Speaking, singing, coughing, laughing, yelling, yawning, chewing—we use our jaws for many purposes. Each action requires a complex coordination of muscles whose activity is managed by neurons in the brain.

But it turns out that the neural circuit behind the jaw movement most essential to survival—eating—is surprisingly simple, as researchers from Rockefeller University recently described in a new paper in Nature. Christin Kosse and other scientists from the Laboratory of Molecular Genetics, headed by Jeffrey M. Friedman, have identified a three-neuron circuit that connects a hunger-signaling hormone to the jaw movements of chewing. The intermediary between these two is a cluster of neurons in a specific area of the hypothalamus that, when damaged, has long been known to cause obesity.

Strikingly, inhibiting these so-called BDNF neurons not only leads animals to consume more food but also triggers the jaw to make chewing motions even in the absence of food or other sensory input that would indicate it was time to eat. And stimulating them reduces food intake and puts a halt to the chewing motions, acting as an effective curb against hunger.

The simple architecture of this circuit suggests that the impulse to eat may be more similar to a reflex than was previously considered—and may provide a new clue about how the initiation of feeding is controlled. 

“It’s surprising that these neurons are so keyed to motor control,” says study first author Christin Kosse, a research associate in the lab. “We didn’t expect that limiting physical jaw motion could act as a kind of appetite suppressant.” 

More than a feeling?

The impulse to eat is driven not just by hunger but by many factors. We also eat for pleasure, community, ritual, and habit; and smell, taste, and emotions can impact whether we eat too. In humans, eating can also be regulated by the conscious desire to consume more or less. The causes of obesity are equally complex, the result of a dynamic interplay of diet, environment, and genes. For example, mutations in several genes, including leptin, the hunger-suppressing hormone, and brain-derived neurotrophic factor (BDNF), lead to gross overeating, metabolic changes, and extreme obesity, suggesting that both factors normally suppress appetite.

When Friedman’s team began this study, they sought to pinpoint the location of the BDNF neurons that curtail overeating. That’s eluded scientists for years, because BDNF neurons, which are also primary regulators of neuronal development, differentiation, and survival, are widespread in the brain. 

In the current study, they homed in on the ventromedial hypothalamus (VMH), a deep-brain region linked to glucose regulation and appetite. It’s well-documented that damage in the VMH can lead to overeating and eventually obesity in animals and people, just as mutated BDNF proteins do. Perhaps the VMH played a regulatory role in feeding behavior.

They hoped that by documenting BDNF’s impact on eating behavior, they could find the neural circuit underpinning the process of transforming sensory signals into jaw motions. They subsequently found that BDNF neurons in the VMH—but not elsewhere—are activated when animals become obese, suggesting that they are activated when weight is gained in order to suppress food intake. Thus when these neurons are missing, or there is a mutation in BDNF, animals become obese.

Chewing without food

In a series of experiments, the researchers then used optogenetics to either express or inhibit the BDNF neurons in the ventromedial hypothalamus of mice. When the neurons were activated, the mice completely stopped feeding, even when they were known to be hungry. Silencing them had the opposite effect: the mice began to eat—and eat and eat and eat, wolfing down nearly 1200% more food than they normally would in a short period of time.

“When we saw these results, we initially thought that perhaps BDNF neurons encode valence,” Kosse says. “We wondered if when we regulated these neurons, the mice were experiencing the negative feeling of hunger or maybe the positive feeling of eating food that’s delicious.”

But subsequent experiments disproved that idea. Regardless of the food given to the mice—either their standard chow or food packed with fat and sugar, like the mouse equivalent of a chocolate mousse cake—they found that activating the BDNF neurons suppressed food intake.

And because hunger is not the only motivation to eat—as anyone unable to skip dessert can attest—they also offered high-palatable food to mice that were already well fed. The animals chowed down until the researchers inhibited the BDNF neurons, at which point they promptly stopped eating. 

“This was initially a perplexing finding, because prior studies have suggested that this ‘hedonic’ drive to eat for pleasure is quite different from the hunger drive, which is an attempt to suppress the negative feeling, or negative valence, associated with hunger by eating,” Kosse notes. “We demonstrated that activating BDNF neurons can suppress both drives.”

Equally striking was that BDNF inhibition caused the mice to make chewing motions with their jaw, directed at any object in their vicinity even when food was not available. This compulsion to chew and bite was so strong that the mice gnawed on anything around them—the metal spout of a water feeder, a block of wood, even the wires monitoring their neural activity.

The circuit

But how does this motor-control switch connect to the body’s need or desire for food?

By mapping the inputs and outputs of the BDNF neurons, the researchers discovered that BDNF neurons are the linchpin of a three-part neural circuit linking hormonal signals that regulate appetite to the movements required to consume it.

At one end of the circuit are special neurons in the arcuate nucleus (Arc) region of the hypothalamus that pick up hunger signals such as the hormone leptin, which is produced by fat cells. (A high amount of leptin means the energy tank is full, while a low leptin level indicates it’s time to eat. Animals with no leptin become obese.) The Arc neurons project to the ventromedial hypothalamus, where their signals are picked up by the BDNF neurons, which then project directly to a brainstem center called Me5 that controls the movement of jaw muscles.

“Other studies have shown that when you kill Me5 neurons in mice during development, the animals will starve because they’re unable to chew solid foods,” says Kosse. “So it makes sense that when we manipulate the BDNF neurons projecting there, we see jaw movements.”

It also explains why damage in the VMH causes obesity, Friedman says. “The evidence presented in our paper shows that the obesity associated with these lesions is a result of a loss of these BDNF neurons, and the findings unify the known mutations that cause obesity into a relatively coherent circuit.”

The findings suggest something deeper about the connection between sensation and behavior, he adds. “The architecture of the feeding circuit is not very different from the architecture of a reflex,” says Friedman. “That’s surprising, because eating is a complex behavior—one in which many factors influence whether you’ll initiate the behavior, but none of them guarantee it. On the other hand, a reflex is simple: a defined stimulus and an invariant response. In a sense, what this paper shows is that the line between behavior and reflex is probably more blurred than we thought. We hypothesize that the neurons in this circuit are the target of other neurons in the brain that convey other signals that regulate appetite.”

This hypothesis is consistent with the work of early 20th century neurophysiologist Charles Sherrington, who pointed out that while cough is regulated by a typical reflex, it can be modulated by conscious factors, such as the desire to suppress it in a crowded theater.

Kosse adds, “Because feeding is so essential to basic survival, this circuit regulating food intake may be ancient. Perhaps it was a substrate for ever-more complex processing that occurred as the brain evolved.”

To that end, in the future the researchers want to explore the brainstem area known as Me5 with the idea that the jaw’s motor controls might be a useful model for understanding other behaviors, including compulsive, stress-related mouth actions such as gnawing on a pencil eraser or strands of one’s hair.

“By examining these premotor neurons in the Me5, we might be able to understand whether there are other centers that project into the region and influence other innate behaviors, like BDNF neurons do for eating,” she says. “Are there stress-activated or other neurons that project into there as well?”