Abstract

Spontaneous correlated activity regulates the precision of developing neural circuits. A synchronized elevation of intracellular calcium ion concentration, [Ca2+]i, occurred in 5–50 adjacent neurons — known as a ‘neuronal domain’ — in developing neocortex. This coordinated response of neuronal cells is mediated by the diffusion of inositol trisphosphate (IP3) via gap-junction channels. In this study, we utilized the N-methyl-d-aspartate (NMDA)-type glutamate receptor ε2 (GluRε2/NR2B)−/− mouse, which does not possess any functional NMDA receptors in the developing neocortex, and showed that NMDA receptors are essential for the generation of ‘neuronal domains’. First, the frequency of spontaneously occurring neuronal domains in brain slices from GluRε2−/− mice was significantly reduced compared to that seen in brain slices from wild-type mice. Secondly, IP3 injection into a single neuron in a cortical slice from a GluRε2−/− brain resulted in very few neuronal domains being observed, but an injection similarly made into a neuron in a wild-type slice promptly resulted in neuronal domains. Even in the GluRε2−/− brain, the elevation of intracellular [Ca2+]i was observed frequently in single neurons and microinjection of IP3 produced an elevation of [Ca2+]i in the injected cells. These results suggest that the diffusion of IP3 into the surrounding neurons via gap junctions is almost completely absent in the GluRε2−/− brain. Our results may reflect the critical role of NMDA receptors in the formation of cortical circuitry, probably via the regulation of gap-junction channels between immature cortical neurons.

Introduction

Spontaneous correlated activity regulates the precision of developing neural circuitry (Wiesel and Hubel, 1965; Rakic, 1976; Shatz and Stryker, 1988; Weliky and Katz, 1997). In macaque monkeys, thalamocortical afferents begin to segregate into stripes before birth (Rakic, 1976) and are arranged into functional columns by birth (Des Rosiers et al., 1978). The development of cortical microcircuitry can be distinguished into the two phases: an early ‘establishment’ phase of cortical connections, where activity-dependent and independent mechanisms could operate, and a later ‘maintenance’ phase, which appears to be controlled by neuronal activity (Yuste and Sur, 1999). The establishment phase of cortical circuits has been examined extensively by the use of optical imaging techniques. Spontaneous correlated activity has been observed in developing retina (Meister et al., 1991) and developing neocortex (Yuste et al., 1992), where coordinated increases in intracellular calcium ion concentration, [Ca2+]i, can occur in 5–50 adjacent neurons, termed a ‘neuronal domain’. It is reasonable to postulate that neuronal domains could be the origin of functional neuronal coupling to establish cortical microcircuitry. Domains occur in the presence of tetrodotoxin but are blocked by gap-junction blockers (Yuste et al., 1995). These spontaneous events appear to involve cell-to-cell diffusion of the second messenger inositol trisphosphate or IP3 (Kandler and Katz, 1998). Therefore, the intracellular signals for the appearance of coordinated neuronal activity based on the functional unit of the ‘neuronal domain’ have mostly been clarified. On the other hand, the mechanism for the formation of this multicellular architecture in developing neocortex is yet to be elucidated.

The N-methyl-d-aspartate (NMDA)-type glutamate receptor is one of the receptor types responsible for cortical plasticity during development, including synaptic formation and synaptic plasticity. In the somatosensory cortex of the rodent, there is an early critical period during which whisker removal prevents the formation of cortical barrels (Woolsey and Wann, 1976; O’Leary et al., 1994). Interestingly, whereas earlier work had suggested that activity might have a minimal role in barrel formation (Chiaia et al., 1992; Henderson et al., 1992), recent work with transgenic mice lacking cortical NMDA receptors indicates that NMDA receptors are indeed crucial for the formation of barrels (Iwasato et al., 2000). Spontaneous neuronal activity in the prenatal stage mediated through NMDA receptors (LoTurco et al., 1991) may result in the formation of cortical microcircuitry. Expression of NMDA receptors in the neocortex of prenatal rodents has been confirmed in a series of reports (LoTurco et al., 1991; Watanabe et al., 1992; Monyer et al., 1994). LoTurco et al. (LoTurco et al., 1991) demonstrated that NMDA receptor channels on cortical plate neurons in embryonic brain slices are activated in the absence of any exogenous source of NMDA receptor agonist. Even though the opening of NMDA channels is blocked by magnesium at the resting membrane potential, ~–60mV (Mayer et al., 1984; Nowak et al., 1984), the membrane potential of neurons at this developmental stage could be elevated by the presence of GABA and glycine. In immature neurons, GABAA receptors and glycine receptors have been found to elicit a depolarizing neuronal membrane potential due to raised ECl (Luhmann and Prince, 1991; Owens et al., 1996; Flint et al., 1998; Miyakawa et al., 2002).

In this paper, we use the NMDA-type glutamate receptor GluRε2−/− mouse model, which does not possess any functional NMDA receptors in the developing neocortex (Kutsuwada et al., 1996), to investigate the relationship between neuronal domains and NMDA receptors. We assessed the frequency of the occurrence of neuronal domains in acute cortical slices from newborn mice. High-speed confocal imaging of fluo-4-stained cortical slices has enabled us to better determine the spatial and temporal resolution of the intracellular Ca2+ signal. We clearly demonstrate here the involvement of NMDA receptors in the functional formation of neuronal domains in developing neocortex of the rodent.

Materials and Methods

Animal Preparation and Genotyping

The NMDA-type glutamate receptor e2 (GluRε2/NR2B) subunit gene null-mutant was generated in a previous study (Kutsuwada et al., 1996). Because homozygous null-mutant mice (GluRε2−/−) die within 24 h of birth, they were generated by means of mating heterozygous (GluRε2±) males and females. Genotyping was performed just after pups were born. In some experiments, we also used ICR mice (Sankyo Laboratory, Tokyo, Japan). All experiments were carried out in accordance with the guidelines for Animal Experiments of the Faculty of Frontier Sciences, The University of Tokyo.

Genotype was checked using a PCR technique. The tail tissues of animals were treated with 0.1 mg/ml proteinase K (Takara, Shiga, Japan) solution at 55°C for 60 min followed by inactivation at 95°C for 15 min, from which supernatants were used as genomic DNA samples for PCR. Primers were 5′-AGA GTC GAC GAG CTG AAG ATG AAG CCC AGC-3′ (E2P1), 5′-CGG GGA ACT ACT GAG AGA TGA TGG AAG TCA-3′ (E2P2) and 5′-GCC TGC TTG CCG AAT ATC ATG GTG GAA AAT-3′ (NeoP1) as previously described (Wainai et al., 2001). Reactions were performed using a HotstarTaq DNA polymerase kit (Qiagen, Valencia, USA) as follows: hot start at 95°C for 15 min followed by 30 cycles of 30 s at 95°C, 30 s at 60°C and 1 min at 72°C. Aliquots of the PCR products were isolated by electrophoresis on 2% agarose gels and stained with ethidium bromide. PCR reactions of wild-type and targeted alleles gave bands of 218 and 377 bp.

Dissection and Fluo-4 Staining

Newborn mice (postnatal day 0: P0) were anesthetized by hypothermia and their brains were removed in cold artificial cerebro-spinal fluid (ACSF) solution. The constituents of the ACSF were as follows: NaCl, 124 mM; KCl, 2.5 mM; NaHCO3, 26 mM; MgCl2, 1 mM; CaCl2, 2 mM; NaH2PO4, 1.25 mM; and d-glucose 10 mM. The ACSF was always saturated with a 95% O2/5% CO2 gas mixture. A coronal section (400 μm thickness) was prepared from the primary somatosensory area of the cerebral cortex using a vibratome (Dosaka, Kyoto, Japan) as described previously (Miyakawa et al., 2002). The section was incubated in ACSF containing 10 μM fluo-4AM (Molecular Probes, Eugene, OR) at 37°C for 90 min. After staining, the section was briefly washed and stored at room temperature until used for experiments.

Confocal Calcium Imaging and the Observation of Neuronal Domains

Fluo-4 fluorescence was detected using a Leica confocal laser scanning microscopy system (TCS-NT or TCS-SL; Leica, Heidelberg, Germany). This system consists of an inverted microscope, a laser scanning unit, a photomultiplier and PC-based software. All fluorescent images were obtained as digital data. A stained section was placed under the microscope and fluorescent images were sequentially obtained. A frame was created from the average of two scans (each scan taking 1.67 s to perform). The interval time between the end of the pre-frame and the end of the post-frame was 6.34 s. All drugs were dissolved in ACSF and bath-applied by switching with normal ACSF. To activate NMDA receptor channels, we used NMDA (100 mM) plus glycine (100 mM), as reported previously (Yamada et al., 1999; Miyakawa et al., 2002). Several antagonists and inhibitors were also used in this study: tetrodotoxin (TTX, 5 μM); 2-amino-5-phosphonopentanoic acid (APV, 100 mM); picrotoxin (100 mM); strychnine (100 mM); halothane (10 mM); and thapsigargin (10 mM). TTX, strychnine and glycine were purchased from Wako (Osaka, Japan). The other compounds were all from Sigma (St Louis, MO).

After measurement, the tempo-spatial dynamics of [Ca2+]i in each neuron were analyzed. When a group of (more than five) neurons exhibited a simultaneously elevated [Ca2+]i, we defined this event as a ‘neuronal domain’, as reported elsewhere (Yuste et al., 1992).

High-speed Confocal Imaging with Microinjection of IP3

Microinjection of IP3 into single cells was done using a micro-electrode amplifier (MEZ-8301; Nihon Kohden, Tokyo, Japan) and an electronic stimulator (SEN-3301; Nihon Kohden). The tip radius of glass microelectrodes was <0.1 μm, giving a tip resistance > 100 MΩ. Microelectrodes were filled with a 1 M LiCl (Wako) solution containing 100 mM IP3 (Sigma). A brain section stained with fluo-4 was positioned under the upright microscope (BX51WI; Olympus, Tokyo, Japan) and a neuron in the developing neocortex was visualized with the aid of an IR-CCD camera (Hamamatsu Photonics, Hamamatsu, Japan). The neuron was impaled with the glass electrode and electrophoretically injected with IP3 by means of a hyperpolarizing current pulse (−50 nA, 100 ms). Fluorescent digital images were taken using a Nipkow disk type confocal laser-scanning unit (CSU-21; Yokogawa, Tokyo, Japan), fitted with a cooled CCD camera (HiSCA; Hamamatsu) and an Aquacosmos interface (Hamamatsu).

Immunohistochemistry

The telencephalons of P0 GluRε2+/+ and GluRε2−/− mice were post-fixed with 4% paraformaldehyde in phosphate-buffered saline (PBS) overnight at 4°C and subsequently subjected to cryoprotection in 30% sucrose (Wako) in PBS for 3 days at 4°C. Tissue samples were embedded in OCT (Sakura, Tokyo) and 40 μm coronal sections cut using a freezing microtome (Microm, Walldorf, Germany). Sections were air dried at 55°C for 15 min, washed in PBS for 5 min and blocked by incubation for 2 h at room temperature in PBS containing 0.1% Triton X-100 and 3% normal goat serum. Incubation was continued at 4°C overnight in fresh blocking solution supplemented with a 1/50 dilution of anti-mouse Cx26 polyclonal antibody (71-0500, Zymed Laboratories Inc., South San Francisco, CA). Sections were then washed in PBS (3 × 20 min) and incubated for 2 h at room temperature in PBS containing anti-rabbit IgG conjugated with Alexa488 (Molecular Probes), followed by washing in PBS (3 × 10 min). Samples were mounted in Immumount plus DABCO and analyzed by laser confocal microscopy.

Results

Observation of neuronal domains in developing cerebral neocortex

We analyzed the dynamics of changes in [Ca2+]i in the developing cerebral neocortex of newborn mice (P0) by means of confocal calcium imaging. In the cortical plate of the primary somatosensory cortex, calcium waves could be observed in 5–50 neighboring neuronal cells as they synchronously increased their [Ca2+]i. These spontaneous coordinated actions of neuronal cells appeared to occur randomly across the cortical plate (Fig. 1). The mean area of calcium waves was 5350 ± 389 μm2 (n = 49, mean ± SEM). As seen in fluorescent images from fluo-4-stained slices, calcium waves frequently originated in neuronal cells in primary layer 2/3.

Given that this type of calcium wave is similar to waves previously described for ‘neuronal domains’ (Yuste et al., 1992), we carried out a series of experiments to confirm this by using a series of antagonists and inhibitors. These were applied to the fluo-4-stained tissue for ~20 min during the Ca2+-imaging protocol. A typical result is shown in Figure 2. TTX did not inhibit the calcium waves; however, the gap-junction inhibitor halothane significantly decreased the frequency of occurrence of calcium waves. Antagonists for the NMDA receptor (APV), the GABAA receptor (picrotoxin) and the glycine receptor (strychnine) did not affect the frequency of appearance of neuronal domains (data not shown), which is consistent with results from previous studies (Yuste et al., 1992, 1995; Kandler and Katz, 1998).

Lack of NMDA-mediated Calcium Influx in GluRε2 Knockout Mouse

The NMDA receptor is a glutamate-gated ion channel and is associated with circuit formation at the developmental stage (Rabacchi et al., 1992; Simon et al., 1992). There are several subunit molecules of the NMDA receptor and their expressions are spatially and temporally regulated (Monyer et al., 1994; Mori and Mishina, 1995). It is known that the GluRζ1 and GluRε2 subunits are predominantly expressed in developing cerebral cortex, meaning that the NMDA receptor is presumably not functional in the neocortex of GluRε2 knockout mice (Kutsuwada et al., 1996; Miyakawa et al., 2002). Neurons in brain slices from wild-type (GluRε2+/+) mice strongly responded with increased [Ca2+]i in the presence of 100 μM NMDA (Fig. 3), which was completely inhibited upon exposure to 100 mM APV. On the other hand, the NMDA-mediated increase in [Ca2+]i was not seen in brain slices from GluRε2−/− mice. This result indicates that calcium signals elicited via NMDA receptor activation are completely absent in GluRε2−/− mice.

Decrease in the Frequency of Neuronal Domains in Brain Slices from GluRε2 knockout mouse

The frequency of occurrence of neuronal domains in brain slices from GluRε2−/− mice was compared to that in wild-type (GluRε2+/+) mice. The GluRε2−/− and GluRε2+/+ mice used in these experiments were littermates. Increased calcium activity in neuronal domains was frequently and spontaneously generated in developing cortical plates from GluRε2+/+ mice, whereas in GluRε2−/− mice the frequency of appearance of neuronal domains was significantly less at around just 10% of that seen in GluRε2+/+ mice (Fig. 4). The size of the infrequent neuronal domains observed in GluRε2−/− mice (5167 ± 823 μm2, n = 20, mean ± SEM) was quite similar to that seen in GluRε2+/+ mice (4970 ± 477 μm2, n = 43). The location of the infrequent neuronal domains in GluRε2−/− mice was also at layer 2/3 of developing cortex, as it was in wild-type mice. Representative observations of wild-type and knockout mice are also shown in a movie file (Supplementary Material, movie 1). This result clearly demonstrates the critical role of NMDA receptors in the generation of neuronal domains in developing cerebral cortex.

Over the time-course of activity in a neuronal domain, it is known that IP3 molecules diffuse from central cells to surrounding cells via gap-junction channels in a wave-like manner (Kandler and Katz, 1998). Therefore, it is expected that at least two mechanisms are necessary in order to generate a neuronal domain. One is an increase of IP3 in a single cell (that is, ‘initiation’ or ‘triggering’) following calcium release from the endoplasmic reticulum (ER). The other is propagation of IP3 via gap-junction channels. In GluRε2−/− mice, although neuronal domains are rarely seen, it was observed that many single cells independently increased their [Ca2+]i spontaneously (Figs 4 and 5; Supplementary Material, movie 1), suggesting that the propagation of IP3 toward neighboring neurons is blocked in GluRε2−/− mice. In order to assess this issue, we attempted to induce neuronal domains by means of IP3 injection into single cells.

The Microinjection of IP3 into a Single Neuron Produces a Neuronal Domain

It is already known that a neuronal domain can be induced by injecting IP3 into a single neuron (Kandler and Katz, 1998). For the injection of IP3, these authors used the natural diffusion of IP3 via a patch-clamp electrode. With a slight modification of this technique, we injected IP3 electrophoretically by means of a glass microelectrode of tip diameter <0.1 μm. This method allows one to produce neuronal domains at a point immediately after the precise injection of IP3. In addition, we used a Nipkow disk type confocal unit (CSU-21; Yokogawa), which enabled us to obtain better time resolution of the observation. It takes only 100 ms to obtain one high-quality confocal image from fluo-4-loaded brain slices. We performed several set-up examinations in order to determine whether our method could be used to evaluate the phenomenon of neuronal domains. First, we observed spontaneous neuronal domains using the experimental set up. As shown in Figure 6a, a calcium wave spreads from central neurons to surrounding cells over a 500 ms period. This result is quite similar to a domain evoked with a temperature drop (Yuste et al., 1995). Next, we injected IP3 into a single cell. Calcium waves that are very similar to spontaneous neuronal domains were observed (Fig. 6b). The properties of calcium waves induced by IP3 microinjection are similar to those of spontaneously occurring neuronal domains in terms of appearance, propagation and propagation velocity.

We further conducted several control experiments. Thapsigargin (10 mM) was used to deplete Ca2+ from the ER by inhibiting the uptake of Ca2+ via the Ca2+-ATPase pump. About 2 h after the application of thapsigargin, there was no evidence of calcium waves induced by IP3 injection, even in the injected cell (data not shown). This indicates that calcium elevation by IP3 injection is derived from calcium stored in the ER. Next, we injected Ca2+ ions by means of a microelectrode filled with a 1 mM CaCl2 solution. Only injected cells responded well, but without any evidence of the spread of a calcium wave (Fig. 6c). These results correspond closely to the results reported by Kandler and Katz (Kandler and Katz, 1998), who noted that neuronal domains are propagated via gap junctions by IP3, rather than by calcium ions. From these results it reasonable to suggest that our method is suitable for evaluating the properties of IP3-driven neuronal domains.

The Microinjection of IP3 into a Single Neuron from a GluRε2−/− Brain Slice

We assessed whether or not the microinjection of IP3 would produce neuronal domains in the developing neocortex of GluRε2−/− neonates. The frequency of IP3-driven neuronal domains in GluRε2−/− cortices was compared to that measured in wild-type cortices (Table 1). In most trials, neuronal domains were induced in brain slices taken from several wild-type (GluRε2+/+) mice. Representative observations from wild-type brain slices are shown in Figure 7a. Neuronal domains were not induced in the majority cases of brain slices from GluRε2−/− mice (Table 1). Instead, only IP3-injected cells showed strong [Ca2+]i elevation (Fig. 7b). The frequencies of IP3-driven neuronal domains in wild-type and GluRε2−/− mice were thus 91 and 20%, respectively. It is conceivable that during cortical development NMDA receptors would deliver the signal which activates functional cell-coupling via gap-junction channels.

In addition, our results imply some compensation for NMDA receptor-related signals, given that the microinjection of IP3 often produced neuronal domains, even in GluRε2−/− mice. The frequency was just 22% (20%/91%) of the wild type. We compared the duration of the fluorescence signal and the propagation area of IP3-driven neuronal domains for the two genotypes (GluRε2−/− mice and wild type). The time for the signal propagation, from the injection of IP3 to the propagation of the neuronal domain to neighboring neurons at the maximum [Ca2+]i level, of GluRε2−/− mice, was 2144 ± 211 ms (n = 4, mean ± SEM), which is equivalent to that of GluRε2+/+ (1906 ± 382 ms, n = 12). In addition, the area of propagation of calcium waves in brain slices from GluRε2−/− mice (7202 ± 1664 μm2, n = 5, mean ± SEM) was similar to that in wild-type mice (8456 ± 801 μm2, n = 20, mean ± SEM). Once a neuronal domain was triggered with IP3 in GluRε2−/− mice, the calcium wave was delivered in a similar manner. Although some compensating mechanism is also suspected, NMDA-type glutamate receptor GluR ζ1–e2 heterodimers in cortical development seem to act as one of the switches, which promote neuron-to-neuron bridges based on gap-junction channels.

Connexin Expression in Developing Neocortex of GluRε2−/− Mice

There is a possibility that the protein responsible for gap junction, connexin (Cx), might be downregulated in GluRε2−/− mice. Sixteen putative Cxs have been described from rodents to date, of which Cxs 26, 32, 36, and 43 are the major isoforms expressed in the developing and adult brain (Nadarajah et al., 1997). Among them, Cx26 expression is highest prenatally and during the first 3 weeks of postnatal life, suggesting that it may take part in the establishment of neuronal coupling in the developing cortex (Bittman et al., 2002). We performed an immunohistochemical study on Cx26 expression in developing neocortex of wild-type and GluRε2−/− mice. As shown in Figure 8, in both types of mice Cx26 was widely expressed in developing cortical plate and no significant differences were found. This suggests that NMDAR does not regulate the expression of Cx26 at developing neocortex.

Discussion

This study provides the first clear demonstration that the presence of NMDA receptors during neocortical development is essential for the generation of ‘neuronal domains’. First, the frequency of neuronal domains occurred spontaneously in wild-type brain slices and was significantly reduced in brain slices from GluRε2−/− mice (Fig. 4). Secondly, injection of IP3 into a neuron in the cortical slice from a GluRε2−/− brain resulted in very few neuronal domains being observed, whereas a similarly performed injection into a neuron in a cortical slice from a wild-type brain promptly resulted in the appearance of many neuronal domains (Fig. 7). The ‘neuronal domain’ is the functional cellular coupling based on cell-to-cell interaction via gap-junction channels (Yuste et al., 1995; Kandler and Katz, 1998). The present experiments show that it is likely that NMDA receptors are involved in the regulation of gap-junction channels between immature neurons.

Our results have clearly shown that NMDA receptors are involved in the establishment of cell-to-cell interactions based on gap-junction channels, rather than potentiating intracellular calcium elevation driven by IP3. The frequency of occurrence of spontaneous neuronal domains was significantly reduced in GluRε2−/− mouse sections, even though the elevation of [Ca2+]i in single cells was frequently observed in brain slices from GluRε2−/− mice (Fig. 5). In addition, very few neuronal domains were observed following IP3 injection into single neuronal cells in GluRε2−/− mice brain slices, even though this injection resulted in the elevation of [Ca2+]i in the injected cell itself.

Kander and Katz (Kander and Katz, 1998) proposed that the propagation of calcium elevation in the neuronal domain is mediated by the passage of IP3 via gap-junction channels. Based on this scenario, the elevation of intracellular IP3 driven by G-protein activation occurs to the same degree in the GluRε2−/− mouse neocortex as in the wild-type mouse, but the propagation of IP3 into the surrounding neurons via gap junctions is virtually absent. Taken together, the decreased frequency of appearance of neuronal domains (10% of that seen in wild-type sections) observed in GluRε2−/− mice suggests that the putative signal driven by the opening of the NMDA receptor channels plays a critical role in the regulation of the generation or the activation of gap-junction channels. Further studies need to be conducted to assess whether NMDA receptor would directly regulate the opening of gap-junction channels between immature neurons. Neurobiotin injection must be a first choice; therefore we have injected neurobiotin along with IP3 into developing cortical neurons from wild-type mice. Unfortunately, we have not been able to detect any dye coupling after the regular staining protocol.

Expression of gap-junction channels in the cortical plate of the mammalian cerebral neocortex has been confirmed in several studies (Peinado et al., 1993a,b; Nadarajah et al., 1997; Sohl et al., 1998). Because gap-junction channels allow the passage of compounds of <1 kDa, as well as of ions, cells coupled via gap-junction channels are able to exchange cations such as sodium and potassium, or second messenger molecules such as calcium and IP3. These physiological events result in the electrical (Gutnick and Prince, 1981; Deans et al., 2001) or metabolic (Saez et al., 1989; Boitano et al., 1992) synchronization of the coupled cells.

With the above studies in mind, we attempted to analyze the expression of gap-junction channels in the developing neocortex of the GluRε2−/− mouse in comparison to the wild-type mouse. Gap-junction channels contain Cx molecules (White and Paul, 1999; Rörig and Feller, 2000). In the developing neocortex, the expression of Cx26, Cx32, Cx36 and Cx43, has been reported (Nadarajah et al., 1997; Naus and Bani-Yaghoub, 1998; Sohl et al., 1998; Bittman et al., 2002). The neuron-specific gap-junction protein Cx36 was recently found and its sequential change of expression was investigated (Sohl et al., 1998). The expression of Cx36 increases to a maximum up to 7 days postnatally and decreases thereafter (Al-Ubaidi et al., 2000; Belluardo et al., 2000; Bittman et al., 2002). Cx26 is expressed in neuronal progenitor cells in the developing neocortex (Bittman and LoTurco, 1999). As for Cx32, Cx36 and Cx43, the analysis is not possible at this stage because of the lack of availability of specific antibodies. We carried out an immunohistochemical study on Cx26 expression in the neocortex of wild-type and GluRε2−/− mice, but no significant differences were found (Fig. 8). It is known that several other factors, such as acidification (pH), [Ca2+]i and protein-phosphorylation regulate the opening of gap-junction channels (Rörig and Feller, 2000). The putative signal driven by NMDA receptor channels may induce the opening of gap-junction channels rather than promote the upregulation of gene expression of these gap-junction channels.

Our current study may indicate that the generation of ‘neuronal domains’ could be separated into two parts. First comes the establishment of cell-to-cell interactions via gap-junction channels, presumably promoted by the signal driven by the opening of NMDA receptor channels. Then comes the execution of coordinated neural activity initiated by the activation of Gq protein as previously proposed (Kandler and Katz, 1998). Our study demonstrates an unexpected function of NMDA receptors, which act as inducers of the establishment of functional gap-junction channels in the developing cerebral cortex of the mammalian brain.

Notes

This work was supported by a Grant-in-Aid for Scientific Research from the Ministry of Education, Science, Sports and Culture of Japan.

Address correspondence to Tatsuhiro Hisatsune, Department of Integrated Biosciences, The University of Tokyo, Bioscience Bldg 402, 5-1-5 Kashiwanoha, Kashiwa, Chiba 277-8562, Japan. Email: hisatsune@k.u-tokyo.ac.jp.

Supplementary Material

Supplementary material can be found at: http://www.cercor.oupjournals.org

Table 1

Frequency of neuronal domains induced by microinjection of IP3

Category Wild-type (GluRe2+/+Knockout (GluRe2−/−
In this study, we used three pairs of a GluRε2 knockout and a wild-type littermate from three individual dams. One slice was prepared from each of the six mice and approximately eight trials were performed for each slice. In most cases, the microinjection of IP3 into neurons of wild-type mice produced a neuronal domain. On the other hand, the microinjection of IP3 into brain slices of GluRε2 knockout mice mostly induced a calcium spike only in the injected cell and this calcium spike did not spread to the neighboring neurons. This event is demonstrated in Figure 7b and call ‘single spike’. Frequency for neuronal domain = (neuronal domain)/(total trial) × 100. 
Total trial 22 25 
Neuronal domain 20 
Single spike 20 
Frequency for neuronal domain (%) 91 20 
Category Wild-type (GluRe2+/+Knockout (GluRe2−/−
In this study, we used three pairs of a GluRε2 knockout and a wild-type littermate from three individual dams. One slice was prepared from each of the six mice and approximately eight trials were performed for each slice. In most cases, the microinjection of IP3 into neurons of wild-type mice produced a neuronal domain. On the other hand, the microinjection of IP3 into brain slices of GluRε2 knockout mice mostly induced a calcium spike only in the injected cell and this calcium spike did not spread to the neighboring neurons. This event is demonstrated in Figure 7b and call ‘single spike’. Frequency for neuronal domain = (neuronal domain)/(total trial) × 100. 
Total trial 22 25 
Neuronal domain 20 
Single spike 20 
Frequency for neuronal domain (%) 91 20 
Figure 1.

Neuronal domains observed in a regular confocal imaging system. Fluorescent images of the primary somatosensory cortex of neonatal mice. Changes in fluo-4 fluorescence were detected using Leica CLSM. An inverted microscope and 20× objective lens were used. (a) Images were obtained from ~15 min (900 s) of scanning of a brain slice. No neuronal domains are apparent in the image taken at time 42 s. Arrowheads in the remaining images indicate typical neuronal domains that occurred spontaneously. Most neuronal domains were generated in primary 2/3 layer. Scale bar = 150 μm. (b) Enlarged time-lapse images from the area in (a) surrounded by a box. The second image (214 s) in (b) is an enlarged image taken at the same time as the image shown in (a). Scale bar = 50 μm.

Figure 1.

Neuronal domains observed in a regular confocal imaging system. Fluorescent images of the primary somatosensory cortex of neonatal mice. Changes in fluo-4 fluorescence were detected using Leica CLSM. An inverted microscope and 20× objective lens were used. (a) Images were obtained from ~15 min (900 s) of scanning of a brain slice. No neuronal domains are apparent in the image taken at time 42 s. Arrowheads in the remaining images indicate typical neuronal domains that occurred spontaneously. Most neuronal domains were generated in primary 2/3 layer. Scale bar = 150 μm. (b) Enlarged time-lapse images from the area in (a) surrounded by a box. The second image (214 s) in (b) is an enlarged image taken at the same time as the image shown in (a). Scale bar = 50 μm.

Figure 2.

Effects of drugs on spontaneous appearance of neuronal domains of brain slices from ICR mice. One rhomboid deflection in the graphs indicates the occurrence of a neuronal domain (coordinated calcium rise in more than five neighboring neurons) and the horizontal axis shows time. The graphs thus indicate frequency of neuronal domains sequentially. All drugs were bath-applied during the time indicated by the bar (~20 min) with the concentrations used given in the experimental procedures. A gap-junction inhibitor, halothane (Halo), significantly blocked the generation of neuronal domains. Cont., control experiment without any drug; TTX, tetrodotoxin.

Figure 2.

Effects of drugs on spontaneous appearance of neuronal domains of brain slices from ICR mice. One rhomboid deflection in the graphs indicates the occurrence of a neuronal domain (coordinated calcium rise in more than five neighboring neurons) and the horizontal axis shows time. The graphs thus indicate frequency of neuronal domains sequentially. All drugs were bath-applied during the time indicated by the bar (~20 min) with the concentrations used given in the experimental procedures. A gap-junction inhibitor, halothane (Halo), significantly blocked the generation of neuronal domains. Cont., control experiment without any drug; TTX, tetrodotoxin.

Figure 3.

Non-appearance of calcium response for exposure to NMDA of brain slices from GluRε2 knockout mice. (a) Neuronal cells responded strongly when NMDA + glycine (Gly) was bath-applied to a wild-type mouse (GluRε2+/+) brain slice. ROI, region of interest, where the averaged data are presented in Figure 3b. Scale bar = 100 μm. (b) Sequential graphs of fluo-4 fluorescent intensity. Response to NMDA of GluRε2+/+ mice brain slice (upper trace) was completely inhibited by APV (lower trace). In GluRε2−/− mice, calcium influx was absent (middle trace). Because glycine was applied with NMDA as a co-agonist of the NMDA receptor, strychnine (Stry) was applied in conjunction with them in order to inhibit any effect via glycine receptor activation. All drugs were used at 100 μM.

Figure 3.

Non-appearance of calcium response for exposure to NMDA of brain slices from GluRε2 knockout mice. (a) Neuronal cells responded strongly when NMDA + glycine (Gly) was bath-applied to a wild-type mouse (GluRε2+/+) brain slice. ROI, region of interest, where the averaged data are presented in Figure 3b. Scale bar = 100 μm. (b) Sequential graphs of fluo-4 fluorescent intensity. Response to NMDA of GluRε2+/+ mice brain slice (upper trace) was completely inhibited by APV (lower trace). In GluRε2−/− mice, calcium influx was absent (middle trace). Because glycine was applied with NMDA as a co-agonist of the NMDA receptor, strychnine (Stry) was applied in conjunction with them in order to inhibit any effect via glycine receptor activation. All drugs were used at 100 μM.

Figure 4.

Reduction of spontaneous neuronal domains in GluRε2 knockout mice. The frequency of spontaneously occurring neuronal domains was compared between GluRε2+/+ and GluRε2−/− mice. These data were obtained from littermates. (a) In GluRε2+/+ mice (upper trace), the frequency was comparable to the results from the ICR mouse strain shown in Figure 2. In GluRε2−/− mice, the frequency of neuronal domains was markedly reduced (lower trace). The imaging data are also available as supplementary movie 1. (b) Mean frequency from the four individual trials is shown with SEM. Longitudinal axis shows the frequency of neuronal domain per minute (P < 0.01, t-test).

Reduction of spontaneous neuronal domains in GluRε2 knockout mice. The frequency of spontaneously occurring neuronal domains was compared between GluRε2+/+ and GluRε2−/− mice. These data were obtained from littermates. (a) In GluRε2+/+ mice (upper trace), the frequency was comparable to the results from the ICR mouse strain shown in Figure 2. In GluRε2−/− mice, the frequency of neuronal domains was markedly reduced (lower trace). The imaging data are also available as supplementary movie 1. (b) Mean frequency from the four individual trials is shown with SEM. Longitudinal axis shows the frequency of neuronal domain per minute (P < 0.01, t-test).

Figure 5.

Independently elicited calcium spikes in neurons from GluRε2 knockout mice. In GluRε2−/− mice, a number of neurons show independently elicited calcium spikes. (a) Five images show typical examples of single calcium spikes. Two cells indicated by arrowheads (second and fourth images) show calcium spikes. (b) Time-lapse value of a calcium flux of the two cells. The upper trace is from the cell marked in the second image and the lower trace from the cell in the fourth image. These two cells did not respond simultaneously. Scale bar = 30 μm.

Figure 5.

Independently elicited calcium spikes in neurons from GluRε2 knockout mice. In GluRε2−/− mice, a number of neurons show independently elicited calcium spikes. (a) Five images show typical examples of single calcium spikes. Two cells indicated by arrowheads (second and fourth images) show calcium spikes. (b) Time-lapse value of a calcium flux of the two cells. The upper trace is from the cell marked in the second image and the lower trace from the cell in the fourth image. These two cells did not respond simultaneously. Scale bar = 30 μm.

Figure 6.

A neuronal domain induced by the microinjection of IP3 into a neuron. (a) As a control observation, a spontaneous neuronal domain was also monitored by the high-speed confocal imaging system. It clearly shows a calcium wave spreading from a central neuron. Scale bar = 40 μm. (b) Calcium dynamics after the microinjection of IP3; −55 ms indicates time before the microinjection. At time 0 ms, IP3 was injected. A number of confocal images were acquired from time 0–110 ms. These images were averaged and shown in one image (indicated as 55 ms). The image at 165 ms shows the clear increase of intracellular calcium in an injected cell. At 495 ms, the neuronal domain has spread to the neighboring cells. Scale bar = 30 μm. (c) When Ca2+ ions are injected into a single cell, a calcium spike is produced in the injected cell, but a neuronal domain is not initiated. Scale bar = 10 μm.

Figure 6.

A neuronal domain induced by the microinjection of IP3 into a neuron. (a) As a control observation, a spontaneous neuronal domain was also monitored by the high-speed confocal imaging system. It clearly shows a calcium wave spreading from a central neuron. Scale bar = 40 μm. (b) Calcium dynamics after the microinjection of IP3; −55 ms indicates time before the microinjection. At time 0 ms, IP3 was injected. A number of confocal images were acquired from time 0–110 ms. These images were averaged and shown in one image (indicated as 55 ms). The image at 165 ms shows the clear increase of intracellular calcium in an injected cell. At 495 ms, the neuronal domain has spread to the neighboring cells. Scale bar = 30 μm. (c) When Ca2+ ions are injected into a single cell, a calcium spike is produced in the injected cell, but a neuronal domain is not initiated. Scale bar = 10 μm.

Figure 7.

Microinjection of IP3 failed to produce neuronal domain in GluRε2−/− mice. (a) In a cortical slice from a GluRε2+/+ mouse, a neuronal domain was produced after microinjection of IP3. The pattern is quite similar to that shown in Figure 6b. Scale bar = 20 μm. (b) In a cortical slice from a GluRε2−/− mouse, although IP3-injected cells responded quite well, the propagation of a calcium wave towards the neighbor cells did not occur. Scale bar = 20 μm.

Microinjection of IP3 failed to produce neuronal domain in GluRε2−/− mice. (a) In a cortical slice from a GluRε2+/+ mouse, a neuronal domain was produced after microinjection of IP3. The pattern is quite similar to that shown in Figure 6b. Scale bar = 20 μm. (b) In a cortical slice from a GluRε2−/− mouse, although IP3-injected cells responded quite well, the propagation of a calcium wave towards the neighbor cells did not occur. Scale bar = 20 μm.

Figure 8.

Immunohistchemical staining of connexin26 in GluRε2+/+ and GluRε2−/−. Immunohistochemical analysis using anti-Cx26 antibody in neonatal wild-type GluRε2+/+ (a) and GluRε2−/− mice (b). In both types of mouse, Cx26 widely expresses in developing cortical plate (cp). There are no significant differences in Cx26 staining between the two genotypes. VZ, ventricular zone. Scale bar = 150 μm.

Figure 8.

Immunohistchemical staining of connexin26 in GluRε2+/+ and GluRε2−/−. Immunohistochemical analysis using anti-Cx26 antibody in neonatal wild-type GluRε2+/+ (a) and GluRε2−/− mice (b). In both types of mouse, Cx26 widely expresses in developing cortical plate (cp). There are no significant differences in Cx26 staining between the two genotypes. VZ, ventricular zone. Scale bar = 150 μm.

References

Al-Ubaidi MR, White TW, Ripps H, Poras I, Avner P, Gomes D, Bruzzone R (
2000
) Functional properties, developmental regulation, and chromosomal localization of murine connexin36, a gap-junctional protein expressed preferentially in retina and brain.
J Neurosci Res
 
59
:
813
–826.
Belluardo N, Mudo G, Trovato-Salinaro A, Le Gurun S, Charollais A, Serre-Beinier V, Amato G, Haefliger JA, Meda P, Condorelli DF (
2000
) Expression of connexin36 in the adult and developing rat brain.
Brain Res
 
865
:
121
–138.
Bittman K, Lo Turco J (
1999
) Differential regulation of connexin 26 and 43 in murine neocortical precursors.
Cereb Cortex
 
9
:
188
–195.
Bittman K, Becker D, Cicirata F, Parnavelas J (
2002
) Connexin expression in homotypic and heterotypic cell coupling in the developing cerebral cortex.
J Comp Neurol
 
443
:
201
–212.
Boitano S, Dirksen ER, Sanderson MJ (
1992
) Intercellular propagation of calcium waves mediated by inositol trisphosphate.
Science
 
258
:
292
–295.
Chiaia N, Fish S, Bauer WR, Bennett-Clarke C, Rhoades RW (
1992
) Postnatal blockade of cortical activity by tetrodotoxin does not disrupt the formation of vibrissa-related patterns in the rat’s somatosensory cortex.
Dev Brain Res
 
66
:
244
–250.
Deans MR, Gibson JR, Sellitto C, Connors BW, Paul DL (
2001
) Synchronous activity of inhibitory networks in neocortex requires electrical synapses containing connexin36.
Neuron
 
31
:
477
–485.
Des Rosiers MH, Sakurada O, Jehle J, Shinohara M, Kennedy C, Sokoloff L (
1978
) Functional plasticity in the immature striate cortex of the monkey shown by the [14C]deoxyglucose method.
Science
 
200
:
447
–449.
Flint AC, Liu X, Kriegstein AR (
1998
) Nonsynaptic glycine receptor activation during early neocortical development.
Neuron
 
20
:
43
–53.
Gutnick MJ, Prince DA (
1981
) Dye coupling and possible electrotonic coupling in the guinea pig neocortical slice.
Science
 
211
:
67
–70.
Henderson T, Woolsey T, Jacquin M (
1992
) Infraorbital nerve blockage from the birth does not disrupt central trigeminal pattern formation in the rat.
Dev Brain Res
 
66
:
146
–152.
Iwasato T, Datwani A, Wolf AM, Nishiyama H, Taguchi Y, Tonegawa S, Knopfel T, Erzurumlu RS, Itohara S (
2000
) Cortex-restricted disruption of NMDAR1 impairs neuronal patterns in the barrel cortex.
Nature
 
406
:
726
–731.
Kandler K, Katz LC (
1998
) Coordination of neuronal activity in developing visual cortex by gap junction-mediated biochemical communication.
J Neurosci
 
18
:
1419
–1427.
Kutsuwada T, Sakimura K, Manabe T, Takayama C, Katakura N, Kushiya E, Natsume R, Watanabe M, Inoue Y, Yagi T, Aizawa S, Arakawa M, Takahashi T, Nakamura Y, Mori H, Mishina M (
1996
) Impairment of suckling response, trigeminal neuronal pattern formation, and hippocampal LTD in NMDA receptor epsilon 2 subunit mutant mice.
Neuron
 
16
:
333
–344.
Lo Turco JJ, Blanton MG, Kriegstein AR (
1991
) Initial expression and endogenous activation of NMDA channels in early neocortical development.
J Neurosci
 
11
:
792
–799.
Luhmann HJ, Prince DA (
1991
) Postnatal maturation of the GABAergic system in rat neocortex.
J Neurophysiol
 
65
:
247
–263.
Mayer ML, Westbrook GL, Guthrie PB (
1984
) Voltage-dependent block by Mg2+ of NMDA responses in spinal cord neurones.
Nature
 
309
:
261
–263.
Meister M, Wong R, Baylor D, Shatz C (
1991
) Synchronous bursts of action potentials in ganglion cells of the developing mammalian retina.
Science
 
252
:
939
–943.
Miyakawa N, Uchino S, Yamashita T, Okada H, Nakamura T, Kaminogawa S, Miyamono Y, Hisatsune T (
2002
) A glycine receptor antagonist, strychnine, blocked NMDA receptor activation in the neonatal mouse neocortex.
Neuroreport
 
13
:
1667
–1673.
Monyer H, Burnashev N, Laurie DJ, Sakmann B, Seeburg PH (
1994
) Developmental and regional expression in the rat brain and functional properties of four NMDA receptors.
Neuron
 
12
:
529
–540.
Mori H, Mishina M (
1995
) Structure and function of the NMDA receptor channel.
Neuropharmacology
 
34
:
1219
–1237.
Nadarajah B, Jones AM, Evans WH, Parnavelas JG (
1997
) Differential expression of connexins during neocortical development and neuronal circuit formation.
J Neurosci
 
17
:
3096
–3111.
Naus CC, Bani-Yaghoub M (
1998
) Gap junctional communication in the developing central nervous system.
Cell Biol Int
 
22
:
751
–763.
Nowak L, Bregestovski P, Ascher P, Herbet A, Prochiantz A (
1984
) Magnesium gates glutamate-activated channels in mouse central neurones.
Nature
 
307
:
462
–465.
O’Leary D, Ruff N, Dyck R (
1994
) Development, critical period plasticity, and adult reorganizations of mammalian somatosensory systems.
Curr Biol
 
4
:
535
–544.
Owens DF, Boyce LH, Davis MB, Kriegstein AR (
1996
) Excitatory GABA responses in embryonic and neonatal cortical slices demonstrated by gramicidin perforated-patch recordings and calcium imaging.
J Neurosci
 
16
:
6414
–6423.
Peinado A, Yuste R, Katz LC (
1993
) Gap junctional communication and the development of local circuits in neocortex.
Cereb Cortex
 
3
:
488
–498.
Peinado A, Yuste R, Katz LC (
1993
) Extensive dye coupling between rat neocortical neurons during the period of circuit formation.
Neuron
 
10
:
103
–114.
Rabacchi S, Bailly Y, Delhaye-Bouchaud N, Mariani J (
1992
) Involvement of the N-methyl-d-aspartate (NMDA) receptor in synapse elimination during cerebellar development.
Science
 
256
:
1823
–1825.
Rakic P (
1976
) Prenatal genesis of connections subserving ocular dominance in the rhesus monkey.
Nature
 
261
:
467
–471.
Rörig B, Feller MB (
2000
) Neurotransmitters and gap junctions in developing neural circuits.
Brain Res Brain Res Rev
 
32
:
86
–114.
Saez J, Connor J, Spray D, Bennett M (
1989
) Hepatocyte gap junctions are permeable to the second messenger, inositol 1,4,5-trisphosphate, and to calcium ions.
Proc Natl Acad Sci USA
 
86
:
2708
–2712.
Shatz C, Stryker MP (
1988
) Prenatal tetrodotoxin infusion blocks segregation of retinogeniculate afferents.
Science
 
242
:
87
–89.
Simon DK, Prusky GT, O’Leary DD, Constantine-Paton M (
1992
) N-methyl-d-aspartate receptor antagonists disrupt the formation of a mammalian neural map.
Proc Natl Acad Sci USA
 
89
:
10593
–10597.
Sohl G, Degen J, Teubner B, Willecke K (
1998
) The murine gap junction gene connexin 36 is highly expressed in mouse retina and regulated during brain development.
FEBS Lett
 
428
:
27
–31.
Wainai T, Takeuchi T, Seo N, Mishina M (
2001
) Regulation of acute nociceptive responses by the NMDA receptor GluRε2 subunit.
Neuroreport
 
12
:
3169
–3172.
Watanabe M, Inoue Y, Sakimura K, Mishina M (
1992
) Developmental changes in distribution of NMDA receptor channel subunit mRNAs.
Neuroreport
 
3
:
1138
–1140.
Weliky M, Katz LC (
1997
) Disruption of orientation tuning in visual cortex by artificially correlated neuronal activity.
Nature
 
386
:
680
–685.
White TW, Paul DL (
1999
) Genetic diseases and gene knockouts reveal diverse connexin functions.
Annu Rev Physiol
 
61
:
283
–310.
Wiesel TN, Hubel DH (
1965
) Comparison of the effects of unilateral and bilateral eye closure on cortical unit responses in kittens.
J Neurophysiol
 
28
:
1029
–1040.
Woolsey T, Wann J (
1976
) Areal changes in mouse cortical barrels following vibrissal damage at different postnatal ages.
J Comp Neurol
 
170
:
53
–66.
Yamada K, Hisatsune T, Uchino S, Nakamura T, Kudo Y, Kaminogawa S (
1999
) NMDA receptor mediated Ca2+ responses in neurons differentiated from p53−/− immortalized murine neural stem cells.
Neurosci Lett
 
264
:
165
–167.
Yuste R, Sur M (
1999
) Development and plasticity of the cerebral cortex: from molecules to maps.
J Neurobiol
 
41
:
1
–6.
Yuste R, Peinado A, Katz LC (
1992
) Neuronal domains in developing neocortex.
Science
 
257
:
665
–669.
Yuste R, Nelson DA, Rubin WW, Katz LC (
1995
) Neuronal domains in developing neocortex: mechanisms of coactivation.
Neuron
 
14
:
7
–17.