Inducible nitric oxide synthase and nitric oxide production by oligodendrocytes

It has been previously demonstrated that microglia and astrocytes produce micromolar amounts of nitric oxide in vitro. In this study, we demonstrate that primary rat oligodendrocytes can be stimulated to produce iNOS mRNA as detected by Northern blot and in situ hybridization analysis and a 131-kDa iNOS protein by Western blot analysis; protein was also detected in cells by single- and double-label immunohistochemistry for iNOS and the oligodendrocyte-specific marker CNPase. NO/NOS are produced as a consequence of activation of the gene encoding the inducible nitric oxide synthase as determined by inhibition with actinomycin D and cyclohexamide. The iNOS is functional, leading to calcium/calmodulin-independent NO production in these in vitro cultures. J. Neurosci. Res. 48:372–384, 1997. © 1997 Wiley-Liss, Inc.     

Colocalization of urocortin and neuronal nitric oxide synthase in the hypothalamus and Edinger-Westphal nucleus of the rat

Different lines of studies suggest that both the corticotropin-releasing hormone-related peptide Urocortin I (Ucn) and the neuromodulator nitric oxide (NO) are involved in the regulation of the complex mechanisms controlling feeding and anxiety-related behaviors. The aim of the present study was to investigate the possible interaction between Ucn and NO in the hypothalamic paraventricular nucleus (PVN), an area known to be involved in the modulation of these particular behaviors. Therefore, we mapped local mRNA and peptide/protein presence of both Ucn and the NO producing neuronal NO synthase (nNOS). This investigation was extended to include the hypothalamic supraoptic nucleus (SON) and the Edinger-Westphal nucleus area (EW), the latter being one of the major cellular Ucn-expressing sites. Furthermore, we compared the two predominantly used laboratory rat strains, Wistar and Sprague-Dawley. Ucn mRNA and immunoreactivity were detected in the SON and in the EW. A significant difference between Wistar and Sprague-Dawley rats was found in mRNA levels in the EW. nNOS was detected in all brain areas analyzed, showing a significantly lower immunoreactivity in the PVN and EW of Sprague-Dawley versus Wistar rats. Contrary to some previous reports, no Ucn mRNA and only a very low immunoreactivity were detectable in the PVN of either rat strain. Interestingly, double-labeling immunofluorescence revealed that in the SON approximately 75% of all cells immunoreactive for Ucn were colocalized with nNOS, whereas in the EW only approximately 2% of the Ucn neurons were found to contain nNOS. These findings suggest an interaction between Ucn and NO signaling within the SON, rather than the PVN, that may modulate the regulation of feeding, reproduction, and anxiety-related behaviors.     

Water-deprivation-induced expression of neuronal nitric oxide synthase in the hypothalamic paraventricular nucleus of rat

This study was conducted to define the molecular mechanism by which dehydration induces expression of neuronal nitric oxide synthase (nNOS) in the hypothalamic paraventricular nucleus (PVN). Rats were deprived from water for 48 hr and then sacrificed immediately or 1 hr after ad libitum access to water. Another group of rats had free access to food and water and was included as euhydrate control group. The PVN sections fixed with 4% paraformaldehyde were processed for nNOS immunohistochemistry and NADPH-diaphorase (NADPH-d)/pCREB or NADPH-d/c-Fos double staining. nNOS-ir neurons significantly increased with water deprivation and decreased with rehydration, both in the posterior magnocellular (pM)- and the medial parvocellular (mP)-PVN. Most NADPH-d histostained neurons in the PVN appeared to exhibit pCREB-ir as well. Water deprivation markedly increased, and rehydration decreased, NADPH-d/pCREB neurons both in the pM- and in the mP-PVN. Gel shift assay demonstrated that dehydration may promote CREB binding to nNOS promoter in the PVN neurons. Significant amounts of NADPH-d-stained neurons in the PVN of water-deprived rats (67-68% in both the mP and the pM) exhibited c-Fos-ir. NADPH-d/c-Fos neurons in the pM-PVN were increased by water deprivation but not changed by rehydration. NADPH-d/c-Fos double-stained neurons in the mP-PVN did not significantly change depending on different water conditions. These results suggest that pCREB may play a role in dehydration-induced nNOS gene expression in the PVN neurons, and c-Fos might not be implicated in the regulatory pathway.     

The effect of lactation on nitric oxide synthase gene expression

Nitric oxide-synthesizing (NOS) enzyme has been identified in several neural populations, including the hypothalamic paraventricular nucleus (PVN). The PVN plays a major role in regulatig milk ejection. In the present study, using in situ hybridization, the effect of lactation on NOS mRNA in the PVN was investigated. A significant increase was seen in the PVN of lactating rats, indicating a possible involvement of nitric oxide in the hypothalamic-pituitary regulation of milk ejection.     

Coexistence of corticotropin-releasing factor and enkephalin in the paraventricular nucleus of the rat

Corticotropin-releasing factor and enkephalin-containing neurons are found in the paraventricular nucleus of the rat hypothalamus. Their immunocytochemical distribution suggests that a subpopulation of neurons of the paraventricular nucleus might contain both peptides. The present study describes the coexistence of corticotropin-releasing factor and enkephalin in parvicellular neurons of the paraventricular nucleus. Immunocytochemical labeling of peptides was combined with in situ hybridization of mRNA to visualize peptide and mRNA labeling in the same tissue section. This resulted in several observations: (1) neurons labeling for the peptide corticotropin-releasing factor also contain the mRNA to synthesize corticotropin-releasing factor, (2) neurons labeling for the peptide enkephalin also contain the mRNA to synthesize the peptide enkephalin, (3) a subpopulation of corticotropin-releasing factor neurons contains the mRNA to synthesize enkephalin, and (4) a subpopulation of enkephalin neurons contains the mRNA to synthesize corticotropin-releasing factor. A large percentage of enkephalin immunoreactive neurons contains corticotropin-releasing factor mRNA, where as a smaller percentage of corticotropin-releasing factor immunoreactive neurons contains enkephalin mRNA. These double-labeled neurons are present throughout the rostral-caudal extent of the paraventricular nucleus; the majority of these neurons is present in the medial parvicellular paraventricular nucleus.     

Refeeding-induced expression of neuronal nitric oxide synthase in the rat paraventricular nucleus

We have previously reported that food deprivation decreases the expression of neuronal nitric oxide synthase (nNOS) in the hypothalamic paraventricular nucleus (PVN) of rats, and this reduction is inhibited by blockade of glucocorticoid receptors. In this study, we examined whether the fasting-induced decrease in nNOS gene expression in the PVN is restored by refeeding. The number of nNOS immunopositive cells in the PVN, which was markedly decreased by 48 h of food deprivation, increased significantly after 6 h of refeeding and was fully restored by 24 h after refeeding. The plasma corticosterone level, which was markedly increased by food deprivation, decreased significantly within 30 min after refeeding and returned to the free fed control level by 6 h. Synthetic glucocorticoid dexamethasone blocked the refeeding-induced nNOS expression in the PVN without suppressing food intake. Refeeding with a non-caloric food mash for 5 h failed to restore the fasting-induced decrease in the PVN-nNOS but did, however, successfully restore the plasma corticosterone level. These results suggest that the refeeding-induced nNOS expression in the PVN is a nutrient-directed event and that plasma glucocorticoids may play an inhibitory role in the regulatory pathway. Additionally, glucocorticoid disinhibition alone does not appear to be sufficient to induce nNOS expression in the PVN; nNOS expression in the PVN upon refeeding may require both nutrient supplementation and glucocorticoid disinhibition.     

Developmental expression of GABAA receptor subunit and GAD genes in mouse somatosensory barrel cortex

In situ hybridization histochemistry with radioactive cRNA probes was used to study patterns of gene expression for α1, α2, α4, α5, β1, β2, and γ2 subunit mRNAs of type A gamma aminobutyric acid (GABA_A) receptors and for 67-kDa glutamic acid decarboxylase (GAD₆₇) mRNA in mouse barrel cortex during the period (postnatal days 1-12; P1-P12) when thalamocortical innervation of layer IV barrels is occurring. The α1, β2, and γ2 subunit mRNAs increased substantially with age, especially in layers V and VI, and throughout the period studied, invariably had the same laminar-specific patterns of expression. All three mRNAs were highly expressed in the dense cortical plate at P1. In layer IV after differentiation of barrels, they were expressed in cells of both barrel walls and hollows but especially in the walls. The α2, α4, α5, and β1 subunit mRNAs were expressed at lower levels and had different laminar patterns of distribution; α2 and α4 showed switches between layers over time; α5 was invariably associated with the subplate or its derivative, β1 with layer IV. Levels of α2 mRNA did not change over time; α4 and β1 mRNAs increased and α5 decreased. GAD₆₇ mRNA was highest in layer I at P1 and progressively increased in other layers. These results suggest that postnatal development of GABA_A receptors is mainly directed at the production of receptors assembled from α1, β2, and γ2 subunits, with β1 contributing in layer IV. Other subunits may be associated with receptors involved in trophic actions of GABA during development and may give GABA_A receptor-mediated responses in the developing cortex their particular physiological profile. J. Comp. Neurol. 383:199-219, 1997. © 1997 Wiley-Liss, Inc.     

Monoaminergic control of vasopressin and VIP expression in the mouse suprachiasmatic nucleus

We studied the effects of serotonin and noradrenaline on the expression of arginine-vasopressin (AVP) and vasoactive intestinal peptide (VIP) in the suprachiasmatic nucleus (SCN). We used transgenic Tg8 mice knockout for the MAO-A (monoamine oxidase A) gene, which are characterized by increased amounts of serotonin and noradrenaline in brain compared to wild-type mice (C3H). The MAO-A deficiency caused an increase in AVP and VIP expression (determined by immunohistochemistry, enzyme immunoassay, and in situ hybridization) compared to C3H mice. The number of peptidergic neurons was also increased. Inhibiting serotonin or noradrenaline synthesis in Tg8 mice by the administration of parachlorophenylalanine or alpha-methylparatyrosine, respectively, the amounts of AVP, VIP and their mRNAs were decreased, but not the number of peptidergic neurons. This study indicates that serotonin and noradrenaline stimulate AVP and VIP expression, and could participate in the differentiation of the neurochemical phenotype in the mouse SCN.     

Subunits of the nitric oxide receptor, soluble guanylyl cyclase, expressed in rat brain

Despite the widespread use of nitric oxide as a signalling molecule in the central nervous system, the molecular makeup of its receptor, soluble guanylyl cyclase (sGC), therein is poorly understood. Accordingly, RT-PCR and in situ hybridization were used to identify sGC subunits expressed in rat brain. In addition to the expected mRNA for alpha 1 and beta1 subunits, message for the beta 2 subunit was detected in the cerebellum at all developmental stages investigated (1--150 days postnatum). The use of degenerate primers allowed the identification of mRNA coding for the rat alpha 2 subunit, which was also expressed at every age studied. All but beta 2 were detected by in situ hybridization in the brains of both 8-day-old and adult rats. The distribution patterns indicated that in some areas, e.g. caudate-putamen and nucleus accumbens, sGC probably exists mainly as the alpha 1 beta 1 heterodimer. In others, e.g. hippocampus and olfactory bulb, alpha 2 beta 1 is likely to be dominant. In the cerebellum, alpha 1 and beta 1 message was strong in the Purkinje cell layer but was not confined to Purkinje cells: smaller cells, presumed to be the Bergmann glia, were also labelled. In contrast, alpha 2 mRNA was concentrated in cerebellar granule cells. Western blotting indicated an excess of alpha 1 over beta 1 protein in the cerebellum, the reverse of what was found in the lung. It is concluded that, in molecular terms, sGC is likely to be more complex and exhibit more regional variation in the brain than previously thought. The functional consequences of this heterogeneity require investigation.     

Changes in expression of PACAP in rat sensory neurons in response to sciatic nerve compression

In the present study, expression of pituitary adenylate cyclase-activating polypeptide (PACAP) in rat dorsal root ganglion (DRG) neurons and sciatic nerve following experimental sciatic nerve compression was studied with the use of quantitative immunohistochemistry and in situ hybridization. Previously, we have investigated changes in PACAP expression after nerve transection and, here, the far more frequently encountered condition of nerve compression injury is examined. Nerve compression was performed unilaterally on the rat sciatic nerve, at mid-thigh level, by application of a narrow silicone tube around the nerve for 3, 7, 14 or 28 days, respectively. We detect a statistically significant upregulation in the number and density of PACAP mRNA expression in both small and large DRG neurons in response to nerve compression. An increased number of PACAP-immunoreactive neurons is also found in the ipsilateral DRG. In addition, PACAP immunoreactivity is observed in the compressed sciatic nerve segment and adjacent nerve tissue after nerve compression. The present findings can be compared with previous studies where we have shown that PACAP expression is upregulated in DRG; in response to peripheral inflammation (primarily in small-medium neurons), and after axotomy (dramatic upregulation in medium-large neurons). In view of the recent findings of an increased PACAP expression in DRG after nerve compression, as well as the previous findings of a modulation of PACAP expression in response to axotomy and inflammation, it is likely that PACAP is also involved in the modulation of the response to peripheral nerve compression.     

Developmental expression of neuronal calmodulin mRNA species in the rat brain analyzed by in situ hybridization.

The temporal and spatial distribution of calmodulin mRNAs which are preferentially expressed in neurons was determined during postnatal development of rat central nervous system. Expression of these mRNAs was strongly detected in the developing neocortex, hippocampus, and cerebellum. Differences in the pattern of expression of a 1.8 and 4.0 kb neuronal calmodulin mRNA species were identified in the developing cerebellum. Expression of the smaller mRNA appeared to correlate with proliferating and developing cerebellar granule neurons while the larger mRNA was present in the mature granule neuron population. A transient elevation in the neuronal calmodulin mRNA species was observed in the superior and inferior colliculus and in the thalamus at postnatal days 5 and 10.     

Differential expression of neurotensin receptor mRNA in the dopaminergic cell groups of the rat diencephalon and mesencephalon

Several lines of evidence support interactions between neurotensin (NT) and dopaminergic (DAergic) neurons in the brain. In order to obtain further knowledge about the anatomical substrate for such interactions, the distribution of cells expressing the cloned neurotensin receptor (NTR) mRNA was examined in relation to tyrosine hydroxylase (TH) mRNA-expressing cells within different subnuclei of the diencephalon and ventral mesencephalon of the male rat. In situ hybridization was performed on consecutive sections labeled with ³³P-labeled oligonucleotide probes. In the hypothalamus, NTR mRNA signals were mostly found in the suprachiasmatic, dorsomedial, dorsal premammillary, and supramammillary nuclei. On the other hand, DAergic cells of the periventricular nucleus of the hypothalamus and dorsal aspect of the arcuate nucleus, revealed by TH in situ hybridization, did not exhibit NTR mRNA even though dense NT binding sites have been previously described in this nuclei. In the zona incerta, TH mRNA-containing cells were concentrated in the medial part, with little overlap with NTR mRNA-expressing cells located mainly in its mediolateral extent. In contrast, the distribution of both markers was superimposable within the different subdivisions of the ventral tegmental area and substantia nigra, as previously suggested, but also in the retrorubral field. These anatomical data further support the NT-dopamine interactions on both mesocorticolimbic and nigrostriatal DAergic systems. Moreover, the results suggest that diencephalic DAergic neurons do not synthesize the cloned NTR mRNA or express it at considerably lower levels than DAergic mesencephalic cells. © 1995 Wiley-Liss, Inc.     

Brainstem afferents to the tuberomammillary nucleus in the rat brain with special reference to monoaminergic innervation

Monoaminergic innervation of a histamine-producing cell group, the tuberomammillary nucleus in the posterior hypothalamus, was investigated in the rat by light and electron microscopic immunohistochemical techniques. Immunohistochemical staining of sections of the posterior hypothalamus was demonstrated afferent fibers immunoreactive to tyrosine hydroxylase in ventral and medial subgroups of the tuberomammillary nucleus afferent fibers immunoreactive to tyrosine hydroxylase (TH), dopamine-beta-hydroxylase (DBH), phenyletanolamine-N-methyltransferase (PNMT), and serotonin (5-HT). TH- and DBH-immunoreactive fibers were similar and were evenly and densely distributed throughout the tuberomammillary nucleus. Fibers stained with 5-HT antibodies were also present throughout the tuberomammillary nucleus but exhibited the densest labeling in the dendritic layer adjacent to the glia limitans in the ventral subgroup. Innervation by PNMT-immunoreactive axons was sparse. Electron microscopic analysis of TH-, DBH-, and 5-HT-immunoreactive fibers in the tuberomammillary nucleus revealed vesicle-containing terminal boutons, which formed synapses with dendrites of varying size. Synaptic contacts with nerve cell bodies were not found. Retrograde transport of the fluorescent dye Fast Blue injected into the tuberomammillary nucleus, combined with immunofluorescent staining with anti-TH, anti-DBH, anti-PNMT, and anti-5-HT antibodies, showed that monoaminergic input to the tuberomammillary nucleus originated mainly from the adrenergic and noradrenergic cell groups C1-C3 and A1-A2, respectively, and from the serotoninergic cell groups B5-B9 as designated by Dahlstr?m and Fuxe ('65). Few double-labeled neurons were found in the nucleus locus coeruleus and the dopaminergic cell groups of the rostral brain stem. The present findings suggest that the activity of the histamine-producing neurons of the tuberomammillary nucleus is influenced by monoaminergic neurons in the ventrolateral and dorsomedial medulla oblongata and the raphe nuclei of the rostral brainstem.     

Localization of the neuromedin K receptor (NK3) in the central nervous system of the rat

The distribution of the neuromedin K receptor (NK₃; NKR) in the central nervous system was investigated in the adult rat by using in situ hybridization and immunohistochemical techniques. The rabbit anti-NKR antibody was raised against a bacterial fusion protein containing a C-terminal portion of NKR and affinity purified with a Sepharose 4B column conjugated to the fusion protein. Immunoblot analysis was performed to test the reactivity and specificity of the antibody. Crude membrane was prepared from cDNA-transfected Chinese hamster ovary (CHO) cells expressing each of the rat NKR, substance P receptor (NK₁; SPR), and substance K receptor (NK₂; SKR) and from the hypothalamus, cerebral cortex, and cerebellum. Immunoreactive bands were observed specifically in the NKR-CHO cells, hypothalamus, and cerebral cortex but not in the SPR- or SKR-CHO cells, nor in the cerebellum. Molecular weights of the immunoreactive bands ranged from 73 to 89 kDa and from 59 to 83 kDa in the NKR-CHO cells and tissues, respectively. The distribution of NKR-like immunoreactivity coincided with that of NKR mRNA. The expression of NKR was indicated on neuronal cell bodies and dendrites. NKR was found to be expressed intensely or moderately in neurons in the glomerular and granule cell layers of the main olfactory bulb; glomerular and mitral cell layers of the accessory olfactory bulb; layers IV and V of the cerebral neocortex; medial septal nucleus; nucleus of the diagonal band; bed nucleus of the stria terminalis; globus pallidus; ventral pallidum; paraventricular nucleus; supraoptic nucleus; zona incerta; dorsal, lateral, and posterior hypothalamic areas; amygdaloid nuclei; medial habenular nucleus; ventral tegmental area; midbrain periaqueductal gray; interpeduncular nuclei; substantia nigra pars compacta; linear, median, dorsal, and pontine raphe nuclei; posteromedial tegmental nucleus; sphenoid nucleus; nucleus of the solitary tract; intermediate and rostroventrolateral reticular nuclei; and lamina II of the caudal spinal trigeminal nucleus and spinal dorsal horn. These findings are discussed in relation to the physiological functions associated with neuromedin K. © 1996 Wiley-Liss, Inc.     

Comparative distribution of myristoylated alanine-rich C kinase substrate (MARCKS) and F1/GAP-43 gene expression in the adult rat brain

Myristoylated alanine-rich C-kinase substrate (MARCKS) and F1/GAP-43 (B-50/neuromodulin) are both major specific substrates for protein kinase C (PKC) and appear to play an important role in the regulation of neuroplastic events during development and in the adult brain. Since PKC isozymes are differentially expressed in brain and the expression of F1/GAP-43 and MARCKS mRNAs are differentially regulated by PKC through posttranslational mechanisms, the present study examined the relative distribution of both mRNAs in the adult rat brain by using in situ hybridization histochemistry. MARCKS hybridization was most pronounced in the olfactory bulb, piriform cortex (layer II), medial habenular nucleus, subregions of the amygdala, specific hypothalamic nuclei, hippocampal granule cells, neocortex, and cerebellar cortex, intermediate in the superior colliculus, hippocampal CA1, and certain brainstem nuclei including the locus coeruleus, and low-absent in regions of the caudate-putamen, geniculate nuclei, thalamic nuclei, lateral habenular nucleus, and hippocampal CA3 pyramidal and hilar neurons. Consistent with previous reports, prominent F1/GAP-43 hybridization was observed in neocortex, medial geniculate, piriform cortex (layer II), substantia nigra pars compacta, hippocampal CA3 pyramidal cells, thalamic and hypothalamic nuclei, lateral habenular nucleus, locus coeruleus, raphe nuclei, and cerebellar granule cells, intermediate in regions of the thalamus, hypothalamus, and amygdala, and low-absent in regions of the olfactory bulb, caudate-putamen, medial habenular nucleus, hippocampal granule cells, and superior colliculus. Overall, F1/GAP-43 was highly expressed in a greater number of regions compared to MARCKS and, in a number of regions, including the hippocampus, habenular complex, ventral tegmentum, geniculate, and certain brain stem nuclei, a striking inverse pattern of expression was observed. These results indicate that MARCKS gene expression, like that of F1/GAP-43, remains elevated in select regions of the adult rat brain which are associated with a high degree of retained plasticity. The potential role of PKC in the regulation of MARCKS and F1/GAP-43 gene expression in brain is assessed. J. Comp. Neurol. 379:48-71, 1997. © 1997 Wiley-Liss, Inc.     

Electronmicroscopic study of somatostatin-containing neurons in rat arcuate nucleus with special reference to neuronal regulation

After an intraventricular administration of colchicine, the arcuate nucleus of rat hypothalamus was examined light and electron microscopically by pre-embedding immunohistochemistry for somatostatin. The arcuate nucleus exhibited numerous immunoreactive cell bodies and dense networks of immunoreactive fibers. The fibers appeared to surround immunonegative cell bodies. The immunoreactive cell bodies were multipolar in shape and projected immunoreactive processes to some extent. The immunoreactive cell bodies and fibers received synaptic contacts by immunonegative fiber terminals containing a large number of synaptic clear vesicles. Similarly, immunoreactive somatostatin fibers appeared to terminate upon other immunonegative cell bodies and fibers. The immunoreactive presynaptic terminals contain several labeled granules and numerous synaptic vesicles. In close proximity to these immunolabeled terminals, non-labeled presynaptic fibers. This suggests that in the arcuate nucleus neurons regulated by somatostatin neurons are also under the control of other types of neurons.     

Downregulation of the alpha5 subunit of the GABA(A) receptor in the pilocarpine model of temporal lobe epilepsy

Specific subunits of gamma-aminobutyric acid (GABA)A receptors may be regulated differentially in animal models of temporal lobe epilepsy during the chronic stage. Although several subunits may be upregulated, other subunits may be downregulated in the hippocampal formation. The alpha5 subunit is of particular interest because of its relatively selective localization in the hippocampus and its potential role in tonic inhibition. In normal rats, immunolabeling of the alpha5 subunit was high in the dendritic layers of CA1 and CA2 and moderate in these regions of CA3. In chronic pilocarpine-treated rats displaying recurrent seizures, alpha5 subunit-labeling was substantially decreased in CA1 and nearly absent in CA2. Only slight decreases in immunolabeling were evident in CA3. In situ hybridization studies demonstrated that the alpha5 subunit mRNA was also strongly decreased in stratum pyramidale of CA1 and CA2. Thus, the alterations in localization of the alpha5 subunit peptide and its mRNA were highly correlated. The large decreases in labeling of the alpha5 subunit did not appear to be related to loss of pyramidal neurons in CA1 or CA2 since these neurons were generally preserved in pilocarpine-treated animals. No comparable decreases in labeling of the alpha2 subunit of the GABA(A) receptor were detected. These findings indicate that the alpha5 subunit of the GABA(A) receptor is capable of substantial and prolonged downregulation in remaining pyramidal neurons in a model of temporal lobe epilepsy. The results raise the possibility that presumptive extrasynaptic GABA(A) receptor subunits, such as the alpha5 subunit, may be regulated differently than synaptically located subunits, such as the alpha2 subunit, within the same brain regions in some pathological conditions.     

Developmental changes of nitric oxide synthase in the rat superior colliculus

We used the activity of nicotinamide adenine dinucleotide phosphate diaphorase (NADPH-d) to detect the presence of nitric oxide synthase (NOS) in the developing rat superior colliculus. Our results showed that NOS is present in cells and nenropil in the developing and adult rat superior colliculus. The first NOS-positive cells were detected at postnatal day 7 and were weakly stained. During the following days the number of stained cells increased markedly, reaching a peak by postnatal day 15, coinciding with the time of eye opening in the rat. By the end of the third postnatal week, both the number and intensity of stained cells showed an adult-like pattern. We conclude that NOS expression lags behind the initial period of reorganization of the retinotectal projection. However, NOS activity could be involved in the subsequent synaptic remodeling and plasticity of the retinocollicular projection. © 1995 Wiley-Liss, Inc.