Photoreceptor cells display a daily rhythm in the orphan receptor Esrrβ

Abstract

Purpose

Nuclear orphan receptors are critical for the development and long-term survival of photoreceptor cells. In the present study, the expression of the nuclear orphan receptor Esrrβ—a transcriptional regulator of energy metabolism that protects rod photoreceptors from dystrophy—was tested under daily regulation in the retina and photoreceptor cells.

Methods

The daily transcript and protein amount profiles were recorded in preparations of the whole retina and microdissected photoreceptor cells using quantitative PCR (qPCR) and western blot analysis.

Results

Esrrβ displayed a daily rhythm with elevated values at night in the whole retina and enriched photoreceptor cells. Daily regulation of Esrrβ mRNA depended on light input but not on melatonin, and evoked a corresponding rhythm in the Esrrβ protein.

Conclusions

The data presented in this study indicate that daily regulation of Esrrβ in photoreceptor cells may contribute to their adaptation to 24-h changes in metabolic demands.

Introduction

In the mammalian retina, nuclear receptors complement other classes of transcriptional regulators in determining the fate of cells [1,2], their function [3,4], and homeostasis [5]. The majority of nuclear receptors found in the retina are so-called orphan receptors that are nuclear receptors with unknown physiologic ligands. Retinal nuclear orphan receptors typically control photoreceptor development and differentiation, and mutations result in dysfunction and degeneration of rod and/or cone photoreceptors [5]. A prominent example is the retinoic acid-related orphan receptor beta (Rorβ; OMIM 601972; Gene ID mouse 225998, Gene ID rat 309288) gene that directs the cell fate of rods and short (s)-opsin-expressing cones (S-cones) and controls the terminal differentiation of photoreceptor cells. Accordingly, mice mutant for Rorβ lack rods and gain primitive S-cones and photoreceptors lacking outer segments [6-8]. Among the nuclear orphan receptors, estrogen-related receptor β (Esrrβ, Nr3b2, OMIM 602167, Gene ID mouse 26380, Gene ID rat 299210) plays a special role. This orphan receptor does not detectably impair development but the long-term survival of photoreceptor cells instead. Thus, mice deficient for Esrrβ show slow and selective degeneration of rod photoreceptors in later life, which is preceded by the loss of rod outer segments [9]. Whereas mutations in nuclear orphan receptor genes often result in photoreceptor or ocular phenotypes, thus far only a few are known to underlie inherited retinal degeneration in humans [10-14]. Mutations in the human Esrrβ gene are associated with autosomal-recessive deafness [15] although investigations regarding vision disturbances have not yet been reported.

Mammalian photoreceptor cells must maintain their function for a lifetime in the face of hazards such as oxidative stress [16] and metabolic/energy challenges [17,18] occurring during the day/night cycle. To avoid age-related dysfunction or death, this may require the ability to adapt the cellular defense mechanisms and metabolism to 24-h changes in the environment [16]. Daily adaptation of photoreceptor cells (and other retinal neurons) is driven by light input and retinal clocks [19-21] through the release of the neuromodulators melatonin and dopamine, both of which play opposing roles in retinal adaptation [22]. Whereas melatonin is released during the dark/night and promotes dark-adaptive mechanisms [23-25], dopamine is released during the light/day and contributes to light adaptation of the photoreceptor cells [26,27].

At the transcriptional level, 24-h changes in the nuclear orphan receptor Rorβ contribute to daily adaptation of the retina and photoreceptor cells [28-30]. The data included in the present study show that daily changes of the nuclear orphan receptor Esrrβ are evident in photoreceptor cells and may contribute to their ability to comply with metabolic demands and thus to the cells’ long-term survival.

Methods

Animals

Animal experimentation was performed in accordance with the European Communities Council Directive (86/609/EEC). The study was approved by the German national investigation office and adhered to the ARVO Statement for Use of Animals in Research. Adult male and female rats (Sprague-Dawley) or mice (melatonin-proficient C3H/He, not carrying the rd mutation; melatonin-deficient C57BL/6Jb) were kept under standard laboratory conditions (illumination with fluorescent strip lights, 200 lux at cage level during the day and dim red light (<3 lux) during the night; 20±1 °C; water and food ad libitum) under 12 h:12 h light-dark (12:12 LD) for 3 weeks. When indicated, after LD treatment the animals were kept for one cycle under dim red light and killed during the next cycle. They were killed at the indicated time points by decapitation following anesthesia with 100% carbon dioxide for approximately 3 min. All dissections during the dark phase were performed under dim red light. The retinas were rapidly removed and immediately processed as follows.

Sample preparation

The sample size for all experiments was n=4, with each n deriving from 4 pooled retinas of 2 animals. The HEPES-glutamic acid buffer mediated organic solvent protection effect (HOPE; DCS, Hamburg, Germany) technique was applied to fix the retinas. Briefly, fixation started with the incubation of fresh retinas in an aqueous protection-solution HOPE I (DCS) for 48 h at 0–4 °C. Retinas were then dehydrated in a mixture of HOPE II solution (DCS) and acetone for 2 h at 0–4 °C, followed by dehydration in pure acetone for 2 h at 0–4 °C (repeated twice). Tissues were then embedded in low-melting paraffin (T_m=52–54 °C). Tissue sections (10 µm) from HOPE-fixed and paraffin-embedded retinas were prepared on membrane-mounted slides (DNase/RNase free PALM MembraneSlides, P.A.L.M., Bernried, Germany). Three sections were placed on each slide. The sections were deparaffinized with isopropanol (2 × 10 min each, at 60 °C). All sections were stained with cresyl violet (1% w/v cresyl violet acetate in 100% ethanol) for 1 min at room temperature, washed briefly in 70% and 100% ethanol, and then air-dried [31].

Laser microdissection and pressure catapulting

To isolate photoreceptor cells and inner retinal neurons from the stained sections in a contact- and contamination-free manner, the laser microdissection and pressure catapulting (LMPC) technique was applied [32]. The LMPC technique was performed using a PALM MicroBeam system (Zeiss MicroImaging, Munich, Germany) with PALM RoboSoftware (P.A.L.M.). Under the 10X objective, the outer nuclear layer bearing the photoreceptor cells and the tissue between the inner part of the outer plexiform layer and the inner ganglion cell layer bearing the inner retinal neurons were selected, cut, and catapulted into the caps of 0.5 ml microfuge tubes with an adhesive filling (PALM AdhesiveCaps, P.A.L.M.) by using a pulsed ultraviolet-A (UV-A) nitrogen laser. Smaller areas of the sections were pooled to reach total average sample sizes of 4 million square microns per tube. Alternatively, the whole retina was excised with a scalpel and collected in a 0.5 ml microfuge tube. Cell lysis for RNA preparation was performed immediately after the sample was collected. To verify the purity of the preparations, they were subjected to molecular analysis with rhodopsin (Rho) and neural retina leucine zipper (Nrl) as markers for photoreceptor cells and tyrosine hydroxylase (Th) and metabotropic glutamate receptor 6 (mGluR6) as markers for inner retinal neurons.

RNA extraction

RNA of the laser-microdissected tissue samples was isolated using the RNeasy Micro Kit (Qiagen, Hilden, Germany) following the manufacturer’s instructions. Briefly, collected cells were lysed in a guanidine-thiocyanate-containing buffer (RLT buffer) supplied by the manufacturer. The lysates were diluted with RNase-free water and treated with proteinase K. The samples were then cleared with centrifugation, diluted with ethanol, and applied to a RNeasy MinElute Spin Column to bind RNA to the silica-gel membrane. After the first washing step, an on-column DNase treatment with RNase-free DNase I was performed as described by the manufacturer. Isolated RNA was eluted in the final volume of 12 µl RNase-free water. The amount of extracted RNA was determined by measuring the optical density at 260 and 280 nm.

Reverse-transcription and quantitative PCR

cDNA was synthesized using the Verso cDNA Kit (Abgene, Hamburg, Germany), following the manufacturer’s instructions. Briefly, 4 µl RNA solution was reverse transcribed using anchored oligo-dT primers supplied with the kit in a final volume of 20 µl. cDNA was then diluted 1:3 in RNase-free water, and aliquots of 5 μl were used for PCR. Quantitative PCR was performed in a total volume of 25 μl containing 12.5 μl ABsolut QPCR SYBR Green Fluorescein Mix (Abgene), 0.75 μl of each primer (10 µM), 6 μl RNase-free water, and 5 μl sample. Primer sequences are listed in Table 1. PCR amplification and quantification were performed in a CFX96 (BioRad, Munich, Germany) according to the following protocol: denaturation for 3 min at 95 °C, followed by 40 cycles of 30 s at 95 °C, 20 s at 60 °C, and 20 s at 72 °C. All amplifications were performed in duplicate. By using agarose gel electrophoresis, the generated amplicons for all genes under examination were shown to possess the predicted sizes (Table 1). The amount of RNA was calculated from the measured threshold cycles (C_t) using a standard curve. The transcript amount of glyceraldehyde-3-phosphate dehydrogenase (Gapdh) was constitutively expressed over the 24-h period in preparations of the whole retina, photoreceptor cells, and inner retinal neurons. Values were then normalized according to the amount of GAPDH mRNA present.

Table 1. Primer sequences.

Gene	Accession number	Primer sequence 5′ to 3′	Product length of product [bp]PCR
r-Crx	NM_021855	F: GGCTGTCCCATACTCAAGTG	107
		R: GGTACTGGGTCTTGGCAAAC
r-Esrrα	AY280663.1	F:GTGGCCGACAGAAGTACAAG	144
		R: CAACCACCAGCAGATGAGAC
r-Esrrβ	NM_001008516	F: CAAGAGACGGCTGGATTCG	105
		R: TCGGCCACCAGTAGATACG
m-Esrrβ	NM_011934	F: CCCTCGCCAACTCAGATTC	104
		R: GCGCTGACTCAGCTCATAG
r-Esrrγ	AY341057	F: CTTGATCCTGGGTGTTGTG	101
		R: GGCCTGCTAATTTGGACTG
r-Gapdh	NM_017008	F: ATGACTCTACCCACGGCAAG	89
		R: CTGGAAGATGGTGATGGGTT
m-Gapdh	BC082592	F: GTCATCCCAGAGCTGAAC	144
		R: CTCAGATGCCTGCTTCAC
r-mGluR6	NM_022920	F: ACAACCGCAGAAACATCTGG	113
		R: TTCCTCACCTGTGCATTTCC
r-NeuroD	AF107728	F: AGACGAGTGCCTCAGTTC	136
		R: CTCCTCCAGATCCTCATCTTC
r-Nr2c1	NM_145780	F: CGGAGAGGAGGAAGTTGTTG	126
		R: CGGGACTGAAGAGAACGATG
r-Nr2e1	NM_001113197	F: GAGTTGACCGCTGTGTCTG	139
		R: CTGATTCGCACACGGACTC
r-Nrl	NM_001106036	F: AGGCCTGGAGGAGCTATAC	134
		R: CATGAGGCCCTTCTACAGAG
r-Nupr1	NM_053611	F: GCCTGGCCCAATCTTATGTC	116
		R: CTTGGTCAGCAGCTTCCTC
r-Otx2	NM_001100566	F: CAGTCAATGGGCTGAGTC	112
		R: GCCCTCGTGAAAGTAGTC
r-Rho	NM_033441	F: ACTCAGAAGGCAGAGAAGG	638
		R: CCTAGACACTTAGCCGTAAGC
r-Rorβ	NM_001270958	F: GCCTGGCTGTTAGAACCAAG	147
		R: GTTGCAGACTGCCGTGATAG
m-Rorβ	NM_001043354	F: CCTGGCTGATCGAACCAAG	144
		R: TGCAGACTGCCGTGATAG
r-Th	NM_012740	F: CCAGGGCCTTTCCCAAAGTC	186
		R: CATGGAGGGCAGGAGGAATG

Western blot analysis

For western blot analysis, samples were loaded on 4–12% NuPAGE Novex Bis-Tris gels (Invitrogen, Carlsbad, CA), separated, and blotted onto polyvinylidene fluoride (PVDF) membrane (Westran S, Whatman Inc., Sanford, ME). For immunodetection, membranes were blocked in 5% skimmed milk powder, and anti-Esrrβ polyclonal antibody (1:500; Abcam, Cambridge, MA; ab19331) was applied overnight at 4 °C. The horseradish-peroxidase-coupled secondary antibodies (goat anti-rabbit-HRP 1:5000; Sigma-Aldrich, St. Louis, MO; A0545) were visualized using an enhanced chemiluminescence (ECL) detection system (GE Healthcare Amersham, Freiburg, Germany). To ensure that immunoreactivity was derived from equal protein amounts of homogenates, staining with rabbit anti-β-actin polyclonal antibody (1:300; Sigma-Aldrich; A2066) was conducted. Densitometric measurement was performed using the ImageJ 1.46o software (National Institutes of Health, Bethesda, MD).

Statistical analysis

All PCR data are expressed as the mean ± standard error of the mean (SEM) of four independent experiments including eight time points. Transcript levels from each sample (consisting of four pooled retinas from two animals) were calculated relative to the average expression of each data set throughout 24 h to plot the temporal expression. Quantitative data were analyzed with ANOVA (one-way ANOVA) to evaluate variations among the groups. Cosinor analysis was used to fit sine-wave curves to the circadian data to mathematically estimate the time of peaking gene expression (acrophase) and to assess the amplitude [33,34]. The model can be expressed according to the following equation: f(t)=A + B cos [2π (t + C) ⁄ T]. The f(t) indicates the relative expression levels of the target genes, t specifies the sampling time (h), A represents the mean value of the cosine curve (mesor; midline estimating statistic of rhythm), B indicates the amplitude of the curve (half of the sinusoid), and C indicates the acrophase (point of time, when the function f(t) is maximum). T gives the time of the period, which was fixed at 24 h for this experimental setting. The significance of the daily regulation was defined by showing p<0.05 in the ANOVA and Cosinor analysis.

Results

Esrrβ mRNA and protein levels are under daily regulation in the retina

Nuclear orphan receptors important for the survival of photoreceptor cells or protective against photoreceptor dystrophy [5,35] were tested under daily regulation. Of the nuclear receptors considered under daily regulation (Figure 1, Table 2), only Esrrβ displayed a daily change in both statistical analyses applied (Figure 2, Table 3). The 24-h pattern of Esrrβ mRNA levels showed peak expression at ZT18.7 and an amplitude of 32.9%. Consistent with the validity of the results, the daily profile of the reference gene Rorβ was similar to those found in previous reports [28,30] (Figure 2, Table 3).

Figure 1.

Twenty-four-hour profiling of various nuclear orphan receptors in preparations of the whole rat retina under 12 h:12 h light-dark cycles. The mRNA levels are plotted as a function of Zeitgeber time (ZT). The blue lines represent the periodic sinusoidal functions (illustrated only for p<0.05 in cosinor analysis). The solid bars indicate the dark period. Data represent a percentage of the average transcript amount during the 24-h period. The value of ZT0 was plotted twice at ZT0 and ZT24. Each value represents mean ± standard error of the mean (SEM; n=4). Statistical analysis of transcriptional profiling is provided in Table 2.

Table 2. Statistical analysis of transcriptional profiling illustrated in Figure 1.

Gene	Whole retina under LD 12:12
	One-way Anova	Cosinor analysis
	P value	P value	Acrophase [h]	Amplitude [%]
Crx	=0.91	>0.05	-	-
Esrrα	=0.90	>0.05	-	-
Esrrγ	=0.86	>0.05	-	-
NeuroD	=0.90	>0.05	-	-
Nr2c1	=0.85	<0.05	16.6	21.8
Nr2e1	=0.95	>0.05	-	-
Nupr1	=0.72	<0.05	6.8	49.8
Otx2	=0.14	<0.05	14.7	28.3

Statistical analysis using one-way Anova and Cosinor analysis are shown for each gene under LD 12:12. (LD 12:12: light/dark 12:12)

Figure 2.

Twenty-four-hour profiling of Esrrβ and the reference gene Rorβ in preparations of whole rat retina under 12 h:12 h light-dark cycles (LD; upper line) or constant darkness (DD; lower line) using qPCR. The mRNA levels are plotted as a function of Zeitgeber time (ZT) and circadian time (CT). The blue lines represent the periodic sinusoidal functions determined with cosinor analysis (only for p<0.05). The solid bars indicate the dark period. Data represent a percentage of the average value of the transcript amount during the 24-h period. The value of ZT0 was plotted twice at ZT0 and ZT24. Each value represents mean ± standard error of the mean (SEM; n=4). Statistical analysis of transcriptional profiling is provided in Table 3. Note that daily rhythmicity of Esrrβ does not persist under DD.

Table 3. Statistical analysis of transcriptional profiling illustrated in Figure 2.

Gene	Whole retina under LD 12:12				Whole retina under DD
	One-way Anova	Cosinor analysis			One-way Anova	Cosinor analysis
	P value	P value	Acrophase [h]	Amplitude [%]	P value	P value	Acrophase [h]	Amplitude [%]
Esrrβ	=0.016	<0.05	18.7	32.9	=0.28	>0.05	-	-
Rorβ	=0.006	<0.05	18.1	53.9	=0.021	<0.05	22.6	15.9

Statistical analysis using one-way Anova and Cosinor analysis are shown for each gene under LD 12:12 and DD. (LD 12:12: light/dark 12:12; DD: constant darkness)

In the context of a putative role of Esrrβ in the protection of the retina against daily changes in the environment, the question of whether daily regulation of the Esrrβ gene results in daily changes in the protein product was addressed using western blotting (Figure 3). The anti-Esrrβ antibody recognizes a band of about 52 kDa, a molecular mass predicted from the Esrrβ gene. The intensity of Esrrβ immunoreactivity tended to increase during the light phase (p=0.041 in one-way ANOVA). This suggests that the daily rhythm in Esrrβ transcript levels evokes corresponding variations in the protein amount, with the temporal lag reflecting the time necessary to translate mRNA into protein.

Figure 3.

Western blot analysis of Esrrβ conducted at different ZTs (Zeitgeber times) during the 24-h cycle in the whole rat retina. The upper lanes show a representative western blot with Esrrβ immunostaining at about 52 kDa. The lower lanes show the β-actin signal to which the Esrrβ immunostaining was normalized. The diagram shows the quantification of immunoreactivity in relation to the corresponding β-actin signal. The solid bar indicates the dark period in the 12 h:12 h light-dark cycles. Data were obtained with densitometric measurement and represent percentages of the overall maximal value. Each value is the mean ± standard error of the mean (SEM; n=4). Note that Esrrβ immunoreactivity is elevated during the light phase (p=0.041 in one-way ANOVA).

Daily regulation of Esrrβ depends on illumination

Daily regulation of Esrrβ may be promoted by light input and/or a circadian clock. To investigate and compare the influence of both parameters between Esrrβ and Rorβ, rats adapted to the 12 h:12 h light-dark cycle were kept in constant darkness (DD) for one cycle and recorded during the subsequent cycle (Figure 2, Table 3). Under these conditions, the daily rhythm in the transcript levels was not detected for Esrrβ. As expected for a component of a circadian clock [23] and as reported previously [30], the amount of Rorβ mRNA persisted to the cycle under DD even though the cycling amplitude was clearly decreased (Figure 2, Table 3).

Daily regulation of Esrrβ evoked by photoreceptor cells

To compare daily regulation of Esrrβ and Rorβ between photoreceptor cells and inner retinal neurons, the LMPC technique was applied [32]. The purity grades of the preparations obtained were verified by using specific gene markers of photoreceptor cells, namely, Rho and Nrl (as markers for rods [36,37]) and of inner retinal neurons, namely, Th (as a marker for amacrine cells [38]) and mGluR6 (as a marker of ON-bipolar cells [39]). Compared to whole retina preparations, in the photoreceptor cells collected with LMPC, the ratio of Rho to Th and mGluR6 increased 34-fold and 85-fold, respectively, and that of Nrl to Th and mGluR6 increased 170-fold and 43-fold, respectively. For the inner retina sample, the ratio of Th to Rho and Nrl increased 23-fold and 29-fold, respectively, and that of mGluR6 to Rho and Nrl increased 15 fold and 18 fold, respectively.

To match the expression of the nuclear orphan receptors in the photoreceptor cells and the inner retinal neurons, mRNA levels were compared at the ZT of respective peak expression (Table 4). Both genes displayed higher transcripts in photoreceptor cells than in inner retinal neurons (Esrrβ: 23-fold; Rorβ: 15-fold). A comparison of the genes within the same tissue revealed that Esrrβ displayed higher transcript levels in photoreceptor cells (30-fold) and inner retinal neurons (15-fold).

Table 4. Comparison of peak mRNA levels between photoreceptor cells (PRC) and the inner retina neurons.

Gene	PRC	Inner retina	PRC/inner retina ratio
Esrrβ	143,000	6.290	22.73
Rorβ	4530	296	15.30

Transcript amount was determined by qPCR and represents the number of transcripts in relation to Gapdh x 10⁵ (mean ± SEM with n=4). (PRC: photoreceptor cells)

To define and compare daily regulation of Esrrβ and Rorβ in photoreceptor cells and inner retinal neurons, the transcript levels were profiled in microdissected preparations of photoreceptor cells and inner retinal neurons. Both nuclear orphan receptors displayed daily rhythms in photoreceptor cells (Figure 4, Table 5) with similar 24-h profiles observed in preparations of the whole retina (Figure 2, Table 3). In contrast, the nuclear orphan receptors showed non-rhythmic expression in the inner retinal neurons (Figure 4, Table 5). These data suggest that the daily regulation of both genes in the retina is evoked primarily by photoreceptor cells.

Figure 4.

Comparative 24-h profiling of Esrrβ and the reference gene Rorβ in preparations of photoreceptor cells (upper line) and inner retinal neurons (lower line) of rats. The mRNA levels are determined by using quantitative PCR (qPCR), and both are plotted as a function of Zeitgeber time (ZT). The blue lines represent the periodic sinusoidal functions determined with cosinor analysis (only for p<0.05). The solid bars indicate the dark period. Data represent a percentage of the average transcript amount during the 24-h period. The value of ZT0 was plotted twice at ZT0 and ZT24. Each value represents mean ± standard error of the mean (SEM; n=4). Statistical analysis of transcriptional profiling is provided in Table 5. Note that daily variations of Esrrβ are evoked by photoreceptor cells.

Table 5. Statistical analysis of transcriptional profiling illustrated in Figure 4.

Gene	PRCs under LD 12:12				IRN under LD 12:12
	One-way Anova	Cosinor analysis			Oone-way Anova	Cosinor analysis
	P value	P value	Acrophase [h]	Amplitude [%]	P value	P value	Acrophase [h]	Amplitude [%]
Esrrβ	=0.017	<0.05	17.3	39.0	=0.56	>0.05	-	-
Rorβ	=0.002	<0.05	17.7	59.3	=0.57	>0.05	-	-

Statistical analysis using one-way Anova and Cosinor analysis are shown for each gene in photoreceptor cells and inner retinal neurons. (LD 12:12: light/dark 12:12; PRC: photoreceptor cells; IRN: inner retinal neurons)

Daily regulation of Esrrβ does not depend on melatonin

To investigate the contribution of melatonin to daily changes in the expression of Esrrβ, the 24-h course of the transcript levels of Esrrβ were compared between melatonin-proficient (C3H/He, not carrying the rd mutation) and melatonin-deficient mice (C57BL/6Jb; Figure 5, Table 6). The daily profiles of Esrrβ were similar in both mice strains and resembled that of rats (Figure 2, Table 3). As with Esrrβ, the reference gene Rorβ cycled in a similar manner in melatonin-proficient and -deficient mice. These observations suggest that daily regulation of Esrrβ and Rorβ does not require a pulsatile melatonin signal.

Figure 5.

Comparative 24-h profiling of Esrrβ and Rorβ in the retina of melatonin-proficient (C3H/He, rd^+/+; upper line) and melatonin-deficient mice (C57BL/6J; lower line). The mRNA levels are determined by using quantitative PCR (qPCR), and both are plotted as a function of Zeitgeber time (ZT). The blue lines represent the periodic sinusoidal functions determined with cosinor analysis (only for p<0.05). The solid bars indicate the dark period. Data represent a percentage of the average value of the transcript amount during the 24-h period. The value of ZT0 was plotted twice at ZT0 and ZT24. Each value represents mean ± standard error of the mean (SEM; n=4). Statistical analysis of transcriptional profiling is provided in Table 6. Note that daily rhythmicity of Esrrβ does not depend on a pulsatile melatonin signal.

Table 6. Statistical analysis of transcriptional profiling illustrated in Figure 5.

Gene	Melatonin proficient (C3H/He, rd+/+)				Melatonin deficient (C57BL/6J)
	One-way Anova	Cosinor analysis			One-way Anova	Cosinor analysis
	P value	P value	Acrophase [h]	Amplitude [%]	P value	P value	Acrophase [h]	Amplitude [%]
Esrrβ	=0.016	<0.05	15.5	53.5	=0.001	<0.05	16.3	31.7
Rorβ	=0.041	<0.05	13.9	34.0	=0.006	<0.05	17.1	29.6

Statistical analysis using one-way Anova and Cosinor analysis are shown for each gene from samples of melatonin proficient and melatonin deficient mice.

Discussion

In this study, Esrrβ expression was observed under daily regulation in the retina. This indicates that cyclicity of Esrrβ is evident not only in peripheral tissues (such as white adipose tissue, brown adipose tissue, the liver, and muscle [40]) but also in the retina and thus in an area of the brain. The temporal timing of Esrrβ expression in the retina is similar to that found in both types of adipose tissues and is phase-advanced to that observed in the liver and muscle [40]. The retina resembles the peripheral tissues with Esrrβ cyclicity in its high metabolic activity and energy demand [41]. Therefore, daily regulation of Esrrβ may be a characteristic feature of tissues with high metabolic demand or may even be unique to them.

In the retina, the daily cyclicity of Esrrβ is evoked by photoreceptor cells. This is evident from the finding that the microdissected photoreceptor cells showed a daily rhythm in Esrrβ, which is similar to that observed in preparations of the whole retina, whereas microdissected inner retinal neurons did not appear to display Esrrβ rhythmicity at all. Although the high-purity grade of the photoreceptor preparations used was proven by the enrichment of transcripts known to be specifically abundant in photoreceptor cells (Rho, Nrl), a limitation of this study is that the photoreceptor preparations were derived from rods and cones. Thus, the gene monitoring performed in the microdissected photoreceptors reflects the average rod and cone values and may not necessarily be valid for each photoreceptor type. However, Esrrβ was observed to be highly enriched in rod photoreceptors [42,43], and the Esrrβ mutation impairs only rod photoreceptors [9]. Therefore, daily regulation of Esrrβ probably occurs in rod photoreceptors.

Daily changes in Esrrβ require light-dark transitions and therefore depend on light input (this study) whereas that of Rorβ is truly circadian [30] (this study). This suggests that retinal nuclear orphan receptors on the whole mediate light-dependent and circadian adaptation. The rhythmicity of Esrrβ and Rorβ continued in melatonin-deficient mice and therefore does not require a pulsatile melatonin signal. As a result, neither the illumination-dependent regulation of Esrrβ nor the circadian regulation of Rorβ is mediated by melatonin.

Esrrβ influences the energy balance in the whole body [44] and is known to regulate key enzymes of fatty acid and carbohydrate metabolism in peripheral tissues [45] and photoreceptor cells [9]. Therefore, the cyclicity of Esrrβ may contribute to daily changes in the energy metabolism of the retina and photoreceptor cells and thus to their ability to adapt to 24-h changes in metabolic demands [17,18]. The fulfillment of metabolic demands appears to be a strong criterion in photoreceptor survival [46-48]. Thus, a role of Esrrβ in metabolic adaptation could (on a long-term basis) contribute to the survival of photoreceptor cells. This suggestion agrees with the observations that the loss of functional Esrrβ on a long-term basis leads to dysfunction and degeneration of rods and that enhanced Esrrβ activity rescues photoreceptor defects [9].

Similar to Esrrβ, the transcriptional coactivator Pgc-1α (i) protects photoreceptor cells from dystrophy [49], (ii) is under daily regulation in photoreceptor cells [50], and (iii) is important for metabolic homeostasis [51]. As a result, Esrrβ and Pgc-1α may act synergistically to increase photoreceptor survival through daily adjustment of its metabolism. Members of the Esrr family are coactivated by Pgc-1α to regulate energy metabolism in peripheral tissues [52]. If this occurs in photoreceptor cells, then Esrrβ and Pgc-1α may act jointly at the promoter level to maintain metabolic homeostasis of photoreceptor cells.

However, Esrrβ is known to target rod-specific genes including Rho [9]. Since Rho transcript levels are under daily regulation [28,53], this is consistent with the concept in which Esrrβ promotes the daily adaptation of rods by regulating rod-specific gene expression including that of Rho.

In conclusion, the data of the present study suggest a daily rhythm in Esrrβ expression that could promote metabolic adaptation of rods and in this way their long-term survival. Future investigations are required to investigate to what extent impairment of daily regulation of Esrrβ expression and energy metabolism may contribute to the various forms of photoreceptor dystrophy—in particular to those that are age-related.