Original Paper Neuroendocrinology 1992;56:18-24
Departments of Neurology and Neurosurgery, Pediatrics and Anatomy, McGill University; Tlie McGill University-Montreal Children’s Hospital Research Institute and Montreal Neurological Institute, Montreal, Canada
Key Words Somatostatin Growth hormone-releasing factor Colocalization Arcuate nucleus Radioautography Immunohistochemistry Growth hormone
Colocalization of Somatostatin Receptors and Growth HormoneReleasing Factor Immunoreactivity in Neurons of the Rat Arcuate Nucleus
Abstract Recent studies from our group have demonstrated an association of [l25I]labeled somatostatin (SRIF)-binding sites with a subpopulation of arcuate (ARC) neurons. The distribution of these cells was similar to that of growth hormone-releasing factor (GRF)-immunoreactive neurons, which led us to propose that at least some SRIF receptors may be directly localized to GRFcontaining cells. To test this hypothesis, we have visualized radiolabeled SRIF-binding sites and GRF immunoreactivity (ir) in adjacent sections of the hypothalamus, by combined radioautography and immunohistochemistry. Adult male rats were sacrificed by decapitation and the brains were rapidly frozen and serially sectioned on a cryostat. Fifteen pairs of adjacent 6-umthick sections, taken at 100-urn intervals through the rostrocaudal extent of the ARC nucleus, were alternately processed for [l25l]-SRIF radioautography and GRF immunohistochemistry. GRF-ir and [125I]-SRIF-labeled cells were mapped at each level and quantified with the aid of a camera lucida. The maps were subsequently superimposed to determine the extent of [l25I]-SRIF/ GRF-ir colocalization. GRF-ir perikarya [13.2 ± 4.4 (mean ± SE) cells per section] were mainly localized in the ventrolateral portion of the ARC nucleus and predominated within the caudal-most tier. [125I]-SRIF-labeled cells (35.6 ± 6.5 cells per section) were more numerous, more evenly distributed, and extended further rostrally and caudally than GRF-ir cells. Superimposition of the camera lucida maps indicated that, overall, 33.5 ± 10.8% of the GRF-ir cells were labeled with [l25I]-SRIF in adjacent sections. Colabeled cells con centrated, rostrocaudally, in the ventrolateral portion of the ARC nucleus; more dorsally placed cells and those of the lateral hypothalamic area were rarely colabeled. Moreover, GRF-ir cells located in extra-ARC regions did not show [l2:T]-SRIF binding. These results demonstrate that a subpopulation of GRF-ir nerve cell bodies detected within the ARC nucleus possesses SRIFbinding sites. The finding of SRIF receptors on GRF-containing ARC neu rons provides strong anatomic evidence to support the physiological concept of a central SRIF-mediated influence on pathways associated with GRF. Such intrahypothalamic interactions between these two neuropeptides may be a vital component in the generation of the ultradian rhythm of GH secretion.
Received: June 18. 1991 Accepted after revision: October 4. 1991
Dr. Gloria S. Tannenbaum Neuropeptide Physiology Laboratory McGill University-Montreal Children’s Hospital Research Institute 2300 Tupper Street. Montreal, Quc. H3H l P3 (Canada)
© 1992 S. Karger AG. Basel 0028-3835/92/0561-0018 S 2.75/0
Downloaded by: Imperial College, School of Medicine, Wellcome Libr. 155.198.30.43 - 1/19/2019 7:53:33 AM
G. Francis McCarthy Alain Beaudet Gloria Shaffer Tannenbaum
Materials and Methods Animals and Experimental Procedure Adult male Sprague-Dawley rats (210-240 g) were obtained from Charles River Canada (St. Constant, Canada) and maintained
on a 12-hour light/12-hour dark cycle (lights on: 06.00 h) in a tem perature (22 ± 1 °C)- and humidity-controlled room. The rats were anesthetized with sodium pentobarbital (Somnotol, 0.12 m l/100 g body weight, i.p.) and administered a single intracerebroventricular injection of colchicine (Sigma Chemical, St. Louis, Mo., USA) into the right lateral ventricle (60 ug/30 pi 0.9% NaCI). Sixty-seven hours later each rat was given a single subcutaneous injection of cysteamine (CSH; Calbiochem, La Jolla, Calif., USA; 300 nig/kg in I ml distilled H :0, pH adjusted to 7.0 with I N NaOH). CSH was used as a tool to optimize detection of GRF-ir within ARC neurons [14]. Five hours after the CSH injection, the animals were sacrificed by decapitation and the brains removed from the skulls, frozen by immersion in liquid isopentane at -40 °C for 30 min and stored at -80 °C until use. The brains were serially sectioned (6 pm thick) on a cryostat (Frigocut, Reichert-Jung) at -18 to -20 °C, from the retrochiasmatic area, rostrally, to the mammillary nucleus, caudally. Fifteen sets of 4 adjacent sections were collected at intervals of 100 pm and mounted on gelatin-subbed slides; sections I and 3 were immediately pro cessed for GRF immunohistochemistry whereas adjacent sections 2 and 4 were stored in airtight containers at -80 °C until processed for SRIF radioautography. GRF Immunohistochemistry' For immunohistochemical detection of GRF-containing neu rons, sections were fixed by immersion in an ice-cold solution of 2% carbodiimide (Sigma Chemical) in HEPES buffer (100 mM, pH 6.0) followed by an equally cold mixture of 4% paraformalde hyde and 15% (vol/vol) saturated picric acid in HEPES buffer (10 niAF, pH 7.2). They were immediately processed according to the indirect peroxidase-antiperoxidase immunohistochemical method of Sternberger [26]. Briefly, the sections were sequentially incubated with: (l) 0.1 M Tris-buffered saline (pH 7.6) containing 0.3% H :0: and 0.1% Na azide (30 min); (2) normal rabbit serum diluted 1:30 (30 min); (3) sheep anti-rGRF(l-29)NH; antibody diluted 1:2,000 in the presence of 0.2% Triton-X (overnight); (4) rabbit antisheep immunoglobulins G (Miles) diluted 1:50 (30 min), and (5) goat peroxidase-antiperoxidase (Dako, Missisauga, Can ada) diluted 1:50 (30 min). All rinses and dilutions were carried out at room temperature in 0.1 M Tris-buffered saline containing 1% normal rabbit serum. The immunoreaction was visualized by treat ing the sections with a 0.05% 3,3'-diaminobenzidine tetrahydrochloride (Sigma Chemical) solution containing 0.01% H2O 2, for 6 min, at room temperature. The reaction product was then silverintensified by the method of Galiyas et al. [27]. The GRF antiserum used has been previously characterized [25, 28]. Immunoreacted sections were mounted, cleared, coverslipped and examined with a Leitz Dialux 20 microscope. Immunopositive cells were mapped at a magnification of 25 x with a camera lucida and counted, I sec tion per level, over 15 different rostrocaudal planes. The mean diameter of GRF-ir neurons was determined with the aid of a Bio quant II image analysis system (R & M Biometrics, Nashville, Tenn., USA). f llflj-SRIF Radioautography Tyr°-D-Trp*-SRIF-14 (Code 8007, Peninsula. Belmont, Calif., USA) was iodinated in the presence of chloramine-T. The iodination mixture conisted of 10 pg SRIF dissolved in 20 pi O.OIA' HCI, 50 ul 0.5 M phosphate buffer (pH 7.5), 1.5 mCi Na[l25i] (IMS-30, Amersham, Oakville, Canada) and 10 ug chloramine-T in distilled
19
Downloaded by: Imperial College, School of Medicine, Wellcome Libr. 155.198.30.43 - 1/19/2019 7:53:33 AM
The pulsatile pattern of growth hormone (GH) secre tion from the pituitary gland is the net result of a delicate interplay between two hypothalamic peptides, the stimu latory GH-releasing hormone, GRF [ 1-3], and the inhibi tory hormone, somatostatin (SRIF) [4], In the male rat, these two neurohormones are released in reciprocal 3- to 4-hour cycles from the median eminence into the hypo physeal portal circulation where they act on the anterior pituitary somatotrophs to generate the ultradian rhythm of GH secretion [5, 6]. A growing body of physiological evidence indicates that, in addition to exerting their characteristic actions at the level of the anterior pituitary, SRIF and GRF may in teract within the central nervous system to modulate GH release [6-14], While these studies, in general, provide support for the concept of a central influence of SRIF on pathways associated with the secretion of GRF, the mech anism and morphological substrate for this action are not yet known. Recent anatomic data suggest that SRIF may directly affect the GRF-containing cells of the arcuate (ARC) nu cleus, the source of most GRF projections to the median eminence [15-17], First, the ARC nucleus contains both a subpopulation of SRIF-immunoreactive (ir) neurons and a dense network of SRIF-positive axon terminals [18-20], some of which appear to originate from within the ARC nucleus and others from the periventricular/preoptic hy pothalamic area [18, 20, 21]. Second, these terminals were shown by double-labeling immunocytochemistry to directly contact GRF-ir cell bodies [22-24], Finally, we have recently demonstrated, using high-resolution radioautography, a selective association of [125I]-labeled SRIF-binding sites with a subpopulation of ARC neu rons, the distribution of which was very similar to that of GRF-ir neurons detected within the same region in a sep arate group of rats [25], These latter results led us to pro pose that some SRIF receptors might be directly localized on the surface or within GRF-containing cells. To test this hypothesis, in the present study, we have visualized and mapped radiolabeled SRIF-binding sites and GRF-ir neurons in adjacent 6-p.m-thick sections throughout the rostrocaudal extent of the ARC nucleus from the same animals.
a Fig. 1. Distribution of GRF-ir (a) and [l35I]-SRIF-labeled neurons (b) in the rat hypothalamic ARC nucleus (Arc), a GRF-ir perikarya (black arrows) are apparent in the ventrolateral component of the nucleus, where, in dark field (b) a group of [l25I]-SRIF-labeled neurons, detected as silver grain aggregates (white arrows), are also local ized. ME = Median eminence; III = third ventricle. Scale bar = 150 pm.
Assessment o f Colabeling The extent of cellular colocalization of [125I]-SRI F-binding sites to GRF-ir neurons was determined by superimposing serial maps of GRF-ir cells on the adjacent maps of radiolabeled SRIF cells. Co labeled neurons in the adjacent section were identified as those lying within one cell diameter ( w 10 urn) from the edge of the GRFir cell. The number of GRF-ir cells identified as SRIF-receptorbearing was documented for each of the 15 rostrocaudal planes in both intra- and extra-ARC regions and the results were expressed as a percentage (mean ± SE) of total GRF-ir cells detected.
20
To validate this approach, a control experiment was carried out in which a single marker (GRF antiserum) was applied to adjacent sections. Fifteen pairs of sections were cut at intervals of 100 pm over the rostrocaudal extent of the ARC nucleus and immediately processed for GRF immunohistochemistry, as described above. Each pair of sections was examined simultaneously with two Leitz Dialux 20 microscopes fitted with a comparison bridge. Between 30 and 50% of detectable GRF-ir ARC neurons were present in the adjacent section.
Results GRF Immunostaining As previously described [25], GRF-ir cells were mainly localized in the ventrolateral portion of the ARC nucleus and predominated within the caudal-most tier (fig. la). These labeled cells were round to fusiform in shape, dis played 2-3 cellular processes and measured 10.1 ± 1.4 pm in mean diameter (fig. 2b). On average, 13.2 ± 4.4 cells were detected per section. Moreover, an average of 4.0 ± 2.2 were detected in extra-ARC regions, most around the ventromedial nucleus of the hypothalamus (VMH) and in the lateral hypothalamic area (LH). f l25IJ-SRIF Radioautography Dark-field examination of light-microscopic radioau tograms of sections incubated with [1251]-SR1F revealed discrete silver grain clusters distributed throughout the ventrolateral portion of the ARC nucleus. These labeled foci stood out against a moderate background reaction (fig. lb) and with, bright-field optics were found to be in
McCarthy/Beaudet/Tannenbaum
Colocalization of SRIF Receptors and GRF Immunoreactivity
Downloaded by: Imperial College, School of Medicine, Wellcome Libr. 155.198.30.43 - 1/19/2019 7:53:33 AM
H :0 at a concentration of 1 mg/ml at room temperature. The re action was stopped after 15 s by addition of 10% BSA in 0.05 M phosphate buffer, pH 7.5 (500 pi). [1251J-SR1F was purified by gel filtration on a Sephadex G-25 (fine) column in 0.087 N acetic acid with 0.1 % BSA. The specific activity of the iodinated ligand was 165 Ci/mmol. For radioautographic visualization of [1251]-SRIF-binding sites, the adjacent section of each set of sections was: (1) preincubated (2 x 15 min at room temperature) in 0.17 A/Tris-HCI buffer (pH 7.4) containing 0.2% BSA (supplemented Tris buffer); (2) incubated (45 min at room temperature) in the same supplemented Tris buffer to which 0.25 nM [l25I]-SRIF, 3 x 10'5 M MgCb and 5 x 10'5 M bacitracin (Sigma Chemical) had been added; additional sections were incubated in the presence of 250 nM nonradioactive Tyr°-DTrps-SR1F-14 for determination of nonspecific binding; (3) tapped dry and rinsed ( 3 x 4 min) in ice-cold baths of supplemented Tris buffer, and (4) fixed immediately in an equally cold solution of 4% glutaraldehyde in 0.05 M phosphate buffer (30 min). After fixation, sections were dehydrated in graded ethanols, de fatted in xylene, rehydrated and coated by dipping in Kodak NTB-2 emulsion diluted 1:1 with distilled water. After 6 weeks of exposure, the radioautograms were developed in Dektol (Kodak), stained with cresyl violet and coverslipped. All [l35I]-SRIF-labeled cells, I section per level, were mapped and quantified as previously de scribed for the GRF series.
Fig. 2. Adjacent-section [l2SI]-SRIF autoradiography (a) and GRF immunocytochemistry (b). Five ['-51]-SRIF-labeled cells are seen to be GRF-iinmunopositive in the adjacent section (arrows). By contrast, two GRF-ir cells are not autoradiographically labeled (arrowheads). Note the greater shrinkage of autoradiographicallylabeled as compared to immunohistochemicallyreacted material. Scale bar = 20 urn.
GRF-ir/f12' IJ-SRIF Colabeling Superimposition of the camera lucida maps of GRF-ir over the [l25I]-SRIF binding maps indicated that, overall, 33.5 ± 10.8% (mean ± SE of 3 animals) of GRF-ir cells within the ARC nucleus were labeled with [1251]-SRIF in the adjacent section (figs. 2, 3). Colabeled cells concen trated, rostrocaudally, in the ventrolateral component of the ARC nucleus; more dorsally placed cells and those of the LH area, were rarely colabeled (fig. 3). Moreover, GRF cells located in extra-ARC regions did not show [I2', I]-SR1F binding. As shown in figure 4, the number of GRF-ir/[,25I]-SRIF colabeled cells did not vary consider ably from the retrochiasmatic area, rostrally (0-300 um),
to the mammillary nucleus, caudally (900-1,200 um). However, at the most caudal component of the ARC nu cleus (1,200-1,500 pm), the number of colabeled cells declined.
Discussion The results of the present study provide the first direct anatomic evidence for an association of high-affinity [l25I]-SRIF-binding sites with GRF-ir neurons in the ARC nucleus of the rat. These results are congruent with those of Bertherat et al. [29], which demonstrate an asso ciation o f[l25I]-SRIF-binding sites with ARC neurons ex pressing GRF mRNA. Taken together, these two sets of experiments provide strong anatomic evidence for the presence of SR1F receptors on a subpopulation of ARC neurons that both contain and express GRF. The combined adjacent-section autoradiographic/immunohistochemical approach used in the present study is a modification of a technique previously developed by us for the localization of high-affinity neurotensin-binding sites to dopaminergic neurons in the midbrain tegmen tum [30] and to vasoactive-intestinal-polypeptide-containing neurons in the suprachiasmatic nucleus of the rat [31]. Admittedly, simultaneous visualization of the two markers on the same tissue sections would have been preferable; however, in initial experiments, [l25I]-SRIF
21
Downloaded by: Imperial College, School of Medicine, Wellcome Libr. 155.198.30.43 - 1/19/2019 7:53:33 AM
direct register with small, round to oval, neuronal perikarya (fig. 2a). An average of 35.6 ± 6.5 SRIF-Iabeled cells were identified per section within the ARC nucleus proper. No labeled cells were present in sections incu bated with an excess of nonradioactive Tyr°-D-Trp8SRIF-14. [l25I]-SRIF-labeIed cells were more numerous, more evenly distributed, and extended further rostrally and caudally than GRF-ir cells. As a whole, however, the distributions of the two populations were similar, particu larly in the ventrolateral aspect of the ARC nucleus. Of the few SRIF-Iabeled cells observed in extra-ARC regions (1.33 ± 0.8), such as in the LH area, none were found in the peri-VMFI position occupied by one third of the GRF-ir cells.
o
• *
Rostro-caudal level (pm)
Fig. 3. Representative superimposed camera lucida drawings of GRF-ir neurons (solid circles), [,25I]-SR1 F-labeled cells (open cir cles) and colabeled GRF-ir/[l25I]-SRIF cells (stars), mapped in each of 4 coronal planes (400, 700, 1,000 and 1,300 um) caudal to the retrochiasmatic area.
22
autoradiographic and GRF immunohistochemical proto cols were found to be mutually exclusive. Nonetheless, the theoretical probability of detecting, and therefore co labeling, the same cell in two adjacent sections is, for neu rons 10 pm in diameter (d) and sections 6 pm in thickness (t), in the order of 45% (p = d /2 t + d), which fits well with our practical observation of 30-50%. It may there fore be assumed that the extent of colocalization reported in the present study is, if anything, underestimated. The monoiodinated probe used here for the visualiza tion of hypothalamic SRIF-binding sites was previously shown to bind with high affinity to the different types of SRIF receptors reported in rat brain [32,33]. Further more, high-resolution autoradiographic data have re cently indicated that SRIF-binding sites associated with
McCarthy/ Beaudet/Tannenbaum
Colocalization of SRIF Receptors and GRF Immunoreactivity
Downloaded by: Imperial College, School of Medicine, Wellcome Libr. 155.198.30.43 - 1/19/2019 7:53:33 AM
Fig. 4. Number (a) of GRF-ir neurons (■ ) and [l25I]-SRIFlabeled cells (□ ); and percent (b) GRF-ir cells colabeled with [ i25I]-SRI F in 5 coronal planes (0-1,500) of the ARC nucleus caudal to the retrochiasmatic area. Mean ± SE of 3 animals; 3 sections per plane.
20, 21]. Furthermore, this region is also the one where, at the ultrastructural level, synaptic junctions between SRIF-ir axon terminals and GRF-ir cell bodies have been identified [22-24], These findings suggest that at least part of the [l25I]-SRIF-binding sites found here to be as sociated with GRF-ir cells are present on the surface of, as opposed to (or in addition to) inside, GRF-producing cells. It should be recalled, however, that an important contingent of non-SRIF-ir terminals, including some identified as containing substance P or enkephalin-8, were also found to synapse upon ARC GRF-ir cells [22]; this is in keeping with our observation that only a subpop ulation of GRF neurons is endowed with SRIF receptors. In summary, the results reported here demonstrate that a subpopulation of GRF-ir nerve cell bodies detected within the ARC nucleus of the rat hypothalamus pos sesses SRIF-binding sites. The finding of SRIF receptors on GRF-containing ARC neurons provides a morpho logical substrate for the concept of direct 'cross-talk' be tween SRI F and GRF neuronal systems within the central nervous system [6-14], Such intrahypothalamic interac tions between these two neuropeptides may be a vital component in the generation and maintenance of the ultradian rhythm of GH secretion.
Acknowledgments We thank Dr. Filoteo Pasquini for technical advice and Julie Temko for expert secretarial assistance. This work was supported by grants MT-6837 (to G.S.T.) and MF-7386 (to A.B.) from the Medi cal Research Council of Canada and grant 911437-102 (to G.S.T. and A.B.) from the Fonds de la Recherche en Santé du Québec. G.F.M. was the recipient of a Postdoctoral Research Fellowship from the McGill University-Montreal Children's Hospital Research Institute. A.B. is a Scientist Awardee from the Medical Research Council of Canada. G.S.T. is a ‘Chercheur-Boursier de Mérite Exceptionnel' of the Fonds de la Recherche en Santé du Québec.
23
Downloaded by: Imperial College, School of Medicine, Wellcome Libr. 155.198.30.43 - 1/19/2019 7:53:33 AM
ARC neurons had the same pharmacological characteris tics as the ones present in the rest of the brain [34], The density of these ARC SRIF-binding sites was found to be unaffected by the type of colchicine pretreatment used here for improved immunohistochemical visualization of GRF-ir neurons [our unpubl. observations]; it was, how ever, shown to be markedly increased in animals pre treated with the drug CSH, administered at the same dose and time interval as in the present study [35]. Interest ingly, this increase appears to mainly affect the density of binding sites per cell as opposed to the number of reac tive cells, suggesting that most, if not all, of the cells bear ing SRIF receptors are being detected under the present pharmacological conditions. In keeping with this inter pretation, the number of [l25I]-SRIF-labeled cells de tected here per section of rat hypothalamus is consistent with that reported earlier in non-colchicine-, non-CSHtreated rats [25]. In contrast, the number of GRF-ir cells detected in the adjacent series was clearly inferior to that previously re ported by us in animals treated with either colchicine [25] or CSH [14] alone, and fixed by intravascular perfusion of aldehydes. This overall reduction in the sensitivity of im munodetection is likely due to both the reduced thickness of the sections and their fixation by immersion in the present study. Immersion fixation of frozen tissue sec tions has indeed been shown to provide poorer retention of peptide antigenicity than that obtained after vascular perfusion [our unpubl. observations]. It is, therefore, likely that only a fraction of ARC GRF-producing cells are being detected under the present experimental condi tions. The quantitative data provided here should there fore be considered more relative than absolute. In parti cular, they do not permit us to state whether or not [1251]-SRIF receptors are also associated with non-GRF neurons; indeed, [125I]-SRIF-labeled cells found to be non-GRF-ir in the adjacent series might have corre sponded to cells that were below threshold of immuno histochemical detection. What can be safely concluded is that a subpopulation of GRF neurons does not bear SRIF receptors, since a number of GRF-positive cells that were recognized by Nissl staining as being present in the adja cent section were clearly free of autoradiographic label. Most conspicuous in this category were GRF-ir cells de tected in the surround of the VMH nucleus which were systematically label-free in the adjacent series. It is interesting to note that most of the GRF-ir/ [l25I]-SRIF colabeled cells were detected in the basolateral region of the ARC nucleus where a dense fiber net work of SRIF-ir terminals has previously been traced [18,
1 Rivier J, Spiess J, Thorner M, Vale W: Charac terization of a growth hormone-releasing fac tor from a human pancreatic islet tumor. Na ture 1982;300:276-279. 2 Guillemin R. Brazeau P, BOhlen P, Esch F, Ling N, Wehrenberg WB: Growth hormonereleasing factor from a human pancreatic tumor that caused acromegaly. Science 1982; 218:585-587. 3 Spiess J, Rivier J, Vale W: Characterization of rat hypothalamic growth hormone-releasing factor. Nature 1983;303:532-535. 4 Brazeau P, Vale W. Burgus R, Ling N, Butcher M, Rivier J, Guillemin R: Hypotha lamic polypeptide that inhibits the secretion of immunoreactive pituitary growth hormone. Science 1973;179:77-79. 5 Tannenbaum GS, Ling N: The interrelation ship of growth hormone (GH)-releasing factor and somatostatin in generation of the ultradian rhythm of GH secretion. Endocrinology 1984;115:1952-1957. 6 Plotsky PM, Vale W: Patterns of growth hor mone-releasing factor and somatostatin secre tion into hypophysial portal circulation of the rat. Science 1985;230:461-463. 7 Abe H, Kato Y, Iwasaki Y, Chihara K, Imura H: Central efTect of somatostatin on the secre tion of growth hormone in the anesthetized rat. Proc Soc Exp Biol Med 1978; 159:346349. 8 Lumpkin MD, Negro-Vilar A, McCann SM: Paradoxical elevation of growth hormone by intraventricular somatostatin: Possible ultra short-loop feedback. Science 1981 ;211: 1072-1074. 9 Thomas CR, Groot K, Arimura A: Antiserum to rat growth hormone (GH)-releasing factor suppresses but does not abolish antisomato statin-induced GH release in the rat. Endocri nology 1985;116:2174-2178. 10 Tannenbaum GS, Patel YC: On the fate of centrally administerated somatostatin in the rat: Massive hypersomatostatinemia resulting from leakage into the peripheral circulation has effects on growth hormone secretion and glucoregulation. Endocrinology 1986; 118: 2137-2143. 11 Murakami Y, Kato Y, Kabayama Y, Inoue T, Koshiyama H, Imura H: Involvement of hy pothalamic growth hormone (GH)-releasing factor in GH secretion induced by intracerebroventricular injection of somatostatin in rats. Endocrinology 1987; 120:31 1-316. 12 Katakami H, Downs TR, Frohman LA: Inhib itory efTect of hypothalamic medial preoptic area somatostatin on growth hormonereleasing factor in the rat. Endocrinology 1988:123:1103-1109. 13 Miki N, Ono M, Shizume K: Withdrawal of endogenous somatostatin induces secretion of growth hormone-releasing factor in rats. J En docrinol 1988;! 17:245-252.
14 Tannenbaum GS, McCarthy GF, Zeitler P. Beaudet A: Cysteamine-induced enhance ment of growth hormone-releasing factor (GRF) immunoreactivity in arcuate neurons: Morphological evidence for putative somatostatin/GRF interactions within hypothala mus. Endocrinology 1990;127:2551-2560. 15 Bloch B, Brazeau P, Ling N, Böhlen P, Esch F, Wehrenberg WH, Benoit R. Bloom F, Guil lemin R: Immunohistochemical detection of growth hormone releasing factor in brain. Na ture 1983;301:607-608. 16 Merchenthaler I, Vigh S, Schally AV, Petrusz P: Immunocytochemical localization of growth hormone-releasing factor in the rat hy pothalamus. Endocrinology 1984; 114:10821085. 17 Sawchenko PE, Swanson LW, Rivier J, Vale WW: The distribution of growth hormone-re leasing factor (GRF) immunoreactivity in the central nervous system of the rat: An immuno histochemical study using antisera directed against rat hypothalamic GRF. J Comp Neu rol 1985;237:100-115. 18 Kawano H, Daikoku S, Saito S: Immunohis tochemical studies of intrahypothalamic so matostatin-containing neurons in rat. Brain Res 1982;242:227-232. 19 Ohtsuka M, Hisano S, Daikoku S: Electronmicroscopic study of somatostatin-containing neurons in rat arcuate nucleus with special reference to neuronal regulation. Brain Res 1983;263:191-199. 20 Willoughby JO, Brogan M, Kapoor R: Hypo thalamic interconnections of somatostatin and growth hormone releasing factor neurons. Neuroendocrinology 1989:50:584-591. 2 1 Makara GB, Palkovits M, Antoni FA, Kiss JZ: Topography of the somatostatin-immunoreactive fibers to the stalk-median eminence of the rat. Neuroendocrinology 1983;37:1-8. 22 Daikoku S, Hisano S, Kawano H, ChikamoriAoyama M, Kagotani Y, Zhang R, Chihara K: Ultrastructural evidence for neuronal regula tion of growth hormone secretion. Neuroen docrinology 1988;47:405-415. 23 Liposits ZS, Merchenthaler I, Pauli WK, Flerko B: Synaptic communication between somatostatinergic axons and growth hor mone-releasing factor (GRF) synthesizing neurons in the arcuate nucleus of the rat. His tochemistry 1988;89:247-252. 24 Horvàth S, Palkovits M, Göres T, Arimura A: Electron microscopic immunocytochemical evidence for the existence of bidirectional syn aptic connections between growth hormone releasing hormone- and somatostatin-con taining neurons in the hypothalamus of the rat. Brain Res 1989;481:8-15.
25 Epclbaum J, Moyse E, Tannenbaum GS. Kordon C, Beaudet A: Combined autoradio graphic and immunohistochemical evidence for an association of somatostatin binding sites with growth hormone-releasing factorcontaining nerve cell bodies in the rat arcuate nucleus. J Neurocndocrinol 1989;1:106-116. 26 Sternberger LA: Immunocytochemistry. New York, Wiley, 1979. 27 Galiyas F. Gores T, Merchenthaler I: High grade intensification of the end-product of the diaminobenzidine reaction for peroxidase his tochemistry. J Histochem Cytochem 1982; 30:183-184. 28 Tannenbaum GS. Lapointe M, Gurd W, Ftnkelstein JA: Mechanisms of impaired growth hormone secretion in genetically obese Zucker rats: Roles of growth hormone-releasing fac tor and somatostatin. Endocrinology 1990; 127:3087-3095. 29 Bertherat J, Doumaud P, Bèrod A, Normand E, Bloch B, Rostène W, Kordon C, Epelbaum J: Growth hormone-releasing hormone-syn thesizing neurons are a subpopulation of so matostatin receptor-labelled cells in the rat ar cuate nucleus: a combined in situ hybridiza tion and receptor light-microscopic radioautographic study. Neuroendocrinology 1992; 56:25-31. 30 Szigethy E, Beaudet A: Correspondence be tween high affinity (lî5Ij-neurotensin binding sites and dopaminergic neurons in the rat sub stantia nigra and ventral tegmental area: A combined radioautographic and immunohis tochemical light microscopic study. J Comp Neurol 1989;279:128-137. 31 François-Bellan AM, Bosler O, Tonon MC, Lin-Tong W, Beaudet A: Association of neu rotensin receptors with VIP-containing neu rons and serotonin-containing axons in the suprachiasmatic nucleus of the rat. Synapse, in press. 32 Krantic S, Martel JC, Weissman D, Quirion R: Radioautographic analysis of somatostatin receptor sub-types in rat hypothalamus. Brain Res 1989;498:267-278. 33 Krantic S, Martel JC, Weissman D, Pujol JF, Quirion R: Quantitative radioautographic study of somatostatin receptors heterogeneity in the rat extrahypothalamic brain. Neurosci ence 1990;39:127-137. 34 Bertherat J, Slama A, Kordon C, Vtdeau C, Epelbaum J: Characterization of pericellular IffyrO DTRP8 somatostatin binding sites in the rat arcuate nucleus by a newly de veloped method: Quantitative high-resolution light microscopic radioautography. Neurosci ence 1991;41:571-579. 35 Bertherat J, Bérod A, Normand E, Bloch B, Rostcne W, Kordon C, Epelbaum J: Somato statin depletion by cysteamine increases so matostatin binding and growth hormone releasing factor messenger ribonucleic acid in the arcuate nucleus. J Neuroendocrino! 1991;3:115-118.
24
McCarthy/Beaudet/Tannenbaum
Colocalization of SRI F Receptors and GRF Immunoreactivity
Downloaded by: Imperial College, School of Medicine, Wellcome Libr. 155.198.30.43 - 1/19/2019 7:53:33 AM
References
Departments of Neurology and Neurosurgery, Pediatrics and Anatomy, McGill University; Tlie McGill University-Montreal Children’s Hospital Research Institute and Montreal Neurological Institute, Montreal, Canada
Key Words Somatostatin Growth hormone-releasing factor Colocalization Arcuate nucleus Radioautography Immunohistochemistry Growth hormone
Colocalization of Somatostatin Receptors and Growth HormoneReleasing Factor Immunoreactivity in Neurons of the Rat Arcuate Nucleus
Abstract Recent studies from our group have demonstrated an association of [l25I]labeled somatostatin (SRIF)-binding sites with a subpopulation of arcuate (ARC) neurons. The distribution of these cells was similar to that of growth hormone-releasing factor (GRF)-immunoreactive neurons, which led us to propose that at least some SRIF receptors may be directly localized to GRFcontaining cells. To test this hypothesis, we have visualized radiolabeled SRIF-binding sites and GRF immunoreactivity (ir) in adjacent sections of the hypothalamus, by combined radioautography and immunohistochemistry. Adult male rats were sacrificed by decapitation and the brains were rapidly frozen and serially sectioned on a cryostat. Fifteen pairs of adjacent 6-umthick sections, taken at 100-urn intervals through the rostrocaudal extent of the ARC nucleus, were alternately processed for [l25l]-SRIF radioautography and GRF immunohistochemistry. GRF-ir and [125I]-SRIF-labeled cells were mapped at each level and quantified with the aid of a camera lucida. The maps were subsequently superimposed to determine the extent of [l25I]-SRIF/ GRF-ir colocalization. GRF-ir perikarya [13.2 ± 4.4 (mean ± SE) cells per section] were mainly localized in the ventrolateral portion of the ARC nucleus and predominated within the caudal-most tier. [125I]-SRIF-labeled cells (35.6 ± 6.5 cells per section) were more numerous, more evenly distributed, and extended further rostrally and caudally than GRF-ir cells. Superimposition of the camera lucida maps indicated that, overall, 33.5 ± 10.8% of the GRF-ir cells were labeled with [l25I]-SRIF in adjacent sections. Colabeled cells con centrated, rostrocaudally, in the ventrolateral portion of the ARC nucleus; more dorsally placed cells and those of the lateral hypothalamic area were rarely colabeled. Moreover, GRF-ir cells located in extra-ARC regions did not show [l2:T]-SRIF binding. These results demonstrate that a subpopulation of GRF-ir nerve cell bodies detected within the ARC nucleus possesses SRIFbinding sites. The finding of SRIF receptors on GRF-containing ARC neu rons provides strong anatomic evidence to support the physiological concept of a central SRIF-mediated influence on pathways associated with GRF. Such intrahypothalamic interactions between these two neuropeptides may be a vital component in the generation of the ultradian rhythm of GH secretion.
Received: June 18. 1991 Accepted after revision: October 4. 1991
Dr. Gloria S. Tannenbaum Neuropeptide Physiology Laboratory McGill University-Montreal Children’s Hospital Research Institute 2300 Tupper Street. Montreal, Quc. H3H l P3 (Canada)
© 1992 S. Karger AG. Basel 0028-3835/92/0561-0018 S 2.75/0
Downloaded by: Imperial College, School of Medicine, Wellcome Libr. 155.198.30.43 - 1/19/2019 7:53:33 AM
G. Francis McCarthy Alain Beaudet Gloria Shaffer Tannenbaum
Materials and Methods Animals and Experimental Procedure Adult male Sprague-Dawley rats (210-240 g) were obtained from Charles River Canada (St. Constant, Canada) and maintained
on a 12-hour light/12-hour dark cycle (lights on: 06.00 h) in a tem perature (22 ± 1 °C)- and humidity-controlled room. The rats were anesthetized with sodium pentobarbital (Somnotol, 0.12 m l/100 g body weight, i.p.) and administered a single intracerebroventricular injection of colchicine (Sigma Chemical, St. Louis, Mo., USA) into the right lateral ventricle (60 ug/30 pi 0.9% NaCI). Sixty-seven hours later each rat was given a single subcutaneous injection of cysteamine (CSH; Calbiochem, La Jolla, Calif., USA; 300 nig/kg in I ml distilled H :0, pH adjusted to 7.0 with I N NaOH). CSH was used as a tool to optimize detection of GRF-ir within ARC neurons [14]. Five hours after the CSH injection, the animals were sacrificed by decapitation and the brains removed from the skulls, frozen by immersion in liquid isopentane at -40 °C for 30 min and stored at -80 °C until use. The brains were serially sectioned (6 pm thick) on a cryostat (Frigocut, Reichert-Jung) at -18 to -20 °C, from the retrochiasmatic area, rostrally, to the mammillary nucleus, caudally. Fifteen sets of 4 adjacent sections were collected at intervals of 100 pm and mounted on gelatin-subbed slides; sections I and 3 were immediately pro cessed for GRF immunohistochemistry whereas adjacent sections 2 and 4 were stored in airtight containers at -80 °C until processed for SRIF radioautography. GRF Immunohistochemistry' For immunohistochemical detection of GRF-containing neu rons, sections were fixed by immersion in an ice-cold solution of 2% carbodiimide (Sigma Chemical) in HEPES buffer (100 mM, pH 6.0) followed by an equally cold mixture of 4% paraformalde hyde and 15% (vol/vol) saturated picric acid in HEPES buffer (10 niAF, pH 7.2). They were immediately processed according to the indirect peroxidase-antiperoxidase immunohistochemical method of Sternberger [26]. Briefly, the sections were sequentially incubated with: (l) 0.1 M Tris-buffered saline (pH 7.6) containing 0.3% H :0: and 0.1% Na azide (30 min); (2) normal rabbit serum diluted 1:30 (30 min); (3) sheep anti-rGRF(l-29)NH; antibody diluted 1:2,000 in the presence of 0.2% Triton-X (overnight); (4) rabbit antisheep immunoglobulins G (Miles) diluted 1:50 (30 min), and (5) goat peroxidase-antiperoxidase (Dako, Missisauga, Can ada) diluted 1:50 (30 min). All rinses and dilutions were carried out at room temperature in 0.1 M Tris-buffered saline containing 1% normal rabbit serum. The immunoreaction was visualized by treat ing the sections with a 0.05% 3,3'-diaminobenzidine tetrahydrochloride (Sigma Chemical) solution containing 0.01% H2O 2, for 6 min, at room temperature. The reaction product was then silverintensified by the method of Galiyas et al. [27]. The GRF antiserum used has been previously characterized [25, 28]. Immunoreacted sections were mounted, cleared, coverslipped and examined with a Leitz Dialux 20 microscope. Immunopositive cells were mapped at a magnification of 25 x with a camera lucida and counted, I sec tion per level, over 15 different rostrocaudal planes. The mean diameter of GRF-ir neurons was determined with the aid of a Bio quant II image analysis system (R & M Biometrics, Nashville, Tenn., USA). f llflj-SRIF Radioautography Tyr°-D-Trp*-SRIF-14 (Code 8007, Peninsula. Belmont, Calif., USA) was iodinated in the presence of chloramine-T. The iodination mixture conisted of 10 pg SRIF dissolved in 20 pi O.OIA' HCI, 50 ul 0.5 M phosphate buffer (pH 7.5), 1.5 mCi Na[l25i] (IMS-30, Amersham, Oakville, Canada) and 10 ug chloramine-T in distilled
19
Downloaded by: Imperial College, School of Medicine, Wellcome Libr. 155.198.30.43 - 1/19/2019 7:53:33 AM
The pulsatile pattern of growth hormone (GH) secre tion from the pituitary gland is the net result of a delicate interplay between two hypothalamic peptides, the stimu latory GH-releasing hormone, GRF [ 1-3], and the inhibi tory hormone, somatostatin (SRIF) [4], In the male rat, these two neurohormones are released in reciprocal 3- to 4-hour cycles from the median eminence into the hypo physeal portal circulation where they act on the anterior pituitary somatotrophs to generate the ultradian rhythm of GH secretion [5, 6]. A growing body of physiological evidence indicates that, in addition to exerting their characteristic actions at the level of the anterior pituitary, SRIF and GRF may in teract within the central nervous system to modulate GH release [6-14], While these studies, in general, provide support for the concept of a central influence of SRIF on pathways associated with the secretion of GRF, the mech anism and morphological substrate for this action are not yet known. Recent anatomic data suggest that SRIF may directly affect the GRF-containing cells of the arcuate (ARC) nu cleus, the source of most GRF projections to the median eminence [15-17], First, the ARC nucleus contains both a subpopulation of SRIF-immunoreactive (ir) neurons and a dense network of SRIF-positive axon terminals [18-20], some of which appear to originate from within the ARC nucleus and others from the periventricular/preoptic hy pothalamic area [18, 20, 21]. Second, these terminals were shown by double-labeling immunocytochemistry to directly contact GRF-ir cell bodies [22-24], Finally, we have recently demonstrated, using high-resolution radioautography, a selective association of [125I]-labeled SRIF-binding sites with a subpopulation of ARC neu rons, the distribution of which was very similar to that of GRF-ir neurons detected within the same region in a sep arate group of rats [25], These latter results led us to pro pose that some SRIF receptors might be directly localized on the surface or within GRF-containing cells. To test this hypothesis, in the present study, we have visualized and mapped radiolabeled SRIF-binding sites and GRF-ir neurons in adjacent 6-p.m-thick sections throughout the rostrocaudal extent of the ARC nucleus from the same animals.
a Fig. 1. Distribution of GRF-ir (a) and [l35I]-SRIF-labeled neurons (b) in the rat hypothalamic ARC nucleus (Arc), a GRF-ir perikarya (black arrows) are apparent in the ventrolateral component of the nucleus, where, in dark field (b) a group of [l25I]-SRIF-labeled neurons, detected as silver grain aggregates (white arrows), are also local ized. ME = Median eminence; III = third ventricle. Scale bar = 150 pm.
Assessment o f Colabeling The extent of cellular colocalization of [125I]-SRI F-binding sites to GRF-ir neurons was determined by superimposing serial maps of GRF-ir cells on the adjacent maps of radiolabeled SRIF cells. Co labeled neurons in the adjacent section were identified as those lying within one cell diameter ( w 10 urn) from the edge of the GRFir cell. The number of GRF-ir cells identified as SRIF-receptorbearing was documented for each of the 15 rostrocaudal planes in both intra- and extra-ARC regions and the results were expressed as a percentage (mean ± SE) of total GRF-ir cells detected.
20
To validate this approach, a control experiment was carried out in which a single marker (GRF antiserum) was applied to adjacent sections. Fifteen pairs of sections were cut at intervals of 100 pm over the rostrocaudal extent of the ARC nucleus and immediately processed for GRF immunohistochemistry, as described above. Each pair of sections was examined simultaneously with two Leitz Dialux 20 microscopes fitted with a comparison bridge. Between 30 and 50% of detectable GRF-ir ARC neurons were present in the adjacent section.
Results GRF Immunostaining As previously described [25], GRF-ir cells were mainly localized in the ventrolateral portion of the ARC nucleus and predominated within the caudal-most tier (fig. la). These labeled cells were round to fusiform in shape, dis played 2-3 cellular processes and measured 10.1 ± 1.4 pm in mean diameter (fig. 2b). On average, 13.2 ± 4.4 cells were detected per section. Moreover, an average of 4.0 ± 2.2 were detected in extra-ARC regions, most around the ventromedial nucleus of the hypothalamus (VMH) and in the lateral hypothalamic area (LH). f l25IJ-SRIF Radioautography Dark-field examination of light-microscopic radioau tograms of sections incubated with [1251]-SR1F revealed discrete silver grain clusters distributed throughout the ventrolateral portion of the ARC nucleus. These labeled foci stood out against a moderate background reaction (fig. lb) and with, bright-field optics were found to be in
McCarthy/Beaudet/Tannenbaum
Colocalization of SRIF Receptors and GRF Immunoreactivity
Downloaded by: Imperial College, School of Medicine, Wellcome Libr. 155.198.30.43 - 1/19/2019 7:53:33 AM
H :0 at a concentration of 1 mg/ml at room temperature. The re action was stopped after 15 s by addition of 10% BSA in 0.05 M phosphate buffer, pH 7.5 (500 pi). [1251J-SR1F was purified by gel filtration on a Sephadex G-25 (fine) column in 0.087 N acetic acid with 0.1 % BSA. The specific activity of the iodinated ligand was 165 Ci/mmol. For radioautographic visualization of [1251]-SRIF-binding sites, the adjacent section of each set of sections was: (1) preincubated (2 x 15 min at room temperature) in 0.17 A/Tris-HCI buffer (pH 7.4) containing 0.2% BSA (supplemented Tris buffer); (2) incubated (45 min at room temperature) in the same supplemented Tris buffer to which 0.25 nM [l25I]-SRIF, 3 x 10'5 M MgCb and 5 x 10'5 M bacitracin (Sigma Chemical) had been added; additional sections were incubated in the presence of 250 nM nonradioactive Tyr°-DTrps-SR1F-14 for determination of nonspecific binding; (3) tapped dry and rinsed ( 3 x 4 min) in ice-cold baths of supplemented Tris buffer, and (4) fixed immediately in an equally cold solution of 4% glutaraldehyde in 0.05 M phosphate buffer (30 min). After fixation, sections were dehydrated in graded ethanols, de fatted in xylene, rehydrated and coated by dipping in Kodak NTB-2 emulsion diluted 1:1 with distilled water. After 6 weeks of exposure, the radioautograms were developed in Dektol (Kodak), stained with cresyl violet and coverslipped. All [l35I]-SRIF-labeled cells, I section per level, were mapped and quantified as previously de scribed for the GRF series.
Fig. 2. Adjacent-section [l2SI]-SRIF autoradiography (a) and GRF immunocytochemistry (b). Five ['-51]-SRIF-labeled cells are seen to be GRF-iinmunopositive in the adjacent section (arrows). By contrast, two GRF-ir cells are not autoradiographically labeled (arrowheads). Note the greater shrinkage of autoradiographicallylabeled as compared to immunohistochemicallyreacted material. Scale bar = 20 urn.
GRF-ir/f12' IJ-SRIF Colabeling Superimposition of the camera lucida maps of GRF-ir over the [l25I]-SRIF binding maps indicated that, overall, 33.5 ± 10.8% (mean ± SE of 3 animals) of GRF-ir cells within the ARC nucleus were labeled with [1251]-SRIF in the adjacent section (figs. 2, 3). Colabeled cells concen trated, rostrocaudally, in the ventrolateral component of the ARC nucleus; more dorsally placed cells and those of the LH area, were rarely colabeled (fig. 3). Moreover, GRF cells located in extra-ARC regions did not show [I2', I]-SR1F binding. As shown in figure 4, the number of GRF-ir/[,25I]-SRIF colabeled cells did not vary consider ably from the retrochiasmatic area, rostrally (0-300 um),
to the mammillary nucleus, caudally (900-1,200 um). However, at the most caudal component of the ARC nu cleus (1,200-1,500 pm), the number of colabeled cells declined.
Discussion The results of the present study provide the first direct anatomic evidence for an association of high-affinity [l25I]-SRIF-binding sites with GRF-ir neurons in the ARC nucleus of the rat. These results are congruent with those of Bertherat et al. [29], which demonstrate an asso ciation o f[l25I]-SRIF-binding sites with ARC neurons ex pressing GRF mRNA. Taken together, these two sets of experiments provide strong anatomic evidence for the presence of SR1F receptors on a subpopulation of ARC neurons that both contain and express GRF. The combined adjacent-section autoradiographic/immunohistochemical approach used in the present study is a modification of a technique previously developed by us for the localization of high-affinity neurotensin-binding sites to dopaminergic neurons in the midbrain tegmen tum [30] and to vasoactive-intestinal-polypeptide-containing neurons in the suprachiasmatic nucleus of the rat [31]. Admittedly, simultaneous visualization of the two markers on the same tissue sections would have been preferable; however, in initial experiments, [l25I]-SRIF
21
Downloaded by: Imperial College, School of Medicine, Wellcome Libr. 155.198.30.43 - 1/19/2019 7:53:33 AM
direct register with small, round to oval, neuronal perikarya (fig. 2a). An average of 35.6 ± 6.5 SRIF-Iabeled cells were identified per section within the ARC nucleus proper. No labeled cells were present in sections incu bated with an excess of nonradioactive Tyr°-D-Trp8SRIF-14. [l25I]-SRIF-labeIed cells were more numerous, more evenly distributed, and extended further rostrally and caudally than GRF-ir cells. As a whole, however, the distributions of the two populations were similar, particu larly in the ventrolateral aspect of the ARC nucleus. Of the few SRIF-Iabeled cells observed in extra-ARC regions (1.33 ± 0.8), such as in the LH area, none were found in the peri-VMFI position occupied by one third of the GRF-ir cells.
o
• *
Rostro-caudal level (pm)
Fig. 3. Representative superimposed camera lucida drawings of GRF-ir neurons (solid circles), [,25I]-SR1 F-labeled cells (open cir cles) and colabeled GRF-ir/[l25I]-SRIF cells (stars), mapped in each of 4 coronal planes (400, 700, 1,000 and 1,300 um) caudal to the retrochiasmatic area.
22
autoradiographic and GRF immunohistochemical proto cols were found to be mutually exclusive. Nonetheless, the theoretical probability of detecting, and therefore co labeling, the same cell in two adjacent sections is, for neu rons 10 pm in diameter (d) and sections 6 pm in thickness (t), in the order of 45% (p = d /2 t + d), which fits well with our practical observation of 30-50%. It may there fore be assumed that the extent of colocalization reported in the present study is, if anything, underestimated. The monoiodinated probe used here for the visualiza tion of hypothalamic SRIF-binding sites was previously shown to bind with high affinity to the different types of SRIF receptors reported in rat brain [32,33]. Further more, high-resolution autoradiographic data have re cently indicated that SRIF-binding sites associated with
McCarthy/ Beaudet/Tannenbaum
Colocalization of SRIF Receptors and GRF Immunoreactivity
Downloaded by: Imperial College, School of Medicine, Wellcome Libr. 155.198.30.43 - 1/19/2019 7:53:33 AM
Fig. 4. Number (a) of GRF-ir neurons (■ ) and [l25I]-SRIFlabeled cells (□ ); and percent (b) GRF-ir cells colabeled with [ i25I]-SRI F in 5 coronal planes (0-1,500) of the ARC nucleus caudal to the retrochiasmatic area. Mean ± SE of 3 animals; 3 sections per plane.
20, 21]. Furthermore, this region is also the one where, at the ultrastructural level, synaptic junctions between SRIF-ir axon terminals and GRF-ir cell bodies have been identified [22-24], These findings suggest that at least part of the [l25I]-SRIF-binding sites found here to be as sociated with GRF-ir cells are present on the surface of, as opposed to (or in addition to) inside, GRF-producing cells. It should be recalled, however, that an important contingent of non-SRIF-ir terminals, including some identified as containing substance P or enkephalin-8, were also found to synapse upon ARC GRF-ir cells [22]; this is in keeping with our observation that only a subpop ulation of GRF neurons is endowed with SRIF receptors. In summary, the results reported here demonstrate that a subpopulation of GRF-ir nerve cell bodies detected within the ARC nucleus of the rat hypothalamus pos sesses SRIF-binding sites. The finding of SRIF receptors on GRF-containing ARC neurons provides a morpho logical substrate for the concept of direct 'cross-talk' be tween SRI F and GRF neuronal systems within the central nervous system [6-14], Such intrahypothalamic interac tions between these two neuropeptides may be a vital component in the generation and maintenance of the ultradian rhythm of GH secretion.
Acknowledgments We thank Dr. Filoteo Pasquini for technical advice and Julie Temko for expert secretarial assistance. This work was supported by grants MT-6837 (to G.S.T.) and MF-7386 (to A.B.) from the Medi cal Research Council of Canada and grant 911437-102 (to G.S.T. and A.B.) from the Fonds de la Recherche en Santé du Québec. G.F.M. was the recipient of a Postdoctoral Research Fellowship from the McGill University-Montreal Children's Hospital Research Institute. A.B. is a Scientist Awardee from the Medical Research Council of Canada. G.S.T. is a ‘Chercheur-Boursier de Mérite Exceptionnel' of the Fonds de la Recherche en Santé du Québec.
23
Downloaded by: Imperial College, School of Medicine, Wellcome Libr. 155.198.30.43 - 1/19/2019 7:53:33 AM
ARC neurons had the same pharmacological characteris tics as the ones present in the rest of the brain [34], The density of these ARC SRIF-binding sites was found to be unaffected by the type of colchicine pretreatment used here for improved immunohistochemical visualization of GRF-ir neurons [our unpubl. observations]; it was, how ever, shown to be markedly increased in animals pre treated with the drug CSH, administered at the same dose and time interval as in the present study [35]. Interest ingly, this increase appears to mainly affect the density of binding sites per cell as opposed to the number of reac tive cells, suggesting that most, if not all, of the cells bear ing SRIF receptors are being detected under the present pharmacological conditions. In keeping with this inter pretation, the number of [l25I]-SRIF-labeled cells de tected here per section of rat hypothalamus is consistent with that reported earlier in non-colchicine-, non-CSHtreated rats [25]. In contrast, the number of GRF-ir cells detected in the adjacent series was clearly inferior to that previously re ported by us in animals treated with either colchicine [25] or CSH [14] alone, and fixed by intravascular perfusion of aldehydes. This overall reduction in the sensitivity of im munodetection is likely due to both the reduced thickness of the sections and their fixation by immersion in the present study. Immersion fixation of frozen tissue sec tions has indeed been shown to provide poorer retention of peptide antigenicity than that obtained after vascular perfusion [our unpubl. observations]. It is, therefore, likely that only a fraction of ARC GRF-producing cells are being detected under the present experimental condi tions. The quantitative data provided here should there fore be considered more relative than absolute. In parti cular, they do not permit us to state whether or not [1251]-SRIF receptors are also associated with non-GRF neurons; indeed, [125I]-SRIF-labeled cells found to be non-GRF-ir in the adjacent series might have corre sponded to cells that were below threshold of immuno histochemical detection. What can be safely concluded is that a subpopulation of GRF neurons does not bear SRIF receptors, since a number of GRF-positive cells that were recognized by Nissl staining as being present in the adja cent section were clearly free of autoradiographic label. Most conspicuous in this category were GRF-ir cells de tected in the surround of the VMH nucleus which were systematically label-free in the adjacent series. It is interesting to note that most of the GRF-ir/ [l25I]-SRIF colabeled cells were detected in the basolateral region of the ARC nucleus where a dense fiber net work of SRIF-ir terminals has previously been traced [18,
1 Rivier J, Spiess J, Thorner M, Vale W: Charac terization of a growth hormone-releasing fac tor from a human pancreatic islet tumor. Na ture 1982;300:276-279. 2 Guillemin R. Brazeau P, BOhlen P, Esch F, Ling N, Wehrenberg WB: Growth hormonereleasing factor from a human pancreatic tumor that caused acromegaly. Science 1982; 218:585-587. 3 Spiess J, Rivier J, Vale W: Characterization of rat hypothalamic growth hormone-releasing factor. Nature 1983;303:532-535. 4 Brazeau P, Vale W. Burgus R, Ling N, Butcher M, Rivier J, Guillemin R: Hypotha lamic polypeptide that inhibits the secretion of immunoreactive pituitary growth hormone. Science 1973;179:77-79. 5 Tannenbaum GS, Ling N: The interrelation ship of growth hormone (GH)-releasing factor and somatostatin in generation of the ultradian rhythm of GH secretion. Endocrinology 1984;115:1952-1957. 6 Plotsky PM, Vale W: Patterns of growth hor mone-releasing factor and somatostatin secre tion into hypophysial portal circulation of the rat. Science 1985;230:461-463. 7 Abe H, Kato Y, Iwasaki Y, Chihara K, Imura H: Central efTect of somatostatin on the secre tion of growth hormone in the anesthetized rat. Proc Soc Exp Biol Med 1978; 159:346349. 8 Lumpkin MD, Negro-Vilar A, McCann SM: Paradoxical elevation of growth hormone by intraventricular somatostatin: Possible ultra short-loop feedback. Science 1981 ;211: 1072-1074. 9 Thomas CR, Groot K, Arimura A: Antiserum to rat growth hormone (GH)-releasing factor suppresses but does not abolish antisomato statin-induced GH release in the rat. Endocri nology 1985;116:2174-2178. 10 Tannenbaum GS, Patel YC: On the fate of centrally administerated somatostatin in the rat: Massive hypersomatostatinemia resulting from leakage into the peripheral circulation has effects on growth hormone secretion and glucoregulation. Endocrinology 1986; 118: 2137-2143. 11 Murakami Y, Kato Y, Kabayama Y, Inoue T, Koshiyama H, Imura H: Involvement of hy pothalamic growth hormone (GH)-releasing factor in GH secretion induced by intracerebroventricular injection of somatostatin in rats. Endocrinology 1987; 120:31 1-316. 12 Katakami H, Downs TR, Frohman LA: Inhib itory efTect of hypothalamic medial preoptic area somatostatin on growth hormonereleasing factor in the rat. Endocrinology 1988:123:1103-1109. 13 Miki N, Ono M, Shizume K: Withdrawal of endogenous somatostatin induces secretion of growth hormone-releasing factor in rats. J En docrinol 1988;! 17:245-252.
14 Tannenbaum GS, McCarthy GF, Zeitler P. Beaudet A: Cysteamine-induced enhance ment of growth hormone-releasing factor (GRF) immunoreactivity in arcuate neurons: Morphological evidence for putative somatostatin/GRF interactions within hypothala mus. Endocrinology 1990;127:2551-2560. 15 Bloch B, Brazeau P, Ling N, Böhlen P, Esch F, Wehrenberg WH, Benoit R. Bloom F, Guil lemin R: Immunohistochemical detection of growth hormone releasing factor in brain. Na ture 1983;301:607-608. 16 Merchenthaler I, Vigh S, Schally AV, Petrusz P: Immunocytochemical localization of growth hormone-releasing factor in the rat hy pothalamus. Endocrinology 1984; 114:10821085. 17 Sawchenko PE, Swanson LW, Rivier J, Vale WW: The distribution of growth hormone-re leasing factor (GRF) immunoreactivity in the central nervous system of the rat: An immuno histochemical study using antisera directed against rat hypothalamic GRF. J Comp Neu rol 1985;237:100-115. 18 Kawano H, Daikoku S, Saito S: Immunohis tochemical studies of intrahypothalamic so matostatin-containing neurons in rat. Brain Res 1982;242:227-232. 19 Ohtsuka M, Hisano S, Daikoku S: Electronmicroscopic study of somatostatin-containing neurons in rat arcuate nucleus with special reference to neuronal regulation. Brain Res 1983;263:191-199. 20 Willoughby JO, Brogan M, Kapoor R: Hypo thalamic interconnections of somatostatin and growth hormone releasing factor neurons. Neuroendocrinology 1989:50:584-591. 2 1 Makara GB, Palkovits M, Antoni FA, Kiss JZ: Topography of the somatostatin-immunoreactive fibers to the stalk-median eminence of the rat. Neuroendocrinology 1983;37:1-8. 22 Daikoku S, Hisano S, Kawano H, ChikamoriAoyama M, Kagotani Y, Zhang R, Chihara K: Ultrastructural evidence for neuronal regula tion of growth hormone secretion. Neuroen docrinology 1988;47:405-415. 23 Liposits ZS, Merchenthaler I, Pauli WK, Flerko B: Synaptic communication between somatostatinergic axons and growth hor mone-releasing factor (GRF) synthesizing neurons in the arcuate nucleus of the rat. His tochemistry 1988;89:247-252. 24 Horvàth S, Palkovits M, Göres T, Arimura A: Electron microscopic immunocytochemical evidence for the existence of bidirectional syn aptic connections between growth hormone releasing hormone- and somatostatin-con taining neurons in the hypothalamus of the rat. Brain Res 1989;481:8-15.
25 Epclbaum J, Moyse E, Tannenbaum GS. Kordon C, Beaudet A: Combined autoradio graphic and immunohistochemical evidence for an association of somatostatin binding sites with growth hormone-releasing factorcontaining nerve cell bodies in the rat arcuate nucleus. J Neurocndocrinol 1989;1:106-116. 26 Sternberger LA: Immunocytochemistry. New York, Wiley, 1979. 27 Galiyas F. Gores T, Merchenthaler I: High grade intensification of the end-product of the diaminobenzidine reaction for peroxidase his tochemistry. J Histochem Cytochem 1982; 30:183-184. 28 Tannenbaum GS. Lapointe M, Gurd W, Ftnkelstein JA: Mechanisms of impaired growth hormone secretion in genetically obese Zucker rats: Roles of growth hormone-releasing fac tor and somatostatin. Endocrinology 1990; 127:3087-3095. 29 Bertherat J, Doumaud P, Bèrod A, Normand E, Bloch B, Rostène W, Kordon C, Epelbaum J: Growth hormone-releasing hormone-syn thesizing neurons are a subpopulation of so matostatin receptor-labelled cells in the rat ar cuate nucleus: a combined in situ hybridiza tion and receptor light-microscopic radioautographic study. Neuroendocrinology 1992; 56:25-31. 30 Szigethy E, Beaudet A: Correspondence be tween high affinity (lî5Ij-neurotensin binding sites and dopaminergic neurons in the rat sub stantia nigra and ventral tegmental area: A combined radioautographic and immunohis tochemical light microscopic study. J Comp Neurol 1989;279:128-137. 31 François-Bellan AM, Bosler O, Tonon MC, Lin-Tong W, Beaudet A: Association of neu rotensin receptors with VIP-containing neu rons and serotonin-containing axons in the suprachiasmatic nucleus of the rat. Synapse, in press. 32 Krantic S, Martel JC, Weissman D, Quirion R: Radioautographic analysis of somatostatin receptor sub-types in rat hypothalamus. Brain Res 1989;498:267-278. 33 Krantic S, Martel JC, Weissman D, Pujol JF, Quirion R: Quantitative radioautographic study of somatostatin receptors heterogeneity in the rat extrahypothalamic brain. Neurosci ence 1990;39:127-137. 34 Bertherat J, Slama A, Kordon C, Vtdeau C, Epelbaum J: Characterization of pericellular IffyrO DTRP8 somatostatin binding sites in the rat arcuate nucleus by a newly de veloped method: Quantitative high-resolution light microscopic radioautography. Neurosci ence 1991;41:571-579. 35 Bertherat J, Bérod A, Normand E, Bloch B, Rostcne W, Kordon C, Epelbaum J: Somato statin depletion by cysteamine increases so matostatin binding and growth hormone releasing factor messenger ribonucleic acid in the arcuate nucleus. J Neuroendocrino! 1991;3:115-118.
24
McCarthy/Beaudet/Tannenbaum
Colocalization of SRI F Receptors and GRF Immunoreactivity
Downloaded by: Imperial College, School of Medicine, Wellcome Libr. 155.198.30.43 - 1/19/2019 7:53:33 AM
References
