Expecting to Eat: Glucagon-Like Peptide-1 and the Anticipation of Meals

We have known for more than a century that the sight, smell, and taste of food can provoke a number of physiological responses referred to as cephalic because they are initiated by the central nervous system (CNS) and not by direct gastrointestinal contact with food (1,2,3). These responses can be elicited by previously arbitrary cues that have become associated with the presentation of food, including sounds or time of day (4). In this sense, cephalic phase responses are a physiological manifestation of an animal’s prediction that food is on the way. Cephalic insulin and pancreatic polypeptide responses have been among the most well studied, and such food-anticipatory secretions are observed in multiple species, including humans (3,5,6). A variety of evidence suggests that the function of these cephalic responses is to better prepare the organism to handle the influx of nutrients about to occur (7). Several studies support the idea that these cephalic responses are indeed required for an animal to be fully capable of ingesting large meals; in their absence, intake is reduced (8,9). In this issue, Vahl et al. (10) reported that we can now add glucagon-like peptide (GLP)-1 to the list. Using rats maintained on a highly restrictive meal-feeding schedule, which tends to exaggerate meal-anticipatory responses, they identified a robust premeal rise in plasma GLP-1 levels. Moreover, they showed that intake during that scheduled meal was reduced by blockade of GLP-1 receptors during the premeal period.

The authors’ logic for investigating the possibility of a meal-anticipatory GLP-1 response is sound. GLP-1 is well known as an incretin hormone and improves glucose tolerance (11,12). These characteristics alone could render it potentially useful in preparing an animal for a large amount of food intake. However, their findings may surprise many who have been following the expanding body of research on GLP-1. It is well established that GLP-1 is secreted in response to the presence of nutrients in the gastrointestinal tract, and studies of the role of GLP-1 in feeding and energy balance have almost exclusively focused on presence during and immediately after a meal (13). The finding that preprandial GLP-1 receptor activation seems to facilitate subsequent food intake is particularly unexpected, given that to date, GLP-1 has been considered a player in satiation and satiety, and almost all previous investigations found GLP-1 to induce anorexia (12). Thus, the results of Vahl et al. (10) are exciting and provocative.

The timing of the preprandial GLP-1 response bears mentioning because it is not coincident with insulin or other cephalic phase responses measured in this model (10,14). Rather, the peak of GLP-1 secretion in this study was observed at about 60 min before the meal began, with a full return to baseline before meal onset. As the authors note, there is at least one other meal-anticipatory response that occurs even earlier: hypothalamic neuropeptide Y (NPY) is elevated approximately 2 h before a scheduled meal (15,16). Among circulating factors, though, the profile for GLP-1 seems unusual. This raises the question of whether biases of common experimental approaches have led us to miss other earlier cephalic phase responses. Blood samples are often taken only a short time before the beginning of a scheduled meal, and others use techniques such as sham feeding to examine responses during orosensory contact with food stimuli (17). At least one study did investigate and fail to observe a cephalic GLP-1 response in humans, but important differences in methods may have led to their negative finding (18). Those subjects were not maintained on a rigid meal schedule for weeks before the test days, and plasma GLP-1 levels were measured starting at only 20 min before test meals. Approaches such as these, although adequate to capture cephalic insulin and many other responses, would naturally fail to capture a considerably earlier response.

It will be of great interest to determine what neural or endocrine pathways mediate the preprandial GLP-1 response. Many other cephalic responses are driven by the vagus nerve (19), and it is not unreasonable to hypothesize that the GLP-1 response is, as well, because efferent vagal transmission is thought to play a role in gastrointestinal nutrient-induced GLP-1 release (13,20). Regardless of whether this turns out to be the case for cephalic GLP-1 release, one may still wonder about mediation in the broader sense. It is worth noting that in general, the CNS mediation of cephalic phase responses is still poorly understood. That the vagus nerve provides the efferent path from the brain to the periphery for many cephalic responses is clear, but where in the brain these responses are initiated is not so clear. At this point, we have little idea of whether the same or different neural circuits initiate different types of cephalic responses, and this is clearly an area in which further research is warranted.

Considering that we are now aware of a number of cephalic responses occurring at various times during the preprandial phase, one might reasonably ask how they relate to one another. In their 1985 review of cephalic phase responses, Powley and Berthoud (17) made the distinction between primary and secondary responses, the idea being that one cephalic response may serve as the stimulus for the next. Given that the premeal GLP-1 peak is temporally situated between the rise in hypothalamic NPY and the later premeal insulin response, one may wonder whether these events are causally linked or if they occur independently of one another. Because of its well-established incretin effect, it is plausible that premeal GLP-1 contributes to the subsequent rise in insulin, although this possibility remains to be experimentally tested. Is it also possible that the earlier rise in NPY contributes to the GLP-1 and/or insulin responses, linking the three in a causal chain? There are several reports that CNS NPY administration can increase insulin release (21,22) as well as at least one study in which hypothalamic NPY increased efferent vagal activity to the gastrointestinal tract (23). It should be noted, however, that some cephalic phase responses are not consistent with the idea of causal interrelationship with GLP-1. For example, the hormone ghrelin also rises shortly before scheduled meals (14), but GLP-1 itself seems not to affect plasma ghrelin levels, and the potent GLP-1 receptor agonist exendin 4 actually reduces circulating ghrelin (24). The timing of each of these responses may ultimately be merely coincidental, and further study will be required to discern any potential links.

Perhaps the most surprising finding in the article by Vahl et al. (10) is that blockade of GLP-1 receptors starting 2 h before scheduled food access increased intake during the meal. The idea that a cephalic response can promote intake is not novel. Cephalic-phase insulin appears to function in the much same way, making it possible for animals to ingest more during the subsequent meal (7). This finding is surprising because previously GLP-1 has been considered an anorexic signal, possibly involved in meal termination, and administration of a GLP-1 receptor antagonist immediately before food access increases intake (25,26). Notably, when Vahl et al. (10) delivered the antagonist 15 min before the scheduled meal, it did increase subsequent intake, as one would expect. With this simple experiment, they showed that scheduled meal feeding does not wholly convert GLP-1 from an anorexic signal into an orexigenic signal but rather that premeal GLP-1 receptor activation promotes intake whereas during or after a meal, GLP-1 receptor activation remains intake-inhibitory. There are other examples of hormones that have such biphasic effects, but these usually occur at doses on opposite ends of the dose-response function or via action on different receptor populations or subtypes, and such opposing effects do not typically appear so close together in time (for examples of such effects of estrogens, see Refs. 27 and 28). How does activation of the same receptor just a few hours later produce the opposite behavioral effect? And what of the many other known effects of GLP-1 receptor activation: are these, too, reversed during the premeal period, or is this change in consequence specific to food intake?

The apparent ability of preprandial GLP-1 to facilitate subsequent food intake is notable not only because of the previous findings of the opposite role for GLP-1 but also simply because we know far less about intake-excitatory factors than we do about those that are intake-inhibitory. Exactly how premeal GLP-1 receptor activation promotes intake will be an important topic for investigation. There are at least two broad possibilities for how this effect may occur. First, GLP-1 may work in a permissive manner, similar to that hypothesized for cephalic phase insulin. That is, preprandial GLP-1 contributes to more efficient glucose disposal during the meal and therefore allows the animal to consume more food. Could faster clearance of incoming nutrients, making room for more—so to speak—translate into reduced satiation and in this way permit further intake? A second possibility is that preprandial GLP-1 is itself intake stimulatory, increasing motivation to eat and/or increasing the palatability of the food once the animal starts eating. This is not as far-fetched as it may sound; there is evidence that GLP-1 signaling can enhance sweet taste (29). Although the permissive factor hypothesis may be more consistent with the traditional proposed function of cephalic phase responses, the second possibility should not be discounted without further study. Moreover, these potential mechanisms are not mutually exclusive. These questions about GLP-1’s mechanism of action underscore how little we truly know about meal-anticipatory signals in general.

The paper by Vahl et al. (10) is an important step forward in understanding the physiology of meal anticipation as well as the role of GLP-1 in feeding and energy balance. Their findings should serve as a reminder of the tremendous influence of schedule, routine, and other external stimuli on homeostatic responses. The suggestion that GLP-1 receptor activation may have a radically different effect on feeding depending on time relative to the start of a meal further highlights the complexity of the biological systems that regulate energy balance. Physiology in day-to-day life, in contrast with that observed under reductionist laboratory conditions, is likely more dynamic and flexible than we realize.