Minireview: Central Sirt1 Regulates Energy Balance via the Melanocortin System and Alternate Pathways

Abstract

In developed nations, the prevalence of obesity and its associated comorbidities continue to prevail despite the availability of numerous treatment strategies. Accumulating evidence suggests that multiple inputs from the periphery and within the brain act in concert to maintain energy metabolism at a constant rate. At the central level, the hypothalamus is the primary component of the nervous system that interprets adiposity or nutrient-related inputs; it delivers hormonal and behavioral responses with the ultimate purpose of regulating energy intake and energy consumption. At the molecular level, enzymes called nutrient energy sensors mediate metabolic responses of those tissues involved in energy balance (1). Two key energy/nutrient sensors, mammalian target of rapamycin and AMP-activated kinase, are involved in the control of food intake in the hypothalamus as well as in peripheral tissues (2,3). The third more recently discovered nutrient sensor, Sirtuin1 (Sirt1), a nicotinamide adenine dinucleotide-dependent deacetylase, functions to maintain whole-body energy homeostasis. Several studies have highlighted a role for both peripheral and central Sirt1 in regulating body metabolism, but its central role is still heavily debated. Owing to the opaqueness of central Sirt1's role in energy balance are its cell-specific functions. Because of its robust central expression, targeting cell-specific downstream mediators of Sirt1 signaling may help to combat obesity. However, when placed in the context of a physiologically relevant model, there is compelling evidence that central Sirt1 inhibition in itself is sufficient to promote negative energy balance in both the lean and diet-induced obese state.

Obesity is recognized as the largest and fastest growing public health problem in developed and developing nations (4, 5). It is a condition that is caused by sustained and unresolved imbalances in energy intake and expenditure (6, 7), resulting in augmented visceral fat, elevated basal glucocorticoids, and insulin and leptin resistance among other features of obesity (8–14). Distinct hypothalamic nuclei, such as the arcuate nucleus (ARC), ventral medial hypothalamus (VMH), dorsal medial hypothalamus (DMH), and the paraventricular nuclei (PVN), are involved in the regulation of feeding and energy expenditure (6, 7). Specifically, the ARC is located in the mediobasal hypothalamus, anteriorly juxtaposing the median eminence. As a circumventricular organ, the ARC is sensitive to peripheral cues such as postprandial fluctuations in hormones, amino acid, and glucose. These molecules signal through their respective cellular sensors, thereby instigating appropriate metabolic, physiological, behavioral, and autonomic responses in nondiseased individuals. In addition, the ARC is sensitive to central cues such as changes in neurotransmitters and neuropeptide hormones. Neurons in the ARC that produce anorexigenic peptides including proopiomelanocortin (POMC) and cocaine- and amphetamine-related transcript and orexigenic peptides including neuropeptide Y (NPY) and agouti-related peptide (AgRP) (15) directly sense peripheral cues and innervate second-order neurons localized in distinct extra-ARC hypothalamic sites, such as the PVN (16–19).

Silent mating type information regulation 2 homolog 1 (Sirt1) is a nicotinamide adenine dinucleotide (NAD+)-dependent class III deacetylase that functions to deacetylate its targets, which include transcription factors, histones, and cofactors. Therefore, Sirt1 regulates gene expression at the transcriptional level by influencing chromatin remodeling, or via posttranslational mechanisms that are mediated via its interactions with transregulators and coregulators (20). Moreover, Sirt1 can also modulate protein activity via its removal of acetyl functional groups. For example, Sirt1-mediated deacetylation of the transcription factor, Forkhead box O1 (FoxO1), increases FoxO1's activity (21, 22). FoxO1 is a metabolic sensor that integrates both leptin and insulin signaling (23, 24). It is abundantly expressed in metabolically relevant hypothalamus nuclei, including neurons of the ARC, DMH, and VMH. Among its metabolic functions, FoxO1 transcriptionally regulates AgRP and NPY expression in a positive manner while transcriptionally repressing POMC, carboxypeptidase E (CPE), and steroidogenic factor 1 (SF1) expression (23–25). In addition, FoxO1 positively regulates Sirt1 expression at the transcriptional level (26). Sirt1 activity is augmented by fluctuations in NAD+, changes in the levels of nicotinamide phosphoribosyl transferase, and nicotinamide mononucleotide adenylyltransferase 1) (two enzymes involved in the biosynthesis of NAD+) (27), changes in its phosphorylation status (28), its association with coregulators (29). Lastly, a number of small chemical compounds are known to augment Sirt1 activity such as the Sirtuin1-activating compounds, resveratrol and sirtuin activator 3 (30), and the Sirt1-specific inhibitors, Ex-527 and sirtinol (31).

Sirt1's dependency on NAD+ supports its role as an energy sensor (32, 33), and recent studies have revealed its involvement in nutrient sensing (33, 34). Sirt1 is abundantly expressed in the periphery, ubiquitously expressed in neurons of the central nervous system (33, 35), and prominently expressed in the hypothalamus ARC, VMH, and PVN (Figure 1A). Neuronal Sirt1's subcellular localization is predominantly nuclear (35); however, Sirt1 can translocate to the cytoplasm in a cell-specific and cell-autonomous manner in response to various physiological stimuli and indices of disease (36). Sirt1 functions in a pleiotropic capacity, including cell survival, apoptosis, proliferation, and metabolism (37–39). In the periphery, Sirt1 has been extensively studied with respect to energy balance and is substantially reviewed by Boutant and Canto (40). In brief, peripheral Sirt1 promotes negative energy balance via its actions in metabolic tissues including liver, pancreas, and adipocyte tissue. For example, Sirt1 improves glucose tolerance and positively regulates insulin secretion in pancreatic beta cells by repressing uncoupling protein 2 (UCP2) (41, 42). In the liver, Sirt1 increases oxidative metabolism and alters lipid metabolism (43, 44), and in adipose tissue, it functions to decrease fat storage, promote lipolysis, and protect against obesity-induced inflammation (45–47).

Figure 1.

Sirt1 is expressed in the brain, is nutrient sensitive, and can influence food intake and body weight. Sirt1 is ubiquitously expressed in the brain and is highly expressed in hypothalamus PVN, VMH, and ARC (A). Immunostaining of POMC (green fluorescence) and Sirt1 protein (brown diaminobenzidine staining) (B). Fasting increases Sirt1 in the hypothalamus and specifically within POMC-expressing neurons of the ARC (C). Enhanced Sirt1 expression in ARC of fasted rodents is associated with reduced ARC levels of the anorectic peptide, α-MSH (D). Data are mean ± SEM. *, P < .05. Reproduced with permission from Cakir et al (34).

Central Sirt1 also has an established role in energy balance. However, whether central Sirt1 functions to promote negative or positive energy balance remains unresolved. Numerous studies have provided conflicting evidence that supports Sirt1's role in mediating orexigenic and anorexigenic, as well as both enhanced and reduced thermogenic responses. This review aims to provide transparency on the role of central Sirt1 as it pertains to energy balance with particular emphasis of its functional role in the fasted and diet-induced-obese (DIO) condition.

Brain Sirt1

Brain Sirt1 and calorie deprivation

Initial studies investigating the role of Sirt1 in calorie restriction-induced longevity established Sirt1 as the molecular link mediating that phenomenon. Subsequently, several studies have demonstrated enhanced Sirt1 content in metabolically relevant hypothalamic nuclei in mice subjected to 30%–40% reduced caloric intake. As observed in other models of Sirt1 overexpression (48), mice with brain-specific Sirt1 overexpression display a metabolic phenotype that closely mimics mouse models of calorie restriction. Moreover, mice with brain-specific Sirt1 overexpression displayed elevated brain Sirt1 levels in a distribution pattern comparable with what is observed in animal models of calorie restriction. These metabolically relevant nuclei include the DMH and the lateral hypothalamus (33, 34, 49–51). Collectively these studies demonstrate a role of Sirt1 in mediating the responses to calorie restriction.

Sirt1 is also elevated in the brain (52) and specifically within the hypothalamus of 24-hour, lean-fasted, male rodents (33, 34). More specifically, Sirt1 levels are enhanced within ARC POMC neurons of fasted rats (Figure 1, B and C), and it is associated with reduced α-MSH levels (Figure 1D). This nutritional state suggests that Sirt1 in POMC neurons may function to promote positive energy balance in an attempt to resolve acute indices of nutrient depletion. Along these lines, p53 (an established target of Sirt1) acetylation is also reduced in the hypothalamus of fasted rats, further supporting enhanced hypothalamus Sirt1 activity as a mechanism involved in the resolution of acute nutrient depletion (31). Whether Sirt1 content or activity is altered in AgRP neurons under conditions of either acute or chronic calorie depletion is currently unknown. Studies investigating the cell-specific fluctuations in Sirt1 levels and/or activity under different nutritional states will prove most beneficial because neurons whose functions oppose or antagonize one another (such as AgRP and POMC neurons) are known to be closely situated in the same anatomical locations. Indeed, cell-specific fluctuations in Sirt1 have major physiological impacts as outlined below.

Neuronal Sirt1 depletion

To date, a number of studies have demonstrated the adverse effect of global Sirt1 knockdown on energy metabolism, specifically in animals challenged with excess dietary fat (53, 54). Consistent with these studies, several Sirt1 global overexpression (brain and peripheral organs) studies have highlighted the beneficial effects of Sirt1 in preventing the development of diet-induced insulin resistance and other features of obesity (55, 56). This finding is consistent with the fact that peripheral Sirt1 has a well-established role in promoting negative energy balance, suggesting that the observed effects were most likely driven by peripheral Sirt1 action. This was confirmed via the use of tissue-specific Sirt1 knockdown models, wherein Sirt1 was selectively diminished in distinct peripheral metabolic sites (40, 42, 45, 57, 58). Evidence suggests that similarly to other metabolic sensors (such as AMP-activated protein kinase), Sirt1's activity in peripheral sites compared to its activity in the hypothalamus is inversely regulated by alterations in hormones, nutrients, and various metabolic states (59, 60), indicating that Sirt1's peripheral functions may oppose its central actions.

To discern the explicit role of neuronal Sirt1 in energy balance, Lu et al (61) generated a Sirt1 neuron-specific knockout (SINKO) mouse model, wherein, SINKO mice were characterized by diminished neuronal Sirt1 mRNA and protein content, specifically in neurons of the central nervous system. Results demonstrated that neuronal Sirt1 deficiency protected against metabolic disease in mice fed a high-fat diet (HFD) (61). In particular, SINKO mice were protected against HFD-mediated adipocyte and hepatocyte inflammation as well as brain and peripheral insulin resistance. Overall, the results from Lu's study indicate a role for neuronal Sirt1 in promoting positive energy balance via a subtle yet effective reduction in autonomic responses and impairments in key signaling pathways that maintain energy homeostasis (eg, insulin resistance).

Sirt1 in ARC POMC- and AgRP-expressing neurons

Pertaining to Sirt1's role in AgRP and POMC neurons, numerous studies have highlighted its role in regulating energy balance via its modulation of the melanocortin system. For example, ARC Sirt1 was shown to function in an orexigenic capacity, as sustained siRNA-mediated reduction in ARC Sirt1 levels resulted in long-term reduced food intake and body weight gain in a male rodent model system (Table 1) (34). Correspondingly, acute inhibition of hypothalamic Sirt1 via intracerebroventricular (ICV) infusion of Ex-527 resulted in decreased food intake and body weight gain in lean, male rats (34, 62). This anorexia-induced reduction in body weight was associated with elevated ARC POMC mRNA and protein content (34, 62) and reduced AgRP mRNA (Table 1) (34). Conversely, ICV infusion of sirtuin activator 3 resulted in an immediate (and short lived) increase in food intake in refed rats (34). These effects were mediated through the central melanocortin system as coadministration of Ex-527 and SHU9119, an antagonist of the melanocortin receptors 3/4 (MC3/4Rs), curtailed the observed anorectic effect resulting from central Sirt1 inhibition. Overall, these results point to an orexigenic role of ARC Sirt1 in the lean condition (34).

Table 1.

Table Outlining the Effect of Sirt1 Manipulation on Energy Balance in the Lean and DIO Condition

	Energy Expenditure	Food Intake	Weight Change	Sirt1 Expression/Activity	POMC, AgRP, α-MSH Expression	Processing Enzymes Affected	Thyroid Axis Affected
Lean + fasted (34, 100)	Decreased	Increased	Increased	Increased in ARC POMC neurons	Decreased POMC and α-MSH	Decreased PC1 and PC2 in PVN	Reduced HPT axis
Lean + ICV EX-527 (16, 62)	Increased	Decreased	Decreased	Decreased central Sirt1 activity	increased POMC and α-MSH reduced AgRP	Increased CPE In ARC	Enhanced HPT axis
Lean + intra-ARC Sirt1 siRNA (16)	?	Decreased	Decreased	Decreased in ARC	?	?	?
DIO (62, 101)	Increased	Increased	Increased	Increased in in ARC and PVNa	Decreased α-MSH	?	Enhanced HPT and HPA axis
DIO + ICV EX-527 (62)	Increased	no change	Decreased	Decreased central Sirt1 activity	Increased POMC and α-MSH, no changes in AgRP	Increased ARC CPE, Decreased PC2 in PVNa	Enhanced HPT axis

Abbreviations: HPA, hypothalamus-pituitary-adrenal; HPT, hypothalamic-pituitary-thyroid; siRNA, small interfering RNA. This is a succinct overview of the effect of Sirt1 manipulation (pharmacological inhibition or siRNA mediated knockdown) on food intake, energy expenditure, weight change, melanocortin activity, and the HPT and HPA axes. The metabolic parameters for lean + fasted, lean + ICV EX-527, and DIO groups were compared against their appropriate lean control animals; the metabolic parameters for lean + intra-ARC Sirt1 siRNA was compared against lean + intra-ARC control siRNA; and the metabolic parameter for DIO + ICV EX-527 was compared against DIO + vehicle. Collectively the results indicate that Sirt1 augments the melanocortin system, processing enzymes, and HPT axis to alter energy balance.

^aUnpublished data.

More recently we demonstrated a similar, yet quite distinct role for ARC Sirt1 in the Sprague Dawley model of DIO. DIO rodents also displayed elevated Sirt1 levels in their ARC (62); and similarly to what was observed in lean rats, DIO rats exhibited weight loss due to acute central Sirt1 inhibition (Table 1). Interestingly, whereas reduced Sirt1 activity reduced food consumption in lean animals (34, 62, 63), DIO animals subjected to central Sirt1 inhibition remained normophagic (Table 1) (62). Hypothalamic AgRP is an established mediator of feeding behavior, as either activating AgRP neurons or increasing AgRP peptide levels, is sufficient to enhance food intake (64–66). As expected, and contrary to the observed effect of central Sirt1 manipulation on AgRP gene expression in the lean state, AgRP mRNA was unchanged between the ARC of DIO rats with or without central Sirt1 activation (62); yet similarly to lean animals, POMC mRNA and protein content was significantly enhanced (Table 1). Instead, DIO rats subjected to central Sirt1 inhibition displayed elevated oxygen consumption relative to their vehicle-infused controls, an effect that was not observed in lean animals (62).

These changes in metabolic rate and body weight, which were attributed to reduced hypothalamic Sirt1 activity, were correlated with the following: increased ARC POMC and CPE protein levels, increased α-MSH in nerve terminals innervating the PVN, elevated TRH in PVN, and elevated serum T₃ levels (62). Of note, enhanced POMC and CPE levels were attributed to a Sirt1-mediated reduction in FoxO1 activity (62). Overall, these studies suggest that Sirt1 in the ARC functions to promote positive energy balance in both lean and DIO rats, whereas central Sirt1 inhibition promotes negative energy balance in both lean and DIO rats, albeit via distinct (behavioral and autonomic) mechanisms.

In support of these findings, Dietrich et al (67) demonstrated that either peripheral or central administration of Ex-527 was sufficient to reduce food intake in both the dark cycle and during ghrelin-induced phagia. This effect on food intake was not associated with concomitant changes in energy expenditure; rather, it was mediated through reduced melanocortin tone and was dependent on the action of UCP2 (67), an inner-membrane mitochondrial protein that has an established role in mediating hypothalamic mitochondrial respiration. Indeed, UCP2 was previously shown to be a critical mediator of ghrelin-induced food intake and ARC-POMC hyperpolarization via enhanced AgRP inhibitory inputs (68). Supporting the findings of Dietrich et al (67), Velasquez et al (31) demonstrated that the orexigenic action of ghrelin is mediated via a Sirt1-p53 pathway. In particular, results from the study of Velasquez et al demonstrated that central Sirt1 inhibition reduced both ghrelin-induced phagia and hypothalamic fatty acid metabolism, and abolished the ghrelin-induced enhanced expression of FoxO1, NPY, and AgRP (31). As previously mentioned, one way in which AgRP abrogates melanocortin activity is via its antagonism of the MC3/4Rs. Alternatively, AgRP-expressing neurons reduce melanocortin activity by tonically inhibiting POMC neurons via γ-aminobutyric acid innervations (69–71). Both central Sirt1 inhibition and depletion of Sirt1 in AgRP neurons caused reduced AgRP neuron activity, resulting in diminished inhibitory synapses onto POMC neurons. The consequences of reduced POMC inhibition included increased melanocortin tone, which was associated with increased POMC mitochondrial density and mitochondrial size (72); of note, mitochondrial dynamics have an established role in the central regulation of whole-body metabolism (68, 72).

Collectively, these results suggest that Sirt1 in AgRP neurons functions through the melanocortin system in an orexigenic capacity to promote positive energy balance (67). Moreover, in a recent publication, Sasaki et al (73) showed that Sirt1 overexpression in AgRP neurons reduced daily food intake in both female and male mice. This effect was associated with reduced body weight, which was significantly pronounced in male mice (73). Male mice with Sirt1 overexpression in AgRP neurons also exhibited enhanced leptin sensitivity, which is known to repress the activity of AgRP neurons; therefore, both Sirt1 ablation (in female mice) and Sirt1 overexpression (in male mice) in AgRP neurons reduces the activity of AgRP neurons, resulting in reduced food intake (63, 73).

Contrary to the above findings, absence of Sirt1 in POMC neurons (POMC-Cre;Sirt1^Loxp/Loxp) resulted in an increased propensity to the development of obesity, specifically in female mice fed a hypercaloric diet (HCD). This effect was a result of reduced energy expenditure due to impaired sympathetic nerve activity (SNA) specifically in perigonadal white adipose tissue (WAT). Partially aligned with those observations, Sasaki et al (73) demonstrated that overexpression of Sirt1 in POMC neurons protected against age-associated weight gain as a result of reduced adipose mass due to enhanced energy expenditure, specifically in male mice fed a standard chow (SC) diet (73). In addition, hypothalamic SIRT1 prevented against age-associated weight gain by improving leptin sensitivity. Of note, Sirt1 overexpression in POMC neurons reduced AgRP and NPY mRNA levels in SC fed male mice that were administered leptin at a time-point prior to the development of their lean phenotype, suggesting that these animals have enhanced ARC leptin sensitivity. However, although Sirt1 overexpression in AgRP neurons also enhanced ARC leptin sensitivity, AgRP and NPY expression was unaltered (73). Collectively, these results suggest that Sirt1 in a cell-specific manner influences the magnitude of leptin sensitivity to augment energy balance.

Consistent with the observed effects in the study by Ramadori et al (74), Sirt1 overexpression in POMC neurons also resulted in enhanced sympathetic activity specifically in the WAT of SC-fed male mice. However, whether the aforementioned effects were exclusively caused by Sirt1 depletion in ARC POMC neurons is unknown because neurons in the caudal portion of the nucleus of the tractus solitarius (NTS) also expresses POMC (75) and Sirt1 (33). In particular, the NTS has an established role in regulating energy balance and SNA (76–78); therefore, it is possible that Sirt1 ablation or overexpression in POMC expressing neurons of the NTS participated to some degree in the augmented metabolic rate observed. Similarly, corticotroph cells of the anterior pituitary also express POMC; however, Sirt1 manipulation in POMC cells did not alter corticosterone levels (73, 74), which is likely due to the near-absent expression of Sirt1 in pituitary corticotrophs (79).

Interestingly, the absence of Sirt1 in POMC neurons resulted in leptin resistance via leptin's inability to signal through phosphoinositide 3-kinase. Impaired leptin responsiveness in POMC neurons of mutant mice resulted in reduced brown adipose tissue-like remodeling of perigonadal WAT (74). Along the lines of the above observation, Sirt1 overexpression in POMC neurons enhanced browning of sc WAT in SC-fed male mice, after a period of prolonged cold exposure (73). Overall, Ramadori et al (74) demonstrated that Sirt1 in POMC neurons is critical for the maintenance of normal body weight via autonomic adaptations, independent of changes in melanocortin signaling, in female mice chronically maintained on a HCD. Surprisingly, in the study by Sasaki et al (73), Sirt1 overexpression in POMC neurons did not protect either male nor female mice from the development of DIO. Instead, high-fat/high-sugar (HFHS) diet-fed mice with Sirt1 overexpression in POMC neurons became obese due to a loss of enhanced energy expenditure. The authors of this study concluded that HFHS feeding suppressed the effect of ARC Sirt1 overexpression via its reduction of Sirt1 protein and hypothalamic NAD⁺ (73).

Reduced ARC Sirt1 content in DIO rodents is contrary to our observations, wherein DIO rats displayed enhanced ARC and PVN Sirt1 protein expression (62) (unpublished observation). The inconsistencies among these studies may be due to temporal fluctuations in Sirt1 protein levels or differences in diet composition; whereas we measured Sirt1 protein content after 12 weeks on a HFD, Sasaki et al (73) measured Sirt1 protein content after 4 weeks on a HFHS diet.

Although Sirt1 levels were elevated in the ARC of DIO rats, whether there is a cell-specific increase and/or decrease of Sirt1 protein in POMC- or AgRP-expressing neurons in the DIO condition is currently unknown. From a syllogistic view, it is likely that ICV infusion of Ex-527 reduces Sirt1 activity in both POMC-expressing and AgRP-expressing neurons of the ARC. Whereas central Sirt1 inhibition enhanced autonomic responses and reduced body weight in DIO rats via the augmentation of the melanocortin system, Sirt1 ablation in POMC neurons of chronically HCD fed female mice predisposed them to obesity, with the former likely being a consequence of Sirt1 inhibition in distinct cell types. As such, although Sirt1 in POMC neurons could logically promote negative energy balance via stimulation of autonomic responses (74), this model does not account for the physiological influences of AgRP action on POMC neurons. Indeed, as demonstrated by Dietrich et al, Sirt1 in AgRP neurons functions to augment POMC neuron activity and melanocortin tone (63). Moreover, ablation of Sirt1 in AgRP neurons resulted in behavioral responses that were closely aligned to the behavioral responses observed in lean rodent models of central (34, 67) and peripheral (mouse only) (63) Sirt1 inhibition. In the broader context, evidence suggests that ARC Sirt1 functions to promote positive energy balance, in part via its regulation of the melanocortin system in lean individuals. Whereas hypothalamus Sirt1 inhibition resulted in reduced POMC mRNA, POMC protein, and diminished α-MSH in studies using the rat-physiological model of DIO, Sirt1 ablation in POMC neurons altered neither ARC POMC mRNA nor protein content, nor were the POMC-derived peptides altered. These major differences are likely responsible for the contradictory effects on energy balance observed in the different models and approaches used.

Sirt1 in the ventromedial hypothalamus

The VMH is another key region involved in the regulation of energy homeostasis. In particular, studies have demonstrated a role for VMH SF1-expressing neurons in regulating feeding and energy expenditure, in part via its responsiveness to leptin (80). As it pertains to Sirt1's function in SF1 neurons, Ramadori et al demonstrated that similarly to what was observed in their POMC-specific Sirt1 knockdown model, mice maintained on a normocaloric diet did not display any changes in the parameters of energy balance due to loss of Sirt1 in SF1 neurons (SF1-Cre;Sirt1^Loxp/Loxp) (81). However, when maintained on a HCD, SF1-Cre;Sirt1^Loxp/Loxp mice were predisposed to the development of obesity. Specifically, mice lacking Sirt1 in SF1 neurons developed enhanced adiposity and hyperleptinemia. HCD-fed SF1-Cre;Sirt1^Loxp/Loxp mice were normophagic; however, they became obese due to a decreased basal metabolic rate, which was mediated in part by impaired leptin signaling. In addition, the absence of Sirt1 in SF1 neurons conferred susceptibility to the development of type 2 diabetes mellitus.

Contrasting these observations, Sirt1 overexpression in SF1 neurons enhanced insulin sensitivity in skeletal muscle, facilitating protection against dietary diabetes. This antidiabetic effect was mediated in part via signaling through orexin A (81), a neuropeptide that has been demonstrated to promote negative energy balance (82). Overall, results from this study demonstrate that Sirt1 in SF1 neurons functions in a protective capacity against the development of DIO, hyperinsulinemia, and hyperglycemia via enhanced leptin sensitivity and enhanced insulin action in skeletal muscle (81). Of note, FoxO1 functions as a transcriptional suppressor of SF1, and inhibits the activity of SF1 neurons thereby promoting positive energy balance. Contrary to the effects of Sirt1 ablation in SF1 neurons, Foxo1 knockdown in VMH SF1 neurons conferred a lean phenotype due to enhanced energy expenditure and protected against the development of DIO.

These changes were associated with increased insulin sensitivity in skeletal muscle and heart and enhanced SNA (83). Presumably, ablation of Sirt1 in SF1 neurons could lead to diminished FoxO1 activity, and as a result, FoxO1's transcriptional suppression of SF1 could be lifted, leading to increased energy expenditure. That Sirt1 ablation in SF1 neurons confers the susceptibility to the development of obesity in HCD-challenged mice via a reduction in metabolic rate (81) suggests a FoxO1-independent pathway in mediating Sirt1's effect on energy homeostasis. However, given that the same parameters of energy balance (insulin sensitivity in skeletal muscle, glucose homeostasis, leptin sensitivity, and basal metabolic rate) were affected, albeit in an opposing manner, collectively, these results suggests that distinct Sirt1-mediated pathways in SF1 neurons play a critical role in instigating different energy balance outcomes.

Sirt1 in the paraventricular nucleus

The PVN is comprised of a heterogeneous subpopulation of specialized neurons, which synthesize and secrete hypophysiotropic peptide hormones. Among their neuroendocrine functions, they are the central regulators of the hypothalamus-pituitary axes. ARC POMC and AgRP neurons innervate PVN neurons including those expressing TRH (84) and CRH (17, 18). Both TRH and CRH neurons express MC3/4Rs; consequentially, melanocortin activity directly influences the thyroid and adrenal axes (16, 19). Activation of the adrenal axis results in enhanced circulating glucocorticoids, which have an established role in energy metabolism. For example, sustained chronic increases of basal glucocorticoids are associated with increased food drive and enhanced abdominal adiposity. Indeed, comparison of basal corticosterone (CORT) levels revealed that DIO Sprague Dawley rats have enhanced CORT in serum compared with their lean counterparts (Toorie, A., N. Cyr, and E.A. Nillni, unpublished data); this was associated with increased Sirt1 expression in the PVN of DIO individuals (Table 1, Toorie, A., N. Cyr, and E.A. Nillni, unpublished data).

This correlation prompted us to investigate the role of PVN Sirt1 in the regulation of the adrenal axis. We found that lean rats subjected to ICV infusion of resveratrol displayed elevated CORT in serum compared with vehicle-infused controls (Toorie, A., N. Cyr, and E.A. Nillni, unpublished data), which indirectly suggests that enhanced central Sirt1 activity is sufficient to activate the adrenal axis. Similarly to all neuropeptide hormones, CRH is synthesized as a larger inactive precursor, preproCRH that undergoes an intracellular series of posttranslational modifications to generate its bioactive form, CRH. The enzymes involved in the maturation of CRH include the prohormone convertases (PCs), PC1 and PC2, and the exopeptidase, CPE. Because central Sirt1 inhibition was previously demonstrated to decrease ARC CPE levels (62), we investigated the PVN-specific effects of central Sirt1 inhibition on the processing enzymes in lean and DIO individuals. Acute central Sirt1 inhibition resulted in decreased PC2 protein levels in the PVN of DIO animals, an effect that was not observed in the lean state, indicating that Sirt1 positively regulates PVN PC2 levels in DIO. Whether Sirt1 regulates PC1 and CPE expression in the PVN is under current investigation. In vitro evidence suggests that Sirt1 can regulate PC1 and PC2 at both the transcriptional and protein level. Furthermore, Sirt1's capacity to alter PC1 and PC2 expression is associated with parallel changes in intracellular and secreted CRH content.

Utilizing the immortalized adrenocorticotropic cell line, AtT20, we demonstrated that cells incubated with resveratrol, a widely used activator of Sirt1, displayed increased PC1 and PC2 protein levels compared with vehicle-treated cells. These effects were associated with a concomitant increase in intracellular CRH and released CRH content (Toorie, A. and E.A. Nillni, unpublished data). Contrary to this finding, AtT20 cells treated with Ex-527 displayed decreased PC1 and PC2 protein content, which was associated with reduced intracellular CRH content (Toorie, A., N. Cyr, and E.A. Nillni, unpublished data). Together these results suggest that PVN Sirt1 has the capacity to activate the adrenal axis and increase basal glucocorticoids levels via its positive regulation of PC2 (and possibly PC1)-mediated processing of proCRH into bioactive CRH. Although this is an appealing concept, further studies are warranted to elucidate the role of PVN Sirt1 in the direct activation of the adrenal axis in both lean and DIO states.

Hypothalamus ARC SIRT1 and FOXO1

Several studies have demonstrated that enhanced Sirt1 activity increases FoxO1's activity and retains its nuclear localization (26, 34, 85). Therefore, Sirt1, via its modulation of FoxO1 activity can inhibit POMC, CPE, and SF1 expression, and enhance AgRP expression. Future studies will need to investigate cell-type specific increases in Sirt1 in the DIO condition and its subsequent effect on FoxO1 (as well as other targets, ie, p53) activity. Nonetheless, as previously reported, inhibition of hypothalamus Sirt1 activity resulted in diminished FoxO1 activity and was associated with increased POMC and CPE levels (34, 62). It is therefore logical to infer that alterations in FoxO1 activity through the actions of Sirt1 may also augment the activity of other metabolically relevant neurons such as those expressing SF1 and AgRP.

Indeed, a similarity in phenotype is observed in cell-specific Foxo1 deletion models, phenocopying, to some degree, what is observed in Sirt1 deletion models. For example, Kim et al (24) previously demonstrated that lean and DIO mice subjected to reduced ARC FoxO1 content reduced their food consumption, which led to a decrease in body weight. Consistent with this observation, mice that were intra-ARC infused with adenovirus-expressing constitutively active FoxO1 displayed increased food intake and reduced energy expenditure alongside an associated increase in body weight. FoxO1's effect on energy balance was mediated in part via leptin and insulin stimulation and was dependent on signaling through the phosphoinositide 3-kinase/AKT pathway. In addition, changes in FoxO1 levels and activity were shown to augment the expression of both AgRP and POMC (24). A similar effect was also noted in DIO Wistar rats, wherein inactivation of Foxo1 diminished food intake, adiposity, and body weight and was correlated with enhanced insulin sensitivity (86).

To assess the function of neuronal FoxO1 in energy balance and to distinguish its cell specific roles, Ren et al (87) analyzed the effects of global neuron-specific FoxO1 knockout, and concurrent FoxO1 knockout in both AgRP and POMC neurons, on food intake and energy expenditure. Global neuron-specific FoxO1 depletion did not alter basal energy homeostasis, nor did it alter body composition. However, this state resulted in a diminished response to fasting-induced phagia, which was attributed to enhance leptin sensitivity (87). Of note, central Sirt1 activation caused an enhanced response to rebound feeding (34), which is consistent with the observed effect of neuronal FoxO1 depletion on fasting-induced rebound feeding (87). Therefore, it is plausible that Sirt1 signals through FoxO1 to modulate rebound-feeding responses. Mice lacking neuronal FoxO1 also displayed increased locomotor activity, resulting from enhanced melanocortin signaling. Furthermore, when maintained on a HCD, mice lacking neuronal FoxO1 were protected from body weight gain and enhanced adiposity, which was partly attributed to increased energy expenditure.

Similarly, simultaneous FoxO1 depletion in POMC and AgRP neurons of lean mice resulted in decreased body weight and fat mass and was associated with enhanced lean mass and elevated energy expenditure (87). As aforementioned, a comparable effect on energy balance was observed in rats subjected to central Sirt1 inhibition, and these effects were shown to be mediated through FoxO1 signaling (34). Collectively these results suggest that ARC Sirt1, via its regulation of FoxO1, signals through the melanocortin system to respond to energy imbalances. However, as supported by Ramadori et al (74), evidence suggests that there are Sirt1 pathways that signal, independent of FoxO1 and the melanocortin system to augment energy balance. In the DIO state, ARC insulin and leptin resistance, coupled with elevated Sirt1 expression and activity, likely creates a self-perpetuating cycle of FoxO1 hyperactivity (Figure 2). Sustained FoxO1 activity has been previously reported to promote positive energy balance, which is in part due to an inability to respond appropriately to fluctuations in peripheral indicators of caloric excess (23, 26, 88).

Figure 2.

Model depicting Sirt1's regulation of the central melanocortin system in hypothalamic POMC and AgRP neurons. ac, acetyl; AKT, protein kinase B; GABA-γ, aminobutyric acid; IRS, insulin receptor substrate; JAK, Janus activated kinase; p, phosphate; PIP₂, Phosphatidylinositol-4,5-bisphosphate; PIP₃, phosphatidylinositol-3,4,5-trisphosphate; STAT3, signal transducer and activator of transcription 3. Broken arrows indicate translocation into or out of the nucleus. Curved arrows indicate removal of acetyl groups. Red arrows and red X indicate diminished signaling. Under normal conditions in nonobese individuals, leptin and insulin signal through their respective receptors to ultimately increase AKT phosphorylation. pAKT translocates to the nucleus and phosphorylates the transcription factor FoxO1, thereby facilitating its inactivation and nuclear exclusion. Consequentially, POMC and CPE transcription is enhanced, resulting in increased α-MSH, whereas AgRP transcription is repressed, resulting in reduced AgRP. In both DIO and fasted conditions, insulin and leptin signaling are impaired via distinct mechanisms, resulting in reduced pAKT levels, which promotes FoxO1 nuclear retention. In addition, Sirt1 protein content is increased within the ARC of both fasted and DIO individuals. Sirt1 via its deacetylation of FoxO1 positively regulates AgRP transcription while negatively regulating POMC and CPE transcription. Reduced POMC and CPE expression results in diminished levels of α-MSH, which signals through MC3/4Rs to exert its potent anorectic effect; AgRP functions as an endogenous antagonist of the MC3/4R. Sirt1 also increases the number of γ-aminobutyric acid secretion synapses onto POMC neurons by AgRP neurons, resulting in the hyperpolarization of POMC neurons and reduced melanocortinergic tone.

In POMC neurons, loss of FoxO1 results in decreased food intake with no effect on energy expenditure. FoxO1 depletion in POMC neurons conferred a lean phenotype, and mutant mice displayed diminished rebound feeding. Moreover, in mice maintained on a HCD, loss of FoxO1 in POMC neurons protected against the development of obesity. These effects were associated with diminished AgRP mRNA and elevated α-MSH, which was a result of enhanced CPE expression (25). Of note, central Sirt1 inhibition in DIO rats also resulted in enhanced α-MSH in nerve terminals innervating the PVN, which was associated with increased ARC CPE content. However, contrary to FoxO1 ablation in POMC neurons, central Sirt1 inhibition resulted in a dramatic increase in oxygen consumption but did not result in changes in food intake or AgRP mRNA (62). Collectively, these results suggest that hypothalamus Sirt1 via its action on FoxO1 promotes cell-specific functions that result in divergent adaptive metabolic responses.

Concluding remarks

As it pertains to energy balance, central Sirt1 has both cell-specific and cell-autonomous functions, as evidenced in the aforementioned studies. As is the case with other metabolic sensors, the role of central Sirt1 in DIO is quite complex. The genetic mice models used in prior studies were advantageous in helping to discern the cell-specific functions of Sirt1. However, data from these studies should be interpreted in the broader context of a suitable physiological model, such as the rat physiological model of DIO. For example, in the mouse model developed by Ramadori et al, Sirt1 was deleted from all POMC cells including neurons from the ARC, the NTS, and the pituitary (74). Therefore, it is difficult to predict what impact the lack of Sirt1 in the NTS or pituitary will have with respect to changes in energy metabolism. There are many advantages of the Sprague Dawley model of DIO. Rats and humans share many characteristics of obesity physiology (89), and rats are considered to be an excellent model to study obesity physiology (90–92). Furthermore, the Sprague Dawley rat model has been extensively characterized for energy-regulating peptide analyses including POMC processing (85, 93–96), which was originally described in the rat (16, 43, 97, 98).

Although considerable knowledge has been acquired through the use of Sirt1 cell-specific genetic models, caution should be exercised when extrapolating the data acquired from these models. For example, as it relates to central Sirt1's role in energy balance, cell-specific Sirt1 studies provide contradictory evidence. However, when extrapolated in the context of a physiological model, wherein cell-cell interactions and distant neural influences remain intact, there is convincing evidence that point to central Sirt1 in promoting positive energy balance. In addition, it is apparent that Sirt1's effector targets, such as FoxO1, play a crucial role in determining whether Sirt1 will signal to promote negative or positive energy balance. Furthermore, the sexual dichotomy evidenced in these genetic models point to the involvement of other relevant, yet currently uninvestigated, gender-specific factors/systems in mediating Sirt1's regulation of energy balance.

As described above, due to the cell specific nature of Sirt1's action in promoting either positive or negative energy balance in response to dietary excess, a more targeted therapy approach against cell-specific downstream mediators of Sirt1 signaling may prove beneficial in combating obesity and the physiological outcomes that predispose to the development of metabolic disorders. Indeed, as demonstrated by Ren et al (99), ablation of GRP17, a G-coupled protein receptor that is a specific downstream target of FoxO1 and is prominently expressed in AgRP neurons, results in enhanced anabolic activity and decreased food intake. Targeted therapeutic approaches may help to circumvent the potentially deleterious and counterproductive mechanisms of global Sirt1 activation/inactivation. Collectively, evidence points to brain Sirt1 in mediating the adaptive responses to nutritional stress via its broad regulation of mechanisms involved in energy homeostasis. It is notable, however, that central Sirt1 inhibition or neuronal Sirt1 ablation is sufficient to promote negative energy balance. Therefore, approaches aimed at enhancing peripheral Sirt1 activity, while reducing central Sirt1 activity, may prove most beneficial in the treatment of obesity.