In the developing cerebral cortex, neuronal nitric oxide synthase (nNOS) is expressed abundantly, but temporarily. During the early postnatal stage, cortical neurons located in the multi-layered structure of the cortical plate start forming well-organized cortical circuits, but little is known about the molecular machinery for layer-specific circuit formation. To address the involvement of nitric oxide (NO), we utilized a new NO indicator (DAR-4M) and developed a protocol for the real-time imaging of NO produced in fresh cortical slices upon N-methyl-D-aspartic acid stimulation. At postnatal day 0 (P0), NO production was restricted to the deep layers (layers V and VI) of the somatosensory cortex where transient synapses are formed. At P10, the production of NO was expanded to layer IV where large numbers of thalamocortical axons form synapses. The pattern of NO production could correspond to active sites for synaptic formation. This study is the first clear demonstration of NO production in the postnatal mouse neocortex. The findings presented may reflect a function of NO in relation to the layer-specific development of neural circuits in the neocortex.
Rodent thalamocortical networks are formed according to genetically oriented and/or activity-dependent processes that take place during postnatal development (Sur and Leamey, 2001; Lopez-Bendito and Molnar, 2003). Thalamocortical axons penetrate through the internal capsule in the prenatal stage, eventually connecting with layer IV cortical neurons (Higashi et al., 1999; Molnar et al., 2000). Molecular mechanisms that govern the projection of thalamocortical axons to the developing subplate have been studied extensively (Lopez-Bendito and Molnar, 2003); however, the ‘machinery’ by which connections between thalamocortical axons and cortical neurons are made is mostly unknown. In thalamocortical connections of the somatosensory system, peripheral activities stimulate the sprouting of thalamocortical axons. The N-methyl-D-aspartic acid (NMDA)-receptor is a candidate for mediating such an activity-dependent development of neural circuitry. Iwasato et al. have developed a cortex-specific deletion of the NMDA-R1 gene, and have demonstrated the disruption of barrel boundaries in the somatosensory cortex of NMDA-R1 knock-out mice (Iwasato et al., 2000). In a subsequent study, defective sprouting of thalamocortical axons in this strain was demonstrated (Datwani et al., 2002).
Nitric oxide (NO) is a gaseous neurotransmitter produced by neuronal nitric oxide synthase (nNOS). In the developing cortex, nNOS is abundantly, but temporarily, expressed in the developing cortical plate (Bredt and Snyder, 1994; Santacana et al., 1998). NOS activity at this stage has been suggested by NADPH-diaphorase (NADPH-d) histochemistry in both experimental animals and human fetus (Bredt et al., 1991; Dawson et al., 1991; Derer and Derer., 1993; Mitrovic and Schachner, 1996; Van Eden et al., 1996; Yan et al., 1996; Vercelli at al., 1999). It is well accepted that nNOS requires elevation of the calcium concentration in order to be activated (Knowles et al., 1989; Bredt and Snyder, 1990). The elevation of intracellular calcium concentration ([Ca2+]i) after the opening of NMDA-receptor channels is regarded as a typical upstream signal leading to NO production (Garthwaite et al., 1988; Bredts and Snyder, 1989; Bohme et al., 1991). During development, NO is thought to act as a retrograde signal, serving to stabilize and refine afferent connections via an activity-dependent mechanism (Wu et al., 1994; Cramer et al., 1996). In both young and adult primates, region- and layer-specific NOS activity in cerebral cortex has been suggested from anti-nNOS staining and NADPH-d histochemistry studies (Rivier and Clarke, 1997; Wong-Riley et al., 1998; Barone and Kennedy, 2000; Wiencken and Casagrande, 2000; Benavides-Piccione and DeFelipe, 2003). As mentioned above, circuit-specific production of NO has been suggested both in developmental and adult cerebral cortex; however, the production of NO from such a cortical tissue had not been demonstrated yet.
Since NO is a gaseous molecule with a very short half-life, the dynamics of NO in developing cerebral cortex have been difficult to measure in a real-time format. DAR-4M is a newly developed NO indicator with a structure based on the rhodamine chromophore, whose signal-to-noise ratio and the pH sensitivity properties have been improved (Kojima et al., 2001) with respect to conventional NO indicators (Kojima et al., 1998a,b). In this communication, we show for the first time the real-time spectrum of NO production in the developing cortical plate, with fine spatiotemporal resolution allowing identification of the layer-specific release of NO.
Materials and Methods
Preparation of Slices
ICR mice [postnatal day (P) 0, P10 and P30; Sankyo Laboratory Service, Tokyo, Japan] were anesthetized by hypothermia or with ether, and whole brains were dissected free and placed in cold low-calcium artificial cerebrospinal fluid (ACSF). All experiments were carried out in accordance with the guidelines for Animal Experiments of the Graduate School of Frontier Sciences, The University of Tokyo. Coronal slices (400 μm) that included the developing somatosensory cortex (Bayer and Altman, 1991; Schambra et al., 1992; Jacobowitz and Abbott, 1998) and tangential slices (400 μm) that included the barrel in layer IV of somatosensory cortex of P10 mice (Welker and Woolsey, 1974) were prepared in low-calcium ACSF at 0°C with the aid of a microslicer (DOSAKA EM, Kyoto, Japan). The slices were then incubated in standard ACSF at 37°C for 30 min. The low-calcium ACSF contained 124 mM NaCl, 2.5 mM KCl, 26 mM NaHCO3, 4.5 mM MgCl2, 1.25 mM NaH2PO4, 0.1 mM CaCl2 and 10 mM glucose. Standard ACSF used in experiments contained 124 mM NaCl, 2.5 mM KCl, 26 mM NaHCO3, 1 mM MgCl2, 1.25 mM NaH2PO4, 2 mM CaCl2 and 10 mM glucose. Each type of ACSF was equilibrated with 95% O2/5% CO2.
Loading of Fluorescence Indicator
In order to investigate the location of NO production, real-time NO imaging was performed using a fluorescent NO-indicator. A selective fluorescent indicator for NO (DAF-2; Kojima et al., 1998a) was developed to visualize the location of NO production (Kojima et al., 1998b), in which DAF-2 traps NO and turns to DAF-2 T (triazole form of DAF-2) emitting a strong fluorescence signal. Initially, we had used DAF-2 in the real-time imaging of NO in the somatosensory cortex of developing mouse, but a satisfactory result could not be obtained, probably due to the insufficient production of NO in this preparation and the low signal-to-noise ratio of the DAF-2 indicator. A further major problem was associated with the fact that the laser light produces a strong fluorescence signal, presumably from photoactivated production of DAF-2 T. Indeed, it was previously reported that DAF-2 is sensitive to photoactivation (Broillet et al., 2001).
For these reasons, we used a newly developed alternative NO-indicator, DAR-4M (Kojima et al., 2001), for NO imaging of developing brain slices. After 30 min incubation, the slices were loaded for 90 min at 37°C with DAR-4M AM (10 μM) in standard ACSF containing 0.01% cremophor EL. Loaded slices were rinsed in standard ACSF for 30 min at 37°C prior to use in imaging experiments. DAR-4M reacts irreversibly with NO to produce the triazole form of DAR-4M, namely DAR-4M T. We did not observe any photoactivated production of strong fluorescence signal. For dual imaging of NO and [Ca2+]i, slices were incubated simultaneously with DAR-4M AM (10 μM) and fluo-4 AM (10 μM) in 0.01% cremophor EL for 90 min at 37°C. Calcium imaging for cortical slices was based on the method reported previously (Okada et al., 2003).
Imaging of Slices by Confocal Laser Scanning Microscopy
Each slice used for imaging experiments was placed in a purpose-built perfusion chamber (depth: 4 mm; diameter: 14 mm: volume: 300 μl, flow rate: 55.56 μl/s) and held in position with a slice anchor. The slices were examined using a confocal laser scanning microscope (TCS-SL, Leica, Mannheim, Germany) equipped with adjustable slits enabling the user to specify the wavelength range for the detection of emitted light. In NO imaging, loaded slices were excited with 543 nm light and the emitted signal was collected in the range of 575–700 nm, for the maximum collection from the DAR-4M T-derived fluorescence signal. In NO and Ca2+ dual imaging, the collection spectrum for DAR-4M T was changed from 575–700 nm to 600–700 nm, because of the weak signal from Ca2+, which is excited at 543 nm and collected at 575–600nm. For the collection from Ca2+-related fluorescence signals, slices were excited at 488 nm and signals were collected at 500–550 nm. In this condition, we did not detect any signal from DAR-4M T. In NO and Ca2+ dual imaging, each scan was performed separately as a Sequential mode (TCS SL system, Leica).
In each imaging protocol, optical signals were collected with a TCS SL system (Leica) and fluorescence images were obtained sequentially at intervals of 6 s. A frame was created from the average of two scans (it took 1.663 s to perform one scan). To evaluate whether or not the NO-indicator DAR-4M was evenly distributed throughout fresh cortical slices, we used a synthetic NO donor, NOC12, and confirmed that the DAR-4M-related fluorescence signals were detected in all cortical layers and areas just after NOC12 application (data not shown). To visualize the activity-dependent production of NO, NMDA (Sigma, St Louis, MO) (100 μM) was bath-applied.
In NO imaging with treatment of inhibitors, DAR-4M-loaded cortical slices were treated with NG-nitro-L-arginine methyl ester (L-NAME; Dojindo, Kumamoto, Japan) (5 mM), NOS universal inhibitor, or 7-nitroindazole (7-NI; Sigma, St Louis, MO) (100 μM), nNOS-specific inhibitor, before starting NO imaging for 30 min. After the treatment, the slices were placed in chamber and perfused with ACSF including inhibitors during NO imaging respectively. NMDA (100 μM) was dissolved in ACSF including inhibitors, and bath-applied. In NO imaging under the condition without inhibitors, all drugs were dissolved in ACSF.
Two regions (region1, region2) of the fluorescence images were selected as upper layers and deep layers of the cortical plate by means of the easily identifiable granule cell zone. Each region was subdivided into five subdivisions (a–e). The fluorescence intensity of each subdivision was converted into values using the Quantification function of the TCS SL system (Leica), following which the values of fluorescence intensity for each region were calculated using Microsoft Excel software as the average of those five subdivisions. Since DAR-4M reacts irreversibly with NO to form DAR-4M T, the level of DAR-4M T production per unit time was obtained by subtracting from the fluorescence intensity at one point from that of the point which preceded it. Time-points were graphed for every five calculated values. Levels of calcium response and NO production were estimated by subtracting the fluorescence intensity at the baseline from that at the peak of the calcium response and at the end of NO imaging respectively. The calculated values of calcium response and NO production were compared between upper and deep layers.
Single Cell Imaging
Coronal slices of P10 mice were prepared as described above. One slice was fixed in a recording chamber (∼0.2 ml volume, RC-26GLP; Warner Instruments, Hamden, CT) under nylon strings attached to a U-shaped platinum frame, then submerged in, and continuously perfused with, the standard ACSF at a flow rate of 1–2 ml/min. Neuronal cells in layer IV could be identified as cells situated between large pyramidal neurons at layer V and small pyramidal neurons at layer II/III.
Patch electrodes were fabricated from borosilicate glass capillaries of 1.2 mm outer diameter (1B120F-4; World Precision Instruments, Sarasota, FL) with a programmable puller (P-87; Sutter Instrument, Novato, CA). The composition of the pipette internal solution was (mM): potassium gluconate, 120; NaCl, 6; CaCl2, 5; MgCl2, 2; MgATP, 2; NaGTP, 0.3; EGTA, 10; HEPES, 10; pH 7.2 with KOH. The intracellular solution also included 10 mM DAR-4M for NO imaging of the cells recorded. The tip resistance of the electrodes ranged from 5 to 10 MΩ when filled with this solution. Data were recorded at room temperature with a CEZ-2400 amplifier (Nihon-kohden, Tokyo, Japan). The membrane current was sampled on-line at 4 kHz (PowerLab; AD Instruments, Grand Junction, CO) after filtering at 2 kHz to achieve synchrony with NO imaging using AQUACOSMOS system (Hamamatsu) (Okada et al., 2003) and stored on the hard disk of a personal computer. The data were analyzed off-line with Igor Pro 4.01 (WaveMetrics, Lake Oswego, OR). NMDA was pressure-applied locally using a PicoPump (PV820; World Precison Instruments) via a glass pipette positioned ∼30 μm from the soma of the recorded cell (Fukuda et al., 2003).
Mice [postnatal day (P) 0] were perfused with phosphate-buffered saline (PBS) (in mM: 137 NaCl, 8.1 NaH2PO4, 2.68 KCl, 1.47 KH2PO4) followed by 4% freshly depolymerized paraformaldehyde in PBS. Brains were removed and postfixed in 4% paraformaldehyde in PBS for 1 h–1 day at 4°C. Tissues were cryoprotected for 1–2 days and then embedded with OCT compound. Coronal sections (40 μm) were then cut on a cryostat (HM505E; MICROM, Walldorf, Germany). For the immunostaining for nNOS protein, the sections were washed in PBS for 10 min and then blocked for 30 min in blocking buffer containing 5% normal goat serum and 0.3% Triton X-100 in PBS. The sections were incubated for 1 day at 4°C with anti-mouse nNOS polyclonal antibody (rabbit IgG, 1:1000 dilution) and then washed with PBS three times and incubated for 2 h at room temperature with Alexa488-conjugated goat anti-rabbit IgG antibody (1:5000 dilution; Molecular Probes, Eugene, OR). Primary antibodies and secondary antibodies were diluted in buffer containing 5% normal goat serum and 0.1% Triton X-100 in PBS. After being washed three times, the sections were mounted with Immu-mount. Fluorescence images were obtained with a TCS SP2 system (Leica).
After NO imaging, the coronal brain slices were used for NADPH-d histochemistry, Toluidine Blue staining and immunohistochemistry for nNOS protein. The slices were postfixed in 4% paraformaldehyde in PBS for 30 min, and then were rinsed in ice-cold PBS for 10 min. The slices were cryoprotected for 1–2 days and then embedded with OCT compound. The embedded slices were resectioned at a thickness of 16 μm on cryostat and resectioned slices were mounted on poly-L-lysine-coated slides. For NADPH-d histochemistry, the resectioned slice was incubated at 37°C for 2–3 h in a solution containing 5 mg nitroblue tetrazolium (NBT; Wako; Osaka, Japan) and 20 mg β-nicotinamide adenine dinucleotide phosphate (NADPH; Oriental Yeast, Tokyo, Japan) in 20 ml of PBS. The slice was then rinsed in PBS, dehydrated, cleared in xylene and coverslipped in Entellan mounting media. Toluidine Blue Nissl staining, when performed, was carried out by following standard protocols (Sugitani et al., 2002). Immunohistochemical staining for nNOS protein was performed as described above.
After the NO imaging of the tangential slices of P10 mouse, the slices were used for cytochrome oxidase (CO) histochemistry. The slices were postfixed in 4% paraformaldehyde in PB for 30–60 min, and were then cryoprotected for 1–2 days. The slices were incubated at 37°C for 1–3 h in a solution containing 10 mg 3,3′-diaminobenzidine tetrahydrochloride (DAB; Sigma), 0.8 g sucrose and 6 mg cytochrome C (Wako) in 20 ml of PB. After the incubation, the slices were rinsed in PBS. Images of sections stained for TB, NADPH-d and CO histochemistry were acquired using microscope (BX-50WI, Olympus, Tokyo, Japan), fitting with a CAMEDIA digital camera (Olympus).
Production of NO in Developing Cerebral Cortex
Previous studies have clearly demonstrated that nNOS is expressed temporarily and abundantly in the cerebral neocortex during development (Bredt and Snyder, 1994; Santacana et al., 1998; Terada et al., 2001), but there is some discrepancy between these studies concerning the spatiotemporal expression of nNOS. In the light of this, we first assessed the expression of nNOS in the developing cerebral cortex by using early postnatal stage (P0) mice. Animals from this developmental stage were chosen because all three of the above studies presented nNOS immunostaining of brain slices from P0 animals. As shown in Figure 1, two types of nNOS staining are present in the cortical plate. One type corresponds to a particular cortical plate neuron expressing nNOS proteins extensively, while in the other, weak levels of staining reflect the perinuclear localization of nNOS in cortical plate neurons as reported by Bredt and Snyder (1994). Similar patterns of NADPH-d staining have been reported in developing rat cortex (Vercelli et al., 1999).
To assess which type of nNOS actually produces NO, real-time NO imaging with high spatiotemporal resolution was used to detect NO in the developing cerebral cortex by means of the NO fluorescence indicator, DAR-4M. A DAR-4M-loaded fresh brain slice that contained the somatosensory cortex was stimulated by the application of NMDA. Very little DAR-4M fluorescence signal could be detected prior to NMDA application (Fig. 2A), but a significant increase in the fluorescence signal within relatively deep cortical layers was detected after NMDA administration (Fig. 2B). In order to analyze the location of NO production, we selected two regions in the cortical plate (region1: upper layers as supragranular layer; region2: deep layers as infragranular layer) and compared the fluorescence intensity of the NO indicator in each region (Fig. 2C,D). NMDA-induced NO production was relatively higher in the deep layers compared to the upper layers (Fig. 2C). Using the NO-imaging technique, we did not detect NO production from any single, strongly stained nNOS-positive neurons, suggesting that such neurons do not produce NO under these conditions. The amount of NO production per unit of time in the deep layers was more than that in the upper layers (Fig. 2D). This is the first visual demonstration of the production of NO in the developing cortical plate.
Recently, it has also been reported that brain endothelial cells expressed NMDA receptors (Sharp et al., 2003). Since endothelial cells possessing endothelial NOS (eNOS) could be activated by NMDA, NO might be produced from eNOS. To verify whether NO was mainly produced from nNOS, we performed NO imaging with a set of inhibitors: L-NAME (a universal NOS inhibitor), and 7-NI (a nNOS-specific inhibitor) (Fig. 3). NO production in the deep layers was observed when no inhibitors were used (control condition), as shown in Figure 2. NO production in slices treated with L-NAME or 7-NI was weaker than in control slices. As shown in Figure3B, NO production in deep layers of slices treated with 7-NI and L-NAME decreased to 31 and 15% of that in control slices respectively. Therefore, these data suggested that the NO production stimulated by NMDA mainly originated from nNOS.
NO–Calcium Dual Imaging Demonstrates Layer-specific Production of Nitric Oxide in Developing Cortical Plate
It is known that nNOS activation requires the elevation of [Ca2+]i (Garthwaite et al., 1988; Bohme et al., 1991) to produce NO. The reason for NO production only in the deep layers of the cortical plate could stem from a layer-specific NMDA receptor-mediated calcium influx and subsequent elevation of [Ca2+]i in neurons of the developing cortical plate. To test this hypothesis, we developed a dual imaging system for the simultaneous measurement of NO and [Ca2+] in order to assess the correlation between calcium influx and NO production in the developing cortical plate.
During the application of NMDA, calcium influx was evoked both in the upper layers and deep layers (Fig. 4A–C), but the response was more robust in the upper cortical layers. However, NO was produced only in the deep cortical layers (Fig. 4E–G), as shown in Figure 2. The dynamics of calcium influx and NO production are shown in Figure 4D,H, where it can clearly be seen that the production of NO was not detected in the upper layers in spite of the rapid elevation of [Ca2+]i in this region. There are several possibilities to explain this finding. (i) This might have been stemmed from the lack of a related molecule in the upper cortical layers. A candidate molecule is calcineurin, which blocks the over-phosphorylation of nNOS (Leamey et al., 2003). (ii) It is possible that there is a Ca2+-independent production of NO at the deep cortical layers in response to NMDA. (iii) A simple explanation might be that there are more NOS-containing processes in the deeper layers, many of which extend long distances away from the NOS somata.
We further evaluated the extent to which NO was produced only in the deep cortical layers of the developing somatosensory cortex. In all experiments, NO production in the upper layers was less than in deep layers, although calcium influx in the upper layers was more prominent than in the deep layers (Fig. 5). The calcium response in the deep layers (arbitrary unit: 20.6 ± 2.27, n = 21, mean ± SEM) corresponded to 47.2% of that in the upper layers (43.6 ± 4.89, n = 21, mean ± SEM) (Fig. 5A). In contrast, NO production in the upper layers (2.62 ± 0.49, n = 25, mean ± SEM) was only 14.6% of that in the deep layers (17.9 ± 0.91, n = 25, mean ± SEM) (Fig. 5B). Therefore, the conclusion can be drawn that the production of NO was limited to the deep layers of developing cortical plates.
Layer-specificity of NO Production in Developing Somatosensory Cortex
In order to examine whether the production of NO is restricted to specific layers in cortical plates of P0 mice, we performed NO imaging in combination with Toluidine Blue staining after resectioning of imaged slices (Fig. 6A–C). In this analysis, we can identify each cortical layer from the size, density and morphology of neurons visualized with Toluidine Blue staining. As described above, NO production after NMDA application was located in layers V and VI of the developing cortical plate at P0 (Fig. 6A,B)—these are the layers of the somatosensory cortex in P0 brains where transient synapses are formed (Higashi et al. 2002). Next, we compared the degree of NO production among the layers of somatosensory cortex of P0 mice brain (Fig. 6D). In P0 mice, the increase in fluorescence in layer V (52.4 ± 7.62, n = 7, mean ± SEM) and VI (36.8 ± 12.8, n = 7, mean ± SEM) was detected more strongly than that in upper layers II–IV (11.9 ± 5.62, n = 7, mean ± SEM) and MZ (9.23 ± 3.35, n = 7, mean ± SEM). At this stage, a boundary of NO production between layer IV and layer V can be distinguished.
We subsequently analyzed the layer-specific production of NO using fresh cortical slices from later developmental stages. NO production in the developing cortical plate of P10 mice was evaluated by NO imaging, and the layer-specificity of NO production was evaluated by Toluidine Blue staining after resectioning imaged slices (Fig. 7). The limit of NO production was shifted to layer IV, where thalamocortical axons reach and form branches and synapses (Fig. 7A–C). To assess the relationship between NO production and nNOS expression, furthermore, we performed immunohistochemistry for nNOS protein and NADPH-d histochemistry with resectioned slices (Fig. 7D,E). The two types of nNOS expressions were observed at the P10 stage (Fig. 7D,E), as well as the P0 stage. Neurons that were strongly positive for both nNOS and NADPH-d were observed and were evenly distributed throughout cortical layers; however, we did not detect any NO production from such a neuron that strongly expressed nNOS. The other type reflects a light and diffuse nNOS expression, which was distributed evenly across layers II–VI. A NO imaging study showed that nNOS proteins expressed lightly at layer IV–VI produce NO, while nNOS in neurons from layer II/III produces only a limited amount of NO. We evaluated the amount of NO production among layers of somatosensory cortex of P10 mice brain (Fig. 7F). In P10 mice, the increases in fluorescent signal at layers IV (88.4 ± 12.7, n = 4, mean ± SEM) and V (79.5 ± 10.1, n = 4, mean ± SEM) and VI (88.2 ± 10.7, n = 4, mean ± SEM) were higher than that in layers II/III (18.8 ± 9.92, n = 4, mean ± SEM) and MZ (8.8 ± 6.15, n = 4, mean ± SEM). This pattern of NO production may reflect the occurrence of layer-specific circuit formation in the developing neocortex (Molnar et al., 2000; Lopez-Bendito and Molnar, 2003).
To assess the relationship between NO production and barrel formation in layer IV of P10 somatosensory cortex, we performed NO imaging with tangential brain slices containing barrel structure in layer IV. We observed elevated fluorescence intensity in layer IV with the shape of barrel (Fig. 8A), which was confirmed by CO staining using imaged slices (Fig. 8B). In addition, in order to evaluate whether NO detected in layer IV corresponds to neuronal cell bodies in layer IV, we performed single-cell imaging using cortical slices of P10 mice (Fig. 9). After a cortical cell was identified as a neuron by its spontaneous synaptic current, we recorded this neuron with glass pipette and evaluated its NO production (Fig. 9A,E). After NMDA application, inward NMDA-activated current was observed in this neuron of layer IV (Fig. 9B), and then the elevation of NO-related fluorescence signals was evaluated at the exact cell simultaneously (Fig. 9B,C). As the time-resolution in this NO imaging is 1.09 s, it can be shown that this layer IV neuron produced NO for 1 s after NMDA stimulation (Fig. 9C).
In order to examine whether NO is still produced in cortical plates of young adult (P30) mice, we performed NO imaging in combination with Toluidine Blue staining (Fig. 10A–C). Weak NO production after NMDA application was observed uniformly in layers II–VI of the developing cortical plate (Fig. 10); however, the degree of NO production is significantly lower than that in P10 as well as P0 cortical plates. Taken together, these results indicate that layer-specific NO production which may affect layer-specific cortical circuit formation is terminated at young adult stage P30.
In this study, we demonstrate for the first time the real-time detection of NO production in the cortical plate of the neonatal mouse brain. We have shown that the production sites of NO in the somatosensory cortex of newborn pups (P0 mice) just after stimulation with NMDA were located in the deep cortical layers (layers V and VI) where transient synapses are formed (Higashi et al. 2002), but not in the upper cortical layers (layers II–IV). As cortical development proceeds, NO production sites spread to layer IV, where thalamocortical axons preferentially make connections with cortical neurons (Higashi et al., 1999; Lopez-Bendito and Molnar 2003). At P30, layer-specific NO production was not detected. Taken together, these results strongly suggest that NO plays an important role in the formation of cortical neuronal networks.
In the present study, we evaluated which type of NOS distribution actively releases NO upon NMDA stimulation. In the developmental stage of the cerebral cortex, a series of reports have suggested a lamina-specific pattern of NADPH-diaphorase activity (Wong-Riley et al., 1998; Vercelli et al., 1999), which would probably reflect NOS activity (Bredt et al., 1991; Dawson et al., 1991). In particular, Vercelli et al. (1999) have carefully analyzed the distribution of NADPH-d staining during rat cortical development (E17–P21) and proposed four types of distribution. (i) A transient and diffuse neuropil staining that is visible in deep layers before birth and then segregates into whisker-specific patches in layer IV until P15. An intense staining of (ii) scattered cells and of (iii) a plexus of fibers distributed in all layers. (iv) Light staining of cortical neurons detected mostly in layers II–IV. We have clearly demonstrated lamina-specific production of NO, which was observed in deep layers at birth and then spreads to layer IV up to P10. This pattern is quite similar to the transient and diffuse neuropil staining obtained by Vercelli et al. (1999). At this stage, several candidates for the site for NO production can be postulated, but the actual release of NO should be limited from NOS molecules, which aredistributed in a transient and diffuse neuropil pattern. Although, in general, NMDA-R channels are blocked by magnesium in a voltage-dependent manner, NMDA-R channels at the early postnatal stage could be activated via the GABAAR and GlyR systems (Ben-Ari et al., 1989; Owens et al., 1996; Miyakawa et al., 2002). It could be postulated that NO production by such locations probably participates in the lamina-specific network formation during the postnatal developmental stages.
In the developing nervous system, NO has been shown to be involved in a variety of processes concerning neural circuit formation, ranging from the refinement of axonal projections to regulating synaptic plasticity (Cramer and Sur, 1996; Gibbs and Truman, 1998; Cramer and Sur, 1999; Ernst et al, 1999; Wu et al, 2000). There is strong evidence showing that NO is involved in neural circuit formation, especially in the chick retinotectal system (Williams et al., 1994; Wu et al., 1994; Ernst et al., 1999), the retinogeniculate projection in ferrets (Cramer et al., 1996; Cramer and Sur, 1999), and the retinocollicular projection in mice (Wu et al., 2000). However, Finney and Shatz (1998) elegantly demonstrated in their study using nNOS knock-out mice and specific antagonists that NO was not involved in barrel formation in the somatosensory cortex during development. A similar result was also obtained by Ruthazer et al. (1996). Therefore, it is reasonable to assume that the putative function of NO would be independent of the formation of cortical barrels.
Recently, McIlvain et al. (2003) have reported that reduced branching of thalamocortical afferent arbors in layer IV of GAP-43 (+/−) mutant cortex at P7, in which the gross formation of barrel structure at the somatosensory cortex developed normally. It is conceivable that axonal sprouting of thalamocortical afferents would be regulated independently of the barrel formation at developing somatosensory cortex. A similar finding of reduced thalamocortical sprouting at layer IV of somatosensory cortex was also demonstrated in a specific mutant strain of cortical excitatory neurons lacking NMDA-R1 (CxNR1KO; Datwani et al., 2002). Since, NMDA-R is a major upstream signal for activating nNOS in excitatory cortical neurons to produce NO (e.g. Fig. 9), the reduced thalamocortical sprouting at layer IV of CxNR1KO might stemmed from the lack of NO production at layer IV of developing somatosensory cortex. In addition, the disappearance of layer-specific NO production after the critical period of cortical plasticity (Fig. 10) seems to follow the laminar-specific shift in NMDA receptor expression on layer IV neurons, from NR2B to NR2A, at the critical period of cortical plasticity (Erisir and Harris, 2003). Taken together, we can propose a possible function of NO production in developing cortex as a promoter of thalamocortical afferent sprouting in layer IV. Such a layer-specific circuit rewiring has been reported in developing visual cortex and adult somatosensory cortex (Diamond et al., 1994; Beaver et al., 2001). Future studies will attempt to clarify the role of NO on the circuit formation during brain development and the importance of cortical plasticity for higher brain function.
We are very grateful to Drs Yoshinobu Sugitani and Masaharu Ogawa for technical advice. This work was supported by a Grant-in Aid for Scientific Research from the Ministry of Education, Science, Sports and Culture of Japan.