Written by Nicholas Dale & Cameron Frayling
University of Warwick, UK

https://doi.org/10.36866/pn.86.26

The brain, and in particular the hypothalamic region, contains neural circuits that control feeding and energy expenditure. Although the inexorable rise of obesity has driven intense study of these neural circuits, the possible roles of non-neuronal cells in this region have not been extensively studied. There is now increasing evidence that hypothalamic tanycytes, cells that lie at the interface between the ventricular cerebrospinal fluid and the brain parenchyma, could be active participants in these hypothalamic networks.

Everyone who has worried about their weight and diet has an intuitive understanding that somehow the brain controls body weight. This control is not just associated with the conscious ‘shall I shan’t I’ internal debate over a tasty and usually highly calorific food item, but is also autonomic i.e. controlled beneath the level of consciousness. The first evidence for this emerged from very early studies that showed that lesions of the ventromedial hypothalamus of rats led to over-eating and obesity (Hetherington & Ranson, 1940) whereas more lateral hypothalamic lesions led to a dramatic drop in food intake (Anand & Brobeck, 1951).

We now understand that neuronal nuclei within the hypothalamus (Fig. 1) are in effect multivalent chemosensors and integrators of information that control: appetite and food intake, the burning of energy and the deposition of fat (three factors that collectively determine energy balance). Physiological control of body weight requires the monitoring and integration of many signals that circulate in the body and provide information from the periphery (stomach, pancreas, fat stores etc.) to the brain (Morton et al. 2006). These key signals include not only circulating metabolites such as glucose, free fatty acids and amino acids, but also specifically secreted hormones such as leptin, grehlin and insulin. Understanding how hypothalamic networks respond to, and integrate, these signals to alter their output is an exciting and highly active area of neuroscience that is extremely relevant to the rapidly increasing incidence of obesity and related illnesses now evident in nations with advanced economies.

At the border of the hypothalamus, lining the wall of the third ventricle, are a group of enigmatic glia-like cells called tanycytes (Fig. 1). These cells have a soma that contacts the ventricular fluid, and a single process that penetrates into the brain parenchyma. They are found from the very ventral part of the ventricle wall at the median eminence b tanycytes) and progress dorsally to about mid-way up the ventricle wall (a tanycytes). The very ventral tanycytes have been of great interest in the context of the regulation of gonadotropin-releasing hormone (GnRH) secretion. The functions of the more dorsal tanycytes remain less clear. The processes of these tanycytes project into the arcuate nucleus and ventromedial hypothalamic nucleus, two key areas associated with energy balance (Fig. 1). This has led to irresistible speculation that tanycytes could be participants in the neural networks that control feeding and energy balance via sensing circulating glucose/nutrients in ventricular cerebrospinal fluid.

There is some indirect correlative evidence that tanycytes may have a physiological role in responses to glucoprivation. Injection of alloxan (an inhibitor of glucokinase) into the third ventricle kills a large number of glial fibrillary acidic protein (GFAP)-positive cells, many of which are tanycytes. These cells eventually regenerate. The glucoprivic response disappears and recovers with a time course similar to that of the disappearance and reappearance of the GFAP-positive cells including the tanycytes (Sanders et al. 2004).

The hypothesised role for tanycytes as glucosensors has been predicated around their possession of the glucose transporter involved in this role in pancreatic b cells, along with K+-ATP channels (KIR 6.1) and the associated sulfonylurea receptor (SUR) subunit as well as glucokinase. This has led to the idea that tanycytes sense glucose via the same mechanism as that of pancreatic b cells. However, the mere presence of these components does not mean that tanycytes are indeed glucosensors; a demonstration that tanycytes respond to variations in glucose concentration over a physiological range is required to strengthen this hypothesis. In a wider context, we can regard tanycytes, placed at the border of the 3rd ventricle, as potential chemosensors able to respond to a variety of signals in the cerebrospinal fluid.

Funded by the MRC via a Milstein award to perform somewhat risky science, we decided to enter the field of tanycyte signalling. Our approach was influenced by the revolution in astrocyte biology that came about through application of intracellular Ca2+ imaging. This demonstrated that astrocytes signal via variations of intracellular Ca2+ rather than membrane potential. Given that tanycytes express glial cell markers and have electrophysiological properties characteristic of glial cells (Jarvis & Andrew, 1988), we decided to apply Fura 2 Ca2+ imaging methods to the investigation of rat tanycyte signalling (Dale, 2011; Frayling et al. 2011).

Our initial investigation of the glucosensitivity of tanycytes was disappointing – when we changed the bathing glucose concentration of our hypothalamic slices, the tanycytes remained stubbornly inactive, although under certain conditions, notably in the presence of modulatory transmitters (5HT and acetylcholine), small intracellular Ca2+ signals in response to alteration of bath glucose concentrations could be observed in tanycytes (Frayling et al. 2011). The revolution in our understanding came when we selectively stimulated the tanycyte cell bodies with glucose by focal applications (an approach that mimics their selective sensing of the cerebrospinal fluid) via a patch pipette. We were then able to see rapid Ca2+ responses to glucose puffs over likely physiological concentration ranges (3–8 mM, Fig. 2A) (Frayling et al. 2011). Interestingly, even non-metabolizable analogues of glucose (2-dexoyglucose and methyl-a-D-glucopyranoside) were able to evoke these Ca2+ responses in tanycytes; however, puffs of saline or sucrose did not evoke responses indicating that these were specific Ca2+ responses to glucose and its analogues. These data suggest therefore that the mechanism of glucosensing in tanycytes may differ from the model of pancreatic b cells, as this model requires the phosphorylation of glucose to glucose-6-phosphate and subsequent metabolism in the Krebs cycle to generating ATP. The resulting alteration of the ATP:ADP ratio ultimately results in closure of the K+-ATP channel and consequent depolarization. Instead the mechanism of glucosensing in tanycytes seems more similar to that in lateral hypothalamic neurons, which can also respond to non-metabolizable glucose analogues (Gonzalez et al. 2009).

Even more interestingly, the tanycyte responses to glucose depended upon ATP signalling via P2Y1 receptors. This observation of ATP receptor-dependent glucosensing has recently been confirmed in primary cultures of tanycytes (Orellana et al. 2012). In keeping with the role of ATP that has been described in other chemosensory processes (Gourine et al. 2005), we were able to use biosensors (Fig. 2B) to show that tanycytes release ATP in response to stimulation with glucose (Frayling et al. 2011). Why tanycytes should respond to selective stimulation of the cell body rather than more general changes of glucose in the bathing medium is unclear and requires further investigation. Nevertheless our data support the idea of tanycytes as polarized chemosensory cells specifically sensing the composition of the cerebrospinal fluid.

Fig 3.

Five major nuclei in the hypothalamus form an interconnected network that regulates energy homeostasis: Arc, arcuate nucleus; VMH, ventromedial hypothalamic nucleus; DMH, dorsomedial hypothalamic nucleus; LH, lateral hypothalamus; and PVN, periventricular nucleus The Arc contains two mutually antagonistic cell types – the neuropeptide Y (NPY)-containing and the pro-opiomelanocortin (POMC)-containing neurons.

Our data also show that tanycytes do much more than respond to glucose. They can be strongly and reliably activated by ATP and neuronally derived transmitters such as histamine and acetylcholine, which are associated with wakefulness and the drive to feed. It is an intriguing question whether tanycytes may also be able to communicate with neurons (Fig. 3). This seems quite plausible as tanycyte processes come into close contact with neurons in the arcuate and ventromedial hypothalamic nucleus (Coppola et al. 2007; Levin et al. 2011) (Fig. 3). Furthermore, tanycytes release ATP, which could potentially activate P2 receptors on the dendrites of neurons (Kittner et al. 2006). Indeed there is some evidence that P2Y1 agonists and antagonists, when injected into the 3rd ventricle can alter food intake (Kittner et al. 2006). While this does not demonstrate either a role for tanycyte–neuron communication or of tanycytes in the regulation of feeding, it is consistent with the idea that tanycytes may contribute to these important hypothalamic networks, a possibility that seems well worth testing directly.

Acknowledgments

We thank the MRC for support.