eNews January 06, 2016

Another reason why your diet is doomed – “Hunger” neurons promote negative feelings

Dieting challenges evolution

In its simplest terms, weight loss occurs when the amount of energy consumed in the form of food is less than the amount of energy burned. This can be accomplished by eating less or exercising more. With either approach, the goal is to create a caloric debt that will be resolved by burning stored carbohydrate, protein, or fat. Challenges to losing the holiday weight (alternatively a beer gut, Freshman Fifteen, etc.) are simple:  eating feels good and being hungry is uncomfortable. Behaviors that evolved as survival mechanisms to ensure that an animal feeds itself become inconvenient and potentially detrimental side effects in industrialized human populations where cardiovascular disease, diabetes, and obesity, rather than starvation, pose greater risks to long-term survival.

To explain the less commonly researched “discomfort when hungry” phenomenon, researchers at the Howard Hughes Medical Institute (HHMI) Janelia Research Campus hypothesized that specific hypothalamic neurons stimulate food-seeking behaviors to eliminate the negative feelings associated with energy deficit (Betley et al. 2015).

alternative mechanisms driving food consumption
Figure 1. Alternative mechanisms for driving food consumption. Eating not only is a necessity, it is a learned response to hunger, too. A) In the positive-valance mechanism, hungry animals learn that eating feels good. B) In the negative-valence mechanism, hunger feels bad to animals, and they learn that eating makes these bad feelings dissipate.

The Hunger Games  – Neural manipulation suggests AGRP signals are negative

Previously demonstrated by other groups, hypothalamic AGRP neurons (agouti-related protein-expressing neurons) fire during energy deficit, such as food restriction, and quickly lead to food-seeking behavior and food-consumption.  To determine whether these neurons were associated with positive signals (such as those associated with food reward) versus negative signals (such as those associated with the uncomfortable feelings of hunger), mice either with photoactivatable AGRP neurons (STOCK Agrptm1(cre)Lowl/J 012899 X B6;129S-Gt(ROSA)26Sortm32(CAG-COP4*H134R/EYFP)Hze/J  012569 F1 hybrids) or with virally-transduced AGRP neurons that could be inhibited pharmacologically were subjected to flavor and place preference tests.

In flavor preference tests, AGRP neuron activation reduced an animal’s preference for previously preferred flavored gel compared to unstimulated controls, suggesting that a once tasty flavor was now a less pleasant experience. In contrast, mouse preferences for a preferred gel increased when AGRP neurons in food-restricted mice were pharmacologically inhibited compared to the preferences of untreated controls, suggesting that AGRP inhibition relieved negative feelings. In complementary place preference experiments in which mice were offered two experimental chambers to explore, well-fed mice avoided chambers in which they had previously received AGRP stimulation, indicating AGRP stimulation was undesirable. Similarly, food-restricted mice spent more time in chambers where they had previously received AGRP inhibition, indicating that relief from AGRP simulation was preferred. Together, these results suggest that AGRP neurons stimulate negative signals and generate Pavlovian responses learned through these negative experiences.

AGRP neurons anticipate feeding and turn off

To support their behavioral studies, the HHMI Janelia group performed deep brain imaging in freely-moving AGRP-specific calcium reporter mice using miniature head-mounted fluorescent microscopes.  As expected, food-restricted mice showed higher fluorescent signals than mice fed ad libitum, demonstrating that food restriction activates AGRP neurons.  Calcium signals quickly diminished when food-restricted mice were allowed to eat. Interestingly, the calcium signals dropped before eating began, including cases where food was visible but not accessible for consumption. In contrast, calcium signals dropped slightly when mice were presented a wooden pellet resembling food, but quickly returned when the mice discovered that the wooden pellet was not food. Together, these experiments indicate that AGRP neuron activity is reduced when feeding cues are present.  Further, the rapidity with which the AGRP signals terminate upon feeding or recommence upon presentation with non-food items strengthen the hypothesis that AGRP neurons promote negative, rather than positive, signals.

Collectively, these results indicate that energy homeostasis depends, in part, on alleviating negative signals produced during hunger. Findings from this study corroborate the negative emotions people experience when dieting and point to alternative avenues to regulate food intake. The negative-valence mechanism that AGRP neurons utilize contrasts with the majority of hunger-associated neurons previously studied, which stimulate reward pathways and result in positive feelings when hunger is satisfied.  Certainly, then, fine control of energy homeostasis is accomplished by possessing both kinds of neurons.