help button home button Endocrine Society Endocrine Reviews
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Purchase Article
Right arrow View Shopping Cart
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow Request Copyright Permission
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Kalra, S. P.
Right arrow Articles by Kalra, P. S.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Kalra, S. P.
Right arrow Articles by Kalra, P. S.
Endocrine Reviews 20 (1): 68-100
Copyright © 1999 by The Endocrine Society

Interacting Appetite-Regulating Pathways in the Hypothalamic Regulation of Body Weight1

Satya P. Kalra, Michael G. Dube, Shuye Pu, Bin Xu, Tamas L. Horvath and Pushpa S. Kalra

Department of Neuroscience (S.P.K., S.P., B.X.), Department of Physiology (M.G.D., P.S.K.), and The University of Florida Brain Institute, University of Florida College of Medicine, Gainesville, Florida 32610; and Department of Obstetrics and Gynecology (T.L.H.), Yale University School of Medicine, New Haven, Connecticut 06510


    Abstract
 Top
 Abstract
 I. Introduction
 II. Neuroanatomical Substrate...
 III. Orexigenic Signals
 IV. Anorexigenic Signals
 V. Leptin
 VI. Summary
 References
 

I. Introduction
II. Neuroanatomical Substrate for Appetite Regulation
A. Arcuate nucleus
B. Ventromedial nucleus and lateral hypothalamus
C. Dorsomedial nucleus
D. Paraventricular nucleus and perifornical hypothalamus
E. Suprachiasmatic nucleus and the timing device
F. Overview of the neuroanatomical substrate associated with appetite control
III. Orexigenic Signals
A. Neuropeptide Y
B. Galanin
C. Endogenous opioid peptides
D. Melanin-concentrating hormone
E. Amino acids: glutamate and {gamma}-aminobutyric acid
F. Hypocretins/orexins
G. Morphological and functional links among orexigenic signals
IV. Anorexigenic Signals
A. Corticotropin-releasing hormone family of peptides
B. Neurotensin
C. Glucagon-like peptide-1
D. Melanocortin and agouti protein
E. Cocaine and amphetamine-regulated transcript
V. Leptin
A. An adipocyte signal
B. Site of action
C. Is leptin a physiological satiety signal?
D. Leptin signal transduction and leptin resistance
VI. Summary


    I. Introduction
 Top
 Abstract
 I. Introduction
 II. Neuroanatomical Substrate...
 III. Orexigenic Signals
 IV. Anorexigenic Signals
 V. Leptin
 VI. Summary
 References
 
THE PAST decade has witnessed an upsurge in our understanding of the hypothalamic regulation of appetite. Expression of appetite or the motivational drive toward an energy source is a highly regulated phenomenon in vertebrates. It is considered a cornerstone for maintenance of energy homeostasis and for rigidly guarding the body weight around a set point. Abnormalities in the onset, periodicity, duration, and magnitude of eating episodes generally underlie augmented appetite (1, 2, 3). Increased appetite, whether temporary, as seen clinically in transient bingeing, or permanent, invariably culminates in an increased rate of body weight gain and obesity (1, 4, 5, 6). On the other hand, anorexia due to psychobiological causes (7, 8) or in response to acute and chronic infections, inflammation, and trauma is followed by severe loss of body weight (9, 10, 11). There is now a growing recognition that expression of appetite is chemically coded in the hypothalamus (1, 12). A perceived corollary is that subtle and progressive derangement in this neurochemical signaling, produced by environmental, genetic, and hormonal factors, impels either hyperphagia or anorexia (1). This conceptual advance has led not only to the identification and characterization of a multitude of neurotransmitters/neuromodulators that either propagate and transmit, or terminate appetite-stimulating impulses, but also to the precise tracking of pathways containing these signal molecules and the intricate interconnections among them. Increased knowledge of the molecular events governing synthesis, release, and signal transduction sequelae of each of these appetite-regulating messengers has enhanced our awareness of the phenomenon of coexistence and corelease of these chemicals in the cross-talk that ensues in response to an ever changing internal milieu (1).

Consequently, information amassed during this decade has revised our views on the hypothalamic control of appetite and helped to detail the mechanistic attributes of locally derived signals in regulating energy homeostasis. These attributes have crystallized into the following broad categories: 1) Embedded in the networks controlling a multitude of hypothalamic functions (1, 13, 14), there is a distinct circuitry regulating appetite. This circuitry is composed of an interconnected network of pathways elaborating and emitting orexigenic and anorexigenic signals (1, 15). 2) The neurons producing these orexigenic and anorexigenic signal molecules are subject to modulation by the internal milieu comprised of a variety of hormonal and other biologically active molecules. In this respect, the recent identification of the adipocyte protein, leptin, has renewed interest in feedback mechanisms between adipocytes and the appetite-regulating hypothalamic circuitry (16, 17). 3) A cascade of temporally related neural events in various components of the appetite-regulating network (ARN) precedes feeding episodes. 4) Emerging evidence supports the involvement of a distinct neural device for the timely onset of appetite expression, and disintegration of this control may result in unregulated food consumption. 5) A deficit in availability of orexigenic signal(s) at the signal transduction level, whether temporary or permanent, can perturb the postsynaptic receptor dynamics that eventuate in hyperphagia and increased body weight gain indistinguishable from that produced by excessive production and release of orexigenic signals (1, 18, 19). 6) Coexistence and corelease of orexigenic signals (20, 21, 22), along with the redundant overlapping and interconnected orexigenic and anorexigenic signaling pathways within the hypothalamus (1, 12, 15, 23), provide a microenvironment wherein subtle perturbations shift signaling in favor of unregulated hyperphagia rather than anorexia. The central theme of this article is to critically review our understanding of these fundamentals underlying neural control of appetite and to collate the growing information on several newly identified messenger molecules. Emphasis is placed on the anatomical distribution of signal-producing pathways in the hypothalamus, the site and mode of action of peripheral signals, and the cellular and subcellular events underlying hyperphagia and obesity in experimental and genetic models. In doing so we will present a conceptual model, which encompasses a broad spectrum of appetite-regulating messenger molecules, to explain the dynamics of the neural circuitry involved in stimulation and inhibition of appetite.


    II. Neuroanatomical Substrate for Appetite Regulation
 Top
 Abstract
 I. Introduction
 II. Neuroanatomical Substrate...
 III. Orexigenic Signals
 IV. Anorexigenic Signals
 V. Leptin
 VI. Summary
 References
 
Historically, a few discrete nuclei in the basal hypothalamus have been viewed as crucial in the regulation of daily energy homeostasis, especially those sites connected with neural mechanisms affecting appetite. The inference that hypothalamic sites, such as the ventromedial nucleus (VMN), dorsomedial nucleus (DMN), paraventricular nucleus (PVN), and lateral hypothalamus (LH), contain neural mechanisms affecting ingestive behavior was based on the results of numerous studies employing either discrete lesions in the hypothalamus or surgical transection of neural pathways (24, 25, 26, 27, 28, 29, 30). With the exception of lesions in the LH, these experimental manipulations within the brain disrupted the daily food intake pattern to produce permanently enhanced appetite or phagia. The inference from these investigations, that appetite-regulatory mechanisms may be confined to a few morphologically well defined regions in the hypothalamus, generated voluminous information on the patterns of ingestive behavior and the attendant shifts in endocrine and autonomic systems in rats undergoing these experimental manipulations (3, 13, 26, 30). Fundamental questions regarding the site(s) of propagation of appetite-stimulating impulses and food-seeking behavior, the transmission pathways and target sites in causation of appetitive behavior, however, remained unanswered for quite some time.

Recent intensive research, first for identification of orexigenic and anorexigenic neurotransmitters in the hypothalamus, followed by identification of the neuronal sites of their production, release, and receptive fields, has changed the landscape of the neuroanatomical substrate underlying ingestive behavior. Also, evidence of the morphological relationships among these neurotransmitter/neuromodulator-producing neurons, and the fact that these neurons can coproduce more than one appetite-regulating signal (1, 13, 20, 21, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41), have strengthened the concept, enunciated earlier (1, 12), that a distinct interconnected circuitry operates locally in the hypothalamus to regulate ingestive behavior. The following are the currently known anatomical components of this hypothalamic circuitry (Fig. 1Go).



View larger version (30K):
[in this window]
[in a new window]
 
Figure 1. A diagrammatic representation of hypothalamic sites associated with appetite-regulating signal pathways on a sagittal section near midline of the rat brain. AC, Anterior commissure; OC, optic chiasm; PC, posterior commissure. Synthesis, Hypothalamic sites involved in synthesis of orexigenic and anorexigenic signals. Action, Hypothalamic areas where orexigenic and anorexigenic messengers act. Regulation, Hypothalamic sites involved in regulation of synthesis, release, and action of orexigenic and anorexigenic signals.

 
A. Arcuate nucleus
The arcuate nucleus (ARC) is located at the base of the hypothalamus on either side of the third cerebroventricle (Fig. 1Go). It extends rostrocaudally from the posterior borders of the optic chiasm to the mamillary bodies. This nucleus has recently gained prominence among the sites associated with the hypothalamic integration of energy balance due to two reasons. It contains a high density of neurons that produce the orexigenic peptides, NPY (21, 41), the opioids, dynorphin, and the POMC-derived peptide, ß-endorphin [ß-END (42, 43, 44)], galanin [GAL (45, 46)], and the amino acids, {gamma}-aminobutyric acid [GABA (47, 48)] and glutamate (49, 50). Interestingly, {alpha}MSH, an anorexigenic peptide derived from the POMC precursor protein, is also coproduced with ß-END in the ARC (51, 52, 53). The agouti-related transcript (ART) that encodes the anorectic agouti-related protein (AgrP), the selective antagonist of MC3-R and MC4-R receptor subtypes (54, 55), is coexpressed with NPY mRNA in the ARC (39, 40). The terminal fields of these orexigenic and anorexigenic producing neurons in the ARC extend into various hypothalamic sites including the VMN, DMN, perifornical hypothalamus (PFH), PVN, and preoptic area (POA) where microinjection of these neurotransmitters/neuromodulators affects feeding behavior (see respective sections dealing with these signal molecules).

Morphological evidence has demonstrated that a subpopulation of neurons in the ARC coexpress NPY and GABA (36, 37, 38), and NPY-producing neurons are synaptically linked with ß-END- (34) and GAL-producing neurons in the ARC (33). Also, GAL-producing neurons establish morphological and functional links with ß-END-producing neurons within the ARC (35). Further, experimental evidence has demonstrated that an interconnected orexigenic network is likely to operate in the hypothalamic control of the daily patterning of food intake in which synergistic action of NPY and GABA (37, 38) and regulation of ß-END (34, 56) and GAL (35) release by NPY may play predominant roles. Recently, cocaine and amphetamine-regulated transcript (CART) mRNA has been localized in the ARC, and CART peptides appear to be physiologically relevant anorexigenic signals (57, 58, 59, 60, 61). In addition, due to the absence of blood brain barrier, the ARC is strategically positioned to be in direct communication with peripheral signals, such as the circulating adrenal and gonadal steroids as well as the large peptides, leptin and insulin, and the signals transported via the cerebrospinal fluid in the cerebroventricular system (14).

B. Ventromedial nucleus and lateral hypothalamus
Lesions in the ventromedial hypothalamus (VMH) that include the ventromedial nucleus (VMN) produce rapid hyperphagia and abnormal body weight gain that persist for a long time (24, 25, 26). The hypotheses that VMN is a "satiety center" exercising constant restraint on feeding and destruction of neuronal elements in the VMN permanently abolishes this restraint to allow sustained, unregulated phagia have been seriously contested over the years (13, 28, 29, 30). Evidence that parasagittal transections of fibers, between the LH, a hunger center, and the VMN extending up to the anterior hypothalamus, reproduced the VMH lesion-type hyperphagia and obesity challenged the role of the VMN in regulation of appetite (28, 30). However, in view of the emerging knowledge of the existence of an ARN extending over several hypothalamic sites, it is highly likely that the hyperphagia and obesity produced by VMH lesions and by transections of discrete pathways traversing the VMN are two distinct syndromes involving disruptions in disparate pathways (18, 19, 62). For example, microinjection of colchicine to disrupt neural signaling in the VMN diminished the availability and release of NPY in the ARC-PVN but augmented the galaninergic signaling in the hypothalamus (63, 64, 65, 66, 67). In addition, NPY gene expression was induced in hypothalamic sites that normally do not express NPY (68). Also, recent evidence clearly shows that VMH lesions disrupt signaling differently in each of the orexigenic pathways (see Section III.A.1 and Refs. 63, 64, 65, 66, 67, 68).

Although there is no evidence yet of production of either orexigenic or anorexigenic signals in the VMN, selective destruction of cell bodies in the VMN with ibotenic acid resulted in hyperphagia and increased rate of body weight gain (69) thus reemphasizing the role of neuronal elements in the VMN in regulation of appetite. On the other hand, microinjection into the VMN of NPY (13, 70), GAL (71, 72, 73), GABA, or GABA-agonist (74, 75, 76, 77) and ß-END (13, 74) stimulated feeding, while injections of leptin inhibited feeding (78, 79). Interestingly, microinjection of urocortin, a potent anorexigenic peptide, into the VMN also inhibited feeding (80). These findings raise the possibility of a receptive field in the VMN for several appetite-regulating signal molecules. The argument that spread of these signals through the extracellular fluid to receptive sites in the neighborhood most likely affected feeding is tenuous because receptors for each of these signals do exist in the VMN (see sections on each of these signals).

Further, the VMN is neurally linked with several hypothalamic sites implicated in the control of ingestive behavior. Although VMN efferents to the ARC have not been described, the VMN receives NPY (21), ß-END (43, 44), and CART (59, 61) containing projections from the ARC. VMN efferents to the DMN and parvocellular subdivision of the PVN (pPVN) have been traced, thereby supporting the possibility that disruption in signaling within the VMN may perturb flow of information to the DMN-PVN axis for the release of orexigenic signals, the consequence of which is unregulated feeding (81, 82, 83, 84).

LH is a contiguous band dorsal and lateral to VMH, extending rostrally from the mesencephalic tegmentum to the lateral preoptic area. In addition to sparsely distributed neuronal subpopulations, it is the site of passage for the medial forebrain bundle and other fibers connecting forebrain and midbrain structures with each other and several hypothalamic sites. Lesions in the LH produced temporary aphagia, adipsia, and loss in body weight (24, 25, 26, 27, 85). The severity of the LH syndrome and near normal recovery of food intake and body weight depended upon the location and the size of the lesion (85). These observations led Anand and Brobeck (24, 25) to conceptualize LH as a "feeding center" for elaboration of "urge to eat" normally restrained by signals from VMH. The fact that daily electrical stimulation of LH produced vigorous feeding leading to increase in body weight (86, 87) suggests the possibility that LH stimulation either restrained the action of an orexigenic signal(s) or activated those orexigenic pathways originating locally or in the vicinity of the LH. Indeed, recent evidence that melanin-concentrating hormone (MCH), orexins, and excitatory amino acid (EAA) glutamate are produced in the LH and microinjection of orexin and glutamate agonist stimulates feeding strengthen the view that LH is an integral part of the ARN in the forebrain (for details see Sections III.D, III.E, and III.F).

C. Dorsomedial nucleus
Electrolytic lesions in the dorsomedial nucleus (DMH) disrupt feeding to a far less extent than lesions in the VMH (81). As in other hypothalamic sites, microinjection of various orexigenic signals in the DMH elicited feeding (13, 70, 71, 72, 73, 74, 75, 76, 77). A crucial role of neurons in the DMN was indicated by the observation that inhibition of NPY-induced feeding by leptin enhanced neuronal c-FOS in the DMN, the protein product of the immediate early gene and a marker of neuronal activation (88, 89, 90). These findings were interpreted by us to infer that the site of NPY and leptin interaction may reside in the DMN and may represent a component of the circuitry involved in either attenuation or inhibition of feeding by leptin (89, 90).

DMN efferents have been traced to the VMN, and the two subdivisions of the PVN, the pPVN and magnocellular PVN (mPVN), with densest projections seen in the pPVN (81, 83, 84, 91, 92). Because there are prominent ARC efferents containing NPY to the DMN (93), it is highly likely that NPY released in this nucleus participates in stimulation of feeding. However, NPY levels in the DMN, unlike that seen in the PVN and ARC, were not elevated in response to fasting (94). Another relevant finding is that although only a few neurons in the DMN normally express NPY (95, 96), NPY gene expression was increased severalfold in the DMN in association with hyperphagia induced by disruption of neural signaling in the VMN (68) and in a genetic model of obesity and in response to lactation-induced hyperphagia (Refs. 97, 98 ; see Section III.A.4). It is not clear whether augmented NPY gene expression is associated with increased NPY availability in DMN efferents to the PVN and surrounding sites in these models of hyperphagia.

D. Paraventricular nucleus and perifornical hypothalamus
The PVN ranks second only to the VMH in terms of investigative interest in the hypothalamic control of appetite (Fig. 1Go). A dense cluster of heterogenous neurons fan out within a well defined boundary of the PVN on either side of the roof of the third cerebroventricle in the hypothalamus (91, 99). Several lines of evidence suggest that neuronal elements in the PVN participate in the control of ingestive behavior. Microinjection into the PVN of virtually all the known orexigenic signals, NPY, GAL, orexins, GABA, opioids, norepinephrine (NE), and epinephrine (E), stimulated feeding (13, 70, 71, 72, 73, 74, 75, 76, 77, 100), thereby implying the existence within the PVN and its vicinity of receptor sites for each of these signals. Additionally, microinjection of the anorexigenic neuropeptides such as CRH, produced locally (101, 102), and leptin (78), attenuated fasting-induced feeding. Furthermore, activation of the early gene marker, c-FOS, was augmented in neurons in the PVN in response to administration of orexigenic (103, 104) and anorexigenic signal molecules (88, 89, 90, 105). Another noteworthy observation is that among other hypothalamic sites tested, such as the VMN, the PVN is the only hypothalamic nucleus in which release of NPY, the most potent orexigenic signal, was augmented both in vivo and in vitro in response to fasting and before initiation of feeding (106, 107). These seminal findings are consistent with the idea that the PVN is one of the crucial sites for the release of orexigenic signals, and also possibly one of the sites of interaction of neurotransmitters/neuromodulators that inhibit feeding by diminishing NPY release.

Whereas evidence from various experimental paradigms reinforces the notion that the PVN is a crucial site for action of orexigenic signals, it is intriguing to find that destruction of the PVN also evoked hyperphagia and abnormal body weight gain accompanied by modifications in endocrine and autonomic profiles that were quite different from those produced by VMH lesions (108, 109, 110). It is possible that, under these conditions, neural elements surrounding the PVN compensate for the PVN loss. The PFH, the neural tissue surrounding the fornix and rostral to the LH, was reported to be relatively more effective in stimulation of feeding than the PVN after the microinjection of orexigenic signals such as NE and NPY (13, 111, 112), and orexin A microinjection into the PFH stimulated feeding (100). However, involvement of the PFH within the orexigenic circuitry is rather tenuous because of the strong possibility of diffusion of NPY and NE from the PFH to the PVN and lack of evidence correlating altered secretion patterns of orexigenic signals in the PFH in response to shifts in energy balance, as seen in the PVN (106, 107).

E. Suprachiasmatic nucleus and the timing device
The pattern of ingestive behavior is a highly regulated phenomenon in all living organisms. The drive to eat is evoked by appetite or the sensation of hunger which, in most vertebrates, is neurally based and entrained to activity-arousal mechanisms in the light-dark cycle. For example, rats consume between 85–90% of their total intake during the lights-off period. Ingestive behavior is initiated soon after onset of the dark phase, which apparently provides either the "trigger" or intensifies the drive for food (1). On the other hand, in nonhuman primates maintained in the laboratory and in humans, signaling of appetite is linked to socially acceptable or individually based requirements. Generally, neural, metabolic, and hormonal signals that influence ingestive behavior are stable during the period preceding and at the onset of the drive to eat (1, 113). The negative energy balance, which occurs during fasting, dieting, or undernourishment, intensifies the appetite to prevent underconsumption, but it is unlikely by itself to be the timely "trigger" for the drive to eat.

In addition, experimental findings invoke the timing mechanism in generation of the sensation of hunger. Discrete lesions in the suprachiasmatic nucleus (SCN), two small round nuclei resting dorsally on the optic chiasm on either side of the third cerebroventricle (Fig. 1Go), result in loss of regulated feeding (114, 115). Destruction of the VMH or disruption of neural signaling in the VMN similarly results in unregulated phagia (18, 19, 26). In humans, several instances of abnormal patterns of feeding characterized by a loss of regulated consumption of meals and appearance of the nighttime binge syndrome are associated with hyperphagia and obesity (2, 116).

In the rat, the photoperiodic-dependent circadian pattern of feeding is undoubtedly linked to information from the SCN to the hypothalamic ARN in two ways (Fig. 1Go). It is likely that initiation of nighttime feeding results from the increased release of orexigenic signals (see Section III.A). This may occur either directly in the orexigenic network or indirectly by restraining the influence of anorexigenic messengers on the release and action of orexigenic signals (see Section IV and Fig. 7Go). Conversely, reinstatement of the anorexigenic influence in a timely manner during the lights-on period may enforce cessation of feeding. The evidence that lesions in the SCN resulted in uninterrupted feeding is in agreement with the notion that the SCN exerts a restraining influence on ingestive behavior through the release of anorexigenic signals. A circadian pattern in hypothalamic gene expression of the orexigenic signals, NPY, GAL, and ß-END-generating precursor (POMC; Refs. 117, 118, 119, 120, 121, 122, 123), is also consistent with a regulated pattern of daily energy management. Interestingly, gene expression of these peptidergic signals peaked between 0700–1500 h during the lights-on period, followed by a trough before and during the dark phase when rats normally eat (Fig. 2Go and Refs. 120, 121, 123). Consequently, whereas this anticipatory augmentation in gene expression of orexigenic peptides may be needed for the supply of orexigenic peptides during the dark phase, undoubtedly feeding is regulated independently by circadian impulses that evoke the release of these peptides. Thus, we propose the existence of distinct neural mechanisms, one for synthesis and the other for release of orexigenic signals, emanating from the SCN and possibly other neural timing devices in the brain.



View larger version (44K):
[in this window]
[in a new window]
 
Figure 7. A schematic representation of the four elements of the ARN in the hypothalamus. The two basic elements of the circuitry are the orexigenic and anorexigenic signal-producing networks. Extensive evidence shows that various orexigenic producing neurons are morphologically and functionally linked to provide overlapping pathways to evoke feeding behavior. The neurosecretory function of these is regulated by information from the brain-timing mechanism, which in a timely fashion either directly, or indirectly through multisynaptic routes possibly traversing the VMH, curtail the restraint of anorexigenic signals. This results in the release of appetite-stimulating signals and feeding. Peripheral hormonal signals such as adipocyte leptin, adrenal glucocorticoids, and gonadal steroids also modulate the output of orexigenic and anorexigenic signals. Coordinated operation of these components of the appetite-regulatory circuitry maintains the daily pattern of food intake. A newly identified feature of this circuitry is the dynamics of receptor activation and sensitivity as directed by the interaction between peripheral hormones and the orexigenic and anorexigenic signals. An analysis of the neurochemical etiology of various experimental models of hyperphagia reveals that both overabundance and low abundance of orexigenic signals (e.g., NPY) at the receptor level cause unregulated eating behavior leading to hyperphagia. For details see text. (+), Stimulatory; (-), inhibitory; arrows depict the direction of information flow. [Adapted with permission from S. P. Kalra: Neuron 19:227–230, 1997 (1 ) © Cell Press].

 


View larger version (72K):
[in this window]
[in a new window]
 
Figure 2. Dynamic changes in hypothalamic NPY, POMC, and GAL mRNA (left panel) and serum leptin, leptin mRNA in adipocytes, and hypothalamic leptin receptor (R) mRNA levels (right panel) over a 24-h period. Shaded columns indicate the dark phase (1900–0500 h). Dissimilar superscripts are significantly different (P < 0.05).

 
It is evident that the SCN efferents shown to terminate in the VMN, DMN, subparaventricular zone, and LH constitute direct lines of communication to components of the orexigenic network for causation of circadian ingestive behavior. These efferent pathways may employ vasoactive intestinal peptide, vasopressin, and GABA as neurotransmitters. The neurons in the SCN and ARC are linked because SCN efferents innervate NPY, GAL, and POMC producing neurons in the ARC. However, it is unknown whether the daily pattern of gene expression of these peptides in the ARC is regulated directly or via projections to the DMN and/or the VMN (Fig. 2Go and Refs. 124, 125, 126, 127, 128, 129, 130, 131, 132, 133, 134, 135).

F. Overview of the neuroanatomical substrate associated with appetite control
Although it is now possible to identify the hypothalamic site or sites involved in regulation of appetite, the precise location of the receptive neural sites for each of the orexigenic and anorexigenic signals has not yet been ascertained (Fig. 1Go). The receptors for these signals are highly concentrated in the PVN, but they are by no means restricted to this site. The evidence that microinjection of orexigenic signals in a number of hypothalamic sites affects feeding argues strongly for a receptive field that extends beyond the PVN. The fact that lesions of the ARC, VMN, DMN, or PVN failed to subdue the drive to eat underscores a widespread receptor field associated with appetite regulation. This observation, coupled with the reports 1) that neural receptive sites for the anorexigenic signals, such as leptin, insulin, CRH, urocortin, and CART, overlap with sites containing receptors and terminal fields of orexigenic signals, including OR (see Sections IV and V) and 2) that orexigenic and anorexigenic neural systems are morphologically and functionally connected, strongly support our proposal of the existence of an interconnected ARN in the hypothalamus. We propose further that the ARN, spanning over several hypothalamic sites including the SCN, is composed of diverse signals (Ref. 1 ; Figs. 1Go and 7Go). Disruptions, either permanent as produced by lesions or neurotoxins in any part of the ARN, or temporary as caused by imbalances in hormonal and autonomic afferent signals, are likely to derange the tight control in the daily management of energy homeostasis leading to hyperphagia, abnormal body weight gain, and obesity.


    III. Orexigenic Signals
 Top
 Abstract
 I. Introduction
 II. Neuroanatomical Substrate...
 III. Orexigenic Signals
 IV. Anorexigenic Signals
 V. Leptin
 VI. Summary
 References
 
A. Neuropeptide Y
In the recent past, no other brain-signaling modality has drawn more attention than neuropeptide Y (NPY) in the control of appetite, body weight gain, and obesity. The potent appetite-stimulating effects of NPY and other members of the pancreatic polypeptide (PP) family were reported (136, 137) soon after Tatemoto (138) isolated and characterized the amino acid sequence of NPY, a member of the PP family. Subsequent research, spanning over a decade, has affirmed the hypothesis that NPY is a naturally occurring appetite transducer, and under normal conditions NPYergic transmissions represent an essential component of the common final pathway in the hypothalamic integration of energy homeostasis (1, 139). The wealth of information regarding NPYergic control of appetite is collated below.

1. Neuroanatomical pathways.
a. Source of NPY.
Although NPY-producing neurons are located in several sites in the brain (21, 139, 140), two subpopulations, one representing the extrahypothalamic cluster in the brainstem (BS) including the locus coeruleus, and the other located in the hypothalamus along the length of the ARC and in the DMN (1, 139), apparently participate in a disparate manner in the daily management of food intake (1, 139, 141).

NPY-producing neurons in the BS innervate various hypothalamic sites including the ARC, VMN, DMN, PVN, and surrounding regions (21, 140, 141, 142). A distinguishing feature of this population is that three other orexigenic signals, the catecholamines, NE and E, and the peptide GAL, are coproduced with NPY (21, 22). Consequently, when coreleased, these orexigenic signals are likely to interact at hypothalamic target sites (143). Experimental evidence suggests an interplay between NPY and the adrenergic system (141, 143). Transection of BS projections to the hypothalamus showed that whereas practically all the NE and E in the hypothalamus originated in the BS, only 40–50% of the NPY in various hypothalamic sites was derived from the BS (141, 142). Elimination of these BS inputs into the hypothalamus, either surgically (141, 142) or by injection of the neurotoxin, 6-hydroxydopamine (6-OHDA; Ref. 62), resulted in hyperphagia, a progressive increase in body weight and obesity. Hyperphagia in these rats was characteristically seen during the dark phase (Ref. 62 and our unpublished results). The cellular and molecular events causing increased phagia and body weight gain in these NPY- and catecholamine-deficient rats have been elucidated recently (Ref. 62 ; see Section III.A.4). The findings are consistent with the view that NPY of extrahypothalamic source along with the coexisting orexigenic signals is engaged in genesis and consolidation of neural stimuli that initiate and regulate nocturnal feeding.

Since the early demonstration that rapid time-dependent changes in NPY levels occurred in various hypothalamic sites in rats in response to fasting and refeeding (94), much attention has been directed toward defining the role of the population of NPY-producing neurons in the hypothalamic ARC. Studies involving lesions in the ARC and retrograde tracing studies demonstrated that approximately 15–20% of ARC NPY neurons innervate the PVN and DMN (93, 144). These quantitative data, together with the reports that the BS neurons contribute 40–50% of NPY in the PVN and other hypothalamic sites (141, 142), imply that the remainder of NPY in hypothalamic sites is produced either locally in the PVN or in neurons located in the vicinity. The DMN may normally be a small source of NPY in the PVN and elsewhere in the hypothalamus (93, 142), but it may contribute heavily in response to challenges that demand increased phagia (68, 97). Similarly, little NPY mRNA is detected in the PVN of normal rats but in response to disruption of neural signaling in the VMN, there is up-regulation of NPY gene expression (68). The possibility that the VMN is also a likely source of NPY for various hypothalamic sites because lesions in the VMN reduced NPY levels in the PVN and other selected hypothalamic sites (145, 146) remains to be investigated.

b. Where does NPY act to stimulate feeding?.
Since administration of NPY into the third cerebroventricle readily stimulated feeding in satiated rats, it was predicted that the site(s) of NPY action may reside in and around the paraventricular region of the hypothalamus (Fig. 1Go; Refs. 136, 137). Indeed, direct application of minute amounts of NPY into various hypothalamic and extrahypothalamic sites stimulated feeding (13, 70), and the PFH, the region lying caudal to the PVN, was reported to be more responsive to NPY than other sites in the neighborhood, including rostral structures such as the POA (70, 112). Also, demonstrations that injection of NPY into the fourth cerebroventricle stimulated feeding raised the possibility that the NPY-receptive field may extend outside the hypothalamus, possibly into the BS (104). Thus, as proposed in the preceding section, it is highly likely that the field of NPY action in stimulation of appetite is widespread in the rat brain (Fig. 1Go). This is supported by the localization of NPY Y1 (18, 19, 147, 148, 149, 150, 151, 152, 153, 154, 155) and Y5 (156, 157, 158, 159, 160, 161) receptor subtypes, putative receptors mediating stimulation of feeding by NPY, in sites corresponding to those where microinjection of NPY stimulated feeding (70).

Neuroactive substances, such as NPY, when microinjected into discrete neural sites or into the cerebroventricle, travel, through volume transmission modality, in the extracellular fluid to remote sites (162, 163). Our findings that NPY injection into either the third or the fourth ventricle stimulated c-FOS in similar forebrain areas (103, 104) argues for this volume transmission modality, i.e., movement of signals in interstitial fluid (163), to transport NPY to distant NPY target sites. Even the endogenously released NPY in the PVN can travel to surrounding hypothalamic sites that contain NPY receptors and are responsive to NPY microinjection. A comparison of the topography of c-FOS activation by NPY in the brain revealed the PVN and DMN as sites closely associated with NPY-induced feeding (103, 104). Indeed, pretreatment of rats with either the NPY Y1 receptor antagonist 1229U91 (164) or leptin (Fig. 3Go; Ref. 90) attenuated NPY-induced feeding and c-FOS activation in the mPVN, and leptin + NPY treatment enhanced c-FOS in the DMN. This line of investigation revealed that NPY targets engaged in stimulation of feeding reside within the DMN-mPVN axis (90, 164).



View larger version (134K):
[in this window]
[in a new window]
 
Figure 3. Photomicrograph showing c-FOS-like immunopositive cells in the parvicellular and magnocellular regions of the PVN (pPVN and mPVN, respectively) of an intraventricularly PBS-injected control (top left), NPY- injected rat (top right), leptin-injected rat (bottom left), and leptin + NPY-injected rat (bottom right). Whereas leptin on its own failed to activate c-FOS, it (LEP+PBS) significantly decreased NPY-induced c-FOS immunoreactivity selectively in the mPVN and food-intake (LEP+NPY). Pe, Periventricular area; V3, third ventricle. [Reproduced with permission from M. Yokosuka et al.: Physiol Behav64:331–338, 1998 (90 ) with permission from Elsevier Science].

 
An additional site of action is apparently the ARC itself, where NPY-producing perikarya abound and where NPY immunoreactive fibers from the BS terminate (21) and NPY collaterals are abundant (165). Intracerebroventricular injections of NPY into the third or fourth cerebroventricle increased the number of c-FOS-immunopositive neurons in the ARC (103, 104), possibly by activating Y1 and/or Y5 receptor subtypes located on peptidergic neurons (148, 149, 156, 157, 166, 167). Recently Rhim et al. (168) reported that NPY elicited a long-lasting synaptic inhibition in the ARC through inhibition of GABAergic and glutaminergic transmission. NPY-producing neurons also synapse in the ARC with neurons producing orexigenic signals such as POMC and GAL (33, 34), and NPY Y1 receptors are located in POMC- and NPY-producing cells (166). Seemingly, NPY released in the ARC not only autoregulates its own release but also the release of other orexigenic signals.

c. Is NPY a physiological appetite transducer?.
Among the known orexigenic signals, NPY is apparently the only messenger molecule that can be considered a physiological appetite transducer in the brain. Critical evidence to satisfy the basic criteria in support of this view is as follows: 1) Central administration of NPY stimulated feeding not only in satiated rats, but it also rapidly enhanced ongoing feeding (169). 2) As expected of a physiological signal, augmentation in food intake was dose dependent, and the response followed a bell-shaped curve; higher doses were less effective and produced a different pattern of feeding reflected in altered local eating rate and time spent eating (137). 3) Continuous central infusion of NPY reproduced the characteristic episodic nocturnal feeding pattern (170) and various components of feeding behavior, such as cumulative food intake, number of feeding episodes, mean episode length, total time spent eating, eating rate (g/min), and interepisode interval, were affected in a dose dependent manner. These similarities between the normally occurring nocturnal feeding behavior and that reproduced by NPY in satiated rats reinforced the view that intermittent nocturnal feeding in rats is chemically coded and presumably dependent upon increased NPY release and action (171). 4) This assumption was underscored by the finding that immunoneutralization of NPY during the dark phase of the light-dark cycle blocked nocturnal feeding (172). 5) Enhanced NPY release, selectively in the PVN, was temporally correlated with food consumption in rats maintained on a scheduled feeding regimen (106). The rate of NPY release in these rats rose just before onset of feeding and progressively decreased as the animal consumed food. Further, these rats continued to hypersecrete NPY if food was withheld, clearly pointing to the participation of NPY in sustaining appetite. 6) Fasting markedly augmented NPY release in the PVN both in vivo and in vitro (106, 107). 7) Fasting and food restriction augmented NPY Y1 receptor mRNA in the hypothalamus (Ref. 173 ; Fig. 4Go). 8) Hyperphagia and obesity in several genetic and experimental models were associated with modifications in NPYergic signaling. Genetically obese ob/ob mice, db/db mice, and fatty Zucker rats (fa/fa) exhibited increased prepro-NPY mRNA in the ARC (174, 175, 176) and NPY levels in various hypothalamic sites and release in the PVN. 9) Augmented NPYergic signaling in the hypothalamus, as evidenced by increases in ARC NPY gene expression and PVN NPY levels (177, 178, 179) and release (179, 180, 181), preceded the onset of hyperphagia induced experimentally by streptozotocin (STZ) treatment and upon extension of the duration of diabetes in these rats; the augmented NPY response spread from the PVN to practically all hypothalamic sites (177, 178, 179, 180, 181). 10) Increased energy demands during lactation, apparently met by hyperphagia, were accompanied by increased NPYergic signaling in the ARC-PVN axis (15, 182). 11) Selective Y1 and Y5 receptor antagonists attenuated both nocturnal feeding and that induced by fasting (18, 19, 150, 151, 152, 155, 158, 159). 12) Studies conducted in mice lacking either Y1 or Y5 receptors also affirm the involvement of NPY in food intake and obesity. In mice deficient in expression of the Y1 receptor subtype, daily food intake and intake in response to fasting was significantly reduced (183). Interestingly, these mice displayed a late onset of increase in body weight. Our observations that fasting normally increases the availability of Y1R (173) and selective Y1 antagonists inhibit feeding (18, 19), taken together with the results from Y1 receptor knockout mice, clearly underscore a role of NPY in control of food intake and obesity, a conclusion in line with earlier results (184). With respect to involvement of other NPY receptor subtypes, it has been shown that Y5R-null mice feed normally early on, display normal feeding response to exogenous NPY, but show late-onset obesity due to increased body weight (185). These observations are consistent with NPY’s participation in stimulation of feeding and reveal that Y1 and Y5R are involved but, as documented earlier (18, 19, 184), signaling via Y1R may be a prominent player in mediating feeding induced by NPY. In view of the several recent reports (186, 187, 188), it seems that either both or another subtype closely related to Y1 and Y5 subtypes may mediate the appetite-stimulating effects of NPY. An exciting new insight from these studies is that not only a partial deficiency in NPY availability induced experimentally in rats (Refs. 18, 19, 63 ; see Section III.A.4b) but now a complete deficiency of either Y1 or Y5 receptor subtypes results in hyperphagia and obesity. Since in NPY-deficient rats, contribution of increased responsiveness to NPY due to up-regulation of Y1R (18, 19, 63) and up-regulation of GAL and GAL responsiveness have been demonstrated (64, 66, 67), a similar pattern of shifts in signal transduction and orexigenic signals responsible for late-onset obesity is highly likely in Y1 and Y5 receptor mutant mice.



View larger version (51K):
[in this window]
[in a new window]
 
Figure 4. Up-regulation of NPY Y1 receptor mRNA in the hypothalamus of food-deprived (Fast) rats and in rats maintained on food restriction (F.rest) as compared with free-fed (FF) rats. Dissimilar superscripts are significantly different (P < 0.05). [Adapted with permission from B. Xu et al.: Regul Pept 75–76:391–395, 1998 (173 )].

 
2. Regulation of hypothalamic NPY secretion. As would be expected of a naturally occurring appetite transducer, multiple neural and hormonal signals affect NPYergic signaling in the hypothalamus. In the majority of studies, NPYergic signaling was evaluated by analysis of either steady-state NPY levels in discrete hypothalamic sites or prepro-NPY mRNA levels in the ARC during the lights-on period when rats do not normally eat. NPY release in discrete hypothalamic sites has been examined only occasionally in correlation with feeding behavior (106, 107, 174). Despite these shortcomings, the accumulated information, as summarized below, indicates that regulation of NPY secretion in the daily management of ingestive behavior by the hypothalamus is complex and multifactorial.

a. Neural control of NPY secretion:
i) Circadian clock. Since eating in rats occurs primarily during the dark phase, a temporally correlated pattern of changes in hypothalamic NPY levels, release, and synthesis was anticipated. Akabayashi et al. (131) reported increased prepro-NPY mRNA in the basal hypothalamus immediately preceding onset of the dark phase, followed by a precipitous drop during the nocturnal period of feeding. However, NPY levels remained elevated in the large dorsomedial hypothalamic area, a part of neural tissue containing PVN and several other adjoining nuclei, during the dark phase. A detailed reexamination by Xu et al. (120, 121, 123) revealed that prepro-NPY mRNA levels increased during the lights-on phase between 0700–1100 h, and then decreased and remained in the basal range throughout the entire dark phase (Fig. 2Go). Evidently, a transient elevation in NPY gene expression occurred several hours before the onset of nocturnal feeding (Fig. 2Go). These findings implied that an independent neural component, driven by the circadian clock, regulated this daily rhythm in gene expression (120, 121, 123). Indeed, evidence showed that up-regulation of ARC NPY gene expression, brought about by food restriction to 4 h during the lights-on period, was sustained throughout the 24-h period despite daily changes in the light-dark cycle (121, 123, 189). This abolition of the daily rhythm in the ARC NPY gene expression suggests uncoupling from the photoperiodic signals emanating from the SCN. The neural pathways that transmit information generated by changes in the pattern of food availability are currently unknown.

ii) Serotonin (5-HT). Guy et al. (190) described nonsynaptic appositions between 5-HT nerve terminals and immunoreactive NPY perikarya and dendrites in the ARC. Subsequently, Dube et al. (191) showed that fenfluramine, which increases serotoninergic signaling and reduces food intake, decreased NPY levels without acutely changing release in the PVN of food-deprived rats. A possible inhibitory influence of 5-HT on NPYergic signaling was later affirmed by a series of studies by Williams and co-workers (192, 193). Pharmacological agents that activated 5-HT output inhibited NPY levels and release in the PVN concurrently with anorexia. Conversely, 5-HT antagonists stimulated feeding in conjunction with increased NPY levels in the ARC and PVN. Although these morphological and pharmacological reports collectively imply an inhibitory role of 5-HT on NPY release, the physiological relevance of the communication between serotoninergic and NPYergic signalings in the daily patterning of food intake warrants further investigation.

3. Hormonal regulation of NPY secretion. Because of its strategic distribution, circulating signals are in direct communication with the NPY system in the hypothalamus. Steroids originating from the gonads and adrenal cortex exert a modulatory influence on NPY synthesis and release (194, 195, 196, 197, 198, 199). In addition, metabolic signals relayed by the circulating peptide hormones, insulin and leptin, and cytokines can affect the NPY system directly and also are suspected to cross the blood-brain barrier and gain access to the cerebrospinal fluid through the circumventricular organs (Refs. 16, 17, 200, 201 ; see Sections III.A and V).

a. Gonadal steroids.
It has been known for a long time that gonadal steroids modulate food intake and body weight gain in rodents and other mammals (15). In the female, ovariectomy generally increased food consumption and body weight gain, and estrogen replacement reinstated the normal pattern of daily intake (15). Gonadal steroids promoted NPY neurosecretion in the hypothalamus (194, 195, 196, 202) as evidenced by the findings that gonadectomy decreased prepro-NPY mRNA levels in the ARC, and testosterone replacement reinstated gene expression and NPY levels in selected hypothalamic sites in male rats. In ovariectomized female rats, estrogen treatment exerted a bimodal effect. Short-term treatment induced daily changes in NPY gene expression in the ARC with a rise in the afternoon sustained for a few hours, a response reminiscent of the pattern of gene expression seen on proestrus in cycling rats (195, 203). In contrast, a prolonged uninterrupted increase in estradiol in physiological levels decreased NPY levels selectively in the PVN and PFH and decreased NPY release from the PVN (204). These results were interpreted to imply that the anorectic effects of estrogen were mediated by decreased NPY release from the PVN (204). Since NPY-producing neurons possess estrogen receptors (194), it is highly likely that chronic estrogen action suppressed both NPY synthesis in the ARC and release in the PVN. Baskin et al. (205) also reported recently that estrogen inhibited the fasting-induced increase in hypothalamic NPY gene expression. Evidently, an underlying role of estrogen is to inhibit food intake via direct action on the NPY pathway in the ARC-PVN of female rats.

b. Adrenal glucocorticoids.
Unlike the hypothalamo-pituitary-gonadal axis, the interaction of the hypothalamic-pituitary-adrenal axis with the NPY system, particularly in relation to nocturnal feeding in the rat, has been unclear despite the concerted efforts of a number of investigators. Experimental evidence invoking the existence of a feedback relationship between the adrenal cortex and hypothalamic NPY system are the following: NPY-producing neurons are synaptically linked with CRH neurons in the PVN (206). Microinjection of NPY into the PVN rapidly activated pituitary ACTH and adrenal corticosterone secretion, suggesting that this communication is excitatory in nature (207). NPY neurons in the ARC are rich in glucocorticoid receptors (197, 198), and treatment of adrenalectomized rats with glucocorticoids increased NPY gene expression in the ARC in association with increased food intake and body weight (197, 199, 208). On the other hand, unlike the effects of gonadectomy, it is intriguing that adrenalectomy either decreased or elicited no effect on hypothalamic NPY levels and NPY gene expression in the ARC (209, 210, 211). Adrenalectomy failed to affect the time of onset and pattern of nocturnal feeding, leading one to infer that NPY-induced nocturnal feeding is not tightly coupled to the daily rhythm in the hypothalamo-pituitary-adrenal axis. Further, in the absence of circulating corticosterone, rats showed only a small decrease in cumulative food intake, and NPY-induced feeding under these conditions was minimally attenuated (211). These unremarkable effects on food intake were attributable to the generally depressed metabolic responses of adrenalectomized animals. Also, contrary to previous reports (209, 210, 211, 212), recent studies showed that the fasting-induced up-regulation of hypothalamic NPY gene expression was not dependent upon adrenal glucocorticoids (213). Evidently, NPY can stimulate the hypothalamo-pituitary-adrenal axis (207), but whether glucocorticoids play a role in governing NPY output for nocturnal feeding is uncertain.

c. Insulin.
Although it has been suggested, but not completely proven, for a long time that pancreatic insulin may act centrally to regulate body weight by restraining food intake (200, 201), the hypothalamic signaling pathway underlying the anorexigenic effects of insulin was unknown. The reports that insulinopenia produced by STZ increased NPY levels in various hypothalamic sites in association with hyperphagia, and insulin replacement normalized NPY levels and blocked hyperphagia (176, 177, 178, 179, 214, 215), provided the breakthrough in firming the concept that peripheral signals exert an inhibitory influence on hypothalamic orexigenic pathways. NPY gene expression and NPY levels in the PVN increased before the onset of hyperphagia in STZ-treated rats (181). Concurrently, NPY release increased selectively in the PVN (180). Importantly, insulin replacement, in doses that did not affect hyperglycemia in STZ-treated rats, markedly decreased NPY release in the PVN (180). Thus, it is highly likely that insulin per se exerted a regulatory influence on NPY release in the PVN. The site of insulin action is apparently the PVN because insulin inhibited NPY release from the PVN in vitro (216). Whether insulin inhibits NPY release directly from NPY nerve terminals or through interneurons has not been delineated. Further, central infusion of insulin attenuated the fasting-induced increase in NPY gene expression (201). Since the effects of insulin on PVN NPY release were rapid (216), it is highly likely that attenuation of the ARC NPY gene expression by central infusion of insulin may be secondary to inhibition of NPY release from axon terminals in the PVN (216). A direct inhibitory action of insulin at the NPY perikaryal level is also doubtful because central infusion of insulin failed to activate c-FOS in the ARC (217), and insulin receptor immunoreactivity (201) has not yet been visualized on NPY neurons in the ARC. These findings are noteworthy when compared with the response elicited by leptin, which also inhibited NPY synthesis (see Section V), but readily stimulated c-FOS expression in several hypothalamic sites (89, 90). Nevertheless, there is a general consensus that insulin deficiency results in NPY hypersecretion which, in turn, induces hyperphagia. The physiological relevance of this postulated restraint of NPY release by insulin in nocturnal feeding is largely unclear. Since insulin rapidly inhibited NPY release from the PVN nerve terminals, it is possible that the postprandial insulin rise (216) exerts a restraint, along with leptin (Fig. 2Go; Refs. 121, 123), on NPY release in the PVN to prevent hyperphagia. This nocturnal restraint on ongoing feeding, we believe, constitutes one of the mechanisms enforcing set point in body weight.

d. Cytokines.
During infection, CNS injury, or other pathological conditions, anorexia leading to a loss in body weight is manifested (9, 10, 11). Recent evidence points to a possible relationship between increased circulating levels of cytokines and hypothalamic NPYergic signaling in induction of anorexia during the pathological state. Interleukin-1 was shown to decrease food intake probably by interfering with NPY’s ability to stimulate feeding (10). Ciliary neurotrophic factor (CNTF), another cytokine, is a neurotrophic factor, but systemic and central injections of CNTF produced severe anorexia (218, 219). Xu et al. (219) reported that CNTF decreased the availability of NPY for release by suppressing NPY gene expression in the ARC. In addition, CNTF markedly inhibited NPY-induced feeding by decreasing Y1 abundance (173). Seemingly, CNTF, and possibly other anorectic cytokines, acts within the hypothalamus through their respective receptor systems to decrease NPY availability and also to suppress the postsynaptic signal transduction processor. A consequence of the interaction of cytokines with the NPY system, thus, is anorexia and loss in body weight during pathological conditions.

4. Role of NPY in the etiology of hyperphagia and obesity in rodents. There are several genetic and experimentally induced models of obesity in rodents (6, 220, 221, 222). Although a hallmark of obesity is an abnormal increase in body weight attributable to multiple metabolic disturbances, partly dependent upon an imbalance in the endocrine and autonomic systems, a primary causal factor is the unrelenting drive to eat. The unregulated feeding pattern manifests either throughout the day or selectively during the dark phase. Based on the new insight that NPY is a primary appetite transducer within the interconnected orexigenic network in the hypothalamus, the possibility that modifications in NPYergic signaling may underlie hyperphagia in various models of obesity has been explored extensively.

a. Genetic models of obesity.
Evidence accumulated to date clearly showed that in genetically obese ob/ob and db/db mice, hyperphagia may be attributed to unabated NPY output in the hypothalamus (174, 175, 176, 223, 224, 225). Whereas the increase in hypothalamic NPYergic signaling in ob/ob mice is due to leptin deficiency (16, 17, 226, 227, 228, 229, 230), in db/db mice a defect in leptin signal transduction due to a point mutation in the leptin receptor is the causal factor (231, 232, 233, 234). Similarly, in fatty Zucker rats (fa/fa), hypothalamic NPYergic signaling is augmented as evidenced by increased NPY levels in various hypothalamic sites, NPY gene expression in the ARC, and NPY release in the PVN (174, 176, 177, 178). The fact that immunoneutralization of NPY in Zucker rats suppressed feeding affirmed an important role of NPY in this model of hyperphagia (235). Leptin resistance is believed to underlie NPY hyperactivity in Zucker rats (236). Also, in corpulent (cp/cp) JCR-LA obese rats, there is evidence of overactive NPYergic signaling in the hypothalamus (237).

b. Experimental models of obesity.
In rats, hyperphagia and abnormal body weight gain leading to obesity can be experimentally induced by disruption of neural pathways either by localized lesions in the hypothalamus, transection of neural pathways in the basal hypothalamus, or by injection of neurotoxins in hypothalamic and extrahypothalamic sites (24, 25, 26, 27, 28, 29, 30, 238). A concerted effort in our laboratory, as summarized below, aimed at elucidating the neurochemical basis of experimentally induced hyperphagia and obesity, revealed disturbances in leptin-NPY signaling (see also Section V).

i) Hyperphagia and obesity due to disruption in VMH signaling. It has been known for more than 50 yr that lesions in the VMH are associated with hyperphagia and abnormal body weight gain in rodents and humans (24, 25, 26, 29). In rats, VMH lesions produced rapid and longlasting hyperphagia resulting in overt obesity (24, 25, 26). Contrary to expectations, prepro-NPY mRNA levels in the ARC and NPY levels and release from the PVN were markedly suppressed in these rats (1, 12, 145, 146). This diminution in hypothalamic NPYergic signaling in hyperphagic VMH-lesioned rats contrasts with the hyperactive NPY system prevailing in genetic models of obesity in rodents (174, 175, 176, 214) and in diabetic rats (177, 178, 179, 180, 181). Interestingly, despite the low levels of NPY secretion, hyperphagia was dependent upon NPY because immunoneutralization of NPY completely suppressed feeding in these rats (239).

In another model in which neural signaling within the VMN was disrupted by microinjection of the neurotoxin colchicine, transient hyperphagia, lasting only for 4–5 days developed rapidly with an attendant increase in body weight (18, 19, 63, 240). As in the VMH-lesioned rats, NPYergic signaling, as demonstrated by diminution in NPY mRNA in the ARC (18, 19, 63) and NPY release in the PVN (65), diminished immediately after colchicine injection. In these rats, hyperphagia was found to be, in part, due to increased sensitivity to NPY (Fig. 5Go) and abundance of Y1 receptors in the hypothalamus, blockade of which by Y1 receptor antagonist inhibited feeding (18, 19, 63). The role of NPY Y5 receptors in this paradigm is not clear. In addition, even this temporary disruption in VMN function rendered rats resistant to the action of leptin (1, 18, 19, 63). Seemingly, either one of the targets of leptin feedback resides in the VMN or alternatively, an intact VMN is required for relay of leptin feedback information to sites outside the VMN. In the other models of obesity, such as that induced by gold thioglucose in mice or by postnatal treatment of rats with monosodium glutamate, hypothalamic NPY gene expression and NPY levels in the PVN were similarly suppressed, and feeding in response to NPY was increased (241, 242).



View larger version (37K):
[in this window]
[in a new window]
 
Figure 5. Increased sensitivity to intracerebroventricular administration of NPY in stimulation of food intake in hyperphagic rats that received colchicine (COL) in the VMN. Each panel compares data from cholchicine- and saline (SAL)-injected rats in response to vehicle (PBS) or NPY (three doses) at various intervals post colchicine. Food intake in response to NPY was consistently greater in colchicine-treated rats vs. SAL-treated rats on the same day. Note that the increased sensitivity to NPY occurred concomitantly with reduced hypothalamic NPY synthesis and release in the PVN (see text). [Reproduced with permission from P.S. Kalra et al.: Regul Pept 72:121–130, 1997 (63 )].

 
A further evaluation recently showed that disruption of signaling by colchicine (COL) in the VMN augmented NPY gene expression in PVN and DMN, the nuclei that normally either do not or only minimally express NPY mRNA. Interestingly, these rats also start to produce excessive GAL and display increased responsiveness to GAL (63, 64, 65, 66, 67, 68).

Taken together, evidence from genetic and experimental models of rodent obesity showed that both over- and underexpression of NPY resulted in unregulated phagia and body weight gain (1). In the case of low NPY abundance, hyperphagia is seemingly due to multiple factors brought about by a shift in the normal pattern of NPYergic signaling in the ARN. These include 1) a rapid increase in NPY Y1 receptors in the hypothalamus leading to increased sensitivity to NPY (18, 19, 63); 2) concurrent development of leptin resistance in these rats, a hypothesis in line with the hypothesis that a normally functioning VMH is necessary for the central inhibitory effect of leptin on food intake (1, 18, 19); 3) up-regulation of NPY in novel hypothalamic sites; and 4) increase in galaninergic signaling.

ii) Hyperphagia and obesity due to interruption of NE input from the BS. Microinjection of the neurotoxin 6-OHDA into the ventral NE bundle at the level of trochlear or oculomotor nuclei induced hyperphagia and obesity (238, 243). Overt hyperphagia manifesting several days after 6-OHDA injection was long lasting and accompanied by increased rate of body weight gain (238). Interestingly, the disruption in NPY-leptin signaling in these rats was markedly different from that seen in the hyperphagia and obesity syndrome produced by either electrolytic lesions in the VMH (1, 12, 145, 146) or colchicine injection (18, 19, 63) in the VMN. The dark-phase hyperphagia in 6-OHDA-treated rats was accompanied by two modifications in NPY-leptin secretion. The normal nocturnal increase in leptin secretion after eating (Fig. 2Go) was abolished by interruption of catecholaminergic afferents to the hypothalamus (62). Further, it was shown that diminution in leptin feedback was responsible for the increase in nocturnal NPY secretion, Y5 receptor mRNA, and hyperphagia (62).

Thus, analysis of the etiology of hyperphagia and obesity produced experimentally revealed a number of loci in NPYergic signaling within ARN that are vulnerable to disruptive insults. A provocative new concept emerging from these efforts is that both low and high abundance of NPY result in hyperphagia and obesity (1, 18, 19, 63). The evidence clearly showed that increased receptor abundance, NPY receptor sensitivity, and up-regulation of NPY in novel hypothalamic sites, together with increased availability of other orexigenic signals, such as GAL, underlie hyperphagia. Shifts in the leptin secretion pattern and development of leptin resistance also contribute toward this altered hyperphagia, producing peptidergic signaling.

B. Galanin
Galanin (GAL), a 29-amino acid peptide, has been studied extensively for its appetite-stimulating action in the brain. Intraventricular or intrahypothalamic injection of GAL stimulated feeding in satiated rats (71, 72, 73, 244, 245, 246, 247, 248), and it appears that receptive sites of GAL may be widely distributed in the rat brain. Microinjection of GAL in the PVN, LH, VMN, and central nucleus of the amygdala stimulated feeding in rats. As compared with NPY, GAL-induced feeding is less remarkable, lasting for only about 30 min with augmented response apparent at the onset of the dark phase (71, 72, 73, 246, 248).

Unlike the discrete distribution of NPY-producing cells (21, 140), several subpopulations of GAL-producing neurons exist in the hypothalamus that innervate most of the hypothalamus (45). A close anatomical and functional relationship exists between GAL and other orexigenic signals producing neurons in the rat brain. NPY-producing neurons were shown to be in direct communication with GAL-producing neurons, especially in the ARC and PVN (33). Thus, GAL may, in part, mediate NPY-induced feeding (1). Further, GAL-immunopositive nerve terminals establish synaptic links with POMC containing dendrites and soma in the ARC (35), raising the possibility that GAL may stimulate the release of ß-END, which also stimulates ingestive behavior (see Section III.C). Pretreatment of rats with naloxone (NAL), an opiate receptor antagonist, markedly attenuated GAL-induced feeding (246). However, because NAL pretreatment failed to completely block the feeding response, it is quite likely that GAL directly stimulates feeding by activating GAL receptors in hypothalamic sites (248, 249, 250, 251). Involvement of NE in GAL-induced feeding is suggested by the reports that GAL stimulated NE release in the PVN, and blockade of {alpha}2-adrenergic receptors with rauwolscine attenuated the feeding stimulated by GAL (244). Furthermore, the two orexigenic signals are coproduced in a subpopulation of neurons in the BS that project into various hypothalamic sites, including the PVN (22), and stimulation of modest feeding under the influence of GAL resembles that induced by NE (72, 73). Thus, release of NE in the PVN and neighboring sites by GAL from nerve terminals emanating from perikarya in the BS may mediate GAL-induced feeding (22). Undoubtedly, feeding elicited by GAL involves an interplay among various orexigenic signals, as documented by the existence of communication lines consisting of NPY->GAL->ß-END and GAL->NE.

Whether GAL constitutes an important orexigenic signal in the daily pattern of feeding has not been clearly established. GAL-producing cells are localized in discrete subpopulations in the ARC, DMN, and PVN, but their participation in normal feeding behavior in the rat is not known. Similarly, high-affinity GAL receptors are distributed in multiple sites including those where GAL microinjection stimulated feeding (252, 253). However, GAL receptor antagonists, including M-40 that readily inhibited GAL-induced feeding, failed to suppress normal feeding in several behavioral paradigms (46, 248, 249). Infusion of GAL antisense oligonucleotides in the PVN inhibited feeding (254) but, unlike NPY, continuous GAL infusion was ineffective in increasing food intake and body weight gain (255).

Attempts to establish a correlation between the daily feeding pattern and changes in GAL synthesis and release in the hypothalamus have not been instructive. Increased appetite, evoked by either restricted feeding or fasting, decreased GAL gene expression in the ARC (256, 257, 258, 259). A small increase in GAL levels in the lateral portions of the pPVN was observed 3 h into the dark phase (118); the physiological significance of this in normal nocturnal feeding is unknown. Interestingly, the daily increases in GAL gene expression in the hypothalamus temporally coincide with those in hypothalamic POMC mRNA but lag slightly behind those in NPY (Fig. 2Go; Refs. 120, 121, 123). Also, unlike NPY (174, 175, 176), GAL levels in the PVN and other hypothalamic sites do not correlate with hyperphagia in Zucker fatty rats (257). Consequently, additional studies are warranted to define the precise physiological role of GAL in the daily pattern of feeding behavior.

C. Endogenous opioid peptides
Although it has been known for a long time that opiates promote phagia (260, 261), interest in the roles of endogenous opioid peptides (EOP) gained momentum after the isolation and characterization of the three biologically active opioids—ß-END, dynorphin A (DYN), and enkephalins—in the hypothalamus (51, 52, 53). The single-copy POMC gene encodes the POMC precursor PP, which yields the opioid, ß-END, and the nonopioid peptides, ACTH and {alpha}MSH (51, 52, 53). Within the hypothalamus, POMC neurons, localized exclusively in the ARC, innervate the VMN, PVN, DMN, and other areas of the hypothalamus (43, 44), where microinjection of ß-END and opiate agonists that bind to the µ-receptor subtype stimulate feeding (13, 23, 74). Neurons producing the two pentapeptides, methionine-enkephalin (met-ENK) and leucine-enkephalin (leu-ENK), are more widely distributed in the hypothalamus, and discrete populations of leu-ENK- or met-ENK-immunopositive perikarya have been visualized in feeding-relevant sites, such as the ARC, VMN, DMN, and PVN, sites that are also richly innervated by ENK-immunopositive fibers (13, 42, 44). Although ENK were ineffective on their own, the long-acting analogs stimulated feeding, possibly involving the {delta}-receptor subtype (13, 261). Central injection of the third hypothalamic opioid, dynorphin A 1–17, derived from the precursor prodynorphin, stimulated feeding by activating the {kappa}-opiate receptor subtype. DYN-producing cells are also found in various regions of the hypothalamus, including the ARC and PVN (13, 44, 261, 262).

A critical evaluation of food intake induced by various EOP showed that opioid-evoked feeding, unlike that evoked by NPY, is generally short lived and relatively modest (261). Extensive evidence from clinical and animal experiments suggested that opiate receptor antagonists, especially the µ- and {kappa}-antagonists, decreased food intake (13, 262, 263, 264, 265, 266). However, whereas fasting and diabetes increased NPY mRNA, POMC mRNA was decreased, but DYN mRNA was augmented in the PVN (258, 267). The daily rhythm in POMC gene expression was similar to that of GAL but lagged behind the rhythm in NPY gene expression by a few hours (Fig. 2Go; Refs. 120, 121, 123).

Although a large body of evidence concurs with the orexigenic nature of EOP, especially ß-END, whether these peptides subserve a relevant physiological role or represent redundant pathways in the daily management of ingestive behavior remains to be experimentally documented. However, the recent anatomical studies establishing morphological links among ß-END, GAL, and NPY and between GABA and ß-END immunoreactive pathways in the hypothalamus are highly suggestive of a role of ß-END within the orexigenic network (1, 33, 34, 35, 36, 37). Horvath et al. (34) showed that NPY immunopositive terminals established synaptic contacts with ß-END-containing dendrites and soma, suggesting that NPY may induce feeding directly on its own and also by stimulating ß-END in relevant sites such as the PVN. Indeed, NPY was shown to stimulate ß-END release in the hypothalamus (56), and NAL pretreatment attenuated feeding in response to NPY (13, 261). On the other hand, NPY decreased POMC mRNA levels in the ARC (268) and the opi