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

DEVELOPMENTAL BRAIN RESEARCH

ELSEVIER

Developmental Brain Research 78 (1994) 279-290

Research Report

Multiple embryonic origins of gonadotropin-releasing hormone (GnRH) immunoreactive neurons R. Glenn Northcutt a, Linda E. Muske b,, '~ Neurobiology Unit - Scripps Institute of Oceanography and Department of Neuroscience - School of Medicine, University o f California at San Diego, La Jolla, 90201 CA, USA h Biology Department, Franklin and Marshall College, Lancaster, PA 17604, USA (Accepted 19 October 1993)

Abstract

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 tb.e 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.

Key words: Luteinizing hormone-releasing hormone; Chicken GnRH II; FMRFamide; Olfactory placode; Nervus terminalis; Brain-pituitary-gonad axis; Posterior tubercle

I. Introduction

The neuropeptide gonadotropin-releasing hormone ( G n R H , L H R H ) plays a critical role in the reproductive cycle of vertebrates. G n R H initiates sexual maturation through stimulation .of gonadotropic cells of the anterior pituitary. G n R H can also act directly on target cells in the brain to facilitate species-typical behaviors associated with reproduction [31,30]. Immunocytochemical (ICC) studies of G n R H immunoreactive (ir) neurons reveal many similarities among vertebrates in their cytoarchitectonic organization [2,35,54]. G n R H - i r cell bodies of the brain-pituitary-gonad (BPG) axis are located principally in the medial septum and anterior preoptic area. These cells give rise to the major G n R H axonal projection through the ventral hypothalamus to

* Corresponding author. Fax: (1) (717) 291-4088. 0165-3806/94/$07.00 © 1994 Elsevier Science B.V. All rights reserved SSDI 01 6 5 - 3 8 0 6 ( 9 3 ) E 0 1 8 4 - M

the median eminence. G n R H - i r perikarya are also present in the nervus terminalis (terminal nerve; TN), a small cranial nerve associated with the olfactory system and ventral forebrain, and in several regions of the posterior diencephalon a n d / o r midbrain [35]. G n R H - i r fibers are extensively distributed throughout all major regions of the central nervous system (CNS). The G n R H peptide occurs in at least seven different molecular forms. N a m e d after the taxa in which the peptides were first found (e.g. salmon G n R H , chicken G n R H I and II), the phyletic distribution of the different G n R H s is widespread, and multiple forms of the peptide occur within .the brain of a single species [for recent reviews, see [23,35,52]]. The diversity in molecular structure, function and anatomical distribution of G n R H raises the question of whether all G n R H neurons are embryonically derived from a single precursor population, or have multiple origins. Based on a number of earlier descriptive stud-

280

R. G. Northcutt, L.E. Muske / Developmental Brain Research, 78 (1994) 279-290

ies, the ganglionic cells of the nervus terminalis are believed to arise embryonically from the olfactory placodes [8,44]. In adult urodele amphibians, GnRH-ir neurons of the TN and septo-preoptic area comprise an anatomical continuum of morphologically homogeneous neurons, which all lie along or close to central projections of the TN [36,37]. On the basis of this finding, Muske and Moore proposed that the two populations share a common embryonic origin in the olfactory placode, and that they migrate centrally during early embryogenesis, along the outgrowing TN. Subsequently, studies of mammals revealed that GnRH-ir cells are present in the nasal region prior to their appearance in the brain, and that the distribution of GnRH-ir cells shifts to progressively more caudal regions as development progresses [50,66]. However, these studies did not consider GnRH-ir neurons located in more posterior brain areas, including the periventricular hypothalamus, posterior tubercle, and midbrain tegrnentum. Posterior GnRH-ir neurons have been identified in all major vertebrate taxa (elasmobranchs [68,69]; teleosts [1,20,29,32,58,67]; amphibians [6,49] reptiles [4,13]; birds [24,26,28]; mammals [11]). Posterior GnRH-ir neurons are not anatomically contiguous with those of the TN/septo-preoptic network, and the two populations differ with respect to location, dendritic morphology, and immunology [35]. These observations suggest that there may be two embryonic sources of GnRH neurons. This hypothesis receives support from developmental studies in anuran amphibians, which showed that the timetable for expression of immunoreactive GnRH in the brainstem and spinal cord differed from that in the TN and septo-preoptic area [38,65]. A second neuropeptide, similar to the molluscan cardioexcitatory peptide Phe-Met-Arg-Phe-NH 2 (FMRFamide), has also been reported in the TN of many vertebrates [25]. Among amphibians, FMRFamide-ir neurons have been identified in the TN of anuran [37,60,63], but not urodele amphibians [37]. FMRFamide-like substance(s) also occur in other regions of the vertebrate CNS [10,14,15,19,48]. In amphibians, these neurons appear to be localized to periventricular areas of the ventral preoptic area and hypothalamus of some amphibians [37]. The present study utilized ICC techniques and a variety of antisera directed against GnRH molecular variants and FMRFamide. Our goals were to determine whether FMRFamide-ir neurons are present in the TN of a urodele amphibian, and to evaluate the hypothesis that neuropeptide-containing cells associated with the TN originate in the olfactory placode, but that periventricular brain neurons expressing the same, or similar peptides have a different embryonic origin. A preliminary report of these results has been previously published [42].

2. Materials and Methods Fertilized eggs of wild-type axolotls, Ambystorna mexicanum, were obtained from the Indiana University Axolotl Colony, and the olfactory placodes were bilaterally extirpated at stage 26 (n = 10) or stage 29 (n = 20). Control embryos from each experimental clutch were demembraned and exposed to the same operating and rearing solutions as the placode-extirpated siblings. Operations were carried out in glass petri dishes lined with a 1:1 solution of beeswax and paraffin, filled with an operating solution containing: NaC1 345 mM, KCI 1.3 mM, and CaC12 0.68 mM and 1.9 ml/liter of the antibiotic Cotrim-ratiopharm (Ratiopharm GmbH and Co., Ulm, Germany). Eggs were placed in operating solution containing 0.33% formalin for 10 rain to kill fungi. They were then transferred to operating dishes, and the vitelline membranes were removed. Olfactory placodes, identified as bilateral thickenings of the ectoderm located medial to the lens placodes, were excised using fine tungsten microneedles. Surgery was performed on at least ten embryos from three different clutches. Following surgery, embryos were transferred to a solution of the same composition as the operating solution, with the salts reduced by 25%. Twenty experimental and twenty control larvae were raised for 4-8 months. They were then anaesthetized and perfused with cold 0.1 M phosphate buffer, followed with buffered 4% paraformaldehyde (pH 7.4). The brains were removed, embedded in 10% gelatin, and postfixed for 12 h. Frozen sections were cut at 25-50 /zm in either the transverse or horizontal plane. Experimental and control brains were paired during embedding and sectioning, and were processed for immunocytochemistry in tandem. Approximately half the brains were processed as follows: freefloating sections were incubated in primary antibody, followed by goat anti-rabbit IgG (1 : 100), rabbit peroxidase-antiperoxidase, and a solution containing 0.035% diaminobenzidine (Sigma), 0.15% hydrogen peroxide, and 2.0% nickel ammonium sulfate. The other brains were thaw-mounted onto gelatinized slides and processed by a commercial avidin-biotin-peroxidase secondary antibody system (Vectastain 'Elite' kit, Vector Labs). Sections were incubated in, sequentially, dilute normal goat serum, primary antibody, biotinylated goat anti-rabbit IgG, and finally a preformed avidin-biotinylated HRP macromolecular complex. After the final antibody was applied, slides were incubated in peroxidase substrate solution containing 0.05% diaminobenzidine (Kierkegaard and Perry, Inc.), 0.04% NiCIe, and 0.02% hydrogen peroxide. Sections were counterstained with 1% Methyl green or Neutral red, dehydrated in alcohols, and mounted. Primary antisera, used at concentrations of 1:t,000 to 1:4,000 were as follows: no. 635-5 (courtesy Dr. Lothar Jennes), and a commercial antibody (Incstar Corp,), both directed against antimammalian GnRH; anti-salmon GnRH no. PBL-L49 (courtesy Dr. Wiley Vale); anti-chicken GnRH II no. 741 (courtesy Drl Judy A. King); anti-FMRFamide (Incstar Corp.) Specificity controls involved pre-incubating the primary antibody overnight in 10 or 100 ~M of the antigen to which the primary antiserum was directed. The normal development of the olfactory nerve, including TN was examined histologically in unoperated subjects.

3. Results

Immunolabeling in normal (unoperated) and control animals was similar. In experimental subjects, immunolabeling varied, depending upon whether regeneration of extirpated olfactory placodes occurred. We discuss the findings of normal and control animals first, followed by the experimental cases.

R.G. Northcutt, L.E. Muske /Deuelopmental Brain Research, 78 (1994) 279-290

281

3.1. Normal and control subjects The first fibers of the olfactory nerve, including the TN [9], entered the brain between stages 31 and 33. By four months of age, a column of G n R H - and FMRFamide-ir perikarya and associated fibers could be observed, extending from distal regions of the olfactory nerve through the ventromedial floor of the telencephalon into the septo-preoptic area (Figs. 1, 2 A - D , 3A,B). This distribution closely matches that of cells and fibers of the TN of urodele amphibians, labeled by applying H R P to the olfactory epithelium [18], and descriptions of the distribution of GnRH-ir cell bodies and fibers in adult Ambystoma tigrinum [64]. The most characteristic type of TN neuron is fusiform bipolar, with primary neurites oriented parallel to other olfactory and terminal nerve fibers (Fig. 3A,B), but round and polygonal immunoreactive profiles were also observed. FMRFamide- and GnRH-ir TN cell bodies were observed in similar numbers. A second GnRH-ir population, comprising magnocellular spherical or pyriform cell bodies, is present in periventricular regions of posterior tubercle, in the caudal thalamus. The primary neurite of these cells extends laterally into the neuropil, where it branches extensively (Fig. 3C). There is also a large population (100 or more cell profiles/40 /xm section) of FMRFamide-ir neurons in the hypothalamus, surrounding the third ventricle, ventral to and separated from the GnRH-ir neurons of the posterior tubercle. FMRFamide-ir hypothalamic neurons are distributed over a broad area, from the level of the anterior commissure to the caudal hypothalamus. Cell bodies are spherical, and in many cases have dendrites extending into the cerebrospinal fluid (CSF; Fig. 3D) Periventricular G n R H - and FMRFamide-ir neurons are not contiguous with G n R H - or FMRFamide-ir neurons in the T N or septo-preoptic regions. There appear to be two classes of immunolabeled fibers, clearly distinguishable on the basis of size. The GnRH-ir fibers associated with TN cell bodies are thick, either smooth or beaded. Thick fibers are restricted to distal and central projections of the TN, and to the G n R H tract that courses from the septo-preoptic area to the median eminence, through the ventrolateral hypothalamus, and to the habenula, via the stria medullaris (Figs. 2, 3A). Immunoreactive G n R H was also detected in an extensive network of very fine diameter beaded fibers, present in all major brain areas. Small diameter GnRH-ir fibers are particularly pronounced within the deeper portions of the cell groups that constitute the ventromedial wall of the cerebral hemispheres (Fig. 2B), and the preoptic and hypothalamic areas (Fig. 2D,E,F). Further caudally, a few small diameter GnRH-ir fibers occur in the optic tectum (Fig. 2F), but

ON A

AOB B C D

E F

AD

G

VII

~gav

H

,X--~VIII

±

~SN Fig. 1. Camera lucida drawing of the ventral surface of the left half of the brain of an adult axolotl, illustrating the position of the GnRH-ir cell bodies of the terminal nerve (solid ovals) and those of the posterior tubercle (asterisks). The levels of the transverse sections shown in Fig, 2 are indicated to the left of the brain.

increased densitites occur in the cerebellum and the medulla (Fig. 2G,H). Within the spinal cord, most small diameter GnRH-ir fibers are restricted to the

282

R.G. Northcutt, L.E. Muske / DeueloprnentalBrain Research, 78 (1994) 279-290

lateral and dorsal funiculi with some fibers ramifying within the spinal gray (Fig. 21). In the TN, FMRFamide-ir fibers are relatively large in diameter, either smooth or beaded, like the GnRH-ir fibers (Fig. 3B). In addition, somewhat smaller diameter fibers immunoreactive to FMRFamide are very abundant in all regions of the brain, particularly the hypothalamus. All four G n R H antibodies yielded positive staining, but the pattern of staining differed. Anti-mammalian antibody 635, and anti-salmon PBL-L49 labeled all of the cell body populations described above, and both thick and thin fiber systems. The commercial antimammalian antibody 'INC' detected only TN and septo-preoptic perikarya, and the thick fiber system. In contrast, anti-chicken G n R H II 741 labeled only G n R H neurons of the posterior tubercle, and the network of fine diameter fibers. The four G n R H antibodies were tested on both control subjects and experimental subjects, and the same staining characteristics were ob-

served. Following pre-incubation of G n R H and FMRFamide antibodies in antigen, immunolabeling was either completely eliminated or markedly reduced. 3.2. Experimental subjects

In ten of the twenty experimental animals, the olfactory placodes regenerated bilaterally, and at the time of sacrifice these subjects exhibited complete olfactory systems (epithelia, nerves and bulbs). Immunolabeling of G n R H - and FMRFamide-ir neurons was indistinguishable from controls, even in subjects (seven of ten), that exhibited some type of developmental abnormality, such as hypertrophy or malformation of the olfactory epithelium, or reduction in size of the olfactory nerve and bulb on one or both sides. Five individuals exhibited only a unilateral olfactory epithelium, nerve and bulb (Fig. 4B). The telencephalic hemisphere ipsilateral to the missing olfactory system was reduced in size, and G n R H - and FMRFamide-ir TN cell bodies

,..'7

,tl, !~'

It~ ¢,e

L'~!'~"'"

_--

..

AU

kk

G I

,~

J

SG' " "

D I I I Fig. 2. Camera lucida drawings of transverse sections through the brain of a non-experimental adult axolotl, illustrating the distribution of GnRH-ir cell bodies and fibers of the terminal nerve (ovals and heavier dashes, respectively)and those of the posterior tubercle (stars and lighter dashes, respectively).Bar = 500 ~m.

R.G. Northeutt, L.E. Muske / Developmental Brain Research, 78 (1994) 279-290

283

Fig. 3. GnRH- and FMRFamide-ir neuronal populations in the nervus terminalis (A and B) and periventricular diencephalon (C and B) in A. mexieanum. A: horizontal section through the septo-preoptic region illustrating GnRH-ir cell bodies (arrows) and associated thick fibers (arrowheads) located on or close to the telecencephalic surface, visualized with anti-mammalian GnRH 'INC'. B: horizontal section through the anterior telecephalon, illustrating FMRFamide-ir terminal nerve neurons at the junction of the olfactory nerve (on) and olfactory bulb (oh). Arrows indicate small diameter FMRFamide-ir fibers of the type found throughout the brain that are not of placodal origin. Bar scale equals 110 ~zm in A and B. C: magnocellular GnRH-ir neurons in the posterior tubercle of an experimental subject with bilateral loss of the olfactory system, and complete absence of ir-GnRH perikarya in the anterior forebrain. Arrows indicate fine beaded fibers found throughout the brain of both control and experimental subjects. Transverse section incubated in anti-mammalian GnRH 635. Bar scale equals 100 ~m. D: horizontal section through the ventral hypothalamus, illustrating the abundance of FMRFamide-ir neurons lining the third ventricle. Note club-shaped dendritic endings extending into the CSF (arrow). Bar scale equals 60/xm.

284

R.G. Northcutt, L.E. Muske / Developmental Brain Research, 78 (1994) 279-290

were found only on the side ipsilateral to the intact olfactory system (Figs. 5, 6A). Contralaterally, a few thick fibers were found decussating from the intact T N system via the anterior and habenular commissures. Five of the experimental animals exhibited no olfactory organs, nerves or bulbs (Fig. 4C). In these individuals, there were no G n R H - or F M R F a m i d e - i r cell bodies in the TN or septo-preoptic area, and thick fibers were completely absent (Fig. 6B). However, unilateral or bilateral loss of the olfactory system failed to reduce or in any way affect labeling of G n R H - i r magnocellular neurons of the posterior tubercle (Fig. 3C), or of F M R Famide-ir cells of the ventral hypothalamus, or of the small diameter fiber systems. In fact, G n R H immunolabeling following bilateral loss of the olfactory system was identical to that of normal or control individuals treated with antibody 741, which selectively labels posterior tubercle cell bodies and the small diameter fiber system.

4. Discussion

Our results support the hypothesis t h a t GnRI-t- and F M R F a m i d e - i r neurons of the nervus terminalis and septo-preoptic regions arise from the olfactory placode.

They corroborate the findings of Murakami et al. [34], who showed that ablation of the olfactory placodes in the newt Cynops pyrrhogaster eliminates G n R H - i r in T N and anterior basal forebrain, and support conclusions drawn from descriptive studies of G n R H ontogeny that septo-preoptic, as well as TN G n R H neurons, migrate centripetally from the placode early in development [33,41,50,66]. The most well-characterized G n R H function, regulation of pituitary gonadotropin (LH, FSH) release, is mediated by septo-preoptic G n R H - i r neurons, which project through the ventral hypothalamus and terminate in neurovascutar endings in the median eminence. The function of G n R H - i r neurons of the T N is much less well understood. In elasmobranchs, which lack the G n R H pathway to the median eminence, extra-cerebral ganglionic cells of the T N may serve as the principal gonadotropin regulator [69]. There is indirect evidence that the T N a n d / o r its G n R H component may be involved in reproductive behavior [12,47,55], but to date, there is no known function for the ~ in adults of any species. Earlier speculations that the TN may be involved in mediation of sex pheromone signals has not been corroborated [17]. The hypothesis that the T N may play a role in the development of the septo-preoptic G n R H system was

Fig. 4. Photomicrographs of rostral portions of brains and olfactory epithelia of control (A) and experimental (B and C) subjects, viewed from dorsal aspect. Brain B has an intact olfactory epithelium and nerve on the left side only. Both hemispheres are reduced in size (compare with A), due to absence or incomplete development of the olfactory bulbs, but the right hemisphere is smaller than the left. In animal C, the olfactory epithelia, nerves and bulbs are absent bilaterally. Bar = 1 mm.

285

R.G. Northcutt, L.E. Muske / Deuelopmental Brain Research, 78 (1994) 279-290

first s u g g e s t e d by studies o f G n R H n e u r o a n a t o m y in a d u l t a m p h i b i a n s . M u s k e a n d M o o r e n o t e d t h a t in many vertebrates, including amphibians and mammals, s e p t o - p r e o p t i c a n d T N G n R H n e u r o n s form an

a n a t o m i c a l c o n t i n u u m of m o r p h o l o g i c a l l y h o m o g e n e o u s cell types, s u g g e s t i n g s o m e type of f u n c t i o n a l a s s o c i a t i o n [36,37]. C l u e s t h a t this a s s o c i a t i o n m i g h t b e d e v e l o p m e n t a l in n a t u r e was s u g g e s t e d by the observa-

A

ON

II

/

E [

I

Fig. 5. Camera lucida drawings of horizontal sections from dorsal (A) to ventral (E) through the brains of an experimental axolotl in which the left olfactory placode was extirpated. Note that GnRH-ir cell bodies and fibers (solid ovals and dashes, respectively) are absent from the extirpated side of the animal. Bar = 500 gm.

286

R.G. Northcutt, L.E. Muske / Det'elopmental Brain Research. 78 (1994) 279-290

Fig. 6. G n R H immunoreactivity in the septo-preoptic area of experimental subjects. Horizontal sections through the ventromedial telencephalon treated with anti-mammalian G n R H 'INC', which selectively labels GnRH-ir neurons in the terminal nerve and septo-preoptic regions. A: unilateral loss of the olfactory system. GnRH-ir terminal nerve cell bodies and thick fibers (arrows) are present only on the side ipsilateral to the intact olfactory system (compare with Fig. 4B). B: bilateral loss of the olfactory system. GnRH immunoreactivity is completely absent. Dark, ovoid profiles in both A and B are staining artifacts. Bar = 90/xm, for both A and B.

tion that in amphibians, TN and septo-preoptic GnRH neurons lie on or close to the brain surface. Since the brains of amphibians contain few migrated nuclei, Muske and Moore argued that the most plausible explanation for the superficial position of GnRH neurons was that they reach their adult destinations by migration, not from the brain ventricles, but from the ectodermal olfactory placode, along the route taken by the developing TN and olfactory nerve. The present study, and that of Murakami et al. [34] provide the first direct experimental test of the hypothesis. The unique role that the TN plays in GnRH neuronal ontogeny sheds light on a number of questions and observations. For example, it helps to explain why in some species, including humans, the TN a n d / o r its GnRH component may be detected during embryonic stages, but not in adults [7,44]. Recently, SchwanzelFukuda and Pfaff [51] showed that in individuals with

Kallman's syndrome, a congenital disorder in humans characterized by sterility and anosmia, growth of olfactory axons and migration of GnRH-ir neurons are blocked. A central challenge for neurobiologists remains to identify the functional role of GnRH-ir TN neurons in adults. Our results are consistent with previous observations that the olfactory system exerts a trophic influence on telencephalic development [8,9]. We found that absence of the olfactory and TN systems was always accompanied by a marked reduction in the size of the ipsilateral telencephalic hemisphere. Our resuits, and those of Murakami et al. [34], suggest that this is due to absence of the olfactory bulb. In normal animals, axons of the olfactory nerve, growing centripetally from the placode, contact the anterior end of the developing telencephalic hemisphere and induce differentiation of the olfactory bulb [9]. Interestingly, we observed several experimental cases in which the olfactory epithelium and some component(s) of the olfactory a n d / o r terminal nerves, including GnRHand FMRFamide-ir neurons, made contact with the rostral pole of the telencephalon, but the hemisphere was reduced in size. These results suggest that the peptide-containing neurons of the TN, and the olfactory neurons which exert atrophic influence on telencephalic development, comprise separate populations which are capable of developing independently of one another. The observation that GnRH-ir neurons in the posterior tubercle are unaffected by olfactory placode extirpation represents a new finding. GnRH-ir neurons have been identified in the posterior diencephalon a n d / o r anterior midbrain tegmentum in species representing all major vertebrate groups [35]. While we cannot be certain that all of the posterior GnRH-ir neuronal populations are homologues, they do share certain salient characteristics. They are topographically isolated from the TN/septo-preoptic complex, they contain many unique cell types, and unlike T N / s e p t o preoptic neurons, GnRH-ir neuronal perikarya in the posterior brain are largely midline systems, lying close to the ventricle. Thus, cytoarchitectonic criteria alone would argue for a separate embryonic origin for the posterior GnRH-ir cell groups. Our results support this hypothesis. The posterior GnRH-ir cell groups also differ immunologically from the TN/septo-preoptic system, implying that the two populations, which appear to have separate embryonic origins, also express a different molecular form of GnRH. It is now well established that GnRH occurs as a family of closely related peptides, and that most vertebrates express more than one form (for recent reviews, see [23,52]). Evidence suggests that one form of GnRH predominates in the TN/septo-preoptic system, while a different form pre-

R.G. Northcutt, L.E. Muske / Det:elopmental Brain Research, 78 (1994)279-290

dominates in extrahypothalamic areas, especially in the caudal brain (reviewed in [35]). Data based on ICC [6,37,38,49,64], and radioimmunoassay [5,16,21,22,53] support the view that the principle gonadotropin regulator in amphibians is mammalian G n R H . Our data, showing that an antibody to mammalian G n R H labels TN-septo-preoptic neurons, and the major G n R H projection to the median eminence, but not G n R H neurons in the posterior tubercle, are consistent with this model. Our data also suggest that G n R H - i r neurons in the posterior tubercle contain a substance similar to chicken G n R H II. Antibody 741, which is directed against chicken G n R H II labeled these cells and the associated small diameter fiber system, but not the TN-septo-preoptic neurons or the large diameter fibers. These observations are consistent with analyses of G n R H neuronal distribution in the newt Taricha granulosa [39], in which a n t i - c G n R H II 741 selectively labeled cell bodies in the caudal diencephalon and an extensive distribution of small diameter fibers. A similar distribution of chicken G n R H II-ir neurons has been found in avian species [26,28], in a teleost [1], and in a m a m m a l [11]. These neuroanatomical observations of chicken G n R H II immunoreactivity are consistent with analyses based on radioimmunoassay of grossly dissected brain tissue [11,27,71]. The tetrapeptide F M R F a m i d e was first sequenced in molluscs [46]. However, the endogenous F M R F amide-like molecules of vertebrates are not the same as F M R F a m i d e (for reviews, see [25,48]). Three related, somewhat more extended molecules have been sequenced in vertebrates [15,70]. All contain an ArgPhe-amide (RFamide) carboxy-terminus, which appears to be the basis of their immunological similarities with F M R F a m i d e . F M R F a m i d e is also structurally and immunologically similar to the carboxy-terminal amino acid sequence of NPY, and antisera to these two peptides may cross-react [40]. In mammals, F M R F amide-like substances have been implicated in a variety of functions, including cardiovascular regulation, neurohypophyseal function, aggressive behavior, and they may interact with opiate systems [25,48]. Terminal nerve neurons containing an F M R F amide-like substance have been identified in diverse fishes (teleosts [3,57,59]; elasmobranchs [10,56]; cyclostomes [19,45]), in anuran amphibians [37,60,63]; and in a bird [62]. Co-localization of i r - G n R H and irF M R F a m i d e to the same TN neurons has been reported in teleosts [3,57,59]. Whether F M R F a m i d e - and G n R H - i r substances are co-localized in the TN of A. mexicanum, or whether the two peptides occur in separate cells cannot be determined from our data. The F M R F a m i d e - i r hypothalamic neurons of A. mexicanum are strikingly similar to a large population of CSF-contacting F M R F a m i d e - i r neurons in an elas-

287

mobranch, the cloudy dogfish Scyliorhinus torzame [10]. In mammals, F M R F a m i d e - i r neurons are distributed somewhat differently. In the rat, for example, the highest concentration of F M R F a m i d e - i r perikarya occurs in the hypothalamus, but cells are also widely distributed in extra-hypothalamic regions, including the cerebral cortex, striatum, septum, thalamus and medulla [61]. Immunoreactive F M R F a m i d e has apparently not been detected in the mammalian TN. Unlike peptidergic neurons of the TN, F M R F amide- and G n R H - i r cells in the periventricular diencephalon comprise separate populations in nearly all species studied. An apparent exception to this rule is a report of co-localization of immunoreactive F M R F amide and G n R H in the dorsal mesencephalic tegmenturn of a teleost, Poecilia latipinna [3]. These neurons appear to be homologous to the G n R H - i r magnocellular population of the axolotl posterior tubercle. Many questions remain concerning the ontogeny and phylogeny of G n R H - and FMRFamide-like substances, and their functional role. What is the biological significance of the association of these two peptides in the TN? Is the FMRFamide-like peptide that arises from the olfactory placode the same as that expressed by neurons in the brain, or, as appears to be the case for G n R H neurons, are these different molecules? Do G n R H - and F M R F a m d e - i r neurons in the periventricular diencephalon develop from the neural plate, or perhaps from another source?

5. Acknowledgements We are grateful to Drs. Wiley Vale, Lothar Jennes, and Judy King for generously donating antisera. The research was supported by grants from the National Science Foundation (IBN 92-09643) to L.E.M. and from the National Institutes of Health (NS24869) to R.G.N.

6. Abbreviations AD AL AM AO AOB AU AV CER CC DH DP DT H HC

anterodorsal lateral line nerve amygdala,pars lateralis amygdala,pars medialis accessoryolfactory nerve accessory olfactory bulb cerebellar auricle anteroventral lateral line nerve cerebral hemisphere cerebellar corpus dorsalhypothalamic nucleus dorsalpallium dorsalthalamus habenular nuclei habenular commissure

R.G. Northcutt, L.E. Muske / Developmental Brain Research, 78 (1994) 279-290

288 IG II LL LP MP NA OB OE ON OT P PL PR PT R SE SG SN ST TE TN V VH VI VII VIII VT IX X XII

internal granular and mitral cell layers of olfactory bulb optic nerve lateral line lobe lateral pallium medial pallium nucleus accumbens main olfactory bulb olfactory epithelium olfactory nerve optic tectum pituitary gland root of postotic lateral line nerves preoptic area posterior tuberculum raphe nuclei septal nuclei spinal grey spinal nerve corpus striatum midbrain tegmentum terminal nerve trigeminal nucleus or nerve ventral hypothalamic nucleus abducent nerve facial nerve octaval nerve ventral thalamus glossopharyngeal nerve vagal nerve hypoglossal nerve

7. References [1] Amano, M., Oka., Y, Aida, K., Okumoto, N., Kawashima, S. and Hasegawa, Y., Immunocytochemical demonstration of salmon GnRH and chicken GnRH-II in the brain of masu salmon, Oncorhynchus masou, J. Comp. Neurol. 314 (1991) 587597. [2] Barry, J., Hoffman, G.E. and Wray, S., LHRH-containing systems. In A. Bjorklund and T. Hokfelt (Eds.), Handbook of Chemical Neuroanatomy, Vol. 4, Elsevier, Amsterdam, 1985, pp. 166-215. [3] Batten, T.F.C., Cambre, M.L., Moons, L. and Vandesande, F., Comparative distribution of neuropeptide immunoreactive systems in the brain of the green molly, Poecilia latipinna, J. Comp. Neurol., 302 (1990) 893-919. [4] Bennis, M., Dubourg, P, Gamrani, H., Calas A. and Kah, O., Existence of a GnRH immunoreactive nucleus in the dorsal midbrain tegmentum of the chameleon, Gen. Comp. Endocrinol., 75 (1989) 195-203. [5] Branton, W.D., Jan, L.Y. and Han, Y.L., Non-mammalian luteinizing hormone-releasing factor (LRF) in tadpole and frog brain, Soc. Neurosci. Abstr., 8 (1982) 14. [6] Branton, W.D., Phillips, H.S. and Jan, Y.N., The LHRH family of peptide messengers in the frog nervous system. In T. Hokfelt, K. Fuxe and B. Pernow (Eds.), Progress in Brain Research, Vol. 68, Elsevier, Amsterdam, 1986, pp. 205-215. [7] Brown, J.W., The nervus terminalis in insectivorous bat embryos and notes on its presence during human ontogeny. In L.S. Demski and M. SchwanzeI-Fukuda (Eds.), The Terminal Nerve (Nervus Terminalis): Structure, Function and Evolution, Ann. N Y Acad. Sci., Vol. 519, 1987, pp. 184-200. [8] Brunjes, P.C. and Frazier, L.L., Maturation and plasticity in the olfactory system of vertebrates, Brain Res. Rev., 11 (1986) 1-45. [9] Clairambault, P., Development of the prosencephalon. In R.

Llinas and W. Precht (Eds.), Frog Neurobiology, Springer-Verlag, Berlin, 1976, pp. 924-945. [10] Chiba, A., Oka, S. and Honma, Y., Immunocytochemical distribution of FMRFamide-like substance in the brain of the cloudy dogfish, Scyliorhinus torzame, Cell Tiss. Res., 265 (1991)243-250. [11] Dellovade, T.L., King, J.A., Millar, R.P. and Rissman, E.F., Presence and differential distribution of distinct forms of immunoreactive gonadotropin releasing hormone in the musk shrew brain, Neuroendocrinology, (in press). [12] Demski, L.S. and Northcutt, R.G., The terminal nerve: a new chemosensory system in vertebrates? Science, 220 (1983) 435437. [13] Demski, L.S. and Wright, D.E., GnRH immunoreactivity in the brain of the brown anole, Anolis sagrei, Soc. NeuroscL Abstr., 18 (1992) 328. [14] Dockray, G.J., Vaillant, C., Williams, R.G., Gayton, R.J. and Osborne, N.N., Vertebrate brain-gut peptides related to FMRFamide and Met-enkphalin-Arg6phe7, Peptides, 2 (1981) 25-30. [15] Dockray, G.J., Vaillant, C., Williams, R.G., Gayton, R.J. and Barnard C.S., A novel active pentapeptide from chicken brain identified by antibodies to FMRFamide, Nature, 305 (1983) 328-330 [16] Eiden, L.E., Loumaye, E., Sherwood, N. and Eskay, R.L., Two chemically and immunologically distinct forms of luteinizing hormone-releasing hormone are differentially expressed in frog neural tissues, PeptMes, 3 (1982) 323-327. [17] Fujita, I., Sorensen, P.W., Stacey, N.E. and Hara, T.J., Chemosensory responses to sex pheromones are carried by the olfactory system not the terminal nerve. Brain Behav. EvoL, 38 (1991) 313-321. [18] Hofmann, M.H. and Meyer, D.L., Central projections of the nervus terminalis in four species of amphibians, Brain Behav. EvoL, 34 (1989) 301-307. [19] Jirikowski, G., Erhart, G., Grimmelikhuijzen, C.J.P, Triepel, J. and Patzner, R.A., FMRFamide-like immunoreactivity in brain and pituitary of the hagfish Eptatreus burgeri (Cyclostomata), Cell Tiss. Res., 237 (1984) 363-366. [20] Kah, O., Brenton, B., Dulka, J.G., Nunez-Rodriguez, J., Peter, R.E., Corrigan, A., Rivier, J.E. and Vale, W.W., A reinvestigation of the Gn-RH (gonadotropin-releasing hormone) systems in the goldfish brain using antibodies to salmon Gn-RH, Cell Tiss. Res., 244 (1986) 327-337. [21] King, J.A. and Millar, R.P., Hypothalamic luteinizing hormonereleasing hormone content in relation to the seasonal reproductive cycle of Xenopus laevis, Gen. Comp. EndocrinoL, 39 (1979) 309-312. [22] King, J.A. and Millar, R.P., Identification of HisS,Trp7,Tyr 8GnRH (chicken GnRH II) in amphibian brain, Peptides, 7 (1986) 827-834. [23] King, J.A. and Millar, R.P., Gonadotropin-releasing hormones. In Vertebrate Endocrinology: Fundamentals and Biomedical Implications, Vol. 4, Part B, Academic Press, New York, 1991, Ch. 8, pp. 1-31. [24] Kuenzel, W.J. and Blahser, S., The distribution of go_ nadotropin-releasing hormone (GnRH) neurons and fibers throughout the chick brain (Gallus domesticus), Cell Tiss. Res., 264 (1991) 481-495. [25] Kyle, A.L., Bao-guo, L., Magnus, T.H. and Ste!l, W.K., Substance P-, F8Famide-and A18Famide-like immunoreactivity in the nervus terminalis and retina of the goldfish, Carassius auratus, Cell Tiss. Res., (in press). [26] Mikami, S., Yamada, S., Hasegawa, Y. and Miyamoto, K., Localization of avian LHRH-immunoreaetive neurons in the hypothalamus of the domestic fowl, Gallus domesticus and the Japanese quail, Coturnix coturnix, Cell Tiss. Res., 251 (1988) 51-58. [27] Millam, J.R., Craig-Veit, C.B., Adams, T.E. and Adams, B.M.,

R.G. Northcutt, L.E. Muske / Developmental Brain Research, 78 (1994) 279-290 Avian gonadotropin-releasing hormones I and II in brain and other tissues in turkey hens, Comp. Biochem. PhysioL, 94A (1989) 771-776. [28] Millam, J.R., El Halawani, M.E. and Phillips, R.E., Immunohistochemical localization of chicken gonadotropin-releasing hormones I and II (cGnRH I and II) in turkey hen brain, Soc. Neurosci. Abstr., 17 (1991) 653. [29] Miller, K.E. and Kriebel, R.M., Peptidergic innervation of caudal neurosecretory neurons, Gen. Comp. Endocrinol., 64 (1986) 396-400. [30] Moore, F.L., Miller, L.J., Spielvogel, S.P., Kubiak, T. and Folkers K., Luteinizing hormone-releasing hormone involvement in the reproductive behavior of a male amphibian, Neuroendocrinology, 35 (1982) 212-216. [31] Moss, R.L. and McCann, S.M., Induction of mating behavior in rats by luteinizing hormone-releasing factor, Science, 181 (19731 177-179. [32] Munz, H., Stumpf, W.E. and Jennes, L., LHRH systems in the brain of platyfish, Brain Res., 221 (1981) 1-13. [33] Murakami, S., Seki, T., Wakabayashi, K. and Arai, Y., The ontogeny of luteinizing hormone-releasing hormone (LHRH) producing neurons in the chick embryo: possible evidence for migrating LHRH neuron from the olfactory epithelium expressing a highly polysialylated neural cell adhesion molecule, Neurosci. Res., 12 (1991) 421-431. [34] Murakami, S., Kikuyama, S. and Arai, Y., The origin of the luteinizing hormone-releasing hormone (LHRH) neurons in newts (Cynops pyrrhogaster): the effect of olfactory placode ablation, Cell Tiss. Res., 269 (1992) 21-27. [35] Muske, L.E., Evolution of gonadotropin-releasing hormone (GnRH) neuronal systems, Brain Behav. Evol., 42 (1993) 215230. [36] Muske, L.E. and Moore, F.L., Luteinizing hormone-releasing hormone-immunoreactive neurons in the amphibian brain are distributed along the course of the nervus terminalis. In L.S. Demski and M. Schwanzel-Fukuda (Eds.), The Terminal Nerve (Nervus Terminalis): Structure, Function and Evolution, Ann. NY Acad. Sci., 519 (1987) 433-446. [37] Muske, L.E. and Moore, F.L., The amphibian nervus terminalis: anatomy, chemistry and relationship with the hypothalamic LHRH system, Brain Behav. Evol., 32 (1988)141-150. [38] Muske, L.E. and Moore, F.L., Ontogeny of immunoreactive gonadotropin-releasing hormone neuronal systems in amphibians, Brain Res., 534 (1990) 177-187. [39] Muske, L.E. and Moore, F.L., Antibodies against different forms of GnRH distinguish different populations of cells and axonal pathways in an amphibian, Taricha granulosa, J. Comp. Neurol., (in press). [40] Muske, L.E., Dockray, G.J., Chohan, K.S. and Stell, W.K., Segregation of FMRF amide-immunoreactive fibers and NPYimmunoreactive amacrine cells in goldfish retina, Cell Tiss. Res., 247 (1987) 299-307. [41] Norgren, R.B. and Lehman, M.N., Neurons that migrate from the olfactory epithelium in the chick express luteinizing hormone-releasing hormone, Endocrinology, 128 (1991)1676-1678. [42] Northcutt, R.G. and Muske, L.E., Experimental embryological evidence of the placodal origin of GnRH and FMRF-amide neurons of the terminal nerve and preoptic area in salamanders, Soc. Neurosci. Abstr., 17 (1991) 321. [43] Oelschlager, H.A., Persistence of the nervus terminalis in adult bats: a morphological and phylogenetical approach, Brain Behay. Evol., 32 (1988) 330-339. [44] Oelschlager, H.A., Buhl, E.H. and Dann, J.F., Development of the nervus terminalis in mammals including toothed whales and humans. In L.S. Demski and M. Sehwanzel-Fukuda (Eds.), The Terminal Nerve (Nervus Terminalis): Structure, Function and Evolution, Ann. NY Acad. Sci., Vol. 519., 1987, pp. 447-464.

289

[45] Ohtomi, M., Fujii, L. and Kobayashi, H., Distribution of FMRFamide-like immunoreactivity in the brain and neurohypophysis of the lamprey, Lampetra japonica, Cell Tiss. Res., 256 (1989) 581-584. [46] Price, D.A. and Greenberg, M.J., The structure of a molluscan cardioexcitatory neuropeptide, Science, 197 (1977) 670-671. [47] Propper, C.R. and Moore, F.L., Effects of courtship on brain gonadotropin hormone-releasing hormone and plasma steroid concentrations in a female amphibian (Taricha granulosa) Gen. Comp, EndocrinoL, 81 (1991) 304-312. [48] Raffa, R.B., The action of FMRFamide (Phe-Met-Arg-PheNH 2) and related peptides in mammals, Peptides, 9 (1988) 915-922. [49] Rastogi, R.K., De Meglio, M. and lela, L., lmmunoreactive luteinizing hormone-releasing hormone in the frog (Rana esculenta) brain: distribution pattern in the adult, seasonal changes, castration effects and developmental aspects, Gen. Comp. Endocrinol., 78 (1990) 444-458. [50] Schwanzel-Fukuda, M. and Pfaff, D.W., Origin of luteinizing hormone-releasing hormone neurons, Nature, 338 (1989) 161164. [51] Schwanzel-Fukuda, J. and Pfaff, D.W., The migration of luteinizing hormone-releasing hormone (LHRH) neurons from the medial olfactory placode into the medial basal forebrain, Experientia, 46 (1990) 956-961. [52] Sherwood, N.M. and Lovejoy, D.A., Gonadotropin-releasing hormone in cartilagenous fish: structure, location and transport, Env. Biol. Fishes, (1993) in press. [53] Sherwood, N.M., Zoeller, R.T. and Moore, F.L., Multiple forms of gonadotropin-releasing hormone in amphibian brains, Gen. Comp. Endocrinol., 61 (1986) 313-322. [54] Silverman, A.-J., The gonadotropin-releasing hormone (GnRH) neuronal systems: immunocytochemistry. In E. Knobil and J. Neill (Eds.), The Physiology of Reproduction, Raven, New York, 1988, pp. 1283-1303. [55] Stacey, N.E. and Kyle, A.L., Effects of olfactory tract lesions on sexual and feeding behavior in the goldfish, PhysioL Behav., 30 (1983) 621-628. [56] Stell, W.K., Luteinizing hormone-releasing hormone (LHRH) and pancreatic polypeptide (PP)-immunoreactive neurons in the terminal nerve of spiny dogfish, Squalus acanthias, Anat. Rec., 208 (1984) 173A-174A. [57] Stell, W.K., Walker, S.E., Chohan, K.S. and Ball, A.K., The goldfish nervus terminalis: a luteinizing hormone-releasing hormone and molluscan cardioexcitatory peptide immunoreactive olfactoretinal pathway, Proc. Natl. Acad. Sci. USA, 81 (1984) 940-944. [58] Subhedar, N. and Rama Krishna, N.S., lmmunocytochemical localization of LH-RH in the brain and pituitary of the catfish, Clarias batrachus (Linn.), Gen. Comp. Endocrinol., 72 (1988) 431-442. [59] Uchiyama, H., Immunohistochemical subpopulations of retinopetal neurons in the nucleus olfactoretinalis in a teleost, the whitespotted greenling (Hexagrammos stelleri), J. Comp. NeuroL, 293 (1990) 54-62. [60] Uchiyama, H., Reh, T.A. and Stell, W.K., lmmunocytochemical and morphological evidence for a retinopetal projection in anuran amphibians, J. Comp. Neurol., 274 (1988) 48-59. [61] Williams, R.G. and Dockray, G.J., Immunohistochemical studies of FMRF-amide-like immunoreactivity in rat brain, Brain Res., 276 (1983) 213-229. [62] Wirsig-Wiechmann, C.R., The nervus terminalis in the chick: a FMRFamide-immunoreactive and acetylcholinesterase-positive nerve, Brain Res., 523 (1990) 175-179. [63] Wirsig-Wiechmann, C.R. and Basinger, S.F., FMRFamide-immunoreactive retinopetal fibers in the frog, Rana pipiens:

290

[64]

[65] [66]

[67]

R.G. Northcutt, L.E. Muske / Developmental Brain Research. 78 (1994) 279-290 demonstration by lesion and immunocytochemical techniques, Brain Res., 449 (1988) 116-134. Wirsig, C.R. and Getchell, T.V., Amphibian terminal nerve: distribution revealed by LHRH and AChE markers, Brain Res., 385 (1986) 10-21. Witkin, J.W., Immunocytochemistry of the GnRH system in developing Xenopus, Soc. Neurosci. Abstr., 16 (1990) 363. Wray, S., Nieburgs, A. and Elkabes, S., Spatiotemporal cell expression of luteinizing hormone-releasing hormone in the prenatal mouse: evidence for an embryonic origin in the olfactory placode, Dev. Brain Res., 46 (1989) 309-318. Wright, D.E. and Demski, L.S., Gonadotropin hormone-releasing hormone (GnRH)-immunoreactivity in the brains of primitive bony fishes, Soc. Neurosci. Abstr., 16 (1990) 127.

[68] Wright, D.E. and Demski, LS., Gonadotropin hormone-releasing hormone (GnRH)-immunoreactivity in the mesencephalon of sharks and rays, J. Comp. Neurol., 307 (1991) 49-56. [69] Wright, D.E. and Demski, L.S., Gonadotropin-releasing hormone (GnRH) pathways and reproductive control in elasmobranchs. Env. BioL Fishes, in press. [70] Yang, H.-Y.T., Fratta, W., Majane, E.A. and Costa, E., Isolation, sequencing, synthesis, and pharmacological characterization of two brain neuropeptides that modulate the action of morphine, Proc. Natl. Acad. Sci. USA, 82 (1985) 7757-7761. [71] Yu, K.L., Sherwood, N.M. and Peter, R.E., Differential distribution of two molecular forms of gonadotropin-releasing hormone in discrete brain areas of goldfish (Carassius auratus), Peptides, 9 (1988) 625-630.