Intact vagal afferent neurons are required for the satiety effects of the intestinal hormone cholecystokinin (CCK) and the orexigenic effects of the gastric regulatory peptide ghrelin. In this study, we examined the localization of ghrelin receptors in nodose ganglia and their function in regulating the expression of other orexigenic receptors, notably cannabinoid (CB)-1 and melanin-concentrating hormone (MCH)-1 receptors. With the use of RT-PCR, transcripts corresponding to both functional [growth hormone secretagogue receptor (GHS-R)1a] and truncated forms (GHS-R1b) of the ghrelin receptor were detected in rat nodose ganglia. There was no difference in expression between rats fed ad libitum or fasted for up to 48 h. Immunohistochemical studies using antibodies directed at GHS-R1a revealed expression in over 75% of neurons also expressing CCK-1 receptors in the mid- and caudal regions of the ganglion. There was also expression in human nodose ganglia. In fasted rats in which CB-1 and MCH-1 receptor expression was increased, administration of ghrelin prevented the downregulation by refeeding. We conclude that the actions of CCK and ghrelin are mediated by a common population of vagal afferent neurons. Ghrelin may act to limit the action of CCK in depressing expression of CB-1 and MCH-1 receptors and other receptors.
- vagus nerve
- melanin-concentrating hormone
the regulation of food intake is determined by a variety of endocrine signals arising in the periphery including the gastrointestinal tract, endocrine pancreas, and adipocytes (2, 13). These signals may act either directly on central nervous system neurons, notably in the hypothalamus, or may act indirectly via vagal afferent neurons and stimulation of ascending pathways from the brain stem to hypothalamus (13). For example, gastric distension inhibits food intake via vagal afferent neurons through mechanisms that appear to be independent of nutrient status. In addition, the intestinal hormone cholecystokinin (CCK), which is released by the ingestion of fat and protein, also inhibits food intake by activation of vagal afferent neurons (12, 14, 23, 25, 27).
Until recently, the main gastrointestinal signals regulating feeding behavior were considered to increase satiety. It is now clear, however, that there are also gastrointestinal signals that stimulate appetite. These include endogenous cannabinoids (CBs), such as anandamide, and peptides, such as ghrelin (3, 15, 19). The latter is produced in X- (or A-)like cells of the gastric corpus (9, 22); it is released during the interdigestive period and before the onset of feeding and falls during the postprandial period (8). Experimental studies (1, 10, 11) in rats have indicated that the orexigenic effects of exogenous ghrelin are dependent on an intact vagus nerve. Thus lesioning of vagal afferent neurons by capsaicin blocked the effect of ghrelin, and electrophysiological studies have suggested that ghrelin decreases the discharge of vagal afferent neurons.
The actions of ghrelin are mediated by the seven-transmembrane domain, G protein-linked, growth hormone secretagogue receptor (GHS-R) encoded by a single gene that is well conserved in the rat, mouse, pig, and human (17, 22, 24, 29). In mice in which the GHS-R gene has been deleted by homologous recombination, exogenous ghrelin does not stimulate food intake (33). Alternative mRNA processing of the GHS-R gene transcript generates two products: a functional receptor of 364 and 366 amino acids (GHS-R1a) in the rat and human, respectively, encoded by two exons; and a truncated receptor of 289 amino acids (GHS-R1b) that is encoded by a single exon, has five transmembrane domains, and is thought to be nonfunctional (24). The 1–265 amino acid sequence is identical in the two receptors, but in GHS-R1b the COOH-terminal 24 amino acids differ from those in GHS-R1a. Expression of GHS-R has been described in rat nodose ganglia compatible with the requirement for intact vagal afferent neurons for the orexigenic effects of intraperitoneal ghrelin (10). In the present study, we sought to define the cell population that expressed the receptor. Because our previous work (4, 6) has shown that the expression of CB-1 and melanin-concentrating hormone (MCH)-1 receptors by vagal afferent neurons is increased by energy restriction and decreased by CCK, we examined whether GHS-R expression was dependent on food intake; we also asked whether ghrelin might regulate the expression of CB-1 and MCH-1 receptors.
MATERIALS AND METHODS
Adult male Wistar rats (250–350 g) maintained on a 12:12-h light-dark cycle were used. Rats were killed by carbon dioxide inhalation, and nodose ganglia were either fixed in 4% paraformaldehyde (PFA) or immersed in RNA-Later (Ambion; Austin, TX).
On the basis of previous studies (4, 6) in which expression of CB-1 and MCH-1 receptors was shown to be increased by food restriction and decreased by food intake, in some experiments, animals were fasted for up to 48 h with water ad libitum before the recovery of nodose ganglia. Some fasted rats were then refed for up to 5 or 16 h; others received rat ghrelin (Tocris Bioscience; Avonmouth, UK) by intraperitoneal injection (4 nmol/kg) or PBS and were killed up to 5 or 16 h later. Nodose ganglia were dissected and processed as described below; in a few experiments, the gastric corpus was taken for comparison.
The human nodose ganglion was obtained from tissue removed during radical dissection of the neck for removal of a glomus tumor (3). Samples of human vagal nerve trunks were obtained during resections for gastroesophageal carcinoma and cleanly dissected from the serosa. After resection, the ganglion or nerve trunks were rapidly divided into several segments and immersed in RNA-Later or fixed in PFA for morphological studies. The work was approved by the Multi-Centre and Local Ethics Committee of Salford and Trafford Authority, and written consent was obtained from all patients.
Total RNA was extracted from pools of rat nodose ganglia (10–12 ganglia) and from human nodose ganglia and vagal trunks in TriReagent (Sigma; Dorset, UK). Briefly, samples were treated with DNAase, reverse transcribed, and processed for RT-PCR using BIOTAQ DNA polymerase (Bioline; London, UK). The primers used are listed in Table 1 (5). The quality of cDNA was confirmed by amplification of samples using primers to the GAPDH gene. PCR products were gel purified (MinElute gele extraction kit, QIAGEN; Crawley, UK) and sequenced directly or after cloning into the pGEM TEasy TA cloning vector (Promega; Southhampton, UK). PCR products or clones were sequenced in both directions using an automated dideoxy method.
Cryostat sections of rat and human nodose ganglia were rinsed in 0.1 M PBS, permeabilized in alcohol, and processed for single- or double-labeling immunofluorescence as described previously (3, 5). For detection of GHS-R1a, we used a rabbit anti-rat GHS-R antiserum (Alpha Diagnostic) raised against a peptide derived from the COOH-terminal region of rat GHS-R1a. In addition, affinity-purified goat polyclonal antibody to GHS-R (Santa Cruz Biotechnology; Santa Cruz, CA) was used, which reacts with the epitope shared by GHS-R1a and GHS-R1b. The CCK-1 receptor was localized using an affinity-purified rabbit polyclonal antibody raised against the NH2-terminal region of the rat CCK-1 receptor (a gift of the late John Walsh, Center for Ulcer Research and Education/Gastroenteric Biology Center, University of California, Los Angeles, CA). The vanilloid receptor (VR)-1 was detected by an affinity-purified goat polyclonal antibody to VR-1 (Santa Cruz Biotechnology). Both affinity-purified goat and rabbit polyclonal antibodies were used to detect CB-1 (Santa Cruz Biotechnology). For detection of the MCH-1 receptor, we used goat polyclonal antibodies raised against the COOH- or NH2-terminus of the human MCH-1 receptor or chicken polyclonal antibody (Acris Antibodies; Hiddenhausen, Germany). MCH was localized using affinity-purified goat polyclonal antibodies (Santa Cruz Biotechnology). Localization of the glial cell marker glial fibrillary acidic protein (GFAP) was made using an affinity-purified goat polyclonal antibody (Santa Cruz Biotechnology). Secondary antibodies were used as appropriate and included fluorescein (FITC)-conjugated (AffiniPure) donkey anti-rabbit, anti-goat, or anti-chicken IgG and Texas red-conjugated (AffiniPure) donkey anti-rabbit or anti-goat IgG (Jackson ImmunoResearch Laboratories; West Grove, PA). Specificity of immunostaining was determined by omitting the primary antibody and by preincubation with an excess of appropriate peptide. Sections were examined using an Axioplan Universal microscope, and images were processed using the Axio Vision 3.0 Imaging system with deconvolution options (Carl Zeiss Vision; Jena, Germany). The quantification of neurons coexpressing GHS-R, VR-1, CCK-1, CB-1, or MCH-1 receptors was made using nodose ganglia from four rats. Sections passing through the full length of the caudal and midregions of the ganglia were selected, and the appropriate double immunostaining and counting was performed on 5 sections/ganglion, with the sections being separated by 90 μm. Results are expressed as percentages of cells counted that expressed one or the other epitope or both epitopes and are presented as means ± SD.
In situ hybridization.
Cryostat sections (10 μm) of quickly frozen (−70°C) rat nodose ganglia were thaw mounted on UV-treated poly-l-lysine-coated slides, fixed in 4% PFA in 1× PBS, and acetylated in 0.25 M acetic anhydride-0.1 M triethanolamine (10 min) as described previously (4). Sections were dehydrated in 70–95% ethanol at 4°C and stored in 95% ethanol at 4°C or dried and stored at −70°C until required. Oligonucleotide probes complementary to bases 4–51, 349–396, and 952–999 of the rat CB-1 receptor (Sygma-Genosys) were 3′-end labeled with [35S]dATP (10 mCi/ml, Amersham Biosciences; Buckinghamshire, UK) and used at a concentration of 3,000 counts·min−1·μl−1 in hybridization buffer [50% formamide, 10% dextran sulphate, 4× SSPE (1× SSPE contained 3 M NaCl, 0.2 M NaH2PO4·2H2O, and 0.2 M EDTA; pH 7.4), 0.2 mg/ml sheared salmon sperm DNA, 0.1 mg/ml polyA RNA, 5× Denhardt's, and 20 mM DDT]. Hybridization with mixtures of labeled oligonucleotide probes (1:1:1) was performed overnight at 42°C; sections were washed in 1× SSC (150 mM NaCl and 15 mM sodium citrate; pH 7.0) at 55°C (30 min), washed again in.1× SSC (30 s), dehydrated through ethanol, air dried, coated with autoradiographic emulsion (LM-1, Amersham Biosciences), and exposed for 4–6 wk before development. Sections were developed, counterstained with hematoxylin-eosin or toluidine blue, dehydrated, and mounted in Histomount (Oncogene Research; Boston, MA). Silver grains were visualized using an Axioplan Universal microscope. Images were processed using the ArxioVision 3.0 Imaging system (Carl Zeiss Vision) combined with dark- and bright-field illumination. Control slides were hybridized with a 100-fold excess of unlabeled oligonucleotides.
Identification of GHS-R1a and GHS-R1b transcripts in rat nodose ganglia.
In initial studies, we characterized GHS-R transcripts in rat nodose ganglia (Fig. 1). Primers specific for GHS-R1b or that did not distinguish between GHS-R1a and GHS-R1b revealed clear bands of the expected size in rat nodose ganglia extracts; in both cases, the bands were of similar intensity using cDNA from the nodose ganglia of either fasted rats or rats fed ad libitum. Primers specific for GHS-R1a also revealed bands of the predicted size, although these were faint compared with those obtained with other primer sets. The identity of the RT-PCR products was verified by sequencing. In contrast, there was no RT-PCR signal in rat nodose ganglia using primers specific for ghrelin (Fig. 1).
Immunohistochemical localization of GHS-R1a and -R1b in rat nodose ganglia.
In immunohistochemical studies, antibodies specific for GHS-R1a or for shared sequences in GHS-R1a and GHS-R1b revealed expression throughout the length of the nodose ganglion (Fig. 2); with the former, 20 ± 2% of all neurons in the mid- and caudal regions of the ganglion exhibited immunoreactivity compared with 25 ± 3% for the latter (means ± SE, n = 4 rats in both cases).
Previous studies (4, 6) have shown that CCK-1 receptor-immunoreactive neurons also express CB-1 and MCH-1 receptors, and, compatible with this, we found that ∼75% of GHS-R1a-immunoreactive neurons also expressed CB-1 and MCH-1 receptors (Figs. 3 and 4). There was no difference using antibodies specific for GHS-R1a or to the common epitope in GHS-R1a and -R1b (Fig. 3).
To characterize the neuronal population expressing GHS-R, we performed colocalization studies using antibodies to the VR-1 receptor. Approximately 50% of neurons expressing VR-1 also expressed GHS-R, and <5% of labeled neurons expressed GHS-R but not VR-1 (Figs. 2 and 4); similar results were obtained using antibodies to either GHS-R1a or to shared epitopes on GHS-R1a and -R1b (not shown). The majority of the GHS-R-immunoreactive neurons also expressed CCK-1 receptors, and this accounted for nearly 75% of the total CCK-1 receptor-immunoreactive population (Figs. 2 and 4).
Human nodose ganglia.
In view of neurochemical differences between rat and human nodose ganglia (3–5), we sought to determine the expression and cellular localization of GHS-R in human nodose ganglia. By RT-PCR, we identified bands of the predicted size corresponding to GHS-R1a and GHS-R1b in human nodose ganglia and the vagal nerve trunk. As in the rat, the intensity of GHS-R1a bands was relatively low (Fig. 5). We then showed by immunohistochemistry that GHS-R-immunoreactivity was localized to cell soma corresponding to populations of both small and large neurons. Interestingly, an antibody specific for GHS-R1a revealed expression only in small neurons. However, unlike receptors for the orexigenic petide orexin (Ox-R1), which are expressed by both neurons and glial cells (3), there was no colocalization with a glial cell marker (GFAP) in human nodose ganglia (Fig. 5).
Effects of ghrelin on CCK-induced gene expression.
Previous work (4, 6) has shown CCK decreases the expression of CB-1, MCH-1, and MCH in nodose ganglia, and because ghrelin inhibits vagal afferent nerve discharge, we then asked whether ghrelin influenced the expression of CB-1, MCH-1, and MCH. Administration of ghrelin to rats fed ad libitum did not increase the abundance of CB-1, MCH-1, or MCH transcripts as determined by RT-PCR either 5 or 16 h later. However, administration of ghrelin inhibited the decrease in abundance of CB-1, MCH-1, and MCH transcripts that occurred when fasted rats were refed. Thus 5 h after refeeding rats that had been fasted for 48 h, ghrelin (administered at the time of refeeding) preserved CB-1 receptor, MCH-1 receptor, and MCH transcripts as shown by RT-PCR (Fig. 6). Similarly, immunohistochemical studies (Fig. 7) showed that CB-1, MCH-1, and MCH protein were depressed 5 h after the refeeding of vehicle-treated animals but preserved by the administration of ghrelin at the time of refeeding. In contrast, the administration of ghrelin to animals fed ad libitum had no effect on the expression of CB-1, MCH-1, or MCH expression as detected by immunohistochemistry (Fig. 7). Moreover, in situ hybridization studies also verified that the loss of CB-1 mRNA in nodose neuron cell soma after the refeeding of fasted rats was inhibited by the administration of ghrelin at the time of refeeding (Fig. 8).
The present study shows that the functional form of the ghrelin receptor, GHS-R1a, is expressed by both human and rat nodose ganglion neurons. The neuronal population expressing the receptor also expresses other orexigenic receptors, notably CB-1 and MCH-1. These neurons have previously been shown to also express CCK-1 and leptin receptors, and they appear to account for roughly 50% of the VR-1-positive neurons in the mid- and caudal regions of the nodose ganglion (3–6). Unlike CB-1 and MCH-1 receptors, the expression of GHS-R appears not to be increased by energy restriction. However, the administration of ghrelin counteracted the downregulation of CB-1 and MCH-1 receptor expression that occurs when fasted rats are refed. The results therefore raise the possibility of a new role for ghrelin in modulating the expression of orexigenic receptors by nodose ganglion neurons.
In recent years, there has been considerable progress in the neurochemical characterization of vagal afferent neurons. In addition to receptors for satiety hormones such as CCK, leptin, and PYY3-36 (5, 7, 21), it is also clear that these neurons express receptors for several orexigenic factors (CB-1, Ox-R1, and MCH-1) as well as GHS-R. Moreover, they also express metabotrophic glutamate receptors and GABAB receptors (20, 28, 31). These neurons are therefore likely to be the site of extensive interactions between different neurotransmitters and hormones. Previous studies (3, 4) have identified several important differences between human and rat nodose ganglia. In particular, in human nodose ganglia, there are distinct populations of both large and small neurons, with extensive satellite cells surrounding the cell soma of large neurons. Both satellite cells and the two populations of neuron appear to express Ox-R1, whereas in the rat, Ox-R1 is expressed only in neurons (3). In the present study, we found both small and large neurons, but not glial cells, expressed GHS-R in human nodose ganglia; this distribution is therefore distinct from that of Ox-R1, CB-1, and MCH-1 receptors (4, 6). These differences in the pattern of expression of receptors associated with orexigenic signaling need to be taken into account in considering the possible implications for humans of experimental studies in the rat.
Previous studies (3, 10, 11) have shown that ghrelin and orexin A have acute effects in inhibiting vagal afferent nerve discharge. Because CCK stimulates vagal afferent discharge (30), the data are compatible with the idea that vagally mediated orexigenic effects of peripheral ghrelin and orexin A involve suppression of the effects of CCK. The present data extend these findings to include stabilization of the expression of CB-1 and MCH-1 receptors and MCH into the postprandial period. They indicate potentially longer-lasting effects on vagal afferent nerve function than are suggested by the acute responses shown by electrophysiological studies.
In the immediate postprandial period, there are declining plasma concentrations of ghrelin and rising concentrations of CCK (8, 25). There is therefore the potential for interactions between the two, and we propose that such interactions determine CB-1 receptor, MCH-1 receptor, and MCH abundance after prolonged fasting. We found no evidence that ghrelin was able to increase the expression of CB-1, MCH-1, or MCH in rats fed ad libitum. It is generally thought that CCK acts in the short term to limit meal size (27). However, the satiety effects of CCK are reduced after prolonged energy restriction (26, 32). Interactions between CCK and leptin at the level of the vagus nerve may partly account for the decreased satiety effects of CCK (26), but increased expression of CB-1 and MCH-1 receptors and MCH in vagal afferent neurons after energy restriction has also been implicated (4, 6). The present data suggest an additional mechanism by which ghrelin inhibition of the vagal actions of CCK would further reduce the satiety response to CCK after limited energy availability.
Stimulation of vagal afferent pathways has been recognized to control autonomic reflexes including, for example, stimulation of the vagovagal reflex leading to relaxation of the body of the stomach and inhibition of gastric emptying. It is notable that ghrelin stimulates both food intake and gastric emptying (18), whereas CCK inhibits food intake and gastric emptying (13). The GHS-R has high constitutive activity (16) and so presumably maintains a tonic restraint on CCK-1 receptor activation that is augmented in the interdigestive and preprandial phases by elevated plasma ghrelin. Our data suggest that GHS-R signaling is also be able to limit the action of CCK in downregulating expression of other receptors associated with orexigenic actions in vagal afferent neurons. Further work is now needed to determine the precise cellular mechanisms involved.
We are grateful to the Biotechnology and Biological Sciences Research Council and the Medical Research Council for grant support.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2006 the American Physiological Society