Direct leptin action on POMC neurons regulates glucose homeostasis and hepatic insulin sensitivity in mice

Abstract

Leptin action on its receptor (LEPR) stimulates energy expenditure and reduces food intake, thereby lowering body weight. One leptin-sensitive target cell mediating these effects on energy balance is the proopiomelano-cortin (POMC) neuron. Recent evidence suggests that the action of leptin on POMC neurons regulates glucose homeostasis independently of its effects on energy balance. Here, we have dissected the physiological impact of direct leptin action on POMC neurons using a mouse model in which endogenous LEPR expression was prevented by a LoxP-flanked transcription blocker (loxTB), but could be reactivated by Cre recombinase. Mice homozygous for the Lepr(loxTB) allele were obese and exhibited defects characteristic of LEPR deficiency. Reexpression of LEPR only in POMC neurons in the arcuate nucleus of the hypothalamus did not reduce food intake, but partially normalized energy expenditure and modestly reduced body weight. Despite the moderate effects on energy balance and independent of changes in body weight, restoring LEPR in POMC neurons normalized blood glucose and ameliorated hepatic insulin resistance, hyperglucagonemia, and dyslipidemia. Collectively, these results demonstrate that direct leptin action on POMC neurons does not reduce food intake, but is sufficient to normalize glucose and glucagon levels in mice otherwise lacking LEPR.

Figure 1. LEPRs are only reactivated in the ARH in Leprlox^TB × POMC-cre mice.

Representative images of leptin-stimulated (5 mg/kg body weight; i.p.) increases in phosphorylation of Stat3 (p-Stat3) immunoreactivity in 18-hour–fasted, 8-week-old, male WT, LEPR-null mice generated by inserting floxed transcription blocking sequences in the LEPR gene (Leprlox^TB) and mice in which LEPRs are selectively reactivated in POMC neurons (Leprlox^TB × POMC-cre). Hypothalamus is shown in A–C. Double immunohistochemistry for p-Stat3 (black) and β-endorphin (brown) is shown in D–F. Insets in D–F show increased magnification of cells, if any, which overlap. Hindbrain is shown in G–I. 3v, third ventricle; VMH, ventromedial hypothalamus; AP, area postrema. n = 8–10 for p-Stat3; n = 3 for β-endorphin. Scale bars: 400 μm (C); 200 μm (F); 10 μm (F, inset).

Figure 2. Body weight is modestly reduced only in male Leprlox^TB × POMC-cre mice after 12 weeks of age compared with Leprlox^TB littermates.

Body weight and composition in WT, WT expressing Cre-recombinase in POMC neurons, LEPR-null mice generated by inserting floxed transcription blocking sequences in the Lepr gene (Leprlox^TB), and littermates in which LEPRs are selectively reactivated in POMC neurons (Leprlox^TB × POMC-cre). Body weight in male and female mice is shown in A and D. Fat and lean mass assessed by NMR in male mice at 12 and 20 weeks are shown in B and C. *P < 0.05; ^†P < 0.05 versus WT controls and Leprlox^TB, respectively. n = 12–15 and 7–9 for male and female mice, respectively. Data are presented as mean ± SEM.

Figure 3. Selective reactivation of LEPR expression in POMC neurons stimulates increased energy expenditure, but does not affect food intake in mice.

Metabolic cage studies in male, 6-week-old WT, WT expressing Cre-recombinase in POMC neurons, LEPR-null mice generated by inserting floxed transcription blocking sequences in the Lepr gene (Leprlox^TB), and littermates in which LEPRs are selectively reactivated in POMC neurons (Leprlox^TB × POMC-cre) assessed using TSE Metabolic Cages. Body weight is shown in A. Oxygen consumption, carbon dioxide production, RER, total activity, and food consumption normalized to total body weight are shown in B–F, respectively. *P < 0.05; ^†P < 0.05 versus WT controls and Leprlox^TB mice, respectively. n = 7–10 mice/genotype. Data are presented as mean ± SEM.

Figure 4. Selective reactivation of LEPR expression in POMC neurons improves hyperglycemia and hyperinsulinemia in mice.

Blood glucose and plasma insulin in 5-hour–fasted male and female mice at various ages are shown in A–D. Plasma leptin in 24-hour–fasted male mice is shown in E. *P < 0.05; ^†P < 0.05 versus WT controls and Leprlox^TB mice, respectively. n = 8–10 mice/genotype. Data are presented as mean ± SEM.

Figure 5. Selective reactivation of LEPR expression in POMC neurons improves hepatic insulin sensitivity in mice.

Hyperinsulinemic-euglycemic (10 mU/kg/min, 150 mg/dl, respectively) clamps of 120 minutes in conscious, chronically catheterized, 4- to 5-hour–fasted, 8-week-old male WT, WT expressing Cre-recombinase in POMC neurons, LEPR-null mice generated by inserting floxed transcription blocking sequences in the Lepr gene (Leprlox^TB), and littermates in which LEPRs are selectively reactivated in POMC neurons (Leprlox^TB × POMC-cre). Body weight on day of experiment is shown in A. Basal and clamp blood glucose, GIR, and plasma insulin are shown in B–D, respectively. Basal and insulin-stimulated (clamp steady-state [t = 80–120 minutes]) glucose production and disposal determined using [3-³H]glucose are shown in E and F. Blood samples were taken from the cut tail. *P < 0.05; ^†P < 0.05 versus WT controls and Leprlox^TB × POMC-cre mice, respectively; ^#P < 0.05 comparing basal to clamp values within a genotype. Data are presented as mean ± SEM.

Figure 6. Inappropriate plasma glucagon levels are corrected in Leprlox^TB × POMC-cre mice, and this contributes to improvements in hepatic insulin action.

(A) Shown are body weight and 12-hour–fasted/carbohydrate-refed plasma glucagon (B) as well as insulin (C) in WT, WT expressing Cre-recombinase in POMC neurons, leptin receptor LEPR-null mice generated by inserting floxed transcription blocking sequences in the Lepr gene (Leprlox^TB), and littermates in which LEPRs are selectively reactivated in POMC neurons (Leprlox^TB × POMC-cre) (n = 8–10 mice/genotype). (D and E) Pancreatic glucagon and insulin content, respectively (n = 12–15 mice/genotype). (F) Blood glucose during 120-minute hyperinsulinemic-euglycemic (25 mU/kg/min, 150 mg/dl, respectively) clamps plus somatostatin (9 ng/kg/min) in conscious, chronically catheterized, 4- to 5-hour–fasted, 8- to 9-week-old male mice (n = 6 mice/genotype). (G) GIR is shown. (H and I) Basal and clamp plasma glucagon and insulin, respectively, are shown. (J and K) Basal and insulin-stimulated (clamp steady-state [t = 80–120 minutes]) glucose production and disposal determined using [3-³H]glucose are shown. *P < 0.05, comparing Leprlox^TB to Leprlox^TB × POMC-cre mice; ^†P < 0.05, comparing WT controls to Leprlox^TB mice; ^#P < 0.05 comparing fed and refed mice; ^‡P < 0.05, comparing clamp to basal with a genotype. Data are presented as mean ± SEM.

Figure 7. Selective reactivation of LEPR expression in POMC neurons improves the lipid profile.

Plasma (A and B) and liver (C and D) TG and cholesterol in 12-week-old, 5-hour–fasted WT, WT expressing Cre-recombinase in POMC neurons, LEPR-null mice generated by inserting floxed transcription blocking sequences in the Lepr gene (Leprlox^TB), and littermates in which LEPRs are selectively reactivated in POMC neurons (Leprlox^TB × POMC-cre). *P < 0.05; ^†P < 0.05 versus WT controls and Leprlox^TB × POMC-cre, respectively. n = 7–12/genotype. Data are presented as mean ± SEM.