• Open Access

Understanding circadian gene function: Animal models of tissue-specific circadian disruption

Authors

  • Tana L. Birky,

    1. Department of Psychology, University of Alabama at Birmingham, Birmingham, AL
    Search for more papers by this author
  • Molly S. Bray

    Corresponding author
    1. Department of Epidemiology, University of Alabama at Birmingham, Birmingham, AL
    2. Department of Nutritional Sciences, University of Texas at Austin, Austin, TX
    • Address correspondence to: Molly S. Bray, Department of Nutritional Sciences, The University of Texas at Austin, T.S. Painter Hall, Room 5.32, 103 W. 24th Street, Austin, TX 78705, USA. Tel.: +1-512-657-1518. Fax: +1-512-495-4945. E-mail: mbray@austin.utexas.edu

    Search for more papers by this author

Abstract

Circadian rhythms are the daily patterns that occur within an organism, from gene expression to behavior. These rhythms are governed not only externally by environmental cues but also internally, with cell-autonomous molecular clock mechanisms present nearly ubiquitously throughout the cells of organisms. In more complex organisms, it has been suggested that the clock mechanisms serve varied functions depending on the tissue in which they are found. By disrupting core circadian gene function in specific tissues of animal models, the various roles of the circadian clock in differing tissues can begin to be defined. This review provides an overview of the model organisms used to elucidate tissue-specific functions of the molecular circadian clock. © 2014 IUBMB Life, 66(1):34–41, 2014.

Introduction

Biologic circadian rhythms are the daily patterns that occur by and within an organism in anticipation of predictable environmental stimuli. These rhythms can include cellular events such as gene expression, bodily events such as hormonal secretion, and organismal behaviors such as sleeping. These patterns are then able to confer the benefit of responding more quickly and appropriately to probable stimuli, resulting in an increase in efficiency and survival for the organism. To maintain these rhythms, cellular molecular clock mechanisms residing within virtually all tissues in the body are coordinated throughout the organism by environmental cues (such as light), neurohumoral factors, or other stimuli [1]. In the mammalian circadian system, central rhythms associated with diurnal behaviors like sleeping/waking and heart rate are regulated in large part by the suprachiasmatic nucleus (SCN), which is located on the anterior hypothalamus and responds to light stimuli detected by the retina [2]. Coordination of oscillatory mechanisms in peripheral tissue is thought to be largely controlled by circulating neurohumoral factors (often associated with feeding and physical activity), including those tightly controlled by the SCN (e.g., melatonin, cortisol) [2]. Increasing evidence has demonstrated that a critical role of the circadian clock is the regulation of metabolic processes, including insulin sensitivity, endocrine regulation, satiety signaling, cellular proliferation, and cellular substrate metabolism. Moreover, metabolic and circadian systems appear to be regulated in a reciprocal fashion, with food intake playing a primary role in driving peripheral rhythms while light-entrainable signals in the brain serve to influence food intake [3].

To maintain biologic rhythms, a molecular circadian mechanism exists within nearly all cells that is characterized by the rhythmic expression and actions of several circadian genes and proteins (Fig. 1) [2, 4]. This clock mechanism is highly conserved across a wide range of organisms and is composed of a combination of positive and negative feedback loops that act to turn gene expression (and ultimately protein production) of target genes on and off in a rhythmic manner. Central to the clock molecular mechanism are the circadian locomotor output cycles kaput (Clock) and aryl hydrocarbon receptor nuclear translocator-like (Arntl or Bmal1) genes, which act as the core component driving these daily rhythms [1]. On translation in the cytoplasm, the CLOCK and BMAL1 proteins form a heterodimer, which is directly and indirectly involved in both the positive and negative transcriptional feedback loops of the circadian clock. CLOCK/BMAL1 heterodimers translocate into the nucleus and activate E-box elements to increase transcription rates of the period (Per) and cryptochrome (Cry) gene families, Bmal1, and other downstream target genes [2, 4]. PER and CRY proteins then form complexes that inhibit the CLOCK/BMAL1 heterodimer's pro-transcriptional activity, thereby decreasing the transcription rates of Bmal1, creating a major negative feedback loop [2, 4]. Other target genes that have CLOCK/BMAL1-induced transcription include nuclear receptor subfamily 1 group D (Nr1d2 or Rev-erbα) and RAR-related orphan receptor alpha (Rora), which both act on the transcription rates of Bmal1 by binding to ROR elements in the Bmal1 promoter [2, 4]. REV-ERBα acts to inhibit the transcription of Bmal1, creating a minor negative feedback loop, while RORA activates Bmal1 transcription [2]. In terms of positive regulation of the circadian clock mechanism, mPER2 may have a positive drive on Bmal1 transcription in the absence of REV-ERBA inhibition [5]. The positive feedback loop also results from the de-repression (and, ultimately, activation) of Bmal1 transcription that occurs with the declining transcription of the Bmal1 negative regulators, Per, Cry, and Rev-Erba resulting from decreasing abundance of the BMAL1/CLOCK dimers [6]. The interaction of the rates of gene transcription and protein degradation creates a pattern whose cycle length approximates 24 h.

Figure 1.

BMAL1 and CLOCK form a heterodimer that acts on E-box elements to drive the transcription of their own genes, as well as the transcription of Period and Cryptochrome family genes, Rora, and Rev-erbα. PER and CRY proteins form a heterodimer that inhibits the pro-transcription activity of BMAL1/CLOCK, thus inhibiting their own expression. REV-ERBα acts on ROR elements to inhibit the transcription of Bmal1, also acting to inhibit its own expression, while RORA activates Bmal1's transcription by acting on ROR elements. [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]

Global Circadian Disruption

Much of what is known about the role of various circadian genes has been elucidated by genetic manipulation in model organisms. Both gene knockout (KO) and dominant negative models have been used, with each type helping to piece together a specific gene's function. Because of the importance of maintaining circadian rhythms, several safeguards are in place if a circadian gene malfunctions. The majority of circadian genes have functionally redundant paralogs that are able to perform adequately in lieu of the primary target gene. For example, the deletion of the Clock gene can be compensated for by its paralog neuronal PAS domain protein 2 (Npas2), which essentially takes over in the absence of a functional Clock gene [7]. Thus, deleting a circadian gene does not necessarily disrupt the clock mechanism within a cell. However, the dependence on a paralog rather than the typical circadian gene may cause a change in circadian parameters of the organism or tissue, including period length, amplitude of expression, and circadian phase, while global circadian rhythms remain intact.

Dominant negative models have been developed to circumvent the problem of functional redundancy by highly overexpressing an inert form of a gene that competes with the functional, endogenous gene. The ClockΔ19 mutation is a widely used dominant negative method to disrupt Clock function. ClockΔ19 is a specific mutation of the Clock gene in which exon 19 is deleted, resulting in a CLOCK protein with a functional binding domain but lacking a transactivation domain [8]. The mutant protein binds to BMAL1 in competition with the endogenous CLOCK and forms a nonfunctional heterodimer that prevents the typical transcription-inducing actions of the BMAL1/CLOCK heterodimer from occurring [8]. Because the expression of the endogenous Clock is not eliminated, transcription of Npas2 is not induced to alleviate the disruption. ClockΔ19/Δ19 mice demonstrate a phenotype mimicking metabolic disorder, including obesity, hyperphagia, hyperlipidemia, hyperleptinemia, hyperglycemia, hypoinsulinemia, and hepatic steatosis, as well as disrupted feeding cycles and decreased energy expenditure [9]. ClockΔ19 mutant mice are hyperphagic across both light and dark periods, whereas wild-type (WT) mice eat 75% of their daily intake during the dark period [10]. These animals are also unable to maintain full-term estrous cycles [4].

Several KO mice have been created by targeting various genes within the circadian clock molecular mechanism. Npas2-deficient mice are nearly indistinguishable from WT littermates but do show deficits in long term memory of cued and contextual fear conditioning [4, 11]. Bmal1−/− mice exhibit a phenotype that mimics premature aging, including early death, metabolic defects, sterility, arthropathy, and osteoporosis, as well as an ablation of several typical diurnal variations, such as glucose and triglyceride levels [4, 12]. The loss of these rhythms may be due to the fact that, although Bmal1 also has a functionally redundant paralog (Bmal2), its transcription is regulated by Bmal1; therefore, Bmal1 KOs act as functional Bmal1/Bmal2 double KOs, disallowing Bmal2 to rescue circadian function [13]. Whole body Bmal1 rescue via bacterial artificial chromosome (BAC) resulted in body weight and rhythmicity of activity similar to WTs, as well as 100% survival rates [14].

Functional redundancy is likely the reason that Per1−/−, Per2−/−, Cry1−/−, and Cry2−/− mice do not have obvious baseline phenotypes [4]. The deletion of either Per1 or Per2 in mice does not disrupt cellular rhythmicity, but Per1/Per2 double KOs exhibit both cellular arrhythmicity and uncontrolled proliferation of osteoblasts, resulting in age-progressive increases in bone mass [4, 13]. Mice homozygously expressing a mutant form of Per2 (Per2-/-) are distinct from Per2 KOs and show an increased rate of mortality, higher overall weight due to hyperphagia during the light period, less activity during the dark period, increased rates of glucose clearance, no corticosterone rhythm, and increased tumorigenesis [10, 15, 16]. This increased rate of tumorigenesis is also present in Per1/Per2 or Cry1/Cry2 double KO mice [17]. Although the effects of global disruption are useful, studies have shown evidence that circadian genes have differing functions depending on the cell and tissue type in which they are expressed. To further understand these tissue-specific cellular clock functions, gene disruption within targeted cell types allows for the examination of circadian gene function within a specific tissue without the presence of confounding global disruption.

Tissue-Specific Circadian Disruption

Retina

The retina is the source of light input information to the SCN through the retinohypothalamic tract and is necessary for mammalian entrainment to light/dark cycles [18]. One of the earliest models of tissue-specific disruption of the circadian clock was performed in a transgenic Xenopus model, in which the ClockΔ19 mutation is driven by the mouse interphotoreceptor retinoid-binding protein (Irbp) promoter, resulting in a circadian clock disruption that is retina-specific [19]. Photoreceptor Clock mutant Xenopus showed arrhythmic melatonin release from their eyecups when placed in total darkness for several days, although the total amount of melatonin released was unchanged [19]. These results suggest that the timing but not the level of melatonin release is under circadian control by the clock within the retina [19]. Hayasaka et al. [20] then created lines that only overexpressed the dominant negative Clock in either the rods or cones of the Xenopus retina, which are roughly equal in proportion. Both cell-specific clock disruptions caused significant abolishment of melatonin release rhythms, but the percentage of arrhythmia was higher in rod-specific versus cone-specific transgenics [20]. However, both cell-specific types showed lower percentages of arrhythmia as compared to the total retina-specific transgenic animals, indicating that both rods and cones play a role in melatonin release, but that rods may have a more dominant role [20].

In a mouse model with retina-specific genetic deletion of Bmal1 (Ret-Bmal1−/−), Storch et al. [21] demonstrated that a functional Bmal1 gene is required for rhythmic retinal processing of visual stimuli, independent of the central clock. While Bmal1 is not necessary for rod photoreceptor electrical responses, it appears to be important for retinal processing of visual stimuli [21]. In mice, Per2 mutants show increased amount of acellular capillaries in retina along with increased retinal permeability [16]; however, this result cannot necessarily be attributed to circadian disruption specifically in the retina but perhaps is a side effect of the global disruption. Recently, Ruan et al. reported that explants of retina and SCN from Per1-, Per2-, Per3-, Cry1-, Cry2-, and Clock-deficient mice showed differential patterns of circadian rhythm alterations within the retinal and SCN tissues, suggesting that “the retinal neural clock has a unique pattern of clock gene dependence at the tissue level that it is similar in pattern, but more severe in degree, than the SCN neural clock, with divergent clock gene regulation of rhythmic period.” [22]. Based on the experiments conducted to date, an intact circadian clock appears to be critical for maintaining normal rhythmic responses to light cues in the retina and for maintaining rhythmicity of melatonin release [23].

Brain

In addition to the metabolic phenotypes observed, additional studies of global Bmal1 KO mice have reported lower body weight, reduced activity levels with abolished activity rhythmicity, and only 29% survival at 10–12 months of age [14], indicating that alterations in energy balance may play a critical role in this overall phenotype. Because inconsistent observations have been made regarding the presence of food-entrainable rhythms in the global Bmal1 KO mice [24, 25], Mieda and Sakurai [26] created a mouse model in which the global Bmal1 KO is driven by the Nestin promoter, which is active in both the central and peripheral nervous system. The nervous system-specific Bmal1 animals lacked food anticipatory behavior and had both reduced total food intake and reductions in body weight, consistent with earlier reports of lower total body weight and adiposity in global KOs, indicating that this component of the global KO is largely neutrally driven [26].

McDearmon et al. [14] used a tissue-specific rescue model to assess the specific role of Bmal1 within the brain; a transgenic mouse model was created on the Bmal1-/- background, with hemagglutinin (HA)-tagged Bmal1 transcription linked to the Scg2 promoter sequence which is known to be expressed throughout the brain and enriched in the SCN. Bmal1 brain-rescued mice showed a partial rescue of wheel running rhythms, although they exhibited less overall activity and a cycle which was an hour shorter than seen in WTs [14]. These mice also weighed less than WTs, had no difference in weight from total Bmal1 KOs, and had increased rates of survival compared to total Bmal1 KOs, with 75% surviving until the end of the study [14]. These findings indicate that Bmal1 in the brain has an effect on behavioral rhythms as well as overall health and aging.

A similar study examined the effects of Clock rescue in ClockΔ19/Δ19 mice, in which mice had strong overexpression of a functional Clock that was linked to the Scg2 promoter sequence and therefore caused recovery of the circadian oscillator within the brain [27]. Brain-rescue of Clock caused restoration of behavioral rhythmicity in the absence of light cues and a slightly decreased period length compared to WTs [27]. Brain-rescued mice also showed a restoration of transcriptional oscillation of some but not all genes in the liver [27]. The strength of the rhythms also exhibited a partial rescue, with brain-rescued mice showing an intermediate amplitude for transcript levels within the liver as compared to WT and ClockΔ19/Δ19 mice [27]. Finally, brain-rescued mice showed a preferential rescue of core clock genes in the SCN, which suggests that these core components are sensitive to signals from the brain and particularly the SCN even without a functional internal cell-autonomous oscillating mechanism [27].

Global Clock KO mice have been reported to exhibit a behavioral phenotype mimicking aspects of bipolar disorder in the manic state, including lower stress and anxiety, less sleep, less depression-like behavior, hyperactivity, and increased reward values for cocaine and sucrose compared to WT animals [28, 29]. Since these animals also demonstrate increased dopamine activity in the ventral tegmental area (VTA) of the brain, Mukherjee et al. [30] examined Clock function specifically in the VTA via RNA interference with shRNA (small hairpin RNA) delivered through stereotaxic injection. This model produced a somewhat inconsistent phenotype to that observed in the global mutant. VTA-specific disruption of the Clock gene was associated with a hyperactive phenotype and a delayed locomotive acclimation activity when exposed to a novel environment; however, when in their home environment, the mice demonstrated lower total activity due to less robust movement in their active period, despite more activity during the typical resting period, compared to WTs [30]. The VTA-specific Clock knockdown mice showed less anxiety than controls across three different behavioral measures by spending more time in the center for the open field test, more time in the open arms in the elevated plus-maze, and more time in the light in the light/dark box than the controls [30]. Unlike the global Clock mutants, the VTA-specific Clock knockdown mice also showed a depressive-like phenotype, having greater immobility in the forced swim test, and greater latency or failure to escape for the learned helplessness challenge [30]. The substantial differences observed in the two types of Clock disruption may be due to the differential activity of dopaminergic cells in the VTA expressing the Clock shRNA compared to global mutants and subsequent alterations of dopamine-related gene expression [30]. The differences may also be due to the non-VTA effects of Clock on the component behaviors that comprise the bipolar phenotype. These findings indicate that Clock function in the VTA helps to regulate dopaminergic output and subsequent behaviors and may be related to the manifestation of symptoms typical of bipolar disorder.

Heart

One of the most well-studied tissue-specific circadian disruption animal models is the cardiomyocyte CLOCK mutant mouse (CCM), in which the expression of the ClockΔ19 mutant gene is driven by the myosin heavy chain (MHC)-α promoter and disrupts clock function only in cardiomyocytes, leaving all other cellular clocks intact. Several studies have been conducted on this model, examining behavioral data, heart function and physiology, and gene expression. Radiotelemetric studies conducted by Bray et al. [31] revealed no significant difference in activity levels throughout the day as measured by beam breaks, but CCM mice showed decreased heart rate overall, especially during their active period. Examination of ex vivo heart function by perfusions with normal and high workload conditions showed consistent results, with a decrease in heart rate present in CCM hearts for both conditions even while separated from neurohumoral factors [31]. It was also shown that CCM hearts lack rhythmicity in cardiac power across different time points while WT mouse hearts have greater cardiac power during their active period [31]. WT mouse hearts show lower oxygen consumption and higher cardiac efficiency compared to CCM animals; these differences were heightened during what would be the active (dark) period, revealing an anticipation of increased workload [31]. The differential cardiac power and efficacy of WT hearts is consistent with previous findings showing that rat hearts isolated during the middle of the dark period, which is the rodent normal active period, have higher steady state cardiac power, oxygen consumption, and carbohydrate oxidation than hearts isolated during the middle of the inactive light period [32]. No similar anticipation or diurnal variation was seen for functional cardiac parameters in CCM hearts for either the normal or high workload condition, indicating that the cardiomyocyte circadian clock plays a role in anticipatory cardiac functions [31]. Gene expression was also investigated using a microarray analysis, with 728 and 296 genes showing oscillatory expression in WT atria and ventricles, respectively, and of those, 548 and 176 showed markedly different expression patterns in corresponding CCM heart tissue [31]. Several of the altered genes are known to be involved in metabolism, signal transduction, protein turnover, transcription, and transport [31]. These findings suggest that Clock in the cardiomyocyte helps to regulate heart rate, cardiac efficiency, responsiveness, and metabolism, as well as governing the daily fluctuations associated with several other aspects of cardiac function.

More recently, CCM mice were demonstrated to have shorter bouts of wheel running with fewer wheel rotations and lower wheel running activity, but maintained circadian rhythmicity and number of bouts, despite the fact that activity as measured by beam breaks was indistinguishable from WT mice [33]. This result suggests that CCM mice are able to sustain low energy activity levels and rhythmicity but seem to have difficulty engaging in long bouts of strenuous activity levels, which could be due to the inability anticipate high cardiac workloads. However, somewhat counterintuitively, it has also been shown that CCM mice are able to increase the muscle mass of their hearts more effectively than WT mice [34]. When subjected to simulated shift work by implementing 12-h biweekly phase shifts for 16 weeks, septal walls of the heart were thicker and biventricular weight to body weight ratios were higher in CCM mice versus WT mice [34]. Mice treated with isoproterenol, a pro-hypertrophic agonist, for 7 days either at the beginning of the light (ZT0) or dark (ZT12) period showed no time dependent hypertrophy, but WT mice showed significantly increased biventricular weight to body weight ratios when isoproterenol is administered at ZT0 [34]. Since the CCM heart appears to have expression levels that mimic WT hearts at ZT0, it is suggested that Clock expression mimicking ZT0 is a pro-hypertrophic state [34]. Additionally, mice lacking a functional Bmal1 gene within cardiomyocytes (CBK mouse) demonstrate an increased biventricular weight phenotype comparable to CCM [34]. It should be noted that these instances refer to physiological rather than pathological hypertrophy, similar to that seen in exercised mice.

Pathways associated with hypertrophy in the heart appear to have several key and modulatory components that are controlled by the cardiomyocyte circadian clock. Hypertrophy requires increased protein synthesis and should show an increase of certain translation initiation factors. Indeed several eukaryotic initiation factors as well as components involved in the ubiquitin and proteasome system are regulated by the cardiomyocyte circadian clock and indicate that hypertrophic states and atrophic states are at least partially controlled by the cardiomyocyte circadian clock [35].

Liver

Increasingly, studies have shown that a key function of circadian clocks is the regulation of a wide range of metabolic functions in multiple tissues. To examine the effect of liver-specific circadian clocks, a conditional Bmal1 KO model was created in which Bmal1 expression is abolished in the presence of cre-recombinase driven by the promoter for Albumin (Alb), which is only expressed in hepatocytes [36]. These liver-specific Bmal1 KO mice had normal activity and feeding behavior, as well as normal body fat content and response to insulin [36]. However, they appeared to demonstrate a fixed minimum expression of glucose transporter 2 (Glut2) transcript and protein, lower fasting/resting blood glucose levels (hypoglycemia), and increased glucose clearance (higher tolerance/lower blood glucose levels in glucose tolerance tests) [36]. These findings suggest that the circadian clocks within hepatocytes act on several aspects of glucose transport.

In another mouse model of liver-specific clock disruption, the expression of Rev-erbα, a negative regulator of Bmal1 is controlled by tetracycline responsive elements, and a tetracycline-dependent transactivator is expressed in hepatocytes [37]. In this model, Rev-erbα is constitutively expressed at high levels in the liver and acts to suppress Bmal1 transcription and Bmal1-regulated elements/target genes [37]. However, when the tetracycline analog doxycycline is administered, the expression of the tagged Rev-erbα is silenced, causing the recovery of circadian rhythms within hepatocytes [37]. In mice not treated with doxycycline, several circadian genes were downregulated, including Bmal1, Cry1, Cry2, and Per1 [37]. For Per2 rhythmicity of in vivo expression appeared to be driven by either the intrinsic circadian oscillator or neurohumoral cues, both of which were sufficient entrainers in lieu of the other [37]. Microarray hybridization revealed over 300 genes whose transcripts were shown to have robust circadian accumulation that was abolished in the absence of doxycycline; of these, 31 genes showed similar rhythmicity, independent of the presence of an intact hepatocyte oscillator [38].

When comparing the levels of gene expression of Rev-erbα KO mice, WT mice, and hepatic-specific Rev-erbα overexpressing transgenic mice, the majority of genes regulated by Rev-erbα in liver tissue were circadian genes [39]. Genes that appear to be regulated by Rev-erbα in the liver include cytochrome P450 7α-hydroxylase (Cyp7a1), which is the rate limiting enzyme in the synthesis of bile acid from cholesterol, along with known targets of the cholesterol sensing factor SREBP, which all appeared to be downregulated in Rev-erbα KO mice [39]. Cyp7a1 mRNA is also downregulated in Rev-erbα KO mice and was also shown to be constitutively overexpressed in transgenic mice with hepatic Rev-erbα overexpression [39]. This evidence further supports previous findings that Rev-erbα KO mice show a 25% decrease in cholic acid concentrations, a type of bile acid synthesized from cholesterol [40]. Expression of Cyp7a1 mRNA has also been shown to be dysregulated in the livers of ClockΔ19 mutant mice, although they have consistently higher expression than their WT counterparts [41]. ClockΔ19 mutant mice also show upregulation of HMG-CoA reductase (Hmgcr) mRNA, which is a key enzyme in bile acid synthesis [41]. Taken together, this evidence provides support that bile acid synthesis is heavily regulated by circadian clocks within the liver.

Circadian rhythms in the liver appear to be strongly entrained by feeding, with the liver being one of the most responsive tissues to inverted feeding schedules independent of light signals compared to most other tissues, including heart, pancreas, and kidney; conversely, altered feeding had little effect on lung tissue and the SCN [38, 42]. The setting of the liver's circadian clock by feeding time is not surprising, since the liver is known to be involved in substrate metabolism, and has actually been shown to regulate many genes encoding relevant macronutrient and micronutrient metabolizing enzymes [38].

Pancreas

A study by Marcheva et al. [43] indicated that certain phenotypic abnormalities such as elevated glucose levels throughout the day observed in ClockΔ19 mice may be due specifically to circadian dysfunction in the pancreas. Pancreatic islets were shown to lose circadian rhythmicity in ClockΔ19 mice, and exhibited impaired insulin release prompted by glucose intake, a decrease in total pancreatic insulin volume, and an approximately 20% decrease in islet size [43]. A similar phenotype was observed in Bmal1−/− mice, showing a 60% decrease in pancreatic insulin release in response to glucose, forskolin, and others, as well as a reduction in the number of large islets observed [43]. To verify that these observations were caused by pancreatic circadian disruption, a pancreas-specific Bmal1 KO mouse was created. Pancreas-specific Bmal1 KO mice had normal levels and rhythms of physical activity, feeding, and normal body weight and body composition [43]. However, they showed impaired glucose tolerance, which was actually more severe than the impairment seen in Bmal1 or Clock global KO or mutant mice. In addition, pancreas-specific Bmal1 KO mice demonstrated both hyperglycemia and hypoinsulinemia by way of decreased secretion, along with the development of all of these symptoms younger than were seen in the global disruption models [43]. These findings indicate that not only do circadian clocks help to control insulin output and glucose tolerance but circadian disruption within the pancreas can cause a more severe phenotype when in the context of an organism with the majority of tissues exhibiting functional circadian oscillators [43].

Skeletal Muscle

Multiple studies have suggested that peripheral circadian clocks within skeletal muscle may play an important role in daily variations of muscle strength, power, responsiveness, and torque seen in humans, as well as other oscillatory changes within skeletal muscles such as glycogen levels, insulin sensitivity, and glucose metabolism changes seen in rodents [44, 45]. However, these studies have not shown the importance of the cell-autonomous clock. McDearmon et al. [14] presented data from Bmal1−/− mice with and without skeletal muscle rescue by tetracycline responsive elements, coupled with doxycycline administration causing the inhibition of Bmal1 expression. In these studies, Bmal1−/− mice have ablated activity rhythmicity, lower overall activity, and lower body weight compared to WTs [14]. While skeletal muscle-rescued mice showed no rhythmicity in activity, they did exhibit a partial rescue of total activity levels approaching WT levels and a partial rescue of body weight [14]. Clearly more studies on skeletal muscle specific rescue or disruption still need to be performed.

Adipose

It has been reported that ClockΔ19 mice have specific phenotypes that appear to mimic metabolic syndrome and obesity, suggesting that the circadian clock within adipose tissue may play an important role in the manifestation of such symptoms [9]. Early studies demonstrated that the core clock gene Bmal1 is a key factor in adipogenesis, which was later shown to be regulated via the Wnt signaling pathway [46, 47]. Global Bmal1 null mice on the C56BL/6J background fail to develop substantial adipose stores, even when fed a high fat diet, and demonstrate elevated levels of circulating lipids and ectopic fat formation in liver and skeletal muscle, suggesting that this central clock gene may play a key role in energy homeostasis [48]. In addition, these global Bmal1 null mice have suppressed expression of several key adipokines, including adiponectin, resistin, and adiponectin receptors 1 and 2 [49]. Conversely, global Bmal1 KOs on a mixed C56BL/6J x 129 background have increased adiposity and impaired glucose tolerance [36], suggesting that genetic background is important in the determination of adiposity phenotypes.

Zvonic et al. [50] reported that circadian rhythmicity is robust in both brown and white adipose and identified 650 transcripts that were expressed in a rhythmic manner in these tissues. In addition, these investigators demonstrated that timed restricted feeding results in a phase-shift of both the central clock genes as well as their downstream target genes [50]. More recently, Paschos et al. [51] used a model of tissue-specific disruption of the Bmal1 gene using a Cre recombinase system driven by the aP2 promoter, which is predicted to disrupt the gene in both brown and white adipocytes as well as macrophages. These investigators report that aP2-specific disruption of the circadian clock results in increased adiposity, which may result from altered patterns of feeding in these animals in which a greater amount of food intake occurred during the light period compared to WT mice, despite equivalent overall food intake [51]. We have recently generated a mouse model in which the overexpression of a dominant negative transgene of the ClockΔ19 mutation is also driven by the aP2 promoter and thus disrupts clock function in adipocytes and potentially in macrophages [52]. Both female and male ACM mice show an increase in weight and body fat percent that is accentuated with age. In addition, the ACM mice exhibit higher mortality rates at younger ages, and show decreased rates of glucose tolerance (Fig. 2). While we are still in the process of characterizing this tissue-specific model of circadian disruption, these findings indicate that many of the components of metabolic disorder may be controlled by the circadian clock within adipocytes.

Figure 2.

No differences were seen between genotypes for total body weight of adult male mice approximately 5 months old or aged mice over 10 months old (A). Significant differences in body fat percentage were observed in aged mice, with aP2-driven Clock mutant mice averaging 29.6% fat while WT mice only had 23.6% (B). **, P < 0.01. [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]

Conclusions

Circadian rhythms have been shown to have substantial effects on human health, especially in cases in which they are repeatedly disrupted or desynchronized. Shift workers, pilots, and flight attendants are prime examples of individuals who are subjected to repeated changes in circadian entrainment factors such as exposure to light and time of eating. These occupations appear to have higher risk for several health issues including obesity, heart disease, diabetes, and overall mortality [4, 52-56]. In animal models and patients with cancer, circadian disruption has been associated with higher rates of tumorigenesis, faster tumor growth, and increased cancer severity. Chronotherapy and chronomodulated drug administration, which take into account the changes in biologic response to drugs as well as the cellular circadian rhythms of the targeted areas, has potential to increase drug efficacy and tolerance [57].

There are a growing number of studies using models of both tissue-specific disruption and tissue-specific rescue of the molecular circadian clock. These models will continue to allow us to parse out the functions of the various circadian genes within a given tissue or cell type. Understanding the mechanisms behind circadian rhythms at the cellular level may help to increase the efficacy of treatments and allow for a greater awareness of the role of circadian biology in disease etiology. More research is needed to elucidate the individual roles of cell-specific clocks in the global regulation of these various biological processes.

Ancillary