7BL/6J and IRS2-KO mice aged 2,4 months, and then, single neurons were prepared, as previously reported. Briefly, the whole ARC was dissected out and incubated with HKRB containing 20 units/ml papain, 1 mM cystein, 0.015 mg/ml deoxyribonuclease, 0.75 mg/ml BSA and 10 mM glucose for 17 min at 36uC, followed by gentle mechanical trituration for 5,10 min. The single neurons obtained were distributed onto coverslips and incubated in the humidified chamber at 30uC for 30 min to 6 hr until use. Measurements of i in ARC neurons were carried out by ratiometric fura-2 fluorescence imaging with 510 nm emission and 340/380 nm excitation by Argus-50 system as reported previously. HKRB containing 10 mM glucose was used as the perfusate. Data Analysis All data were shown as means 6 SEM. Statistical analysis was performed by one-way ANOVA followed by Tukey’s or Dunnett’s multiple comparison tests, unpaired t-test, or chi-square test using the Prism 4. P,0.05 was considered significant. Measurements 17149874 of Insulin Concentration in Pancreatic Artery, Pancreatic Vein, and Portal Vein in Rats Male Wistar rats after overnight fasting or with ad lib condition were used. The blood samplings were performed at 10:00,11:00, as previously reported. Briefly, after injection of pentobarbital anesthesia to the rats, blood samples were collected from the pancreatic artery and vein and portal vein. Plasma insulin concentrations were measured using an enzyme-linked immunosorbent assay kit. Results Insulin Concentration-dependently Increases i in NG Neurons We measured direct effect of insulin on i in the single NG neurons isolated from ICR mice. Insulin at 10213,1026 M was sequentially applied to NG neurons with 4 min pulses and at least 8 min washing periods. As shown in Fig. 1A and B, a NG neuron responded to insulin at 10210, 1029 and 1027 M, while another NG neuron responded to insulin at 1027 and 1026 M. Insulin at 10212,1026 M, but not 10213 M, increased i. The incidence and amplitude of i responses took plateau with insulin at 10211,1028 M, and the Insulin Activates 10712926 Vagal Afferent Neurons incidence further increased at 1027 M and 1026 M insulin, thus showing a concentration-dependency with two components. Relationship between Insulin-, CCK- and CAP-responsive NG Neurons CCK-8 and capsaicin are the well characterized substances that directly activate NG neurons. We examined whether insulin-responsive neurons are distinct from or 1702259-66-2 overlap with CCK-8- and/or CAP-responsive neurons in the NG. NG neurons were exposed sequentially to insulin, CCK-8, CAP, and KCl . We previously reported that 1028 M for CCK-8 and 1027 M for CAP are the maximal doses for inducing i increases in NG neurons. Insulin, CCK-8, and CAP raised i in 53 of 500, 207 of 500, and 307 of 500 
neurons examined, respectively. Among 53 NG neurons that responded to insulin, 37 neurons responded to CCK-8 and all to CAP. Thus, the insulin responders were largely CCK responders and entirely CAP responders. Neurochemical Characterization of Insulin-responsive Neurons The NG neurons highly express CART. We examined whether insulin-responsive NG neurons contain CART. The effects of insulin on i were measured in single NG neurons, which were subsequently immunostained with antibody against CART. The neuron exemplified in Fig. 1F responded to insulin with an increase in i and was subsequently proved to be immunoreactive to CART. Out of 23 neurons that responded to 4 Insulin Activates Vagal Afferent Neurons insulin,
