Eckel-Mahan and Sassone-Cors PhysRev 2013_circadian Clock

8/10/2019 Eckel-Mahan and Sassone-Cors PhysRev 2013_circadian Clock

1/29

METABOLISM AND THE CIRCADIAN CLOCK CONVERGEKristin Eckel-Mahan and Paolo Sassone-Corsi

University of California, Irvine, California

LEckel-Mahan K, Sassone-Corsi P. Metabolism and the Circadian Clock Converge.Physiol Rev 93: 107135, 2013; doi:10.1152/physrev.00016.2012.Circadianrhythms occur in almost all species and control vital aspects of our physiology, fromsleeping and waking to neurotransmitter secretion and cellular metabolism. Epidemio-logical studies from recent decades have supported a unique role for circadian rhythm

in metabolism. As evidenced by individuals working night or rotating shifts, but also by rodentmodels of circadian arrhythmia, disruption of the circadian cycle is strongly associated withmetabolic imbalance. Some genetically engineered mouse models of circadian rhythmicity areobese and show hallmark signs of the metabolic syndrome. Whether these phenotypes are due to the loss of distinct circadian clock genes within a specic tissue versus the disruption of rhythmicphysiological activities (such as eating and sleeping) remains a cynosure within the elds ofchronobiology and metabolism. Becoming more apparent is that from metabolites to transcriptionfactors, the circadian clock interfaces with metabolism in numerous ways that are essential formaintaining metabolic homeostasis.

I. INTRODUCTION 107II. SYSTEMS HOMEOSTASIS AND... 111III. CIRCADIAN RHYTHMS IN METABOLIC... 114IV. MOLECULAR RHYTHMS OF... 124 V. FROM METABOLITES TO PHYSIOLOGY 129

I. INTRODUCTION

Circadian rhythms control a wide variety of physiologi-cal events, including metabolism, in all organisms. In-grained in our modern life-style is the exibility to eat,sleep, socialize, and exercise around the clock, yet theseallowances correlate with rising metabolic disorders andobesity. Increasingly evident is that metabolic homeosta-sis at the systems level relies on accurate and collabora-tive circadian timing within individual cells and tissues of the body. At the center of these rhythms resides the cir-cadian clock machinery, an incredibly well-coordinatedtranscription-translation feedback system that incorpo-rates a changing landscape of mRNA expression, protein

stability, chromatin state, and metabolite production,utilization, and turnover to keep correct time. Recentndings show that regulation of metabolism by the cir-cadian clock and its components is reciprocal. Speci-cally, components of the circadian clock sense alterationsin the cells metabolism. Understanding more fully theties that exist between cellular metabolism and the circa-dian clock will provide not only needed insights aboutcircadian physiology, but also novel approaches regard-ing both pharmacological and nonpharmacologicaltreatment of metabolic disorders.

A. A Brief Overview of Circadian Physiology

The word circadian derives from the Latin circa (around)and dies (day). Oscillations of 24-h periodicity are re-ferred to as circadian (FIGURE 1 A ). The molecular and phys-iological oscillations discussed here may vary in amplitudeand even phase; however, in common is their 24-h peri-odicity (or tau, ), which temporally follows the earthsrotation around its axis. Zeitgebers (the German word for

time givers) are signals that help synchronize the bodyscircadian clock with the environment (TABLE 1) . In mam-mals, light is one such zeitgeber and functions as such byactivating a small region of the hypothalamus located justabove theoptic chiasm. Lesion studies identied this region,the suprachiasmatic nucleus (SCN), as being light respon-sive as lesions within this area abolished rhythmic circadianbehavior in both locomotion and food consumption (224).The SCN is located in the anterior hypothalamus and iscomprised of a meager 20,000 neurons, which receive pho-tic information from the environment via neurons tran-scending from the retina via the retino-hypothalamic tract.The circadian system is organized hierarchically, meaningthat while molecular oscillations occur in most cells andtissues of the body, the SCN functions as the master regu-lator, synchronizing the phase of other oscillating tissues(95, 204). Its superiority in this sense has been demon-strated in grafting experiments; SCN grafts from wild-type(WT) mice into a genetically arrhythmic mouse can restorenormal circadian rhythmicity (227). Examples of SCN-di-rected circadian rhythms include feeding, the sleep-wakecycle, glucose metabolism, insulin secretion, and even learn-ing and memory (95). Various techniques have been em-ployed to measure and depict the properties of these

Physiol Rev 93: 107135, 2013doi:10.1152/physrev.00016.2012

1070031-9333/13 Copyright 2013 the American Physiological Society
http://-/?-http://-/?-


2/29

rhythms including the simple actogram. Actograms graph-ically portray these rhythms, plotting activity in such a waythat both the phase and period of the oscillation can beobtained. FIGURE 1 B shows the actogram of a nocturnalanimals activity during the 24-h day, with dark shadingrepresenting their activity (in this case, wheel running) andblank areas reecting the animals rest or sleeping period.

Central to the molecular rhythmicity of SCN neurons aswell as other oscillating cells are transcription factors thatdrive expression of their own negative regulators (204).This property of the clock results in a negative transcrip-tional and translational feedback loop that perpetuates os-cillations in gene expression that occur every 24 h (FIGURE1 C ). This program is highly conserved across species. FIG-

A

C

D

B

Period(or tau, )

Activators

Kai (A,B,C)

***

*** The bacterial oscillator depends on rhythmic KaiC phosphorylation, rather than thetranscriptional feedback regulation provided by the other circadian clock systems.

Negative

Positive

Bacteria

CCA1, LHY

TOC1

Plant

FRQ

WC-1, WC-2

Neurospora

PER, TIM

dCLK, CYC

Drosophila

PER, CRY

CLK, BMAL1

Mammals

Repressors

Feedback

loop

Amplitude

Phase

mRNA

++

FIGURE 1. Basic characteristics of circadian rhythms. Circadian rhythms have a period of 24 h. A: theamplitude of various biological processes, such as eating, locomotion, gene expression, etc., vary considerablyacross organisms and physiological events, as does the phase of the rhythm. B : actograms are used to depict the rhythm of an organism over the 24-h cycle and typically consist of digitized activity values that are presentedas a double plot, where a line contains data for that day as well as the proceeding day. In this example, the redarrow denotes the activity of a mouse in a 12-h light/12-h dark cycle, while the blue arrow denotes the activityof the animal as a result of a switch to constant darkness (DD). Such activity in DD is referred to as freerunning. C : the circadian clock depends in part on negative transcriptional/translational feedback mecha-nisms in which proteins participate in the production of their own negative-feedback regulators. D : thepositive and negative regulators in the core clock system. Positive factors TOC (Timing of Cab Expres-sion1), WC-1 (White Collar-1) and WC-2 (White Collar-2), CLK (CLOCK), CYC (Cycle), CLK, BMAL1 (Brainand Muscle Arnt-like Protein-1) regulate the transcription of their own negative regulators, CCA1 (Circa-dian Clock-Associated1), LHY (Late Elongated Hypocotyl), FRQ (Frequency), PER (Period), TIM (Timeless)and PER, CRY (Cryptochrome). The bacterial system does not rely on a negative transcriptional feedback mechanism but rather relies on three Kai proteins. The rhythmicity governing this system relies onrhythmic phosphorylation and can be recapitulated in vitro by the combination of the three Kai proteins andATP (164).

KRISTIN ECKEL-MAHAN AND PAOLO SASSONE-CORSI

108 Physiol Rev VOL 93 JANUARY 2013 www.prv.org


3/29

URE 1 D demonstrates the corresponding positive- and neg-ative-feedback proteins of this molecular system across sev-eral species. With the exception of the cyanobacteria circa-dian clock, the proteins in this table participate in thenegative transcriptional translational feedback loop de-scribed. In mammals, two basic helix-loop-helix (bHLH)transcription factors, CLOCK and BMAL1 (also known asMOP3 or ARNTl), heterodimerize and subsequently bindto conserved E-box sequences in target gene promoters. Inthis way, they drive the rhythmic expression of mammalianPeriod (Per1 , Per2 , and Per3 ) and Cryptochrome (Cry1 andCry2 ) genes (FIGURE 2) . PER and CRY proteins form acomplex that translocates back to the nucleus to inhibitCLOCK:BMAL1-mediated gene expression. This deceiv-ingly simple transcriptional feedback loop is regulated byhighly complex mechanisms (such as posttranslationalmodication of circadian proteins and additional interlock-ing feedback loops of gene transcription) that mandate ne-tuning of the clock and yet provide for it plasticity by whichit can adjust to changes in the environment (21, 38). Whilethe clock genes are necessary for circadian physiology, a

number of clock controlled genes (CCGs) also contribute.These are genes that are regulated by the circadian clock(and therefore oscillate with 24-h periodicity) but are notnecessarily conserved across tissues.

While the view that this transcriptional and translationalfeedback loop is essential for timekeeping in cells has heldpreeminence, recent studies may revise the classical view of how rhythmicity is maintained within the cell. New datareveal that circadian rhythms can persist even in theabsenceof transcription. Specically, the posttranslational modi-

cation of some proteins can occur in a transcription-inde-pendent manner (167, 168). This has been demonstrated byrhythmicity in the oxidation of peroxiredoxin proteins, forexample, which is temperature compensated and entrainedby zeitgebers, two fundamental qualities of circadianrhythms. While these transcription-independent oscilla-tions occur, they do appear to exist in collaboration withthe classical transcriptional loops described, as nucleatedmouseembryonicbroblasts from circadian arrhythmic an-imals show altered but present rhythmic peroxiredoxin ox-idation. Thus there appears to be direct interaction betweennuclear and cytoplasmic circadian rhythms when there is anucleus present (168). Interestingly, the circadian oxida-tion-reduction cycles of peroxiredoxin are very highly con-served, even more so than the conservation that exists forthecircadian clock components of FIGURE 1 D (62). The ideathat oscillations can persist in the absence of a nucleus isimportant because it means that there are likely numerouszeitgebers for the clock that are yet unidentied as such. Infact, the metabolome in its entirety may need to be ad-dressed as possible zeitgebers and may feed into the clock

system in specic ways, affecting phase, amplitude, or pe-riod of existing transcriptionally dependent oscillations.The concept of metabolites as critical modulators of thecircadian clock will be further addressed in section IV B.

Circadian rhythmicity can be seen in many physiologicalprocesses. Examples include body temperature, activity,sleep, metabolism, heart rate, blood pressure, and hormoneand neurotransmitter secretion (95). This being the case, itis perhaps not surprising that many cellular processes, in-cluding gene expression, oscillate within the cell. Approxi-

Table 1. Denitions for common circadian terminology and the core clock components of the mammalian circadian clock

Term Acronym Denition

Circadian Refers to the 24-h nature of an event. Circadian is derived from the Latin roots circa(about) and diem (day).

Entrainment Adaptation over time to an imposed cue such as light or food.Zeitgeber ZT German word meaning time giver. A cue (such as light or food) that entrains the

circadian clock. ZT refers to zeitgeber time.Suprachiasmatic nucleus SCN A small region of the anterior hypothalamus consisting of bilateral nuclei that

coordinately synchronize rhythmicity within other tissues.Circadian clock-controlled

geneCCG A gene expressed in a circadian-dependent manner, usually under the control of a

promoter that contains an E-box, D-box, or RRE.Brain and muscle Arnt-like

protein-1BMAL1 The mammalian bHLH-PAS transcription factor that dimerizes with CLOCK and NPAS2

to activate gene transcription.Circadian locomotor output

cycles kaputCLOCK The mammalian bHLH-PAS transcription factor that dimerizes with BMAL1 to activate

promoters that contain E-boxes (CACGTG).Cryptochrome proteins CRY Transcriptional repressors that dimerize with PER to inhibit CLOCK:BMAL1-mediated

gene transcription. In plants and invertebrates, these function as light-responsiveavoproteins.

Neuronal PAS domainprotein 2

NPAS2 A transcription factor similar to CLOCK and highly expressed in the forebrain. NPAS2dimerizes with BMAL1 to activate gene transcription.

Period homolog proteins PER PAS domain containing proteins which dimerize with TIM (in Drosophila ) or CRY (in

mammals) to inhibit CLOCK:BMAL1-induced gene transcription. The Drosophila Periodgene was the rst circadian gene discovered.

METABOLISM AND THE CIRCADIAN CLOCK CONVERGE

109Physiol Rev VOL 93 JANUARY 2013 www.prv.org


4/29

mately 10% of gene transcripts oscillate in the cell; how-ever, a much larger percent of the proteome oscillates ineither expression or activity (25, 57, 149, 180, 190). Thelarge number of transcripts that oscillate in the cell dependsin part on circadian changes in chromatin remodeling, per-petuated by rhythmic alterations in histone modicationssuch as phosphorylation or acetylation (161). Histoneacetyltransferase and deacetylase activity at oscillatinggenes serves as a preamble for much of the rhythmic geneexpression in vivo. Some of the chromatin modifying en-zymes responsible for this activity may associate directlywith circadian clock proteins (6, 162). CLOCK itself con-

tains histone acetyltransferase activity which contributes tothe rhythmic acetylation of K14 of histone H3 and K537 of its binding partner, BMAL1 (99). Recently, metabolitesthat fuel these rhythmic events have been observed to oscil-late, sometimes producing rhythmic enzymatic activity evenwhen the expression of a substrates enzyme doesnt oscil-late (163, 189). As more metabolites that contribute toclock maintenance are being identied, it is becoming evi-dent that metabolism and circadian rhythmicity are inter-twined molecularly and that one process cannot be ade-quately studied in isolation of the other. While much re-

mains to be revealed in terms of their interactions, howoscillators in various compartments, from cells to tissues,interact to control metabolic physiology is beginning to beunderstood more fully.

B. Metabolic Homeostasis

Western (and what we generally consider as westernized)societies are experiencing a dramatic rise in metabolic dis-orders (14, 32, 61, 222). This rise is not limited to the adultpopulation, as evidenced by an increase in overweight chil-

dren and adolescents (222). The production of high-energyfoods that are nutritionally wanting as well as the decreasein energy expenditure often associated with jobs that allowa sedentary life-style have certainly contributed to the oc-currence of metabolic disorders. Genetics also contributesto alterations in metabolic function (88). In mammals, bothenvironmentally induced circadian disruption as well as ge-netic aberrations in circadian clock machinery can lead tometabolic disorders (reviewed in Ref. 75). Why circadiandisruption has this effect on metabolic homeostasis appearsto be complex.

CRY PER CRY PER

Degradation

BMAL1

BMAL1

AMPK CKI

P P

Rev-erb Rev-erb

Bmal1

ROR

Bmal1

E-Box

CLOCK

BMAL1

E-Box

CLOCK

FIGURE 2. Additional inputs to the core circadian loop. The basic positive and negative transcriptionalfeedback loop of the circadian clock is elaborate, with external loops and posttranslational modications, suchas phosphorylation contributing to maintenance of the core oscillatory players. 5 = AMP-activated protein kinase(AMPK) and casein kinase I epsilon (CKI ) contribute to phosphorylation and degradation of the CRY and PERproteins, respectively, thus regulating the negative-feedback potential of these proteins on the CLOCK:BMAL1complex. The transcription of Bmal1 is negatively regulated by one of its own gene targets, Rev-erb , theprocess of which controls the amount of BMAL1 protein available for CLOCK binding. Retinoic acid receptor-related orphan receptor alpha (ROR ) exerts opposite effects to REV-ERB on the Bmal1 promoter.


http://-/?-http://-/?-


5/29

Energy balance is when energy (food) intake is equal toenergy expenditure. Major sources of metabolic fuels in-clude glucose, fatty acids, and ketone bodies. If not con-sumed, metabolic fuels are converted into metabolic storesincluding liver glycogen, muscle protein, and the triglycer-ides found in adipose tissue. Energy expenditure is a com-posite of three components: basal metabolic rate (energy tokeep the nervous system working and the vital organs func-

tioning properly), thermogenesis (the energy required forthe absorption and storage of food), and physical activity(the most variable). Whether one is a 1.58 m, 86.2 kg indi-vidual or a 1.85 m, 72.6 kg individual, the body is remark-able in its ability to maintain energy equilibrium.

While genetics indubitably contributes to the bodys abilityto maintain energy balance (61, 88, 98), numerous studieshave revealed that environmental cues including those af-fecting ones circadian rhythms also play essential roles inmetabolic homeostasis (71, 97). With modern technology(which makes provisions such as exogenous lightingthroughout earths 24-h day), night shift and rotating shiftwork have become increasingly common. Such alterationsbeyond the conventional workday alter not only temporalaspects of work and social activities but also physiologicaland molecular rhythms in the body. Compelling evidencethat circadian rhythm disruptions contribute to metabolicdisorders has been observed in experiments centered onnight shift and rotating shift workers. For example, an as-sociation between circadian disturbance and cardiovasculardisease, increased body mass, and elevated plasma glucoseand lipid levels has been observed in humans subjected tonighttime shift work (118, 120, 182, 234). Rodent studiesalso support this link as simply altering their normal light-

dark cycle to one in which dim light replaces the normaldark period causes changes in metabolism that are observedon a physiological level (70). The integration of these pro-cesses at the cellular and systems level will be discussedfurther in later sections.

Rhythms in energy intake must coincide with endogenousuctuations in gene expression for metabolic homeostasisat the cellular level to occur. At the cellular level, energyfrom our diet is used to generate macromolecules such asproteins, DNA, membrane components, and polysaccha-rides. As energy intake is a circadian activity, it drives uc-

tuations in the rate of these activities in different tissues.Furthermore, many of the oscillating gene transcripts in agiven tissue are tissue specic or at least oscillate in a tissue-specic fashion (226). Tissues thereby meet their metabolicdemands using regulatory molecules that are temporally orspatially distinct from other tissues. For example, manyhormones that control food intake including insulin, gluca-gon, peptide YY, GLP-1, corticosterone, leptin, and ghrelindepend on energy intake or oscillate in a circadian manner,and are largely secreted in a tissue-specic fashion (re-viewed in Ref. 75). The contrasting prole of nuclear recep-

tor expression across different tissues also emphasizes thetissue-specicnature of metabolic gene expression. Numeroustranscription factors, including metabolic nuclear receptors,display 24-h periodicity (250). These includereceptors suchasFXR, LXR, HNF-4 , PPAR , PPAR , NUR77, and manyothers ( www.nursa.org/10.1621/datasets.02001 ) that aremore highly expressed in some metabolic tissues over others.

In summary, evidence that circadian and metabolic pro-cesses interact at the cellular levels is gaining strength andmay help provide molecular explanations for the growingnumber of metabolic disorders associated with circadiandisruption. Studies addressing this link at both the systemand cellular levels are being performed and are illuminatingsome surprising effects of the clock on physiology.

II. SYSTEMS HOMEOSTASIS ANDCIRCADIAN RHYTHMICITY OFPROCESSES INVOLVED IN ENERGY BALANCE

A. Feeding Is a Circadian Rhythm

The circadian regulation of energy intake is consistentacross mammals. Mammals tend to consume the vast ma-jority of food and water during their waking period. Thisoccurs during the day for diurnal mammals and during thenight for nocturnal animals and coincides with food-seek-ing activity. In mammals, feeding is under homeostatic con-trol and involves humoral factors acting on hypothalamicneurons that ultimately control the urge to feed or not tofeed. This is accomplished via endocrine molecules includ-ing leptin, ghrelin, and peptide YY acting on neurons withinthe hypothalamic arcuate nucleus (63). Neurons of theparaventricular nucleus both sense and integrate the orexi-genic and anorexigenic signals that result from these hor-mones (reviewed in Ref. 181). These and other biologicalsignals that control energy intake will be discussed in moredetail in section III A2. Early lesion studies identied theSCNs contribution to circadian rhythmicity in eating anddrinking as ablation of the SCN destroys rhythmicity inboth eating and drinking (224). While it has been arguedthat early lesion studies may have involved tissue injury thatextended beyond the SCN to other hypothalamic circuits,other models of circadian arrhythmicity also appear to be

arrhythmic in energy intake when fed ad libitum (237). Tosome extent then, feeding is controlled directly or indirectlyboth by circadian and homeostatic processes.

Feeding that follows a typical pattern of daytime eating fordiurnal organisms or nighttime eating for nocturnal organ-isms seems to be important for metabolic homeostasis. Onerecent study addressing the role of the circadian clock inmetabolic function showed a somewhat surprising result:simply replacing the dark period of a rodents circadiancycle with very dimlight produced a pronounced increase in


http://www.nursa.org/10.1621/datasets.02001http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://www.nursa.org/10.1621/datasets.02001


6/29

the rodents body mass (70). Using a 16-h light/8-h darkcycle paradigm for control animals, the experimenters ex-posed some mice to a 16-h light/8-h dim light or a consti-tutively light paradigm. Interestingly, exposure to dim lightat night caused increased energy intake during the daytimehours compared with animals in normal conditions, al-though total energy intake was unchanged between the twogroups. An increase in body mass was detected in both

experimental groups after only 1 wk of exposure to the newlighting paradigm, and this increase continued throughoutthe 8-wk program in the new lighting conditions. Epididy-mal fat pads were enlarged in experimental groups, indicat-ing that changing the lighting conditions had an effect onwhite adipose deposition. As locomotion, corticosteronelevels, and total 24-h energy intakewere allunaltered in oneor both of the experimental groups, only the increased per-centage of energy intake that occurred during the day couldaccount for the increased body mass observed in these ro-dents. When dim light-exposed animals were restricted toonly nighttime feeding, the body mass alterations were pre-

vented. While this study presents evidence that the normalcircadian rhythmicity in food intake is important for meta-bolic homeostasis in rodents, the results support humanstudies in which night-eating syndrome (NES) is associatedwith obesity and circadian misalignment (170, 202, 231).

NES and sleep-related eating disorder (SRED) are two re-lated eating disorders that include consciousor unconsciousfood consumption at night. They are typically associatedwith morning anorexia and evening hyperphagia as well asperpetual awakenings during the night that often involveeating (reviewed in Ref. 169). NES-inicted individuals

may wake from one to four times at night, and 74% of allawakenings are associated with food consumption (170).NESpatients show a delayed melatonin rhythmand a phaseadvance of orexigenic ghrelin (which is released predomi-nantly by thegastrointestinal tract) rhythms. Theamplitudeof ghrelin rhythms is also decreased in NESindividuals (81).While plasma glucose levels are antiphase to control humansubjects, insulin levels are phase delayed and are also con-siderably reduced in amplitude in NES-aficted subjects.As NES reects a phase delay in the acquisition of food,potential circadian mechanisms underlying the diseaseare under consideration. Sertaline, a selective serotoninreuptake inhibitor (SSRI), has been shown to reduce theamount of nighttime calories consumed in NES-afictedindividuals and may do so by altering rhythms at the levelof the SCN (171). It is still unclear whether NES individ-uals represent a situation in which uncoupling of periph-eral oscillators with the central clock occurs or whetherindividual peripheral oscillators are functioning out of phase with each other independent of central clock co-herence. NES-aficted individuals, however, supportother evidence that implicates clock disturbances withmetabolic disorder.

B. Food as a Zeitgeber, a MetabolicCircadian Cue

While originally thought to be limited to the brain, theoccurrence of circadian rhythms has been noted throughouttissues of the body. In fact, most tissues studied to dateshow robust oscillations in gene expression (254). There arefew known zeitgebers, however, that can entrain the circa-dian clock in vivo. Those identied include light and, morerecently, food. As a light pulse during the subjective nightcan phase advance or delay the SCN circadian clock, foodcan function as a potent zeitgeber for peripheral tissues,underscoring the important relationship between circadianand metabolic processes. The circadian rhythm of the SCN,which responds robustly to light, is largely unaffected bychanges in feeding patterns, while oscillations within someperipheral tissues, such as the liver, appear to be dependenton communication with a rhythmic brain but principally onthe feeding cycle. For diurnal animals, like humans, feedingoccurs during the day, whereas for nocturnal organisms,food intake takes place predominantly at night, when they

are awake and active. Feeding and circadian rhythms ingene expression are so tightly linked that, when food isrestricted to a precise time of the day, the expression of alarge number of hepatic genes is altered, with a new rhythmthat follows that of the feeding cycle (37, 225). One studydemonstrating the zeitgeber property of food involved micethat were restricted to 4 h of feeding during their rest pe-riod. As early as 2 days after restricted feeding, the livershowed a 10-h phase change in circadian rhythmicity. Anal-ysis of several tissues from animals that were fasted follow-ing a week of restricted feeding showed that not only theliver but also the lung had been entraining to the new re-stricted feeding schedule. Conversely, in a liver which lacksnormal circadian rhythmicity, the majority of hepatic geneexpression rhythms can be restored by exposure to a tem-porally restricted feeding schedule (237). Technically, foodcan entrain both the periphery and the brain, as a free-running animal that is exposed to a restricted feeding sched-ule can show signs of SCN entrainment (as measured bywheel running, for example) that may be caused directly orindirectly by food and the anticipatory activity involved inits administration (reviewed in Ref. 156). The restriction of feeding to a few hours during an animals normal restingperiod results in food anticipatory activity (FAA), in whichlocomotion patterns deviate from normal circadian cycles

(64). Such enhanced activity during the rest period serves asa preamble to the feeding event, indicative of the animalsanticipation of food consumption. While still under inves-tigation, a functional clock in the dorsomedial hypothala-mus (DMH) has been implicated in FAA (77). Bmal1knockout animals and Clock mutant animals, both of which show complete arrhythmia, however, can still showFAA under some experimental conditions (183, 185), so itis unclear what role the circadian clock might play in thecentral nervous system for FAA to occur. Entrainment tofood can also occur in rodents with SCN lesions (127, 223),


http://-/?-http://-/?-http://-/?-http://-/?-


7/29

indicating that the neuronal locations governing FAA are atleast partially distinct from those which participate in lightentrainment. Interestingly, ghrelin seems to stimulate theappetitive and consummatory aspects of food intake duringfeeding restriction, and ghrelin receptor knockout animalsshow a reduction in FAA (141). While the arcuate nucleus isone of several ghrelin-expressing regions of the brain thatcould account for the change in FAA in ghrelin receptor

knockout animals, other molecules in satiety centers alsocontribute to this process. The melanocortin 3 receptor, forexample, is highly expressed in food-responsive regions of the hypothalamus and responds to anorexigenic hormonesso as to lower energy intake and increase energy expendi-ture. Melanocortin 3-receptor knockout mice have in-creased adiposity (27) and, interestingly, melanocortin 3knockout mice are immune to FAA (16, 228). Knockout of themelanocortin 3 receptor in mice demonstrates its impor-tance for the circadian timing of food consumption but alsofor entrainment of anticipatory behavior to feeding time.

In summary, feeding is a circadian event. However, it serves

not only as an output of the clock but also as a clock inputmechanism, particularly for peripheral tissues. Because pe-ripheral tissues communicate back to the brain via ghrelin,leptin, glucose, insulin, etc., circadian feeding contributes toan intertwining of theclock and metabolism that appears tobe crucial for metabolic homeostasis.

C. Circadian Rhythmicity of Energy Expenditure

Energy expenditure involves the maintenance of ones basalmetabolic rate, the maintenance of resting metabolic rate,physical exercise, and the maintenance of core body tem-perature. The vast majority of energy expended in an or-ganism goes towards maintaining ones basal metabolicrate, and this directly affects body temperature. Core bodytemperature oscillates in a circadian fashion with a peakthat occurs during the early evening in humans and a troughduring the early morning hours. Human subjects isolated infree running conditions (i.e., in the absence of zeitgeberssuch as light) still show oscillations in core body tempera-ture and sleep/wake cycles (259). Heart rate follows thetemporal prole of body temperature, with beats per min-ute reaching a nadir around ZT3 in human subjects. Inter-

estingly, thermoregulatory changes appear to be tied intothe sleep/wake cycle and are important for the circadianmodulation of sleepiness and propensity to sleep (126). Afraction of spent energy is in the form of physical exercise,and this type of energy expenditure is highly variable acrossthe population. Somehuman performance measures includ-ing muscle strength, anaerobic power output, and joint-exibility follow the circadian pattern of core body temper-ature (reviewed in Ref. 192). Therefore, studies have ad-dressed whether sports performance, for example, orexercise is optimal at specic zeitgeber times. An interesting

concept is that rigorous exercise might be part of the recip-rocal interplay between the circadian clock and metabo-lism. Recentattention hasbeen paid to theidea that exercisemight actually function as a zeitgeber for the body. Thispromises to be an intriguing area for future study.

D. Malfunctioning Clocks and MetabolicDisease

Recent data from night shift workers and individuals withsleep disorders provide evidence that metabolism and circa-dian rhythms are tightly linked in vivo. The implications of this are probably widespread. With the use of the denitionthat shift work is that which occurs during nontraditionalworking hours (usually from 10 p.m. to 6 a.m.), almost20% of workers in industrialized nations are considered tobe shift workers (5). Experiments focused on the effects of shift work indicate that shift workers show an increasedprevalence of obesity and that they gain more weight thanworkers engaged in a conventional workday. In one study,

shift workers showed a higher body mass index (BMI) com-pared with normal shift workers, an increase that did notcorrelate with age or length of time at the nonconventionalshift work (44). Other studies generally support this trend,although in some cases a correlation has been found be-tween length of time at the shift work and BMI (reviewed inRef. 58). One study tracked male, Japanese workers for upto 27 years to assess whether the night shift populationwithin the manufacturing sector showed an increased riskof obesity relative to their male counterparts who workedonly the day shift (128). An increase in obesity was ob-served in night shift workers, and after 10 years of followup, a pronounced risk was observed. A similar study de-signed to assess the effects of alternating shift work on bodyweight showed that job schedule was associated with BMIbut that drinking habits and age were negatively correlatedwith increases in BMI (229). Cross-sectional data accumu-lated from the Vsterbotten intervention program supportsthese results. Data collected from shift workers participat-ing in health surveys at ages 30, 40, 50, and 60 years of ageindicate that obesity is more prevalent in female shift work-ers of all ages. Men in some age groups also showed apropensity for obesity (117). Furthermore, high-density li-poproteins were generally decreased in younger shift work-ers and after adjusting for age and socioeconomic factors, a

high level of circulating triglycerides was a risk factor inboth sexes. These studies clearly demonstrate a correlationbetween night shift work and metabolic disturbance, butmay even underestimate the damaging effects of circadiandysrhythmia as workers having difculty adapting to nightshifts are often moved back to daytime shifts and are there-fore removed from analysis such as these (128).

Unarguably, shift workers have a hard time adjusting com-pletely to a new phase of zeitgebers (198). A new phase of schedule means that the phase of endogenous rhythms will


http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-


8/29

need time to adjust. For example, food consumption gener-ally takes place during the waking nighttime hours in nightshift workers, often resulting in an additional meal con-sumed during the 24-h day. This requires peripheral clocksto adapt and alter the phase of humoral rhythms such asthose of ghrelin. The brain must also adapt to a new sleep-ing schedule during the day, a task that is arduous as someshift workers report (53). Many shift workers complain of

fatigue, jetlag type symptoms, gut disturbances, amongother circadian-related maladies, implying that desyn-chrony might last a very long time, because the internalclocks of the body are not cohesive with the new environ-ment (53). Besides the obvious physical manifestations of fatigue, pain, anddiscomfort, there appear to be further andyet unexplored ramications of an internal clock that is outof phase with its environment.

One concern with this internal and external desynchrony isthat the levels of coronary artery disease associated withshift workers are rising. One study looking at United Statesfemale nurses who worked rotating night shifts (dened asequal to or greater than three nights per month as well asadditional day and evening shifts) reveals that women whoworked 6 years or more of shift work had an increased riskof coronary heart disease after correcting for smokingamong other risk factors (120). A recent summary of 17prior studies focused on shift work and cardiovascular dis-eases (reviewed in Ref. 122) reveals that shift workers had a40% increased risk of cardiovascular disease relative toexclusively daytime workers. It is well known that severalheart-related conditions manifest more frequently at partic-ular times of the day. Ventricular tachycardia, cardiac ar-rest, acute mycoradial infarctions, and myocardial ischemia

have all been reported to manifest in a circadian fashion(reviewed in Ref. 52).

To better understand the mechanisms behind internal syn-chronization, experiments with rodents have been designedto mimic human night shift work. One such study used awheel-running task imposed on nocturnal rats either duringtheir normal sleep phase or during their normal wake phase.Interestingly, plasma glucose rhythmicity was lost entirelyin rodents that worked during their sleep phase and se-rum triglyceride (TAG) levels were reversed from those of control animals, with peaks occurring during the sleep

phase (201). To address whether internal desynchrony waspresent, PER1 and PER2 protein expression was analyzedin the SCN in control and working groups where it wasfound both in phase and of comparable amplitude in theSCN of sleep phase-working animals relative to controls.Corticosterone levels were also invariant, other than aninitial rise when animals began their work phase, regardlessof zeitgeber time. After 34 wk of sleep-phase work, ratssubjected to work during the sleep phase gained moreweight than their counterparts that worked during thewak-ing phase, supporting human studies showing a similar pro-

le of adiposity and weight gain in shift workers. It is likelythat such desynchrony also exists in humans. If pharmaco-logical treatments could be devised to help assist in internaland external synchronization, perhaps some of the negativeeffects associated with shift work might be avoided.

Perhaps a more common form of circadian desynchronycomes in the form of social jetlag, when individuals are not

sleeping within the normal circadian sleep window due totime schedules imposed by work or shool, social events, etc.Recent studies addressing this common problem show thateven social jetlag is associated with obesity (195). Impor-tantly, sleep timing appears to be as important a regulatorof body mass index as sleep duration.

III. CIRCADIAN RHYTHMS IN METABOLICTISSUES

A. How the Brain Talks to the Periphery and Vice Versa

As the synchronizer of rhythms throughout the body, theefferents of the SCN play a paramount role in the circula-tion of humoral factors that target peripheral tissues. ManySCN efferents are contained within regions of the hypothal-amus itself (FIGURE 3) . The core region of the SCN, whichassimilates photic and nonphotic input from the retinohy-pothalamic tract and the raphe, respectively, appears toproject predominantly to the lateral subparaventricularzone (207).Conversely, the shell of theSCN, which receivesinput from the limbic forebrain and other hypothalamicloci, projects predominantly to the medial subparaventricu-lar zone and the dorsomedial hypothalamus (139). In acircuit that includes the paraventricular nucleus and theintermediolateral nucleus in the thoracic spinal cord, theSCN communicates with the pineal gland, using sympa-thetic neurons of the superior cervical ganglion, thus regu-lating the rhythmic secretion of melatonin. Melatoninshows peak expression during the dark phase in both noc-turnal and diurnal animals (157) and directly feeds backinto the clock by targeting receptors in the SCN. The effectsof melatonin on physiology are diverse. For example, mel-atonin appears to have a strong effect on blood pressure,and its receptors are found throughout the blood vessels of many arterial beds (reviewed in Ref. 193). Oscillating mel-

atonin levels probably contribute to the normal circadianvariations in human blood pressure. As ambulatory heartrate is a circadian controlled function, melatonin has beenused to reduce blood pressure in patients with hypertension(203). The decrease in human blood pressures in bothnormo- and hypertensive patients after melatonin adminis-tration is consistent with rodent studies where melatonintherapy is an effective treatment for rats with spontaneouslyhigh blood pressure (230). Melatonin probably providesthis effect, in part, because it relaxes the smooth musclewall.




9/29

The SCN uses the autonomic nervous system, includingboth the parasympathetic and sympathetic systems to reg-ulate theperiphery. Glucocorticoids, which are released in acyclical fashion from the adrenal cortex, are controlled bythe circadian clock. Glucocorticoids oscillate in both anultradian and circadian fashion, with circadian peaks oc-curring during the early morning for diurnal animals andearly evening for nocturnal animals (reviewed in Ref. 46).Adrenocorticotropic hormone (ACTH), which is releasedby the corticotrope cells of the pituitary, shows a similarprole and stimulates the release of the steroid hormonecorticosterone from the adrenal cortex. Light inhibits therelease of corticosterone, and in SCN-lesioned animals,light does not produce the normally observed depression of corticosterone release. Corticosterone production is linkedto the environment indirectly and probably depends on

SCN relays to the paraventricular nucleus (22). The circa-dian regulation of corticosterone is important for physiol-ogy, as corticosterone serves not only as a precursor foraldosterone,but it also functions in rodents and other mam-mals as a glucocorticoid, playing central roles in liver met-abolic function. Interestingly, in SCN-lesioned animals,where at least some of rhythmic liver gene expression is lost,rhythmicity can be restored by the administration of gluco-corticoid receptor activation, suggesting the critical natureof SCN-regulated glucocorticoid release in the regulation of peripheral oscillations (191).

Other humoral factors such as vasopressin and acetylcho-line are also used to communicate to the periphery by theclock in the brain. Perhaps one of the best known oscillatinghumoral factors is arginine vasopressin (AVP). Vasopressinis released into the bloodstream by the pituitary, and itaffects the periphery by regulating blood pressure and bydecreasing water elimination from the kidney during peri-ods of dehydration. Its circadian release by the pituitaryinto the cerebrospinal uid is SCN-dependent (208), but itis also released by the SCN itself where it plays an importantrole in neuronal synchronization within the structure (8,111). Acetylcholine is another oscillating neurotransmitterthat affects the periphery by inducing skeletal muscle con-traction while inhibitingcardiac muscle contraction. In vivomicrodialysis techniques demonstrate that there is a circa-dian release of acetylcholine (113). Acetylcholine serves as

the primary neurotransmitter for preganglionic sympa-thetic neurons, ultimately controlling processes as disparateas heart contraction and gluconeogenesis in the liver. Ace-tylcholine contributes to timekeeping in the central pace-maker as well and can phase shift SCN rhythms via SCNmuscarinic acetylcholine receptors (145). Finally, as dem-onstrated in viral tracking experiments, the SCN uses thesympathetic nervous system to communicate with a num-ber of peripheral tissues. Peripheral injections of pseudo-rabies virus (useful for the analysis of identifying tran-synaptic circuits) in brown adipose and white adipose

Posterior hypothalamic area,tuberomammillary nucleus

Hypothalamic paraventricularand dorsomedial nucleus

Preopticarea

Subparaventricularzone

Paraventricularthalamic nucleus

White adipose tissue

Brown adipose tissue

Thyroid gland

Kidney

Bladder

Adrenal gland

Spleen

Thyroid

Liver

Pancreas

Submandibular gland

Parasympatheticnervous system

Sympatheticnervous system

SCN

FIGURE 3. Main outputs of the SCN as revealed by anterograde labeling experiments. SCN neurons project to numerous other regions of the central nervous system, the large majority of which are hypothalamic. SCNefferents include the paraventricular thalamic nucleus, the subparaventricular zone, the hypothalamic paraventricular and dorsomedial nuclei, the tuberomammillary nucleus, and the medial preoptic area. The SCNcommunicates to the periphery in part via humoral signals released by other tissues of the central nervoussystem and via the sympathetic and parasympathetic nervous system. Some SCN-targeted peripheral tissuesinclude white and brown adipose tissue, the gallbladder, the thyroid gland, the kidney, the spleen, the adrenalgland, the thyroid, the liver, the pancreas, and the submandibular gland (1, 13, 207, 239).




10/29

tissues have revealed that the SCN projects via the sym-pathetic nervous system to these tissues among manyothers (reviewed in Ref. 13).

Therole of communication provided by theperiphery to thebrain in controlling energy homeostasis is paramount be-cause the periphery releases a large number of factors suchas adipokines and hormones that communicate back to the

brain. Such signals provide information to the central ner-vous system that the peripheral demands have been met orneed to be met. Leptin and ghrelin are two such hormones.Ghrelin, which is predominantly secreted by a small frac-tion of cells in the stomach, signals to the brain that thebody needs to be fed. It acts on receptors in the brain in amanner that opposes the actions of leptin and stimulatesfood consumption by driving feelings of appetite and hun-ger. Ghrelin levels plummet after a meal, whereas in antic-ipation of a meal, levels rise again. This oscillatory activityproduces changes in ghrelin levels of approximately sixfoldin the blood (147). Ghrelin is modied by an eight-carbonfatty acid residue that is central to ghrelins effects in the

brain, and the unmodied and octanoylated forms of ghre-lin circulate in the blood (110). Ghrelin is essential forgrowth hormone release and is highly conserved acrossmammals. This is accomplished via hypothalamic arcuateneurons that release growth hormone releasing factor(GHRH), which then activates the release of growth hor-mone from the pituitary. Ghrelin receptors are also presentin the pituitary and can contribute to the release of growthhormone during times of starvation (82). Also released bythe periphery, glucose and insulin play seminal roles in thecentral nervous systemto regulate energy intakeand metab-olism. Recent work on insulin signaling in the brain high-lights neurons of theventromedial hypothalamusas insulin-responsive regulators of diet-induced obesity. While the ab-lation of the insulin receptor in steroidogenic factor 1 (SF-1)cells of the ventromedial hypothalamus (VMH) has no ef-fect on animals fed a normal diet, when challenged with ahigh-fat diet, animals devoid of the insulin receptor in thisregion are partially protected from obesity and show en-hanced peripheral glucose metabolism compared with WTcontrol mice. Insulin appears to induce hyperpolarizationof VMH SF-1 neurons, which reduces their glutamatergicoutput to proopiomelanocortin (POMC)-expressing neu-rons of the arcuate nucleus, thereby inhibiting anorexigenicoutput. When insulin signaling is impaired in this region,

anorexigenic output is enhanced, leading to protectionagainst adiposity.

B. Circadian Rhythmicity Within MetabolicTissues

While once thought to be restricted to the SCN, it is nowclear that peripheral tissues host clocks and engage in pre-cise and yet entrainable timekeeping necessary for circadianphysiology. As measured in rodents using luciferase activity

driven by the Per2 promoter, most tissues show circadianrhythmicity that dampens over time in the absence of theSCN (254). Interestingly, rhythmicity in gene expressionwithin tissues is generally unique, generating phase and pe-riod-specic oscillations over the circadian cycle. The brainalso shows tissue-specic oscillations in gene expression,which can be independent of SCN rhythmicity (83, 84).While some rhythmic genes are tissue specic, others are

shared across many tissues. Some of the clock genes ( Per2 ,for example) show rhythmicity across many different tis-sues as do Bmal1 , Rev-erb , and the Cry genes (249). Someunlikely molecular candidates linking the circadian clock tospecic metabolic tissues are being revealed, however, andare illuminating new ways by which circadian physiologymay be controlled in tissue-specic ways.

The sirtuin family of proteins is becoming recognized for itsperipheral and central role in both circadian rhythmicityand metabolism. The sirtuin family is composed of sevenfamily members, some of which are mitochondrial (SIRT3,SIRT4, and SIRT5) and others that are principally cytoplas-mic (SIRT2) nuclear (SIRT1, SIRT6, and SIRT7) or ex-pressed in more than one compartment within the cell (91).Sir2 (silent information regulator 2, the homolog of themammalian SIRT1) is an NAD -dependent histonedeacetylase, perhaps best known initially for its role in lon-gevity (135, 144), although it is not clear that it contributesto a longer life span in all species (24). Consequently, themammalian ortholog of Sir2, SIRT1, has emerged as a cen-tral component linking the circadian clock to metabolism.SIRT1 is a class III histone deacetylase and differs from theclass I and II deacetylases in that it requires NAD as acofactor for its enzymatic activity. SIRT1 breaks down

NAD during the process of lysine deacetylation producingO -acetyl-ADP-ribose. During fasting, levels of NAD arehigh, and the activity of SIRT1 is elevated (194). However,when energy is in excess, NAD is depleted because therampant ux through theglycolytic cyclepromotes thecon-version of NAD to NADH. Recent studies demonstratethat SIRT1 directly interacts with circadian clock machin-ery and that its enzymatic activity contributes to robustoscillations of rhythmic genes in vivo (17, 162). Some stud-ies show that the enzyme is constitutively expressed, whileothers demonstrate rhythmicity in expression (17, 162).What is clear, however, is that its enzymatic rhythmicity is

due in part to an oscillation in the levels of the enzymescofactor NAD . In search of the source of SIRT1s oscillat-ing activity, it was discovered that CLOCK:BMAL1 di-rectly regulate the nicotinamide phosphoribosyltransferase(Nampt ) gene promoter, the activation of which providesits expression (163, 189). This activation is remarkable inthat NAMPT provides the rate-limiting step in the NADsalvage pathway. In this way, the classical circadian tran-scriptional loop is linked to an enzymatic feedback loop inwhich SIRT1s own activator is produced in a circadianmanner by the circadian clock machinery (FIGURE 4) (59).




11/29

While initially identied as a histone deacetylase, SIRT1also targets non-histone proteins. In the fasted sate, SIRT1activity affects the activity of numerous target proteins in-cluding many involved directly or indirectly in metabolichomeostasis. These target proteins include PGC-1 ,FOXO, IRS1/2, LXR, HNF-4 , FXR, RAR, TORC2,BMAL1, eNOS, LKB1, AMPK, and SREBP (reviewed inRefs. 68, 142). SIRT1-mediated deacetylation of PGC-1activates the protein, and therefore promotes gluconeo-genic gene transcription and the inhibition of glycolyticgene transcription in the liver (194). SIRT3 appears to bethe major mitochondrial protein deacetylase where itdeacetylates targets such as acetyl-CoA synthetase, lecithin-cholesterol aceyltransferase, and 3-hydroxy-3-methylglu-taryl-CoA synthase 2 (HMGCS2) and thereby controls thelevels of ketone body production (100, 209, 213). Whilelittle is known about the circadian regulation of SIRT3, as aconsumer of NAD , it is likely that SIRT3 activity, likeSIRT1, is activated in a circadian manner. The full extent towhich the sirtuin family affects tissue-specic oscillationsremains to be seen; however, the ubiquitous expression of the sirtuin proteins across metabolic tissues indicates thattheir roles in the circadian clock are likely to be importantfor metabolic homeostasis.

Oscillations within the family of peroxisome proliferator-activated receptor (PPAR) genes have been shown to beprominent in several tissues, including the liver, muscle,brown adipose tissue and white adipose tissue (250). Asregulators of lipid storage and lipogenesis, hepatic fatty acidoxidation, and ketogenesis, this family of proteins providesan important linkbetween peripheral rhythmicityin metabolictissues. A key regulator of gluconeogenesis and glycolysis, theperoxisome proliferator-activated receptor gamma coactiva-tor 1-alpha (PGC-1 ) provides an additional contribution toclock ticking in metabolically active tissues via its regulatory

role in Bmal1 gene transcription. Specically, PGC-1 actsin concert with ROR to activate Bmal1 gene transcriptionand, in so doing, links metabolism to the circadian clock(146). In Pgc-1 null mice, circadian clock gene expressionis altered, and the circadian oscillation amplitudes of oxy-gen consumption are considerably dampened while totallevels of VO2 are elevated across both day and night (146).

1. Circadian rhythmicity in the brain

A) CIRCADIAN OSCILLATIONS IN NEURONS OF THE CENTRAL PACE -MAKER . The mammalian SCN responds to light via neuronsthat extend from the retina of the eye and through theretinohypothalamic tract. The light responsiveness of theseneurons provides synchronization of the organism to thesurrounding light-dark cycles. When an organism is put inan environment that deviates from the 24-h cycle, the en-dogenous circadian clock must reset during a processknown as entrainment. The entrainment process takes time(days to weeks) and temporarily renders a desynchroniza-tion between the brain and peripheral clocks. For example,the disruption of the endogenous clock during travel acrosstime zones gives one theexperience of jetlag, as homeostaticand circadian cues are temporarily out of sync. The endog-enous circadian clock must entrain to the new light/dark

conditions, a process that relies on distinct signaling trans-duction pathways that lie upstream of the transcriptionaltranslational loops of the clock machinery already de-scribed.

While the clock machinery is required to maintain circadianoscillations in the SCN (i.e., Bmal1 knockout animals andClock mutant mice have altered circadian rhythms in SCNtissue), much of the complexity that lies upstream of thesetranscriptional activators has been revealed. The SCN re-ceives nonphotic information from the geniculohypotha-

NAMPT

Nmnat1Nmnat2Nmnat3

NAD+

CRY PER

BMAL1

NAMPT CLOCK

E-Box

SIRT1

Transcription Feedback Loop Enzymatic Feedback Loop

NAMPT

NAM

NMN

SIRT1

FIGURE 4. The sirtuins link circadian rhythmicity to metabolism. The SIRT1:CLOCK:BMAL1 complex drivesexpression of Nampt , the rate-limiting enzyme in the salvage pathway for SIRT1s own cofactor, NAD . The twoloops depicted in the gure demonstrate the mechanisms of both the transcriptional feedback circuit as wellas the enzymatic feedback circuit.




12/29

lamic tract, the dorsal raphe nucleus, and the median raphenucleus (reviewed in Ref. 45). While some redundancy ex-ists, generally the photic input relayed via the retinohypo-thalamic tract is independent of rods and cones but relies onphotosensitive retinal ganglion cells (18). Photic and non-photic inputs are integrated in the SCN by various signaltransduction pathways. Neurons of the SCN respond topituitary adenylyl cyclase activating peptide (PACAP) dur-

ing the day, which produces neuronal depolarization (93).At night, neurons of the SCN respond to acetylcholine aswell as other cGMP-activating analogs (79). While non-photicSCN resetting can occur viaserotonergic innervationof the SCN (217), light exposure during the night can alsoreset the SCN clock by triggering the release of glutamatefrom retinal ganglion cells and a resulting activation of NMDA receptors on SCN neurons. The subsequent depo-larization of SCN neurons leads to activation of calcium-sensitiveadenylyl cyclases in theSCNand the production of cAMP. cAMP activates proteins such as the guanine nucle-otide exchange factors (EPAC proteins) as well themitogen-activated protein kinase (MAPK) signaling cascade, whichin neurons couples depolarization to transcription via acti-vation of cAMP response element binding protein (CREB).Indeed, organisms exposed to a light pulse at night show arapid and robust MAPK phosphorylation in the SCN aswell as phosphorylation of CREB and activation of CRE-mediated gene transcription (172, 173). There is direct ev-idence that cAMP oscillations are required for normal cir-cadian rhythmicity (166). Noncompetitive inhibition of cAMP-producing adenylyl cyclase enzymes in the SCN se-verely prolongs locomoter periods in rodents, an eventwhich is abrogated by mutations in the central clock ma-chinery. Furthermore, adenylyl cyclase activators can reset

the clock, producing robust changes in the amplitude of gene expression, dependent on the circadian time at whichactivators are administered.

The neuronal signaling that occurs in response to lightwould be useless without the ability of SCN neurons toproperly synchronize with each other. SCN neurons in cul-ture show circadianrhythmicity in ring, although synapsesbetween the neurons are not necessary or sufcient for thisrhythmicity (241). In fact, cultured SCN neurons that oscil-late in ring rate can vary drastically in phase from neigh-boring cells. Phase coherence in vivo is thought to be medi-

ated at least in part via the secretion of vasointestinal poly-petide (VIP), which is released from 25% of SCN neurons(8). VIP appears to be dually important for rhythmicity insome of the SCN pacemaker cells and synchrony in others.Interestingly, one of the receptors for VIP, the G protein-coupled VIPR2 (or VPAC2), not only contributes to main-tenance of normal circadian rhythmicity but also to meta-bolic homeostasis. Loss of this receptor interferes with cir-cadian rhythmicity in rodents (94, 102, 212) and reducesmetabolic output during the active cycle (15). Vipr2 knock-out animals show a reduction in both oxygen consumption

and carbon dioxide output during the nighttime wakinghours, and much of their food consumption is advancedinto the resting period. Additional studies also support thelink between VIP signaling and metabolic homeostasis. Forexample, VIP knockout mice have elevated plasma glucose,insulin, and leptin levels, probably due to the expression of VIP receptors on taste cells which in the absence of VIPaffects the tongues role as a sensory gate for energy intake

(153).

The SCN modulates the circadian release of multiple neu-rotransmitters and hormones but responds to many of themin turn. For example, the SCN is required for rhythmicmelatonin release from the pineal gland, but the SCN alsoresponds to melatonin, and can be reset by the activation of its own melatonin receptors (178). Similar regulation byand feedback to the SCN is observed with other hormonesand neurotransmitters as well. While peripheral tissues re-spond acutely to glucose demands, the brain anticipatesglucose demands. It accomplishes this, in part, by commu-nicating with the periphery to release glucose and insulin ina circadian manner. In humans, insulin release is highestduring the early morning hours, when the body anticipatesupcoming glucose metabolism (19, 131, 232). There is evi-dence that the SCN participates in maintaining this balanceas lesions of the SCN eliminate plasma glucose and insulinrhythmicity (159, 160). These oscillations appear to be in-dependent of a loss of rhythmicity in food intake (131).Circadian oscillations in glucose and insulin production arealso controlled by inhibitory and excitatory inputs into thepreautonomic neurons of the paraventricular nucleus(PVN) via the SCN. Stimulation of preautonomic neuronsof the PVN via the GABAergic antagonist bicuculline pro-

duces hyperglycermia but only during the light period,while exciting neurons of the PVN via the glutamatergicagonist NMDA does not show this time-of-day effect (112).The SCN appears to use rhythmic GABAergic projectionsto control the activity level of sympathetic preautonomicneurons of the PVN as SCN-lesioned animals show no cir-cadian variation in hyperglycemia after GABAergic antag-onism.

B) CIRCADIAN OSCILLATIONS IN FOOD ENTRAINABLE ENSEMBLES OFTHE BRAIN . As the hypothalamus plays a central role in feed-ing, how the circadian clock affects hypothalamic function

has been a recent focus of interest. Lesion experiments thatpredate the 1950s reveal that while disruption of the medialhypothalamus causes hyperphagia and obesity, lesions of the lateral hypothalamus lead to a cessation of food con-sumption (reviewed in Ref. 47). The arcuate nucleus of thehypothalamus plays a unique role in metabolic homeosta-sis. Within this structure are two neuronal populations thatcontribute profoundly to energy homeostasis: one popula-tion consists of orexigenic neurons [expressing the neuro-peptide Y (NPY) and agouti-related peptide (AgRP)] andanother population consists of anorexigenic neurons




13/29

[which express proopiomelanocortin-derived peptides in-cluding the -melanocyte stimulating hormone (MSH) andcocaine- and amphetamine-regulated transcript (CART)proteins]. Both of these neuronal populations project to thePVN, where AgRP antagonizes melanocortin receptors and

-MSH functions as an agonist at melanocortin receptors(66, 150). Thus injection of NPY into the paraventricularhypothalamus causes an increase in food intake (220). In-

terestingly, NPY has been shown to oscillate in a circadianfashion, an oscillation which is controlled in part by bothserotonin and GABA and is abolished by feeding restriction(80, 246). Serotonin enhances NPY release while GABA(which also oscillates in a circadian fashion) inhibits NPYrelease in the SCN. The adenylate cyclase-coupled melano-cortin receptors respond to these peptides, and knockoutsof these receptors have demonstrated their importance inenergy homeostasis. Melanocortin 3 (MCR-3) and melano-cortin 4 (MCR-4) receptor knockout animals are bothobese, with MCR-3 rodents being obese without hyperpha-gia. MCR-4 mutations are known to cause obesity in bothhumans and rodents (67, 105). The melanocortin-secreting

cells of the arcuate nucleus are highly responsive to leptin(184). The precise mechanisms underlying leptins effectson body weight are still unclear, but leptin receptors inGABA-secreting neurons appear to contribute greatly tobody weight regulation. Specically, the elimination of lep-tin receptors from all GABA-secreting neurons but not glu-atamatergic neurons promotes dramatic increases in foodintake and body fat mass (238).

Ghrelin produces its orexigenic effects by triggering the ac-tivation of the growth hormone secretagogue receptor(GHSR). GHSR was rst identied as ghrelin-responsive inthe pituitary, where its activation triggers the release of growth hormone (123). Ghrelins receptors are also abun-dant in the hypothalamus, and the GHSR-mediated activa-tion of AMPK in hypothalamic neurons culminates in in-creased mitochondrial fatty acid oxidation within NPY-expressing cells and GABA-mediated inhibition of NPY/ AgRP neighborPOMC neurons (124).As a result, feeding isincreased in response to ghrelin release from the stomach.This chain of events bears circadian inuence at multiplesteps. First of all, ghrelin oscillates in a circadian fashion(141). Second, the increased fatty acid oxidation dependson cofactors that have been shown to oscillate in othertissues, such as NAD (163, 189). As the molecular ma-

chinery required for the feedback loops depicted in FIGURE4 are intact in hypothalamic circuits, it is quite possible thata similar circadian prole can be expected for NAD -regu-lated processes in melanocortin neurons.

2. The hepatic circadian clock

Due to the relative importance of the liver in glucose andlipid homeostasis, its participation in the circadian clock isof central importance to metabolic physiology. In fact, theliver has been a central target in the study of circadian

rhythmicity as it displays robust oscillations in circadianoutput genes as well as in genes specic to the hepatic sys-tem (2). Hepatic rhythmicity depends in part on the rhyth-mic intake of food which is preserved in constant darkconditions. Interestingly, even under the inuence of a func-tional central clock, the vast majority (over 80%) of hepaticgenes are rhythmic in response to food intake (237). Con-versely, in mice devoid of circadian rhythmicity, such as is

the case in Cry1 /

/ Cry2 /

double knockout mice, re-stricted feeding can restore rhythmicity to many genes in theliver that are otherwise arrhythmic in expression. WhileSCN lesions generally abolish circadian rhythmicity in theliver, strong transcriptional activators, such as the gluco-corticoid receptor, can confer rhythmicity to gene expres-sion made arrhythmic by SCN lesions (191). In spite of theinuence of food intake on the hepatic clock, the core clockproteins still hold a prominent role in maintaining the he-patic clock. Clock mutations or deletions such as theClock 19 mutation and the Clock exon 5 deletion severelyaffect the hepatic circadian system as demonstrated in nu-merous rodent studies (41, 42, 154). Bmal1 knockout miceare arrhythmic in the brain and the liver as are the Clock 19

mice (154, 236). Interestingly, Clock knockout mice showsome arrhythmicity in the liver while remaining rhythmic inthe brain, underscoring the tissue-specic role of CLOCKprotein in the liver. While NPAS2 protein can compensatefor loss of Clock in the brain, it does not compensate for lossof CLOCK function in the liver (42, 43).

These results demonstrate rhythmicity of gene expression inthe liver, although not to be ignored is the fact that as muchas 20% of liver-soluble proteins are subject to circadianregulation (190). In fact, for almost 50% of rhythmic pro-

teins identied in the liver, no oscillation in the correspond-ing mRNA levels can be observed, underscoring the impor-tance of posttranslational and translational modicationsin maintaining hepatic rhythmicity. The hepatic proteomealso shows some dependence on the circadian clock ma-chinery as livers from Clock mutantmice and mPer2 ldc mice(11) show dampened oscillatory activity for some proteins(190).

Recent work on the hepatic circadian clock has led to anemerging theme in the circadian eld; several circadianclock proteins actually have numerous intracellular roles in

addition to contributing to the classical loop of FIGURE 2 .As many genes show liver-specic oscillations, it is perhapsnot surprising that some members of the clock machineryinteract with non-core clock proteins such as tissue-specicnuclear factors (FIGURE 5) . Nuclear receptors are central toliver metabolism, regulating genes involved in glucose andlipid metabolism among others (251). Recently, PER2 andCRY1 have been observed to have functions independent of their CLOCK:BMAL1 repression (86, 255). Thus far, PER2seems particularly promiscuous, binding to PPAR , PPAR ,and REV-ERB (87, 206), thereby controlling white adipose




14/29

and liver tissue metabolic processes. PER2 appears to beunique amongthe PERproteins in itsability to bind core clockmachinery as well as thesenuclear receptors.Studies in the lastdecadehave conrmed the circadian rhythmicityof numerous

nuclear receptors, the oscillations of which show tissue speci-city in some cases (250).

CRY1 also appears to have dual roles in regulating thehepatic clock. In addition to its role as a negative-feedbackregulator of CLOCK:BMAL1-dependent gene transcrip-tion, recently demonstrated is its ability to suppress hepaticgluconeogenesis. CRY1 protein oscillates, with elevatedlevels in the liver occurring during the nighttime in noctur-nal rodents. CRY1 appears to block cAMP production inthe liver, probably by binding to G s , thereby preventingcoupling of G protein-coupled receptors to adenylyl cycla-ses (255). As forskolin, a general adenylyl cyclase activator,is able to overcome inhibition by CRY1 of gluconeogenicgene expression [phosphoenolpyruvate kinase (PEPCK)and glucose-6-phosphatase (G-6-Pase), specicially] in thepresence of more than one G protein-coupled receptor, it ispossible that CRY1 is a general inhibitor of G s .

3. Circadian rhythmicity in metabolic peripheral tissues

Most tissues display circadian rhythms. In addition to thehepatic clock, other metabolic tissues show strong oscilla-tions in gene expression. Adipose tissue is among these.

Adipocytes host molecular clocks, and their proliferation isunder the inuence of the transcription factor CCAAT en-hancer binding protein beta (C/EBP ), the expression of which oscillates in adipose tissue and by the circadian pro-tein REV-ERB , which suppresses anti-adipogenic genes(reviewed in Ref. 20). Adipose tissue is particularly relevantto metabolic homeostasis because leptin and adiponectin,two hormones central to the control of metabolism andcardiovascular disease, are released from adipocytes in acircadian fashion. The adipokine adiponectin is also re-leased from adipocytes, and its plasma levels fall during the

nighttime hours (78). This circadian property of adiponec-tin release has been a topic of interest as adiponectin, alsoknown for its anti-inammatory and antiatherogenic prop-erties, is tightly linked to metabolism and body weight reg-

ulation in humans. For example, decreases in adiponectinhave been observed in individuals with disorders associatedwith insulin resistance including obesity and type 2 diabetes(101, 242). The oscillation of adiponectin may also be im-portant for target tissues such as muscle where adiponectinsignaling affects the number of mitochondria. Recent stud-ies looking at adiponectin signaling in muscleshow that lossof the adiponectin receptor 1 in muscle results in a decreasein exercise capacity (107). Mediated through AMPK andSIRT1 (two proteins that contribute to circadian rhythmic-ity), adiponectin appears to modulate muscle insulin sensi-tivity by modulating both PGC-1 expression (via CaMKand CREB) and its activity (via SIRT1-mediated deacetyla-tion). These rodent studies compliment what is seen in hu-mans, namely, a reduction in adiponectin receptor 1 andPGC-1 activity in individuals with type 2 diabetes (158).

Leptin secretion by adipocytes is also a circadian regulatedevent. Leptins contribution to body weight regulation wasrst observed in the ob/ob and db/db mice, two obese, mu-tant mice generated at Jackson laboratories (31, 104, 106).Subsequent studies revealed that the obese ( ob) gene was aleptin-encoding gene, and db / db mice were found to lackthe leptin receptor (28, 256). These leptin-decient mice arehyperphagic, show mild hyperglycemia and severe obesity,

and serve as a model for type 2 diabetes. While alteredleptin signalinghasaccountedforonly a small percentage of human obesity, its link to circadian rhythms may contributeto themetabolicphysiology of these individuals. In humans,plasma leptin levels oscillate with a maximum level occur-ring in the late night and early morning hours (inverse toadiponectin). Leptin levels generally correlate with in-creased adiposity, and leptin levels fall after weight loss inboth mice and humans (151). Leptin levels as well as oscil-lation amplitudearegreatly enhanced in obese women com-pared with normal-weight control women (136).

CREB HNF-4

CRY PER PER2

CCGsClock output

Adipocytedifferentiation

Inflammation?

CRY1

Gluconeogenesis

BMAL1CLOCK

GR

CRY1

PPAR

PER2

Crtc2

FIGURE 5. Circadian clock proteins engage in additional functions when they collaborate with proteinsoutside of the core clock machinery. While CLOCK:BMAL1 activate numerous clock output genes, CRY1 andPER proteins bind to other nuclear receptors or intracellular proteins to regulate disparate functions such asadipogenesis and gluconeogenesis.




15/29

Aside from the humoral oscillations generated by adiposetissue, recent work shows that PER2 plays a key role inadipocyte gene expression via a direct interaction withPPAR (87). In adipose tissue, the binding of PER2 toPPAR recruits it away from target gene promoters.Studies performed on Per2 / mice (10), which have re-duced adiposity, show that in the absence of PER2, genesnormally expressed in brown adipose tissue begin to be

expressed in white adipose tissue (87), essentially trans-forming white adipose tissue into a highly oxidative,brown adipose-like tissue, enhancing energy output andthereby producing a lean mouse.

The pancreas hosts an autonomous clock, and recent workon pancreatic islet cells clocks demonstrates their impor-tance in insulin production and blood glucose maintenance(189, 199). Specically, elimination of Clock or Bmal1function specically in islet cells of the pancreas causes re-duced glucose tolerance, impaired insulin secretion, andalterations in both the size and proliferation of islet cells(152). Importantly, insulin resistance in circadian mutantmice is dissociable from obesity as restoration of islet cellactivity can relieve thesymptom of insulin resistance in spiteof persisting obesity (152).

As a regulator of uid volume and mineral and electrolytecomposition in the body, the kidney has also been studied ina circadian context. The renal tubular NHE3 gene (whichencodes the Na /H exchanger) is CLOCK:BMAL1 re-sponsive, containing an E box that interacts directly withthe CLOCK:BMAL1 complex (200). Furthermore, the cir-cadian uctuation of gene expression in the kidney appearsto be both light and to a lesser extent, food responsive,

although both appear to be necessary for rapid entrainmentof the kidney clock (244). Less is known about the physio-logical consequences of circadian disruption in the kidney.Also a central regulator of uid balance, the adrenal glandsappear to be important physiological mediators of theclock. The adrenal gland releases among other moleculesmineralocorticoids and glucocorticoids and thereby partic-ipates in glucocorticoid-responsive gene expression in theliver. Approximately 5% of the adrenal transcriptome isunder control of the circadian clock (176, 177), and recentstudies reveal that theamong these aregenes that contributeto aldosterone release and, therefore, blood pressure ho-

meostasis (51).

C. Clock Gene Function Within MetabolicTissues

An exciting pursuit in circadian physiology is how the indi-vidual clock genes function in different metabolic tissues.To better understand how proteins of the circadian clockcontribute to metabolism in metabolic tissues, a number of mutant and transgenic mouse strains have been studied,

resulting in a better resolution picture of how circadian andmetabolic processes interact to control physiology.

Circadian rhythmicity within the central nervous system ishighly resilient. Even when core components of the clockmachinery are genetically disturbed, rhythms typically per-sist in LD and even DD conditions. This speaks to the per-sistence of the clock and also to the dependence organisms

have on a system that can synchronize with the environ-ment. While disruption of circadian rhythmicity in DD con-ditions occurs in genetically modied Bmal1 knockout ro-dents (23), both rodents and humans that harbor certaindeletions in, overexpression of, or mutation in other clockgenes have produced some (though less pronounced) alter-ations in the endogenous time-keeping capacities of theclock. Moresubtle but still debilitating disturbances includeloss of rhythmicity in free-running conditions, impaired re-sponse to light shift paradigms (i.e., jetlag scenarios), andvery advanced or delayed sleep cycles. Some of the trans-genic and knockout animals made to address the role of circadian clock genes in vivo as well as their correspondingcircadian and metabolic phenotypes (or lack thereof) aredepicted in TABLE 2 . The number of genetically modiedmice demonstrating both circadian and metabolic abnor-malities is growing and includes global and sometimes tis-sue-specic alternations in Bmal1 , Clock , Npas2 , and iso-forms of the Ck1 , Cry, Per, Pgc-1, Rev-erb , and Ror genes(23, 35, 41, 125, 130, 134, 152, 199, 233, 236, 258) (3, 12,26, 29, 33, 42, 51, 5456, 65, 85, 86, 92, 138, 140, 143,146, 148, 175, 186, 188, 211, 219, 235, 248, 257).

The Clock gene was originally identied in a screen de-signed to look for endogenous mutations that result in a

circadian phenotype (236). The Clock19

animals werefound to harbor a deletion in the Clock gene that involves a51-amino acid deletion in the transactivation domain, amutation that is now known to be antimorphic to CLOCKfunction (121). This Clock mutant mouse (or Clock 19

mouse) has been well studied and is of particular interestfrom a metabolic perspective. Clock 19 mice are obese on ahigh-fat diet, insulin resistant, and show altered plasmatriglyceride levels on the two separate backgrounds studied(130, 174, 233). Interestingly, another metabolic pheno-type caused by the Clock 19 mutation has also been ob-served. While in a C57/bL6 background, the Clock 19 mu-

tation confers obesity on a high-fat diet, Jcl:ICR back-ground mice harboring the Clock 19 mutation show noobesity (174), but show abnormal lipid partitioning. Whenfed a high-fat diet, the Clock mutation in an ICR back-ground reduces hepatic triglyceride levels. Oscillations inserum FFA levels are ablated by this mutation, and underhigh-fat feeding conditions, serum FFA levels rise (130).While plasma glucose and triglycerides have been repeat-edly observed to oscillate in WT animals, oscillations aredisrupted in Clock 19 mutant mice (197, 205, 210). Thehypertriglyceridemia in Clock 19 mice may be attributable




16/29

to the inability of the circadian clock to adequately regulatediurnal variations in plasma triglycerides. Specically, os-cillating ApoB lipoprotein-carrying triglycerides depend onthe microsomal triglyceride transfer protein (MTP) chaper-one to shuttle between membranes. SHP(small heterodimerpartner) regulates MTP, and the recent discovery of a func-tional CLOCK-responsive E box in SHP led to studies

which revealed its diurnal variation and contribution tonormal circadian uctuations in plasma triglyceride levels(24, 179). Aberrant expression of other metabolism-regu-lating genes has also been observed in the Clock 19 mice onone or the other background, including hypocretin neuro-peptide precursor ( Hcrt , which encodes for orexin A andorexin B), ghrelin, cocaine and amphetamine regulated

Table 2. Genetically modied circadian mutants and their corresponding circadian and metabolic phenotypes

GeneTissue-Specic

Disruption? Circadian Phenotype Metabolic Phenotype

Bmal1 (Arntl) No Arrhythmia in DD, altered light response Reduced activity in LD, reduced lifespanand accelerated aging

Bmal1 (Arntl) Yes, pancreas Impaired GT, reduced islet insulin secretionBmal1 (Arntl) Yes, liver Fasting-induced hypoglycemiaClock 19 No Long period in DD, then loss of

rhythmicityIncreased serum TGA, impaired glucose

sensitivity, obesity, low liver TG on highfat diet, reduced islet size andproliferation

Clock decient No Slightly shorter period in DD, impairedlight response and entrainment in LD

Partial diabetes insipidus, decrease inblood pressure, slight decrease inlifespan

Clock/Npas2 No Arrhythmicity in DDCk1 (tau

mutation)No Shortened period in DD, impaired

entrainmentCk1 mutant No Shortened periodCk1 No Embryonic lethalityCk1 Yes, l iver Local period lengtheningCry1, (Tg) NoCry1 (AP-Tg) No Long period in DD, splitting, impaired

entrainmentElevated glucose in serum and urine

Cry1/Cry2 No Arrhythmia in DD, altered light response Salt-sensitive hypertensionNoc (Nocturnin) No Resistant to diet-induced obesity (19)Npas2 No Enhanced adaptation to light

entrainmentHigher nocturnal wheel running, delayed

FAA activityPer1 No Shortened period in DD, impaired light

responsePer2 No Shortened period followed by arrhythmia

in DD, impaired light responseImpaired lipid metabolism

Per3 No Mild period shortening in DD Increased adipose mass, glucoseintolerance

Per1/Per2 No Arrythmia in DDPgc-1 No Mildly longer period in DD Resistance to HF diet-induced obesity,

reduced thermogenesis, defective thermogenesis

Pgc-1 No Reduced activity in the dark Reduced serum TG and FFA levels,elevated hepatic lipid accumulationduring HF feeding

Rev-erb No Mildly shortened activity period in DD,impaired light response

High plasma LDL, altered hepatic TGAlevels, low bile acid accumulation

Rev-erb , Rev-erb(inducible)

No Advanced phase angle of entrainment,short period in DD

Increased circulating glucose and TGAlevels, reduced circulating FFA

Ror (staggerersg/sg)

No Mildly shortened period in DD Increased HDL and hepatic TGA,accelerated development ofatherosclerosis, resistance to HF diet-induced obesity

Ror No Long period in DD

The circadian mutation as well as the affected tissue is documented along with the resulting circadianphenotype. If metabolic phenotypes for the corresponding mutant animal have been observed, they are listedin the last column.




17/29

transcript ( Cart ), fatty acid binding protein1 ( Fabp 1), acyl-CoA synthetase 4 ( Acsl4), 3-hydroxy-3-methylglutaryl-CoA reductase ( Hmgcr ), the low-density lipoprotein recep-tor (Ldlr ), and cytochrome P-450 family 7 subfamily Apolypeptide 1 ( Cyp7a1 ) (reviewed in Ref. 129).

The Clock 19 mouse demonstrates theglobal importance of CLOCK on metabolism, and subsequent studies have been

undertaken to identify the function of CLOCK and othercircadian clock genes in different areas of the brain and theperiphery. Additional mouse models made to address therole of the CLOCK protein in vivo have revealed someinitially surprising results. First, the complete loss of CLOCK protein does not induce arrhythmia (41), althoughexplant experiments from Clock knockout animals alsodemonstrate a dependence of peripheral oscillators onCLOCK protein (43). Studies addressed to determine whythe absence of clock protein is unable to perturb the brainsability to tell time revealed that Npas2 (the Clock paralogwhich is highly expressed in the brain and able to bind toCLOCKs transcriptional partner, BMAL1) is able to com-pensate for CLOCK in the brain but not the periphery. TheClock / Npas2 double knockout animal, which has no func-tional CLOCK or NPAS2 protein show arrhythmia, aswould be expected if the presence of either CLOCK orNPAS2 protein wasessential for maintaining rhythmicity intissues of thecentral nervous system(42). The Clock knock-out mouse is not completely immune to disturbances in thebrains clock, however. These animals show an impairedresponse to light shifting. Specically, while light exposureduring theearly night (ZT1216)phase delays thecircadianclock in WT animals, Clock knockout animals are notphase delayed by light exposure (41). CLOCK and NPAS2

both appear to contribute to sleep homeostasis. Clock mu-tant mice sleep less and show reduced rapid eye movement(REM) sleep recovery after sleep deprivation than their WTcounterparts (165). NPAS2 also has an effect on neuronalenergetics as theloss of NPAS2 produces alterations in sleephomeostasis, altering the electrophysiological properties of neurons after sleep deprivation (55, 74).

The phenotype of Bmal1 knockout mice is unique in that itis the only circadian mutant that harboring a single genemutation, shows complete arrhythmia (23, 125). Bmal1knockout mice are completely arrhythmic in DD and dis-

play a severe metabolic phenotype. The fat, muscle, bone,spleen, kidney, testis, heart, and lung of Bmal1 knockoutanimals all show an age-dependent reduction in size, whichis consistent with their elevated ROS levels (125). UnlikeWT animals, glucose and triglyceride levels do not oscillatein Bmal1 knockout animals, and while WT animals typi-cally show a circadian rhythmicity in blood glucose recov-ery following insulin injection, such rhythmicity is absent inBmal1 knockout mice, which show instead pronounced hy-poglycemia around the clock in response to insulin (197).These knockouts also show impaired gluconeogenesis,

probably due in part to the lack of oscillatory PEPCK ac-tivity in their hepatocytes.

The Cry1 / Cry2 double knockout animals demonstrate thenecessity forcryptochromeprotein function in thecircadianclock in the absence of zeitgebers. These animals show com-plete arrhythmia in free running conditions (235). Thesemutants also show disrupted rhythms after changes in the

lighting conditions. This particular

Documents

Eckel-Mahan and Sassone-Cors PhysRev 2013_circadian Clock