Ultrastructure of luteinizing hormone-releasing hormone (LHRH) neurons and their projections in the golden hamster

Brain Research

Bulkin,

0361-9230/88 $3.00 + .OO

Vol. 20, pp. 211-221. DPergamon Press plc, 1988.Printed in the U.S.A.

Ultrastructure of Luteinizing Hormone-Releasing Hormone (LHRH) Neurons and Their Projections in the Golden Hamster MICHAEL

N. LEHMAN*

AND ANN-JUDITH

~ILVERMAN~

“Department of Anatomy & Cell Biology, University of Cincinnati College of Medicine 231 Bethesda Avenue, Cincinnati, OH 45247 TDepartment of Anatomy & Cell Biology, Columbia University College of Physicians & Surgeons 630 West f68th St., NeFt>York, NY 10032 Received

14 October

1987

LEHMAN, M. N. AND A.-J. SILVERMAN. ~lt~~str~l~tur~ uf lt~r~~~iziffghltr~ff~~-rel~~asi~g h~t~~~ttz~~ (LHRH) tmrms golden ~ff~srer. BRAIN RES BULL 20(2)21 l-221, f988.-Seasonaf breeding mammals exhibit a reversible annual cycle of fertility, and thus represent valuable models for investigating the organization of luteinizing hormone-releasing hormone (LHRH) neurons which mediate the neuroendocrine control of reproduction. Electron microscopic immunocytochemistry was used to examine the ultrastructure of LHRH neurons and their projections in a seasonal breeder, the golden hamster, housed under photoperiodic conditions which are reproductively stimulatory. LHRH perikarya in the diagonal band, medial septum, and preoptic area were bipolar or unipolar cells which possessed nuclei with prominent indentations and multiple nucleoh. LHRH reaction product within these cells was associated with neurosecretory granules and the rough endoplasmic reticulum (RER), particularly those portions of RER adjacent to the outer nuclear envelope. LHRH cell bodies and dendrites in the hamster received an extremely limited amount of synaptic input; of a total of twenty-five cells analyzed, we found only five instances of morphologically identified synapses onto immunoreactive dendrites. Occasionally close somatic appositions were seen between LHRH cells and non-immunoreactive neurons, and in one instance between two LHRH somas. In the organum vasculosum of the lamina terminalis (OVLT), LHRH fibers appeared as isolated strings of varicosities. In contrast, within the median eminence immunoreactive axonal profttes were frequently bundled together in fascicles although synaptic specializations were not observed between individual axons. The existence of contacts between LHRH axons in the median eminence, coupled with the extreme paucity of synaptic inputs onto LHRH neurons in this species, suggests that the median eminence may be a site where neural or hormonal signals could influence the coordinated release of LHRH from cells of dispersed origin. und fhtlirpr~~~~,ri~t~sin the

LHRH Hamster Jmmunocytochemistry Organum vasculosum of the lamina terminalis

Etectron microscopy Median eminence

Medial septum

Preoptic area

role not only in seasonal breeding, but also in neuroendocrine mechanisms underlying puberty and the estrous cycle [12,18]. While recent work has provided a detailed understanding of the manner in which photoperiod regulates the activity of the LHRH pulse generator in seasonal breeders [11,29], much less is known about the neural substrate upon which these signals act. Light microscopic immunocytochemic~ studies in the hamster 18,221 and sheep [15] have revealed an organization of LHRH cells and pathways similar to that seen in non-photoperiodic species [24,34]. In both hamsters and sheep, a majority of LHRH cells form a loose continuum extending from the diagonal band and medial septum, rostraily, to the preoptic area and anterior hypothalamus, caudally. In addition, clusters of LHRH neurons can be found within the intracerebral and extracerebraf portions of the terminal ganglion in the hamster [8,32].

breeding mammals, such as hamsters and sheep, have a reversible annual cycle of fertility which ensures that offspring are born at a time of year advantageous to their survival [I 11. As such, seasonal breeders represent valuable models for understanding the neuroendocrine basis of reproduction, and exploring the potential for plasticity in the neural mechanisms governing reproductive physiology and behavior [21]. In both hamsters and sheep, the major environmental cue controlling the expression of seasonal breeding is daylength [11,29]. The central mechanism upon which photoperiodic signals conveying daylength act, consists, in part, of neurons which synthesize and secrete the decapeptide, luteinizing hormone-rele~ing hormone (LHRH). LHRH cells and their projections to the median eminence comprise a neural pulse generator which controls the episodic release of pituitary gonadotropins [I, 2, 171. The so-called “LHRH pulse generator” pfays a pivotal SEASONALLY

211

LEHMAN

AND SILVERMAN

FIG. I. (A) Immunoreactive LHRH cells and fibers in the medial septum. Bar=50 Km; (B) Low power electron micrograph of an LHRH perikarya in the medial septum. Note prominent indentations in the outer nuclear envelope with associated reaction product. Curved arrow indicates a labelled neurosecretory vesicle. nc=nucleus.

In the present study, electron microscopic (EM) immunocytochemistry was used to characterize the ultrastructure and possible synaptic inputs of LHRH cells in the adult male golden hamster. We examined immunoreactive perikarya in the diagonal band, the medial septum and the medial preoptic area. We also examined the ultrastructure of LHRH fibers and terminals in the median eminence, as well as in the organum vasculosum of the lamina terminalis (OVLT), another neurohaemal target of LHRH neurons in this species [8] and others [15,34]. Because of the potential importance of contacts between LHRH neurons in coordinating their pulsatile activity, we looked carefully for instances of close appositions between individual LHRH neurons or fibers. When male hamsters are exposed to less than 12.5 hours of light per day their testes regress [3,4]. Brains analyzed in this study were taken from intact male hamsters housed under reproductively stimulatory photoperiods, and therefore protide baseline data for possible future comparisons between hamsters housed under either stimulatory or inhibitory photoperiods. METHOD

Adult male golden hamsters (n=12) were obtained from Engle Laboratories (Farmersburg, IN) and group housed for at least 10 weeks under a 14 L: 10 D (1ight:dark) photoperiod. All hamsters used in this study had large testes (3.11 g, mean combined testes weight) consistent with exposure to a reproductively stimulatory photoperiod [3,4]. Hamsters were heavily anesthesized with sodium pentobarbitol (Nembutal, 75 mgikg) and perfused intracardially with 500 ml of 4% paraformaldehyde, 0.2% glutaraldehyde in 0.1 M phosphate buffer, pH 7.3, with 0.1% sodium nitrite added to this fiiative as a vasodilator. Following perfusion, the brain and attached pituitary were carefully removed from the cranium and post-fixed for an additional 2-3 hours.

Brains were blocked and cut coronally on a vibratome at 50 Wm. Sections were collected and washed several times in 0.1 M phosphate buffer, pH 7.3, and then treated with 0.5% hydrogen peroxide to reduce endogenous peroxidase activity. The primary antiserum against LHRH used in this study (LR-1, kindly provided by Dr. Robert Benoit) recognizes amino acids 3, 4, 7, 8, 9, 10 of the decapeptide and no other identified neuropeptide. Sections were incubated for 48-72 hours at 4°C in primary antiserum diluted l:lO,OOO in phosphate buffer with 0.02% saponin and 1% normal goat serum. Preabsorption of the primary antiserum with l-10 nanomolar concentrations of LHRH (Peninsula Laboratories) for 24 hr at 4°C completely eliminated all staining in the hamster brain. After incubation with primary antiserum, sections were reacted, sequentially, with a biotinylated goat anti-rabbit IgG and an avidin-biotin-HRP conjugate (Vectastain, Vector Laboratories). Each solution was applied for 1 hour at room temperature at the dilutions recommended by the supplier. The HRP was demonstrated using 3,3’ diaminobenzodine (DAB) and the glucose oxidase method for generating the hydrogen peroxide substrate [7]. In alternate sections, the reaction product was intensified using a silver-gold substitution procedure [193. Regions containing immunoreactive cells in the diagonal band, medial septum, and preoptic area, and immunoreactive fibers in the median eminence and OVLT were then dissected out. These tissue pieces were postfixed in 2% osmium tetroxide containing 1.5% potassium ferricyanide [ 131, dehydrated and flat embedded in Epon 812. Semi-thin (1 pm) sections were cut and examined for the presence of reaction product. Ultra-thin sections (70 nm) were cut from the remaining block, mounted on formvar coated grids, and stained with ethanolic uranyl acetate and lead citrate [31]. Serial ultra-thin sections were used to find the active zones of presumptive axon terminals.

UL~ASTRUCTUR~

OF LHRH NEURONS

IN THE HAMSTER

213

FIG. 2. Electron micrograph showing the localization of silver-gold intensified reaction product (arrowheads) within the lumen of the rough endoplasmic reticufum (RER) of an LHRH perikarya. 49,000~.

Sections were examined with a Jeol 100s electron microscope, and observations were made from electron micragraphs at final magnifications of 10,000 to 40,000~. Measurements of vesicle and granule diameter, and the iengths of plasma membrane bearing synaptic densities, were made using a Jandel Scientific digitizer pad and Sigma&an software fun on an IBM PC computer. Uitra-thin sections of twentyfive LHRH cells (approximately two from each brain) were quantitatively analyzed for percentage of plasma membrane occupied by synaptic modifications; serial sections were not included in this analysis. RESULTS

At a light microscopic level, immu~~~~active LHRH cells in the diagonaI band, medial septum, and medial preoptic area were fusiform with one or two major dendritic processes, and gave rise to axons which had the appearance of beaded varicosities (Fig. IA). In ultra-thin (70 nm) sections, LHRH perikarya of similar shape were identified by the presence of granular DAB reaction product which was unevenly distributed throughout the cytoplasm (Fig. 1B). In sectians which were silver-gold intensified, metallic particles were plated upon this granular reaction product (Figs. 2, 3) and aided considerably its subcellular localization. Reaction product in LHRH perikarya was most frequently associated

with the lumen of the rough endoplasmic reticulum (RER) (Fig. 2), particularly that portion of the RER adjacent to the nucleus (Figs. IB, 4). Reaction product was also associated with neurosecretory granules (100-130 nm diameter) but usually was absent from the prominent Golgi apparatus seen in these cells. The sole instance we found of light reaction product associated with a portion of a Golgi apparatus is shown in Fig. 3. Interestingly, in the same section an immunoreactive neurosecretory granule can be seen emanating from the edge of the labelled Golgi saccule. Of particular interest, the nuclei of LHRH perikarya in all brain regions possessed several prominent indentations and contained multiple nucleoli (Figs. 18, 4, 6). Dense reaction product was often seen in aggregates along the cytoplasmic face of the outer nuclear envelope and within its indentations. Contacts between LHRN cell bodies or dendrites were rare, although immunoreactive perikarya were occasionally seen in close contact with non-identi~ed neurons (see Fig_ 6). In one instance (Fig. 4), we observed a close somatic contact between two LHRH perikarya in the medial septum that was evident both at a light microscopic (see inset Fig. 4) and ultrastructural level. Nigher power examination of this contact revealed a direct apposition between the plasma membranes of these two cells (Fig. 5). Although synaptic specializations were not seen in this or adjacent serial sections, at one portion of this contact the plasma membranes of

214

LEHMAN

AND SILVERMAN

FIG. 3. Electron micrograph showing silver-gold intensified reaction product associated with a Go&i saccule (g) (curved arrow) and with two neurosecretory vesicles (short arrows), one of which appears to be emanating from the edge of the labelled Golgi saccule. 36,000~

both cells appeared to disappear and clear spherical vesicles were seen clustered nearby (solid arrows, Fig. 5) possibly due to the tangential plane of this section. Synaptic inputs were rarely observed onto either the cell bodies of dendrites of immunoreactive LHRH cells, regardless of their anatomical location. Of a total of 1325 pm plasma membrane examined from 25 LHRH cells, we found only five instances of morphologically identified synaptic inputs. The mean percentage of LHRH plasma membrane bearing synaptic mod~cations was therefore extremely low (0.11%). One example of synaptic input onto the dendrite of an immunoreactive cell is shown in Fig. 6 and 7. In this LHRH cell and others (Fig. 3), reaction product was restricted to that area adjacent to the outer nuclear envelope. Consequently, identification of the dendrite of this cell depended upon following the outline of this cell’s plasma membrane. At higher magn~cations, an axon terminal containing both clear, spherical vesicles (30-50 nm diameter) and larger dense-core vesicles (80-100 nm diameter) can be seen forming synaptic contact upon a portion of this dendrite (boxed area, Fig. 6; Fig 7A). Serial sections revealed a complex formation of active zones at multiple synaptic clefts (Fig. 7A, B). In association with one of these active zones, a small number of clear vesicles within the postsynaptic dendrite were clustered perpendicular to the synaptic cleft. The plasma membranes of LHRH cell bodies and dendrites were most frequently separated from the adjacent neuropil by thin glial lamellae (100-300 nm in thickness)

emanating from neighboring astrocytes. Similar glial processes almost entirely ensheathed LHRH axons, as they coursed through the medial septum and preoptic area (also see below). LHRH Fibers in the OVLT and Median Eminence

At a light microscopic level, numerous immunoreactive LHRH fibers converged upon the organum vasculosum of the lamina terminalis (OVLT) (Fig. 8A) which in this species is located at the rostral edge of the preoptic recess of the third ventricle. When viewed at an ultrastntctural level, immunoreactive fibers in the OVLT were seen as isolated strings of varicosities located between ependymal cells at the border of the ventricle (Fig. 8B). These varicosities contained immunoreactive neurosecretory vesicles (10&130 nm diameter) and non-immunoreactive mit~hond~a. LHRH varicosities in the OVLT were rarely seen in contact with vascular elements, and were often surrounded by thin glial lamellae which separated them from adjacent ependymal cells. In contrast to the isolated appearance of immunoreactive fibers in the OVLT, LHRH varicosities in the median eminence were often clustered together in direct plasma membrane contact (Fig. PA, B). Synaptic or other membrane specializations were not seen between individual axonal profiles. Further, we could not determine, from analysis of serial sections, whether these clustered LHRH varicosities

ULTRASTRUCTURE

OF LHRH NEURONS

IN THE HAMSTER

FIG. 4. Low power electron micrograph of two septal LHRH cells in direct somatic contact. Note the multiple nucleoi (asterisks) and prominent identations in the nuclear membrane. nc=nucleus. Boxed area is shown in higher magnification in Fig. 5. Inset (lower right) shows thick (1 pm) section of the same two cells in close contact (curved arrow) taken before thin sectioning. Bar= 10 pm.

had arisen from a single axon or several different axons. LHRH varicosities in the median eminence contained immunoreactive neurosecretory vesicles, mitochondria, and often clusters of non-immunoreactive clear spherical vesicles (30-50 nm diameter). As in the OVLT and elsewhere, LHRH profiles in the median eminence were frequently separated from the adjacent neuropil by thin glial lamellae. DISCUSSION

LHRH cell bodies and dendrites in the hamster receive an extremely limited amount of synaptic input, particularly when compared to LHRH cells of other species (see below). Of the 25 LHRH cells analyzed in this study, we found only five instances of morphologically identified synapses onto immunoreactive dendrites. The few LHRH cells for which we did observe synaptic input were not anatomically segregated in any particular brain region but evenly distributed throughout the diagonal band, medial septum, and preoptic

area. The mean percentage of plasma membrane of LHRH cells (both soma and dendrites) bearing a synaptic modiflcations was O.ll%, a percentage considerably less than that previously reported for LHRH somas and dendrites in the submitted manuscript), or for sheep (0.62%; Lehman, LHRH dendrites alone in the rat (0.38%; [35]) or mouse (3.12%; Silverman, unpublished observations). It should be noted that in both the sheep and rat, LHRH neurons are significantly less densely innervated than adjacent nonidentified preoptic neurons (sheep soma + dendrite: 1.98% of plasma membrane occupied by synaptic specializations, Lehman, submitted manuscript; rat dendrites: 6.61% [35]); we have not analyzed non-identified preoptic cells in the hamster. The results indicate considerable variations among mammals in the density of synaptic input which LHRH cells receive, and suggest potential differences between species in the physiological relevance of those inputs. Since LHRH cells in the hamster project to a number of other targets besides the median eminence [8,22], the ques-

216

LEHMAN

FIG. 5. arrows) replaced adjacent

AND SILVERMAN

Higher magnification of boxed area in Fig. 4 showing direct contact between the plasma membranes (open of two adjacent LHRH perikarya. At one point, adjacent plasma membranes become indistinct and appear to be by a cluster of small clear vesicles (solid arrows). A neurosecretory vesicle (curved arrow) can be seen to this contact.

tion arises as to whether the cells observed in the present report are representative of those which project to the median eminence, and therefore constitute the LHRH pulse generator. We recently determined which LHRH cells project to the median eminence in the hamster using a combination of retrograde tract tracing and immunocytochemical techniques [14]. The results parallel recent findings in the rat [25] and mouse [9] and suggest that LHRH cells which project to the median eminence are not localized to any specific brain region, but instead are intermingled with LHRH cells that do not project to the median eminence and presumably innervate different targets. Consequently, the sample of cells we analyzed in this study probably included a subpopulation of LHRH neurons which projected to the median eminence. We cannot exclude the possibility that LHRH cells which project to the median eminence differ quantitatively in their innervation from those which do not, although the overall paucity of synapses seen in this study would argue against this. Future studies using a combination of retrograde tract tracing techniques and immunocytochemistry at an EM level could resolve this issue. The use of a silver-gold intensification procedure [19] in our immunocytochemical protocol allowed us to easily visualize the subcellular localization of reaction product within

LHRH perikarya. Silver-gold intensified reaction product was most frequently associated with the lumen of rough endoplasmic reticulum (RER), suggesting that the antisera used in this study (LR-I of Dr. Robert Benoit) recognizes the LHRH peptide precursor as well as the mature decapeptide. However, consistent with reports in the rat [35], mouse [27] and rhesus monkey [33] using this antiserum, we found only one instance of labelling in the Golgi apparatus (Fig. 3) even though both RER and neurosecretory granules were labelled in the same section. It may be that either the transit of the LHRH precursor through the Golgi appartus occurs too rapidly for immunocytochemical detection, or that during its movement through the Golgi apparatus the precursor is modified and assumes a conformation in which antigenic sites are unavailable. LHRH cells in the hamster exhibit several morphological features not reported in studies of LHRH neurons in other mammals. LHRH neurons in the hamster, for example, differ from those in the sheep (Lehman, submitted manuscript) or rat [ 10,341 in possessing nuclei with prominent indentations and multiple nucleoli. LHRH reaction product was often densest in RER directly adjacent to the nucleus and within the nuclear indentations (see Fig. 4), perhaps reflecting a particularly active state of protein synthesis in these cells.

ULTRASTRUCTURE

OF LHRH

NEURONS

IN THE HAMSTER

217

FIG. 6. Low power electron micrograph of an LHRH perikarya in direct somatic contact (arrowheads) with a non-immunoreactive neuron. A portion of this immunoreactive cell (asterisks) forms a dendrite onto which a synaptic input was found (boxed area). nc=nucleus.

Immunoreactive LHRH neurons in the hamster, like those cells in the sheep (Lehman, submitted manuscript) or rhesus monkey [33], are distinct from neighboring nonimmunoreactive cells in the extent to which they are ensheathed by glial processes. Thin glial lamellae almost entirely surrounded the plasma membranes of LHRH perikarya and axons, regardless of their anatomical location. Glial elements play an important role in regulating the physFOLLOWING

iological activity of magnocellular neurosecretory cells and their terminals in the rat [6, 28,301. It is conceivable that glial processes which surround LHRH neurons and their axons may influence their activity either by modifying the density of synaptic input they receive [20], or perhaps by regulating the access of LHRH terminals to blood vessels in neurohaemal targets, such as the median eminence. In contrast to the isolated appearance of LHRH fibers in

PAGES

FIG. 7. Synaptic input to an LffRH dendrite. (A) Higher magnification of the boxed area in Fig. 6 showing an axon terminal (at) containing clear spherical vesicles (v) and larger, dense-core vesicles (solid arrow), synapsing upon an LHRH dendrite (d). Two active zones (straight open arrows) can be seen; in the postsynaptic dendrite, several small vesicles (curved open arrow) are lined up perpendicular to one of these active zones. (B) Serial section adjacent to that seen in A showing the same axon terminal, in which small, clear vesicles are associated with one of the two active zones shown in A (arrow). 49,000x. FIG. 8. LHRH fibrrs in the organum vasculosum of the lumina terminals (OVLT). (A) Darktield light microscopic photograph of immunoreactive fibers in the preoptic area and fibers and terminals within the OVLT (arrow) oc=optic chiasm. (B) Low power electron micrograph of an isolated LHRH fiber coursing between ependymal cells (ep) adjacent to the preoptic recess of the third ventricle (v).

218

LEHMAN FIG. 7.

AND SILVERMAN

ULTRASTRUCTURE FIG. 8.

OF LHRH NEURONS IN THE HAMSTER

219

220

FIG.

LEHhMN

AND SILVERMAN

and rerminuis in rhr median em&we. (A$) Electron micrographs; (3 is a photomontage) showing LHRH termm&, appear in direct membrane contact Qqxm arrows), most of which contain immunoreactive neurosecretory granllles (n) and

9. LHRHfibers

some of which

non-immunoreactive

mitochundria (m)

and

&ear spherical vesicles (v). B&I glial lamelk

the OVLT, immunoreactive axonal profiles in the median eminence were frequently bundled into fascides. We recently described similar contacts in the median eminence of the sheep (Lehman, submitted manuscript), although they have not been observed in the rat 151.As in the sheep, membrane specializations were not seen along these contacts: although the use of detergents (i.e., saponin) in our immunocytochemical procedures makes the preservation of fine membrane structures {e-g,, gap junctions) unlikely. Further, we could not determine whether these clustered LHRH varieosities had arisen from a single axon or from several axons representing dif%rent cells of origin. If contacts between LHRH axons in the median eminence represent functional associations between different cells, then the median eminence in the sheep and hamster may be a site where neural an&or hormonal signals could activate or inhibit the coordinated refease of LHRH. With the exception of one instance of direct contact between two LHRH somas (Figs. 4, 5), we found little evidence for contacts between LHRH neurons in the preoptic area which might serve as a morphological basis for the LHRH pulse generator. In contrast to observations in the rat [ 16,341 we found no instances of LHRH axon terminals contacting other LHRH neurons. Unlike the rat [35] and guinea pig [23,261, we found no evidence of presynaptic LHRH terminals in preoptic area which might serve as a basis for local circuit connections between LHRH cells and their

(g?) frequently surround these terminals. 36,ooOx.

neighbors. Occassional close somatic rtppositions were abserved between LHRH cells and adjacent non-immun~r~active ceUs (see Fig. 6) but r15 mo~hol~gical specializations could be found along these areas of contact. The lack of LHRH-LHRH contacts at the level of their cell bodies, together with the paucity of synaptic inputs to these cells, suggests that the putative pulsatile activity of LHRH neurons in the hamster may be regulated either by a relatively small number of synapses, by extrinsic humoral signals, or by interactions at the fevel of the median eminence. Finally, it is important to note that the present observations are based solely on male hamsters housed under reproductively stimulatory photoperiods [3,4]. It is possible that changes in photoperiodic signals may induce alterations in either the density of synaptic input to LHRH neurons in the hamster, or in other aspects of their ultrastntcturaf organization. We have recently observed significant seasonal changes in the density of synaptic input onto LHRH cells in the ewe (Karsch er nl., Sot Neur~~i Absrr 13: 1527, 1987), although we have not determined whether this alteration is due to changes in photoperiodic and/or hormonal signals which accompany the transition from the breeding season to anestrus. Future experiments comparing the ultrastructure of LHRH neurons in hamsters housed under either stimulatory or inhibitory photoperiods could examine the possibility of photoperiodically-induced plasticity in the LHRH system.

ULTRASTRUCTURE

OF LHRH

NEURONS

221

IN THE HAMSTER

ACKNOWLEDGEMENTS

We thank Ms. Joni Vest, Mrs. Ying Tsai and Ms. Katharine Bock for technical assistance. Supported by NIH grants HD 21968 (M.N.L.) and HD 10665 (A.J.S.). Portions of this research were presented at the 15th annual meeting of the Society for Neuroscience (Sot Neurosci Ahsrr 9: 318, 1985). REFERENCES I. Carmef, P. W., S, Araki and M. Ferin. Pituitary stalk portal blood collections in rhesus monkeys: Evidence for pufsatile release of gonadotropin release hormone (GnRH). Endocrinn/o,q> 99: 243-248, 1976. 2. Clarke, I. J. and J. T. Cummins. The temporal relationship

between gonadotropin releasing hormone (GnRHf and futeinizing hormone (LH) secretion in ovariectomized ewes. E~?d~)~ri~ology 111: 1737-1739, 1982. 3. Elliott, J. A. Circadian rhythms and photoperiodic time measurement in mammals. Fed Proc 35: 2339-2346, 1976. 4. Gaston, S. and M. Menaker. Photoperiodic control of hamster testis. Srirnce 158: 925-928, 1976. _ 5. Goldsmith. P. C. and W. F. Ganone. Ultrastructural localization of futeinizing hormone-releasing hormone in the median eminence of the rat. Brain Res 97: 181-193, 1975. 6. Hatton, G. I., L. S. Perlmutter, A. K. Safm and C. D. Tweedle. Dynamic neuronal-glial interactions in hypothalamus and pituitary: implications for control of hormone synthesis and release. Pcptides 5: Suppl 1, 121-138, 1984. 7. Itoh, K., A. Konishi, S. Nomura, N. Mizuno, Y. Nakamura and T. Sugimoto. Application of coupled oxidation reaction to efectron microscopic demonstration of horseradish peroxidase: cobalt-glucose oxidase method. Cell Tissw Rcs 198: 381-395, 1979. 8. Jennes, L. and W. E. Stumpf. LHRH-systems in the brain ofthe golden hamster. Cell Tissur Res 209: 239-256, 1980. 9. Jennes, L. and W. E. Stumpf. Gonadotropin-releasing hormone immunoreactive neurons with access to fenestrated capillaries in mouse brain. Neuroscimcc 18: 403-436, 1986. 10. Jennes, L., W. E. Stumpf and M. E. Sheedy. Ultrastructural characterization of gonadotropin-releasing hormone (GnRH)producing neurons. .I Camp Nvurol232: 534547, 1985. 11. Karsch, F. J., E. L. Bittman, D. L. Foster, R. L. Goodman, S. J. Legan and J. E. Robinson. Neuroendocrine basis of seasonal breeding. Ret Prog Harm Res 40: 184-225, 1984. 12. Knob& E. The neuroendocrine control of the menstrual cycle. Ret, Prog Harm Res 36: 53-88. 1980. 13. Langford, L. A. and R. E. Cogneshall. The use of uotassium ferricyanide in neural fixation. &zr Ret 197: 297-303, 1980. 14. Lehman, M. N., S. W. Newman and A. J. Silverman. Luteinizing hormone-releasing hormone (LHRH) in the vomeronasal system and terminal nerve of the hamster. Ann NY Acad Sci, in press, 1987. 1s. Lehman, M. N., J. E. Robinson, F. J. Karsch and A. J. Silverman. Immunocytochemic~ localization of luteinizing hormone-releasing hormone (LHRH) pathways in the sheep brain during anestrus and the mid-futeal phase of the estrous cycle. J Camp Neurni 244: 19-35, 1986. 16. Leranth, Cs., L. M. G. Segura, M. Palkovits, N. J. MacLusky, M. Shanabrough and F. Naftolin. The LHRH-containing neuronal network in the preoptic area of the rat: demonstration of LHRH-containing nerve terminals in synaptic contact with LHRH neurons. Brnin Rrs 34% 332-336, 1985. 17. Levine, J. E., K.-Y. F. Pau, V. D. Ramirez and G. L. Jackson. ^. .+. *. Stmutcaneous measurement 01 lutemizing hormone-releasing hormone and luteinizing hormone release in unanesthetized, ovariectomized sheep. Endocrinology 111: 1449-1455, 1982. 18. Lincoln, D. W., H. M. Fraser, G. A. Lincoln, G. B. Martin and A. S. McNeilly. Hypothalamic puIse generators. Ret Prag Harm Res 41: 369-411, 1985.

19. Liposits, Zs.. T. Gores, F. Gaflyas, B. Kosaras and Gy. Setalo. Improvement of the electron microscopic detection of peroxidase activity by means of silver intensification of the diaminobenzidine reaction in the rat nervous system. Ncurosci Letf 31: 7-11, 1982. 20. Mesh&, C. K.. F. J. Seil and R. M. Herndon. Astrocytes pIay a role in regulation of synaptic density. Brain Rm 402: 139-14.5. 1987. 21. Nottebohm, F. A brain for all seasons: cyclic anatomical changes in song control nuclei of the canary brain. Science 214: 1368-1370, 1981. 22. Phillips. H. S., G. Hostetter, B. Kerdefhue and G. P. Kozlowski. Immunocytochemical localization of LHRH in central olfactory pathways of the hamster. Brain Rcs 193: 574-579, 1980.

23. Silverman, A. J. Luteinizing hormone-releasing hormone concentrating synapses in the diagonal band and preoptic area of the guinea pig. ./ Camp Ncwrol 227: 452-458, 1984. 24. Silverman, A. J., J. L. Antunes, G. M. Abrams, G. Nilaver, R. Thau, J. A. Robinson, M. Ferin and L. C. Krey. The luteinizing hormone-releasing hormone pathways in rhesus (Mrrt.cfcn ma~atfa) and pigtailed (~%f~aucrtnf,~z~.~tril~~t~ monkeys: new observations on thick, unembedded sections. J Cotnp Neural 211: 309-317, 1982. 25. Silverman, A. J., J. Jhamandas and L. Renaud. Localization of luteinizing hormone-releasing hormone (LHRH) neurons that project to the median eminence../ Nearosci 7: 2312-2319, 1987. 26. Silverman, A. J. and J. W. Witkin. Synaptic interactions of luteinizing hormone-releasing hormone (LHRH) in the guinea pig preoptic area. J ~;.sf~~~i~f,~z~~~o~~ern 33: 69-72, 1985. 27. Silverman, A. J., E. A. Zimmerman, G. J. Kokoris and M. J. Gibson. Ultrastructure of gonadotropin-releasing hormone neuronaf structures derived from normal fetal preoptic area and transplanted into hypogonadal mutant (hpg) mice. J Neurosci 6: 2090-2096, 1986. 28. Theodosis. D. T. and D. A. Poulain. Evidence for structural plasticity in the supraoptic nucleus of the rat hy~thalamus in relation to gestation and lactation. ~~~fr~).s~~~~~z~e II: 183-193, 1984. 29. Turek. F. W., J. Swann and D. J. Earnest. Role of the circadian system in reproductive phenomena. Rw Prog Harm Res 40: 143-177. 1984. 30. Tweedle, C. D. and G. 1. Hatton. Morphological adaptability at neurosecretory axonal endings on the neurovascular contact zone of the rat neurohypophysis. ,~ef~r~)~s~~~~n~e 20: 241-246, 1987. 31. Venabfe, J. H. and R. A. Coggeshaff. A simplified lead citrate stain for use in electron microscroov. __ J Cell Biol 25: 405-408, 1965. 32. Wirsig, C. R. and C. M. Leonard. The terminal nerve projects centripetally in the hamster. Neuroscience 19: 709-717. f986. 33. Witkin, J. W., M. Fetin and A. J. Silverman. Ultr&t~cturaf analysis of LHRH neurons in female rhesus macaques. Sac .Neurosti Abstr 13: 18, 1987. 34. Witkin, J. W., C. M. Paden and A. J. Silverman. The luteinizing hormone-releasing hormone (LHRH) systems in the rat brain. Neuroendocrinolagy

35: 429-438,

1982.

35. Witkin, J. W. and A. J. Silverman. Synaptology of luteinizing hormone-releasing hormone neurons in the rat preoptic area. Peptides 6: 263-271, 1985.