D into the left lateral ventricle using the following coordinates from Bregma: 1.0 mm lateral, 0.46 mm posterior and 2.2 mm ventral. For third ventricle cannulations the following coordinates from Bregma were used: 0.0 mm lateral, 1.3 mm posterior and 5.7 mm ventral. The guide cannula was secured to the skull surface with dental cement (GC Europe N.V., Leuven, Belgium) and the anesthesia was antagonized using 2.5 mg/kg BW Antipamezol (Pfizer, Capelle a/d IJssel, The Netherlands), 0.5 mg/kg BW Flumazenil (Roche, Mijdrecht, The Netherlands) and 1.2 mg/kg BW Naloxon (Orpha, Purkersdorf, Austria). Animals were single housed after the surgery.vehicle (PBS, 100 mL). Both drugs were tested once, in the number of mice indicated. Blood samples were taken from the tail tip into chilled heparin-coated capillaries (Vitrex Medical, Herlev, Denmark) at the indicated time points 
up to 90 min after tyloxapol injection. The tubes were kept on ice after which they were centrifuged (12.000 rpm for 5 min at 4uC). Plasma TG concentration was determined using a commercially available kit according to the instructions of the manufacturer (no. 11488872, Roche Molecular Biochemicals, Indianapolis, IN) At 120 min, the animals were sacrificed and blood was collected by orbital puncture for isolation of VLDL by density gradient ultracentrifugation [36]. 35S-activity was measured in the VLDL fraction and VLDL-apoB production rate was calculated as dpm.h21 [37].Verification of Cannula PositionAfter termination of mice, brains were taken out and fixed by submerging in 4 paraformaldehyde for 48 hours (Sigma-Aldrich, Zwijndrecht, the Netherlands) followed by 30 sucrose (SigmaAldrich, Zwijndrecht, the Netherlands) in PBS for at least 24 hours, until the brain has sank to the bottom of the container. Cannula position was verified in 30 mm thick brain cryosections mounted on microscopic slides. The sections were fixated and defatted in CARNOY solution (100 ethanol, chloroform and acetic acid in a 6:3:1 ratio), hydrated by descending ethanol concentrations (10096-70 ) in MilliQ (MQ) water, and a Nissl staining was performed using 301-00-8 biological activity cresyl violet (Sigma-Aldrich, Zwijndrecht, the Netherlands): 0.9 g cresyl violet, 300 mL MQ, 2.25 mL 10 acetic acid, pH 4.5. The sections were then dehydrated in ascending ethanol concentrations (70-96-100-100 ) followed by 2 times isopropanol and 2 times Histo-Clear (National 
diagnostics, Atlanta, USA). Cover slips were mounted using xylene, and the cannula position was verified by locating the cannula track in the tissue. When the cannula track ended within the Felypressin site respective ventricle, the cannula was considered to be positioned correctly. The average success rates of LV and 3V cannulation were ,85 and ,60 respectively.Food Intake MeasurementAfter a recovery period of at least 1 week, the mice received a pre-weighed amount of food after which basal food intake was measured for two hours, starting from 09:00 a.m. One day later, mice received an i.c.v. injection of NPY (0.2 mg/kg in 1 mL of artificial cerebrospinal fluid, aCSF) under light isoflurane anesthesia (1.5 in air). Food was weighed before and one and two hours after waking up from the anesthesia to determine NPYinduced food intake.Hepatic VLDL-TG and VLDL-apoB ProductionIn experiments performed under complete anesthesia, 4 h fasted mice were anesthetized with 6.25 mg/kg Acepromazine (Alfasan, Woerden, The Netherlands), 6.25 mg/kg Midazolam (Roche, Mijdrecht, The Netherlands), and 0.3.D into the left lateral ventricle using the following coordinates from Bregma: 1.0 mm lateral, 0.46 mm posterior and 2.2 mm ventral. For third ventricle cannulations the following coordinates from Bregma were used: 0.0 mm lateral, 1.3 mm posterior and 5.7 mm ventral. The guide cannula was secured to the skull surface with dental cement (GC Europe N.V., Leuven, Belgium) and the anesthesia was antagonized using 2.5 mg/kg BW Antipamezol (Pfizer, Capelle a/d IJssel, The Netherlands), 0.5 mg/kg BW Flumazenil (Roche, Mijdrecht, The Netherlands) and 1.2 mg/kg BW Naloxon (Orpha, Purkersdorf, Austria). Animals were single housed after the surgery.vehicle (PBS, 100 mL). Both drugs were tested once, in the number of mice indicated. Blood samples were taken from the tail tip into chilled heparin-coated capillaries (Vitrex Medical, Herlev, Denmark) at the indicated time points up to 90 min after tyloxapol injection. The tubes were kept on ice after which they were centrifuged (12.000 rpm for 5 min at 4uC). Plasma TG concentration was determined using a commercially available kit according to the instructions of the manufacturer (no. 11488872, Roche Molecular Biochemicals, Indianapolis, IN) At 120 min, the animals were sacrificed and blood was collected by orbital puncture for isolation of VLDL by density gradient ultracentrifugation [36]. 35S-activity was measured in the VLDL fraction and VLDL-apoB production rate was calculated as dpm.h21 [37].Verification of Cannula PositionAfter termination of mice, brains were taken out and fixed by submerging in 4 paraformaldehyde for 48 hours (Sigma-Aldrich, Zwijndrecht, the Netherlands) followed by 30 sucrose (SigmaAldrich, Zwijndrecht, the Netherlands) in PBS for at least 24 hours, until the brain has sank to the bottom of the container. Cannula position was verified in 30 mm thick brain cryosections mounted on microscopic slides. The sections were fixated and defatted in CARNOY solution (100 ethanol, chloroform and acetic acid in a 6:3:1 ratio), hydrated by descending ethanol concentrations (10096-70 ) in MilliQ (MQ) water, and a Nissl staining was performed using cresyl violet (Sigma-Aldrich, Zwijndrecht, the Netherlands): 0.9 g cresyl violet, 300 mL MQ, 2.25 mL 10 acetic acid, pH 4.5. The sections were then dehydrated in ascending ethanol concentrations (70-96-100-100 ) followed by 2 times isopropanol and 2 times Histo-Clear (National diagnostics, Atlanta, USA). Cover slips were mounted using xylene, and the cannula position was verified by locating the cannula track in the tissue. When the cannula track ended within the respective ventricle, the cannula was considered to be positioned correctly. The average success rates of LV and 3V cannulation were ,85 and ,60 respectively.Food Intake MeasurementAfter a recovery period of at least 1 week, the mice received a pre-weighed amount of food after which basal food intake was measured for two hours, starting from 09:00 a.m. One day later, mice received an i.c.v. injection of NPY (0.2 mg/kg in 1 mL of artificial cerebrospinal fluid, aCSF) under light isoflurane anesthesia (1.5 in air). Food was weighed before and one and two hours after waking up from the anesthesia to determine NPYinduced food intake.Hepatic VLDL-TG and VLDL-apoB ProductionIn experiments performed under complete anesthesia, 4 h fasted mice were anesthetized with 6.25 mg/kg Acepromazine (Alfasan, Woerden, The Netherlands), 6.25 mg/kg Midazolam (Roche, Mijdrecht, The Netherlands), and 0.3.
