Immunohistochemical subpopulations of retinopetal neurons in the nucleus olfactoretinalis in a teleost, the whitespotted greenling (Hexagrammos stelleri)

Neural Control of Sexual Behavior in Fish

Many vertebrate species show breeding periods and exhibit series of characteristic species-specific sexual behaviors only during the breeding period. Here, secretion of gonadal sex hormones from the mature gonads has been considered to facilitate sexual behaviors. Thus, the sexual behavior has long been considered to be regulated by neural and hormonal mechanisms. In this review, we discuss recent progress in the study of neural control mechanisms of sexual behavior with a focus on studies using fish, which have often been the favorite animals used by many researchers who study instinctive animal behaviors. We first discuss control mechanisms of sexual behaviors by sex steroids in relation to the anatomical studies of sex steroid-concentrating neurons in various vertebrate brains, which are abundantly distributed in evolutionarily conserved areas such as preoptic area (POA) and anterior hypothalamus. We then focus on another brain area called the ventral telencephalic area, which has also been suggested to contain sex steroid-concentrating neurons and has been implicated in the control of sexual behaviors, especially in teleosts. We also discuss control of sex-specific behaviors and sexual preference influenced by estrogenic signals or by olfactory/pheromonal signals. Finally, we briefly summarize research on the modulatory control of motivation for sexual behaviors by a group of peptidergic neurons called terminal nerve gonadotropin-releasing hormone (TN-GnRH) neurons, which are known to be especially developed in fishes among various vertebrate species.

The centrifugal visual system of vertebrates: a comparative analysis of its functional anatomical organization

The present review is a detailed survey of our present knowledge of the centrifugal visual system (CVS) of vertebrates. Over the last 20 years, the use of experimental hodological and immunocytochemical techniques has led to a considerable augmentation of this knowledge. Contrary to long-held belief, the CVS is not a unique property of birds but a constant component of the central nervous system which appears to exist in all vertebrate groups. However, it does not form a single homogeneous entity but shows a high degree of variation from one group to the next. Thus, depending on the group in question, the somata of retinopetal neurons can be located in the septo-preoptic terminal nerve complex, the ventral or dorsal thalamus, the pretectum, the optic tectum, the mesencephalic tegmentum, the dorsal isthmus, the raphé, or other rhombencephalic areas. The centrifugal visual fibers are unmyelinated or myelinated, and their number varies by a factor of 1000 (10 or fewer in man, 10,000 or more in the chicken). They generally form divergent terminals in the retina and rarely convergent ones. Their retinal targets also vary, being primarily amacrine cells with various morphological and neurochemical properties, occasionally interplexiform cells and displaced retinal ganglion cells, and more rarely orthotopic ganglion cells and bipolar cells. The neurochemical signature of the centrifugal visual neurons also varies both between and within groups: thus, several neuroactive substances used by these neurons have been identified; GABA, glutamate, aspartate, acetylcholine, serotonin, dopamine, histamine, nitric oxide, GnRH, FMRF-amide-like peptides, Substance P, NPY and met-enkephalin. In some cases, the retinopetal neurons form part of a feedback loop, relaying information from a primary visual center back to the retina, while in other, cases they do not. The evolutionary significance of this variation remains to be elucidated, and, while many attempts have been made to explain the functional role of the CVS, opinions vary as to the manner in which retinal activity is modified by this system.

GnRH-immunoreactive centrifugal visual fibers in the Nile crocodile (Crocodylus niloticus)

Thin varicose centrifugal visual fibers, between 30-45 in number and displaying cGnRH-I immunoreactivity, were identified in Crocodylus niloticus. Approximately 80% of these fibers were also FMRF-amide-like immunoreactive. The cGnRH-I fibers extended from the preoptic region to the retina where they appeared to terminate in the external portion of the inner plexiform layer. The location of their neurons of origin could not be determined precisely following the intraocular injection of the retrograde axonal tracer RITC. Nevertheless, the presence of cGnRH-I-immunoreactive neurons exclusively within the complex comprising the terminal nerve and the septo-preoptic region, and of several retinopetal fibers labelled retrogradely with the axonal tracer at the septo-preoptic junction, indicates that the cGnRH-immunoreactive centrifugal visual system originates from within this complex.

Terminal nerve and vision

The vertebrate retina receives efferent input from different parts of the central nervous system. Efferent fibers are thought to influence retinal information processing but their functional role is not well understood. One of the best-described retinopetal fiber systems in teleost retinae belongs to the terminal nerve complex. Gonadotropin-releasing hormone (GnRH) and molluscan cardioexcitatory tetrapeptide (FMRFamide)-containing fibers from the ganglion of the terminal nerve form a dense fiber plexus in the retina at the border of the inner nuclear and inner plexiform layer. Peptide-containing fibers surround and contact perikarya of dopaminergic interplexiform cells in teleost retina. In vitro experiments demonstrated that exogenously supplied GnRH mediates dopaminergic effects on the membrane potential and on the morphology of dendritic tips (spinules) of cone horizontal cells. These effects can be specifically blocked by GnRH-antagonists, indicating that the release of dopamine and dopamine-dependent effects on light adaptation of retinal neurons are affected by the terminal nerve complex. Recent data have shown that olfactory information has an impact on retinal physiology, but its precise role is not clear. The efferent fiber of the terminal nerve complex is one of the first retinopetal fiber systems for which the sources of the fibers, their cellular targets, and several physiological, morphological, and behavioral effects are known. The terminal nerve complex is therefore a model system for the analysis of local information processing which is influenced by a distinct fiber projection.

Gonadotropin-releasing hormone neuronal development during the sensitive period of temperature sex determination in the pejerrey fish, Odontesthes bonariensis

The development of gonadotropin-releasing hormone (GnRH) neurons was studied in relation to the sensitive period of thermolabile sex determination in the pejerrey Odontesthes bonariensis, an atherinid fish from South America. Fish were raised from hatching at three different temperatures: 17 degrees C (100% females), 24 degrees C (70% females), and 29 degrees C (100% males). Three groups of immunoreactive GnRH (ir-GnRH) neurons were identified at the terminal nerve ganglion (TNG), the midbrain tegmentum (MT), and the preoptic area (POA). Immunoreactive GnRH (ir-GnRH) neurons were identified in the TNG at hatching (day 0) and in the MT at day 3 at all the experimental temperatures. In the POA ir-GnRH neurons were identified in the nucleus preopticus periventricularis simultaneously with the first appearance of ir-GnRH fibers in the pituitary on days 11, 14, and 17 for larvae kept at 29, 24, and 17 degrees C, respectively. The number of ir-GnRH neurons in the TNG did not show any statistical difference between temperatures. The number of ir-GnRH neurons in the MT increased in number during the experiment for larvae kept at 17 and 24 degrees C but decreased between days 17 and 31 in larvae kept at 29 degrees C. The number of ir-GnRH neurons in the POA increased during development with a peak at day 28 for all temperatures studied and the magnitude of this peak showed a correlation with incubation temperature. These results reinforce the notion that the hypothalamus-pituitary-gonadal axis is active during sex determination in pejerrey suggesting a possible role of the central nervous system and GnRH in this process. It is also suggested that GnRH neurons located in the preoptic area might be the physiological transducers of temperature during the temperature sensitive period in this species.

The terminal nerve ganglion cells project to the olfactory mucosa in the dwarf gourami

Single- and double-label immunocytochemical studies were conducted using antisera to salmon gonadotropin-releasing hormone (sGnRH) and molluscan cardioexcitatory peptide (FMRFamide) to determine whether terminal nerve ganglion cells project to the olfactory mucosa in the dwarf gourami, Colisa lalia. Both peptides were present in terminal nerve ganglion perikarya and fibers in brain and nasal cavity. Labeled fibers were present in the olfactory nerve and could be traced to the olfactory mucosa. All terminal nerve ganglion cells contained both sGnRH and FMRFamide-like peptides. This study suggests that the terminal nerve ganglion cells can influence both brain and chemoreceptive structures.

Ontogenetic organization of the FMRFamide immunoreactivity in the nervus terminalis of the lungfish, Neoceratodus forsteri

The development of the nervus terminalis system in the lungfish, Neoceratodus forsteri, was investigated by using FMRFamide as a marker. FMRFamide immunoreactivity appears first within the brain, in the dorsal hypothalamus at a stage around hatching. At a slightly later stage, immunoreactivity appears in the olfactory mucosa. These immunoreactive cells move outside the olfactory organ to form the ganglion of the nervus terminalis. Immunoreactive processes emerge from the ganglion of the nervus terminalis in two directions, one which joins the olfactory nerve to travel to the brain and the other which courses below the brain to enter at the level of the preoptic nucleus. Neither the ganglion of the nervus terminalis nor the two branches of the nervus terminalis form after surgical removal of the olfactory placode at a stage before the development of FMRFamide immunoreactivity external to the brain. Because this study has confirmed that the nervus terminalis in lungfish comprises both an anterior and a posterior branch, it forms the basis for discussion of homology between these branches and the nervus terminalis of other anamniote vertebrates.

Comparative immunocytochemical study of FMRFamide neuronal system in the brain of Danio rerio and Acipenser ruthenus during development

The distribution of FMRFamide-like immunoreactive (ir) neurons and fibers was investigated in the central nervous system of developing zebrafish and juvenile sturgeon (sterlet). Adult zebrafish was also studied. In zebrafish embryos FMRFamide-ir elements first appeared 30 h post-fertilization (PF). Ir somata were located in the olfactory placode and in the ventral diencephalon. FMRFamide-ir fibers originating from diencephalic neurons were found in the ventral telencephalon and in ventral portions of the brainstem. At 48 h PF, the ir perikarya in the olfactory placode displayed increased immunoreactivity and stained fibers emerged from the somata. At 60 h PF, bilaterally, clusters of FMRFamide-ir neurons were found along the rostro-caudal axis of the brain, from the olfactory placode to rostral regions of the ventro-lateral telencephalon. At 60 h PF, numerous ir fibers appeared in the dorsal telencephalon, optic lobes, optic nerves, and retina. Except for ir fibers in the hypophysis at the age of 72 h PF, and a few ir cells in the nucleus olfacto-retinalis (NOR) at the age of 2 months PF, no major re-organization was noted in subsequent ontogenetic stages. The number of stained NOR neurons increased markedly in sexually mature zebrafish. In adult zebrafish, other ir neurons were located in the dorsal zones of the periventricular hypothalamus and in components of the nervus terminalis. We are inclined to believe that neurons expressing FMRFamide originate in the olfactory placode and in the ventricular ependyma in the hypothalamus. On the same grounds, a dual origin of FMRFamide-ir neurons is inferred in the sturgeon, an ancestral bony fish: prior to the observation of ir cells in the nasal area and in the telencephalon stained neurons were noted in circumventricular hypothalamic regions.

Development of immunoreactivity to neuropeptide Y in the brain of brown trout (Salmo trutta fario)

The development of neuropeptide Y-immunoreactive (NPY-ir) neurons in the brain of the brown trout, Salmo trutta fario, was studied by using the streptavidin-biotin immunohistochemical method. Almost all NPY-ir neurons found in the brain of adults already appeared in embryonic stages. The earliest NPY-ir neurons were observed in the laminar nucleus, the locus coeruleus, and the vagal region of 9-mm-long embryos. In the lateral area of the ventral telencephalon, habenula, hypothalamus, optic tectum, and saccus vasculosus, NPY-ir cells appeared shortly after (embryos 12-14 mm in length). The finding of NPY-ir cells in the saccus vasculosus and the vagal region expand the NPY-ir structures known in teleosts. Among the regions of the trout brain most richly innervated by NPY-ir fibers are the hypothalamus, the isthmus, and the complex of the nucleus of the solitary tract/area postrema, suggesting a correlation of NPY with visceral functions. Two patterns of development of NPY-ir populations were observed: Some populations showed a lifetime increase in cell number, whereas, in other populations, cell number was established early in development or even diminished in adulthood. These developmental patterns were compared with those found in other studies of teleosts and with those found in other vertebrates. J. Comp. Neurol. 414:13-32, 1999.

Preoptic gonadotropin-releasing hormone (GnRH) neurons innervate the pituitary in teleosts

In most teleosts, there are three groups of gonadotropin-releasing hormone (GnRH) neurons. In this study we addressed the question of GnRH neuronal innervation of the pituitary in the dwarf gourami and the tilapia using immunocytochemistry combined with biocytin tract tracing. Biocytin was applied to the pituitary attached to the brain in vitro. Similar results were obtained in both species. GnRH neurons retrogradely labeled with biocytin were observed only in the preoptic area. These results indicate that preoptic GnRH neurons innervate the pituitary. Negative labeling of biocytin in the terminal-nerve and midbrain GnRH neurons suggests that these two GnRH neuronal populations do not project to the pituitary. Biocytin-positive but GnRH-negative neurons were also observed in the preoptic area and the ventromedial parts of the hypothalamus, suggesting neuropeptidergic and aminergic innervation of the pituitary besides GnRH.

Immunohistochemical and Ultrastructural characterization of the terminal nerve ganglion cells of the ayu, Plecoglossus altivelis (Salmoniformes, Teleostei)

BACKGROUND

Little is known on the cytological properties of the terminal nerve ganglion (TNG) cells in teleosts (Demski, 1993. Acta Anat., 148:81-95).

MATERIALS AND METHODS

To characterize the TNG cells of a salmonoid fish, Plecoglossus altivelis, we adopted immunohistochemistry and transmission electron microscopy (TEM).

RESULTS

The majority of the TNG cells formed a compact mass halfway between the olfactory sac and the olfactory bulb, whereas a few cells were scattered in the ventromedial region of the olfactory bulb. The cell had a voluminous perikaryon that was positive to antisera against gonadotropin-releasing hormone (GnRH), molluscan cardioexcitatory tetrapeptide (FM-RFamide), and neuropeptide Y (NPY). Immunostaining of consecutive sections with each antiserum showed the coexistence of these antigens in the same cells and their processes. Most of the processes originating from the cells projected centrally to the basal forebrain, including the optic nerve. With TEM, the cells revealed a peptidergic nature, i.e., the presence of abundant granular endoplasmic reticula and well-developed Golgi bodies in association with vesicles that were 70-100 nm in diameter. Occasionally, the cells adjoined one another directly without the intervention of glial sheets. Synaptic contacts were frequent in the proximal region of the processes, where thin lateral processes of the cells and axon terminals of unknown origin were intermingled with each other. Terminal buttons being engulfed by the soma were commonly seen.

CONCLUSIONS

The TNG cells of the salmonoid fish share many cytological characteristics with the cells of the nucleus olfactoretinalis of advanced teleosts such as acanthopterygians.

Characterization of the RFamide-like neuropeptides in the nervus terminalis of the goldfish (Carassius auratus)

FMRFamide-immunoreactivity has been demonstrated in the CNS of many vertebrate species. We sought to further characterize this immunoreactivity in nervus terminalis retinal efferents of the goldfish using an antiserum raised against a bovine morphine modulating peptide (A18Famide). This antiserum robustly labels nervus terminalis efferents to the retina, as well as a sub-population of retinal amacrine cells. Under immunocytochemical conditions the antiserum cross-reacted with neuropeptide Y-like as well as A18Famide-like peptides, but under conditions of radioimmunoassay it was highly specific for A18Famide-like peptides. High pressure liquid chromatography, gel permeation chromatography and radioimmunoassay showed that at least two different RFamide-like peptides, approximately the same size as the bovine RFamide-like peptides, are present in the goldfish nervus terminalis.

Substance P-, F8Famide-, and A18Famide-like immunoreactivity in the nervus terminalis and retina of the goldfish Carassius auratus

We re-investigated the occurrence of substance P-like immunoreactivity in the retina of the goldfish Carassius auratus using antisera to substance P and other tachykinins. Most antisera labelled a previously described single class of mono-stratified amacrine cells arborizing in layer 3 of the inner plexiform layer. Preabsorption experiments showed that these amacrine cells contained at least one tachykinin-like peptide. One antiserum (INC 353) to substance P labelled not only these amacrine cells but also fibres in layer 1 of the inner plexiform layer and fibres in the optic nerve. These fibres were identified as retinopetal projections of the nervus terminalis, in part because of colocalized labelling with antisera against gonadotropin-releasing hormone and FMRFamide. Preabsorption experiments showed that the substance P-immunoreactive material in the nervus terminalis was not substance P or any other typical tachykinin. Labelling of the nervus terminalis with INC 353 was blocked by preabsorption with two bovine FMRF-amide-like peptides, F8Famide and A18Famide, which contain a substance P(4-7)-like region. Antisera to F8Famide and A18Famide strongly labelled ganglia of the nervus terminalis and retinopetal fibres. We suggest that labelling of the nervus terminalis by antisera to substance P and FMRFamide occurs because of homologies between these antigens and a non-tachykinin, endogenous peptide that is similar to F8Famide and A18Famide.

Nervus terminalis ganglion of the bonnethead shark (Sphyrna tiburo): evidence for cholinergic and catecholaminergic influence on two cell types distinguished by peptide immunocytochemistry

The nervus terminalis is a ganglionated vertebrate cranial nerve of unknown function that connects the brain and the peripheral nasal structures. To investigate its function, we have studied nervus terminalis ganglion morphology and physiology in the bonnethead shark (Sphyrna tiburo), where the nerve is particularly prominent. Immunocytochemistry for gonadotropin-releasing hormone (GnRH) and Leu-Pro-Leu-Arg-Phe-NH2 (LPLRFamide) revealed two distinct populations of cells. Both were acetylcholinesterase positive, but LPLR-Famide-immunoreactive cells consistently stained more darkly for acetylcholinesterase activity. Tyrosine hydroxylase immunocytochemistry revealed fibers and terminal-like puncta in the ganglion, primarily in areas containing GnRH-immunoreactive cells. Consistent with the anatomy, in vitro electrophysiological recordings provided evidence for cholinergic and catecholaminergic actions. In extracellular recordings, acetylcholine had a variable effect on baseline ganglion cell activity, whereas norepinephrine consistently reduced activity. Electrical stimulation of the nerve trunks suppressed ganglion activity, as did impulses from the brain in vivo. During electrical suppression, acetylcholine consistently increased activity, and norepinephrine decreased activity. Muscarinic and, to a lesser extent, alpha-adrenergic antagonists both increased activity during the electrical suppression, suggesting involvement of both systems. Intracellular recordings revealed two types of ganglion cells that were distinguishable pharmacologically and physiologically. Some cells were hyperpolarized by cholinergic agonists and unaffected by norepinephrine; these cells did not depolarize with peripheral nerve trunk stimulation. Another group of cells did depolarize with peripheral trunk stimulation; a representative of this group was depolarized by carbachol and hyperpolarized by norepinephrine. These and other data suggest that the bonnethead nervus terminalis ganglion contains at least two cell populations that respond differently to acetylcholine and norepinephrine. The bonnethead nervus terminalis ganglion appears to differ fundamentally from sensory and autonomic ganglia but does share some features with the neural circuits of forebrain GnRH systems.

Multiple embryonic origins of gonadotropin-releasing hormone (GnRH) immunoreactive neurons

Experiments were conducted to test the hypothesis that gonadotropin-releasing hormone immunoreactive (GnRH-ir) and FMRFamide-ir neurons present in the brain and nervus terminalis originate in the embryonic olfactory placode. The olfactory placodes were bilaterally extirpated in stage 26 or stage 29 embryos of the axolotl, Ambystoma mexicanum, which were then reared for 4-8 months before they were examined immunohistochemically. In experimental subjects with bilateral loss of olfactory epithelia, nerves and bulbs, there was complete absence of GnRH- and FMRFamide-ir neurons in the terminal nerve, and in septal and preoptic areas, and complete absence of large diameter peptidergic fibers associated with the TN-septo-preoptic system. However, GnRH-ir perikarya in the posterior tubercle, and FMRFamide-ir perikarya in the ventral hypothalamus, and small diameter peptidergic fibers were not affected by placodal ablation. These results support the hypothesis that contrary to recent reports, GnRH-ir neurons have more than one embryonic origin. Region-specific patterns of staining with antisera directed against different molecular forms of GnRH support the interpretation that GnRH-ir neurons of placodal origin express mammalian GnRH, whereas GnRH-ir neurons of non-placodal origin, in the posterior tubercle, express chicken GnRH II.

Neuroanatomical pathways linking vision and olfaction in mammals

Retinal projections to several telencephalic structures have been demonstrated in a wide range of mammalian species following intraocular injections of tritiated amino acids and cholera toxin subunit-B conjugated to horseradish peroxidase. Since these regions are also innervated by olfactory fibers, we investigated the distribution of convergent projections using simultaneous injections of different anterograde tracers in the eye and olfactory bulbs. Convergent projections from the retina and from the olfactory bulbs were observed in the piriform cortex, olfactory tubercle, the cortical region of the medial amygdala, lateral hypothalamus, and the bed nucleus of the stria terminalis. A few retinal fibers also invade the nucleus of the lateral olfactory tract, the bed nucleus of the accessory olfactory bulb and the diagonal band of Broca. Injections of retrograde tracers in the medial amygdala, the bed nucleus or the lateral hypothalamus shows that the visuo-olfactory convergence mainly involves projections originating from the accessory olfactory bulb, and to a lesser extent from the ventromedial region of the main olfactory bulb. Fewer than 20 retinotelencephalic ganglion cells were identified in the retina, mainly located contralateral to the injection site. Ganglion cells were medium sized and possessed two long slender opposing dendrites. These retinal and olfactory projections could provide an anatomical substrate for the modulation of gonadotropin hormone levels and the olfactory influence on light mediated rhythms related to reproductive physiology.

FMRFamide-like immunoreactivity in the brain of the Pacific hagfish, Eptatretus stouti (Myxinoidea)

The distribution of FMRFamide-like immunoreactivity was investigated in the brain of a myxinoid, the Pacific hagfish, Eptatretus stouti, by means of immunocytochemistry. In the forebrain, labelled cell bodies occurred in the infundibular nucleus of the hypothalamus and some closely adjacent nuclei. Labelled fibers formed a diffuse network in the forebrain, but there was no evidence for the presence of intracerebral ganglionic cells of the terminal nerve or a central projection of the terminal nerve. In the hindbrain, a group of labelled cells was found in the trigeminal sensory nucleus. A distinct terminal arborization occurred in the ventrally adjacent nucleus A of Kusunoki and around the nuclei of the branchial motor column. These findings suggest that FMRFamide may play a role in the central control of branchiomotor activity.

Distribution of FMRFamide-like immunoreactivity in the forebrain of the catfish, Clarias batrachus (Linn.)

The anatomical distribution of FMRFamide-like immunoreactivity in the forebrain and pituitary of the catfish, Clarias batrachus, was investigated. Immunoreactive cells were observed in the ganglion cells of the nervus terminalis (NT) and in the medial olfactory tracts. In the preoptic area, FMRFamide-containing perikarya were restricted to the lateral preoptic area, paraventricular subdivision of the nucleus preopticus, nucleus suprachiasmaticus and nucleus preopticus periventricularis posterior. In the postoptic area, some cells of the nucleus postopticus lateralis and nucleus of the horizontal commissure showed moderate immunoreactivity. In the tuberal area, immunoreactivity was observed in few cells of the nucleus hypothalamicus ventralis and nucleus arcuatus hypothalamicus (NAH). Nucleus ventromedialis thalami was the only thalamic nucleus with FMRFamide immunoreactivity. Immunoreactive processes were traceable from the NT through the medial as well as lateral olfactory tracts into the telencephalon and the area ventralis telencephali pars supracommissuralis (Vs). Further caudally, the immunoreactive fibers could be traced into discrete areas, including habenular and posterior commissures, neurohypophysis and pituitary; isolated fibers were also observed in the pineal stalk. A loose network of immunoreactive processes was observed in the olfactory bulbs and the entire telencephalon, with higher densities in some areas, including Vs. A dense plexus of immunoreactive fibers was seen in the pre- and postoptic areas and around the paraventricular organ, while relatively few were observed in the thalamus. A high concentration of fiber terminals was found in the caudal tuberal area.

Ultrastructure of the ganglion cells of the terminal nerve in the dwarf gourami (Colisa lalia)

In our previous light microscopic studies (Oka et al., Brain Res. 367: 341-345, '86; Oka and Ichikawa, J. Comp. Neurol. 300: 511-522, '90), we reported that there are at least two types of terminal nerve (TN) cells based on cell size and immunoreactivity: type I cells had large cell bodies, while type II cells had smaller cell bodies. Type I TN cells were immunoreactive to gonadotropin-releasing hormone (GnRH) and may be the major source of GnRH-immunoreactive fibers that are widely distributed throughout the brain. Type II TN cells, on the other hand, were not immunoreactive to GnRH. In the present paper, we examined the cytology and synaptology of these two types of TN cells with electron microscopy. Type I TN cell bodies were found to have morphological characteristics similar to those of other peptide-synthesizing neurons and are likely to be actively synthesizing GnRH. The frequent occurrence of coated vesicles close to the plasma membrane of the cell body was suggestive of membrane retrieval following exocytosis of the vesicular contents from the cell surface. Neighboring TN cells were either in direct juxtaposition with one another or made specialized "glomeruloid" cell-to-cell contacts; these specializations may be relevant for nonsynaptic intercellular communications among the TN cells. Within these glomeruloid complexes, the somatic processes of TN cells received inputs from two types of synaptic terminals: one containing only spherical synaptic vesicles and another containing a small number of dense-cored vesicles in addition to the spherical synaptic vesicles. Axosomatic synapses were rare on type I TN cell bodies. In contrast, type II TN cell bodies had morphological characteristics similar to those of neurons in other brain regions. These receive axosomatic inputs from synaptic terminals containing only spherical synaptic vesicles and those with a small number of dense-cored vesicles in addition to the spherical synaptic vesicles. Thus, each type of TN cell has unique fine structural characteristics which may correlate to their different functional roles.

Gonadotropin-releasing hormone (GnRH) immunoreactive system in the brain of the dwarf gourami (Colisa lalia) as revealed by light microscopic immunocytochemistry using a monoclonal antibody to common amino acid sequence of GnRH

The present paper aims to give a morphological basis for the study of the terminal nerve system and its relation to the whole gonadotropin-releasing hormone (GnRH) immunoreactive (ir) neuronal system. We examined the GnRH-ir neuronal system of a tropical fish, the dwarf gourami, by using a recently developed monoclonal antibody against GnRH (LRH13) which recognizes the amino acid sequence common to all known variants of GnRH (Park and Wakabayashi, Endocrinol. Jpn. 33:257-272, '86). The ganglion cells of the terminal nerve (TN-ggl cells) in the transitional area between the olfactory bulb and the telencephalon reacted strongly with the LRH13. A distinct bundle of axons emanating from the TN-ggl cells ran caudally through the ventral telencephalon and the preoptic area. Some of these axons entered the optic nerve and innervated the retina. The remaining axons continued caudally to enter the hypothalamus and the midbrain. A second group of GnRH-ir cell bodies was found in the preoptic area. A distinct bundle of GnRH-ir fibers originating from these cell bodies innervated the pituitary. This pathway is equivalent to the preoptico-infundibular pathway of other vertebrates, and the GnRH in this pathway is presumed to function as hypophysiotrophic hormone to facilitate the release of gonadotropins from the pituitary. The distribution of GnRH-ir fibers in the brain was extensive. Most fibers apparently originated from the TN-ggl cells and covered various brain regions from the olfactory bulb to the spinal cord. They were especially abundant in the olfactory bulb, ventral telencephalon, preoptic area, optic tectum, and some hypothalamic areas. Thus, GnRH might function as a neuromodulator and/or neurotransmitter in these areas. The abundant GnRH-ir fibers in the ventral telencephalon and the preoptic area might affect some aspects of sexual behavior, since these areas have been suggested to be involved in the control of sexual behavior in teleosts.

Electron microscopic study of immunocytochemically labeled centrifugal fibers in the goldfish retina

The centrifugal fibers innervating the goldfish retina were studied quantitatively by light and electron microscopy. These fibers originating from cell bodies in the olfactory bulb were labeled by antiserum to the tetrapeptide Phe-Met-Arg-Phe-NH2 (FMRFamide). The number of FMRFamide-immunoreactive (ir) centrifugal fibers in each eye of the adult goldfish (body length: 12-15 cm) was 65 +/- 14 (mean +/- S.D., n = 7). All of these fibers in the optic nerve and the retina were unmyelinated. Each FMRFamide-ir centrifugal fiber runs along the optic fiber layer and gives several terminal arborizations in the outermost layer (layer 1) of the inner plexiform layer. Layer 1 is, therefore, densely covered by a plexus of terminal arborizations. Along these terminal arborizations, we found output synapses characterized by a cluster of small clear vesicles (40 nm in diameter) at the presynaptic site and a thickened membrane in the apposed retinal cell processes. In a sample area of 2,000 microns 2, such synapses occurred at a density of one per 105 microns 2, or about 13,000 per centrifugal fiber. Thus, the FMRFamide-ir centrifugal fibers are likely to modulate retinal cell activity through an estimated total of 840,000 output synapses per retina.