Sex difference in physical activity, energy expenditure and obesity driven by a subpopulation of hypothalamic POMC neurons

Abstract

Objective: Obesity is one of the primary healthcare challenges of the 21st century. Signals relaying information regarding energy needs are integrated within the brain to influence body weight. Central among these integration nodes are the brain pro-opiomelanocortin (POMC) peptides, perturbations of which disrupt energy balance and promote severe obesity. However, POMC neurons are neurochemically diverse and the crucial source of POMC peptides that regulate energy homeostasis and body weight remains to be fully clarified.

Methods: Given that a 5-hydroxytryptamine 2c receptor (5-HT2CR) agonist is a current obesity medication and 5-HT2CR agonist's effects on appetite are primarily mediated via POMC neurons, we hypothesized that a critical source of POMC regulating food intake and body weight is specifically synthesized in cells containing 5-HT2CRs. To exclusively manipulate Pomc synthesis only within 5-HT2CR containing cells, we generated a novel 5-HT 2C R (CRE) mouse line and intercrossed it with Cre recombinase-dependent and hypothalamic specific reactivatable Pomc (NEO) mice to restrict Pomc synthesis to the subset of hypothalamic cells containing 5-HT2CRs. This provided a means to clarify the specific contribution of a defined subgroup of POMC peptides in energy balance and body weight.

Results: Here we transform genetically programed obese and hyperinsulinemic male mice lacking hypothalamic Pomc with increased appetite, reduced physical activity and compromised brown adipose tissue (BAT) into lean, healthy mice via targeted restoration of Pomc function only within 5-HT2CR expressing cells. Remarkably, the same metabolic transformation does not occur in females, who despite corrected feeding behavior and normalized insulin levels remain physically inactive, have lower energy expenditure, compromised BAT and develop obesity.

Conclusions: These data provide support for the functional heterogeneity of hypothalamic POMC neurons, revealing that Pomc expression within 5-HT2CR expressing neurons is sufficient to regulate energy intake and insulin sensitivity in male and female mice. However, an unexpected sex difference in the function of this subset of POMC neurons was identified with regard to energy expenditure. We reveal that a large sex difference in physical activity, energy expenditure and the development of obesity is driven by this subpopulation, which constitutes approximately 40% of all POMC neurons in the hypothalamic arcuate nucleus. This may have broad implications for strategies utilized to combat obesity, which at present largely ignore the sex of the obese individual.

Keywords: 5-HT2c receptor; Brown adipose tissue; Energy expenditure; Hyperinsulinemia; Hypothalamus; Obesity; Pro-opiomelanocortin (Pomc); Sexual dimorphism.

Graphical abstract

2

3

Figure 1

Generation of mice with restored ARC Pomc function in 5-HT_2CR expressing cells. (A,B) ARC POMC neurons (POMC-IR, red) co-express 5-HT_2CRs (GFP-IR, green; co-labeled, yellow). Scale bar, 50 μm (inset) and 100 μm. (C) POMC-IR and 5-HT_2CR (GFP-IR) co-expression by ARC bregma level. (D) Schematic of wild-type allele (Pomc^WT) containing both neuronal Pomc enhancers, nPE1 and nPE2 (green); 5-HT_2CR^CRE inserted after ht2rc exon 7 (red); a disrupted Pomc^NEO knockout allele carrying nPE2 deletion and a loxP-flanked-mediated disruption of nPE1 transcriptional activation function (gray); and a re-activated Pomc/5-HT_2C allele (Pomc^5−HT2CR) (yellow). (E)Pomc re-expression in 5-HT_2CR neurons in male (F_3,19 = 22.40, P < 0.0001) and (F) female mice (F_3,15 = 18.40, P < 0.0001) normalized to 36β4 mRNA, relative to Pomc^WT, in arbitrary units (AU). *P < 0.05, **P < 0.01, ***P < 0.001 as indicated.

Figure 2

Subpopulation of Pomc differentially modulates physical activity and energy expenditure in male and female mice. 24 h food intake was significantly elevated in (A) male (F_3,12 = 13.98, P < 0.001) and (B) female (F_3,11 = 6.55, P < 0.001) ARC Pomc null (Pomc^NEO) mice, and this was normalized by restoration of Pomc in 5-HT_2CR neurons in Pomc^5−HT2CR mice. 24 h locomotor activity was significantly reduced in (C) male (F_3,17 = 3.68, P < 0.05) and (D) female (F_3,18 = 5.36, P < 0.05) Pomc^NEO mice, and this was normalized in male, but not female, Pomc^5-HT2CR mice. General linear model in (E) male and (F) female mice illustrating reduced total daily energy expenditure in Pomc^NEO and Pomc^5-HT2CR mice (solid line) compared to control siblings (dashed line) in females (group effect P < 0.01) but not in males (group effect P > 0.05). (G) General linear model revealing Pomc^NEO and Pomc^5-HT2CR females (solid line) display reduced resting metabolic rate compared to control siblings (dashed line) (group effect P < 0.01). (H,I) Lipid accumulation in interscapular brown adipose tissue (BAT; F_3,17 = 11.61, P < 0.001) and relative expression of mitochondrial genes important for BAT thermogenesis, (K)Pgc-1α (F_3,18 = 8.638, P < 0.001) and Elovl3 (F_3,18 = 4.147, P < 0.05), were fully normalized by restoration of Pomc in 5-HT_2CR neurons (Pomc^5-HT2CR) in male mice. By contrast, Pomc^5-HT2CR female mice display an increase in (H,J) BAT lipid content (F_3,10 = 40.00, P < 0.0001) and a downregulation in (L)Pgc-1α(F_3,12 = 8.941, P < 0.01) and Elovl3 (F_3,18 = 8.941, P < 0.01) compared to littermate controls. Data expressed as arbitrary units and expression of target genes corrected to the geometric average of four housekeeping genes:18s, 36β4, Gapdh and βactin. Scale bar, 2000 μm *P < 0.05, **P < 0.01, ***P < 0.001 compared to all other genotypes, except in D, J and L compared to Pomc^WT and 5-HT_2CR^Cre mice.

Figure 3

Subpopulation of Pomc differentially modulates body weight and adiposity in male and female mice. (A,B) Body weight (male, F_3,20 = 114.35, P < 0.001; female F_3,19 = 60.80, P < 0.001), (C, F, G) fat mass (male F_3,20 = 9.43, P < 0.001; female F_3,22 = 42.04, P < 0.001), (H–J) gonadal white adipocyte diameter (male F_3,14 = 13.98, P < 0.001; female F_3,13 = 11.87, P < 0.001) were increased in ARC Pomc null (Pomc^NEO) mice and normalized by restoration of Pomc in 5-HT_2CR neurons (Pomc^5-HT2CR) in male but not female mice. However, female Pomc^NEO mice still exhibited significantly greater body weight and fat mass, but not adipocyte size, compared with Pomc^5-HT2CR mice. (D, E) Both male and female Pomc^NEO mice exhibited significantly greater lean mass compared to Pomc^5-HT2CR mice and controls. Scale bar, 2000 μm *P < 0.05, **P < 0.01, ***P < 0.001 compared to control Pomc^WT and 5-HT_2CR^CRE siblings, except where noted in D, E, G and I.

Figure 4

Subpopulation of Pomc modulates insulin sensitivity in male and female mice. (A,B) Plasma insulin (male, F_3,21 = 63.20, P < 0.0001; female F_3,23 = 20.33, P < 0.0001) was increased in ARC Pomc null (Pomc^NEO) mice and normalized by restoration of Pomc in 5-HT_2CR neurons (Pomc^5-HT2CR). (C,D) Fasting blood glucose was not statistically different among genotypes (male, F_3,15 = 0.70, P > 0.05; female F_3,18 = 2.30, P > 0.05). Both male (E,F) and female (G,H)Pomc^NEO mice exhibited impaired responses in an insulin tolerance test compared with control littermates, which were normalized by restoration of Pomc in 5-HT_2CR neurons (Pomc^5-HT2CR); (E,G) Insulin tolerance tests (male, F_3,104 = 33.09, P < 0.0001; female F_3,102 = 24.39, P < 0.0001) and (F,H) their respective areas under the curve (AUC) (male, F_3,18 = 18.06, P < 0.0001; female F_3,17 = 6.56, P < 0.01). *P < 0.05, **P < 0.01, ***P < 0.001 compared to all other genotypes.