Designing and Evaluating Circadian Experiments on Mouse Skin

Abstract

All skin layers and cutaneous appendages harbor a robust circadian clock, whose phase is under the influence of light through the central clock in the suprachiasmatic nucleus. The skin clock coordinates fundamental biological processes, including metabolism and stem cell activation. It also prominently modulates activity of skin-resident immune cells and the inflammatory response. Numerous diurnally regulated genes in the skin have been implicated in skin diseases in GWASs. Therefore, the mouse skin is a powerful model for understanding the diverse roles of circadian biology in maintaining tissue health and the initiation and propagation of disease states. When planning experiments to study the circadian biology of mouse skin, multiple technical and biological factors must be carefully considered. In this paper, we provide comprehensive guidance on the general circadian experimental design and associated housing for the mice. We highlight the importance of aligning sample collection with the desired hair cycle stage and animal age. We introduce methods to disrupt the clock in the skin, including altering light and feeding schedules as well as using transgenic mouse models. Finally, we discuss the use of transcriptomic data, both bulk and single cell, for circadian studies.

INTRODUCTION

All terrestrial organisms have evolved an intrinsic, autonomous circadian clock with a periodicity of approximately 24 hours. This clock coordinates our physiology with environmental changes caused by the daily rotation of the Earth around its axis. Within cells, the circadian clock is driven by transcription—translation loops. In the core clock network, BMAL1 and CLOCK are activators that drive expression of their inhibitors, PERs and CRYs (Figure 1a). The central clock, which resides in neurons of the suprachiasmatic nucleus (SCN) of the hypothalamus, receives photic input from the retina, allowing for its phase to be adjusted to the environmental light—dark (LD) cycle (Cox and Takahashi, 2019; Takahashi, 2017). Ablation of the SCN disrupts the synchrony of circadian clock in cells body-wide (Izumo et al, 2014; Yoo et al, 2004), pointing to the dominant influence of the SCN on the peripheral circadian clocks, including skin cells (Tanioka et al, 2009).

Figure 1. The circadian clock is robust in the skin.

(a) The core clock gene network is a transcription–translation feedback loop. (b) Cell proliferation and metabolism are regulated by the circadian clock in the skin.

In addition to the influence from the central clock, peripheral clocks can be altered by other clock-entraining inputs, such as time of feeding and exercise. Studies have shown that approximately 10% of the transcriptome has diurnal pattern of expression in each organ (Rijo-Ferreira and Takahashi, 2019; Zhang et al, 2014). The genes with diurnal expression pattern in different organs are largely distinct; the core clock genes are the only genes commonly diurnal in all tissues. The evolution of the clock is thought to provide a survival advantage, and disease-associated genetic alterations are enriched in diurnally expressed genes in the relevant organ (Zhang et al, 2014).

Numerous studies have revealed essential roles for the circadian clock in the skin. Examples of biological skin phenomena with a diurnal pattern and evidence for regulation by the circadian clock include initiation of hair follicle growth (anagen) (Lin et al, 2009), proliferation of interfollicular epidermal stem cells (Figure 1b) (Andersen et al, 2023; Geyfman et al, 2012), activation of hair follicle stem cells (Janich et al, 2011), metabolism (Figure 1b) (Stringari et al, 2015), the immune response (Greenberg et al, 2020), the wound-healing response (Hoyle et al, 2017), the inflammatory response to sunburn (Gaddameedhi et al, 2015), sensitivity to γ radiation (Plikus et al, 2013), and UVB-induced DNA damage and skin carcinogenesis (Gaddameedhi et al, 2011; Geyfman et al, 2012). Gene sequence variations linked to inflammatory skin diseases have also been shown to be overrepresented in diurnal skin genes (Duan et al, 2024; Greenberg et al, 2020). There is evidence that circadian disruption is associated with skin diseases (Duan et al, 2021) and that clock gene expression is altered in skin diseases such as psoriasis (Greenberg et al, 2020). Furthermore, proper central and peripheral clock synchronization and communication are essential for maintaining optimal function of the skin (Mortimer et al, 2024). There is also evidence that daily administration of drugs can be timed for maximum efficacy and minimum toxicity (Duan et al, 2021).

Although humans are diurnal and mice are nocturnal, mice are a valuable model for circadian studies owing to their accessibility and the translatable insights they provide. It is hypothesized that many of the patterns observed in mouse skin are antiphasic in human skin (Dakup and Gaddameedhi, 2017; Geyfman et al, 2012). However, some genes are expressed with the same phase in human and mouse skin, including KLF9 (Spörl et al, 2012; Wang et al, 2017), which has been implicated in the regulation of epidermal cell proliferation. Although circadian transcriptomic datasets from various human tissues (Ruben et al, 2018) and cells (Janich et al, 2013) are available, there are very few available from human skin (Spörl et al, 2012; Wu et al, 2018). Hence, in this paper, we will review the best practices for designing and performing experiments to study the circadian biology of skin, with a particular focus on using mice as the model organism.

GENERAL EXPERIMENTAL DESIGN FOR CIRCADIAN STUDIES IN THE SKIN

In circadian experiments, time is denoted by either zeitgeiber time (ZT) or circadian time (CT) (Vitaterna et al, 2001). Under artificial light schedules, such as in an animal vivarium, ZT0 indicates lights turning on, and ZT12 indicates lights turning off. Samples collected across the day from animals housed in a 12-hour:12-hour LD condition reveal diurnal phenomena, which may be driven by the animals’ internal circadian clocks and/or other rhythmic external stimuli. Because the 12-hour:12 hour LD condition simulates the natural day—night cycles, many circadian studies are carried out under this condition. Alternatively, when mice are in “free-running” conditions in constant darkness, CT indicates the time relative to an animal’s intrinsic clock, where CT0 indicates the start of the day, and CT12 indicates the start of the night. In nocturnal animals such as mice, CT12 is used to represent the onset of activity. The circadian clock is autonomous, and circadian phenomena such as sleep and wakefulness persist without light cues, although in mice, the period is slightly shorter than 24 hours in total darkness (Jud et al, 2005; Vitaterna et al, 2001). Diurnal biological phenomena in samples collected from animals housed in constant darkness distinguish circadian variations driven by the animal’s internal circadian clock from those caused by rhythmic external stimuli.

Different strains of mice have slightly different lengths of the “free-running” locomotion period. For example, BALB/cByJ mice have a “free-running” locomotion period that is around 50 minutes shorter than in C57BL/6J mice (Schwartz and Zimmerman, 1990). Specifically, in the skin, no studies have defined strain-attributed differences in the circadian rhythms and function. Mice with C57BL/6J, B6;129, and SKH-1 backgrounds have been used to study circadian functions in the skin.

Sample collection for circadian studies should span the entire circadian cycle, with at least 2 replicates per ZT/CT time point, either from the same day (eg, ZT0-Replicate 1 and ZT0-Replicate 2) or from different and ideally consecutive days (eg, ZT0 and ZT24). When generating at least 2 replicates for every ZT/CT is not feasible, it is recommended to have 2 replicates for the initial time point. For example, if the collection starts at ZT0 with 4-hour intervals in a system with a 24-hour period, we recommend sampling at ZT0, ZT4, ZT8, ZT12, ZT16, ZT20, and ZT24 (as a replicate of ZT0) to better capture features that peak or trough between ZT20 and ZT0. Intervals between sample collection time points should be carefully considered during experimental design. Samples are usually collected at a fixed time integral, with at least 6 samples covering the circadian period. In a 24-hour system, this means sampling every 4 hours (Hughes et al, 2017). Researchers may balance the number of replicates per ZT/CT and the time interval to achieve their experimental goals in a cost-efficient way. Typically, shorter sample collection time interval can compensate for fewer replicates and vice versa. ANOVA test and t-test can be used to confirm the significance of differences between time points, and additional algorithms are available to determine rhythmicity in the data (discussed later in the section on transcriptomic data).

When the goal is to compare the skin’s response to outside stimuli (chemicals, wounding, irradiation, etc) during the day with that of the night, collecting samples with replicates at 2 time points 12 hours apart representing day and night may be sufficient. In addition, if the researchers have prior knowledge of the system and the oscillatory pattern of their phenomenon of interest, sample collection from 2 time points with the greatest diurnal variation can be sufficient. However, it should be noted that diurnal variation can be missed if the 2 collection time points happen to be 6 hours past the peak and trough (Figure 1b).

ANIMAL HOUSING CONDITIONS FOR CIRCADIAN EXPERIMENTS

The circadian clock in the skin is sensitive to the environment, so special attention should be given to mouse housing conditions to ensure that the clock is unperturbed by the beginning of an experiment. We recommend using designated rooms for circadian experiments, where the light schedule can be freely altered. If available, vivarium equipped with secondary rooms that transition into the dark room with a revolving door to prevent any unintended light exposure of animals should be used. Always ensure that the room changes from light to dark at appropriate times by witnessing the light phase change at the set times; a camera can be set up in the vivarium to monitor room lighting. Alongside a room with a standard LD schedule, an inverted LD room allows sample collection from all time points within 12 hours. Researchers should minimize any light disturbances using a dim red light while working on the mice during the dark hours. However, extended red light exposure can also be disruptive, so researchers should still limit the overall time they spend in a dark room (Peirson et al, 2018). In addition to light, the skin’s circadian clock and the diurnal transcriptome are affected by timing of food intake (Wang et al, 2017) and calorie restriction (Solanas et al, 2017). The clock is also affected by high-fat diet (Kohsaka et al, 2007). Therefore, skin circadian experiments not focused on the role of food intake should be performed with standard chow diets and water provided ad libitum.

The most widely used experimental conditions involve a 12-hour:12-hour LD cycle, standard temperature (20—22 °C), standard humidity (40—60%), and unlimited access to regular chow and water. It should be noted that this commonly used temperature in the vivaria is below the thermoneutral temperature for mice, which is 30 °C, and may be lower than optimal (Speakman and Keijer, 2012). Cotton nestlets should be provided as standard enrichment, minimizing stress to mice during experiments. To synchronize clocks of all mice intended for the study, animals should be kept under these conditions for at least 2 weeks before starting a circadian experiment.

Because periodic exercise is able to influence the circadian clock on its own, locomotor activity of animal should be monitored for circadian disruption experiments (Hughes et al, 2021). General movement can be monitored with a camera, but running wheels are the standard for quantifying voluntary activity and exercise in circadian experiments. The wheels should be connected to a tool that measures the number of revolutions in a given interval of time, and the resulting running data can be plotted over time as an actogram (Jud et al, 2005). Actograms are able to show immediate behavioral changes due to circadian disruption and are an essential way to demonstrate shifts to an animal’s active and rest phases. In the skin, one of the most reliable methods to establish the status of the clock is to measure mRNA for core clock genes across 24 hours. In this sense, Bmal1 and Per2, with their antiphasic expression pattern, are commonly used as markers for assessing the status of the skin clock (Figure 1b).

ENVIRONMENTAL FACTORS THAT DISRUPT THE CLOCK IN THE SKIN

External light exposure

Changing the light schedule alters the circadian clock in the skin. Jet lag models can be used to disrupt the clock, simulating circadian disruption caused by travel or shift work (Figure 2). Here, the LD phases in windowless vivarium facilities are shifted daily or every other day (McGowan and Coogan, 2013). Mice under jet lag schedules have been reported to develop spontaneous dermatitis (Kettner et al, 2016) and worsen radiation-induced dermatitis (Dakup et al, 2020). Koritala et al (2023) have established shift-work-like conditions for mice, simulating short-term rotating shifts, long-term rotating shifts, and chronic jet lag. Each condition distinctly alters properties of the skin circadian clock as demonstrated with gene expression of the core clock genes. Mice can also be exposed to constant light or dim light at night, modeling light pollution in modern society (Walker et al, 2019). The length of external light and dark phases can also be altered, such as 14-hour:10-hour or 16-hour:8-hour LD schedules, which models seasonal changes in light exposure (Benloucif et al, 1999; Redlin et al, 2005). Mouse cages usually receive 150—300 lux during the light phase and 0 lux during the dark phase, and altering the dimness of vivarium lights can disrupt the clock as well (Fonken et al, 2013). The lux meters should be placed at the level of each cage when determining the amount of light delivered to mice for dim light conditions.

Figure 2. The circadian clock in the skin can be disrupted in various ways.

Surgical SCN destruction, altered light schedules, altered feeding schedules, and genetic manipulation targeting the core clock genes can disrupt the circadian clock in the skin. SCN, suprachiasmatic nucleus.

Timing of food intake

Mice are nocturnal and consume most of their food during the dark phase of the day when they are active. Limiting food to the daytime changes properties of the clock, including its phase, in the skin and other peripheral tissues (Figure 2) (Khapre et al, 2014; Wang et al, 2017). Day-time—restricted feeding dampens circadian gene expression and alters the diurnal transcriptome, diurnal DNA damage repair, and the immune response in the skin (Greenberg et al, 2020; Wang et al, 2017). Locomotor activity correlates with food access, which can further contribute to circadian disruption (Gelegen et al, 2006; Haupt et al, 2021) caused by time-restricted feeding. When manipulating food intake, mouse and food weight should be recorded throughout an experiment. During fasting periods, mice should be placed in a new cage to eliminate their access to buried food or stool. To accurately measure food intake of individual mice as well as to avoid fighting-related injuries, mice should be housed and handled separately during the experiments. In addition, indirect calorimetry is recommended for tracking energy expenditure over time by housing mice in metabolic cages, where animal activity, water intake, and respiration can also be monitored across time points (Koronowski and Sassone-Corsi, 2022).

INTERNAL FACTORS THAT INFLUENCE THE CLOCK IN THE SKIN

Age

In rodents, the circadian clock develops during the first 3 weeks of postnatal life (Canal-Corretger et al, 2001), suggesting that mice should be at least 3 weeks old for circadian skin studies. Disruption of the circadian clock through a Bmal1 mutation has been shown to promote premature aging phenotypes (Kondratov et al, 2006). However, in aged wild-type (WT) mice, the core circadian machinery in the skin retains rhythmicity of young mice, whereas the diurnal transcriptome in the epidermis exhibits striking differences between the aged and young mice (Benitah and Welz, 2020; Solanas et al, 2017). Therefore, researchers studying the circadian clock and its functions in the skin need to take into consideration the age of the mice and ensure that it is controlled in all experiments. Generally, we recommend studying mice between the ages of 1 and 6 months as a model for circadian regulation in adult skin. However, more precise age limits may be necessary to exclude effects of the hair cycle as outlined below.

Hair cycle

Skin with its hair follicles in the resting phase (telogen) expresses higher number of diurnal genes and shows higher amplitude of core clock genes than skin with actively growing (anagen) follicles (Geyfman et al, 2012; Lin et al, 2009). Consequently, it is necessary to control the hair growth cycle phase in circadian experiments. In studies that do not explicitly focus on the hair follicles or their growth cycle, it is recommended to harvest skin samples during the second telogen phase, the longest partially synchronized resting hair growth cycle phase in adult mice. After second telogen, hair growth across skin occurs in distinct body-wide waves that are coordinated but not synchronized (Geyfman and Andersen, 2010; Plikus et al, 2008), and this could introduce confounding biological variations to the experiments. Thus, it is recommended to plan experiments in the middle of the second telogen, specifically between postnatal days 45 and 65 (Lin et al, 2009), and to have both control and experimental skin samples collected together.

COMMON MISTAKES TO AVOID IN CIRCADIAN EXPERIMENTS

A common error in circadian studies is a failure to ensure proper room conditions before and during an experiment. Fluctuations in humidity, temperature, cage-level light exposure, and food content disturb normal mouse activity, which can disrupt the circadian clock. If researchers do not verify settings of room conditions and automated light schedules, they may miss a malfunction or incorrect input for the LD schedule. Another error is to immediately initiate experiments with mice on arrival from a vendor. At that time, mice are often agitated and unacclimated to the room conditions, resulting in aberrant baseline activity during measurements. In addition, the vendor may use a light cycle that differs from the 12-hour:12-hour LD schedule. For example, the Jackson Laboratory uses a 14-hour:10-hour LD cycle, with lights on early in the east coast time zone. Therefore, we recommend a 2-week minimum time for re-entrainment to the laboratory 12-hour:12-hour LD cycle.

In addition to aberrant housing conditions, researchers may agitate the mice too much during an experiment. Mice are nocturnal, so they are expected to be active and resting during the dark and light phases, respectively. Hence, it is better to handle mice and cages during the active (dark) phase. Regular disturbances in the day can interrupt sleep, cause stress, and disrupt the clock. These disturbances can include jostling cages, making loud sounds, and roughly handling mice. In addition, when researchers are in a dark room, they may spend too much time using excess red light, dim light, hood light, or room lights, which can result in circadian disruption. Frequently testing, measuring, or collecting mouse samples can also stress mice, so it is important to minimize room entry while also ensuring appropriate resolution of rhythmic changes. Many procedures such as obtaining skin tissue cannot be done serially on the same mouse because wounding can affect the clock.

GENETICALLY MODIFIED MICE FOR STUDYING THE CIRCADIAN CLOCK IN THE SKIN

Genetically engineered mouse models that alter the circadian clock serve as invaluable tools, offering precise insights into the roles of both systemic and tissue-intrinsic circadian clocks in the skin (Figure 2). Transgenic mouse models that target members of the core molecular clock network, including Bmal1, Clock, Pers, and Crys genes, are available and have been used to study the circadian clock in the skin (Table 1). Although both global and cell-type—specific Bmal1 mutation models are available, only global knockout models for Clock, Pers, and Crys have been used in skin studies. Various aspects of the circadian clock in the skin have been studied using these genetic models, with clock circuit disruption targeting its positive or negative arm typically resulting in opposing effects. Note that because many clock proteins have additional noncircadian roles, it is most optimal to use distinct transgenic models for validation. For example, observations attributed to clock disruption in Bmal1^−/− mice can be additionally validated in Per1/2^−/− mice.

Table 1.

Transgenic Mouse Models Targeting the Core Clock Gene Network Are Used to Study the Circadian Regulation of Skin Physiology and Function

Gene	Manipulation	Reference
Bmal1	Systemic knockout	Geyfman et al (2012), Greenberg et al (2020), Kirchner et al (2023), Kowalska et al (2013), Li et al (2022), Lin et al (2009), Silveira et al (2020), Watabe et al (2013), and Welz et al (2019)
	Keratinocyte-specific knockout	Geyfman et al (2012), Janich et al (2011), and Plikus et al (2013)
	Blood endothelial cell and skin lymphatic endothelial cell—specific knockout	Holtkamp et al (2021)
	Skin lymphatic endothelial cell—specific knockout	Holtkamp et al (2021)
	Skin conventional dendritic cell—specific knockout	Holtkamp et al (2021)
	Keratinocyte-specific expression	Mortimer et al (2024) and Welz et al (2019)
Clock	Systemic knockout	Dubrovsky et al (2010)
	Point mutation	Ando et al (2015), Hashikawa et al (2017), Kirchner et al (2023), Lin et al (2009), Matsunaga et al (2014), and Takita et al (2013)
Per	Systemic Per2 knockout	Ando et al (2015), Nakamura et al (2011), Yujra et al (2024), and Zagni et al (2017)
	Systemic Per1 and Per2 double knockout	Dakup et al (2020), Gaddameedhi et al (2015), Janich et al (2011), and Kowalska et al (2013)
Cry	Systemic Cry1 knockout	Destici et al (2013)
	Systemic Cry2 knockout	Destici et al (2013)
	Systemic Cry1 and Cry2 double knockout	Gaddameedhi et al (2015, 2011), Lee et al (2013), Ozturk et al (2009), Plikus et al (2013), and Tanioka et al (2009)
Npas2	Systemic knockout	Sasaki et al (2020) and Shibuya et al (2022)

Transgenic mouse models targeting other clock circuit members, such as Rev-Erb and Ror genes, are also available. However, because studies to date primarily employed them for studying tissues outside of skin or without the focus on their circadian function, they are not covered in this review.

Bmal1

Bmal1-mutated mice, global or cell-type specific, have been used to establish a role for the circadian clock in diurnal phenomena in the skin. Constitutive Bmal1^−/− eliminates behavioral rhythmicity in mice (Bunger et al, 2000; Ko and Takahashi, 2006), leads to premature aging, and reduces lifespan (Kondratov et al, 2006). In the skin, lack of Bmal1 abolishes clock function, resulting in the loss of rhythmicity in the expression of core clock genes and clock-controlled genes (Mortimer et al, 2024; Welz et al, 2019).

Although global Bmal1-knockout mice can provide insights regarding the clock’s function in the skin, it is challenging to distinguish the regulations driven by the skin-intrinsic clock from those driven by the indirect effect of clocks in other tissues. Furthermore, because of the premature aging phenotype of global Bmal1 mutants, phenotypes could be caused by other systemic effects, and investigators should use mice in the first 2 months of age.

Crossing Bmal1^flox/flox and K14Cre mice produces Krt14Cre;Bmal1^flox/flox mice with Bmal1 specifically knocked out in the interfollicular epidermis and in hair follicle epithelial cells (Geyfman et al, 2012; Janich et al, 2011). Tamoxifen-inducible K14Cre^ER;Bmal1^flox/flox mouse model allows studies of conditionally deleted Bmal1 in keratinocytes. Mouse models with tamoxifen-inducible and cell-typ—specific knockout of Bmal1 were also used to study immune cell infiltration in the skin. Dendritic cells migrate into the skin through lymphatic vessels in a diurnal manner, and such migration pattern is lost in Cdh5Cre^ERT2;Bmal1^flox/flox and Prox1Cre^ERT2;Bmal1^flox/flox mice, in which Bmal1 is deleted in the vascular endothelial cells and/or lymphatic endothelial cells. Clec9aCre;Bmal1^flox/flox mice with deletion of Bmal1 specifically in the dendritic cells also lack such diurnal pattern of dendritic cell migration, suggesting that functional circadian clocks in both the cutaneous vascular cells and dendritic cells are essential for regulating the diurnal infiltration of mouse skin (Holtkamp et al, 2021).

The use of global and cell-type—specific knockout of Bmal1 has proved a powerful tool for studying the role of cellular and tissue circadian clocks in maintaining skin health. Nonetheless, knockout of Bmal1 or other circadian clock genes does not allow us to determine the sufficiency of the clock for driving the rhythmicity of biological processes. Moreover, such approaches are inadequate for dissecting interactions between circadian clocks in different cell or tissue types (so called ‘clock communication’). To overcome this limitation, genetic approaches have been developed to allow reconstitution of Bmal1 expression in a single-cell or tissue type in an otherwise Bmal1-null mouse. To recover Bmal1 expression specifically in keratinocytes, Bmal1-stopFL mice were crossed with Krt14Cre mice to create K14Cre;Bmal1stop^{stopFL/stopFL} mice. These mice show systemic circadian disruption indicated by the lack of behavioral rhythmicity in both regular LD cycles and in constant darkness. Interestingly, some rhythms in the epidermis were maintained, with the amplitudes of oscillatory clock gene expressions dampened, when the animals were kept in LD condition. Such weak rhythmicity was eliminated when the mice were in constant darkness, implying that external stimuli such as light can directly but partially entrain peripheral clock in the skin (Mortimer et al, 2024; Welz et al, 2019). To study the communication between the SCN clock and skin clock, K14Cre;Bmal1stop^{stopFL/stopFL} and Syt10Cre;Bmal1stop^{stopFL/stopFL} mice were crossed to reconstitute clocks in the brain and the epidermis. Unlike the epidermis of K14Cre;Bmal1stop^{stopFL/stopFL} mice, the epidermis of these double-reconstituted mice showed circadian rhythms in clock gene expressions that are comparable with those of the control even when mice were kept in constant darkness. It is also found that brain-to-skin clock communication plays a key role in regulating cell cycle in the skin and that autonomous clock in the skin gates signals from the brain to optimize DNA replication timing to avoid oxidative conditions (Mortimer et al, 2024).

Clock

Owing to overlapping roles of Clock and Npas2 (DeBruyne et al, 2007), Clock-deficient mice still display robust behavioral rhythmicity, and the molecular clocks in the central and peripheral tissues continue to function, although slight differently from the ones in WT mice (DeBruyne et al, 2006). However, Clock^−/− mice are prone to developing dermatitis (Dubrovsky et al, 2010).

Clock^Δ19-mutant mice (King et al, 1997b), which still produce CLOCK proteins but with an internal deletion that acts in a dominant-negative fashion (antimorphic) (King et al, 1997a), have been more commonly used in skin research. These mice are able to maintain behavioral rhythmicity in the regular LD condition but become arrhythmic in constant darkness (Ko and Takahashi, 2006; Vitaterna et al, 1994). The circadian clock in the skin is disrupted in Clock^Δ19 mice, as indicated by abnormal expression of core clock and clock-controlled genes (Lin et al, 2009). Clock^Δ19 mice express low and constant level of Aqp3 in the epidermis, leading to a consistently reduced level of stratum corneum hydration compared with that in WT mice (Matsunaga et al, 2014). These mice also experience delayed anagen, although less prominent than Bmal1^−/− mice and without hair follicle deformity (Lin et al, 2009). Researchers interested in studying the immune response in Clock^Δ19 mice may find it relevant that these mice express lower levels of antiviral proteins in both intact and wounded skin (Kirchner et al, 2023) and that their T cells express lower levels of Il23r, which alleviates imiquimod (IMQ)-induced dermatitis, than WT mice (Ando et al, 2015). When challenged, Clock^Δ19 mice experience contact hypersensitivity inflammation that is accompanied by an increased number of mast cells in the skin (Takita et al, 2013). Researchers investigating the relationships between carcinogenesis and circadian clock should note that Clock^Δ19 mice develop fewer tumors than WT mice when treated with carcinogenic chemical on the dorsal skin (Hashikawa et al, 2017).

NPAS2

NPAS2, an analog of CLOCK, is a member of the clock network and shows robust circadian expression in WT mouse dermal fibroblasts (Sasaki et al, 2020). Researchers planning to use Npas2^−/− mice to investigate the circadian clock in the skin should note that NPAS2 and CLOCK have similar functional roles in circadian regulations (DeBruyne et al, 2007; Landgraf et al, 2016) and that the diurnal expression of the core clock genes in dermal fibroblasts isolated from Npas2^−/− is almost identical to that in WT mice except for Npas2 and Per2 (Sasaki et al, 2020). For those who are particularly interested in studying the association between the circadian clock and wound healing in the skin, it is notable that Npas2^−/− has a beneficiary effect on wound healing by promoting fibroblast proliferation, migration, and the ability to form well-organized collagen fibers (Sasaki et al, 2020).

PERs

Single mutation of either Per1 or Per2 leads to shorter period length of behavioral rhythmicity, but Per2-mutant mice eventually develop behavioral arrhythmicity (Bae et al, 2001; Ko and Takahashi, 2006), making them more commonly used in circadian research. Per2^−/− leads to changes in the cellular composition of the skin as well as in cutaneous responses to perturbations. The intact skin of Per2^−/− mice harbors more epidermal stem cells (Zagni et al, 2017). When wounded, Per2^−/− mice experience faster wound healing accompanied by more abundant proliferative and migratory epithelial cells, more abundant myofibroblasts, and more mature collagen fibers at the wound sites than WT mice (Yujra et al, 2024). When treated with IMQ, Per2^−/− mice experience more severe dermatitis along with upregulation of Il23r in the T cells, an observation opposite to the one found in Clock^Δ19 mice, because PER2 inhibits CLOCK (Ando et al, 2015). When challenged with an allergen to induce passive cutaneous anaphylactic reactions at different times of the day, Per2^−/− mice show similar symptom severity in the skin irrespective of the time of challenge, whereas WT mice show diurnal pattern of symptom severity (Nakamura et al, 2011).

Double knockout of Per1 and Per2 immediately abolishes the circadian clock (Bae et al, 2001; Ko and Takahashi, 2006), and skin cells in Per1/2^−/− mice show elevated proliferation. Hair follicle bulge cells and epidermal stem cells in Per1/2^−/− mice are more proliferative than those in WT mice, with the bulge cells showing increased responsiveness to hair growth activation signals (Janich et al, 2011). Overabundance of fibroblasts is also found at wound sites on Per1/2^−/− mice on day 6 after wounding (Kowalska et al, 2013). In addition, Per1/2^−/− mice do not show time-of-day–dependent sensitivity to radiation that is found in WT mice (Dakup et al, 2020; Gaddameedhi et al, 2015).

CRYs

For researchers considering disrupting the clock by targeting CRY, it is important to knock out both Cry1 and Cry2, because a single knockout of either Cry1 or Cry2 does not eliminate behavioral rhythmicity in mice maintained in the regular LD cycle or in constant dark conditions. Double mutation Cry1/2^−/− leads to immediate arrhythmicity in constant darkness (van der Horst et al, 1999). Similarly, whereas single knockout of either Cry1 or Cry2 only advances or delays the circadian clock in the skin when mice are kept in the regular LD cycles (Destici et al, 2013), Cry1/2^−/− double knockout disrupts the clock in the skin when mice are maintained in constant darkness, as indicated by consistently high level of Per2 and Dbp expression (Tanioka et al, 2009). Specifically, in the hair follicles of the Cry1/2^−/− mice, PER2 and CLOCK protein levels are consistently elevated, and mitosis levels are maintained at the peak value found in WT mice. Those interested in using Cry1/2^−/− mice in experiments involving radiation should be aware that Cry1/2^−/− abolishes skin’s diurnal sensitivity to γ and UVB radiation (Gaddameedhi et al, 2015, 2011; Plikus et al, 2013).

STUDYING THE CIRCADIAN CLOCK IN THE SKIN WITH TRANSCRIPTOMIC DATA

To systemically identify genes with a diurnal or circadian expression rhythm in a given tissue, time series of bulk RNA sequencing or microarray data are often collected. The same guidance for general circadian study design as detailed in the previous sections also applies to collection of transcriptomic data. There are a few computational algorithms available for identification of rhythmic genes, including but not limited to Lomb-Scargle (Ruf, 1999), ARSER (Yang and Su, 2010), JTKCycle (Hughes et al, 2010), RAIN (Thaben and Westermark, 2014), meta2d (Wu et al, 2016), and BioCycle (Agostinelli et al, 2016). The reader is referred to Wu et al (2016) for a discussion about the advantages and disadvantages of the mentioned algorithms. To compare the transcriptomic rhythmicity between different conditions, researchers may consider using LimoRhyde (Singer and Hughey, 2018) and DryR (Weger et al, 2021) for statistical rigorousness. Many existing transcriptomic datasets lack information about the time of collection. Such data can nevertheless be used for circadian analysis because multiple computational pipelines can infer CT stamps from transcriptomic data. These algorithms include ZeitZeiger (Hughey et al, 2016), BIO_CLOCK (Agostinelli et al, 2016), and tau-Fisher (Duan et al, 2024). In particular, tauFisher does not require the training and test data from the same assay platform, and its performance has been validated on transcriptomic data from mouse skin (Duan et al, 2024).

Besides transcriptomic data collected using bulk assay methods, single-cell RNA sequencing (scRNA-seq) allows identification of rhythmic genes for specific cell groups. Although the sample preparation protocols for bulk transcriptomic assays and scRNA-seq differ in aspects such as incubation temperature and tissue-processing time, previous research in skin (Duan et al, 2024) has shown consistent clock behavior between bulk transcriptomic data and pseudobulk scRNA-seq data. This implies that the clocks captured in both bulk transcriptomics assay and scRNA-seq reflect the clock at the time of sample isolation.

Owing to the high sequencing dropout (a gene being expressed by a cell but not being picked up by the technology) rate in scRNA-seq data, studying the circadian clock at the single-cell resolution remains challenging. As a result, the pseudobulk approach is usually employed by aggregating and normalizing the total number of reads for each gene in a group of cells, such as cells from the same cell type. The algorithms designed for bulk-level transcriptomic datasets mentioned earlier can successfully identify rhythmic genes in the time series of pseudobulk data, allowing identification of diurnal genes at a finer resolution. Studying the circadian clock in a cell-type—specific manner provides added insights over bulk transcriptomics studies. Thus, in the mouse dermal skin, genes related to metabolism (Adh1, Ndufs8, etc) and migration (Ilk, Vegfa, etc) are rhythmically regulated in the fibroblasts, whereas those related to immune responses (Ifitm3, Cd84, etc) are rhythmically regulated in the immune cells (Duan et al, 2024). Only a few genes are found to be diurnal in both cell types, and most of the shared rhythmic genes are core circadian regulators (Nr1d2, Clock, Arntl, etc). It is worth mentioning that the pseudobulk data can become noisy when the number of captured cells of interest is low at the majority of the sampled time points. In this case, researchers can refer to the expression patterns of the core clock genes to evaluate whether the pseudobulk data are too noisy to be reliable.

CONCLUSION

We have outlined the general experimental design for studying the circadian clock in the skin. For describing diurnal phenomena in the skin, we highlight the importance of establishing and maintaining consistent light and feeding schedules to prevent circadian disruption. In addition, we point to the importance of conducting experiments during the second telogen to minimize influence of the hair cycle. We also discussed methods to disrupt the circadian clock in the skin, either by making genetic mutations in core clock genes or by altering light or feeding schedules. We provide a comprehensive summary of skin findings from transgenic mouse models targeting core clock genes that have been utilized in skin research.

Supplementary Material

Supplementary material is linked to this paper. Teaching slides are available as supplementary material.

Abbreviations:

CT

circadian time

IMQ

imiquimod

LD

light–dark

SCN

suprachiasmatic nucleus

scRNA-seq

single-cell RNA sequencing

WT

wild-type

ZT

zeitgeiber time