Expression of antimicrobial peptide genes oscillates along day/night rhythm protecting mice skin from bacteria

Abstract

Antimicrobial peptides (AMPs) are important components of the innate immune system and are involved in skin protection against environmental insults and in wound healing. Herein, we assessed the gene expression of chemerin (Rarres2), cathelicidin CRAMP (Camp), and three β-defensins (Defb1, Defb3, and Defb14) in mouse skin during light/dark cycle (LD 12:12) and constant darkness (DD). Next, we examined the survival of bacteria applied on the skin at specific times during the day. We found that the expression of Rarres2, Camp, and Defb1 was the highest at 4 h after the beginning of darkness, during high activity of mice. These rhythms, however, were not maintained under DD in the skin but were present in the liver. This indicated that in the case of skin, a circadian input was masked by daily changes of light in the environment. In contrast, Defb3 and Defb14 showed the highest mRNA levels when the mice slept, and these rhythmic mRNA oscillations were maintained under DD. This shows that Rarres2, Camp, and Defb1 levels in the skin are correlated with high locomotor activity in mice and they are controlled by daily changes of light and dark. Alternatively, oscillations in the mRNA levels of Defb3 and Defb14 seem to protect skin and heal wounds during sleep. These rhythms are maintained under DD, indicating that they are regulated by a circadian clock. Our study suggests that daily AMP expression affects the survival of bacteria on the surface of skin, which depends on the phase of AMP cycling.

1 |. INTRODUCTION

Most physiological processes in animals show circadian rhythms, including circadian oscillations in blood pressure, heart rate, energy metabolism, DNA repair, and memory, as well as in wakefulness and sleep.^1–3 Temporal segregation of physiological processes allows organisms to adapt to cyclic changes in light and other environmental cues (Zeitgebers). These regulations are crucial for organism survival and are generated by direct daily changes in environmental cues and/or by internal endogenous mechanisms called circadian clocks. They generate circadian rhythms with a period longer or shorter than 24 h, which are synchronised by Zeitgebers to rhythms with a 24-h period.

All organisms possess a circadian system composed of a central clock or pacemaker in the brain and peripheral clocks located in various cells, tissues, and organs throughout the body.^4,5 Circadian rhythms have also been detected in gene expression and cellular processes in animals and humans, including the immune system and skin.^6–9 In mammals, the central clock is located in the suprachiasmatic nucleus (SCN) of the hypothalamus, while peripheral clocks reside in various tissues, including skin.^5,9 On the cellular level, circadian rhythms are generated by circadian clocks. The molecular mechanism of the clock is based on several autoregulatory feedback loops. In the main negative feedback loop, the expression of clock gene Periods (Per1, Per2, and Per3) and Cryptochromes (Cry1 and Cry2) is activated by transcription factors Clock and Arntl (Bmal1). They are repressed by clock proteins Per and Cry, which, after translocation to the nucleus, inhibit the expression of their own genes. This maintains self-sustainable circadian oscillations in clock genes, their proteins, as well as in clock-controlled genes.¹⁰ Peripheral clocks with cyclic expression of clock genes have been detected in the skin of both humans^7,8 and mice.⁹ Interestingly, the skin has the ability to synthesise and metabolise melatonin, which is a night hormone.^11–13 During the day and night cycle, skin is exposed to different environmental factors including ultraviolet radiation, extreme temperature ranges, risk of physical injuries, toxins, or pathogens, and all these factors depend on the time of the day.^11,14–16 Resultant physiological processes in the skin are clock-controlled to adapt to the external environment.¹⁷ It has been demonstrated that over 1400 genes in the skin are clock-controlled,¹⁸ including genes responsible for immune processes. Among many functions,¹⁷ the circadian clock in the skin increases immunity at a particular time of the day to protect the skin from infection and to minimise the autoimmune responses.¹⁴ Several mechanisms in the skin help to protect the whole organism against infection, one of which is the synthesis of antimicrobial peptides (AMPs), which are involved in innate immune responses.¹⁹ AMPs have a broad antimicrobial potential but also act as multifunctional effector molecules. They are also involved in immune cell migration to the site of infection/inflammation, their proliferation and differentiation, and cytokine production, and they play a key role in linking innate and adaptive immune responses.²⁰ AMPs have been detected in many organisms, including bacteria, fungi, plants, and animals. For instance, in insects that possess only an innate immune system, AMPs are essential for protection against bacteria, viruses, and fungi.²¹ Among the thousands of AMPs identified in organisms, there are over 100 human AMPs; defensins and cathelicidins have been most intensively studied.^22–24 In humans, only one cathelicidin has been described, the peptide LL-37, which consists of 37 amino acids and two leucine residues on its N-terminus.^22,23 Indolicidins, a subfamily of mammalian cathelici dins, are active against fungi, bacteria, and viruses and can be used to treat diseases induced by various microbes.²⁵ AMPs can be especially effective in skin defense against microbiota.¹⁹

The vertebrate liver is also involved in innate immune responses to infection. The “acute phase” response to infection or inflammation increases hepatic synthesis of many secreted proteins involved in host defense mechanisms.²⁶

Herein, we examined the expression of the following genes: Rarres2, Camp, Defb1, Defb3, and Defb14, which encode chemerin, CRAMP, mBD-1, mBD-3, and mBD-14, respectively. We found that the expression of genes encoding chemerin, CRAMP, and mBD-1 in skin is mainly regulated by daily changes in light and darkness, while their expression in the liver is clock-controlled. Defb3 and Defb14 in the skin seem to be regulated by a circadian clock. Moreover, we found that the susceptibility of the skin to bacterial infections depended on the time of the day and was correlated with the expression level of AMP genes.

2 |. MATERIALS AND METHODS

2.1 |. Animals

C57BL6 strain mice (males, 5–6 weeks old) used in our studies were habituated for 2 weeks in LD12:12 (12 h of light and 12 h of darkness) conditions (light intensity: 60 lx) at 22°C and 50% humidity and then divided into two groups, namely LD (n = 80) and DD (n = 80). Animals in the first group were maintained for an additional 10–15 days in LD12:12, while animals in the second group were habituated for the next 2 weeks under the same temperature and humidity but in constant darkness (DD). The animals were fed with a standard diet and had access to water ad libitum. A small, dorsal area of the skin was shaved two days before the mice were killed. Skin and liver samples were collected every 3 h over a 24-h period at ZT1 (ZT—Zeitgeber time), ZT4, ZT7, ZT10, ZT13, ZT16, ZT19, and ZT22, where ZT0 is the beginning of the day and ZT12 is the beginning of the night, and at CT1 (CT—circadian Time), CT4, CT7, CT10, CT13, CT16, CT19, and CT22, where CT0 and CT12 denote the beginning of the subjective day and the subjective night in DD, respectively. All experimental procedures were performed in accordance with the principles of European animal research laws (European Communities Council Directive 2010/63/EU) and were approved by and were in compliance with the guidelines of the Second Local Ethical Committee on Animal Testing at the Institute of Pharmacology, Polish Academy of Sciences in Krakow.

2.2 |. RNA isolation, cDNA synthesis, and quantitative PCR

Dorsal skin fragments were quickly dissected and collected in stay-RNA™ (A&A Biotechnology, Poland). Liver samples were prepared, frozen in liquid nitrogen, and subjected to total RNA isolation using a Total RNA Mini Kit (A&A Biotechnology). cDNA synthesis was carried out using NxGen M-MuLV Reverse Transcriptase (Qiagen GmbH, Hilden, Germany) with random primers (Thermo Fisher, Waltham, MA, USA). Gene expression was examined using a StepOnePlus Real-Time PCR System and SYBR Green containing universal PCR master mix (A&A Biotechnology) in the presence of primer sequences, as listed in Table 1. Amplification was performed using 10 min of initial denaturation at 95°C followed by a three-step PCR of 40 cycles of 15 s at 95°C (denaturation), 30 s at 59°C (annealing), and 30 s at 72°C (extension). Product specificity was assessed by melting curve analysis, and selected samples were run on 1% agarose gels for size assessment. A standard curve was used to calculate the gene expression level.

TABLE 1.

Primer sequences used for quantitative PCR (qPCR) assessment

Gene symbol	Protein name	Function	Primer sequences	Exon location	Primer efficiency	Ampiicon size
Eif2a	eukaryotic translation initiation factor 2A	protein translation	F: CAACGTGGCAGCCTTACA R: TTTCATGTCATAAAGTTGTAGGTTAGG	4 5	1.07	74
Utp6c	U3 small nucleolar RNA-associated protein 6 homolog	Rn18 s biogenesis	F: TTTCGGTTGAGTTnrCAGGA R: CCCTCAGGTTTACCATCTTGC	17 18	1.15	75
Rarres2	retinoic acid receptor responder protein 2 (chemerin)	multifunctional, antimicrobial activity	F: TACAGGTGGCTCTGGAGGAGTTC R: CTTCTCCCGTTTGGTTTGATTG	2 3 and 4	1.13	195
Defb1	beta-defensin 1	antimicrobial activity	F: GGTGTTGGCATTCTCACAAG R: TTTACAATCCATCGCTCGTC	1 and 2 2	1.07	196
Defb3	beta-defensin 3	antimicrobial activity	F: GTCAGATTGGCAGTTGTGGA R: GCTAGGGAGCACTTGTTTGC	2 2	1.05	170
Defb14	beta-defensin 14	antimicrobialactivity	F: CTTGTTCTTGGTGCCTGCT R: CGACCGCTATTAGAACATCGAC	1 2	1.08	144
Camp	Cathclicidin antimicrobial peptide	antimicrobial activity	F: CTTCAACCAGCAGTCCCTAGACA R: TCCAGGTCCAGGAGACGGTA	1 1	1.09	51
Perl	Period circadian protein homolog 1	Clock gene	F: GAAAGAAACCTCTGGCTGTTCCTA R: TGGTTGTACTGGGAATGTTGCA	15 16	1.00	92
Per2	Period circadian protein homolog 2	Clock gene	F: TGCTGGCAGAGAGGGTACACT R: GGTTGTTGTGAAGATCCTCTTCTCA	8 9	1.07	73
Cry1	Cryptochrome-1	Clock gene	F: CACCATCCGCTGCGTCTATA R: CTCAAGACACTGAAGCAAAAATCG	1 1 and 2	1.11	94
Cry2	Cryptochrome-2	Clock gene	F: CGTGGAGGTGGTGACTGAGA R: CTGCCCATTCAGTTCGATGA	3 4	1.06	70
Arntl	Aryl hydrocarbon receptor nuclear translocator-like protein 1	Clock gene	F: GAAGGTTAGAATATGCAGAACACCAA	transcipt variant 5&4: exons 7 and 8; transcipt variant 3& 1: exons 6 and 7; transcript variant 2:exons 3 and 4	1.02	82
			R: TCCCGACGCCTCTTTTCA	transcipt variant 5&4: exon 9; transcipt variant 3&1: exon 8; transcript variant 2:exon 5
Clock	circadian locomoter output cycles protein kaput	Clock gene	F: ACGGCGAGAACTTGGCATT	transcript variant 3: exon14; transcript variant 2&1: exon 15	1.08	83
			R: TGATACGATTGTCAGACCCAGAA	transcript variant 3: exon15; transcript variant 2 & 1: exon 16

2.2.1 |. Preparation of quantification standards

PCR products of amplified primers (listed in Table 1) resulting from the conventional PCR reaction were gel-excised and purified using a gel purification kit (Qiagen). DNA concentration was measured using a NanoDrop 2000 Spectrophotometer (Thermo Fisher). The number of copies of target template was later calculated on the basis of the Avogadro number assumption, and dilutions, ranging from 10¹⁰ to 10⁵ of these standards, were prepared. The copy number values for the unknown samples were then quantified by comparison with the obtained curve.

The Microsoft Excel-based application NormFinder was used to analyse the expression stability of commonly used reference genes. Based on this analysis, Eif2a and Utp6c were selected as housekeeping genes for normalising RNA expression in RT-qPCR.²⁷ The number of target gene copies was normalised to the geometric mean of these housekeeping genes.

2.3 |. Topical skin infection

Seven- to ten-week-old male mice were housed under pathogen-free conditions in a facility at the Malopolska Centre of Biotechnology, Jagiellonian University, Krakow. Mice were topically infected with Staphylococcus aureus strain 8325–4 as per previously described methods.²⁷ We used S. aureus strain 8325–4 as a topical skin infection model in the context of antimicrobial chemerin activity.^28,29 Therefore, to correlate skin susceptibility to infection with skin chemerin levels, it is important to use the same strain.

Briefly, the mice were anaesthetised by an intraperitoneal injection of ketamine and xylazine mixture in sterile water (ketamine 100 mg/kg; xylazine 10 mg/kg). The backs of the mice were shaved with an electric razor, and the remaining hair was removed by the application of a depilatory cream. The skin was rinsed with sterile water and dried with a paper towel. The next day, mice were anaesthetised as described above, and a small dorsal area of the skin was sterilised with ethanol. Two rubber rings with 7 mm inner diameters were subsequently attached to the backs of mice below the shoulder blades using an ethyl cyanoacrylate-based adhesive. In this model, two rubber rings on either side of the midline allow the application of bacteria and PBS to each mouse (each mouse as its own control). The skin within the side of the rings was punctured six times using a syringe needle (0.3 × 8 mm), and the rings were covered with Opsite Flexigrid. S. aureus (CFU: 1 × 10⁴, 5 × 10⁴, or 1 × 10⁵) in 50 μl of PBS was injected through the Opsite Flexigrid into the cavity formed by the rubber rings. The rings injected with sterile PBS were used as controls. The following two groups of mice were subjected to the treatment: the first group was treated at ZT16 and the second at ZT22 and mice were killed after 6 h (at ZT22 and ZT4, respectively) by intraperitoneal injection of a 1:1 (vol/vol) mixture of ketamine and xylazine, followed by cervical dislocation. We performed three independent experiments, each involving 2–4 mice per experimental point, and two treatment sites (two independent cavities, one containing PBS) per mouse.

To evaluate the CFU, bacterial suspensions and PBS were collected using a syringe needle from the cavity formed by the rubber rings; then, they were serially diluted, plated onto tryptic soy agar (3% tryptic soy broth and 1.5% agar in water), grown at 37°C, and counted after 18 h. No bacteria were retrieved from the skin treated with PBS only (CFU = 0).

2.4 |. Statistical analysis

For statistical analysis, one-way ANOVA followed by the Tukey's post hoc test, non-parametric ANOVA followed by the Kruskal–Wallis test, Dunn's post hoc test, or two-tailed Studenťs t test were performed using the GraphPad Prism 7.05 software (La Jolla, CA, USA).

3 |. RESULTS

3.1 |. AMP gene expression in the skin under light-dark and constant darkness conditions

Expression profiles of the Rarres2 gene encoding chemerin and genes of CRAMP and β-defensin-1, 3, 14 (Defb1, Defb3, Defb14) in the skin under 12-h light/12-h dark conditions (LD12:12) (Figure 1A) showed that the expression of Rarres2 was rhythmic with a peak at 4 h from the beginning of the dark/night (ZT16) and is minimum at the end of the night (ZT22). Moreover, a similar cyclic expression was observed in the case of Defb1 and Camp with highest mRNA levels at 4 h after the beginning of the night (ZT16) and lowest levels at ZT22 (Defb1) and ZT7 (Camp). In the case of Rarres2 mRNA, the second peak was observed at ZT1. Since chemerin exhibits many functions, we hypothesise that it may also have important functions during sleep, such as wound healing.³⁰

FIGURE 1.

Antimicrobial peptide gene expression in the skin under light–dark (A) and constant darkness conditions (B). The skin mRNA levels of the following genes encoding AMPs were examined: chemerin (Rarres2), β-defensin 1 (Defb1), β-defensin 3 (Defb3), β-defensin 14 (Defb14), and cathelin-related antimicrobial peptide (Camp). The results are expressed as mean + SD (n = 10 mice per every time point). On the y axis, we present ratio of the number of copies of target genes (quantity) to geometric means of housekeeping genes. A standard curve was used to calculate the gene expression level. Statistically significant differences between time points are indicated as ***p < 0.001; **p < 0.01; *p < 0.5

The expression profile of the Defb14 mRNA was maximum at 4 h after the beginning of day/light period (ZT4) with a minimum expression together with Rarres2 and Defb1 mRNAs at the end of the dark period (ZT22). There was a lack of statistically significant changes in Defb3 mRNA level during 24 h in LD12:12. The same antimicrobial peptide genes were examined in constant darkness (Figure 1B). The chemerin-, mBD-1-, and CRAMP-encoding genes did not show the rhythmicity in mRNA level under DD. The expression profile of Defb14 was rhythmic in DD with maximum at CT7 and minimum at CT16. The robust circadian rhythm in mRNA level was also observed in the case of Defb3 with maximum and minimum at CT7 and CT22, respectively.

To explore if the molecular clock in the skin was functional in the animals studied, we examined expression of the following core clock genes: Per1, Per2, Cry1, Clock, and Arntl (Figure S1). For all these genes, the expression profiles showed oscillations in both LD12:12 (Figure S1A) and DD (Figure S1B) conditions; however, differences in Per1 mRNA level between time points were not statistically significant. In the case of Per2 and Cry1, their highest mRNA levels were observed at ZT22 (the additional Cry1 peak was also observed at ZT16) in LD12:12 and at CT13 and CT16 in DD Arntl and Clock mRNA levels were highest at ZT22 (LD12:12) and at CT19 and CT22 (DD). These changes were consistent with the previously published expression patterns of the same clock genes.⁹

3.2 |. Antimicrobial peptide gene expression in the liver under light-dark and constant darkness conditions

Subsequently, we examined expression profiles of the same AMP genes in the liver and only two genes, encoding chemerin and mBD-1, showed daily and circadian changes in mRNA expression levels (Figure 2). The expression of Rarres2 was rhythmic in LD12:12 and DD. Under both conditions, maximum expression was observed at ZT16/CT16, minimum expression was observed in the middle of the day (ZT4-ZT7) in LD12:12, and at the end of the subjective night (CT22) in DD. The maximum expression of Debf1 gene was detected at the beginning of the day (ZT1) in LD12:12 and at the beginning of the subjective day (CT1) in DD. These results indicate that some of the AMP genes expressed in the liver are not constitutively expressed but show daily oscillations. These rhythms are clock-controlled since they are maintained in constant darkness. Since a peripheral circadian clock resides in the liver, we also examined expression of the two clock genes Cry1 and Arntl in liver samples (Figure S2). The expression profiles of both genes were similar in LD12:12 and DD conditions, with maximum of Cry1 mRNA at ZT19 (LD12:12) and at CT16–19 (DD) and Arntl at ZT19 (LD12:12) and at CT19–22 (DD). These results were consistent with the previously published Cry1 and Arntl expression patterns in the liver of mice.^31,32

FIGURE 2.

Antimicrobial peptide gene expression in the liver under light–dark (A) and constant darkness conditions (B). The liver mRNA levels of chemerin (Rarres2) and β-defensin 1 (Defb1) were examined. The results are expressed as mean + SD (n = 10 mice per every time point). On they axis, we present ratio of the number of copies of target genes (quantity) to geometric means of housekeeping genes. A standard curve was used to calculate the gene expression level. Statistically significant changes are indicated as ***p < 0.001; **p < 0.01; *p < 0.5

3.3 |. Bacteria survival in the experimental topical skin infection

To determine if the daily rhythm of three AMP genes correlated with skin defense against bacterial infection, mice were topically infected with S. aureus. For this treatment, we selected two time points in LD12:12, ZT16 and ZT22, which represented respective maximum and minimum Rarres2, Camp, and Defb1 mRNA levels in the skin. Bacterial loads that were recovered from the skin surface 6 h later, after the application at ZT16 or ZT22, were measured by a colony-forming assay (Figure 3). Based on pilot results, we selected three bacterial inocula (10⁴, 5 × 10⁴, and 10⁵ CFU) that allowed the mice to partly, but not completely limit bacteria growth on the skin at ZT16, 6 h after infection. As shown in Figure 3, the survival of S. aureus was significantly reduced at ZT16 compared to the survival at ZT22 in all tested concentrations. Locomotor activity of mice shows the robust circadian rhythm with high activity during the night when mice are at increased risk of skin injury and consequent bacterial infection.³³ We conclude that the increased level of AMPs in the skin during high locomotor activity of mice enables higher skin defense against bacterial skin infections.

FIGURE 3.

Bacterial survival in experimental topical skin infection. Mice were topically infected with 10⁴ CFU (A), 5 × 10⁴ CFU (B), and 10⁵ CFU (C) of S. aureus 8325–4 at ZT16 and ZT22. Data points indicate the colony-forming units of bacteria recovered from the skin surface 6 h after application, with each data point representing one cavity and a horizontal line indicating the mean value in each group (n = 5–10 total number of mice per experimental group from three independent experiments). Statistically significant differences are indicated as *p < 0.05; **p = 0.0054

4 |. DISCUSSION

Skin is permanently exposed to a wide variety of harmful microorganisms and, when intact, it provides protection from infection. The risk to pathogen exposure depends on time of the day, and one of the lines of immunological defense in the skin includes antimicrobial peptides. Genes for a few of them are constitutively expressed, while production of other antimicrobial peptides is a response to bacterial infection or proinflammatory cytokine exposure.^14,19,34 Herein, we demonstrated that expression of several skin AMP genes showed daily and circadian changes to adapt animals to cyclical changes in the environment. In case of Rarres2, Defb1, and Camp, their mRNA levels were highest at 4 h from the beginning of the night (ZT16), which correlated with the high locomotor activity of C57BL6 mice.³³ However, the rhythms in their mRNA levels were not maintained in DD, thereby indicating that these rhythms were exogenous or the clock input was weak because of masking by light. Since locomotor activity and exposure to pathogens are the same in DD as in LD, these rhythms are probably controlled mainly by daily changes of light and darkness.

Rarres2 mRNA also exhibited a second peak at the beginning of the day, which could be connected to other functions of chemerin rather than its antibacterial activity because we did not observe increased locomotor activity at ZT1. However, Sheiermann et al (2018) have reported that chemerin simulates the migration of immune cells and the immune system shows daily and circadian oscillations in many processes and cells.^35,36

In contrast to Rarres2, Defb1, and Camp, the Defb3 mRNA rhythm was suppressed in LD12:12 but observed in DD. The mRNA level of Defb14 was the highest during the day in LD12:12, when mice slept, and this rhythm was maintained in DD.

The locomotor activity of mice is regulated by the circadian clock, and, in the C57BL6 mouse strain, it starts just after the beginning of night and continues for about 6–8 h during the night.³³ It is possible that some of the AMPs may protect animals during sleep and predict changes in the environment that are provided by the circadian clock. It is interesting that Rarres2, Defb1, and Camp genes show daily rhythms in the expression, while the rhythms of Defb3 and Defb14 mRNAs are circadian. This indicates that the immune system is under the control of direct changes in the environment and by the clock. The expression of Rarres2, Defb1, and Camp genes in the skin is an example of processes affected by daily changes of light. Daily and circadian changes in immune responses evolve to be correlated when the highest risk for skin injury and infection is encountered during the active phase, while in the rest phase they are decreased for tissue repair after inflammation and for energy conservation.³⁷ However, some AMPs are important during this time too. It has been demonstrated that several processes in the immune system work in a cyclical manner including phagocytosis, cytokine expression, and immune cell traffic.^19,38,39

AMPs act as skin protection factors either constitutively or in a regulated manner.⁴⁰ β-defensin 1 has been suggested to be the most important antibacterial defensin because of its constitutive expression in most tissues.^41,42 It has also been demonstrated that up-regulation of human DEFB1 is achieved by non-inflammatory pathways.⁴³ Moreover, Sherman and Froy have shown that regulation of hBD-1 in human cells is mediated by the circadian clock.⁴¹ Our results showed that, in the murine skin, Defb1 exhibited a peak of mRNA at ZT16, which was at the same time as that observed with two other AMP genes, Rarres2 and Camp. These data suggest a similar regulatory mechanism for these three peptides. CRAMP belongs to a family of cathelicidins¹⁹ that are highly expressed during bacterial skin infections or after physical skin injury. However, the molecular mechanism regulating this process is unclear.⁴³ Under normal conditions, CRAMP is expressed only at low concentrations in epithelia and is upregulated in response to external insults like wounds, UV irradiation, epidermal permeability, and barrier abrogation.^19,44,45 Here, we found that Camp expression oscillated during the day without exposure to bacteria or other harmful factors.

Chemerin is a multifunctional protein that has structural similarity to cathelicidins and is recognised as an antimicrobial protein.⁴⁶ In normal skin, chemerin is primarily expressed in epidermal keratinocytes and is upregulated by bacteria and the acute phase of cytokines.²⁹ Chemerin is also present in plasma in a nanomolar range.⁴⁷ It shows daily oscillations in human blood⁴⁸ and expresses the day/night pattern in murine blood with a trough during the night and peak during the day.⁴⁹ We also assessed chemerin mRNA levels in the serum of C57BL6 mice and observed the daily rhythm (data not shown).

The AMPs tested in the present study demonstrated a strong antibacterial potential against S. aureus. Previously, we demonstrated that recombinant human chemerin and chemerin-derived antimicrobial peptide exhibited antimicrobial activity against different strains of S. aureus,^28,46 whereas mice deficient with chemerin demonstrated higher counts of viable S. aureus associated with the epidermis in an experimental model of skin infection.²⁸ It has also been reported that CRAMP inside neutrophils and mBD-1 synthetic peptide acts against S. aureus.^50,51

Similar to the case of other circulating serum proteins, the liver may be a primary source of circulating blood chemerin. We found that Rarres2 expression in the liver exhibited a peak of mRNA expression at ZT16, same as that observed in the skin. We also observed the daily rhythm of Defb1 mRNA in the liver with its maximum at ZT1. The liver is perceived as an organ required for metabolic activity, nutrient storage, and detoxification, but not as an organ of the immune system. However, the liver is responsible for the production of acute phase proteins, cytokines, chemokines, and complement components. It is also an organ where diverse populations of immune cells reside. Even in healthy organisms, the liver is constantly exposed to dietary and commensal bacterial products with inflammatory potential.⁵² Several processes in the liver such as cholesterol synthesis/metabolism, amino acid regulation, drug and toxin metabolism, the citric acid cycle, glycogen and glucose metabolism, and a wide variety of cellular processes have been studied, and their circadian rhythms have been reported.⁵³

Several studies have shown that the infection susceptibility in different epithelial organs is dependent on time of the day. It has been reported that lung epithelium cyclically responds to lipopolysaccharide and Streptococcus pneumonia challenges.⁵⁴ Additionally, the gut epithelium defense against microbes is dependent on the circadian expression of defensins⁵⁵ and the pattern recognition receptors in intestinal epithelial cells peak at the transition between the night and the day.⁵⁶ Moreover, it has been reported that mice exposed to Salmonella during the day showed increased colonisation levels and pathology scores compared to the mice exposed to the pathogen at night.⁵⁷ In our study, we found that the susceptibility to skin infections depended on time of the day. We observed that bacterial survival in a topical skin infection model was higher after the treatment at ZT22 (minima of Rarres2, Defb1, and Camp gene expression) than that at ZT16 (maxima of Rarres2, Defb1, and Camp gene expression). These findings support the theory that direct light exposure promotes immunity during the night, when mice are active exploring the environment and are most likely to encounter pathogenic microorganisms.

In conclusion, in the present study, we showed that expression of genes encoding AMPs in the skin oscillated during the day, and these rhythms were generated by a circadian clock and daily changes of light and dark conditions. This mechanism of AMP gene cyclic expression affects the survival of bacteria on the surface of the mouse skin, thus protecting animals during their high locomotor activity. Another group of AMP genes with higher expression during sleep may be involved in healing skin damage, which appear during the exploration and high activity time before sleep. These results add more evidence which indicate that the innate immune system is important in antibacterial defense and this system is more efficient when is controlled by the circadian clock. Similar oscillations in immune responses to bacterial infections have been described in Drosophila and mice.⁵⁸ Moreover, disruption of the circadian clock due to clock gene mutations exerts effects on immunity.⁵⁸ In humans, disruption of the clock by shift work may disturb processes in the immune system and antibacterial immunity, thereby leading to skin infections.

Similar antibacterial protection by AMPs may also exist in humans; however, mice are nocturnal animals and humans are active during the day. Thus, the pattern of AMP expression will be different. Circadian rhythms in cell proliferation have been found in skins of mice, rats, and humans.⁵⁹ Hundreds of rhythmically expressed genes have been identified in the epidermis of mice and humans,⁶⁰ but expression of the genes occurs in different phases of the day/night cycle. Wu et al (2018) compared human cyclically expressed genes with published time-series skin data from mice and found a strong correlation in circadian phase across species for both transcripts and pathways. For example, Arntl peaks at 8:00–9:00 PM in humans and at ZT19–ZT23 in mice (Wu et al, 2018, our study). It indicates that treatment of skin diseases should be correlated with circadian rhythms of treated individuals, and the rhythms of pathogens may also be considered. Although circadian rhythms have not been detected in S. aureus, other prokaryotes and fungi show endogenous oscillations.⁶¹ This implies that circadian rhythms should be studied not only for elucidation of host response to the microbiota but also for pathogens and for exploration of new methods of skin disease treatment, especially those involving drug-resistant microorganisms.

Supplementary Material

Supplemental fig.1

FIGURE S1. Clock gene expression in the skin under light-dark (A) and constant darkness conditions (B). The skin mRNA levels of the following clock genes were examined: Per1, Per2, Cry1, Arntl, and Clock. The results are expressed as mean + SD (n = 10 mice) per each time point. A standard curve was used to calculate the gene expression level. Statistically significant differences are indicated in the table below figures as ***p < 0.001; **p < 0.01; *p < 0.5.

Supplemental fig. 2

FIGURE S2. Clock gene expression in the liver under light–dark (A) and constant darkness conditions (B). The liver mRNA levels of the following clock genes were examined: Per1, Per2, Cry1, Arntl, and Clock. The results are expressed as mean + SD (n = 10 mice) per each time point. A standard curve was used to calculate the gene expression level. Statistically significant differences are indicated in the table below figures as ***p < 0.001, **p < 0.01, *p < 0.5.