2.2 CNS Circuits Integrating Circadian and Metabolic Processes

Circadian and metabolic processes interact at both the neuroanatomic and neuroendocrine levels to regulate overall metabolic homeostasis. In order to better appreciate how the circadian clock network within brain regulates whole body metabolism, it is important to understand the anatomic connections between brain centers essential for circadian rhythmicity and those that control appetite and energy expenditure (Fig. 2) (reviewed in Horvath 2005; Horvath and Gao 2005; Saper et al. 2005; Slat et al. 2013; Kalsbeek and Fliers 2013). Classical lesioning studies performed in the 1970s first determined that the “master pacemaker” neurons are localized to the suprachiasmatic nuclei (SCN), which consist of bilateral nuclei within the anterior hypothalamus that receive environmental light input via the retinohypothalamic tract (RHT). SCN ablation abolishes 24-h rhythms of locomotor activity, feeding, drinking, and sleep. Subsequent studies in the 1990s revealed that transplantation of a short-period (tau) mutant SCN into an SCN-lesioned wild-type hamster results in the wild-type hamster having a period length identical to that of the tau mutant donor SCN (Ralph et al. 1990). However, 24-h rhythms of glucocorticoid and melatonin oscillations were not restored. These experiments established that direct neuronal projections, in addition to secreted factors, are necessary for SCN regulation of behavioral and metabolic homeostasis (Cheng et al. 2002; Kramer et al. 2001; Meyer-Bernstein et al. 1999; Silver et al. 1996). Elegant tracing studies have revealed that neurons from the SCN predominantly synapse directly on cell bodies within the ventral and dorsal subparaventricular zones (vSPZ and dSPZ, respectively) and the dorsomedial hypothalamus (DMH). vSPZ neurons are involved in regulation of sleep/wake and activity cycles, but not body temperature cycles, while neurons within the dSPZ control temperature rhythms and have minimal impact on sleep and activity cycles (Lu et al. 2001). The SPZ also projects to the DMH, a brain region important for rhythms of the sleep/wake cycle, locomotor activity, body temperature, food intake, and corticosteroid secretion (Chou et al. 2003). Projections from neurons within the DMH target neuronal centers involved in the regulation of sleep/wake cycle (ventrolateral preoptic nucleus—VLPO), corticosteroid release (paraventricular nucleus—PVN), and feeding and wakefulness (lateral hypothalamus—LHA). As such, the DMH acts as a relay center, amplifying circadian signals from the SCN to multiple regions of the brain involved in regulation of sleep, activity, and feeding.

Open in a separate window

Fig. 2

Neural circuits linking hypothalamic regions important in circadian and energetic control. Signals from the exogenous environment (i.e., light) and endogenous metabolism (i.e., hormones and metabolites) are integrated in a network of hypothalamic neuronal centers (see text for details), which in turn impart rhythmicity on behavioral and metabolic outputs, including sleep, feeding and activity behavior, thermogenesis, and hormone secretion [Figure modified from Huang et al. (2011)]

The DMH and LHA also receive input from the arcuate nucleus (ARC), which plays a well-characterized role in the regulation of feeding and appetite. The ARC contains orexigenic neurons that express neuropeptide Y (NPY) and the agouti-related protein (AgRP), as well as neurons expressing anorexigenic peptides including pro-opiomelanocortin (POMC) and cocaine- and amphetamine-regulated transcript (CART). Because the choroid plexus allows passage through the blood–brain barrier, the ARC is uniquely positioned to integrate humoral signals from the periphery with neuronal signals within the hypothalamus. For example, leptin, a hormone secreted in proportion to fat mass by adipose tissue, stimulates production of α-melanocyte-stimulating hormone (α-MSH) from POMC/CART neurons while simultaneously inhibiting NPY/AgRP production within the ARC (Cowley et al. 2001; Frederich et al. 1995). This in turn decreases production of the orexigenic peptides orexin (ORX) and melanin-concentrating hormone (MCH) within the LHA and suppresses appetite and food intake. When leptin levels are low, the orexigenic neurons in the ARC produce NPY and AgRP, which stimulate hunger and decrease energy expenditure via signaling to the LHA. Interestingly, in addition to regulation by nutrient status, leptin also displays a circadian pattern of expression. While our knowledge of circadian regulation of circulating hormones from the periphery is still quite incomplete, it will ultimately be important to identify how signals regulated by the nutritional status of the organism communicate with and co-regulate brain centers involved in control of activity, feeding, sleep, and metabolism. In this regard it is of note that NPY, AgRP, and orexin display circadian patterns of expression within the hypothalamus, with peaks around the beginning of the active period, while α-MSH levels are highest at the beginning of the inactive period (Kalra et al. 1999; Lu et al. 2002). Thus, understanding the neural networks integrating centers involved in regulation of circadian rhythms, sleep, and energy homeostasis may shed light on the interplay between anticipatory and adaptive behaviors involved in long-term energy constancy.

3.2 Circadian Gene Polymorphisms and Metabolic Phenotypes in Humans

In addition to the recent findings in animal models, mounting evidence suggests that genetic variation in circadian genes also influences metabolic parameters in humans. A number of genome-wide association studies (GWAS) have uncovered links between polymorphisms in CLOCK and susceptibility to hypertension, obesity, and metabolic syndrome (Garaulet et al. 2009; Sookoian et al. 2008, 2010). Single-nucleotide polymorphisms (SNPs) in CLOCK are also associated with high plasma ghrelin concentrations, short sleep duration, and altered eating behaviors, leading to higher total energy intake, decreased compliance with prescribed diet plans, and, ultimately, resistance to weight loss (Garaulet et al. 2010b, 2011). Case–control studies have further revealed a correlation between common genetic variations of CLOCK and the incidence and severity of nonalcoholic fatty liver disease (NAFLD), one of the most common disorders observed in obese persons (Sookoian et al. 2007).

Genetic variants in BMAL1 have also been linked with the development of hypertension and type 2 diabetes, and SNPs in loci in or near CRY2 have been associated with fasting glucose concentrations (Dupuis et al. 2010; Woon et al. 2007). PER2 polymorphisms have also been linked with hyperglycemia, abdominal obesity, and unhealthy feeding behavior phenotypes, while expression of PER2 mRNA may also correlate with metabolic markers in humans (Ando et al. 2010; Englund et al. 2009; Garaulet et al. 2010a; Gomez-Abellan et al. 2008). Conversely, a rare variant of NAMPT1, a gene involved in a clock negative feedback loop, is associated with protection from obesity (Blakemore et al. 2009).

GWAS studies have also recently indicated that MTNR1A and MTNR1B, G-protein-coupled receptors for melatonin, a circadianly regulated circulating hormone implicated in the regulation of sleep and glucose homeostasis, are associated with high fasting plasma glucose levels (Dubocovich et al. 2003; Li et al. 2011; Ling et al. 2011; Peschke and Muhlbauer 2010; Ronn et al. 2009; Takeuchi et al. 2010; Tam et al. 2010). Genetic variants in and near MTNR1B are further associated with impaired insulin secretion and increased risk of developing type 2 and gestational diabetes mellitus, while the SNP rs2119882 in the MTNR1A gene is linked to insulin resistance and susceptibility to polycystic ovary syndrome, a hormonal disorder in women often associated with obesity, type 2 diabetes, and heart disease (Kim et al. 2011; Kwak et al. 2012; Li et al. 2011; Lyssenko et al. 2009; Staiger et al. 2008).

While most GWAS analyses do not take into account the added complexity of gene–gene interactions, one study demonstrated the existence of a synergistic effect of specific polymorphisms in PER2, PER3, CLOCK, and BMAL1 that accounts for morning or evening activity preference in humans (Pedrazzoli et al. 2010). The extensive cross talk between circadian and metabolic networks further adds to the complexity of analysis of human genetic studies. For instance, environmental dysregulation of metabolic homeostasis, such as overnutrition with western diet, may reciprocally feedback to impair circadian homeostasis (Kohsaka et al. 2007). Further, carriers of the CLOCK SNP rs4580704 display lower glucose levels and improved insulin sensitivity only when on a diet high in monounsaturated fatty acids, while an association between the CLOCK SNP rs1801260 and increased waist circumference is evident only in the presence of saturated fatty acids (Garaulet et al. 2009). A complete understanding of circadian gene function in humans will ultimately require consideration of both genetic and nutritional variables. As new evidence for interactions between clocks and metabolism emerges from epidemiologic and association studies, it underscores the clinical importance of rhythmicity for the maintenance of metabolic homeostasis and emphasizes the significance and implications of a better understanding of the interconnections between the two systems.

4.2 Circadian Control of NAD⁺ Biosynthesis and Sirtuin/PARP Activity

One potential molecule that has been implicated as a mediator between circadian and metabolic pathways is nicotinamide adenine dinucleotide NAD⁺, a key cofactor involved in cellular redox reactions. The molecular clock directly regulates transcription of nicotinamide phosphoribosyltransferase (NAMPT), a key rate-limiting enzyme in the NAD⁺ salvage pathway (Nakahata et al. 2009; Ramsey et al. 2009). Consistent with circadian regulation of NAMPT, NAD⁺ levels also oscillate in peripheral tissues such as liver and adipose tissue, even when animals are maintained in constant darkness (Ramsey et al. 2009; Sahar et al. 2011). Importantly, mice with mutations in the activator genes Clock and Bmal1 exhibit constitutively low NAD⁺ levels, while those deficient in the clock repressor genes Cry1 and Cry2 have elevated NAD⁺, indicating direct regulation of NAD⁺ by the clock (Ramsey et al. 2009).

In addition to its role in redox reactions, NAD⁺ also acts as a cofactor for several enzymatic reactions, including NAD⁺-dependent deacetylation and ADP-ribosylation. The circadian clock was recently shown to modulate the activity of the metabolic enzyme, SIRT1, a class III protein deacetylase and a member of the sirtuin family of NAD⁺-dependent deacetylases. SIRT1 resides primarily in the nucleus and targets several transcription factors involved in the maintenance of nutrient flux, including Peroxisome Proliferator-Activated Receptor Gamma Coactivator 1-alpha (PGC-1α), Forkhead Box Protein O1 (FOXO1), Transducer of Regulated CREB Protein 2 (TORC2), Sterol Regulatory Element-Binding Protein 1c (SREBP-1c), and Signal Transducer and Activator of Transcription 3 (STAT3) (Haigis and Sinclair 2010; Rodgers et al. 2005; Sahar et al. 2011; Saunders and Verdin 2007). SIRT1 is a critical regulator of metabolic processes such as gluconeogenesis, lipid metabolism, and insulin sensitivity, as well as lifespan (Haigis and Sinclair 2010; Sahar et al. 2011; Saunders and Verdin 2007). Through rhythmic NAD⁺ biosynthesis, the circadian clock modulates SIRT1 activity, which then coordinates the daily transitions between the periods of fasting and feeding. Of note, SIRT1 also modulates CLOCK/BMAL1 activity, generating a negative feedback loop in which circadian control of NAD⁺-mediated sirtuin activity in turn regulates the clock itself through interaction with PER2, CLOCK, and BMAL1 (Asher et al. 2008; Grimaldi et al. 2009; Nakahata et al. 2008). Thus, the cross talk between the biological clock and the NAMPT/NAD⁺/SIRT1 pathway provides a nexus linking the circadian system and nutrient-sensing pathways.

It will be of great interest to determine whether circadian control of NAD⁺ biosynthesis affects other NAD⁺-dependent metabolic enzymes, including the six other mammalian sirtuin homologs (SIRT2–7) or the poly (ADP-ribose) polymerase (PARP-1). SIRT3–5 are primarily localized to mitochondria and have recently been characterized as important regulators of oxidative (fasting) metabolic pathways, including fatty acid oxidation, TCA cycle, ketogenesis, urea cycle, and oxidative phosphorylation (Haigis et al. 2006; Hallows et al. 2006, 2011; Hirschey et al. 2010; Huang et al. 2010; Nakagawa and Guarente 2009; Schwer et al. 2006; Shimazu et al. 2010; Someya et al. 2010). While little is known about circadian rhythms of NAD⁺ in mitochondria, it is intriguing to speculate that the circadian clock may influence mitochondrial sirtuins. PARP-1, which also plays an important role in the response to metabolic stress, has recently been identified as a regulator of clock gene expression through direct ADP-ribosylation of CLOCK and inhibition of CLOCK/BMAL1 DNA-binding activity (Asher et al. 2010). Interestingly, PARP-1 activity in mice is rhythmic, suggesting circadian control of ADP-ribosylation (Asher et al. 2010; Kumar and Takahashi 2010). However, at present, there is no evidence to suggest that these oscillations are mediated by clock-dependent control of NAD⁺ biosynthesis.

4.3 A Role of Redox in Circadian Rhythms

Prior to the discovery that the molecular clock directly regulates rhythms of NAD⁺, in vitro studies by McKnight and coworkers revealed that cellular redox status affects the circadian clock. Increased levels of oxidized cofactors (NAD⁺ or NADP⁺) decrease the ability of CLOCK/BMAL1 and NPAS2/BMAL1 to bind to DNA in purified systems, suggesting that cellular redox changes may be sufficient to entrain clocks (Rutter et al. 2001).

Interestingly, recent studies have further identified ~24-h rhythms in the cellular redox state, which control circadian oscillations of the oxidation state of the peroxiredoxin family of antioxidant enzymes (Kil et al. 2012; O'Neill and Reddy 2011; O'Neill et al. 2011; Vogel 2011). Oscillating peroxiredoxin redox state affects its antioxidant activity, generating self-sustained rhythms of cellular redox status even in the absence of transcriptional control of circadian gene expression (both in anucleate human red blood cells and in single-celled alga Ostreococcus tauri treated with inhibitors of transcription and translation) (O'Neill and Reddy 2011; O'Neill et al. 2011). More recent studies have revealed that rhythms of peroxiredoxin oxidation exist in all domains of life and that these rhythms persist even in the presence of genetic clock disruption in mammalian cells, fungi and flies (Edgar et al. 2012; O'Neill and Reddy 2011). These studies raise the possibility that oscillations in cellular redox state may be an integral mechanism by which circadian rhythms of metabolic processes are controlled and that these rhythms may be maintained independently from the molecular clock transcriptional feedback loop.

4.4 A Link Between Circadian Rhythms and Nuclear Hormone Receptor Pathways

Several recent studies have focused on the role of nuclear hormone receptors as potential nutrient-sensing factors linking metabolic and circadian pathways. NHRs comprise a large family of proteins, containing both DNA- and ligand-binding domains that regulate their activities as transcriptional activators and/or repressors. Known NHR ligands include a wide range of molecules (e.g., steroid hormones, fatty acids, heme, sterols), though many NHRs are still classified as “orphan” since their endogenous ligands have not yet been identified (Sonoda et al. 2008). Interestingly, more than half of the ~50 known NHRs display rhythmic expression patterns in peripheral tissues and are thus attractive candidates as integrators of circadian and nutrient-sensing pathways (Asher and Schibler 2011; Teboul et al. 2009; Yang et al. 2006, 2007).

Some NHRs, such as members of the REV-ERB and ROR families, are both transcriptional targets of the CLOCK/BMAL1 complex as well as transcriptional regulators of clock genes themselves. In the SCN and most metabolic tissues, REV-ERBs repress and RORs activate Bmal1 transcription, generating a short negative feedback loop (Akashi and Takumi 2005; Duez and Staels 2009; Preitner et al. 2002; Sato et al. 2004) (refer also to Sect. 3.1). As sensors of metabolites including heme, fatty acids, and sterols, REV-ERBα/β and RORα integrate nutrient signals with transcriptional regulation of the clock (Jetten 2009; Kallen et al. 2004; Yin et al. 2007). REV-ERBs and RORs also directly interact with other important metabolic factors, including the transcriptional regulator PGC-1α, which coactivates Bmal1 with RORα and β and also plays a key role in regulation of mitochondrial oxidative metabolism (Grimaldi and Sassone-Corsi 2007; Liu et al. 2007). Interestingly, activity of PGC-1α is regulated by SIRT1-mediated NAD⁺-dependent deacetylation, suggesting that circadian regulation of NAD⁺ may represent yet another metabolic feedback loop involving PGC-1α (Rodgers et al. 2005).

In addition to REV-ERBs and RORs, other NHRs are important for the coordination of molecular clocks and nutrient-sensing pathways. For example, members of the PPAR family of NHRs are also regulators of clock gene expression. PPARα activates both Bmal1 and Rev-erbα in liver (Li and Lin 2009; Schmutz et al. 2010; Yang et al. 2007). Conversely, several proteins, including PPARα, γ, and δ, are transcriptionally regulated by the circadian clock (Li and Lin 2009). Ligands for PPARs include various types of lipids, including the circulating gut metabolite oleylethanolamide (OEA), which is generated and released from the small intestine and suppresses food intake during the rest period in a PPARα-dependent manner (Fu et al. 2003; Rodriguez de Fonseca et al. 2001). In addition, PPARγ maintains daily rhythms of blood pressure and heart rate in blood vessels via regulation of Bmal1 expression (Wang et al. 2008). Further studies will be necessary to determine the importance of other lipid-derived PPAR ligands as effectors of the molecular clock.

The glucocorticoid receptor (GR) is another important NHR involved in the cross talk between central and peripheral circadian clocks. GR is expressed in many metabolic tissues, including liver, skeletal muscle, heart, and kidney, and activates several metabolic pathways including lipid metabolism and gluconeogenesis (Dickmeis and Foulkes 2011). GR is activated by glucocorticoid (GC) steroid hormones, which are produced and secreted in a circadian manner from the adrenal cortex. Timed release of GC from the adrenal cortex is generated through relays from the SCN to the hypothalamic-pituitary-adrenal axis (HPA). As such, GC rhythms are entrained by light and peak in the early morning in humans (Chung et al. 2011; Oster et al. 2006). Rhythmic GC release affects GR activity in peripheral tissues and thereby acts to synchronize peripheral clocks with the SCN (Teboul et al. 2009). Indeed, liver-specific GR knockout mice display accelerated clock phase-shifting in response to daytime food restriction, suggesting a lack of entrainment by the central clock (Le Minh et al. 2001). Furthermore, pharmacological activation of GR by dexamethasone resets the peripheral clock in liver, heart, and kidney, presumably by direct GR regulation of Rev-erbα and Per1 expression (Balsalobre et al. 2000; Torra et al. 2000; Yamamoto et al. 2005). While detailed mechanisms underlying clock resetting by GR have not been fully defined, GR appears to be a key entrainment signal of peripheral clock rhythms, allowing for the coupling of food- and SCN-derived resetting cues.

C.B. Peek, A. Affinati and B. Marcheva are supported by NIH grants F32 DK092034-01, 1F30DK085936-01A1, and T32 DK007169, respectively. J. Bass is supported by NIH grants R01 HL097817-01, R01 DK090625-01A1, and P01 AG011412, the American Diabetes Association, the Chicago Biomedical Consortium Searle Funds, and the University of Chicago Diabetes Research and Training Center (grant P60 DK020595).