Biocytin Filling of Adult Gonadotropin-Releasing Hormone Neurons in Situ Reveals Extensive, Spiny, Dendritic Processes
http://www.100md.com
内分泌学杂志 2005年第3期
Centre for Neuroendocrinology and Department of Physiology, Otago School of Medical Sciences, University of Otago, 9001 Dunedin, New Zealand
Address all correspondence and requests for reprints to: Allan E. Herbison, Centre for Neuroendocrinology, Department of Physiology, University of Otago School of Medical Sciences, P.O. Box 913, Dunedin, New Zealand. E-mail: allan.herbison@stonebow.otago.ac.nz.
Abstract
Ultrastructural studies suggest that GnRH neurons receive relatively few synaptic inputs. However, these techniques are biased toward the analysis of portions of the neuron containing GnRH peptide. Using acute brain slices prepared from transgenic GnRH-green fluorescent protein mice, individual fluorescing GnRH neurons were identified, patched, and filled with the small-molecular-weight dye biocytin. Cells were subsequently visualized with an avidin-conjugated fluorophore, and their morphological characteristics were analyzed by confocal microscopy. In total, 45 GnRH neurons from seven adult male and eight diestrus female mice were examined. Unexpectedly, we found that GnRH neurons possess remarkably long dendritic processes, in some cases extending over 1000 μm distal to the cell body. The somata and dendrites of all GnRH neurons were decorated with an assortment of spine-like protrusions, including filopodia, in an heterogeneous manner. Overall, GnRH neurons had a mean dendritic spine density of 0.4 spines/μm, with the highest densities found in the first 50 μm of the dendrite. GnRH neurons with dendrites running in a horizontal orientation had significantly (P < 0.05) more spines than dendrites with a vertical orientation. The comparison of male and female GnRH neurons revealed no sexually differentiated characteristics of somal or dendritic spine density. Using a technique in which the full extent of the GnRH neuron can be visualized, we demonstrate here a previously unrecognized GnRH neuron morphology of long dendrites covered in spines. These observations suggest that GnRH neurons are not poorly innervated and that they receive abundant excitatory synaptic inputs.
Introduction
GnRH, SYNTHESIZED AND secreted from a diffuse population of neurons in the medial forebrain, drives the central regulation of reproductive function and fertility. Despite their scattered distribution, GnRH neurons function in synchrony to release pulsatile GnRH at the median eminence, which subsequently regulates gonadotrophin release from the pituitary and downstream gonadal events (1). Proper functioning of the GnRH neurons, and resultant fertility, depends upon the integration of both internal and external physiological signals transmitted through the circuits of the GnRH neuronal network. Although numerous neurotransmitters have been reported to be involved in the central regulation of fertility (2, 3, 4, 5), little is known about the specific elements that make up the GnRH neuronal network and how they influence, directly or indirectly, GnRH neurons. A clear understanding of these interactions is essential to expand our knowledge of GnRH neuron physiology.
The morphology of the GnRH neuron has been studied intensively at both light and electron microscopic levels (6). Classically, GnRH neurons in most species have been described as fusiform or bipolar in shape with simple, unbranched dendritic extensions at one or both poles. In addition, the recognition of two distinct morphologies of GnRH neurons has emerged from studies in which their cell bodies and proximal dendrites were described as having either a smooth or spiny profile (7, 8, 9, 10, 11, 12, 13, 14, 15). Importantly, electron microscopic studies have described GnRH neurons as unusual in that they receive a relatively sparse synaptic input (11, 16, 17, 18). Ultrastructural serial reconstructions of GnRH neurons in the rat and monkey found between two and 12 synapses on each cell body (19). This is clearly different to other neuronal cell types in the brain where individual neurons display anywhere between 2,000 and 16,000 synaptic inputs (20, 21). However, the reported paucity of synapses on GnRH neurons is somewhat incongruent with recent electrophysiological data that suggest GnRH neurons receive extensive glutamatergic and -aminobutyric acid (GABA)-ergic synaptic input (22). Although electron microscopy represents a definitive tool by which synapses may be identified, such investigations are limited to examining regions of the cell where GnRH peptide is abundant, typically limiting labeling to the cell body, proximal dendrites, and nerve terminals.
Recently, transgenic mouse models in which the GnRH neurons fluoresce have provided a means through which the scattered GnRH cell bodies can be readily identified and investigated within acute brain slices (23, 24). By patching onto individual GnRH neurons in situ, cells can be filled with highly diffusible, low-molecular-weight molecules such as biocytin and subsequently labeled with a fluorophore for analysis by confocal microscopy. The advantage of visualizing GnRH neurons in this way is that the entire cytoplasmic volume of the neuron can be assessed, including elements such as dendritic spines that cannot be detected reliably with labeling of GnRH peptide alone. Dendritic and somal spines are now established as representing the postsynaptic sites at which the great majority of excitatory inputs make synapses (25, 26). In the present study, we questioned whether an assessment of GnRH neuron spine density with this cell-filling approach might provide a novel way to reevaluate the issue of synaptic input density to these cells.
Materials and Methods
Experimental animals
Male and female homozygous GnRH-GFP (green fluorescent protein) (27) and GnRH-EGFP-mut5 (28) mice were housed under conditions of 12 h of light (lights on at 0700 h) with ad libitum access to food and water. Daily vaginal smears from female mice were taken, and diestrus female mice were used. All experimentation was approved by the University of Otago Animal Welfare and Ethics Committee (Dunedin, New Zealand) under Project 66/02.
Biocytin filling of GnRH neurons
Both lines of mice were treated in the same manner. Between 60 and 75 d of age, male and diestrous female mice were killed by cervical dislocation, and their brains rapidly removed and placed in ice-cold, bicarbonate-buffered, artificial cerebrospinal fluid of the following composition: 118 mM NaCl, 3 mM KCl, 0.5 mM CaCl2, 6.0 mM MgCl2, 11 mM D-glucose, 10 mM HEPES, and 25 mM NaHCO3 (pH 7.4 when bubbled with 95% O2 and 5% CO2). Brains were blocked and glued with cyanoacrylate to the chilled stage of a vibratome, and 200- to 300-μm-thick coronal slices containing the medial septum (MS) and preoptic area were prepared. The slices were then incubated at 30 C for 1 h in oxygenated recording artificial cerebrospinal fluid consisting of: 118 mM NaCl, 3 mM KCl, 2.5 mM CaCl2, 1.2 mM MgCl2, 11 mM D-glucose, 10 mM HEPES, and 25 mM NaHCO3 before being transferred to the recording chamber, held submerged, and continuously superfused at a rate of 2 ml/min at room temperature. Slices were initially examined under fluorescence with a low-power objective to determine the distribution of fluorescent cells. A single, randomly chosen GnRH neuron was then brought into focus under the high-power objective using fluorescence for 5–10 sec before switching to differential interference contrast optics. After patching of the cell under differential interference contrast optics, it was then briefly (5 sec) examined under fluorescence again to confirm its fluorescent identity. Patch pipettes were pulled from thin-wall borosilicate glass capillary tubing (GC150TF, 1.5-mm outer diameter, Harvard Apparatus Ltd., Edenbridge, UK) on a vertical pipette puller. The tip resistance of the electrodes was 4–6 M. The patch pipette solution was passed through a disposable 0.22-μm filter before use and contained: 130 mM KCl, 5 mM NaCl, 0.4 mM CaCl2, 1 mM MgCl2, 10 mM HEPES, and 1.1 mM EGTA, with pH adjusted to 7.3 with KOH. Biocytin was added at a final concentration of 0.5%. After achieving whole-cell mode, the pipette was kept attached to the GnRH neuron for 5 min to allow diffusion of biocytin. After detaching the pipette from the cell, slices were maintained in the recording chamber for 30 min and then placed in ice-cold 4% paraformaldehyde for 4 h at 4 C. Typically, only one cell was filled in each brain slice. Slices were then transferred to Tris-buffered saline (TBS, pH 7.6) and kept at 4 C until processing for immunocytochemistry.
Immunocytochemistry and confocal analysis
Floating slices were washed in TBS at room temperature to remove any residual paraformaldehyde. Slices were then incubated with Texas Red-conjugated avidin (2.6 μl/ml, Vector Laboratories, Inc., Peterborough, UK) in TBS containing 0.3% Triton X-100 and 0.3% BSA for 4 h at room temperature in darkness. Slices were then washed in TBS over 1 h, mounted onto glass slides, and coverslipped with Vectashield aqueous mountant (Vector Laboratories, Inc.).
Forty-five biocytin-filled GnRH neurons were imaged and analyzed on a Zeiss 510 LSM upright confocal laser scanning microscope system using LSM 510 control software (version 3.2) (n = 23 neurons from seven male mice and n = 22 neurons from eight diestrus female mice). Stacks of confocal images were captured using the following objectives, 40x Plan Neofluar (numerical aperture, 1.3), 63x PlanApochromat (numerical aperture, 1.4) with 2x zoom function. A red helium neon laser exciting at 633 nm was used to image the Texas Red fluorophore. A series of images at 0.23-μm intervals throughout the entire filled neuron were collected for analysis. Images are presented as projections of optical images stacks. The brightness and contrast of the images were adjusted in Photoshop (Adobe Systems, San Jose, CA) to match microscope visualization. For two GnRH neurons, a three-dimensional image was rendered for diagrammatic purposes using 3D for LSM (version 1.04.0060, Carl Zeiss GmbH, Jena, Germany).
Using LSM 510 control software (version 3.2), spines were identified as protrusions from the somata or dendrite of less than 5 μm in length (29). Thin, nonbeaded spiny protrusions greater than 5 μm in length from the soma or proximal dendrites were identified as filopodia (29). The dendrite with the greatest circumference extending from the filled GnRH soma was designated the primary dendrite. Spine numbers were recorded for each labeled GnRH soma and for 50-μm intervals of the entire visible length of the primary dendrite. The analyzed intervals of the primary dendritic length were assigned as follows: first 50 μm proximal to soma followed by subsequent 50-μm portions every 100 μm so that spine numbers were recorded at 100- to 150-, 250- to 300-, 400- to 450-μm intervals, etc., to the natural end of the dendrite or until the dendrite was observed to extend out of the slice. For bipolar neurons, the spine numbers of the secondary dendrite were determined by counting spines occurring within the first 50 μm. Where the secondary dendrite was shorter than 50 μm, the number of spines per 10 μm was determined and compared with the same region on the primary dendrite. The orientation and morphological characteristics of each neuron were recorded, and the entire dendritic length was traced and measured. Labeled GnRH neurons were classified by brain region [preoptic area/diagonal band of Broca (POA/DBB) or MS], overall morphological profile (unipolar, Fig. 1A; bipolar, Fig. 1B; or irregular, Fig. 1C); the latter being defined as having more than two dendrites within 50 μm of the cell body or a very proximally branched primary dendrite) and primary dendritic orientation (vertical or horizontal/diagonal). An investigator blinded to experimental groups counted spines in two dimensions using collected z series confocal images.
FIG. 1. Confocal images showing high-power images of biocytin-filled GnRH neurons displaying unipolar (A), bipolar (B), and irregular (C) morphologies and both horizontally (D) and diagonally (E) orientated dendrites. The convoluted termination of a GnRH dendrite is shown in F (arrowheads). G, Montage of confocal images to show the full extent of a GnRH dendrite that runs for 538 μm in the vertical plane. Inset, Terminal portion of the dendrite with some of its spines indicated by arrowheads. Note again the convoluted termination of the dendrite (yellow arrowhead). Scale bars represent 10 μm in all plates.
Statistical analysis
The number of neurons analyzed is presented by n and experimental data are expressed as mean ± SEM. Data were analyzed by one-way ANOVA followed by Tukey’s post hoc test using InStat software (GraphPad Software, San Diego, CA). P < 0.05 was considered significant.
Results
The 45 biocytin-filled GnRH neurons analyzed represented a random population from the rostral to caudal extent of the GnRH neuron population. The majority of filled neurons (n = 34; 75%) were located within the POA/DBB region, nine neurons (20%) were located in the MS, and two filled cells (4%) were located in the lateral anterior hypothalamic area. Fifteen percent of filled neurons were unipolar (Fig. 1A), and 65% were bipolar with two dendrites extending from opposite poles of the soma (Fig. 1B), whereas 25% had three or more dendrites (Fig. 1C). Most GnRH neurons (71%) exhibited a vertical orientation with vertical dendritic extensions, 15% had dendritic extensions in the horizontal plane within the slice (Fig. 1D), and 13% had dendrites arranged in a diagonal fashion (Fig. 1E). Although the orientation of each neuron was classified according to the direction of its primary dendrite, many dendrites were observed to change direction (Fig. 1G). Filled neurons from the different transgenic mouse models, GnRH-GFP and GnRH-EGFP-mut5, were indistinguishable from one another.
Biocytin filling revealed that GnRH neurons extend long primary dendrites (Fig. 1G) with an average length of 307 ± 49 μm. However, half (23 of 45) of these cells had primary dendrites that exited the coronal brain slice, and their true length was not established. Notably, three GnRH neurons displayed dendrites extending over 1000 μm in length, and one GnRH dendrite measured 1627 μm from cell body to its point of exit from the slice. In cases where the proximal dendrite ended within the brain slice, a characteristic, extensive looping and tufted morphology was observed (Fig. 1, F and G, insert).
All 45 biocytin-filled GnRH neurons exhibited somal and dendritic spines with densities ranging up to a maximum of 1 spine/μm plasma membrane. The majority of spines observed on GnRH neurons were simple sessile spines, synaptic protrusions without a neck constriction; however, thin pedunculated spines and mushroom-like spines (30) were also observed (Fig. 2, C and D). Approximately half of the GnRH neurons (23 of 45) had long, thin filopodia extending from their cell bodies and proximal dendrites (Fig. 2D). Although some neurons had up to three filopodia, on average, GnRH neurons displayed less than one filopodium (0.73 ± 0.13) per neuron.
FIG. 2. High-power confocal images of spines and filopodia on biocytin-filled GnRH neurons. A and B, Note the large numbers of spines in the two dendrites. C and D, Three-dimensional render-ing of confocal images of two biocytin-filled GnRH cell bodies showing spines and filopodia. Black arrows, Examples of mushroom-shaped spines; white arrows, simple spines. Two filopodia are seen in D and marked with arrowheads. Scale bars represent 5 μm in all plates.
The numbers of spines detected on GnRH soma were observed to vary considerably ranging from one to 41 spine-like processes, with a mean of 13 ± 10 spines/soma. An analysis of spine density along the length of the primary dendrite revealed that spines were quantitatively most abundant within the first 50 μm, with an average of 24 ± 2 spines in the first 50 μm of the dendrite (Fig. 3A). The secondary dendrite exhibited a mean spine density of 0.5 ± 0.1 spines/μm within the first 50 μm, compared with 0.4 ± 0.1 spines/μm for the same regions of the designated primary dendrite (n = 10). Although there was a trend for a gradual reduction in spine numbers with increasing distance from the cell body, spines were still observed even on the most distal elements of the primary dendrite (Figs. 1G and 3A). On an individual neuron basis, the density of dendritic spines ranged from 0.07–1.04 spines/μm with a mean of 0.4 spines/μm of dendritic membrane.
FIG. 3. A, Histogram showing the mean + SEM number of spines detected upon the cell somata and dendrites of biocytin-filled mature GnRH neurons (n = 45). Dendritic spine numbers are shown per 50-μm segment running distally from the cell body. B, Histogram showing the mean + SEM numbers of spines detected upon the cell soma and dendrites of biocytin-filled adult male (white box, n = 22) and female (black box, n = 23) GnRH neurons.
Because a previous electron microscopic study had demonstrated that the synaptic input onto the cell body of male GnRH neurons was only half that of female GnRH neurons in the rat (31), we examined whether sex differences were also apparent in spine densities. No significant differences were found in spine number between male (n = 22) and diestrus female (n = 23) GnRH neurons in any neuronal segment (Fig. 3B).
In an effort to reveal any trends in spine density among the GnRH neurons, spine numbers were compared between neurons classified by neuronal morphology, anatomical location, and dendritic orientation. Spine numbers were not significantly different among GnRH neurons classified as classically unipolar, bipolar, or irregular in morphological profile (Fig. 4A). Likewise, no significant differences in spine number were found between GnRH neurons located in the POA/DBB compared with the MS (Fig. 4B). However, neurons with dendrites extending horizontally or diagonally had significantly more spines on their proximal dendrites than those of cells with vertically orientated dendrites (Fig. 4C). Although not significantly different (P = 0.0514), a similar trend of increased spine density in horizontal/diagonal dendrites was also seen at greater dendritic lengths (Fig. 4C).
FIG. 4. A, Histogram showing the mean + SEM number of spines detected upon the soma and first 450 μm of primary dendrite of GnRH neurons exhibiting unipolar (black bar, n = 6), bipolar (gray bar, n = 20), and irregular (white bar, n = 7) morphologies. B, Histogram showing the mean + SEM number of spines detected upon the soma and first 450 μm of primary dendrite of POA/DBB (black bar, n = 32) and MS (white bar, n = 11) GnRH neurons. C, Histogram showing the mean + SEM number of spines detected upon the soma and first 150 μm of primary dendrite exhibiting a diagonal/horizontal (white bars, n = 13) or vertical (black bars, n = 32) orientation. *, P < 0.05.
Discussion
The present study has used a dye-filling approach in GnRH-GFP transgenic mice to reevaluate GnRH neuron morphology. We report here that the primary dendrites of GnRH neurons are much longer than had been recognized previously, and that GnRH neuron dendrites and somata exhibit numerous spine-like elements. Ultrastructural investigations will be required to establish with certainty that these spine-like structures represent points of synaptic input to GnRH neurons. However, elsewhere in the nervous system, spines such as those visualized here have always been found to be associated with excitatory synaptic transmission (29). Together, these results extend prior observations of subpopulations of spiny GnRH neurons using immunocytochemistry (7, 8, 9, 10, 11, 12, 13, 14, 15, 18) and suggest that the density of synaptic input to GnRH neurons has been underestimated.
Given that the average dendritic spine density of a GnRH neuron was 0.4 spines/μm of membrane, and that the mean primary dendrite length was 350 μm, we estimate that an average GnRH neuron has approximately 160 spines. However, this estimate is very conservative. For one, we have only been able to evaluate spine density in a two-dimensional manner and all spines perpendicular to the plane of the slice have not been counted. Secondly, we have only performed a detailed analysis on the primary dendrite and so bipolar and irregular GnRH neurons will have more spines per cell. Furthermore, the actual length of the primary dendrite in over half of the GnRH neurons examined was longer than the recorded value because it exited the brain slice before terminating. Three GnRH neurons were found to have primary dendrites more than 1000 μm in length. For these reasons, we believe that individual GnRH neurons are, in reality, very likely to receive hundreds of individual excitatory inputs, and that this number will be determined principally by their dendritic architecture. Given that this estimate does not include the substantial GABAergic inputs, which appear to greatly out number excitatory inputs in GnRH neurons (22, 32), or certain neuropeptidergic and aminergic afferents to the GnRH neuron, the total number of synapses upon a GnRH neuron could easily be within the range of 500-1000. This indicates that GnRH neurons are not a neuronal phenotype with relatively sparse synaptic input. Indeed, our conservative estimate of 0.4 dendritic spines/μm for GnRH neurons shows that these cells receive more excitatory inputs than do unidentified neurons in the preoptic area (0.1 dendritic spines/μm) (33), ventromedial nucleus (0.2 dendritic spines/μm) (34), and arcuate nucleus (0.25 dendritic spines/μm) (35) of the rat hypothalamus.
Although providing a different view of GnRH neuron morphology, the present findings are not necessarily discordant with prior ultrastructural studies. When comparing results at the level of the cell body, it is interesting to note that we found an average of 13 spines/cell body compared with two to 12 synapses detected in serial ultrastructural reconstructions of GnRH cell bodies in the rat and monkey (19). Admittedly, spine numbers will only represent a subset of all synaptic inputs to the GnRH cell bodies, but, given this, the values are not substantially different. A recent electron microscopic study reported that monkey GnRH cell bodies and proximal dendrites had a synapse density of 0.1 μm/μm plasma membrane (32). However, there appear to be clear differences between the ultrastructural and cell-filling data in terms of dendritic morphology, with many more synaptic inputs suggested by the latter method. One reason for this difference may be that major species differences exist between GnRH neurons in the mouse compared with rats, sheep, and monkeys. Another possibility may be that spine numbers are altered in GnRH neurons as a result of brain slicing. A 40% increase in spine numbers on hippocampal neurons has been reported 1 h after slice preparation (36). However, even if this was to be the case, our data would still suggest that GnRH neurons receive substantially more inputs than has been recognized previously. We believe that the most likely reason for the discrepancy, however, lies in the ability of the cell-filling approach to reveal the whole of the dendritic tree including elements of the GnRH neuron that do not contain GnRH peptide. Although dendritic spines have been observed by electron microscopy on GnRH neurons (11, 17), it is likely very difficult to visualize structures protruding up to 5 μm away from the cell membrane with this approach.
We also find evidence here for filopodium-like structures extending from around half of all GnRH neurons. Similar structures have been observed in ovine GnRH neurons (13). Filapodia are thought to be involved in synaptogenesis by locating and guiding appropriate axons back to the dendrite (37, 38, 39). As such, filopodia are highly motile and can extend and retract within minutes (40). However, filopodia are thought to be rare in the mature brain (29), and their observation here on mature GnRH neurons is intriguing and may suggest ongoing synaptogenesis in these cells. In addition, it is interesting to note that we found moderate numbers of spine-like structures on the soma themselves. With some exceptions (41), somatic spines are not common in the adult nervous system.
With the exception of the orientation of the primary dendrite, we failed to find any topographical or morphological parameter that correlated with the range of spine densities observed on GnRH neurons. The nature and functional significance of horizontally orientated dendrites having more excitatory inputs than vertical dendrites can only be speculated upon. It has been suggested that dendritic orientation reflects, in part, trophic influences of excitatory afferents during development (42). It is possible that horizontally orientated dendrites experience a different pattern of excitatory input during development and that this then remains in the form of increased spine numbers and synaptic input during adulthood.
We also failed to detect any significant sex differences in spine densities upon GnRH cell bodies and dendrites. This was surprising considering a previous electron microscopic study reported that GnRH neurons in female rats had twice the number of synapses as those found on GnRH neurons in males (31). Nevertheless, an earlier immunocytochemical study had failed to detect sex differences in the number of spiny verses smooth GnRH neurons found in the rat (43). As noted above, spines may be particularly difficult to evaluate with electron microscopy compared with synapses that exist directly upon on the somal plasma membrane. Thus, it remains possible that excitatory inputs to GnRH neurons via spines are not sexually differentiated, whereas nonspinous (e.g. GABAergic) inputs may be more abundant in females. There is some evidence to support this idea as GABAergic inputs to mouse GnRH neurons are modified by masculinizing prenatal androgen exposure in females (44), and sex differences have been observed in GABAA receptor subunit mRNA expression in mouse GnRH neurons (45).
In summary, we report here a previously unsuspected dendritic morphology for the GnRH neuronal phenotype with long primary dendrites that may extend for up to 1500 μm. We also show that all GnRH neurons have the potential to receive substantial numbers of excitatory inputs. Such findings are in good agreement with the recent electrophysiological (27, 46, 47) and receptor-labeling (48, 49) studies that have demonstrated direct glutamatergic neurotransmission in mouse GnRH neurons. The highest density of excitatory input to a GnRH neuron was found within the first 50 μm of proximal dendrite. Together, these findings indicate that, like other final output neurons (50), the GnRH neuron receives many synaptic inputs and, as such, represents a major site of neuronal integration within the GnRH neuronal network.
Acknowledgments
We thank Dr. C. Jasoni for critical appraisal of an earlier draft.
References
Levine JE 2003 Gonadotropin-releasing hormone (GnRH). In: Henry H, Norman A, eds. Encyclopedia of hormones. San Diego: Academic Press; 157–165
Kalra SP 1993 Mandatory neuropeptide-steroid signaling for the preovulatory luteinizing hormone-releasing hormone discharge. Endocr Rev 14:507–538
Herbison AE 1998 Multimodal influence of estrogen upon gonadotropin-releasing hormone neurons. Endocr Rev 19:302–330
Smith MJ, Jennes L 2001 Neural signals that regulate GnRH neurones directly during the oestrous cycle. Reproduction 122:1–10
Petersen SL, Ottem EN, Carpenter CD 2003 Direct and indirect regulation of gonadotropin-releasing hormone neurons by estradiol. Biol Reprod 69:1771–1778
Silverman A, Livne I, Witkin JW 1994 The gonadotrophin-releasing hormone (GnRH), neuronal systems: immunocytochemistry and in situ hybridization. In: Knobil E, Neill JD, eds. The physiology of reproduction. 2nd ed. New York: Raven; 1683–1706
Kozlowski GP, Chu L, Hostetter G, Kerdelhue B 1980 Cellular characteristics of immunolabeled luteinizing hormone releasing hormone (LHRH) neurons. Peptides 1:37–46
Krisch B 1980 Two types of luliberin-immunoreactive perikarya in the preoptic area of the rat. Cell Tissue Res 212:443–455
King JC, Tobet SA, Snavely FL, Arimura AA 1982 LHRH immunopositive cells and their projections to the median eminence and organum vasculosum of the lamina terminalis. J Comp Neurol 209:287–300
Witkin JW, Paden CM, Silverman AJ 1982 The luteinizing hormone-releasing hormone (LHRH) systems in the rat brain. Neuroendocrinology 35:429–438
Jennes L, Stumpf WE, Sheedy ME 1985 Ultrastructural characterization of gonadotropin-releasing hormone (GnRH)-producing neurons. J Comp Neurol 232:534–547
Liposits Z, Setalo G, Flerko B 1984 Application of the silver-gold intensified 3,3'-diaminobenzidine chromogen to the light and electron microscopic detection of the luteinizing hormone-releasing hormone system of the rat brain. Neuroscience 13:513–525
Lehman MN, Robinson JE, Karsch FJ, Silverman AJ 1986 Immunocytochemical localization of luteinizing hormone-releasing hormone (LHRH) pathways in the sheep brain during anestrus and the mid-luteal phase of the estrous cycle. J Comp Neurol 244:19–35
Wray S, Hoffman G 1986 Postnatal morphological changes in rat LHRH neurons correlated with sexual maturation. Neuroendocrinology 43:93–97
Witkin JW 1996 Effects of ovariectomy on GnRH neuronal morphology in rhesus monkey (Macaca mulatta). J Neuroendocrinol 8:601–604
Witkin JW, Ferin M, Popilskis SJ, Silverman A 1991 Effects of gonadal steroids on the ultrastructure of GnRH neurons in the rhesus monkey: synaptic input and glial apposition. Endocrinology 129:1083–1092
Witkin JW, Silverman AJ 1985 Synaptology of luteinizing hormone-releasing hormone neurons in rat preoptic area. Peptides 6:263–271
Lehman MN, Karsch FJ, Silverman AJ 1988 Ultrastructure and synaptic organization of luteinizing hormone-releasing hormone (LHRH) neurons in the anestrous ewe. J Comp Neurol 273:447–458
Witkin JW, O’Sullivan H, Silverman AJ 1995 Novel associations among gonadotropin-releasing hormone neurons. Endocrinology 136:4323–4330
Gulyas AI, Megias M, Emri Z, Freund TF 1999 Total number and ratio of excitatory and inhibitory synapses converging onto single interneurons of different types in the CA1 area of the rat hippocampus. J Neurosci 19:10082–10097
Matyas F, Freund TF, Gulyas AI 2004 Convergence of excitatory and inhibitory inputs onto CCK-containing basket cells in the CA1 area of the rat hippocampus. Eur J Neurosci 19:1243–1256
Herbison AE 2003 GnRH neuron. In: Henry H, Norman A, eds. Encyclopedia of hormones. San Diego: Academic Press; 171–177
Spergel DJ, Kruth U, Shimshek DR, Sprengel R, Seeburg PH 2001 Using reporter genes to label selected neuronal populations in transgenic mice for gene promoter, anatomical, and physiological studies. Prog Neurobiol 63:673–686
Herbison AE, Pape JR, Simonian SX, Skynner MJ, Sim JA 2001 Molecular and cellular properties of GnRH neurons revealed through transgenics in the mouse. Mol Cell Endocrinol 185:185–194
Sorra KE, Harris KM 2000 Overview on the structure, composition, function, development, and plasticity of hippocampal dendritic spines. Hippocampus 10:501–511
Harris KM, Kater SB 1994 Dendritic spines: cellular specializations imparting both stability and flexibility to synaptic function. Annu Rev Neurosci 17:341–371
Spergel DJ, Kruth U, Hanley DF, Sprengel R, Seeburg PH 1999 GABA- and glutamate-activated channels in green fluorescent protein-tagged gonadotropin-releasing hormone neurone in transgenic mice. J Neurosci 19:2037–2050
Han SK, Todman MG, Herbison AE 2004 Endogenous GABA release inhibits the firing of adult gonadotropin-releasing hormone neurons. Endocrinology 145:495–499
Fiala JC, Spacek J, Harris KM 2002 Dendritic spine pathology: cause or consequence of neurological disorders? Brain Res Brain Res Rev 39:29–54
Harris KM, Jensen FE, Tsao B 1992 Three-dimensional structure of dendritic spines and synapses in rat hippocampus (CA1) at postnatal day 15 and adult ages: implications for the maturation of synaptic physiology and long-term potentiation. J Neurosci 12:2685–2705
Chen WP, Witkin JW, Silverman AJ 1990 Sexual dimorphism in the synaptic input to gonadotropin releasing hormone neurons. Endocrinology 126:695–702
Zsarnovszky A, Horvath TL, Garcia-Segura LM, Horvath B, Naftolin F 2001 Oestrogen-induced changes in the synaptology of the monkey (Cercopithecus aethiops) arcuate nucleus during gonadotropin feedback. J Neuroendocrinol 13:22–28
Amateau SK, McCarthy MM 2002 A novel mechanism of dendritic spine plasticity involving estradiol induction of prostaglandin-E2. J Neurosci 22:8586–8596
Calizo LH, Flanagan-Cato LM 2000 Estrogen selectively regulates spine density within the dendritic arbor of rat ventromedial hypothalamic neurons. J Neurosci 20:1589–1596
Mong JA, Glaser E, McCarthy MM 1999 Gonadal steroids promote glial differentiation and alter neuronal morphology in the developing hypothalamus in a regionally specific manner. J Neurosci 19:1464–1472
Kirov SA, Sorra KE, Harris KM 1999 Slices have more synapses than perfusion-fixed hippocampus from both young and mature rats. J Neurosci 19:2876–2886
Fiala JC, Feinberg M, Popov V, Harris KM 1998 Synaptogenesis via dendritic filopodia in developing hippocampal area CA1. J Neurosci 18:8900–8911
Wong WT, Wong RO 2001 Changing specificity of neurotransmitter regulation of rapid dendritic remodeling during synaptogenesis. Nat Neurosci 4:351–352
Maletic-Savatic M, Malinow R, Svoboda K 1999 Rapid dendritic morphogenesis in CA1 hippocampal dendrites induced by synaptic activity. Science 283:1923–1927
Dailey ME, Smith SJ 1996 The dynamics of dendritic structure in developing hippocampal slices. J Neurosci 16:2983–2994
Jakab RL, Leranth C 1990 Somatospiny neurons in the rat lateral septal area are synaptic targets of hippocamposeptal fibers: a combined EM/Golgi and degeneration study. Synapse 6:10–22
Wu GY, Zou DJ, Rajan I, Cline H 1999 Dendritic dynamics in vivo change during neuronal maturation. J Neurosci 19:4472–4483
Wray S, Gainer H 1987 Effect of neonatal gonadectomy on the postnatal development of LHRH cell subtypes in male and female rats. Neuroendocrinology 45:413–419
Sullivan SD, Moenter SM 2004 Prenatal androgens alter GABAergic drive to gonadotropin-releasing hormone neurons: implications for a common fertility disorder. Proc Natl Acad Sci USA101:7129–7134
Pape JR, Skynner MJ, Sim JA, Herbison AE 2001 Profiling -aminobutyric acid (GABA(A)) receptor subunit mRNA expression in postnatal gonadotropin-releasing hormone (GnRH) neurons of the male mouse with single cell RT-PCR. Neuroendocrinology 74:300–308
Kuehl-Kovarik MC, Pouliot WA, Halterman GL, Handa RJ, Dudek FE, Partin KM 2002 Episodic bursting activity and response to excitatory amino acids in acutely dissociated gonadotropin-releasing hormone neurons genetically targeted with green fluorescent protein. J Neurosci 22:2313–2322
Suter KJ 2004 Control of firing by small (S)--amino-3-hydroxy-5-methyl-isoxazolepropionic acid-like inputs in hypothalamic gonadotropin releasing-hormone (GnRH) neurons. Neuroscience 128:443–450
Gore AC 2002 Gonadotropin-releasing hormone (GnRH) neurons: gene expression and neuroanatomical studies. Prog Brain Res 141:193–208
Jennes L, Lin W, Lakhlani S 2002 Glutamatergic regulation of gonadotropin-releasing hormone neurons. Prog Brain Res 141:183–192
Napper RM, Harvey RJ 1988 Quantitative study of the Purkinje cell dendritic spines in the rat cerebellum. J Comp Neurol 274:158–167(Rebecca E. Campbell, Seon)
Address all correspondence and requests for reprints to: Allan E. Herbison, Centre for Neuroendocrinology, Department of Physiology, University of Otago School of Medical Sciences, P.O. Box 913, Dunedin, New Zealand. E-mail: allan.herbison@stonebow.otago.ac.nz.
Abstract
Ultrastructural studies suggest that GnRH neurons receive relatively few synaptic inputs. However, these techniques are biased toward the analysis of portions of the neuron containing GnRH peptide. Using acute brain slices prepared from transgenic GnRH-green fluorescent protein mice, individual fluorescing GnRH neurons were identified, patched, and filled with the small-molecular-weight dye biocytin. Cells were subsequently visualized with an avidin-conjugated fluorophore, and their morphological characteristics were analyzed by confocal microscopy. In total, 45 GnRH neurons from seven adult male and eight diestrus female mice were examined. Unexpectedly, we found that GnRH neurons possess remarkably long dendritic processes, in some cases extending over 1000 μm distal to the cell body. The somata and dendrites of all GnRH neurons were decorated with an assortment of spine-like protrusions, including filopodia, in an heterogeneous manner. Overall, GnRH neurons had a mean dendritic spine density of 0.4 spines/μm, with the highest densities found in the first 50 μm of the dendrite. GnRH neurons with dendrites running in a horizontal orientation had significantly (P < 0.05) more spines than dendrites with a vertical orientation. The comparison of male and female GnRH neurons revealed no sexually differentiated characteristics of somal or dendritic spine density. Using a technique in which the full extent of the GnRH neuron can be visualized, we demonstrate here a previously unrecognized GnRH neuron morphology of long dendrites covered in spines. These observations suggest that GnRH neurons are not poorly innervated and that they receive abundant excitatory synaptic inputs.
Introduction
GnRH, SYNTHESIZED AND secreted from a diffuse population of neurons in the medial forebrain, drives the central regulation of reproductive function and fertility. Despite their scattered distribution, GnRH neurons function in synchrony to release pulsatile GnRH at the median eminence, which subsequently regulates gonadotrophin release from the pituitary and downstream gonadal events (1). Proper functioning of the GnRH neurons, and resultant fertility, depends upon the integration of both internal and external physiological signals transmitted through the circuits of the GnRH neuronal network. Although numerous neurotransmitters have been reported to be involved in the central regulation of fertility (2, 3, 4, 5), little is known about the specific elements that make up the GnRH neuronal network and how they influence, directly or indirectly, GnRH neurons. A clear understanding of these interactions is essential to expand our knowledge of GnRH neuron physiology.
The morphology of the GnRH neuron has been studied intensively at both light and electron microscopic levels (6). Classically, GnRH neurons in most species have been described as fusiform or bipolar in shape with simple, unbranched dendritic extensions at one or both poles. In addition, the recognition of two distinct morphologies of GnRH neurons has emerged from studies in which their cell bodies and proximal dendrites were described as having either a smooth or spiny profile (7, 8, 9, 10, 11, 12, 13, 14, 15). Importantly, electron microscopic studies have described GnRH neurons as unusual in that they receive a relatively sparse synaptic input (11, 16, 17, 18). Ultrastructural serial reconstructions of GnRH neurons in the rat and monkey found between two and 12 synapses on each cell body (19). This is clearly different to other neuronal cell types in the brain where individual neurons display anywhere between 2,000 and 16,000 synaptic inputs (20, 21). However, the reported paucity of synapses on GnRH neurons is somewhat incongruent with recent electrophysiological data that suggest GnRH neurons receive extensive glutamatergic and -aminobutyric acid (GABA)-ergic synaptic input (22). Although electron microscopy represents a definitive tool by which synapses may be identified, such investigations are limited to examining regions of the cell where GnRH peptide is abundant, typically limiting labeling to the cell body, proximal dendrites, and nerve terminals.
Recently, transgenic mouse models in which the GnRH neurons fluoresce have provided a means through which the scattered GnRH cell bodies can be readily identified and investigated within acute brain slices (23, 24). By patching onto individual GnRH neurons in situ, cells can be filled with highly diffusible, low-molecular-weight molecules such as biocytin and subsequently labeled with a fluorophore for analysis by confocal microscopy. The advantage of visualizing GnRH neurons in this way is that the entire cytoplasmic volume of the neuron can be assessed, including elements such as dendritic spines that cannot be detected reliably with labeling of GnRH peptide alone. Dendritic and somal spines are now established as representing the postsynaptic sites at which the great majority of excitatory inputs make synapses (25, 26). In the present study, we questioned whether an assessment of GnRH neuron spine density with this cell-filling approach might provide a novel way to reevaluate the issue of synaptic input density to these cells.
Materials and Methods
Experimental animals
Male and female homozygous GnRH-GFP (green fluorescent protein) (27) and GnRH-EGFP-mut5 (28) mice were housed under conditions of 12 h of light (lights on at 0700 h) with ad libitum access to food and water. Daily vaginal smears from female mice were taken, and diestrus female mice were used. All experimentation was approved by the University of Otago Animal Welfare and Ethics Committee (Dunedin, New Zealand) under Project 66/02.
Biocytin filling of GnRH neurons
Both lines of mice were treated in the same manner. Between 60 and 75 d of age, male and diestrous female mice were killed by cervical dislocation, and their brains rapidly removed and placed in ice-cold, bicarbonate-buffered, artificial cerebrospinal fluid of the following composition: 118 mM NaCl, 3 mM KCl, 0.5 mM CaCl2, 6.0 mM MgCl2, 11 mM D-glucose, 10 mM HEPES, and 25 mM NaHCO3 (pH 7.4 when bubbled with 95% O2 and 5% CO2). Brains were blocked and glued with cyanoacrylate to the chilled stage of a vibratome, and 200- to 300-μm-thick coronal slices containing the medial septum (MS) and preoptic area were prepared. The slices were then incubated at 30 C for 1 h in oxygenated recording artificial cerebrospinal fluid consisting of: 118 mM NaCl, 3 mM KCl, 2.5 mM CaCl2, 1.2 mM MgCl2, 11 mM D-glucose, 10 mM HEPES, and 25 mM NaHCO3 before being transferred to the recording chamber, held submerged, and continuously superfused at a rate of 2 ml/min at room temperature. Slices were initially examined under fluorescence with a low-power objective to determine the distribution of fluorescent cells. A single, randomly chosen GnRH neuron was then brought into focus under the high-power objective using fluorescence for 5–10 sec before switching to differential interference contrast optics. After patching of the cell under differential interference contrast optics, it was then briefly (5 sec) examined under fluorescence again to confirm its fluorescent identity. Patch pipettes were pulled from thin-wall borosilicate glass capillary tubing (GC150TF, 1.5-mm outer diameter, Harvard Apparatus Ltd., Edenbridge, UK) on a vertical pipette puller. The tip resistance of the electrodes was 4–6 M. The patch pipette solution was passed through a disposable 0.22-μm filter before use and contained: 130 mM KCl, 5 mM NaCl, 0.4 mM CaCl2, 1 mM MgCl2, 10 mM HEPES, and 1.1 mM EGTA, with pH adjusted to 7.3 with KOH. Biocytin was added at a final concentration of 0.5%. After achieving whole-cell mode, the pipette was kept attached to the GnRH neuron for 5 min to allow diffusion of biocytin. After detaching the pipette from the cell, slices were maintained in the recording chamber for 30 min and then placed in ice-cold 4% paraformaldehyde for 4 h at 4 C. Typically, only one cell was filled in each brain slice. Slices were then transferred to Tris-buffered saline (TBS, pH 7.6) and kept at 4 C until processing for immunocytochemistry.
Immunocytochemistry and confocal analysis
Floating slices were washed in TBS at room temperature to remove any residual paraformaldehyde. Slices were then incubated with Texas Red-conjugated avidin (2.6 μl/ml, Vector Laboratories, Inc., Peterborough, UK) in TBS containing 0.3% Triton X-100 and 0.3% BSA for 4 h at room temperature in darkness. Slices were then washed in TBS over 1 h, mounted onto glass slides, and coverslipped with Vectashield aqueous mountant (Vector Laboratories, Inc.).
Forty-five biocytin-filled GnRH neurons were imaged and analyzed on a Zeiss 510 LSM upright confocal laser scanning microscope system using LSM 510 control software (version 3.2) (n = 23 neurons from seven male mice and n = 22 neurons from eight diestrus female mice). Stacks of confocal images were captured using the following objectives, 40x Plan Neofluar (numerical aperture, 1.3), 63x PlanApochromat (numerical aperture, 1.4) with 2x zoom function. A red helium neon laser exciting at 633 nm was used to image the Texas Red fluorophore. A series of images at 0.23-μm intervals throughout the entire filled neuron were collected for analysis. Images are presented as projections of optical images stacks. The brightness and contrast of the images were adjusted in Photoshop (Adobe Systems, San Jose, CA) to match microscope visualization. For two GnRH neurons, a three-dimensional image was rendered for diagrammatic purposes using 3D for LSM (version 1.04.0060, Carl Zeiss GmbH, Jena, Germany).
Using LSM 510 control software (version 3.2), spines were identified as protrusions from the somata or dendrite of less than 5 μm in length (29). Thin, nonbeaded spiny protrusions greater than 5 μm in length from the soma or proximal dendrites were identified as filopodia (29). The dendrite with the greatest circumference extending from the filled GnRH soma was designated the primary dendrite. Spine numbers were recorded for each labeled GnRH soma and for 50-μm intervals of the entire visible length of the primary dendrite. The analyzed intervals of the primary dendritic length were assigned as follows: first 50 μm proximal to soma followed by subsequent 50-μm portions every 100 μm so that spine numbers were recorded at 100- to 150-, 250- to 300-, 400- to 450-μm intervals, etc., to the natural end of the dendrite or until the dendrite was observed to extend out of the slice. For bipolar neurons, the spine numbers of the secondary dendrite were determined by counting spines occurring within the first 50 μm. Where the secondary dendrite was shorter than 50 μm, the number of spines per 10 μm was determined and compared with the same region on the primary dendrite. The orientation and morphological characteristics of each neuron were recorded, and the entire dendritic length was traced and measured. Labeled GnRH neurons were classified by brain region [preoptic area/diagonal band of Broca (POA/DBB) or MS], overall morphological profile (unipolar, Fig. 1A; bipolar, Fig. 1B; or irregular, Fig. 1C); the latter being defined as having more than two dendrites within 50 μm of the cell body or a very proximally branched primary dendrite) and primary dendritic orientation (vertical or horizontal/diagonal). An investigator blinded to experimental groups counted spines in two dimensions using collected z series confocal images.
FIG. 1. Confocal images showing high-power images of biocytin-filled GnRH neurons displaying unipolar (A), bipolar (B), and irregular (C) morphologies and both horizontally (D) and diagonally (E) orientated dendrites. The convoluted termination of a GnRH dendrite is shown in F (arrowheads). G, Montage of confocal images to show the full extent of a GnRH dendrite that runs for 538 μm in the vertical plane. Inset, Terminal portion of the dendrite with some of its spines indicated by arrowheads. Note again the convoluted termination of the dendrite (yellow arrowhead). Scale bars represent 10 μm in all plates.
Statistical analysis
The number of neurons analyzed is presented by n and experimental data are expressed as mean ± SEM. Data were analyzed by one-way ANOVA followed by Tukey’s post hoc test using InStat software (GraphPad Software, San Diego, CA). P < 0.05 was considered significant.
Results
The 45 biocytin-filled GnRH neurons analyzed represented a random population from the rostral to caudal extent of the GnRH neuron population. The majority of filled neurons (n = 34; 75%) were located within the POA/DBB region, nine neurons (20%) were located in the MS, and two filled cells (4%) were located in the lateral anterior hypothalamic area. Fifteen percent of filled neurons were unipolar (Fig. 1A), and 65% were bipolar with two dendrites extending from opposite poles of the soma (Fig. 1B), whereas 25% had three or more dendrites (Fig. 1C). Most GnRH neurons (71%) exhibited a vertical orientation with vertical dendritic extensions, 15% had dendritic extensions in the horizontal plane within the slice (Fig. 1D), and 13% had dendrites arranged in a diagonal fashion (Fig. 1E). Although the orientation of each neuron was classified according to the direction of its primary dendrite, many dendrites were observed to change direction (Fig. 1G). Filled neurons from the different transgenic mouse models, GnRH-GFP and GnRH-EGFP-mut5, were indistinguishable from one another.
Biocytin filling revealed that GnRH neurons extend long primary dendrites (Fig. 1G) with an average length of 307 ± 49 μm. However, half (23 of 45) of these cells had primary dendrites that exited the coronal brain slice, and their true length was not established. Notably, three GnRH neurons displayed dendrites extending over 1000 μm in length, and one GnRH dendrite measured 1627 μm from cell body to its point of exit from the slice. In cases where the proximal dendrite ended within the brain slice, a characteristic, extensive looping and tufted morphology was observed (Fig. 1, F and G, insert).
All 45 biocytin-filled GnRH neurons exhibited somal and dendritic spines with densities ranging up to a maximum of 1 spine/μm plasma membrane. The majority of spines observed on GnRH neurons were simple sessile spines, synaptic protrusions without a neck constriction; however, thin pedunculated spines and mushroom-like spines (30) were also observed (Fig. 2, C and D). Approximately half of the GnRH neurons (23 of 45) had long, thin filopodia extending from their cell bodies and proximal dendrites (Fig. 2D). Although some neurons had up to three filopodia, on average, GnRH neurons displayed less than one filopodium (0.73 ± 0.13) per neuron.
FIG. 2. High-power confocal images of spines and filopodia on biocytin-filled GnRH neurons. A and B, Note the large numbers of spines in the two dendrites. C and D, Three-dimensional render-ing of confocal images of two biocytin-filled GnRH cell bodies showing spines and filopodia. Black arrows, Examples of mushroom-shaped spines; white arrows, simple spines. Two filopodia are seen in D and marked with arrowheads. Scale bars represent 5 μm in all plates.
The numbers of spines detected on GnRH soma were observed to vary considerably ranging from one to 41 spine-like processes, with a mean of 13 ± 10 spines/soma. An analysis of spine density along the length of the primary dendrite revealed that spines were quantitatively most abundant within the first 50 μm, with an average of 24 ± 2 spines in the first 50 μm of the dendrite (Fig. 3A). The secondary dendrite exhibited a mean spine density of 0.5 ± 0.1 spines/μm within the first 50 μm, compared with 0.4 ± 0.1 spines/μm for the same regions of the designated primary dendrite (n = 10). Although there was a trend for a gradual reduction in spine numbers with increasing distance from the cell body, spines were still observed even on the most distal elements of the primary dendrite (Figs. 1G and 3A). On an individual neuron basis, the density of dendritic spines ranged from 0.07–1.04 spines/μm with a mean of 0.4 spines/μm of dendritic membrane.
FIG. 3. A, Histogram showing the mean + SEM number of spines detected upon the cell somata and dendrites of biocytin-filled mature GnRH neurons (n = 45). Dendritic spine numbers are shown per 50-μm segment running distally from the cell body. B, Histogram showing the mean + SEM numbers of spines detected upon the cell soma and dendrites of biocytin-filled adult male (white box, n = 22) and female (black box, n = 23) GnRH neurons.
Because a previous electron microscopic study had demonstrated that the synaptic input onto the cell body of male GnRH neurons was only half that of female GnRH neurons in the rat (31), we examined whether sex differences were also apparent in spine densities. No significant differences were found in spine number between male (n = 22) and diestrus female (n = 23) GnRH neurons in any neuronal segment (Fig. 3B).
In an effort to reveal any trends in spine density among the GnRH neurons, spine numbers were compared between neurons classified by neuronal morphology, anatomical location, and dendritic orientation. Spine numbers were not significantly different among GnRH neurons classified as classically unipolar, bipolar, or irregular in morphological profile (Fig. 4A). Likewise, no significant differences in spine number were found between GnRH neurons located in the POA/DBB compared with the MS (Fig. 4B). However, neurons with dendrites extending horizontally or diagonally had significantly more spines on their proximal dendrites than those of cells with vertically orientated dendrites (Fig. 4C). Although not significantly different (P = 0.0514), a similar trend of increased spine density in horizontal/diagonal dendrites was also seen at greater dendritic lengths (Fig. 4C).
FIG. 4. A, Histogram showing the mean + SEM number of spines detected upon the soma and first 450 μm of primary dendrite of GnRH neurons exhibiting unipolar (black bar, n = 6), bipolar (gray bar, n = 20), and irregular (white bar, n = 7) morphologies. B, Histogram showing the mean + SEM number of spines detected upon the soma and first 450 μm of primary dendrite of POA/DBB (black bar, n = 32) and MS (white bar, n = 11) GnRH neurons. C, Histogram showing the mean + SEM number of spines detected upon the soma and first 150 μm of primary dendrite exhibiting a diagonal/horizontal (white bars, n = 13) or vertical (black bars, n = 32) orientation. *, P < 0.05.
Discussion
The present study has used a dye-filling approach in GnRH-GFP transgenic mice to reevaluate GnRH neuron morphology. We report here that the primary dendrites of GnRH neurons are much longer than had been recognized previously, and that GnRH neuron dendrites and somata exhibit numerous spine-like elements. Ultrastructural investigations will be required to establish with certainty that these spine-like structures represent points of synaptic input to GnRH neurons. However, elsewhere in the nervous system, spines such as those visualized here have always been found to be associated with excitatory synaptic transmission (29). Together, these results extend prior observations of subpopulations of spiny GnRH neurons using immunocytochemistry (7, 8, 9, 10, 11, 12, 13, 14, 15, 18) and suggest that the density of synaptic input to GnRH neurons has been underestimated.
Given that the average dendritic spine density of a GnRH neuron was 0.4 spines/μm of membrane, and that the mean primary dendrite length was 350 μm, we estimate that an average GnRH neuron has approximately 160 spines. However, this estimate is very conservative. For one, we have only been able to evaluate spine density in a two-dimensional manner and all spines perpendicular to the plane of the slice have not been counted. Secondly, we have only performed a detailed analysis on the primary dendrite and so bipolar and irregular GnRH neurons will have more spines per cell. Furthermore, the actual length of the primary dendrite in over half of the GnRH neurons examined was longer than the recorded value because it exited the brain slice before terminating. Three GnRH neurons were found to have primary dendrites more than 1000 μm in length. For these reasons, we believe that individual GnRH neurons are, in reality, very likely to receive hundreds of individual excitatory inputs, and that this number will be determined principally by their dendritic architecture. Given that this estimate does not include the substantial GABAergic inputs, which appear to greatly out number excitatory inputs in GnRH neurons (22, 32), or certain neuropeptidergic and aminergic afferents to the GnRH neuron, the total number of synapses upon a GnRH neuron could easily be within the range of 500-1000. This indicates that GnRH neurons are not a neuronal phenotype with relatively sparse synaptic input. Indeed, our conservative estimate of 0.4 dendritic spines/μm for GnRH neurons shows that these cells receive more excitatory inputs than do unidentified neurons in the preoptic area (0.1 dendritic spines/μm) (33), ventromedial nucleus (0.2 dendritic spines/μm) (34), and arcuate nucleus (0.25 dendritic spines/μm) (35) of the rat hypothalamus.
Although providing a different view of GnRH neuron morphology, the present findings are not necessarily discordant with prior ultrastructural studies. When comparing results at the level of the cell body, it is interesting to note that we found an average of 13 spines/cell body compared with two to 12 synapses detected in serial ultrastructural reconstructions of GnRH cell bodies in the rat and monkey (19). Admittedly, spine numbers will only represent a subset of all synaptic inputs to the GnRH cell bodies, but, given this, the values are not substantially different. A recent electron microscopic study reported that monkey GnRH cell bodies and proximal dendrites had a synapse density of 0.1 μm/μm plasma membrane (32). However, there appear to be clear differences between the ultrastructural and cell-filling data in terms of dendritic morphology, with many more synaptic inputs suggested by the latter method. One reason for this difference may be that major species differences exist between GnRH neurons in the mouse compared with rats, sheep, and monkeys. Another possibility may be that spine numbers are altered in GnRH neurons as a result of brain slicing. A 40% increase in spine numbers on hippocampal neurons has been reported 1 h after slice preparation (36). However, even if this was to be the case, our data would still suggest that GnRH neurons receive substantially more inputs than has been recognized previously. We believe that the most likely reason for the discrepancy, however, lies in the ability of the cell-filling approach to reveal the whole of the dendritic tree including elements of the GnRH neuron that do not contain GnRH peptide. Although dendritic spines have been observed by electron microscopy on GnRH neurons (11, 17), it is likely very difficult to visualize structures protruding up to 5 μm away from the cell membrane with this approach.
We also find evidence here for filopodium-like structures extending from around half of all GnRH neurons. Similar structures have been observed in ovine GnRH neurons (13). Filapodia are thought to be involved in synaptogenesis by locating and guiding appropriate axons back to the dendrite (37, 38, 39). As such, filopodia are highly motile and can extend and retract within minutes (40). However, filopodia are thought to be rare in the mature brain (29), and their observation here on mature GnRH neurons is intriguing and may suggest ongoing synaptogenesis in these cells. In addition, it is interesting to note that we found moderate numbers of spine-like structures on the soma themselves. With some exceptions (41), somatic spines are not common in the adult nervous system.
With the exception of the orientation of the primary dendrite, we failed to find any topographical or morphological parameter that correlated with the range of spine densities observed on GnRH neurons. The nature and functional significance of horizontally orientated dendrites having more excitatory inputs than vertical dendrites can only be speculated upon. It has been suggested that dendritic orientation reflects, in part, trophic influences of excitatory afferents during development (42). It is possible that horizontally orientated dendrites experience a different pattern of excitatory input during development and that this then remains in the form of increased spine numbers and synaptic input during adulthood.
We also failed to detect any significant sex differences in spine densities upon GnRH cell bodies and dendrites. This was surprising considering a previous electron microscopic study reported that GnRH neurons in female rats had twice the number of synapses as those found on GnRH neurons in males (31). Nevertheless, an earlier immunocytochemical study had failed to detect sex differences in the number of spiny verses smooth GnRH neurons found in the rat (43). As noted above, spines may be particularly difficult to evaluate with electron microscopy compared with synapses that exist directly upon on the somal plasma membrane. Thus, it remains possible that excitatory inputs to GnRH neurons via spines are not sexually differentiated, whereas nonspinous (e.g. GABAergic) inputs may be more abundant in females. There is some evidence to support this idea as GABAergic inputs to mouse GnRH neurons are modified by masculinizing prenatal androgen exposure in females (44), and sex differences have been observed in GABAA receptor subunit mRNA expression in mouse GnRH neurons (45).
In summary, we report here a previously unsuspected dendritic morphology for the GnRH neuronal phenotype with long primary dendrites that may extend for up to 1500 μm. We also show that all GnRH neurons have the potential to receive substantial numbers of excitatory inputs. Such findings are in good agreement with the recent electrophysiological (27, 46, 47) and receptor-labeling (48, 49) studies that have demonstrated direct glutamatergic neurotransmission in mouse GnRH neurons. The highest density of excitatory input to a GnRH neuron was found within the first 50 μm of proximal dendrite. Together, these findings indicate that, like other final output neurons (50), the GnRH neuron receives many synaptic inputs and, as such, represents a major site of neuronal integration within the GnRH neuronal network.
Acknowledgments
We thank Dr. C. Jasoni for critical appraisal of an earlier draft.
References
Levine JE 2003 Gonadotropin-releasing hormone (GnRH). In: Henry H, Norman A, eds. Encyclopedia of hormones. San Diego: Academic Press; 157–165
Kalra SP 1993 Mandatory neuropeptide-steroid signaling for the preovulatory luteinizing hormone-releasing hormone discharge. Endocr Rev 14:507–538
Herbison AE 1998 Multimodal influence of estrogen upon gonadotropin-releasing hormone neurons. Endocr Rev 19:302–330
Smith MJ, Jennes L 2001 Neural signals that regulate GnRH neurones directly during the oestrous cycle. Reproduction 122:1–10
Petersen SL, Ottem EN, Carpenter CD 2003 Direct and indirect regulation of gonadotropin-releasing hormone neurons by estradiol. Biol Reprod 69:1771–1778
Silverman A, Livne I, Witkin JW 1994 The gonadotrophin-releasing hormone (GnRH), neuronal systems: immunocytochemistry and in situ hybridization. In: Knobil E, Neill JD, eds. The physiology of reproduction. 2nd ed. New York: Raven; 1683–1706
Kozlowski GP, Chu L, Hostetter G, Kerdelhue B 1980 Cellular characteristics of immunolabeled luteinizing hormone releasing hormone (LHRH) neurons. Peptides 1:37–46
Krisch B 1980 Two types of luliberin-immunoreactive perikarya in the preoptic area of the rat. Cell Tissue Res 212:443–455
King JC, Tobet SA, Snavely FL, Arimura AA 1982 LHRH immunopositive cells and their projections to the median eminence and organum vasculosum of the lamina terminalis. J Comp Neurol 209:287–300
Witkin JW, Paden CM, Silverman AJ 1982 The luteinizing hormone-releasing hormone (LHRH) systems in the rat brain. Neuroendocrinology 35:429–438
Jennes L, Stumpf WE, Sheedy ME 1985 Ultrastructural characterization of gonadotropin-releasing hormone (GnRH)-producing neurons. J Comp Neurol 232:534–547
Liposits Z, Setalo G, Flerko B 1984 Application of the silver-gold intensified 3,3'-diaminobenzidine chromogen to the light and electron microscopic detection of the luteinizing hormone-releasing hormone system of the rat brain. Neuroscience 13:513–525
Lehman MN, Robinson JE, Karsch FJ, Silverman AJ 1986 Immunocytochemical localization of luteinizing hormone-releasing hormone (LHRH) pathways in the sheep brain during anestrus and the mid-luteal phase of the estrous cycle. J Comp Neurol 244:19–35
Wray S, Hoffman G 1986 Postnatal morphological changes in rat LHRH neurons correlated with sexual maturation. Neuroendocrinology 43:93–97
Witkin JW 1996 Effects of ovariectomy on GnRH neuronal morphology in rhesus monkey (Macaca mulatta). J Neuroendocrinol 8:601–604
Witkin JW, Ferin M, Popilskis SJ, Silverman A 1991 Effects of gonadal steroids on the ultrastructure of GnRH neurons in the rhesus monkey: synaptic input and glial apposition. Endocrinology 129:1083–1092
Witkin JW, Silverman AJ 1985 Synaptology of luteinizing hormone-releasing hormone neurons in rat preoptic area. Peptides 6:263–271
Lehman MN, Karsch FJ, Silverman AJ 1988 Ultrastructure and synaptic organization of luteinizing hormone-releasing hormone (LHRH) neurons in the anestrous ewe. J Comp Neurol 273:447–458
Witkin JW, O’Sullivan H, Silverman AJ 1995 Novel associations among gonadotropin-releasing hormone neurons. Endocrinology 136:4323–4330
Gulyas AI, Megias M, Emri Z, Freund TF 1999 Total number and ratio of excitatory and inhibitory synapses converging onto single interneurons of different types in the CA1 area of the rat hippocampus. J Neurosci 19:10082–10097
Matyas F, Freund TF, Gulyas AI 2004 Convergence of excitatory and inhibitory inputs onto CCK-containing basket cells in the CA1 area of the rat hippocampus. Eur J Neurosci 19:1243–1256
Herbison AE 2003 GnRH neuron. In: Henry H, Norman A, eds. Encyclopedia of hormones. San Diego: Academic Press; 171–177
Spergel DJ, Kruth U, Shimshek DR, Sprengel R, Seeburg PH 2001 Using reporter genes to label selected neuronal populations in transgenic mice for gene promoter, anatomical, and physiological studies. Prog Neurobiol 63:673–686
Herbison AE, Pape JR, Simonian SX, Skynner MJ, Sim JA 2001 Molecular and cellular properties of GnRH neurons revealed through transgenics in the mouse. Mol Cell Endocrinol 185:185–194
Sorra KE, Harris KM 2000 Overview on the structure, composition, function, development, and plasticity of hippocampal dendritic spines. Hippocampus 10:501–511
Harris KM, Kater SB 1994 Dendritic spines: cellular specializations imparting both stability and flexibility to synaptic function. Annu Rev Neurosci 17:341–371
Spergel DJ, Kruth U, Hanley DF, Sprengel R, Seeburg PH 1999 GABA- and glutamate-activated channels in green fluorescent protein-tagged gonadotropin-releasing hormone neurone in transgenic mice. J Neurosci 19:2037–2050
Han SK, Todman MG, Herbison AE 2004 Endogenous GABA release inhibits the firing of adult gonadotropin-releasing hormone neurons. Endocrinology 145:495–499
Fiala JC, Spacek J, Harris KM 2002 Dendritic spine pathology: cause or consequence of neurological disorders? Brain Res Brain Res Rev 39:29–54
Harris KM, Jensen FE, Tsao B 1992 Three-dimensional structure of dendritic spines and synapses in rat hippocampus (CA1) at postnatal day 15 and adult ages: implications for the maturation of synaptic physiology and long-term potentiation. J Neurosci 12:2685–2705
Chen WP, Witkin JW, Silverman AJ 1990 Sexual dimorphism in the synaptic input to gonadotropin releasing hormone neurons. Endocrinology 126:695–702
Zsarnovszky A, Horvath TL, Garcia-Segura LM, Horvath B, Naftolin F 2001 Oestrogen-induced changes in the synaptology of the monkey (Cercopithecus aethiops) arcuate nucleus during gonadotropin feedback. J Neuroendocrinol 13:22–28
Amateau SK, McCarthy MM 2002 A novel mechanism of dendritic spine plasticity involving estradiol induction of prostaglandin-E2. J Neurosci 22:8586–8596
Calizo LH, Flanagan-Cato LM 2000 Estrogen selectively regulates spine density within the dendritic arbor of rat ventromedial hypothalamic neurons. J Neurosci 20:1589–1596
Mong JA, Glaser E, McCarthy MM 1999 Gonadal steroids promote glial differentiation and alter neuronal morphology in the developing hypothalamus in a regionally specific manner. J Neurosci 19:1464–1472
Kirov SA, Sorra KE, Harris KM 1999 Slices have more synapses than perfusion-fixed hippocampus from both young and mature rats. J Neurosci 19:2876–2886
Fiala JC, Feinberg M, Popov V, Harris KM 1998 Synaptogenesis via dendritic filopodia in developing hippocampal area CA1. J Neurosci 18:8900–8911
Wong WT, Wong RO 2001 Changing specificity of neurotransmitter regulation of rapid dendritic remodeling during synaptogenesis. Nat Neurosci 4:351–352
Maletic-Savatic M, Malinow R, Svoboda K 1999 Rapid dendritic morphogenesis in CA1 hippocampal dendrites induced by synaptic activity. Science 283:1923–1927
Dailey ME, Smith SJ 1996 The dynamics of dendritic structure in developing hippocampal slices. J Neurosci 16:2983–2994
Jakab RL, Leranth C 1990 Somatospiny neurons in the rat lateral septal area are synaptic targets of hippocamposeptal fibers: a combined EM/Golgi and degeneration study. Synapse 6:10–22
Wu GY, Zou DJ, Rajan I, Cline H 1999 Dendritic dynamics in vivo change during neuronal maturation. J Neurosci 19:4472–4483
Wray S, Gainer H 1987 Effect of neonatal gonadectomy on the postnatal development of LHRH cell subtypes in male and female rats. Neuroendocrinology 45:413–419
Sullivan SD, Moenter SM 2004 Prenatal androgens alter GABAergic drive to gonadotropin-releasing hormone neurons: implications for a common fertility disorder. Proc Natl Acad Sci USA101:7129–7134
Pape JR, Skynner MJ, Sim JA, Herbison AE 2001 Profiling -aminobutyric acid (GABA(A)) receptor subunit mRNA expression in postnatal gonadotropin-releasing hormone (GnRH) neurons of the male mouse with single cell RT-PCR. Neuroendocrinology 74:300–308
Kuehl-Kovarik MC, Pouliot WA, Halterman GL, Handa RJ, Dudek FE, Partin KM 2002 Episodic bursting activity and response to excitatory amino acids in acutely dissociated gonadotropin-releasing hormone neurons genetically targeted with green fluorescent protein. J Neurosci 22:2313–2322
Suter KJ 2004 Control of firing by small (S)--amino-3-hydroxy-5-methyl-isoxazolepropionic acid-like inputs in hypothalamic gonadotropin releasing-hormone (GnRH) neurons. Neuroscience 128:443–450
Gore AC 2002 Gonadotropin-releasing hormone (GnRH) neurons: gene expression and neuroanatomical studies. Prog Brain Res 141:193–208
Jennes L, Lin W, Lakhlani S 2002 Glutamatergic regulation of gonadotropin-releasing hormone neurons. Prog Brain Res 141:183–192
Napper RM, Harvey RJ 1988 Quantitative study of the Purkinje cell dendritic spines in the rat cerebellum. J Comp Neurol 274:158–167(Rebecca E. Campbell, Seon)