Circadian Regulation of IOP Rhythm by Dual Pathways of Glucocorticoids and the Sympathetic Nervous System

Five-week-old male C57BL/6J and Balb/c mice (N = 130; Japan SLC Inc., Shizuoka, Japan) were purchased and housed in plastic cages (170 wide × 240 deep × 125 high mm, Clea, Tokyo, Japan) under a 12-hour light (200 lux)/12-hour dark cycle (12L12D, 0800 light ON, 2000 light OFF) under constant temperature (23 ± 1°C). For real-time bioluminescence recording, Period 2::Luciferase fusion protein (Per2::Luc) knock-in mice23 were provided by Dr. Joseph Takahashi (Southwestern Medical Center, Texas University, TX). Food (CE-2; CLEA) and water were provided ad libitum. All animal experiments were approved by the Committee of Animal Care and Use of Aichi Medical University and Kindai University. All experimental procedures were conducted in accordance with institutional guidelines for the use of experimental animals and the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research.

SCGX and ADX were performed in 5-week-old C57BL/6J mice and Per2::Luc heterozygous mice under general anesthesia, as previously reported.17^,24 Because ADX causes mineralocorticoid deficiency, ADX mice were given 0.9% saline ad libitum instead of water after surgery. IOP measurements were started 2 weeks after surgery to allow for the IOP rhythm to disrupt.

IOP measurements were performed with a tonometer (Icare TonoLab, TV02; Icare Finland Oy, Espmoo, Finland), as previously reported.12 All mice were kept under 12L12D conditions for more than 2 weeks before taking IOP measurements and then transferred to constant dark (DD) conditions. Unanesthetized mice were gently held using a sponge. The IOP rhythm was obtained by measuring IOP at 4-hour intervals under 12L12D conditions and on the first day of DD; at zeitgeber time 2 (1000), 6 (1400), 10 (1800), 14 (2200), 18 (0200), and 22 (0600); and circadian time (CT) 2 (1000), 6 (1400), 10 (1800), 14 (2200), 18 (0200), and 22 (0600), respectively. The time of light ON defines as zeitgeber time 0 (0800). CT was based on the internal free-running period. By convention, CT0 is the beginning of the subjective day and CT12 is the beginning of the subjective night. CT12 (2000) defined the onset of activity in mice. IOPs were measured during the light phase under light (200 lux) conditions and during the dark phase under dim red-light conditions. At each time point, four IOP measurements (five per eye, first measurement was discarded to exclude handling-induced high IOP) were averaged and used for further analysis of rhythmicity and statistics.

When statistically significant differences in IOP rhythm were observed by one-way ANOVA, cosinor curve fitting was performed in Graph Pad Prism 6 (GraphPad Software Inc., San Diego, CA) as previously described.25 Temporal profiles with values of R² that were greater than 0.2 were classified as rhythmic.

A single drop (30 µL) of NE solution (1 mg/mL; # 081-104758, Daiichi-Sankyo, Tokyo, Japan) and CORT (1 mg/mL saline; C0388, TCI, Tokyo, Japan) was instilled onto bilateral eyes during CT0 to CT1 or CT12 to CT13 on the first day of DD in 8-week-old male sham and ADX+SCGX mice. During instillation, mice were gently restrained by their necks held back, under dim red-light conditions.

Paraffin sections (5 µm) of paraformaldehyde-fixed eyes of albino Balb/c mice and frozen sections (14 µm) of paraformaldehyde-fixed eyes of C57BL/6J mice were used for immunohistochemistry.25 We used rabbit polyclonal antibody against Adrb2 (1:2000, bs-0947, Bioss), glucocorticoid receptor (GR) (1:200, sc-1004, Santa Cruz Biotechnology, Dallas, TX), guinea pig polyclonal antibody against Per1 (1:200, PM091, MBL, Nagoya, Japan) and mouse monoclonal antibody against Bmal1 (1:1000, sc-365645, Santa Cruz Biotechnology, Dallas, TX), at 4°C for 48 hours. This was followed by incubation with biotinylated goat antirabbit IgG (120 minutes; 1:1000; PK-6101, Vector Laboratories, Burlingame, CA), and then avidin–biotin complex solution (120 minutes; 1:400; Vectastain Elite ABC Kit, Vector Laboratories) before substrate visualization using 3,3′-diaminobenzidine (349-00903, Dojindo, Kumamoto, Japan). For immunofluorescence, sections were incubated with donkey polyclonal AlexaFluor 488-conjugated antirabbit IgG (1:400; R37118↗, Thermo Fisher Scientific, Waltham, MA) and goat polyclonal DyLight 549-conjugated antiguinea pig IgG (1:400; 606-142-129, Rockland Immunochemicals Inc, Pottstown, PA) secondary antibody for 120 minutes (1:400; R37118↗, Thermo Fisher Scientific).

We performed real-time bioluminescence recording of circadian rhythms as previously reported.26 Briefly, brains and ciliary body/iris complex of adult heterozygous Per2::Luc fusion reporter mice were quickly removed under general anesthesia and immersed in ice-cold Hank's buffered saline (09735-75, Nacalai tesque, Kyoto, Japan). The hypothalamus, containing the central part of SCN was frontally sectioned at 200 µm. Then, the bilateral SCN and iris/ciliary body were placed on a culture membrane (Millicell-CM PICM0RG50, Merck Millipore, Darmstadt, Germany) in a 35-mm petri dish and cultured at 37°C with 1.2 mL of medium containing Dulbecco's modified Eagle medium (Gibco, Carlsbad, CA), B-27 supplement (Gibco), 100 U/mL penicillin, and 100 µg/mL streptomycin, and were then transferred to the culture medium containing D-luciferin potassium salt (200 µM, Wako, Tokyo, Japan) and sealed using parafilm to prevent evaporation. Luminescence emission of the whole explant was examined using a photomultiplier tube (AB-2550, Kronos Dio, ATTO, Tokyo, Japan) for 1 minute at 10-minute intervals. Data analyses were performed using the Kronos Data Analysis Software (ATTO), Clock Lab (ver. 6) (Actimetrics, Wilmette, IL), and Microsoft Excel. Iris/ciliary body cultures underwent medium changes at 7- to 8-day intervals.

For analysis of CORT-induced Per2 expression, iris/ciliary body cultures were treated with the GR agonist dexamethasone (DEX) (100 nM; TCI, D1961) at medium change. After 30 minutes, these agents were washed out.

To generate Chx10-Bmal1 knockout mice, Chx10-Cre mice were mated to mice with a conditional Bmal1 allele (Bmal1^flox/flox) (N > 9 backcrossed to C57BL/6J, Jackson Laboratory, Bar Harbor, ME, #007668).27Bmal1^flox/flox mice were provided by Michihiro Mieda (Kanazawa University, Ishikawa, Japan). We purchased the sperm of Chx10-Cre (B6.Cg-Tg [Chx10-cre] G5 Tfur/TfurRbrc mice) mice from RIKEN BioResource Research Center (Tsukuba, Japan) that were generated in Takahisa Furukawa (Osaka University).28 Genotyping was performed using multiplex PCR and primers; for Bmal1-flox, 5′-ACTGGAAGTAACTTTATCAAACTG-3′ (forward L1), 5′-CTGACCAACTTGCTAACAATTA-3′ (forward L2), and 5′-CTCCTAACTTGGTTTTTGTCTGT-3′ (reverse R4); for Chx10-Cre, 5′-GTCTCCTAGCCTTTGCGTTCAGAC-3′ (forward), and 5′-TTCGGCTATACGTAACAGGG-3′ (reverse).

All data are shown as mean ± SEM. Statistical comparisons were made using GraphPad Prism 6 (GraphPad Software Inc., San Diego, CA) or Excel-Toukei 2012 software (Social Survey Research Information Co. Ltd., Osaka, Japan). Student's t-tests were used to compare two groups and one-way ANOVA with Tukey's multiple comparison test for more than two groups. Time and group interactions or genotype and group interactions were compared by two-way ANOVA with Tukey's multiple comparison test. Differences with a P values of less than 0.05 were considered statistically significant.

To address the effects of glucocorticoids and sympathetic pathways on IOP rhythm, we performed time course IOP measurements under 12L12D in the mice after bilateral ADX and SCGX. Initially, a normal circadian locomotor activity rhythm under 12L12D and DD conditions was performed in all mice (Supplementary Fig. S1A) and completely bilateral ADX mice (Supplementary Fig. S1B). Sham mice showed a clear IOP rhythm: lower in the day and higher at night (P < 0.001) (Fig. 1A). ADX affected IOP at some time points (P < 0.05 at zeitgeber time 18 [0200]), but IOP rhythm was retained (Fig. 1A, Supplementary Fig. S1C). SCGX also had a minor effect on the IOP rhythm (Fig. 1A, Supplementary Fig. S1C). In contrast, the dispersion of IOP rhythm peaks in ADX or SCGX mice varied between the 12L12D and the DD conditions (F < 0.05) (Fig. 1B, Supplementary Fig. S1D).

We next performed both ADX and SCGX in mice to confirm this relationship. This suppressed nocturnal IOP increase under the 12L12D condition, but yielded few rhythm changes (P = 0.05) (Fig. 1A, Supplementary Fig. S1C). We found similar results under the DD condition (P > 0.05, ADX+SCGX) (Fig. 1A, Supplementary Fig. S1C). We could not detect differences in IOP rhythm between 12L12D and DD except in the sham group (P > 0.05 for ADX, SCGX, and ADX+SCGX, P < 0.05 for the sham group) (Supplementary Fig. S1E). ADX and SCGX also dispersed the IOP rhythm peaks in both conditions (Fig. 1B, Supplementary Fig. S1D). A decreased amplitude and correlation coefficient (R²) of rhythmicity was observed under both 12L12D (Figs. 1C, 1D) and DD (Figs. 1E, 1F) conditions. In each mouse, circadian changes in IOP were arrested (Supplementary Fig. S2A–S2D). These results suggest that the transmission pathway of IOP rhythm is composed of two components and involves adrenal glucocorticoids and the sympathetic nervous system.

To investigate the modulation of IOP rhythm by CORT and NE, we administered eyedrops containing CORT and NE, respectively, into bilateral eyes of ADX+SCGX mice, at 0 to 1, that is, during the physiologic antiphase, on the second day of DD. Administration of CORT and NE induced a weak but significant antiphase peak in IOP rhythm, as compared with the arrhythmicity of the control group (P < 0.0001 for sham [Veh], and ADX+SCGX [CORT, NE]; P > 0.05 for ADX+SCGX [Veh]) (Fig. 2A, Supplementary Fig. S3A). Moreover, the dispersed phase caused by ADX+SCGX was phase advanced owing to morning NE and CORT instillation, when compared with sham mice (Fig. 2B). NE and CORT almost restored the decreased amplitude and R² (but not completely) to normal levels (Figs. 2C, 2D). These data suggest that CORT and NE may be involved in circadian regulation of the IOP rhythm. In contrast, the administration of CORT and NE with the correct physiologic phase, CT12 to CT13, also increased IOP (Supplementary Fig. S3A). The peak after instillation and IOP amplification effects were not different from the CT0 to CT1 group (Supplementary Fig. S3B–S3D), indicating a null time-dependent response.

To confirm whether the IOP rhythm is an autonomous rhythm or a driven output, we measured the IOP rhythm until the second day after instillation. The CORT- and NE-induced IOP changes were not sustained until the second day (E), suggesting that IOP rhythms may be a driven output, independent of circadian clock regulation, and not self-sustainable after a single stimulation. We next administered eyedrops containing CORT and NE, respectively, into the bilateral eyes of sham mice at CT0 to CT1. A slight daytime increase of IOP and decrease of amplitudes was detected; however, the nocturnal IOP increase was maintained and the phases did not shift (G–J). These findings indicate that the IOP rhythm is perhaps purely a systemic-driven response. Supplementary Fig. S3 Supplementary Figs. S3 S3

As the expression of β2-adrenergic receptors (Adrb2) in the nonpigmented epithelium (NPE) of the human ciliary body is higher than that of β1-adrenergic receptors,29 and because GR are also expressed in human30 and chicken ciliary bodies,31 we sought to confirm the localizations of expression for GR and Adrb2. Immunohistochemistry revealed strong expression of GR and Adrb2 in the pars plana of the ciliary NPE and weak expression in the retinal ganglion cell layer, inner nuclear layer, and pars plicata of the ciliary NPE, of albino Balb/c (Figs. 3A, 3B) and C57BL/6J mice (Supplementary Fig. S4A). The circadian clock protein Bmal1 and Per1 were identified in the NPE, but not in the trabecular meshwork in Balb/c mice (Figs. 3C, 3D) and C57BL/6J mice (Supplementary Fig. S4B). These findings indicate that CORT and NE act on the ciliary body and regulate IOP rhythm with a nocturnal increase by mediating the production of aqueous humor.

To understand the effect of ADX and SCGX on circadian clock rhythm in the ciliary body, we next performed real-time monitoring of Per2::Luc in a culture of the iris/ciliary body and SCN slice from ADX and SCGX Per2::Luc knock-in mice. As expected, the bioluminescence rhythm was markedly damped and phase-advanced by approximately 4 hours in the iris/ciliary body culture (Figs. 4A, 4B), whereas no differences were observed in the SCN slice cultures (Supplementary Fig. S5). Although ADX and SCGX did not affect the periodicity of rhythm in the iris/ciliary body culture (Fig. 4C), its significantly decreased amplitude (Fig. 4D), indicating circadian rhythm disruption in the iris/ciliary body by ADX and SCGX. Furthermore, Per2::Luc rhythm in the cultured iris/ciliary body could be induced and entrained by DEX administration (Figs. 4A, 4B). The rhythms, even those in ADX and SCGX mice, were self-sustaining after a single DEX stimulation, in contrast with the IOP rhythm after CORT instillation (Fig. 2A), indicating independent regulation of the IOP rhythm by the local clock. Furthermore, when we examined the individual effects of ADX and SCGX on the ciliary clock, ADX markedly attenuated Per2::Luc rhythm and resulted in a phase delay of approximately 8 hours, whereas SCGX slightly damped the rhythm and resulted in a marked phase delay of approximately 7 hours (Figs. 4E–4H). ADX or SCGX alone did not affect the period (Fig. 4G). It is possible that CORT and NE act additively on the ciliary body to generate rhythmicity and achieve an appropriate phase of IOP rhythm; however, mediation by ciliary clock remains unclear.

To identify the role of the ciliary clock in IOP rhythm generation, we generated retina-ciliary epithelium-specific Bmal1 knockout mice (cKO) by crossing Bmal1 flox (wild type) and Chx10-Cre mice (Supplementary Fig. S6A, S6B). In cKO mice, Bmal1 expression decreased in the ciliary body NPE, retinal ganglion cells, and the inner nuclear layer of the retina, but was retained in the SCN (Fig. 5A). cKO mice also maintained normal circadian rhythmicity of locomotor activity under 12L12D and DD conditions (Supplementary Fig. S6C, S6D). These findings indicate a tissue-specific loss of Bmal1 function. However, the IOP rhythm of cKO mice showed almost no difference from the wild-type mice (two-way ANOVA, P > 0.05 vs. wild-type) (Fig. 5B). Phases, amplitudes, and R² were also almost normal (Figs. 5C–5E). These findings suggest that the IOP rhythm is directly driven by CORT or NE, but not via the ciliary clock. However, the temporal patterns of IOP rhythm were slightly different between the 12L12D and DD conditions (Fig. 5B). The ciliary clock may, therefore, be necessary for achieving appropriate IOP rhythm through light stimuli.