expression in the hypothalamic arcuate nucleus following systemic

GH-deficient dw/dw rats and lit/lit mice show increased Fos expression in the hypothalamic arcuate nucleus following systemic injection of GH-releasing peptide-6 S L

Dickson,

O Doutrelant-Viltart and G

Leng

Neuroendocrinology, The Babraham Institute, Babraham, Cambridge CB2 4AT, UK (O Doutrelant-Viltart is now at Laboratoires Neurosciences du Comportement, Université de Liile, 59655 Villeneuve d'Ascq Cedex, France) (G Leng is now at Department of Physiology, Medical School, University of Edinburgh, Teviot Place, Edinburgh EH8 9AG, UK) (Requests for offprints should be addressed to S L Dickson who is now at Anatomy and Human Biology Croup, King's College London, Strand, Laboratory

of

London VVC2R 2LS, UK)

Abstract

synthetic GH secretagogue GH-releasing acts centrally to activate a subpopulation of arcuate neurones as reflected by increased electrical activation and by the detection of Fos protein in cell

the number of cells expressing Fos protein in the arcuate nucleus (or in any other hypothalamic structure studied). These results support our hypothesis that GHRP-6 has a central site and mechanism of action and provide evidence to suggest that the activation of arcuate neurones by GHRP-6 is not mediated by a central action of GH or IGF-I. Furthermore, since the lit/lit mouse pituitary does not release GH following GHRP-6 administration, our finding that the central actions of GHRP-6 remain intact in these animals suggests the possible existence of two subpopulations of putative GHRP-6 receptors. journal of Endocrinology (1995) 146, 519\p=n-\526

Introduction

effects

In the rat, the

peptide (GHRP-6)

nuclei. Since GHRP-6 also induces GH secretion via a direct action on the pituitary, we set out to determine whether the central actions of GHRP-6 are mediated by GH itself. First, we demonstrated that peripherally administered GHRP-6 induces Fos expression in the arcuate nucleus of GH-deficient animals (dw/dw rats and lit/lit mice). Secondly, in dw/dw rats, neither intracerebro\x=req-\ ventricular injection of 15 \g=m\grecombinant bovine GH nor 1 \g=m\grecombinant human IGF-I resulted in an increase in

studies

demonstrated that

systemically hormone-releasing peptide (GHRP-6) activates a subpopulation of hypothalamic arcuate neurones, as reflected by increased electrical activity and by the increase in Fos-like immunoreactivity in this area (Dickson et al. 1993, 1995). Here we set out to investigate the possibility that the activation of arcuate neurones by GHRP-6 is indirect and mediated by a central feedback action of growth hormone (GH). Sys¬ temic administration of a large dose of GH induces c-fos expression in the arcuate nucleus of hypophysectomized rats (Minami et al. 1993); this may result from a direct central action of GH itself or may be mediated by the In

previous

administered

we

growth

GH-induced release of some other peripheral factor, for example, the GH-induced increase in plasma levels of

factors (IGFs). administration of the synthetic GHRP-6 elicits a large increase in plasma GH concentration in all species studied (Bowers et al. 1984). The actions of GHRP-6 appear to be selective for GH with only slight

insulin-like

Systemic

growth

to

increase

plasma adrenocorticotrophin

in

some

al. 1990, Hickey et al. 1994). The site and mechanism of the action of GHRP-6 (and of various peptide and non-peptide mimetics) is currently the focus

species (Bowers

et

of extensive investigation since they appear to act via a different mechanism from that of the two principal neuroendocnne systems controlling GH secretion, the stimula¬ tory GH-releasing hormone (GRF) neurones (Guillemin et al. 1982, Rivier et al. 1982) and the inhibitory somato¬ statin neurones (Brazeau et al. 1973). The neurochermcal identity of the arcuate neurones activated by GHRP-6 is unresolved, but the distribution of these cells suggests that they include GRF neurones (Jacobowitz et al. 1983). Systemic injection of the GHRP-6 mimetic hexarelin stimulates GRF release into the portal blood of the sheep (Guillaume et al. 199A) and, in rats, the infusion of GRF antiserum attenuates the GH response to GHRP-6 (Clark et al. 1989) or hexarehn (Conley et al. 1995). At the pituitary, GHRP-6 binds to a different receptor site from GRF (Codd et al. 1989) and synergizes with GRF when administered either in vitro (Cheng et al. 1989) or in vivo (Bowers et al. 1984). However, the

administration of GRF antiserum attenuates the in vivo GH response to GHRP-6 (Clark et al. 1989), indicating that the actions of GHRP-6 in vivo are not completely independent of GRF. Furthermore, the GH-deficient dwarf lit/lit mouse, a mutant having a point mutation in the N-terminal ligand-binding domain of the GRF receptor (Lin et al. 1993), does not respond to GRF (Jansson et al. 1986e) or to GHRP-6 (jansson et al. 1986è), suggesting that the pituitary actions of GHRP-6 require the presence of a functionally intact GRF receptor. Many studies investigating the central feedback actions of GH have employed animals with GH deficiency resulting from a spontaneous mutation, such as the lit/lit mouse (Eicher & Beamer 1976) or the dw/dw rat (Charlton et al. 1988) since, unlike the hypophysectomized rat model, the deficiency is selective for GH. The pituitary deficiency in the dw/dw rat appears to be different from that of the lit/lit mice; the dw/dw rats have low but detectable levels of GH (Charlton et al. 1988) and are able to elicit very tiny GH responses to GRF (Carmignac & Robinson 1990) and GHRP-6 (ICAF Robinson, per¬ sonal communication) suggesting that the GRF receptor and the putative GHRP receptor must be functionally intact in these animals. The aim of the present study was twofold: (1) to test the hypothesis that the hypothalamic actions of GHRP-6 are mediated by a central feedback action of GH, and (2) to determine whether the lit/lit mouse hypothalamus is unresponsive to GHRP-6 (as reported previously for the pituitary of this mutant; Jansson et ai. 19866). Thus, we investigated whether systemically administered GHRP-6 results in an increase in the number of cells expressing Fos protein in the arcuate nucleus of GH-deficient animals (lit/lit mice and dw/dw rats). Furthermore, we sought evidence for a direct central action of GH or IGF-I by examining the distribution of Fos protein in the hypothala¬ mus of dw/dw rats injected intracerebroventricularly (i.c.v.) with either of these agents. Materials and Methods Two strains of animals

were used (both from the colonies): (1) adult male GH-deficient dwarf rats (dw/dw; 150—220 g body weight; 16—26 weeks old) and (2) adult male homozygous GH-deficient little mice (lit/lit; 25-40 g body weight; 20-30 weeks old) and their non-GH-deficient heterozygous littermates (+ /lit; 35—50 g body weight; 20—30 weeks old). All animals were maintained by non-barrier methods in a controlled

Babraham

figure 1. Fos-like immunoreactivity in the non-GH-deficient + Hit mice killed 90 min Section thickness=40 µ .

environment (14 h hght: 10 h darkness, 21-22 °C) and pelleted food and water were available ad libitum. In the first study, dw/dw rats were anaesthetized with an i.p. injection of tribromoethanol/amyl hydrate (8 ml/kg) for placement of jugular catheters. Two days later, conscious rats were injected i.v. with 50 µg GHRP-6 (dissolved in saline; a gift from Dr R Smith, Merck Research Laboratories, New York, NY, USA) or an equal volume of saline vehicle (0-2 ml). The lit Hit mice and their +/lit littermates were injected i.p. with either 300 µg GHRP-6/kg or an equal volume of sahne vehicle (0-2 ml). In the second study, dw/dw rats were anaesthetized with tribromoethanol/amyl hydrate (8 ml/kg i.p.) for placement of i.c.v. cannulae in the lateral ventricle (co-ordinates: 16 mm lateral, 0-6 mm posterior to bregma, 4-5 mm below skull; Paxinos & Watson 1982). Five days later, the conscious dw/dw rats were injected i.c.v. with 15 µg recombinant bovine GH (rbGH; a gift from Dr F Adnaens, Monsanto pic, Basingstoke, Hants, UK), 1 µg recombinant human IGF-I (rhIGF-I; Ciba Geigy, St Aubin, Switzerland) or an equal volume (2 µ ) of vehicle (0-05 m acetic acid, 0·05 mg BSA/ml in saline). Ninety minutes after injection, all animals were terminally anaesthetized with sodium pentobarbitone (60 mg/kg i.p.) and perfused transcardially with heparin¬ ized isotonic saline (5 min) followed by 4% paraformalde¬ hyde in 0-1 m phosphate buffer (PB; pH 74; 15-20 min). Brains were removed and post-fixed for 2 h in the same fixative containing 15% sucrose. They were then trans¬ ferred to a 30% sucrose solution in PB for 24 h. Brains sections were cut on a shding were frozen and 40 µ microtome. Endogenous peroxidases were deactivated in PB containing 20% methanol, 0-2% Triton-XlOO and 1-5% hydrogen peroxide for 15 min. Sections were incu¬ bated with a rabbit polyclonal anti-fos antibody (directed against amino acids 4 to 17 of human Fos protein; Ab-2; PC05 from Oncogene Science, New York, NY, USA; 1:1000 diluted in 1% normal sheep serum/0-3% TritonX100/0-1 m PB) for 24 h at 4 °C. Following this, sections were incubated in peroxidase-labelled anti-rabbit IgG (Vector, Peterborough, Cambs, UK; 1:500 diluted in 1% normal sheep serum/0-3% Triton-X100/01 m PB) for 24 h at 4 °C. The reaction product was visualized using a nickel intensified diaminobenzidine reaction (adapted from Shu et al. 1988). In the control for the experiment the primary antibody was omitted. Preabsorption of this antibody with the N-terminal Fos peptide has been reported previously (Murphy et al. 1991). All chemicals for

immunocytochemistry

were

supplied by Sigma (Poole,

Dorset, UK) unless otherwise stated. mean

number of

Fos-positive

For each rat, the nuclei per section was

nucleus of the hypothalamus of conscious male GH-deficient lit/lit mice and following an i.p. injection of 300 µg GHRP-6/kg or sahne vehicle. 3v=third ventricle. arcuate

2. Fos-like immunoreactivity in the arcuate nucleus of the hypothalamus of conscious male GH-deficient dwldw rats killed 90 min following an i.v. injection of (A) GHRP-6 or (B) saline vehicle. 3v=third ventricle. figure

calculated (1) through the entire length of the arcuate nucleus (11—15 sections per rat, and 6-10 sections per mouse), extending forward to where the median eminence joins the third ventricle and posteriorly to the region of the postenor periventricular nucleus. The mean count per section was then averaged for each experimental group and expressed as mean ± s.e.m. nuclei/section per animal. The non-parametric Mann-Whitney U-test was used to make a statistical comparison of the mean nuclei/ section per animal for control vs GH secretagogue-injected experimental groups. For each brain, the paraventricular nucleus of the thalamus was also examined as a positive control area, which invariably expresses a few Fos-positive nuclei under all experimental conditions.

Results In GH-deficient lit/lit mice and non-GH-deficient +/lit

mice, i.p. injection of GHRP-6 induced

expression nucleus,

as

c-fos gene in a subpopulation of cells in the arcuate reflected by the detection of Fos protein in cell

nuclei (Fig. 1). The distribution of Fos-positive nuclei was similar to that described previously for the rat (Dickson et al. 1993, see Fig. 2 for dw/dw rat): nuclei were scattered throughout the arcuate nucleus, extending forward to the region where the third ventricle joins the median eminence and posteriorly towards the posterior peri¬ ventricular nucleus. Within the arcuate nucleus, a notable cluster of nuclei was observed in the ventromedial portion, at the tip of the third ventncle. Unlike the rat, there was a fine ridge of Fos-positive nuclei extending along the interface between the third ventricle and the median eminence. Importantly, the number of Fos-positive nuclei counted was almost identical between the lit/lit and the +/lit mice: mean ± s.e.m.=63-5 ±23-2 nuclei/section per mouse (tt=4) and 78-9 ±23-1 nuclei/section per mouse (n=A) respectively. In each case the incidence of Fos-immunoreactive nuclei was significantly higher than for vehicle-injected controls (6-4 ±4-3 nuclei/section per mouse for lit/lit mice (n=A) and 3-7 ±2-4 nuclei/section per mouse for +/lit mice (n=4); P