The gut peptide ghrelin has been shown to stimulate food intake after both peripheral and central administration, and the hypothalamic arcuate nucleus has been proposed to be the major site for mediating this feeding stimulatory action. Ghrelin receptors are widely distributed in the brain, and hindbrain ghrelin administration has been shown to potently stimulate feeding, suggesting that there may be other sites for ghrelin action. In the present study, we have further assessed potential sites for ghrelin action by comparing the ability of lateral and fourth ventricular ghrelin administration to stimulate food intake and alter patterns of hypothalamic gene expression. Ghrelin (0.32, 1, or 3.2 nmol) in the lateral or fourth ventricle significantly increased food intake in the first 4 h after injection, with no ventricle-dependent differences in degree or time course of hyperphagia. One nanomole of ghrelin into either the lateral or fourth ventricle resulted in similar increases in arcuate nucleus neuropeptide Y mRNA expression. Expression levels of agouti-related peptide or proopiomelanocortin mRNA were not affected by ghrelin administration. These data demonstrate that ghrelin can affect food intake and hypothalamic gene expression through interactions at multiple brain sites.
- neuropeptide Y
ghrelin is a 28-amino acid peptide first identified in the rat stomach as an endogenous ligand for the growth hormone secretagogue receptor (GHS-R). It is secreted primarily from the gastrointestinal tract and circulates in human blood (10, 15). A role for ghrelin in the control of food intake is supported by data demonstrating that injections of ghrelin into the brain or the periphery induce food intake (15–17, 21, 25). Consistent with these data, plasma ghrelin levels peak immediately before the onset of a meal and fall rapidly once a meal has begun, indicating that endogenous ghrelin may be involved in meal initiation (8).
Although the GHS-R is expressed in multiple brain areas (13), the arcuate nucleus of the hypothalamus (ARC) has been suggested to be a direct central site for the feeding stimulatory actions of ghrelin. Evidence for such a site of action includes data demonstrating that the GHS-R is colocalized with arcuate neuropeptide Y (NPY) and agouti-related protein (AgRP), potent orexigenic peptides coexpressed in subsets of ARC neurons (5, 23). Ghrelin administration increases NPY and AgRP gene expression in the ARC, and, as demonstrated by electrophysiological studies (6, 7), ghrelin directly activates NPY/AgRP neurons. Ghrelin-induced feeding can be blocked by antibodies against NPY and AgRP or by a NPY Y1 receptor antagonist (16, 18). Finally, very low doses of ghrelin administered directly into the ARC stimulate food intake, and ARC-ablated rats do not increase food intake in response to exogenous ghrelin administration (9, 20).
Whether circulating ghrelin affects food intake by directly accessing arcuate sites remains unclear. Although the ARC is adjacent to pituitary sites without a blood-brain barrier, there are no data demonstrating that circulating peptides have direct arcuate access. Furthermore, although a blood-brain barrier transport system for ghrelin has been identified, this system favors a brain-to-blood rather than a blood-to-brain direction of transport (3). Finally, ghrelin-containing neurons within the hypothalamus have been identified, suggesting that local rather than circulating ghrelin may be the source for interactions with hypothalamic GHS-R.
Other potential sites for ghrelin action have been suggested. Ghrelin has been demonstrated to inhibit vagal activity, and vagotomized rats do not respond to the feeding stimulatory effects of ghrelin (1, 11, 18, 24), suggesting that ghrelin acts via the vagal afferents to stimulate food intake. Finally, Faulconbridge et al. (12) have demonstrated that ghrelin injected into the third or fourth ventricle stimulates food intake at identical doses and that both routes of administration result in decreased latency to the onset of feeding, suggesting that there may be a distributed system for ghrelin feeding actions. The present experiments were undertaken to further characterize the effects of hindbrain stimulation of GHS-Rs and determine whether stimulation of these receptors affects hypothalamic gene expression in ways similar to what has been reported following forebrain ghrelin administration. Similar patterns of gene expression following hind and forebrain administration of ghrelin would further support distributed sites of ghrelin action with a common hypothalamic mediation.
MATERIALS AND METHODS
Animals and surgery.
Male Sprague-Dawley rats (Charles River) weighing 250–300 g were individually housed in hanging wire cages and maintained on a 12:12-h light-dark schedule at constant temperature (lights off at 12:00 PM). All rats received ad libitum access to pelleted chow (Prolab RMH 1000) and water for 1 wk after shipment to allow for acclimation to the new environment and stabilization of growth rates. All procedures were approved by the institutional animal care and use committee at Johns Hopkins University.
After 1 wk, rats were weight matched and divided into two groups. One group was stereotaxically implanted with cannulas aimed at the lateral ventricle, and the other was implanted with cannulas aimed at the fourth ventricle. Briefly, rats were anesthetized with 1 ml/kg of a 4:3 mixture of ketamine (100 mg/ml) and xylazine (20 mg/ml) and placed in a stereotaxic instrument. For lateral ventricular cannulations, a 23-gauge guide cannula (10 mm in length) was inserted 1.8 mm posterior to bregma, 1.66 mm lateral to midline (with bregma and lambda at the same vertical coordinate), and 3.3 mm ventral to dura for the lateral ventricle group. Fourth ventricular cannulas were 23-gauge, 14 mm in length, and implanted on the midline of the skull, 11.6 mm posterior to bregma and 7.8 mm below the surface of the skull (18a). A 30-gauge, 11.5-mm (lateral) or 14.5-mm (fourth) obturator was placed in the guide cannula. The cannula was secured to the skull with four screws and dental acrylic. After surgery, rats received intramuscular penicillin (60,000 units), an intramuscular injection of the analgesic banamine (1.1 mg/kg), and 10 ml of saline subcutaneously. Testing did not commence until body weights reached presurgical levels. Lateral ventricular cannula placement was verified by assessing a drinking response to angiotensin II (ANG II) administration. Rats were intracerebroventricularly injected with 5 nmol ANG II, after which water intake was measured. Rats that drank at least 5 ml more than they drank after an intracerebroventricular saline injection (in 30 min) were deemed to have correct cannula placement and were used in the experiments (n = 13). Fourth ventricular cannula placement was verified through measurement of the hyperglycemic response to 210 μg of 5-thio-d-glucose in 2 μl of saline (18b).
Feeding responses to intraventricular ghrelin.
After recovery of presurgical body weights, each rat received injections of saline, as well as 0.32, 1.0, and 3.2 nmol of rat ghrelin in 2 μl of saline (Bachem). Doses were delivered in a counterbalanced order on different occasions with each injection separated by 5 days. Saline or ghrelin was delivered with a 30-gauge injector via Gilmont syringe. All injections were delivered at 8:00 AM (4 h before the onset of the dark phase). Before experimental testing, rats received two saline injections to habituate them to the procedure. On testing days, food was removed at 7:00 AM. After injection, preweighed food was returned and papers were placed under the cages to catch spillage. Food hoppers were weighed 1, 2, 4, and 24 h later. Food weights were adjusted to account for spillage collected at each time point.
One week after the final assessment of the feeding responses to ghrelin following lateral or fourth ventricular administration, rats were weight matched and divided into two groups per cannula location (2 groups with lateral and 2 groups with fourth ventricular cannulas) for assessment of the effects of ghrelin administration on hypothalamic gene expression. On the day of death, food was removed and one-half of the rats received an intraventricular injection of saline (2 μl), whereas the other half received an intraventricular ghrelin injection (1 nmol in 2 μl of saline). Two hours after injection, rats were killed. Brains were removed and rapidly frozen for subsequent analysis of hypothalamic gene expression using in situ hybridization techniques.
Cryosections and riboprobes.
Coronal sections (14 μm) through the paraventricular nucleus and ARC were taken via cryostat, mounted on Superfrost Plus slides (Fisher Scientific), and fixed with 4% paraformaldehyde. Three sections per brain were anatomically matched among animals for each hybridization assay in the same condition. The plasmids of NPY, proopiomelanocortin (POMC), and AgRP were linearized using appropriate restriction enzymes. Antisense riboprobes were labeled with 35S-labeled UTP (Amersham Pharmacia Biotech) using in vitro transcription systems with appropriate polymerases according to the manufacturer’s protocols (Promega) and purified using Quick Spin RNA columns (Roche Diagnostics) to yield a specific activity of 5 ×108 cpm/μg.
In situ hybridization.
Sections were treated with acetic anhydride and incubated in hybridization buffer containing 50% formamide, 0.3 M NaCl, 10 mM Tris·HCl (pH 8.0), 1 nM EDTA (pH 8.0), 1× Denhardt’s solution (Eppendorf), 10% dextran sulfate, 10 mM DTT, 500 μg/ml yeast tRNA, and 108 cpm/μl 35S-UTP. Treated sections were incubated at 55°C overnight. After hybridization, sections were washed three times in 2× SSC twice at 55°C and then washed twice for 15 min/wash in 0.1× SSC at 55°C. Slides were dehydrated and exposed with BMR-2 film (Kodak). Exposure time was 1 day.
In situ hybridization quantitative analysis.
Images obtained by in situ hybridization were analyzed using the National Institutes of Health Scion Image software. Autoradiographic images were first scanned using an Epson Professional Scanner and then quantified with Scion Image software that utilized autoradiographic 14C-microscales (Amersham Pharmacia Biotech) as a standard. Data for each animal were means of the product of hybridization area times density, with the background density subtracted from the three sections, reflecting the level of gene expression in a specific region. Data for each animal were normalized to those of saline-treated controls as 100%, and all of the data are expressed as means ± SE.
Food intake data were analyzed using two-way (ventricle × drug condition) repeated-measures ANOVA to evaluate overall treatment effects. Repeated-measures ANOVA on cumulative intakes with time and dose as repeated factors were followed by planned t comparisons such that the effects of each dose were compared with intake after saline injection. Data are represented as means ± SE. The in situ hybridization data were analyzed by comparing the level of expression to that of saline-treated controls by one-way ANOVA. Levels of expression following lateral and fourth ventricular ghrelin administration also were analyzed by one-way ANOVA to compare the effects of lateral vs. fourth ventricular ghrelin administration.
Cumulative food intake.
Ghrelin administration increased cumulative food intake in the first 4 h after injection regardless of injection site (P < 0.02). The cumulative food intake responses 1, 2, and 4 h after ghrelin administration in the lateral or fourth ventricles are depicted in Fig. 1, A and B, respectively. Rats given ghrelin in the lateral or fourth ventricle significantly increased food intake at all doses (0.32, 1.0, and 3.2 nmol) compared with their food intake after saline injection (P < 0.05 in all cases). There were no differences in intake 24 h after ghrelin administration compared with intake 24 h after saline treatment (data not shown).
Hypothalamic gene expression effect of lateral and fourth ventricular ghrelin administration.
Similar patterns of hypothalamic gene expression were obtained after lateral or fourth ventricular ghrelin administration. As depicted in Fig. 2, NPY mRNA expression in the ARC was significantly elevated after both lateral and fourth ventricular ghrelin administration. Lateral ventricular ghrelin resulted in a 72% increase in NPY mRNA expression levels relative to saline-treated levels (Fig. 2, A and B, respectively), and fourth ventricular ghrelin resulted in a 59% increase compared with NPY mRNA expression after saline injection in the fourth ventricle (Fig. 2, C and D, respectively). The magnitude of increase between the lateral ventricular and fourth ventricular ghrelin did not differ significantly from each other (Fig. 3). There were no significant differences in AgRP mRNA expression levels in the ARC afterghrelin administration into the lateral or fourth ventricle (Fig. 4). Similarly, there were no significant differences in POMC expression after ghrelin administration regardless of injection site (Fig. 5).
Our results provide additional evidence to support the view that there are both forebrain and hindbrain contributions to ghrelin-induced hyperphagia. Lateral and fourth ventricular ghrelin produced similar increases in short-term food intake, and both routes of ghrelin administration resulted in increased NPY mRNA expression levels in the ARC. Thus these data provide support for the hypothesis put forth by Faulconbridge et al. (12) that the ghrelin feeding system is a distributed one involving the brain stem GHS-R, in addition to forebrain GHS-R. The data further demonstrate that activation of either forebrain or caudal hindbrain GHS-R results in increased peptide expression in hypothalamic arcuate neurons, suggesting a final common pathway underlying the orexigenic effects of lateral and fourth ventricular ghrelin administration.
The ability of ghrelin to induce increased NPY mRNA expression in arcuate neurons is consistent with prior data demonstrating both c-fos activation in NPY-containing neurons (18, 22) and elevated arcuate NPY expression after both central and peripheral ghrelin administration (1, 18, 19, 22). Such data had been interpreted to suggest that the arcuate is a direct target of both peripheral and central ghrelin. The current demonstration that fourth ventricular ghrelin also induces increased arcuate NPY mRNA expression and does so to a similar extent as lateral ventricular administration is a novel finding and demonstrates that direct access to arcuate neurons is not necessary for ghrelin-induced increases in arcuate NPY mRNA expression.
A role for NPY in mediating the orexigenic actions of ghrelin has been demonstrated from both pharmacological and genetic studies. Prior administration of NPY antagonists specific for Y1 (16) or Y5 (2) significantly attenuates the ability of peripheral or central ghrelin to inhibit food intake. Furthermore, although an initial report had demonstrated the ability of chronic ghrelin administration to increase food intake and body weight in NPY-deficient mice (21), recent work (4) has demonstrated that the feeding stimulatory actions of ghrelin are significantly reduced in NPY knockout mice.
We did not measure significant increases in arcuate AgRP mRNA expression in response to either lateral or fourth ventricular ghrelin administration. Although there was a relative increase in arcuate AgRP after lateral ventricular administration, this increase did not reach statistical significance. Other investigators have reported significant increases in AgRP in response to central ghrelin administration (14, 19). This apparent discrepancy may be due to differences in ghrelin dosage and the timing of death. A role for AgRP signaling in mediating the orexigenic actions of ghrelin has been suggested by data demonstrating the ability of melanocortin agonists to block the increase in food intake and the lack of feeding response to ghrelin in combined MC3 and MC4 receptor knockout mice. However, ghrelin retains significant efficacy in AgRP knockout mice, suggesting that increased AgRP signaling itself may not be critical (4).
Dependence on arcuate signaling for the feeding stimulatory actions of lateral ventricular ghrelin has been demonstrated from experiments examining the effects of neonatal monosodium glutamate (MSG) lesions that ablate the arcuate nucleus. MSG-lesioned rats do not increase food intake in response to ghrelin administration (20). The necessity of the arcuate nucleus for the feeding stimulatory action of peripheral or fourth ventricular ghrelin has not been assessed.
The site of action for ghrelin-induced feeding remains unclear. Multiple mechanisms have been proposed. Data demonstrating the presence of GHS-R within the ARC and the ability of ghrelin to directly alter the electrophysiological activity of NPY- and POMC-containing arcuate neurons in in vitro preparations have led to the suggestion that both peripherally circulating and centrally administered ghrelin have direct access to arcuate nucleus sites to exert their effects on food intake (6, 7, 13, 23). However, distributed actions for distinct ghrelin pools also are a possibility. Thus peripheral ghrelin has been shown to reduce activity in the gastric afferent vagal fibers, and vagotomy blocks the ability of peripheral ghrelin to stimulate food intake, suggesting a peripheral vagal mediation underlying the actions of peripheral ghrelin (11). Ghrelin-containing neurons have been identified in a variety of brain areas, including a population of neurons within the periventricular hypothalamic area that innervate several hypothalamic nuclei, including the arcuate nucleus (7). Furthermore, ghrelin-like immunoreactive terminals have been shown to synapse on NPY-containing neurons (13). Thus the endogenous ligand for interacting with arcuate GHS-R may be of local hypothalamic origin. Whether there also are ghrelin-containing neurons within the dorsal hindbrain that may provide endogenous ligand for GHS-R around the fourth ventricle has yet to be determined.
The current results support the view that the ghrelin feeding system is a distributed one with multiple potential sites of interaction. The demonstration of comparable activation of arcuate NPY mRNA expression by forebrain and hindbrain ghrelin administration suggests that arcuate NPY activation may be a common final pathway in mediating the orexigenic actions of ghrelin but one that does not depend on direct ghrelin hypothalamic interactions.
This work was supported by National Institute of Diabetes and Digestive and Kidney Diseases Grant DK-19302 (to T. H. Moran).
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