Reelin signaling modulates GABA_B receptor function in the neocortex

Abstract

Reelin is a protein that is best known for its role in controlling neuronal layer formation in the developing cortex. Here, we studied its role for postnatal cortical network function, which is poorly explored. To preclude early cortical migration defects caused by Reelin deficiency, we used a conditional Reelin knock-out (Reln^cKO) mouse, and induced Reelin deficiency postnatally. Induced Reelin deficiency caused hyperexcitability of the neocortical network in vitro and ex-vivo. Blocking Reelin binding to its receptors ApoER2 and VLDLR resulted in a similar effect. Hyperexcitability in Reln^cKO organotypic slice cultures (OTCs) could be rescued by co-culture with wildtype OTCs. Moreover, the GABA_B receptor (GABA_BR) agonist baclofen failed to activate and the antagonist CGP35348 failed to block GABA_BRs in Reln^cKO mice. Immunolabelling of Reln^cKO cortical slices revealed a reduction in GABA_BR1 and GABA_BR2 surface expression at the plasma membrane and western blot of Reln^cKO cortical tissue revealed decreased phosphorylation of the GABA_BR2 subunit at serine 892 and increased phosphorylation at serine 783, reflecting receptor deactivation and proteolysis. These data show a role of Reelin in controlling early network activity, by modulating GABA_BR function.

Graphical Abstract

We suggested the following modification of synaptic function after conditional reelin knockout in the cerebral cortex: After Reelin deficiency, the presynaptic reelin signaling cascade via the reelin receptors ApoER2 and Vldlr and the adaptor protein Dab1, is no longer activated. This causes a decreased phosphorylation of presynaptic GABA_B-receptors at S892, while phosphorylation at S783 increases, likely induced by AMP-activated protein kinase (AMPK). As a consequence, GABA_B-receptors are degraded, calcium influx via NMDA-receptors and synaptic vesicle release is increased. We believe that our findings revealed a new synaptic function of reelin expressed by cortical interneurons during cortical development.

Introduction

Reelin is an extracellular glycoprotein that controls several aspects of mammalian brain development and function. The most prominent role of Reelin is the control of neuronal migration and layer formation in the developing cerebral cortex, as evidenced by numerous histological studies on reeler mutant mice that lack Reelin expression due to a defect of the Reelin gene (Caviness 1976; Curran and D'Arcangelo 1998; Lambert de Rouvroit and Goffinet 1998). Through signaling via its membrane receptors (Bock and May 2016; Cooper and Howell 1999), Reelin guides the migration of newborn neurons and orchestrates the development of cortical layers. In the developing and adult brain, the majority of GABAergic interneurons in the neocortex expresses Reelin. The canonical Reelin signaling cascade involves direct binding of Reelin to ApoER2 and VLDLR and subsequent activation of the intracellular adapter protein Disabled-1 (DAB1) by tyrosine phosphorylation (Cooper and Howell 1999; D'Arcangelo et al. 1999; Hiesberger et al. 1999; Howell et al. 2000; Trommsdorff et al. 1999).

In the absence of Reelin or its receptors, the process of neuronal migration is compromised, which causes severe abnormalities in cortical lamination. The resulting phenotype was initially described as an inversion of the layers (Caviness 1982; Caviness and Sidman 1973). Although it had already been shown that deviant barrel formation in the somatosensory cortex of the reeler is compromised (Caviness 1976; Welt and Steindler 1977), it has been recently demonstrated that the barrel field retains its proper somatotopic organization (Guy et al. 2015; Wagener et al. 2010). Besides its cortical migration defects, the reeler mutants exhibit cerebellar hypoplasia and a neurological phenotype that is characterized by ataxia ( (Miyata et al. 1997). A similar phenotype was observed in human patients carrying homozygous mutations in the Reelin gene, resulting in lissencephaly and cerebellar hypoplasia (Hong et al. 2000). Moreover, in humans, heterozygous Reelin mutations were described that cause autosomal-dominant temporal lobe epilepsy (Dazzo et al. 2015).

Gamma-aminobutyric acid (GABA) is the major inhibitory neurotransmitter in the mammalian CNS and plays a key role in modulating neuronal activity. GABA mediates its action via two different classes of receptors, ionotropic receptors (GABA_A, GABA_C) and metabotropic GABA_B receptors. The GABA_B receptors are guanine nucleotide-binding protein (G protein)-coupled metabotropic receptors that modulate Ca²⁺- and potassium (K⁺) – channels, and elicit both presynaptic and slow postsynaptic inhibition (Benarroch 2012; Bettler et al. 2004). In particular, presynaptic GABA_BRs are coupled to Ca²⁺ channels, regulating the release of neurotransmitters, while postsynaptic GABA_BRs are coupled to K⁺ inward rectifying (Kir) channels (Kir3), regulating postsynaptic slow inhibition (Pinard et al. 2010). Remarkably, GABA does not mediate hyperpolarization-dependent inhibition during early development, since GABA_AR signaling is mainly depolarizing and excitatory during this period, while GABA_BRs are uncoupled from G-proteins and Kir3 channels until the end of the first postnatal week (Ben-Ari et al. 2012; Fukuda et al. 1993; Owens and Kriegstein 2002). Moreover, GABA is abundant in the neonatal nervous system and activates GABA_BRs (Ben-Ari et al. 2007). In the cortex, GABA_BR1 and GABA_BR2 were detected by immunocytochemistry as early as E14, and GABA_B1Rs were found to colocalize with GABA_B2Rs in neurons of the marginal zone and the subplate, indicating that these proteins are coexpressed and could be forming functional GABA_BRs during prenatal development in vivo (López-Bendito et al. 2002). In the murine neocortex, Reelin-haploinsufficiency has been shown to disrupt the developmental trajectory of GABA excitation/inhibition balance (Bouamrane et al. 2016), suggesting a possible interaction between Reelin and the GABA receptors.

Adult conditionally induced Reln^cKO mice exhibited altered hippocampal LTP and a subtle behavioural phenotype while the cortical architecture was shown to be indistinguishable from their wild-type littermate (Lane-Donovan et al. 2015). The question whether Reelin deficiency affects early cortical neuronal activity remains to be solved. To address this question, we investigated the effect of postnatally induced Reelin deficiency on Ca²⁺ signaling and on synaptic function in the neocortex of Reln^cKO mice. Since synaptic release is mainly controlled by GABA_BRs that mediate inhibitory GABA effects around the second postnatal week, we focused here on a potential interaction between Reelin signaling and GABA_BRs function.

Material and Methods

Reelin conditional knockout mice (Reln^cKO)

Animals were housed in a standard 12-h light cycle and fed ad libitum with standard mouse chow. All care and use of experimental animals were respected according to the Federal German law with permission Nr. 84-02.04.2016.A383 and the ARRIVE guidelines. The generation of the conditional Reln^cKO line was previously described (Lane-Donovan et al. 2015). To obtain conditional Reelin knockout mice (Reln^flox/flox CAG-Cre^ERT2 mice), we crossed Reln^flox/flox mice with hemizygous tamoxifen-inducible Cre recombinase expressing mice (CAG-Cre^ERT2) (Hayashi and McMahon 2002). For the experiments, only Reln^flox/flox CAG-Cre^ERT2 male mice were selected and then crossed with Reln^flox/flox female mice to generate Reln^{flox/flox wildtype} (WT) and Reln^flox/flox CAG-Cre^ERT2 (Reln^cKO) siblings as verified by PCR. The cKO mouse line ubiquitously expresses a fusion protein comprising Cre recombinase and a mutated form of the estrogen receptor (Cre-ERT2). Tamoxifen administration induces nuclear Cre activation and knockout of the floxed Reelin gene. 20 μl tamoxifen (Cat# T5648, Sigma, Deisenhofen, Germany) was dissolved in corn oil at a concentration of 20 mg/ml by shaking overnight at 37°C and fed to newly born pups at P1 for 5 consecutive days. At P14 animals were deep anesthetized with isoflurane CP® (CP-Pharma, Burgdorf, Germany) to minimize suffering and killed by decapitation. Ex-vivo slices from somatosensory cortex were explanted from each animal for acute slice preparation at P14 or at P0 for OTCs preparation by respecting the coordinates indicated in Allens mouse brain atlas. A graphical flow chart for ex-vivo experimental procedures is shown in (Fig. 1A). For each outcome measure, we used 5 Reln^wt and 5 Reln^cKO mice (each experiment was repeated 3 times). In total 390 animals were used for this study. From each animal we obtained 3-4 slices. We recorded from each slice 3 different areas of interests (ROI). In each area of interest, we averaged amplitude, frequency and Ca²⁺ transient half-width of 6 cells. The average of the 3 ROI has been plotted in the box-plot as single value. This study was not pre-registered. No exclusion criteria were pre-determined and the study was exploratory. No blinding, randomization was performed to allocate subjects in the study. No sample size calculation was performed in this study.

Fig. 1. Recording of neuronal activity in Reln^cKO mice.

(A) Graphical flow chart of experimental procedures. In each experiment, slices were prepared from 5 wt and 5 Reln^cKO mice. (B): Confocal image examples of OGB-1 AM loaded P14 acute neocortical wt and Reln^cKO slices. Scale bars: 20 μm. (C): Typical example of spike waveform from a recorded cell which shows amplitude and Ca⁺² transient (spike) half-width. (D): The box-plot in the graph represents the values of maximal increase in Ca²⁺-signal amplitude in control wt and Reln^cKO acute slices expressed as ΔF/F⁰, which is unaltered. (E): The Ca²⁺ frequency is significantly higher in the Reln^cKO when compared to wt, and (F) the Ca⁺² transient half-width is significantly higher in Reln^cKO when compared to wt. Mann–Whitney U test; ***P<0.001. The number of acute slices analysed is indicated above the box-plots in (D). To assess pyramidal cells and interneurons separately, acute slices were transfected with GCaMP6s plasmid to visualize the cell type (G-J). (G): Confocal images of a GCaMP6s transfected pyramidal cell in a wt P14 acute neocortical slice during Ca²⁺-imaging at resting fluorescence (F⁰) and at maximal amplitude peak. Scale bars: 20 μm. (H): The box-plot in the graph shows no change in maximal Ca²⁺-amplitude signals between control wt and Reln^cKO groups for both cell types, but shows (I and J) a significant increase in Ca²⁺ frequency and Ca⁺² transients half-width in Reln^cKO recorded slices in comparison to wt for both pyramidal cells and interneurons. One-way ANOVA on Ranks followed by Dunn’s Multiple Comparison Test, ***p < 0.001 and **P<0.01. The number of single cells in (H-J) is indicated above the box-plots in (H). The data is obtained from 3 independent acute slice preparations.

PCR and genotyping

All samples were stored at −20°C until PCR analysis. DNA from samples of ear, tail and brain tissue were isolated with ReliaPrep gDNA kit (Cat# A205, Promega, Mannheim, Germany). All procedures were performed according to protocols provided by the manufacturer. The amounts of DNA isolated from the various samples were determined by spectrophotometry with the Genova Nano system (Jenway). DNA was amplified by PCR. PCR reactions were performed in a total volume of 50 μl reaction mixture containing 200 ng of template DNA, Soriano buffer (0.67 M Tris, 0.16 M ammonium sulphate, 67 mM MgCl₂, 67 μM EDTA and 50 mM β-Mercaptoethanol), Taq polymerase, 2 μL DMSO and 10 mM dNTPs. For genotyping we used the following primers: Wildtype mice, forward primer 5′-ATAAACTGGTGCTTATGTGACAGG-3′, reverse primer 5′-AGACAATGCTAACAACAGCAAGC-3′ (450 bp). Reln^flox/flox mice, forward primer 5′-GCTCTGGCCAAGCTTTATC-3′, reverse primer 5′- CGCGATCGATAACTTCGTATAGCATAC-3′ (1200 bp). For detection of CAG-Cre^ERT2, forward primer 5′-ATTGCTGTCACTTGGTCGTGG-3′, reverse primer 5′-GGAAAATGCTTCTGTCCGTTTGC-3′ (200 bp). The amplification products were verified on a 2% agarose gel in TBE buffer.

Organotypic cultures and pharmacological treatment

OTCs were prepared from newborn postnatal day 0 (P0) mice. 3-4 OTCs from somatosensory cortex were collected from each animal. All solutions used for OTCs preparation were sterile and all preparations were performed in a laminar air flow bench with horizontal counter flow (Horizontal Flow, ICN, Biomedicals, Eschwege; Germany). The mice were briefly anesthetized and decapitated. The skull was removed gently, and the cortex was placed on the chopper plate (McIllwain, Bad Schwalbach, Germany). Somatosensory cortex was cut into 350 μm thick slices, and the slices were gently transferred into ice cold buffered salt solution (GBSS, Cat# 24020117, Gibco, Eggenstein, Germany) containing 25 mM D-glucose. After 30 minutes of recovery, the slices were transferred onto coverslips (12 x 24 mm, Kindler). A mixture of chicken plasma (Sigma) and GBBS/thrombin (Merck) was mixed in a proportion of 2:1 and then allowed to coagulate for 45 min. The coverslip was transferred into a roller tube (Nunc) filled with 750 μl semiartificial medium and placed in a roller incubator at 37 °C. To induce the knockout, the OTCs were directly stimulated after preparation with 1 μM (Z)-4-hydroxytamoxifen (4-OHT) (Cat# 3412, Tocris, Wiesbaden, Germany) for 5 consecutive days and kept for experiments until DIV14. A graphical flow chart for OTCs experimental procedures is shown in (Fig. 2A). For the Ca²⁺ imaging experiments, the following drugs were used: Baclofen (10 μM, Cat# 0417, Tocris, Wiesbaden, Germany), CGP35348 (10 μM, Cat# 1245/10, Tocris, Wiesbaden, Germany), PP2 (1 μM, Cat# 1407/10, Tocris, Wiesbaden, Germany), pertussis toxin (PTX, 500 ng/mL, Cat# 3097/50U, Tocris, Wiesbaden, Germany), CNQX (10 μM, Cat# 0190/10, Tocris, Wiesbaden, Germany) and APV (50 μM, Cat# 0190/10, Tocris, Wiesbaden, Germany). As an inhibitor of low-density lipoprotein receptor–related proteins, we used the recombinant Mouse LRPAP Protein (LDL receptor-related protein-associated protein 1; also named receptor-associated protein (RAP), 50 ng/ml, Cat# 4480-LR, R&D Systems). RAP serves as a molecular chaperone for LDL receptor family proteins including VLDL and APOER2 and prevents interaction of ligands with these receptors. Therefore, RAP is used to block the canonical Reelin pathway by blocking Reelin from binding to VLDL and APOER2 (Bu and Schwartz 1998; Gong et al. 2007; Herz et al. 1991).

Fig. 2. Secreted Reelin rescues impaired Ca²⁺-frequency in Reln^cKO OTCs.

(A) A graphical flow chart for OTCs experimental procedure. (B) Schematic representation of the co-culture experimental approach. Left: Two co-cultured wt OTCs. Middle: A wt OTC co-cultured together with a Reln^cKO OTC. Right: Two co-cultured Reln^cKO OTCs. (C) Co-culture medium lysates from DIV14 OTCs (40 OTCs obtained from 5 mice for each experimental group) were analysed by western blot. (D) Quantification of the western blot 180-kDa Reelin signals. An actin antibody was used as loading control. The densiometric quantification showed a decreased Reelin protein level in the wt + Reln^cKO co-culture medium and almost complete absence in the Reln^cKO + Reln^cKO co-culture medium. One-way ANOVA followed by Holm-Sidak Multiple Comparison Test, ***p < 0.001 wt + Reln^cKO compared to wt + wt; ***P<0.001 wt + Reln^cKO compared to Reln^cKO + Reln^cKO. At 14 DIV, OTCs were loaded with OGB-1 AM and recorded. (E) The box-plot in the graph represents the values of maximal increase in Ca²⁺ signal amplitude, which shows no change in all experimental groups. (F and G) The box-plot shows a significant increase in Ca²⁺ frequency and Ca⁺² transient half-width in Reln^cKO co-cultured OTCs. Note that the enhanced frequency and Ca⁺² transients half-width observed in the Reln^cKO OTC are restored to a normal level when it is co-cultured together with a wt OTC. The number of recorded OTC is indicated above the box-plots in (E). One-way ANOVA on Ranks followed by Dunn’s Test, ***p < 0.001. The data is obtained from 3 independent OTCs preparations.

Expression plasmids and biolistic transfection

Transfection was performed using a Helios Gene Gun (Bio-Rad) as described previously (Wirth and Wahle 2003). In brief, cartridges were prepared by coating 10 mg gold particles (Ø = 1μm; Bio-Rad) with genetically encoded Ca²⁺ indicator pGP-CMV-GCaMP6s (GCaMP6s). GCaMP6s was a gift from Douglas Kim (RRID: Addgene_40753) (Chen et al. 2013). To prevent excitotoxicity during transfection, glutamate receptors were temporarily blocked with 3 mM kynuric acid (Cat# K3375, Sigma, Deisenhofen, Germany) and 50 mM APV (Cat# A5282, Sigma, Deisenhofen, Germany) before blasting. The blockers were washed out 3 hours after transfection.

Ca²⁺ imaging using Spinning disc laser microscopy

For studying network activity, Ca²⁺ imaging was performed in somatosensory cortex of supragranular layers II/III with the Ca²⁺ indicator Oregon Green BAPTA-1 Acetoxymethyl ester (OGB-1 AM) (Cat# O6807, Molecular Probes, Eugene, OR, USA). Acute slices (P14) or OTCs DIV14 were loaded according to our previously published protocol (Hamad et al. 2015). In brief, a solution of 20% PF127 (Cat# P2443, Sigma, Steinheim, Germany) dissolved in DMSO (w/v) (J.T. Baker) was prepared to dissolve OGB-1 AM. The final loading solution concentration was 1 μM OGB-1 AM. After loading, the slices were washed several times to remove excess dye and allowed to recover for an hour. The slices were then transferred to the recording chamber mounted on a fixed stage of an inverted microscope, and perfused with 95% O₂, 5% CO₂-bubbled artificial cerebrospinal fluid (ACSF; 3–5 ml/min) at 32 ± 2 °C. Composition of ACSF: (125 mM NaCl, 5 mM KCl, 2 mM CaCl₂, 1 mM MgSO₄, 25 mM NaHCO₃, 1.25 mM NaH2PO4, 25 mM glucose, pH 7.4). The osmolality was 295±5 mOsm as determined with a cryoscopic osmometer (Osmomat 030, Gonotec, Berlin, Germany). Fluorometric Ca²⁺-recordings were made using a Visiscope spinning-disc confocal system CSU-W1 (Visitron, Munich, Germany) featuring a spinning disk unit CSU-W1-T2 and a sCMOS digital scientific grade camera (4.2 Mpixel rolling shutter version) on an inverted Nikon Ti-E motorized microscope using a CFI P-Fluor 20× objective (NA 0.5, WD = 2.10 mm). Images were acquired at 3 frames per second with exposure times of 330 ms with VisiView image acquisition software (Visitron, Munich, Germany). The Ca²⁺ dye OGB-1 AM and the biosensor GCaMP6s were excited at 488 nm. Emitted fluorescence was collected through ET 525/50 filter for OGB-1 AM and GCaMP6s. Fluorometric data are expressed as ΔF/F⁰ (background-corrected increase in fluorescence divided by the resting fluorescence). Raw data delivered in the form of a linear 16-bit intensity scale were plotted as fluorescence intensity versus time. Pyramidal cell or interneuron somata were chosen as the region of interest (ROI). The background fluorescence measured near a ROI was then subtracted from these raw data. The baseline fluorescence (F⁰) was calculated as an average of 20 frames in a time window without neuronal activity (as judged by visual inspection). Subsequently, data were normalized to the mean fluorescence intensities [ΔF/F⁰ =(F–F⁰)/F⁰], allowing the comparison of data across experiments. Spike (calcium transient) half-width was calculated as the width of the spike at half-maximal amplitude (Weir et al. 2014).

Immunohistochemistry and surface labelling

At DIV10, OTCs were fixed with 4% paraformaldehyde (PFA) in 0.1 M phosphate buffer pH 7.4 and warmed to 36°C for 1 hour. After washing twice in TBS (Tris buffered Saline: 50 mM Tris, 150 mM NaCl, pH 7.6) and permeabilization in TBST (TBS, 0.1% Triton X), OTCs were blocked for 1 hour with 1% normal goat serum in TBST. Primary antibodies diluted in TBST were incubated for 24 h at room temperature. The following antibodies were used: Mouse anti-Reelin G10 (Cat# MAB5364, RRID: AB_2179313, 1:1000) , rabbit anti-Cre recombinase (Cat# 69050 , RRID: AB_2314229, 1:2000), rabbit anti-Cux-1 (CULT-1, Cat# ab140042, Abcam, Cambridge, UK, 1:500), rabbit anti-Wolframin 1 (WFS 1, Cat# 11558-1-AP, RRID: AB_2216046, 1:2000), mouse anti-Parvalbumin (Parv, Cat# 235, RRID: AB_10000343, 1:3000), rabbit anti-Glutamate Decarboxylase-65 (GAD65, Cat# G5038, RRID: AB_259920, 1:1000). After washing twice in TBS, the secondary antibodies were added accordingly: Goat anti mouse or goat anti rabbit biotinylated (Cat# E043201-8, Dako, Hamburg, Germany, 1:300). After several washes in TBS buffer, the slices were subjected to ABC-horseradish peroxidase method using diaminobenzidine as chromogen. For immunofluorescence, the slices were subsequently washed 3 × 15 min with TBST and incubated with secondary antibodies (anti-mouse IgG coupled to Alexa594 and anti-rabbit IgG coupled to Alexa488 in TBS (ThermoFisher, Darmstadt, Germany, 1:300) for 30 min at room temperature. After three repetitive washings with TBS-Tween, the OTCs were mounted with sRIMS mounting medium (70% sorbitol w/v in 0.02 M phosphate buffer with 0.01% sodium azide, pH 7.5). Fluorescence was analysed with the Visiscope spinning-disc confocal system as mentioned above. For GABA_BR1 und GABA_BR2 surface labelling, P14 wt and Reln^cKO acute slices were fixed with 4% paraformaldehyde. After several washes in TBS, a group of acute slices was permeabilized in TBST to detect total GABA_BRs and the other group was not permeabilized to detect only surface expression. The slices were blocked and incubated with the following primary antibodies that recognize only the N-terminus (extracellular) domain of GABA_BR1 (mouse anti-GABA_BR1, Cat# ab55051, RRID: AB_941703, 1:500) or the N-terminus (extracellular) domain of GABA_BR2 (Rabbit anti-GABA_BR2, Cat# 4819, RRID: AB_2108339, 1:750) for 24 h at room temperature. The slices were then washed and incubated with appropriate secondary fluorescent antibodies for 30 min. After repetitive washings with TBS, the slices were mounted with sRIMS buffer and mean fluorescence intensity was measured from layers II/III somatosensory cortices with the Visiscope spinning-disc confocal system.

Western blotting

Equal amounts of cortical tissue samples (22.5 μg) were homogenized in urea assay lysis buffer (100 mM Tris-HCl, pH= 7.5; 12 mM magnesium acetate tetrahydrate and 6M urea) with protease and phosphatase inhibitors and then centrifuged at 14,000 rpm for 15 min to remove debris and nuclei. The SDS loading buffer was run on a 10% SDS-polyacrylamide gel and transferred to a nitrocellulose membrane for 1 h 40 min in transfer buffer (25 mM Tris/Base, 192 mM glycine). Membranes were then blocked for 1 h at room temperature (RT) while shaking in blocking buffer with TBS (Cat# P/N 927-50000, Li-Cor, Bad Homburg, Germany). Membranes were incubated overnight (O/N) at 4 °C with the primary antibody and for 2h at RT with the secondary antibody. After each incubation step, membranes were washed for 3 × 15 min in TBS with 0.1% Tween 20. Membranes were imaged with the Odyssey immunoblot software (Lincoln). The Odyssey software system was also used for densitometric analysis. The following antibodies were used: Rabbit anti-GABA_BR2 (Cat# 4819, RRID: AB_2108339, 1:500) , mouse anti-GABA_BR1 (Cat# ab55051, Abcam, Cambridge, UK, 1:500), rabbit anti-GABA_BR2 p-Ser 783 (Cat# TA309142, OriGene Technologies, Herford, Germany, 1:1000), p-GABA_BR2 S892 (Cat# PPS073, RRID: AB_2108325, 1:1000), mouse anti-Reelin G10 ((Cat# MAB5364, RRID: AB_2179313, 1:1000), and ß-Actin (Cat# ab8227, RRID:AB_2305186,1:10000).

Statistical analysis

Statistical analysis was performed with Sigma Stat 12 (SPSS Incorporated). The data are presented as box plots with median (center line), minimum, and maximum (whiskers), and 25th–75th percentiles (box). Comparisons between two groups were performed with Students unpaired T-test when normality test (Shapiro-Wilk) passed, otherwise the Mann–Whitney test. Comparisons between groups larger than two were performed with one-way-ANOVA and a Holm-Sidak Multiple Comparison Test for post-hoc analysis if normality test passed. If normality failed, we run One-Way-Anova on Ranks followed by Tukey’s multiple comparison test for post hoc analysis to isolate the significant groups. If the sample sizes were unequal, the Dunn`s multiple comparison test was used to isolate the significant groups. Results were considered statistically significant at p<0.05.

Results

Cortical layering is not altered after postnatal loss of Reelin in Reln^cKO mice

To bypass the cortical layer malformations that are present in the reeler mutant, tamoxifen was fed to newborn pups (~P1) for 5 consecutive days. Around P14, animals were sacrificed, and sections of cortical tissue were subjected to immunohistochemical staining (Fig. S1). First, we confirmed that tamoxifen administration induces nuclear Cre activation and knockout of the floxed Reelin gene. We found that tamoxifen administration at P1 induced complete elimination of Reelin immunostaining at P14 (Fig. S1A). The pattern of immunostaining against wolframin (Wfs1), a protein expressed by layer II and IV neurons in the mouse somatosensory cortex, was indistinguishable from wt (Fig. S1B), while the layer-specific distribution of this protein is disrupted in the reeler cortex (Boyle et al. 2011). Next, we examined the distribution of cut-like homeobox 1 (Cux-1) protein, a transcription factor expressed in neurons throughout layers II–IV. In the cortex of adult reeler mice, most Cux-1-positive neurons are found in cortical layers V and VI and do not respect the strict boundary that characterizes wildtype cortical layering (Nieto et al. 2004). The distribution of Cux-1 immunostaining in the cortex of Reln^cKO mice was found in the supragranular layers II-IV and was indistinguishable from wt (Fig. S1C). We then examined the distribution of the fast-spiking interneuron marker Parvalbumin (Parv) and found it to be unchanged in the Reln^cKO when compared to wildtype mice (Fig. S1D). These findings confirm that postnatal Reelin elimination did not alter the distribution of layer specific expression markers in the neocortex.

Excessive Ca²⁺ spike frequency in early postnatally Reelin-deficient Reln^cKO mice

Using Ca²⁺ imaging, we compared spontaneous activity in acute slices prepared from wt and Reln^cKO littermates at P14 (both pup groups were fed with tamoxifen). In our experience (Hamad et al. 2015) as well as in other studies (Yuste et al. 2011), imaging a large network of neuronal populations loaded with the calcium indicator OGB-1 AM under a spinning-disk confocal microscope, equipped with fast cameras, can be used to image thousands of neurons simultaneously without significant photobleaching, with good signal-to-noise ratio and minimized cellular damage because it does not require electrodes penetrating the tissue. To ensure that tamoxifen per se did not alter basic synaptic transmission, we recorded Ca²⁺ amplitude and frequency in DIV14 (days in vitro 14) OTCs loaded with OGB-1 AM from wt mice which were stimulated for 5 consecutive days either with 4-hydroxytamoxifen (4-OHT) or vehicle (DMSO) as a control group (Fig. S2). Neither 4-OHT nor vehicle groups showed any difference in Ca²⁺ amplitude, frequency or Ca²⁺ transient half-width when compared to the untreated wt control group (Fig. S2). Next, we compared Ca²⁺ signal amplitude, frequency and Ca²⁺ transient half-width in wt and Reln^cKO slices using the Ca²⁺ indicator OGB-1 AM (Fig. 1B-F). Ca²⁺ amplitudes were not altered in Reln^cKO when compared to the wt control group (Fig. 1D). Strikingly, the Ca²⁺ frequency and Ca²⁺ transient half-width were significantly increased in Reln^cKO (Fig. 1E and F). Recordings were performed in acute slices prepared at P14 and transfected with a genetic construct encoding the Ca²⁺ indicator GCaMP6s (Chen et al. 2013). GCaMP6s is very sensitive to Ca²⁺ changes and we found it to be distributed in soma, dendrites and axon (Fig. 1G). Criteria to distinguish pyramidal cells and interneurons were based on dendritic and axonal patterns (Hamad et al. 2011; Hamad et al. 2014; Karube et al. 2004; Kawaguchi et al. 2006). To confirm the cell type, after Ca²⁺ recording, acute slices were fixed and stained against glutamate decarboxylase-65 (GAD-65) which labels the GABA biosynthesis enzyme in the soma of all GABAergic interneurons but does not label glutamatergic pyramidal cells (Fig. S3). Our results confirmed an unchanged amplitude in Reln^wt compared to Reln^cKO in both pyramidal cells and interneurons (Fig. 1H) and revealed an enhanced frequency and Ca²⁺ transient half-width in both pyramidal cells and interneurons in the Reln^cKO (Fig. 1I and J), suggesting that Reln^cKO mice exhibit defects at the presynaptic level. Moreover, Ca²⁺ amplitude and Ca²⁺ transient half-width did not differ between pyramidal cells and interneurons in both wt and Reln^cKO slices during GCaMP6s recordings, confirming a previously published study (Weir et al. 2014). Therefore, the subsequent recording experiments were performed with the calcium dye OGB-1 AM.

Wildtype Reelin rescues Ca²⁺ frequency in Reln^cKO neurons

Next, we investigated whether secreted Reelin from wt OTCs might rescue the abnormal neuronal activity observed in the Reln^cKO OTCs. To address this question, we co-cultured OTCs as follows: wt + wt, or Reln^cKO + Reln^cKO, or wt + Reln^cKO (Fig. 2B). To detect the abundance of secreted Reelin in the culture medium under these three conditions, we quantified secreted Reelin protein with western blot from culture medium. As expected, the amount of secreted Reelin was abundant in the wt + wt, a lesser amount was detectable in the wt + Reln^cKO, whereas Reelin was almost undetectable in the Reln^cKO + Reln^cKO co-culture medium (Fig. 2C and D). Ca²⁺ recordings experiments showed that the Ca²⁺ amplitude did not change under any of the three experimental conditions (Fig. 2E) confirming our interpretation that Reelin does not affect postsynaptic activity. However, in the Reln^cKO OTCs co-cultured with wt, the Ca²⁺ frequency and Ca²⁺ transient half-width in the recorded Reln^cKO OTC was reduced to the same level as in the wt OTC (Fig. 2F and G). These results suggest that Reelin secreted from wt OTCs into the incubation medium restored presynaptic activity in Reln^cKO OTCs to a basic level comparable to the wt control group.

GABA_BR function is impaired in Reln^cKO mice

GABA_BRs are present in GABAergic neuronal terminals (as autoreceptors) and in glutamatergic and other terminals (heteroreceptors) (Benarroch 2012). Around the second postnatal week, the only source of Reelin in the neocortex are inhibitory interneurons (Pohlkamp et al. 2014). We speculated that the increased Ca²⁺ spiking frequency in Reln^cKO mice in pyramidal cells and interneurons that we observed, was likely due to a defect of GABA_BR function at the GABAergic or glutamatergic presynaptic terminal. If presynaptic GABA_BRs at the GABAergic neuronal terminals were defective in Reln^cKO mice, activated GABA_BRs would no longer be able to block Ca²⁺ channels because presynaptic GABA_B receptors inhibit N type (Cav2.2) or P/Q type (Ca_v2.1) Ca²⁺ channels, resulting in reduced neurotransmitter release (Benarroch 2012). In turn, Ca²⁺ channels trigger the release of GABA, and thereby inhibit neighbouring cells. However, Reln^cKO mice exhibited a higher Ca²⁺ frequency when compared to wt control (Fig. 1), suggesting that presynaptic GABA_BRs at GABAergic neuronal terminals were not defective in Reln^cKO mice. Therefore, we tested a second possibility, i.e. defective presynaptic GABA_BRs at glutamatergic terminals in Reln^cKO mice. To test this possibility, we recorded wt and Reln^cKO OTCs in the presence of the AMPA- and kainate receptor blocker CNQX and the NMDARs antagonist APV at DIV14 (Fig. S4). Bath application of 10 μM CNQX immediately decreased the Ca²⁺ amplitude, frequency and Ca²⁺ transient half-width in Reln^cKO to the same extent as in the wt control group (Fig. S4 A and B). Similarly, APV application elicited the same effect (Fig. S4 D-F). The observation that the excessive Ca²⁺ frequency in the Reln^cKO can be blocked with CNQX or APV suggests that glutamatergic signaling in the absence of Reelin is still functional. Taken together, we conclude from our findings that most likely GABA_BRs at glutamatergic presynaptic terminals are defective in the Reln^cKO. Moreover, We preclude a possible defect of GABA_BRs at the postsynaptic site because no significant change in the Ca²⁺ amplitude was observed (Fig. 1).

To examine a possible crosstalk between Reelin signaling and GABA_BR function, we prepared OTCs from P0 mice. OTCs were treated with 4-OHT at DIV1-5 to induce Reelin deficiency. Subsequent Reelin knockout, Ca²⁺-recordings were performed around DIV14 (Fig. 3). First, to confirm conditionally induced Reelin knockout, OTCs were fixed at DIV14 and immunostained against Cre-recombinase and Reelin (Fig. S5). While pronounced Reelin staining was seen in wt OTCs, it was absent in Reln^cko OTCs. Bath application of the GABA_BR agonist baclofen (10 μM) immediately activated GABA_BRs and strongly decreased the Ca²⁺ amplitude by 6-fold in wt OTCs, but to a lesser extent (1.6-fold) in Reln^cKO OTCs (Fig. 3A), suggesting that postsynaptic GABA_BRs were still functioning in Reln^cKO OTCs. However, we do not preclude a postsynaptic functional difference between wt and Reln^cKO. Moreover, baclofen induced a profound reduction of the Ca²⁺ frequency and Ca²⁺ transient half-width in wt OTCs (Fig. 3B and C and Movie S1), but no effect was observed in Reln^cKO OTCs (Fig. 3B, C and Movie S2). This suggests an altered GABA_BR function at presynaptic sites in Reln^cKO, probably due to a lack of trafficking and localization of GABA_BRs at the plasma membrane in Reln^cKO mice. Strikingly, the GABA_BR antagonist CGP35348 (10 μM) did not influence the Ca²⁺ amplitude, neither in wt nor in Reln^cKO OTCs (Fig. 3D). However, blockade of GABA_BRs in wt OTCs dramatically increased the Ca²⁺ signaling frequency and Ca²⁺ transient half-width in wt but not in Reln^cKO OTCs (Fig. 3E and F). Taken together, these results indicate a presynaptic dysfunction of GABA_BRs in the Reln^cKO OTCs.

Fig. 3. GABA_BRs are dysfunctional in Reln^cKO.

To test the functionality of the GABA_BRs at DIV14, OTCs were loaded with OGB-1 AM, spontaneous activity was first recorded and then the GABA_BR agonist baclofen (10 μm) was added. In each experiment, OTCs were prepared from 5 wt and 5 Reln^cKO mice. (A): The application of baclofen immediately activates GABA_BRs and thereby strongly decreases Ca⁺² amplitude in wt OTCs (One-way ANOVA on Ranks followed Tukey’s Test, ***p < 0.001) but to a lesser extent in Reln^cKO OTCs (*p < 0.05). Ca²⁺-frequency (B) and Ca²⁺ transient half-width (C) were significantly decreased in the wt but not in Reln^cKO OTCs (One-way ANOVA on Ranks followed Tukey’s Test, ***p < 0.001). The GABA_BR antagonist CGP35348 (10 μM) did not alter Ca²⁺ amplitudes in both wt and Reln^cKO OTCs (D). The application of CGP35348 significantly enhanced Ca²⁺ frequency (E) and Ca²⁺ transient half-width (F) in wt but not Reln^cKO OTCs. One-way ANOVA on Ranks followed by Dunn’s Test, ***p < 0.001. The number of OTCs analysed is indicated above the box-plot in (D). The data is obtained from 3 independent OTCs preparations.

Presynaptic GABA_BRs are modulated by Reelin signaling

Based on our observation that presynaptic GABA_BRs do not properly function in the absence of Reelin, we wondered whether the function of GABA_BRs might be modulated by canonical Reelin signaling. To address this question, we treated OTCs with RAP, an inhibitor of low-density lipoprotein receptor–related proteins (including VLDL and APOER2; (Herz et al. 1991; Willnow et al. 1996)). Reelin acts via binding to VLDLR and ApoER2 to regulate Dab1 tyrosine phosphorylation (Cooper and Howell 1999; Trommsdorff et al. 1999). We have found that chronic RAP treatment (50 ng/ml, over 5 days) did not affect the Ca²⁺ amplitude in Reln^cKO and wt OTCs (Fig. 4A). However, RAP blockade in wt OTCs enhanced the Ca²⁺ frequency and Ca²⁺ transient half-width to a comparable extent as seen in the Reln^cKO without RAP treatment (Fig. 4B and C). As expected, Reln^cKO OTCs treated with RAP did not exhibit any changes in frequency and Ca²⁺ transient half-width (Fig. 4B and C). These observations suggest that GABA_BRs require crosstalk with Reelin signaling for their proper function.

Fig. 4. Effect of chronic blockade with the LDL-receptor chaperone RAP.

Reln^cKO or wt OTCs chronically treated with RAP (50 μg) were loaded with OGB-1 AM and recorded. In each experiment, OTCs were prepared from 5 wt and 5 Reln^cKO mice. Control wt or Reln^cKO OTCs were treated with H₂O. (A) RAP blockade did not affect Ca²⁺-amplitude in wt and Reln^cKO OTCs. However, RAP significantly increased Ca²⁺-frequency (B) and Ca²⁺-transient half-width (C) in wt but not Reln^cKO OTCs. One-way ANOVA on Ranks followed by Dunn’s Test, ***p < 0.001. The number of OTCs analysed is indicated above the box-plots in (A). The data is obtained from 3 independent OTCs preparations.

Reelin activates members of Src family tyrosine kinases (SFKs) via ApoER2 and VLDLR and Dab1 (Bock and Herz 2003; Cooper and Howell 1999). Therefore, we addressed the question whether inhibition of Src family kinases with the pharmacological blocker PP2 might affect GABA_BR function. Our results show that the application of 1 μM PP2 at DIV14 did not alter the Ca²⁺ amplitude neither in wt nor in Reln^cKO OTCs (Fig. 5A). In turn, the Ca²⁺ frequency and Ca²⁺ transient half-width were increased in wt only but not Reln^cKO OTCs (Fig. 5B and C). To test whether the inhibition of Src family kinases with the pharmacological blocker PP2 might also affect GABA_BR surface expression, OTCs from wt and Reln^cKO mice were treated with 1 μM PP2 and immunostained against GABA_BR1 and GABA_BR2 under non-permeabilized condition to detect their surface expression (Fig. S6). The surface expression of GABA_BR1 and GABA_BR2 was significantly reduced in wt OTC but remained unaltered in the Reln^cKO. Together, these results suggest that Reelin signaling through Src is important to maintain GABA_BRs function and surface expression. In line with these findings, a previous study had shown that the expression of SNAP25, a protein that is involved in the control of transmitter release, was decreased in the hippocampus of reeler mutant mice, but not altered in ApoER2-, VLDLR-, or Dab1-deficient mice (Hellwig et al. 2011). Moreover, GABA_BRs were found to be downregulated in the reeler mutant (Cremer et al. 2011). Taken together, our findings suggest that presynaptic GABA_BR function is modulated through the Reelin signaling cascade.

Fig. 5. Effect of acute Src kinase inhibition on wt and Reln^cKO Ca²⁺ activity.

Acute application of PP2 (1 μM) did not affect Ca²⁺-amplitude in both wt and Reln^cKO OTCs (A). However, PP2 application enhances Ca²⁺ frequency (B) and Ca²⁺ transient half-width (C) in wt but not Reln^cKO OTCs (B). In each experiment, OTCs were prepared from 5 wt and 5 Reln^cKO mice. One-way ANOVA on Ranks followed by Dunn’s Test, ***p < 0.001. The number of OTCs analysed is indicated above the box-plots. The data is obtained from 3 independent OTCs preparations.

Reelin signaling alters GABA_BR2 cell surface expression and its phosphorylation at S783 and S892

Previous studies have demonstrated that GABA_BRs do not undergo agonist-induced internalization (Terunuma, Pangalos et al. 2010), and that GABA_BR cytoplasmic domains contain multiple phosphorylation sites (Couve et al. 2002; Fairfax et al. 2004). To find out whether Reelin deficiency might affect GABA_BR cell surface expression, we performed immunofluorescence staining of GABA_BR1 and GABA_BR2 with specific antibodies that recognize only the extracellular N-terminal domain of the receptor subunits (Fig. 6). Mean fluorescence intensity analysis revealed that both GABA_BR1 and GABA_BR2 surface staining was significantly reduced in non-permeabilized cells of Reln^cKO acute slices in comparison to wt (Fig. 6), suggesting a role of Reelin in regulating GABA_BR cell surface expression. Next, we investigated whether secreted Reelin from wt OTCs might rescue the reduction of GABA_BR1 and GABA_BR2 surface expression in the Reln^cKO. To address this question, we co-cultured OTCs from wt + Reln^cKO (Fig. S7). Our results show that Reelin secreted from wt OTCs into the incubation medium maintained GABA_BR1 and GABA_BR2 expression in Reln^cKO at a level comparable to the wt control group.

Fig. 6. Surface GABA_BR1 and GABA_BR2 expression are reduced in the Reln^cKO slices.

Acute slices (P14) from wt and Reln^cKO mice were immunostained against total (permeabilized) and surface (non-permeabilized) GABA_BR1 and GABA_BR2 expression. In each experiment, slices were prepared from 5 wt and 5 RelncKO mice. Confocal image example at 40 x magnification from layers II/III somatosensory cortices of wt (A) and Reln^cKO (B) acute slices stained against GABA_BR1 or GABA_BR2 (C and D) under permeabilization condition to detect total receptor expression. Confocal image example of wt (E) and Reln^cKO (F) acute slices stained against GABA_BR1 or GABA_BR2 (G and H) under non-permeabilization condition for receptor surface expression. (I) The box-plots represent values of average pixel intensity of total (permeabilized) and surface (non-permeabilized) GABA_BR1 expression. (J) The box-plots represent values of average pixel intensity of total (permeabilized) and surface (non-permeabilized) GABA_BR2 expression. One-way ANOVA on Ranks followed by Dunn’s Test, ***p < 0.001. The number of acute slices analysed is indicated above the box-plots. The data is obtained from 3 ex-vivo acute slice preparations. Scale bar: 10 μm.

Finally, we performed western blotting, to assess the amount of expressed GABA_BR1 and GABA_BR2 protein, which however, did not differ significantly between wt and Reln^cKO, though the amount of both receptors showed a tendency to decrease in the Reln^cKO (Fig. 7A and B). To find out whether Reelin deficiency might affect GABA_BR phosphorylation, we analysed cortical lysates from wt and Reln^cKO P14 mice by western blotting with antibodies against GABA_BR phosphorylation sites. GABA_BR2 phosphorylation at Ser892 is mediated by cyclic AMP (cAMP)–dependent protein kinase (PKA) and stabilizes cell surface expression and coupling to G-proteins of the receptors (Couve et al. 2002). Western blotting revealed that in Reelin deficient tissue, S892 phosphorylation of GABA_BR2 was significantly reduced when compared to wt (Fig. 7C), suggesting that Reelin signaling is important to stabilize GABA_BR2 function at the cell surface. Another GABA_BR2 phosphorylation site at S783 is required for degradation, and its phosphorylation is mediated by 5ʹ-AMP-dependent protein kinase (AMPK) (Terunuma, Vargas et al. 2010). GABA_BR2 recycling and degradation is controlled via phosphorylation site S783 (Gassmann and Bettler 2012). Western blot analysis revealed that GABA_BR2 phosphorylation at S783 in Reln^cKO mice was in creased when compared to wt, and that phosphorylated GABA_BR2s were degraded in Reln^cKO mice only (Fig. 7D). Taken together, these findings indicate that Reelin is required to stabilize the cell surface expression of functional GABA_BR2 receptors, probably by inhibiting receptor degradation.

Fig. 7. Reelin signaling regulates GABAB-receptor phosphorylation.

Cortical lysates from P14 wt and Reln^cKO mice were analysed by western blot (n = 6 animals from each group; The data is obtained from 3 different tissue explants). Total GABA_BR1 (A) and GABA_BR2 (B) were unaltered in Reln^cKO in comparison to wt control. (C): Phosphorylation of GABA_BR2 at S892 was significantly reduced in Reln^cKO (T-test, ***p < 0.001), whereas phosphorylation at S783 was significantly increased (D) when compared to wt (T-test, ***p < 0.001), and appearance of several bands suggests proteolysis of the receptor (Gassmann and Bettler 2012; Maier et al. 2010).

GABA_BR intracellular signaling via Gαi/o proteins is affected in Reln^cKO mice

To further investigate a potential crosstalk between Reelin, GABA_BR signaling and G-protein coupled receptors, we recorded OTCs in the presence of pertussis toxin (PTX). Stimulation of GABA_BRs with baclofen inhibits adenylyl cyclase (Xu and Wojcik 1986), which in turn can be blocked by pertussis toxin (PTX), known to inhibit G-protein function (Karbon and Enna 1985; Nishikawa and Kuriyama 1989). Application of PTX at DIV14 enhanced the Ca²⁺ amplitude, frequency and Ca²⁺ transient half-width in wt but not in Reln^cKO OTCs (Fig. S8). These results further support the interpretation that Ca²⁺-signalling is unaltered by the treatment in Reln^cKO OTCs because GABA_BR receptors are degraded in the absence of Reelin. It also confirms that Reelin is important to maintain proper GABA_BR signaling.

Discussion

In the present study, we characterized a novel role of Reelin in early postnatal cortical function. We found that postnatally induced Reelin deficiency caused abnormal Ca²⁺ spike frequency and Ca²⁺ transient half-width in cortical neurons. RAP blockade of Reelin binding to its receptors ApoER2 and VLDLR mimicked this effect. There are no specific selective antagonists for ApoER2 and VLDLR, therefore we used RAP to broadly block LRP family members. Co-culture of Reln^cKO OTCs with Reelin secreting wt OTCs rescued deficient network activity in Reln^cKO OTCs. While the GABA_BR agonist baclofen failed to activate GABA_BRs, the GABA_BR antagonist CGP35348 failed to inhibit GABA_BRs in Reln^cKO OTCs, indicating deactivation and/or degradation of GABA_BRs in Reln^cKO OTCs. Immunostaining of non-permeabilized cells with antibodies against GABA_BR1 and GABA_BR2 extracellular domains was significantly reduced in Reln^cKO slices in line with our pharmacological experiments, and confirming a previous study that reported on a downregulation of GABA_BRs in the reeler mutant (Cremer et al. 2011). When analysing cortical tissue of conditionally induced Reelin deficient mice by western blotting, we found a decreased GABA_BR2 phosphorylation at S892 and an increased phosphorylation at S783, both indicative of GABA_BR2 degradation (Gassmann and Bettler 2012), and proteolytic processing of GABA_BR2. Accordingly, in Reln^cKO slices we observed a decrease in surface expression of GABA_BR1 (approx. 35% decrease) and GABA_BR2 (approx. 39% decrease) respectively. The enhanced Ca⁺² frequency in the Reln^cKO slices can be attributed to the degradation (downregulation) of the GABA_BRs at the presynaptic sites. The main function of presynaptic GABA_BRs at the presynaptic site is to inhibit calcium channels, which results in inhibition of neurotransmitter release (Benarroch 2012). Together, our findings suggest a novel role for Reelin in controlling cortical neuronal network maturation, by regulating Ca²⁺ spiking in glutamatergic neurons via presynaptic GABA_BRs.

Prolonged NMDA-R stimulation triggered both the endocytosis of GABA_BRs (Terunuma, Vargas et al. 2010) and activation of PP2A. PP2A in turn favors dephosphorylation of S783 in GABABR2 and redirection of the endocytosed pool from recycling to degradation. However, Surprisingly, despite the observed dephosphorylation of S783 in GABAergic neurons, no increase in the lysosomal degradation of GABA_BRs has been observed by (Padgett et al. 2012). The identification of S783 on the GABA_BR2 subunit as an AMPK substrate points to a link between the induction of ischemia and increased phosphorylation of GABA_BRs in the hippocampus (Kuramoto et al. 2007). The increase in GABAB receptor phosphorylation, evident during ischemia, could provide a potentially neuroprotective mechanism that may limit neuronal exposure to excitotoxicity. In absence of Reelin, we also observed enhanced Ca2+ signalling in the Reln^cKO and increased phosphorylation at Ser783, which might similarly be important to protect neurons from excitotoxicity.

By performing immunohistochemical staining with antibodies against layer-specific markers, we confirm that cortical neurons in postnatal Reln^cKO did not alter their characteristic layer specific marker expression in the absence of Reelin (Boyle et al. 2011). There are two populations of Reelin expressing neurons in the neocortex. At early embryonic stages, Reelin is secreted by Cajal-Retzius (CR) cells. The number of CR cells declines postnatally, and these cells disappear almost completely around P14 by undergoing selective cell death through apoptosis (Anstötz et al. 2014). Moreover, almost all GABAergic interneurons in the neonatal cortex express Reelin, suggesting that in the current study Reelin was mainly expressed by interneurons.

A role of Reelin in modulating glutamatergic activity has been previously reported (Chen et al. 2005). Reelin-haploinsufficiency has been shown to disrupt the developmental course of GABA excitatory/inhibitory balance (Bouamrane et al. 2016). In the current experiments, we studied spontaneous neuronal calcium activity in the Reln^cKO compared to wt littermates. While the Ca²⁺ amplitude was unaltered in RelncKO, the frequency of Ca²⁺ spiking and Ca²⁺ transient half-width was dramatically increased in Reln^cKO when compared to wt. The Ca²⁺ spike half-width narrows around the second postnatal week (14DIV or P14) and this observation has been previously attributed to an increase in cell size and maturation of ion channels (Gold et al. 2007). Thus, the increased Ca²⁺ spikes half-width in Reln^cKO may suggest an immature ion channels condition in these mice. Along this line, related observations were reported in the reeler hippocampus. Thus, using electron microscopy, it has been shown that the number of presynaptic vesicles was significantly increased in hippocampal CA1 synapses of reeler mutant mice when compared to wt (Hellwig et al. 2011). In contrast, acute Reelin application to dissociated hippocampal neurons enhanced spontaneous neurotransmitter release without affecting properties of evoked neurotransmission (Bal et al. 2013). The discrepancy here may be due to the different models and methodologies used. For instance, acute application of 5 nM Reelin (Bal et al. 2013) versus analysis of the reeler mutant (Hellwig et al. 2011) or the Reln^cKO (present study). In neocortical neurons, neurotransmitter release is controlled by GABA_BR function (see Introduction). Moreover, perturbations of GABAB receptor signaling during development may shift the excitatory/inhibitory balance, culminating in various neurological dysfunctions, including typical and atypical absence seizures (Han et al. 2012). To find out whether Reelin function might interfere with GABA_BR function, we manipulated GABA_BR physiology in Reln^cKO tissue using pharmacological agents. Our observation that a GABA_BR agonist failed to activate GABA_BRs and an antagonist failed to inhibit GABA_BRs in the Reln^cKO, points to a novel function of Reelin as a modulator of GABA_BRs during early cortical function. Taken together, our data suggest that in the absence of Reelin, GABA_BRs are downregulated, as shown by GABA_BR cell surface immunolabelling experiments, and by western blotting that revealed an altered GABA_BR phosphorylation status as well as proteolytic receptor processing.

Next, we investigated possible mechanisms underlying cross-talk of Reelin and GABA_BR signaling. Binding of Reelin to its receptors induces phosphorylation of Dab1 protein by Src kinase (Bock and Herz 2003). Blocking of wt OTCs with a Src inhibitor increased the Ca²⁺ signaling frequency to a comparable level as in Reln^cKO. Moreover, the Src inhibitor decreased GABA_BRs surface expression in wt slices but it did not affect GABA_BRs surface expression in the Reln^cKO. These observations suggest that Reelin signaling through Src is important to maintain GABA_BRs function. Src may directly interact with the α subunits of G-proteins (Ma et al. 2000) and recently it has been shown that Src/Gαo interactions provide evidence for a novel type of cross-talk between a non-receptor tyrosine kinase stimulated by Reelin and heterologous G protein-coupled receptors (GPCR) (Cho et al. 2015). Since GABA_BRs, are members of the G-protein coupled receptors (GPCRs), it is very well possible that metabotropic GABA_BRs are modulated by Reelin through Src-kinase.

Stimulation of GABA_BRs with baclofen can inhibit both basal, and forskolin stimulated adenylyl cyclase (Xu and Wojcik 1986), which in turn can be blocked by pertussis toxin (PTX), reflecting the involvement of the Giα- and Goα- subunit of the G-protein (Karbon and Enna 1985; Nishikawa and Kuriyama 1989). Moreover, immunoprecipitation assays disclosed the possibility that Reelin might increase the active forms of both Src and Gαo and promote their direct association (Cho et al. 2015). Thus, our finding that PTX increases Ca²⁺ frequency in wt OTCs but not in Reln^cKO, indicates the possibility of a disrupted Src-Gαi/o interaction in the absence of Reelin. For instance, based on our experiments it appears more likely that reduced cell surface expression of GABA_BR is the reason why Ca²⁺ signalling is unaltered in Reln^cKO OTCs after PTX application.

Functional deficits associated with decreased Reelin expression have often been attributed to the well-known developmental neuronal migration defects in Reelin deficient mice (Eastwood and Harrison 2003). Moreover, both Reelin deficiency and cortical network dysrhythmias caused by aberrant interneuron activity were also discussed as possible factors contributing to cognitive and behavioural deficits in Alzheimer disease (AD) (Palop and Mucke 2016; Xiao et al. 2017). Besides Reelin, also different ApoE isoforms act as lipoprotein receptor ligands and are known to influence AD pathogenesis, with ApoE4 being the most important genetic risk factor for AD (Huang et al. 2017). GABA_BRs were repeatedly shown to be implicated in synaptic plasticity (Davies et al. 1991; Mott and Lewis 1991). Until recently, it remained unclear whether GABA_BRs can influence neuronal plasticity through GABA_BRs signaling. It has been demonstrated that G-protein mediated signaling through GABA_BRs delays the recruitment of synaptic vesicles during sustained activity and after short-term depression (Sakaba and Neher 2003). This delay occurs through a reduction of cAMP, which in turn blocks the stimulatory effect of the increased Ca²⁺ concentration on vesicle recruitment (Sakaba and Neher 2003). In conclusion, we provide evidence for a novel role of Reelin signaling in controlling early neuronal network activity by regulating neurotransmitter release through GABA_BRs.

Supplementary Material

figures S1-S8

Fig. S1. Expression of neocortical markers in wt and Reln^cKO mice.

Fig. S2. Effect of chronic 4-OHT treatment on Ca²⁺ spontaneous activity in OTCs

Fig. S3. Immunolabelling of GCaMP6s transfected cells with GAD-65.

Fig. S4. Effect of glutamatergic transmission blockade on spontaneous activity in Reln^cKO mice.

Fig. S5. Cre recombinase and Reelin immunostaining from DIV14 4-OHT treated OTCs.

Fig. S6. PP2 reduces surface GABA_BR1 and GABA_BR2 expression in wt OTCs.

Fig. S7. Secreted Reelin rescues surface expression of GABA_BR2 in Reln^cKO OTCs.

Fig. S8. Effect of acute PTX application on Ca²⁺ signaling in Reln^cKO OTCs.

movie S1

Movie S1. Baclofen inhibits Ca²⁺ spike activity in wt OTCs.

movie S2

Movie S2. Baclofen did not affect Ca²⁺ spike activity in Reln^cKO OTCs.

Abbreviations:

ACSF

artificial cerebrospinal fluid

ANOVA

Analysis of variance

APV

(2R)-amino-5-phosphonopentanoate

CNQX

6-cyano-7-nitroquinoxaline-2,3-dione

Cux-1

cut-like homeobox 1

DIV

Days in vitro

DAB1

Disabled-1

F⁰

baseline fluorescence

GAD65

Glutamate Decarboxylase-65

GABA

gamma-aminobutyric acid

LRPAP Protein = RAP

LDL receptor-related protein-associated protein 1

OGB-1 AM

Oregon Green BAPTA-1 Acetoxymethyl

OTCs

Organotypic cultures

4-OHT

(Z)-4-hydroxytamoxifen

PFA

paraformaldehyde

P

postnatal day

Parv

Parvalbumin

PTX

pertussis toxin

Reln^cKO

Reelin conditional knock-out mice

ROI

region of interest

RRID

research resource identifier

TBS

Tris buffered Saline

WFS

Wolframin 1

WT

Reln^flox/flox wildtype