Zhao, Z.; Xu, B.; Anthony, S.; Subramanian, S.; Granger, B.; Von-Walter, C.; Mizrachi, E.; Kidd, M.; Srigiriraju, A.; McKie, I.; Li, Z.; Bolton, M. M.; Berto, S.; Stern, S. A.

The insular cortex (insula) is an interoceptive hub, which senses internal states such as hunger, thirst, pain, and emotions. Previous studies, however, suggest that the insula directly senses internal states, but the mechanisms remain elusive. We identified a novel population of glutamatergic leptin receptor-positive (LepR+) cells with a unique morphology in the insula (INSLepR). Based on leptin\'s known role in signaling adiposity levels, we hypothesized that INS-LepR neurons detect internal states and thus regulate food intake and body weight. Accordingly, we found that leptin administration in the insula increases intrinsic excitability of INSLepR cells, suppresses feeding, and reduces body weight. Moreover, optogenetic stimulation of INSLepR cells suppresses learned operant feeding and leads to avoidance in a real time place preference context. We also found that INSLepR neurons encode operant feeding bouts, and their activity is modulated specifically by hunger, but not thirst, states. Single-nuclei sequencing and ribosome profiling revealed that INSLepR is expressed on both Layer 6 neurons and on vascular cells, suggesting that leptin can be delivered directly through receptor-mediated transport through the blood brain barrier into the insula. . INSLepR neurons form mainly local connections within the insula and an external projection to the basolateral amygdala (BLA), and optogenetic stimulation of BLA terminals also regulated learned operant feeding. Moreover, we found that administration of leptin alters insula neural dynamics in response to feeding, but not drinking, behavior, and reshapes the transcriptome, suggesting that internal state information provided by leptin is used by the insula to coordinate feeding behavior. Taken together, our data supports a model for direct interoceptive input to the insula, in which INSLepR cells integrate adiposity level signals to regulate feeding and body weight in a learned manner.