Cortical parvalbumin GABAergic deficits with α7 nicotinic acetylcholine receptor deletion: Implications for schizophrenia

Abstract

Dysfunction of cortical parvalbumin (PV)-containing GABAergic interneurons has been implicated in cognitive deficits of schizophrenia. In humans microdeletion of the CHRNA7 (α7 nicotinic acetylcholine receptor, nAChR) gene is associated with cortical dysfunction in a broad spectrum of neurodevelopmental and neuropsychiatric disorders including schizophrenia while in mice similar deletion causes analogous abnormalities including impaired attention, working-memory and learning. However, the pathophysiological roles of α7 nAChRs in cortical PV GABAergic development remain largely uncharacterized. In both in vivo and in vitro models, we identify here that deletion of the α7 nAChR gene in mice impairs cortical PV GABAergic development and recapitulates many of the characteristic neurochemical deficits in PV-positive GABAergic interneurons found in schizophrenia. α7 nAChR null mice had decreased cortical levels of GABAergic markers including PV, Glutamic Acid Decarboxylase 65/67 (GAD65/67) and the α1 subunit of GABA_A receptors, particularly reductions of PV and GAD67 levels in cortical PV-positive interneurons during late postnatal life and adulthood. Cortical GABAergic synaptic deficits were identified in the prefrontal cortex of α7 nAChR null mice and α7 nAChR null cortical cultures. Similar disruptions in development of PV-positive GABAergic interneurons and perisomatic synapses were found in cortical cultures lacking α7 nAChRs. Moreover, NMDA receptor expression was reduced in GABAergic interneurons, implicating NMDA receptor hypofunction in GABAergic deficits in α7 nAChR null mice. Our findings thus demonstrate impaired cortical PV GABAergic development and multiple characteristic neurochemical deficits reminiscent of schizophrenia in cortical PV-positive interneurons in α7 nAChR gene deletion models. This implicates crucial roles of α7 nAChRs in cortical PV GABAergic development and dysfunction in schizophrenia and other neuropsychiatric disorders.

Introduction

Nicotinic receptor mediated neurotransmission has been implicated in many CNS processes including neuronal plasticity, learning and memory as well as in neurological, neurodevelopmental and psychiatric disorders (Dani and Bertrand, 2007; Lin et al., 2010; Ross et al., 2010; Miwa et al., 2011; Griguoli and Cherubini, 2012; Wallace and Bertrand, 2013; Freedman, 2014; Lin et al., 2014). Some of the most powerful evidence comes from genetic studies in which a homozygous microdeletion of human15q13.3, including the α7 nAChR gene, is associated with cortical circuit disorders including schizophrenia, autism, mental retardation, epilepsy, and intellectual disability (Freedman et al., 1997; Sharp et al., 2008; Helbig et al., 2009; Miller et al., 2009; Shinawi et al., 2009; Endris et al., 2010; Lepichon et al., 2010; Ancin et al., 2011; Liao et al., 2011; Spielmann et al., 2011; Yasui et al., 2011; Hoppman-Chaney et al., 2012; Freedman, 2014). Based on the size and the overlap of deletions in such individuals, loss of the α7 nAChR gene alone appears sufficient to cause the majority of clinical features in such patients (Shinawi et al., 2009; Endris et al., 2010; Liao et al., 2011; Hoppman-Chaney et al., 2012). In humans reduced levels of α7 nAChR have been found in the prefrontal cortex and hippocampus of schizophrenia patients in postmortem studies (Freedman et al., 1995; Leonard et al., 1996; Freedman et al., 1997; Freedman et al., 2000; Freedman and Leonard, 2001; Freedman and Goldowitz, 2010; Ross et al., 2010). In animals loss of the α7 nAChR causes analogous abnormalities including impaired attention, working-memory and learning (Fernandes et al., 2006; Young et al., 2007; Levin et al., 2009; Brown et al., 2010; Young et al., 2011; Hellier et al., 2012). These findings suggest crucial roles of α7 nAChR in cortical circuit function and dysfunction.

Cortical parvalbumin (PV)-containing GABAergic interneurons have also been implicated in cortical circuit function and dysfunction in schizophrenia (Lewis et al., 2012; Nakazawa et al., 2012; Inan et al., 2013; Jiang et al., 2013; Le Magueresse and Monyer, 2013). A major role of GABA-mediated neurotransmission is to synchronize network oscillations essential for normal cognition (Gonzalez-Burgos et al., 2011), and impaired GABAergic neurotransmission in schizophrenia may underlie impairments in cortical gamma oscillations associated with cognitive dysfunction (McNally et al., 2013). In particular, the PV-positive subtype of GABAergic interneurons have uniquely fast membrane and synaptic properties that are crucial for cortical circuit functions such as gamma oscillations (Lewis et al., 2012). The cortical PV-positive interneurons consist of two main types: chandelier and basket cells. Cortical gamma oscillations require strong inhibitory inputs to pyramidal neurons from the PV-positive basket cells and are mediated through postsynaptic α1 subunit-containing GABA_A receptors in pyramidal neurons (Lewis et al., 2012; Inan et al., 2013). Dysfunction of cortical PV-positive basket cells has been implicated in aberrant gamma oscillations and cognitive deficits in schizophrenia due to reduced levels of PV and the 67-kDa isoform of GABA synthesizing enzyme glutamic acid decarboxylase (GAD67) as well as lower GABA_A receptor α1 subunit levels in pyramidal neurons postsynaptic to basket cells (Curley and Lewis, 2012; Lewis et al., 2012). These observations implicate cortical PV-positive basket cells and their postsynaptic pyramidal cells in cortical circuit function and dysfunction in schizophrenia.

Thus, both α7 nAChR and cortical PV GABAergic abnormalities are associated with dysfunction in schizophrenia. Apart from decreased immunoreactivities of GAD65 and GABA_A receptor binding in the hippocampus of heterozygous α7 nAChR knockout mice, the pathophysiological processes linking nicotinic receptor and cortical PV GABAergic abnormalities remain largely uncharacterized (Adams et al., 2012). In the present studies in α7 nAChR gene deletion models, we demonstrate impaired cortical PV GABAergic development and multiple characteristic neurochemical deficits that are reminiscent of cortical PV GABAergic deficits in schizophrenia. These include reduced PV and GAD67 levels in PV-positive GABAergic interneurons, reduced GABA_Aα1 levels in pyramidal neurons as well as perisomatic GABAergic synaptic deficits. Our findings thus implicate α7 nAChR in cortical PV GABAergic circuit development and dysfunction in schizophrenia and other neuropsychiatric disorders.

Materials and Methods

Materials

Timed-pregnant C57BL/6 mice were purchased from Charles River Laboratories and bred from α7 nAChR knockout mice (B6.129S7-Chrna7^tm1Bay/J, Jackson Laboratory). Biochemicals and antibodies included kynurenic acid (Sigma), α-GABA_Aα1 (Millipore, rabbit polyclonal), α-somatostatin (Millipore, rabbit polyclonal), α-parvalbumin (Millipore, mouse monoclonal; Swant, rabbit polyclonal), α-GAD65/67 (Millipore, rabbit polyclonal), α-GAD65 (Developmental Studies Hybridoma Bank, mouse monoclonal), α-VGAT (Millipore, rabbit polyclonal), α-VGLUT1 (Synaptic Systems, rabbit polyclonal), α-GAPDH (Novus Biologicals, mouse monoclonal), α-actin (Sigma, rabbit polyclonal).

Neuronal cultures

Primary cortical cultures from E17-19 wild-type or α7 nAChR knockout mice were prepared as described (Dong et al., 2004). Briefly, the cortex was dissected, gently minced, trypsinized (0.027%, 37 °C; 7% CO₂ for 20 min), and then washed with 1× HBSS. Neurons were seeded to a density of 3 × 10⁵ viable cells in 35-mm culture dish with five 12-mm glass coverslips or a density of 1.6 × 10⁶ viable cells in 60-mm culture dishes. The culture dishes were coated with poly-D-Lysine (100 μg/ml) prior to seeding neurons. Neurons were maintained at 37°C with 5% CO₂ in Neurobasal medium with B27 supplement. Neurobasal medium contains choline chloride, a selective agonist at α7 nAChR (Zhang and Warren, 2002). In addition, cholinergic neurons are present in cortical cultures as identified by ChAT immunostaining as described by Abcam in their manufacturer’s instructions (Lin et al., 2010). At 21–28 days in vitro (DIV), cultures were subject to western blotting analysis, immunocytochemistry or patch clamp recording.

For cell lysate preparation, cultures were lysed in lysis buffer (150 mM NaCl, 1 mM EDTA, 100 mM Tris-HCl, 1% Triton X-100, 1% sodium deoxycholate and 1% SDS, pH 7.4, supplemented the day of use with 1:500 EDTA-free protease inhibitor cocktail III (Calbiochem) for 1 hr at 4°C and collected. Debris was cleared by centrifugation at 16,100 × g for 20 mins at 4°C. Supernatants were stored at −80°C until use.

Tissue preparation

For tissue homogenate preparation, the age-matched wild-type (WT) and α7 nAChR null (α7-KO) mouse littermates at postnatal day (P1-P56) and 9 months older of either sex were anesthetized with isoflurane before decapitation in accordance with protocols approved by The Children’s Hospital of Philadelphia Animal Care and Use Committee. The mouse brain was rapidly removed, and the cortex was dissected under Leica EZ4 stereomicroscope and immediately transferred to dry ice. The cortex was homogenized in 20ml lysis buffer per 1 g weight, and lysed for 1 hr at 4°C. Lysis buffer contained 150 mM NaCl, 1 mM EDTA, 100 mM Tris-HCl, 1% Triton X-100, and 1% sodium deoxycholate, pH 7.4, supplemented the day of use with 1:500 EDTA-free protease inhibitor cocktail III (Calbiochem). Debris was cleared by centrifugation at 39,000 × g for 1 hr at 4°C. Supernatants were stored at −80°C until use.

For immunohistochemical studies, the age-matched male WT and α7-KO mice at P56-P90 were anesthetized with isoflurane and ketamine/xylazine mixture, and cardiac perfused with 10 ml of PBS, followed by 20 ml of PBS containing 4% paraformaldehyde in accordance with protocols approved by The Children’s Hospital of Philadelphia Animal Care and Use Committee. Brains were excised and immersed overnight in 4% paraformaldehyde, washed in PBS, dehydrated and embedded in paraffin. A series of brain coronal sections (5 µm thick) were obtained using microtone at The Children’s Hospital of Philadelphia Pathology Core Facility.

Western blotting analysis

Western blotting was performed as described previously (Dong et al., 2004). Protein content was determined using BCA Protein Assay (Thermo Scientific). Equal amounts of total protein (5µg cell lysate or 30µg cortex homogenate per lane) was subjected to 4–12% NuPAGE Gel for electrophoresis and transferred to nitrocellulose membranes. Membranes were blocked with 3% nonfat milk and incubated with primary antibody overnight at 4 °C. Blots were then incubated with appropriate horseradish peroxidase, HRP-conjugated secondary antibodies (Cell Signaling) for 2 hr at room temperature and then washed; Reaction bands were visualized using a luminol-enhanced chemiluminescence (ECL) HPR substrate (Thermo Scientific). Each blot was then incubated with stripping buffer (2% SDS, 50 mM Tris, pH 6.8, and 100 mM β-mercaptoethanol) for 45 mins at room temperature to remove the signals and reprobed for other proteins including actin and GAPDH as internal controls. Reaction product levels were quantified by scanning densitometry and normalized by the levels of actin or GAPDH using NIH Image J software. Since the cortical actin and GAPDH levels alter during the time course of postnatal development, the protein levels were normalized by internal control at the same time-point between WT and KO mice.

Immunocytochemistry and immunohistochemistry

Primary cultured cortical neurons were fixed for 20 min at 4°C with 4% paraformaldehyde in phosphate-buffered saline (PBS) (pH 7.4), and then subject to immunostaining procedure. The paraffin-embedded brain sections were deparaffined, rehydrated, and antigen-retrieved in antigen unmasking solution (Vector Laboratories) and then subject to immunostaining procedure. For immunostaining procedure, after blocking with 5% normal goat serum and 1 % bovine serum albumin in combination with 0.3% (vol/vol) Triton X-100 in PBS at room temperature for 1 hr, the coverslips or slides were incubated with primary antibodies at 4°C overnight and then secondary antibodies conjugated to Alexa fluor 488 or 594 (Invitrogen) at room temperature for 60–90 mins. Following several washes with PBS, cells or slides were mounted with Vectashield with DAPI (Vector Laboratories).

Fluorescence imaging

Fluorescence images were obtained with Olympus FluoView laser scanning confocal microscope. For mouse cortex, the brain sections were stained for GABAergic markers (PV, SST, GAD65/67, GABA_Aα1). The infralimbic, prelimbic, or anterior cingulate regions of the mouse medial frontal cortex are collectively referred to as mouse prefrontal cortex (Rotaru et al., 2011). Confocal scans were performed in layer 2–6 of mouse prefrontal cortex, and the imaging parameters were identical for wild-type and α7 nAChR knockout mice. Wild-type controls were always included in all experiments to normalize for expected variations in antibody staining intensity performed on different days. The confocal images were acquired at the focal plane with maximal number of synapses from at least 3 sections per animal and from at least 3 animals per group. NIH Image J software was used to quantify VGAT- or GAD65-positive GABAergic synapses in the acquired confocal images. Thresholds were set at 3 SDs above the mean staining intensity of six nearby regions in the same visual field. Thresholded images present a fixed intensity for all pixels above threshold after having removed all of those below and VGAT- or GAD65-labeled puncta in the thresholded images were quantified (Lozada et al., 2012). For quantification of the immunofluorescence intensity of PV and SST in the soma of PV- or SST-positive interneurons, the acquired confocal images were quantified using NIH image J software in the thresholded images from 6–8 sections per animal and from at least 3 animals per group. For quantification of the number of the PV-positive GABAergic interneurons in mouse prefrontal cortex, a blind observer counted the PV-positive interneurons in mouse prefrontal cortex under the epifluorescence microscopy from 7–8 sections per animal and from at least 3 animals per group. For cortical cultures, neurons were stained for GABAergic markers (VGAT, GAD65, GABA_Aα1 and PV). For quantification analysis, the confocal images were acquired under 63× objectives from 4–6 samples and 4–5 different cultures for quantification of VGAT- or PV-labeled puncta in wild-type and α7 nAChR knockout cortical cultures. NIH Image J software and the thresholded images were used to quantify the number of GABAergic synapses in cortical cultures.

Whole-cell patch-clamp recording

Cortical pyramidal neurons were identified based on their morphological properties and firing pattern. Specifically, besides the typical morphology of cultured cortical pyramidal neurons, pyramidal neurons were also confirmed by their action potential firing patterns in respond to intracellular injection of depolarized current using whole cell patch-clamping techniques (Rotaru et al., 2011). Since spontaneous IPSC recording was recorded using a CsCl-based intrapipette solution and resting membrane potential of recorded cells would be affected under current clamp mode of recording, an initial screening recording was conducted to reduce the data contamination by interneurons using potassium gluconate-based intrapipette solution. The recording electrodes were filled with potassium gluconate-based intrapipette solution (in mM: 145 K-gluconate, 2 MgCl₂, 2 ATP·Mg, 0.5 GTP·Tris, 0.1 BAPTA, 2.5 KCl, 2.5 NaCl, and 10 HEPES), and the recording chamber was perfused with artificial CSF (aCSF) with composition (in mM): 130 NaCl, 3 KCl, 2 CaCl₂, 1 MgCl₂, 1.25 NaH₂PO₄, 26 NaHCO₃ and 10 glucose. Cultured cortical neurons were recorded under current clamp mode and action potentials were induced by intracellular injection of depolarized current. Matched with the morphology, 100% of selected cells (n = 15) for an initial screening recording were cortical pyramidal neurons judged by their action potential firing patterns (Rotaru et al., 2011). Also, an example of cortical pyramidal neuron responding to intracellular injection (700 msec) of depolarized current is shown in Supplemental Figure 1.

Spontaneous inhibitory postsynaptic current (sIPSC) recording was performed using standard whole-cell voltage clamp techniques in cultured mouse cortical neurons grown on coverslips for 21–28 DIV. Cells were transferred to a recording chamber at room temperature (23 ± 2°C) containing an extracellular solution composed of (mM): 155 NaCl, 3 KCl, 3 CaCl₂, 1 MgCl₂ 10 HEPES, pH 7.4. Recording electrodes were pulled on a Flaming-Brown micropipette puller (Model P-97, Sutter Instrument Co.) and filled with a solution containing (mM): 130 CsCl, 2 MgCl₂, 10 HEPES, 0.8 CsOH, 2 Mg-ATP, pH 7.35, 280–300 mOsm. In order to isolate the inhibitory synaptic neuronal events, the glutamate receptor-mediated sEPSCs were blocked by bath application of glutamate receptor antagonist kynurenic acid (2 mM) in extracellular solution. The holding potential for sIPSC recording was −70 mV. Recorded signals were amplified using an Axopatch-1D amplifier, filtered at 5 kHz and then saved as data files using pCLAMP 9.01 software (Axon Instrument Inc., Hamden, CT) for off-line analysis. The synaptic events were analyzed using Mini-Analysis software to calculate the sIPSC frequency, amplitude as well as current decay time constant.

Statistical analysis

Data were shown as the mean ± S.E.M. Experiments were analyzed using Mann-Whitney Rank Sum test or Student’s t-test to compare two conditions or ANOVA followed by planned comparisons of multiple conditions. Significance was set at P < 0.05.

Results

Cortical levels of GABAergic markers were decreased in α7 nAChR null mice during late postnatal development and adulthood

Our previous studies have demonstrated that α7 nAChR gene deletion leads to specific loss of synaptic NMDA receptors and their coagonist, D-serine, as well as glutamatergic synaptic deficits in mouse cortex (Lin et al., 2014). In the present study, we explored whether α7 nAChR gene deletion affects GABAergic neuronal development by examining the expression of parvalbumin (PV), the α1 subunit of GABA_A receptor (GABA_Aα1), and two isoforms of the GABA synthesizing enzyme, glutamic acid decarboxylase 65/67 (GAD65/67) as markers of GABAergic interneurons in the cortex of α7 nAChR null (α7-KO) mice and wild-type (WT) littermates at P1-P56 and P9M. In WT mice, expression of PV remained low until P21, peaked at P56, and remained high in adulthood (Fig. 1A). Levels of GAD65 remained low until P14, increased by P21 and reached its maximum at P56 (Fig. 1B), while the levels of GABA_Aα₁ remained low until P8, began to increase by P14, peaked at P56 and remained high in adulthood (Fig. 1C). The results suggest that the neurochemical maturation of mouse cortical GABAergic circuitry occurs by P56, shortly after sexual maturation, similar to development of GABAergic circuitry in monkeys (Hoftman and Lewis, 2011).

Figure 1. Cortical levels of GABAergic markers were decreased in α7 nAChR null mice during late postnatal development and adulthood.

Western blotting analysis of whole cortex homogenates (30µg per lane) showing PV (A), GAD65/67 (B), GABA_A α1 (C) expression in the cortex of wild-type (WT) and α7 nAChR null (KO) mice at postnatal days P1, P4, P8, P14, P21, P56 and P9M (>9 months old) (*p<0.05 **p<0.01 KO vs. WT; n=6; t-test).

Remarkably, in the neocortex of α7-KO mice expression levels of these markers of GABAergic maturation were significantly reduced. PV was modestly but significantly reduced at P21, an effect that persisted into adulthood (Fig. 1A). For both GAD65 and GABA_Aα₁, there was a trend for a decrease during the second postnatal week and in adults, with a statistically significant decrease present at P21 and P56 (Fig. 1B and 1C). Similar to GAD65, GAD67 expression levels were also decreased in the cortex of α7-KO mice compared with WT littermates (Fig. 1B). These findings suggest that α7 nAChR deletion alters cortical GABA synthesis, PV regulation, and GABA_Aα₁ subunit-containing receptor presence during late postnatal development.

Neurochemical deficits of cortical PV-positive GABAergic interneurons in the prefrontal cortex of α7 nAChR null mice

We then examined the effects of α7 nAChR gene deletion on cortical PV-positive GABAergic interneurons in mouse prefrontal cortex by examining PV, GAD65/67 and GABA_Aα1 immunoreactivities microscopically at P56-P90. In the prefrontal cortex of WT mice, the PV-positive GABAergic interneurons were scattered in layers 2–6 (Fig. 2A, 2G). GAD67 was localized in the soma of PV-positive GABAergic interneurons, and GAD65 in the GABAergic synapses. GAD65-positive GABAergic synapses were widely and extensively distributed in the neuropil surrounding the PV-positive GABAergic neurons and pyramidal neurons (Fig. 2B). The GABA_Aα₁-containing receptors were found in both PV-positive GABAergic neurons and pyramidal neurons (Fig. 2H). In α7-KO mice, the overall levels of PV, GAD65/67 and GABA_Aα₁ immunoreactivities (Fig. 2D–2F and 2J–2L) were decreased compared with WT littermates (Fig. 2A–2C and 2G–2I), demonstrating that deletion of the α7 nAChR gene reduced PV, GAD65/67 and GABA_Aα₁ expression during late postnatal development. We further examined the specificity of PV alteration in α7-KO mice by examining somatostatin (SST) immunoreactivity, a marker of a different subclass of GABAergic interneurons, in the mouse cortex. In the prefrontal cortex of WT mice, the SST-positive GABAergic interneurons were scattered in all layers (Fig. 3B), while the PV-positive GABAergic interneurons were scattered in layers 2–6 (Fig. 3A). The overall level of somatostatin was not altered in the prefrontal cortex of α7-KO mice (Fig. 3D–3F and 3J–3L), whereas that of PV was markedly reduced compared with WT littermates (Fig. 3A–3C and 3G–3I). Quantification analysis showed significant reduction of PV levels in PV-positive GABAergic interneurons (Fig. 3M) and similar levels of SST in SST-positive GABAergic interneurons (Fig. 3N) in the prefrontal cortex of α7-KO mice compared with WT mice. Although PV levels were decreased in the prefrontal cortex, the number of PV-positive GABAergic neurons in the prefrontal cortex of α7-KO mice was not significantly decreased compared with that of WT littermates (Fig. 3O). Our findings suggest that deletion of the α7 nAChR gene reduces cortical PV levels but not PV interneuron number, reflecting a reduction of PV synthesis rather than a failure of PV-positive neurons to develop.

Figure 2. The levels of PV, GAD65/67 and GABA_Aα1 were decreased in the prefrontal cortex of α7 nAChR null mice during late postnatal development and adulthood.

A–F. Confocal images of PV (green), GAD65/67 (red) immunofluorescence and merged images with DAPI-stained nucleus showing a reduction in the overall level of PV and GAD65/67 immunoreactivities in the prefrontal cortex of α7 nAChR null (α7-KO) mice (D–F) compared with WT littermates at P56-P90 (A–C). G–L. Confocal images of PV (green), GABA_Aα1 (red) immunofluorescence and merged images with DAPI-stained nucleus showing a reduction in the overall level of PV and GABA_Aα1 immunoreactivities in the prefrontal cortex of α7-KO mice (J–L) compared with WT littermates at P56-P90. Scale bars as indicated.

Figure 3. The levels of PV, but not somatostatin, immunoreactivities were decreased in the prefrontal cortex of α7 nAChR null mice during late postnatal development and adulthood.

A–L. Confocal images of PV (green), somatostatin (red) immunofluorescence and merged images with DAPI-stained nucleus showing a reduction in the overall level of PV, but not somatostatin, immunoreactivities in the prefrontal cortex of α7-KO mice (D–F, J–L) compared with WT littermates (A–C, G–I) at P56-P90. M. Quantification analysis showing reduced PV levels in the soma of PV-positive GABAergic interneurons of α7-KO mouse prefrontal cortex compared with WT littermates.(***p<0.001 KO vs. WT; 6–8 sections per animal, 3 animals each group; t-test). N. Quantification analysis showing similar SST levels in the soma of SST-positive GABAergic interneurons of α7-KO mouse prefrontal cortex compared with WT littermates.(p>0.05 KO vs. WT; 6–8 sections per animal, 3 animals each group; t-test). O. Quantification analysis showing that the number of PV-positive GABAergic interneurons in the prefrontal cortex was not significantly reduced in α7-KO mice compared with WT littermates at P56-P90 (p>0.05; 7–8 sections per animal, 3 animals per group; t-test). Scale bars as indicated.

Moreover, PV and GAD67 levels were reduced in the soma of PV-positive GABAergic interneurons(Fig. 4D–4F, white arrows) as were the number of GAD65-positive GABAergic synapses surrounding the pyramidal neurons (Fig. 4D–4F grey arrows) in the prefrontal cortex of α7-KO mice compared with WT littermates (Fig. 4A–4C). The GABA_Aα₁-containing receptors were localized in both PV-positive GABAergic interneurons (Fig. 4G–4I, white arrows) and pyramidal neurons (Fig. 4G–4I, grey arrows), with higher levels in PV-positive GABAergic interneurons in the prefrontal cortex of WT mice. GABA_Aα1 levels were reduced in both PV-positive GABAergic interneurons (Fig. 4J–4L, white arrows) and pyramidal neurons (Fig. 4J–4L, grey arrows) in the prefrontal cortex of α7-KO mice compared with WT littermates (Fig. 4G–4I). Our findings thus demonstrated multiple characteristic neurochemical deficits reminiscent of schizophrenia in PV-positive GABAergic interneurons in mouse prefrontal cortex with α7 nAChR deletion.

Figure 4. PV and GAD67 levels in PV-positive GABAergic interneurons as well as GABA_Aα1 levels in pyramidal neurons were reduced in mouse prefrontal cortex with α7 nAChR deletion during late postnatal development and adulthood.

A–F. Confocal images of PV (green), GAD65/67 (red) immunofluorescence and merged images with DAPI-stained nucleus showing reduced PV and GAD67 levels in PV-positive GABAergic neurons (white arrows) and reduced GAD65-positive GABAergic synapses surrounding pyramidal neurons (grey arrows) in the prefrontal cortex of α7-KO mice (D–F) compared with WT littermates at P56-P90 (A–C). G–L. Confocal images of PV (green), GABA_Aα1 (red) immunofluorescence and merged images with DAPI-stained nucleus showing reduced GABA_Aα1 levels in pyramidal neurons (J–L, grey arrows) and in PV-positive GABAergic interneurons (J–L, white arrows) in the prefrontal cortex of α7-KO mice compared with WT littermates at P56-P90 (G–I). Scale bars as indicated.

Cortical GABAergic synapse formation was impaired in prefrontal cortex and in cortical cultures of α7 nAChR null mice

We then examined whether α7 nAChR gene deletion affects GABAergic synapse formation by examining the immunoreactivity of a GABAergic specific presynaptic marker, the vesicular GABA transporter (VGAT), in the prefrontal cortex. In the prefrontal cortex of WT mice, similar to GAD65-positive GABAergic presynaptic terminals, VGAT-positive GABAergic presynaptic terminals were extensively distributed in the neuropil surrounding the PV-positive GABAergic interneurons and pyramidal neurons (Fig. 5A–5C). In α7-KO mice, the overall levels of VGAT immunoreactivity (Fig. 5D–5F) were reduced compared with WT littermates (Fig. 5A–5C), and the number of VGAT-positive GABAergic presynaptic terminals surrounding the pyramidal neurons (Fig. 5J–5L, grey arrows) were reduced in the prefrontal cortex of α7-KO mice compared with WT littermates (Fig. 5G–5I, grey arrows). Quantification of VGAT-positive GABAergic presynaptic terminals identified a significant decrease in the prefrontal cortex of α7-KO mice compared with WT littermates (Fig. 5M). Thus, deletion of the α7 nAChR decreased cortical GABAergic synapse formation in mouse prefrontal cortex.

Figure 5. GABAergic synapse formation was impaired in the prefrontal cortex of α7 nAChR null mice during late postnatal development and adulthood.

A–F. Confocal images of PV (green), VGAT (red) immunofluorescence and merged images with DAPI-stained nucleus showing a reduction in the overall levels of PV and VGAT in the prefrontal cortex of α7-KO mice (D–F) compared with WT littermates at P56-P90 (A–C). G–L. The PV levels in PV-positive GABAergic neurons (white arrows) and VGAT-positive GABAergic presynaptic terminals surrounding pyramidal neurons (grey arrows) were reduced in the prefrontal cortex of α7-KO mice (J–L) compared with WT littermates at P56-P90 (G–I). M. Quantification of the number of VGAT-positive terminals per 1000 µm² showing a significant decrease in the number of GABAergic presynaptic terminals in the prefrontal cortex of α7 KO mice compared with WT littermates at P56-P90 (***p<0.001 KO vs. WT; 4–5 sections per animal, 3 animals per group; t-test). Scale bars as indicated.

We further examined the effects of α7 nAChR gene deletion on GABAergic synaptic development in cortical cultures. Double immunostaining with GABAergic presynaptic marker GAD65 and postsynaptic marker GABA_Aα1 antibodies identified GAD65- and GABA_Aα1-positive GABAergic synapses and demonstrated a reduction in the number of GABAergic synapses and level of GABA_Aα1 in α7-KO cultures (Fig. 6D–6F) compared with WT cultures (Fig. 6A–6C). However, high magnification of confocal images showed that 100% of GAD65-positive puncta were colocalized with or adjacent to one or more GABAAα1-positive puncta in the dendrites of cortical pyramidal neurons as shown in Fig.6A–6F, the number of GABAergic specific presynaptic marker VGAT-positive puncta thus accurately reflects the number of GABAergic synapses on the dendrites of pyramidal neurons. Immunostaining with VGAT antibody identified specific presynaptic terminals of GABAergic synapses in WT (Fig. 6G) and α7-KO (Fig. 6H) cultures and demonstrated a significant reduction in the number of GABAergic synapses on the dendrites of pyramidal neurons in α7-KO cortical cultures compared with WT cultures based on quantification (Fig. 6I). Whole-cell voltage clamp recording showed significant reduction in sIPSC frequency and decay time (Fig. 6J–6L), but not amplitude (Fig. 6M), in cultured cortical pyramidal neurons, consistent with the reduction of GABAergic synapses in α7-KO cortical cultures. The findings thus demonstrate that deletion of the α7 nAChR gene in mice severely impairs cortical GABAergic synapse formation in vivo and in vitro.

Figure 6. GABAergic synapse formation and neurotransmission were reduced in α7-KO cortical cultures.

A–F. Confocal images of GAD65 (green), GABAAα1 (red) immunofluorescence and merged images showing reduced number of GABAergic synapses and levels of GABAAα1 in α7-KO cortical cultures (D–F) compared with WT cultures (A–C). High magnification of confocal images showing that 100% of GAD65-positive puncta (A, D) were colocalized with or adjacent to one or more GABAAα1 puncta (B, E) on the dendrites of cortical neurons in WT and α7-KO cultures (C, F). G, H. Confocal images of VGAT (red) immunofluorescence showing reduced GABAergic synapses in α7-KO cortical cultures (H) compared with WT cultures (G). I. Quantification of the number of VGAT-positive puncta per 10µm length dendrite of cortical neurons showing a significant decrease in the number of GABAergic synapses on the dendrites of pyramidal neurons in α7 KO cultures compared with WT cultures. (***p<0.001 KO vs. WT; 5–6 neurons per culture, 4 different cultures each group; t-test). J–M. Spontaneous IPSC (sIPSC) properties for WT and α7-KO cultured cortical pyramidal neurons (21–28 DIV). Cells were voltage-clamped at −70mV and whole cell recording of sIPSCs was performed. J. Representative traces for sIPSC for both WT and α7-KO cultured cortical pyramidal neurons. sIPSC analysis demonstrated that the sIPSC frequency was decreased (K), current decay time constant was faster (L), but no change in amplitude (M) in α7-KO cultures compared with those in WT cultures. Box-Whisker graphs were plotted due to the non-parametric data distribution. The data were from 4–5 independent experiments. (**p<0.01 and *** p<0.001 vs. WT; n=25; Mann-Whitney Rank Sum test). Scale bars as indicated.

The development and maturation of cortical PV-positive basket cells were impaired in α7 nAChR null cortical cultures

We further explored whether deletion of the α7 nAChR impaired the development of cortical PV-positive GABAergic interneurons by examining PV and GAD65 immunoreactivities in cortical cultures at 21–28 DIV. In WT cortical cultures, PV immunoreactivity was widely distributed in the soma, dendrites, axons and synapses of cortical PV-positive GABAergic interneurons (Fig. 7A–7C and 7G–7I). The axons of an individual PV-positive basket cells formed perisomatic GABAergic synapses surrounding the multiple pyramidal neurons. The PV-positive perisomatic GABAergic synapses were positive for GAD65 immunoreactivity (Fig. 7A–7C and 7G–7I, white arrows). In α7-KO cultures, the development of PV-positive basket cells and the formation PV- and GAD65-positive perisomatic GABAergic synapses were impaired (Fig. 7D–7F and 7J–7L, white arrows) compared with WT cultures (Fig. 7A–7C and 7G–7I, white arrows). The overall PV immunoreactivity in the soma of PV-positive GABAergic interneurons was significantly decreased in α7-KO cultures compared with WT cultures (Fig. 7M–7O), matching the PV reduction in prefrontal cortex of α7-KO mice.

Figure 7. The development and maturation of PV-positive basket cells were impaired in α7-KO cortical cultures.

A–L. Confocal images of PV (red) and GAD65 (green) immunofluorescence and merged images showing PV immunoreactivities in the soma, dendrites, axons and synapses of PV-positive basket cells in WT and α7-KO cultures (A, D, G and J). The axons of an individual PV-positive basket cell formed perisomatic GABAergic synapses surrounding the multiple pyramidal neurons (A–L, white arrows). The development of PV-positive basket cells and perisomatic GABAergic synapses was impaired in α7-KO cortical cultures (D–F and J–L) compared with WT cultures (A–C and G–I). M, N. Confocal images of PV (red) immunofluorescence showing PV immunoreactivity in PV-positive GABAergic interneurons in WT (M) and α7-KO (N) cortical cultures. O. Quantification analysis showing reduced PV levels in the soma of PV-positive GABAergic interneurons in α7-KO cultures compared with WT cultures. (***p<0.001 KO vs. WT; 5–6 neurons per culture, 4 different cultures; t-test). Scale bars as indicated.

We further examined the maturation of PV-positive basket cells by examining the formation of PV-positive perisomatic GABAergic synapses surrounding the pyramidal neurons in cortical cultures. In WT cultures, the PV- and GAD65-positive perisomatic synapses formed well surrounding the soma and proximal dendrites of pyramidal neurons (Fig. 8A–8C and 8G–8I). In α7-KO cultures, the formation of PV- and GAD65-positive perisomatic synapses was markedly impaired (Fig. 8D–8F and 8J–8L), and there was a significant reduction of the number of PV-positive perisomatic synapses in α7-KO cultures compared with WT cultures (Fig. 8M).The findings thus suggest that deletion of the α7 nAChR impaired the development and the perisomatic synapse formation of cortical PV-positive GABAergic interneurons in cortical cultures.

Figure 8. The PV-positive perisomatic GABAergic synapses were reduced in α7-KO cortical cultures.

A–L. Confocal images of PV (red) and GAD65 (green) immunofluorescence and merged images showing the reduced PV- and GAD65-positive perisomatic synapses of PV-positive basket cells surrounding the soma and proximal dendrites of pyramidal neurons in α7-KO cultures (D–F and J–L) compared with WT cultures (A–C and G–I). M. Quantification of the number of PV-positive perisomatic GABAergic synapses surrounding the soma and proximal dendrites of pyramidal neurons showed significant reduction of PV-positive perisomatic synapses in α7-KO cultures compared with WT cultures (***p<0.001 KO vs. WT; 5–6 neurons per culture, 4 different cultures; t-test). Scale bars as indicated.

NMDA receptor expression was reduced in cortical GABAergic interneurons in α7-KO cultures

As our previous findings demonstrated that NMDA receptor (NMDAR) expression and function was reduced in cortical α7-KO pyramidal neurons (Lin et al., 2014), we further examined if NMDAR was reduced in cortical GABAergic interneurons by examining NR1 and GABAergic neuronal marker GABA immunoreactivities in cultures. Double immunostaining with α-GABA and α-NR1 antibodies identified GABA-positive interneurons and NR1-positive puncta in the soma and dendrites of cortical GABAergic interneurons in WT and α7-KO cultures (Fig. 9A–9L, white arrows). The levels of NR1 immunoreactivity in GABAergic neurons were reduced in α7-KO cultures (Fig. 9D–9F and 9J–9L) compared with WT cultures (Fig. 9A–9C and 9G–9I), suggesting a reduction and hypofunction of NMDAR in cortical GABAergic interneurons.

Figure 9. Reduction of NMDAR in GABAergic interneurons in α7-KO cortical cultures compared with WT cultures.

A–F, Confocal imaging of NR1 (green) and GABA (red) immunofluorescence showing NR1 reduction in the soma and dendrites of GABA-positive interneurons (white arrows) in α7-KO cortical cultures (D–F) compared with WT cultures (A–C). G–L. High magnification of confocal imaging of NR1 (green) and GABA (red) immunofluorescence showing reduced NR1 puncta in the soma and dendrites of GABAergic interneurons in α7-KO cortical cultures (J–L) compared with WT cultures (G–I). Scale bar as indicated.

Discussion

The present study uses both in vivo and in vitro models to identify multiple deficits of cortical parvalbuminergic/GABAergic circuits from α7 nAChR deletion. The loss of α7 nAChR decreases cortical levels of PV and GABAergic markers, and alters the development of GABAergic synapses and circuits, particularly the development of PV-containing GABAergic interneurons and perisomatic synapses during late postnatal life and adulthood. Moreover, multiple neurochemical deficits in cortical PV-containing GABAergic interneurons with α7 nAChR deletion are reminiscent of characteristic features in schizophrenia. This implicates α7 nAChR-mediated control of cortical PV GABAergic development as crucial components in cortical circuit function and dysfunction in schizophrenia and other neuropsychiatric disorders. The present results thus provide a rationale for the association of α7 nAChR deletion in humans with a spectrum of neurodevelopmental and neuropsychiatric disorders. Relative loss of α7 receptors by genetic or epigenetic regulation could represent a fundamental pathogenic mechanism in such disorders.

In our studies, cortical GAD65/67, PV and GABA_Aα1 levels were decreased in α7-KO mice during postnatal development. GABA-mediated neurotransmission is crucial for shaping the functional maturation of cortical circuits (Le Magueresse and Monyer, 2013). GABA is synthesized by two isoforms of glutamic acid decarboxylase (GAD), GAD65 and GAD67. Both isoforms are present in most GABAergic interneurons, with GAD67 distributed throughout the soma and GAD65 in GABAergic synapses. The development of GAD65/67 expression is thus a reliable indicator of the maturation of cortical GABAergic interneurons’ ability to produce GABA. Reduced cortical GAD65/67 levels and decreased spontaneous IPSC frequency suggest deficits in cortical GABA synthesis and maturation of GABAergic circuits with α7 nAChR deletion. As GAD67 is a key regulatory sensor of cortical circuit activity (McNally et al., 2013), reduced GAD67 levels in PV-containing interneurons thereby further imply cortical PV GABAergic dysfunction with α7 nAChR deletion. Moreover, since PV expression is essential for synchronizing GABA release and network gamma oscillatory activities (Volman et al., 2011; McNally et al., 2013), reduced cortical PV levels could thus contribute to abnormal synchronization and oscillations in the cortex with α7 nAChR deletion. GABA_Aα1-containing receptors are crucial for mediating strong and fast inhibition of PV-positive basket cells onto pyramidal neurons and gamma oscillations (Hajos and Paulsen, 2009; Curley and Lewis, 2012; Lewis et al., 2012). Reduced GABA_Aα1 levels in α7-KO pyramidal neurons, along with reduced PV and GAD67 levels in PV-positive GABAergic interneurons may contribute to the compromised inhibitory control of network activity and could lead to abnormalities in gamma oscillations, working-memory, attention, sensory gating and learning.

Cortical PV levels in PV-positive GABAergic interneurons are specifically affected with α7 nAChR deletion, while the somatostatin immunoreactivity remains unaltered, suggesting that the development of neocortical PV-positive interneurons may be specifically regulated by α7 nAChR. In the developing brain, α7 nAChRs are expressed at high levels during the first postnatal week and reach peak levels at P7 in the prefrontal cortex and hippocampus, when GABAergic signaling changes from excitatory to inhibitory (Adams et al., 2002). α7 nAChR null mice indeed exhibited delays in the switch over from GABA-mediated excitation to inhibition during brain development (Liu et al., 2006). In addition, α7 nAChRs colocalize with GABA_A receptors in hippocampal interneurons and are critical in hippocampal gamma oscillations (Song et al., 2005; Le Magueresse et al., 2006; Liu et al., 2006; Bruel-Jungerman et al., 2011; Adams et al., 2012). The PV-positive, fast-spiking basket cells have uniquely fast membrane and synaptic properties that provide strong and fast inhibitory control of pyramidal neurons and gamma oscillations (Hajos and Paulsen, 2009; Curley and Lewis, 2012; Lewis et al., 2012). Furthermore, loss of the α7 nAChR impaired the development of PV-positive GABAergic interneurons and perisomatic synapses in cortical cultures. Indeed, α7 nAChRs have been found in mouse hippocampal interneurons and rat cortical PV GABAergic interneurons (Adams et al., 2001; Murakami et al., 2013). Our findings, along with other studies, thus suggest an intrinsic role of α7 nAChR in mediating cortical PV GABAergic neuronal and circuit development. Moreover, NMDAR expression in cortical GABAergic interneurons was reduced, suggesting NMDAR hypofunction in cortical GABAergic interneurons of α7-KO mice. Ablation of NMDARs in parvalbumin-positive interneurons of the cerebral cortex results in diminished theta oscillations, impaired spatial representation, and defective memory (Korotkova et al., 2010; Carlen et al., 2012). Similarly, postnatal ablation of NR1 in cortical and hippocampal GABAergic neurons leads to schizophrenia-related symptoms after adolescence, whereas a conditional NR1 knockout in which NMDAR deletion occurring after adolescence does not result in such abnormalities (Belforte et al., 2010). Loss or reduction of α7 nAChRs may lead to such profound glutamatergic hypofunction in PV-positive interneurons through neuregulin 1, NRG1/ErbB4 pathways in schizophrenia, since both α7 nAChRs and NRG1 are schizophrenia risk genes that associate with NMDARs and excitatory synapse formation on PV-positive interneurons (Hahn et al., 2006; Hancock et al., 2008; Ting et al., 2011). Our findings thus implicate NMDAR hypofunction in GABAergic interneurons, which may contribute to cortical dysfunction leading to impaired attention working-memory and learning in α7 nAChR deletion models.

α7 nAChR null mice show many neurochemical findings that are analogous to features observed in schizophrenia. First, α7 nAChR levels are reduced in multiple brain regions including the prefrontal cortex in schizophrenia patients (Freedman et al., 1995; Breese et al., 1997; Freedman et al., 2000; Mathew et al., 2007; Freedman and Goldowitz, 2010; Ross et al., 2010; Olincy and Freedman, 2012). . Though the transient upregulation of another nicotinic receptor subtype α4β2 in α7 nAChR null mouse cortex and hippocampus during postnatal development may have some compensatory effects on brain development (Orr-Urtreger et al., 1997; Yu et al., 2007), the changes of α4β2 nAChRs and their association with schizophrenia patients remain less clear and inconsistent in the literatures (Adams and Stevens, 2007). In addition, α7 nAChR selective agonists have demonstrated efficacy in clinical trials of schizophrenia patients and in animal models of positive and negative symptoms and cognitive deficits of schizophrenia (Hajos and Rogers, 2010; Tregellas et al., 2011; Wallace and Porter, 2011; Cannon et al., 2013; Freedman, 2014). Finally, cortical PV GABAergic circuitry is altered in schizophrenia patients in a manner similar to that observed in α7-KO mice. Deficient cortical GABA synthesis is a conserved feature and loss of GAD67 is prominent in the PV-positive GABAergic interneurons in schizophrenia patients (Fung et al., 2010; Lewis et al., 2012). Postsynaptic GABA_Aα1 expression in cortical layer 2/3 pyramidal neurons is reduced in schizophrenia patient dorsolateral prefrontal cortex (Lewis et al., 2012). Furthermore, the level of PV in PV-positive interneurons, but not the number of PV interneurons, is reduced in the prefrontal cortex of schizophrenia patients (Lewis et al., 2012). Thus the neurochemical abnormalities of the cortical PV GABAergic circuit in α7-KO mice mirror the pathological alterations in schizophrenia patients. Our findings thus demonstrate impaired cortical PV GABAergic development and multiple neurochemical deficits in α7 nAChR gene deletion models that are reminiscent of PV GABAergic deficits in schizophrenia, thus implicating crucial roles of α7 nAChR in cortical PV GABAergic circuit development and dysfunction in schizophrenia and other neuropsychiatric disorders.

α7 nAChR heterozygous and null mutants in different mouse strains showed similar GABAergic deficits in the cortex and hippocampus (Adams et al., 2008; Adams et al., 2012). While our findings showed cortical PV GABAergic deficits in the prefrontal cortex of C57BL/6 strain of α7 nAChR null mutant mice, we also observed the reduction of PV and GABAergic markers in other cortical areas such as somatosensory cortex of α7 nAChR null mice (Supplemental Fig.2). Other lines of investigation have shown decreased levels of GAD65 and GABAA receptors in the hippocampus as well as abnormal hippocampal auditory gating in C3H α7 nAChR heterozygous mutant mice (Adams et al., 2008; Adams et al., 2012). Moreover, both the α7 nAChR null mutant and an inbred mouse strain that has lower expression of α7 nAChRs showed deficits in hippocampal (P20/N40) inhibitory sensory gating in DBA/2 mice (Stevens et al., 1998; Freedman, 2014). The P50 auditory sensory gating deficit has been considered a leading endophenotype in schizophrenia, and polymorphisms in the 5’ promoter of CHRNA7 and in a second gene CHRFAM7A with a partial duplication of CHRNA7 are associated with schizophrenia and the P50 sensory gating deficit (Leonard et al., 2002; Sinkus et al., 2009; Ross et al., 2010; Freedman, 2014). Taken together, loss or reduction of α7 nAChRs may contribute to the totality of deficits expected in GABA-mediated inhibition both in the PV-positive interneurons on prefrontal cortex executive function deficits as well as hippocampal auditory gating deficits in schizophrenia patients.

Supplementary Material

01. Supplemental Figure 1.

Representative morphology of a cultured cortical pyramidal neuron and typical firing pattern of a recorded pyramidal neuron responding to intracellular injection (700 msec) of depolarized current.

02. Supplemental Figure 2.

Reduction of PV and VGAT levels in the somatosensory cortex of α7-KO mice (D–E) compared with WT mice (A–C). Scale bar as indicated.

Abbreviations

nAChR

nicotinic acetylcholine receptor

PV

parvalbumin

SST

somatostatin

GAD65/67

Glutamic Acid Decarboxylase 65/67

GABA_Aα1

α1 subunit of GABA_A receptors

VGAT

vesicular GABA transporter

NMDAR

N-methyl-D-aspartate receptor

VGLUT1

vesicular glutamate transporter 1

WT

wild-type

α7-KO

α7 nAChR null