Circadian Rhythms and Metabolic Syndrome: From Experimental Genetics to Human Disease

Forward genetics and positional cloning enabled identification of the first mammalian circadian clock gene and provided an entry point into a molecular understanding of the clock mechanism.^51,52 The core molecular clock is composed of a transcription–translation feedback loop that oscillates with 24-hour rhythmicity (Figure 1). The driving force is the positive limb of the clock comprised of the bHLH-PAS (basic helix-loop-helix–Period-Arnt-Single-minded) transcription factors CLOCK (circadian locomotor output cycles kaput), and its paralog NPAS2 (neuronal PAS domain protein 2), and BMAL1/ARNTL (brain and muscle aryl-hydrocarbon receptor nuclear translocator-like 1/aryl-hydrocarbon receptor nuclear translocator-like). CLOCK or NPAS2 and BMAL1 heterodimerize and activate the rhythmic transcription of downstream target genes that contain E-box cis-regulatory enhancer sequences, including the Period (Per1, Per2, and Per3) and cryptochrome (Cry1 and Cry2) genes.^51,53–56 Following translation, PER/CRY dimerize and translocate back to the nucleus where they directly inhibit the CLOCK/BMAL1 complex, effectively repressing their own transcription.^57–59

Additional regulatory loops are interconnected with the core loop described above, providing multiple layers of control of the core circadian clock.^60,61 In addition to the Per and Cry genes, CLOCK/BMAL1 also activate transcription of the retinoic acid-related orphan nuclear receptors Rev-erbα and Rorα.^62–65 REV-ERBα binds to the retinoic acid–related orphan receptor (ROR) response element (RORE) in the Bmal1 promoter resulting in inhibition of transcription and this action is opposed by RORα, which activates the RORE.^62–64,66 In addition to the nuclear hormone receptor feedback loop, PAR domain basic leucine zipper transcription factors (PAR bZIP), including DBP (D-site binding protein), TEF (thyrotroph embryonic factor), HLF (hepatic leukemia factor), and the cAMP pathway (CREB-ATF-CREM) also feedback on the clock, acting through cognate D box and CREB elements respectively.^{60,61,67–70} Posttranslational modification, including phosphorylation and ubiquitination, provide further regulation of the clock network. Casein kinase 1ε and -δ (CK1ε and CK1δ) phosphorylates PER and CRY, tagging them for polyubiquitylation by the E3 ubiquitin ligase complexes β-TrCP1 (β-transducin repeat containing protein 1) and FBXL3 (F-box and leucine-rich repeat protein 3), respectively, ultimately leading to their degradation by the 26S proteosome.^71–78 In addition to phosphorylation mediated by the casein kinase family, in Drosophila and mammalian cells, a role for GSK3-β signaling has been established.^78a,79 As a result of this degradation, CLOCK/BMAL1 is released from repression, activating the forward limb of the 24-hour cycle. Lastly, epigenetic regulation has emerged as an additional node in circadian systems (discussed further below), including the possibility that CLOCK participates directly in protein acetylation and chromatin modification.⁸⁰

Genetic mouse models have revealed key roles for each of the core clock genes in the generation and maintenance of circadian rhythms. Bmal1 knockout mice display a complete loss of circadian rhythmicity in constant darkness.⁵³ Mice with a dominant-negative Clock mutation have an approximately 4-hour lengthening of their free-running period and become arrhythmic in constant darkness.⁵² Of note, Clock knockout mice have normal locomotor activity rhythms, because of developmental compensation by NPAS2.^81,82 Compensation can also be demonstrated within the negative limb of the clock, as both Per1/Per2 and Cry1/Cry2 knockout mice display a much more pronounced loss of circadian rhythmicity compared to their single mutant counterparts.^56,83–86 Although the aforementioned studies have focused on overt locomotor activity rhythms, it remains uncertain as to whether compensation extends to functions of the clock within peripheral tissues. It is likely that future genetic studies will continue to identify additional regulators and modifying loops of the core clock mechanism.

Many metabolic functions occur at specific times of the day. Indeed, understanding the effects of molecular clock gene disruption on organismal physiology can be advanced through consideration of the molecular and hierarchical organization of the clock. The location of the master neural clock in mammals was originally discovered through classical lesioning studies within pacemaker neurons of the brain: the suprachiasmatic nucleus (SCN) of the hypothalamus (for more complete review, see^87,88). The SCN controls physiological and behavioral circadian rhythms and coordinates peripheral clocks through hormonal and neural signals.⁸⁷ The master role played by SCN was demonstrated by transplantation experiments in hamster. Neural grafts from the suprachiasmatic region restored circadian locomotor and feeding activity to arrhythmic animals whose own SCN had been ablated.⁸⁹ The restored rhythms of the host always matched the rhythms of the donor, demonstrating the strong impact of this nucleus in circadian activity. However, despite the restoration of locomotor rhythmicity, melatonin and glucocorticoids remained arrhythmic, suggesting that neural connections must be critical for the generation of certain circadian rhythms.⁹⁰ Interestingly, this master pacemaker has anatomic connections with several regions of the CNS involved in the control of appetite, energy expenditure regulation and behavioral activity, namely with the supraventricular area, the arcuate nucleus and the lateral hypothalamic area.⁹¹

The SCN clock is entrained by light through the retinohypothalamic tract (Figure 2). Photic input provides a dominant time-keeping signal (zeitgeber), orienting the animal each day to geophysical time. Endogenous period length is not precisely 24 hour (humans are longer, whereas mice are shorter), thus the daily entrainment to light is a critical mechanism to maintain organismal synchrony with the external environment (Figure 2). Perception of light occurs through activation of a population of directly light-sensitive ganglion cells within the eye, the melanopsin cells; these regulate both circadian rhythms and melatonin synthesis.^92,93 Direct output of the SCN, and the entrainment of the SCN axis to light, plays a key role in synchronizing endogenous hormonal rhythms including the glucocorticoid rhythm.⁹⁴ In turn, light-induced entrainment of the glucocorticoid rhythm may maintain phase coherence of multiple cellular oscillators, such as those in fibroblasts and liver.^95,96

In addition to photic entrainment, food also entrains circadian processes in neural and peripheral cells (Figure 2). Food restriction to the normal rest period in rodents also induces a burst of anticipatory activity, an effect that is altered in some experimental systems following lesioning of the dorsomedial nucleus.^97–99 However, there remains controversy concerning the involvement of circadian oscillators in food anticipatory activity (FAA) because the behavior persists in Bmal1 nullizygous mice.^97,100,101 Interestingly, FAA appears to involve the melanocortin signaling pathway, because restriction failed to increase wakefulness before food presentation in melanocortin-3 receptor–null mice.¹⁰² Rather than localization to a single nucleus of hypothalamus, the food entrainable oscillator may in fact involve a more dispersed network of cell groups.¹⁰³ Furthermore, the FAA may constitute a metabolic oscillator responsive to peripheral neural or circulating signals elicited by food ingestion.^104,105 An interesting question remains concerning whether macronutrient flux in the postprandial state may participate in establishing FAA (reviewed in^104,105). A related question is whether nutrient signaling per se may affect core properties of the SCN pacemaker.

Molecular analyses have revealed that the clock network is also widely expressed throughout nearly every tissue/cell type in vertebrates (Figure 3).^106,107 Original studies by Schibler and colleagues demonstrated cell autonomous clock gene oscillation within fibroblasts ex vivo.¹⁰⁸ Following this discovery, in addition to the master clock in the SCN, independent circadian oscillators have been found in a number of peripheral tissues in mammals. Gene expression profiling has shown that 3% to 20% of genes display a 24-hour rhythmic expression, and a large proportion of these genes have a role in metabolic processes (reviewed previously⁸). The circadian rhythms of peripheral organs are also self-sustained, as demonstrated using a mouse line in which luciferase expression is driven from the endogenous Per1 or Per2 loci.^106,107 Variation in temporal gene expression was reported to play an important role in tissues implicated in glucose and lipid metabolism, such as fat, liver, cardiac, and skeletal muscle.^109–117 Many nuclear receptors expressed in liver and white and brown adipose tissues also display rhythmic patterns of expression.^118–120 Therefore, the nuclear receptors may link clock genes to metabolism by integrating energy flux with varying physiological demands across the light-dark cycle. In this way, circadian patterns of metabolic gene expression may optimize the switch between daily anabolic and catabolic states corresponding with periods of feeding and fasting. For example, the cyclic expression of gastrointestinal tract enzymes may ensure that factors involved in nutrient absorption are expressed in anticipation of daily episodes of food ingestion.¹⁰⁵ In addition, adipose enzymes involved in fatty acid storage peak coincident with feeding (reviewed previously¹²¹). Moreover, components of gluconeogenesis, glycolysis, and fatty acid metabolism cycle with a peak during the subjective night in mouse liver.¹¹² Coordinating gene expression patterns according to the varying metabolic demands across the active and rest period is also important in muscle, where elaboration of aerobic and anaerobic enzymes varies during the sleep-wake cycle.^111,122 Thus, peripheral oscillators are self-sustained, cell autonomous and tissue-specific, yet a major question is: what are the mechanisms involved in maintaining synchrony within and between peripheral tissue clocks? A related question is whether misalignment of local circadian oscillation within and between peripheral tissues contributes to cardiovascular and metabolic pathologies.

To discern whether the rhythmic expression of genes in peripheral organs is driven by local (cell autonomous) oscillators or by circadian systemic signals, Schibler and colleagues have recently exploited the tetracycline-inducible system of Bujard, enabling conditional Rev-erbα overexpression within liver.¹²³ In this model, REV-ERBα represses the transcription of the essential core clock gene Bmal1 in a doxycycline-dependent manner. Among 351 genes with rhythmic expression revealed in the doxycycline-fed mice, only 31 genes, including the core clock gene mPer2, oscillated robustly irrespective of whether the liver clock was running or not. These studies suggest that the rhythmicity of metabolic liver genes is driven by both cell-autonomous and nonautonomous signals.¹²³

Multiple signals related to feeding, and even fasting, may entrain peripheral clocks. Indeed, in in vitro experiments, a bewildering variety of stimuli can induce or reset circadian gene expression. These factors include chemical activators of protein kinase A (forskolin, butyryl cAMP), protein kinase C, and/or mitogen-activated protein kinase (phorbol esters, fibroblast growth factor, endothelin) and glucocorticoid receptors (dexamethasone) and even glucose; dissecting how these signaling pathways converge on the clock remains an area of intensive investigation (reviewed previously¹⁰⁸).

How might circadian misalignment impact the metabolic comorbidities of obesity, diabetes, and cardiovascular disease? Several lines of evidence suggest that circadian dysregulation may exert a broad impact not only on glucose control, but also on inflammation, fibrinolysis, fluid balance, and vascular reactivity. A central node linking metabolic and circadian pathways involves the nuclear receptor superfamily, including those downstream of REV-ERBα and the RORs that modulate the core clock and diverse metabolic processes ranging from adipogenesis to inflammation and thrombosis (reviewed previously¹⁴⁰). Experimental models have helped to demonstrate the impact of the clock network in metabolic gene expression and provide evidence that this clock disruption leads to metabolic abnormalities. Homozygous Clock^Δ19 mutant mice, which express a loss of function mutation in Clock, have yielded new insight in this field.¹⁴¹ In addition to disruptions in sleep and circadian behavior, these mice also develop hyperphagia early in life, with subsequent development of hyperlipidemia, hyperleptinemia, and hypoinsulinemic hyperglycemia, indicating that this animal exhibits features of the metabolic syndrome.¹⁴¹ The feeding rhythm in these mice is damped, with increased food intake during the day, and, in addition, these mice have significantly increased food intake overall. High-fat feeding studies revealed exaggerated weight gain of Clock^Δ19 mutant mice, and dual-energy X-ray absorptiometry scanning and fat pad weight both demonstrated significant increases in fat and lean mass relative to controls following high fat feeding.¹⁴¹ It is likely that the obese phenotype results, at least in part, from altered rhythms of neuropeptides in the hypothalamus, because ghrelin, CART (cocaine- and amphetamine-regulated transcript), and orexin are all expressed at constitutively low levels in the Clock^Δ19 mutant mice.¹⁴¹ In addition, the anorectic neuropeptide POMC was decreased throughout the entire light/dark cycle in hypothalami of young Clock^Δ19 mutant before the onset of weight gain and overt diabetes and is consistent with a deficit in the central homeostatic regulation of weight constancy. Because the original Clock^Δ19 mutant was developed in a melatonin-deficient strain (C57BL/6J), Kennaway et al evaluated the contribution of melatonin deficiency on glucose metabolism by crossing Clock^Δ19 mutant mouse with the melatonin-producing CBA strain to produce the “Clock^Δ19+MEL” mouse.¹⁴² Interestingly, in this model, the restoration of melatonin did not rescue gene expression rhythms in liver or muscle.¹⁴³ Such studies underscore the importance of strain background in the evaluation of metabolic phenotype. For example, when introgressed onto the ICR strain, the Clock^Δ19 mutation results in malabsorption of lipid and thus resistance to diet-induced obesity, thus primary effects of the Clock mutation on energy balance and fuel homeostasis cannot be evaluated in the ICR strain.¹⁴⁴ Disruption of other circadian clock genes also leads to metabolic alterations. For example, gene disruption in Bmal1 induces an abnormal metabolic phenotype characterized by impaired gluconeogenesis, hyperleptinemia, glucose intolerance, and dyslipidemia.^145–147 In addition, Per2 knock-out mice develop increased weight gain on high-fat diet (HFD).¹⁴⁸ Conversely, mice deficient in the circadian deadenylase nocturin remain lean and resistant to hepatic steatosis when fed a HFD despite equivalent caloric intake, similar metabolic rates, and reduced activity compared with control mice.¹⁴⁹

Although clock genes impact metabolic homeostasis, a reciprocal effect of metabolic disruption on circadian rhythms also exists, because diet-induced obesity per se alters circadian behavioral and molecular rhythms in C57BL/6J mice.⁹ Indeed, HFD also attenuates the amplitude of diurnal rhythms of feeding and locomotor activity, as high fat fed mice increase their food intake during their normal rest (light) period.⁹ Interestingly, genetically obese animals are resistant to weight gain when feeding is restricted to the active (dark) phase.¹⁵⁰ In agreement with these observations, recent evidence demonstrated that circadian timing of food intake contributes to weight gain.¹⁵¹ Indeed, mice fed a HFD only during the 12-hour light phase gain significantly more weight compared to isocalorically fed mice provided food only during the 12-hour dark phase.¹⁵¹ Further studies are necessary to understand how the timing of food intake impacts energy constancy. Interestingly, a recent study demonstrated that treatment with an antagonist of T-type calcium channel, which is involved in sleep-wake regulation, improved HFD-induced behavioral alterations, including both a decrease in inactive phase activity, core body temperature, feeding and adiposity.¹⁵² Taken together, these observations largely based on animal studies, raise important questions concerning the impact of circadian misalignment and clock gene disruption on obesity and its metabolic complications, and suggest avenues for future investigation in human subjects.

Excess adipose tissue and altered body fat distribution, rather than adiposity per se, is an important risk factor for obesity-related disorders. Excess intra-abdominal fat rather than subcutaneous fat (central versus peripheral obesity) is associated with MS and cardiovascular disease.³ However, the mechanisms responsible for this association, and its causality, remain uncertain. Emerging evidence from both cell-based and human studies suggests that expression of the circadian clock transcription network within adipose tissue may influence both adipogenesis and the relative distribution of subcutaneous versus visceral depots.^121,153,154

In adipose tissue, the clock machinery controls the expression of a large array of enzymes involved in lipid metabolism. Indeed, adenovirus-mediated expression of BMAL1 in 3T3-L1 adipocytes resulted in induction of several factors involved in lipogenesis, whereas BMAL1 deletion in adipose cell lines resulted in impaired adipogenesis.¹⁴⁷ Furthermore, heme, the REV-ERBα/β natural ligand, has long been known to enhance adipocyte differentiation in vitro.¹⁵⁵ Activation of SIRT1, which regulates the clock network, may increase insulin sensitivity and reduce the inflammatory response in adipocytes,^156,157 however it is unclear whether the effect is direct or not.

Experiments in mice have revealed that temporally restricted feeding causes a coordinated phase-shift in circadian expression of core clock genes and their downstream targets in adipose tissues.¹¹⁷ In addition, HFD also alters the cyclic expression and function of core clock genes and clock-controlled genes in adipose tissue, resulting in disrupted fuel use.^9,158 Of further interest, clock gene disruption targeted to the fat body in flies is sufficient to induce increased food consumption, decreased glycogen levels, and increased sensitivity to starvation.¹⁵⁹ At least in flies, these findings suggest involvement of a peripheral tissue clock in neural energy homeostasis.¹⁵⁹

Several teams have recently started to examine the potential relationship between clock gene expression and metabolic syndrome parameters in humans. Expression levels of the core molecular clock genes in cultured visceral and subcutaneous fat explants obtained from morbidly obese subjects correlated with certain metabolic syndrome parameters, such as waist circumference, sagittal diameter, and body mass index.^153,154,160 However, biopsies from human fat likely represent a heterogenous mixture of adipose cells in addition to macrophages, thus conclusions must be viewed with caution regarding the contribution of adipose tissue to the observed circadian patterns of gene expression. Interestingly, circadian rhythms of gene expression are sustained ex vivo in human fat explants,^160,161 including the rhythmic oscillation of genes involved in glucocorticoid turnover.¹⁶¹ In agreement, human adipose biopsies removed at different zeitgeber times reflect different levels of gene expression consistent with the observed circadian rhythmicity found in cell-culture studies.¹⁶² Further studies are necessary to gain more detailed insight into the relationship between temporal patterns of gene expression in adipose tissue and development of MS.

In addition to effects of circadian transcription on intracellular metabolic pathways, clock dysregulation in adipose tissue and/or misalignment with meal times may lead to inappropriate expression patterns of enzymes involved in lipid metabolism such as lipoprotein lipase.¹²¹ For example, misalignment between the fasting/feeding cycle and lipogenic and/or lipid catabolic gene expression pathways may perturb fatty acid flux and contribute to lipotoxicity. Indeed, circadian synchrony may play a distinct role not only within different tissue types (liver versus muscle) but also within distinct adipose depots (visceral versus subcutaneous).¹⁶⁰ It is further possible that differences of circadian gene expression patterns within visceral adipose tissue and subcutaneous adipose tissue depots may contribute to cell-autonomous differences in inflammatory, lipogenic, and/or lipolytic pathways within these locales.^163,164 The limited storage-capacity of fat and/or increased lipolysis results in an overflow of fatty acids to ectopic sites such as liver, muscle, and islets (reviewed previously^165,166). Interestingly, both have been proposed to be involved in the etiology of the MS.^3,167

Another important function of adipose tissue is its secretion of numerous bioactive peptides or proteins, collectively named “adipokines.” These may play a central role in energy and vascular homeostasis, as well as immunity, and are fundamental to the pathogenesis of the MS (reviewed previously¹⁶⁸). Because obesity-related inflammation is receiving increased attention for its potential role in the pathogenesis of MS, steatosis and cardiovascular disease, it may be opportune to consider the impact of circadian systems at the level of adipokine regulation. In mice, leptin exhibits rhythmic production across the light-dark cycle that appears to be dependent of the feeding rhythm.¹⁶⁹ Interestingly, in obese humans, disruption of the 24 hour profiles of leptin and adiponectin was observed compared to healthy lean subjects.^170,171 These adipokines play major roles in fuel partitioning and insulin sensitivity but also regulate immunity.^168,172 Indeed, leptin was the first adipokine found to control energy balance.¹⁷³ The metabolic effects of leptin are thought to primarily involve its actions within brain,^174–176 whereas adiponectin functions primarily within peripheral target tissues. Many of the metabolic effects of leptin and adiponectin involve activation of AMP-activated protein kinase signaling in muscle/liver.^177,178 Following the discovery of leptin, a growing list of adipokines has been identified, some of which also exert proinflammatory roles. Because many adipokines are expressed in a circadian fashion in humans,¹⁶² it is tempting to speculate that regulation of adipokine oscillation may be important in metabolic homeostasis. In addition, certain adipokines may also interact with or modulate sleep. For instance, leptin is involved in sleep regulation as demonstrated by EEG monitoring of sleep in leptin deficient and leptin resistant mice.^179,180 Interleukin-6 and tumor necrosis factor-α plasma levels, which are increased in obesity^181,182 may also impair circadian clock gene oscillations and promote sleep.^183,184

Circadian regulation may extend to effects within adipose tissue on endoplasmic reticulum (ER) stress, an important component of the inflammatory response in this tissue.¹⁸⁵ Obesity results in conditions that increase demand on the ER in metabolic tissues including liver, adipocytes and pancreas, resulting in a persistent inflammatory state.¹⁸⁶ For example, accumulation of reactive oxygen species, which are abundantly produced by both the ER and the mitochondria during conditions of stress are increased in metabolic organs in MS.^185,186 In adipose tissue, ER stress is involved in adipogenesis and adipokine oversecretion.^187,188 Interestingly, the endoplasmic reticulum chaperone protein BiP, a key protein involved in the ER stress response, is expressed in a circadian manner in flies.¹⁸⁹ It has also been reported that clock genes may influence the production of reactive oxygen species.^8,190,191 Thus, disrupted synchrony of stress response gene expression may alter adipose function and thereby directly contribute to insulin resistance. ER stress may also be induced in brain following high fat feeding, thereby contributing to leptin resistance¹⁹² and perhaps circadian and sleep disturbances (reviewed previously¹⁹³).

In addition to the impact of white adipose tissue excess in MS pathogenesis, several independent groups recently demonstrated that brown adipose tissue is present and active in adult humans, and its presence and activity are inversely associated with adiposity and indexes of the metabolic syndrome.^194–196 As numerous genes including nuclear receptors exert circadian expression profiles in brown adipose tissue,^116,117 alterations of circadian oscillator genes in fat may have significant metabolic implications.

Disruption in the normal cyclic pattern of glucose tolerance is a hallmark of type 2 diabetes,¹⁹⁷ and as such, understanding the circadian control of glucose metabolism is critical for delivering the best clinical diabetes management. Strong evidence from human studies demonstrates rhythmic variation in glucose tolerance and insulin action across the day.^198–202 For example, oral glucose tolerance is impaired in the evening compared to the morning,^203,204 an effect which is believed to be attributable to a combination of both decreased insulin secretion and altered insulin sensitivity in the evening.^{200,205–210} The ‘dawn phenomenon’ is also a well-described phenomenon where glucose levels are known to peak before the onset of the active period.^211,212 Furthermore, studies in rats have revealed that the SCN is critical for the maintenance of diurnal variations in glucose metabolism.²⁰⁷

Although these studies indicate a role for circadian systems in the control of glucose metabolism, the molecular mechanisms underlying these phenomena are not well understood. Recently, human genome-wide association studies and experimental mouse model systems have begun to provide clues as to the nature of the molecular links between rhythms and glucose metabolism. As described above, data from several independent groups has now demonstrated that genetic variants of the melatonin receptor may be involved in abnormal glucose homeostasis^47,48 and that melatonin treatment of pancreatic β-cells inhibits glucose-induced insulin release.⁴⁹ Furthermore, mice that have a mutation in the Clock gene develop hypoinsulinemic hyperglycemia, and mice nullizygous for Bmal1 have impaired glucose tolerance,^141,145,146 suggesting that a functional clock network is required for the maintenance of glucose homeostasis. The cellular etiology of impaired glucose tolerance in circadian mutant animals remains an important yet unresolved area of research.

Finally, it is interesting to speculate that the NAD biosynthetic enzyme NAMPT and SIRT1 may play an important role in the circadian control of glucose metabolism. As described above, NAMPT and SIRT1 are regulated by CLOCK/BMAL1 and constitute a negative feedback loop within the core circadian network. NAMPT and SIRT1 have been demonstrated to be involved in a myriad of metabolic functions, including regulation of gluconeogenesis in liver and glucose-stimulated insulin secretion in islets.^44,45 Indeed Nampt-deficient (Nampt^+/−) mice showed impaired glucose tolerance attributable to a defect in glucose-stimulated insulin secretion, which was corrected by intraperitoneal administration of NMN,¹³⁴ and mice overexpressing SIRT1 specifically in their β-cells displayed improved glucose tolerance and increased glucose-stimulated insulin secretion.²¹³ As NAMPT and SIRT1 function are impaired in circadian mutant mice, these data suggest that circadian rhythms of NAMPT and SIRT1 may act in β cells to regulate the daily cycles of insulin secretion and that NAD⁺ might function as an oscillating metabolite linking circadian and metabolic cycles.

Circadian variation in endogenous factors such as autonomic nervous system function, blood catecholamine concentrations, coagulability, heart rate, blood pressure regulation, and platelet aggregability have been suggested to explain the morning onset of myocardial infarction.²³ Conversely, pressure overload–induced hypertrophy and diabetes mellitus result in alterations in the circadian clock within the heart.^214,215 In this regard, it was suggested that alterations in circadian control of cardiac fuel handling may contribute to myocardial contractile dysfunction (reviewed previously²¹⁶). The identification of genes that exhibit circadian regulation within large vessels of the mouse^217–220 has provided clues as to the impact of the circadian system within vasculature. Indeed, clock gene expression within vasculature has been shown to impact blood pressure and thromboocclusive response.^219,221,222 Clock genes may influence the temporal incidence of clinical cardiovascular events by regulating the magnitude of the early morning rise in blood pressure.²¹⁹ Indeed, the circadian variation in blood pressure and heart rate is disrupted in mice with deleted or mutated core clock genes, a phenomenon that may be partially explained by the altered diurnal variation in epinephrine and norepinephrine in these mice.²¹⁹ Interestingly, the mutated mice also showed a reduced response to immobilization stress compared to wild-type mice. Thus, expression of the core clock within peripheral vasculature may modulate the capacity to respond to environmental stressors at different times of day.²¹⁹ Effects of the clock system on blood pressure may also involve the modulation of aldosterone biosynthesis by Per1.²²³ Furthermore, Anea et al demonstrated that Bmal1-knockout and Clock mutant mice present a loss of vascular adaptation and predisposition to thrombosis,²²² both hallmarks of endothelial dysfunction.²²⁴ The endothelial dysfunction in Bmal1-knockout mice has been related to defects in Akt and nitric oxide signaling.²²² Interestingly, the defects in endothelium-dependent arterial relaxation of Clock mutant mice were normalized by entrainment to light, indicating that the vascular phenotype is not simply a consequence of Clock mutation or Bmal1 deficiency but rather the result of behavioral disruption in these animals.²²²

As noted above, NAD⁺ regulation has recently emerged as a major factor coupling circadian rhythms and metabolic signaling pathways. Because NAMPT-mediated NAD biosynthesis has also been shown to impact cardiomyocyte survival pathways, it will be important to ascertain whether dysregulation of NAD contributes to the adverse cardiovascular consequences of circadian disruption.²²⁵

Cardiomyocytes must adapt rapidly to changes in circulating fatty acid, the primary fuel source for contraction. In this regard, PPAR signaling is important in the control of cardiac energy metabolism.²²⁶ Of note, it was also demonstrated that the circadian clock within the cardiomyocyte is essential for responsiveness of the heart to fatty acids.^227–229 To address the circadian function of vascular tissue and the role of PPARγ in the vascular clock, conditional deletion of PPARγ targeted to this tissue was performed.²³⁰ These mice developed abnormalities in blood pressure and heart rate in parallel with a reduction of diurnal variation in the sympathetic nerve activity.²³⁰ Furthermore, vascular PPARγ exhibits a robust cyclic expression, whose rhythmic phase may be reset by changes in feeding time as well as changes in the photoperiod.²³⁰ Thus, the temporal environment may be integrated within the heart by PPARγ.²³⁰ In agreement, PPARγ agonists were found to shift the circadian fluctuation of blood pressure in patients with type 2 diabetes, indicating that vasculoprotective actions of thiazolidinediones may in part involve effects on the clock transcription network.²³¹

Emerging clinical evidence has uncovered unique actions of the PPARα agonist fenofibrate in the circadian control of blood pressure and heart rate in diabetic subjects. In particular, fenofibrate exerted its most marked antihypertensive effects at night.²³² In contrast, only modest decrease antihypertensive effects were detected in studies involving a single measurement during the day.^233,234 In addition, fenofibrate lowered heart rate throughout the 24-hour period.²³² Taking together, these data suggest an interaction between PPARα, blood pressure control and circadian rhythms in diabetes.

A comprehensive temporal map of the nuclear receptor transcriptome provides additional clues concerning the circadian control of cardiovascular physiology.¹²⁰ For instance, RARα and RXRα interact with CLOCK and MOP4 resulting in repression of CLOCK/MOP4: BMAL1 activity in vascular cells.²¹⁸ Indeed, the ligation of retinoic acid, the oxidized form of vitamin A, to its receptors can phase shift Per2 mRNA rhythm in vivo and in smooth muscle cells in vitro.²¹⁸ Whether additional nuclear hormone receptor agonists impact circadian regulation of vascular tone remains a question for future investigation.

Finally, the role of circadian rhythms in the time of onset of thrombotic events has been recognized for many years. Numerous coagulation/ fibrinolytic factors, such as protein C, antithrombin, factor VII, protein S, and fibrinogen, have been demonstrated to fluctuate in a circadian manner in humans. Among these factors, plasminogen activator inhibitor type 1, the most important physiological inhibitor of plasminogen activation, peaks in the early morning, explaining, at least in part, the occurrence of hypofibrinolysis and of prothrombotic state.²³⁵ Circadian control of plasminogen activator inhibitor type 1 gene expression by the REV-ERBα in liver may contribute to the circadian variation in fibrinolysis,²³⁶ an effect that may also involve interactions between the cycle-like factor (CLIF) and CLOCK.²³⁷ Thus, further studies on the circadian gene control of fibrinolysis may shed new light on factors contributing to the prothrombotic state.

Liver also plays a key role in the development of metabolic syndrome (reviewed previously²³⁸). BMAL1 and CLOCK control gene expression of enzymes critical in liver and influence both glucose and lipid homeostasis (reviewed previously²³⁹). A recent study reported that mice with a liver-specific deletion of Bmal1 exhibited hypoglycemia during fasting, indicating a role for the liver clock in maintaining euglycemia during rest.¹⁴⁶ Some of the effects of BMAL1 and CLOCK in liver may involve direct regulation of phosphoenolpyruvate carboxykinase (Pepck).¹⁴⁵ In addition, circadian gene expression in hepatocytes is altered in mouse models of type 2 diabetes²⁴⁰ and by high fat feeding.^9,241 Moreover, HFD induced a phase delay of components of the adiponectin signaling pathway.²⁴¹ Alterations in circadian control of adiponectin signaling may reduce its protective effects²⁴¹ and thereby increase susceptibility to steatosis, a major risk factor in cardiovascular disease.²⁴²

Hepatic clock gene expression also modulates both bile acid and apolipoprotein biosynthesis, raising the possibility that clock disruption may impact multiple components of hepatic lipid homeostasis.²⁴³ For example, several proteins involved in lipid metabolism (such as hepatic cytochrome P450 cholesterol 7 α-hydroxylase, 3-hydroxy-3-methylglutaryl coenzyme A [HMG-CoA] reductase, or apolipoprotein AIV) show diurnal variation in both humans and rodents.¹⁰⁵ Interestingly, Rev-erbα was recently found to play an important role in the control of bile acid metabolism via the regulation of the neutral bile acid synthesis pathway.²⁴⁴ In mouse liver, Rev-erbα expression levels are high during the late light phase, leading to the repression of both small heterodimer partner (SHP) and E4 promoter binding protein 4 (E4BP4) hepatic expression. Reduced levels of SHP and E4BP4 may counter the suppressive effects of bile acids on the cholesterol 7α-hydroxylase (CYP7A1) gene transcription, thereby contributing to the circadian regulation of bile acid and cholesterol homeostasis.²⁴⁴ In addition, Rev-erbα also controls the daily expression of genes involved in cholesterol and lipid homeostasis through circadian modulation of SREBP (sterol regulatory element-binding protein) signaling.²⁴⁵

In addition to effects on lipogenesis, lipid catabolism and thrombosis, the circadian system may also promote inflammatory pathways that contribute to the development of cardiovascular disease. At the molecular level, the circadian transcription factor Rev-erbα, which is expressed in cells from the immune system such as macrophages and other cell types, may impact the inflammatory response¹⁴⁰ (also see elsewhere²⁴⁶). Intriguingly, REV-ERBα increases the tumor necrosis factor-α–induced nuclear factor-κB response, whereas RORα impedes it.¹⁴⁰ As rhythmic mRNA expression of the clock genes is dampened in peripheral leukocytes of patients with type 2 diabetes, this impairment might be involved in its pathogenesis.²⁴⁷