Monoamines such as serotonin and dopamine have been shown to regulate cortical interneuron migration but very little is known regarding noradrenaline. Similarly to other monoamines, noradrenaline is detected during embryonic cortical development and adrenergic receptors are expressed in transient embryonic zones of the pallium that contain migrating neurons. Evidence of a functional role for the adrenergic system in interneuron migration is lacking. In this study we first investigated the expression pattern of adrenergic receptors in mouse cortical interneuron subtypes preferentially derived from the caudal ganglionic eminences, and found that they expressed different subtypes of adrenergic receptors. To directly monitor the effects of adrenergic receptor stimulation on interneuron migration we used time-lapse recordings in cortical slices and observed that alpha2 adrenergic receptors (adra2) receptor activation inhibits the migration of cortical interneurons in a concentration-dependent and reversible manner. Furthermore, we observed that following adra2 activation the directionality of migrating interneurons was significantly modified, suggesting that adra2 stimulation could modulate their responsiveness to guidance cues. Finally the distribution of cortical interneurons was altered in vivo in adra2a/2c-knockout mice. These results support the general hypothesis that adrenergic dysregulation occurring during embryonic development alters cellular processes involved in the formation of cortical circuits.
In rodents, cortical interneurons are mainly generated in the medial and caudal ganglionic eminences of the subpallium and migrate tangentially to reach the developing cortex (Wonders & Anderson, 2006; Gelman & Marin, 2010; Rudy et al., 2011). The specification and migration of cortical interneurons is controlled by a combinatorial cascade of transcription factors which regulates a variety of receptors and effectors required for their proper response to cell-extrinsic cues (Flames & Marin, 2005; Chedotal & Rijli, 2009). Among these external cues, monoamines such as serotonin and dopamine have been shown to regulate cortical interneuron migration (Crandall et al., 2007; Riccio et al., 2009). Similarly to serotonin and dopamine, noradrenaline is another monoamine which is detected during cortical development and has been suggested as modulating cellular processes involved in the formation of cortical circuits (Lidow & Rakic, 1994). Support for a role for noradrenaline in this process comes from the fact that noradrenergic fibres reach the rodent developing cortex at late embryonic day (E)16 (Levitt & Moore, 1979), at a time period when cortical interneurons are in the process of actively invading the intermediate zone and cortical plate. A potential role for noradrenaline in neuronal migration is not restricted to rodents. In humans and non-human primates, noradrenaline fibres have been shown to reach the early developing cortex during a period of intense neuronal migration (Lidow & Rakic, 1994; Zecevic & Verney, 1995; Wang & Lidow, 1997). Further support for a developmental role of noradrenaline comes from studies demonstrating that adrenergic receptors are strongly expressed during embryonic cortical development (Lidow & Rakic, 1994; Wang & Lidow, 1997; Winzer-Serhan & Leslie, 1999). Alpha1 adrenergic receptors (adra1), alpha2 adrenergic receptors (adra2) and beta adrenergic receptors (adrb) display distinct and restricted temporospatial expression throughout the transient embryonic zones of the macaque and rodent pallium (Lidow & Rakic, 1994; Wang & Lidow, 1997; Winzer-Serhan & Leslie, 1999). The expression pattern of adrenergic receptors in the developing pallium has led to the hypothesis that these receptors could regulate different developmental processes including neuronal migration (Wang & Lidow, 1997). Interestingly, in non-neuronal systems, adrenergic modulation regulates the migration of different cell types including hematopoietic progenitor cells (Spiegel et al., 2007), corneal epithelial cells (Pullar et al., 2007), keratinocytes (Pullar et al., 2006), vascular smooth muscle cells (Johnson et al., 2006) and different types of cancer cells (Masur et al., 2001; Bastian et al., 2009). In the neocortex, evidence of a functional role for the adrenergic system in the migration of cortical neurons is lacking. Early studies suggested that the destruction of noradrenergic innervation during the early postnatal period could affect the maturation of the cerebral cortex (Maeda et al., 1974; Felten et al., 1982; Brenner et al., 1985). However, no studies have directly tested the effects of adrenergic stimulation on cortical interneuron migration. In this study we investigated the expression pattern of adrenergic receptors in embryonic cortical interneuron subtypes preferentially derived from the caudal ganglionic eminences, and used time-lapse recordings to directly monitor the consequences of adrenergic receptor pharmacological manipulation on interneuron migration in control and adra2a/2c-knockout (ko) mice. Finally we investigated the positioning of cortical interneurons in adra2a/2c-ko mice in vivo at postnatal day 21.
Materials and methods
All animal experiments were conducted according to relevant national and international guidelines and approved by the local Geneva animal care committee. The day of the vaginal plug detection was counted as E0.5. To monitor cortical interneurons, we used transgenic mice expressing GFP under the control of the GAD65 promoter (GAD65-GFP mice; Riccio et al., 2009). Previously characterised adra2a-, adra2c- and adra2a/2c-ko mice (Hein et al., 1999) were crossed to GAD65-GFP mice to generate adra2a-ko GAD65-GFP, adra2c-ko GAD65-GFP, adra2a/2c-ko GAD65-GFP mice.
In utero electroporation, cortical slice preparation and drugs
To label pyramidal neurons and interneurons, GAD65-GFP+ embryos from timed pregnant E14.5 dams were electroporated with a pRIX plasmid expressing a red fluorochrome (TOM+) under the regulation of the ubiquitin promoter in the ventricular zone (VZ) of the lateral pallium. For details of the construct see Dayer et al., 2007. After in utero electroporation, dams were killed at E17.5 by intraperitoneal (i.p.) pentobarbital injection (50 mg/kg), pups were killed by decapitation and brains were dissected. Cortical slices (200 μm thick) were cut on a Vibratome (Leica VT100S; Nussloch, Germany), washed in a dissection medium (minimum essential medium, 1×; Tris, 5 mm; and penicillin–streptomycin, 0.5%) for 5 min, placed on porous nitrocellulose filters (Millicell-CM; Millipore. Zug, Switzerland) in 60-mm Falcon Petri dishes and kept in neurobasal medium (Invitrogen, Lucerne, Switzerland) supplemented with B27 (Invitrogen), 2%; glutamine, 2 mm; sodium pyruvate, 1 mm; N-acetyl-cysteine, 2 mm; and penicillin–streptomycin, 1%. Drugs were obtained from Tocris (Abingdon, UK): medetomidine, cirazoline, guanfacine and isoproterenol hydrochloride (all diluted in H2O; stock 100 mm) and (R)-(+)-m-nitrobiphenyline oxalate (diluted in DMSO; stock 50 mm).
Tissue processing and immunohistochemistry
Animals were deeply anesthetised with pentobarbital injected i.p (50 mg/kg), and killed by intracardiac perfusion of 0.9% saline followed by cold 4% paraformaldehyde (PFA; pH 7.4). Brains were post-fixed over-night in PFA at 4 °C and coronal sections were cut on a Vibratome (Leica VT100S; Nussloch, Germany; 60-μm-thick sections) and stored at 4 °C in 0.1 m phosphate-buffered saline (PBS). For free-floating immunohistochemistry, sections were washed three times with 0.1 m PBS, incubated overnight at 4 °C with a primary antibody diluted in PBS with 0.5% bovine serum albumin (BSA) and 0.3% Triton X-100, washed in PBS, incubated with the appropriate secondary antibody for 2 h at room temperature, counterstained in Hoechst 33258 (1 : 10 000) for 10 min and then mounted on glass slides with Immu-Mount™ (Thermo Scientific, Erembodegem, Belgium). Primary antibodies were the following: rabbit anti-calretinin (1 : 1000; Swant, Switzerland), mouse anti-parvalbumin (1 : 5000; Swant), rat anti-somatostatin (1 : 100; Millipore, Zug, Switzerland), rabbit anti-NPY (1 : 1000; Immunostar, Losone, Switzerland), rabbit anti-VIP (1 : 1000; Immunostar) and mouse anti-reelin (1 : 1000; Medical Biological Laboratories, Nagoya, Japan). Secondary Alexa-568 antibodies (Molecular Probes, Invitrogen, Lucerne, Switzerland) raised against the appropriate species were used at a dilution of 1 : 1000.
cDNA synthesis and quantitative polymerase chain reaction (PCR)
E17.5 cortical slices from GAD65-GFP+ pups electroporated at E14.5 were prepared and kept in vitro for 24 h. The lateral cortex containing TOM+ pyramidal neurons and GAD65-GFP+ interneurons were trypsinised in Hanks’ medium for 10 min at 37 °C. After centrifugation the pellet was filtered using 40-μm-pore filters (Falcon). GFP+ and TOM+ cells were sorted using fluorescence-activated cell sorting (FACS). Total RNA from the sorted cells was extracted, amplified (MessageAMP™ II aRNA Amplification kit; Ambion, Zug, Switzerland) in order to obtain at least 50 ng of RNA, and converted into cDNA. PCR was done using a REDtaq Ready-Mix (Sigma, Buchs, Switzerland) and PCR products were electrophoresed in a 2% agarose gel. For quantitative PCR, PCR reactions were performed in triplicate on cDNA from TOM+ cells and GAD65-GFP+ cells using SYBR green PCR Master Mix (Applied Biosystems, Rotkreuz, Switzerland) in an ABI Prism 7900 Sequence Detection system (Applied Biosystems). Four genes were used as internal controls: beta-actin (actb), gamma-actin (actg1), eukaryotic elongation factor-1 (eef1a1) and beta-glucuronidase (Gusb). Primers for the different adrenergic receptors were designed using the Ensembl database and the Primer3 software. Primer sequences were as follows: adra1a forward, 5′-CTGCCATTCTTCCTCGTGAT-3′ and reverse 5′-GCTTGGAAGACTGCCTTCTG-3′, adra1b, forward, 5′-AACCTTGGGCATTGTAGTCG-3′ and reverse 5′-CTGGAGCACGGGTAGATGAT-3′ adra1d forward, 5′-TCCGTAAGGCTGCTCAAGTT-3′ and reverse, 5′-CTGGAGCAGGGGTAGATGAG-3′, adra2a forward, 5′ TGCTGGTTGTTGTGGTTGTT-3′ and reverse, 5′-GGGGGTGTGGAGGAGATAAT-3′, adra2b, forward 5′-GCCACTTGTGGTGGTTTTCT-3′, reverse, 5′- TTCCCCAGCATCAGGTAAAC-3′, adra2c forward, 5′-TCATCGTTTTCACCGTGGTA-3′ and reverse, 5′-GCTCATTGGCCAGAGAAAAG-3′, adrb1 forward, 5′-TCGCTACCAGAGTTTGCTGA-3′ and reverse, 5′-GGCACGTAGAAGGAGACGAC-3′, adrb2, forward. 5′-GACTACACAGGGGAGCCAAA-3′, and reverse, 5′-TGTCACAGCAGAAAGGTCCA-3′, adrb3 forward, 5′-TGAAACAGCAGACAGGGACA-3′, reverse 5′-TCAGCTTCCCTCCATCTCAC-3′.
Image acquisition and data analysis
Cortical slices were imaged in a thermoregulated chamber maintained at 37 °C and CO2 at 5% as previously described (Riccio et al., 2009). Time-lapse movies were acquired in parallel using two fluorescent microscopes (Eclipse TE2000; Nikon, Egg, Switzerland) equipped with a Nikon Plan 10×/0.30 objective connected to a digital camera (Retiga EX). Time-lapse imaging was performed 3–4 h after slice preparation over a period of 24 h. Images were acquired using the Open-lab software (version 5.0; Schwerzenbach, Switzerland) every 5 min for 200 min in short time-lapse sequences and for 600 min in washout experiments. A control time-lapse sequence of 95 min was acquired in each condition before the treatment condition. Time-lapse stacks were generated and analysed using Metamorph software (version 7.4; Visitron, Puchheim, Germany). GAD65-GFP+ cells (n =40 cells per slice in at least three independent experiments) located in the intermediate zone and the cortical plate of the prospective somatosensory cortex were randomly selected in the control condition and single-cell tracking was performed blind to the treatment condition on migrating cells (velocity of at least 15 μm/h). The mean velocity over a 90-min recording period was calculated in the control and treatment condition. To measure a change in the directionality of migrating interneurons after treatment conditions, the angle change between the track path of the control condition and of the wash condition was calculated.
For quantification of the distribution of GAD65-GFP+ interneurons, sections from GAD65-GFP mice and adra2a/2c-ko GAD65-GFP mice were obtained at P21 and quantified in the somatosensory cortex (bregma -1.34; mouse brain atlas, Paxinos and Franklin, 2001). Composite epifluorescent images (Nikon Plan 10× objective) were obtained with GAD65-GFP+ and Hoechst labelling, a grid was apposed on the corresponding somatosensory cortex using the Metamorph software (version 7.4) and GAD65-GFP+ cells were manually counted in the different cortical layers (n =6 GAD65-GFP+ brains, total of 881 cells; n =6 adra2a-ko GAD65-GFP+ brains, total of 1015 cells).
Epifluorescent images (Nikon Plan 10× objective) were taken at the level of the somatosensory cortex to quantify the percentage of GAD65-GFP+ interneurons located in upper (I–IV) and lower (V and VI) cortical layers and expressing VIP (n =3, 529 cells), reelin (n =3, 685 cells), NPY (n =3, 644 cells), calretinin (n =3, 673 cells), parvalbumin (n =3, 726 cells) and somatostatin (n =3, 623 cells).
Statistical analysis (GraphPad prism software, version 4.0) was done using unpaired Student’s t-test, one-way anova with Tukey’s multiple comparison test, or χ2 test. Statistical significance was defined at *P <0.05, **P <0.01. Values given are means ± SEM.
Expression of adrenergic receptors in migrating GAD65-GFP+ cortical interneurons
Transgenic mice expressing GFP under the control of the GAD65 promoter were used to study cortical interneuron migration as previously described (Riccio et al., 2009). Given the high subtype diversity of cortical interneurons, we first characterised the identity of GAD65-GFP interneurons using molecular markers. As previously reported (Lopez-Bendito et al., 2004; Riccio et al., 2011), we found that GAD65-GFP+ interneurons preferentially express markers that label cortical interneurons derived from the caudal ganglionic eminences but not the medial ganglionic eminences (Fig. S1). Quantification at postnatal day 21 in the somatosensory cortex revealed that GAD65-GFP+ cortical interneurons hardly expressed parvalbumin or somatostatin (Fig. S1), which are classical markers of cortical interneuron subtypes derived from the medial ganglionic eminences (Rudy et al. 2011). In contrast, GAD65-GFP+ interneurons expressed markers such as reelin, NPY, VIP and calretinin, which preferentially label cortical interneuron subtypes derived from the caudal ganglionic eminences (Fig. S1; Rudy et al. 2011). Migration of GAD65-GFP+ interneurons was monitored between E17.5 and E18.5, a developmental time period when GAD65-GFP+ cortical interneurons exit the tangential stream of the subventricular zone (SVZ) and invade the intermediate zone and cortical plate. To first determine whether migrating GAD65-GFP+ interneurons expressed adrenergic receptors, we used FACS to isolate a population of GAD65-GFP+ cortical interneurons from cortical slices. To label and isolate excitatory pyramidal precursors using FACS, in utero electroporation of a TOM+-expressing plasmid in the ventricular zone of the dorsal pallium was performed at E14.5 (Fig. 1A). This method is widely used to specifically label excitatory pyramidal neurons in vivo (Chen et al., 2008). Electroporation in the GAD65-GFP+ mice confirmed that TOM+ cells did not overlap with cortical interneurons (Fig. 1A). Real-time PCR performed on amplified mRNA extracted from FACS-isolated GAD65-GFP+ cells revealed that GAD65-GFP+ cells expressed a pattern of adrenergic receptors: the alpha1d adrenergic receptor (adra1d), the alpha2a adrenergic receptor (adra2a), the alpha2c adrenergic receptor (adra2c) and the beta1 adrenergic receptor (adrb1; Fig. 1B). None of the other adrenergic receptor subtypes were detected an the mRNA level (data not shown). Quantitative PCR did not reveal any major differences between the expressions of adrenergic receptors in FACS-isolated GAD65-GFP+ interneurons and TOM+ pyramidal neurons (Fig. 1C), indicating that adrenergic receptor numbers are not specifically raised in GAD65-GFP+ cortical interneurons. Among the four adrenergic receptors expressed in GAD65-GFP+ cells, adra2a, adra2c and adrb1 were expressed at higher levels than adra1d (Fig. 1D).
Activation of adrenergic receptors decreased cortical interneuron migration
To determine whether migrating interneurons could respond to adrenergic stimulation, we used time-lapse imaging of GAD65-GFP interneurons in cortical slices combined with pharmacological drug applications. Imaging of cortical interneurons was performed between E17.5 and E18.5. Migrating cortical interneurons located in the cortical plate and intermediate zone were randomly selected and tracked initially during a control period of 95 min. After 95 min of time-lapse imaging, drugs targeting adrenergic receptors expressed in GAD65-GFP+ cells were applied to the bath medium and effects on migration were analysed. Using this slice assay, application of an adrb agonist (isoproterenol, 500 μm) did not significantly modify the mean speed of neuronal migration whereas application of an adra1 agonist (cirazoline 500 μm) and an adra2 agonist (medetomidine 500 μm) significantly reduced the mean migratory speed of GAD65-GFP interneurons (P <0.01 for both drugs vs. control, one-way anova, Tukey multiple comparison test; Fig. 1, E1, E2–G and Movies S1 and S2). Application of cirazoline and medetomidine shifted the speed distribution of GAD65-GFP+ interneurons to lower migratory speeds and a greater proportion of cells migrating at < 15 μm/h were observed during exposure to medetomidine and cirazoline than during control conditions (Fig. 1G).
Activation of alpha2 adrenergic receptors decreased cortical interneuron migration in a concentration-dependent and reversible manner
To further investigate the effects of adra2 agonist stimulation on interneuron migration, we applied the agonists medetomidine or guanfacine at different concentrations (P <0.01 for all concentrations tested vs. control, one-way anova, Tukey’s multiple comparison test; Fig. 2A–C). A concentration-dependent effect of medetomidine on migratory speed was observed (Fig. 2B). This concentration-dependent effect could be detected after application of guanfacine, an agonist with some selectivity for the adra2a subtype (P <0.01 for all concentrations tested vs. control, one-way anova, Tukey’s multiple comparison test; Fig. 2A–C, Movies S3) and (+)-m-nitrobiphenyline oxalate, a more specific adra2c agonist (P <0.01 for all concentrations tested vs. control, one-way anova, Tukey’s multiple comparison test; Fig. 2D), further confirming that activation of adra2a and adra2c affects the migratory speed of GAD65-GFP+ cortical interneurons. To test whether these drugs altered cortical interneuron migration by specifically acting on adra2a and adra2c receptors, time-lapse imaging was performed on cortical slices of adra2a/2c-ko GAD65-GFP mice (Hein et al., 1999). No basal differences in the mean migratory speeds were observed in adra2a/2c-ko GAD65-GFP cells compared to control GAD65-GFP+ cells. Single-cell tracking revealed that guanfacine (300 μm) and medetomidine (300 μm) significantly decreased the migration speed of GAD65-GFP+ interneurons compared to adra2a/2c-ko GAD65-GFP+ interneurons (P <0.01 for guanfacine in controls vs. guanfacine in adra2a/2c-ko and P <0.01 for medetomidine in controls vs. medetomidine in adra2a/2c-ko, one-way anova, Tukey’s multiple comparison test; Fig. 2E and F), indicating that the effects of these drugs on GAD65-GFP+ migrating interneurons are dependent on the activation of adra2a and adra2c receptors. It should be noted, however, that guanfacine decreased the migratory speed of adra2a/2c-ko GAD65-GFP+ cells (P <0.05, one-way anova, Tukey’s multiple comparison test), suggesting that guanfacine could partially act independently of adra2a/2c receptor activation.
To test whether adra2 agonist stimulation produced persistent effects on interneuron migration, medetomidine (500 μm) was applied in the bath medium for > 6 h. Using this protocol, we observed that long-term application of medetomidine (> 6 h) almost completely halted the migration of cortical interneurons without inducing toxic effects such as cell death (Fig. 3A and C, Movies S4). In contrast, when medetomidine was washed out of the medium after a shorter time period of drug application (95 min), the effects of adra2 activation on the speed of interneuron migration were reversible (Fig 3B and C, Movies S5). Single-cell tracking revealed that after washing out medetomidine, the migratory speed of GAD65-GFP+ interneurons significantly increased and gradually reached control values (P <0.01 at the first time interval after the drug washout when comparing medetomidine vs. no-wash medetomidine, one-way anova, Tukey’s multiple comparison test; Fig. 3B and C). These effects were also observed after application of guanfacine. When guanfacine was washed out from the medium, the speed of interneuron migration significantly increased and gradually reached control values (P <0.01 at the first time interval after the drug wash when comparing guanfacine vs. no-wash medetomidine, one-way anova, Tukey’s multiple comparison test; Fig. 3C). Interestingly, we observed that although the migratory speed of GAD65-GFP+ cells was gradually restored during the removal of either medetomidine or guanfacine, the directionality of GAD65-GFP+ cells was modified by adra2 stimulation (Fig. 3B). Quantification revealed that during the washout period a significant proportion of GAD65-GFP+ cells modified their directionality following medetomidine or guanfacine application. The percentage of GAD65-GFP+ interneurons that made directionality changes in the range > 120–180° after the medetomidine wash or the guafancine wash was significantly increased compared to control GAD65-GFP+ interneurons (P <0.01 for guanfacine compared to control and P <0.05 for medetomidine vs. control, one-way anova Tukey’s multiple comparison test; Fig. 3D), suggesting that adrenergic stimulation of cortical interneurons may alter their responsiveness to guidance cues.
Cortical interneuron positioning was altered in adra2a/2c-ko mice
To determine whether cortical interneuron migration is altered in adra2a/2c-ko mice, we analysed the cortical distribution of GAD65-GFP+ interneurons at postnatal day 21 in GAD65-GFP mice and in adra2a/2c-ko GAD65-GFP mice. Quantification revealed that the distribution of GAD65-GFP+ cortical interneurons in the somatosensory cortex was significantly altered in adra2a/2c-ko mice (n =6) compared to the control mice (n =6; P <0.05, χ2 test; Fig. 4). A significant increase in the percentage of GAD65-GFP+ cells was observed in upper cortical layers II/III in adra2a/2c-ko mice (P <0.05, unpaired t-test), indicating that adrenergic receptors are necessary for the proper positioning of cortical interneurons in vivo. Quantification of the distribution of GAD65-GFP+ cells at P21 in the somatosensory cortex of adra2a-ko or of adra2c-ko mice was not significantly different from control GAD65-GFP+ mice (data not shown), suggesting that constitutive deletion of adra2a or adra2c during development may be compensated for by the presence of the other subtype.
In this study we found that migrating cortical interneuron subtypes preferentially derived from the caudal ganglionic eminences express a specific pattern of adrenergic receptors and that pharmacological activation of these receptors affects the dynamic migration of cortical interneurons as they invade the developing cortical plate. Effects of adrenergic stimulation were most effective after adra2 stimulation, and they were concentration-dependent and reversible. Furthermore, effects of adra2 activation on the migration of cortical interneurons were significantly reduced in adra2a/2c-ko mice. Altogether these data strongly suggest a role for adrenergic stimulation in cortical interneuron migration.
A role for noradrenaline during cortical development has been hypothesised on the basis that noradrenergic fibres originating from the locus coeruleus (LC) reach the cortical anlage during the embryonic period in rodents, macaques and humans (Levitt & Moore, 1979; Zecevic & Verney, 1995; Wang & Lidow, 1997). During embryonic cortical development, fibres from the LC express dopamine-beta-hydroxylase, the rate-limiting enzyme for noradrenaline, and are thus likely to release noradrenaline in the extracellular space of the cortical anlage (Wang & Lidow, 1997). An alternative source of noradrenaline could be the cerebrospinal fluid where high levels of noradrenaline have been detected during the embryonic period (Masudi & Gilmore, 1983). Noradrenaline in the CSF could originate from the fetal blood by passing through the immature blood–brain barrier, diffuse from the CSF into the ventricular wall and regulate cellular processes involved in the formation of cortical circuits, including neuronal migration. A role for noradrenaline during embryonic cortical development is further supported by the fact that different subtypes of adrenergic receptors are dynamically expressed across species during cortical development and follow a restricted temporal and spatial pattern of expression. Initial binding studies revealed that adra1, adra2 and adrb1 are highly expressed in the developing cortical plate and transient embryonic zones of the non-human primate brain (Lidow & Rakic, 1994). A more detailed study on adra2a indicated that this receptor is expressed at E70, E90 and E120 throughout the macaque embryonic wall (Wang & Lidow, 1997). Interestingly, this study revealed that migrating neurons in the intermediate zone and cortical plate expressed high levels of adra2a, suggesting that this receptor could play a role in regulating neuronal migration (Wang & Lidow, 1997). A role for adra2a in neuronal migration is further supported by the fact that strong adra2a expression is detected in the cortical plate, intermediate and subventricular zones of the embryonic rat cortex (Winzer-Serhan & Leslie, 1997; Winzer-Serhan & Leslie, 1999).
The group of adra2 receptors is composed of three highly homologous subtypes (adra2a, adra2b and adra2c). In this study we found that migrating cortical interneurons expressed adra2a and adra2c but not adra2b, and that activation of adra2a and adra2c affects neuronal migration. Interestingly, it has been recently reported that adra2 receptors regulate adult hippocampal neurogenesis, a developmental process that persists in the adult brain (Yanpallewar et al., 2010). Progenitor cells in the hippocampus express adra2a, adra2b and adra2c subtypes and adra2 stimulation inhibits the proliferation of granule cell progenitors in the dentate gyrus, leading to decreased levels of adult hippocampal neurogenesis (Yanpallewar et al., 2010). Knockout mice for adra2a and adra2c have been generated (Hein et al., 1999) and the role of these receptors in the cardiovascular system has been studied in detail (Knaus et al., 2007a,b). Briefly, adra2a/2c-ko mice display elevated plasma concentrations of catecholamines, increased blood pressure and cardiac hypertrophy in adulthood (Hein et al., 1999; Knaus et al., 2007a,b). The developmental consequences of constitutive deletions of adra2a, adra2c and adra2a/2c in the central nervous system are not striking and the brains of these animals appear to be grossly normal. Quantification of the distribution of GAD65-GFP+ interneurons in adra2a-ko or adra2c-ko mice did not reveal any significant changes in the distribution of cortical interneurons at P21, suggesting compensatory regulatory mechanisms following constitutive developmental deletion of either of these receptors. Interestingly a significant increase in the percentage of GAD65-GFP+ cells in upper cortical layers II/III were detected in the somatosensory cortex of adra2a/2c-ko mice, indicating that combined deletion of adra2a and adra2c receptors significantly modifies the distribution of cortical interneurons in vivo.
The intracellular mechanism mediating the effects of adra2 stimulation on interneuron migration is likely to involve different transduction pathways. Adra2 are G-protein-coupled receptors negatively coupled to adenylate cyclase, and modifications in the levels of cAMP could thus constitute a downstream effector of adra2 stimulation. Cyclic AMP is a key molecule regulating growth cone dynamics (Song & Poo, 2001), and experimental manipulation of the ratio of cAMP to cGMP determines the responsiveness of axonal growth cones to guidance cues (Nishiyama et al., 2003). In the embryonic brain cAMP is critical for proper axonal pathfinding of olfactory sensory neurons (Chesler et al., 2007). In migrating neurons, alteration in the levels of cAMP decreases the migratory speed of cerebellar granule cells (Cuzon et al., 2008) and modulates the effects of serotonin on migrating cortical interneurons (Riccio et al., 2009). Interestingly, there is a functional pathway linking adra2a, cAMP and hyperpolarization-activated cyclic nucleotide-gated cation channels (HCN channels; Wang et al., 2007). HCN channels have been shown to regulate axonal targeting of olfactory sensory neurons during development (Mobley et al., 2010) and thus represent an attractive downstream developmental target of cAMP that could regulate interneuron migration. Calcium could also be another downstream effector mediating the effects of adra2 activation on migrating interneurons. In other cellular systems, it has been shown that adra2a stimulation regulates intracellular calcium levels through the modulation of voltage-gated N-type calcium channels and that this process occurs independently of cAMP modulation (Lipscombe et al., 1989; Ikeda, 1996). As intracellular calcium levels regulate the migration of cerebellar granule cells (Komuro & Rakic, 1996), adra2a-induced calcium changes could potentially regulate cortical interneuron migration. Finally, an interesting observation in this study is that adra2 stimulation affected not only the migratory speed of cortical interneurons but also their directionality. When adra2 agonist was removed from the bath medium, cortical interneurons resumed a normal migratory speed but the directionality of migration was significantly modified in a fraction of cells compared to the control situation. These results suggest that changes in cAMP levels through adra2 stimulation could modify the responsiveness of cortical interneurons to guidance cues. Support for this possibility comes from the observation that in other systems manipulation of cAMP levels can modify the responsiveness of thalamocortical axons to guidance cues through the monoaminergic activation of G-protein-coupled receptors negatively linked to adenylate cyclase (Bonnin et al., 2007).
In this study the effects of adrenergic stimulation on interneuron migration were detected using several different drugs at relatively high concentrations. However, it must be noted that in this slice culture system drugs reached the migrating cells by passively diffusing through the pores of the Millipore inserts. It is thus likely that the cortical interneurons migrating in the slice are exposed to lower drug concentrations. Importantly, application of adra2a/2c agonists significantly decreased the migratory speed of wildtype cortical interneurons compared to adra2a/2c-ko cortical interneurons. These results strongly indicate that the effects of adra2a/2c stimulation on cortical interneurons are dependent on the activation of these receptors. It should be noted, however, that guanfacine slightly affected the migratory speed of GAD65-GFP+ interneurons in adra2a/2c-ko mice, suggesting that this drug could also act independently of adra2a/2c activation. Interestingly, a study using adra2a/2b/2c triple-ko mice has revealed that clonidine, an adra2 agonist, could modulate heart reactivity by directly acting on HCN (Knaus et al., 2007b). Finally, although adrab1 was found to be expressed in GAD65-GFP+ cells, application of an adrb1 agonist at relatively high concentration failed to modify the migration of interneurons, suggesting that this receptor may not be functional at this embryonic timepoint.
In conclusion, we report that several adrenergic receptors are expressed in migrating cortical interneurons, particularly the adra2a and adra2c subtypes. Using time-lapse imaging we have demonstrated that activation of adra2 affects cortical interneuron migration in a reversible manner. Finally, the distribution of cortical interneurons was altered in vivo in adra2a/2c-ko mice. These results support the hypothesis that adrenergic dysregulation induced by exposure during pregnancy to drugs that block adrenergic receptors may affect cellular processes involved in the assembly of cortical circuits. These results also open the possibility that a pathological overstimulation of adrenergic receptors due to excessive levels of norepinephrine during development could lead to alterations in cortical circuit formation.
We wish to thank C. Aubry for technical lab assistance and the Geneva Genomics Platform for qPCR assistance (Christelle Barraclough and Patrick Descombes). This work was supported by a Swiss National Foundation grant (PP00P3_128379), the NCCR Synapsy, the Thorn Foundation, the Mercier Foundation, NARSAD (The brain and Behaviour Research Fund; A.D.) and by BIOSS Centre for Biological Signalling Studies (EXC 294, in support of L.H.).