WITHDRAWN BY AUTHOR: Central Sirt1 Regulates Body Weight and Energy Expenditure Along With the POMC-Derived Peptide α-MSH and the Processing Enzyme CPE Production in Diet-Induced Obese Male Rats

Abstract

In the periphery, the nutrient-sensing enzyme Sirtuin 1 (silent mating type information regulation 2 homolog 1 [Sirt1]) reduces body weight in diet-induced obese (DIO) rodents. However, the role of Sirt1 in the brain, particularly the hypothalamus, in body weight and energy balance regulation is debated. Among the first studies to reveal that central Sirt1 regulates body weight came from experiments in our laboratory using Sprague Dawley rats. In that study, central inhibition of Sirt1 decreased body weight and food intake as a result of a Forkhead box protein O1 (FoxO1)-mediated increase in the anorexigenic proopiomelanocortin (POMC) and decrease in the orexigenic Agouti-related peptide in the hypothalamic arcuate nucleus. Here, we demonstrate that central inhibition of Sirt1 in DIO decreased body weight and increased energy expenditure at higher levels as compared with the lean counterpart. Brain Sirt1 inhibition in DIO increased acetylated FoxO1, which, in turn, increased phosphorylated FoxO1 via improved insulin/pAKT signaling. Elevated acetylated FoxO1 and phosphorylated FoxO1 increased POMC along with the α-MSH maturation enzyme carboxypeptidase E, which resulted in more of the bioactive POMC product α-MSH released into the paraventricular nucleus. Increased in α-MSH led to augmented TRH levels and circulating T₃ levels (thyroid hormone). These results indicate that inhibiting hypothalamic Sirt1 in DIO enhances the activity of the hypothalamic-pituitary-thyroid axis, which stimulates energy expenditure. Because we show that blocking central Sirt1 causes physiological changes that promote a negative energy balance in an obese individual, our results support brain Sirt1 as a significant target for weight loss therapeutics.

Obesity is a major health concern that has reached epidemic proportions in developed countries (1). In the United States, obesity is associated with an estimated 300 000 deaths per year (2). Despite efforts, the development of safe and effective antiobesity drugs has been largely unsuccessful. Therefore, understanding the physiological mechanisms controlling energy balance and body weight is timely.

Sirt1 (silent mating type information regulation 2 homolog 1) is a member of the Sirtuins family of proteins that are NAD⁺-dependent deacetylases, and its enzymatic activity is regulated by NAD⁺, nicotinamide phosphoribosyltransferase, nicotinamide mononucleotide adenylyltransferase l, and posttranslationally modified addition and/or removal of functional groups (ie, phosphorylation). Sirt1, among its different roles, is defined as an energy and nutrient sensor regulating body weight and metabolism (3). Several studies demonstrate that Sirt1 regulates body metabolism in peripheral organs including the liver (eg, References 4 and 5), adipose tissue (6), and pancreas (7). Moreover, recent evidence from our laboratory (8) and others (9–12) reveal a central role for Sirt1 control of body metabolism. Because changes in brain Sirt1 can reflect changes in the animal's nutrient status (eg, fasting) and respond by altering energy balance (8, 10), recent reviews by Schug and Li (13)as well as Coppari (14) advocate central Sirt1 as a target for the treatment of obesity and its associated comorbidities. Furthermore, each review calls for more research on how to better understand how central Sirt1 regulates metabolism particularly in the hypothalamus, which is considered a control center for body weight and energy expenditure (13, 14).

The primary hypothalamic appetite and energy expenditure regulators are the anorexigenic pro-opiomelanocortin (POMC) and the orexigenic Agouti-related peptide (AgRP) produced in distinct neurons of the arcuate nucleus (ARC) (15). POMC and AgRP exert their actions by binding melanocortin 3/4 receptors (MC3/4R) in second-order target neurons where the POMC-derived bioactive peptide α-MSH is a MC3/4R agonist and AgRP is an MC3/4R reverse agonist (16). Sirt1 regulates Forkhead box protein O1 (FoxO1) (17). Deacetylated FoxO1 blocks POMC transcription and enhances AgRP transcription (18), suggesting that hypothalamic Sirt1 may regulate POMC and AgRP via FoxO1. Indeed, we previously demonstrated that small interfering RNA silencing of hypothalamic Sirt1 or inhibition of Sirt1 activity with Ex-527 increased POMC and decreased AgRP caused by elevated acetylated (ac) FoxO1 and resulted in decreased food intake and body weight (8). FoxO1 activity is controlled by deacetylation-dependent retention in the nucleus and phosphorylation-dependent nuclear exclusion. An association of acetylated state FoxO1 with enhanced sensitivity of FoxO1 to Akt-mediated phosphorylation and nuclear exclusion has been demonstrated (19). Another study revealed that pAkt and phosphorylated FoxO1 (pFoxO1) levels were greater in the hypothalamus of high-fat diet (HFD)-fed mice lacking neuronal Sirt1 (12). Therefore, Sirt1's actions on POMC via FoxO1 may be an important mechanism regulating body weight, not only in the lean but also in the obese condition.

The present study investigated the role of Sirt1 in POMC regulation as well as downstream changes in body weight and energy expenditure in the Sprague Dawley rat model of diet-induced obesity (DIO). In a previous study (8), we demonstrated that Sirt1 regulated POMC protein and mRNA in lean rats but did not explore whether Sirt1 altered the production of the POMC-derived anorexigenic α-MSH. To generate mature α-MSH, POMC must undergo a series of proteolytic cleavages initially catalyzed by the enzymes prohormone convertase 1 (PC1) (20) and PC2. PC1 cleaves POMC to generate ACTH after which PC2 cleaves ACTH to form ACTH _(1–17) and corticotropin-like intermediate peptide (CLIP). Carboxypeptidase E (CPE) removes the C-terminal basic residues from ACTH _(1–17), peptidyl-glycine alpha-amidating monooxygenase enzyme acts to generate desacetyl α-MSH, followed by N-acetyltransferase conversion of desacetyl α-MSH to acetyl α-MSH (21–23). POMC processing is critical in body weight control; for example, a patient with defective POMC processing was severely obese (24). Obesity has also been associated with PC1 and CPE mutations (Cpe^fat/Cpe^fat mouse) (25–27). In the current study we showed that Sirt1 regulates the POMC-derived peptide α-MSH production through a posttranslational processing mechanism and its consequent effect on the TRH neuron located in the paraventricular nucleus (PVN) of the hypothalamus, essential in the regulation of the hypothalamic-pituitary-thyroid (HPT) axis. In addition, we showed that inhibition of hypothalamic Sirt1 in DIO causes a decreased body weight and profound augmentation in oxygen consumption as compared with lean controls.

Materials and Methods

Animals and diets

Male Sprague Dawley rats (22 days old; Harlan Laboratories) and fed standard chow (Purina Lab Chow; catalog no. 5001) or HFD (Rodent Chow; catalog no. D12492; Research Diets) for 12 weeks. Regular diet provided 3.3 kcal/g of energy (59.8% carbohydrate, 28.0% protein, and 12.1% fat). HFD provided 5.24 kcal/g of energy (20.0% carbohydrate, 20.0% protein, and 60.0% fat). DIO rats were individuals fed HFD for 12 weeks whose body weight exceeded the mean plus 3 SDs of the lean group fed standard chow. Rats on HFD for 12 weeks whose body weight did not exceed the mean plus 3 SDs of the lean group fed standard chow were considered DIO resistant (28) and were excluded from our studies. We previously characterized the rat DIO model (29). Food and water were available ad libitum unless otherwise indicated. Body weights were measured weekly. The Institutional Animal Care and Use Committee of Rhode Island Hospital/Brown University approved all experimental protocols and euthanasia procedures.

Surgeries and infusions

Animals were weighed and anesthetized with ip injections of 50 mg/kg Ketamine and 0.25 mg/kg Dormitor and prepared for stereotaxic implantation of an intracerebroventricular (icv) 21-gauge stainless steel guide cannula (Plastic One), according to coordinates obtained from Paxinos and Watson atlas into the lateral ventricle: (anteroposterior, −0.8 mm; lateral, −1.2 mm; and ventral, −3.6 mm). Guide cannulas were assessed for correct placement by monitoring water intake upon administration of angiotensin II (40 ng/rat; Sigma) and verified by India ink test. All infusions were performed on free-moving animals using a 30-gauge needle where the injection tip was connected by polyethylene tubing to a 25 μL Hamilton syringe. Animals were given 7 days to recover from surgery for all experiments.

Ex-527 (Sigma-Aldrich), which has been shown to very specifically inhibit Sirt1 enzymatic activity and to elicit changes in POMC and body weight when given icv in rats (8), was used for Sirt1 inhibition experiments. For these experiments, lean and DIO rats were icv infused with either vehicle control (2 μL dimethylsulfoxide + 3 μL artificial cerebrospinal fluid) or Ex-527 (5 μg/rat in vehicle). Injections were given at 3:00 pm on day 1 and again the following morning at 9:00 am. Food was removed at 3:00 pm on day 1 because food intake can affect Sirt1 levels, and fasting increases Sirt1 in the rat ARC (8). Animals were euthanized 3 hours following the second injection (at 12:00 pm on day 2). For the food intake experiment only, food was not removed. In another experiment, we infused the Sirt1 activator resveratrol (5 μg/rat in 6 μL dimethylsulfoxide vehicle) icv in lean and DIO fed rats. We then collected ARC samples for mRNA analysis of AgRP.

Sample collection and preparation

PVN (Bregma, −1.3 to −2.3 mm) and ARC (Bregma, −2.5 to −3.5 mm) samples were microdissected and frozen immediately in liquid nitrogen and placed in −80°C (30). Blood samples were taken to analyze plasma T₃ and T₄ levels. Liver and brown adipose tissue was also collected. Samples were subjected to the following: peptide extraction with 2 N acetic acid supplemented with a protease inhibitor cocktail to measure peptides by specific RIA, protein extraction with radioimmune precipitation assay buffer (50 mM Tris-HCl, pH 7.4, 150 mM NaCl, 0.5% sodium deoxycholate, 0.1% SDS, 1% Nonidet P-40) supplemented with protease inhibitor cocktail for Western blot analysis, RNA isolation using TRIzol reagent protocol (Molecular Research Center) for quantitative real-time PCR (qrtPCR). All analyses (RIA, Western blot, and qrtPCR) were performed on samples taken from individual animals. Samples were not pooled.

Oxygen consumption

Oxygen consumption was measured using a customized Oxymax system (Columbus Instruments). Rats were placed in individual calorimetry chambers 2 hours before sampling; the first 2 hours of sampling data were excluded from data analysis; therefore rats were acclimated to chamber for 4 hours prior to data collection. Samples were taken every 6 minutes thereafter. For Ex-527 experiments, rats were placed in chambers immediately following the first infusion at 3:00 pm when food was removed. Rats were taken out of the chamber the next morning at 9:00 am for their next infusion. Animals were allowed free access to water at all times. Oxygen consumption was compared between lean rats with and without Sirt1 inhibition and between DIO rats with and without Sirt1 inhibition separately. Thus, within each diet group, surface area and body composition did not differ, which is known to affect energy expenditure (31). Oxygen consumption data were normalized to body weight for each individual.

In vitro studies

The mouse corticotropic tumoral AtT-20 cell line was used for in vitro studies; these cells abundantly express POMC, and when transfected with PC2 cDNA, α-MSH release can be measured (32). AtT-20 cells were cultured in DMEM (Invitrogen-Life Technologies) supplemented with 10% of fetal bovine serum, 100 U/mL penicillin, and 10 μg/mL streptomycin. To measure α-MSH release, cells were transiently transfected with a cytomegalovirus plasmid containing PC2 cDNA. For Sirt1 knockdown, we transfected AtT-20s with control or Sirt1 small hairpin RNA (shRNA) kindly provided by Dr Pere Puigserver (Dana-Farber Cancer Institute, Department of Cell Biology, Harvard Medical School, Boston, MA) and described in Ref. 33. To enhance Sirt1 levels, we transfected AtT-20s with Sirt1 cDNA (Addgene). cDNA and shRNA transfections were performed using lipofectamine 2000 (Invitrogen) following the manufacturer's instructions. Briefly, cells were incubated with cDNA (4 μg)-lipofectamine complex for 6 hours in serum-free media. The medium was changed to DMEM medium for overnight incubation. For knockdown studies, we used 1 μg shRNA and antibiotic/serum-free DMEM. We also conducted a time-response study in which 72 and 96 hours posttransfection with Sirt1 shRNA yielded a significant reduction in Sirt1 protein levels. Cells were collected in radioimmune precipitation assay buffer supplemented with protease inhibitor cocktail for Western blot analysis, and media were collected to measure α-MSH by RIA.

RIA

RIA analysis of α-MSH, TRH, insulin T₃, and T₄ are routinely done and developed in our laboratory (34–38). Briefly, the α-MSH RIA was performed using an anti-α-MSH antiserum and 5000 cpm of [¹²⁵I]desacetyl α-MSH tracer. The α-MSH antibody used in this assay can detect 100% of desacetyl α-MSH, but does not cross-react with the free acid nonamidated form of α-MSH, ACTH, or corticotropin-like intermediate peptide (CLIP) (39). The sensitivity of the assays was approximately 10.0 pg/tube, and the intra- and interassay variability was approximately 5%–7% and 10%–11%, respectively. The tracer was iodinated using the Chloramine T method followed by HPLC purification. The sensitivity was 2.0 pg/tube, and the intra- and interassay variability was 5%–6% and 9%–12%, respectively. Serum thyroid and insulin hormones were measured using commercial RIA kits from MP Biomedicals Diagnostic Division (as described in Reference 29) according to manufacturer's instructions. The sensitivity of the T₃, T₄, and insulin assays were 25 ng/dL, 1.2 mg/dL, and 0.081 ng/mL, and the intra- and interassay variability was approximately 5%–7%, 10%–11%, and 3%–9%, respectively (see Supplemental Table 1).

Deiodinase 2 activity assay

Deiodinase 2 activity was performed according to the method of Perello et al (29). Briefly, extracted brown adipose tissue (BAT) was homogenized in P100E2D1. Total protein concentrations were determined using a Bradford assay. D2 activity in BAT was assayed using P100E2D25 and [3′5′-¹²⁵I]T₄ (3000 cpm) with 1 mM 6-propyl-2-thiouracil (Sigma p3755; to block D1 activity) in a total reaction volume of 0.5 mL. Samples were incubated at 37°C for 30 minutes, and reactions were stopped with 0.2 mL ice-cold methanol. Samples were then centrifuged at 3200 rpm for 30 minutes, and 0.5 mL of supernatant was removed and counted for radioactivity. All samples were run in the same assay on the same day.

Glucose assay

Serum glucose levels were measured using the QuantiChrom glucose assay kit (BioAssay Systems), according to manufacturer's instructions.

Western blotting

Protein samples (30 μg) were separated on a 8% acrylamide gel and transferred onto a polyvinylidene difluoride membrane. Precision Plus Protein standards were used as molecular weight markers (Bio-Rad Laboratories). Membranes were washed in PBS with 0.1% Tween 20 and then blocked with 5% BSA for 60 minutes. Membranes were then probed with a primary antibody overnight at 4°C. Refer to the antibody table (Supplemental Table 1) for a complete list of antibodies and their specificity (37, 40–42). The following day the membranes were washed in PBS with 0.1% Tween 20 (3 times for 5 minutes) and incubated in the appropriate secondary antibody for 60 minutes. Membranes were incubated with enhanced chemiluminescence buffer for 1 minute and visualized using the Alpha Innotech imaging system (ProteinSimple). Band density was analyzed using ImageJ (National Institutes of Health).

qrtPCR

Total RNA was extracted from ARC and PVN samples. cDNA was prepared from 1 μg of total RNA using random hexamer primers and SuperScript III reverse transcriptase (Invitrogen). qrtPCRs included 100 ng of cDNA as template, 200 nM of our target gene or tubulin/hprt primers, and Power SYBR Green PCR Master Mix (Applied Biosystems-Life Technologies). An ABI Prism 7500 FAST sequence detector (Applied Biosystems) was used to amplify reactions and generate standard curves. Values were normalized to tubulin/hprt. The ΔΔCT method was used for relative quantification and statistical analysis. Primer sequences used for qrtPCR are available upon request.

Statistical analysis

Results are presented as mean ± SEM. A two-tailed t test was used to analyze differences between 2 groups. An ANOVA, followed by Tukey's HSD post hoc test, was used to analyze differences between more than 2 groups. A Shapiro-Wilk W goodness-of-fit test was used to test for normal distributions, and Levene's test (2-group comparisons) or Bartlett's test (comparisons between more than 2 groups) was used to test for homogenous variances in all variables. For variables with nonnormal distributions and unequal variances, a nonparametric Mann Whitney U-test or Kruskal-Wallis test was used. Differences were considered to be significant at P < .05. Prism (version 4.0b GraphPad Software, Inc). Statistics for each comparison are presented in the figure legends.

Results

In vitro studies support Sirt1 regulation of POMC, CPE, and FoxO1

Previously, we demonstrated that Sirt1 activation, in vitro, caused reduced POMC mRNA and protein levels, whereas small interfering RNA-mediated Sirt1 knockdown caused increased POMC protein levels in AtT-20 cells. In the present study, we investigated a mechanism by which Sirt1 alters POMC levels and also evaluated whether, and by what mechanism, Sirt1 manipulation alters levels of α-MSH. We first conducted a time response experiment using 1 μg of Sirt1 shRNA, which is a dose shown previously to successfully knock down Sirt1 (5). Sirt1 shRNA significantly reduced Sirt1 protein at both 72 and 96 hours posttransfection (ANOVA, F = 16.35; P < .0001; Figure 1A) and significantly increased POMC (t_{(1, 10)}, 3.8; P < .004; Figure 1B) and CPE (t_(1,10), 3.027; P < .013; Figure 1C) protein levels. In another set of experiments, AtT-20 cells transfected with PC2 cDNA show that α-MSH release from the cells into the media was elevated with Sirt1 knockdown (t_(1,22), 2.54; P < .019; Figure 1D). This was correlated with altered Akt and FoxO1 phosphorylation, as phosphatidylinositol 3-kinase (PI3K) (t_(1,13), 2.54; P < .005; Figure 1E), pAKT (t_(1,10), 2.26; P < .05; Figure 1F), and pFoxO1 (t_(1,10), 2.76; P < .021; Figure 1G) each increased significantly with Sirt1 shRNA. We next showed that overexpressing AtT-20 cells with Sirt1 cDNA (Sirt1 transfection, 2.534 ± 0.5359; control transfection, 0.4872 ± 0.02772; P < .0001; Figure 2A) decreases POMC (t_(1,20), 5.58; P < .0001; Figure 2B) and CPE (Mann-Whitney U, 3.0; P < .0001; Figure 2C). Taken together, in vitro results suggest that Sirt1 acts to decrease pAKT, pFoxO1, POMC, CPE, and α-MSH.

Figure 1.

A, Western blot of Sirt1 protein levels in ARC of lean (black bar) and DIO rats (gray bars) (n = 6). For B–D, lean or DIO rats were fasted and treated icv with Ex-527 (5 μg; gray bar) or vehicle control (black bar) (see Materials and Methods). For E and F, lean or DIO rats were fed and treated icv with resveratrol (5 μg; gray bar) or vehicle control (black bar). B, Weight gain in grams in vehicle control and Ex-527-treated rats (n = 12). C, Western blot of ARC POMC protein levels in vehicle (n = 6) and Ex-527-treated lean rats (n = 5). D, Western blot of ARC POMC protein levels in vehicle (n = 6) and Ex-527-treated DIO rats (n = 5). E, AgRP mRNA in the ARC of vehicle (n = 3) and resveratrol-treated lean rats. ARC was measured by RT-PCR (n = 4). F, AgRP mRNA in the ARC of vehicle (n = 3) and resveratrol-treated DIO rats. ARC was measured by RT-PCR (n = 4). Data are mean ± SEM. *, P < .05 **, P < .01 vs lean control (A) or vehicle control (B–F).

Figure 2.

A, The POMC processing cascade is modeled here and has been described previously. Western blot of Sirt1 protein levels in lean and DIO rats (n = 6). For B–G, lean or DIO rats were fasted and treated icv with Ex-527 (5 μg; gray bar) or vehicle control (black bar) (see Materials and methods). PC1 protein levels in the ARC lean (B) and DIO (C) treated with and without Ex-527. PC2 protein levels in the ARC lean (D) and DIO (E) treated with and without Ex-527. CPE protein levels in the ARC lean (F) and DIO (G) treated with and without Ex-527 (n = 6–9). Data are mean ± SEM. *, P < .05 vs vehicle control.

Sirt1 regulates body weight and POMC in DIO rats

Our earlier studies demonstrated that Sirt1 increased body weight in lean rats in part by decreasing ARC POMC and increasing AgRP (8). We currently show that Sirt1 protein levels are elevated in the ARC of DIO rats compared with lean controls (t_(1,10), 4.09; P < .003; Figure 3A). Acute central Sirt1 inhibition resulted in significant weight loss in our DIO animals compared with their vehicle-infused controls (t_{(1, 10)}, 3.72; P < .002; Figure 3B). Sirt1 inhibition also increased POMC levels in the ARC of lean (t_{(1, 10)}, 4.30; P < .002; Figure 3C) and DIO (Mann-Whitney U, 2.0; P < .018; Figure 3D) rats. We previously showed elevated AgRP mRNA levels with Ex-527 in lean animals, but here we did not detect any changes in ARC AgRP mRNA levels of DIO rats treated with Ex-527 or vehicle control (t_{(1, 6)}, 0.978; P > .35; data not shown). We explored whether a central infusion of resveratrol, a commonly used potent activator of Sirt1 that was recently demonstrated to bind to and allosterically activate Sirt1 protein (43–45), would alter ARC AgRP levels. Resveratrol significantly increased AgRP mRNA in the ARC of lean rats (t_(1,5), 3.22; P < .024; Figure 3E), but did not alter AgRP in the ARC of DIO rats (t_(1,5), 68, P > .5; Figure 3F). We, therefore, focused further studies on understanding how Sirt1 regulates POMC in the DIO condition.

Figure 3.

A, DIO rats were fasted and treated icv with Ex-527 (5 μg; gray bar) or vehicle control (black bar) (see Materials and Methods). Western blot analyses of DIO ARC samples are as follows: A, acFoxO1 protein in vehicle (n = 6) and Ex-527-treated rats (n = 5); B, pFoxO1 protein in vehicle (n = 6) and Ex-527-treated rats (n = 6); C, pAKT protein in vehicle (n = 7) and Ex-527-treated rats (n = 7); D, pStat3 protein in vehicle and Ex-527-treated rats (n = 4). Data are mean ± SEM. *, P < .05 vs vehicle control.

Sirt1 regulates α-MSH maturation in DIO rats

POMC undergoes a series of endoproteolytic cleavages to produce smaller bioactive peptides; and PC1, PC2, and CPE are the primary enzymes that catalyze the conversion of POMC to the anorectic α-MSH (Figure 4A; Refs. 21–23). Whether Sirt1 regulates the POMC-processing enzymes is unknown. We therefore investigated the effect of icv Ex-527 infusion on these enzymes in the ARC of both lean and DIO rats. Although inhibition of Sirt1 did not alter PC1 or PC2 (all P > .1; Figure 4, B-E); CPE protein levels were significantly higher in the ARC of Ex-527-treated rats compared with controls in both the lean (t_(1,10), 3.56; P < .006; Figure 4F) and DIO conditions (Mann-Whitney U, 6.0; P < .05; Figure 4G). Collectively, results indicate that ARC Sirt1 regulates POMC protein and posttranslational processing in both lean and DIO rats.

Figure 4.

Food intake was measured in lean (A) or DIO (B) animals treated with or without Ex-527. C–F, Overnight oxygen consumption in fasted lean and DIO rats infused with Ex-527 or vehicle control (n = 6). C and D indicate the average overnight oxygen consumption, and E and F depict the oxygen consumption over each hour measured (not including the initial period to adjust to the chambers) (n = 6–9). Data are mean ± SEM. *, P < .05 vs vehicle control.

Sirt1 regulates acetylated FoxO1 (acFoxO1) and pFoxO1 in the DIO ARC

Sirt1 is known to deacetylate FoxO1 (17), and deacetylated FoxO1 blocks both POMC and CPE transcription (18). Consistent with our previous findings in the ARC of lean rats (8), we found that Sirt1 inhibition elevated acFoxO1 in the DIO ARC (t_(1,9), 2.69; P < .03; Figure 5A). FoxO1 activity is also mediated by phosphorylation-dependent nuclear exclusion, and both pAkt and pFoxO1 levels are elevated in the hypothalamus of HFD-fed mice lacking neuronal Sirt1 (12). Therefore, we investigated whether Sirt1 inhibition altered pAkt and pFoxO1 levels in the DIO ARC. Both pFoxO1 (t_(1,10), 2.83; P < .02; Figure 5B) and pAkt (t_(1,12), 2.18; P < .05; Figure 5C) levels were higher in the ARC of Ex-527-treated DIO rats compared with controls. Phosphorylation of another transcription factor, signal transducer and activator of transcription (Stat)3, positively regulates POMC in the ARC. However, ARC pStat3 levels of DIO rats treated with Ex-527 or control were similar (t_(1,6), 0.045; P > .9; Figure 5D). These results reveal that Sirt1 regulates acFoxO1 and pFoxO1, but not pStat3 in the ARC of DIO rats.

Figure 5.

Lean or DIO rats were fasted and treated icv with Ex-527 (5 μg; gray bar) or vehicle control (black bar) (see Materials and Methods). RIA analyses are as follows: A, PVN α-MSH peptide in vehicle (n = 6) and Ex-527-treated lean rats (n = 5); B, PVN α-MSH peptide in vehicle (n = 4) and Ex-527-treated DIO rats (n = 5); C, PVN TRH peptide in vehicle (n = 6) and Ex-527-treated lean rats (n = 5); D, PVN TRH peptide in vehicle (n = 4) and Ex-527-treated DIO rats (n = 5); E, plasma T₃ levels in vehicle and Ex-527-treated lean rats (n = 7); F, plasma T₃ levels in vehicle (n = 6) and Ex-527-treated DIO rats (n = 7); G, plasma T₄ levels in vehicle and Ex-527-treated lean rats (n = 7); H, plasma T₄ levels in vehicle (n = 6) and Ex-527-treated DIO rats (n = 7). Finally the ratio of T₃:T₄ in lean (I) and DIO (J) rats with and without Ex-527 (n = 6–7). K, PRCP in lean and DIO treated with Ex-527. Data are mean ± SEM. *, P < .05; **, P < .01 vs vehicle control.

Sirt1 inhibition increases oxygen consumption and HPT axis activity in DIO more than in lean rats

POMC controls energy balance through changes in food intake and energy expenditure. Thus, we measured both overnight (ie, during the active period for the rat) food intake and oxygen consumption in lean and DIO rats icv infused with Ex-527 or vehicle control. Lean rats treated with Ex-527 ate significantly less than their controls (t_{(1, 10)}, 3.347; P < .007; Figure 6A) as previously reported (8); yet, overnight food intake in DIO rats treated with Ex-527 did not differ from that of controls (t_(1,5), 0.408; P > .69; Figure 6B). Instead, Ex-527-treated DIO rats showed a significant increase in overnight oxygen consumption compared with controls (Mann-Whitney U, P < .02; Figure 6D), an effect not observed in lean rats (t_(1,10), 0.611; P > .55; Figure 6, C and E). The increase in oxygen consumption in DIO animals treated with Ex-527 was sustained throughout the overnight period (Figure 6F), and the magnitude of the response varied among individuals (Levine's test F_{(3, 5)}, 293.6; P < .0001). We next examined BAT for uncoupling protein (UCP)1 levels, an enzyme known to facilitate energy expenditure, but we were unable to demonstrate changes in UCP1 mRNA and UCP1 protein levels in BAT of rats with central Sirt1 inhibition compared with their vehicle infused controls (Supplemental Figure 2). However, we cannot exclude changes in uncoupling proteins activity and levels in alternate metabolic tissues (such as liver and skeletal muscle), as mediators of increased energy expenditure in DIO rats with central Sirt1 inhibition.

Figure 6.

Sirt1 was knocked down in AtT-20 cells using shRNA targeting Sirt1 and compared with cells treated with control shRNA. A, Time/response for Sirt1 shRNA (1 μg) knockdown of Sirt1 protein. B, Western blot analyses of AtT-20 cells with the Sirt1 or control shRNA for the following: B, POMC; C. CPE; D, α-MSH; F, PI3K; and G, pFoxO1 (n = 6–9). Data are mean ± SEM. *, P < .05; **, P < .01; ***, P < .001 vs shRNA control.

α-MSH increases energy expenditure primarily by activating TRH in the PVN, which regulates the HPT axis and ultimately controls circulating thyroid hormone levels (46–49). Because we observed a significant increase in energy expenditure with hypothalamic Sirt1 inhibition in the DIO but not the lean condition, we investigated the effect of hypothalamic Sirt1 inhibition on TRH and thyroid hormone levels in lean and DIO rats. We measured the amount of α-MSH in the ARC, where it is produced, as well as the amount of α-MSH that reaches the PVN in axon terminals. We did not detect differences in α-MSH (all P > .05; data not shown) or the α-MSH clearance enzyme, prolylcarboxypeptidase (PRCP), in either the lean or DIO (all P > .05; Figure 6K) ARC of rats treated with or without Ex-527. However, Sirt1 inhibition increased α-MSH in the PVN of DIO rats compared with controls (t_(1,7), 3.578; P < .037; Figure 7B), which was not observed in the PVN of lean rats (t_{(1 9)}, 0.319; P > .75; Figure 7A). Similarly, we observed an increase in both PVN TRH levels (t_{(1, 9)}, 3.33; P < 0.013; Figure 7D) and circulating T₃ levels (t_{(1, 11)}, 2.57; P < .027; Figure 7F) in DIO rats treated with Ex-527, but observed no changes in TRH or T₃ levels (all P > .5; Figure 7, C and E) with Sirt1 inhibition in lean rats. Although circulating T₄ levels of Ex527-treated rats were similar to controls in both the lean and DIO states (all P > .5; Figure 7, G and H), the ratio of T₃:T₄ was greater in Ex527-treated DIO rats (t_(1,11), 2.291; P < .05; Figure 7J), and no change was detected in the ratio of T₃:T₄ lean rats (t_(1,11), 0.281; P > .7; Figure 7D). We examined BAT for changes in deiodinase 2, one of the deiodinases that facilitates the conversion of T₄ to T₃, but found no change in either deiodinase 2 activity or level in DIO rats with central Sirt1 inhibition compared with their vehicle-infused controls (Supplemental Figure 2). However, we cannot exclude changes in deiodinase activity or levels in other metabolic tissues (ie, liver and thyroid) as mediators of elevated T₃ in DIO rats with central Sirt1 inhibition. Overall, data indicate that hypothalamic Sirt1 inhibition in the DIO rat increased POMC in the ARC, leading to elevated α-MSH and TRH in the PVN causing increased T₃ and increased energy expenditure. However, increased POMC in the lean condition appears to alter food intake but not the amount of α-MSH or TRH in the PVN nor circulating T₃ and increased energy expenditure (Figure 8).

Figure 7.

Sirt1 was enhanced in AtT-20 cells by transiently transfecting Sirt1 cDNA and compared with cells treated with control cDNA. A, Overexpression of Sirt1 increased Sirt1 protein. B, Western blot analyses of AtT-20 cells with the Sirt1 or control transfection for the following: B, POMC; and C, CPE (n = 11–12). Data are mean ± SEM. ****, P < .0001 vs control.

Figure 8.

Model predicting proposed action of Sirt1 in the ARC. In the ARC, Sirt1 deacetylates FoxO1, which decreases production of POMC and the POMC-processing enzyme CPE, causing less α-MSH to reach MC3/4 receptors on target tissues such as the PVN. Lower α-MSH in the PVN reduces TRH, T₃, and energy expenditure. Arrow from pStat3 is meant to point to POMC transcription, not CPE. IRS, insulin receptor substrate; PIP₂, phosphatidylinositol 4,5-bisphosphate; PIP₃, phosphatidylinositol (3,4,5)-trisphosphate.

Discussion

Our major findings are that central Sirt1 inhibition decreased body weight and increased ARC POMC and CPE in the ARC of DIO rats, with correlative increases in α-MSH and TRH in the PVN of DIO rats as well as circulating T₃ levels and oxygen consumption. The effects of Sirt1 inhibition on α-MSH, TRH, T₃, and oxygen consumption were not detected in lean rats. These results support the hypothesis that central Sirt1 promotes positive energy balance, as shown earlier (8), and blocking Sirt1's activity can promote a negative energy balance and weight loss.

We previously demonstrated that brain Sirt1 inhibition reduced body weight and food intake along with increased POMC and decreased AgRP in lean ARC (8). Subsequent studies investigating the affect of Sirt1 on energy balance have reported conflicting results. Supporting our findings, Dietrich et al (11) showed that both central infusion of Ex-527 and Sirt1 knockout in AgRP neurons decreased body weight by reducing AgRP-induced food intake. This was mediated, in part, by a reduced tonic inhibition on POMC neurons by AgRP neurons (11). This study and our initial study demonstrated that the orexigenic effects of Sirt1 were mediated by the central melanocortin system. Further supporting the hypothesis that brain Sirt1 is orexigenic, Lu et al (12) found that knocking out neuronal Sirt1 partially protected mice from weight gain on a HFD. In contrast, Ramadori et al (50) found that mice with Sirt1 knocked out specifically in POMC or SF1 (51) neurons were more sensitive to DIO due to decreased energy expenditure. The change in energy expenditure in mice lacking Sirt1 in POMC neurons was due to a reduction in sympathetic nerve activity and a change to brown fat-like characteristic in perigonadal white adipose tissues such as a decrease in UCP1. The contradictory effects on energy balance reported by Satoh et al (10) and Ghosh et al (43) in their POMC Sirt1-specific knockdown model does not account for the cumulative influence of central Sirt1 inhibition on hypothalamic control of energy balance. In addition, Sasaki et al (9) reported that adenovirus-associated overexpression of Sirt1 in the mediobasal hypothalamus can suppress food intake but only when hyperphagia was induced by adenovirus-associated overexpression of FoxO1. The inconsistency among these studies may be partially due to different approaches. For example, Sasaki et al (9) used adenovirus-associated Sirt1 infusion coadministered with FoxO1 overexpression, whereas other studies used either Ex-527 to reduce Sirt1 activity or cell type-specific Sirt1 knockout mouse models. Interestingly, central infusion of Ex-527 in either mice or rats resulted in reduced body weight. ICV infusion of Ex-527 is expected to inhibit Sirt1 in both POMC and AgRP neurons; therefore, it is possible that tonic inhibition of POMC neuron activity by Sirt1-mediated changes in AgRP function may play a critical role in HPT regulation in DIO.

In the present study, we determined the role of Sirt1 in ARC POMC regulation using the Sprague Dawley rat model of DIO. Rats and humans share many characteristics of obesity physiology (24), and rats are considered to be an excellent model in which to study obesity physiology (17, 25, 26). This laboratory has extensively characterized the Sprague Dawley rat model for energy-regulating peptide analyses including POMC processing (16, 18, 19, 27–29), which was originally described in the rat (30–33). Moreover, we previously demonstrated changes in POMC processing in the rat ARC under nutritional changes (37). Also, the rat provides a larger source of material for peptide analysis compared with other species such as mice. Lastly, Sirt1 knockout studies in mice show opposite results in terms of body weight regulation (11, 50), potentially indicating that long-term loss of Sirt1 causes compensatory responses (13). Therefore, the physiological rat model of DIO might prove useful in understanding how Sirt1 regulates POMC in individuals made obese by eating an HFD. Our results reveal significantly elevated Sirt1 protein levels in the ARC of DIO rats compared with their lean counterparts. This is consistent with a prior study in mice in which Sirt1 was reported to be elevated in the hypothalamus of db/db mice, which are obese due to a mutation of the leptin receptor (9). Although the same study did not detect any change in hypothalamic Sirt1 levels in mice fed a standard or high-fat-high-sugar diet (9). The disparity in this study and ours may be due to slight differences in the HFDs, or importantly because Sasaki et al (9) measured Sirt1 protein in the entire hypothalamus whereas we measured Sirt1 protein specifically in the hypothalamic ARC. We also show that inhibition of Sirt1 in DIO rats caused weight loss and increased ARC POMC levels, which parallels results in lean rats. Interestingly, we did not detect changes in AgRP mRNA levels with Sirt1 manipulation in DIO rats, suggesting that POMC is the principal melanocortin signal promoting positive energy balance via hypothalamic Sirt1 in the DIO rat. However, we cannot definitively exclude the possibility that AgRP peptide or activity is unchanged because of Sirt1 inhibition, because enumerable studies have demonstrated fluctuations in protein or peptide levels in the face of unchanged mRNA. For example, Sirt1 protein increased in the brain of fasted rodents without accompanying changes in Sirt1 mRNA (52, 53).

Unlike in the lean condition, central Ex-527-induced weight loss in the DIO rat was not caused by changes in food intake. POMC increased in both the lean and DIO ARC. However, Ex-527 inhibition of Sirt1 decreased AgRP only in the lean condition. Similarly, enhanced activity of central Sirt1 with resveratrol augmented AgRP mRNA levels in the ARC of lean animals, but not DIO animals. A recent study supports the importance of AgRP in feeding behavior by demonstrating that activation of a subset of AgRP neurons using cell type-specific photostimulation elicited a strong feeding response within minutes (54). Moreover, Dietrich et al (11) demonstrated that Sirt1 inhibition with centrally or peripherally infused Ex-527 decreased food intake during the dark cycle along with ghrelin-induced food intake in mice. Sirt1 knockout in hypothalamic AgRP neurons, in mice fed regular chow, also decreased electric responses of AgRP neurons to ghrelin, resulting in reduced food intake. Taken together, these results suggest that central Sirt1 control of food intake in lean rats is due, at least in part, to changes in ARC AgRP activity. The current study demonstrated that manipulation of central Sirt1 did not alter AgRP levels in DIO, which could explain the discrepancy in Sirt1's affects on food intake in lean and DIO rats. Among the main factors contributing to the body's maintenance of obesity is that obese individuals are resistant to the anorectic effects of peripherally produced hormones such as leptin and insulin in specific tissues controlling energy balance including the ARC in which leptin and insulin normally act to regulate the production of neuropeptides such as AgRP and α-MSH (18). As a result, the levels of these energy-regulating peptides are altered during DIO. For example, levels of the orexigenic peptides AgRP and Neuropeptide Y are higher, whereas levels of the anorexigenic α-MSH are lower in the DIO ARC compared with lean controls (55–58). Consequently, these peptide changes caused by obesity may override the effects of Sirt1 inhibition on food intake in the DIO condition. Our future studies will explore this possibility.

In our study and in the study of Dietrich et al (11), lean rodents subjected to central Sirt1 inhibition lost weight because of a decrease in food intake, without any changes in energy expenditure. Here, we found that icv infusion of Ex-527 significantly decreased body weight in DIO rats not due to decreased food intake, but instead due to increased oxygen consumption. POMC-derived α-MSH is known to regulate energy expenditure and is a key regulator of TRH that is produced in the hypothalamic PVN (36, 48). TRH is essential for energy balance because it regulates the HPT axis, which stimulates energy expenditure. Sirt1 inhibition did not alter ARC α-MSH levels, which may be due to different clearance or release rates in animals treated with or without Ex-527. To begin to explore this possibility, we investigated whether Sirt1 inhibition altered levels of the α-MSH clearance enzyme PRCP. We found that the protein level expression of this enzyme did not change by inhibiting Sirt1. However, we do not know whether PRCP levels or activity changes at α-MSH target sites such as the PVN with central Sirt1 manipulation. Central Sirt1 inhibition does appear to alter α-MSH release differently in lean and DIO conditions, at least to the PVN, because RIA analysis showed elevated levels of α-MSH released/available to the PVN in DIO animals treated with Ex-527, but no changes were found in lean animals treated with and without Ex-527. That central Sirt1 inhibition augmented the amount of α-MSH released to the PVN in DIO, but not lean, rats may explain why TRH levels were elevated in the PVN of DIO rats along with circulating levels of the active thyroid hormone T₃, a generally accepted indicator of energy expenditure (59), yet no changes in TRH or T₃ were detected in lean rats treated with or without Ex-527. Overall, the results suggest that regulation of POMC by Sirt1 may differentially affect energy balance in the lean and DIO conditions wherein central inhibition of Sirt1 reduces food intake in the lean condition but increases HPT axis activity and energy expenditure in the DIO condition. A previous study showed that Sirt1 positively regulates TSH exocytosis from pituitary thyrotropes (60). Together, results suggest that Sirt1 regulates the HPT axis at various levels although that regulation may depend on nutritional status. α-MSH can affect energy expenditure through several mechanisms (61), and our future studies will explore other possible mechanisms underlying the increase in energy expenditure with central Sirt1 inhibition.

POMC processing is known to be a critical mechanism to generate the anorexigenic α-MSH (21, 23). Here, we show, for the first time, that the α-MSH maturation enzyme CPE is regulated by Sirt1 in both the lean and DIO ARC. POMC and CPE produced in the ARC have each been shown to be important in obesity regulation. For example, Zhan et al (62) showed that ablating POMC neurons specifically in the ARC using pharmacogenetic techniques caused obesity and metabolic disorders. Likewise, mutations in CPE are known to cause obesity, and enhancement of CPE levels in the ARC has been shown to cause anorexigenic effects (27, 63). Evidence suggests that the mechanism linking Sirt1's actions on POMC and CPE and subsequently body weight is through changes in the transcription factor FoxO1. For example, Sirt1 is known to deacetylate FoxO1 (17), which in turn blocks both POMC and CPE transcription (18). We demonstrated previously that the Ex-527-induced increase in ARC POMC was dependent on FoxO1 (8). As predicted, acFoxO1 also increased in the ARC of DIO rats infused with the Sirt1 inhibitor Ex-527. In addition, Sirt1 CPE is reduced in the ARC of DIO mice, and deletion of FoxO1 in POMC neurons (POMC-FoxO1^−/−) rescued that decrease in CPE and protected against weight gain. Although we did not observe a significant change in CPE levels in the ARC of our DIO rats in the present or previous studies (56), we did find that CPE in the ARC of our DIO rats is sensitive to changes in Sirt1.

Leptin and insulin are important hormones that elicit a negative energy balance in the lean condition, in part, by regulating ARC FoxO1 and thus POMC (18). For example, leptin and insulin each increase POMC transcription by activating pAKT that in turn increases pFoxO1, which promotes nuclear exclusion and inactivation of FoxO1 (Figure 1) (18, 64). However, POMC mRNA levels are similar in lean and DIO rodents despite hyperinsulinemia and hyperleptinemia, indicating insulin and leptin resistance in the ARC (55, 56). Using HEK-293 cells, Qiang et al (19) showed that mutating FoxO1 to mimic acFoxO1 increased FoxO1's sensitivity to AKT-mediated phosphorylation and nuclear exclusion. Therefore, it is possible that the increase in ARC acFoxO1 induced by inhibiting Sirt1 altered pAKT and pFoxO1 levels. In fact, we found a significant increase in both pAKT and pFoxO1 in the ARC of our Ex-527 icv-infused DIO rats compared with vehicle controls, signifying that central Sirt1 inhibition can sensitize pAKT signaling in the ARC of DIO rats, which leads to elevated POMC and CPE levels in the DIO hyperinsulinemic/hyperleptinemic state.

Our in vitro studies support our in vivo findings. Enhanced expression of Sirt1 significantly decreased POMC and CPE whereas Sirt1 shRNA knockdown significantly increased POMC and CPE. In vivo and in vitro data reveal that both pharmacologic and genetic manipulation of Sirt1 leads to similar effects on POMC and CPE. In addition, Sirt1 shRNA knockdown increased PI3K, pAKT, and pFoxO1 levels in AtT-20 cells, supporting the idea that blocking Sirt1 enhances pAKT/pFoxO1 signaling. As mentioned above, the increase in ARC Akt phosphorylation during Sirt1 inhibition could be caused by increased acFoxO1. However, the increase in PI3K with Sirt1 shRNA suggests that Sirt1 could mediate insulin signaling, because insulin activates the PI3K/pAkt pathway (Figure 8). This idea is further supported by a recent study in which Sirt1 neuron-specific knockout mice exhibited greater insulin sensitivity compared with controls (12). Neuronal Sirt1 deficiency enhanced hypothalamic pAKT and protected mice from insulin resistance caused by feeding a HFD. These Sirt1 neuron-specific knockout mice also gained less weight than controls when fed a HFD. In the present study, we examined serum for circulating glucose and insulin levels but found no changes in their levels between rats with central Sirt1 inhibition and their vehicle-infused controls (Supplemental Figure 3). Enhanced Akt phosphorylation during Sirt1 inhibition could also be caused by changes in leptin sensitivity because leptin can elevate PI3K (Figure 8). Although we did not detect any changes in Stat3 with Sirt1 manipulation, it is possible that leptin can alter Akt without any changes in Stat3 because other factors such as the tyrosine phosphatase TCPTP are known to regulate Stat3 in the hypothalamus (65). Future work will determine the mechanism(s) by which Sirt1 regulates Akt phosphorylation. Taken together, results suggest that central Sirt1 acts to increase body weight, in part, by FoxO1-mediated regulation of POMC and CPE (see Figure 8).

In summary, we showed previously that central inhibition of Sirt1 evokes weight loss in lean rats via behavioral modifications (8); and the current study reveals that central inhibition of Sirt1 can also induce weight loss in the Sprague Dawley rat model of DIO via autonomic changes. Our results demonstrated that central Sirt1 inhibition in DIO rats augmented the levels of phosphorylated Akt and phosphorylated FoxO1 in the ARC, leading to elevated levels of the anorectic peptides POMC, TRH, and T₃, which resulted in increased energy expenditure and weight loss. Given these results, we propose a model describing mechanisms by which central Sirt1 regulates POMC in a FoxO1-dependant manner to control body weight in the DIO condition (Figure 8). Although the role of brain Sirt1 in body weight and energy balance regulation has proven complicated, the current study adds to a growing body of evidence that blocking central Sirt1 can improve metabolic disorders and body weight in the obese condition.