Maternally involved galanin neurons in the preoptic area of the rat.

Recent selective stimulation and ablation of galanin neurons in the preoptic area of the hypothalamus established their critical role in control of maternal behaviors. Here, we identified a group of galanin neurons in the anterior commissural nucleus (ACN), and a distinct group in the medial preoptic area (MPA). Galanin neurons in ACN but not the MPA co-expressed oxytocin. We used immunodetection of phosphorylated STAT5 (pSTAT5), involved in prolactin receptor signal transduction, to evaluate the effects of suckling-induced prolactin release and found that 76 % of galanin cells in ACN, but only 12 % in MPA were prolactin responsive. Nerve terminals containing tuberoinfundibular peptide 39 (TIP39), a neuropeptide that mediates effects of suckling on maternal motivation, were abundant around galanin neurons in both preoptic regions. In the ACN and MPA, 89 and 82 % of galanin neurons received close somatic appositions, with an average of 2.9 and 2.6 per cell, respectively. We observed perisomatic innervation of galanin neurons using correlated light and electron microscopy. The connection was excitatory based on the glutamate content of TIP39 terminals demonstrated by post-embedding immunogold electron microscopy. Injection of the anterograde tracer biotinylated dextran amine into the TIP39-expressing posterior intralaminar complex of the thalamus (PIL) demonstrated that preoptic TIP39 fibers originate in the PIL, which is activated by suckling. Thus, galanin neurons in the preoptic area of mother rats are innervated by an excitatory neuronal pathway that conveys suckling-related information. In turn, they can be topographically and neurochemically divided into two distinct cell groups, of which only one is affected by prolactin.

Efferent projections of the suprachiasmatic nucleus based on the distribution of vasoactive intestinal peptide (VIP) and arginine vasopressin (AVP) immunoreactive fibers in the hypothalamus of Sapajus apella.

The suprachiasmatic nucleus (SCN), which is considered to be the master circadian clock in mammals, establishes biological rhythms of approximately 24 h that several organs exhibit. One aspect relevant to the study of the neurofunctional features of biological rhythmicity is the identification of communication pathways between the SCN and other brain areas. As a result, SCN efferent projections have been investigated in several species, including rodents and a few primates. The fibers originating from the two main intrinsic fiber subpopulations, one producing vasoactive intestinal peptide (VIP) and the other producing arginine vasopressin (AVP), exhibit morphological traits that distinguish them from fibers that originate from other brain areas. This distinction provides a parameter to study SCN efferent projections. In this study, we mapped VIP (VIP-ir) and AVP (AVP-ir) immunoreactive (ir) fibers and endings in the hypothalamus of the primate Sapajus apella via immunohistochemical and morphologic study. Regarding the fiber distribution pattern, AVP-ir and VIP-ir fibers were identified in regions of the tuberal hypothalamic area, retrochiasmatic area, lateral hypothalamic area, and anterior hypothalamic area. VIP-ir and AVP-ir fibers coexisted in several hypothalamic areas; however, AVP-ir fibers were predominant over VIP-ir fibers in the posterior hypothalamus and medial periventricular area. This distribution pattern and the receiving hypothalamic areas of the VIP-ir and AVP-ir fibers, which shared similar morphological features with those found in SCN, were similar to the patterns observed in diurnal and nocturnal animals. This finding supports the conservative nature of this feature among different species. Morphometric analysis of SCN intrinsic neurons indicated homogeneity in the size of VIP-ir neurons in the SCN ventral portion and heterogeneity in the size of two subpopulations of AVP-ir neurons in the SCN dorsal portion. The distribution of fibers and morphometric features of these neuronal populations are described and compared with those of other species in the present study.

The role of orexin-1 receptors in physiologic responses evoked by microinjection of PgE2 or muscimol into the medial preoptic area.

The medial preoptic area (mPOA) of the hypothalamus has long been thought to play an important role in both fever production and thermoregulation. Microinjections of prostaglandin E2 (PgE2) or the GABA(A) agonist muscimol into the mPOA cause similar increases in body temperature, heart rate, and blood pressure. Microinjections of these compounds however evoke different behavioral responses with muscimol increasing and PgE2 having no effect on locomotion. The purpose of this study was to determine the role of orexin-1 receptors in mediating these dissimilar responses. Systemic injections of the orexin-1 receptor antagonist SB-334867 reduced temperature and cardiovascular responses produced by microinjections of muscimol, but had no effect on either response produced by PgE2. SB-334867 did not significantly decrease locomotion evoked by microinjections of muscimol into the mPOA. These data suggest that there are two central nervous system circuits involved in increasing body temperature, heart rate and blood pressure: one circuit activated by muscimol, involving orexin neurons, and a separate orexin-independent circuit activated by PgE2.

Neural interaction between galanin-like peptide (GALP)- and luteinizing hormone-releasing hormone (LHRH)-containing neurons.

Galanin-like peptide (GALP), commonly known as an appetite-regulating peptide, has been shown to increase plasma luteinizing hormone (LH) through luteinizing hormone-releasing hormone (LHRH). This led us to investigate, using both light and electron microscopy, whether GALP-containing neurons in the rat brain make direct inputs to LHRH-containing neurons. As LHRH-containing neurons are very difficult to demonstrate immunohistochemically with LHRH antiserum without colchicine treatment, we used a transgenic rat in which LHRH tagged with enhanced green fluorescence protein facilitated the precise detection of LHRH-producing neuronal cell bodies and processes. This is the first study to report on synaptic inputs to LHRH-containing neurons at the ultrastructural level using this transgenic model. We also used immunohistochemistry to investigate the neuronal interaction between GALP- and LHRH-containing neurons. The experiments revealed that GALP-containing nerve terminals lie in close apposition with LHRH-containing cell bodies and processes in the medial preoptic area and the bed nucleus of the stria terminalis. At the ultrastructural level, the GALP-positive nerve terminals were found to make axo-somatic and axo-dendritic synaptic contacts with the EGFP-positive neurons in these areas. These results strongly suggest that GALP-containing neurons provide direct input to LHRH-containing neurons and that GALP plays a crucial role in the regulation of LH secretion via LHRH.

Peptidergic and catecholaminergic synaptic contacts onto nucleus preopticus medianus neurons projecting to the subfornical organ in the rat.

The nucleus preopticus medianus (POMe) is known to be a key site in regulation of cardiovascular and body fluid homeostasis. To clarify the regulation mechanism to the POMe, the innervation pattern of synapses made by axon terminals immunoreactive to beta-endorphin, neuropeptide Y and tyrosine hydroxylase onto POMe neurons projecting to the subfornical organ (SFO) was investigated in the rat. After injection of a retrograde tracer, wheat germ agglutinin-conjugated horseradish peroxidase-colloidal gold complex, into the SFO, many neurons were retrogradely labeled in the POMe, more frequently in its dorsal part. Electron microscopy of the POMe revealed that beta-endorphin- and tyrosine hydroxylase-immunoreactive axon terminals formed predominantly axo-somatic synapses, and neuropeptide Y-immunoreactive axon terminals formed more axo-dendritic than axo-somatic synapses with retrogradely labeled neurons. The present localization patterns of POMe neurons retrogradely labeled from the SFO and the type of synapses of axon terminals immunoreactive to three neurochemical markers on these neurons were compared to those of POMe neurons retrogradely labeled from the paraventricular hypothalamic nucleus demonstrated in our previous report.

Sexually dimorphic hormonal regulation of the gap junction protein, CX43, in rats and altered female reproductive function in CX43+/- mice.

Astrocytic gap junctional communication is important in steroid hormone regulation of reproductive processes at the level of the hypothalamus, including estrous cyclicity and sexual behavior. We examined the effects of estradiol and progesterone on the abundance of the gap junctional protein, connexin 43 (CX43), which is highly expressed in astrocytes. Gonadectomized rats received hormone treatments that induce maximal sexual behavior and gonadotropin surges in females (estrogen for 48 h followed by progesterone, estrogen alone or progesterone alone). Control animals received vehicle (oil) injections. In the female rat preoptic area (POA), containing the gonadotropin-releasing hormone (GnRH) cell bodies, treatment with estrogen, progesterone or estrogen + progesterone significantly increased CX43 protein levels in immunoblots. In contrast, estrogen + progesterone significantly decreased CX43 levels in the male rat POA. This sexually dimorphic hormonal regulation of CX43 was not evident in the hypothalamus, which contains primarily GnRH nerve terminals. Treatment with estrogen + progesterone significantly decreased CX43 levels in both the male and female hypothalamus. To examine the role of CX43 in female reproductive function, we studied heterozygous female CX43 (CX43+/-) mice. Most mutant mice did not show normal estrous cycles. In addition, when compared to wild type females, CX43+/- mice had reduced lordosis behavior. These data suggest that hypothalamic CX43 expression is regulated by steroid hormones in a brain-region-specific and sexually dimorphic manner. Therefore, gap junctional communication in the POA and hypothalamus may be a factor regulating the estrous cycle and sexual behavior in female rodents.

Evidence for vesicular glutamate transporter synapses onto gonadotropin-releasing hormone and other neurons in the rat medial preoptic area.

The medial preoptic area is a key structure in the control of reproduction. Several data suggest that excitatory amino acids are involved in the regulation of this function and the major site of this action is the medial preoptic region. Data concerning the neuromorphology of the glutamatergic innervation of the medial preoptic area are fragmentary. The present investigations were focused on: (i) the morphology of the vesicular glutamate transporter 1 (VGluT1)- and vesicular glutamate transporter 2 (VGluT2)-immunoreactive nerve terminals, which are considered to be specific to presumed glutamatergic neuronal elements, in the medial preoptic area of rat; and (ii) the relationship between these glutamate transporter-positive endings and the gonadotropin-releasing hormone (GnRH) neurons in the region. Single- and double-label immunocytochemistry was used at the light and electron microscopic level. There was a weak to moderate density of VGluT1- and a moderate to intense density of VGluT2-immunoreactive elements in the medial preoptic area. Electron microscopy revealed that both VGluT1- and VGluT2-immunoreactive boutons made asymmetric type synaptic contacts with unlabelled neurons. VGluT2-labelled, but not VGluT1-labelled, axon terminals established asymmetric synaptic contacts on GnRH-immunostained neurons, mainly on their dendrites. The present findings are the first electron microscopic examinations on the glutamatergic innervation of the rat medial preoptic area. They provide direct neuromorphological evidence for the existence of direct glutamatergic innervation of GnRH and other neurons in the rat medial preoptic area.

Multiple mechanisms control brain aromatase activity at the genomic and non-genomic level.

Evidence has recently accumulated indicating that aromatase activity in the preoptic area is modulated in parallel by both slow (hours to days) genomic and rapid (minutes to hours) non-genomic mechanisms. We review here these two types of control mechanisms and their potential contribution to various aspects of brain physiology in quail. High levels of aromatase mRNA, protein and activity (AA) are present in the preoptic area of this species where the transcription of aromatase is controlled mainly by steroids. Estrogens acting in synergy with androgens play a key role in this control and both androgen and estrogen receptors (ER; alpha and beta subtypes) are present in the preoptic area even if they are not necessarily co-localized in the same cells as aromatase. Steroids have more pronounced effects on aromatase transcription in males than in females and this sex difference could be caused, in part, by a sexually differentiated expression of the steroid receptor coactivator 1 in this area. The changes in aromatase concentration presumably control seasonal variations as well as sex differences in brain estrogen production. Aromatase activity in hypothalamic homogenates is also rapidly (within minutes) down-regulated by exposure to conditions that enhance protein phosphorylation such as the presence of high concentrations of calcium, magnesium and ATP. Similarly, pharmacological manipulations such as treatment with thapsigargin or stimulation of various neurotransmitter receptors (alpha-amino-3-hydroxy-methyl-4-isoxazole propionic acid (AMPA), kainate, and N-methyl-D-aspartate (NMDA)) leading to enhanced intracellular calcium concentrations depress within minutes the aromatase activity measured in quail preoptic explants. The effects of receptor stimulation are presumably direct: electrophysiological data confirm the presence of these receptors in the membrane of aromatase-expressing cells. Inhibitors of protein kinases interfere with these processes and Western blotting experiments on brain aromatase purified by immunoprecipitation confirm that the phosphorylations regulating aromatase activity directly affect the enzyme rather than another regulatory protein. Accordingly, several phosphorylation consensus sites are present on the deduced amino acid sequence of the recently cloned quail aromatase. Fast changes in the local availability of estrogens in the brain can thus be caused by aromatase phosphorylation so that estrogen could rapidly regulate neuronal physiology and behavior. The rapid as well as slower processes of local estrogen production in the brain thus match well with the genomic and non-genomic actions of steroids in the brain. These two processes potentially provide sufficient temporal variation in the bio-availability of estrogens to support the entire range of established effects for this steroid.

Androgen receptor immunoreactivity in forebrain axons and dendrites in the rat.

As members of the steroid receptor superfamily, androgen receptors (ARs) have been traditionally identified as transcription factors. In the presence of ligand, ARs reside in the nucleus, where, upon ligand binding, the receptors dimerize and bind to specific response elements in the promoter region of hormone-responsive genes. However, in this report, we describe the discovery that ARs are also present in axons and dendrites within the mammalian central nervous system. AR expression in axons was identified in the rat brain at the light microscopic level using two different antibodies directed against the N terminus of the AR protein and nickel intensified 3'-3'-diaminobenzidine, and also using fluorescence methods and confocal microscopy. This distribution was confirmed at the ultrastructural level. In addition, AR immunoreactivity was identified in small dendrites at the ultrastructural level. AR-immunoreactive axons were observed primarily in the cerebral cortex and were rare in regions where nuclear AR expression is abundant. The observation that ARs are present in axons and dendrites highlights the possibility that androgens play an important and novel extra-nuclear role in neuronal function.

Distribution of 1,25-dihydroxyvitamin D3 receptor immunoreactivity in the limbic system of the rat.

We used immunocytochemistry to obtain a complete cellular and subcellular mapping of the 1,25-dihydroxyvitamin D3 receptor protein (VDR) in the rat limbic system. We observed specific VDR immunostaining in the nucleus as well as in the perinuclear cytoplasm of neuronal cells. The limbic system consists of a variety of neuronal structures, and is known to have influence on memory, behavior, emotions and reproduction. In the hippocampal formation, we found strong nuclear staining as well as less distinguished cytoplasmic VDR staining in CA1, CA3 and CA4. The CA2 area showed a unique cytoplasmic predominance of VDR. The amygdala was found to exhibit specific patterns of VDR distribution in the various regions of the nucleus. We observed distinct differences of VDR localization within the limbic preoptic areas of the hypothalamus. Further parts of the brain we analyzed included the mammillary bodies, the indusium griseum and the cingulate cortex. The subcellular distribution of VDR in regions of the limbic system suggests a specific functional role of the receptor protein and indicates a role for calcitriol as a neuroactive steroid.

A confocal microscopic study of synaptic inputs to gonadotropin-releasing hormone cells in mouse brain: regional differences and enhancement by estrogen.

Even though the cells producing gonadotropin-releasing hormone (GnRH) are scattered in the basal forebrain, a large proportion of them is present in the organum vasculosum of the lamina terminalis (OVLT) and in the preoptic area. The present studies were undertaken to investigate whether there is any difference in the number of synaptic inputs between GnRH cells located in the OVLT and those located at more anterior levels of the brain. Immunohistochemical staining for the synaptic marker synaptophysin coupled with confocal microscopy was employed to analyze synaptic inputs to GnRH cells located at the two levels examined. The results indicate that GnRH cells in the OVLT region receive a greater number of synaptophysin-immunoreactive appositions as compared with those located in the anterior septum. This supports the existence of subsets among the GnRH cells located in the basal forebrain. The effect of estradiol on the number of synaptophysin-immunoreactive appositions onto GnRH cells was also studied. Treatment of ovariectomized mice with estradiol significantly enhanced the number of synaptophysin-immunoreactive appositions to GnRH cells located at both levels examined. Thus the effect of estrogen on GnRH cells may be mediated in part by changes in the number of synaptic contacts.

Increased synaptic input to gonadotropin releasing hormone cells in preoptic area grafts that support reproductive development in female hypogonadal mice.

Ultrastructural studies have established that gonadotropin releasing hormone (GnRH) neuronal cell bodies receive sparse synaptic input compared to other neuronal cell types. In the present studies, immunocytochemistry for the presynaptic marker synaptophysin, coupled with confocal microscopy, was employed to evaluate whether there was a difference in synaptic input to GnRH cells within preoptic area grafts (hypogonadal, HPG; preoptic area, POA) in hypogonadal female mice that did or did not show ovarian development. GnRH cells in HPG/POA mice with ovarian development exhibited significantly higher numbers of synaptophysin immunoreactive (syn-IR) appositions as compared with HPG/POA mice without ovarian development. This suggests that synaptic input to the grafted GnRH cells is important for the correction of reproductive functions in HPG/POA mice. Following mating, Fos immunoreactivity was present in several GnRH cells in HPG mice with successful POA grafts, indicating the establishment of neuronal projections conveying somatosensory information to the GnRH cells in these mice. The presence of a higher number of syn-IR appositions to GnRH cells in the successful grafts supports this hypothesis.

Low-sexually performing rams but not male-oriented rams can be discriminated by cell size in the amygdala and preoptic area: a morphometric study.

Brain regions of male sheep behaviorally classified as high-sexually performing (n=10), low-sexually performing (n=8) or male-oriented (n=9) were examined to determine if differences in reproductive behavior were associated with differences in density or sizes of neurons. High-sexually performing rams actively mounted estrous ewes, low-sexually performing rams failed to mount or had long latencies to mounting estrous ewes, and male-oriented rams mounted other rams in preference to ewes in estrus. Cell densities and sizes were quantified in Nissl stained sections through the medial amygdala (meAMY), preoptic area (POA), bed nucleus of the stria terminalis (BNST), ventromedial hypothalamic nucleus (VMH), lateral geniculate nucleus (LG) and medial geniculate nucleus (MG). Multivariate discriminant analysis based on soma sizes within nuclei of known importance for reproductive behavior and/or gonadotropin release (meAMY, POA, BNST and VMH) discriminated (Wilks Lambda P<0.05) low-performing rams from high-performing and male-oriented rams, but did not discriminate (Wilks Lambda P=0.14) between high-performing and male-oriented rams. Cell size in the parvocellular and magnocellular layers of the LG along with cells of the MG, structures without a specific role in reproduction, did not discriminate any of the three behaviorally defined groups of rams (Wilks Lambda P=0.57). Density of cells present in structures important for the display of reproductive behavior (POA, meAMY, BNST) and/or gonadotropin release (POA, VMH) had no discriminating power nor did density of cells in structures important for the processing of visual (LG) or auditory (MG) stimuli. In conclusion, significant differences in sizes of cells located within nuclei that are specifically important for the display of male reproductive behavior were found in low-sexually performing rams compared to high-sexually performing and male-oriented rams. These differences may result from neuron development in utero or occur later as a consequence of endocrine factors or behavioral experience. Neuronal cell size is a critical variable that determines excitability to synaptic inputs because cell surface area varies exponentially with cell diameter. Relatively small differences in neuron diameter could relate to functionally important differences in neuronal excitability.

Neuroendocrine mechanisms for reproductive senescence in the female rat: gonadotropin-releasing hormone neurons.

Reproductive aging in female rats is characterized by profound alterations in the neuroendocrine axis. The preovulatory luteinizing hormone (LH) surge is attenuated, and preovulatory expression of the immediate early gene fos in gonadotropin-releasing hormone (GnRH) neurons is substantially reduced in middle-aged compared with young rats. We tested the hypothesis that alterations in GnRH gene expression may be correlated with the attenuation of the LH surge and may be a possible mechanism involved in neuroendocrine senescent changes. Sprague-Dawley rats ages 4 to 5 mo (young), 12-14 mo (middle-aged), or 25 to 26 mo (old) were killed at 10:00 AM or 3:00 PM on proestrus, the day of the LH surge, or diestrus I in cycling rats, and on persistent estrus or persistent diestrus in acyclic rats. RNase protection assays of GnRH mRNA and GnRH primary transcript were performed. GnRH mRNA levels increased significantly with age, whereas GnRH primary transcript levels, an index of GnRH gene transcription, decreased in old compared to young and middle-aged rats. This latter result suggests that an age-related change in GnRH mRNA levels occurs independently of a change in gene transcription, indicating a potential posttranscriptional mechanism. On proestrus, GnRH mRNA levels increased significantly from 10:00 AM to 3:00 PM in young rats. This was in contrast to proestrous middle-aged rats, in which this afternoon increase in GnRH mRNA levels was not observed. Thus, the normal afternoon increase in GnRH mRNA levels on proestrus is disrupted by middle age and may represent a substrate for the attenuation of the preovulatory GnRH/LH surge that occurs in rats of this age, prior to reproductive failure.

Prenatal testosterone masculinizes synaptic input to gonadotropin-releasing hormone neurons in sheep.

In sheep, the control of tonic and surge GnRH secretion is sexually differentiated by testosterone in utero. However, GnRH neurons are not sexually dimorphic with respect to number, distribution, or gross morphology. Therefore, this study tested the hypothesis that prenatal steroids influence synaptic input to GnRH neurons. We compared the number of synapses on GnRH neurons from male, female, and androgenized female lambs (n = 5 each). Androgenized females were exposed to testosterone during mid-gestation. Yearling lambs were perfused, and GnRH neurons were visualized using the LR-1 antibody. Five to seven GnRH neurons from the rostral preoptic area in each animal were viewed at the ultrastructural level. Afferent synapses and glial ensheathment on each neuron were counted in a single section through the plane of the nucleus. GnRH neurons from females received approximately twice as many contacts (3.6 +/- 0.7 synapses/100 microm plasma membrane) as those from male lambs (1.6 +/- 0.3; p < 0.05), similar to previous reports in rats. In addition, the number of synapses on GnRH neurons from androgenized female lambs (1.5 +/- 0.5) was similar to that from male lambs, suggesting that prenatal steroids give rise to sex differences in synaptic input to GnRH neurons.

Axon sparing lesions of the medial preoptic area block female sexual behavior.

The role of the medial preoptic area (mPOA) in regulating female musk shrew sexual behavior was assessed with excitatory neurotoxin, N-methyl-D-aspartate (NMDA) lesions. Ovariectomized, testosterone-implanted females that received lesions in the mPOA were statistically less likely to show complete sex behavior as compared to controls. These data suggest that the mPOA plays an activational role in testosterone-induced female sexual behavior.

Characterization of pars intermedia connections in amphibians by biocytin tract tracing and immunofluorescence aided by confocal microscopy.

Biocytin, recently introduced in neuroanatomical studies, was used as a retrograde tract tracer in combination with immunofluorescence in order to analyse the neurochemical characters of some central neuronal projections to the pars intermedia in two amphibian species, the anuran Rana esculenta and the urodele Triturus carnifex. After biocytin insertions in the pars intermedia, neurons became retrogradely labelled in the suprachiasmatic hypothalamus and the locus coeruleus of the brainstem in both species. Some scattered biocytin-labelled neurons were observed in the preoptic area. Moreover, working on the same sections, immunofluorescence revealed a number of codistributions and, in some cases, colocalization in the same neurons of biocytin labellings and immunopositivity for (1) tyrosine hydroxylase in the suprachiasmatic hypothalamus and the locus coeruleus of Rana and Triturus, (2) gamma-aminobutyric acid in the suprachiasmatic hypothalamus of Rana and Triturus and (3) neuropeptide Y in the suprachiasmatic hypothalamus of Rana. The specificity of such colocalizations was fully confirmed using dual-channel confocal laser scanning microscopy analysis.

Distribution of vasotocin- and mesotocin-like immunoreactivities in the brain of typhlonectes compressicauda (Amphibia, gymnophiona): further assessment of primitive and derived traits of amphibian neuropeptidergic systems.

To further assess primitive and derived conditions, we have studied the vasotocinergic (AVT) and mesotocinergic (MST) systems by immmunohistochemistry in the brain of Typhlonectes compressicauda. This species belongs to a separate order of amphibians which differs in several morphological and behavioral aspects from anurans and urodeles which have been studied previously. Nevertheless, the vasotocinergic and mesotocinergic systems of T. compressicauda are largely comparable to those of other amphibians. Apart from a well-developed hypothalamo-hypophyseal system, extrahypothalamic AVT-and MST-immunoreactive groups of cells and extensive networks of fibers were found. A major difference, however, is that neuropeptidergic cells in the caudal hypothalamus and the midbrain tegmentum of T. compressicauda contain MST, whereas those in corresponding locations contain AVT in anurans and urodeles. This suggests that certain neuropeptidergic cell groups in the gymnophionan brain have switched from AVT to MST gene expression, and, thereby, offers a new view on the functional significance of these neuropeptidergic systems.

Nonspecific esterase activity in astrocytes of the goldfish brain.

We examined the distribution of nonspecific esterase (NSE) activity in the brain of the goldfish, Carassius auritus, and found that the enzyme is expressed at high levels in cells that appeared to be radial astrocytes. Several instances in which neurons expressed NSE activity were also seen. To confirm the identity of the radial profiles as astrocytes, similar sections were labeled with antiserum against goldfish glial fibrillary acidic protein (GFAP). The concordance between the NSE and the anti-GFAP data in both the visual system and the telencephalon was essentially complete, confirming that the NSE reaction was labeling astrocytes in these structures. The two methods also gave similar results in both the cerebellum and the vagal lobes, although the concordance between them in these instances was somewhat less complete. Both the NSE reaction and immunohistochemistry with anti-GFAP serum revealed labeled nonradial cells lying free in the cerebellar molecular layer. We suggest that these cells may represent free astrocytes, a cell type that has not previously been reported in morphological studies of the teleostean brain. On the basis of our observations, we suggest that the NSE reaction may be a useful adjunct in morphological studies of teleost astroglia. Finally, we propose that the expression of NSE activity in goldfish astrocytes may he related to their ability to internalize neural debris during Wallerian degeneration.