Unraveling the Neural Circuitry of Energy Balance with Molecular Connectomics

The brain meticulously regulates the balance between energy intake and expenditure to support survival. The arcuate nucleus of the hypothalamus (Arc) plays a central role in this process, primarily through two functionally opposing neuronal populations: AgRP and POMCArc neurons. AgRP neurons, often referred to as the brain’s “hunger” neurons, are activated during energy deficits, driving food-seeking behavior and consumption. Conversely, POMCArc neurons promote satiety and increase energy expenditure in energy-sufficient states. The interplay between these populations is essential for maintaining energy homeostasis and is governed by both feedback and feedforward signals that monitor and anticipate metabolic needs. Feedback signals, such as circulating hormones from peripheral organs, reflect the body’s current energy reserves. Feedforward signals, conveyed by upstream neuronal networks, enable rapid modulation of AgRP and POMCArc neuron activity in response to expected changes in energy availability. The ability of these neurons to process feedforward information is crucial for adapting feeding behaviors and fine-tuning metabolic responses. However, the upstream neuronal populations that provide these signals remain largely unidentified.

To molecularly identify presynaptic partners of AgRP neurons and POMC neurons, we developed RAMPANT (Rabies Afferent Mapping by Poly-A Nuclear Transcriptomics), a technique that combines rabies virus-based circuit mapping with single-nuclei RNA sequencing to molecularly identify synaptically connected neurons. Using RAMPANT, we mapped the presynaptic partners of AgRP and POMCArc neurons, revealing numerous inputs from across the hypothalamus.

Molecular profiling of these presynaptic partners enables further investigation of their functional roles. For example, RAMPANT identified a previously uncharacterized arcuate afferent population expressing thyrotropin-releasing hormone (Trh), glucagon-like 1 receptor (Glp1r), and leptin receptor (Lepr). Further characterization of this TRH arcuate (TRHArc) population revealed that these neurons provide direct GABAergic input to AgRP neurons, and their activation suppresses food intake. We also found that TRHArc neurons respond to liraglutide, a glucagon-like 1 receptor agonist, and that silencing these neurons significantly reduces the drug’s satiating and weight-loss effects. These findings suggest a potential mechanism for the therapeutic efficacy of GLP-1 receptor agonists in weight loss.

Additionally, our study identified POMCArc neurons as presynaptic partners to AgRP neurons, a potential connection that has been largely overlooked. This reciprocal connectivity may help explain how AgRP and POMCArc neurons maintain their opposing activity to regulate energy balance. Furthermore, we found that AgRP and POMCArc neurons receive input from largely the same local neuronal populations, raising the possibility that shared afferents coordinate their opposing activity. These afferents may release distinct signals to drive opposing effects or use the same signal via different receptors and intracellular pathways.

Overall, our study leverages a novel molecular connectomics approach to advance our understanding of the neural circuits governing systemic energy metabolism. These findings provide new insights into the regulation of energy balance and may inform the development of therapeutic strategies for obesity and metabolic disorder.