Identification of adropin as a secreted factor linking dietary macronutrient intake with energy homeostasis and lipid metabolism

Abstract

Obesity and nutrient homeostasis are linked by mechanisms that are not fully elucidated. Here we describe a secreted protein, adropin, encoded by a gene, Energy Homeostasis Associated (Enho), expressed in liver and brain. Liver Enho expression is regulated by nutrition: lean C57BL/6J mice fed high-fat diet (HFD) exhibited a rapid increase, while fasting reduced expression compared to controls. However, liver Enho expression declines with diet-induced obesity (DIO) associated with 3 months of HFD or with genetically induced obesity, suggesting an association with metabolic disorders in the obese state. In DIO mice, transgenic overexpression or systemic adropin treatment attenuated hepatosteatosis and insulin resistance independently of effects on adiposity or food intake. Adropin regulated expression of hepatic lipogenic genes and adipose tissue peroxisome proliferator-activated receptor gamma, a major regulator of lipogenesis. Adropin may therefore be a factor governing glucose and lipid homeostasis, which protects against hepatosteatosis and hyperinsulinemia associated with obesity.

Figure 1. Enho mRNA Encodes a Highly Conserved Secreted Peptide

(A) Enho mRNA sequence with the conserved ORF underlined. (B) Conservation of adropin sequence in mammals. (C) Presence of FLAG-IR in cell lysate (L) and conditioned media (M) of HEK293 cells transfected with vectors expressing adropin with or without a C-terminal FLAG. (D) Western blot of FLAG-IR in sera collected from B6 mice 1-21 days after injection with 10⁸ pfu of adenovirus expressing FLAG-tagged adropin. (E) Northern blots showing distribution of Enho mRNA in human (upper panel) and mouse tissues (lower panel). Tissues were obtained from lean B6 (L) or obese leptin-deficient (O) mice. (F) A series of autoradiographs summarizing Enho mRNA expression sites across the B6 mouse brain: 7N, facial nucleus; AP, area postrema; Cu, cuneate nucleus; DVC, dorsal vagal complex; HDB, nucleus of the horizontal limb of the diagonal band; LGN, lateral geniculate nucleus; MHb, medial habenula; MnR, median raphe; MS, medial septum; MVe, medial vestibular nucleus; PAG, periaqueductal gray; PMD, dorsal premamillary nucleus; Pn, pontine nuclei; Po, posterior thalamic nuclear group; Pr, prepositus nucleus; PVP, paraventricular thalamic nucleus; RN, red nucleus; RtTg, reticulotegmental nucleus of the pons; SNC, substantia nigra pars compacta; Sp5, spinal trigeminal nucleus; TgN, tegmental nucleus; VDB, nucleus of the vertical limb of the diagonal band; VP, ventral posterolateral thalamic nucleus.

Figure 2. Regulation of Liver Enho by Obesity and Nutrition

(A) Liver Enho expression in 6-month-old male Mc3r^-/-, Mc4r^-/-, and Lep^ob/Lep^ob mice compared to lean controls. The Mc3r^-/- and Mc4r^-/- mice used for this analysis were backcrossed >10 generations onto the B6 background. * p < 0.05 compared to WT, n = 5-6 per group. (B) Liver Enho mRNA is also reduced in A^y/a mice compared to lean B6 controls, which develop obesity due to hyperphagia associated with expression of an Mc4r antagonist. The reduced expression of Enho in liver is not observed in A^y/a mice subject to calorie restriction (CR), suggesting that the decline is secondary to obesity and not a specific defect due to reduced melanocortin receptor signaling (n = 4 per group). * p < 0.05 versus obese A^y/a mice and lean controls. (C) Feeding lean male B6 mice HFD (60% kJ/fat) for 2 days is associated with increase liver Enho expression compared to mice fed LFD (10% kJ/fat) * p < 0.01. (D) Differential effects of purified LFD or HFD on the expression of liver Enho mRNA in male B6 mice after 2, 7, 14, or 28 days. * p < 0.05 compared to LFD within time point (n = 4 per group). (E) Fasting is associated with reduced liver Enho expression compared to controls that had been fed a purified LFD (10% kJ/fat) for 10 days. The rebound in expression is affected by dietary fat content, with mice refed LFD (open bars) exhibiting a delayed response compared to mice refed with HFD (black bars) (n = 5-6 per group). * p < 0.05, ns, no significant difference. Note that the mice used for the studies shown in (C) and (D) were 3 months old, and had been weaned onto standard rodent chow. (F) Reduced Enho mRNA in HepG2 cells after 48 hr treatment with LXR agonists (GW3965 [GW]; TO9 [TO]; both 1 μM for 24 hr). (G) An siRNA targeting LXRα increases Enho mRNA in HepG2 cells treated with GW-3965. (H) Liver Enho mRNA in 12-week-old B6 mice is significantly reduced 4 hr after intravenous injection with 10 mg/kg GW3965 (n = 4, *p < 0.05). (I) The FXR agonist GW-4064 (1 μM for 24 hr) increased expression of Shp1 but did not effect Enho expression in HepG2 cells. Values shown in each panel are mean ± SEM.

Figure 3. Adr-Tg Mice Exhibit Attenuated Metabolic Distress Associated with Chronic Exposure to HFD

(A) Schematic showing the transgene construct with the adropin ORF ligated into a vector containing the human β-actin promoter. (B) Representative RT-PCR analysis using transgene-specific primers; expression was observed in liver, heart, and whole brain.

Figure 4. Analysis of Whole Body Energy Metabolism in Chow-Fed Adr-Tg and WT Mice

(A) Total 24 hr energy expenditure (TEE) of male and female Adr-Tg mice compared to controls. (B and C) TEE expressed per g body weight (B) or FFM (C) (*p < 0.05). (D) RQ of male and female Adr-Tg and WT mice (*p < 0.05). (E and F) Spontaneous locomotory activity of female Adr-Tg and WT mice (*p < 0.05). In (E), the data are presented as mean X beam breaks over 24 hr, during the lights-on period (0600-1800 hr; L or white bar) and during the dark period (1800-0600 hr; D or black bar), by gender. All data are mean ± SEM and are the mean from 3 days of recordings; n = 7-8 each genotype and sex. (C) Quantitative analysis of Enho mRNA in liver, muscle, and retroperitoneal WAT of WT and Adr-Tg mice (n = 3 per group, * p < 0.05). (D) Western blot analysis of adropin IR in control HEK293 cells (no DNA), cells infected with adenovirus vector expressing adropin (Ad treated), and liver and muscle tissue lysates from Adr-Tg (Tg) or WT mice. (E) Body weight of WT and Adr-Tg mice fed HFD for 6 weeks (N = 3-8 for each genotype and sex). * p < 0.05 versus WT (within gender). (F) Reduced FM and normal FFM for male mice shown in (E) after 6 weeks on HFD. *p < 0.05 for FM (n = 3 for WT, 8 for Adr-Tg). (G) Blood chemistries for Adr-Tg mice fed HFD for 6 weeks. Transgenics exhibited reduced insulin, normal blood glucose, increased serum adiponectin (in females), and reduced serum TG compared to WT mice. (H-J) Improvements in glucose homeostasis in Adr-Tg mice are independent of reduced weight gain. Body weight and composition data from male Adr-Tg and WT mice aged 6-7 months and maintained on chow or HFD are shown (60% kJ/fat) (H). Adr-Tg mice exhibited protection from hyperinsulinemia and hyperglycemia associated with obesity; reduced HOMA-IR values also suggested improved insulin sensitivity. Serum leptin levels were affected by diet (p < 0.01), correlating with increased adiposity, but not by genotype (I). Glucose clearance was also significantly improved in Adr-Tg mice relative to controls, irrespective of diet (J). All data are mean ± SEM.

Figure 5. Attenuated Hepatic Steatosis and Altered Expression of Genes Involved in Lipid Metabolism in Adr-Tg Mice

(A) Hematoxylin- and eosin-stained liver sections showing less severe macrovesicular fatty change in livers of Adr-Tg mice compared to WT. (B) Liver TG content is lower in Adr-Tg mice (n = 3-8 each genotype and sex, *p < 0.05 versus WT). (C) Expression of stearoyl-CoA desaturase-1 (Scd1), fatty acid synthase (Fas), and lipoprotein lipase (lpl) mRNA is reduced in liver of male and female Adr-Tg mice compared to WT (n = 3-8 each genotype and sex, *p < 0.05 versus WT). Liver data shown in (A)-(C) was from mice fed HFD for 3 months. (D) Expression of genes involved in fatty acid metabolism in WAT from male and female Adr-Tg and WT mice. * p < 0.05. Values shown are mean ± SEM from mice fed HFD for 6 weeks.

Figure 6. Adropin^(34-76) Therapy Affects Energy Metabolism in DIO B6 Mice and Cultured Adipocytes

(A-J) Administration of adropin^(34-76) by i.p. injection at 0900 and 1800 hr for 14 days is associated with reduced food intake (A and B) and weight loss (C). Adropin treatment reduced fasting insulin (D) without affecting blood glucose (E), and increased serum adiponectin (F), suggesting improved insulin sensitivity. Adropin treatment improves hepatic steatosis (G) and significantly reduces liver TG accumulation (H), associated with reduced expression of genes involved in lipid synthesis in liver (I). Pparg expression in WAT of mice treated with adropin^(34-76) was not significantly different from controls treated with saline (J). The values shown are mean ± SEM with N = 12/treatment group.

Figure 7. Short-Term Treatment with Synthetic Adropin^(34-76) Improves Glucose Homeostasis in DIO B6 Mice Independently of Weight Loss or Reduced Food Intake

(A-D) Treatment of male DIO mice with 90 nmol/kg/day adropin^(34-76) administered as two i.p. injections/day over 2 days did not significantly affect food intake (A) or weight gain (B) (n = 8/group). Treatment followed a 5 day lead-in period, with the injection period marked by the gray area. Following treatment, there were significant improvements in glucose homeostasis, indicated by reductions in fasting insulin and HOMA-IR values (C) and improved glucose tolerance (D). * p < 0.05 compared to saline. However, there were no changes observed in serum TG (C). (E-J) A second study assessed a dose-response to 2 days of injections on glucose homeostasis (F-H) and liver metabolism (I, J) (n = 6/group). There was no significant effect of treatment on body weight (E, the white and solid bars are body weight pre- and posttreatment). There was evidence for improved glucose homeostasis (F-H), with significant reductions in fasting glucose (F) and HOMA-IR (H). While liver TG content was not significantly affected by treatment (I), there were changes in the expression of genes involved in fatty acid metabolism, which was statistically significant for Scd1 and diacylglycerol O-transferase 2 (Dgat2) (J). * p < 0.05 versus saline. The values shown are mean ± SEM.