Different activity patterns control various stages of Reelin synthesis in the developing neocortex

Abstract Reelin is a large extracellular matrix protein abundantly expressed in the developing neocortex of mammals. During embryonic and early postnatal stages in mice, Reelin is secreted by a transient neuronal population, the Cajal–Retzius neurons (CRs), and is mostly known to insure the inside-out migration of neurons and the formation of cortical layers. During the first 2 postnatal weeks, CRs disappear from the neocortex and a subpopulation of GABAergic neurons takes over the expression of Reelin, albeit in lesser amounts. Although Reelin expression requires a tight regulation in a time- and cell-type specific manner, the mechanisms regulating the expression and secretion of this protein are poorly understood. In this study, we establish a cell-type specific profile of Reelin expression in the marginal zone of mice neocortex during the first 3 postnatal weeks. We then investigate whether electrical activity plays a role in the regulation of Reelin synthesis and/or secretion by cortical neurons during the early postnatal period. We show that increased electrical activity promotes the transcription of reelin via the brain-derived neurotrophic factor/TrkB pathway, but does not affect its translation or secretion. We further demonstrate that silencing the neuronal network promotes the translation of Reelin without affecting the transcription or secretion. We conclude that different patterns of activity control various stages of Reelin synthesis, whereas its secretion seems to be constitutive.


Introduction
Reelin is a large secreted protein highly expressed by Cajal-Retzius neurons (CRs) and GABAergic neurons in the developing neocortex. Reelin fulfills a variety of developmental processes, including regulation of neuronal migration and formation of cortical layers in an inside-out manner (Caviness and Rakic 1978;D'Arcangelo et al. 1995;Ogawa et al. 1995), neurite outgrowth and branching (Del Río et al. 1997;Niu et al. 2004Niu et al. , 2008Borrell et al. 2007), and maturation of neurotransmitter receptors (Groc et al. 2007;Hamad et al. 2021), before being downregulated (Schiffmann et al. 1997;Ringstedt et al. 1998). In line with this variety of developmental functions, dysregulation of Reelin levels is associated with multiple neurodevelopmental diseases. Mice heterozygous for the reelin gene (Rakic and Caviness 1995) are used as a model for schizophrenia (Ogawa et al. 1995;Tueting et al. 1999;Tsuneura et al. 2021) and in humans, patients suffering from autistic spectrum disorder or schizophrenia present decreased levels of reelin mRNA (Impagnatiello et al. 1998;Fatemi et al. 2005;Eastwood and Harrison 2006;Kushima et al. 2017). Inversely, experimental data from rodents and humans indicate that an aberrant survival of CRs and thus a failure to downregulate the overall level of Reelin is associated with epilepsy (Eriksson et al. 2001;Blümcke et al. 2002;Savell et al. 2019;Riva et al. 2023).
The Reelin protein presents 2 cleavage sites, generating 6 predicted proteins (Lambert de Rouvroit et al. 1999;Ignatova et al. 2004). While many studies have focused on the role of proteolytic cleavage on Reelin protein activity (Trotter et al. 2014;Ogino et al. 2017;Okugawa et al. 2020), little is known about the mechanisms regulating its expression. To our knowledge, only one study has investigated whether Reelin is expressed in a constitutive manner, or whether environmental factors control its rate of expression in the developing neocortex. This early study demonstrates that brain-derived neurotrophic factor (BDNF) promotes the downregulation of Reelin expression in embryonic CRs (Ringstedt et al. 1998). More recently, our lab and others have revealed the role of electrical activity in the cell death of CRs (Blanquie et al. 2016;Riva et al. 2019), providing an additional mechanism leading to the overall developmental downregulation of Reelin.
Here, we investigated whether neuronal activity regulates Reelin synthesis and secretion in developing neocortical neurons. In the first part of the work, we quantified the density of Reelin-expressing neuronal subtypes in the marginal zone of the neocortex during the first 3 weeks in vivo. In accordance with previous reports (Ogawa et al. 1995;Alcántara et al. 1998;Chowdhury et al. 2010;Ma et al. 2014;Ledonne et al. 2016), we show that CRs represent the main source of Reelin at birth and that between postnatal day (P) 0 and P5, the density of Reelinpositive neurons drops and the proportion of GABAergic neurons expressing Reelin rises 3-fold. In the second part of the work, we used neuronal cultures to investigate the role of electrical activity on the regulation of Reelin synthesis. We show that the GABA A receptor antagonist Gabazine (Gbz) increases neuronal activity and promotes reelin transcription via the BDNF/TrkB pathway, but that this effect does not translate at the protein level. Inversely, silencing the neuronal network does not affect reelin transcription but promotes Reelin protein translation. In addition, for both activity patterns, we observed that the secretion rate of Reelin follows the rate of synthesis. These results demonstrate that different activity patterns have the ability to affect Reelin expression at specific stages of its synthesis, whereas the secretion of Reelin appears to be constitutive.

Animals
All experiments were conducted in accordance with National and European (2010/63/EU) laws for the use of animals in research. Neonatal C57BL/6NRj mice, born and housed in the local animal facility, were used for neuronal cultures. For quantification of Reelin-positive cells in slices, GAD67-GFP knock-in mice were used. These mice are heterozygous for the expression of GFP in cells positive for GAD67 (Tamamaki et al. 2003).

Dissociated cortical cell cultures
C57BL/6 mice were decapitated at postnatal day (P) 0, brains were removed from the skull and quickly transferred into icecold Ca 2+ -and Mg 2+ -free HBSS (Gibco, Thermo Fisher Scientific) supplemented with 50 units/mL penicillin/streptomycin (Gibco), 11 mg/mL sodium pyruvate (Sigma-Aldrich), 0.1% glucose (Sigma-Aldrich) and 10 mM HEPES (Sigma-Aldrich). The hindbrain was removed, the 2 cerebral hemispheres were separated by a sagittal cut and the neocortex was isolated from subcortical structures. Cortices were then washed 3 times with HBSS and immersed in 0.05% trypsin/EDTA at 37 • C. After 20 minutes (min), 200 U/mL of DNase (Sigma-Aldrich) were added to the solution. After 5 min, cells were rinsed with a plating medium composed of Minimal Essential Medium (Gibco) supplemented with 10% horse serum and 0.6% glucose. Mechanical dissociation of neocortical cells was performed by trituration through fire-polished pipettes. Once the tissue was dissociated, the solution was filtered using a cell strainer (Greiner Bio-One), living cells were counted using trypan blue staining (Sigma-Aldrich) and neurons were plated on coverslips or on 12-well plates (Greiner CELLSTAR) coated with polyornithine at an initial density of 1,500 cells/mm 2 , or on microelectrode arrays (MEA) coated with polyethyleneimine (0.05% in borate-buffered solution, Sigma-Aldrich) at an initial density of 4,500 cells/mm 2 . After 30 min, the plating medium was replaced by a medium containing Neurobasal medium A (Gibco) supplemented with 2% B27 (Gibco) and 1 mM L-glutamine (Gibco). Cells were incubated at 37 • C in 95% air and 5% CO 2 for 7 days in vitro (DIV).

Pharmacology
At DIV 2-3, cultures were treated with 5 μM Ara-C. At DIV 7, 1 to 10 μL of drugs were added to cell cultures. For western blot experiments, medium was exchanged with phenol red-free medium the day of treatment. The following final concentrations were applied:

Fluorescent in situ hybridization
Fluorescent in situ hybridization (FISH) was performed using the ViewRNA ISH Cell Assay Kit (Thermo Fisher Scientific, Cat No: QVC0001) with probes targeting the coding sequences of c-fos (VB1-11874-VC), reelin (VB6-3198420-VC), and gad1 (VB4-19846-VC). The manufacturer's protocol was applied with the following modifications: cells were fixed in formaldehyde 4% (ROTIHistofix, Carl Roth) for 10 min, dehydrated in 50%, 70%, and 100% EtOH for 15 min, and stored in 100% EtOH at 4 • C. The protease solution was used at a concentration of 1:6,000 and probes were used at a concentration of 1:800.

Western blot
Cells were lysed in lysis buffer (Roche Applied Science, Complete), 1/4 volume of 4x lithium dodecyl sulfate (LDS) buffer (Thermo Fisher Scientific, Cat No: NP0007) and 1/10 volume of sample reducing agent 10x (50 mM DTT) were added and samples were boiled at 95 • C for 3 min. Culture medium was concentrated using a 30-kDa cut-off filter (Millipore, Amicon Ultra-2 Centrifugal Filter Unit, Cat No: UFC203024), 1/100 volume of protease inhibitor (Thermofisher Scientific, Halt Protease-Inhibitor-Coktail x100, Cat No: 78440) was added and a Bicinchoninic acid (BCA) test was performed before adding 4x LDS buffer and sample reducing agent. Total protein concentrations were equalized using 1x LDS. Medium samples were boiled at 95 • C for 3 min. Protein samples and molecular mass markers (Biorad, were separated by Sodium dodecyl-sulfate polyacrylamide gel electrophoresis (SDS-PAGE) on 3-8% tris-acetate gels (Invitrogen) at 150 V for 1.5 h (cell lysate) or 90 V for 2 h (culture medium). Proteins were transferred to PVDF membranes (Roth) and electro-transferred to 0.45 μm nitrocellulose membranes (Amersham, GE Healthcare, Germany) for 1.5 h. After 1 h blocking at room temperature with 4% (w/v) milk in Tris-buffered saline (50 mM Tris, 150 mM NaCl, pH 7.6) containing 0.1% (v/v) Tween 20 (TBS-T), membranes were incubated overnight at 4 • C in mouse anti-Reelin G10 (MAB5364, Millipore 1:1,000) or goat anti-GAPDH (AF5718, R&D System, 1:5,000). After 3 10-min washes in TBS-T, membranes were incubated with an horseradish peroxidase (HRP)-conjugated secondary antibodies (Jackson Immunoresearch; 1:10,000) diluted in TBS-T with 4% milk for 1 h. Enhanced chemiluminescence (ECL) detection was performed in a Chemi-Doc XRS+ system and ImageLab software (Biorad). The densities of protein bands were quantified in ImageJ. The bands of the cellular fraction were normalized to the ones of GAPDH while the bands of the culture medium were normalized to a pre-stained marker (Thermo Fisher Scientific, Cat No: 26612) added to the samples.

RNA extraction and qPCR
RNA was isolated using the RNeasy Mini Kit (Qiagen). mRNA was reverse-transcribed using the Transcriptor High Fidelity cDNA Synthesis Kit (Roche Applied Science). qRT-PCR was performed using the iTaq Universal SYBR Green Supermix (Bio-Rad) in a StepOne Plus Real-Time PCR System (Thermo Fisher Scientific). Primer sequences (all in 5 -3 orientation) of target genes and probes are as follows: GAPDH (ATGCCAGT-GAGCTTCCCGTTCAG and CATCACTGCCACCCAGAAGACTG); reelin (CCAGTCTCATGAAGAACTGCAC and GCTTGCGCATGC-TAGTAACAC), bdnf (TGCAGGGGCATAGACAAAAGG and CTTAT-GAATCGCCAGCCAATTCTC). The qRT-PCR crossing points were used for relative quantification based on the Ct method and GAPDH was used as a reference gene.

Microelectrode array recordings
Extracellular electrical recordings from cortical neurons were performed on microelectrode arrays (MEAs) containing 120 planar titanium nitrite electrodes with 4 internal references (120tMEA100/30iR-ITO-gr, Multi Channel Systems, Harvard Bioscience). MEAs had an electrode diameter of 30 μm and an interelectrode spacing of 100 μm. Signals from 120 recording electrodes were recorded with MC_Rack 4.6 software in a MEA 2100 system (Multi Channel Systems) at a sampling rate of 50 kHz and high-pass filtered at 200 Hz. Spikes were detected using a threshold-based detector set to a threshold of 7× the SD of the noise level (MC_Rack, Multi Channel Systems). Electrophysiological recordings were performed for 10 min in culture medium maintained at 37 • C with a temperature controller (TC02, Multi Channel Systems) and in a humidified atmosphere with 5% CO 2 controlled with a gas mixer (CO 2 -Controller 2000, PeCon). Spike datasets before and after pharmacological treatment were merged with MC_DataTool (Multi Channel Systems) and sorted with Off line Sorter (Plexon Inc., Dallas, Texas, United States) using K-means scan (KMS) method with a unit range of 1-4, followed by manual curation. Sorted electrophysiological units were then imported into Matlab 9.8 (The MathWorks Inc., Natick, Massachusetts, United States) for analysis using a custom written routine. Only units with at least 1 spike/min were considered as active units. The network firing rate was calculated as the sum of all neuron firing rates (i.e. the number of spikes recorded across all electrodes per second). Representative filtered traces were converted to HDF5 format with Multi Channel DataManager (Multi Channel Systems) and displayed in Matlab 9.8 (The MathWorks Inc.).

Fluorescence microscopy and analysis
Immunostaining pictures of slices were taken with an IX81 epif luorescent microscope (Olympus) using a 20x objective. For image analysis, 2 hemispheres per animal with 4 fields of view each were analyzed. Based on DAPI staining, one field of view (FOV) was chosen per hemisphere and per cortical area, i.e. primary motor cortex (M1), primary somatosensory cortex (S1), primary auditory cortex (Au1), and primary visual cortex (V1) (Franklin et al. 2008;Paxinos et al. 2020). The marginal zone was identified based on the DAPI staining as the low-density layer located between the pial surface and the layers II-IV (see Fig. 1A1). Measurement of the marginal zone area and cell quantification were performed manually using ImageJ software. All pictures were analyzed in blind conditions. FISH pictures were taken with a TCS SP5 confocal microscope (Leica) using a 40x objective. Z-stacks with 5 frames and a z-size of 0.75 μm. For cFOS analysis, 3 FOVs per culture and per condition were taken based on DAPI staining. For the quantification of neuronal subtypes, 6 FOVs per culture were taken. Images were analyzed with a custom-written routine in ImageJ. All pictures were taken and analyzed in blind conditions.

Statistical analysis
All statistical tests were performed using GraphPad Prism 9.3. Normality of sample distributions was tested with Shapiro-Wilk test. When the samples were normally distributed, parametric 2-tailed paired or unpaired Student t-test was applied for comparison of 2 experimental groups; if more than 2 groups were compared, 1way or 2-way Analysis of Variance (ANOVA) was performed and differences between groups were analyzed by a Tukey post-hoc test. If the data were not normally distributed, Wilcoxon test was performed for paired direct comparison. For MEA recording analysis, outlier testing was performed to remove units with extreme values of firing rate; rout method (Q = 5%) was applied after logtransforming the sample distributions. Normally-distributed data are presented as mean ± standard error of the mean (SEM) and non-normally distributed data as median with interquartile range (IQR). Data significance was considered at P-values < 0.05.

Results
The density of neurons expressing Reelin declines in the marginal zone of the neocortex during the first 3 postnatal weeks and the source of Reelin switches from CRs to GABAergic neurons.
We further investigated whether the decrease in the number of Reelin-positive cells varies across functional cortical areas (A1) Coronal sections from P0 to P20 GAD67-GFP mice were stained against Reelin, allowing the quantification of Reelin-positive GABAergic neurons (labeled in green and magenta) and CRs (labeled in magenta only). Representative images were taken with a TCS SP5 confocal microscope (Leica) using a 40x objective. A maximum projection was performed on 11 z-stacks acquired with a z-resolution of 1 μm. Scale bar is 20 μm. (A2) Quantification of the number of Reelin-positive cells in the marginal zone shows that the density of cells expressing Reelin drops during the postnatal period. At birth, CRs represent the major source of Reelin. At P5/6, the density of CRs has dramatically decreased and the proportion of GABAergic neurons expressing Reelin reaches almost half of the Reelinexpressing cells. This proportion keeps increasing until P19/20, where GABAergic neurons are the only neuronal type positive for Reelin. (B) Although the number of CRs is higher in S1 compared to M1 and V1, the dynamics of Reelin expression follows the same trend throughout the cortex, and the proportion of CRs to GABAergic neurons is similar during the first 3 postnatal weeks in all 4 cortical areas. The marginal zone/layer I is located between the 2 yellow, dotted lines. M1: primary motor cortex, S1: primary somatosensory cortex, Aud1: primary auditory cortex, and V1: primary visual cortex.
( Fig. 1B and Supplementary Fig. S1). Although the density of CRs is higher in the S1 (772.0 ± 119.8 CRs/mm 2 in S1) than in the other cortical areas at birth (M1: 481.7 ± 89.5 CRs/mm 2 , 2-way ANOVA, P < 0.005; V1: 488.5 ± 67.5 CRs/mm 2 , P < 0.01), the drop in the density of Reelin-positive cells occurs concurrently across all cortical areas investigated, and the proportion CRs to GABAergic neurons remains similar. These results indicate that one or several common factor(s) similarly regulate(s) the decrease in Reelin expression across the neocortex.

Gbz application increases the level of activity in immature cortical cultures and promotes reelin transcription
To investigate whether electrical activity regulates reelin mRNA expression at a single cell level, we used primary cortical cultures and pharmacologically modulated neuronal firing. The large majority of Reelin-positive neurons in cultures undergo cell death until DIV 9 (Blanquie et al. 2016), and firing events can be first detected with the establishment of synaptic contacts towards the end of the first week in culture (Sun et al. 2010;Verstraelen et al. 2014). Therefore, to reveal activity-dependent effects on gene transcription, experiments were performed at DIV 7. Neuronal cultures deriving from GAD67-GFP mice did not exhibit f luorescence signal at this early stage and immunostainings against GAD67 did not allow the post-hoc identification of GABAergic neurons. To quantify the relative proportion of Reelin-positive neurons, FISH was performed against reelin and gad 1. At DIV 7, 63.8 ± 7% of reelin-positive neurons are GABAergic and 36.2 ± 7% CRs (n = 5 cultures).
The effect of Gbz on the neuronal network was assessed by performing MEA recordings before and after treatment ( Supplementary Fig. S2A). As a control, the same volume of vehicle (1 μL H 2 O) was applied in sister neuronal cultures. As shown in Fig. 2A and B and Supplementary Fig. 2B and C, before treatment, neuronal cultures display sparse neuronal activity with a median number of active neurons of 4 (IQR = 2-7, n = 27 cultures) and a median network firing rate of 1.76 Hz (IQR = 0.62-3.41). After 6 h of H 2 O application, the number of active neurons ( Fig. 2A-2, H 2 O: 3.5 active neurons, IQR = 2-8.75, Wilcoxon test, P > 0.05, n = 12) and their network and single neuron firing rates ( Fig. 2A-3 (Fig. 2B-2, 7 active neurons, IQR = 2-14, Wilcoxon test, P < 0.001, n = 15) and the network displays only a slight but non-significant increase in the firing rate ( Fig. 2B-3, Gbz: 0.24, IQR = 0.06-0.79, Wilcoxon test, P < 0.001, n = 73). Spike sorting analysis demonstrates that among the neurons originally active, application of Gbz significantly increases the firing rate ( Supplementary Fig. 2C-1, Ctrl: 0.11 Hz, IQR = 0.04-0.49; Gbz: 0.24 Hz, IQR = 0.06-0.79, Wilcoxon test, P < 0.01, n = 73), but the firing frequency of activated units remains lower compared to the originally active neurons within the same dish, thereby explaining the absence of significant increase in the overall firing rate ( Supplementary Fig. 2C-2, originally active neurons: 0.91 ± 0.17 Hz, activated: 0.1 ± 1.9 × 10 −2 Hz, Student t-test, P < 0.001). Notably, the vast majority of activated neurons by Gbz (60 out of 66) is detected in previously silent channels spatially distributed across the MEA recording area. Therefore, in our model the application of Gbz not only increased the firing rate of already active neurons, but also caused a network-wide disinhibition leading to the activation of previously inactive neurons.
To confirm that the depolarizing effect of Gbz could induce changes on the transcriptional level, we performed a FISH and quantified the level of mRNA for the immediate early gene c-fos. Compared to control conditions (Fig. 2C, 6.7 ± 1.1 c-fos puncta/cell, n = 26 FOV from 4 cultures), Gbz-treated cultures display almost a 3-fold increase in c-fos mRNA (18.2 ± 2.4 c-fos puncta/cell, Student t-test, P < 0.005, n = 27 FOV from 4 cultures), confirming that Gbz application promotes neuronal activity.
We next investigated whether Gbz has any effect on the overall level of reelin mRNA by performing a PCR in untreated versus Gbztreated cultures. Upon Gbz application, the level of reelin mRNA is doubled compared to control conditions (Fig. 2D, 194.5 ± 35.8%, Student t-test, P < 0.05, n = 6 samples from 3 cultures per condition). Thus, Gbz application promotes neuronal activity and increases the transcription of reelin mRNA.

Gbz-mediated increase in reelin transcription is mediated by the BDNF/TrkB pathway
In the rodent embryonic cortex, BDNF signaling has been shown to downregulate reelin mRNA transcription in CRs without affecting their cell death (Ringstedt et al. 1998). Since Gbz promotes neuronal activity in our culture model and BDNF is released in an activity-dependent manner (Brigadski and Leßmann 2020), we investigated whether and how BDNF signaling pathway could be involved in the observed Gbz-dependent increase in reelin transcription.
We started to assess the effect of Gbz on the levels of bdnf mRNA. As shown in Fig. 3A, Gbz-treated cultures present a 4fold increase in bdnf mRNA levels compared to control cultures (405.3 ± 96.2%, Student t-test, P < 0.05, n = 4 cultures). We then tested whether the Gbz-induced elevation in reelin mRNA and bdnf mRNA are only correlated or causally-related. Neuronal cultures were treated with DHF, an agonist of the BDNF cognate receptor TrkB, and the effect on reelin mRNA expression was assessed. Similar to Gbz, TrkB activation leads to elevated levels of reelin mRNA (Fig. 3B, 164 ± 12.6%, Student t-test, P < 0.01, n = 4 cultures). To confirm that the Gbz-mediated increase in reelin expression is mediated by the BDNF/TrkB pathway, we applied Gbz while concomitantly blocking TrkB with a TrkB-Fc. As depicted in Fig. 3C, blocking the BDNF/TrkB pathway abolishes the effect of Gbz on reelin expression (1-way ANOVA, Gbz: 182 ± 26.1%, P < 0.05; Gbz + TrkB-Fc: 97.4 ± 10.6%, P > 0.05, n = 4 cultures), confirming that Gbz promotes reelin mRNA expression by activating the BDNF/TrkB pathway. We conclude that increasing neuronal activity with the GABA A receptor antagonist Gbz leads to an increased expression and secretion of BDNF, which then promotes reelin transcription upon binding on its receptor TrkB.

Pharmacological increase of neuronal activity leads to uncoupled effects on reelin mRNA and Reelin protein levels
To investigate whether the high amounts of reelin mRNA transcripts found in Gbz-treated cultures are translated into proteins and secreted in the extracellular environment, western blots were performed on the cell lysate and on the culture medium of control and Gbz-treated cultures. In the cellular fraction as well as in the supernatant, cultures treated with Gbz did not present a different amount of Reelin protein compared to untreated cultures ( Fig. 4A and B, cellular fraction: 103.0 ± 16.4%, Student t-test, P > 0.05, n = 5 cultures; culture medium: 115.2 ± 8.9%, Student t-test, P > 0.05, n = 4 cultures), indicating that the extra reelin mRNA molecules  To rule out the possibility that the time window of the experiment is too small to allow reelin mRNA to be translated (Campo et al. 2009), we performed the same set of experiment after 9 h of pharmacological treatment. Again, after 9 h, the amount of Reelin protein is similar in Gbz-treated and untreated cultures, in the cellular fraction as well as in the medium ( Fig. 4C and D, cellular fraction: 98.6 ± 14.5%, Student t-test, P > 0.05, n = 7 cultures; culture medium: 123.7 ± 16.1%, Student t-test, P > 0.05, n = 4 cultures). Thus, neuronal depolarization promotes the transcription of reelin, but the excess reelin mRNA molecules are not translated into proteins.

Silencing the neuronal network does not affect reelin transcription but promotes a transient increase in Reelin translation
We next investigated whether silencing neuronal activity has the opposite effect of Gbz on Reelin expression, i.e. leading to decreased reelin mRNA levels without affecting Reelin protein levels. To answer this question, DIV 7 neuronal cultures were silenced with the sodium channel blocker TTX for 6 h and levels of reelin mRNA and Reelin proteins were quantified.
We next performed a PCR and quantified the levels of reelin mRNA in the 2 conditions. We found a similar level of reelin mRNA in silenced cultures and control cultures (Fig. 5C, 112.5 ± 17.4% of control, Student t-test, P > 0.50, n = 4 cultures). Thus, silencing the neuronal network does not affect reelin transcription.
To test whether neuronal silencing has any effect on the expression and/or secretion of the Reelin protein, neuronal cultures were treated with either TTX or H 2 O, and the amount of protein was quantified in the cellular fraction and in the supernatant. After 6 h, the mean amount of Reelin protein in the cellular fraction increases by 78.6 ± 31.3% compared to untreated cultures (Fig. 6A, Student t-test, P < 0.05, n = 5 cultures), whereas it remains unchanged in the supernatant (Fig. 6B, 100.5 ± 14.4%, Student t-test, P > 0.05, n = 5 cultures). We next performed a western blot after a 9-h treatment period. After 9 h of TTX treatment, the level of Reelin protein in the cellular fraction is back to control levels (Fig. 6C, 99.8 ± 14.2%, Student t-test, P < 0.01, n = 5 cultures), whereas the level of Reelin protein in the supernatant is doubled (Fig. 6D, 212.7 ± 34.5%, Student t-test, P < 0.05, n = 5 cultures). Thus, the extra Reelin proteins present in the cellular fraction after 6 h of pharmacological treatment are secreted in the medium after 9 h of TTX treatment, showing that the rate of secretion of Reelin strictly follows its rate of translation. Altogether, these results demonstrate that neuronal silencing does not affect reelin transcription but promotes Reelin translation. In addition, upon TTX application, the secretion rate of Reelin follows its translation rate, indicating that Reelin secretion is constitutive.

Discussion
Reelin is a secreted protein that plays major developmental roles throughout embryonic and early postnatal stages in the cortex. Here, we established a detailed pattern of Reelin expression in the developing neocortex and used dissociated cortical neurons to investigate the role of electrical activity in the regulation of Reelin protein synthesis or Reelin secretion.
We started to establish the developmental pattern of Reelin expression in the marginal zone of the neocortex. In accordance with previous reports (Schiffmann et al. 1997;Alcántara et al. 1998;Ma et al. 2014), the density of neurons secreting Reelin declines during the postnatal period and the identity of neurons synthesizing Reelin switches from mostly CRs at birth to almost exclusively GABAergic neurons after 3 weeks postnatal (Schiffmann et al. 1997;Alcántara et al. 1998;Chowdhury et al. 2010;Ma et al. 2014;Ledonne et al. 2016). Strikingly, the density of CRs drops between P0 and P5/6. These data reveal that the reduction in the density of CRs in the marginal zone occurs at the same time than the swift developmental upregulation of Reelin in deep layer GABAergic neurons (Schiffmann et al. 1997;Alcántara et al. 1998). We further found that the number of Reelin-positive neurons and the proportion of each neuronal type at each developmental stage is highly conserved across all cortical areas investigated. Altogether, these data show that the pattern of Reelin expression is tightly controlled in a time-and cell type-specific manner, and that the intrinsic and/or environmental factors controlling this developmental pattern are conserved throughout the neocortex.
During the first postnatal weeks, electrical activity inf luences several developmental mechanisms, including the cell death of CRs (Blanquie et al. 2016). To investigate whether the expression of Reelin is also regulated by electrical activity, pharmacological treatments were applied to dissociated cortical cultures, a model that preserves developmental features of neuronal networks and enables the control and monitoring of network activity levels (Weir et al. 2015;Warm et al. 2022). In this in vitro model, spontaneous activity can first be observed at the end of the first week (Sun et al. 2010;Verstraelen et al. 2014). At this early stage, however, neuronal cultures deriving from GAD67-GFP mice did not exhibit f luorescence signal and immunostainings against GAD67 did not allow the post-hoc identification of GABAergic neurons. Only FISH performed against reelin and gad 1 allowed the quantification of the relative proportion of neuronal subtypes and showed that at DIV 7, the majority (64%) of reelin-positive neurons are GABAergic (gad 1-positive) and 36% are CRs (gad 1negative).
At DIV 7, application of the GABA A receptor antagonist Gbz induces neuronal depolarization throughout the network  as quantified by MEA recordings and causes elevated mRNA levels of the immediate early gene c-fos. Interestingly, in TTXtreated cultures, where action potential generation is completely prevented, c-fos levels are unchanged compared to untreated cultures. Next, we measured the level of reelin mRNA in control, Gbz-or TTX-treated cultures. Whereas Gbz leads to a 2-fold increase in reelin mRNA levels, TTX does not affect the level of reelin mRNA transcription. Since neuronal depolarization induced by Gbz leads to elevated levels of c-fos and reelin mRNA whereas TTX does not affect the level of either of these genes, we hypothesize that electrical activity has the ability to affect gene transcription only if a certain level of activity has been reached. We cannot exclude that in vivo the difference in activity levels between physiological conditions and a silenced network would allow a significant change in either c-fos or reelin mRNA levels.
Experiments performed in embryonic BDNF −/− mice have demonstrated that the developmental upregulation of BDNF is causal for reelin mRNA downregulation in neocortical CRs (Ringstedt et al. 1998). Interestingly, although expression and secretion of BDNF is enhanced by electrical activity (Balkowiec and Katz 2000), inducing epileptiform activity for 24 h using either elevated potassium or kainate does not modify the overall level of reelin mRNA and Reelin protein in organotypic hippocampal slices (Tinnes et al. 2011). These last results are in accordance with data showing that Reelin expression is independent of activity in cerebellar neurons (Lacor et al. 2000). Here, we investigated the role of increased electrical activity and the subsequent release of BDNF on reelin mRNA expression in cortical neurons. We found that elevated levels of activity are associated with increased levels of reelin mRNA. We further found that upon Gbz application, increased reelin mRNA levels do not only positively correlate with bdnf mRNA levels, but that BDNF signaling is causal for the increase in reelin mRNA expression. One possibility for these discordant results is that the regulation of reelin mRNA expression is cell type-dependent. In our cortical cultures, about two-thirds of the Reelin-positive neurons are GABAergic and a third are CRs. The work of Ringstedt et al. focuses on the embryonic neocortex, where the large majority of Reelin-secreting cells are CRs. We hypothesize that the opposite effect of BDNF observed in embryonic CRs versus cortical cultures is due to its antithetic effect in CRs versus GABAergic neurons: BDNF might prevent reelin mRNA expression in CRs (Ringstedt et al. 1998) whereas it promotes reelin mRNA transcription in GABAergic neurons. The putative cell-type specific mechanism underlying Reelin expression is further supported in the work of Tinnes et al., where a transient increase of electrical activity does not affect Reelin protein expression in CRs, but leads to a 30% decrease in Reelin immunoreactivity in interneurons a week later.
Interestingly, silencing the neuronal network with TTX does not affect reelin mRNA transcription but promotes a transient increase in Reelin protein translation. Thus, increasing neuronal activity and silencing the neuronal network do not show opposite effects but rather appear to act on different steps of protein synthesis. We hypothesize that neuronal activity leads to increased reelin mRNA via the BDNF/TRkB pathway, thus activating transcription factors such as CREB (Shaywitz and Greenberg 1999). On the other side, neuronal silencing leads to an overall decrease in gene expression and might lead to an increased availability of translation machinery elements such as transportation elements, ribosomes, or tRNA (Metzl et al. 2020). This increased availability of translation machinery elements could allow an accelerated translation of reelin mRNA molecules at first (seen in the western blot as an increased protein levels) before reaching a steady-state level matching the transcriptional speed.
We further investigated whether changing activity patterns leads to a modification in the rate of Reelin secretion. After 6 h of pharmacological treatment to either silence the neuronal network or promote neuronal depolarization, no change in Reelin levels was observed in the culture medium. However, the increased levels of Reelin protein observed in the cellular fraction after 6 h of TTX application are followed by elevated Reelin levels in the culture medium 3 h later, indicating that Reelin secretion follows the level of Reelin protein translation. These data are in accordance with previous experiments performed in adult cerebellar neurons showing that Reelin secretion is not regulated by activity-dependent exocytosis (Lacor et al. 2000). On the other side, experimental data performed in a mouse model temporal lobe epilepsy demonstrate that epileptiform activity leads to an intracellular accumulation of Reelin in adult hippocampal CRs (Duveau et al. 2011). It is possible that under our conditions, the levels of activity in Gbz-treated cultures are not elevated enough to lead to a modification of Reelin protein expression and subsequent secretion.
Here, we demonstrated that the expression of Reelin is mostly ensured by CRs until the first postnatal days in vivo, and that in neuronal cultures, various electrical activity patterns control different stages of Reelin synthesis in developing cortical neurons. These data hint at a cell-type specific mechanism by which electrical activity switches off the synthesis of Reelin in CRs and upregulates Reelin expression in GABAergic neurons. In animal models and tissues from human patients, a defect in Reelin expression and/or secretion leads to severe neurodevelopmental disorders. It will be interesting to investigate in more details the mechanisms allowing electrical activity to modify the levels of Reelin synthesis in developing CRs and GABAergic neurons under physiological and pathological conditions. For instance, the hypermethylation of the reelin promoter in adult GABAergic neurons represents an important mechanism leading to the downregulation of reelin transcription in patients suffering from schizophrenia (Abdolmaleky et al. 2005;Grayson et al. 2006;Magwai et al. 2021). Whether electrical activity or neurotrophic factor signaling have any effect on the methylation of reelin promoter in developing cortical neurons would be an interesting avenue for future research (Moore et al. 2013). These findings will be important to better understand the physiological steps controlling cortical development and to gain insight on the etiology of developmental disorders associated with pathological levels of Reelin.