A simple perusal of the wonderful illustrations in the Netter Atlases showing the effects of various hypothalamic lesions on food intake and energy homeostasis makes it very clear that there exists a long history of research directed toward understanding the integrated central nervous system (CNS)1 circuitries through which the brain is able to achieve such a precise balance of energy intake with energy requirements. The second session of this symposium is four presentations on the “Neurobiology of Obesity,” highlighting some very exciting new technologies and some very important considerations that will clearly need to be a part of the continued effort to understand not only how the brain achieves this balance under normal circumstances, but ultimately to identify the pathological changes that lead to the many clinical abnormalities associated with abnormal regulation of energy homeostasis.

Many presentations at this meeting emphasized the potential importance of neurons in the arcuate nucleus to the regulation of feeding and metabolism. Harvey Grill convinced at least this part of the audience that, despite the recognized roles of this hypothalamic region, the “arcuate model” does not take into account the very extensive evidence, suggesting important roles for other hypothalamic, medullary, and midbrain centers in contributing to a “distributed system” for control of energy balance. Interesting reversed parallels exist here in the literature describing neural mechanisms of cardiovascular control. Thus, the medullary dorsal vagal complex was for many years thought to be the primary dominant cardiovascular control center. Our understanding of this system has now evolved to a clearer perspective, which, while still recognizing important roles for this region of the brain, has also identified the additional sophistication of this control system associated with the contributions of midbrain, hypothalamic, and even cortical cardiovascular control centers.

A variety of techniques (lesion, stimulation, recording, imaging, and more recently, knockout animals) have been used in attempts to understand the roles of specific neuronal subgroups (anatomical or chemical) in the control of energy homeostasis. These studies, while providing important information, have suffered from many difficulties of interpretation including those associated with fibers of passage, anatomical vs. functional subgroups of neurons; multiplicity of systems directed toward essential “end points,” developmental vs. adult functions in genetically modified animals, and in vivo compared with in vitro effects.

Dr. Nina Balthasar's paper in this supplement describes her recent work using sophisticated modern genetic tools to conditionally compromise the function of very specific neuronal subgroups involved in energy homeostasis and, thus, permit very precise analysis of the contributions of very specific components of the integrated neuronal circuitry. In this work, by specifically targeting leptin receptor knockout on proopiomelanocortin (POMC) neurons, she suggests that the relatively mild obesity that results (only ∼20% of that observed in full-blown leptin receptor or leptin gene knockouts) argues persuasively that leptins’ role in regulation of this population of neurons cannot be the primary determinant leading to the development of obesity. Similar observations (partial development of obesity) for leptin receptor knockout specific to agouti-related peptide (AgRP) arcuate neurons or steroidogenic factor 1–expressing neurons in the ventromedial hypothalamus similarly argue in favor of multiple sites of action for central leptin in controlling body weight.

Similar conclusions emerge from a second equally elegant set of melanocortin-4 receptor (MC4R) manipulations. Knockout mice were developed without the MC4R receptor, which was reinserted specifically into the hypothalamic paraventricular nuclei/amygdala by selective reactivation by Cre-recombinase, whose expression is driven by the Sim-1 promoter. In this case, while these animals showed reversal of the effects of MC4R knockout on food intake (no hyperphagia), effects on energy expenditure remained and could not be normalized by the melanocortin receptor agonist melanotan II injection. Thus, this approach again highlights the conclusion that separate subsets of MC4R neurons contribute to the control of different components of the apparently integrated final effects on energy homeostasis.

Thus, the picture of a complex, distributed, integrated circuitry underlying the control of energy homeostasis continues to emerge. Dr. Tamas Horvath, however, added yet another intriguing level of complexity to this overall perspective. As you will read in his paper, these circuitries also seem to show what we would describe as true neuronal plasticity, because there are both changes in numbers of synaptic densities and the occurrence of postsynaptic currents in neuropeptide Y/AgRP and POMC neurons of the arcuate in ob/ob mice. Importantly, these modifications in “synaptology” seem to be rapidly reversed (<6 hours) by exogenous leptin before effects of this peptide on food intake. These observations suggest that the changes in circuitry may, in fact, be a prerequisite for the behavioral effects of leptin.

Such synaptic remodeling does not seem to be unique to actions of leptin, because either systemic ghrelin and/or overnight food deprivation have also been shown to influence synaptic inputs to arcuate (neuropeptide Y/AgRP or POMC) and lateral hypothalamic (orexins) neurons, respectively. The precise mechanisms underlying such rapid synaptic remodeling obviously represent an important area for future attention.

In the final presentation of this session, Dr. Bill Banks focused on another important issue, that of how the many circulating energy/satiety signals influence energy homeostasis through actions on the brain, despite the fact that they do not diffuse across the normal blood–brain barrier. This barrier is also of significance in providing a clear distinction between what is present in the circulation and what is seen in the brain. The importance of this compartmentalization was also stressed by the perspective that insulin actions in the circulation should not necessarily be expected to be precisely coordinated with insulin actions behind the blood–brain barrier. In fact, it may be that the barrier allows such differentiation of function in periphery vs. brain for many other (perhaps nearly all?) signaling molecules such that they may be used as hormones in the circulation and neurotransmitters within the CNS.

Dr. Banks highlighted the intriguing roles of specific transporters for some of these signaling molecules (e.g., leptin) in permitting access to CNS regions behind the endothelial cell barrier lining the vasculature. He also provided evidence that distribution of these transporters is site specific and potentially regulated. There was considerable vigorous discussion of the position of the arcuate nucleus and the degree of protection afforded this region by the blood–brain barrier. Consensus around the view that the arcuate does in fact have a normal intact anatomical blood–brain barrier was reached, a perspective that is not in conflict with data showing effects of various circulating signals that do not diffuse across the barrier in arcuate. Possible alternate mechanisms for access were discussed, including modulation of barrier function as well as potential roles for the circumventricular organs of the brain as sensory windows in the barrier where these signals could be detected.

Collectively, the papers that follow this perspective emphasize, perhaps more than anything, that we are still a long way from understanding the complex interactions through which the brain controls food intake and energy balance. However, it is clear that the CNS is dependent on signals from the periphery—both neural and “hormonal”—to sculpt the global outputs of the complex distributed systems that integrate the control of energy balance. The approaches outlined in this supplement have the real potential to significantly enhance our knowledge of these complex systems over the next 5 to 10 years.

  • 1

    Nonstandard abbreviations: CNS, central nervous system; POMC, proopiomelanocortin; AgRP, agouti-related peptide; MC4R, melanocortin-4 receptor.