Amygdala NPY circuits are critical for the development of accelerated obesity under chronic stress

Neuropeptide Y (NPY) exerts powerful feeding related functions in the hypothalamus. However, NPY is also present in extra-hypothalamic nuclei, however their influence on energy homeostasis is unclear. Here we uncover a previously unknown feeding stimulatory pathway that is activated under conditions of stress in combination with calorie dense food with NPY neurons in the central amygdala (CeA) being responsible for an exacerbated response to a combined stress and high fat diet intervention. CeA NPY neuron specific Npy overexpression mimics the obese phenotype seen in a stress/HFD model, which is prevented by the selective ablation of Npy. Using food intake and energy expenditure (EE) as readout we demonstrate that selective activation of CeA NPY neurons results in increased food intake and a decrease in EE, which requires the presence of NPY. Mechanistically it is the diminished insulin signalling capacity on CeA NPY neurons under stress combined with HFD conditions that leads to the exaggerated development of obesity. RNA from both sides of the amygdala of NpyCre/+;TRAPlox/lox mice were immunoprecipitated by using the TRAP protocol described above. Input RNA and immunoprecipitated (IP) RNA from three mice were pooled together to give enough material for RNA-sequencing. Two replicates were included for both IP and input samples. A total of 200 ng of the IP RNA and its corresponding input control was used for RNA-sequencing experiment. Illumina HiSeq HT chemistry sequencing with 100 base pair single-end read was employed. RNA quality and quantity were determined by Bioanalyzer Nanokit (Agilent). The same amount of RNA from both the immunoprecipitated samples and the input samples were used for library preparation (Illumina) in the Australian Genome Research Facility. Sequencing quality check of the raw sequencing reads was performed using the FastQC tool (Andrew S. 2010) through the Galaxy Project platform. No over-represented sequences were detected, indicating that adapter sequences did not affect the overall dataset. Read alignment was performed using TopHat (Trapnell et al., 2012) with default parameter, and iGenome (http://cole-trapnell-lab.github.io/cufflinks//igenome_table/ index.html) mm10 UCSC transcriptome annotation. Differential expression was determined by Cufflinks and Cuffdiff analysis, and a false discovery rate of <0.05 were considered to be statistically significant.