play a role in thyroid-stimulating hormone release. On the release during stress, mediated at least partly by suppression of

Transcription

1 Proc. Nail. Acad. Sci. USA Vol. 89, pp , December 1992 Physiology The role of endogenous atrial natriuretic peptide in resting and stress-induced release of corticotropin, prolactin, growth hormone, and thyroid-stimulating hormone (third ventricular I Jection/atrial natriuretic peptide antiserum) CLSO R. FRANCI*, JANT A. ANSLMO-FRANCI*, AND SAMUL M. MCCANNt4 *DePartamento de Fisiologia, Faculdade de Medicina de Ribeirao Preto, Universidade de Sao Paulo, Ribeirao Preto, Sao Paulo, Brazil; and tneuropeptide Division, Department of Physiology, The University of Texas Southwestern Medical Center, 5323 Harry Hines Boulevard, Dallas, TX Contributed by Samuel M. McCann, August 28, 1992 ABSTRACT Our previous studies have shown that stimulation of the anteroventral third ventricle region increases atrial natriuretic peptide (ANP) release, whereas lesions of the anteroventral third ventricle or median eminence block the release of ANP from blood volume expansion, suggesting a critical central nervous system participation in this response. ANP is also produced within neurons that have cell bodies in the rostral hypothalamus and axons that extend to the median eminence and neural lobe. In addition to its natriuretic effect, the peptide can inhibit the release of corticotropin (ACTH) and prolactin, anterior pituitary hormones that are released during stress. To determine the physiologic sinfca of ANP in the control of basal and stress-induced release of anterior pituitary hormones, highly specific antiserum against the peptide (AB- ANP) was microinjected into the third cerebral ventricle of conscious freely moving male rats to Immunoneutralize hypothalamic ANP. In the initial experiment, the antiserum or control normal rabbit serum () was injected into the third cerebral ventride to determine the effect of the antiserum on basal release of pituitary hormones. The antiserum had no effect on the concentrations of plasma ACTH, prolactin, or thyroid-stimulating hormone for 3 hr after the injection; however, plasma growth hormone concentration, although unchanged for 2 hr, was markedly elevated at 3 hr. These results indicate that although ANP appears to have no effect on the basal release of the other hormones, it has a physiologically significant inhibitory effect on growth hormone release. The delay of the effect is probably related to the time required for the antiserum to diffuse to the site of action of the peptide, presumably at some distance from the ventricle. Since this effect was demonstrable only after 3 hr, in the stress experiment, the antiserum or was microinjected into the third ventricle 3 hr prior to application of ether stress. The rapid elevation of plasma ACTH in -injected rats was markedly augmented by AB-ANP. ther also induced a rapid increase in plasna prolactin in the -injected animals, as expected. Contrary to the ACTH response, the maximal increase in plasma prolactin after ether was attenuated in animals preinjected with AB-ANP. In the -injected animals, there was a significant decline in plasma growth hormone after the application of ether that was significantly accentuated by AB-ANP, but this was probably the result of the her initial levels of plasma growth hormone in the ANP-AB group followed by its disappearance with a half-time similar to that of the -injected group. The decline in plasma thyroidstimulating hormone after ether stress was unaltered in the animals injected with AB-ANP. The results of these immunoneutralization studies suggest that endogenous ANP does not The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C solely to indicate this fact play a role in thyroid-stimulating hormone release. On the other hand, the endogenous peptide appears to have a physiologically significant inhibitory role in suppressing ACTH release during stress, mediated at least partly by suppression of vasopressin release. ndogenous ANP has a pathophysiologic role in augmenting the prolactin release in stress either by inhibiting release of prolactin-inhibiting factors or, alternatively, by enhancing release of prolactin-releasing factors. ndogenous ANP appears to inhibit resting, without altering stress-induced inhibition of growth hormone release by stimulating somatostatin release and/or inhibiting growth hormone-releasing hormone release or by both actions. Atrial natriuretic peptide (ANP) produced and released from atrial myocytes induces natriuresis, diuresis, and lowering of blood pressure (1, 2). In addition, it decreases water and salt intake (3-6). All of these actions tend to promote a decrease in body fluid volume and they appear to be homeostatic responses to increased extracellular fluid volume. Furthermore, the peptide suppresses vasopressin release from the neural lobe (7), which can cause renal water loss and inhibits release corticotropin (ACTH) (8) and prolactin (9) from the anterior lobe of the pituitary gland. The decreased ACTH release would reduce the secretion of aldosterone from the adrenal glomerulosa, which is further reduced by direct action of ANP (10), leading to additional renal sodium loss. The decreased prolactin release could further augment natriuresis since prolactin has a direct salt-retaining action on the kidney (11). Stress increases ACTH and prolactin and inhibits the release of growth hormone (GH) and thyroid-stimulating hormone (TSH) in the rat (12) and the actions of ANP to alter the release of these hormones suggest that it may influence the hypothalamic pituitary response to stress. These effects of ANP may be mediated via a hypothalamic ANP neuronal system with cell bodies that extend from the anteroventral third ventricle region to the paraventricular nucleus and have axons that project to the median eminence and neural lobe of the pituitary (13, 14). In the median eminence the peptide could enter hypophyseal portal veins and be delivered to the anterior pituitary to affect its function. Alternatively, the ANP neurons might alter the release of hypothalamic releasing and inhibiting hormones via a hypothalamic action that would, in turn, alter release of anterior pituitary hormones. The physiologic significance of the actions of ANP on anterior pituitary hormone release is unknown. Consequently, in the present study, we evaluated the effects of Abbreviations: ANP, atrial natriuretic peptide; ACTH, corticotropin;, normal rabbit serum; TSH, thyroid-stimulating hormone; GH, growth hormone; 3V, third cerebral ventricle; CRH, corticotropinreleasing hormone. tto whom reprint requests should be addressed.

2 11392 Physiology: Franci et al. Proc. Nadl. Acad. Sci. USA 89 (1992) passive immunoneutralization of brain ANP by microinjecting highly specific antiserum developed against the peptide into the third cerebral ventricle (3V) and measuring the effect on plasma pituitary hormone concentrations in conscious, normal, and stressed male rats. Preliminary reports of this work have been presented (15, 16). MATRIALS AND MTHODS Male Harlan-Sprague-Dawley rats, g, were housed in a constant light-dark cycle (lights on between the hours of 0500 and 1900) and temperature (23-250C)-controlled room. Laboratory chow and water were freely available. One week before the experiments, a stainless steel guide cannula was implanted into the 3V as described (17) under anesthesia induced by 2.5% tribromoethanol [Aldrich; 1 mg/100 g (body weight), i.p.]. Animals that had regained preoperative weight were submitted to an additional operation 24 hr before the experiments during which a catheter was introduced into the external jugular vein under tribromoethanol anesthesia (18). erimental Procedure. All experiments began between the hours of when a heparinized blood sample (0.6 ml) was collected from the jugular catheter immediately before 3V microinjection of normal rabbit serum () or antiserum against ANP (AB-ANP). Additional blood samples (0.6 ml) were collected from unstressed animals 15, 30,, 120, and 180 min after the microinjections. After removal of each blood sample, 0.6 ml of 0.91% NaCl was injected to replace the volume of blood removed. Rats to be subjected to stress received microinjections of or AB-ANP just after withdrawal of the initial blood sample (0.6 ml). They were allowed to rest for 180 min. Further samples (0.6 ml) were then drawn 5 min before and 2, 5, 15, 30, and min after exposure to ether for 1 min as described (19). The microinjections were carried out with a Hamilton microsyringe connected by P10 polyethylene tubing to a needle filled with the solution to be injected. Antiserum (2 1d) diluted 1:2 with 0.9%o NaCl or similarly diluted was microinjected over 1 min. The AB-ANP (Peninsula Laboratories) crossreacts 1000% with rat ANP {human [Ile12]ANP}, human ANP, ANP48-33), human [IBe12]ANP-(3-29), and rat atriopeptin III. It also crossreacts % with ANP-(18-28), 10% with auriculin A, 5% with rat atriopeptin II, and 1% with ANP-(13-28) and crossreacts negligibly with oxytocin, [Arg8]vasopressin, somatostatin, and rat atriopeptin I (H. Chang, Peninsula Laboratories, personal communication). At the end of the experiment, the brains were removed and the site of microinjections was determined by microscopic examination of frozen coronal sections. Radlonimunoassay. Plasma prolactin, GH, and TSH were measured using RIA kits provided by the National Institute of Diabetes, Digestive and Kidney Disease (NIDDK). Results were expressed in terms ofthe RP-2 reference standards provided with the kits. Plasma ACTH was measured with a method (20) using commercially available antiserum (IgG Corp., Nashville, TN) and human ACTH provided by the NIDDK pituitary hormone program. Statistics. The values of plasma hormones were compared by an ANOVA followed by the Newman-Keul test. Significance of differences between two means, as for the areas under two curves ofplasma hormone values, was determined by Student's t test. RSULTS Table 1 shows plasma concentrations of GH, TSH, prolactin, and ACTH immediately before (time 0) and 15, 30,, 120, and 180 min after 3V microinjection of or AB-ANP in unstressed rats. Hormone values in the -injected animals were not altered except for a significant decline in plasma TSH by 180 min. There were no significant differences in hormone concentrations between the control group (3V microinjection of ) and the group that received 3V microinjection of AB-ANP, except that the plasma GH concentrations, although unaltered earlier, were significantly higher 180 min after microinjection of AB-ANP than. There was a significant elevation of plasma ACTH concentration 2 min after the onset of ether stress and a peak at 5 min in the animals previously injected with (Fig. 1). The levels declined precipitously at 15 min and were no longer elevated 30 min after ether. In contrast, the response in the animals previously injected with AB-ANP was magnified and significantly greater than that in the -injected animals 2, 15, and 30 min after the onset of ether exposure. The maximum increment in plasma ACTH in the AB-ANPtreated rats was significantly greater (P < 0.01) than that in -injected controls (Fig. 2). The area under the curve of plasma ACTH was also significantly (P < 0.01) increased in the AB-ANP-treated rats when compared to that in the -injected controls (data not shown). As expected, ether anesthesia induced a sharp elevation in plasma prolactin concentrations in -injected rats; the concentration reached a peak at 2 min and had markedly declined by 15 min (Fig. 3). Values were no longer elevated at 30 and min. Although they were somewhat lower in the AB-ANP-injected group than in the -injected group 2, 5, and 15 min after ether, these changes were not statistically significant at any time; however, the maximum increment in plasma prolactin after ether was significantly lower (P < 0.01) in the rats preinjected with AB-ANP than in those given (Fig. 4). Plasma GH was significantly increased 3 hr after 3V microinjection of AB-ANP, S min before application of ether stress (Fig. 5), as in the first experiment. After ether stress, plasma GH decreased in both groups, but it remained significantly higher in the group injected with AB-ANP. The decline in plasma GH in the AB-ANP-treated group was greater (P < 0.01) than that in the -treated group; Table 1. Plasma GH, TSH, prolactin, and ACTH in normal rats immediately before (0 min) and 15, 30,, 120, and 180 min after microinjection of or AB-ANP into the 3V Hormone Serum 0 min 15 min 30 min miin 120 miin 180 min GH, ng/ml (9) 44.8 ± ± ± ± ± ± 3.4 AB-ANP (9) 45.2 ± ± ± ± ± ± 8.2* TSH, ng/ml (8) ± ± ± ± ± ± 6.8 AB-ANP (11) ± ± ± ± ± ± 13.5 Prolactin, ng/ml (10) 11.6 ± ± ± ± ± 1.5 AB-ANP (10) 10.8 ± ± ± ± ± 1.3 ACTH, pg/ml (9) ± ± ± ± ± 1.2 AB-ANP (8) ± ± ± ± 1.9 Values are mean ± SM. Number in parentheses indicates group size. *P < 0.01 vs. control value ().

3 Physiology: Fmnci et al. Proc. Nati. Acad. Sci. USA 89 (1992) n 0- HT- C) rt Mi m) CL 150o 130 [ I 1 0F 90 v 70 h 50 i- 30 r j l D0-c-- I U ~~~(9) D ABANPo /' \ T'T/ ; K \ 40G L o MU NR. 1 C. l AB-ANP i_ i,r'. 3 Tirne.!-tS Time, min FIG. 1. ffects of 3V microinjection of and AB-ANP on plasma ACTH in rats exposed to ether stress. In this and subsequent figures, numbers in parentheses indicate the size of the group. *, P < 0.05; **, P < 0.02 vs. control values (). however, the half-time of its disappearance (7 min) was not altered from that of the -injected rats. After ether stress, plasma TSH levels were reduced to near minimal values at 5 min and remained low for the -min duration of the experiment in the -injected animals (Fig. 6). There was no difference between the results in the AB-ANP and -injected groups. DISCUSSION Although intraventricular injection of AB-ANP had no effect on resting ACTH levels, it markedly augmented the ACTH release caused by ether stress. This result indicates that endogenous central ANP plays an inhibitory role in the control of ACTH release in response to stress. We speculate that the antiserum immunoneutralizes the ANP within the < 80 qco _ F-_ l_ *.-.. *... ::-: *. * :.: : *... * :.: :---.. *.- - * * AB-ANP FIG. 2. Maximal increase in plasma ACTH after inhalation of ether by - and AB-ANP-injected rats. **, P < 0.01 vs. injected rats. a Time, rein FIG. 3. ffects of 3V microinjection of and AB-ANP on plasma prolactin in rats submitted to ether stress. hypothalamus and thereby relieves an inhibition of stressinduced vasopressin release produced by ANP. Because vasopressin is released into the hypophyseal portal vessels during stress and directly stimulates ACTH release by the corticotrophs of the anterior pituitary (21, 22), this could account for the inhibitory action of ANP on stress-induced ACTH release. Vasopressin also appears to increase release of corticotropin-releasing hormone (CRH) and to potentiate stimulation of ACTH release by CRH (21-24). ANP may also directly inhibit CRH release but this has not been demonstrated. ANP has also been reported to suppress ACTH release via a direct action on the pituitary (8), and ifthe antisera reached the pituitary in sufficient concentrations in the present experiments, it might have blocked this inhibition as well. Our results indicate that ANP has a physiologically significant action to suppress stress-induced ACTH release by hypothalamic and/or pituitary action. ANP has been shown to act centrally to suppress prolactin release (9) via stimulation of dopamine release into the portal vessels, which then inhibits release of prolactin from the lactotrophs. The inhibitory action of centrally injected ANP on prolactin release appears not to be physiologically significant in the control of resting prolactin release since the AB-ANP did not alter resting prolactin values. However, 0) LĒ 20 C 0cut 10_ AB-ANP FIG. 4. Maximal increase in plasma prolactin after inhalation of ether by - and AB-ANP-injected rts. **, P < 0.01 vs. injected rats. **

4 11394 Physiology: Franci et al. 130 U:D 90- (5 In C- 70i 50j F 3 Time. hr -7 C),.. (10) Li F--. AB-ANP (12) ~~~~~~~~~~~~~~~ \ lr Time, min FIG. 5. ffects of 3V microinjection of and AB-ANP on plasma GH in rats submitted to ether stress. *, P < 0.001; **, P < 0.01 vs. control values (). although the ANP antiserum had no effect on resting prolactin levels, the elevation of plasma prolactin after ether stress was decreased below that in -injected animals. This is consistent with a physiologically significant stimulatory action of ANP on prolactin release during stress. During stress, this action is presumably mediated by its ability to stimulate the release of a prolactin-releasing factor, such as oxytocin (25), or to inhibit the release of a prolactin release inhibitory factor, such as dopamine (26). Our results demonstrate a clear elevation in plasma GH, which was present only at 3 hr after 3V injection of AB-ANP. The time required to observe effects after intraventricular injection of antisera to various peptides has ranged from 10 min to 4 hr (19, 27). Presumably, the variability reflects differences in the time required for the different antisera to penetrate into hypothalamic tissue and to diffuse to the site of action of the immunoneutralized peptide; the longer the time, presumably, the further the distance of the site from the ventricular wall. We hypothesize that the site of action of AB-ANP is within the hypothalamus at some distance from the ventricular wall to inactivate the action of the endogenous go 1[ o L r: 1 c00[ Me 80L (-'0 40 :0!,re h S U~(9) AB-ANP C) L (1 0) 7- -a. A ~ - ~~ is Time. mir FIG. 6. ffects of 3V microinjection of and AB-ANP on plasma TSH in rats submitted to ether stress. Proc. Natl. Acad. Sci. USA 89 (1992) peptide that is suppressing GH secretion. This suppression could be mediated either by stimulation of somatostatin release into portal vessels, which would then block release of GH from the somatotrophs, by inhibition of GH-releasing hormone release into the portal vessels, or by a combination of these two actions. Alternatively, if ANP has a direct effect to suppress GH release at the pituitary (8), the antiserum might immunoneutralize the peptide and block this inhibitory action after its uptake by portal veins and transport to the pituitary. Our studies indicate that =% % of ANP injected into the 3V reaches the anterior pituitary (J. Antunes-Rodrigues, J. Gutkowska, and S.M.M., unpublished data). The long time course for the antiserum to be effective is consistent with the possibility that it might be immunoneutralizing ANP, which is inhibiting GH release from the pituitary directly. Thus, it is clear that ANP exerts a tonic inhibitory control over GH release, probably at the hypothalamus, but possibly also directly at the pituitary. Although the lowering of plasma GH that followed stress was significantly greater in the animals injected 3 hr before with AB-ANP than in the -injected animals, we believe that this greater decline in GH in AB-ANP-injected rats is simply a reflection of their higher prestress levels ofgh since the half-time of disappearance of the peptide was not altered. Thus, stress suppresses GH release similarly in both groups. Consequently, we conclude that endogenous ANP does not alter the inhibitory effect of stress on GH release, which is mediated by increased somatostatin release (28). There have been few studies of the effects of ANP on GH release. In one report, intraventricular injection of ANP increased plasma GH probably by promoting a decrease of somatostatin release rather than an increase of GH-releasing hormone release (29). Thus, in that experiment, the peptide itself had the same effect as the antiserum directed against it in our experiments. The explanation for the apparent discrepancy between their results and ours (nay be related to injection of the peptide into the right lateral ventricle, rather than the 3V as we used. Lateral ventricle injection would lead to asymmetric delivery of the peptide to other than hypothalamic and, particularly, cortical structures. Our experiments appear to indicate that the increased release of ANP during stress dampens a number ofresponses to the stressful stimulus. For example, the inhibition of vasopressin secretion stimulated by stress induced by ANP would increase renal water loss and the diminished ACTH release induced by the peptide would lead to reduction in aldosterone secretion thereby diminishing the sodium retention, which characterizes stress retention. The ANP released in stress probably acts intrahypothalamically to suppress vasopressin secretion, and since vasopressin augments ACTH secretion during stress, this diminution in vasopressin release would lead to the decreased ACTH secretion as illustrated in Fig. 7. The increased prolactin release in stress induced by the increased endogenous ANP release would not augment sodium excretion since, at least under certain circumstances, it appears to have a direct antinatriuretic effect on the kidney (11). ACTH, via its stimulation of adrenal glucocorticoid secretion, has immunosuppressive action (12). Therefore, the action of ANP in stress would reduce this immunosuppressive effect. Contrariwise, prolactin enhances immune responses and the increased prolactin release induced by ANP would enhance immunity (12). Thus it would appear that the secretion of ANP within the hypothalamus during stress tends to alter the stress response. In contrast to these results with GH, ACTH, and prolactin, there was not only no effect of the antiserum on resting TSH values but also no alteration in the suppression of TSH release induced by ether stress. In other studies we have indicated (19) the physiological significance of the suppres-

5 Physiology: Franci et al. FIG. 7. Schematic diagram of the presumed influence of ANP released from hypothalamic neurons on suppression of the release of vasopressin, and consequently, the release of ACTH from the anterior lobe of the pituitary gland. The decrease in stress-induced vasopressin release would decrease the antidiuresis that occurs in stress, whereas the decrease of ACTH release would lead to decreased aldosterone release and diminished salt retention. Vasopressin released into the long portal veins (LPV) and into the short portal vessels (SPV) stimulates ACTH release from the corticotropes directly and amplifies the response to CRH. ANPn, ANP neuron; VPn, vasopressin n; OC, optic chiasm; M, median eminence of the tuber cinereum; MB, mammillary bodies; S, pituitary stalk; V, vein; NL, neural lobe; AC, adrenal cortex; K, kidney; Aldo, aldosterone; Na', Na+ excretion rate; UV, urine excretion rate; -, inhibition; +, excitation;, increased; j, decreased. sive action of ANP on basal leuteinizing hormone but not follicle-stimulating hormone release. In conclusion, these immunoneutralization experiments support the concept that ANP acts centrally during stress to dampen the stress-induced stimulation of ACTH and augment the stress-induced prolactin release that characterize the hypothalamic pituitary stress response in the male rat. The stress-induced inhibition of TSH and GH release is unmodified, but ANP released tonically inhibits GH release but not that of the other hormones. This work was supported by National Institutes of Health Grants HD09988 and DK Proc. Natl. Acad. Sci. USA 89 (1992) DeBold, A. J., Borenstein, H. B., Veress, A. T. & Sonnenberg, H. (1981) Life Sci. 28, DeBold, A. J. (1982) Proc. Soc. xp. Biol. Med. 170, Antunes-Rodrigues, J., McCann, S. M., Rogers, L. C. & Samson, W. K. (1985) Proc. Natl. Acad. Sci. USA 82, Antunes-Rodrigues, J., McCann, S. M. & Samson, W. K. (1986) ndocrinology 118, Franci, C. R., Kozlowski, G. P. & McCann, S. M. (1989) Proc. Nati. Acad. Sci. USA 86, Masoto, C. & Negro-Vilar, A. (1985) Brain Res. Bull. 15, Samson, W. K. (1985) ndocrinology 117, Shibasaki, T., Naruse, M., Yamauchi, N., Masuda, A., Imaki, T., Naruse, K., Demura, H., Ling, N., Inagami, T. & Shizume, K. (1986) Biochem. Biophys. Res. Commun. 135, Samson, W. R. & Bianchi, R. (1988) Can. J. Physiol. Pharmacol. 66, Goodfriend, T. L., lliott, M.. & Atlas, S. A. (1984) Life Sci. 35, Stier, C. T., Cowden,. A., Friesen, H. G. & Allison, M.. M. (1984) ndocrinology 115, McCann, S. M., Rettori, V., Milenkovic, L., Jurcovicova, J. & Gonzalez, M. C. (1990) in Circulating Regulatory Factors and Neuroendocrine Function, eds. Porter, J. C. & Jezova, D. (Plenum, New York), Vol. 274, pp Jacobowitz, D. M., Skofitsch, G., Keiser, H. R., skay, R. L. & Zamir, N. (1985) Neuroendocrinology 40, Nakamaru, M., Takayanagi, R. & Inagami, T. (1986) Peptides 7, McCann, S. M., Rettori, V., Wenger, T. & Dalterio, S. (1987) in Neuro. ndocrinologia, Proceedings of the National Academy ofsciences ofargentina, ed. Arguilles, A. (National Acad. Press, Buenos Aires, Argentina), pp McCann, S. M. (1990) in Neuroendocrinology: New Frontiers, eds. Gupta, D., Woolmann,. & Ranke, W. (Brain Research Promotion, Tubingen, F.R.G.), pp Antunes-Rodrigues, J. & McCann, S. M. (1970) Proc. Soc. xp. Biol. Med. 133, Harms, P. G. & Ojeda, S. R. (1974) J. Appl. Physiol. 36, Franci, C. R., Anselmo-Franci, J. A. & McCann, S. M. (1989) Neuroendocrinology 51, Nicholson, W.., Davis, D. R., Sherrell, B. J. & Orth, D. N. (1984) Clin. Chem. 30, McCann, S. M., Lumpkin, M. D. & Samson, W. K. (1982) in Neuroendocrinology of Vasopressin, Corticoliberin and Opiomelanocortins, eds. Baertsche, A. J. & Dreifuss, J. J. (Academic, London), pp Ono, N., Lumpkin, M. D., Samson, W. K., McDonald, J. K. & McCann, S. M. (1984) Life Sci. 35, Ono, N., Bedran de Castro, J. & McCann, S. M. (1985) Proc. Natl. Acad. Sci. USA 82, Ono, N., Samson, W. K., McDonald, J. K., Lumpkin, M. D., Bedran de Castro, J. C. & McCann, S. M. (1985) Proc. Natl. Acad. Sci. USA 82, Samson, W. K., Lumpkin, M. D. & McCann, S. M. (1986) ndocrinology 119, McCann, S. M., Lumpkin, M. D., Mizunuma, H., Khorram, O., Ottlecz, A. & Samson, W. K. (1984) Trends Neurosci. 7, Ottlecz, A., Snyder, G. D. & McCann, S. M. (1988) Proc. Natl. Acad. Sci. USA 85, Aguila, M. C., Pickle, R. L., Yu, W. H. & McCann, S. M. (1991) Neuroendocrinology 54, Murakami, Y., Kato, Y., Tojo, K., Inoue, T., Yanaihara, N. & Imura, H. (1988) ndocrinology 88,