Haploinsufficiency of GABAA Receptor-Associated Clptm1 Enhances Phasic and Tonic Inhibitory Neurotransmission, Suppresses Excitatory Synaptic Plasticity, and Impairs Memory
- Djavad Mowafaghian Centre for Brain Health and Department of Psychiatry, University of British Columbia, Vancouver, British Columbia V6T 2B5, Canada
-
Author contributions: Y.G. and A.M.C. designed research; Y.G. performed research; Y.G. analyzed data.
Abstract
The mechanisms utilized by neurons to regulate the efficacy of phasic and tonic inhibition and their impacts on synaptic plasticity and behavior are incompletely understood. Cleft lip and palate transmembrane protein 1 (Clptm1) is a membrane-spanning protein that interacts with multiple γ-aminobutyric acid type A receptor (GABAAR) subunits, trapping them in the endoplasmic reticulum and Golgi network. Overexpression and knock-down studies suggest that Clptm1 modulates GABAAR-mediated phasic inhibition and tonic inhibition as well as activity-induced inhibitory synaptic homeostasis in cultured hippocampal neurons. To investigate the role of Clptm1 in the modulation of GABAARs in vivo, we generated Clptm1 knock-out (KO) mice. Here, we show that genetic KO of Clptm1 elevated phasic and tonic inhibitory transmission in both male and female heterozygous mice. Although basal excitatory synaptic transmission was not affected, Clptm1 haploinsufficiency significantly blocked high-frequency stimulation-induced long–term potentiation (LTP) in hippocampal CA3→CA1 synapses. In the hippocampus-dependent contextual fear-conditioning behavior task, both male and female Clptm1 heterozygous KO mice exhibited impairment in contextual fear memory. In addition, LTP and contextual fear memory were rescued by application of L-655,708, a negative allosteric modulator of the extrasynaptic GABAAR α5 subunit. These results suggest that haploinsufficiency of Clptm1 contributes to cognitive deficits through altered synaptic transmission and plasticity by elevation of inhibitory neurotransmission, with tonic inhibition playing a major role.
- Clptm1
- GABA receptor
- GABAA receptor alpha5
- inhibitory synapse
- long-term potentiation
- tonic inhibition
Significance Statement
The CLPTM1 gene was originally identified as disrupted in a family with cleft lip and palate. At the molecular level, Clptm1 interacts with multiple GABAA receptor subunits to limit their surface expression. Here, we generated Clptm1 knock-out mice to uncover its functions in vivo. Clptm1 not only limited hippocampal inhibitory phasic and tonic transmission; it was also required for excitatory synaptic plasticity and hippocampus-dependent cognitive function. A modulator of extrasynaptic GABAA receptors rescued the deficits in plasticity and behavior in Clptm1 heterozygous knock-out mice, indicating the importance of tonic inhibition. These findings reveal a role for Clptm1 in balancing inhibitory strength and raise the possibility that disruptions of Clptm1 function may contribute to synaptic and cognitive deficits in neurological diseases.
Introduction
γ-Aminobutyric acid type A receptors (GABAARs) are the major mediators of inhibitory neurotransmission in the mammalian central nervous system. Native GABAARs are constituted from 19 subunits (α1–6, β1–3, γ1–3, δ, ε, θ, π, and ρ1–3) and mediate two forms of inhibition: phasic and tonic (Farrant and Nusser, 2005). Phasic inhibition is mediated by postsynaptic GABAARs that are mainly composed of two α1, α2, or α3 subunits together with two β2 or β3 subunits and a single γ2 subunit, whereas tonic inhibition is generated primarily by extrasynaptic GABAARs that contain either the δ subunit or α5 subunits (Farrant and Nusser, 2005; Glykys et al., 2008). Precise control of GABAergic strength is critical for maintaining accurate neuronal network activity and normal brain function, and impairment of GABAergic transmission has been implicated in the pathologies of many brain and neurodevelopmental disorders including epilepsy (Sperk et al., 2004; Bryson et al., 2023), depression (Luscher et al., 2011, 2023), anxiety (Lydiard, 2003), autism spectrum disorders (Zhao et al., 2021), and schizophrenia (de Jonge et al., 2017).
Using a combined transgenic and proteomic approach, we have isolated a GABAAR-interacting protein, cleft lip and palate transmembrane protein 1 (Clptm1; Ge et al., 2018). The CLPTM1 gene was originally identified as being disrupted in a family with cleft lip and palate (Yoshiura et al., 1998). However, our study revealed that Clptm1 interacts with multiple GABAAR subunits and modulates surface levels of GABAARs by trapping them in the endoplasmic reticulum and Golgi complex (Ge et al., 2018). Altering expression levels of Clptm1 occludes some forms of homeostatic activity regulation and affects tonic as well as phasic inhibitory transmission with no impact on excitatory transmission (Ge et al., 2018). Clptm1 is relatively unique in targeting all GABAAR subunits tested and is a negative modulator with the function of inhibiting excessive surface expression of GABAARs. Overexpression of Clptm1 in the lateral habenula downregulates synaptic inhibition and reduces reward anticipatory behavior (Lalive et al., 2022). Clptm1 was further reported to regulate epileptic seizures by modulating GABAAR-mediated inhibitory synaptic transmission in a pentylenetetrazol-induced epilepsy model in rats (X. L. Zhang et al., 2023), and de novo CLPTM1 variants have been found to be associated with epilepsy in patients (Liu et al., 2023). A multitissue neocortical transcriptome-wide association study found that downregulation of Clptm1 is significantly associated with Alzheimer's disease (Gockley et al., 2021). However, all of the studies to date on Clptm1were either correlative or performed using knock-down or overexpression; the function of Clptm1 has not been validated by a targeted genetic approach in vivo.
We further suggest that Clptm1, as a GABAAR-associated protein modulating GABAAR-mediated inhibitory neurotransmission, may modulate activity-dependent excitatory synaptic plasticity and memory, a direction that has not yet been explored. Activity-dependent excitatory synaptic plasticity, particularly long-term potentiation (LTP), has long been believed to be a fundamental subcellular mechanism of learning and memory (S. J. Martin et al., 2000; Takeuchi et al., 2014; Humeau and Choquet, 2019). While the pharmacological manipulation of GABAAR function using agonists or antagonists has been found to impair or enhance synaptic plasticity and memory behavior (C. Castellano and McGaugh, 1990; Luft et al., 2004; D. H. Kim et al., 2012), the endogenous mechanisms utilized by neurons to regulate the efficacy of phasic and tonic inhibition and their impacts on synaptic plasticity and behavior are still poorly understood.
Here, we used a novel Clptm1 knock-out (KO) mouse model to study whether genetic reduction of Clptm1 alters hippocampal inhibitory neurotransmission and how this affects LTP and contextual fear memory.
Materials and Methods
Mice
All experimental procedures with animals were conducted following the guidelines of the Canadian Council for Animal Care and were approved by the University of British Columbia Animal Care Committee.
The Clptm1tm1a(KOMP)Wtsi (MGI:4419487) stem cell line (EPD0395_5_D10) was obtained from UC Davis KOMP Repository (Skarnes et al., 2011). The stem cells were injected into albino C57BL/6 blastocysts to generate chimeras in the Mouse Genetics Core, Washington University, in St. Louis. Founder chimeras were backcrossed to C57BL/6J mice at least eight times prior to any experiments.
Clptm1tm1a(KOMP)Wtsi, the “KO-first” (KOfirst) allele, contains a cassette composed of flippase recombination target sites, a lacZ sequence, a neomycin resistance gene under the control of the human β-actin promoter, SV40 polyA, and loxP sites flanking exon 6 of the Clptm1 gene (Skarnes et al., 2011). Clptm1 heterozygous KOfirst mice were crossed with a FLP1-expressing deleter line [B6; SJL-Tg(ACTFLPe)9205Dym/J; Jackson Laboratory] to generate Clptm1tm1c(KOMP)Wtsi (flox) mice.
Brain immunofluorescence and imaging
For confocal imaging, 22-d-old male Clptm1 KOfirst/+ mice were anesthetized with 20% urethane and perfused transcardially with 4% paraformaldehyde in PBS. The brains were then collected and postfixed with 4% paraformaldehyde in PBS overnight. The fixed brains were rinsed with PBS three times and then embedded in 3% low-melting agarose. Sagittal sections were cut at 100 µm thickness using a vibratome (VT1000s, Leica Camera). The sections were incubated in blocking solution (5% BSA, 5% normal goat serum, and 0.25% Triton X-100 in PBS) for 1 h at room temperature followed by primary antibodies overnight at 4°C. The following primary antibodies were used: anti-β-gal (1:500; Abcam catalog ab9361; RRID, AB_307210); anti-NeuN (1:500; Merck Millipore catalog #MAB377; RRID, AB_2298772). The sections were then washed with PBS and incubated with appropriate Alexa-conjugated secondary antibodies (1:500; Invitrogen) for 1 h at room temperature. For fluorescent labeling of nuclei, 4′,6-diamidino-2-phenylindole (100 ng/ml) was included with the secondary antibodies. The sections were washed with PBS and mounted in Elvanol [Tris–HCl, glycerol, polyvinyl alcohol, 2% 1,4-diazabicyclo(2,2,2)octane]. Images were captured using a Zeiss LSM 700 confocal microscope with a 25 × 0.8 numerical aperture lens and customized filter sets.
Neuron culture and transfection
Dissociated hippocampal neuron cultures were prepared from pooled male and female Embryonic Day (E)18 mice. The method was essentially as previously described for rat neurons (Kaech and Banker, 2006) except mouse neurons were dissociated using papain (20 units/ml, 15 min, 37°C) and cultured with additional insulin in the media. The hippocampal cells were plated on glass coverslips coated with poly-L-lysine (PLL) at a density of 300,000 per 60 mm dish. The coverslips were inverted over a feeder layer of astrocytes in Neurobasal Medium (Invitrogen) supplemented with B27 (Invitrogen or STEMCELL Technologies) and 25 mg/ml bovine pancreatic insulin (Sigma-Aldrich). Cytosine arabinoside (5 mM) was added to neuron culture dishes at 2 d in vitro (DIV) to prevent overgrowth of glial cells. For GABAAR γ2 surface staining and miniature inhibitory postsynaptic current (mIPSC) recordings in Clptm1 flox/flox neuron culture, hippocampal neurons were transfected with pCAG-Cre (Addgene #13775) and pFB-hSyn-Dio-GFP (P. Zhang et al., 2018) or pLL3.7-hSyn-YFP (Ge et al., 2018) as control at 0 DIV using nucleofection (Amaxa Biosystems), and electrophysiological recordings or immunostaining were performed at 14 DIV.
For the Western blotting assay, Clptm1 flox/flox mouse cortical neurons were plated on 12-well plates coated with PLL at a density of one million per well and cultured in the same medium as hippocampal neurons without the addition of cytosine arabinoside. AAV-GFP-Cre or AAV-GFP was added in the medium at 3 DIV, and cell lysates were harvested for Western blotting at 14 DIV.
Western blotting
Cultured cortical neurons (Fig. 2A) were homogenized in ice-cold RIPA buffer containing 50 mM Tris–HCl, pH 7.4, 1% Triton X-100, 150 mM NaCl, 1 mM EDTA, 0.5% deoxycholic acid sodium, and a cocktail of protease inhibitors (Roche). After incubation on ice for 30 min, the homogenates were centrifuged at 14,000 rpm at 4°C for 15 min. The supernatants were collected, and the total protein concentrations were measured using DC Protein Assay Kit (Bio-Rad).
Brains collected from male and female embryos (Fig. 1B) or mice at different developmental stages (Fig. 5A) were homogenized in a homogenate buffer [320 mM sucrose, 5 mM HEPES, 1 mM MgCl2, 0.5 mM CaCl, pH 7.4; supplemented with protease inhibitor cocktail (Roche)] using a glass Teflon homogenizer. The homogenates were spun at 1,400 × g for 10 min at 4°C, and the supernatants were transferred to new tubes. Protein concentration was determined using DC Protein Assay Kit (Bio-Rad).
Equal amounts of total protein (20–30 µg) in the homogenates were heated in a sample buffer at 60°C for 5 min, separated on a 10% sodium dodecyl sulfate-polyacrylamide gel (prepared using TGX FastCast Acrylamide Kit, Bio-Rad), and then transferred to a polyvinylidene difluoride membrane (Merck Millipore). The membrane was blocked with 5% milk for 1 h at room temperature, immunoblotted with primary antibodies overnight at 4°C, and then incubated with HRP-conjugated secondary antibodies (1:10,000, SouthernBiotech) for 1 h at room temperature. Blots were developed using the Immobilon Western Chemiluminescent HRP Substrate Kit (Merck Millipore) and imaged with the Bio-Rad gel imaging system (Bio-Rad). Protein band intensities were quantified with the Image Lab software (Bio-Rad).
The following primary antibodies were used: anti-Clptm1 (1:2,000; Abcam catalog #ab174839), anti-GFP [1:2,000; Thermo Fisher Scientific catalog #A-11122 (also A11122); RRID, AB_221569], anti-GABAAR γ2 (1:1,000; Novus International catalog #NB300-190; RRID, AB_2294344), anti-α-Tubulin (1:10,000; Merck Millipore catalog #05-829; RRID, AB_310035).
Immunocytochemistry
For detecting surface GABAAR γ2 or α5, cultured hippocampal neurons were live stained with anti-GABAAR γ2 antibody (1:500; Synaptic Systems catalog #224 003; RRID, AB_2263066) or anti- GABAAR α5 antibody (1:500; Synaptic Systems catalog #224 503; RRID, AB_2619944) in conditioned media for 60 min at 37°C. Neurons were then fixed in a parafix solution (4% paraformaldehyde and 4% sucrose in PBS), pH 7.4, for 15 min followed by permeabilization with 0.25% Triton X-100 in PBS for 5 min. They were then incubated with blocking solution (3% bovine serum albumin and 5% normal goat serum in PBS) for 45 min at 37°C, followed by incubation with primary antibodies [anti-vesicular GABA transporter (VGAT; 1:3,000; Synaptic Systems catalog #131 004; RRID, AB_887873) or anti-gephyrin (1:300; Synaptic Systems catalog #147 021; RRID, AB_1279448) and anti-MAP2 (1:4,000; Abcam catalog #ab5392; RRID, AB_2138153)] in a blocking solution (overnight, 4°C) and the appropriate Alexa-conjugated secondary antibody (1:500; Invitrogen; 45 min, 37°C). Images were obtained on a Zeiss LSM 700 microscope with a 40 × 1.4 numerical aperture oil objective, an Orca-Flash4.0 CMOS camera (Hamamatsu Photonics), and customized filter sets. Measures were made using ImageJ (NIH). For quantification of fluorescence intensity and puncta number of synaptic GABAAR γ2, synapses were identified as clusters with pixel overlap between the separately thresholded VGAT and GABAAR γ2 channels, and the number of synapses was normalized to the area of MAP2 positive dendrites that were randomly selected.
AAV injection
The AAV was purchased from the University of Pennsylvania Vector Core. On the day of birth [Postnatal Day (P)0], Clptm1 flox/flox neonates were anesthetized using isoflurane and bilaterally injected with 0.5 µl of viral vector into each cerebral lateral ventricle through a finely drawn glass micropipette as described previously (J. Y. Kim et al., 2014). Electrophysiological recording was performed at postnatal day15–19.
Electrophysiological recording
For recording in cultured hippocampal neurons, whole-cell recordings were made from the cultures at room temperature (20–22°C) at 14 DIV. Neurons were visualized using infrared differential interference contrast and fluorescence microscopy (Olympus) and only GFP- or YFP-positive neurons were chosen for recording. Neurons were continuously perfused (1 ml/min) with the extracellular solution containing the following (in mM): 140 NaCl, 1.3 CaCl2, 5.4 KCl, 1 MgCl2, 25 HEPES, 33 glucose, pH 7.35 (osmolality 300–310 mosmol/Kg). mIPSCs were recorded at a holding potential of −60 mV and in the presence of 0.5 µM tetrodotoxin, 10 µM 6-cyano-7-nitroquinoxaline-2,3-dione (CNQX), and 100 µM DL-2-amino-5-phosphonopentanoic acid sodium salt (APV). The patch pipettes were pulled from borosilicate glass capillary tubes (World Precision Instruments) using a PP-830 pipette puller (Narishige). The resistance of pipettes was 4–6 MΩ. The patch pipette solution contained the following (in mM): 140 CsCl, 0.1 CaCl2, 2 MgCl2, 10 HEPES, 10 BAPTA, 4 ATP, and 5 QX314, pH 7.2 (osmolality 280–290 mosmol/Kg). The series resistance in these recordings varied from 6 to 10 MΩ, and recordings where series resistance varied by >10% were rejected. No electronic compensation for series resistance was employed.
For recording in hippocampal slices, the mice were anesthetized by isoflurane inhalation, and the brain was rapidly removed and immersed in ice-cold artificial cerebrospinal fluid (ACSF) containing the following (in mM): 125 NaCl, 2.5 KCl, 2 CaCl2, 2 MgCl2, 1.25 NaH2PO4, 26 NaCO3, and 25 glucose (osmolality 310–320 mosmol/Kg) which was continuously bubbled with carbogen (95% O2/5% CO2) to adjust the pH to 7.35. The brain was then cut into 400-µm-thick coronal slices containing the hippocampus with a vibratome (VT1000s, Leica Camera). Freshly cut slices were placed in a recovery chamber with carbogenated ACSF at 31°C for 60 min. Slices were then maintained at room temperature prior to recording. For whole-cell recording, slices were transferred to a recording chamber continuously perfused with carbogenated ACSF (2 ml/min). The patch pipettes were pulled from borosilicate glass capillary tubes (World Precision Instruments) using a PP-830 pipette puller (Narishige). The resistance of pipettes was 4–6 MΩ. The holding potential was −60 mV. AAV-infected neurons were visualized using infrared differential interference contrast and fluorescence microscopy (Olympus), and only GFP-positive neurons were chosen for recording. For mIPSC recordings, the same blockers were supplemented in the perfusion ACSF, and the same patch pipette solutions were used as in cultured hippocampal neurons. Miniature excitatory postsynaptic currents (mEPSCs) were recorded in the presence of 0.5 µM tetrodotoxin, 10 µM bicuculline methiodide, and 100 µM APV, and the patch pipette solution contained the following (in mM): 122.5 Cs gluconate, 17.5 CsCl, 2 MgCl2, 10 HEPES, 10 BAPTA, 4 ATP, and 5 QX314, pH 7.2 (osmolality 280–290 mosmol/Kg). To record tonic GABAergic currents, we recorded hippocampal CA1 pyramidal neurons in ACSF supplemented with 5 µM GABA, 10 µM CNQX, and 100 µM APV. The GABAAR-competitive antagonist bicuculline (20 μM) was added in perfusion solution after obtaining a stable baseline recording. The amplitude of the tonic current was calculated as the difference between the holding current measured before and after the application of bicuculline. The holding current was calculated from segments containing no synaptic currents. The series resistance in these recordings varied from 6 to 15 MΩ, and recordings where series resistance varied by >10% were rejected. No electronic compensation for series resistance was employed.
Evoked inhibitory postsynaptic currents (eIPSCs) were recorded in the presence of 10 µM CNQX and 100 µM APV, while evoked excitatory postsynaptic currents (eEPSCs) were recorded in the presence of 10 µM bicuculline methiodide and 100 µM APV. The same patch pipette solutions were used as for mIPSC or mEPSC recordings. The holding potential was −60 mV. For evoked whole-cell recordings, the stimulating electrode was placed in the stratum radiatum within 100–200 µm from the whole-cell–patched CA1 pyramidal neuron. For input/output curves, eIPSCs or eEPSCs were recorded in response to a series of increasing stimulation intensities with one-time stimulation intensity being the minimal intensity to elicit a detectable response (∼30 pA for eIPSC and ∼8 pA for eEPSC). Presynaptic function was assayed through application of paired pulses while varying the interpulse interval and expressing the second pulse as a ratio of the first. The series resistance in these recordings varied from 6 to 15 MΩ, and recordings where series resistance varied by >10% were rejected. Electronic compensation for series resistance was employed.
Extracellular field EPSPs (fEPSPs) were elicited by using two bipolar nickel–chromium electrodes placed in the stratum radiatum of area CA1 and recorded with a glass microelectrode filled with ACSF (resistances, 2–3 MΩ). Input/output curves were generated by monitoring changes in fEPSP slopes in response to increasing stimulation intensities. Input/output curves were plotted as a function of presynaptic fiber volley amplitude. The initial slope of the fEPSP was measured to quantify synaptic strength. Baseline responses were set at 30% of maximum and taken for 20 min prior to inducing LTP. In high-frequency stimulation consisting of 1 × 100 Hz, 1 s duration was used to induce LTP. For the pharmacological studies, L-655, 708 (10 nM), or vehicle (0.1% DMSO) was added in the ACSF after obtaining 10 min stable baseline recording, and 10 min later, LTP was induced by high-frequency stimulation.
All recordings were performed using a MultiClamp 700A amplifier under the voltage-clamp mode (whole-cell recordings) or I = 0 mode (fEPSP; Molecular Devices), filtered at 2 kHz, and acquired with a pCLAMP10 software (Molecular Devices). Tonic GABAergic currents, eIPSC amplitude, eEPSC amplitude, and fEPSP slope were analyzed using the Clampfit10 software (Molecular Devices). mIPSCs or mEPSCs recorded in a 2 min period of each condition were analyzed using the Mini 6.0 software (Synaptosoft). The trigger level for detection of events was set approximately three times higher than the baseline noise, and manual inspection of the raw data was used to eliminate any false event. The decay time constants were analyzed from averaged mIPSCs for each cell and calculated over 90–37% of the decay phase using the Mini 6.0 software. The cumulative probability distribution plots were generated from randomly selected 50 mIPSC or 20 mEPSC events from each cell in a given condition using the GraphPad Prism software. A Kolmogorov–Smirnov (K-S) test was used to test whether there were significant differences between each group.
Fear conditioning
Mice were tested at 8–12 weeks of age. Thirty minutes before training, L-655,708 (0.7 mg/kg, i.p.) or the vehicle control (3% DMSO, i.p.) was injected. Each mouse was placed in a conditioning chamber (40 × 16 × 22 cm) for training. Animals were allowed to become acclimatized (freely move around) to the chamber for 3 min, and the freezing behavior (the absence of movement except for respiration) was monitored. After the acclimatization period, mice received three 2 s footshocks of 0.5 mA intensity with 60 s intervals. The animal remained in the chamber for an additional 60 s before being removed. All testing took place between 09:00 and 12:00 to minimize the influence of circadian variations. To test contextual fear memory, we returned the mice to the original chamber on the second day for a 5 min interval during which no footshock was applied. The freezing behavior was analyzed using the ToxTrac software (Rodriguez et al., 2017, 2018).
Quantification and statistical analysis
Statistical analysis and plotting were performed with Prism 9 (GraphPad Software). Values are expressed as mean ± SEM. For comparisons between two groups, if data passed a D'Agostino and Pearson’s test for normality, then Student's t test was used. Otherwise, the nonparametric Mann–Whitney test was used. For comparisons among multiple groups and factors, one-way, two-way, and three-way ANOVA followed by post hoc tests were used as indicated. Statistical significance is noted as *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001. Each experiment was repeated at least in two independent experiments, and n represents the number of cells, brain slices, or mice as indicated. Detailed information can be found in the figure legends.
Results
Clptm1 is ubiquitously expressed in mouse brain and KO is lethal
In order to study the function of Clptm1 in vivo, we generated Clptm1tm1a(KOMP)Wtsi “KO-first” (KOfirst) mice (Fig. 1A) using ES cells from the Mutant Mouse Resource and Research Center (Skarnes et al., 2011). The KOfirst allele was designed by inserting a lacZ trapping cassette and a floxed promoter-driven neo cassette into an intron along with floxing an exon of the CLPTM1 gene, which was expected to disrupt gene function (Skarnes et al., 2011). Disruption of Clptm1 protein in homozygous KOfirst mice was confirmed by Western blotting of brain homogenates collected at E18 (Fig. 1B). The litter size when female KOfirst/+ mice were crossed with male KOfirst/+ mice was smaller than when they were crossed with male WT mice (Fig. 1C), suggesting premature death of pups. Further investigation found that breeding of female KOfirst/+ mice with male WT mice generated pups with the expected genotype distribution (Fig. 1D), whereas breeding of female KOfirst/+ mice with male KOfirst/+ mice did not generate any homozygous KOfirst pups and failed to meet Mendelian expectations (Fig. 1E). Since the tail tissues for genotyping were collected at P9–11, which might pass the time point of premature death, we genotyped embryos at E18. About one-fourth of embryos were homozygous KOfirst, which met Mendelian expectations (Fig. 1E), suggesting that homozygous KOfirst mice can survive in the embryonic stage but die shortly after birth.
Generation and lethality of Clptm1 KO mice. A, The targeted locus in Clptm1tm1a(KOMP)Wtsi KOfirst mice (Skarnes et al., 2011). The KOfirst allele can be converted to the floxed (Flox) allele Clptm1tm1c(KOMP)Wtsi by Flp recombinase expression, and subsequent Cre expression results in a KO allele Clptm1tm1d(KOMP)Wtsi. B, Disruption of Clptm1 was confirmed by Western blotting of embryonic brain tissue lysates. The faint signal in the homozygous KOfirst sample may be a cross-reacting band, as a similar band was reported by Abcam for this antibody in a homozygous Clptm1 KO cell line. n = 4 mice; p < 0.0001, one-way ANOVA. C, The litter size was reduced in heterozygous KOfirst mouse crosses (n = 17 litters) compared with female heterozygous KOfirst and male WT crosses (n = 19 litters); p = 0.0423; t test. D, The genotype distribution of the pups from female heterozygous KOfirst and male WT crosses met expectations (n = 19 litters); p = 0.5856; binomial test. The tail tissues were collected for genotyping at P9–11. The dashed line represents the expected Mendelian distribution. E, The genotype distribution of the pups from heterozygous KOfirst crosses met expectations at E18 (6 litters; X2 = 1.500; p = 0.4724; χ2 test) but not P9–11 (17 litters; X2 = 29.72; p < 0.0001; χ2 test). The dashed lines represent the expected Mendelian distribution. F, Expression patterns in the brains of heterozygous Clptm1 KOfirst mice were examined by immunostaining for lac Z expression (β-gal). Scale bar represents 1 mm. G, Composite images through the cortex and hippocampus show widespread expression of lacZ from the Clptm1 KOfirst locus. β-Gal appears to be expressed by both NeuN-positive neurons and NeuN-negative non-neuronal cells. Scale bar represents 50 µm.
Clptm1 heterozygous KOfirst/+ mice were viable and fertile on a C57BL/6J background, with no obvious defects in gross brain morphology into adulthood (data not shown). Since Clptm1 KOfirst/+ mice were generated by knock-in of lacZ (Fig. 1A), we assessed the expression pattern of Clptm1 by immunostaining for lacZ expression (β-gal; Fig. 1F,G). At P22, lacZ at the Clptm1 locus was ubiquitously expressed in the mouse brain, particularly high in olfactory bulb, cortical layers, hippocampus, striatum, and cerebellar Purkinje cells (Fig. 1F). At the cellular level, in the hippocampal CA1 region based on cell positions and NeuN costaining, β-gal fluorescence appeared to be expressed by both pyramidal cells and interneurons (Fig. 1G). These expression patterns from the Clptm1 locus are consistent with previous reports for Clptm1 (Lein et al., 2007; Allen Mouse Brain Atlas, https://mouse.brain-map.org/experiment/show?id=71608394).
Inhibitory synaptic transmission was increased in Clptm1 KO hippocampal neurons
Since Clptm1 homozygous KO was lethal to mice at the early postnatal stage, in order to study the function of Clptm1, we next generated cultured embryonic hippocampal neurons lacking Clptm1. We first generated conditional KO (Clptm1tm1c(KOMP)Wtsi, flox) mice by crossing KOfirst/+ mice with FLP1-expressing mice and then prepared cultures of homozygous flox hippocampal neurons, mixed cultures from males and females. Equivalent sister Clptm1 homozygous flox culture dishes could then be compared as controls or upon Cre expression for Clptm1 homozygous KO (Clptm1tm1d(KOMP)Wtsi, KO). We confirmed by Western blotting that efficient AAV-mediated expression of Cre resulted in loss of detectable Clptm1 protein (Fig. 2A). For the phenotypic assays, we chose to generate sparse cell KO to assess cell-autonomous postsynaptic effects and avoid potential additional consequences of global KO. We thus used nucleofection to express Cre-dependent GFP (DIO-GFP) along with Cre and assessed GFP-expressing Clptm1 KO neurons compared with control Clptm1 flox neurons expressing YFP alone. We previously reported that knock-down of Clptm1 induced scaling up of synaptic GABAARs and inhibitory transmission (Ge et al., 2018). In this study, to assess whether KO of Clptm1 increases synaptic GABAARs, we immunostained for surface GABAAR γ2 and for VGAT, an inhibitory presynaptic marker. Consistent with our previous findings, KO of Clptm1 increased the intensity but not the number of postsynaptic surface clusters of GABAAR γ2 (Fig. 2B). We then tested whether inhibitory synaptic transmission was affected in Clptm1 KO neurons by measuring mIPSCs. The amplitude of mIPSCs was significantly increased in Clptm1 KO cultured hippocampal neurons relative to control Clptm1 flox neurons (Fig. 2C), consistent with the increase in synaptic GABAAR intensity. There were no changes in mIPSC frequency (Fig. 2C), consistent with no change in postsynaptic GABAAR cluster density and suggesting that KO of Clptm1 does not affect inhibitory synaptic number.
KO of Clptm1 increases inhibitory synaptic strength. A, Clptm1 flox/flox cultured hippocampal neurons were transduced with AAV-GFP or AAV-GFP-Cre at 3 DIV. Lysates were collected at 14 DIV for Western blotting. No signal for Clptm1 was detected in the AAV-GFP-Cre–transduced sample. B, C, Clptm1 flox/flox cultured hippocampal neurons were transfected with YFP or Cre + DIO-GFP at 0 DIV. mIPSC recordings and immunostaining were performed at 14 DIV. B, Neurons were immunostained live using anti-GABAAR γ2 antibody, followed by fixation and immunostaining for VGAT and the dendritic marker MAP2. KO of Clptm1 significantly increased synaptic surface γ2 intensity without changing postsynaptic γ2 puncta number. Scale bar represents 10 µm. n = 31 cells each from two independent experiments; for puncta intensity, ****p < 0.0001, t test; for puncta number, p = 0.0592, t test. C, KO of Clptm1 significantly increased mIPSC amplitude without changing mIPSC frequency. n = 19 cells for YFP and 20 cells for Cre + DIO-GFP from two independent experiments; for cumulative probability of interevent interval, p = 0.5819, K-S test; for cumulative probability of amplitude, p < 0.0001, K-S test; for frequency, p = 0.7584, Mann–Whitney test; for amplitude, *p = 0.0448, t test. D, Clptm1flox/flox mice were injected with AAV-GFP or AAV-Cre + AAV-DIO-GFP at P0. mIPSC recordings were performed from GFP-positive CA1 pyramidal neurons at P15–19. KO of Clptm1 increased mIPSC amplitude without changing mIPSC frequency. AAV-GFP, n = 24 cells from three mice; AAV-Cre + AAV-DIO-GFP, n = 18 cells from four mice; for cumulative probability of interevent interval, p = 0.8900, K-S test; for cumulative probability of amplitude, p < 0.0001, K-S test; for frequency, p = 0.9476, t test; for amplitude, ****p < 0.0001, Mann–Whitney test. E, Quantification of averaged mIPSC events from cultured hippocampal neurons in C showed no significant changes in decay kinetics. p = 0.4842, t test. F, Quantification of averaged mIPSC events from acute hippocampal slices in D showed no significant changes in decay kinetics. p = 0.1313, t test.
To determine whether KO of Clptm1 regulates inhibitory synaptic transmission in vivo, we injected AAV vectors expressing Cre recombinase bilaterally into the brain ventricles of Clptm1 homozygous flox newborn mice. mIPSC recordings were performed on hippocampal CA1 pyramidal neurons in acute brain slices at P15–19. For sparse KO to prevent potential lethality and to isolate cell-autonomous postsynaptic effects, AAV-expressing Cre driven by the CMV promoter was used, along with AAV-expressing DIO-GFP driven by the synapsin promoter to mark the transduced neurons. Consistent with the elevation of mIPSC amplitude in cultured hippocampal neurons, mIPSC amplitude was significantly increased in Cltpm1 KO CA1 pyramidal neurons expressing Cre and DIO-GFP compared with Clptm1 flox neurons expressing GFP control (Fig. 2D). Analysis of mIPSCs recorded from hippocampal cultured neurons and acute brain slices showed no significant differences in decay time between Clptm1 KO and WT neurons (Fig. 2E,F), which is consistent with our previous findings that Clptm1 interacts with all GABAAR subunits tested (Ge et al., 2018). Altogether, these results suggest that endogenous Clptm1 is required to sequester GABAARs intracellularly, maintaining inhibitory postsynaptic strength at submaximal level; loss of Clptm1 expression increased GABAAR surface expression, causing elevated inhibitory postsynaptic transmission.
Inhibitory, but not excitatory, synaptic transmission is enhanced in the juvenile Clptm1 heterozygous KO mouse hippocampus
Many mouse mutants which are lethal in homozygous form show phenotypes in heterozygous form due to haploinsufficiency (White et al., 2013). To determine if synapse function is altered as a result of Clptm1 haploinsufficiency, we performed whole-cell recordings from CA1 pyramidal neurons in brain slices from Clptm1 KOfirst/+ and WT littermate mice at P15–18. To examine sex differences, we included both male and female mice in mIPSC recordings. mIPSC amplitude was significantly increased in both male and female KOfirst/+ mouse CA1 pyramidal neurons with no changes in mIPSC frequency or decay time (Fig. 3A,B). We previously found that Clptm1 knock-down altered tonic as well as phasic inhibitory currents (Ge et al., 2018). To determine whether GABAAR-mediated tonic currents were altered in CA1 pyramidal neurons in brain slices from Clptm1 KOfirst/+ compared with WT littermate mice, we recorded the change in baseline current upon application of bicuculline. Heterozygous KO of Clptm1 significantly increased the tonic GABAergic currents in both male and female mice (Fig. 3C). Thus, Clptm1 haploinsufficiency caused increased GABAergic phasic and tonic inhibition. As there were no sex differences between male and female mice, we used only male mice for subsequent assays. To test evoked transmission, we generated input/output curves by performing whole-cell recording of eIPSCs while stimulating the stratum radiatum. An enhanced input/output relationship was observed from CA1 pyramidal neurons in brain slices from Clptm1 KOfirst/+ mice compared with WT littermates (Fig. 3D), whereas no differences were observed in paired-pulse facilitation (Fig. 3E), suggesting input/output deficits were not due to changes in neurotransmitter release probability. These results indicate a function of Clptm1 in suppressing inhibitory neurotransmission and are consistent with a postsynaptic mechanism.
Heterozygous Clptm1 KO increases phasic and tonic inhibition in juvenile mice. A, mIPSC recordings were collected from CA1 pyramidal neurons at P15–18. Both male and female heterozygous Clptm1 KOfirst mice had increased mIPSC amplitude without changing mIPSC frequency. Male WT, n = 17 cells from two mice; male KOfirst/+, n = 17 cells from two mice; female WT, n = 16 cells from two mice; female KOfirst/+, n = 16 cells from three mice; for cumulative probability of interevent interval, male p = 0.5621; female p = 0.8013, K-S test; for cumulative probability of amplitude, male p < 0.0001; female p < 0.0001, K-S test; for frequency, sex p = 0.0122; genotype p = 0.9585; sex and genotype interaction p = 0.5792, two-way ANOVA; for amplitude, sex p = 0.0165; genotype p < 0.0001; sex and genotype interaction p = 0.8885; two-way ANOVA and **p < 0.01 post hoc Tukey's test. B, Quantification of averaged mIPSC events showed no significant changes in decay kinetics. Sex p = 0.1833; genotype p = 0.1044; sex and genotype interaction p = 0.0512, two-way ANOVA. C, Recordings were collected from CA1 pyramidal neurons at P15–18. Bicuculline (20 µM) was added into the recording solution to record tonic currents. Both male and female heterozygous Clptm1 KOfirst mice had increased GABAAR-mediated tonic currents. Male WT, n = 13 cells from three mice; male KOfirst/+, n = 11 cells from two mice; female WT, n = 16 cells from four mice; female KOfirst/+, n = 16 cells from four mice; sex p = 0.5698; genotype p < 0.0001; sex and genotype interaction p = 0.2267, two-way ANOVA and **p < 0.01; ***p < 0.001 post hoc Tukey's test. D, eIPSCs were recorded from CA1 pyramidal cells while stimulating the stratum radiatum with increasing multiples of threshold intensity at P16–18. eIPSCs were significantly elevated at stimulation intensities greater than four times threshold in heterozygous KOfirst mice. WT, n = 10 cells from three mice; KOfirst/+, n = 12 cells from two mice; stimulation intensity p < 0.0001; genotype p < 0.0001; stimulation intensity and genotype interaction p < 0.001; two-way ANOVA and *p < 0.05; ***p < 0.001 post hoc Sidak's test. E, No differences were detected in the paired-pulse ratio of eIPSCs recorded from CA1 pyramidal cells with multiple interpulse intervals at P16–18. WT, n = 10 cells from three mice; KOfirst/+, n = 12 cells from two mice; interpulse interval p = 0.5239; genotype p = 0.4711; interpulse interval and genotype interaction p = 0.9999, two-way ANOVA.
We next examined whether excitatory synaptic transmission was changed in Clptm1 KOfirst/+ mice. In contrast to the elevation of mIPSC amplitude, no differences were detected in mEPSC frequency or amplitude (Fig. 4A). In addition, there were no changes in input/output curve (Fig. 4B) or paired-pulse facilitation (Fig. 4C) of eEPSCs recorded from CA1 pyramidal neurons in brain slices from Clptm1 KOfirst/+ and WT littermates, suggesting the function of Clptm1 is specific for the modulation of inhibitory neurotransmission. In addition to the whole-cell recording which records responses from a single neuron, we also recorded fEPSPs which measure extracellular field potentials from a population of neurons. fEPSPs were recorded with electrodes positioned in the stratum radiatum while stimulating Schaffer collateral fibers. The initial slope of the fEPSP, which is less subject than its amplitude to contamination from other sources of current flow in the slice, was quantitated relative to the amplitude of the fiber volley, which is an indication of the presynaptic action potential arriving at the recording site. Consistent with no changes in input/output curve of eEPSCs, no differences in fEPSP input/output relationship were detected in Clptm1 KOfirst/+ mice compared with WT littermates (Fig. 4D).
Heterozygous Clptm1 KO interferes with excitatory synaptic plasticity without changing basal excitatory transmission. A, mEPSC recordings were collected from CA1 pyramidal neurons at P16–17. No differences were detected in mEPSC amplitude or frequency. WT, n = 19 cells from two mice; KOfirst/+, n = 22 cells from four mice; for cumulative probability of interevent interval, p = 0.3818, K-S test; for cumulative probability of amplitude, p = 0.1644, K-S test; for frequency, p = 0.6201, t test; for amplitude, p = 0.6417, t test. B, eEPSCs were recorded from CA1 pyramidal cells while stimulating Schaffer collateral fibers with increasing multiples of threshold intensity at P15–18. No differences were detected in eEPSC input/output relationship. WT, n = 10 cells from four mice; KOfirst/+, n = 11 cells from two mice; stimulation intensity p < 0.0001; genotype p = 0.9605; stimulation intensity and genotype interaction p > 0.9999, two-way ANOVA. C, No differences were detected in paired-pulse ratio of eEPSCs recorded from CA1 pyramidal cells with multiple interpulse intervals at P15–18. WT, n = 12 cells from four mice; KOfirst/+, n = 12 cells from two mice; interpulse interval p < 0.0001; genotype p = 0.6656; interpulse interval and genotype interaction p = 0.9857, two-way ANOVA. D, No differences were detected in input/output relationship of the slope of the CA1 fEPSP in response to increasing stimulation of the Schaffer collateral fibers at P15–16. WT, n = 11 slices from five mice; KOfirst/+, n = 14 slices from six mice; stimulation intensity p < 0.0001; genotype p = 0.1017; stimulation intensity and genotype interaction p = 0.9340, two-way ANOVA. E, LTP was induced by HFS (1 s, 100 Hz) of Schaffer collateral fibers after obtaining 20 min stable baseline recordings. Bar graphs of the slope of the fEPSP recorded 0–2 and 55–60 min after HFS show that heterozygous Clptm1 KOfirst mice were deficient in the induction and expression of LTP. WT, n = 5 slices from five mice; KOfirst/+, n = 5 slices from three mice; *p < 0.05, Mann–Whitney test.
Although the basal level of excitatory neurotransmission was not changed, activity-dependent synaptic plasticity of excitatory neurotransmission might be affected by Clptm1 haploinsufficiency. To determine whether synaptic plasticity was altered, we tested LTP in Schaffer collateral to CA1 synapses in hippocampal brain slices from Clptm1 KOfirst/+ and WT littermate mice at P16–19. While high-frequency stimulation (HFS, 100 Hz, 1 s) induced a long-lasting (60 min) potentiation of fEPSPs in WT mice, the same stimulation protocol failed to induce LTP in Clptm1 KOfirst/+ mice (Fig. 4E). The potentiation of fEPSPs was significantly reduced at 0–2 min as well as at 55–60 min poststimulation in Clptm1 KOfirst/+ mice (Fig. 4E), suggesting a deficit in induction of LTP. Thus, Clptm1 haploinsufficiency blocked this major form of excitatory synaptic plasticity without affecting basal excitatory synaptic transmission.
Inhibitory synaptic transmission is enhanced in the adult Clptm1 heterozygous KO mouse hippocampus
Since our findings so far were all obtained from 2- to 3-week-old mice, we next tested whether the function of Clptm1 was altered in adult mice. We first examined the expression levels of Clptm1 at different developmental stages using Western blotting. The expression levels of Clptm1 increased from P1 to P14, plateaued between P14 and P21, and declined after P21 (Fig. 5A). In contrast, the expression levels of GABAAR γ2 increased from P1 to P42 (Fig. 5A). The unique pattern of developmental changes in Clptm1 expression raised the question whether the function of Clptm1 was also changed with development. We next examined inhibitory synaptic neurotransmission by recording mIPSCs from CA1 pyramidal neurons in brain slices from adult Clptm1 KOfirst/+ and WT littermates at P42–49. Consistent with the findings from juvenile mice, the mIPSC amplitude was significantly increased in adult Clptm1 KOfirst/+ mice compared with WT littermates, with no differences in mIPSC frequency or decay time (Fig. 5B,C). In addition, GABAAR-mediated tonic currents were significantly increased in adult Clptm1 KOfirst/+ mice compared with WT littermates (Fig. 5D). These results suggest although the expression levels of Clptm1 were decreased from juvenile to adult, the function of Clptm1 in the suppression of inhibitory neurotransmission was retained.
Heterozygous Clptm1 KO increases phasic and tonic inhibition in adult mice. A, Expression levels of Clptm1 and GABAAR γ2 subunits in WT mouse brain were detected and compared at different developmental stages using Western blotting. Expression of Clptm1 increased from P1 to P14, plateaued between P14 and P21, and declined after P21. In contrast, expression of GABAAR γ2 increased from P1 to P42. n = 3 mice; Clptm1, p = 0.0034; GABAAR γ2, p < 0.0001, one-way ANOVA. B, mIPSC recordings were collected from CA1 pyramidal neurons at P42–49. Heterozygous Clptm1 KOfirst mice had increased mIPSC amplitude with no change in mIPSC frequency. WT, n = 10 cells from two mice; KOfirst/+, n = 10 cells from two mice; for cumulative probability of interevent interval, p = 0.7699, K-S test; for cumulative probability of amplitude, p < 0.0001, K-S test; for frequency, p = 0.2975, t test; for amplitude, ***p = 0.0009, t test. C, Quantification of averaged mIPSC events showed no significant change in decay kinetics. p = 0.8363, Mann–Whitney test. D, Recordings were collected from CA1 pyramidal neurons at P43–50. Bicuculline (20 µM) was added into the recording solution to record tonic currents. GABAAR-mediated tonic currents were significantly increased in Clptm1 heterozygous KOfirst mice compared with WT mice. WT, n = 8 cells from three mice; KOfirst/+, n = 9 cells from four mice; ***p = 0.0003, t test.
Clptm1 heterozygous KO mice exhibit deficits in hippocampus-dependent cognitive function which are restored by GABAAR α5 modulation
To determine whether synaptic plasticity was altered in adult Clptm1 KOfirst/+ mice, we tested LTP in Schaffer collateral to CA1 synapses in hippocampal brain slices from Clptm1 KOfirst/+ and WT littermate mice at P43–49. Consistent with the results from juvenile mice, a significant reduction in the magnitude of LTP was observed in KOfirst/+ slices relative to WT (Fig. 6A). The deficits in LTP might be related to elevated tonic inhibition. Pharmacological elevation of tonic inhibition has been reported to impair HFS-induced LTP (Cheng et al., 2006; L. J. Martin et al., 2009; Whissell et al., 2013). In the hippocampus, the majority of extrasynaptic GABAARs that mediate tonic inhibition contain the α5 subunit (Caraiscos et al., 2004; Glykys and Mody, 2006; Glykys et al., 2008). Surface immunostaining for GABAAR α5 revealed that surface levels of α5 were significantly increased in KOfirst/+ cultured hippocampal neurons compared with WT (Fig. 6B), suggesting that Clptm1 regulates tonic inhibition by limiting α5 surface expression. To test whether α5-containing GABAARs are involved in modulating LTP in Clptm1 KOfirst/+ mice, we used L-655,708, an α5-GABAAR-selective negative allosteric modulator with an affinity for α5 50–100-fold greater than for α1, α2, or α3 subunit-containing GABAARs (Quirk et al., 1996). L-655,708 (10 nM) was previously reported to block tonic currents ∼40% without affecting phasic conductance (L. J. Martin et al., 2010). While bath application of L-655,708 (10 nM) had no effect on baseline fEPSPs and the magnitude of LTP in WT slices, L-655,708 rescued the magnitude of LTP in Clptm1 KOfirst/+ slices to a WT level (Fig. 6A). These results suggest that the deficits in LTP observed in Clptm1 KOfirst/+ slices may be mainly caused by the increase in tonic rather than phasic inhibition.
Clptm1 regulation of synaptic plasticity and hippocampus-dependent contextual fear memory is modulated by GABAAR α5. A, CA1 fEPSP was recorded while stimulating the Schaffer collateral fibers at P43–49. The GABAAR α5-negative modulator L-655,708 (10 nM) or vehicle (0.1% DMSO) was added in the ACSF after obtaining 10 min stable baseline recording, and 10 min later LTP was induced by HFS (1 s, 100 Hz). Bar graphs of the slope of the fEPSP recorded 0–2 and 55–60 min after HFS showed that heterozygous Clptm1 KOfirst mice had a deficit in the expression of LTP. Adding L-655,708 in the perfusion solution had no effects in WT slices but rescued LTP in KOfirst/+ slices. WT, n = 8 slices from seven mice; KOfirst/+, n = 7 slices from four mice; WT L-655,708, n = 8 slices from seven mice; KOfirst/+ L-655,708, n = 9 slices from six mice; 0–2 min, drug p = 0.4543; genotype p = 0.1672; drug and genotype interaction p = 0.1650, two-way ANOVA; 55–60 min, drug p = 0.0377; genotype p = 0.0140; drug and genotype interaction p = 0.0474, two-way ANOVA and *p < 0.05 post hoc Tukey's test. B, Immunostaining of surface GABAAR α5 in cultured hippocampal neurons showed surface levels of GABAAR α5 were significantly increased in KOfirst/+ neurons compared with WT neurons. Scale bar represents 10 µm. n = 36 cells each from three independent experiments; ****p < 0.0001, Mann–Whitney test. C, A schematic representation showing the timing for the fear-conditioning protocol. Thirty minutes before fear conditioning, mice were intraperitoneally injected with L-655,708 (0.7 mg/kg) or vehicle (3% DMSO). The baseline activity during the 3 min preceding the conditioning procedure was recorded. Three 2 s, 0.5 mA foot shocks, separated by 60 s intervals, were used for contextual fear conditioning. The after-shock activity for 1 min immediately after the footshocks was recorded. The contextual fear memory was tested by placing mice in the same chamber 24 h later. D, There were no differences in baseline freezing between WT and Clptm1 heterozygous KOfirst mice. Furthermore, L-655,708 had no effect on the freezing response. WT male vehicle, n = 11; WT female vehicle, n = 11; WT male L-655,708, n = 10; WT female L-655,708, n = 10; KOfirst/+ male vehicle, n = 10; KOfirst/+ female vehicle, n = 10; KOfirst/+ male L-655,708, n = 10; KOfirst/+ female L-655,708, n = 9; drug p = 0.6393; genotype p = 0.6876; sex p = 0.2783; drug and genotype interaction p = 0.8549; dug and sex interaction p = 0.6480; genotype and sex interaction p = 0.0929; drug, genotype, and sex interaction p = 0.7152, three-way ANOVA. E, There were no differences between WT and Clptm1 heterozygous KOfirst mice in the freezing behavior immediately after footshocks. Furthermore, L-655,708 had no effect on the freezing response. WT male vehicle, n = 11; WT female vehicle, n = 11; WT male L-655,708, n = 10; WT female L-655,708, n = 10; KOfirst/+ male vehicle, n = 10; KOfirst/+ female vehicle, n = 10; KOfirst/+ male L-655,708, n = 10; KOfirst/+ female L-655,708, n = 9; drug p = 0.9093; genotype p = 0.6954; sex p = 0.0864; drug and genotype interaction p = 0.1042; dug and sex interaction p = 0.9799; genotype and sex interaction p = 0.8030; drug, genotype, and sex interaction p = 0.6484, three-way ANOVA. F, The contextual fear memory was impaired in vehicle-treated Clptm1 heterozygous KOfirst mice compared with WT mice. In addition, L-655,708, while having no effects on the performance of WT mice, improved the performance of Clptm1 heterozygous KO mice to the level observed in WT mice. WT male vehicle, n = 11; WT female vehicle, n = 11; WT male L-655,708, n = 10; WT female L-655,708, n = 10; KOfirst/+ male vehicle, n = 10; KOfirst/+ female vehicle, n = 10; KOfirst/+ male L-655,708, n = 10; KOfirst/+ female L-655,708, n = 9; drug p < 0.0001; genotype p = 0.0746; sex p = 0.2552; drug and genotype interaction p < 0.0001; dug and sex interaction p = 0.7823; genotype and sex interaction p = 0.7146; drug, genotype, and sex interaction p = 0.7085; three-way ANOVA and **p < 0.01; ****p < 0.0001 post hoc Sidak's test.
We anticipate that the alteration of synaptic plasticity in Schaffer collateral to CA1 synapses would be associated with changes in hippocampus-dependent learning and memory. To test this idea, we conducted contextual fear conditioning using Clptm1 KOfirst/+ mice and WT littermates. This is a task in which mice learn to associate aversive mild footshock with a specific environment, and it is dependent on the function of hippocampal CA1 neurons (Maren, 2001; Goshen et al., 2011; Mocle et al., 2024). To test the modulatory role of α5-containg GABAAR–mediated tonic inhibition, L-655,708 or vehicle control was injected intraperitoneal into mice 30 min before being placed in the fear-conditioning chamber on Day 1 (Fig. 6C). Freezing behavior was recorded and quantified during a baseline habituation period, an immediate after-shock period on Day 1 and a contextual recall period on Day 2 (Fig. 6C). No differences were observed in the freezing behavior between Clptm1 KOfirst/+ mice and WT littermates during the baseline and immediate after-shock periods (Fig. 6D,E), suggesting Clptm1 haploinsufficiency has no effects on basal motor activity and response to fearful stimuli. However, Clptm1 KOfirst/+ mice showed significant deficits in contextual fear memory 24 h later (Fig. 6F). While L-655,708 had no effects on the freezing behavior during the baseline and immediate after-shock periods for both Clptm1 KOfirst/+ and WT mice, it rescued the subsequent deficit in contextual memory in Clptm1 KOfirst/+ mice to WT level (Fig. 6D–F). Both male and female mice were included in the contextual fear memory test, and no sex differences were observed. Thus, a hippocampal LTP deficit was associated with memory impairment in Clptm1 KOfirst/+ mice, and both phenotypes were rescued by pharmacological inhibition of α5-containg GABAARs.
Discussion
In this study, we used a new Clptm1 KO mouse model to uncover the function of Clptm1 in the modulation of inhibitory neurotransmission, synaptic plasticity, and contextual fear memory. Haploinsufficiency of Clptm1 specifically upregulated GABAAR-mediated phasic inhibition and tonic inhibition without affecting excitatory synaptic transmission. Moreover, Clptm1 haploinsufficiency impaired HFS-induced LTP as well as hippocampus-dependent contextual fear conditioning. Importantly, negatively modulating GABAAR α5, an extrasynaptic subunit that mediates tonic inhibition, rescued LTP and contextual fear memory in Clptm1 heterozygous KO mice. Thus, this study furthers our understanding of the roles of GABAAR-associated proteins in controlling phasic and tonic inhibition and their modulation of excitatory synaptic plasticity and behavior.
The finding that Clptm1 homozygous KOfirst mice died at early postnatal stages suggests a critical role of Clptm1. Lethality may be due to a strong upregulation of surface GABAARs, which at the early postnatal stage would lead to excessive depolarization and potentially excitation (Ben-Ari, 2002). Considering the ubiquitous expression of Clptm1 in the brain and other tissues (Yoshiura et al., 1998), Clptm1 might also be involved in other unknown functions contributing to survival. Although the KOfirst cassette is >38 kb from the nearest neighbor gene, there is also a possibility that altered expression of neighboring genes could be involved in the lethality. In contrast, there is a high degree of confidence that all the other phenotypes reported here for Clptm1 KOfirst are due to disruption of Clptm1, for two reasons. First, the effects on inhibitory synaptic transmission were also shown in Clptm1 KO neurons generated by exon deletion using the Cre loxP system (Clptm1tm1d(KOMP)Wts, Fig. 2). Second, the deficits in excitatory synaptic plasticity and memory in Clptm1 KOfirst heterozygous mice were rescued by treatment with the GABAAR α5 modulator (Fig. 6). We did not see any evidence of a cleft lip or palate in the E18 Clptm1 homozygous KOfirst mice. Although CLPTM1 was originally identified at a breakpoint in a family with cleft lip and palate (Yoshiura et al., 1998), it was not listed among >40 significant risk genes for cleft lip and palate in recent reviews (Nasreddine et al., 2021; Robinson et al., 2024).
These data showing restorative effects of GABAAR α5 modulation upon Clptm1 KO contribute to growing evidence that tonic inhibition plays an important role in the regulation of network activity, synaptic plasticity, and cognitive function. In hippocampal pyramidal neurons, the α5 subunit is predominantly expressed at extrasynaptic regions and mediates the majority of tonic inhibition (Caraiscos et al., 2004; Glykys and Mody, 2006; Glykys et al., 2008). Pharmacological manipulations of GABAAR α5 in some ways parallel our findings. Etomidate, an anesthetic which increases tonic inhibition by promoting the function of α5-containing GABAARs, impaired HFS-induced LTP and contextual fear conditioning, and these effects were blocked by pretreatment with L-655,708 (Cheng et al., 2006; L. J. Martin et al., 2009). A recent study suggests that the effects of etomidate on fear conditioning depend on GABAAR α5 in both pyramidal neurons and interneurons (Zhu et al., 2023). Also consistent with our findings, decreasing tonic inhibition by L-655,708 treatment alone or KO of GABAAR α5 had no effects on HFS-induced LTP and contextual fear conditioning, although they could enhance other forms of plasticity and memory (Collinson et al., 2002; Atack et al., 2006; Cheng et al., 2006; L. J. Martin et al., 2009, 2010; Saab et al., 2010). Importantly, Clptm1 interacts with all GABAAR subunits tested to date (Ge et al., 2018) and elevates phasic as well as tonic inhibitory transmission (Figs. 2, 3, 5), yet negative modulation of GABAAR α5 alone was sufficient to restore the deficits in LTP and fear memory in Clptm1 heterozygous KO mice (Fig. 6). These findings emphasize the importance of GABAAR α5-mediated tonic inhibition in plasticity and cognition. Indeed, negative modulators of GABAAR α5 have shown promise in mouse models of brain disorders and are in clinical trials for treatment of cognitive impairment in Down syndrome and schizophrenia (Maramai et al., 2020; Goeldner et al., 2022). In addition, while our study supports cross talk between GABAergic and glutamatergic systems in one direction, from GABAAR α5 and tonic inhibition to excitatory synaptic plasticity, there is also cross talk in the other direction. Subtype-dependent NMDA receptor activation regulates GABAAR α5 surface expression and tonic inhibition (K. Wu et al., 2021a), and LTP triggers a partial redistribution of GABAAR α5 from extrasynaptic to synaptic sites through dephosphorylation of its extrasynaptic scaffold radixin (C. M. Davenport et al., 2021).
Clptm1 is one among several transmembrane proteins that interact with GABAARs. Like Clptm1, Shisa7 interacts with multiple GABAAR subunits but promotes their trafficking in the opposite direction, increasing synaptic and extrasynaptic levels and enhancing phasic and tonic inhibition (Han et al., 2019; K. Wu et al., 2021b). Shisa7 has multiple additional effects, including modulating gating to accelerate GABAAR channel deactivation kinetics and enhancing diazepam-induced potentiation of GABAARs (Han et al., 2019; D. Castellano et al., 2022). Through a phosphorylation-regulated interaction with GABAAR α5, Shisa7 is critical for homeostatic potentiation and sleep/wake cycle regulation of tonic inhibition (K. Wu et al., 2021b). In contrast, the transmembrane protein lipoma HMGIC fusion partner-like 4 (LH4) interacts specifically with synaptic GABAARs as an accessory subunit, forming a tripartite complex with synaptic GABAARs and neuroligin-2 (E. C. Davenport et al., 2017; Yamasaki et al., 2017; M. Wu et al., 2018). LH4 is critical for synaptic clustering of GABAARs, neuroligin-2, and gephyrin and for phasic but not tonic inhibitory transmission (E. C. Davenport et al., 2017; Yamasaki et al., 2017; M. Wu et al., 2018). LH4 KO mice die prematurely and show increased seizure susceptibility and deficits in cerebellum-related motor behaviors at P17–18 (M. Wu et al., 2018). Interestingly, KO of LH4 reduces phasic inhibitory transmission in CA1 pyramidal cells but not in hippocampal interneurons (E. C. Davenport et al., 2017), indicating a cell-type–specific role of LH4. Yet another class of transmembrane GABAAR-interacting proteins resides in the endoplasmic reticulum, like Clptm1, to regulate GABAAR proteostasis, i.e., their synthesis, folding, assembly, trafficking, and degradation (Wang et al., 2022). Perhaps this process is best understood for GABAAR α1, which is processed sequentially by BiP and calnexin, Grp94, OS-9, and Hrd1 to deliver misfolded subunits to the proteasome for degradation (Di et al., 2016).
Altogether, these studies of other GABAAR-interacting proteins raise interesting directions for future studies on Clptm1. Given the localization of Clptm1 to the endoplasmic reticulum and its ability to trap GABAARs there (Ge et al., 2018), further studies on its functions in GABAAR proteostasis may be warranted. Phosphorylation of Clptm1 has been detected in multiple proteomic studies (Bian et al., 2014; Tsai et al., 2015), raising possibilities for regulated interaction and modulation of tonic or phasic inhibition. Another interesting open question is the role of Clptm1 during different stages of development. Although the function of Clptm1 in the modulation of inhibitory neurotransmission was consistent from juvenile to adulthood, the expression of Clptm1 peaked at 1–3 weeks after birth (Fig. 5), a critical period that is important for synaptogenesis and for the shift of GABAARs from depolarizing to hyperpolarizing. In adult mice, Clptm1 haploinsufficiency increased inhibitory phasic and tonic currents without affecting excitatory transmission, resulting in an excitation/inhibition imbalance, which further caused impairment of HFS-induced LTP as well as hippocampus-dependent contextual fear conditioning. A similar process may be occurring during downregulation of Clptm1 associated with Alzheimer's disease (Gockley et al., 2021); the reduction of Clptm1 may contribute to the loss of synaptic plasticity and cognitive function. Our studies of Clptm1 using a targeted genetic model are also consistent with the beneficial role reported for Clptm1 knock-down on phasic inhibitory transmission and seizure severity in a rat model of epilepsy (X. L. Zhang et al., 2023) but further suggest caution due to expected effects on tonic inhibitory transmission, excitatory synaptic plasticity, and cognition.
Footnotes
- Received March 15, 2024.
- Revision received June 17, 2024.
- Accepted June 20, 2024.
-
We thank Xiling Zhou and Nazarine Fernandes for their excellent technical assistance and Mia Wallace of the Washington University Mouse Genetics Core for generating the mutant mouse line from the ES cells. This work was supported by Canadian Institutes of Health Research FDN-143206 (to A.M.C.).
-
The authors declare no competing financial interests.
- Correspondence should be addressed to Yuan Ge at geyuan{at}mail.ubc.ca or Ann Marie Craig at acraig{at}mail.ubc.ca.
- Copyright © 2024 the authors
