channels (KCa) families (Choe 2002; Enyedi and Czirják 2010; Ryoo and Park 2016). We added inhibitors for each of the four types of K⁺ channels (TEA, Ba2+, apamin, and spadin) to cultured astrocytes. We only used astrocytes within 7 days of primary culture at P1, and as expected from the known expression level, we confirmed that spadin was most effectively inhibited (Figure 2A,B and Figure S2A,B). This was consistent with previous findings from our group (Hwang et al. 2014; Bae et al. 2020). Next, we investigated the effect of spadin on K⁺ currents modulated by NHERF‐1. K⁺ currents were significantly reduced in GFP‐NHERF‐1 transfected cultured astrocytes compared to GFP control transfected cells. Interestingly, spadin had relatively little effect on K⁺ currents in astrocytes overexpressing NHERF‐1, but effectively reduced the current in control cells (Figure 2C,D). We measured spadin‐sensitive K⁺ currents in astrocytes by subtracting the pre‐K⁺ current from the K⁺ current after treatment with spadin. Spadin‐sensitive currents were smaller in GFP‐NHERF‐1 transfected astrocytes than in control cells (Figure 2E,F). Together, these data indicate that NHERF‐1 is primarily involved in regulating TWIK‐1/TREK‐1 heterodimer channel‐mediated K⁺ currents in cultured astrocytes.

FIGURE 2.

NHERF‐1 primarily contributes to TWIK‐1/TREK‐1‐mediated K⁺ currents in cultured astrocytes. (A) Normalized traces of various K⁺ channel inhibitors‐sensitive currents of cultured astrocytes. (B) The bar graph shows normalized various K⁺ channel inhibitors‐sensitive currents densities at +50 mV and −150 mV. (C) Current densities showing before and after application of spadin to GFP control (GFP) or GFP‐NHERF‐1 transfected astrocytes. (D) Summary bar graph of C at +50 mV and −150 mV. (E) Spadin‐sensitive currents were calculated by difference between currents before and after 10 μM spadin application in C. (F) Summary bar graph of GFP control (GFP/spadin sensitive) or GFP‐NHERF‐1 (GFP‐NHERF‐1/spadin sensitive) spadin‐subtracted currents plotted at +50 mV and −150 mV. The number on each bar indicates n for each condition. All values are mean ± SEM. p‐values were obtained using one‐way ANOVA followed by Bonferroni‘s post hoc test or Student's t‐test. ns: not significant, *p < 0.05, **p < 0.01, and ***p < 0.001.

3.3. Identification of Binding Regions Between NHERF‐1 and TREK‐1

In a previous study, we confirmed that only the TWIK‐1/TREK‐1 heterodimer mainly exists in astrocytes (Hwang et al. 2014), so we tested whether NHERF‐1 would interact with either TREK‐1 or TWIK‐1 channels. We applied bioluminescence resonance energy transfer (BRET) analysis to determine the interaction between the two proteins. The BRET signal ratio of TREK‐1/NHERF‐1 was significantly higher than that of TWIK‐1/NHERF‐1 (Figure S3A–C). These data suggest that NHERF‐1 interacts better with TREK‐1 than with TWIK‐1.

Next, to find the detailed binding site of TREK‐1 on NHERF‐1, we generated vector constructs expressing only the N‐ and C‐terminal regions (TREK‐1‐N or TREK‐1‐C). The N and C termini of K2P channels are present in the cytoplasmic region of the cell (Lesage et al. 1996; Enyedi and Czirják 2010). Therefore, by performing a yeast two‐hybrid assay, we confirmed that NHERF‐1 is interacting with the N‐terminus of TREK‐1 (Figure 3A). Co‐IP experiments using the same strategy also clearly showed that NHERF‐1 binds to the N‐terminus of TREK‐1 (Figure 3B). To reconfirm the interaction between NHERF‐1 and TREK‐1‐N, Bi‐fluorescence complementation (BiFC) constructs expressing N‐ or C‐terminal region deleted forms of TREK‐1 (TREK‐1‐ΔN or TREK‐1‐ΔC) was produced (Figure 3C). Cells co‐transfected with TREK‐1/NHERF‐1 and TREK‐1‐ΔC/NHERF‐1 showed strong fluorescence, whereas cells co‐transfected with TREK‐1ΔN/NHERF‐1 did not detect BiFC signal (Figure 3D). To further determine the binding site of NHERF‐1 within TREK‐1, the N‐terminus of TREK‐1 was divided into four sections (ΔN1, ΔN1,2, ΔN1,2,3, or ΔN) (Figure S4A) (Kim et al. 2022). BiFC signals were detected for TREK‐1 ΔN1, TREK‐1 ΔN1,2, and TREK‐1 ΔN1,2,3 using NHERF‐1, but not for TREK‐1 ΔN using NHERF‐1 (Figure S4B). Additionally, we generated a novel substitution mutation (KWK → AWA) in TREK‐1 through sequence comparison among TREK family members. This approach was taken to eliminate the possibility of structural misfolding or trafficking abnormalities that could arise from the deletion of the N‐terminal region of TREK‐1 (Figure S4C,D). NHERF‐1 has been shown to interact with the N4 region of the N‐terminus of TREK‐1. We therefore performed Co‐IP and BiFC analyses using the N4 region deletion or AWA mutant construct (TREK‐1 ΔN4 or TREK‐1‐AWA). As expected, TREK‐1 ΔN4 and TREK‐1‐AWA showed low interaction across these assays (Figure 3E,F). We also investigated whether each TREK‐1 mutant (ΔN4 or AWA) exhibited increased membrane potential and enhanced conductance as a functional potassium channel due to impaired binding to NHERF‐1. Indeed, the TREK‐1 mutants showed larger potassium currents than the wild type (WT), even in the presence of NHERF‐1. Colocalization analysis using membrane markers further confirmed that relatively greater amounts of protein were localized to the membrane (Figure S5A–G). These results indicate that NHERF‐1 binds to the N4 region of TREK‐1.

FIGURE 3.

NHERF‐1 directly binds with N4 of N‐terminus of TREK‐1. (A) Top; diagram of the full‐length (FL), N‐, C‐terminus of TREK‐1. Bottom; yeast two‐hybrid assays show direct binding between NHERF‐1 and N‐terminal of TREK‐1. (B) Co‐IP experiments suggest that NHERF‐1 interacts with N‐terminal of TREK‐1. (C) Diagram of the TREK‐1 FL and TREK‐1 N (ΔN) or C (ΔC) terminus deletion mutants and schematic diagram depicting the bimolecular fluorescence complementation (BiFC). (D) BiFC experiments were performed in HEK293T cells transfected with VC‐NHERF‐1 and VN‐TREK‐1, VN‐TREK‐1 ΔN, or VN‐TREK‐1 ΔC. Scale bar, 10 μm. Data obtained from three independent experiments. The number on each bar indicates n for each condition. (E) Top; diagram of the TREK‐1 FL TREK‐1 ΔN4 or TREK‐1(KWK → AWA). Bottom; BiFC experiments with VC‐NHERF‐1 and VN‐TREK‐1, TREK‐1 ΔN4 or TREK‐1(KWK → AWA). Scale bar, 10 μm. (F) Co‐IP experiments to assess the interaction between NHERF‐1 and TREK‐1 FL TREK‐1 ΔN4 or TREK‐1(KWK → AWA).

Conversely, to identify the binding region of NHERF‐1 to TREK‐1, we also generated several mutant constructs of NHERF‐1 (Figure 4A). NHERF‐1 contains two PDZ domains and one C‐terminal ezrin‐binding domain (EBD) (Bhattacharya et al. 2019). We tested an interaction between TREK‐1 and several domains of NHERF‐1 by showing a positive yeast colony in the Y2H assay. The Y2H assay showed that TREK‐1 binds to the PDZ1 or PDZ2 domain of NHERF‐1 (Figure 4A). Co‐IP data also showed that TREK‐1 strongly interacts with the PDZ1 domain of NHERF‐1 and weakly binds with PDZ2 domain (Figure 4B). To confirm the interaction between PDZ1 domain of NHERF‐1 and TREK‐1, we produced BiFC constructs expressing vector PDZ1, PDZ2, or C‐terminal region deleted forms of NHERF‐1 (NHERF‐1 ΔPDZ1, NHERF‐1 ΔPDZ2, NHERF‐1 ΔC, NHERF‐1 ΔPDZ1,2) (Figure 4C). We performed a BiFC assay to confirm these interactions. The BiFC data showed that the fluorescence signal was detected in cells co‐transfected with TREK‐1/NHERF‐1, TREK‐1/NHERF‐1 ΔPDZ1, TREK‐1/NHERF‐1 ΔPDZ2, and TREK‐1/NHERF‐1 ΔC, whereas not detected in cells co‐transfected with TREK‐1/NHERF‐1 ΔPDZ1 and 2 (Figure 4D). These results indicate that PDZ domains of NHERF‐1 bind to the N4 region of TREK‐1.

FIGURE 4.

TREK‐1 directly binds with PDZ domains of NHERF‐1. (A) Top; diagram of the full length (FL) and PDZ1, PDZ2, or C tail of NHERF‐1. Bottom; yeast two‐hybrid assays show interactions between TREK‐1 and FL, PDZ1, PDZ2, or C tail of NHERF‐1. (B) Co‐IP data show also interactions between TREK‐1 and FL, PDZ1, PDZ2, or C tail of NHERF‐1. (C) Diagram of the FL of NHERF‐1 and various deletion mutants of NHERF‐1. (D) BiFC experiments between VN‐TREK‐1 and VC‐NHERF‐1 or VC‐NHERF‐1 deletion mutants. Scale bar, 10 μm.

3.4. NHERF‐1 Regulates the Surface Expression of TREK‐1 via Their Interaction in Cultured Astrocytes

To identify whether NHERF‐1 endogenously interacts with TREK‐1 in astrocytes, we performed a Co‐IP and Duolink proximity ligation assay (PLA). Co‐IP experiments showed an interaction between TREK‐1 and NHERF‐1 with anti‐NHERF‐1 antibody in cultured astrocytes (Figure 5A). This interaction was emphasized by the Duolink PLA signals. A strong Duolink PLA signal from TREK‐1 and NHERF‐1 was observed in astrocytes compared to only the exposed anti‐TREK‐1 antibody (Figure 5B).

FIGURE 5.

NHERF‐1 regulates the surface expression of TREK‐1 in astrocytes. (A) Co‐IP data shows interactions between TREK‐1 and NHERF‐1 in astrocytes. Whole‐cell lysates were immunoprecipitated using anti‐NHERF‐1 antibody and then blotted using anti‐TREK‐1 antibody. (B) Left; representative images of Duolink PLA assays in astrocytes. Scale bar, 20 μm. Right; PLA signals were counted using ImageJ software and average numbers of spots per cell are presented in the graph. (C) Cell surface biotinylation results from astrocytes transfected with scrambled shRNA (Sc shRNA) or Nherf‐1 shRNA. (D) The summary bar graph shows the summary of C. (E) Cell surface biotinylation experiments with astrocytes transfected with GFP control (GFP) or GFP‐NHERF‐1. (F) Data from three independent experiments are presented in the normalized bar graph as in (E) data obtained from three independent experiments. The number on each bar indicates n for each condition. All values are mean ± SEM. p‐values were obtained using Student's t‐test. *p < 0.05 and ***p < 0.001.

Channel activity is associated with surface expression in cells, and this concept commonly serves as a mechanism for regulating channel activity (Solé et al. 2009; Lee et al. 2016). To determine the effect of NHERF‐1 on the surface expression of the TREK‐1 channel, we preferentially confirmed the surface expression of TREK‐1 in NHERF‐1 overexpressed or NHERF‐1 knocking down cultured astrocytes. Cell surface biotinylation assay indicated that the surface expression of TREK‐1 was markedly enhanced in the NHERF‐1 knockdown condition using Nherf‐1 shRNA in cultured astrocytes (Figure 5C,D). We confirm the reverse effect by testing it under the NHERF‐1 overexpressed condition. NHERF‐1 overexpression significantly reduced the surface expression of TREK‐1 (Figure 5E,F). These results suggest that NHERF‐1 inversely regulates cell surface expression of TREK‐1 through physical interactions in cultured astrocytes.

3.5. Astrocytic Passive Conductance Is Enhanced in Heterozygous NHERF‐1 Knockout Mice

To clarify whether NHERF‐1 mediates astrocytic passive conductance via TWIK‐1/TREK‐1 heterodimeric channels in vivo, we used the NHERF‐1 gene knockout (KO) animal model. The genomic illustration of NHERF‐1 KO mice showed that homologous recombination replaced the first exon containing the beginning of the transcription of the mouse Nherf‐1 gene with the neomycin sequence (Figure S6A) (Shenolikar et al. 2002). To verify the genetic deletion of NHERF‐1 mice, we performed PCR‐based genotyping of tail samples from NHERF‐1 WT and heterozygous (Het) mice (Figure S6B). qRT‐PCR analysis revealed the mRNA expression levels of Nherf‐1. In NHERF‐1 Het mice, qRT‐PCR results confirmed that Nherf‐1 mRNA was reduced by almost half (Figure S6C,D). We also examined the expression level of NHERF‐1 protein in brain tissue using a specific anti‐NHERF‐1 antibody (Figure S6E).

To determine the contribution of NHERF‐1 to TWIK‐1/TREK‐1 heterodimeric channel‐mediated astrocytic passive conductance, we aimed to measure the effects of NHERF‐1 on astrocytic passive conductance in hippocampal slices of NHERF‐1 Het mice. The electrophysiological recording clearly showed that passive conductance in NHERF‐1 Het mice was more enhanced than that in WT mice. The application of spadin was significantly modulated in both WT and Het mice (Figure 6A). The linear current‐to‐voltage (I‐V) relationship of passive conductance reflects the intrinsic properties of mature astrocytes (Zhou et al. 2009; Hwang et al. 2014; Bae et al. 2020). NHERF‐1 Het mice also showed an increased linear I‐V compared with WT mice (Figure 6B,C). The spadin‐sensitive currents were consistent with these results in the NHERF‐1 WT and Het astrocytes (Figure 6D,E). Together, these results suggest that NHERF‐1 regulates TWIK‐1/TREK‐1 heterodimeric channels‐mediated astrocytic passive conductance.

FIGURE 6.

Heterozygous NHERF‐1 KO mice showed an enhanced TREK‐1‐mediated astrocytic passive conductance. (A) Representative traces of passive conductance induced by voltage stepping from −160 mV to +40 mV in hippocampal astrocytes from NHERF‐1 WT or Het mice treated with or without 10 μM spadin (SP). (B) I‐V curves of passive conductance in A. (C) Bar graphs showing averaged currents at −160 mV and +40 mV. (D) SP‐sensitive currents were calculated by difference between currents before and after SP application in B. (E) Summary bar graph of SP‐subtracted currents plotted at −160 mV and +40 mV (N = 4, respectively). All values are mean ± SEM. p‐values were obtained with one‐way ANOVA followed by Bonferroni's post hoc test or Student's t‐test. ns: not significant, *p < 0.05, **p < 0.01, and ***p < 0.001.

3.6. Dysfunction of Passive Conductance by Astrocyte‐Specific NHERF‐1 Overexpression Accelerates Seizure Susceptibility

Among the previous reports, the expression of NHERF‐1 is exhibited in various cells of the brain such as ependymal epitheliums (Treat et al. 2016). Because the NHERF‐1 KO can't exclude the effects of another cell type, we injected the NHERF‐1 overexpressing virus using astrocyte‐specific truncated GFAP (GfaABC1D) promoter into the hippocampal CA1 SR regions for 2 weeks (Figure 7A). The expression of mCherry protein was observed only in astrocytes, not neurons, of the hippocampal tissue infected with the AAV‐GFAPp‐NHERF‐1‐2A‐mCherry. This result was confirmed using NHERF‐1 antibodies (Figure 7B and Figure S7A). In western blot experiments conducted under the same experimental conditions with NHERF‐1‐specific antibodies, both endogenous NHERF‐1 and overexpressed NHERF‐1‐2A were detected at approximately 50 kDa. Additionally, only the overexpressed NHERF‐1‐2A was selectively identified using 2A‐specific antibodies (Figure S7B). Next, we investigated whether NHERF‐1 regulated passive conductance in hippocampal astrocytes. The electrophysiological recording exhibited a huge passive conductance in astrocytes infected AAV‐GFAPp‐mCherry, but when NHERF‐1 was overexpressed by AAV‐GFAPp‐NHERF‐1‐2A‐mCherry, passive conductance was dramatically decreased. The reduction of astrocytic passive conductance by NHERF‐1 overexpression had no effect after spadin treatment (Figure 7C–E). The spadin‐sensitive currents were coherent with these results (Figure 7F,G).

FIGURE 7.

Dysfunction of passive conductance by astrocyte NHERF‐1 overexpression accelerates seizure susceptibility. (A) Diagram showing injection locations of AAV‐GFAPp‐mCherry or AAV‐GFAPp‐NHERF‐1‐2A‐mCherry in hippocampal CA1 SR. (B) Left; representative fluorescence images of dual immunostaining of MAP2 (green), AAV‐GFAPp‐NHERF‐1‐2A‐mCh (red) and GFAP (purple) with DAPI staining of nuclei (blue) in mouse hippocampus. Scale bar, 100 μm. Right; enlarged images of B. Scale bar, 50 μm. (C) Representative traces of passive conductance induced by voltage stepping from −160 mV to +40 mV in AAV‐GFAPp‐mCherry or AAV‐GFAPp‐NHERF‐1‐2A‐mCherry expressing astrocytes of SR with or without 10 μM spadin (SP) respectively. (D) I‐V curves of passive conductance in C. (E) Bar graphs showing averaged currents at −160 mV and +40 mV. (F) SP‐sensitive currents were calculated by difference between currents before and after spadin application in D. (G) Summary bar graph of SP‐subtracted currents plotted at −160 mV and +40 mV (N = 3 or 4, respectively). (H) Schematic representation of the experimental timeline. Each virus was injected into mice, and 2 weeks later, an additional injection of 35 mg/kg KA was administered intraperitoneally (ip) into the abdominal cavity. (I) Time course of mean behavioral seizure score following KA injection. (J) Total cumulative seizure score for each group. (K) Latency to Score 3 seizure for each group. (L) Representative Fluoro‐Jade B staining images of hippocampal tissues injected with AAV‐GFAP‐mCh or AAV‐GFAPp‐NHERF‐1‐2A‐mCh. (M) Summary bar graph of Fluoro‐Jade B + neuronal cell counts. (N) Summary bar graph of DAPI + neuronal cell counts. The number on each bar indicates n for each condition. All values are mean ± SEM. p‐values were obtained with one‐way ANOVA followed by Bonferroni's post hoc test or Student's t‐test. ns: not significant, *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001.

Dysregulation of extracellular K⁺ concentration via astrocytic potassium buffering involved in epileptic seizure (David et al. 2009; Wang et al. 2020). To identify the seizure sensitivity when NHERF‐1 overexpressed in astrocytes, which showed decreased astrocytic passive conductance, we observed seizure behavior following the intraperitoneally (ip) injection of 35 mg/kg KA in the AAV‐GFAPp‐mCherry or AAV‐GFAPp‐NHERF‐1‐2A‐mCherry‐infected mice (Figure 7H). The behavioral seizure responses were monitored using a modified Racine score for 90 min after the KA injection (Racine 1975). KA‐induced seizure responses were bigger in AAV‐GFAPp‐NHERF‐1‐2A‐mCherry‐infected mice than in control mice during 90 min (Figure 7I). We also found that the cumulative score of NHERF‐1 overexpressed mice was higher than control mice. But latency to Score 3 was not altered in between both mice. (Figure 7J,K). Furthermore, we investigated whether overexpression of NHERF‐1 enhanced cell degeneration or death in hippocampal tissue, which is highly sensitive to seizures. We established a KA model in mice infected with either AAV‐GFAPp‐mCh or AAV‐GFAPp‐NHERF‐1‐2A‐mCh and performed Fluoro‐Jade B staining on the hippocampal tissue. While the number of DAPI (+) neurons did not significantly decrease, the number of Fluoro‐Jade B (+) neurons was significantly increased in the group with overexpressed NHERF‐1 in astrocytes (Figure 7L–N). These data suggested that When NHERF‐1 is overexpressed only in astrocytes, the seizure response is severe because of the reduction of astrocytic passive conductance.

3.7. Additional Overexpression of TREK‐1 N4 Effectively Reverses the Effects of NHERF‐1 Overexpression

Next, we asked whether TREK‐1 N4, the interaction region of TREK‐1 and NHERF‐1, prevented the dysfunction of passive conductance and KA‐induced seizure susceptibility in NHERF‐1 overexpression mice. Before the in vivo experiments, we confirmed the recovery ability by TREK‐1 N4 of K⁺ currents under the cultured system. K⁺ currents of mCherry‐NHERF‐1 transfected astrocytes were decreased, whereas they recovered by HA‐TREK‐1‐2A‐GFP overexpression (Figure S8A,B). Next, we generated TREK‐1 N4 overexpression virus using GFAP (GfaABC1D) promoter and injected AAV‐GFAPp‐NHERF‐1‐2A‐mCherry and/or AAV‐GFAPp‐TREK‐1 N4 2A‐GFP into the hippocampal CA1 SR regions for 2 weeks (Figure 8A). The expression mCherry and GFP protein was observed only in astrocytes, not neurons, of the hippocampal tissue infected with the AAV‐GFAPp‐NHERF‐1‐2A‐mCherry and/or AAV‐GFAPp‐TREK‐1 N4‐2A‐GFP (Figure 8B and Figure S9A,B). Same as the cultured system, reduced TWIK‐1/TREK‐1‐mediated passive conductance in NHERF‐1 astrocyte‐specific overexpressed mice was rescued by AAV‐GFAPp‐TREK‐1 N4‐2A‐GFP infection (Figure 8C–G). The KA‐induced behavioral seizure responses during the 90 min observation indicated that AAV‐GFAPp‐NHERF‐1‐2A‐mCherry and AAV‐GFAPp‐TREK‐1 N4‐2A‐GFP‐infected mice showed delayed seizure onset compared to AAV‐GFAPp‐NHERF‐1‐2A‐mCherry only infected mice (Figure 8H,I). The cumulative score for 20 min also decreased in AAV‐GFAPp‐NHERF‐1‐2A‐mCherry and AAV‐GFAPp‐TREK‐1 N4‐2A‐GFP‐infected mice (Figure 8J). However, latency to Score 3 was not changed between both mice (Figure 8K). These results highlight that the decreased potassium conductance of astrocytes and increased seizure susceptibility caused by NHERF‐1 overexpression were restored by additional TREK‐1 N4.

FIGURE 8.

NHERF‐1 overexpression‐mediated astrocytes dysfunction and seizure behavior were regained by TREK‐1 N4. (A) Diagram showing injection locations of AAV‐GFAPp‐NHERF‐1‐2A‐mCherry and/or AAV‐GFAPp‐TREK‐1 N4‐2A‐GFP in hippocampal CA1 SR. (B) Left; representative fluorescence images of dual immunostaining of AAV‐GFAPp‐TREK‐1 N4‐2A‐GFP (green), AAV‐GFAPp‐NHERF‐1‐2A‐mCh (red) and GFAP (purple) with DAPI staining of nuclei (blue) in mouse hippocampus. Scale bar, 50 μm. Right; enlarged images of B. Scale bar, 20 μm. (C) Representative traces of passive conductance induced by voltage stepping from −160 mV to +40 mV in AAV‐GFAPp‐NHERF‐1‐2A‐mCherry and/or AAV‐GFAPp‐TREK‐1 N4‐2A‐GFP expressing astrocytes of SR with or without 10 μM spadin (SP) respectively. (D) I‐V curves of passive conductance in C. (E) Bar graphs showing averaged currents at −160 mV and +40 mV. (F) SP‐sensitive currents were calculated by difference between currents before and after spadin application in D. (G) Summary bar graph of SP‐subtracted currents plotted at −160 mV and +40 mV (N = 2 or 3, respectively). (H) Schematic representation of the experimental timeline. Each virus was injected into mice, and 2 weeks later, an additional injection of 35 mg/kg KA was administered intraperitoneally (ip) into the abdominal cavity. (I) Time course of mean behavioral seizure score following KA injection. (J) Cumulative seizure score over 20 min for each group. (K) Latency to Score 3 seizure for each group. The number on each bar indicates n for each condition. All values are mean ± SEM. p‐values were obtained with one‐way ANOVA followed by Bonferroni's post hoc test. ns: not significant, *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001.

4. Discussion

In this study, we found that NHERF‐1 is abundantly expressed in astrocytes and not only regulates intrinsic K⁺ currents in cultured cells (Figure 1) but also contributes to TWIK‐1/TREK‐1‐mediated passive conductance in adults (Figure 2). Interestingly, we discovered NHERF‐1 as a novel binding protein for the TREK‐1 channel and confirmed its protein–protein interaction in vitro (Figures 3 and 4). Moreover, NEHRF‐1 was associated with endogenous TREK‐1 in cultured astrocytes, and overexpression or knockdown of NHERF‐1 regulated the surface expression of TREK‐1 (Figure 5). Astrocyte passive conductance was increased in NHERF‐1 het KO mice with reduced expression of NHERF‐1 (Figure 6), whereas NHERF‐1 overexpression had the opposite effect. Astrocyte‐selective overexpression of NHERF‐1 reduced astrocyte passive conductance and predisposed KA‐evoked seizure responses (Figure 7). However, this phenomenon is significantly restored by TREK‐1 N4, which interferes with the binding of NHERF‐1 to TREK‐1 (Figure 8). The present study demonstrates that NHERF‐1 can act as an effective regulator of TREK‐1‐mediated passive conductance in glial cells. Additionally, it provides evidence that changes in NHERF‐1 expression can influence KA‐induced seizure sensitivity through alterations in astrocytic passive conductance.

One of the unique functions of astrocytes is the regulation of ionic homeostasis in the brain, such as extracellular K⁺ buffering through K⁺ channels. Astrocytes show a linear I‐V relationship with a negative membrane potential (Vm), called passive conductance. This phenomenon results from leakage of K⁺ membrane conductance, and it is not affected by time and voltage (Nwaobi et al. 2016; Ryoo and Park 2016; Bellot‐Saez et al. 2017). Although the exact molecular mechanism is not yet known, it appears to be partly due to weakly rectifying K⁺ (Kir) channels and two‐pore domain K⁺ (K2P) channels, especially the TWIK‐1/TREK‐1 heterodimer channel (Zhou et al. 2009; Hwang et al. 2014; Olsen et al. 2015; Nwaobi et al. 2016). These changes in ionic homeostasis can induce a variety of symptoms. Mice lacking Kir4.1 in astrocytes show no change in passive conductance but rapidly hyperpolarize the membrane potential of astrocytes and develop severe spontaneous seizures after birth (Djukic et al. 2007). However, TWIK‐1/TREK‐1 double KO mice not only did not show the expected changes in astrocytic passive conductance but also showed no pathological behavior (Du et al. 2016). Since this may be due to a compensatory effect during development, we thought that experiments using viruses in adults would be a more reliable alternative to confirm the role of effective passive conductance. In this study, we used an NHERF‐1 overexpressing virus in adults and observed that it can regulate astrocytic passive conductance through interaction with TREK‐1. NHERF‐1 is a member of the PDZ family, which represents the most common protein–protein interaction domain (Feng and Zhang 2009; Dunn and Ferguson 2015; Bushau‐Sprinkle and Lederer 2020). Subcellular localization, endocytosis, transport, and signaling of many receptors and channels are regulated through interaction with PDZ domains in various cells or tissues (Lee and Zheng 2010; Dunn and Ferguson 2015). We demonstrated that TREK‐1 binds to the PDZ domain of NHERF‐1 and that this interaction regulates spadin‐sensitive astrocytic passive conductance (Figure 8). These findings suggest the possibility that TREK‐1 may interact with other families of proteins that possess PDZ domains in addition to NHERF‐1.

TREK‐1 has a short N‐terminus consisting of 46 amino acids and most of the reported TREK‐1 binding proteins bind to its relatively long C‐terminus. Interestingly, in this study, NHERF‐1 was identified as the protein that binds to the N‐terminus of TREK‐1, next to β‐COP. β‐COP was identified as an interacting protein of TREK‐1 in astrocytes, binding to the N‐terminal domain and increasing both surface expression and channel activity (Kim et al. 2010, 2022). However, NHERF‐1 is bound to the N‐terminal domain of TREK‐1 but reduces both surface expression and channel activity. Because both of these proteins bind to the TREK‐1 N‐terminal domain, especially the N4 region, and induce opposite effects, it may be important which protein binds predominantly in a physiological condition. We observed no changes in K⁺ currents when only N4 was expressed in cultured astrocytes (data not shown). If NHERF‐1 expression is relatively increased in certain circumstances, such as cancer development, a decrease in TREK‐1‐mediated K⁺ currents in astrocytes is expected (Figure 7). It has been reported that many glioblastoma patients suffer from headaches and seizures (Pace et al. 2009; Sizoo et al. 2010) and one of the reasons for this may be the increased seizure susceptibility caused by NHERF‐1 overexpression in astrocytes, according to our results. Further studies are needed to determine the causal relationship between increased expression of NHERF‐1 and pathological conditions.

NHERF‐1 is widely distributed in the rodent brain and is abundantly expressed in astrocytes rather than neurons. This is crucial because astrocytes are a specific group of glial cells that have functions such as the control of ionic and osmotic composition of the extracellular environment. NHERF‐1 in astrocytes supports a role for the PDZ domain as a possible regulator of membrane transporters and channels. In a previous report, excitatory amino acid transporter 1 (EAAT1), called GLAST, colocalized and bound with NHERF‐1 in astrocytes. This suggests that NHERF‐1 may regulate glutamate homeostasis through GLAST in astrocytes (Lee et al. 2007). Additionally, glial fibrillary acidic protein (GFAP), a characteristic intermediate filament protein of astrocytes, stabilizes and maintains GLAST at the plasma membrane through a series of intermediate linking proteins, including NHERF‐1 and ezrin (Sullivan et al. 2007). Therefore, this report demonstrates a novel function for NHERF‐1 in regulating TWIK‐1/TREK‐1‐mediated astrocytic passive conductance. Despite the many roles of NHERF‐1 in astrocytes, there are no reports of NHERF‐1 being significantly involved in brain function, other than reports that NHERF‐1 KO mice exhibit hydrocephalus (Georgescu et al. 2015; Treat et al. 2016). In this respect, this report provides new compelling evidence on the function of NHERF‐1 in the brain, especially in astrocytes.

We also found that the reduced astrocyte passive conductance under NHERF‐1 overexpression was restored by TREK‐1 N4 addition, and KA‐evoked seizure behavior was also significantly rescued (Figure 8). These findings suggest that dysfunction of astrocyte passive conductance plays an important role in epileptic seizures and that dissociation of TREK‐1 and NHERF‐1 complexes should be considered a promising target for seizure therapy. This new target may resolve various side effects caused by existing antiseizure drugs targeting nerve cells. In conclusion, we have identified NHERF‐1 as a novel binding protein for TWIK‐1/TREK‐1 channels and studied the regulatory mechanisms between them, resulting in a better understanding of the function of astrocytes. Additionally, we proposed a new therapeutic strategy targeting astrocytes for the treatment of seizures.

Author Contributions

G.‐S.Y., J.‐Y.P. and E.M.H. designed this study, while J.‐Y.P. and E.M.H. supervised and edited the manuscript. Y.B. conducted most of the experiments and acquired the data. A.K. injected AAV and performed behavioral experiments. S.L. conducted most of the revised experiments, and J.N. performed the whole‐cell patch experiments using a newly generated mutant of TREK‐1 (AWA mutant). S.L. conducted most co‐IP experiments using overexpressed cells, and S.‐S.K. performed co‐IP experiments using TREK‐1 Δ4 mutant overexpressed cells. K.R. conducted co‐IP experiments using primary cultured astrocytes. S.P. performed binding analyses using a yeast two hybrid system.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Data S1.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.