III. Circadian rhythms and metabolism

Genetic manipulations of circadian genes reveal their role in the regulation of energy metabolism. Bmal1-/- mice expressing a dominant negative allele of Clock (ClockΔ19 corresponding to the deletion of exon 19) have impaired glucose tolerance and defective
islet function. In addition, these mice have arrhythmic levels of serum free fatty acids and decreased lipolysis rates, which contribute to the development of obesity. The overexpression of Cry1 leads to the development of features characteristics of diabetes, such as high plasma glucose levels, polydipsia, and polyuria. Cry1/Cry2-null mice have a malfunctioning renin-angiotensin-aldosterone system linked with saltsensitive
hypertension. Genetic deletion of Rev-erb α largely abolishes the oscillation of brown adipose tissue activity and whole-animal core temperature, suggesting that clock control of brown adipose tissue is a major driver of the circadian rhythm of body temperature.1,2,5
These mutant mice also exhibit defects in cardiovascular function. In humans, a specific haplotype in the Clock gene has been associated with the prevalence of metabolic syndrome. Bmal1 is associated with susceptibility to hypertension and type 2 diabetes, while the expressions of Per2, Cry1, and Bmal1 in adipose tissue are associated with features of the metabolic syndrome in man (Eleanor Scott Lecture).

1. Synchronization of the clock
Desynchronization of the SCN and metabolic alterations

Shift work, lights at night, and jet lag have been associated with increased body mass index, altered plasma lipids, and altered glucose metabolism. In addition, changes in the amplitude and in the synchrony of circadian rhythms have been observed in
ageing. Light is the main cue that induces the resetting of the core circadian clock to external variations. Different protocols manipulating the environmental light conditions have been applied to rodent models to explore the effects of the light environment on metabolism. Repeated shifts in the light-dark cycle might mimic shift work, while continuous exposure to light induces a desynchronization of the SCN pacemaker
activity, such as in ageing. In rodents, both protocols reduce the rhythmicity and the period of the electrical activity of the SCN (Figure 5). In parallel, continuous light exposure or shifts in the light-dark cycle increase visceral adiposity and reduce the activity of brown adipose tissue. Moreover, such changes in light environment accelerate the development of diabetes in the human islet amyloid polypeptide transgenic
(HIP) rat, which is a diabetes-prone rodent model. The desynchronization of neuronal activities in the SCN decreases the amplitude of the circadian rhythm and is associated with metabolic alterations (Figure 5).

The exposure of animals to a standard light-dark cycle with the introduction of dim light at night (dLAN) mimics the widespread increase in the use of artificial lighting at night in modern society. This exposure alters food intake and reduces glucose tolerance in mice; more precisely, it reduces the amplitude of PER1 and PER2 levels in the SCN and the rhythmicity of clock-controlled genes in the liver. The alteration of liver gene expression by specific light exposures seems to be induced by both hormonal and neural pathways. Interestingly, these metabolic alterations are prevented when mice are fed exclusively in the dark period, suggesting that the metabolic alterations are caused by a desynchronization between food intake and behavioral activity as a result of light exposure at night.
Exposure to shift-work or jet-lag paradigms in mice results in metabolic alterations, which depend on diet. Mice fed a high-fat diet and exposed to these paradigms reduce their daily locomotor activity and food intake, which results in an increased body weight gain. Mice fed a high-fat, high sucrose diet and exposed to these paradigms have no change in their body weight or locomotor activity, but develop hypercholesterolemia,
hyperglycemia, and glucose intolerance. Shift-work paradigms induced by forced activity in the resting phase lead to a circadian misalignment: food is taken during the active phase, whereas hormonal rhythms are not affected. This misalignment occurs in parallel with an increase in body weight and abdominal fat mass and a reduction glucose tolerance. It induces hepatic desynchronization characterized by different circadian phases of clock genes and metabolic genes (Johanna Meijer lecture).

Mechanisms involved in the synchronization of the SCN and peripheral clocks.

Peripheral clocks are not induced by light, but are regulated by peripheral signals controlled directly or indirectly by the SCN. The SCN controls the rhythmicity of peripheral clocks. The development of specific devices using bioluminescence to visualize the oscillations of clock genes in vivo can pinpoint this regulation. Indeed, lesion of the SCN induces the loss of phase coherence of PER2 between tissues. However, the rhythmicity of Rev-erb α is maintained in the liver. This synchrony has been shown to involve glucocorticoids, whose plasma concentration oscillates in a daily rhythm. Glucocorticoids induce the synchronization of circadian rhythms in cell culture and phase shifts in the liver, independently of a resetting of the SCN phase. Glucocorticoid signaling may be one of the more direct pathways used by the SCN to synchronize liver clocks. At the molecular level, this regulation may involve the regulation of Per1
through direct binding of the glucocorticoid receptor (GR) on its promoter. Interestingly, the absence of GR in the SCN might explain the resilience of this nucleus to the entrainment induced by glucocorticoids (Figure 6).
While in vitro experiments demonstrate that a large variety of redundant metabolites and signaling pathways are able to synchronize circadian clocks in individual cells, in vivo experiments confirm that feeding is the main timing cue for the synchronization of peripheral clocks. Inverting the feeding and activity schedules uncoupled the peripheral clocks (aligned with the feeding phase) from the central SCN pacemaker
(aligned with the light phase). Cyclic metabolites, such as NAD+, are plausible timing cues for peripheral tissues. NAD+ controls the activity of deacetylases such as sirtuins and poly [ADP-ribose] polymerase 1 (PARP1) (Figure 1 and 6). The activities of both enzymes exhibit daily oscillations and regulate the transactivation capacities of CLOCK:BMAL1. Whereas changes in Sirt1 expression alter BMAL1 circadian amplitude,
the involvement of this protein in the entrainment of the peripheral clock by feeding has not yet been demonstrated. In contrast, deletion of Parp1 moderately affects the amplitude of circadian gene oscillations and delays the phase adaptation to inverted feeding rhythms. The real-time recording of circadian oscillations highlights the involvement of the SCN in the feeding-induced phase shift of peripheral clocks. Indeed,
in SCN-lesioned mice, liver clocks change their phase more rapidly than those in intact animals, suggesting that signals sent by the SCN slow down the phase change induced by feeding rhythms. The precise molecular nature of the signals controlled by feeding rhythms is still unknown, but the signals seem to be more related to the fasting period than the feeding period. This would be in keeping with the role of AMPK in phase
resetting, through the phosphorylation and destabilization of CRY proteins. (Figure 1)
Body temperature is also a strong Zeitgeber for peripheral clocks. In spite of the small amplitude in temperature cycles, mimicking daily body temperature rhythms can drive clock oscillations in cells. This regulation involves the heat shock factor 1 (HSF1) protein and the old-inducible RNA binding protein (CIRP). Whereas HSF1 levels are maximal at the highest temperature, CIRP levels are maximal at the lowest; regulation occurs at the posttranslational level. Indeed, temperature has been shown to control the activity of the spliceosome by enhancing the splicing efficiency of CIRP pre-mRNA (Figure 6, Ueli Schibler lecture).

2. Mechanisms coupling the clock and metabolism

Redox cycles coupled to energy metabolism highlight the interplay between clock and cell metabolism. In addition to the molecular clock, redox couples—such as NADP+:NADPH, glutathione:glutathione disulphide (GSH:GSSG), and cysteine:cystine (Cys:CySS)—undergo circadian oscillations. These couples are linked to energy metabolism by their regulation of many biochemical reactions, and their oscillations may be driven by food intake or by the molecular clock. The deletion of Bmal1 results in an increase in ROS production and premature ageing, which is reversed by the administration of antioxidants. In fly and mouse, the molecular clock seems to control glutathione
synthesis. Finally, several examples illustrate that redox state may control the transcriptional activity of clock repressors (PER/CRY) or factors involved in peripheral clock synchrony (GR) (Akhilesh Reddy Lecture).

NAD+ and mitochondrial respiration

Local organ clocks impact metabolic function by anticipating the varying requirements for anabolic and catabolic processes across the daily fasting-feeding/sleep-wake cycle. Using tissue-specific invalidation of clock genes, the role of the clock in energy and glucose homeostasis in different tissues (particularly in the liver, muscle, and pancreas) has been defined. Surprisingly, circadian mutant mice do not resist prolonged fasting.

The metabolism of these mice cannot adapt in response to glucose deprivation or to oxidative damage and cell stress due to a lack of NAD+ bioavailability. More precisely, circadian disruption induced by the deletion of Bmal1, Cry1, or Cry2 impairs electron transfer from lipids to the TCA cycle in parallel with increased ROS production, which increases sensitivity to genotoxic stress. This mitochondrial failure has a profound effect on respiration and exercise tolerance in muscle, but can be rescued by NAD+ supplementation. Controlling NAD+ levels might thus be a key mechanism by which the clock regulates fuel switching from glycolytic to oxidative substrate. Indeed, the liver and myoblasts exhibit an autonomous rhythm of oxygen consumption, glucose oxidation, and mitochondrial lipid catabolism that is directly linked to an autonomous rhythm of NAD+ metabolism (Joseph Bass Lecture).

Rev-erb α

Rev-erb α belongs to the repressor arm of the clock, which inhibits transcription through the recruitment of NCOR and HDAC3 to target promoters. However, several lines of evidence illustrate the role of Rev-erb α in mediating the crosstalk between circadian rhythms and tissue-specific metabolism. The synthesis of heme, the Rev-erb α substrate, is controlled by clock components, which are inhibited by Rev-erb α. In brown adipose tissue, the decrease in Rev-erb α levels induced by cold temperature exposure facilitates the maximal induction in Ucp1 and thermogenesis. In accordance with a major role of Rev-erb α in the regulation of circadian rhythm in brown adipose tissue, the oscillation of brown adipose tissue activity and whole-animal core temperature is abolished in Rev-erb α-/- mice. The reversible induction of hepatic lipogenic genes prior to an animal waking and feeding is primarily controlled by the derepression of genes targeted by the complex Rev-erb α/NCOR-HDAC3. Rev-erb α also controls the rhythmicity of cholesterol metabolism through the sterol regulatory element-binding protein (SREBP) and bile acid synthesis through the regulation of cholesterol 7 alpha-hydroxylase (Cyp7a1). In the liver, Rev-erb α coordinates reversible changes in the liver epigenome through the recruitment of the complex NCOR-HDAC3, but also contributes to liver gene transcription via its association with liver-specific transcription factors, such as hepatocyte nuclear factors 6 (HNF6) and 4 (HNF4). Rev-erb α thus uses different pathways to control the molecular clock (through direct competition with ROR) and metabolism (indirectly by binding to tissue-specific transcription factors) (Figure 7). In muscle, Rev-erb α controls lipid utilization, at least in part, through the regulation of lipoprotein lipase as evidenced in Rev-erb α-/- mice. Rev-erb α also impacts exercise capacity and oxidative metabolism by interfering with the LKB1-
AMPK-SIRT1-PGC-1α axis (Mitchell Lazar Lecture).

Peroxiredoxins : the biochemical loop of the clock Daily redox rhythms also take place in cells unable to perform transcription. The antioxidant proteins peroxiredoxins (PRDXs) undergo such circadian oscillations. These proteins are involved in hydrogen peroxide metabolism and signaling and are expressed in almost all aerobic species. The circadian oscillations of the 2-Cys PRDXs can occur in nontranscriptional systems, persist in constant conditions or over a range of physiological temperatures, and have the ability to entrain the clock. PRDXs thus exhibit hallmarks of biochemical circadian oscillators. Interestingly, PRDXs are conserved during evolution and still cycle when clock genes are deleted in mice, suggesting a parallel evolution of the transcriptional components and the biochemical components of
the clock. However, both components are tightly coupled in the regulation of the clockcontrolled genome. For example, PRDX3 has been linked to the rhythmic production of corticosterone from the mouse adrenal cortex (Akhilesh Reddy Lecture).

“The Islet and Metabolism Keep Time”
I. The circadian system
II. Circadian regulation of the transcriptional network
III. Circadian rhythms and metabolism
IV. Clock and the islet
V. Conclusion