NGN3-DPP2 kd mice have normal levels of active GLP-1. The active GLP-1 concentration was analyzed by bioassay, described previously (21). The experiment was repeated three times with similar results, and a representative experiment is shown (n = 4 in each group). White bars, Control groups; black bars, NGN3-DPP2 kd groups.

A, Food intake was increased in NGN3-DPP2 kd mice compared with control littermates (n = 10 per group in HFD group, n = 6 per group in LFD group). *, P < 0.038. B, Total body weight in NGN3-DPP2 kd mice and littermate controls. Body weight was monitored weekly throughout the course of the experiment (P = 0.935). C, Epididymal adipose tissue mass was increased in NGN3-DPP2 kd mice at 7 months of age on LFD (*, P < 0.045) and HFD (*, P < 0.034); perirenal adipose tissue mass and adipose tissue mass was increased in NGN3-DPP2 kd mice on HFD (*, P < 0.043), compared with control littermates. Bars, mean ± sd of data obtained with n = 3 in each LFD group and n = 4 animals in each HFD group. D, The RER of NGN3-DPP2 kd mice (7 months of age, on LFD, n = 4–5/group) was significantly higher than those of the control littermates, as measured in metabolic chambers. Data were analyzed using SAS (version 9.3.1).

Production of NGN3-DPP2 kd mice. A, Conditional vector used to express shRNA for lentiviral infection of embryonic stem cells. shRNA in pSico is not transcribed until a Cre-mediated recombination occurs, which is marked by the loss of enhanced GFP and expression of shRNA (adapted from Ref. 4). B, NGN3-GFP expression profile in brain. Frozen sections of the hypothalamus were immunolabeled, using anti-GFP antiserum. A high concentration of GFP⁺ cells is seen in the hypothalamic VMN, with smaller populations in the hypothalamic ARC. The organization of the ARC and VMN is shown with a dashed line (1). Arrows (1, 2) indicate GFP⁺ cells of the VMN; arrowheads (1,2) indicate GFP⁺ cells in the ARC. Fibers are present in the hypothalamic paraventricular nucleus (PVN) (3). No immunolabeling is seen in the hypothalamus of NGN3-Cre⁻ GFP-loxP mice (4). III, Third ventricle; DMN, dorsomedial nucleus; ME, median eminence. Scale bar, 200 μm (1) and 100 μm (2–4). C, NGN3-GFP expression profile in different tissues. Frozen sections of small intestine (1,2), pancreas (3,4), and liver (5,6) were immunolabeled, using FITC-conjugated anti-GFP antiserum (1, 3, and 5, phase contrast; 2, 4, and 6, immunofluorescence). D, DPP2 mRNA is specifically kd in the NGN3-rich pancreas but not in the liver or muscle. Total mRNA was isolated from pancreas, liver, and muscle of NGN3-DPP2 kd mice (n = 4) and control mice (n = 4: NGN3-Cre, shDPP2-loxP). DPP2 levels were measured by Taqman qRT-PCR. *, P = 0.047.

NGN3-DPP2 kd mice present with fasting hyperglycemia, glucose intolerance, hyperinsulinemia, and insulin resistance. A, The fasting glucose level is higher, and the glucose tolerance in OGTT is lower in NGN3-DPP2 kd mice. Blood glucose concentration in NGN3-DPP2 kd and control littermates was measured at the indicated times after administration of 2 mg/kg of glucose by gavage. Combined results from several experiments are shown for each data point (mean ± sem). The statistical significance for the difference between NGN3-DPP2 kd and control groups was calculated, based on AUC, with corrections for baseline differences. P < 0.002 in LFD; P < 0.04 in HFD. White circles, Control group; black circles, NGN3-DPP2 kd group in all figures. B, NGN3-DPP2 kd mice have higher baseline insulin levels and increased production of insulin in response to glucose (Glu) challenge. Plasma concentration of insulin was determined in the morning hours (baseline) and 15 min after glucose challenge. Bars, mean ± sd of data obtained with four animals in each group. *, P < 0.041 for baseline level; *, P < 0.008 in response to challenge in both age groups on LFD; *, P < 0.028 for baseline in all age groups on HFD; *, P < 0.041 for response to challenge in HFD group age 4 months and up. C, Glucose fall in response to insulin in ITT is reduced in NGN3-DPP2 kd mice on LFD and HFD diets. A representative experiment (n = 4 in each group) from three repeats is shown (mean ± sem). The statistical significance for the difference between NGN3-DPP2 kd and control groups was calculated, based on AUC.

Steatosis is more enhanced in NGN3-DPP2 kd mice. A, Livers from NGN3-DPP2 kd and control animals were weighed (n = 4 per age group). Bars, data obtained as mean ± sem. Increased weight of livers of NGN3-DPP3 kd mice was observed at both ages in the HFD group (*, P < 0.033). B, Frozen liver sections of 7-month-old mice on LFD and HFD (n = 4 per group) were stained with Sudan Black. The sections contain numerous positively stained blue-black round globules (lipid) throughout the parenchyma. NGN3-DPP2 kd mice demonstrated mild to moderate steatosis compared with controls, even when fed LFD. As expected, severe liver steatosis developed in NGN3-DPP2 kd and control mice, which received HFD. However, the section from the NGN3-DPP2 kd mouse presents with macrosteatosis (hepatocytes contain larger fat vacuoles) compared with those of the control mouse (microsteatosis). Scale bar, 100 μm. Representative sections are shown. C, The percentage areas of Sudan Black stain were quantified, using image analytical software MetaXpress 2.0 (Molecular Devices). Percent area positively stained with Sudan Black B is presented (mean ± sd), *, P < 0.05. This represents 39 and 56% increase in the area positively stained for lipid in the NGN3-DPP2 kd compared with the control mouse on HFD and LFD, respectively (n = 4/group).