Modulation of Circadian Rhythms Affects Corneal Epithelium Renewal and Repair in Mice

Abstract

Purpose: In mammalian corneal epithelium, mitosis shows a distinct circadian pattern. However, how this circadian pattern is maintained, and how it or its disruption influence renewal and regeneration remain unclear.

Methods: C57BL/6 mice were maintained under 12-hour light/12-hour dark (LD), 12-hour light/12-hour light (LL), 12-hour dark/12-hour dark (DD), or reversed LD (DL, 12-hour dark/12-hour light; jet-lag defined as a shift of 12 hours) conditions. Mitotic cells in corneal epithelium were enumerated and analyzed via immunofluorescence at different zeitgeber times (ZTs). Expression of core clock genes (Clock, Bmal1, Period2, Cry1, and Rev-erbα) was qualified via quantitative RT-PCR. The rate and quality of healing at different ZT times and after administration of two small-molecule modifiers of the circadian clock, KL001 and SR8278, was evaluated.

Results: In this study, photic cues were found to influence the 24-hour rhythm of corneal clock gene expression and epithelial cell mitosis in mice. Disruption of the circadian clock by exposure to constant light, constant dark, or jet-lag conditions modified the normal 24-hour patterns of corneal epithelial mitosis and corneal clock gene expression. The time of day of wound occurrence affected the rate and quality of corneal healing, with both of these parameters peaking during the more mitotically active hours of the morning. The two small-molecule modifiers of the circadian clock, KL001 and SR8278, had negative and positive effects on corneal wound healing, respectively.

Conclusions: Circadian rhythms significantly influence corneal epithelium renewal and repair in mice. Our findings reveal possible opportunities for biological rhythm-based interventional strategies to control corneal healing and restore corneal homeostasis.

The circadian clock is an evolutionarily conserved mechanism that generates the endogenous ∼24-hour rhythm in organisms. Nearly all tissues and cell types possess intrinsic molecular clocks. The basis of molecular clocks is a set of genes called the clock genes. The proteins transcribed and translated from these clock genes form a negative oscillator loop that creates a 24-hour oscillation. Within this oscillator is a heterodimeric core transcription factor consisting of two basic helix-loop-helix PER-ARNT-SIM (bHLH-PAS)-domain protein subunits that are known as CLOCK and BMAL1. This heterodimer binds the enhancer box, a DNA response element found in some eukaryotes that acts as a protein-binding site, and induces the expression of the repressors Period (Per) 1 and 2 and Cryptochrome (Cry) 1 and 2. After translation, Per and Cry translocate to the nucleus, where they disrupt the BMAL1:CLOCK complex and inhibit their own expression, resulting in rhythmic gene expression.^1,2 Other regulatory loops have also been described, including one that involves the nuclear receptors Rev-erbα and Rorα. Rev-erbα binding downregulates Bmal1 expression, while Rorα binding upregulates Bmal1 expression.³ 

The cornea, as a peripheral tissue, is the transparent anterior segment of the eyeball that controls and focuses light entering the eyeball, contributing 65% to 75% of the total refractive power of the eyeball.⁴ Corneal epithelium consists of five to seven layers of nonkeratinized squamous epithelial cells that form a physical barrier against microbial invasion. The mammalian cornea is in a state of repetitive renewal, and cells lost from the corneal epithelium must be promptly replaced by dividing epithelial cells from below. Thus, the maintenance of normal corneal epithelium and rapid healing after wounding are critical for normal corneal integrity and transparency. Earlier research utilizing ³H-thymidine incorporation and the mitotic index confirmed that corneal epithelial mitosis is influenced by the 24-hour circadian clock.^5–9 Furthermore, some drugs, such as isoproterenol and dexamethasone, act to phase-shift or suppress the rhythm of corneal epithelial mitosis.^10–12 The time of day has been shown to affect the time course of corneal epithelial cell mitosis and the process of corneal epithelial wound healing.¹³ In addition, several recent studies have shown that the cornea expresses vital clock genes, such as Per1 and Per2, and exhibits circadian oscillations both in vivo and ex vivo.¹⁴ Melatonin, a clock-controlled humoral factor, is involved in the entrainment of circadian rhythms¹⁵ and was found to affect the rhythm of expression of a Per2-luciferase reporter construct in the cornea through interaction with receptors expressed in the corneal epithelium.¹⁶ Glucocorticoids also entrain Per1 expression in the cornea.¹⁷ However, it is unknown if interruption of the physiological circadian rhythm, for example by jet lag, causes changes in mitosis and the expression patterns of other main clock gene components in normal corneal tissues. 

Recent research highlights a close association between circadian rhythms and tissue development, physiological turnover, and regenerative repair.¹⁸ Clock gene-deficient animals show abnormal wound repair of the skin¹⁹ and muscle²⁰ and interference with normal rhythms significantly delays wound healing of bowel anastomosis.²¹ Recently, we found that constant exposure to light significantly delays the process of re-epithelialization and leads to greater inflammation of the cornea after corneal epithelial abrasion.²² However, it is not known how corneal epithelial renewal and regeneration change after disturbance of the normal circadian rhythms. 

In recent years, many small molecules that modulate clock functions have been discovered or designed. These small-molecule modifiers have been used as tools to explore basic circadian biology in attempts to control endogenous circadian rhythms and have been considered as potential treatment options for diseases related to clock dysfunction.^23–25 However, it remains unclear whether these drugs affect wound healing of the cornea through their alteration of the circadian clock. 

Although a considerable body of knowledge on the mitotic rhythm of the corneal epithelium is available, it remains unclear how this circadian pattern is maintained, and how it, or its disruption, influences renewal and regeneration. In this study, we examined whether the rhythmic control of mitosis and corneal clock gene expression are affected by various light schedules (constant light, constant dark, reversed light/dark) and jet-lag conditions. Additionally, we checked how circadian timekeeping influences the healing rate following mechanical wounding of the cornea at different times of the day. Finally, we determined the effects of small molecular circadian modifiers KL001 and SR8278 on inhibition or promotion of corneal wound healing. Our results support the hypothesis that interference with normal light cycles has significant negative effects on the homeostasis of epithelial mitosis. Further, we found that small-molecule modifiers of the circadian clock have the potential to modify the processes of corneal wound healing. Therefore, our data provide compelling evidence that tampering with normal environmental lighting cycles influences the homeostasis of the corneal epithelium. 

Materials and Methods

Animals

Male C57BL/6 mice (free of eye disease) were purchased from the Guangdong Province Animal Experimental Center. All animals used in the present study were 8 to 12 weeks old and were housed in our facility following the guidelines described in the ARVO Statement for the Use of Animals in Vision and Ophthalmic Research. All experimental protocols were approved by the Jinan University Medical School Animal Care and Use Committee. Control mice were maintained at 20° to 22°C on a schedule of lights on at 6 AM/lights off at 6 PM, with ad libitum access to standard rodent chow. Light intensities ranged from 110 to 210 μW/cm². Experimental mice were housed either under conditions of 12-hour light/12-hour dark (LD), 24-hour light (LL), 24-hour dark (DD), or reversed LD (DL, 12-hour dark/12-hour light, 12-hour shift jet-lag by advancing the LD cycle by 12 hours; see Figs. 2C, 3B, bottom). Mice were kept on the LD schedule for 7 days and then transferred to DD or LL cycles for an additional 3 days. For DL cycle experiments, animals were switched to a short 12-hour time difference maintained for 24 hours, to mimic acute jet-lag conditions, or for 3 weeks to investigate the possibility of adjustments in the circadian rhythm to long-term jet-lag conditions. The time of day is expressed in zeitgeber time (ZT; where lights are turned on at ZT0 and turned off at ZT12). 

Figure 1

 

Mitosis in mouse corneal epithelial cells on the LD light schedule. (A) Total mitotic cells in mouse corneal epithelium maintained on an LD schedule were counted from limbus to limbus every ZT point at 1-hour intervals for 24 hours (n = 6 corneas at each time point). Active and resting phases of mitosis are highlighted in panel A. (B) Representative figure showing mitotic cells labeled with DAPI in the paralimbal field during active (ZT5; left) and resting (ZT17; right) phases (40×, scale bar = 50 μm). (C) Distribution of mitotic epithelial cells across the corneal diameter at time points during the active phase (n = 6 corneas at each time point).

Figure 2

 

The effects of photic cues on the circadian rhythm of corneal mitosis. (A, B) Circadian fluctuations in the number of mitotic cells. Mice were placed on LD, LL, or DD schedules. The number of dividing cells in the corneas (A) (n = 6 corneas at each time point) and the total number of dividing cells counted per 24-hour cycle (B) (n = 6) are shown. (C, D) Circadian fluctuations in the number of mitotic cells under LD (red), 12-hour jet lag (blue), and 3-weeks jet lag (green) (C) (n = 6 corneas at each time point), and the total number of dividing cells counted per 24-hour cycle (D) (n = 6). The arrowhead in C shows the initiation of the light shift. *P < 0.05, **P < 0.01.

Figure 3

 

Circadian patterns of clock gene expression after exposure to different lighting conditions. (A, B) Quantitative RT-PCR of clock genes in mouse corneas exposed to conditions of LD (red), DD (blue), and LL (green) (A), and in mouse corneas exposed to conditions of LD (red), 12-hour jet lag (blue), and 3-weeks jet lag (green) (B). Gene expression levels relative to the housekeeping gene Gapdh mRNA levels. Gray areas indicate dark phases. *P < 0.05, **P < 0.01. n = 3 independent samples containing 6 mice each.

Whole-Mount Immunofluorescence and Imaging

Immunostaining was performed on whole-mount corneas as described previously.^26–30 Briefly, normal or wounded corneas with the complete limbus were excised, fixed in 2% paraformaldehyde for 40 minutes, and permeabilized in 0.1% Triton X-100/1% bovine serum albumin for 20 minutes followed by three washes in phosphate-buffered saline (PBS). Radial cuts were made in the cornea so that it could be flattened under a cover slip, and the cornea was mounted overnight with Mowiol mounting medium (Sigma-Aldrich Corp., St. Louis, MO, USA) containing 3 μM 4′,6-diamidino-2-phenylindole (DAPI; Sigma-Aldrich Corp.) to label nuclei. Mitotic cells were also stained with DAPI.^31,32 Images were analyzed using a DeltaVision Elite Microscopy System (GE Healthcare Life Sciences, Pittsburgh, PA, USA). 

Reverse Transcription Quantitative PCR (qRT-PCR)

Total RNA was extracted from fresh corneas containing complete limbus using the RNAsimple Total RNA Kit (Tiangen Biotech, Beijing, China), and cDNA was synthesized using the ReverTra Ace qPCR RT Kit (Toyobo, Osaka, Japan) following the manufacturers' instructions. Quantitative PCR was conducted on an Applied Biosystems 7900HT Fast Real-Time PCR System with SYBR Green Master Mix (Toyobo) and 1 mM primers (Table). Thermal cycles included denaturation for 1 minute at 95°C, followed by 45 cycles of 10 seconds at 95°C, 15 seconds at 60°C, and 15 seconds at 72°C. Relative expression levels were determined using the comparative C_T method, and target gene mRNA levels were normalized to those of Gapdh. 

Table

 

Primers for qRT-PCR

 

Corneal Wound-Healing Model

The central cornea was mechanically abraded as previously described.^31,33,34 Briefly, animals were anesthetized using an intraperitoneal injection of sodium phenobarbital (25–50 mg/kg; Beyotime Institute of Biotechnology, China). The central epithelium was marked with a 2-mm trephine and removed using a golf club spud for refractive surgery (Accutome, Malvern, PA, USA) under a dissecting microscope. Wound-healing rates were measured by counting the number of dividing epithelial cells, photographing the ocular surface at a known scale labeled with sodium fluorescein, and calculating the surface area of the wound with Photoshop CS4 at each time point. The numbers of nuclei appearing as pairs were used to quantify dividing cells in normal or wounded corneal epithelia. Nine fields spanning from limbus to limbus were counted using a DeltaVision Elite Microscopy System (GE Healthcare Life Sciences) at a magnification of 40×. At least six corneas were counted at each time point. The mean number of mitotic cells observed on the two lines of each cornea was used as the average number of dividing cells for that cornea, as described previously.^27,29,30 At 6 to 48 hours after injury, animals were killed, and the cornea with the complete limbus was excised and processed for immunohistological analysis as summarized above. 

Pharmacologic Agent Administration

KL001 and SR8278 were purchased from Tocris Bioscience (Bristol, United Kingdom). Both compounds were dissolved in a small volume of dimethyl sulfoxide (DMSO)³⁵ and diluted with PBS. Chemicals were administered intraperitoneally 6 hours before sample collection, and corneal wounding was conducted on an LD schedule. KL001 was injected at a dose of 5 mg/kg, and SR8278 was injected at 3 mg/kg, with each chemical being delivered in a total volume of ∼200 μL per mouse. Control animals were injected with PBS containing the equivalent volume of DMSO in place of the dissolved compound. 

Data Analysis

Data analysis was performed using SPSS 21.0. The 24-hour rhythms of mitotic cells were analyzed using 1-way analysis of variance (ANOVA). A factorial-design ANOVA with post hoc Tukey's tests was performed to compare data from independent samples. Wound closure rates and the numbers of dividing cells after injury were analyzed using 2-tailed Student's t-tests. P < 0.05 was considered statistically significant. Data are presented as the mean ± standard deviation. 

Results

Photic Cues Modulate Corneal Epithelial Cell Mitosis

To gain a comprehensive understanding of the mitotic rhythm of the corneal epithelium, mice were entrained to LD (12-hour light/12-hour dark cycles, lights turned on at 6 AM and turned off at 6 PM), and corneas were collected every hour after the 24-hour circadian cycle. The number of mitotic corneal epithelial cells, observed by microscopy as pairs of nuclei, fluctuated markedly between ZT19 and ZT7 (Fig. 1A, P < 0.01 in 1-way ANOVA). Five peaks in the number of dividing cells were observed, at ZT19, ZT22, ZT1, ZT5, and ZT7, with an overall maximum at ZT5 (Fig. 1A). The number of mitotic cells from ZT19 to ZT7 accounted for 78.88% of total mitotic cells in a 24-hour LD cycle; therefore, we defined the period from ZT19 to ZT7 as the active mitotic phase (Fig. 1A, green area). By examining nine microscopic fields along the diameter of the cornea, we found that the distribution of dividing cells in the cornea varied with time in the circadian cycle (Fig. 1C). At ZT22 and ZT1, more mitotic cells were observed in paralimbal fields (Field 2); the ratios of mitotic cells in Field 2 alone to the total mitotic cells from the nine fields from limbus to limbus were 39.40% and 38.40%, respectively. At ZT5 and ZT7, however, dividing cells were evenly distributed across all fields from limbus to limbus (Fig. 1C). In contrast to the fluctuating pattern in the active phase, the number of mitotic cells gradually declined from ZT7 to ZT18 (Fig. 1A). We defined this stage as the resting mitotic phase (Fig. 1A, orange area). 

To assess the effects of different lighting schedules on corneal epithelial mitosis, following 2 weeks of entrainment to LD conditions mice were exposed to constant dark (DD) or constant light (LL) conditions for 3 days (72 hours). Although the total number of dividing cells in the next 24-hour DD cycle after the 72-hour DD entrainment significantly declined when compared to the normal LD control group (Fig. 2B, 2-tailed t-test, P < 0.01), the oscillation curve for mitotic cells maintained a similar pattern (Fig. 2A). This observation is consistent with free-running circadian activity such as heart rate, motility, body temperature,³⁶ and vascular development³⁷ observed during constant darkness. Thus, this indicates that the mitotic circadian rhythm persists under conditions of constant darkness. In contrast, the oscillation of mitosis at ZT19-ZT4 was significantly suppressed relative to normal LD cycle treatments in the LL group (Fig. 2A, factorial-design ANOVA, P < 0.05), while the oscillation for mitotic cells from ZT8 to ZT18 was similar to LD control group (Fig. 2A). 

To test whether changes in light input are sufficient to alter the rhythm of corneal epithelial mitosis, we subjected mice to acute jet-lag stress by advancing the LD cycle by 12 hours³⁸ (Fig. 2C, bottom). During the next 24-hour cycle following the 12-hour jet lag, the number of epithelial mitotic events was significantly decreased at ZT4 and ZT22 compared to the control LD-entrained condition (Fig. 2C, factorial-design ANOVA, P < 0.01 and P < 0.05, respectively). In addition, the total number of mitotic cells per 24-hour cycle (counted at ZT1, 4, 7, 11, 14, 18, and 22) remained significantly lower than that seen in the control group (Fig. 2D, 2-tailed t-test; 12-hour jet-lag group compared with LD group, P < 0.01). Therefore, a short change in the light cycle was sufficient to induce marked alterations in the mitotic rhythm of the corneal epithelium. To observe whether longer-term changes in light input alter the rhythm of corneal epithelial mitosis, we subjected mice to jet-lag stress by advancing the LD cycle by 12 hours for 3 weeks (Fig. 2C, bottom module). Based on the ZT timing of the normal LD group, the diurnal oscillation in epithelial mitosis was phase-shifted forward by approximately 14 hours (normal mitotic peak time shifted from ZT4 to ZT18) in the 3-week jet-lag group relative to the control LD group (Fig. 2C). In addition, the total number of mitotic cells per 24-hour cycle (counted at ZT1, 4, 7, 11, 14, 18, and 22) was still significantly lower than that in the control group (Fig. 2D, 2-tailed t-test, 3-week jet-lag group compared with LD group, P < 0.01). These data suggest that after long-term exposure to jet-lag stress conditions, mitotic rhythms in the corneal epithelium adapt to the new light conditions. 

Photic Cues Modulate Corneal Clock Gene Expression

Given the circadian oscillation and photic modulation in corneal epithelial mitosis, we reasoned that clock genes controlling circadian physiology might be differentially expressed at different times of day and that their expression might be modified by photic cues. Thus, we examined the expression profiles of five main clock genes in the corneal epithelium of mice exposed to LD conditions at 4-hour intervals over a 24-hour period via qRT-PCR. Clock, Bmal1, Cry1, Per2, and Rev-erbα mRNA levels all clearly showed diurnal changes (Fig. 3, 1-way ANOVA, P < 0.01). Clock, Bmal1, and Cry1 mRNA levels peaked at the dark-to-light transition (approximately ZT1) and reached their lowest levels at the light-to-dark transition (approximately ZT13; Fig. 3). However, peak Per2 and Rev-erbα expression was observed during the light period (approximately ZT9), with the lowest expression levels during the dark period (Fig. 3). Taken together, these findings demonstrate that clock gene expression oscillates in the cornea under stable light-dark cycles. 

To test the hypothesis that photic cues affect clock gene expression, we exposed animals to LL and DD for 3 days. We then analyzed the diurnal expression of Clock, Bmal1, Cry1, Per2, and Rev-erbα mRNAs in the cornea (Fig. 3A). In response to LL and DD conditions, the peak Clock, Bmal1, Per2, and Rev-erbα expression levels were severely attenuated, whereas Cry1 levels were only slightly attenuated (Fig. 3A). Therefore, different light schedules result in CCG-specific changes in expression. Interestingly, following long-term (3-week) jet-lag conditions, the diurnal mRNA expression of Clock and Bmal1 remained low (approximately ZT1, ZT5, and ZT21; Fig. 3B). However, the diurnal expression of Cry1 and Rev-erbα recovered and adapted to the new light cycle (Fig. 3B). These findings, therefore, suggest that the exposure of mice to LL, DD, short-term DL, and long-term DL impair expression of the core clock genes in mouse cornea. Clock genes Cry1 and Rev-erbα may be critical for rapid adaptation to new environmental light conditions. However, future work will be required to establish any direct mechanistic links between changes in clock gene expression and changes in corneal mitosis or physiology. 

Time of Day at Wounding Influences Corneal Healing Dynamics

To determine if the time of day at which wounding occurs affects mitosis in the corneal epithelium, we wounded the corneas of mice exposed to control LD conditions in the morning (ZT21, ZT24, and ZT3) and afternoon/evening (ZT9, ZT12, and ZT15). We collected tissues at 3-hour intervals from 12 to 48 hours after injury and compared the number of dividing cells at the different time points. Significant differences in the number of dividing cells as a function of time after wounding was observed. Corneal epithelial cell mitosis peaked at 24 hours after wounding in the ZT24 group and at 30 hours after wounding in the ZT21 and ZT3 groups (Fig. 4A). In the ZT9, ZT12, and ZT15 groups, cell division peaked at 36 hours after wounding (Fig. 4A). Figure 4B displays the total number of mitotic epithelial cells counted at the seven time points from 12 to 48 hours after wounding. The numbers of mitotic cells in the ZT21, ZT24, and ZT3 groups were higher than those in the ZT9, ZT12, and ZT15 groups (Fig. 4B, 2-tailed t-test, P < 0.05). 

Figure 4

 

Effects of wounding time of day on healing under LD condition. (A) Circadian fluctuations in mitotic cell numbers in the cornea at the indicated time points after corneal abrasion wounding in the morning (ZT21, ZT24, and ZT3) or afternoon/evening (ZT9, ZT12, and ZT15) (n = 6 corneas at each time point). (B) Total number of dividing cells counted from 12 to 48 hours in corneas wounded in the morning or afternoon/evening (n = 6). (C) Fluorescence micrographs of healing over time following corneal epithelial abrasion at the indicated times (n = 6 mice). (D) Typical wound closure dynamics in mice wounded at the indicated times (n = 6 corneas at each time point). *P < 0.05, **P < 0.01.

Re-epithelialization is the process by which a wound is resurfaced with new epithelium, and the timing of this process dictates healing outcomes.³⁹ Moreover, rapid re-epithelialization prevents scar formation, resulting in full regeneration. Epithelial cell division is the main source of new cells for re-epithelialization after wounding. To evaluate the effects of wounding time of day on the epithelialization (wound closure) process, corneas were wounded by abrasion at the same time points mentioned above. Re-epithelialization was complete at 18 hours after morning wounding, but was not complete until 24 hours after afternoon/evening wounding (Fig. 4C). Wound size differed significantly between the morning and evening groups at 18 hours and 24 hours (Figs. 4C, 4D, 2-tailed t-test, P < 0.05). Therefore, consistent with earlier work in rats, the point at which wounding occurs during the circadian cycle affects corneal healing dynamics.¹³ Specifically, wounding in the morning leads to more rapid healing than wounding in the afternoon/evening. 

Pharmacologic Interference With the Circadian Clock Modulates Corneal Epithelial Mitosis and Wound Healing

KL001 and SR8278 are two recently discovered small molecules that interfere with circadian activity through different mechanisms. KL001 specifically acts on CRY proteins and results in the lengthening of the circadian period by blocking the proteasome-mediated degradation of CRY1 and CRY2 and stabilizing cryptochrome proteins.²⁵ In contrast, SR8278, a synthetic antagonist of Rev-erbα, inhibits Rev-erbα-mediated transcriptional repression.⁴⁰ To explore whether corneal epithelial mitosis could be influenced by pharmacologic interference with the circadian circuitry, mice were treated with KL001 or SR8278 under LD conditions. Mitotic dynamics and re-epithelialization after wounding were then examined. The number of mitotic cells was significantly reduced at five time points (ZT1, 5, 7, 19, 22), from ZT19 to ZT7, in the KL001 group compared to the vehicle-treated control group (Fig. 5A, factorial-design ANOVA, P < 0.01). Peak numbers were phase-shift advanced by 6 hours in the KL001 group, with peaks occurring at ZT23 (rather than at ZT5 as was observed in the control group). In the SR8278 treatment group, the numbers of mitotic cells at ZT19, ZT22, ZT1, and ZT7 were significantly reduced compared to those in the control group (Fig. 5A, factorial-design ANOVA, P < 0.01). At the other time points, however, the number of mitotic cells and the phase oscillation pattern of mitosis were not significantly changed (Fig. 5A). We also compared the total number of mitotic cells counted from ZT19 to ZT18 in the control and treatment groups (Fig. 5B). The total number of mitotic cells in the KL001-treated group was significantly lower than that in the control group (Fig. 5B, 2-tailed t-test, P < 0.01), whereas in the SR8278-treated group the number was not significantly changed. 

Figure 5

 

Effects of KL001 and SR8278 on mitosis and re-epithelialization after wounding under LD condition. (A) Circadian fluctuations in mitotic cell numbers in mouse corneas treated with DMSO in PBS (control), KL001, and SR8278 (n = 6 corneas at each time point). (B) Total number of dividing cells counted over a 24-hour cycle beginning at 6 hours after drug administrations (n = 6). (C) Fluorescence micrographs of corneal epithelial wound areas in the treatment groups at the indicated times after abrasion of the corneal epithelium (n = 6 mice). (D) Wound-closure dynamics and (E) mitotic cell numbers after wounding in the treatment groups at the indicated times after wounding (n = 6 corneas at each time point). (F) Total number of dividing cells counted at 12, 18, 24, 30, 36, 42, and 48 hours after wounding (n = 6). *P < 0.05, **P < 0.01.

We next tested the effects of the same two small molecules on mitosis and re-epithelialization after corneal epithelial abrasion. The corneas of mice housed under normal LD cycles were wounded at ZT12. In the KL001-treated group, the areas of the unrepaired wounds after 18 and 24 hours were significantly larger than those in the control group (Fig. 5C, 5D, 2-tailed t-test, P < 0.05). In the SR8278-treated group, the unrepaired areas at 12 hours after wounding were significantly smaller than those in the control group (Fig. 5C, 5D, 2-tailed t-test, P < 0.01). In addition, cell division peaked at 36 hours after wounding in the control group, but at 30 and 24 hours after wounding in the KL001- and SR8278-treated mice, respectively (Fig. 5E). Compared to the control group, the number of dividing cells was significantly decreased in the KL001 group at 36 hours after wounding (Fig. 5E, 2-tailed t-test, P < 0.01); however, it was significantly increased in the SR8278 group at 24 and 30 hours after wounding (Fig. 5E, 2-tailed t-test, P < 0.01). Additionally, we found that the total number of mitotic cells counted at the seven time points (12–48 hours after wounding) was significantly increased in the SR8278-treated group and significantly decreased in the KL001-treated group (Fig. 5F, 2-tailed t-test, P < 0.01). In summary, the data suggested that KL001 impairs mitosis and re-epithelialization after wounding, whereas SR8278 enhances healing dynamics. 

Discussion

In this study, we demonstrated a substantial role for the circadian clock in regulating corneal epithelial mitosis. Firstly, ambient environmental light modulated the time course and pattern of corneal epithelial mitosis and the expression of core clock genes. Secondly, the time of day at which wounding occurred affected the rate and quality of corneal healing. Lastly, two small-molecule modulators of circadian circuitry altered the circadian rhythm of corneal epithelial cell mitosis and affected wound healing. These observations provide insights into how the circadian clock might synchronize homeostatic mitosis in the cornea and wound repair after corneal epithelial abrasion. 

Alterations in ambient environmental light cycles not only affect the rhythmic expression of clock genes, but also interfere with diurnal physiological functions in peripheral tissues.^41–44 For example, exposure to LL and DD conditions can disrupt Bmal1 and Per2 expression and impair developmental neovascularization in zebrafish.³⁷ When mice are subjected to continuous light or to conditions of 12-hour jet lag, the circadian release of hematopoietic stem cells into the peripheral blood is markedly altered through sympathetic nervous system via circadian-controlled noradrenaline secretion.³⁸ Likewise, early studies in rats⁴⁵ and Japanese quail,⁴⁶ along with our present data, show that in a constant light environment, oscillations in corneal epithelial mitosis are significantly altered. These phase changes under different lighting conditions may be due to free-running of the clock or a change in the phase angle of the mitosis. Additionally, the decrease in number of mitotic cells under the constant condition (LL) may also be due to the visible light. Some evidence show that photons of visible light can damage DNA, membranes, and intracellular organs such as mitochondria and affect cellular respiration by producing reactive oxygen species and destroying cytochromes.^47,48 Whether visible light can inhibit the mitosis of corneal epithelial cells and its possible range of strength need to be further assessed. Based on these findings, we propose that when environmental lighting is switched from day to night, the mitotic rhythm and clock gene expression in the corneal epithelium can be adapted through ambient lighting adjustments. 

Several studies highlight the role of circadian rhythms in tissue regeneration.¹⁸ Genetic evidence shows that the pattern of wound healing in the skin is altered in circadian clock gene-deficient Per1/Per2^mut mice and Bmal1^−/− mice.¹⁹ The elimination of Period (clock-repressor) proteins results in fibroblast and keratinocyte hyper-proliferation, whereas the elimination of the Bmal1 and clock-activator protein results in decreased epidermal cell proliferation and highly disorganized tissue granulation.¹⁹ Stem-cell regeneration in fruit fly intestines and mouse skin tissues follows a circadian rhythm.^49,50 Disruption of clock components leads to division arrhythmia, stem cell aging, and delays in the healing process.¹⁸ In our study, we found that interruption of normal 24-hour rhythms by exposure to continuous light or a jet-lag model, even over a short time, leads to a significant reduction and pattern changes in corneal epithelial mitosis. These results indicate that environmental 24-hour cycles are essential to maintain a healthy cornea. Moreover, regardless of the time of injury, the majority of corneal repair occurs during the late phase of the light/dark cycle (Fig. 4). In fact, all groups exhibited the highest level of cell divisions immediately before the onset of darkness. This suggests that corneal wound healing occurs in a clock-controlled manner; however, further experiments will be required to determine the precise underlying mechanisms. 

The time of day at which an injury occurs also affects the healing processes. For example, when a partial hepatectomy is performed at different times of the day (ZT24 and ZT8), a distinct 8-hour time lag in mitosis, from 40 to 48 hours, is observed.⁵¹ The results from a kinetic analysis of zebrafish tail-fin cell proliferation after truncation also showed that the healing rate depends on the time of day of the injury. After an amputation at ZT24, the number of dividing cells 10 hours later was significantly greater than when amputation was performed at ZT12.³⁹ Similarly, the number of epithelial cells undergoing mitosis after corneal injury in rats induced at noon (ZT6) was significantly higher than the number observed after a corneal injury induced at midnight (ZT18).¹³ Our data expand on previous studies of the cornea, and we found that wounding the cornea during the active phase of corneal epithelial mitosis (ZT19–ZT7) is followed by rapid healing, including rapid re-epithelialization, and early entry into mitosis. In contrast, corneal wounding during the resting phase (ZT7–ZT18) is associated with slow healing (Fig. 6). Moreover, our study demonstrates that the process of corneal wound healing is remarkably influenced by an intrinsic circadian clock that allows corneal epithelial cells to synchronize waves of mitotic divisions and wound repair (Fig. 6). However, while mice are nocturnal, humans are diurnal. Some evidence indicates that the core circadian clock and clock output exhibit opposite phases in humans and mice.^52,53 Therefore, we predict that humans will exhibit a faster repair rate in the evening. Before recommendations can be made with regard to the preferred time of day for performing human corneal surgeries, tests on human corneal activity should be carried out with a particular focus on cell division as a function of the time of day. 

Figure 6

 

Proposed model of the circadian clock in corneal epithelial mitosis and wound healing. The level of DNA replication in corneal epithelial cells in mice is high in the morning and low in the afternoon/evening. As a consequence, injury in the morning is expected to lead to accelerated entry into mitosis and wound repair compared to injury in the afternoon/evening. Shading indicates dark periods.

KL001 and SR8278 are two small molecule circadian-interference modifiers that act upon different signaling pathways. We observed that KL001 exhibited a negative regulatory effect on the mitotic rhythm of the corneal epithelium and suppressed postwounding mitosis and re-epithelialization. In contrast, while we did not observe a clear effect of SR8278 on the mitotic rhythm, this agent unexpectedly promoted corneal wound healing. The mechanisms underlying this effect will require further investigation. Given the close association between Rev-erbα and the inflammatory response,^54,55 we hypothesized that KL001 and SR8278 may also have some influence on the inflammatory process after corneal wounding. Inflammation after corneal wounding has a decisive influence on perfective healing. More work is needed to determine the mechanical mechanism of this effect, including the possibility that these small molecules affect mitosis and wound healing by targeting pathways other than the circadian clock. 

In conclusion, we have demonstrated the critical role of the circadian rhythm system in corneal epithelial mitosis and wound healing and the potential of pharmacologic inhibition or promotion of corneal wound healing by interfering with the molecular clock circuitry. This new knowledge may provide opportunities for circadian-targeted strategies to rapidly restore vision and shorten the time window of possible infection, ranging from the selection of corneal surgery timing to pharmacologic targeting of molecular clock circuitry. Further studies will be aimed at deciphering the cellular mechanisms and signaling circuits within the corneal epithelium to determine how the circadian clock influences mitosis and wound healing, possibly offering opportunities to design strategies for alternative therapies to accelerate healing. Furthermore, it will be interesting to investigate why such regulation has evolved (i.e., why it might be beneficial for mitosis and wound repair in the corneal epithelium to vary over a 24-hour circadian cycle), or whether these changes are secondary consequences of circadian rhythms. 

Acknowledgments

Supported by grants from the National Natural Science Foundation of China (grants no. 30672287, 30772387, and 81470603). 

Disclosure: Y. Xue, None; P. Liu, None; H. Wang, None; C. Xiao, None; C. Lin, None; J. Liu, None; D. Dong, None; T. Fu, None; Y. Yang, None; Z. Wang, None; H. Pan, None; J. Chen, None; Y. Li, None; D. Cai, None; Z. Li, None