Evidence that the nucleus accumbens shell, ventral pallidum, and lateral hypothalamus are components of a lateralized feeding circuit

Abstract

Pronounced feeding can be elicited by injections of the GABA_A agonist muscimol into the medial shell region of the nucleus accumbens (AcbSh). This region of AcbSh has been shown to project to both the lateral hypothalamus (LH) and the medial ventral pallidum (VPm). The current study examined the effects of unilateral LH or VPm lesions on the ingestive responses induced by injections of muscimol into the AcbSh on either the same or the opposite side of the brain. We found that lesions of either of these structures drastically attenuated feeding induced from the ipsilateral, as compared to the contralateral, AcbSh. The “ipsilateral/contralateral disruption design” employed here virtually rules out the possibility that the suppressive effects of the lesions were nonspecific and suggests that the VPm and LH play essential roles in mediating the ingestive effects of inactivation of the AcbSh.

Introduction

Although the nucleus accumbens is often assumed to exert a generalized influence on motivational or reward mechanisms, studies have demonstrated that pharmacological inactivation of the shell subregion of the accumbens (AcbSh) with GABA agonists or glutamate receptor antagonists induces a large, and remarkably specific, increase in feeding behavior in satiated rats [1], [2], [3], [4]. The active site in the ventral striatum at which these drugs elicit feeding has been localized to the rostral medial AcbSh [1], [5], [6], [7] and the fact that large, behaviorally specific, increases in food intake can be elicited by blocking the actions of endogenous glutamate, or by increasing levels of endogenous GABA [1], indicates that the AcbSh is likely to play a role in the normal physiological control of feeding behavior. The importance of the AcbSh in the physiological control of feeding behavior is underscored by the fact that changes in food intake have also been observed after local injections of a variety of feeding related compounds including nociceptin, cocaine- and amphetamine-related transcript protein, melanin-concentrating hormone, orexin, opioids, amylin, and endocannabinoids (see [8] for review).

An obviously important line of investigation involves the determination of the functional circuitry through which the AcbSh mediates food intake. The AcbSh sends substantial projections to at least two brain regions known to play a role in the control of feeding; the lateral hypothalamus (LH) and medial ventral pallidum (VPm) [9], [10], [11], [12], [13]. While the role played by the VP in the expression of feeding induced by manipulations of amino acid-coded circuits in the AcbSh has never been investigated, the available data suggest the occurrence of a functional relationship between the AcbSh and the LH. Thus, feeding induced by inactivation of the AcbSh can be significantly attenuated by acute suppression of LH activity through local drug injections [2], [14]. While these findings suggest that the feeding behavior elicited by inhibition of neurons in the AcbSh is dependent on activation of LH neurons, they are, unfortunately, open to alternative explanations. It is quite possible, for example, that the LH does not play a direct role in mediating effects obtained from the AcbSh, but that the LH manipulations performed in these studies may have suppressed feeding as a result of some “nonspecific” effect, such as sedation, malaise, alterations in hormonal or autonomic reactivity, or changes in gustatory or olfactory processing.

One way to circumvent these interpretive difficulties is through the use of what could be called an “ipsilateral/contralateral disruption design” (ICD) which allows for an evaluation of whether lesion effects are due to a specific disruption of lateralized circuitry activated by the AcbSh injections. This approach requires that the AcbSh on one side of the brain exerts a larger influence on the ipsilateral LH than on the contralateral LH. The available anatomical data supports this possibility with respect to direct projections, since very few fibers originating in the AcbSh appear to cross the midline to terminate in the contralateral LH [13], [15], [16]. The uncrossed functional nature of these pathways is further supported by the observation that unilateral injections of muscimol into the AcbSh produce a much larger increase in the number of cells expressing the nuclear protein Fos, a marker of neural activity [17], [18], in the ipsilateral LH than the contralateral LH [19]. If muscimol-induced feeding is mediated through a predominantly unilateral activation of the LH, one would expect that lesions of this structure on the same side as the muscimol injection would produce a much greater disruption of the feeding response than would lesions on the contralateral side. Conversely, if AcbSh-mediated food intake were suppressed because a particular lesion elicited a “nonspecific” response, such as sedation or malaise, one would expect that the “nonspecific” response would be elicited in a similar manner by lesions on either side of the brain and therefore, that equivalent reductions in food intake would be observed after either ipsilateral or contralateral lesions.

The AcbSh also projects heavily to the VPm and both lesioning and drug microinjection studies strongly implicate this region in the control of feeding [12], [20], [21]. Considerably fewer Fos-immunoreactive cells are seen in the VPm than the LH after intra-AcbSh muscimol injections, but those cells that are present are located primarily ipsilateral to the injection site [19]. Anatomical data also suggest that the direct projection from the AcbSh to the VPm is strongly unilateral [16]. These findings suggest that the ICD experimental design may again be useful for evaluating the role of the VPm in mediating the feeding induced by GABAergic inactivation of the AcbSh.

In light of these considerations, we conducted three experiments to evaluate the connections through which the AcbSh is able to influence food intake. First, we examined feeding following unilateral injections of muscimol into the AcbSh, to verify that unilateral inactivation of the AcbSh is able to induce feeding in our current testing situation. We then examined the effects of unilateral lesions of the LH and of the VPm using the ICD approach outlined above.