Distinct Forebrain and Caudal Brainstem Contributions to the Neuropeptide Y Mediation of Ghrelin Hyperphagia

Male Sprague-Dawley rats (Charles River Laboratories, Wilmington, MA) weighing 295–330 g at surgery were housed in hanging stainless steel cages under a 12-h light/12-h dark cycle (lights on at 9:00 a.m.). Pelleted Chow (Ralston Purina, St. Louis, MO) and water were available ad libitum. The experimental protocols used conform to institutional standards of animal care and the Guide for the Care and Use of Laboratory Animals (National Research Council, 1996).

Rats were anesthetized with a mixture of ketamine (9 mg/kg) and xylazine (1.5 mg/kg i.m.). Rats received a 22-gauge guide cannula (Plastics One, Roanoke, VA) implanted 2 mm above an injection site in the third and/or fourth ventricle (experiments 1 and 4), and for experiments 2, 3, 5, and 6, an additional 19-gauge guide cannula was implanted into the cerebral aqueduct. The third-ventricle cannula was positioned on the midline, 2.0 mm posterior to bregma and 7.7 mm below the skull surface. Fourth ventricle coordinates were on the midline, 2.5 mm anterior to the occipital suture and 5.2 mm below the skull surface. The aqueduct cannulas were angled 11° in the lateral-midline direction and guided to a position on the midline, 7 mm posterior to bregma and 5 mm below the dura. The cannulas were attached to the skull with jeweler’s screws and dental acrylic and closed with obturators.

Ventricular cannula placements were evaluated functionally after at least 7 days of recovery from surgery through measurement of the sympathetically mediated hyperglycemic response to 210 μg of 5-thio-d-glucose in 2 μl of artificial CSF (31). Only rats that showed at least a twofold increase of plasma glucose level in response to this treatment were used in the experiments.

Rat ghrelin (Phoenix Pharmaceuticals, Belmont, CA) and the NPY receptor antagonists 1229U91 (Neosystems) and d-Tyr^27,36, Thr³² NPY_27–36 (Bachem, King of Prussia, PA) were dissolved in artificial CSF and stored at −80°C. The 150-pmol ghrelin dose is known to be a moderate, supra-threshold dose for the feeding response upon forebrain ventricular delivery (5). The 5-nmol dose of 1229U91 had been shown to reverse melanin-concentrating hormone-induced hyperphagia by 86% (38); a lower dose (2 nmol) reduced NPY-induced feeding by 48% (39). The 10-μg dose of d-Tyr^27,36, d-Thr³² NPY_27–36 was shown to reverse ghrelin-induced hyperphagia when delivered to the lateral ventricle (21).

Drugs and vehicle were delivered via a microsyringe (Hamilton, Reno, NV) with a 28-gauge injector that extended beyond the guide cannula and into the ventricle. All treatment conditions consisted of two injections: ghrelin + vehicle, NPY receptor antagonist + vehicle, vehicle + vehicle, or NPY antagonist + ghrelin, delivered in two 1-μl bolus injections separated by a 30-s pause. The injector was kept in place for an additional 30 s after the last bolus. All drug injections were delivered in the 2nd h of the light phase. On at least two occasions before experimental testing, rats were given vehicle injections to habituate them to the injection procedure. In all experiments, food intake (dish weight difference minus spillage) was measured 1.5, 3, and 24 h after injection.

For experiments in which cerebral aqueduct occlusion was employed (experiments 2, 3, 5, and 6), 5 μl of a 50:50 ratio of High Vacuum grease (Dow Corning) and Dialectic Connector grease (VersaChem) mixed with 0.1 μg of fast green dye (for visualization) was injected via a 100-μl Hamilton syringe into the cerebral aqueduct 1 h before drug injections on the aqueduct occlusion test day.

Aqueduct occlusion was verified functionally by 5-thio-d-glucose injection after testing. Rats with third intracerebroventricular cannulas were included in the analysis only when the injection failed to increase plasma glucose. For histological verification, black India Ink (2 μl) was injected through the cannula just before the rats were killed. Brains were removed and postfixed in a 10% sucrose-formalin solution. Mid-sagittal sections were cut and then examined under a dissection microscope for verification of aqueduct occlusion. Only data for animals showing no ink in the opposite ventricle were included in the analysis.

The study consisted of six experiments in two series of three experiments each: the first three (experiments 1, 2, and 3) used 1229U91, and the second three (experiments 4, 5, and 6) used d-Tyr^27,36, d-Thr³² NPY_27–36. The initial experiment in each series (experiments 1 and 4) consisted of two groups of rats, one with third intracerebroventricular cannulas (experiment 1a, n = 10; experiment 4a, n = 7) and one with fourth intracerebroventricular cannulas (experiment 1b, n = 13; experiment 4b, n = 7). All rats in each group received each of four treatment combinations (ghrelin + vehicle, NPY antagonist + vehicle, vehicle + vehicle, or NPY antagonist + ghrelin) delivered in counterbalanced order with 2 days between test sessions.

The second in each series (experiments 2 and 5) consisted of two groups of rats, one with third ventricle and cerebral aqueduct cannulas (experiment 2a, n = 12; experiment 5a, n = 14) and one with fourth ventricle and aqueduct cannulas (experiment 2b, n = 16; experiment 5b, n = 9). All rats received vehicle + vehicle as a baseline condition before insertion of the aqueduct occlusion. For the postocclusion treatment conditions, rats were divided into two subgroups receiving either ghrelin + vehicle or ghrelin + NPY antagonist.

The third in each series (experiments 3 and 6) consisted of two groups of rats, with all rats receiving three cannulas positioned, respectively, above the third and fourth ventricles and into the cerebral aqueduct. In experiment 3a, rats received ghrelin into the third ventricle and 1229U91 into the fourth ventricle (n = 6) or ghrelin into the third ventricle and vehicle into the fourth ventricle (n = 6) after aqueduct occlusion. In experiment 3b, rats received ghrelin into the fourth ventricle and 1229U91 into the third ventricle (n = 8) or ghrelin into the fourth ventricle and vehicle into the third ventricle (n = 8) after aqueduct occlusion. In experiment 6a, animals received ghrelin into the third ventricle and d-Tyr^27,36, d-Thr³² NPY_27–36 into the fourth ventricle (n = 5) or ghrelin into the third ventricle and vehicle into the fourth ventricle (n = 4). In experiment 6b, animals received ghrelin into the fourth ventricle and d-Tyr^27,36, d-Thr³² NPY_27–36 into the third ventricle (n = 6) or ghrelin into the fourth ventricle and vehicle (n = 7) into the third ventricle.

For experiments 1 and 4, two-way (ventricle × drug condition) repeated-measures ANOVAs were used to evaluate overall treatment effects, and Tukey’s honestly significant difference post hoc tests were used to evaluate pairwise intake differences among the four conditions. Separate analyses were conducted for intake measured 1.5 and 3 h after treatment. For experiments 2, 3, 5, and 6, reversal of ghrelin hyperphagia was assessed by a two-tailed independent-samples t test comparing the size of the ghrelin effect in rats receiving ghrelin alone after aqueduct occlusion (ghrelin + vehicle minus vehicle + vehicle) with the intake difference between the ghrelin + Y receptor antagonist condition and vehicle + vehicle baseline.

Figure 1 shows the values for the four treatment conditions (vehicle + vehicle, ghrelin + vehicle, 1229U91 + vehicle, and ghrelin + 1229U91) delivered to the third (experiment 1a) and fourth (experiment 1b) ventricles. The respective ANOVAs (for each ventricle placement and each of the two time points) revealed a significant main effect of injection condition in all cases (P < 0.01). A hyperphagic response to ghrelin was obtained in each case (vehicle + ghrelin versus vehicle + vehicle; P < 0.01). At the 1.5-h time point for both third and fourth ventricle groups, the antagonist reversed the effect of ghrelin (i.e., ghrelin + 1229U91 < ghrelin + vehicle [P < 0.015], and there were no significant differences between ghrelin + 1229U91 and vehicle + vehicle conditions). For the fourth intracerebroventricular experiment, this reversal was sustained at the 3-h time point, whereas for the third intracerebroventricular conditions, there was no longer a statistically significant reversal.

The results with aqueduct occlusion (Fig. 2A) were in marked contrast to those of the open-ventricle study (experiment 1a). The ghrelin effect was apparent in both cases, but with occlusion, the antagonist failed to reverse the hyperphagic response; that is, the comparisons between the difference scores (ghrelin + vehicle minus vehicle + vehicle versus ghrelin + 1229U91 minus vehicle + vehicle) were not significant (1.5 h: t(10) = −0.04, NS; 3 h: t(10) = −0.60, NS). This result suggests that the Y1 receptors critical for the reversal of the third intracerebroventricular ghrelin response in the open-ventricle case reside in the CBS, an inference supported by the results of experiment 3a.

The pattern of results was comparable with that obtained from the corresponding open-ventricle study (experiment 1b). The reversal of ghrelin hyperphagia in this experiment (Fig. 2B) is shown by the significant difference between the size of the ghrelin effect and the intake difference between ghrelin + 1229U91 and its vehicle control values (1.5 h: t(14) = 3.38, P < 0.005; 3 h: t(14) = 2.95, P < 0.02).

Figure 3A (right) shows the reversal of the hyperphagic response to third intracerebroventricular ghrelin delivery at the 3-h time point with Y1 receptor antagonist delivery to the fourth ventricle (t(7) = −3.04, P < 0.02). For the 1.5-h time point, the difference was not statistically significant (P = 0.09).

The lack of a significant difference between values for the ghrelin + vehicle group and the ghrelin + Y1 receptor antagonist group (1.5 h: t(8) = −0.75, ns; 3 h: t(8) = −0.47, NS) (see Fig. 3B) indicates that forebrain Y1 receptor activation is not required for the hyperphagia obtained with CBS ghrelin delivery.

Figure 4 shows the results for the four treatment conditions delivered to the third (experiment 4a) and fourth (experiment 4b) ventricles. The results with the present antagonist were generally consistent with those described for 1229U91 (experiment 1). The ANOVAs (for each ventricle placement and each of the two time points) indicated a significant main effect of injection condition in all cases (P < 0.001). A hyperphagic response to ghrelin was obtained in each case (vehicle + ghrelin versus vehicle + vehicle; P < 0.01). For third intracerebroventricular delivery, the antagonist reversed the hyperphagic response to ghrelin at the 1.5-h time point (P < 0.03), but at 3 h, there was no longer a significant difference between the ghrelin + d-Tyr^27,36, d-Thr³² NPY_27–36 and ghrelin + vehicle conditions (P = 0.66). For the fourth ventricle (Fig. 4B), the trend toward reversal approached but did not reach statistical significance (P = 0.07) 1.5 h after treatment. By the 3-h time point, a highly significant reversal was obtained (P = 0.008).

As shown in Fig. 5, the antagonist coadministered with ghrelin significantly reversed the hyperphagic response for both third intracerebroventricular (experiment 5a: 1.5 h, t(12) = −3.70, P < 0.002; 3 h, t(12) = −3.82, P < 0.003) and fourth intracerebroventricular (experiment 5b: 1.5 h, t(6) = −2.81, P < 0.04; 3 h, t(6) = −2.29, P = 0.06) placements. The third intracerebroventricular results for d-Tyr^27,36, d-Thr³² NPY_27–36 contrast with those described above for 122U91 (experiment 2a), in which that antagonist failed to reverse ghrelin hyperphagia when the aqueduct was occluded.

For all cross-ventricle combinations, d-Tyr^27,36, d-Thr³² NPY_27–36 failed to reverse the hyperphagic response when the antagonist was delivered to the fourth ventricle and ghrelin to the third ventricle (experiment 6a) and when delivered to the third ventricle along with fourth intracerebroventricular ghrelin administration (experiment 6b) (Fig. 6). An anomalous finding was obtained for the 3-h intake measurement in experiment 6b, where the hyperphagic response was apparently enhanced by antagonist treatment (P < 0.03).

The present results confirm an important role for NPY systems in the downstream mediation of ghrelin hyperphagia but strongly qualify the often-posited view of an intrinsic hypothalamic circuit sufficient to explain the ghrelin-NPY interactions of relevance to food intake control. A broader neuroanatomical perspective is compelled by the existence of separate GHS receptor “triggers” for the hyperphagic response in the CBS as well as in the hypothalamus (see also 6), and the clear indication that NPY receptor stimulation within the CBS is required for the expression of the hyperphagic response to central ghrelin administration.

Perhaps the most salient counterexamples to the intrinsic hypothalamic model were provided by studies with the NPY-Y1 receptor antagonist 1229U91. A reversal of the hyperphagic response to ghrelin by 1229U91 delivery to the third ventricle was obtained, but only when the caudal flow of CSF was not obstructed (experiment 1a). Ghrelin hyperphagia persisted with aqueduct occlusion (experiment 2a), but the response was no longer affected by administration of the antagonist. It appears, then, that activation of hindbrain Y1 receptors is critical for the expression of the response to forebrain GHS receptor stimulation. This inference was further supported by the cross-ventricle study (experiment 3a) in which the hyperphagia triggered by third intracerebroventricular ghrelin was reversed by 1229U91 delivery to the fourth ventricle. Activation of CBS Y1 receptors (19) also appears to be necessary for the ingestive response stimulated by CBS ghrelin administration. This was indicated by complete reversal of the hyperphagic response when ghrelin and 1229U91 were coadministered to the fourth ventricle in experiments both with (experiment 2b) and without (experiment 1b) aqueduct occlusion. These results, taken together, indicate that endogenous activation of Y1 receptor within the CBS is necessary for the hyperphagic response triggered by either forebrain or hindbrain ghrelin treatment. Activation of forebrain Y1 receptors is not required for ingestive response to ghrelin delivery at either placement.

Although ghrelin-NPY interactions within the hypothalamus are not sufficient to fully account for the forebrain-elicited ghrelin response, work with the second receptor antagonist d-Tyr^27,36, d-Thr³² NPY_27–36 indicates that there is a within-hypothalamus interaction that is necessary for this response. In contrast to the result obtained with 1229U91, d-Tyr^27,36, d-Thr³² NPY_27–36 coadministered with ghrelin to the third ventricle with aqueduct occlusion (experiment 5a) reversed the hyperphagic response. Thus, the response arising from GHS receptor stimulation within the forebrain, interestingly, appears to require the simultaneous activation of NPY receptors in both forebrain and CBS but with different subtypes of relevance in the two regions: Y1 receptors within the CBS and a different NPY receptor subtype (36), perhaps the Y5 receptor, within the forebrain (25,40).

A different set of conclusions is reached for the NPY mediation of ghrelin hyperphagia elicited from the CBS placement. First, there appears to be no requirement for activation of forebrain NPY receptors; results of the cross-ventricle studies (experiments 3b and 6b) were negative. Within the CBS, activity at both sets of NPY receptors bound by the respective antagonists is necessary for the hyperphagic response. The fourth intracerebroventricular–elicited ghrelin response was reversed by Y1 receptor antagonist delivery to the CBS. This, in fact, is the common feature for response reversal for both intracerebroventricular placements, although it is not clear whether activation of the same or different subsets of CBS Y1 receptors is relevant to the responses driven by forebrain and CBS GHS receptor stimulation. For receptors bound by d-Tyr^27,36, d-Thr³² NPY_27–36, those relevant for the CBS-elicited response reside within the CBS proper. This was the clearest contrast between the respective responses to CBS and forebrain ghrelin delivery. This antagonist was effective when coadministered with ghrelin through the same intracerebroventricular cannula (fourth or third intracerebroventricular), but not effective when the ligands were delivered to different sides of the occluded aqueduct (experiments 6a and 6b).

Our analysis of the NPY receptors underlying ghrelin hyperphagia raises a set of questions concerning the location of the NPY neurons driven, respectively, by forebrain and CBS GHS receptor stimulation. It is possible that ARC neurons are required for all centrally elicited responses, regardless of the trigger site. The account would be straightforward for the third intracerebroventricular–elicited ghrelin response. Thus, ARC projections to neurons expressing receptors bound by d-Tyr^27,36, d-Thr³² NPY_27–36 (in, for example, PVN or lateral hypothalamus) could anchor an intrinsically hypothalamic component of the mediating circuit. Descending projections from ARC, although considered weak and not far reaching (41), may be sufficient to provide the requisite stimulation of Y1 receptors expressing neurons in the CBS. An ARC-based account of the fourth intracerebroventricular–elicited response, however, would appear less parsimonious. It would require “a long-loop arrangement,” given that the GHS and all relevant NPY receptors reside within the CBS. Here, ascending projections of unknown phenotype must engage ARC NPY neurons that would then, via descending projections, provide the presynaptic drive for both classes of relevant NPY Y receptor subtypes within the CBS.

Another limitation for the ARC NPY model for the mediation of ghrelin hyperphagia is the exclusion of a potential contribution from NPY neurons within the CBS (42,43), some of which, particularly those in the nucleus of the solitary tract, are of demonstrated relevance to the control of ingestive behavior (44,45). It is possible that the fourth intracerebroventricular ghrelin response is mediated by a circuit contained entirely within the CBS and also that hindbrain NPY neurons, by virtue of their local and ascending projection systems (46), contribute in some measure to the hyperphagic response to forebrain GHS receptor stimulation. A number of questions must be addressed before the mediating circuitry for centrally elicited ghrelin hyperphagia can be elucidated. Further attention should be directed toward the NPY neurons in the hindbrain, including investigation of possible GHS receptor coexpression, and of the location and neurochemical properties of behaviorally relevant projection sites elsewhere within the CBS. Functional-anatomical studies may prove particularly useful if coordinated with strategies allowing stimulation of restricted subpopulations of GHS receptors. For example, it should be possible in rats with cerebral aqueduct occlusion to ascertain whether NPY neurons in either or both CBS and ARC are activated after delivery of ghrelin to the third or fourth ventricle and to characterize the neuraxial location and phenotypes of NPY receptor–expressing neurons that are activated by these treatments.

In summary, the present results introduce a considerable degree of complexity for the NPY mediation of centrally elicited ghrelin hyperphagia, with the subtype and location of NPY receptors critical for the behavioral response differing, depending on whether the response was elicited by forebrain or hindbrain GHS receptor stimulation. The area of overlap for hyperphagia triggered in the separate locations was a requirement for activation of NPY-Y1 receptors in the CBS, whereas critical receptors bound by d-Tyr^27,36, d-Thr³² NPY_27–36 (possibly the Y5 receptor) were always local to (or at modest distances from) the site of ghrelin delivery. The cells of origin for the endogenous ligand mediating CBS- and forebrain-elicited ghrelin hyperphagia remain to be determined, with likely candidates including NPY neurons in the ARC and those in CBS nuclei such as the nucleus of the solitary tract. Also open is the question of which NPY neurons and receptor subpopulation(s) are most relevant under physiological conditions where both forebrain and CBS receptors are likely to be stimulated simultaneously. Some of the considerations raised in the present study are likely to be relevant for analysis of various neurochemical systems implicated in food intake and body weight control, particularly when a given hormone (e.g., leptin [47]) or peptide mediator (e.g., melanocortins [48,49]) yields comparable effects when delivered selectively to different regions of the brain.

While the current report was in process, we completed a full set of experiments with a third NPY receptor antagonist, l-152-804, a selective antagonist at the Y5 receptor subtype (Kantani et al., Biochem Biophys Res Comm 272:169–173, 2000). The results were comparable with those reported for d-Tyr^27,36, d-Thr³² NPY_27–36 in experiments 4–6 (i.e., reversal of ghrelin hyperphagia in same-ventricle studies and lack of effect with cross-ventricle administration). The results are consistent with the hypothesis that in addition to the Y1 receptor requirement, endogenous activation of the Y5 receptor (at the anatomical level of the pharmacologically stimulated GHS receptor) is critical for the expression of ghrelin hyperphagia.

H.J.G. has received support from National Institutes of Health Grant R01-DK-21397. J.M.K. has received support from National Institutes of Health Grant DK-42284.

We thank Rebecca Stack, Gina Gubitosa, and Joanna Wapner for their technical contribution to the studies.